I 0 5 83 A World Bank S y m p o s i u m AGRICULTURE AND THE ENVIRONMENT Perspectives on Sustainable Rural Development 4; Edited by Ernst Lutz with the assistance of Hans Binswanger. Peter Hazell. and Alexander McCalla Agriculture and the Environment A World Bank Symposium Agriculture and the Environment Perspectives on Sustainable Rural Development Edited by Ernst Lutz with the assistance of Hans P. Binswanger, Peter Hazell, and Alexander McCalla The World Bank Washington, D.C. © 1998 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W., Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of America First printing October 1998 The findings, interpretations, and conclusions expressed in this study are entirely those of the authors and should not be attributed in any manner to the World Bank, to its affiliated organi- zations, or to members of its Board of Executive Directors or the countries they represent. The cover was designed by Tomoko Hirata and Alan Kahan, using photographs by Curt Carnemark (center), Ernst Lutz (top right and bottom left, in Togo), and Stefano Pagiola (top left and bottom right, in Bangladesh). The material in this publication is copyrighted. Requests for permission to reproduce portions of it should be sent to the Office of the Publisher at the address shown in the copyright notice above. The World Bank encourages dissemination of its work and will normally give permis- sion promptly and, when the reproduction is for noncommercial purposes, without asking a fee. Permission to copy portions for classroom use is granted through the Copyright Clearance Center, Inc., Suite 910, 222 Rosewood Drive, Danvers, Massachusetts 01923 U.S.A. For a copy of the "Update" describing new publications, contact the Office of the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433 U.S.A. A catalog and order- ing informnation are also available on the Internet at http://www.worldbank.org. Library of Congress Cataloging-in-Publication Data Agriculture and the environment: perspectives on sustainable rural development / edited by Ernst Lutz; with the assistance of Hans P. Binswanger, Peter Hazell, and Alexander McCalla. p. cm. - (A World Bank symposium) Includes bibliographical references. ISBN 0-8213-4249-5 1. Agriculture-Environmental aspects-Developong countries. 2. Ru:ral development-Developing countries. 3. Sustainable development-Developing countries. 4. Agriculture and state- Developing countries. 5. Rural development-International cooperation. 6. Sustainable development-International cooperation. I. Lutz, Ernst, 1949- . II. World Bank. HI. Series. S589.76.D44A46 1998 338. ]-dc2l 98-36512 CIP Contents Foreword ix Ismail Serageldin and Ian Johnson Contributors xi Acknowledgments xiv 1 Introduction 1 Ernst Lutz, Hans P Binswanger, Peter Hazell, and Alexander McCalla Part I. Policy Perspectives 3 2 Integrating Environmental and Sustainability Concerns into Rural Development Policies 9 Peter Hazell and Ernst Lutz 3 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia 22 John Heath and Hans P Binswanger 4 Toward More Conducive Policies for Sustainable Agriculture 35 Jules N. Pretty 5 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management in Developing Countries 50 Stein T Holden and Hans P Binswanger 6 Agricultural Trade Reforms, Research Initiatives, and the Environment 71 Kym Anderson 7 Property Regimes in Economic Development: Lessons and Policy Implications 83 Daniel W Bromley v vi Contents 8 Equity and Environmental "Modifiers" for Rural Development Policies 92 Peter Hazell Part II. Institutional and Social Perspectives 95 9 Participatory Rural Development 103 Deepa Narayan 10 Involving Local Organizations in Watershed Management 118 Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg 11 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers 130 John Farrington and Graham Thiele 12 People-Centered Agricultural Development: Principles of Extension for Achieving Long-Term Impact 145 Roland Bunch 13 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South A:frica 156 Klaus Deininger 14 Decentralization and Biodiversity Conservation 175 Julian Caldec ott and Ernst Lutz 15 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development 186 Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick Part III. Technical Issues and Perspectives 203 16 The Costs and Benefits of Soil Conservation in Central America and the Caribbean 215 Ernst Lutz, Stefano Pagiola, and Carlos Reiche 17 Toward a Strategy for Mainstreaming Biodiversity in Agricultural Development 230 Jitendra Srivastava, Nigel Smith, and Douglas A. Forno 18 Integrated Pest Management: Strategies and Policies for Effective Implementation 242 Tjaart W. Schillhorn van Veen, Douglas A. Forno, Steen Joffe, Dina L Umali-Deininger and Sanjiva Cooke 19 Toward Reduced Pesticide Use for Cereal Crops in Asia 254 Prabhu L. Pingali and Roberta V Gerpacio 20 The Economic and Environmental Impacts of Irrigation and Drainage in Developing Countries 271 Ujjayant Chakravorty 21 Livestock and the Envirornment: Issues and Options 283 Henning Steinfeld, Cornelis de Haan, and Harvey Blackburn Contents vii 22 Public Policies to Reduce Inappropriate Tropical Deforestation 303 David Kaimowitz, Neil Byron, and William Sunderlin 23 Costs, Benefits, and Farmer Adoption of Agroforestry 323 Dean Current, Ernst Lutz, and Sara Scherr Part IV. Conclusions 345 Index 361 Foreword Agriculture in developing countries has been remark- developed a new one. The vision and specific steps to ably successful during the past few decades in enabling be taken are outlined in Rural Development: From food production to keep pace with the growing popula- Vision to Action. A Sector Strategy (1997). With a tion. This success has been achieved in part at the cost renewed commitmnent to sound development in the of placing more stress on natural resources and the rural sector, the Bank seeks to play its part in a broad- environment. As we look to the future, the need to con- ly based effort encompassing farmers, their families, tinue increasing food production, while at the same communities, and policymakers at the national levels, time minimizing environmental damages, conserving as well as the donor community. the resource base, and reducing poverty, hunger, and This volume brings together state-of-the-art malnutrition, poses an enormous challenge. applied, practical research related to agriculture and During the 1970s, in part through the leadership environment in the developing world. It distills cur- of Robert McNamara, the World Bank placed empha- rent knowledge and summarizes it for development sis on reaching the rural poor through integrated rural practitioners. Where possible, authors use specific development projects. But as such projects produced examples to indicate which approaches have worked mixed results and prices of agricultural commodities and which have not, under which conditions, and why. fell in real terms, international interest in agricultural The general observations and findings in this vol- and rural matters waned during the 1980s. This trend ume indicate that, in the future, environmentally was reflected in a decline in the World Bank's rural sound increases in productivity may be harder to activities. Nevertheless, roughly 70 percent of the achieve than ini the past. But a significant potential for poorest people in developing countries still live in achieving sustainable increases still exists. To realize rural areas, and focusing on the rural sector must that potential, it is becoming even more important to again become a top priority of development assis- undertake necessary reforms-even politically sensi- tance. Therefore, rather than abandon rural develop- tive ones like land reforms-so that policy and ment work because earlier work had limited success, institutional frameworks fully support sustainable we must learn from past mistakes and proceed with agricultural intensification by creating a conducive renewed vigor. This means reinventing rural devel- environment in which appropriate technological inno- opment to fully integrate the environmental and vation can flourish. social dimensions that were given insufficient atten- A conducive policy and institutional framework tion earlier. includes the political will for reform and a comniit- Over the past two years, the World Bank has inten- ment by those in power to seek the common good sively revisited its rural development strategy and rather than let policy be driven by special interests ix x Foreword and influenced by corruption. Such reform needs to be Where externalities are involved, appropriate gov- supported by broadly based policy analysis that not emient interventions will be needed to achieve more only considers agriculture narrowly defined but also optimal results for the whole society. integrates natural resources and the environment as Although we emphasize work on the policy and well as the social dimensions of rural development institutional side rather than "technological fixes," work. we also stress the continued need for top-quality, It is clear that, in contrast to some of the earlier long-term research such as that undertaken over the centralized efforts, the new sustainable rural past several decades by the various international cen- development approach must be much more decen- ters supported by the Consultative Group for tralized and participatory. It will entail more on-the- International Agricultural Research (CGIAR). Such ground testing with pilot projects before scaling up to basic research, now with a more serious effort to inte- larger efforts. Farmers, local communities, non- grate environmental, social, and small-farmer con- governmental organizations, and other parts of civil cerns, remains essential for the future. society will need to be involved in rural development The challenges for the rural sector loom large, and activities to a greater extent than before, the constraints are serious, but with good governance Farmers, although operating with incomplete and a real commitment to environmentally and knowledge and discounting the future in accordance socially sustainable rural development at all levels, with their level of income and wealth, will be viewed much can be achieved. We at the World Bank and as rational managers who seek to achieve many those of us involved with CGIAR will do everything objectives (such as increased food security and we can to alleviate rural poverty and hunger and to income and decreased risks) given the many con- assist with participatory rural development, while straints they face. They therefore need to be given safeguarding natural resources and the environment. choices and encouraged to innovate and experiment rather than be pushed to adopt preselected technolo- Ismail Serageldin gies by extension services that often lack cutting- Vice Presidentfor Special Programs edge knowledge or messages. Many of the sustain- and Chairman, CGIAR able increases in productivity are expected to come The World Bank from innovative farmers working within a policy and institutional framework that provides the right incen- Ian Johnson tives for environmentally sound rural development, Vice President such as more efficient use of inputs. In many cases, Environmentally and Socially this will be good for productivity, input cost, and the Sustainable Development government budget, as well as the environment. The World Bank Contributors KYM ANDERSON NEIL BYRON University of Adelaide Center for International Forestry Research (CIFOR) Adelaide, Australia Bogor, Indonesia JACQUELINE A. ASHBY JULIAN CALDECOTT Center for Tropical Agriculture (CIAT) Consultant Cali, Colombia Salisbury, United Kingdom HANS P. BINSWANGER UJJAYANT CHAKRAVORTY World Bank University of Hawaii and East-West Center Washington, D.C. Honolulu HARVEY BLACKBURN Sanjiva Cookeb Agricultural Research Service World Bank United States Department of Agriculture Washington, D.C. Washington, D.C. DEAN CURRENT DANIEL W. BROMLEY Tropical Agricultural Research University of Wisconsin and Higher Education Center (CATIE) Madison Turrialba, Costa Rica LYNN R. BROWN CORNELIS DE H4AAN International Food Policy Research Institute World Bank Washington, D.C. Washington, I).C. ROLAND BUNCH KLAUS DEININGER Asociaci6n de Consejeros para una Agricultura World Bank Sostenible, Ecol6gica y Humana (COSECHA) Washington, D.C. Tegucigalpa, Honduras xi xii Contributors JOHN FARRINGTON DEEPA NARAYAN Overseas Development Institute World Bank London Washington, D.C. DOUGLAS A. FORNO STEFANO PAGIOLA World Bank World Bank Washington, D.C. Washington, D.C. ROBERTA V. GERPACIO PRABHU L. PINGALI University of California International Maize and Wheat Improvement Davis Center (CIMMYT) Mexico City, Mexico LAWRENCE HADDA1) International Food Policy Research Institute JULES N. PRETTY Washington, D.C. University of Essex Colchester, United Kingdom PETER HAZELL International Food Policy Research Institute AGNES R. QUISUMBING Washington, D.C. International Food Policy Research Institute Washington, D.C. JOHN HEATH World Bank HELLE MUNK RAVNBORG Washington, D.C. Center for Development Research Copenhagen, Denmark STEIN T. HOLDEN Agricultural University of Norway CARLOS REICHE As Interamerican Institute for Agricultural Cooperation (IICA) STEEN JOFFE San Jose, Costa Rica Consultant Linton, Kent, United Kingdom SARA SCHERR University of Maryland DAVID KAIMOWITZ Washington, D.C. Center for International Forestry Research (CIFOR) Bogor, Indonesia TJAART W. SCHILLHORN VAN VEEN World Bank EDWIN B. KNAPP Washington, D.C. Center for Tropical Agriculture (CIAT) Cali, Colombia NIGEL SMITH University of Florida ERNST LUTZ Gainesville World Bank Washington, D.C. JITENDRA SRIVASTAVA World Bank ALEXANDER MCCALLA Washington, D.C. World Bank Washington, D.C. HENNING STEINFELD Food and Agriculture Organization of the United RUTH MEINZEN-DICIK Nations (FAo) International Food Policy Research Institute Rome Washington, D.C. Contributors xiii WILLIAM SUNDERLIN DINA UMALI-DE[ININGER Center for International Forestry Research (CIFOR) World Bank Bogor, Indonesia Washington, D.C. GRAHAM THIELE Prograrna de Investigaci6n de la Papa (Potato Research Program; PROIMPA) Cochamba, Bolivia Acknowledgments We are grateful to the Swiss Trust Fund for Studies for volume. For most of them, this was a volunteer effort cofinancing the production of this compendium. based on each author's commitment to rural development. Special thanks go to Paul Egger, director of the Editing and proofreading by Elizabeth R. Forsyth, Agricultural Service in the Swiss Agency for composition by Patricia Hord, and indexing by Development Cooperation, who supported the idea of Patricia Deminna are gratefully acknowledged. a volume on sustainable rural development and gave Special thanks go to James Wolfensohn, president of valuable advice. the World Bank, for supporting a renewed emphasis We would also like to thank each contributor to the on the rural sector. xiv 1 Introduction Ernst Lutz, Hans P Binswanger, Peter Hazell, and Alexander McCalla Even though urbanization is increasing at a rapid rounding the concept of sustainability, surveys pace, most of the world's poor people still live in rural resource problems and indicative cost estimates, and areas. Thus anyone concerned for the poor should broadly discusses ways in which these problems can have a strong interest in rural development. A second be addressed. In chapter 3, Heath and Binswanger reason for giving appropriate attention to the rural show how natural resource degradation can often be sector is the enormous challenge facing developing policy-induced. Therefore, as a matter of priority, policy countries, which have to satisfy the nutritional needs distortions need to be removed to achieve better of their growing populations. incentives for higher growth and, in many cases, for The production increases of recent decades have environmental reasons. Pretty, in chapter 4, discusses been significant, as a result in part of expansion in in more detail what a policy framework looks like that area and in part of increases in productivity. However, is conducive to sustainable agriculture. As is argued, negative side effects-deforestation, desertification, given the diversity of situations, one does not need a salinization-may impair the capacity to produce fixed model; rather the social process of innovation increasing amounts of food in the future. Clearly, itself must be made sustainable. In chapter 5, Holden much more attention needs to be paid to increasing and Binswanger extensively survey recent research productivity without negatively affecting natural findings on decisionmaking by small farmers, given resources and the environment. In this book, we have the market imperfections and risks they face, and included chapters that summarize applied research on examine the irnplications for sustainable rural devel- ways to meet the challenge. opment policies. Trade liberalization is one part of We have grouped the chapters into three parts. The improving the overall incentive framework. first one takes a policy perspective, the second takes Anderson, in chapter 6, discusses the economic and an institutional and social perspective, and the third environmental impacts of this. In chapter 7, Bromley looks at rural development mainly from a technical distills key lessons from research and project experi- perspective. In actuality, all three are intertwined. ence dealing with property regimes and development Each of the three parts opens with an introduction and and assesses the policy implications. His chapter summary. would fit equally well into part II. A short chapter by The chapters in part I discuss policy perspectives Hazell (chapter 8) concludes part I. It is almost too and make the case that an appropriate policy frame- short to be a free-standing chapter, but it serves to work is essential to achieving sustainable rural devel- summarize part I by discussing the key ingredients of opment. We start with a chapter by Hazell and Lutz sustainable ruiral development policies in terms of (chapter 2) that deals with definitional issues sur- growth, equity, poverty, and environmental concerns. 1 2 Introduction The chapters in part II approach rural development institutional frameworks as well. It starts with a chap- mainly from an institutional and social perspective. ter by Lutz and others (chapter 16) that looks at soil They make the case that sustainable rural develop- conservation from a farmer's perspective. It argues ment must be participatory and decentralized and that for more empirical work that assesses the on-farm it must address gender issues (for productivity and profitability of soil conservation measures and for less other reasons) and not shy away from sensitive issues rhetoric. Srivastava and others (chapter 17) assess the like land reform. Narayan (chapter 9) outlines broad- importance of biodiversity conservation for agricul- ly what participatory rural development looks like and tural development and propose ways to mainstream what the project experience has been. Her principles agrobiodiversity into agricultural development. are illustrated and applied to watershed management Overuse and misuse of pesticides have been a major by Ashby and others in chapter 10. Similarly, reason for negative environmental impacts connected Farrington and Thiele (chapter 11) examine what with productivity increases. The ways in which pest local, participatory approaches to technology genera- management can be improved with integrated pest tion and dissemination look like and how they could management and minimal use of pesticides are be scaled up. Bunch (chapter 12) applies participato- addressed in chapter 18 by Schillhom and others and ry principles to innovative extension. Deininger in chapter 19 by Pingali and Gerpacio. Overuse of (chapter 13) explains the rationale for market-based irrigation water and insufficient drainage have led to or "negotiated" land reform and provides actual expe- waterlogging and salinity problems in many parts of riences from Colombia, Brazil, and South Africa. the world. Chakravorty, in chapter 20, makes the case Caldecott and Lutz (chapter 14) discuss decentraliza- that more market-based and basinwide approaches tion trends in developing countries and their implica- would use the increasingly scarce water more opti- tions for biodiversity conservation and, in particular, mally. A comprehensive treatment of livestock- the conditions under which a certain degree of decen- environment interactions is presented by Steinfeld tralization might be beneficial to conservation efforts. and others in chapter 21. How to distinguish between Part II concludes with a survey of applied research by appropriate and inappropriate deforestation and how Quisumbing and others (chapter 15) that shows the to reduce inappropriate deforestation are discussed by importance of properly addressing gender issues for Kaimowitz and others in chapter 22. Lessons from productivity growth and well-being. agroforestry projects in Central America and the Part III of the volume addresses a number of top- Caribbean are presented by Current and others in ics mainly from a technical perspective, while demon- chapter 23. We draw conclusions from the whole vol- strating the importance of appropriate policy and ume in part IV. Part I Policy Perspectives This part of the volume includes chapters that primar- are particularly acute when inappropriate incentive ily take a policy perspective, even though institution- systems encourage the overuse of such inputs. In con- al ones inevitably arise because policy and institu- trast, in many rainfed agricultural areas, the main tional issues are often intertwined. Here we describe problem is resource degradation. the contents of the chapters of part I (as a guide to the In the absence of comprehensive estimates of the reader) and summarize key points made by the vari- costs of resource degradation in agriculture in devel- ous authors. oping countries, the authors present indicative esti- Choosing among the many possible definitions of mates of the costs of deforestation, water depletion, sustainable development, Hazell and Lutz, in chapter waterlogging and salinization, fish stock depletion, 2, define sustainable agricultural development as agri- soil degradation, health effects, and biodiversity loss- cultural income traditionally measured in national es. They then discuss policy, institutional, and technical income accounts corrected for changes in the value of options that would reduce those costs. In particular, the stock of capital, which includes natural, man- they consider factors that form the overall incentive made, and human capital. According to their defini- system for natural resource management: prices and tion of sustainability, capital can be changed from one subsidies, nonprice interventions, externalities, infor- category into another, while keeping the total con- mation, poverty reduction through investments in stant. Yet they also recognize that there may be criti- resource-poor areas, property rights, costs of collec- cal levels of each type of capital (such as critical tive action, and technology design and research. They ecosystems) beyond which capital should not be sub- conclude that environmental problems are not the stituted. The chapter discusses externality issues, inevitable outcome of agricultural growth. Instead, environmental valuation techniques, and the design of they are usuallv associated with inappropriate incen- appropriate incentive schemes that internalize exter- tive systems, insufficient investment in resource-poor nal costs. areas, inadequate attention to poverty and social con- Although there is great diversity in agroecological cerns, and political economy systems where the rich conditions around the developing world, the authors and powerful extract rents. contrast two important types of systems because the In chapter 3, Heath and Binswanger state that environmental problems related to them are quite dif- reducing population growth and assisting poor farm- ferent. In irrigated agriculture, productivity has ers with resource management problems are urgently grown rapidly as a result of increased use of modem needed in many countries. But they argue that the inputs including irrigation water, high-yielding vari- removal of adverse policies should be given even eties, fertilizers, pesticides, and machinery. Problems higher priority, because those policies may, as their 3 4 Part 1. Policy Perspectives joint consequence, reduce economic efficiency, In chapter 4, Jules Pretty makes a case for sustain- increase poverty, and cause natural resource degrada- able intensification of agriculture. He describes five tion. These policies also tend to favor the moderniza- main groups of people with different views of future tion of large-scale farming at the expense of more agricultural development: business-as-usual opti- efficient and employment-intensive family farms. mnists, environmental pessimists, industrial-world-to- Their removal thus would be a win-win-win situation the-rescue group, new modernists, and the group in which the same policy changes could lead to more arguing for sustainable intensification of agriculture. growth, less poverty, and more sustainable natural As part of the last group, Pretty shows that regenera- resource management. tive and low-input (but not necessarily zero-input) Heath and Binswanger begin by discussing the agriculture can be highly productive if farmers partic- Boserup hypothesis, which suggests that, under the ipate fully in all stages of technology development right conditions, Ihigher population and market access and extension. He also suggests that the productivity can lead to improvements in natural resources rather of agricultural and pastoral lands is as much a func- than their deterioration. They note that the Boserup tion of human capacity and ingenuity as it is of bio- effects are far from automatic. Rather they are an out- logical and physical processes. Such agriculture seeks come of investment decisions made by farmers, who to integrate the use of a wide range of pest, nutrient, require a positive incentive regime and access to soil soil, and water management technologies. It also aims and water resources and markets. If adverse policy to create diverse enterprises within farms so that by- regimes deny these conditions, impoverished peasants products or wastes from one component or enterprise are forced to mine the land resources rather than aug- become inputs to another, thereby reducing the nega- ment them. This iis especially damaging if they only tive impacts on the environment. have access to marginal land or to humid tropical for- Pretty stresses the importance of not prescribing a est frontiers. Although the authors discuss the impor- concretely defined set of technologies, practices, or tance of agricultuiral policies, they also stress the policies, because, as conditions and knowledge importance of proper infrastructure (related to change, so must farmers and communities be encour- Boserup's market access condition), education (in aged and allowed to change. In Pretty's view, there- particular the education of girls), health, and macro- fore, sustainable agriculture is not simply a model to economic stability. be imposed; it is more a process for learning, an idea Using recent studies of the Machakos district in that is also supported by Bunch (chapter 12). What Kenya and soil degradation in Ethiopia, the authors needs to be made sustainable is the social process of illustrate the depenidence of the Boserup effects on the innovation itself, which implies an enhanced capacity policy environment that governs the investment to adapt to unexpected changes and emerging uncer- incentives of farmers. They also examine growth, the tainties. For such innovation to happen, it is important use of land and labor, and the degradation of margin- to have an enabling policy environment. In this con- al natural resources in Colombian agriculture. text, Pretty recognizes that in practice, policy is the Heath and Binswanger admit that eliminating the net result of the actions of different interest groups adverse policies a:nd privileges of rural elites, which pulling in complementary and opposing directions, reduce efficiency and increase poverty and resource often making the outcome uncertain. degradation, is an extremely difficult task. These poli- In chapter 5, Holden and Binswanger analyze cies did not come about by accident, but as the conse- small-farmer decisionmaking in the context of imper- quence of a historical evolution that involved bargain- fect markets and its implications for efficiency and for ing among politically strong groups looking out for sustainability of natural resource management. They their own interests at each point in time. Analyzing the argue that new policies to stimulate sustainable rural joint consequences of the policies may support the development are much needed and that, to succeed, reform effort but is unlikely to be sufficient. Peasants these policies have to build on an understanding of the have a low collective action potential, and steps must decisionmaking environment and behavioral re- be taken to provide them with a greater input into pol- sponses of small farmers. icymaking. Unless some political actors with enlight- Small farmers represent a major link between the ened hearts and minds see it as in their interest to bring economy and the environment because their liveli- poor peasants more strongly into the policy process, hoods depend directly on the use of land resources policy and land reforms will continue to be slow. (soil and vegetation). The fact that small farmers rep- Part I. Policy Perspectives 5 resent both production and consumption units compli- tainable intensification given stagnant technology. cates matters, particularly when market imperfections These rural pockets appear to be in a poverty- cause their production and consumption decisions to environment trap that resembles a neo-Malthusian be nonseparable. Nonseparability implies that con- scenario. They represent one of the most significant sumption needs and asset distribution may have sig- challenges to the developing world. Induced techno- nificant impacts on production decisions and thus the logical and institutional innovation appears to be the management of natural resources. best means, together with stable macroeconomic con- Even though small farmers tend to be only partly ditions, of switching from a Malthusian to a integrated into markets, they are affected by market Boserupian development path.. Agricultural research imperfections. These include missing markets, partly and extension, infrastructure development, basic edu- missing markets (rationing, seasonality), thin markets cation, and family planning are also important ele- (imperfect competition), and interlinked markets. ments, together with land tenure policies and other Covariate risk, asymmetric information, and transac- measures to provide secure livelihoods. tion costs are the causes of pervasive market imper- Given the complexity of these issues, Holden and fections. These problems are even more acute in areas Binswanger suggest a strategy with pilot projects and with low population density where the population flexible project: design accompanied by systematic derives a living from marginal land resources. monitoring and evaluation. This approach would Small farmers largely are rational and respond to allow for maximum learning, adjustment, and scaling changes in the set of constraints and opportunities up as confidence is gained. they face. But inefficiencies may still accrue due to In chapter 6. Anderson notes that the world, during economies of scale, poverty and subsistence con- the past decade, at last began reducing distortions to straints, unequal access to credit and markets, and agricultural incentives in significant ways. The long- various policies disfavoring small farmers. It is thus term growth in agricultural protection in high-income possible that small farmers are too poor to be efficient countries slowed, and the heavy direct and indirect from society's perspective. These inefficiencies may taxation of agriculture was reduced in numerous also lead to underinvestments in conservation. lower-income countries. Market imperfections are more likely to hurt small Anderson addresses the concern that agricultural and poor farmers than large and rich farmers because trade liberalization in industrial countries may have it is more probable that poor farmers will be rationed negative effects on rural and natural environments, on out of credit markets, for example, and will be less agriculture's resource base, and on food safety. He is able to solve their problems through consumption confident that the removal of trade distortions will in smoothing (coping strategies). This explains why the almost all situations boost economic welfare. In the most vulnerable households are most likely to diver- case of agricultural policy reform, it would also sify their production. There may thus be a case for improve the environment on balance in many intervention on the basis of efficiency as well as equi- instances. Where it would not, the reason is common- ty. The possibility of a reinforcing relationship ly that governments do not have appropriate environ- between poverty and environmental degradation mental policies in place. But Anderson also shows strengthens the case for appropriate intervention. that many policy changes can have negative environ- According to Holden and Binswanger, population mental effects even when the net welfare effect might growth may stimulate sustainable intensification in cer- be positive. H:ence, more attempts will need to be tain conditions, while in others it may lead to land made to quantify the environmental changes likely to degradation. They believe that incentive structures and be associated wvith trade and other economic reforms. how these are affected by policies have a strong impact Another concern is expressed by the aid communi- on whether small farmers are able to choose a sustain- ty, which is worried that the reduction in government able and welfare-improving development path or are attention to agriculture in high-income countries is forced onto a nonsustainable and welfare-reducing one. contributing to a loss of interest in funding agricultur- Particularly in Africa where policy regimes have al developmen,t assistance programs, including inter- been very adverse and in other areas that are poorly national agricultural research. To date, comparative integrated into markets, small farmers appear to static economic modelers have paid little attention to degrade their environment because population growth these environmental or research and development fragments scarce land resources and leads to unsus- effects of reform. 6 Part 1. Policy Perspectives Chapter 6 then examines the consequences for er institutional structure. Property regimes are struc- incentives to invest in agricultural research of a unilat- tural attributes of an economy that provide agents eral removal of policies that keep domestic food prices with domains of choice within which they may act; artificially low. At higher food prices, the environmen- property regimes must be understood as social con- tal cost-degradation of the farm sector's natural structs whose nature and existence are antecedent to resource base-of not reducing distortions in the mar- what we ordinarily regard as "economic" behavior. kets for farm chemicals and irrigation water is higher They must be seen as occurring logically and socially than it would be if product prices and hence input prior to the economizing behavior we attempt to ana- intensities were lower. Another implication is that the lyze and explain. returns to developing countries from investing in agri- All property regimes require external legitimacy. cultural research will be higher in the future as a result To have a property right is to know that the authority of those higher prices. The international agricultural system will defend you against the predatory behavior research centers too will contribute more from their of others. Degradation of natural resources often research. Insofar as developing countries' reforms occurs as a result of a breakdown of authority over the involve raising prices not only for farm products but resource, caused by internal conflict or by intrusion also for chemical inputs and water, the rewards will be into the area by outsiders. especially high from generating environmentally Ambiguous rights regimes have ambiguous effi- friendlier precision technologies as distinct from those ciency and distributional consequences. Ambiguity of that use heavy doses of chemicals and water. some general meager resources can work to the Anderson expects that environmental groups will advantage of the poor. However, when development pressure the World Trade Organization to place the projects increase the income flow from natural issue of trade and environment linkages on the agen- resources over which rights were previously ambigu- da of its next round of talks. It is too early to tell what ous, then powerful individuals at the local level will form that might take, but it is unlikely to result in usually figure out a way to expropriate at least a agreement to do anything as extreme as allow coun- part-if not the majority-of these new income tries to use trade measures against countries with less- streams. stringent environmental policies. Even calls to have Property relations must be specified prior to the an environmental impact assessment of what is nego- implementation of development projects. As a general tiated in the way of reducing trade restrictions are rule, projects are placed in poor areas that warrant impractical, because, if such an assessment were to outside assistance precisely on account of the inco- produce amendments, the whole agreement could herence and contentious nature of local institutional unravel. Such calls will be less likely, however, as arrangements-of which property relations are of more empirical research demonstrates the (expected) paramount importance. Therefore, every development absence of significant adverse environmental conse- intervention must be preceded by a concerted effort to quences of trade liberalization. ensure that the institutional arrangements have been In chapter 7, Bromley addresses property relations modified so as to ensure that the benefits go to the in the context of developmental efforts. His thesis is intended beneficiaries. that property relations are central in development Resource degradation is the result ofproblems that because they connect people to one another with precede property regimes. The most serious resource respect to land and related natural resources. Property degradation problems occur in situations of open relations are simply socially constructed contractual access in which no property rights exist. But such sit- arrangements among a group of people for mediating uations are themselves manifestations of larger insti- individual and collective behaviors regarding objects tutional problems, only one of which may relate to the and circumstances of value to members of the com- particular property regime in place. munity. Development projects are sometimes less Ecological variability demands flexible institu- successful than they would otherwise be precisely tions and actions. Ecological settings that exhibit because the existing property relations have been great variability require property regimes that allow ignored or misunderstood. quick and low-cost adaptations to new circumstances. Bromley offers nine main lessons for development Livestock, because of their mobility, provide this flex- practitioners. The first one is quite obvious but merits ibility in a way that land, which is immobile, cannot being recorded: property regimes are part of the larg- possibly equal. Part I. Policy Perspectives 7 Resource degradation is contextual. Once we rec- rience can be sumamarized as the five "I"s for agricul- ognize that the social and economic meaning of vari- tural growth: innovation, infrastructure, inputs, insti- ous assets differs across ecological and social circum- tutions, and incentives. It later became evident that stance, then it follows automatically that the social governments also need to promote the role of the pri- and economic meaning of "resource" also differs. vate sector in agricultural production, marketing, and Security of tenure is necessary for efficient invest- service provision. ments, but a title, as such, is not necessary. Security of In the 1970s and 1980s, attention focused on tenure is a necessary condition for investment in, or using agricultural development to reduce poverty and the "wise" management of, land and related natural food insecurity as well as to spur growth. The lessons resources. The contrary proposition-investment in that emerged from that era can be summarized as fol- land is a necessary condition to secure "title" in that lows: (a) broadly based agricultural development land-is equally probable. should be promoted, (b) participation of affected Mobilizing local interests can improve the chances stakeholder groups is essential, (c) land reforms may of a program's success. In natural resource projects, be necessary, (d) investments in human capital are there is growing recognition that "community-based very important, (e) women's concerns must be conservation" is an important innovation in creating addressed, and (f) the rural nonfarm economy should the conditions for sustaining both natural resources be actively encouraged. and local commitment to that sustainability. But The new priority for environmental sustainability although creating a sense of local ownership in par- that has emerged in the 1990s does not negate the ticular resources can solve much of the incentive need for agriclulture to continue contributing to problem, such arrangements may still require an over- growth, poverty alleviation, and increased food secu- sight role for some higher authority if there is a rity. It is just thal agriculture is now required to do this national or international interest in the particular in ways that do not degrade the environment. Hazell resource or if externalities are present. argues that two ifndamentally different types of envi- Bromley, in chapter 7, essentially argues that the ronmental problems are associated with agriculture. outcomes of development activities are fundamental- The first type arises in intensive farming systems and ly dependent on institutional arrangements in general is associated with the misuse of modern inputs (irri- and property relations in particular. gation water, fertilizers, and pesticides). The second Hazell, in chapter 8, states that agricultural devel- type arises in extensive farming systems and is asso- opment should simultaneously contribute to four prin- ciated with rapid population growth, poverty, and cipal goals: growth, poverty reduction, food security, growth in agricultural productivity insufficient to and sustainable natural resource management. meet the increasing need for food and livelihood Experience suggests that there can be a high degree of (elaborated in chapter 2). complementarity between these four goals if agricul- A general finding of the part on policy perspec- tural development (a) is broadly based and involves tives is that significant strides can be made toward small- and medium-size farms, (b) is market driven, sustainability. But clearly, appropriate policies and (c) is participatory and decentralized, and (d) is driven institutions are a precondition. These are important by productivity-enhancing technological change that for growth, equity, and environmental reasons; they does not degrade the resource base. are also needed to find appropriate technical solu- Back in the 1950s and 1960s, agricultural devel- tions. A key question that remains is whether the opment experts tended to be interested in agricultural political process can produce the needed policies and growth, and the lessons that emerged from their expe- institutions. 2 Integrating Environmental and Sustainability Concerns into Rural Development Po lficies Peter Hazell and Ernst Lutz Many developing countries have achieved impressive poverty and hunger, driving rural people to expand growth rates in agriculture in recent decades. India, cultivation into less-favored, often environmentally for example, which was threatened by hunger and fragile areas, such as forests, hillsides, and wetlands, mass starvation in the 1960s, is now self-sufficient in and to reduce fallow periods to the point of depress- staple foods even though her population has more ing soil fertility. than doubled. Yet in spite of this success, serious con- Continued agricultural growth is a necessity, not cerns remain for the future. First, hunger and malnu- an option, for most developing countries. Further, this trition persist in many countries, often because past growth must be achieved on a sustainable basis so as patterns of agricultural growth failed to benefit the not to jeopardize the underlying base of natural poor adequately. Second, agricultural demand will resources, and it must be equitable if it is to contribute grow along with population growth and rising per to the alleviation of poverty and food insecurity. capita incomes, and this will require continuing Government policies, institutional development, agri- increases in agricultural productivity. Yet growth in cultural research, and projects at local, regional, and yields appears to be slowing while the prospects for national levels need to be designed and implemented expanding cropped and irrigated areas are limited. with these objectives in mind. Third, if not checked, environmental problems associ- ated with agriculture could threaten future levels of What Is Sustainable Agricultural Development? agricultural productivity as well as the health and well-being of rural people. There are many alternative definitions of sustainabili- Two basic types of environmental problems are ty and of sustainable agricultural development associated with agriculture. Most of the successful (Pezzey 1992). Conceptually, those most relevant for breakthroughs in productivity have occurred in more- our purposes are measures of agricultural income that favored agroecological zones and have been based on are corrected for changes in the value of the capital intensive use of irrigation water, fertilizers, pesticides, stock, especially the stock of natural resources used in and other modern inputs. Agriculture based on inten- agricultural production. As Sir John Hicks has argued, sive use of these inputs is prone to mismanagement sustainable income is the maximum value that a per- that leads to environmental degradation (particularly son or society can consume during a specific period of when the system of incentives is inappropriate). time and still expect to be as well off at the end of the However, where governments have neglected to period as at the beginning (Hicks 1946: 172). intensify agricultural production through the use of Applying this concept, an agricultural system can be modern technology, population growth has worsened said to be sustainable if the amount of income extract- 9 10 Integrating Environmental and Sustainability Concerns into Rural Development Policies ed for consumption each year can be sustained over this way raises three sets of issues. First, at what level time. This requires that the value of the capital stock of society are environmental costs and benefits to be not be depleted over time; hence sufficient income measured? Because resource degradation often has must be set aside or otherwise forgone to replenish spillover or externality costs for nonusers (as when any capital depreciation or losses incurred in the pro- the degradation of watershed protection areas leads to duction process. For this purpose, capital is defined to the silting of irrigation systems downstream or when include all natural resource, human, and man-made water pollution causes health problems throughout capital assets. river basin systems), the more aggregated the concept A further distinction is made between weak and of society used, the greater the total environmental strong sustainability, where the difference depends on costs. Thus levels of environmental degradation that whether the defiinition permits substitution between may be considered acceptable to farmers or rural different types of capital assets. With weak sustain- communities may be unacceptable at the national or ability, substitution of different forms of capital is international levels if the externality costs are signifi- allowed, but the total value of capital must remain cant. The reverse can also be true: resources that gen- intact or increase. In contrast, with strong sustainabil- erate environmental benefits may be undervalued at ity, substitution of different forms of capital is not the farmer or community levels because their benefits allowed, and an effort is made to keep constant the are captured at broader national or international lev- stock of each form of capital, including the value of els. For example, farmers may convert forests in natural resources (Pearce, Markandya, and Barbier excess because they do not capture the forest's full 1989). Serageldin (1996) has suggested a compromise value in modulating climate or in preserving definition called sensible sustainability. This allows biodiversity. Clearly, assessing sustainability at the for capital transformations-that is, of natural capital farm or community levels is not sufficient in the pres- into human capital (through education) or into man- ence of important externalities, and regional and made capital-but also recognizes that critical levels national assessments are also needed to guide policy of each type of capital may exist beyond which con- decisions. International or global assessments are also cerns about substitutability can arise.' Because the required where important international externalities exact boundaries of the critical limits for each type of are involved (such as the impacts of agriculture on capital are not known, "it behooves the sensible per- global climate change), even though these may be son to err on the side of caution in depleting resources perceived as less relevant for decisionmakers in indi- (especially natural capital)" (Serageldin 1996: 8). vidual countries. "Weak" and "sensible" definitions of sustainabili- The second issue concerns methods of valuing ty have the advantage that they do not suggest freez- environmental costs and benefits. Environmental val- ing or preserving all natural resources at existing lev- uation techniques have been increasingly used in els. Rather, they permit tradeoffs between growth and empirical studies during the past decade or two. environmental objectives: resources can be degraded Values can be grouped into direct use values, indirect or depleted to increase production, if compensatory use values, option values, and existence values. Direct investments are made in other forms of capital to sus- use values are associated with food production and tain the stream of consumable income over time. consumption, biomass, recreation, and health. Indirect Decisions about the degradation or depletion of nat- use values are associated, for example, with ecologi- ural resources should, therefore, depend on whether cal functions, flood control, and storm protection. the social benefits accruing to the users are greater or Option values are associated with the preservation of smaller than the social costs (Turner 1993: figs. 1.2 resources that have probable but uncertain value in and 1.3). For example, because some resources are the future, for example, biodiversity conservation. renewable (trees, soil nutrients) or have adequate sub- Finally, existence values include the value that people stitutes (farm trees can replace woodlands for fuel- place on the mere existence of resources from which wood), they need not all be preserved at current levels they may never directly or indirectly benefit (from the or at all points in time. In cases of critical habitats, existence of parks or species, which they may never however, the costs of degradation are likely to be so actually visit or see). As a rule, the more tangible the high that conversion to agriculture is not a socially costs and benefits, the more reliable the estimates that optimal decision. can be produced. Various methods have been devel- To achieve agricultural sustainability defined in oped for estimating environmental costs and benefits Peter Hazell and Ernst Lutz 1] for use in project and policy analyses or environmen- * Perennial flooding of rice paddies and continuous tal accounting (see, for example, Dixon and others rice culture lead to a buildup of micronutrient defi- 1992, Lutz 1993, Lutz and Munasinghe 1994, and ciencies and soil toxicities, formation of hard pans Hamilton and Lutz 1996). in the soil, and a reduction in the nitrogen-carrying Empirical valuation work in developing countries capacity of the soil. Work at the International Rice at the conjunction of agriculture and the environment Research Institute shows that farmers are having to is still quite limited. Continued development of data- increase the amount of fertilizer they use over time bases and practical methods of valuing environmental simply to maintain existing yields in intensive costs and benefits at the farm, community, and nation- paddy fields (Pingali 1992). al levels will be important for conducting a rigorous * Excessive and inappropriate use of pesticides deteri- analysis of policy and agricultural research priorities orates the quality of water, poses health hazards for for promoting sustainable agricultural development. humans, and leads to resistance of pests to pesti- The third issue concerns the design of incentive cides. Farmers can become trapped into using more schemes that induce farmers and communities to fac- and more frequent sprays to control pest damage. tor important externality costs or benefits into their * An increasing reliance on a few carefully bred own resource management decisions. As a general crop varieties contributes to a loss of genetic diver- finding, undistorted macroeconomic and sectoral sity and to a common vulnerability to the same policies tend to provide better incentives from an pest- and weather-related risks. In some cases, mil- environmental perspective than highly distorted poli- lions of hectares of land have been planted to the cies (Warford, Munasinghe, and Cruz 1997 and Lutz same wheat ar rice varieties, and widespread loss- and Young 1992). Subsidies for agricultural inputs es have occuiTed because of the outbreak of a sin- tend to be particularly costly in terms of economic gle pest or disease. The loss of traditional varieties efficiency, government budgets, and the environment also reduces the pool of genes available for breed- (World Bank 1997: chap. 4). ing plants capable of resisting evolving plant pests. Environmental Problems in Agriculture Rainfed Agriculture Despite considerable diversity in agroecological con- Population growth and poverty in many rainfed lands ditions around the developing world, it is useful to are causing serious resource degradation. Until distinguish between the environmental problems recently, natural resources were generally abundant in related to intensive, irrigated farming areas and those these areas, and, once used, farmers could allow related to more extensive, rainfed farming areas. resources time to recover through rotations and shift- ing cultivation. Moreover, many of the more fragile Irrigated Agriculture lands were not farmed at all. Today, rainfed lands fre- quently must support moderate to high population Productivity has grown fastest in irrigated agriculture densities, providing not only increasing amounts of because of the increased use of modern inputs: irriga- food but basic essentials such as fuelwood, water, and tion, fertilizers, pesticides, high-yielding varieties, housing. In the absence of adequate increases in agri- and machinery. However, this intensification has also cultural productivity to secure their livelihoods, farm- increased the potential for the inappropriate use of ers reduce needed fallows and expand into new areas, modern inputs, particularly when inappropriate incen- many of which are environmentally fragile and easily tives prevail. degraded. Environmental problems associated with The following major environmental problems are rainfed farming include (Scherr and Hazell 1994): associated with intensification in irrigated areas * Conversion of primary forest to agriculture, with (Pingali and Rosegrant 1994): loss of biodiversity, climate change, and exposure * Intensive use of irrigation water in areas with poor of fragile soils drainage can lead to waterlogged soils and a rise in * Expansion into steep hillsides causing soil erosion the water table. In arid and semiarid areas, this in and lowland flooding turn causes salt to build up in the soil. Salinization * Degradation of watershed protection areas, with reduces yields and can eventually lead to abandon- downstream siltation of dams and irrigation sys- ment of land. tems and increased flooding 12 Integrating Environmental and Sustainability Concerns into Rural Development Policies * Shortened fallows with loss of soil nutrients and ages are often compounded by wasteful and ineffi- organic matter, resulting in declining yields cient use and by pollution from the dumping of * Increased pressure on common property resources human and animal wastes and industrial pollutants. (woodlands and grazing areas), with breakdown of Water quality is inadequate for an estimated 1 billion indigenous institutions that regulate and manage people, and related health costs and fatalities are high. these resources, leading to open-access regimes Agriculture is the major user of water in many devel- and resource degradation oping countries and is a significant contributor to * Declining resilience in ecosystems, with reduced water quality problems. ability to rebound from stresses such as droughts. Water problems are in large part policy-induced: Overuse of fertilizers and pesticides is much less cost recovery rates in irrigation amount to between 20 common in rainfed agriculture because the current and 25 percent of true costs, and subsidies cost gov- levels of use are still low. Exceptions include a few emments some $20 billion to $25 billion each year high-value cash crops such as tobacco and vegetables. (de Moor and Calamai 1997). Electricity for pumping A bigger problem is insufficient use of fertilizers to water is also subsidized in many countries. These sub- replace soil nutrients lost through increased cropping sidies encourage excessive and inefficient use of and reduced fallows. water, and this not only aggravates water scarcity problems but also leads to waterlogging and saliniza- Costs of Resource Degradation tion of irrigated land. In China, for example, salinization already affects 23 percent of the total irrigated area The full extent of environmental damage from agri- (Umali 1993: table 3.1). In Pakistan, which has the culture is difficult to assess. Here, we attempt to sum- world's largest contiguous surface distribution sys- marize indicative estimates from various studies. As tem, an estimated 3.5 million hectares (or 25 percent will be shown throughout this volume, many of these of the irrigated area) is affected by waterlogging and costs are associated with market, institutional, and salinity, and 8 percent of these lands are seriously policy failures. degraded (Pinstrup-Andersen and Pandya-Lorch 1994). The detrimental effects on crop productivity Deforestation and Forest Degradation can be serious. Yields of major crops in the affected areas of Egypt and Pakistan have declined signifi- According to the Food and Agriculture Organization cantly (Barghouti and Le Moigne 1991). Further, of the United Nations, deforestation is the sum of all there are also negative social impacts because rich transitions from n,atural forest classes (both continuous farmers generally benefit from water-related subsidies and fragmented) to all other classes. Based on this def- more than the poor. inition, deforestation during 1980-90 amounted to 8.2 percent in Asia, 6.1 percent in Latin America, and 4.8 Fish Stock Depletion percent in Africa (see chapter 22 in this volume). Not all of this can be attributed to agricultural conversion. Fish protein is an important part of the diet in many Forest degradation represents a decrease of density or countries. Fish production from aquaculture is increas- increase of disturbance in forest classes. Even when ing, but output from the marine fisheries sector appears there may be little deforestation as such, degradation can to have peaked at around 85 million to 90 million tons. imply major losses of forest products and ecological ser- Fish stocks are being depleted due to ineffective man- vices, such as reductions in watershed and climate services, agement of common property resources. An important biodiversity, and carbon storage. Because private problem is that in many countries, the government sub- agents generally do not consider these costs in their sidizes the maintenance of excessively large fishing decisions about logging and forest conversion, rates fleets, and this encourages overfishing. Estimates of the of depletion can be excessive from a societal view. total subsidies involved range from $11 billion to $54 billion (Milazzo 1997). Uncontrolled and wasteful fish- Water Depletion, Water Degradation, ing (including the throwing away of unwanted catches) and Waterlogging and Salinization have depleted marine stocks, driving once-common species like cod and halibut to commercial extinction Water is a vital economic asset that is already becom- and threatening the livelihood of millions of people. ing scarce in many countries. Moreover, water short- Water pollution also affects fisheries. Peter Hazell and Ernst Lutz 13 Soil Degradation of these costs are relatively easy to measure, such as the costs of dredging water channels or the costs of Good soil is a crucial factor of production in agricul- reduced fish catclhes as a result of water with high tur- ture, and maintaining the depth and quality of soil is bidity. In other cases, costs are more hidden (like sed- important for preserving agricultural productivity in imentation of reservoirs used for hydro power or irri- the future. The costs of soil degradation can be mea- gation) or not easily accounted for (like the damage to sured by lower yields or, where farmers can compensate coral reefs from silt deposits). through more intensive use of manure and fertilizers, In assessing the depreciation of a capital stock-in by higher production costs. What is not so easily mea- this case the stock of soil-by discounting the future sured is the reduction in value of the stock of this nat- income losses, the discount rate assumed is highly ural capital. Inasmuch as a reduction in value has a influential on the actual results. For nations as a whole direct impact on future agricultural production, the and for national income and accounting purposes, one loss in future yields can be measured. But it is not just may assume discount rates of 5 to 10 percent. Higher the amount of soil eroded that is the key to future rates are relevant for individual farmers, particularly yield losses; the amount (depth) and quality of the soil for those with small farms (Cuesta, Carlson, and Lutz that remains, including the soil that is deposited in 1997). other fields used for agriculture, are also important. Research on soil erosion and degradation is com- Other important factors that are generally not plicated by the influence of stochastic variables, such recorded also affect agricultural productivity (such as as rainfall intensity. Much more work is needed, par- soil compaction, pH, and organic matter content). ticularly work that extends the focus from plot-level Field studies showing the relationships between soil experiments to the watershed and from purely techni- loss, degradation, and deposition and agricultural pro- cal aspects to economic and social aspects as well. ductivity in developing countries are rare and, even where such information exists, cannot easily be Health Effects extrapolated because soil degradation differs from site to site because of differences in the type of soil, slope, Injudicious use of pesticides and a lack of safe spray vegetative cover, and cultivation practices. Not sur- equipment and protective clothing suitable for tropi- prisingly, therefore, estimates of the costs of soil ero- cal conditions are causing significant short-term as sion and degradation, particularly at national levels, well as long-term chronic health problems. Studies in are few and vary widely. Indonesia, Malaysia, and Sri Lanka in the early 1980s In work on Indonesia, Repetto and others (1989) suggest that 12-15 percent of farmers who use pesti- have estimated that capitalized losses for future pro- cides have been poisoned at least once in their careers ductivity are approximately 40 percent of the annual (Jeyaratnam, Lun, and Phoon 1987). But long-term value of upland farm production in Java, or between health effects can also be significant (Pingali and $340 million to $406 million a year. Of this amount, Roger 1995). In addition to exposure during spraying, $315 million are estimated to be on-farm losses of the unsafe storage, handling, and disposal of pesti- productivity and the remaining $25 million to $80 cides subject fanners and farm families to high levels million to be downstream damages from siltation. For of health risk and can contaminate the ecosystem. Mali, Bishop and Allen (1989), assuming a 10-year time horizon and a 10 percent discount rate, have esti- Biodiversity Losses mated that the present value of current and future net farm income forgone nationwide due to one year of There has been serious loss of traditional crop vari- average soil loss amounts to between 4 and 16 percent eties, while large areas are planted to a few modem of agricultural gross domestic product. The World varieties (for example, IR36 rice has been planted on Resources Institute (1991) has estimated that soil more than 10 mnillion hectares in Asia). This varietal depreciation amounts to almost 10 percent of Costa specialization has occasionally led to widespread crop Rica's annual agricultural production. But in our losses due to outbreaks of diseases and pests (brown view, all of these estimates appear to be high; in any planthopper in rice) and requires devoting more case, more empirical work is needed. shares of research resources to "maintenance" In addition to on-site effects of soil erosion and research and germ plasm conservation programs degradation, one must consider off-site effects. Some (Anderson, Hazell, and Evans 1987). 14 Integrating Environmental and Sustainability Concerns into Rural Development Policies Policy, Institutional, and Technical Options Policy Issues and Options to Reducing Resource Degradation Several factors impinge on the incentives for manag- In thinking about possible solutions, it should be rec- ing natural resources: technology design, poverty, ognized that resources cannot be protected or con- property rights, externalities, costs of collective served without finding acceptable means of liveli- action, prices, government nonprice interventions, hood for the people who use them. This is particularly and access to information about the condition of true in fragile, rainfed areas where poverty is a major resources. Focusing on these factors provides a useful force driving the degradation of many resources. way of discussing the kinds of changes needed to Sustainable and poverty-reducing agricultural inten- move toward environmentally sustainable agricultur- sification in problem areas is often key to solving al development. In the following paragraphs, we dis- resource degradation problems, particularly because cuss these factors in the same policy, institutional, and solutions involviing interregional migration and eco- technology groupings used in this book. nomic diversification into nonfarm activities are lim- ited. Prices and Subsidies It also needs to be recognized that future food needs will be difficult to meet if countries revert to Inappropriate prices for inputs and outputs can low-input, low-yield agricultural technologies. encourage farmers to degrade resources by making Continued increases in yield will be critical, with unsustainable practices more profitable. Inappropriate limited possibilities for reducing dependence on prices can arise from externality problems that distort water and fertilizers. Moreover, short of any major market prices from their socially correct values. For biotechnology breakthroughs, the scope for meeting example, rural market prices for fuelwood and char- all additional food needs from irrigated lands is lim- coal generally undervalue the true cost of the wood, ited, and rainfed lands will need to become increas- because the market price does not capture the envi- ingly important sources of agricultural output. This ronmental benefits of trees for soil conservation, will require that increasing shares of the resources modulation of local climate, and so forth. But govern- available for agricultural research be devoted to rain- ments also distort market prices. Inputs (especially fed areas and that these areas receive higher priority water, fertilizers, and pesticides) are often subsidized, in public investment programs. encouraging excessive use. De Moor and Calamai Policy interventions that seek to overcome envi- (1997) offer the following estimates for developing ronmental problems in agriculture need to be based countries: water subsidies, $20 billion to $25 billion; on a proper understanding of why farmers degrade fertilizer subsidies, about $6 billion; pesticide subsi- natural resources. Why, for example, do farmers dies, about $4 billion. Many governments have also often seem to overgraze rangeland, deplete soil nutri- kept agricultural output prices too low through export ents and organic matter, and overuse irrigation water, taxes and overvalued exchange rates, reducing farm- pesticides, and nitrogen, when these actions cause ers' profits and their returns to investing in the con- health problems and reduce future incomes for them- servation and improvement of natural resources. selves, their children, and the communities in which (Fortunately, the average degree of taxation of agri- they live? culture in developing countries has been reduced dur- The answer lies with incentives including dis- ing the past decade, see chapter 6.) count rates. Farmers are not irrational. To the con- In many cases, market liberalization, removal of trary, they maximize income and minimize risks in a subsidies, and realignment of domestic and border dynamic context and often under harsh conditions prices will improve incentives for sustainable farming and serious constraints. So, for example, they practices. A good example is the favorable impact in degrade resources when there are good economic and Indonesia of removing subsidies on pesticides. Not social reasons for doing so (when the benefits they only has pesticide use dropped dramatically, but the obtain exceed the perceived costs that they, as indi- adoption of integrated pest management practices has viduals, must bear). If the management of natural spread widely and yields have increased (Ruchijat and resources is to be improved, these economic and Sukmaraganda 1992). Yet market solutions alone are social incentives will need to be changed in appro- not always adequate. Externality problems, as for priate ways. example with tree products, may require adjustments Peter Hazell and Ernst Lutz 15 in domestic prices from border prices to reflect envi- Externalities ronmental costs and benefits to the country. But this is not always practical, especially for export products, Externalities arise when the costs associated with unless competing countries also make similar price resource degradation are not fully borne by the indi- adjustments or importing countries are willing to pay a viduals causing the problem. For example, removing price premium through green labeling schemes. This trees that protect watersheds may be privately benefi- requires more effective international collaboration on cial to the individual farmers who do it but can lead to environmental problems than has been achieved so far. soil erosion and downstream flooding and siltation of irrigation works that are costly to society. Externalities Nonprice Government Interventions are prevalent in. agriculture and are a particularly important factor in explaining water pollution, Government nonprice interventions can have signifi- overuse of agrochemicals, deforestation, and loss of cant effects on the incentives and opportunities avail- biodiversity. able to farmers in making choices about technology In countries with well-developed institutions, care- and resource use. Public investments in, for example, fully crafted taxes, subsidies, and government regula- rural roads, schools, clean water, health centers, fain- tions can be effective in overcoming externality prob- ily planning, and soil erosion control can create new lems. These instruments are typically less effective in opportunities in farm and nonfarm activities for rural developing countries, particularly in remote rural people and reinforce positive incentives for sustain- areas. Some externalities can be reduced by pricing able resource management. Education and improved marketed goods appropriately. But in most cases, health can also help increase opportunities for migra- solutions must be sought through local governments tion, reducing the population pressure on resources and organizations. Central governments can help by and providing capital flows through remittances for providing environmental guidelines, contributing to investments in agriculture. Drought-relief interven- resource monitoring systems (providing aerial pho- tions, such as food-for-work programs, can be partic- tographs, remote-sensing data, downstream water ularly helpful in relieving the pressure on resources testing), and protecting designated conservation areas when they are most vulnerable. (parks and other environmentally valued sites). But But inappropriate government interventions can effective action requires the joint involvement of be environmentally destructive. Construction of new people who misuse resources and of people who are roads or settlement schemes in environmentally frag- affected most immediately by that misuse. Empower- ile areas can be destructive (for example, some of ment of local action groups is important, and this can Indonesia's transmigration projects). Rigid regula- sometimes be reinforced by appropriate changes in tion of land and forest use can also inhibit develop- property rights. For example, ownership rights over ment of sustainable but more profitable patterns of watershed protection areas and waterways can be resource use (farmers in the Philippines and Thailand bestowed on communities or local organizations were initially prevented from investing in new types rather than on individuals or the public sector. of productive trees in protected hillside areas). Drought-relief interventions can also backfire if they Information on the Condition of Natural Resources are heavily subsidized. Subsidized drought insur- ance, for example, increases the profitability of more We believe that farmers respond rationally to eco- risky farming practices, some of which may be envi- nomic and social incentives but that sometimes they ronmentally unsuitable for drought-prone areas. may be poorly informed about the consequences of A key issue for policy is determining the relative their actions. For example, they may not be fully weight to be given to resource-poor areas in allocat- aware of the lon,ger-term consequences of soil degra- ing public investments among rural areas. Resource- dation of particular farming practices, the effect of poor areas have been sadly neglected in the past, yet removing trees on soil erosion on neighboring farms, the prevalence of poverty and environmental degra- or the effect of siltation problems downstream. dation in those areas implies that increasing the Integrated pest management programs have shown social and environmental benefits of public invest- that many farmers cannot distinguish between harm- ments may help to offset their lower efficiency ful and beneficial insect species and are often inclined returns. to spray both. 16 Integrating Environmental and Sustainability Concerns into Rural Development Policies Better information needs to be provided; programs costs are considered. At the same time, the returns to increase public awareness that use schools, non- from investments in resource-poor areas are becom- governmental organizations (NGOs), and the media ing more attractive. Recent developments in research can be effective. Simple resource-monitoring systems on farming systems, soil management, and agro- also need to be established to help communities track forestry suggest that there may be potential for greater important changes. These might include systems that productivity in many resource-poor areas than previ- track biodiversity, beneficial species, and soil erosion. ously thought (Scott and Scherr 1995), and this is sup- Such monitoring systems need to be maintained by ported by recent research in India (Fan and Hazell the community itself, perhaps by a farm cooperative 1996). Moreover, the worsening poverty and resource or the village school. degradation that are occurring in many resource-poor areas as populations reach critical levels suggest that Institutional Issues and Options the social and environmental benefits from investing in these areas are now potentially attractive. In this section, we discuss select institutional issues Unfortunately, investments in resource-poor areas that are addressed much more extensively in later are complicated and risky. The diversity of agrocli- chapters. matic, social, and economic conditions that exist demands considerable site specificity in approach. Poverty Reduction through Investments Moreover, genetic improvement of individual crops is in Resource-Poor Areas not likely to be sufficient, and productivity growth must typically be achieved through improved man- Poor people are more desperate and more likely to agement of landscapes and farming systems with lim- trade off tomorrow's production in order to eat today ited reliance on purchased inputs. Successful than people who are not poor. Chronic poverty, exac- approaches, as exemplified by the efforts of several erbated by population growth and occasional NGOS, are based on participatory approaches with droughts, underlies much of the resource degradation farmers and local organizations, not on the more tra- observed in rainfed areas. It is much less of a factor in ditional top-down approach that characterized many intensive agricultural areas, although it can still be a integrated rural development projects of the past. problem where there is a high incidence of landless- Success is also difficult to measure in the short run, ness or a highly inequitable distribution of land. not only because of a multiplicity of economic, social, Alleviating poverty is fundamental to redressing and environmental goals, but also because resource- the environmental problems of most resource-poor poor areas are typically subject to considerable cli- areas. Appropriate strategies need to include targeted matic risk from year to year. assistance programs for the poor, economic diversifi- cation strategies for rural areas, assistance programs Property Rights for small farms, and public investments in rural infra- structure, health facilities, schools, and the like. The property rights that farmers have over natural Agricultural growth also needs to play a major role resources can be important in determining whether in the development of many resource-poor areas. they take a short- or long-term perspective in manag- Despite outmigration, the absolute number of people ing resources. For example, farmers who feel that living in these areas will continue to grow in the next their tenure is insecure, with or without formal rights, decades. Rural nonfarm economies are too small and are less likely to be interested in conserving resources too constrained by local demand to create significant or in making investments that improve the long-term employment opportunities in the absence of agricul- productivity of resources. tural growth. Property rights are often problematic during the There was some hesitancy in the past to invest in transition from extensive to intensive agricultural sys- resource-poor areas, primarily because the returns to tems, when they typically must evolve from indige- investment were perceived to be much higher in nous, community-based tenure systems to registered resource-rich areas. But these higher returns are less and legally recognized, private property arrange- assured today. Productivity growth in intensive agri- ments. Property rights are also becoming more prob- cultural systems is becoming more difficult and cost- lematical in the management of common property ly to achieve, particularly when the environmental resources, such as open rangeland or forest, because Peter Hazell and Ernst Lutz 17 the local institutions that traditionally control and reg- effective organizations are needed to manage them. ulate the use of such resources are breaking down. Often, governments have undermined indigenous Research has shown that secure property rights institutions by nationalizing important common prop- over land are an important factor in determining land- erty resources, such as forests and rangelands, and improving investments, such as tree planting, contin- then failed to manage them effectively. As a result, uous manuring, and terracing and contouring. these resources have degenerated into open-access However, this does not necessarily imply that govern- areas. The most successful institutions for managing ments should immediately invest in ambitious land common properties are local organizations dominated registration programs where this has not already been by the resource users themselves. It helps if local done. Many of the indigenous, community-based land organizations can be built on thne vestiges of indige- tenure systems that exist in long-settled areas can be nous rules and institutions and if the group as a whole surprisingly effective, and these systems are sponta- has secure rights to own the resource. Successful neously evolving, with the growth of populations and examples include social forestry groups in India and commercialization, toward systems of privatized land rangeland management groups in parts of West rights (Migot-Adholla and others 1991). These tradi- Africa. tional systems can also be effective in recognizing Property righiTs over irrigation water are in most multiple-user rights over the same land and are often need of reform in many countries. The ownership and more equitable than fully privatized systems. The allocation of water continue to be dominated by the appropriate role for governments in these cases is not public sector, and a pernicious combination of to replace the indigenous systems abruptly, but to seek bureaucratic ine:rtia, low water charges (including ways of strengthening them and facilitating their limited cost recovery), and subsidized electricity for adaptation to changing circumstances. Legal registra- pumping is responsible for considerable inefficiency tion of blocks of community or village-held land and in the use of water for agriculture as well as inappro- simple voluntary systems for recording land transac- priate water management practices that lead to pollut- tions may sometimes increase security by reducing ed water and waterlogged and salinized land. land disputes between and within communities. Evidence is growing that more appropriate pricing In contrast, land registration may be economically of water and electricity can significantly reduce water worthwhile in areas of high population density or and environmental problems in irrigated agriculture commercialized agriculture, particularly when formal as well as provide new sources of revenue for main- lending institutions are also well developed and land taining and improving irrigation infrastructure. is already effectively privatized. Feder and others Unfortunately, progress has been slow because of (1988) have found this to be true for Thailand. farmer resistance and vested political interests. In Traditional management systems for many com- most countries, cost recovery does not cover expendi- mon property resources face increasing challenges as tures for operations and maintenance. rural populations grow. Not only must more local More promising approaches lie in the devolution users be accommodated, but common property of many water allocation decisions to water user resources are increasingly threatened by encroach- groups and in the creation of tradable water rights for ment and unregulated use by outsiders. Full privatiza- farmers. Water rights are particularly attractive tion of common property resources is sometimes the because they lead to the full economic pricing of relevant solution, but often there are good economic water; water is valued at its full opportunity cost, not and social reasons for keeping some resources as just a charge, based on cost recovery, set by an irriga- common property. A good example is open rangeland tion authority. Moreover, creation of property rights in in drought-prone areas. The ability to move animals water bestows a new asset on farmers and is therefore over wide areas is an integral part of risk manage- more likely to be accepted than a new water charge or ment, especially in drought years. The costs of fenc- tax. ing and the existence of limited water holes also con- strain privatization. Some resources are better man- Cost of Collective Action aged as common property because of economies of scale or because this is an effective way of managing Conserving or improving natural resources often multiple-user rights. requires collec tive action by groups of users. Where resources are to remain common property, Examples include the management of common prop- 18 Integrating Environmental and Sustainability Concerns into Rural Development Policies erty resources or the organization of adjacent farmers Technological Issues and Options to invest labor in land terracing or bunding. Organizing farmers into effective and stable groups When poorly designed, or inappropriately used, tech- for collective action is difficult, and success is condi- nologies can lead farmers to increase production in tioned on a range of physical, social, and institutional ways that degrade natural resources. New technolo- factors (Uphoff 1986, Ostrom 1994, and Rasmussen gies have often been developed with a narrow focus and Meinzen-Dick 1995). on short-term profitability to farmers and without due A resource that is to be managed or improved col- consideration to their longer-term sustainability. For lectively should be reasonably small, naturally bound- example, the development of powerful pesticides and ed, and accessible to group members to facilitate the herbicides reduced costs and improved yields but control and exclusion of outsiders. It helps too if use often had negative effects on the environment and on by one member has limited effect on the availability long-term yields. A related problem is the spread of of the resource to other members (low "subtractabili- new technologies from the agroclimatic zones or ty"). Greater social cohesion within the group is facil- farming systems for which they were developed to itated if the number of users is small, the members are other, less-suitable areas where they may degrade homogeneous in their values and economic depen- resources. dence on the resource, and the net benefits from group Many national agricultural research and extension membership are substantial and distributed equitably. systems have yet to integrate environmental concerns Institutional design is also important. Ostrom successfully into their agenda. Too little consideration (1994) has identified five principles for designing is given to the sustainability features of recommended effective local organizations: (a) the members and the technologies, to broader aspects of natural resource boundaries of the resource to be managed or management, and to the technology and management improved must be clearly defined, (b) a clear set of problems of more fragile rainfed areas where resource rules and obligations should be established that are degradation is considerable. adapted to local conditions, (c) members should col- Public expenditure on agricultural research has lectively be able to modify those rules to changing declined in recent years, and there are serious ques- circumstances, (d) an adequate monitoring system tions about whether the national, regional, and inter- should be in place, with enforceable sanctions, prefer- national research systems that serve developing coun- ably graduated to match the seriousness and context tries have the capacity to rise to the productivity and of the offense, and effective mechanisms for resolving environmental challenges that lie ahead. During conflicts, and (e) the organization, if not empowered recent years, these systems have been reduced in size or recognized by government authorities, should at and effectiveness through declining financial support, least not be challenged or undermined by them. they have limited capacity to capitalize on modern An important line of research on the determinants biotechnology research, and many need major institu- of successful collective action highlights the changing tional reform if they are to respond to the concerns of dynamics of the net benefits. In areas with low popu- natural resource management and the needs of rainfed lation densities, resource shortages and degradation farming. are rare, and there is little incentive to organize for This decline has not been, and will not be, ade- investments in improving and conserving resources. quately offset by private sector investment. The pri- But in areas with growing populations, resources vate sector makes money from focusing on the prob- become scarcer and, in the absence of investments, lems of commercial farmers and higher-value crops. It are subject to worsening degradation. Eventually, the has little or no incentive to work on the problems of economic and social benefits of organizing for invest- small-scale farmers growing basic food crops, largely ment exceed the costs, and user groups can be expect- for home consumption, or to address many of the ed to emerge spontaneously to make the necessary environmental problems in agriculture that have off- investments. Within the context of this "induced inno- site rather than on-site costs. Publicly funded national vation" theory, it is futile to try to organize farmers agricultural research systems must be reformed to prematurely. Moreover, the purpose of outside inter- address natural resource management problems more ventions should be to create enabling conditions for effectively within the context of agricultural intensifi- the emergence of spontaneous group action, not to cation. supplant it with externally planned organizations. Research at the International Food Policy Peter Hazell and Ernst Lutz 19 Research Institute and elsewhere shows that the eco- Evans. 1987. "Variability of Cereal Yields: nomic payoff from public investments in agricultural Sources of Change and Implications for Agricul- research is high. On economic grounds alone, a good tural Research and Policy." Food Policy 12(3, case can be made for investing more in agricultural August): 199-212. research. And this does not consider the social bene- Barghouti, Shawki, and Guy Le Moigne. 1991. fits that properly directed agricultural research can "Irrigation and the Environmental Challenge." generate by helping to reduce poverty and environ- Finance and Development 28(June):32-33. mental degradation. Bishop, Joshua, and Jennifer Allen. 1989. "The On- World Bank grant support for the Consultative Site Costs of Soil Erosion in Mali." Environment Group for International Agricultural Research grew Working Paper 21. Environment Department, gradually from $12 million in fiscal 1980 to $45 mil- World Bank, Washington, D.C. Processed. lion in fiscal 1997, with additional grant support pro- Cuesta, Mauricio, Gerald Carlson, and Ernst Lutz. vided during the fiscal 1994-95 period. Bank funding 1997. "An Empirical Assessment of Farmers' of national research systems during the same period Discount Rates in Costa Rica." Work in Progress fluctuated within a range of between $130 million and Paper. Environment Department, World Bank, $390 million, with no discernible trend. Washington, D.C. Processed. de Moor, A., and P. Calamai. 1997. "Subsidizing Conclusions Unsustainable Development: Undermining the Earth with Public Funds." Earth Council, San Jose, Past patterns of agricultural growth have sometimes Costa Rica. Processed. been associated with negative environmental effects Dixon, John, Richard A. Carpenter, Louise Fallon, and with inequities that have pushed small farmers Paul B. Sheiman, and S. Manipomoke. 1992. Eco- and landless persons to the margin of society. But this nomic Analysis of the Environmental Impacts of is not an inevitable outcome of agricultural growth. Development Projects. London: Earthscan. Rather, we believe, it reflects inappropriate economic Fan, Shenggen, and Peter B. R. Hazell. 1996. "Should incentives for managing modern inputs in intensive the Indian Government Invest More in Less- farming systems, insufficient investment in many Favored Areas?" Environment and Production resource-poor areas and in social and poverty con- Technology Division Discussion Paper 33. cerns, and political economy systems that allow the International Food Policy Research Institute, rich and powerful to extract rents by legal or illegal Washington, D.C. Processed. means. The chapters in this volume address these pol- Feder, Gershon, T. Unchain, Y. Chalamwong, and C. icy issues, institutional approaches, and technical Hongladaron. 1988. Land Policies and Farm issues. The challenges are great, and the constraints Productivity in Thailand. Baltimore, Md.: Johns are significant, but much can be done with good gov- Hopkins University Press. ernance and a real commitment to environmentally Hamilton, Kirk, and Ernst Lutz. 1996. "Green and socially sustainable rural development. National Accounts: Policy Uses and Empirical Experience." Environment Department Paper 39. Note Environment Department, World Bank, Washington, D.C. July. Processed. 1. The main problem is not with transformations as Hicks, John. 1946. Value and Capital, 2d ed. Oxford: such. It is that assets are sometimes drawn down or Oxford University Press. liquidated and the revenues consumed rather than Jeyaratnam, J., K. C. Lun, and W. 0. Phoon. 1987. invested in other productive assets. "Survey of Acute Pesticide Poisoning among Agricultural Workers in Four Asian Countries." Bulletin of the World Health Organization References 65(4):521-27. Lutz, Ernst, ed. 1993. Toward Improved Accounting The word "processed" describes informally repro- for the Environment. An UNSTAT-World Bank duced works that may not be commonly available Symposium. Washington, D.C.: World Bank. through libraries. Lutz, Ernst, and Mohan Munasinghe. 1994. Anderson, Jock R., Peter B. R. Hazell, and L. T. "Integration of Environmental Concerns into 20 Integrating Environmental and Sustainability Concerns into Rural Development Policies Economic Analyses of Projects and Policies in an Agriculture, and the Environment Discussion Operational Context." Ecological Economics Paper 1. Washington, D.C.: International Food 10:37-46. Policy Research Institute. Lutz, Ernst, and M. Young. 1992. "Integration of Rasmussen, L. N., and Ruth Meinzen-Dick. 1995. Environmental Concerns into Agricultural Policies Local Organizations for Natural Resource of Industrial and Developing Countries." World Management: Lessons from Theoretical and Development 20(2):241-53. Empirical Literature. Environment and Production Migot-Adholla, Frank, E. Shem, Peter B. R. Hazell, Technology Division Discussion Paper 11. Benoit Blarel, and Frank Place. 1991. "Indigenous Washington, D.C.: International Food Policy Land Rights Systems in Sub-Saharan Africa: A Research Institute. Constraint on Productivity?" The World Bank Repetto, Robert, William Magrath, Michael Wells, Economic Review 5(1):155-75. Christine Beer, and Fabrizio Rossini. 1989. Milazzo, M. J. ]L997. "Reexamining Subsidies in Working Assets: Natural Resources in the National World Fisheries." Office of Sustainable Fisheries, Income Accounts. Washington, D.C.: World U.S. Department of Commerce, Washington, D.C. Resources Institute. Processed. Ruchijat, E., and T. Sukmaraganda. 1992. "Rational Ostrom, Elinor. 1994. Neither Market nor State: Integrated Pest Management in Indonesia: Its Governance ojf Common-Pool Resources in the Successes and Challenges." In P. Ool, G. Lim, T. Twenty-First Century. Lecture Series 2. Ho, P. Manalo, and J. Waage, eds., Integrated Pest Washington, D.C.: International Food Policy Management in the Asia-Pacific Region. Oxford: Research Institute. C.A.B International. Pearce, David, Anil Markandya, and Edward Barbier. Scherr, Sara, and Peter B. R. Hazell. 1994. Sustainable 1989. Blueprint for a Green Economy. London: Agricultural Development Strategies in Fragile Earthscan Publications. Lands. Environment and Production Technology Pezzey, John. 1992. "Sustainable Development Division Discussion Paper 1. Washington, D.C.: Concepts: An Economic Analysis." Environment International Food Policy Research Institute. Paper 2. Environment Department, World Bank, Serageldin, Ismail. 1996. Sustainability and the Washington, D.C. Processed. Wealth of Nations: First Steps in an Ongoing Pingali, Prabhu L. 1992. "Diversifying Asian Rice- Journey. Environmentally Sustainable Development Farning Systeims: A Deterministic Paradigm." In Studies and Monographs Series 5. Washington, Shawki Barghouti, L. Garbux, and Dina Umali, D.C.: World Bank. eds., Trends in Agricultural Diversification: Templeton, Scott R., and Sara Scherr. 1995. Regional Perspectives, pp. 107-26. World Bank Population Pressure and the Microeconomy of Technical Paper 180. Washington, D.C.: World Land Management in Hills and Mountains of Bank. Developing Countries. Environment and Produc- Pingali, Prabhu L., and P. A. Roger, eds. 1995. Impact tion Technology Division Discussion Paper 26. of Pesticides on Farmer Health and the Rice Washington, D.C.: International Food Policy Environment. Norwell, Mass.: Kluwer Academic Research Institute. Publishers; Los Bafios, Philippines: International Turner, Kerry. 1993. Sustainable Environmental Rice Research Institute. Economics and Management: Principles and Pingali, Prabhu L., and M. W. Rosegrant. 1994. Practice. London: Belhaven Press. Confronting the Environmental Consequences of Umali, Dina L. 1993. Irrigation-Induced Salinity: A the Green Revolution in Asia. Environment and Growing Problem for Development and the Production Technology Division Discussion Paper Environment. World Bank Technical Paper 215. 2. Washington, D.C.: International Food Policy Washington, D.C.: World Bank. Research Institute. Uphoff, Norman. 1986. Local Institutional Pinstrup-Andersen, Per, and R. Pandya-Lorch. 1994. Development: An Analytical Sourcebook with Alleviating Poverty, Intensifying Agriculture, and Cases. West Hartford, Conn.: Kumarian Press. Effectively Managing Natural Resources. Food, Warford, Jeremy, Mohan Munasinghe, and Wilfrido Peter Hazell and Ernst Lutz 21 Cruz. 1997. The Greening of Economic Policy Development. Washington, D.C. Reforn. 2 vols. Washington, D.C.: World Bank. World Resources Institute. 1991. Accounts Overdue: World Bank. 1997. Expanding the Measure of Wealth: Natural Resource Depreciation in Costa Rica. Indicators of Environmentally Sustainable Washington, D.C. 3 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia John Heath and Hans P Binswanger Observers are often struck by the joint occurrence of Colombia has been driven in highly inefficient direc- growing rural populations, rural poverty, and degra- tions by a variety of agricultural, land, and rural dation of the natural resource base used by the poor. finance policies and programs. These have premature- As a result, population growth and poverty are often ly and dramatically reduced employment opportuni- seen as the causes of natural resource degradation. ties in the sector and, as a result, have concentrated Reducing the rates of population growth and migra- poverty in rural areas and increased resource degrada- tion to urban areas is seen as a possible solution, along tion on hillsides and on the Amazon frontier. with assisting farmers with soil conservation. Binswanger (1989) has described similar phenomena Of course, reducing population growth and assist- in the Brazilian Amazon. We show that labor policies ing poor farmers with resource management problems have not contributed in major ways to these adverse are urgently needed in many countries. In this chapter, trends but instead that the misallocation of land and however, we argue that many other policy options labor and an exceptionally high female unemploy- exist. They consist of the removal of adverse policies ment rate in rural Colombia are consequences of the that have as their joint consequence a reduction in same policy factors. We then turn to various policy economic efficiency, an increase in poverty, and the options for correcting the misallocation of resources, degradation of natural resources. Their removal thus reducing poverty, and relieving the pressure of unsus- would be a win-win-win situation in which the same tainable farming on hills and in tropical forest areas policy changes could lead to more growth, less pover- with marginal land resources. ty, and more sustainable natural resource management. We start with a discussion of the Boserup hypoth- The Boserup Effects esis, which suggests that higher population and mar- ket access lead to improvements in-not deterioration Judgments about the sustainability of farming often of-natural resources. Using recent studies of the give too much weight to assumptions about the "car- Machakos district in Kenya and soil degradation in rying capacity" of the resource base (see Cleaver and Ethiopia, we illustrate the dependence of the Boserup Schreiber 1994). In fact, there is no clear evidence effects on the policy environment that governs the about what constitutes a threshold level of population investment incentives of farmers. pressure. Nevertheless, it is indisputable that popula- With this background, we examine growth, the use tion growth influences farning techniques, invest- of land and labor, and the degradation of marginal nat- ment in land, and land use on the toposequence. ural resources in Colombian agriculture. We demon- As a statement about the nature of the population strate with data that the use of land and labor in dynamic in agricultural systems, Boserup's hypothe- 22 John Heath and Hans P Binswanger 23 sis is compelling: as land becomes more scarce in ies documenting the Boserup sequences. Machakos is relation to labor, and access to markets improves, a semiarid district in Kenya with poor to middling agriculture is intensified, with the net result being agroclimatic conditions. In Kenya, the best land was higher agricultural production per unit of area reserved for white settlers. Colonial policy thereby (Boserup 1965). Rather than deteriorating, the base of obliged the native population to derive their food sup- land resources improves in the process. ply from a greatly reduced land base. The natives The Boserup effects of population growth and were forbidden to grow the most remunerative cash improved market access lead to: crops and therefore lacked incentives from the mar- * The intensification of land use ket. In the 1930s, the region was characterized by * A shift from hand hoes to plows heavy soil erosion and declining yields. In contrast, * The increasing use of organic and inorganic fertil- today Machakos supports a population almost six izer times as large ats it was in 1932, and agricultural out- * The shift to integrated crop-livestock systems put per unit of area (in constant maize units) has * Investment in land and irrigation facilities increased almost tenfold. Crop yields have risen. Cash * An increase in the use of agricultural labor crops and horticultural products have been success- * Higher agricultural production per unit of area. fully introduced. There are more trees, more soil con- This hypothesis is consistent with much research servation works, and greater use of organic manure on the development of farming systems and has now than in the 1930s. received empirical support from an Africa-wide study, The investment incentives needed to bring about a study of erosion in the rangelands of Botswana and these effects are associated with the following condi- Tanzania and of smallholder settlement areas of tions, which have been assured since independence: Zimbabwe, and an in-depth longitudinal study of the * An agricultural policy that, compared with that of Machakos district of Kenya (Ruthenberg 1980, other African countries, taxes the sector lightly Pingali, Bigot, and Binswanger 1987, and Tiffen, * Access to international markets for coffee and Mortimore, and Gichuki 1994). domestic markets for other cash crops Much of the literature on the Boserup effects treats * Construction of infrastructure associated with these beneficial relationships as if they were mechan- rural development projects ical, brought about by compelling physical and bio- * Access to nonfarm and urban employment oppor- chemical relationships and by individual utility- tunities in Nairobi maximizing behavior within the constraints set by the * Ability to finance investments from sales revenues natural world. We argue that the Boserup effects are and labor incomes far from automatic. We view them as the outcome of * Security of tenure, provided initially by the tradi- investment decisions made by farmers. In order to tional communal tenure and later via land titles come about, the investments require a positive incen- * New food production technology, especially for tive regime, access to soil and water resources, and maize access to markets. If adverse policy regimes deny * Locally adapted soil conservation technology and these conditions, impoverished peasants are forced to farmer-led irnitiatives to implement it. mine, rather than augment, the land resources. This is Recent cross-sectional studies of Ethiopia, in con- especially damaging if they only have access to mar- trast, find that areas where population density signifi- ginal land or to humid tropical forest frontiers. cantly exceeds carrying capacity have high indexes of Various authors have analyzed soil conservation as soil degradation (Grepperud 1994).1 Longitudinal an investment strategy, with consideration of the eco- studies similar to those undertaken in Machakos nomic and policy factors underlying erosion. would be required to prove that these areas have not Anderson and Thampapillai (1990) and Barbier and experienced the beneficial sequence of Boserup Bishop (1995) review these factors. In this chapter, effects. Nevertheless, a comparison of policies and the discussion of policy effects is consistent with this programs prevailing in the two countries over the past tradition. 30 to 40 years is instructive. Before turning to Colombia, we use an African Ethiopian farmers were heavily taxed throughout comparison of the workings of the Boserup effects. the period via a great diversity of methods. The con- The Machakos study of Tiffen, Mortimore, and struction of rural infrastructure was limited. Access to Gichuki (1994) is the most recent of a wealth of stud- international rnarkets and even to domestic markets 24 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia was often disrupted. Employment opportunities in the access to land elsewhere, indigenous farmers are carv- rural nonfarm and urban economies were extremely ing up their land into microfundias. In the Amazon limited by the lack of agricultural and economywide and Orinoco basins, and on the Pacific coast, pressure growth. Periodic famines depleted assets and further is rising to put unstable lands of intrinsically limited undermined the peasant's ability to mobilize invest- fertility under annual crops. ment resources. According to Currie (1965), 40 years ago the prob- There was no security of tenure during the period. lem was caused by too many poor, inefficient farmers Under Haile Selassie, farmers were tenants at will working on the slopes. The solution-both to rural rather than holders of secure ownership or usufruct poverty and to resource degradation-was to encour- rights. Under the communist regime, the state owned age migration of the surplus rural population to the the land, usufruct rights were never secure, and peas- towns, leaving the land to be worked by fewer, more ants were dislocated as a result of villagization pro- technically sophisticated farmers. More than 40 years grams and forced migration and were pressured to later, the population of poor people in rural Colombia join collectives. Irnproved technology in food grains has more than doubled, despite massive rural-urban was scarce. Very few programs were aimed at devel- migration (World Bank 1995a). Three out of four poor oping and disseminating improved soil conservation people in Colombia now reside in rural areas. Poor techniques among peasant farmers. farmers are still working the slopes. The sharp contrast in policy regimes and natural Recently the natural rate of increase of the rural resource outcomes between Kenya and Ethiopia population has slowed substantially, helping to raise shows that Boserup effects are not an automatic wages and reduce the absolute level of pressure on the response to population growth or market access. resource base. Since .1985 the rural population has Instead they require a policy and institutional regime begun to decline for the first time in absolute terms. that provides peasants and commercial farmers with But the overall pattern of growth has failed to absorb favorable incentives as well as opportunities to earn what Currie perceived as "surplus rural labor." Poor income on and off the farm to enable farmers to invest farmers continue to have limited access to good land in production inputs, technology, and conservation and therefore continue to exert pressure on the slopes activities. With this background, we now turn to (and on equally fragile land in the Amazon-Orinoco Colombia, our main example of the impact of policy basin). on land degradation and poverty. Are there other solutions? Improving the tech- niques that hillside farmers use on marginal land may The Case of Colombia be part of the story. However, although low-cost tech- nologies exist, little progress has been made in diffus- In Colombia, unsustainable farming of the Andean ing them, largely because the incentives are lacking: slopes has long been recognized as a problem.2 In on marginal lands small farmers may have insufficient 1950 Lauchlin Currie led a World Bank mission to incentives to invest in soil conservation (Ashby Colombia. This mi.ssion "noted for the first time on a 1985). Even if better, more cost-effective techniques national scale the extent to which flat, apparently rich, for hillside farming are developed, they will be a pal- bottomlands were occupied by low-intensity livestock liative at best. To devise an appropriate solution, it is ranching estates, while slopes steep enough to make necessary to look beyond the hillsides. The challenge cultivation a hazard to life were occupied for crop is to improve the rural poor's access to less fragile, farming" (Blakemore and Smith 1971: 232; see also more fertile lands at lower elevations and to extend Currie 1965). The situation has not improved since the opportunities for off-farm rural employment. We then. In many areas, the Andean slopes are being show here that it is necessary to change the agricul- denuded of vegetative and soil cover, and the result- tural policy framework that influences the operation ing loss of moisture retention is having an adverse of land and labor markets. effect on the flow of streams and reducing the avail- Currie's prescription of accelerating rural-urban ability of water for agriculture, both for poor farmers migration and turning farming over to large-scale, on the slopes and for richer farmers located in the val- modernized commercial farmers has now been pur- ley bottoms. In the scattered indigenous reserves sued in Colombia for more than 40 years. It has not (what remains from the colonial resguardos), there is reduced rural poverty significantly but rather has fur- an acute problem of fragmented holdings: deprived of ther concentrated the poor in the countryside. It has John Heath and Hans P Binswanger 25 also led to an extremely low use of both land and been extremely capital- and labor-intensive. Between labor in this country and to low overall productivity of 1950 and 1987, agriculture's annual growth rate aver- the agricultural sector compared with its enormous aged an impressive 3.5 percent. Capital inputs to agri- potential. The urban migration approach to solving culture grew at an average annual rate of 2.8 percent, rural poverty and resource degradation has been a land area devoted to agriculture and livestock grew complete failure. 1.4 percent, and employment grew only 0.6 percent (see table 3.1). The Low Propensity of Colombian Agriculture The sector's, relatively low propensity to absorb to Absorb Labor labor is reflected in the pattern of land use. Crop farm- ing has captured a relatively small proportion of the Agricultural development in Colombia has involved natural resource base: 16 percent of Colombia's land substantial misallocation of resources: land has been area is suitable for crops, but less than 4 percent is very unevenly exploited, and the growth of farm out- cultivated. Livestock rearing is overextended: 13 per- put has absorbed less labor than might have been cent of the territory is deemed appropriate for pasture, expected. Over the past 40 years or so, compared with but 35 percent is put to this use. Small farmers with other countries in the same range of per capita limited access to good-quality flatland may end up income, the growth of farm employment has been deforesting marginal land on the Andean slopes. Only exceptionally low in Colombia (Misi6n de Estudios about half of Colombia's agricultural and forestry del Sector Agropecuario 1990, which extends earlier lands is still forested, whereas two-thirds are only work by Syrquin and Chenery 1989). Between 1925 suitable for fo:restry and should be left under tree and mid-century, by international standards, the cover (IGAC 1988). decline in Colombian agriculture's share of the gross Agricultural policymaking in Colombia shows domestic product (GDP) and its share of the labor force substantial elements of "large-farmer bias." First, the was comparable to that of other countries in the same patterns of public investment and orientation of the range of per capita income. But from the early 1950s trade regime have combined to favor livestock and onward, this picture changed abruptly: the relative grain crops, neither of which uses labor intensively. importance of farm employment fell off at a much Second, credit policies have tended to discriminate faster rate than projected on the basis of international against small farmers. Only about one-third of small evidence concerning the correlation between employ- farmers are able to obtain a loan from the formal sec- ment and GDP shares (see figure 3.1). tor. There is no sign that past practices favoring large- In Colombia, the growth path of agriculture has scale farms in credit policies are being corrected.3 Figure 3.1. Share of Primary Sector in Total Employment in Colombia, 1925-86 Percent 75 _ 70 _ 65 AN__ 60 55 50 ' _ _4% 45 Expected 40 35 _ Observed 30 n in \t \t E ( E 00 00 Source: Misi6n de Estudios del Sector Agropecuario 1990. These findings extend earlier work by Syrquin and Chenery 1989. 26 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia Table 3.1. Sources of Farm Sector Growth in Colombia, 1950-87 Growth rates Contributions to growth Period Fann GDP Area Capital Employment Area Capital Employment Allfactors Productivity 1950-55 3.03 1.15 -0.37 2.09 0.22 -0.13 0.96 1.05 1.98 1955-60 4.08 0.68 1.26 0.21 0.13 0.44 0.10 0.67 3.41 1960-65 2.77 1.31 2.44 0.67 0.22 0.81 0.34 1.36 1.41 1965-70 4.94 2.17 4.68 1.29 0.41 1.87 0.53 2.81 2.13 1970-75 4.33 2.03 6.51 -3.91 0.38 2.60 -1.60 1.39 2.94 1975-80 4.58 1.67 4.96 2.75 0.28 2.18 1.05 3.51 1.07 1980-84 0.89 1.80 0.93 0.66 0.27 0.45 0.24 0.96 -0.07 1984-87 3.53 0.48 1.97 0.76 0.07 0.95 0.28 1.30 2.23 Average, 3.52 1.41 2.80 0.57 0.25 1.15 0.24 1.63 1.89 1950-87 Source: World Bank 1995b. Third, tax policies, have converted agriculture into a ventions) that the government conferred on a group of tax shelter for both income and capital gains taxation, nine farm commodities (World Bank, LACTAD 1994). providing incentives for holding land as a tax shelter Expansion of cattle rearing has been land-extensive, rather than for agricultural production. favoring the creation of large estates: the rate of growth By subsidizing capital inputs and the livestock sec- of the herd has only slightly exceeded the rate of tor, the Colombian government has sponsored a strat- growth of the area in pasture. The off-take rate remains egy of agricultural development that discriminates very low-equal to 50 or 60 percent of that achieved in against small farmers. Tractor subsidies have encour- Argentina and the United States. The extensive live- aged shedding of labor: already, by the early 1970s, stock sector occupies large amounts of arable land on one-quarter of the cultivated area was mechanized. "lower slopes," which could be developed for intensive Public irrigation schemes have tended to favor larger agriculture by the construction of road infrastructure. rather than small producers; and, because much of the Much of these lands will also need drainage. The large land in the irrigation districts is devoted to pasture and estates favored by policy simply do not have the labor grain crops-which use less labor than higher-margin needed to make these investments. crops-employment generation has been much less Narrow access to land and employment in rural significant than it might have been. areas has increased the propensity for violence. Cattle rearing spread rapidly in the immediate Although narrow access has not been the only cause postwar period. The area under pasture grew from of rural violence, it has contributed significantly to 12.1 million hectares in 1950 to 26.7 million hectares Colombia's extremely violent history over the past in 1986. (In the same period, the area under crops half century. The propensity for violence has a nega- increased from 2.6 to 4.3 hectares). The Caribbean tive feedback on investment and employment. Rural region accounted for 38 percent of the growth in pas- insecurity probably reduces the incentive to invest in tures, while the Andean valley bottoms accounted for agriculture and, more important, skews the pattern of a further 30 percent. There was also rapid growth of investment toward activities that are relatively non- the pastoral area in the eastern savanna (centered on intensive in the use of labor. The supervision costs Meta): by the mid.-1980s, no less than half of all cat- and incentive losses that are typically associated with tle pasture was located in the sparsely populated area hired labor are even greater in an environment where east of the Andean ranges (Misi6n de Estudios del hired workers may initiate, or collude in, violent Sector Agropecuario 1990). reprisals against employers. This disincentive may The livestock sector has been more protected than encourage landowners in areas of arable potential to the crop sector, helping to account for its rapid expan- invest in livestock rather than crops; and it may rein- sion. Between 1980 and 1992, beef and milk absorbed force the tendency for larger irrigated holdings to be 82 percent of total support (price plus nonprice inter- placed under pasture or grain crops rather than high- John HIeath and Hans P Binswanger 27 margin crops that require careful supervision and are land reform agency, INCORA, have done little to intensive in the use of hired labor. change the overall shape of the agrarian structure. Since its inception in 1961, INCORA has redistributed The Negative Impact of Past Agricultural relatively little ]iand in areas of established settlement, and Land Policies concentrating its efforts on the colonization and titling of frontier lands. Because land on the frontier has A number of policy interventions have circumscribed tended not to be handed out in small parcels, colo- the poor's access to land. First, the scope for tenancy nization initiatives on this agricultural margin have has been reduced. Beginning with Law 200 of 1936, failed to alter the overall picture of concentration. and culminating with the Ley de Aparceria of 1975, a Between 1960 and 1988, the area occupied by hold- series of legal measures have had the effect-inten- ings under 5 hectares declined from 6 to 5 percent, the tional or otherwise-of reducing the incentive for area in mediurn-size farms (5-50 hectares) rose from large landowners to lease out land to tenants (see table 24 to 26 percent, and the area in larger farms (more 3.2). The right of landowners to employ sharecrop- than 50 hectares) fell from 70 to 69 percent. pers, other tenants, and colonos was formally out- The lack of tenancy and land reform options lawed by Law la of 1968. This led to a precipitous encouraged land invasion. During the 1970s, there fall in the use of sharecroppers and colonos, although was a wave of illegal farmn occupations, affecting they had not completely disappeared even by 1988. A 1,500-2,000 farms and roughly two-thirds of depart- less dramatic reduction occurred in the number of ments. From this point on, INCORA'S work tended to other tenants. These trends occurred even while agri- center on regularizing the claims of illegal invaders. cultural area expanded in the country as a whole. But even this avenue for land acquisition was closed: Much of the decline in sharecropping centered on the Law 30 of 1988 banned INCORA from acquiring ille- coffee-growing region (associated with an increase in gally occupied land. the importance of small owner-operated farms) and The literature on farm size and productivity has on areas of the Caribbean lowlands, following clear- established thait factor productivity on farms operated ance of the land for cattle rearing. primarily with family labor is typically larger than The steepest fall in rural employment-3.9 percent that of larger farms operated primarily with hired a year-occurred between 1970 and 1975. This sug- labor or tenants. Around 1970, the situation in gests that the 1968 law was the critical event and that, Colombia was consistent with this generalization (see by the time the ban on sharecropping was reaffirned Berry and Cline 1979). Moreover, for any given level in 1975, most of the tenants and squatters had already of resource use and output, the small family-based been expelled. sector generates much more employment than the The negative impact on employment that resulted large-scale sector. When there is technical change, the from closing the tenancy option could potentially optimal size of family-based operations tends to have been offset by redistributing land toward small- increase. But once the size of farms has adjusted to er farmers. But colonization and redistribution by the the new technologies, the typical negative relation- Table 3.2. Changes in Land Tenure in Colombia, 1960-88 (thousands of hectares unless otherwise noted) Growth Tenure 1960 1988 (percent) Owners 18,995 29,117 53.3 Sharecroppers 1,100 273 -75.2 Other tenants 1,231 829 -32.7 Colonosa 2,889 554 -80.8 Otherb 526 1,123 113.5 a. Occupants of untitled land (internal and external frontier). b. Includes squatters. Source: World Bank 1995b. 28 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia ship reappears.4 municipios located in distinct agricultural and agrari- Between 1976 and 1988, the yields of small farm- an systems in Colombia (Suarez and others 1993). ers increased on average by 82 percent, compared The study demonstrates that the market for land is with an increase of only 2 percent for medium and highly segmented, failing to transfer land from large larger farmers. Much of this small-farmer yield to small farmers. There is a very active land market response was based on the adoption of scale-neutral among large landowners, more active indeed than in input packages (primarily involving agrochemicals) the United States or other countries.5 There is also successfully promoted by the extension component of an active market for the sale of small parcels the integrated rural development program that was between smallholders. But there are few transactions launched in 1976. The yield response is striking given between the two groups. However, the evidence pre- that the volume of public resources channeled to the sented in this study may not be statistically represen- small-farm sector was small in relation to the volume tative of the different regions covered. Thus, there are garnered by larger farms. The level and speed of the good grounds for making a closer, more precise eval- response by small farmers suggest that the relation- uation of this phenomenon of segmentation. ship between farm size and productivity observed for One serious flaw of past approaches to land reform the 1960s and early 1970s continues to hold. in Colombia and elsewhere was failure to recognize This raises an important question. If small family that the market segmentation and inflated prices of fanns are more efficient than large farms relying on land are themselves policy consequences. They are hired labor, why do large farmers not find it more affected by macroeconomic conditions, nonfarm profitable to subdivide their properties and rent or sell investment options, tax policies, and agricultural poli- these parcels to smaller farmers? The restriction on cies (for a consideration of these issues in another leasing has already been examined. Large farms in context, see Binswanger 1989). Correcting the mal- Colombia generally are not parceled out and sold to distribution of land in Colombia and elsewhere must small farmers because the market price of agricultur- start by eliminating these distortions. Doing so would al land typically exceeds the capitalized value of farm enable land markets to work better and would make profits. This occurs because the value of farmland is any land reform program more affordable to govem- only partly based on its agricultural potential. In all ments and beneficiaries. In Colombia, for example, areas, land serves as a hedge against inflation. Its improving the collection of agricultural income taxes immobility makes it a preferred form of collateral in and capital gains taxes on farmland would reduce the credit markets, conferring additional utility, particu- attractiveness of land as a tax shelter, therefore tend- larly where production risk cannot be insured. In peri- ing to lower its price. urban areas, land holds out the prospect of higher A further problem is that past agricultural credit returns from real estate development than from farm- policies and recent credit initiatives are not conducive ing. Finally, credit subsidies and tax write-offs are either to holding down the price of land or to bringing likely to be capitalized into land values (Binswanger, more land onto the market. The agricultural invest- Deininger, and Feder 1995). Many of these nonfarm- ment incentives recently adopted (Law 101 of 1993) ing benefits do not apply to small farmers, who rarely will mainly be captured by large farmers, giving them have access to subsidized credit and do not benefit a competitive edge in relation to small farmers. Past from the value of landownership as a tax shelter. instances of blanket rescheduling of credit have If the market price of land exceeds the capitalized helped to keep alive inefficient large farms that could value of farm profits, a poor smallholder or landless fruitfully be parceled out and sold off to small farmer worker will not be able to finance the purchase of land groups. A better approach would be to provide finan- out of farm profits, even if the owner, or a mortgage cial assistance for restructuring the assets and liabili- bank, is willing to advance him or her a loan covering ties of potentially viable large farms, coupled with the full purchase price of the land. This means that financial incentives for the owners of nonviable farms purchase will only be feasible if the productivity dif- to leave the agricultural sector, making their land ferential between small and large farms is huge, if available for purchase by small farmers. recourse is made to nonfarm income, or if the pur- An important factor tending to push up the price of chaser is willing to exploit unpaid family labor, devot- land, one that is specific to Colombia, is the launder- ing the imputed labor earnings to the purchase of land. ing of money from drug trafficking. It is estimated This hypothesis is borne out by a study of 15 that, in the late 1980s, traffickers were investing John Heath and Hans P Binswanger 29 between 8 and 23 percent of their cocaine revenues in find work. In towns, the gender disparity was much the purchase of land and had accumulated nearly 1 narrower: 6 percent of men and 9 percent of women million hectares (equivalent to 3 percent of the farmed were unemployed. The rural male-female unemploy- area). Most of this land was located in the depart- ment differential is also extremely high by interna- ments of C6rdoba, northern Antioquia, Caquetd, tional standards. Meta, Sucre, Atlantico, and Casanare and was primar- There is a long tradition of women working as ily used for cattle rearing (Bejerano 1988). One way day laborers (in the coffee harvest). Also, between to tackle this problem would be to place the onus on 1973 and 1985, the share of women in the economi- would-be buyers of large tracts to demonstrate that cally active population of rural areas rose from 14 to the cash used to buy land was obtained by licit means; 32 percent (Berry 1992). However, females are less failure to provide satisfactory proof would result in an able to respond with alacrity to demand for short- embargo on the sale of land. term workers, because they often must care for young children, which tends to reduce their mobili- Labor Policies ty. Although women have less propensity than men to remain in rural areas, among those who do remain To what extent have inflexible and distorted labor the rate of unemployment is much higher than it is markets contributed to the misallocation of land and for men. labor in Colombia and therefore aggravated the dete- The particularly high unemployment rate for iural rioration of resources on slopes and at the frontiers of women (but not for men) is not a consequence of the the rain forest? Labor markets may be said to be effi- low labor absorption of Colombian agriculture, which cient when there are signs of integration between affects both men and women. It is partly a conse- urban and rural sectors and between different rural quence of social norms that discourage participation localities. Integration suggests a trend toward lower- in off-farm work, except in specific operations such as ing the wage differentials between localities and sec- the coffee harvest. However, the unemployment of tors for workers with comparable skills. women may be aggravated by inadequate access to The urban-rural wage differential in Colombia is land. This hits relatively immobile women more low: wages in construction are only about 10 percent severely than rmore mobile males. The usual response higher than those in agriculture. The recent social to slack labor (lemand is to increase the use of unpaid security reform will probably tend to push up the cost family labor on the farm. For families with insuffi- of hiring workers in the urban formal sector. This will cient farmland, either rented or owned, this option reduce the take-up from the informal sector, a trend does not exist. that may induce a greater slackness in the urban labor This brief review suggests that rural labor markets market, tending to keep wages for unskilled work or labor market policies are not a major cause of the close to those prevailing in the rural sector. misallocation of land and labor in Colombia. Instead, Interregional rural to rural migration has also been the disquieting feature of extremely high female substantial, and there is little evidence of regional seg- unemployment may be a consequence of agricultural mentation of rural labor markets. In addition, a sub- and land market policies that cause low labor absorp- stantial amount of temporary interregional migration tion and inefficient land use. is associated with the contracting of casual labor for harvesting sugarcane, coffee, cotton, and other crops, Temporary Shocks in Rural Labor Markets but this involves men much more than women. As a consequence, male rural unemployment rates are very Although the rural labor market appears reasonably low, on the order of 2 percent. flexible, temporary shocks to the farm sector can still In contrast, rural female unemployment rates are have major adverse effects on employment. In very high. According to a 1993 household survey, in 1992-93, the combination of severe drought and a the countryside, 12 percent of women looking for sharp downturn in world prices seriously depressed work were unable to find it compared with only 2 per- the sector. The employment. impact was probably cent of men (results of the CASEN survey are reported greater than it would have been if the rural demo- in Reyes and Martinez 1994). Rural unemployment graphic transition had already been fully negotiated. was particularly acute among young women: 25 per- Although the rural population has declined in absolute cent of women under 25 years of age were unable to as well as relative terms since 1985, the population of 30 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia working age increased at an annual average of 2.9 labor in periods of a temporary downturn will do lit- percent between 1985 and 1990, much higher than in tle to cure the fundamental misallocation of land and previous periods. This reflects the relatively high rural labor in Colombian agriculture. Nor will interventions birth rates of the late 1960s. to improve the productivity of hillside agriculture of Between 1990 and 1993, the number employed in poor farmers be sufficient. More radical approaches agriculture fell from 2.3 million to 2.2 million. The are required. A strong case can be made for reorient- loss of agricultural employment (54,000 jobs lost) ing and revitalizing the Colombian land reform. Three was more than offset by the growth of employment in reasons are paramount. First, migration away from the the "rural nonfarm" sector (163,000 jobs created; see countryside has not-contrary to the initial expecta- Reyes and Martinez 1994). Within the farm sector, tions of Colombian policymakers-eliminated rural employment trends varied significantly by type of poverty, nor has it reversed the tendency of small crop. Between 1990 and 1993, nonperennial crops farmers to cultivate steep slopes in an unsustainable contracted 19 percent, most of this related to the manner. Small farmers need to get access to some of declining profitability of cotton, rice, and vegetables the more fertile bottomland currently used for exten- (and, to a lesser extent, maize, sorghum, soybeans, sive livestock grazing. Second, putting land into the and wheat). Coffee contracted 2.4 percent. Other hands of the rural poor will help to defuse one of the perennials grew 9 percent: falls in the employment causes of rural violence and may help to promote generated by cassava, panela, and cocoa were more investment in agriculture. Third, the evidence sug- than offset by increases in bananas, plantains, oil gests that small farmers in Colombia have high pro- palm, and sugarcane (this refers to changes in the ductivity if given access to land, markets, and number of man-days-jornales-worked in each crop agricultural services. They are capable of making sig- per year; see Suarez and others 1993). nificant productivity gains. In other words, putting The impact of contraction would have been felt more land in the hands of small farmers is likely to most strongly by households that depend significant- generate efficiency as well as equity gains, making ly on rural wage incomes. In 1993, 57 percent of rural Colombian agriculture more competitive. household heads had worked for a wage, and wage Land reform is a highly appropriate response to earnings made up 46 percent of rural household this problem of segmentation. An alternative would income. A larger proportion of the rural labor force be to create incentives for larger estates to hire in was self-employed (58 percent), but incomes from more labor-for example, by reducing capital invest- these activities accounted for only 37 percent of rural ment and livestock subsidies. But rural violence household income. Wage employment was a much would still reduce the propensity for large farmers to more significant source of income in the Caribbean hire labor, and, even in the absence of violence, the region than in other rural areas (Reyes and Martinez supervision costs associated with using hired labor 1994). would make these estates less efficient than small This brief overview suggests that labor market farms using family labor. Therefore, there are good rigidities are not a major problem in Colombia. grounds for facilitating the subdivision and sale of However, there is an urgent need to find adequate estates to small farmers. Leasing to small farmers ways of responding to sharp downturns in rural labor could also be encouraged, but incentive problems will demand occasioned by the periodic agricultural crises make these farms somewhat less efficient than small associated with import surges and price slumps. The owner-operated farms. second issue is how to ease female unemployment. Land reform will only work if complementary Policy interventions could usefully stress the creation measures are designed to reduce the "nonfarm" incen- of an enabling environment for rural nonfarm enter- tives for holding large estates-measures to reduce prises, including broader opportunities for female the wedge between the price of land and the capital- employment, and the promotion of public works pro- ized value of farm profits. grams to alleviate temporary downturns in the agri- In addition, INCORA, the parastatal land reform cultural demand for labor. agency, has been inefficient in carrying out much land reform. This ineffectiveness can be attributed to sev- The Case for Land Reform eral causes: * The major episode of redistributive land reform Government interventions to drive up the demand for occurred in the 1960s, initiated at the end of the John HYeath and Hans P. Binswanger 31 civil war known as La Violencia in 1958. After that direct negotiation between buyer and seller: INCORA initial spurt, budget allocations for the purpose of facilitates these negotiations but does not buy land, land reform dried up. except as a last resort. INCORA may compulsorily pur- • Land acquisition was based on expropriation of chase land (a) on behalf of indigenous or other desig- insufficiently used farms, a divisive process nated target groups, (b) to protect the environment, fraught with endless litigation. As a consequence, and (c) when beneficiaries are unable to reach agree- INCORA was a slow and poor buyer of land, acquir- ment with landowners who previously had agreed to ing mostly already invaded and marginal lands, sell to them. often at inflated prices. The grant for land purchase amounts to 70 percent * Selection of land reform beneficiaries, land alloca- of the cost of buying a family-farm holding; beneficia- tion, and farm development were carried out by a ries cover the remaining 30 percent of the purchase centralized bureaucracy in paternalistic ways. price through loans that INCORA helps them to negoti- Beneficiaries received incomplete rights to land ate. A substantial grant element seems justified in and poor services. order to offset the wedge between the market price of * INCORA continued to allocate frontier land to land and the capitalized value of farm profits. The size favored individuals in large units, rather than as of the wedge is a function of the severity of policy dis- family-farm units, thus reproducing the unfavor- tortions; complementary measures are needed to able agrarian structure at the frontier. Its settle- reduce these distortions. Under the new arrangement, ments of smallholders on the frontier suffered the loan compornent is not so large as to saddle benefi- from centralized and paternalistic bureaucratic ciaries with a high debt-asset ratio that would compro- approaches similar to those of its land reform pro- mise their ability to secure credit for working capital. gram. However, a 70 percent subsidy may be too large. Given this less-than-encouraging record, future First, it entails a major fiscal cost, which the state may land reform efforts clearly must be based on radically be poorly placed to bear. Second, it could be argued new approaches that are cheaper, faster, more decen- that if the policy distortions are removed, a smaller pur- tralized, and more participatory. chase subsidy would be needed. Also, the size of the The government of Colombia has recently taken subsidy may encourage persons to apply for it, even steps to revitalize land reform and to change how it is when they have no intention of continuing to fann. administered. It has introduced new legislation (Law In the past, thle agrarian reform ceded ownership of 160 of 1994) that proposes to use the land market for large tracts of public lands on the frontier, thus failing land reform.6 Beneficiaries buy land directly from to redress the concentration of rural property. Under large farmers, using grant and loan resources provid- the new law, land on the frontier is parceled out in ed by the government. Land allocation on the frontier family-farm units. Persons whose net worth exceeds is restricted to family-farm plots for poorer groups. the Col$1,000 minimum wage will not qualify for The legislation also decentralizes land reform admin- these homesteads; neither will agrarian reform offi- istration and makes it much more participatory. Thus cials. These provisions represent a significant break the legislation has many positive features. However, with the past, holding out the prospect of a more equi- little consideration has so far been given to the policy table distribution of land on the frontier. issues that affect the efficiency with which land mar- There are a number of weaknesses in the new law. kets work. Next, we briefly discuss the strengths and First, holdings that are sold off under the auspices of shortcomings of the new legislation. the land reform may not be sold or rented for 12 years after the transfer: this moratorium seems excessive. Evaluating Law 160 The stipulation is unwise because it is unlikely to be enforceable. To the extent that it is enforced, it will The new law provides for the poor to be given grants needlessly restrict the flexibility of land and labor to buy land. Persons deemed eligible for grants are markets. Some of the newly created farm enterprises allowed to identify land that they want to buy and to will inevitably fail, and farmers who fail must be negotiate a price with the owner. This marks a clear allowed to exit. break from the previous arrangement whereby INCO- Second, it is not clear that government needs to RA bought land and transferred it to beneficiaries. The make the award of grants conditional on purchase of centerpiece of the new initiative is the provision for a "family-farm-size" holding or to restrict the scope 32 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia for subsequent subdivision of such holdings. The pro- mortgage, and land sales contracts. posal denies the option to the poor of buying a "sub- Although the law promises a new wave of land family farm" and supplementing farm revenue with reform, significant policy problems in the law itself, income from off-farm sources. Arguably, there is no and in other agricultural policies that favor large- need to fix a minimum acceptable holding size. INCO- scale farms, will have to be addressed. Only then will RA could simply pay 70 percent of the cost of land land reform's potential to reduce both poverty and land purchase up to a maximum amount rather than fix a degradation be fully realized. target farm size. This would incidentally remove the need for time-consuming and costly surveys to deter- Conclusions mine, for each locality, what area of land is needed to support a farm family. This article has argued that rural poverty, inefficient Third, the present law reopens the possibility of resource allocation, and natural resource degradation regularizing the tenure status of persons who have are joint phenomena, often induced by a common invaded land. This is valid: past invasions need to be nexus of policy failures that favor the modernization resolved. But, in regulating the law, it will be impor- of large-scale farming at the expense of more efficient tant to create incentives to acquire land through the and employment-intensive family farms. The policy market rather than invasion, to ensure that title is nexus has as an additional adverse consequence the granted more quickly if land is not invaded, and to use concentration of impoverished populations with few the provision of infrastructure as leverage-concen- investment resources on marginal lands, at tropical trating investment on areas where land has been trans- forest frontiers, and on erodible hillsides. Eliminating ferred through market negotiation, not invasion. the adverse policy nexus should be given higher pri- Fourth, private property rights are deemed to lapse ority than reducing population growth, relocating the if land is abandoned, unsustainably exploited, or used impoverished population to the cities, or improving for illicit purposes (cultivation of narcotics). The new farming techniques on marginal lands. law indicates, for example, that "mere tree felling" is Of course, eliminating the privileges of rural elites not a sufficient indicator of "adequate exploitation." who are embedded in these policies is an extremely However, there is much room for discretionary inter- difficult task. These policies did not come about by pretation of what constitutes proper use of the land. accident, but as the consequence of a historical evolu- The best way to get around this problem is to remove tion that involved bargaining among politically strong the policy distortions that encourage people to hang groups who protected their interests at each point in on to large tracts of land: changing the incentive struc- time. Analyzing the joint consequences of the poli- ture will work better than trying to regulate land use. cies may help in the reform effort but is unlikely to Fifth, the new law makes no attempt to alleviate be sufficient. Peasants have been poorly articulated the severe restrictions on leasing. These restrictions in the Colombian political process, and steps must entail substantial equity and efficiency losses and be taken to provide them with a greater input into have resulted in so much premature eviction of labor policymaking. from the sector. Opposition to leasing is based on the There are very few checks and balances when it perception that, in the past, such arrangements (par- comes to agricultural policymaking. The power of ticularly sharecropping) were exploitative. Evidence large farmers-articulated through the commodity from around the world suggests that sharecropping is associations (gremios) and through congress-is in no often a way for differently endowed enterprises to way offset by countervailing peasant or consumer pool their resources to mutual benefit, overcoming lobbies. It is particularly striking that government credit constraints and helping to manage risk. gives peasant organizations very little scope to inter- Safeguards could be adopted to protect small farmers vene in the policy dialogue. The Ministry of from the possibility of exploitation. Minimum stan- Agriculture is dominated by the gremios and by con- dards for tenancy contracts may be appropriate. These gress and spends much of its time responding to contracts could be enforced by municipal arbitration demands from commercial farmers-primarily, pro- councils whose judgment would, in the last resort, ducers of importable goods-for greater protection allow for appeal th[rough the courts. Arbitration coun- and subsidies. Policymaking, therefore, tends to focus cils would provide statutory representation to all par- on short-term issues of crisis management; there is lit- ties involved in disputes concerning labor, tenancy, tle pressure on the government to address issues of John Heath and Hans P Binswanger 33 long-term strategy or poverty reduction. Unless some will reap the lion's share of the subsidy; the law political actors see that it is in their interest to bring implicitly discriminates against small farmers. poor peasants more strongly into the policy process, Chronic farmer indebtedness is a major problem, one land and policy reforms will continue to be a slow that was exacerbated by the drought and downturn in process. world prices of 1992. There has been very little suc- cess in recovering overdue loans from large producers Notes who have borrowed money from Caja Agraria, the state-owned agricultural credit agency: of the total This chapter was originally published in 1996 as John value of overdue loans (Col$150 billion), half is Heath and Hans Binswanger, "Natural Resource attributable to large producers; most of the money Degradation Effects of Poverty and Population recovered has come from small producers. Growth Are Largely Policy-Induced: The Case of Refinancing schemes have been poorly targeted and Colombia," Environment and Development may have helped to postpone necessary adjustment in Economics l(1):65-83. The authors are grateful to the sector, setting a precedent that might discourage Cambridge University Press for permission to reprint larger producers form repaying loans. this edited version. 4. Recent studies confirm the earlier findings and 1. Specifically, Grepperud (1994) states that "This provide a theoretical rationale for the observed rela- paper tests the population pressure hypothesis for the tionship based on incentives issues and missing mar- Ethiopian highlands using quantitative methods. The kets. This literature is summarized by Binswanger, hypothesis posits that, under comparable physical Deininger, and Feder (1995). conditions, heavily eroded areas occur in highly pop- 5. In Colombia, about 5 percent of farmland was ulated regions. A soil erosion severity index, a proxy sold in 1990 and 1991 (Suarez and others 1993). The variable for soil erosion, was chosen as the dependent percentage of farmland transferred on average each variable. Because the dependent variable is categori- year is 3 percent of the total in the United States, 1.5 cal and ordinal, an ordinal cumulative logit model was percent in Great Britain and in the white sector of chosen for the analysis. Two alternative variables South Africa, and 0.5 percent in Ireland and Kenya were applied to reflect population pressure: the ratio (Moll 1988: 354). of rural population to arable land and the ratio of the 6. Between 1995 and 1998, the intention is to facil- population-support capacity to rural population. For itate the purclhase of 1 million hectares by 69,900 the first ratio, the effect on the level of soil erosion families and to title 5 mnillion hectares of frontier land, seemed linear and weak, because substantial increas- benefiting 178,600 families. The total number of ben- es in population density were needed to increase the eficiaries is estimated at 1.2 million, or 13 percent of odds of an area being classified as more seriously the rural population. eroded. For the second ratio, a reciprocal transforma- tion improved the explanatory power of the model, References suggesting a hyperbolic relation between the ratio and the soil erosion index. The probability of an area The word "processed" describes informally repro- being classified above any given level of soil erosion duced works that may not be commonly available increases rapidly as a rural population exceeds the through libraries. population-supporting capacity of its region." Anderson, Jock R., and Jesuthason Thampapillai. 2. Taking Mexico, Central America, the Andean 1990. Soil Conservation in Developing Countries: countries, and the Caribbean as a whole, Posner Project and Policy Intervention. Policy and (1981) estimates that 30 percent of peasant house- Research Series 8. Washington, D.C.: World Bank. holds farm steep slopes, generating 20-40 percent of Ashby, J. A. 1985. "The Social Ecology of Soil the output of annual crops. Erosion in a Colombian Farming System." Rural 3. Law 101 of 1993 offers a mandate for continu- Sociology 50(3):377-96. ing credit subsidies and low interest rates for agricul- Barbier, Edward B., and Joshua T. Bishop. 1995. ture indefinitely: the policies do not facilitate small "Economic Values and Incentives Affecting Soil farmer access to credit, most of the subsidies being and Water Conservation in Developing Countries." captured by larger farmers. Because it is mostly large Journal of Soil and Water Conservation 50(2, farmers who qualify for working capital loans, they March-April):133-37. 34 Policy-Induced Effects of Natural Resource Degradation: The Case of Colombia Bejerano, J. A. 1988. "Efectos de la violencia en la African Reserves." World Development 16(3): producci6n agropecuaria." Coyuntura Econ6mica 349-60. 18(3, September):25-45. Pingali, Prabhu, Yves Bigot, and Hans Binswanger. Berry, R. A. 1992. "Agriculture during the Eighties' 1987. Agricultural Mechanization and the Recession in Colombia." InA. Cohen and F. R. Gunter, Evolution of Farming Systems in Sub-Saharan eds., The Colombian Economy: Issues of Trade and Africa. Baltimore, Md.: Johns Hopkins University Development. Boulder, Colo.: Westview Press. Press. Berry, A., and W. Cline. 1979. Agrarian Structure and Posner, J. L. 1981. The Densely Populated Steep Productivity in Developing Countries. Baltimore, Slopes of Tropical America: Profile of a Fragile Md.: Johns Hopkins University Press. Environment. New York: Rockefeller Foundation. Binswanger, Hans. 1989. "Brazilian Policies That Reyes, Alvaro, and Jaime Martinez. 1994. Encourage Deforestation in the Amazon." World "Funcionamiento de los mercados de trabajo Development 19(7):821-29. rurales en Colombia." In Clara Gonzalez and Binswanger, Hans P., Klaus Deininger, and Gershon Carlos Felipe Jaramillo, eds., Competitividad sin Feder. 1995. "Power, Distortions, Revolt, and pobreza. Bogoti: Tercer Mundo Editores. Reform in Agricultural Land Relations." In J. Ruthenberg, Hans. 1980. Farming Systems in the Behrman and T. N. Srinivasan, eds., Handbook of Tropics. New York: Oxford University Press. Development Economics. Vol. 3B. Amsterdam: Suarez, R., G. Hurtado, L. A. Pacheco, and E. Segura. Elsevier. 1993. El mercado de tierras y la formaci6n de Blakemore, H., and C. T. Smith. 1971. Latin America: propietarios en Colombia. Bogota: Centro de Geographical Perspectives. London: Methuen. Estudios Ganaderos y Agrfcolas. Boserup, Ester. 1965. The Conditions of Agricultural Syrquin, Moshe, and H. Chenery. 1989. Patterns of Growth: The Economics of Agrarian Change Development, 1950-83. Washington, D.C.: World under Population Pressure. New York: Aldine. Bank; Cambridge, Mass.: Harvard International Cleaver, Kevin M., and Gotz A. Schreiber. 1994. Institute for Development. Reversing the Spiral: The Population, Agriculture, Tiffen, Mary, Michael Mortimore, and Francis and Environment Nexus in Sub-Saharan Africa. Gichuki. 1994. More People, Less Erosion: Washington, D.C.: World Bank. Environmental Recovery in Kenya. Chichester, Currie, Lauchlin. 1965. El manejo de cuencas en U.K.: John Wiley. Colombia: Estudio sobre el uso de las tierras. World Bank. 1995a. "Colombia: Poverty Assess- Bogotd: Tercer Mundo. ment." Country Department III, Natural Resources Grepperud, Sverre. 1994. "Population-Environment Management and Rural Poverty Division, Latin Links: Testing a Soil Degradation Model for American and the Caribbean Region, Washington, Ethiopia." Divisional Working Paper 1994-46. D.C. Processed. Pollution and Environmental Economics Division, . 1995b. "Colombia: Review of Agricultural World Bank, Washington, D.C. Processed. and Rural Development Strategy." Report 13437. IGAC (Instituto Geogrdfico Colombiano). 1988. Natural Resources Management and Rural Poverty Suelos y bosques de Colombia. Bogota. Division, Latin America and the Caribbean Misi6n de Estudios del Sector Agropecuario. 1990. El Region, World Bank, Washington, D.C. Processed. desarrollo agropecuario en Colombia. Vol. 1. World Bank, LACTAD (Latin America and the Bogota: Ministerio de Agricultura, Departamento Caribbean Technical Advisory Group). 1994. "Trade Nacional de Planeaci6n. Surveillance Study: Colombia." Washington, D.C. Moll, P. G. 1988. "Transition to Freehold in the South Processed. 4 Toward More Conducive Policies for Sustainable Agriculture Jules N. Pretty As this century draws to a close, agricultural develop- The business-as-usual optimists, with a strong ment faces unprecedented challenges. By the year belief in the power of the market, say that supply will 2020, the world may have to support some 8.4 billion always meet increasing demand and that world food people. Even though enough food is being produced production will continue to grow alongside expected in aggregate to feed everyone, some 800 million peo- reductions in population growth (Rosegrant and ple still do not have access to sufficient food. This Agcaolli 1994, Mitchell and Ingco 1993, and FAO includes 180 million underweight children suffering 1993). They argue that food prices are falling (down from malnutrition. The gap between the wealthy and 50 percent in the past decade for most commodities), the poor has widened. Despite a doubling in global indicating that there is no current crunch in demand. income in the past three decades, the number of peo- Food production will continue to expand as the fruits ple living in poverty continued to rise, from 944 mil- of biotechnology research ripen, boosting plant and lion to 1,300 million. animal productivity, and as the area under cultivation Recent models constructed to investigate changes expands, probably some 20-40 percent by 2020 (this in agricultural production and food security over the may mean that an extra 79 million hectares of uncul- next quarter to half century all conclude that food pro- tivated land are converted to agriculture in Sub- duction will have to increase substantially (IFPRI Saharan Africa alone). They expect that population 1995, Crosson and Anderson 1995, Leach 1995, growth will slow and that developing countries will CGIAR 1995, and FAO 1993, 1995). But the views on substantially increase food imports from industrial- how to proceed vary enormously. Some are opti- ized countries (perhaps as much as fivefold by 2050). mistic, even complacent; others are darkly pes- The environmental pessimists contend that ecolog- simistic. Some indicate that not much needs to ical limits to growth are being approached or have change; others argue that agricultural and food sys- already been reached (Harris 1995, Brown 1994, tems need fundamental reforms. Some indicate that a CGIAR 1995, Kendall and Pimentel 1994, Brown and significant growth in food production will occur only Kane 1994, and Ehrlich 1968). Following a neo- if new lands are taken under the plow; others suggest Malthusian line, these pessimists claim that popula- that there are sufficient social and technical solutions tions continue to grow too rapidly, while growth in the for increasing yields on existing farmland. yield of major cereals will slow or even fall, particu- There are five distinct schools of thought for future larly because of growing production constraints in the options in agricultural development (see also McCalla form of resource degradation (soil erosion, land 1994, Hazell 1995, Hewitt and Smith 1995, Pretty degradation, forest loss, excessive use of pesticides, 1995b, and Hinchcliffe, Thompson, and Pretty 1996). and excessive exploitation of fisheries). Dietary 35 36 Toward More Conducive Policies for Sustainable Agriculture shifts, especially increasing consumption of livestock 1995a, 1995b, Hazell 1995, McCalla 1994, 1995, products, are an emerging threat, because this results Scoones and Thompson 1994, Northwest Area in the consumption by livestock of an even greater Foundation 1994, and Hewitt and Smith 1995). share of cereal products. They do not believe that new Proponents argue that empirical evidence now indi- technological breakthroughs are likely. Solving these cates that regenerative and low-input (but not neces- problems means that seeking some form of population sarily zero-input) agriculture can be highly produc- control should be the first priority. tive, provided farmers participate fully in all stages of The industrialized-world-to-the-rescue lobby technology development and extension. This evi- believes that developing countries will never be able dence also suggests that the productivity of agricul- to feed themselves, for a wide range of ecological, tural and pastoral lands is as much a function of institutional, and infrastructural reasons, and that the human capacity and ingenuity as it is of biological looming food gap will have to be filled by modem- and physical processes. Such sustainable agriculture ized agriculture in the industrialized countries (Avery seeks the integrated use of a wide range of pest, nutri- 1995, Wirth 1995, DowElanco 1994, Carruthers 1993, ent, soil, and water management technologies. It aims and Knutson and others 1990). Increasing production for an increased diversity of enterprises within farms in large, mechanized operations will force smaller and combined with increased linkages and flows between more "marginal" farmers to go out of business, taking them. By-products or wastes from one component or the pressure off natural resources. These resources enterprise become inputs to another. As natural can then be conserved in protected areas and wilder- processes increasingly substitute for external inputs, nesses. The larger producers will then be able to trade the impact on the environment is reduced. their food with those who need it or have it distributed by famine relief or food aid. They vigorously argue What Is and What Is Not Sustainable Agriculture? that any adverse health and environmental conse- quences of chemically based agricultural systems are Sustainability is a word that has entered common use mninor in comparison with those wrought by the in recent years. Since the Brundtland Commission put expansion of agriculture into new lands. External "sustainable development" on the map in the middle inputs (especially pesticides and fertilizers) and free to late 1980s, close to 100 definitions of sustainabili- trade are said to represent a crucial part of any strategy ty have been published. Each emphasizes different for feeding the world (see Avery 1995, in particular). values, priorities, and practices. Clearly no reasonable One group, what we might call the new mod- person is opposed to the idea. But what does it mean? ernists, argues that increases in biological yield are To some it implies the capacity of something to con- possible on existing lands, but that this food growth tinue unchanged for a long time. To others, it implies can only come from "modem" high-extemal-input not damaging natural resources. To others still, it is farming (Borlaug 1992, 1994a, 1994b, Sasakawa just accounting for the environment while continuing Global 2000 1993-95, World Bank 1993, Paarlberg on a business-as-usual track. Does any of this help in 1994, Winrock International 1994, and Crosson and the context of farming? We all know that sustainabil- Anderson 1995). This group argues that farmers sim- ity represents something good, but what exactly? And, ply use too few fertilizers and pesticides, which are more important, has the notion of sustainable agricul- said to be the only way to improve yields and keep the ture contributed to better farm practices, or is the term pressure off natural habitats. This repeat of the green too easily hijacked? revolution model iis widely termed science-based agri- In any discussion of sustainability, it is important culture, the objective being to increase farmers' use of to clarify what is being sustained, for how long, for fertilizers and pesticides. It also argues that high-input whose benefit and at whose cost, over what area, and agriculture is more environmentally sustainable than measured by what criteria. Answering these questions low-input agriculture, which makes intensive use of is difficult, because it means assessing and trading off local resources that may be degraded in the process. values and beliefs (Pretty 1995b, Viederman 1994, Another group, though, is making the case for the Steer and Lutz 1993). It is critical, therefore, that sus- benefits of sustainable intensification, on the grounds tainable agriculture not prescribe a concretely defined that substantial growth is possible in currently unim- set of technologies, practices, or policies. This would proved or degraded areas, while at the same time pro- serve only to restrict the future options of farmers. As tecting or even regenerating natural resources (Pretty conditions and knowledge change, so must farmers Jules N. Pretty 37 and communities be encouraged and allowed to farmers change and adapt. Sustainable agriculture is, there- * The enhancement of wildlife and other public fore, not a simple model or package to be imposed. It goods of the countryside. is more a process for learning (Pretty 1995a and Sustainable agriculture seeks the integrated use Roling 1994). of a wide range of pest, nutrient, soil, and water During the past 50 years, agricultural and rural management technologies. It aims for an increased development policies have successfully emphasized diversity of enterprises within farms combined with external inputs as the means to increase food produc- increased linkages and flows between them. By- tion. This has produced remarkable growth in global products or waLstes from one component or enter- consumption of pesticides, inorganic fertilizer, animal prise become inputs to another. As natural processes foodstuffs, and tractors and other machinery. increasingly sulbstitute for external inputs, the nega- However, these external inputs have substituted for tive impacts oni the environment are reduced, and natural control processes and resources, rendering positive contributions are made to regenerate natural them more vulnerable. Pesticides have replaced bio- resources. logical, cultural, and mechanical methods for control- ling pests, weeds, and diseases; inorganic fertilizers Current Extent and Impact have substituted for livestock manure, compost, nitrogen- of Sustainable Agriculture fixing crops, and fertile soils; information for man- agement decisions comes from input suppliers, Increasing evidence shows that regenerative and researchers, and extensionists rather than from local resource-conserving technologies and practices can sources; and fossil fuels have substituted for locally bring both environmental and economic benefits for generated energy sources. What were once valued farmers, communities, and nations. The best evi- local resources often now become waste products. dence comes from countries of Africa, Asia, and The basic challenge for sustainable agriculture is Latin America, where the concern is to increase food to make better use of available physical and human production in the areas where farming has been large- resources. This can be done by minimizing the use of ly untouched by the modern packages of externally external inputs, by regenerating internal resources supplied technologies. In these lands, farming com- more effectively, or by combining the two in various munities adopting regenerative technologies have ways. This ensures the efficient and effective use of substantially improved agricultural yields, often what is available and keeps any dependencies on using only a few or no external inputs (Bunch 1990, external systems to a reasonable minimum. 1993, GTZ 1992, UNDP 1992, Krishna 1994, Shah A more sustainable agriculture is any food produc- 1994, SWCB 1994, Balbarino and Alcober 1994, de tion system that systematically pursues the following Freitas 1994, and Pretty 1995b). goals: But these are not the only sites for successful sus- * A thorough integration of natural processes such as tainable agriculture. In the high-input and generally nutrient cycling, nitrogen fixation, and pest-predator irrigated lands., farmers adopting regenerative tech- relationships into agricultural production processes, nologies have maintained or improved yields while so ensuring profitable and efficient food production substantially reducing their use of inputs (Bagadion * A minimization of the use of those external and and Korten 1991, Kenmore 1991, van der Werf and nonrenewable inputs with the potential to damage de Jager 1992, UNDP 1992, Kamp, Gregory, and the environment or harm the health of farmers and Chowhan 1993, and Pretty 1995b). And in the indus- consumers and targeted use of the remaining trialized countries, farmers have been able to main- inputs used with a view to minimizing costs tain profitability, even though input use has been cut * The full participation of farmers and other rural dramatically, such as in the United States (Liebhardt people in all processes of problem analysis and and others 1989, National Research Council 1989, technology development, adaptation, and exten- Hanson and others 1990, Faeth 1993, Northwest Area sion, leading to an increase in self-reliance among Foundation 1994, and Hewitt and Smith 1995) and in farners and rural communities Europe (El Titi and Landes 1990, Vereijken 1990, * A greater productive use of local knowledge and Jordan, Hutcheon, and Glen 1993, Pretty and Howes practices, including innovative approaches not yet 1993, Reus, Weckseler, and Pak 1994, Somers 1997, fully understood by scientists or widely adopted by and Pretty 1998). 38 Toward AMlore Conducive Policies for Sustainable Agriculture Current Extent is that these represent what is possible on a wider scale. It could be argued, however, that they are The International Institute for Environment and only successful because they have occurred where Development has examined the extent and impact of there is a combination of the least resistance and sustainable agriculture in a select number of coun- most opportunity, although the sheer diversity of tries (Pretty, Thompson, and Hinchcliffe 1996). The approaches and contexts represented undermines government and nongovernmental programs and such an assertion. Moreover, many of the improve- projects included in this analysis share important ments are occurring in difficult, remote, and characteristics. They have: resource-poor areas that have commonly been * Made use of resource-conserving technologies in assumed to be incapable of developing innovations conjunction with group or collective approaches or producing food surpluses. to agricultural improvement and natural resource management Contested Views * Put participatory approaches and farmer-centered activities at the center of their agenda, so that This empirical evidence is still contested. In the these activities are occurring on local people's United States, some 82 percent of conventional terms and are more likely to persist after the proj- farmers believe that low-input agriculture will ects and programs have ended always be low output, even though the top quarter * Not used subsidies or food-for-work to "buy" the of sustainable agriculture farmers obtain better participation of local people or to encourage yields and gross margins than conventional farmers them to adopt. particular technologies, so that (Hewitt and Smith 1995 and Northwest Area improvements are unlikely to fade away or sim- Foundation 1994). Influential politicians continue ply disappear at the end of the project or program to reinforce these beliefs. In 1991 the former U.S. * Supported the active involvement of women as secretary of agriculture, Earl Butz, said, key producers and facilitators We can go back to organic agriculture in this * Emphasized "adding value" to agricultural prod- country if we must-we once farmed that ucts through agroprocessing, marketing, and other way 75 years ago. However, before we move off-farm activities, thus creating employment, in that direction, someone must decide which income-generating opportunities, and surplus 50 million of our people will starve. We sim- retention in the rural economy. ply cannot feed, even at subsistence levels, Two types of transition to sustainable agriculture our 250 million Americans without a large were assessed: from modern or conventional high- production input of chemicals, antibiotics, external-input agriculture (such as farming in green and growth hormones. (Quoted in Schaller revolution lands or in the industrialized countries) 1993.) and from traditional, rainfed agriculture where cereal Yet a selection of recent evidence shows that at yields have largely remained constant over cen- least 40,000 farmers in 32 states are using sustain- turies. Because these transitions are recent (within able agriculture technologies and have cut their use the past 5-10 years), they provide evidence that of external inputs substantially. This includes 2,800 similar improvements could occur elsewhere and sustainable agriculture farmers in the northwestern could be repeated on a larger scale. states, who grow twice as many crops as conven- In the 20 countries of the South (and the total of tional farmers, use 60-70 percent less fertilizer, 63 projects) examined and analyzed, some 2.38 mil- pesticide, and energy, and obtain roughly compara- lion households are farming 4.37 million hectares ble yields; they also spend more money on local with sustainable agriculture technologies and practices goods and services (Northwest Area Foundation (see table 4.1). The data in table 4.1 do not represent a 1994). comprehensive survey of sustainable agriculture in Despite the growing number of successful sus- each of the countries. They do illustrate, however, tainable agriculture initiatives in different parts of what has been achieved by specific projects and what the world, it is clear that most of these are still only could be replicated elsewhere. Most of these "islands of success." There remains a huge chal- improvements have occurred in the past 10 years lenge to find ways to spread or scale up the process- (many in the past two to five years). The assumption es that have brought about these transitions. Jules N. Pretty 39 Table 4.1. Examples of the Extent and Impact of Sustainable and People-Centered Agriculture in Different Agricultural Systems Number Cereal yield offarming Number of improvement households hectares Dominant factor Countries reported reported cereal crop (percent) Rainfed systems Brazil 223,000 1,330,000 Maize, wheat 198-246 Burkina Faso 22,500 37,360 Sorghum, millet 250 Ethiopia 24,175 21,850 Maize 154 Guatemala 17,000 17,000 Maize 250 Honduras 27,000 42,000 Maize 250 India 307,910 993,410 Sorghum, millet 288 Kenya 222,550 250,000 Maize 200 Mexico 7,400 23,500 Coffee 140 Nepal 3,000 1,300 Maize, wheat 164-307 Philippines 850 920 Upland rice 214 Senegal 200,000 400,000 Sorghum, millet 300 Uganda 9,426 21,379 Maize 150 Zambia 6,300 6,300 Sorghum, millet 200 Total 1,146,111 3,257,519 Irrigated systems Bangladesh 11,025 4,772 Rice 110 China 47,000 12,000 Rice 111 India 50,000 71,300 Rice 108 Indonesia 400,000 267,000 Rice 107 Malaysia 2,500 3,925 Rice 108 Philippines 175,000 385,000 Rice 112 Sri Lanka 100,000 95,350 Rice 117 Thailand 500 2,040 Rice 109 Vietnam 6,600 3,540 Rice 108 Total 792,625 844,927 Industrialized systems Germany (integrated) 75,000 200,000 Wheat, barley 90 Netherlands (integrated) 500 - Wheat, barley 85 United States (integrated) 40,000 632,000 Wheat, barley 95 European Union (organic) 50,000 1,200,000 Wheat, barley 80 Total 165,500 1,832,000 - Not available. Note: Improvements are measured against nonsustainable farming equivalents, which are taken to be 100 per- cent. Thus an improvement of 200 percent implies a doubling of yields; one of 90 percent implies a fall in yields of 10 percent. The time frame for these improvements is the life of program activities, usually less than five years. Some improvements are expected to occur in the season following the adoption of sustainable agriculture, and these tend to increase over time. Source: See Pretty, Thompson, and Hinchcliffe 1996 for details of data sources and surveys. 40 Toward More Conducive Policies for Sustainable Agriculture The Record of Conventional Agricultural Projects The Problems with Comprehensive Technology Packages When the recent record of development assistance is considered, it is clear that sustainability has been Modernist agricultural development has begun with poor. There is a widespread perception among both the notion that there are "on-the-shelf' technologies multilateral and bilateral organizations that agricultur- that work, and it is just a matter of inducing or per- al development is difficult, that agricultural projects suading farmers to adopt them. Yet few farmers are perform badly, and that resources may best be spent in able to adopt whole packages of conservation tech- other sectors. Reviews by the World Bank, the nologies without adjusting their own practices and European Comrnission, the Danish International livelihood systems. For example, alley cropping, an Development Agency, and the British Department for agroforestry system comprising rows of nitrogen- International Development (formerly the Overseas fixing trees or bushes separated by rows of cereals, Development Aclministration) have all shown that has long been the focus of research (Kang, Wilson, agricultural and natural resource projects both per- and Lawson 1984 and Attah-Krah and Francis 1987). formed worse in the 1990s than in the 1970s-80s and Many productive and sustainable systems, needing worse than projects from other sectors (World Bank few or no external inputs, have been developed. They 1993, Pohl and Mihaljek 1992, European stop erosion, produce food and wood, and can be Commission 1994, DANIDA 1994, and Dyer and cropped over long periods. But the problem is that Bartholomew 1995). They are also less likely to con- very few, if any, farmers have adopted these alley tinue achievements beyond the provision of aid inputs. cropping systems as designed. Despite millions of A recent analysis of the evaluations of 95 agricultural dollars of research expenditure over many years, projects recorded on the database of the Development many of the systems produced are suitable only for Assistance Committee of the Organisation for Economic research stations (Carter 1995). There has been some Co-operation and Development (OECD) shows a disturb- success, however, where farmers have been able to ing rate of failure, with at least 27 percent of projects adapt one or two components of alley cropping to having nonsustainable structures, practices, or institu- their own farms. In Kenya, for example, farmers tions and 10 percent causing significantly negative planted multiple rows of leguminous trees next to environmental impact (Pretty and Thompson 1996). field boundaries or single rows through their fields; The reasons given for failure include an emphasis on and in Rwanda, extension workers planted alleys that external technologies alone, lack of participation by soon became dispersed through fields (Kerkhof local people, ineffective training of professionals, 1990). and institutions that were working with no view But the prevailing view tends to be that farmers toward the diversity of local conditions and needs of should adapt to the technology. Of the Agroforestry local people. Outreach Project in Haiti, the evaluators said that This evidence from completed agricultural devel- Farmer management of hedgerows does not opment projects suggests three reasons for the lack of conform to the extension program ... Some sustainability: fanners prune the hedgerows too early, others * If coercion or financial incentives are used to too late. Some hedges are not yet pruned by two encourage adoption, sustainable agriculture tech- years of age, when they have already reached nologies, such as soil conservation, alley cropping, heights of 4-5 meters. Other hedges are pruned and integrated pest management, are not likely to too early, mainly because animals are let in, or persist. the tops are cut and carried to animals ... * If imposed, new institutional structures, such as Finally, it is very common for farmers to allow cooperatives or other groups at the local level or some of the trees in the hedgerow to grow to project management units and other institutions at pole size. (Bannister and Nair 1990.) the project level, rarely persist beyond the project. This could be read as a great success-farmers were * If introduced with no thought to how they will be clearly adapting the technology to their own special paid for, expensive external inputs, including needs. But it was not. The language of the evaluators subsidized inputs, machinery, or high-technology is quite clear: this was considered a failure. hardware, generally do not persist beyond the In Laos, one project used food for work to encour- project. age shifting agriculturalists to settle and adopt contour Jules N. Pretty 41 farming with bench terraces (Fujisaka 1989). But back from capital cities. There are also benefits to these fields became so infested with weeds that farm- the forests. Farmners say they no longer need to cut the ers were forced to shift to new lands, and the struc- forests, because they have the technologies to farm tures were so unstable in the face of seasonal rains permanently the same piece of land. Before the pro- that they led to worsened gully erosion. Farmers then grams, national park authorities sought to keep vil- refused to do further work when the incentives were lagers out of the forests; now there is no such concern gone. because the forests are no longer threatened. What are the implications for sustainable agricul- Few published studies give evidence of impacts ture? How should we proceed so as to ensure that years after outside interventions have ended. In 1994, farmers are fully involved in developing and adapting however, staff oi the Honduran organization COSECHA these sustainable and productive technologies? (Asociaci6n de Consejeros para una Agricultura Sostenible, Ecol6gica y Humana) returned to the three Enhancing Farmers' Capacity to Innovate program areas and used participatory methods with local communities to evaluate subsequent changes Important evidence comes from a variety of soil con- (Bunch and Lopez 1996). servation and agricultural regeneration programs in They first divided all 121 villages into three cate- Central America (Bunch and L6pez 1996). The gories, according to where they felt there had been Guinope (1981-89) and Cantarranas (1987-91) pro- good, moderate., and poor impact. Twelve villages grams in Honduras and the San Martfn Jilotepeque were sampled from these-four from each program program in Guatemala (1972-79) were collaborative comprising one of the best, two of the moderate, and efforts between World Neighbors and other local one poor. These villages had sorne 1,000 families agencies. All began with a focus on soil conservation (with a range of 30 to 180 per village). The first major in areas where maize yields were very low (400 to 660 finding was that crop yields and adoption of conserv- kilograms per hectare) and where shifting cultivation, ing technologies continued to grow after the project malnutrition, and outmigration prevailed. All show ended (see table 4.2). the importance of developing resource-conserving Surprisingly, though, many of the technologies practices in partnership with local people. known to be successful during the project had been There were several common elements. All forms superseded by new practices. Had the original tech- of paternalism were avoided, including giving things nologies been poorly selected? ][t would appear not, as away, subsidizing farmer activities or inputs, or doing many that had been dropped by farmers are still very anything for local people. Each started slowly and on successful elsewhere. The explanation appears to be a small scale, so that local people could meaningfully that changing external and internal circumstances had participate in planning and implementation. They reduced or eliminated their usefulness, such as chang- used technologies, such as green manure, cover crops, ing markets, droughts, diseases, insect pests, land contour grass strips, in-row tillage, rock bunds, and tenure, labor availability, and political disruptions. animal manure, that were appropriate to the local area Altogether, some 80-90 successful innovations and were finely tuned through experimentation by and were documented in these 12 villages. In one with farmers. Extension and training were done large- Honduran village, Pacayas, there had been 16 innova- ly by villager farmers who had already experienced tions, including four new crops, two new types of success with the technologies on their own farms. green manure, two new species of grass for contour Each program substantially improved agricultural barriers in vegetables, chicken pens made of king yields, increasing output per area of land from some grass, marigolds for nematode control, lablab and vel- 400-600 kilograms per hectare to 2,000-2,500 vet bean as cattle and chicken feed, nutrient recycling kilograms per hectare. Altogether improvements into fish ponds, human wastes placed in composting have been made in some 120 villages. Over time, soils latrines, napier grass to stabilize cliffs, and home- have not been simply conserved but regenerated, with made sprinklers for irrigation. depth increases from 0.1 meter to 0.4-1.3 meters not Technologies, had been developed, adopted, adapt- uncommon. These programs have also helped to ed, and dropped. The study concluded that the half- regenerate local economies. Land prices and labor life of a successiul technology in these project areas is rates are higher inside the project areas than outside. six years. Quite clearly the technologies themselves There are housing booms, and families have moved are not sustainable. As Bunch and L6pez (personal 42 Toward AMore Conducive Policies for Sustainable Agriculture Table 4.2. Changes in the Adoption of Resource-Conserving Technologies, Maize Yields, and Migration Patterns in Three Programs in Central America during and after Projects Technology At initiation At termination' In 1994 Number offarme.rs with the technology Contour grass barriers 1 192 280 Contour drainage ditches 1 253 239 Contour rows 0 100 245 Green manure 0 35 52 Crop rotation 12 209 254 No burning fields or forests 2 160 235 Organic matter as fertilizer 44 195 397 Yields of maize (kilograms per hectare) 1. San Martin, Guatemala (1972-79) 400 2,500 4,500 2. Guinope, Honduras (1981-89) 600 2,400 2,730 3. Cantarranas, Honduras (1987-91) 660 2,000 2,050 Migration (number of households) 1. San Martin: 2 villages San Antonio Correjo 65 n.d. 4 Las Venturas 85 n.d. 4 2. Guinope: 3 villages 38 0 (2)b 3. Cantarranas: 3 villages n.d. 10 (6)b n.d. No data. a. Program termination dates were San Martin, 1979; Guinope, 1989; Cantarranas, 1991. b. Numbers in parentheses refer to negative outmigration, that is, families returning to their villages. Source: Bunch and L6pez 1996. communication, 1996) have put it, "What needs to be the Department of Primary Industry, using very sim- made sustainable is the social process of innovation ple learning tools that enabled farmers to investigate itself." Sustainability does not equal fossilization or the impact of rainfall on their soil, have encouraged continuation of a thing or practice forever: rather it more than 80 percent of farmers to adopt conservation implies an enhanced capacity to adapt in the face of technologies. Many of these have gone on to develop unexpected changes and emerging uncertainties. and adopt new and different technologies for their A similar picture has emerged in Gujarat, where own farms, and they now fully support the values and many farmers have developed new technical innova- principles they once would have opposed (Hamilton tions after support from the Aga Khan Rural Support 1995). Program for undertaking simple conservation mea- Another example comes from Thailand, where the sures. Farmers have introduced planting of grafted four phases of the Thai-German Highland mango trees and bamboo near embankments, making Development Project clearly illustrate the importance full use of residual moisture near gully traps. They of genuine participation with local people (Thai- have also introduced cultivation of vegetables, such German Highland Development Project 1995 and as brinjal and lady's finger, other leguminous crops, Steve Carson, personal communication, 1996). The and tobacco in the newly created silt traps. This has project has been working with upland communities in increased production substantially, particularly in northern Thailand to support the transition toward years of poor rainfall. Most of these innovations and sustainable agriculture. The resource-conserving adaptations have been introduced and sustained with technologies developed and adapted for local use support from the local network of village extension- include hedgerows on contours, buffer strips, new ists (Shah 1994). crop rotations, integrated pest management, crop In South Queensland, Australia, extensionists from diversification, and livestock integration. The Jules N. Pretty 43 approach, however, has changed significantly since them on to consutmers, and would be forced to switch the mid-1980s (see table 4.3). In the first phase, cash to more resource-conserving technologies. This incentives and free inputs were used to encourage notion is captured in the polluter-pays principle, a adoption of these technologies; as a result adoption concept used for many years in the nonfarm sector rates were high, although there was little or no adap- (OECD 1989). However, beyond the notion of encour- tation of the technologies by the farmers. In 1990 all aging some internalization of costs, it has not been of the incentives were stopped when the project adopted practical use for policy formulation in agriculture. a participatory approach; adoption rates fell immedi- In general, farmers are entirely rational to contin- ately, and withdrawal increased threefold. By ue using high-input degrading practices under current 1993-94, the participatory village planning had fully policies. High prices for particular commodities, such involved communities, and the ratio of adopters to as key cereals, have discouraged mixed farming prac- withdrawers was equal. Most recently, the number of tices, replacing them with monocultures. This is strik- farmers using sustainable technologies has grown ingly clear in the United States, where commodity rapidly, and, crucially, farmers are now actively programs inhibit the adoption of sustainable practices adapting those technologies and innovating new ones by artificially making them less profitable to farmers. to satisfy their particular needs (Steve Carson, per- In Pennsylvania, the financial returns to monocropped sonal communication, 1996). and continuous maize are about the same as those to sustainable agriculture involving mixed rotations The Policy Environment in Industrial Countries (Faeth 1993). But continuous maize attracts about twice as much d.irect support in the form of deficien- Current policies often do not reflect the long-term cy payments. In addition, continuous maize farms social and environmental costs of resource use. The require much more nitrogen fertilizer, erode more external costs of modern farming, such as soil erosion, soil, and cause three to six times as much damage to health damage, or polluted ecosystems, generally are off-site resources. Quite clearly, a transition to not incorporated into individual decisionmaking by resource-conserving rotations would substantially farmers. In this way, resource-degrading farmers bear benefit both farmers and the national economy. neither the costs of damage to the environment or In this context of systemic support for high-input economy nor those incurred in controlling the pollut- agriculture, many countries have sought to "bolt" ing or damaging activity (Pretty 1996). In principle, it conservation goals onto these policies. These coun- is possible to imagine pricing the free input to farm- tries have tended to rely on conditionality, such as ing of the clean, unpolluted environment. If charges cross-compliance, whereby farmers receive support were levied in some way, then degraders or polluters only if they adopt certain types of resource-conserving would have higher costs, would be forced to pass technologies and practices. Such cross-compliance Table 4.3. Changing Phases in the Thai-German Highland Development Project: The Case of 113 Villages in Nam Lang, Northern Thailand Ratio of adoption Phase and date Characteristics to withdrawal 1. 1987-90 Cash incentives and free inputs; high adoption, but little 5:1 or no adaptation of technologies 2. 1991-92 All incentives stopped; beginning of participatory work; 1:2 adoption rates fell to 25 percent of phase 1; withdlrawal immediately increased threefold 3. 1993-94 Participatory village planning; communities fully involved; 1:1 adopters and withdrawers now equal 4. 1995-96 Adopters increasing; farmers adapting technologies and 3:1 diversifying (pineapple strips, lemongrass, cash c:rops, soil and water conservation) Source: Steve Carson, personal communication, 1996. 44 Toward More Conducive Policies for Sustainable Agriculture occurs widely in the South too. The process of agri- In 1989 the Group Farming for Rice Program was cultural modernization has widely involved encourag- launched. Local committees comprising all rice farm- ing farmers to adopt modern practices through the ers were formed to chart a detailed plan of fanning. In linkage of credit or other benefits. If farmers wish to this group, activities, such as water management and receive one type of support, they must adopt a partic- labor operations, are agreed jointly. Costs are reduced ular set of technologies and practices. In some cases, through community nursery raising of rice, fertilizer coercion has been used to achieve certain levels of applications on the basis of soil testing, the introduc- adoption. The problem with these cross-compliances tion of integrated pest management and minimum use is that they can create long-term resentment and lead of pesticides, and the formation of plant protection to a reversal of practices when policies change or the squads. The average cost reduction to farmers has money runs out. been Rsl,000 per hectare, and rice yields have Although a growing number of policy initiatives improved 500 kilograms per hectare. are oriented specifically toward improving the sus- In China, agricultural policy is encouraging farm- tainability of agriculture, most have focused on input ers to grow green manure in the rice fields (Yixian reduction strategies. Only a few as yet represent 1991). During the 1980s, continuous and monocrop- coherent plans and processes that clearly demonstrate ping of rice caused widespread soil fertility and pest the value of integrating policy goals. A thriving and problems. The agricultural ministry set up multiplica- sustainable agricultural sector requires both integrated tion bases for green manure, which are expected to action by farme:rs and communities and integrated produce 5.5 million kilograms of seed each year. In action by policymakers and planners. This implies some regions, farmers selling green manure seed to both horizontal integration with better linkages state-run farm cooperatives receive fertilizers at lower between sectors and vertical integration with better prices. Green manure and plant residues are now used linkages between the micro and macro level. on 68 percent of the 22 million hectares of rice fields. Integration has been the policy buzz of the 1990s. Similar successes have been observed in Indonesia But putting this desired integration into practice has where, during the late 1980s, pesticide subsidies were been much more difficult. There have been substantial cut from 85 percent to zero by January 1989 and differences in the views of major policy actors, such farmer field schools were established for integrated as those representing the interests of farmers, envi- pest management (Kenmore 1991, Matteson 1995, ronmentalists, and treasuries. Other problems include FAO 1994, and Roling and van der Fliert 1997). The market failure, such as nonpayment or underpayment country saves some $130 million-$160 million each for resource degradation, or intervention failure, such year, and pesticide production fell nearly 60 percent as undersupply of public goods. Nonetheless, small between 1985 and 1990. Rice yields have continued steps are occasionally being taken both to penalize to improve, despite the cut in inputs. This raises a cru- polluters and to encourage resource conservers, par- cial policy issue when it comes to cutting inputs. ticularly in industrial countries where levels of facili- Farmers who depend on external inputs need support ties and pesticide applications are much higher than in to make the transition to a more sustainable agricul- developing nations. ture. As this program demonstrates, if this support increases farmers' capacity to learn and act on their Policy Initiatives in the South own fanns, the substitute is more than adequate. The FAO is supporting similar integrated pest management The Group Farming Initiative of the Kerala State gov- programs for rice in eight other countries of South and ermment in India is a good example of how coordinat- Southeast Asia (FAo 1994). Together, these programs ed action within the agricultural sector can have a sig- are training hundreds of thousands of farmers and nificant impact on farming practices (Sherief 1991). have saved many millions of dollars in pesticides. Land reform in the 1970s led to the formation of a new class of small farmers, with some 70 percent Policy Processes: Conditionality or Participation? owning less than 2 hectares. But as the costs of inputs and pest and disease control spiraled, the area under Policy reform has been under way in many countries, rice fell from 810,000 to 570,000 hectares between with some new initiatives supporting elements of a 1975 and 1988. Small farmers were unable to adopt more sustainable agriculture. Most of these have the whole technological package. focused on input reduction strategies, because of con- Jules N. Pretty 45 cerns over foreign exchange expenditure or environ- sustainable agriculture, are impossible. Sustainability mental damage. Only a few as yet represent coherent itself is a complex and contested concept. Sustainable plans and processes that clearly demonstrate the value agriculture should not, therefore, be seen as a set of of integrating policy goals. practices to be fixed in time and space. It implies the Governments can do much with existing resources capacity to adapt and change as external and internal to encourage and nurture the transition from modern- conditions change. Yet there is a danger that policy, as ized systems toward more sustainable alternatives it has tended to do in the past, will prescribe the prac- (see table 4.4). The first action that governments can tices that farmers should use rather than create the take is to declare a national policy for sustainable enabling conditions for locally generated and adapted agriculture. This would help to raise the profile of technologies. these processes and needs as well as give explicit Environmental policy has sometimes tended to value to alternative societal goals. It would also estab- take the view that mral people mismanage natural lish the necessary framework within which the more resources. The history of soil and water conservation, specific actions listed in table 4.4 can fit and be sup- rangeland management, protected area management, ported. irrigation development, and modern crop dissernina- It is important to be clear about just how policies tion shows a common pattern: technical prescriptions should be trying to address issues of sustainability. As are derived from controlled and uniform conditions, suggested at the beginning of this chapter, precise and supported by limited cases of success, and then absolute definitions of sustainability, and therefore of applied widely with little or no regard for diverse Table 4.4. Polices That Work toward Sustainable Agriculture Policy Intent Policy 1 Declare a national policy for sustainable agriculture Encourage resource-conserving technologies and practices Policy 2 Establish a national strategy for integrated pest management Policy 3 Prioritize research into sustainable agriculture Policy 4 Grant farmers appropriate property rights Policy 5 Promote farmer-to-fanner exchanges Policy 6 Direct limited grants toward sustainable technologies Policy 7 Link support payments to resource-conserving practices Policy 8 Provide better information for consumers and the public Policy 9 Adopt natural resource accounting Policy 10 Establish appropriate standards and licensing for pesticides Support local groups for community action Policy 11 Encourage the formation of local groups Policy 12 Foster rural partnerships Policy 13 Support training and field schools for farmers Policy 14 Provide incentives for on-farm employment Policy 15 Permit groups to have access to credit Reform external institutions and professional approaches Policy 16 Encourage the formal adoption of participatory methods and processes Policy 17 Support information systems to link research, extension, and farmers Policy 18 Rethink the project culture Policy 19 Strengthen the capacities of NGOS to scale up Policy 20 Foster stronger NGO-govemment partnerships Policy 21 Reform teaching and training establishments Policy 22 Develop capacity in planning for conflict resolution and mediation Source: Adapted from Pretty 1995b. 46 Toward More Conducive Policies for Sustainable Agriculture local needs and conditions (Pretty and Shah 1994, for the Hillsides of Haiti: Experience of an Benhke and Scoones 1992, and Pimbert and Pretty Agroforestry Outreach Project." American Journal 1995). Differences in receiving environments and ofAlternative Agriculture 5(2):51-59. livelihoods often make the technologies unworkable Beaumont, Peter. 1993. Pesticides, Policies, and and unacceptable. When they are rejected locally, People. London: The Pesticides Trust. policies shift to seeking success through the manip- Benhke, Roy, and Ian Scoones. 1992. Rethinking ulation of social, economic, and ecological envi- Range Ecology: Implications for Rangeland ronments and, in some cases, through outright Management in Africa. Drylands Program Issues enforcement. Paper 33. London: International Institute for For sustainable agriculture to spread widely, poli- Environment and Development. cy formulation must not repeat these mistakes. Borlaug, Norman. 1992. "Small-Scale Agriculture in Policies will have to arise in a new way. They must be Africa: The Myths and Realities." Feeding the enabling and create the conditions for sustainable Future (newsletter of the Sasakawa Africa development based on locally available resources and Association) 4:2. local skills and knowledge. Achieving this will be dif- . 1994a. "Agricultural Research for Sustain- ficult. In practice, policy is the net result of the actions able Development." Testimony before the U.S. of different interest groups pulling in complementary House of Representatives Committee on Agricul- and opposing directions. It is not just the normative ture, Washington, D.C., 1 March. Processed. expression of governments. Effective policy will have . 1994b. "Chemical Fertilizer 'Essential to recognize this and seek to bring together a range of [Letter]."' International Agricultural Development actors and institutions for creative interaction and (November-December):23. joint learning. Brown, Lester R. 1994. "The World Food Prospect: Entering a New Era." In Assisting Sustainable References Food Production: Apathy or Action? Arlington, Va.: Winrock International. The word "processed" describes informally repro- Brown, Lester R., and Hal Kane. 1994. Full House: duced works that may not be commonly available Reassessing the Earth's Population Carrying through libraries. Capacity. New York: W. W. Norton and Co. Attah-Krah, A. N., and P. A. Francis. 1987. "The Role Bunch, Roland. 1990. Low-Input Soil Restoration in of On-Farm lTrials in the Evaluation of Composite Honduras: The Cantarranas Farner-to-Farmer Technologies: The Case of Alley Fanning in Southern Extension Program. Gatekeeper Series SA23. Nigeria." Agricultural Systems 23:133-52. London: International Institute for Environment Avery, Denis. 1995. Saving the Planet with Pesticides and Development, Sustainable Agriculture and Plastic. Indianapolis, Ind.: The Hudson Program. Institute. . 1993. "EPAGRI'S Work in the State of Santa Bagadion, B. U., and F. F. Korten. 1991. "Developing Catarina, Brazil: Major New Possibilities for Irrigators' Organisations: A Learning Process Resource-Poor Farmers." Asociaci6n de Approach." In Michael M. Cemea, ed., Putting Consejeros para una Agricultura Sostenible, People First, 2d ed. New York: Oxford University Ecol6gica y Humana, Tegucigalpa. Processed. Press. Bunch, Roland, and Gabino L6pez. 1996. Soil Balbarino, Edwin A., and Doris L. Alcober. 1994. Recuperation in Central America: Sustaining "Participatory Watershed Management in Leyte, Innovation after Intervention. Gatekeeper Series Philippines: ]Experience and Impacts after Three SA 55. London: International Institute for Years." Paper prepared for the International Environment and Development, Sustainable Institute for Environment and Development/ Agriculture Program. ActionAid conference new horizons: the social, Carruthers, Ian. 1993. "Going, Going, Gone! Tropical economic, and environmental impacts of participa- Agriculture as We Knew It." Tropical Agriculture tory watershed development, Bangalore, India, 28 Association Newsletter 13(3):1-5. November-2 December. Processed. Carter, Jane. 1995. Alley Cropping: Have Resource- Bannister, M. E., and P. K. R. Nair. 1990. "Alley Poor Farmers Benefited? Natural Resource Cropping as a Sustainable Agricultural Technology Perspectives 3. London: Overseas Development Jules N. Pretty 47 Institute. Philippines, and the United States. Washington, CGIAR (Consultative Group on International D.C.: World Resources Institute. Agricultural Research). 1995. "Sustainable FAO (Food and Agriculture Organization of the United Agriculture for a Food Secure World: A Vision for Nations). 1993. Strategies for Sustainable International Agricultural Research." Expert panel Agriculture cnd Rural Development (SARD): The of CGIAR, Washington, D.C., and Swedish Role of Agriculture, Forestry, and Fisheries. International Development Authority, Stockholm. Rome. Processed. . 1994. "Intercountry Program for the Crosson, Pierre, and J. R. Anderson. 1995. Development and Application of Integrated Pest "Achieving a Sustainable Agricultural System in Control in Rice in South and Southeast Asia, Phase Sub-Saharan Africa." Building Block for Africa I and II. Project Findings and Recommendations." Paper 2. AFTES Africa Technical Department, Terminal report. Rome. Processed. Environment and Social Division, World Bank, . 1995. World Agriculture: Towards 2010. Washington, D.C. Processed. Rome. DANIDA (Danish International Development Agency). Fujisaka, Sam. 1989. "The Need to Build on Farmer 1994. Agricultural Sector Evaluation: Lessons Practice and Knowledge: Reminders from Learned. Copenhagen: Ministry of Foreign Selected Upland Conservation Projects and Affairs. Policies." Agroforestry Systems 9:141-53. de Freitas, Hercilio V. 1994. "EPAGRI in Santa GTZ (German Agency for Technical Cooperation). Catarina, Brazil: The Micro-Catchment 1992. The Spark Has Jumped the Gap. Eschborn. Approach." Paper prepared for the new horizons Hamilton, Gus. 1995. "Learning to Learn with conference, International Institute for Farmers." Ph.D. diss., Wageningen Agricultural Environment and Development, Bangalore, India, University, WTageningen, Netherlands. November. Processed. Hanson, J. C., D. M. Johnson, S. E. Peters, and R. R. DowElanco. 1994. The Bottom Line. Indianapolis, Janke. 1990. "The Profitability of Sustainable Ind.: DowElanco. Agriculture on a Representative Grain Farm in the Dyer, Nick, and A. Bartholomew. 1995. "Project Mid-Atlantic Region, 1981-1989." Northeastern Completion Reports: Evaluation Synthesis Study." Journal of Agriculture and Resource Economics Evaluation Report Ev583. Overseas Development 19(2):90-98. Administration, London. Processed. Harris, J. M. 11995. World Agriculture: Regional Ehrlich, Paul. 1968. The Population Bomb. New York: Sustainability and Ecological Limits. Discussion Ballantine. Paper 1. Medford, Mass.: Tufts University, Center El Titi, Adel, and H. Landes. 1990. "Integrated for Agriculture, Food, and Environment. Farming System of Lautenbach: A Practical Hazell, Peter. 1995. "Managing Agricultural Contribution toward Sustainable Agriculture." In Intensification." IFPRI 2020, Brief 11. International C. A. Edwards, R. Lal, P. Madden, R. H. Miller, Food Policy Research Institute, Washington, D.C. and G. House, eds., Sustainable Agricultural Processed. Systems. Ankeny, Iowa: Soil and Water Hewitt, Tracy ][., and Katherine R. Smith. 1995. Conservation Society. Intensive Agriculture and Environmental Quality: European Commission. 1994. Evaluation des projets Examining the Newest Agricultural Myth. de developpement rural finances durant les con- Greenbelt, Md.: Henry Wallace Institute for ventions de Lome L II, et 111. Brussels. Alternative Agriculture. Eveleens, Kees G., Richard Chisholm, Elske van der Hinchcliffe, Fiona, John Thompson, and Jules N. Fliert, M. Kato, Pham Thi Nhat, and Peter Pretty. 1996. "Sustainable Agriculture and Food Schmidt. 1996. Mid-Term Review of Phase III FAO Security in East and Southern Africa." A report Intercountry Programme for the Development and prepared for the Swedish International Application of Integrated Pest Control in South Development Cooperation Agency, Stockholm, and South East Asia. GCP/RAS/145-147/NET. and International Institute for Environment and Rome: Food and Agriculture Organization. Development, London. Processed. Faeth, Paul, ed. 1993. Agricultural Policy and IFPRI (International Food Policy Research Institute). Sustainability: Case Studies from India, Chile, the 1995. A 2020 Vision for Food, Agriculture, and the 48 Toward More Conducive Policies for Sustainable Agriculture Environment. Washington, D.C. Activities of the World Bank." Paper prepared for Jordan, Vic W. L., J. A. Hutcheon, and D. M. Glen. the fifteenth annual agricultural symposium, 1993. Studies in Technology Transfer of Integrated World Bank, Washington, D.C., 4-6 January. Farming Systems: Considerations and Principles Processed. for Development. Bristol: Institute of Arable Crops Mitchell, D. O., and M. D. Ingco. 1993. "The World Research, Long Ashton Research Station. Food Outlook." International Economics Kamp, K., R. Gregory, and G. Chowhan. 1993. Department, World Bank, Washington, D.C. "Fish Cutting Pesticide Use." ILEIA Newsletter Processed. 2(93):22-23. National Research Council. 1989. Alternative Kang, B. T., G. F. Wilson, and T. L. Lawson. 1984. Agriculture. Washington, D.C.: National Academy Alley Cropping: A Stable Alternative to Shifting Press. Agriculture. Ibadan: International Institute for Northwest Area Foundation. 1994. A Better Row to Tropical Agriculture. Hoe: The Economic, Environmental, and Social Kendall, H. W., and D. Pimentel. 1994. "Constraints Impact of Sustainable Agriculture. St. Paul, Minn. on the Expansion of the Global Food Supply." OECD (Organisation for Economic Co-operation and Ambio 23:198-205. Development). 1989. Agricultural and Environ- Kenmore, Peter. 1991. How Rice Farmers Clean up mental Policies. Paris. the Environment, Conserve Biodiversity, Raise Paarlberg, Roy L. 1994. "Sustainable Farming: A More Food, Make Higher Profits. Indonesia's Political Geography." IFPRI 2020, Brief 4. IPM-A Model for Asia. Manila: Food and International Food Policy Research Institute, Agriculture Organization of the United Nations. Washington, D.C. Processed. Kerkhof, Paul. 1990. Agroforestry in Africa: A Survey Pimbert, Michel A., and Jules N. Pretty. 1995. Parks, of Project Experience. London: Panos Institute. People, and Professionals: Putting 'Participation' Knutson, R. D., J. B. Taylor, J. B. Penson, and E. G. into Protected Area Management. UNRISD Smith. 1990. Economic Impacts of Reduced Discussion Paper 57. Geneva: United Nations Chemical Use. College Station: Texas A&M Research Institute for Social Development and University. Worldwide Fund for Nature International. Krishna, Anirudh. 1994. "Large-Scale Government Pohl, G., and D. Mihaljek. 1992. "Project Evaluation Programs: Watershed Development in Rajasthan, and Uncertainty in Practice: A Statistical Analysis India." Paper prepared for the new horizons con- of Rate-of-Return Divergences of 1,015 World ference, International Institute for Environment Bank Projects." The World Bank Economic Review and Development, Bangalore, India, November. 6(2):255-77. Processed. Pretty, Jules N. 1995a. "Participatory Learning for Leach, Gerry. 1995. Global Land and Food in the 21st Sustainable Agriculture." World Development Century. Polestar Series Report 5. Stockholm: 23(8):1247-63. Stockholm Environment Institute. . 1995b. Regenerating Agriculture: Policies Liebhardt, W., R. W. Andrews, M. N. Culik, R. R. and Practice for Sustainability and Self-Reliance. Harwood, R. R. Janke, J K. Radke, and S. L. London: Earthscan Publications; Washington, Rieger-Schwartz. 1989. "Crop Production during D.C.: National Academy Press; Bangalore: Conversion from Conventional to Low-Input ActionAid. Methods." Agronomy Journal 81(2):150-59. . 1996. "A Three-Step Framework for Matteson, Patricia. 1995. "The 50 Percent Pesticide Agricultural Change." Pesticides News 32(June): Cuts in Europe: A Glimpse of Our Future?" 6-8. American Entomologist 41(4):210-20. . 1998. The Living Land: Agriculture, Food, McCalla, Alex. 1994. "Agriculture and Food Needs to and Community Regeneration for Rural Europe. 2025: Why We Should be Concerned." Sir John London: Earthscan Publications, Ltd. Crawford Memorial Lecture, October 27. CGIAR Pretty, Jules N., and Rupert Howes. 1993. "Sustainable Secretariat, World Bank, Washington, D.C. Agriculture in Britain: Recent Achievements and Processed. New Policy Challenges." IIED Research Series. . 1995. "Trowards a Strategic Vision for the International Institute for Environment and Rural/Agricultural/Natural Resource Sector Development, London. Processed. Jules N. Pretty 49 Pretty, Jules N., and Parmesh Shah. 1994. Soil and Somers, B. M. 1997. "Learning about Sustainable Water Conservation in the 20th Century: A History Agriculture: The Case of Dutch Arable Farmers." of Coercion and Control. Rural History Centre In Niels Roling and W. A. E. Wagemakers, eds., Paper 1. Reading, U.K.: University of Reading. Social Learning for Sustainable Agriculture. Pretty, Jules N., and John Thompson. 1996. Cambridge, U.K.: Cambridge University Press. "Sustainable Agriculture at the Overseas Steer, A., and Ernst Lutz. 1993. "Measuring Development Administration." Report for Natural Environmentally Sustainable Development." Resources Policy Advisory Department, Overseas Finance and Development (December):20-23. Development Administration, London. Processed. SWCB (Soil and Water Conservation Branch). 1994. Pretty, Jules N., John Thompson, and Fiona The Impact of the Catchment Approach to Soil and Hinchcliffe. 1996. Sustainable Agriculture: Water Conservation: A Study of Six Catchments in Impacts on Food Production and Challenges for Western, Rift Valley, and Central Provinces, Food Security. Gatekeeper Series SA 60. London: Kenya. Nairobi: Ministry of Agriculture. International Institute for Environment and Thai-German Highland Development Project. 1995. Development. "Thai-German Highland Development Project Reus, J. A. W. A., H. J. Weckseler, and G. A. Pak. Annual Report." Chiang Mai, Thailand. Processed. 1994. Towards a Future EC Pesticide Policy. UNDP (United Nations Development Programme). Utrecht: Centre for Agriculture and Environment 1992. The Benefits of Diversity: An Incentive (CLM). toward Sustainable Agriculture. New York. Roling, Niels. 1994. "Platforms for Decisionmaking van der Werf, E., and A. de Jager. 1992. Ecological about Ecosystems." In L. Fresco, ed., The Future of Agriculture in South India: An Agro-Economic the Land. Chichester, U.K.: John Wiley and Sons. Comparison and Study of Transition. The Hague: Roling, Niels, and Elske van der Fliert. 1997. Landbouw-Economisch Institut; Leusden: ETC "Transforming Agricultural Extension for Foundation. Sustainable Agriculture: The Case of Integrated Vereijken, Peter. 1990. "Research on Integrated Pest Management in Rice in Indonesia." In N. G. Arable Farming and Organic Mixed Farming in Roling and W. A. E. Wagemakers, eds., Social the Netherlands." In C. A. Edwards, R. Lal, P. Learning for Sustainable Agriculture. Cambridge, Madden, R. H. Miller, and G. House, eds., U.K.: Cambridge University Press. Sustainable Agricultural Systems. Ankeny, Iowa: Rosegrant, M. W., and M. Agcaolli. 1994. Global and Soil and Water Conservation Society. Regional Food Demand, Supply, and Trade Viederman, S. 1994. "Ecological Literacy: Can Prospects to 2010. Washington, D.C.: International Colleges Save the World?" Keynote address to Food Policy Research Institute. Associated Colleges of the Midwest Conference Sasakawa Global 2000. 1993-95. Annual Reports. on Ecological Education, Beloit College, March Tokyo: Sasakawa Africa Association. 11. Jessie Smith Noyes Foundation, New York. Schaller, N. 1993. "Sustainable Agriculture: Where Processed. Do We Go Now?" Forum for Applied Research Winrock International. 1994. Assisting Sustainable and Public Policy (fall):78-82. Food Production: Apathy or Action? Arlington, Scoones, Ian, and John Thompson, eds. 1994. Beyond Va. Farmer First: Rural People's Knowledge, Wirth, Tim. 1995. "U.S. Policy, Food Security, and Agricultural Research, and Extension Practice. Developing Countries." Presentation of the under- London: Intermediate Technology Publications. secretary of state for global affairs to the Committee Shah, Parmesh. 1994. "Village-Managed Extension on Agricultural Sustainability for Developing Systems in India: Implications for Policy and Countries, Washington, D.C. Processed. Practice." In Ian Scoones and John Thompson, World Bank. 1993. "Agricultural Sector Review." eds., Beyond Farmer First. London: Intermediate Agriculture and Natural Resources Department, Technology Publications. Washington, D.C. Processed. Sherief, A. K. 1991. "Kerala, India: Group Farming." Yixian, G. 1991. "Improving China's Rice Cropping AERDD Bulletin (University of Reading) 32:14-17. Systems." Shell Agriculture 10:28-30. 5 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management in Developing Countries Stein T Holden and Hans P Binswanger This chapter analyzes the decisionmaking of small corrective measures to deal with these problems have farmers in the context of imperfect markets and its to be defended on efficiency, equity, or sustainability implications for efficiency and for sustainability of grounds that incorporate the costs of intervention. natural resource management. We examine the basis A world with transaction costs and imperfect infor- for market imperfections and their structure, perva- mation is almost always constrained Pareto-ineffi- siveness, and implications for behavior (and vice cient (Greenwald and Stiglitz 1986). Unfortunately, it versa). We discuss policies in relation to market is not obvious which interventions are most effective imperfections with some historical hindsight, com- in reducing transaction costs, eliminating or reducing plemented by empirical studies and some theoretical market failures (inefficiencies), and solving equity and applied model studies. We end by discussing and sustainability problems. We prefer to distinguish promising approaches to promoting sustainable between market imperfections and market failures intensification in small-farm agriculture. More and define market failures as a subcategory of market research is required to develop and test these imperfections that implies efficiency losses. In a approaches. world with transaction costs, not all market imperfec- Small farmers represent the majority of the popu- tions may imply inefficiencies because the costs of lation in developing countries. They represent even a correcting the imperfections may be higher than the larger share of t]he population below the poverty line benefits of doing so. Efficiency and equity concems because rural poverty is more extensive than urban can no longer be separated, because redistributive poverty. They also represent a major link between the policies may sometimes be defended on efficiency economy and the environment because their liveli- grounds alone. We relate this to the debate on the rel- hoods depend so directly on the use of land resources ative efficiency of small versus large farners. We ask (soil and vegetation). Small farmers are usually only whether small farmers are too poor to be efficient and partly integrated into markets. Typical market imper- to manage natural resources carefully. fections include missing markets, partly missing mar- Policy failures in the form of neglect and severe kets (rationing, seasonality), thin markets (imperfect taxation of the agricultural sector are largely to blame competition), and interlinked markets. In a world with for poverty, economic stagnation, and decline in many such market imperfections, incorrect or missing price low-income countries. New polices that stimulate signals may accrue from society's perspective and rural development and promote sustainable manage- possibly result in inefficiencies. Possible outcomes ment of natural resources are badly needed. To suc- include the too rapid extraction of and too low invest- ceed, these policies have to build on an understanding ment in natural resources. However, interventions or of the decisionmaking environment of small farmers. 50 Stein T Holden and Hans P. Binswanger 51 Policymakers have to experiment with the behavioral fication and investment in agricultural development. responses of these rural households, but they have a Particularly in Africa, where policy environments limited number of policy instruments. The fact that have been adverse to agriculture, and in areas that are small farmers are rural households that are both pro- poorly integrated into markets, small farmers appear duction and consumption units complicates the analy- to degrade their environment. Population growth is sis, particularly when market imperfections cause seen as fragmenting scarce land resources and leading their production and consumption decisions to be non- to nonsustainable intensification under stagnant tech- separable (Singh, Squire, and Strauss 1986 and de Janvry, nology conditions (Grepperud 1994, World Bank Fafchamps, and Sadoulet 1991). Nonseparability implies 1996, Lele and Stone 1989, and Cleaver and that consumption needs and asset distribution may Schreiber 1994). These rural pockets appear to be in a have significant impacts on production decisions and poverty-environment trap resembling a neo- thus the management of natural resources. Malthusian scenario. They represent one of the largest Schultz (1964) has characterized small farmers as challenges to the developing and industrial world "poor but efficient." Research that has followed has because they can lead to large-scale famines, forced shown that small farmers largely are rational and migration, enormous refugee problems, and war. respond to changes in the set of constraints and oppor- Induced technological and institutional innovation tunities they face. (The transaction cost and imperfect appear to be the best means, together with stable information school sees agents as acting in a way that macroeconomic conditions, of switching from a advances their objectives, given the information and Malthusian to a Boserupian development path opportunities that they have; see Hoff, Braverman, (Hayami and Ruttan 1985). Agricultural research and and Stiglitz 1993.) Inefficiencies may still accrue due extension, infrastructure development, basic educa- to economies of scale, poverty, and subsistence con- tion, and family planning are important elements, straints, unequal access to markets, and various poli- together with land tenure policies and other measures cies disfavoring small farmers. It is thus possible that to provide secure livelihoods. Environmental degra- small farmers are too poor to be efficient from soci- dation necessitates a better integration of agricultural ety's perspective. These inefficiencies may also lead and environmenital policies and calls for a new and to underinvestments in conservation. An important greener green revolution. question is therefore whether small farms are operat- Covariate risk, asymmetric information, and trans- ed less sustainably than large farms. If small farmers action costs lead to pervasive market imperfections. are poor and poverty leads to myopic behavior, this These problems are even more acute in areas with low could be a threat to sustainability. There may thus be population density where the population derives a liv- grounds for interventions from the viewpoints of effi- ing from marginal land resources (Binswanger and ciency, equity, as well as sustainability. Rosenzweig 1986a, Binswanger, McIntire, and Udry The population-poverty-agriculture-environment 1989, and Mclntire 1993). Market imperfections in nexus is complex and subject to much debate these types of areas may not be due primarily to market- (Malthus 1987, Boserup 1965, Lele and Stone 1989, distorting policy failures because they will persist Cleaver and Schreiber 1994, Tiffen, Mortimore, and after the market-distorting policies have been Gichuki 1994, Kates, Hyden, and Turmer 1993, Heath removed. and Binswanger 1996, and Holden and Sankhayan Interlinkage of markets may reduce the costs of 1997). It appears that population growth under certain information and transaction costs associated with conditions may stimulate sustainable intensification, such imperfections (Cheung 1969, Stiglitz 1974, while in others it may lead to land degradation. The Binswanger and. Rosenzweig 1984, Williamson 1985, incentive structures and how these are affected by and Bardhan 1989). Braverman and Stiglitz (1986) policies appear to have a strong impact on whether have argued that interlinkages rnay affect the adoption small farmers are able to choose a sustainable and of new technologies. We may thus look at interlink- welfare-improving development path or are forced ages as endogenous institutional response that may onto a nonsustainable and welfare-reducing one. have the potential to be powerful policy tools for Population growth and higher population densities influencing the adoption and development of technol- may reduce transaction costs and facilitate better mar- ogy (Holden and Shanmugaratnam 1995). Further ket integration when intensification paths are open, research is required to investigate the potential of and market integration may further stimulate intensi- interlinkage, cross-compliance, or cost-sharing poli- 52 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management cies as tools for promoting sustainable natural result may be endogenous village prices that may be resource management. We discuss some of the issues observed in local markets or be shadow prices that here. may not be observed and are intemal to households and various contractual arrangements. Dynamic Market Imperfections in Rural Economies processes as well as policy interventions may change the positions of villages in the diagram. Villages in In this section, we provide an overview of market land-abundant economies (low population density) structures and their determinants in rural economies may be located close to the upper-left-hand corner, dominated by rainfed agriculture, drawing heavily on while population growth tends to create a move in the Binswanger and Rosenzweig (1986a), Binswanger, southeastern direction in the diagram. Wealth accu- McIntire, and Udiy (1989), and Hoff, Braverman, and mulation and specialization usually lead to a move- Stiglitz (1993). The focus is both on material and behav- ment from left to right (increased differentiation), ioral determinants of institutional structures and on par- while a land reform or a progressive tax may cause a ticular markets. We start out by using a rough typology move in the opposite direction. Infrastructure invest- of village economies (Holden, Taylor, and Hampton, ments and pan-territorial pricing may reduce trans- forthcoming), which is illustrated in figures 5.1 and portation costs and the degree of isolation from the 5.2. Village economies are characterized along two outside world for remote villages (downward move- main dimensions, in relation to transaction costs or ment), while the removal of pan-territorial prices and isolation from the outside world and to the degree of reduced investments in infrastructure may lead to an internal differentiation of access to resources. This upward movement (Holden 1997). diagram gives four extreme corner situations, with the The following material conditions contribute to neoclassical model as the best representation when market imperfections (Binswanger, McIntire, and transaction costs are zero or minimal. Local trade is Udry 1989): only relevant when there is an internal differentiation * Dispersed, low-density population in access to resources and when local transaction * High transportation costs costs do not prohibit local trade. With isolation, due to * Seasonal rainfall causing seasonal demand for long distances, poor infrastructure, and poor access to agricultural labor transportation technologies, external markets may * Simple technologies without significant economies become missing or only seasonal in character. The of scale. Figure 5.1. Village Economy Typology Differentiation 0 Low High A High . Isolated village Isolated village No local or Local trade only extemal trade Transaction costs Specialized Neoclassical neoclassical economy economy No local trade with local trade Low Stein T Holden and Hans,P Binswanger 53 The following behavioral and ability assumptions relevant exte:rnalities) and institutional responses are important: to reduce their costs. * Individuals are interested in their own well-being * New sorts of market failures may imply new sorts and value consumption. of policy sol utions. * Individuals have limited access to information * Local institutions may play a significant role in about their environment, and obtaining informa- implementing policies aimed at eliminating mar- tion is costly. ket failures. * Individuals are averse to drudgery above a certain * High weather risk causes covariate crop yields and level of effort. the failure of intertemporal markets. * Individuals face risks and are in general averse to * Poverty and imperfections in intertemporal mar- risk. The degree of risk aversion varies among kets may lead to intertemporal extemalities that individuals and between levels of wealth. require remecial action on sustainability grounds. * Individuals have minimum subsistence require- The government may have a central role to play in ments. stimulating social innovation and cooperative solu- The following implications of the behavioral and tions to natural resource management in rural areas ability assumptions are the most important (Hoff, and in imposing sustainability constraints to protect Braverman, and Stiglitz 1993; Binswanger and the interests of fiuture generations. Rosenzweig 1986a): In the following section, we outline typical market * Asymmetric information leads to problems of imperfections that are found in rural areas with low- adverse selection and moral hazard, which and high-density populations in developing countries, increase the costs of contract formation, monitor- interactions among imperfect markets, nonmarket ing, and enforcement. institutions, and dynamic processes affecting the dis- * Economies in general are not constrained Pareto- tribution of resources and human welfare. efficient. Individual rational behavior does not necessarily lead to socially optimal outcomes. How Imperfect Are Intertemporal Markets? • Allocational efficiency itself depends on the distri- bution of resources. There are sound a priori reasons for supposing that, in * Asymmetric information and transaction costs a low-income rural setting, credit and crop insurance lead to new types of market failures (Pareto- markets will be either highly imperfect or simply Figure 5.2. Typology of Village Economy Models Differentiationi Low High A k High Village CGE models with nonseparable Nonseparable farm household models farn Transaction household costs models Village CGE models with separable farrm household models Separable farm household models Low 54 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management nonexistent (Binswanger and Rosenzweig 1986a). To clients gravitated toward formal sector loans. begin with, there is a problem of moral hazard. It is However, both traditional and modern credit arrange- hard for providers of credit or insurance to screen ments failed to serve producers with few assets. applicants for creditworthiness or moral integrity and Moneylenders were only prepared to lend for the costly to administer credit and insurance extended to duration of a crop season or less, offering small a multitude of small producers. Information is asym- amounts (much smaller than the expected value of metric: clients are able to mislead providers about key output). The takeup of poor farmers by the formal sys- indicators such as crop yields, thus altering in their tem was impeded by the unsatisfactory repayment favor the terms of contracts or the amount of com- record of bank clients (inferior to that of the money- pensation paid. lenders' clients), threatening the solvency of these Collateral can be used to circumvent moral hazard. institutions. The high default rate had many causes, However, farmers may lack assets or their assets may including inadequate screening procedures, scarcity be hard to collateralize (such as movable goods), of adequate collateral, political impediments to the making it less likely that lenders will deem them cred- seizure of assets, and the high level of agroclimatic itworthy. risk in certain regions. Second, high covariance discourages the provision Another indication that intertemperal markets in of both credit and insurance: spatially in relation to these villages are highly imperfect comes from an yield risk and temporally in relation to the sequence experimental study of individual-specific internal of depositing and borrowing. In the event of drought, rates of discount (Pender 1996). Farmers were given a large number of farmers in the affected region will the option of choosing bags of rice today or larger simultaneously default on their loans and submit bags of rice 12 months from today. About one-third insurance claims. Also, because all farmers are bound of the sample always chose rice today, no matter by the seasons to a fairly uniform cropping calendar, what the tradeoff, thereby exhibiting internal rates of deposits and withdrawals from the credit fund will be discount in excess of 100 percent. The mean value of synchronized. To deal with the resulting covariance of internal rates of discount for the other two-thirds deposits, withdrawals, and claims, credit and insur- ranged from 0.33 to 0.41, depending on the precise ance providers must set aside larger cash reserves than nature of the games offered. These internal rates of they would if there were lower covariance between discount were well in excess of the highest interest the borrowing and claiming activities of their clients. rates paid by each of the farmers. The results indi- Crop insurance (other than for very specific risks) cate severe credit rationing. Internal rates of dis- is therefore not commercially viable (Hazell, count were much higher for the poor than for the Pomareda, and VaLdes 1986), and rural credit markets wealthier farmers. are usually poorly developed. Empirical evidence These credit and insurance markets imply that assembled by vil]Lage-level studies of the International inherited endowments are central to the ability of Crop Research Institute for the Semi-Arid Tropics farmers to acquire productive assets out of equity (IcRIsAT) lends additional support to the a priori argu- and to engage in additional borrowing for invest- ments (Binswanger and others 1985 and Binswanger ments and for working capital. Therefore, inherited and Rosenzweig 1986b). The research examines lon- endowments, not just abilities, determine the capital gitudinal data from village households that ICRISAT stock with which an individual has to work, and monitored between 1978 and 1985. The survey vil- inefficiency ensues. lages were selected on the basis of agroclimatic risk to This inefficiency is consistent with the results of allow for controlled comparison of the impact of shal- an investigation of whether poverty is transmitted low soils and unreliable rainfall on the provision of across generations via imperfect markets (Singh and credit. Formal credit facilities were found to be more Binswanger 1988). A comparison was made of the developed in low-risk than in high-risk villages. This assets, income, and inherited goods of two genera- finding was confirmed independently by analyzing tions (valued at 1983 prices), controlling for the level district-level data from 80 districts in India of schooling. The panel format (unlike cross-section- (Binswanger and Khandker 1989). al data) makes it possible to estimate the policy effect In the ICRISAT villages, the growth of bank lending of providing individuals with additional schooling or was associated with a reduction in the relative impor- assets, such as land: unobservable ability variables do tance of moneylenders after the most creditworthy not contaminate the results. The research findings Stein T h'olden and Hans P Binswanger 55 demonstrate that the level of inherited wealth and risks are more likely to invest in longer-distance schooling strongly influences the individual's capaci- migration and marriage arrangements. The research ty to earn and accumulate capital. In contrast, with findings suggest that agricultural technical change perfect markets, the capacity to earn and accumulate may significantly affect spatial marriage patterns by should depend on ability, not inherited wealth. altering the spatial covariance and level of risk, ren- Having established the absence or imperfection of dering the assessment of risk and the establishment intertemporal markets, the next step is to ask whether pro- of implicit risk arrangements more difficult. If formal ducers adopt risk-diffusing strategies and whether credit markets are improved so that they facilitate such strategies offer full compensation for the inade- consumption srnoothing, risk will play a smaller role quacy of credit and insurance markets, in the sense in marital arrangements and rural migration, perhaps that they enable producers to maxirnize expected diminishing rural mobility. profits. Third, the nmembership, size, and composition of Do informal mechanisms for diffusing risk signifi- households and cross-household kinship ties con- cantly reduce consumption risk? ICRISAT village-level tribute further to smoothing consumption and mitigat- studies and other studies have described many of the ing risk. The membership of household units typically ex ante and ex post risk diffusion mechanisms used by spans the generations, allowing knowledge of optimal farmers. These include crop diversification, intercrop- productive packages to be passed down from young to ping, storage of fodder and grain, and sharecropping old. Households entering into new, implicit contracts (Jodha 1978, Walker and Jodha 1986, and Walker and with other households reduce their participation in Ryan 1990). In addition to these practices, there are sharecropping, the previous form of risk protection, other methods of diffusing risk involving capital accu- and increase tlhe extent to which transfers, most of mulation and social relations. which originate outside the village, mitigate the First, storage and savings are significant forms of volatility of intertemporal consumption. The intergen- risk diffusion. The ownership of bullocks emerged as erational and spatial extension of households helps to the preferred means of accumulating "buffer stock," reduce their risk ex post, thereby making them more presenting low risks because these animals are fairly able to bear risk ex ante through the organization of resistant to disease and drought. However, the trans- productive resources (Rosenzweig 1988a, 1988b). portability of livestock reduces its value as collateral, Does the undiffused risk that remains affect con- so the ownership of bullocks does not greatly facili- sumption, farm investment, and its efficiency? The tate borrowing. Also, the absence of a rental market research found that although farmers adopt these mul- for bullocks (owing to the synchronicity of demand tiple risk-mitigating strategies, risk is not sufficiently for their services) prevents the owner from realizing diffused to permit farmers to maintain consumption cash through leasing. These factors account for the steady over tirme and to maximize profits. In other frequent sale of bullocks, which tends to coincide words, the straiegies fail to compensate fully for the with periods of low income and enables households to efficiency-inhibiting impact of nonexistent or imper- smooth consumption over time (Rosenzweig and fect markets for credit and insurance. If risk diffusion Wolpin 1989). is not sufficient, farmers tend to invest too heavily in Second, it is important to consider the impact of inputs that reduce risk and to invest too little in high- female marriage contracts (Rosenzweig and Stark risk inputs. This produces a factor mix that is ineffi- 1989). The researchers hypothesized that marrying cient from the standpoint of profit maximization. daughters to locationally distant but kinship-related The ICRISAT data provide a unique opportunity to households, as is typical in South India, is an attempt test these hypotheses. Rosenzweig and Binswanger to mitigate income risks and smooth consumption in (1994) first shIow that individual-specific income an environment characterized by high information shocks do not affect food consumption of households. costs and spatially covariant risks. The ICRISAT vil- However, covariate village-specific income shocks do lage data indicate that marriage and migration signif- affect household food consumption. This suggests icantly reduce the variability of a household's food that the informal mechanisms for risk diffusion are consumption because they pool income between insufficient to fully diffuse risks, which therefore are households in different regions and the risks attached likely to spill over into investrnent behavior. to the various sources of income are not highly To test this hypothesis the authors estimate a prof- covariant. Households exposed to higher income it function, usinig a flexible functional form, incorpo- 56 Small-Farner Decisionmaking, Market Imperfections, and Natural Resource Management rating nine capital inputs, two rainfall measures, and India concludes that risk aversion varies inversely with personal characteristics such as family composition, wealth. Poor farmers are less able to cope with uncer- age, and education. The estimation procedure exploits tain weather. One of the most important findings is the panel data and allows for the possibility that that, contrary to Schultz (1964), poor farmers are not unmeasured farmer attributes and abilities influence able to maximize profits: specifically, in the ICRISAT the composition and level of investment. For each of villages, for every rupee invested, poor farmers obtain the investment inputs, the marginal contributions to 30 percent less than they would if they had been able profits and riskiness of the portfolio are computed. to adopt a profit-maximizing portfolio (Binswanger The test statistics (Binswanger and Rosenzweig and Rosenzweig 1984). Thus poor farmers are too 1989: table 3) show that the investment portfolios of poor to be efficient. poor farmers are not allocated to maximize profits. Indeed, uninsured weather risk is a significant cause Small Farms, Efficiency, and Sustainability of lower efficiency and lower average incomes: a decrease of 1 standard deviation in the timing of the Efficiency and equity discussions have been at the rainy season would raise average profits by up to 35 heart of the analysis of interlinked markets or contracts, percent among farmers in the poorest quintile. while sustainability issues rarely have been addressed. Wealthier farmers, in contrast, are able to absorb risk, and the profitability of their investment is not Interlinkage and Efficiency affected by an increase in riskiness. The results indicate that improving the ability of The theories of interlinkage emerged from efforts to poor farmers to smooth consumption, perhaps via understand the persistence of share tenancy, which increased consumption credit, would increase the Marshall (1920) claims is an inefficient institutional profitability of their agricultural investments. arrangement because the tenant receives only a share of Similarly, providing insurance against failure of the his marginal benefit but has to cover the entire margin- monsoon, if feasible, would raise the overall level of al cost. The works of Cheung (1969) and Stiglitz profit and decrease the inequality of earnings in high- (1974) claim that share tenancy may be a superior solu- risk areas. The research also demonstrates that differ- tion to fixed-rent and owner-operator (with hired labor) ences in investment portfolios across wealth classes contracts because it provides a combination of risk- do not arise from technical scale economies. Thus dif- sharing and incentive effects that represent a compro- ferences in the composition of investment principally mise solution. reflect differences in the capacity of farmers to absorb Interlinked contracts may implicitly provide credit risk and differences in objective risk (variable weather). and insurance to tenants in environments where The investment portfolio also reflects the liquidity credit and insurance markets are imperfect (Otsuka, of the assets, that is, their usefulness in permitting the Chuma, and Hayami 1992). Cost-sharing arrange- farmer to compensate for unforeseen losses. In partic- ments may relieve the working capital constraints of ular, farmers frequently sell bullocks in the event of tenants. Nevertheless, adjusting for quality differ- drought-induced c:rop failure. In South India, the fre- ences in plots, farmers in the ICRISAT villages had a quent selling of bullocks (to permit consumption 16 percent lower efficiency on their sharecropped smoothing) means that average bullock stocks are typ- plots than their own plots. For further review of ically below the level necessary to ensure maximum the empirical literature, see Otsuka, Chuma and profitability and efficiency offarms. Simulations based Hayami, 1992. on parameters estimated from a dynaric structural Share tenancy represents primarily the interlinkage model applied to the ICRISAT data demonstrate that of labor and land markets (and the risk and insurance augmenting sources of off-farm income (such as wel- market) but may also include credit and input markets fare payments, rural industry, and public works pro- (with a choice of technology implications). Other grams) reduces the farmer's need to sell bullocks and examples of interlinked markets include credit and boosts the stock of bullocks, which could raise agri- land (collateral) markets; credit, input, and output mar- cultural productivity (Rosenzweig and Wolpin 1989). kets (often government-sponsored programs); labor To sum up, poor farmers are less likely to diffuse and output markets (food-for-work contracts), which risk successfully because they have fewer assets for often provide insurance as well (food security); and collateral and fewer "buffer" stocks. The research in simple barter trade (linking two commodity markets). Stein T hIolden and Hans P Binswanger 57 Small Farms and Efficiency close coordination of harvesting, processing, and mar- keting have led to the development of plantations for A common belief is that small farmers eventually will sugarcane, tea, oil palm, and banana. However, con- become either landless workers or commercial farm- tract farming may eliminate the link of farm produc- ers and thus only represent a temporary mode of pro- tion with the processing plant (where the economies duction (Ellis 1988 and Sadoulet and de Janvry 1995). of scale are located) via joint ownership in a planta- This belief has led to an overemphasis on the devel- tion. Examples are sugarcane in India and Thailand, opment of large mechanized farms and collective tea in Sri Lanka, and oil palm in Indonesia. The plan- farms. However, these attempts to benefit from tation mode of production has therefore declined economies of scale frequently have been countered by sharply at the expense of contract farming with small- large moral hazard problems. holders (Binswanger, Deininger, and Feder 1995). Even though we have seen in the last section that Rosenzweig and Binswanger (1994) showed that in small farmers have great difficulties to diffuse covari- the ICRISAT villages there are no technical economies ate risk in poorly developed intertemporal markets, of scale at the farm level. They also showed that, many studies indicate that small farms may be more despite their inferior risk diffusion capacity discussed efficient than large farms. Few of these studies have in the last section, small farmers have a higher profit used a comprehensive total factor productivity rate than large farmers, indicating no productivity approach for the estimation. Some studies in develop- advantages. Indeed, together with the inefficiencies ing countries hold a higher methodological quality associated with sharecropping, these results indicate (Berry and Cline 1979 and Prosterman and Riedinger that a redistribution of assets (mainly land) could 1987). They also show an inverse relation between increase productivity. farm size and productivity. Differences in land quali- However, the empirical literature also suggests ty and use of capital cannot explain all of the negative that there may be economies of scale at very small relationship. Moral hazard problems favor family farm sizes and diseconomies of scale at larger farm farms over large farms depending on hired labor. sizes. Thus excessive fragmentation could be harmful Berry and Cline (1979) find that the second smallest for efficiency. If population growth is pushed as far as class of farm, which includes the smallest full-time to lead to excessive fragmentation and poverty, popu- farmers, has the highest output per unit of area. lation growth may be bad for efficiency, leading The optimal size of family farms may have to a population-poverty trap and reinforcing the increased as a result of economies of scale due to Malthusian scenario. This outcome could be prevented mechanization, an increase in the acceptable level of by further intensification, diversification to high- family income, and a fall in the price of farm output, income crops, off-farm opportunities (market integra- which could be avoided by engaging in some off- tion), migration, income-increasing technologies, and farm activities. Cash scarcity, lack of insurance family planning. against production and market risks, poorer access to credit markets, higher interest rates, and indivisibili- Small Farms and Sustainability ty of input and capital investments may decrease the use of inputs and increase the use of less-efficient Missing credit markets in combination with poverty technologies on small farms. Imperfections in more and risk may cause poor farmers to have very high than one market are required to obtain the inverse discount rates, as shown by Pender (1996) in India relationship between farm size and productivity. and Holden, Shiferaw, and Wik (1998) in Ethiopia, Srinivasan (1972) has shown that a combination of Indonesia, and Zambia. Pender finds that as much as missing land (rental) and insurance markets under one-third of the population has discount rates above conditions of risk and risk-averse producers (with 100 percent. Holden, Shiferaw, and Wik find average nonincreasing absolute risk aversion and nondecreas- rates of 93 percent among transmigration settlers in ing relative risk aversion in wealth) will yield the Seberida, Sumatra, 105 percent among peasant house- type of inverse relationship observed. Similarly, holds in northern Zambia, and 53 percent among Feder (1985) and Carter and Kalfayan (1989) have small farmers in a high-potential cereal zone in the shown that imperfections in credit and labor markets highlands of Ethiiopia. Average incomes per capita are may yield similar results. $107 in Sumatra, $108 in Zambia, and $196 in Economies of scale in processing and the need for Ethiopia. Shortage of cash appears to have a strong 58 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management impact on the discount rates in Sumatra and Zambia, Does improved market access make small farm where the average income levels are lowest. Wealth in production more sustainable? The vast literature on the form of livestock and total income is most strong- Boserup effects clearly shows that this is not gener- ly correlated (negatively) with the discount rate in ally the case. Higher profitability from market access Ethiopia. tends to induce more land investment. For, example, These studies have shown a remarkable consisten- improved market access may create incentives for cy across sites, vvhich supports the hypothesis that intensification by encouraging farmers to switch poverty combined with liquidity constraints leads to from systems with low to systems with high external high discount rates. This implies that poverty reduc- inputs. An example is the promotion of maize pro- tion itself or profit-increasing reform may reduce the duction in Zambia, which reduced the extent of intertemporal externality, because the discount rates shifting cultivation (Holden 1991, 1997). of the poor will decline as incomes increase. However, in some cases, improved market access However, in locations where severe environmental leads to more rapid deforestation and extraction of degradation is taking place, the high discount rates forest products due to higher output prices. Under found in these studies point to the need for additional conditions of uncertain or missing tenure rights, interventions to stimulate conservation. improved market access may initiate more clearing What are the implications of high discount rates of land as a way to obtain property rights to land for investment in conservation technologies? This (homesteading; Angelsen 1997). The immediate intertemporal externality may create incentives for effect may thus be more deforestation. nonsustainable resource extraction as a short-term survival strategy. Models of Farm Households Under conditions of poor social organization and/or adverse po:licies, population pressure may also Further insight on these issues can be gained by draw- lead to encroachment into and production of food ing on modeling efforts. Chayanov (1966) was the crops on fragile steep slopes, which may be rapidly first to develop farm household models. He did this degraded. Livestock may serve as insurance, causing for Russian rural society, which was characterized by farmers to build up their herds beyond the carrying land abundance and no labor market during the 1920s. capacity of the land, leading to overgrazing and ineffi- His models later were formalized into a neoclassical cient use of grazing lands. theoretical framework by Nakajima (1986), who has Are small farms more efficient in some cases also developed models for a number of specific mar- because they are rriore intensively run, making higher ket structure settings. efficiency no more than a rapid mining of the soil and The neoclassical farm household model developed vegetation? The Indonesian rubber farmers with small by Barnum and Squire (1979) is a separable model plots and few trees have managed to increase their that contains only one missing market (for land) and rubber production per tree significantly in the short ignores intertemporal markets (credit, risk and insur- run by tapping their rubber trees as frequently as ance, and investment decisions). Production decisions every day despite the recommendation to tap young can still be treated as independent of consumption trees only twice a week (Holden, Hvoslef, and decisions. Simanjuntak 1995). However, in the chitemene shift- The farm household model becomes nonseparable ing cultivation system in northern Zambia, the level of in the presence of a missing labor market, significant income or material welfare depends directly on the risk (and lack of insurance), credit market imperfec- extent of cultivation by the individual household. tions, and/or binding subsistence requirements. With Labor-rich households therefore have higher incomes nonseparability, a farm household approach becomes but also cause the most deforestation when the car- essential for policy analysis because production deci- rying capacity of the system has been exceeded. sions depend on the consumption side of the model Female-headed households and labor-poor house- and vice versa. Endogenous shadow prices link the holds require only one-third to half of the area per sides, and fewer qualitative results may be drawn capita required by labor-rich households (Holden unambiguously from theoretical models. 1991). In this case, better market access or higher De Janvry, Fafchamps, and Sadoulet (1991) have income could imply more rapid deforestation. developed a general farm household model with miss- Stein T Holden and Hans P Binswanger 59 ing markets. They have developed a numerical model credit. Historically, peasants in northern Zambia have for a typical African peasant household that simulates been very responsive to the possibility of producing situations with missing markets for food and labor, this cash crop if they are given access to input and labor only, food only, and no missing markets. They credit markets. The introduction of structural adjust- illustrate the effects of a price change for a cash crop ment policies removed the input, output, and credit in each of these four cases. They show that the supply subsidies and the implicit transportation subsidies of response for the cash crop is substantially reduced in the pan-territorial pricing system. This re.,ulted in a the case with market imperfections. The reduction is less favorable ratio of fertilizer/output prices for peas- greatest in the absence of both labor and food crop ants in remote locations in northern Zambia, reducing markets. The reduction is less when only the labor the price of the cash crop. The model is able to trace market is missing because households can substitute the consequences of any policy change to the expan- from the production of food crops to the production of sion of shifting cultivation of food crops, which can cash crops and purchase more of their food. With a lead to deforestation and land degradation, if carrying missing food crop market, this substitution is much capacity has been exceeded. less possible, while hired labor may substitute for fam- The model is then used to simulate this price ily labor in the production of cash crops. response. A 1 0 percent reduction in the price of maize De Janvry and others (1992) use nonseparable reduces the profit in maize production by 16-19 per- household models for small and medium farms in cent, reduces marketed surplus of maize by 2-3 per- Morocco to simulate the effects of an agricultural cent, increases shifting cultivation by 2-6 percent, structural adjustment program in a case with a missing increases marketed surplus of food crops by 14-44 market for child labor and milk and a credit constraint. percent, and recluces full income by 1-4 percent. This Children are used to herd animals in communal graz- occurs under the assumption that there is no change in ing lands. The results show that the credit constraint the rationed supply of credit, implying no change in severely limits the ability of farmers to hire labor, rent fertilizer use. The privatization of transportation costs machinery, and buy fertilizer and thus lowers the sup- may thus result in more extensive cultivation prac- ply response in agricultural production. The adjust- tices and more deforestation and land degradation in ment program favored the production of crops over remote regions. These model results compare well livestock, which reduced the production of milk and with empirical findings (Holden 1997). beef. Rising forage prices, however, caused a substitu- Some recent. theoretical developments have signif- tion into more communal grazing and more use of icance for how risk may influence the behavior of child labor, with increased overgrazing and school small farmers. Sandmo (1971) has established that a absenteeism as negative side effects. This is an exam- risk-averse firm facing output risk will produce less ple of how market imperfections can contribute to output than a risk-neutral firm. Small farmers have unpredictable and paradoxical effects. been found to be risk-averse (Binswanger 1980 and Holden, Taylor, and Hampton (forthcoming) have Binswanger and Sillers 1983). Credit market imper- developed a village economy model for a typical fections and the inability of poor farmers to pool risks remote village in northern Zambia, consisting of two across time may be complementary explanations for independent nonseparable farm household models high discount rates and lower use of cash inputs representing two groups of households, one more (Binswanger 1980, Binswanger and Sillers 1983, market oriented than the other. Unlike the model of de Eswaran and Kotwal 1990, and Morduch 1995). Janvry and others, these models have a credit con- Risk has also been used to explain the heterogene- straint, the cash crop (maize) is partly consumed by ity of production (crop diversification) that may the households, and the food crops (millet, ground- reduce overall production risk. However, other expla- nuts, cassava, and beans) are produced in an extensive nations, like the need for seasonally more even distri- system of shifting cultivation and are partly marketed. bution of labor, heterogeneity of soils (Bellon and Credit institutions provide formal credit for the pur- Taylor 1993), and commodity market imperfections chase of fertilizer and improved seeds of maize and combined with heterogeneous demand for food (com- require that the maize be marketed through them plementary commodities), may be important as well (interlinkage of input, credit, and output markets). (Holden 1991 and Wik and Holden 1996). The model is used to analyze the responses to changes Finkelshtain and Chalfant (1991) have argued that in cash crop price, fertilizer price, and provision of a peasant household that consumes part of its produc- 60 Small-Fanner Decisionmaking, Market Imperfections, and Natural Resource Management tion and faces output price risk really faces multivari- dictions about behavioral responses to policy inter- ate risk because the price of a consumption good is ventions. random. They show that such a household, which is Most dynamic models of soil degradation and con- risk-averse, may even choose to produce more than a servation have assumed separability (McConnel risk-neutral profit-maximizer when the price is ran- 1983, Barbier 1990, LaFranche 1992, and Shiferaw dom. Sandmo's qualitative result therefore does not 1996). As noted by McConnel, when markets are per- always hold for marketed surplus-producing house- fect, the private intertemporal path for soil use will holds. Sandmo's result does hold if the good produced mimic the socially efficient path. Thus there are no is a normal good. However, if the good is an inferior reasons to be concerned with land degradation unless good, the household may increase its production with some market imperfections are Pareto-relevant or rel- increasing risk. For a net buyer of food, an inferior evant to social welfare for other reasons. good will be produced at higher quantities under risk Pagiola (1995) has developed a dynamic farm than under certainty. For a normal good, the Sandmo household model with subsistence requirements and result may or not lhold for net buyers. uses it to explore the conditions under which subsis- Wik and Holden (1996) have tested the relevance tence requirements prevent the adoption of sustain- of the Finkelshtain and Chalfant result for peasant able land use practices. He discusses four cases and households in northern Zambia. Adjustment policies shows that under certain conditions (severe poverty), imply a removal of price controls for maize, the main poor farmers may have more incentives to adopt sus- cash crop there, and this results in more variable tainable practices than other farmers because the prices. Because maize is found to be a normal good future disutility of degrading the resource is poten- and most producers are net sellers and are risk-averse, tially unbounded. the Sandmo result should hold; the increase in risk as Conversely, he also looks at the case where sub- a consequence of the introduction of the adjustment sistence constraints prevent farmers from adopting policy (assuming no change in average relative sustainable practices. Farmers may be too poor and prices) should resualt in lower production of maize. lack the access to credit markets that would enable They also show that the increase in price risk is like- them to make the necessary investments. When sus- ly to lead to more credit rationing because it will lead tainable practices do not exist or are unattainable, to higher default rates and lower expected returns to poor farmers without migration options degrade the the lenders. This has been confirmed by Tviland soil more slowly than households with migration (1996), who finds that, in a sample of 161 households, options. the share of peasants who received formal credit Holden, Taylor, and Hampton (forthcoming) have dropped from 56 percent in 1993 to 14 percent in also shown that households with high (constant) dis- 1995. This is not explained by a change in credit count rates may save and invest more to secure future demand, as approximately 60 percent of the peasants survival as they become poorer but that their level of applied for credit in both years. The default rate for savings and investment is low at higher income levels. those who received credit increased from 15 to 35 However, if the discount rate is increasing with falling percent from 1994 to 1995. (However, this may also income, this pulls in the opposite direction. be a result of a collapsed credit system that did not Households may reach a switch point at which they consist of rational ]lenders.) see no hope for survival in the future. Grepperud (1997) has developed a dynamic model Models of Deforestation and Land Degradation of soil depletion choices under production and price or Conservation uncertainty. It is a model with one market imperfec- tion only (risk and insurance). The model has three Much of tropical deforestation is caused by agents control variables; productivity increasing but degrad- other than small farmers as shown by FAO (1996, cited ing input, conservation but productivity-reducing in chapter 22 of this volume). Nevertheless, labor input, and "win-win" input, which both increases pro- constraints, cash constraints, limited access to input, ductivity and conserves the soil. The model illustrates credit, and output markets, and subsistence require- why risk and risk aversion have unclear effects on ments may be highly relevant for the decisions of conservation (Anderson and Thampapillai 1990 and small farmers to clear forests. Models that ignore the Ardilla and Innes 1993). Detailed knowledge of the resulting nonseparabilities may fail to give good pre- farming systems and sources of risk is required to pre- Stein T holden and Hans P. Binswanger 61 dict the direction of effects. One of the conclusions of force, nevertheless indicating imperfections in labor the model is that if output price is uncertain, risk aver- markets. The study does not analyze how investment sion may induce farmers to use less of all inputs. per unit of area changes with farm size, however. The Another is that uncertain production may induce farm- study finds that conservation investment increases ers to use fewer degrading inputs and more conserva- with debt, which may be a sign that those with low tion inputs. Thus production uncertainty may be more debt are credit constrained. Savings affect conserva- favorable for conservation than price uncertainty. tion investments positively in only one of the villages, Pender and Kerr (1996) have developed a theoret- indicating some credit market imperfections. ical model for farm household conservation decisions Shiferaw and Holden (1997b) have developed an with market imperfections. It is a two-period model applied farm household model incorporating the incorporating several ca.egories of assets. They ana- user's cost of land and the conservation decisions of lyze the four cases: perfect markets, a missing labor farm households in the Ethiopian highlands. The market, a missing credit market, and both labor and model includes subsistence requirements, liquidity, credit markets missing. They also have a category of and credit market constraints and is thus a nonsepara- nontradable assets. They assume a strongly concave ble farm household model. Short- and long-term production function with all cross-partial derivatives responses are analyzed for cases when conservation as non-negative and a strictly concave utility function. technologies reduce, make no difference, and increase They derive comparative statics results and show that yields. Low (10 percent) and average (53 percent) dis- conservation investments are independent of labor count rates are used to estimate user costs of land and endowments and savings but may be influenced by then are compared with results when user costs are the level of fixed assets and the initial level of con- ignored. The results indicate that conservation tech- servation investments. With a missing labor market, nologies will be adopted when they do not negatively investment in conservation will increase with the affect yields. (See also chapter 16.) However, if they labor endowment of households. The impact of other have a negative effect on yields in the short run, categories of assets is negative or ambiguous. With a which is usually assumed to be the case in the missing credit market, investment in conservation Ethiopian highlands, these technologies will not be will increase with the household's amount of saving. adopted unless the farmer has a low discount rate (10 Effects of labor endowments and other assets are percent) or is provided incentives to reduce his short- ambiguous. If both labor and credit markets are miss- term costs or increase his short-term benefits. If the ing, conservation investment will still increase with social rate of discount is low (less than 10 percent) savings, while the impact of labor endowments and compared with the private rate (53 percent), this is an other categories is ambiguous. intertemporal externality that calls for intervention. Pender and Kerr (1996) have conducted an econo- Shiferaw and Holden (1997a) have used the same metric study of conservation investment decisions in model as a baseline for testing the efficiency of various three (ICRISAT) villages in India. They test for the pos- policy measures to reduce land degradation and increase sible impact of market imperfections using their own conservation investments. We return to the results of theoretical model and find that imperfect land markets this exercise when we discuss various policy measures cause farmers to invest less in conservation on leased to promote sustainable natural resource management. land in two of the study villages (Aurpalle and Shirapur). Owners invest more in their land than tenants Policy and Market Imperfections do. Indian land tenure laws, which undermine long- term leases, may explain this effect, because a long-term Ideology had a considerable impact on policy design lease would give the tenant right to the land. In two of in the past. Leftists thought that markets were "evil the villages (Aurpalle and Kanzara), conservation capitalistic creatures" that should be condemned and investments decrease with farm size, while in the third replaced with more fair systems of exchange and village (Shirapur) conservation investments increase sharing. The influence of leftist ideologies or of rent- with farm size. These effects may also be signs of land seeking behavior in developing countries caused the and other market imperfections, leading to endoge- implementation of policies aimed at replacing mar- nous land prices at the farm level. Conservation kets to ensure a more fair distribution but neglecting investment increases with the adult male work force the implications for efficiency. The neglect of behav- but, strangely, decreases with the adult female work ioral conditions undermined not only efficiency but 62 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management also equity because the policies frequently had a itive effects on security of ownership or access to strong urban bias and reinforced rural underdevelop- credit. Feder (1993), however, has found that in ment. Parallel "black markets" developed and Thailand land titles have a positive effect on access to allowed the elite and corrupt persons to make fortunes credit, since land can be used as collateral in a well- at the expense of the poor. developed credit system. By facilitating access to There are also many examples of well-intended credit, land titles have had a positive effect on the but unsuccessful interventions to promote agricultur- level of investment and output and on market values al development, including price, credit, input, market- of land. ing, and research and extension policies. These unfor- Can we trust that removing policy distortions will tunate developments were partly due to a lack of solve the problems? Will first-best policies work in a understanding of the functioning of public institu- second-best world? There is no straightforward tions. Improper incentives and control mechanisms answer to this. The need for complementary policies resulted in problems related to moral hazard, corrup- has to be assessed in each case. High transportation tion, and growing rather than declining transaction costs, poor infrastructure, inelastic demand, droughts, costs. With the economic decline and debt crisis in the and poorly developed information systems have early 1980s, the role and functioning of the state were added to the difficulties of designing good food poli- assessed more critically, and state activities were reduced cies. Removing price controls may thus increase the and reorganized. output price risk, and this may have adverse effects on Some policies were aimed at correcting market efficiency and the functioning of credit markets (Wik imperfections, which were believed to represent mar- and Holden 1996). Investments in infrastructure and ket failures, but more frequently interventions were systems for storage may improve both food and cred- made on grounds other than efficiency. Interventions it markets. to establish credit markets where they were missing or where interest rates in informal credit markets were Policy and Project Design Implications very high may have been implemented on efficiency as well as equity grounds. In many cases, they were The research in India suggests that the concern of pol- not successful. Price regulation in food crop markets icymakers in developing countries, the World Bank, was another key arena for interventions with limited and the rural development community at-large with success, often causing severe consequences such as providing access to crop insurance and credit for poor stagnant agricultural and rural development, farmers not only has an equity basis but also may be increased rural-urban migration, and huge efficiency justified on efficiency grounds, at least in the context losses. This, in addition to the debt crisis, economic of semiarid India. The problem is that almost all past stagnation, and decline facing many developing coun- public sector programs failed (Hazell, Pomareda, and tries, is the context in which adjustment policies were Valdes 1986). For small-farmer credit, the number of implemented beginning in the early 1980s. successful cases is very limited as well. But failing to State property rights to land have frequently resulted provide poor farmers with access to more efficient in de facto open access and degradation of natural forms of savings may have a cost. Paying financial resources by undermining customary rights and informal intermediaries a fee for serving these groups may be cooperative management systems. These traditional amply justified to restore efficiency. Such a scheme systems often provide security of access to land. should, of course, not subsidize interest rates because Institutionalization of private property rights has not that would provide perverse incentives to serve large succeeded in many developing countries due to the farmers. Instead, it might be better to pay institutions high costs of implementation and maintenance. The for rendering specific services to small farmers, irre- introduction of forrmal title has not improved land pro- spective of the size of transaction. Further experimen- ductivity in Africa (Migot-Adholla and others 1993), tation and research are clearly warranted. and the absence of formal title is therefore not con- Although failures in the credit and insurance mar- sidered a constraint to development. Rather, there is kets appear to be at the root of many inefficiencies, evidence that customary rights regimes in Africa are this does not imply that the second-best intervention becoming more privatized as population pressure and is to be found in these markets. It could be that the commercialization grow. The titling program in moral hazard and covariance problems that led to the Kenya has been expensive and had no significant pos- market failure in the first place cannot be overcome Stein T Holden and Hans P2 Binswanger 63 directly with a public policy intervention, however and nonmarket institutions because they were intro- well designed and well intended. Instead, research on duced before the major developments in the econom- other policy instruments should take account of the ics of rural organization. possible impact of more stable income or consump- By drawing on these new theoretical develop- tion on investment efficiency. ments, it may be possible to identify good ideas for Alternatively, with respect to credit and insurance, successful institutional innovation and to reduce the there may be a case for redistributing assets to enable number of counterproductive interventions. For the poor to obtain loans and to insure against income example, the use of group lending and peer monitor- fluctuations. However, land reform poses severe ing to control moral hazard problems in credit mar- financing problems of its own (Binswanger and Elgin kets is shown to be justified by this literature. The 1990). successes of the Grameen Bank and Aga Khan Rural The research results suggest that it often makes Support Programme have shown that it may be possi- sense to provide poor producers with transfers (food ble to reach poor people with credit and retain low subsidies) or publicly sponsored employment (food- default rates. Otier lessons include the desirability of for-work programs). So far, such measures have been communal management of local commons. There evaluated by looking at the consumption and nutrition often are good grounds for cooperative solutions benefits derived by beneficiaries. Much concern has when the group is fairly small and homogeneous, been expressed about the disincentives to work creat- mutual trust ex.ists, communication and monitoring ed by such transfers and the diversion from more opportunities are good, a system of responsibilities remunerative employment that may follow from job and sanctions is in place, and cooperation is like a creation schemes. No consideration has been given to repeated game (Dutta, Ray, and Sengupta 1989, the possibility that providing the poor with more Ostrom 1990, and Balland and Plateau 1996). secure incomes may enable them to make better deci- sions regarding the formation of both physical and Interlinked Markets and Contracts human capital. As much of the research surveyed in this chapter Holden and Shanmugaratnam (1995) have forwarded has amply demonstrated, questions concerning the the idea that interlinkage mechanisms may be power- role of financial markets and risk in influencing the ful tools for promoting sustainable agricultural devel- efficiency of agricultural investment and productivi- opment. They use the example of land degradation ty can best be addressed using panel data. It is nec- linked to poverty and missing or imperfect markets to essary to follow the same household over a relative- argue that this may represent a Pareto-relevant exter- ly long period in order to detect fluctuations in risk, nality. Poverty may prevent poor farmers from invest- in asset holdings, and in the pattern of investment. It ing in land conservation due to imperfections in cred- is therefore regrettable that the ICRISAT data set it markets and high subsistence requirements. Poor (which surveyed the same households frequently farmers may use agricultural practices that cause large over a 10-year period) should remain the only one of net losses in the stock of soil nutrients. Under such its kind. conditions, one may argue for a Pigouvian subsidy on fertilizers to recduce the nutrient loss externality that Interlinkage and Induced Innovation would negatively affect future productivity. However, a fertilizer subsidy alone may not be sufficient to Agricultural development has been promoted through internalize the externality because cheap fertilizers the provision of interlinked input (fertilizer and seeds) may become a substitute for good management and credit packages, which efficiently promoted the (Repetto 1987 and Low 1986). But if a subsidy is used adoption of technology (green revolution) in many in relation to an, interlinked contract where the farm- cases. These efforts were efficiency-enhancing ele- ers receive a combination of fertilizer, high-yielding ments of induced technological and institutional inno- seeds, and eventually credit, if they implement certain vations (Hayami and Ruttan 1985). They followed or conservation practices (terracing), it may be possible were accompanied by public investments in agricul- to internalize a larger share of the externality because tural research and extension, infrastructure, and mar- this package will ensure better nutrient recycling and keting development. However, they were not based on more efficient use of nutrients. a thorough analysis and understanding of rural market The problem may not be solved by introducing a 64 Small-Fanrer Decisionmaking, Market Imperfections, and Natural Resource Management credit market, however, because improved access to Targeted Credit in Surplus-Producing Areas credit may not lead to more sustainable resource extraction and investments in conservation. Credit Targeting of credit for productivity-increasing inputs may also be used to increase resource extraction and and conservation investments is likely to function best consumption in the short run. Targeting credit to in surplus-producing areas because the activity investments in conservation may be preferable, but financed must produce a surplus that has to be used in not without prob]Lems (Holden and Shanmugaratnam part to pay back the loan. It is more likely to succeed 1995). where production and market risks, probability of Shiferaw and Holden (1997b) use an applied non- leakages, and monitoring costs are low. This suggests separable household model from the Ethiopian high- it may be difficult to reach poor, deficit-producing lands to investigate the plausible responses to various farmers, living in risky environments with poor mar- policy measures, including seed, fertilizer, and labor ket access. The state or implementing agency may subsidies (cost sharing) linked to the installation of have to be prepared to absorb a large share of the costs conservation structures on erodible land. These pack- in such areas. If successful, such programs may, how- ages were designed for cases in which the conserva- ever, change farmers and societies from subsistence tion technology alone had a negative effect on yields and deficit-producing to surplus-producing, thus (because it took up space and increased pest prob- reducing the need for outside support. Where policies lems) or had no effect on yields (but required addi- are not the main reason for the poverty and degrada- tional investments, primarily labor). It has been found tion trap, a temporary cost-sharing scheme may thus that input subsidies linked to conservation invest- break the vicious spiral of poverty and environmental ments can increase the adoption of conservation tech- degradation and lift society onto a sustainable devel- nologies even when conservation technologies reduce opment path. Family planning and slower population yields by 20 percent. However, the subsidy rates have growth may be necessary to sustain the improvement. to be very high. T]he benefit/cost ratio is less than 1 for It is usually assumed that making preventive a fertilizer subsidy (75 and 90 percent) linked to con- investments to avoid land degradation is cheaper than servation, while it is higher than 1 (1.61 and 1.66) for rehabilitating degraded lands. Yet it is possible for a seed subsidy (90 and 100 percent) for improved rehabilitation to be more realistic if no interventions to seeds of teff, assuming that the organizational costs of promote conservation have been made. Population intervention are 10 percent of the costs of direct sub- pressure increases the value of land and may thus sidy. However, the seed subsidy initiates the conser- induce incentives to rehabilitate degraded lands where vation of only a. small share (12 percent) of the lower population pressure (and land values) does not erodible uplands. provide sufficient incentives for undertaking preven- If the conservation technologies have no negative tive conservation investments. Credit constraints, high impact on yields in the short run, lower subsidy rates private discount rates, and other land market imper- are sufficient to create incentives for adoption, and fections (insecure tenure) may be other explanations. both fertilizer subsidies and seed subsidies may give positive benefit/cost ratios although the seed subsidies Food-for-Work Programs in Deficit-Producing Areas give the highest ratios. Linking fertilizer and seed subsidies with conser- Food for work is a common policy instrument. It may vation requirements yields positive adoption responses be seen as linking labor and food markets and as pro- and benefit/cost ratios above 1 at lower subsidy levels viding poor unemployed people with employment and when conservation investments reduce yields by 20 food security, while contributing to economic growth, percent. The model is also used to test the response to if the work represents productive investments. If such taxation (20 percent) of the erosive crop (teff) and investments are made to improve the productivity or subsidization of soil-improving crops (pulses). the resource base in low-resource economies, they Taxation of teff reduces the teff area considerably may contribute to efficiency as well as equity (if the (more than 50 percent) and reduces soil erosion by 15 retums to investment are higher than for alternative percent, but this policy instrument does not create uses). The negative environmental externalities of not incentives to implement other conservation tech- intervening may strengthen the basis for intervention. niques and has a severe, negative effect on household So do negative extemalities in the form of famines, income (-11 percent). forced migration, social conflicts, and human suffer- Stein T Hlcden and Hans P. Binswanger 65 ing, which may be reduced by providing food assis- Conclusions tance and employment to drought- and poverty- stricken societies (von Braun 1995). These programs Standard tools or rules of thumb from environmental work best if they set the wage rate (or payment in economics or imicroeconomics in general may not kind) low to ensure targeting of the poor (self-selec- always hold when applied to small farmers. Taxing tion). the output from their land-degrading activities (the polluter-pays principle) may induce poor small farm- The Importance of Local Participation ers to increase their land-degrading activities. Such taxation may worsen their poverty and cash liquidity Local participation may have several meanings and constraints and reduce their ability to invest in con- levels, ranging from forced participation to self-mobi- servation. Policies aimed at the input side are more lization with many intermediate forms. Active partic- likely to have the expected effects. Interlinkage and ipation and broad collective mobilization provide a targeting policies may be useful in resource-poor good basis for successful programs for natural remote areas and where poverty causes nonsustain- resource management. The local leadership and com- able resource management. More empirical research niitment can have a large impact on the opportunities is needed to investigate the relationships among for participatory organization of activities. Food-for- poverty, farm size, market imperfections, and natural work programs may not have the intended investment resource management. (conservation) effects if participation is primarily for Market imperfections are more likely to hurt small employment and food security reasons and the pro- and poor farmers than large and rich farmers because gram does not represent local priorities. The owner- poor farmers are more likely to be rationed out of ship and future responsibility for maintenance and credit markets, for example, and to be less able to derivation of benefits must be clear and have local solve their problems through consumption smoothing support. Education, information, and communication (coping strategies). They may therefore have to go for systems are crucial to improving the basis for local ex ante income smoothing. This explains why the participation. One important strategy may be to facil- most vulnerable households are the most likely to itate learning across communities by supporting net- diversify their production (Morduch 1995). A case working and exchange programs. Collective action can therefore be made for intervening on the basis of may have more chances to succeed if it is led by rela- efficiency as well as equity. The possibiLity that tively young, literate persons who have been exposed poverty reinforces environmental degradation (and to the outside world, know the traditional society, and vice versa) strengthens the case for intervention. have integrity (Balland and Plateau 1996). The big ancL difficult question is how best to inter- vene in such cases. What is the optimal mix of incen- The Importance of Local Institutions tive-based ancL command-and-control mechanisms? and Decentralization And what role should markets have? The difference between "market" and "nonmarket" institutions Governments, donors, and nongovernmental organi- becomes blurred because various forms of interlinked zations should cooperate to strengthen local institu- contracts or markets may be perceived as either one. tions. Decentralization of power is often a prerequi- We believe that there are good hopes for improving site for success. Likewise, if the benefits are short market performance and that there are many win-win- term, local people should be willing to invest their win gains from doing so. However, particularly in time and energy because they tend to have high dis- remote and marginal regions and regions with severe count rates. Sometimes, local institutions may take on poverty, we need to look for supplementary policy new responsibilities; other times, it may be necessary measures. Interlinked contracts between local com- to establish new institutions because existing ones are munities, which have better local information unsuitable and difficult to modify. Forming new insti- (adverse selection, moral hazard, and enforcement tutions will require time to build a platform for coop- problems), and outside institutions, which face less eration as well as knowledge and managerial skills. It covariate risk problems, may provide win-win-win may be wise to start with the tasks that are easiest and solutions. Such contracts should be formed on the most likely to succeed and to expand to more difficult basis of local participation and local priorities that are tasks as experience is gained. compatible with efficiency, equity, and sustainability 66 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management objectives. A stepwise procedure should be chosen in Degradation of Natural Resources. Is There a Role the design and expansion of such contracts to allow for Rural Communities? Oxford: Clarendon Press. for learning on both sides, to achieve cooperative Barbier, Edward B. 1990. "The Farm-Level solutions, and to reduce information asymmetries. Economics of Soil Conservation: The Uplands of This may only be achieved if the principal is per- Java." Land Economics 66(2):199-211. ceived to be benevolent. Bardhan, P., ed. 1989. The Economic Theory of The disadvantage of this approach is that it Agrarian Institutions. Oxford: Clarendon Press. demands knowled,ge on both sides, requires continu- Barnum, H. N., and Lyn Squire. 1979. A Model of an ity and stability in leadership, and is easily under- Agricultural Household. Washington, D.C.: World mined by power struggles. Contracts should be Bank. designed to safeguard against this. Group responsibil- Bellon, M., and E. Taylor. 1993. "'Folk' Soil ity may be extended to areas other than shared respon- Taxonomy and the Partial Adoption of New Seed sibility for paying back loans. It could also be linked Varieties." Economic Development and Cultural to the management of local resources, where a share Change 41(4):763-86. of the costs or insurance against weather risk, for Berry, R. A., and W. R. Cline. 1979. Agrarian Structure example, could be provided from the outside (princi- and Productivity in Developing Countries. Geneva: pal). Food security for sustainable management of Intemational Labor Organization. local resources may be one option. Family planning Binswanger, Hans P. 1980. "Attitudes towards Risk: may have to be introduced in situations where the Experimental Evidence from Rural India." ratio between resources and population is unfavor- American Journal of Agricultural Economics able. Support from the outside should be made condi- 62:395-407. tional on local performance, creating incentives for Binswanger, Hans P., V. Balaramaiah, V. B. Rao, M. J. cooperation. Bhende, and K. V. Kashirsagar. 1985. "Credit Agency theory, contract theory, and mechanism Markets in Rural India: Theoretical Issues and design theory, combined with solid knowledge of Empirical Analysis." Discussion Paper ARU 45. local conditions, may provide a basis for institutional Agriculture and Rural Development Department, innovation by building on local priorities and partici- World Bank, Washington, D.C. Processed. pation. The complexity of the issues points toward a Binswanger, Hans P., K. Deininger, and Gershon strategy with pilot projects and flexible project design Feder. 1995. "Power, Distortions, Revolt, and accompanied by systematic research to allow for Reform in Agricultural Land Relations." In J. maximum learning and consequent adjustment and Behrman and T. N. Srinivasan, eds., Handbook of upscaling as confidence is gained. Development Economics. Vol. 3B. Amsterdam: North-Holland. References Binswanger, Hans P., and M. Elgin. 1990. "What Are the Prospects for Land Reform?" In Allen The word "processed" describes informally repro- Maunder and Alberto Valdes, eds., Agriculture and duced works that may not be commonly available Governments in an Interdependent World. through libraries. Proceedings of the twentieth international confer- Anderson, Jock R., and Jesuthason Thampapillai. ence of agricultural economists, 24-31 August 1990. Soil Conservation in Developing Countries: 1988. Buenos Aires: Dartmouth Publishing Co. Projects and Policy Intervention. Policy and Binswanger, Hans P., and S. R. Khandker. 1989. Research Report 8. Washington, D.C.: World Bank. "Agroclimatic Potential, Infrastructure, and the Angelsen, A. 1997. "Tropical Agriculture and Expansion of the Financial System in Rural India." Deforestation: Economic Theories and a Study Latin America and the Caribbean Country from Indonesia." Ph.D. diss. Norwegian School of Department II, World Bank, Washington, D.C. Econormics and Business Administration, Bergen. Processed. Ardilla, S., and R. Innes. 1993. "Risk, Risk Aversion, Binswanger, Hans P., John McIntire, and C. Udry. and On-Farm Soil Depletion." Journal of 1989. "Production Relations in Semi-Arid African Environmental Economics and Management Agriculture." In P. Bardan, ed., The Economic 25:27-45. Theory of Agrarian Institutions. Oxford: Balland, J.-M., and J.-P. Plateau. 1996. Halting Clarendon Press. Stein T Holden and Hans P Binswanger 67 Binswanger, Hans P., and M. R. Rosenzweig. 1984. Dutta, B., D. Ray, and K. Sengupta. 1989. "Contracts Contractual Arrangements, Employment, and with Eviction in Infinitely Repeated Principal- Wages in Rural Labor Markets in Asia. New Agent Relationships." In P. Bardhan, ed., The Haven, Conn.: Yale University Press. Economic Theory ofAgrarian Institutions. Oxford: 1986a. "Behavioral and Material Clarendon Press. Determinants of Production Relations in Ellis, F. 1988. Peasant Economics: Farm Households Agriculture." Journal of Development Studies and Agrarian Development. Cambridge, U.K.: 22(3):503-39. Cambridge University Press. 1986b. "Credit Markets, Wealth, and Eswaran, M., and A. Kotwal. 1990. "Implications of Endowments in Rural South India." Discussion Credit Constraints for Risk Behavior in Less Paper ARU 59. Agriculture and Rural Development Developed Economies." Oxford Economic Papers Department, World Bank, Washington, D.C. 42(April):473-82. Processed. FAO (Food and Agriculture Organization of the United . 1989. "Wealth, Weather Risk, and the Nations). 1996. Forest Resources Assessment Composition of Agricultural Investments." 1990: Survey of Tropical Forest Cover and Study Agriculture Department, World Bank, Washington, of Change Processes. FAO Forestry Paper 130. D.C. Processed. Rome. Binswanger, Hans P., and D. Sillers. 1983. "Risk- Feder, Gershon. 1985. "The Relation between Farm Aversion and Credit Constraints in Farmers' Size and Farm Productivity: The Role of Family Decisionmaking: A Reinterpretation." Journal of Labor, Supervision, and Credit Constraints." Development Studies 20:5-21. Journal of Development Economics 18(August): Boserup, Ester. 1965. The Conditions of Agricultural 297-313. Growth: The Economics of Agrarian Change . 1993. "The Economics of Land and Titling in under Population Pressure. London: Allen and Thailand." In Karla Hoff, Avishay Braverman, and Unwin. Joseph E. Stiglitz, eds., The Economics of Rural Braverman, Avishay, and Joseph E. Stiglitz. 1986. Organization: Theory, Practice, and Policy. New "Landlords, Tenants, and Technological Inno- York: Oxford University Press. vations." Journal of Development Economics Finkelshtain, Israel, and J. A. Chalfant. 1991. 23:313-32. "Marketed Surplus under Risk: Do Peasants Agree Carter, M. R., and J. Kalfayan. 1989. "A General with Sandmo?" American Journal of Agricultural Equilibrium Exploration of the Agrarian Question." Economics 73(August):557-67. Madison, Wis. Processed. Greenwald, Bruce C., and Joseph E. Stiglitz. 1986. Chayanov, A. V. 1966. The Theory of Peasant "Externalities in Economies with Imperfect Economy. ed. by Daniel Thorner, Basile Kerblay, Information and Incomplete Markets." Quarterly and R. E. F. Smith. Homewood, Ill.: Richard Irwin Journal of Economics 101(May):229-64. for American Economic Association. Grepperud, Sverre 1994. "Population-Environment Cheung, S. N. S. 1969. The Theory of Share Tenancy. Links: Testing a Soil Degradation Model for Chicago: University of Chicago Press. Ethiopia." Divisional Working Paper 1994-46. Cleaver, K. M., and G. A. Schreiber. 1994. Reversing Pollution and Environment Economics Division, the Spiral: The Population, Agriculture, and World Bank. Washington, D.C. Processed. Environment Nexus in Sub-Saharan Africa. . 1997. Soil Depletion Choices under Washington, D.C.: World Bank. Production and Price Uncertainty. Discussion de Janvry, Alain, M. Fafchamps, M. Raki, and E. Paper 186. Oslo: Statistics Norway. Sadoulet. 1992. "Structural Adjustment and the Hayami, Yujiro. and V. W. Ruttan. 1985. Agricultural Peasantry in Morocco: A Computable Household Development: An International Perspective. Model." European Review of Agricultural Baltimore, Md.: Johns Hopkins University Press. Economics 19(4):427-53. Hazell, Peter, C. Pomareda, and Antonio Valdes, eds. de Janvry, Alain, M. Fafchamps, and E. Sadoulet. 1986. Crop Insurance for Agricultural Develop- 1991. "Peasant Household Behavior with Missing ment: Issues and Experience. Baltimore, Md.: Markets: Some Paradoxes Explained." Economic Johns Hopkins University Press. Journal 101:1400-07. Heath, J., and Hans P. Binswanger. 1996. "Natural 68 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management Resource Degradation Effects of Poverty and Prices Lead to More or Less Soil Degradation?" Population Growth Are Largely Policy-Induced: Australian Journal of Agricultural Economics The Case of Colombia." Environment and 36(1):57-82. Development Economics 1(1):65-83. Lele, U., and S. W. Stone. 1989. "Population Pressure, Hoff, Karla, Aviishay Braverman, and Joseph E. the Environment, and Agricultural Intensification: Stiglitz, eds. 1993. The Economics of Rural Variations on the Boserupian Hypothesis." MADIA Organization: Theory, Practice, and Policy. New Discussion Paper 4. Agriculture Department, York: Oxford lJniversity Press. World Bank, Washington, D.C. Processed. Holden, S. T. 1991. "Peasants and Sustainable Low, A. 1986. Agricultural Development in Southern Development: The Chitemene Region of Zambia. Africa: Farm Household Economics and the Food Theory, Evidence, and Models." Ph.D. diss. Crisis. London: James Currey. Department oi Economics and Social Sciences, Malthus, T. R. 1987. An Essay on the Principles of Agricultural University of Norway, As. Population. Vol. 1, ed. by Patricia James. 1997. "Adjustment Policies, Peasant Cambridge, U.K.: Cambridge University Press. Household Resource Allocation, and Deforestation Marshall, Alfred. 1920. Principles of Economics. 8th in Northem Zambia: An Overview and Some ed. London: Macmillan. Policy Recommendations." Forum for Devel- McConnel, K. E. 1983. "An Economic Model of Soil opment Studies 1: 117-34. Conservation." American Journal of Agricultural Holden, S. T., H. Hvoslef, and R. Simanjuntak. 1995. Economics 65(1):83-89. "Transmigration Settlements in Seberida, McIntire, John. 1993. "Markets and Contracts in Sumatra: Deterioration of Farming Systems in a African Pastoralism." In Karla Hoff, Avishay Rainforest Environment." Agricultural Systems Braverman, and Joseph E. Stiglitz, eds., The 49:237-58. Economics of Rural Organization: Theory, Holden, S. T., and P. Sankhayan. 1997. "Population Practice, and Policy. New York: Oxford Pressure, Agricultural Change, and Environmental University Press. Degradation in the Himalayan Region: A Migot-Adholla, S., Peter B. Hazell, B. Blarel, and F. Conceptual, Historical, and Methodological Place. 1993. "Indigenous Land Right Systems in Basis." Discussion Paper D-05/1997. Department Sub-Saharan Africa: A Constraint on of Economics and Social Sciences, Agricultural Productivity?" In Karla Hoff, Avishay Braverman, University of Norway, As. Processed. and Joseph E. Stiglitz, eds., The Economics of Holden, S. T., and N. Shanmugaratnam. 1995. Rural Organization: Theory, Practice, and Policy. "Structural Adjustment, Production Subsidies, and New York: Oxford University Press. Sustainable Land Use." Forum for Development Morduch, J. 1995. "Income Smoothing and Studies 2:247-66. Consumption Smoothing." Journal of Economic Holden, S. T., B. Shiferaw, and M. Wik. 1998. Perspectives 9(3):103-14. "Poverty, Market Imperfections, and Time Nakajima, C. 1986. Subjective Equilibrium Theory of Preferences: Of Relevance for Environmental the Farm Household. Amsterdam: Elsevier. Policy?" Environment and Development Economics Ostrom, Elinor. 1990. Governing the Commons: The 3(1):83-104. Evolution of Institutions for Collective Action. Holden, S. T., .l. E. Taylor, and S. Hampton. Cambridge, U.K.: Cambridge University Press. Forthcoming. "'Structural Adjustment and Market Otsuka, Keijiro, H. Chuma, and Y. Hayami. 1992. Imperfections: A Stylized Village Economy-Wide "Theories of Share Tenancy: A Critical Survey." Model with Nonseparable Farm Households." Economic Development and Cultural Change 37 Environment and Development Economics. (October):31-68. Jodha, N. 1978. "Effectiveness of Farmers' Pagiola, Stefano. 1995. "The Effect of Subsistence Adjustment to Risk." Economic and Political Requirements on Sustainable Land Use Practices." Weekly 13(25)::38-48. Paper presented at meetings of the American Kates, R. W., G. Hyden, and B. L. Turner II, eds. Agricultural Economnics Association, Indianapolis, 1993. Population Growth andAgricultural Change Ind., 6-9 August. Processed. in Africa. Gainesville: University Press of Florida. Pender, John L. 1996. "Discount Rates and Credit LaFranche, J. T. 1992. "Do Increased Commodity Markets: Theory and Evidence from Rural India." Stein T Holden and Hans P Binswanger 69 Journal of Development Economics 50(2):257-96. Economnic Incentives for Soil Conservation: The Pender, John L., and J. Kerr. 1996. Determinants of Case of Highland Smallholders in Ethiopia." Farmers'Indigenous Soil and Water Conservation Discussion -Paper D-09/1997. Department of Investments in India's Semi-Arid Tropics. EPTD Economics and Social Sciences, Agricultural Discussion Paper 17. Washington, D.C.: University of Norway, As. Processed. International Food Policy Research Institute. . 1997b. "A Farm Household Analysis of Prosterman, R. L., and J. M. Riedinger. 1987. Land Resource Use and Conservation Decisions of Reform and Democratic Development. Baltimore, Smallholders: An Application to Highland Md.: Johns Hopkins University Press. Farmers in Ethiopia." Discussion Paper D- Repetto, Robert. 1987. "Economic Incentives for 03/1997. Department of Economics and Social Sustainable Production." Annals of Regional Sciences, Agricultural University of Norway, As. Science, special edition on environment manage- Processed. ment and economnic development 21(3):44-59. Singh, Inderjit, Lyn Squire, and John A. Stauss, eds. Rosenzweig, Mark R. 1988a. "Risk, Implicit 1986. Agricultural Household Models: Contracts, and the Family in Rural Areas of Low- Extensions, Applications, and Policy. Baltimore, Income Countries." The Economic Journal 98 Md.: Johns Hopkins University Press. (December):1148-70. Singh, R. P., and Hans P. Binswanger. 1988. "Income . 1988b. "Risk, Private Information, and the Growth in Poor Dryland Areas of India." Family." American Economic Review 78(2): Agriculture and Rural Development Department, 245-50. World Bank., Washington, D.C. Processed. Rosenzweig, Mark R., and Hans P. Binswanger. 1994. Srinivasan, T. N. 1972. "Farm Size and Productivity: "Wealth, Weather Risk, and the Profitability of Implications of Choice under Uncertainty." Agricultural Investment." Economic Journal Sankhya, The Indian Journal of Statistics (series B) 103:56-78. 34(4):409-20. Rosenzweig, Mark R., and 0. Stark. 1989. Stiglitz, Joseph E. 1974. "Incentives and Risk- "Consumption Smoothing, Migration, and Sharing in Sharecropping." Review of Economic Marriage: Evidence from Rural India." Journal of Studies 41(2):219-55. Political Economy 97(4):905-26. Tiffen, M., M. Mdortimore, and F. Gichuki. 1994. More Rosenzweig, Mark R., and K. I. Wolpin. 1989. People, Less Erosion: Environmental Recovery in "Credit Market Constraints, Consumption Kenya. New York: John Wiley and Sons. Smoothing, and the Accumulation of Durable Tviland, M. 1996. "Credit to Small-Scale Farmers in Production Assets in Low-Income Countries: Northern Zambia: Reasons for High Default Rates Investment in Bullocks in India." Department of and Means of Reducing Them." MSc. thesis. Agricultural Economics, University of Minnesota, Department of Economics and Social Sciences, St. Paul. Processed. Agricultural University of Norway, As. Processed. Sadoulet, E., and Alain de Janvry. 1995. Quantitative von Braun, Joachim, ed. 1995. Employment for Developnment Policy Analysis. Baltimore, Md.: Poverty Reduction and Food Security. Washington, Johns Hopkins University Press. D.C.: International Food Policy Research Institute. Sandmo, A. 1971. "On the Theory of the Competitive Walker, T. S., and N. S. Jodha. 1986. "How Small Firm under Price Uncertainty." American Farm Households Adapt to Risk." In Peter Hazell, Economic Review 6:65-73. C. Pomareda, and Antonio Valdes, eds., Crop Schultz, T. W. 1964. Transforming Traditional Insurance for Agricultural Development: Issues Agricutlture. New Haven, Conn.: Yale University and Experience. Baltimore, Md.: Johns Hopkins Press. University Press. Shiferaw, B. 1996. "Economic Analysis of Land Walker, T. W., and J. G. Ryan. 1990. Village and Degradation and Incentives for Soil Conservation Household Economies in India's Semi-Arid Tropics. in Smallholder Farming: A Theoretical Baltimore, Md.: Johns Hopkins University Press. Development." Discussion Paper D-30/1996. Wik, M., and S. T. Holden. 1996. "Risk and Peasant Department of Economics and Social Sciences, Adaptation: Do Peasants Behave According to the Agricultural University of Norway, As. Processed. Sandmo Model?" Discussion Paper D-24/1996. Shiferaw, B., and S. T. Holden. 1997a. "Analysis of Department of Economics and Social Sciences, 70 Small-Farmer Decisionmaking, Market Imperfections, and Natural Resource Management Agricultural University of Norway, As. Processed. World Bank. 1996. Towards Environmentally Williamson, 0. 1985. The Economic Institutions of Sustainable Development in Sub-Saharan Africa: Capitalism. New York: Free Press. A World Bank Agenda. Washington, D.C. 6 Agricultural Trade Reforms, Research Incentives, and the Environmnent Kym Anderson In recent years the world has, for the first time this ment. It takes as an example the General Agreement on century, begun reducing distortions to agricultural Tariffs and Trade (GATT) / World Trade Organization incentives in significant ways. The long-term growth (WTO) Uruguay Round agricultural agreement to in agricultural protection in high-income countries reduce high-income countries' import protection and (see McCalla 1969 and Lindert 1991) has slowed, and producer and export subsidies for farm products. The the heavy direct and indirect taxation of agriculture is opposite type of change in developing countries (that being reduced in numerous lower-income countries. is, the unilateral removal of policies that keep domestic Economists have hailed this as a breakthrough that food prices artificially low) is reinforcing the reloca- will lead to more efficient use of global resources, and tion brought on 1Dy the Uruguay Round of farm pro- they have provided many empirical estimates of the duction away from protected high-income countries. comparative static effects of these reforms on prices, An important feature of those reforms, though, is production, trade, and economic welfare (Tyers and their impact on incentives to invest in agricultural Anderson 1992 and Martin and Winters 1996). research. This is the subject of the second section of Others, however, have been less positive. Some envi- the chapter. The concluding section draws out some roDmental groups, for example, claim that economic implications of the analysis for developing countries modelers give too little attention to the (in their view) and for aid agencies. It also mentions areas where adverse effects that trade liberalization may have on further empirical economic research is needed. our rural and natural environments, on agriculture's Economists are confident that the removal of trade resource base, and on food safety. The aid communi- distortions will in almost all situations boost eco- ty is another example: it is worried that the reduction nomic welfare and will in many situations also in government attention to agriculture in high-income improve the environment, particularly if govern- countries may be contributing to a loss of interest in ments have in p lace and enforce optimal environ- funding agricultural development assistance pro- mental policies. Nonetheless, the chapter makes clear grams, including international agricultural research. that the environmental effects of many policy To date, comparative static economic modelers have changes can have both positive and negative ele- paid little attention to the effects of reform on the ments. Hence quantifying the main environmental environment or on research and development (R&D). effects of economic reform is needed to convince This chapter looks at both of these concerns. It skeptics that o.n balance trade liberalization does begins by exploring the claim that the income growth indeed, as espoused in the preamble to the agreement and resource reallocations that follow farm price and establishing the WTO, contribute to sustainable global trade policy reforms will adversely affect the environ- development. 71 72 Agricultural Trade Reforms, Research Incentives, and the Environment The Environmental Consequences of Reducing that can raise those incomes. The main exceptional Agricultural Protection in Industrial Countries circumstance is when the environmental effects of greater spending spill over national boundaries. The Concerns about the environmental effects of reducing spillovers could be physical, as with carbon emissions agricultural protection stem in large part from the fact and large-scale deforestation (climate change, that reducing t:he distortions to international trade reduced biodiversity) and with chlorofluorocarbons increases real incomes in the countries whose exports (cFcs) and halons (ozone depletion). Or the spillovers expand and changes the mix and international loca- could be of an emotional kind, as concern for the wel- tion of production and consumption. Both of these fare of animals. One possible solution when there are effects worry some environmentalists in high-income international spillovers is to seek international envi- countries. That, in turn, concerns people in poorer ronmental agreements (the Montreal Protocol on CFC countries, who fear that environmental groups may phaseout and the Convention on International Trade use their political power to convince industrial- in Endangered Species). But such agreements typical- country governraents to impose trade or other sanc- ly are very difficult to conclude, not least because of tions on poorer countries deemed to have inadequate large differences across countries in incomes and environmental a:nd food safety policies. hence in the ability of people to place a high value on the goods and services of the natural environment. In Effects of Boosting Incomes those circumstances, wealthier and more-concerned environmentalists will look for other ways to influ- With respect to the effect of trade reform on raising ence the environmental damage of other countries (as incomes in developing countries, many environmen- they perceive it), and using their own country's trade talists assume that the consequent increase in spend- policy as a stick or carrot to influence behavior in ing will place greater demands on the environment. other countries may be one of the few options they That assumption, however, is questionable, because have (Anderson 1997b). income growth brings with it at least three important changes in behavior. One is that population growth Effects of Relocating Production tends to decline as incomes rise, which reduces one important source of pressure on both urban and rural With respect to changes in the mix and international environments of developing countries. Another is that location of production that would accompany agricul- education investment expands with income, and with tural trade liberalization, environmentalists have it comes more skillful management of all resources, expressed at least two concerns. They fear that in the including private use of the natural environment. And highly protected countries of Western Europe the rural third, the governments of modernizing communities countryside and villages will be less visually attrac- with rising incomes and improving education tend tive and less populated as farmers respond to lower eventually to improve private property rights and to domestic food prices by "getting bigger or getting put more-stringent environmental policies in place out." And they fear that higher food prices in interna- (Radetzki 1992, Grossman 1995, and Grossman and tional markets, following reduced exports and Krueger 1995). Not only does the demand for pollu- increased imports by the highly protectionist tion abatement policies appear to be quite income economies, will raise agricultural land prices in tropi- elastic (at least beyond a certain threshold), but also cal and Southern Hemisphere countries. That, in turn, the cost of compliance falls, because trade liberaliza- will stimulate greater net deforestation to expand the tion expands the opportunities to acquire more envi- area of agricultural land (thereby reducing the forest's ronmentally benign production technologies, inputs, absorption of carbon dioxide from the atmosphere and and consumer products. its contribution to biodiversity) and the use of heavier Environmentalists may be disappointed that gov- doses of agricultural chemicals (which not only emments adopt less-stringent environmental stan- degrade the local environment but also increase the dards and charges than they would like; but the appro- amount of chemical residues in food that the reform- priate response in most circumstances is for them to ing countries might import). These concerns are advocate tougher domestic environmental standards understandable, but they are based on only a small in their own and other countries as incomes rise, number of direct and indirect environmental effects rather than to argue against price and trade reforms involved in the relocation of production that would be Kym Anderson 73 induced by trade reform. It is possible to examine a that and will be phased in over a six-year period for few additional influences by using some estimates of industrial countries and up to 10 years (to 2004) for the production effects of a multilateral reforn as sim- poorer countries (Martin and VWinters 1996). Even ulated by a model of world food markets (for some assuming such a huge liberalization and instanta- early studies, see Abler and Shortle 1992, Barbier and neous adjustment, the estimated impact on world Burgess 1992, Lutz 1992, and Runge 1994). food output in aggregate is estimated to be negligi- The estimates of Tyers and Anderson's (1992) ble and the relocation of production only minor: model are from an extreme simulation: it assumes grain and meat production would be 5 or 6 percent complete removal of all farmer support policies in lower in industrial countries and 3 to 8 percent high- all industrial countries and U.S. land set-asides as of er in developing countries. The big declines in out- 1990 and full adjustment within a year. In fact, the put would occur in Japan and Western Europe, but Uruguay Round cuts will be less than one-fifth of they would be partly offset by increases in North Figure 6.1. Shares of Various Countries in Global Changes in Food Production Following Full Liberalization of Industrial-Country Food Policies (percent) a Meat < t~ Africa/Middle East m Grain _ Other Asia South Asia China Other Latin America Brazil =_____ Argentina Australia -_ _ North America -_____ Japan ____ EFTA EC-12 -80 -60 -40 -20 0 20 40 Source: Anderson and Strutt 1996. Figure 6.2. Proportional Changes in Food Production of Various Countries Following Full Liberalization of Industrial-Country Food Policies (percent) * Meat Africa/Middle East M Grain Asia Latin America Australia North America Japan Western Europe -60 -50 -40 -30 -20 -10 0 20 40 Source: Anderson and Strutt 1996. 74 Agricullural Trade Reforms, Research Incentives, and the Environment Figure 6.3. Relationship between Agricultural Producer Subsidy Equivalent (PSE) for 1979-89 and the Use of iChemical Fertilizers and Pesticides per Hectare of Arable and Pasture Land 400 *6 350 A : 300 4 250 f 200 3 P X150.X/ / 2 2 o 10 loo~~~~~~~~~~~~~~~~~~~~ -40 -20 0 20 40 60 80 Agricultural PSE (percent) Source: Anderson and Strutt 1996, based on World Resources Institute 1990: table 18.2. America and Australasia. The rich regions in North chemicals depends on more than just the level of dis- America and Australasia are estimated to account tortion in output price, but previous econometric work for more than a quarter of the increases, with devel- suggests that a thorough multivariate regression oping countries providing the balance (see figures analysis that includes additional variables such as 6.1 and 6.2). wage rates is still likely to find a strong positive cor- What would such a relocation do for the use of relation between chemical use and output prices. pollutive farm inputs, for land use, and for the use of Furthermore, land-scarce Western Europe and labor and capital? We consider each in turn. Japan crop twice as much of their total land area as does the rest of the world on average, so the extent of Changes in the use of environmentally unfriendly contamination of their soil, water, and air from the use farm inputs. Figure 6.2 suggests that for regions of farm chemicals is even greater than figure 6.3 sug- where production would fall, the reductions would gests, relative to other countries. That is, the reloca- constitute a fairly large proportion of output. But for tion of crop production from densely populated pro- regions where production would expand, the increases tectionist countries to the rest of the world would would constitute a relatively small proportion of cur- cause a much larger reduction in degradation in the rent output, especially for grain. The difference is former compared with any increased degradation in because the price decreases in the former regions are the latter, where chemical use would expand from a much larger than the price increases in the latter low base and to still-modest levels. Should the regions. This has important implications for environ- expanding region have or choose to introduce an opti- mental degradation and chemical residues in food, mal tax on the sale of farm chemicals at the time of because the contracting regions are relatively densely the rise in food prices following the reform abroad, populated with high land prices and so use farm that tax would not eliminate externalities from chem- chemicals and intensive livestock methods much ical use but would at least ensure that at the margin more than the expanding regions. In the case of fertil- the benefit to farmers from applying more farm chem- izer and pesticides, for example, the highly protected icals would just equal the external cost that those countries use more than 10 times as much per agri- chemicals impose on others. Rather than taxing the cultural hectare as Australasia and most developing pollution caused by irrigation and fertilizer and pesti- countries (figure 6.3). Of course, the demand for farm cide use in farmning, many developing-country govern- Kym Anderson 75 ments have contributed to environmental damage by and Ringler 1997). subsidizing the use of those inputs in the past. Reforms In rich countries where price supports are lowered are occurring though. For example, Bangladesh and and there are no land set-aside policies, land values Indonesia have chosen to phase out such subsidies in would fall, which would probably further reduce the recent years, even if they have yet to add pollution use of land substitutes such as farm chemicals, irriga- taxes; and already the reductions in environmental tion water, and feed concentrates. (The word "proba- degradation are evident (World Bank 1997). bly" is necessary because some offsetting forces make What about the effect of trade policy reform on the it uncertain whether reducing output prices leads to use of inputs in the livestock sector? The relocation of more or less care of farmland, particularly because the meat and milk production from the most densely pop- altered incentives to avoid on-site damage also affect ulated rich countries to relatively lightly populated off-site damage; see, for example, Barrett 1991, poorer countries would be associated with a decline in Clarke 1992, and LeFrance 1992.) Much of that the extent to which the world's livestock is fed grain would happen immediately, but over time their use and supplements rather than pasture. With this would would fall even more as the decline in land price come a decline in the use of growth hormones and reduces the incentive to seek land-saving technologies veterinary medicines, partly because animals are less (Hayami and Ruttan 1985). The use of some land valuable in less-protected economies and partly would switch to products whose domestic prices have because the risk of the spread of disease is lower in not fallen or have fallen less, such as the subset of range feeding than in intensively housed conditions. fresh fruits and vegetables that are not easily tradable The greater use of less-intensive methods would internationally because of high transport costs. Such reduce not only air, soil, and water contamination products may gen,erate higher or lower negative exter- associated with the disposal of animal effluent but nalities than the ones previously produced. But other also the chemical additives in the food produced. land might revert to recreational and other nonfarm Moreover, insofar as this relocation leads to greater uses (golf courses, for example), whose positive use of crop and leguminous pasture rotation methods, externalities may be more or less for nonfarmers than there will be less need for chemical fertilizer and the aesthetic pleasure they derive from seeing the land hence less water pollution from nitrates. being farmed. Changes in the use of land. What about land degra- Changes in the use of labor and capital. The fall in dation? Think first of the nonsubsidizing countries, farm profitability in industrial countries would where product prices would rise slightly following encourage more labor and capital to be employed in reforms abroad. The lightly populated countries such nonfarm sectors, most likely the dominant services as Argentina, Australia, and New Zealand, where sector, which is relatively unpollutive. By contrast, in most of the potential farming land is already cleared the developing countries the greater profitability of for agriculture, would see a more-intensive use of that farming because of higher intemational and domestic land, but from its current relatively low base. In trop- farm product prices would encourage more labor and ical countries, some land might be attracted from crop capital to be employed in commercial agriculture. plantation areas (coffee, palm oil, and so forth), which Those productive factors would come from other may cause more or less degradation than the now- activities where they may well be contributing more more-profitable alternative. But conservationists are environmental damage than they would in an expand- concerned that more tropical forests would be felled. ed commercial agricultural sector. One possibility is How much is an empirical question, about which manufacturing, which in many newly industrializing there is relatively little evidence at present. It is worth countries can be quite pollutive until incomes rise suf- noting, however, that during the period 1961-91 cere- ficiently for people to demand stricter enforcement of al production in developing countries increased 165 environmental policies. The other possibility, particu- percent, but the area of arable land increased only 17 larly in less-advanced economies, is that underem- percent (Alexandratos 1995), and that area is project- ployed labor will be attracted to commercial farming. ed to expand only 13 percent during the subsequent Whether such workers come from the urban slums or three decades, even though per capita food supplies from rural areas, they are likely to do less environ- will expand and cereal prices in international markets mental damage in their new job. In the case of work- are expected to be 20 percent lower by 2020 (Rosegrant ers who would otherwise be eking out a subsistence 76 Agricultural Trade Reforms, Research Incentives, and the Environment income by squatting on marginal hillsides, the result Environmental instruments in lieu of price sup- would be less deforestation and soil degradation on ports. The second example has to do with environ- those hillsides. As well, the increased value of rural mental measures per se. An important consequence of labor would raise the opportunity cost of collecting and making traditional barriers to international trade less chopping wood for fuel. Cleaner fuels such as kerosene acceptable is that industry assistance in politically would then be used instead, and forests would be correct green garb is likely to be used more. Hence it depleted less as a result. This positive effect on the is not surprising that there are already European environment could be substantial, given that four-fifths Union Council Regulations (for example, no. of logs felled in developing countries are used as fuel 2078/92) allowing environmental subsidies to agricul- (World Resources Institute 1996: table 9.3). ture, including subsidies to reduce the use of fertilizer and pesticides, to promote environmentally sound Effects of Policy Reinstrumentation methods of production, to encourage extensive agri- cultural techniques, to maintain practices that are What about other policy responses to trade policy already compatible with the environment, and to reform under the Uruguay Round agreement on agri- assist organic farming. These are allowable ("green culture? Because the agreement bans the use or box") measures in the Uruguay Round agreement on requires less use of some but not all policy instru- agriculture and hence are likely to become more com- ments, there is the possibility that some protection mon in Europe and elsewhere over time, substituting reinstrumentation will take place as governments of somewhat for traditional economic protectionism. industrial countries attempt to reduce the economic These measures may well improve the environment loss to powerful producer groups. Four examples but are likely to be excessive if used as substitutes for serve to illustraite the extra uncertainty that this can farm price supports-and they are not necessarily the introduce to predicting the environmental conse- first-best ways to achieve particular environmental quences of trade policy reforms: output-limiting outcomes. Hence monitoring their rise may be just as instruments, environmental measures, R&D invest- important as monitoring the promised declines in tra- ments, and quarantine import restrictions. ditional border protection and other means of farm price support, with a view to arguing against exces- Production-limiting policy instruments. In the case sive or inappropriate interventions. of the Uruguay Round agreement on agriculture, price supports for products subject to acreage set-asides and Agricultural research investments. A decline in the other production-limiting instruments were not count- profitability of farming in industrial countries because ed in the aggregate measure of support (AMS). Instead of cuts in protection would lead one to expect the they were classified as "blue box" instruments, profitability of agricultural research to fall as well. whereby U.S. and European Union governments, in However, because much of that research is supported trying to minimize the political cost of meeting com- by government and because public agricultural mitments to reduce the AMS along with cuts in export research funding appears in the WTO'S "green box" subsidies and increases in import market access, were and so is not counted as part of a country's AMS, pub- allowed to maintain price supports for products with lic research investment may rise rather than fall. an output-limiting instrument attached. However, Conceivably that rise could more than offset the like- there is a cons:iderable degree of leakage in such ly fall in private investment in agricultural research. instruments. For instance, farners may be allowed to And with environmental policies in these countries set aside their worst-yielding acres, which reduces altering farmers' incentives in ways that discourage their output by only a fraction of the reduction in the use of environmentally damaging practices, new acreage. Yet the fact that land is set aside makes the investments in public agricultural research are likely remaining land used more valuable, which encour- to reinforce that trend. ages higher doses of land-augmenting farm chemicals per hectare-not an obvious improvement from an Quarantine import restrictions. Quarantine import environmental viewpoint. If more price support pro- restrictions have always been a potential substitute for grams were to use this type of policy instrumentation, more direct measures of assistance to farmers. this could reduce future environmental gains in farm Agricultural exporting countries were well aware of assistance-reducing countries. the possibility that the Uruguay Round agreement on KKym Anderson 77 agriculture could encourage greater use of quarantine export markets abroad and to reduce the returns to import restrictions as a new form of protectionism producers who have enjoyed protection from import and, for that reason, ensured that an agreement on the bans on quarantine grounds that cannot be justified application of sanitary and phytosanitary (sps) mea- scientifically. In the latter cases, removal of unjusti- sures (to protect human, animal, or plant life or fied import barriers could boost the welfare of domes- health) was part of the Uruguay Round package. tic consumers by nore than it would harm the welfare Under that agreement, members of the WTO have a of domestic producers (James and Anderson 1998). right to apply their own standards, provided they are Only time will tell. whether the sPs agreement is tough not unduly trade-restrictive. The sps agreement estab- enough to prevenlt reinstrumentation toward a greater lishes detailed rights and obligations for environmen- use (abuse) of quarantine measures as a means of sup- tal health and food safety as well as measures to pre- porting farm incomes. vent the spread of pests or diseases among animals and plants. It outlines procedures for product inspec- The nonagricultural Uruguay Round agreements. tion, treatment, and processing, risk assessment, and Trade reforms are typically not confined to just one allowable maximum levels of pesticide residues and sector. The multilateral Uruguay Round involved certain food additives based on agreed international more than a dozen agreements, of which agriculture standards (the Codex Alimentarius Commission for was just one. Likewise, regional agreements tend to food safety, the International Office of Epizootics for involve many, if not all, sectors. And unilateral eco- animal health, and the International Plant Protection nomic reforms are often across-the-board or at least Convention for plant health). Members can adopt not focused on the farm sector alone. Hence relative higher standards if they wish, but only with appropri- prices of farm products are affected by the change in ate scientific justification. The sps agreement allows nominal price not only of those goods but also of governments to challenge, via the WTO'S dispute set- other goods affected by the reform package. tlement body, another country's food safety or other The Uruguay Round, for example, not only goes technical requirements based on evidence showing only a small fraction of the way toward freeing farm whether the measure is justified. Canada, for instance, trade (less than one-fifth) but also frees up manufac- is challenging Australia's justification for sanitary turing and other trade as well. Hence real intemation- measures banning salmon imports. Another important al food prices are expected to be raised by no more food case before the WTO in 1997 had to do with the than 1 percent, mainly because the international price European Union's ban on the import of beef produced of some manufactures also would rise (Goldin and with the use of growth hormones. The European van der Mensbrugghe 1996). For countries in East Union claimed that the ban applies to domestic pro- Asia, in particular, the Uruguay Round agreement on ducers and so does not contravene the national treat- textiles and clothing ensures that domestic prices will ment provision of the GATT, whereas the United States rise more for those products than for farm products argued (successfully, as it happened) that the hor- over the Uruguay Round's implementation period. As mones are not harmful to humans and hence the a result, the share of agriculture in the Association of European Union standard is excessive and labeling Southeast Asian Nation's (ASEAN) gross domestic should be sufficient protection for consumers. For the product (GDP) is projected to fall by 27 percent instead first time, the WTO called on scientific experts to help of 21 percent between 1992 and 2005, the share of resolve that case. other primary products is projected to fall by 17 This is clearly an area where trade tensions could instead of 13 percent, while the share of light manu- escalate. It is telling that during the 47 years of the factures in ASEAN'S GDP iS expected to rise by 42 per- GATT, virtually no formal trade disputes arose over sPs cent instead of 16 percent, according to the Global measures, because a country's import restrictions to Trade Analysis Project (GTAP) modeling results protect human, health, and plant life were difficult to reported in Anderson and others (1997). The estimat- challenge under previous GATT agreements. In the ed environmental implications of the Uruguay Round first 18 months of the WTO'S formation, by contrast, for that group of countries clearly would be very dif- seven formal complaints were lodged under WTO dis- ferent if the focus were only on the agriculture agree- pute settlement procedures. The sps agreement is thus ment rather than on the full package of agreements. likely to help those agricultural exporters who have With this in mind, Strutt (forthcoming) uses an been facing unduly restrictive barriers in potential economywide approach to examine the environmental 78 Agricultural Trade Reforms, Research Incentives, and the Environment effects of the UJruguay Round and of the proposed reform. Indeed, a recent empirical study that attempt- Asia-Pacific Economnic Community trade liberaliza- ed to place a monetary value on estimated changes in tions for the case of Indonesia. The global GTAP model pollution following the Uruguay Round finds that the is enhanced with environmental modules to examine changes reduced developing countries' estimated eco- air, soil, and water pollution over the next two nomic welfare gains from the Uruguay Round by less decades without and with those trade reforms. It pro- than 2 percent while raising the welfare gains to some duces striking results in several respects. First, that advanced economies (Cole, Rayner, and Bates 1997). computable general equilibrium (CGE) model suggests And of course if the trade reform were accompanied that only some rather than all of Indonesia's agricul- by reform that brought environmental policies closer tural subsectors will expand as a result of the Uruguay to optimal regulations, the likelihood of an even bet- Round. Second, it suggests that while the expansion ter outcome would be enhanced, as Beghin and others in aggregate output and the shift in composition of (1997) demonstrate quantitatively in a case study of that output because of the Uruguay Round would add Mexico. slightly to soil erosion, the cost of the damage caused This review focuses only on the comparative stat- by that increased erosion is minuscule, amounting to ic effects of farm policy reform (that is, the one-off less than 0.2 percent of the national economic welfare increase in income, resource adjustment, and produc- gain (as traditionally measured) from Uruguay Round tion relocation). Yet such policy reforms can also liberalization. Third, the trade policy reforms slated affect incentives to invest, including R&D efforts, that for the next two decades are shown to reduce many can permanently boost economic growth rates forms of air and water pollution, water use, and defor- (Grossman and Helpman 1991). Because faster rates estation in Indonesia, and even the forms of pollution of economic growth tend to speed up improvements that are projected to increase would add only very in the management of environmental policies, they slightly to environmental degradation-even without are desirable both from an economic and an environ- toughening the enforcement of existing environmen- mental viewpoint. The next section of this chapter tal regulations or adding new ones. And fourth, for the therefore examines the impact of a rise in farm prod- pollution levels that do rise with trade reform, the uct prices in developing countries on the incentive to extent of the rise in all cases is a tiny fraction (less invest in agricultural research, because such research than 2 percent) of the rise projected to accompany is arguably one of the fastest ways to boost the wel- Indonesia's economic growth (again assuming no fare of the world's poorest people (Schultz 1980). toughening of environmental regulations) over the period of implementation (see Strutt and Anderson The Impact of Hfigher Real Farm Product Prices 1998). in Developing Countries on Returns to Agricultural Research Conclusion Traditionally, many (especially food-surplus) devel- In short, the environmental consequences of farm oping countries have adopted price, trade, and price and trade policy reforms are very complex, with exchange rate policies that have discriminated heavi- lots of offsetting forces at work, of which only a sub- ly against the rural sector (Krueger, Schiff, and Valdes set will ever be easily quantified. The reforms' effects, 1988 and Tyers and Anderson 1992). Independent of both on boosting income and hence encouraging more the Uruguay Round, those distortionary policies have output and on relocating production globally, could been coming under scrutiny in recent years and are increase or decrease global environmental damage. beginning to be dismantled unilaterally in a growing But what this makes clear is that (a) there is much number of developing countries. Such unilateral more likelihood that the reforms will have a positive reforms could have a much larger impact-by raising net effect on the environment than is often admitted agricultural prices in these countries-than the reduc- by environmental groups and (b) where the environ- tions in agricultural protection resulting from the mental effect is negative, the extra damage caused by Uruguay Round. They will reinforce the global trade reforms is, only a minuscule fraction of that income growth and relocation of farm production which would occur as a result of normal economic away from protected high-income countries brought growth, whose adverse welfare effects are likely to be on by the Uruguay Round and so tend to amplify the far more than offset by the welfare gains from trade effects on the environment outlined in the previous Kym Anderson 79 section. But because the rise in domestic farm prices If developing countries were to reduce their policy could be very substantial in many cases (in China, discrimination against agriculture, there would be prices rose at least 50 percent in the late 1970s, early larger national agricultural research systems to adapt 1980s,, for example), effects on economic welfare and and promote technologies developed at international the environment, through altered agricultural agricultural research centers. Hence the measured research incentives and rewards, also are worthy of returns to investment in those development assistance consideration. agencies would also rise following such reforms. The A simple way to explore this issue is to consider a latter should help to offset the aid community's con- market for a single commodity (say a staple such as cern that the reduction in government support for maize) in a small country. Assume that, for the bene- agriculture in high-income countries, following the fit of poor net buyers of this food staple (albeit at the Uruguay Round, will contribute to a loss of interest expense of net sellers, even though they may be the in funding agricultural development assistance pro- majority of the poor population), the country has kept grams, including those at international agricultural the domestic price low via a ban on exports of this research centers. good. If the ban were lifted, this would become a big- The boost to agricultural research in developing ger industry in terms of gross output or value added, countries that would be stimulated by unilateral and hence the scope for levying the income of pro- price and trade policy reforms, together with a pos- ducers to fund national agricultural research would sible reduction in national agricultural research in rise. In a numerical example, Anderson (1997a) those high-income countries that reduce their high shows that even if the export ban depressed the food prices to rneet their Uruguay Round commit- domestic price by just one-third (not unusual in ments, would tend to amplify the relocation of agri- developing countries; see Krueger, Schiff, and Valdes cultural production discussed in connection with fig- 1988), in the case where the supply and demand ures 6.1 and 6.2. One might therefore expect it to curves had 450 slopes, it would reduce: amplify the environmental consequences of that relo- * The scope for raising national agricultural research cation. However, two caveats should be kept in mind. funds via a grower levy by between two-thirds and One is that the research itself may lead to the three-fourths (depending on whether funds are to adoption of new technologies that have different be levied as a percentage of farm income gross or environmental effects than present practices. Indeed net of farm input costs) there have been calls for more development and dis- * The benefit to producers from (and hence their semination of precision technologies to reduce envi- incentive to seek) the new technology by between ronmental degradation, but that is unlikely to be 75 and 88 percent (depending on whether the sup- heeded until more developing-country governments ply curve shift is parallel or proportional) have removed subsidies on farm inputs and taxed * The net economic benefit from the new technolo- their pollutive content (Zilberman, Khanna, and gy (from which national agricultural research costs Lipper 1997). have to be subtracted) by between 50 and 63 per- The other caveat is that any stimulus to interna- cent and therefore tional agricultural research centers may well have a * The internal rate of return from national agricultur- positive effect on advanced economies also. Byerlee al research investment by as much as 63 percent. and Traxler (1996) find that a substantial proportion In short, reforming the policies that depress food of the productivity gains associated with past invest- prices, other things being equal, is going to raise ments in international agricultural research centers much more than proportionately the rates of return have accrued in the donor countries, because of their from agricultural research for that reforming econo- well-developed capacity to import and adapt new my. It is also going to alter the relative returns to technologies. The more that happens, the more research among the various farm products to the world output will grow, putting downward pressure extent that they face different degrees of distortion or on agricultural prices in international markets. Those reform. Should that policy reform also involve raising developing countries that are net food importers will the prices of environmentally damaging farm inputs, enjoy a terms-of-trade gain, while the net food it would induce more of a search for new precision exporters will benefit so long as their growth in farm technologies that use those inputs less intensively productivity more than offsets the downturn in (Hayami and Ruttan 1985). prices of their exports. 80 Agricultural Trade Reforms, Research Incentives, and the Environment Policy Implications and Areas for Further rise is that the returns from investing in agricultural Quantitative Economic Research research will be much higher in the future. Available empirical studies suggest that developing countries Economists are confident that the removal of trade have underinvested in this activity even at low prices distortions will in almost all situations boost econom- (Alston, Craig, and Pardey 1996). In recent years, ic welfare. The discussion in this chapter suggests that govermments have sought to downsize, so the degree in the case of agricultural policy reform it would in of underinvestment will be even greater at less- many situations also make a net improvement in the depressed price levels. environment. Where it would not, the reason is com- The international agricultural research centers too monly that governments do not have appropriate envi- will enjoy higher rates of return to their research on ronmental policies in place. Having said that, it is behalf of refonring developing countries. Insofar as clear that the environmental effects of many policy those countries' reforms involve raising prices not changes can have some negative elements even when only for farm products but also for chemical inputs the net welfare effect might be positive. Hence before and water, the rewards will be especially high from it is possible to convince skeptics that trade liberal- generating environmentally friendlier precision tech- ization does indeed-as espoused in the preamble to nologies as distinct from ones using heavy doses of the agreement establishing the WTO-contribute to chemicals and water (Zilberman, Khanna, and Lipper sustainable global development, more attempts will 1997). These changes should make it easier for inter- need to be made to quantify the environmental national agricultural research centers to entice financial changes likely to be associated with trade and other support from donors, not least because the precision economic reforms. The papers by Beghin and others technologies are also becoming more sought-after in (1997), Cole, Rayner, and Bates (1997), and Strutt the advanced economies supplying development and Anderson (1998) are a small start to that process. assistance funds. This suggests yet another topic for For developing countries, several policy implica- empirical modeling work, namely, projecting the tions follow from the analysis in this chapter. One is effects of new research on the relocation of the that raising domestic prices of farm products up world's agricultural production, and the associated toward those in international markets is highly desir- impact on the environment, under various assump- able but needs to be accompanied by other policy tions about the degree of spillover adoption by changes to ensure that the country's agricultural advanced economies. development is sustainable. Specifically, at those Finally, what might be in store for agriculture as the higher output prices, the cost of not reducing distor- World Trade Organization approaches the next centu- tions in markets for farm chemicals and irrigation ry? Clearly much remains to be done before agricul- water-degradation of the farm sector's natural tural trade is as fully disciplined as under the WTO or resource base-will be much greater than when prod- as free as world trade in manufactures. The Cairns uct prices and hence input intensities are lower. If the Group of agriculture-exporting countries is planning country seeks to export some of that extra output to simultaneous action on numerous fronts (Anderson high-income countries, the cost of not reducing those 1998). Their first priority is to secure an early com- distortions might be high in two other senses as well. mitment to begin a new round of multilateral trade One is that the food safety standards of the importing negotiations at the turn of the century, one that is com- countries would need to be met, including limits on prehensive enough to allow intersectoral tradeoffs. the level of chemical additives in food. The other is The second priority is to ensure that all possible areas that production deemed to be environmentally irre- for opening agricultural markets are on the table. In sponsible could trigger trade sanctions, as Mexico addition to reductions in production and export subsi- found when it continued to catch dolphin in tuna nets dies and bound tariffs, this includes removing the and sought to export the tuna to the United States, "blue box" in the Uruguay Round agreement on agri- which had banned that form of netting. Such sanctions culture, which has allowed the United States and are much less likely in the future, however, following European Union to continue subsidizing industries the GATT dispute settlement finding in favor of subject to set-aside and other production-limiting mea- Mexico on that issue. sures, expanding tariff rate quotas, demonopolizing Another implication for developing countries that state trading enterprises, and phasing out special safe- allow their low clomestic prices of farm products to guard provisions. It also includes tightening up the Kym Anderson 81 provisions in the sps agreement to reduce further the Adelaide: University of Adelaide. scope for countries to use (misuse) quarantine restric- Anderson, Kym, Betina Dimaranan, Tom Hertel, and tions to raise their farmers' incomes and striving to get Will Martin. 1997. "Economic Growth and Policy socialist economies in transition (most notably China, Reforms in the APEC Region: Trade and Welfare Russia, Ukraine, and Vietnam) to reform sufficiently Implications by 2005." Asia-Pacific Economic to be welcomed into WTO membership. Review 3(1, April): 1-21. On the specific issue of trade and environment Anderson, Kynm, and Anna Strutt. 1996. "On linkages, there will be pressure from environmental Measuring the Environmental Impact of groups to have the matter on the agenda of the next Agricultural Trade Liberalization." In M. E. WTO round of talks. It is too early to tell what form Bredahl, N. Blallenger, J. C. Dunmore, and T. L. that discussion might take, but it is unlikely to result Roe, eds., Agriculture, Trade, and the Environment: in agreement to do anything as extreme as allow Discovering and Measuring the Critical Linkages, countries to use trade measures against other coun- ch. 11. Boulder, Colo.: Westview Press. tries with less-stringent environmental policies. Even Barbier, Edward B., and Joanne C. Burgess. 1992. calls to have an environmental impact assessment of "Agricultural Pricing and Environmental any agreements that further reduce trade restrictions Degradation." Background paper for the World will be seen as impractical, for if any such assessment Development Report 1992. World Bank, would lead to required amendments to the negotiated Washington, I).C. Processed. agreement, the whole agreement would unravel. Such Barrett, Scott. 1991. "Optimal Soil Conservation and calls will be less likely, however, as more empirical the Reform of Agricultural Pricing Policies." research fails to demonstrate the significant adverse Journal of Development Economics 36(2): 167-87. environmental consequences of trade liberalization. Beghin, John, Sebastian Dessus, David Roland-Holst, and Dominique van der Mennsbrugghe. 1997. "The References Trade and Environment Nexus in Mexican Agriculture: A General Equilibrium Analysis." The word "processed" describes informally repro- Agricultural Economics 17(2-3, December): duced works that may not be commonly available 115-31. through libraries. Byerlee, Derek, and Greg Traxler. 1996. "The Role of Abler, David G., and John S. Shortle. 1992. Technology Spillovers and Economies of Size in "Environmental and Farm Commodity Policy the Efficient Design of Agricultural Research Linkages in the U.S. and the E.C." European Systems." Paper presented to the global agricultur- Review of Agricultural Economics 19(2): 197-217. al science policy conference, 26-28 August, Alexandratos, Nikos, ed. 1995. World Agriculture: Melbourne. Processed. Towards 2010. New York: John Wiley and Sons Clarke, Harry. 1992. "The Supply of Nondegraded for the Food and Agriculture Organization of the Agricultural Land." Australian Journal of United Nations. Agricultural Economics 36(1, April): 31-56. Alston, Julian, Barbara Craig, and Philip Pardey. Cole, M. A., Anthony J. Rayner, and J. M. Bates. 1997. 1996. "Research Lags and Research Returns." The Environmental Impact of the Uruguay Round. Paper presented to the global agricultural science CREDIT Research Paper 97/20. Nottingham: policy conference, 26-28 August, Melbourne. University of Nottingham. Processed. Goldin, Ian, and Dominique van der Mensbrugghe. Anderson, Kym. 1997a. The Changing International 1996. "Assessing Agricultural Tariffication under the Context of Australian Aid. CIES Policy Discussion Uruguay Round." In Will Martin and L. Alan Winters, Paper 97/06. Adelaide: University of Adelaide. eds., The Uruguay Round and the Developing --. 1997b. "Environmental Standards and Countries. Carnbridge, U.K.: Cambridge University International Trade." In M. Bruno and B. Press. Pleskovic, eds., Annual World Bank Conference on Grossman, Gene M. 1995. "Pollution and Growth: Development Economics 1996, pp. 317-38. What Do We Know?" In 1. Goldin and L. A. Washington, D.C.: World Bank. Winters, eds., The Economics of Sustainable . 1998. Agriculture and the WTO into the 21st Development.. Cambridge, U.K.: Cambridge Century. CIES Policy Discussion Paper 97/06. University Press. 82 Agricultural Trade Reforms, Research Incentives, and the Environment Grossman, Gene M., and Elhanan Helpman. 1991. Trade and the Environment. Discussion Paper 159. Innovation and Growth in the Global Economy. Washington, D.C.: World Bank. Cambridge, Mass.: MIT Press. Rosegrant, Mark W., and Claudia Ringler. 1997. Grossman, Gene M., and Alan B. Krueger. 1995. "World Food Markets into the Twenty-First "Economic Growth and the Environment." Century: Environmental and Resource Constraints Quarterly Journal of Economics 110(2, and Policies." Australian Journal of Agricultural May):353-78. and Resource Economics 41(3, September): Hayami, Yujiro, and Vernan W. Ruttan. 1985. 401-28. Agricultural Development: An International Runge, C. Forde. 1994. "The Environmental Effects Perspective, rev. ed. Baltimore, Md.: Johns of Trade in the Agricultural Sector." In The Hopkins University Press. Environmental Effects of Trade, pp. 19-54. Paris: James, Sallie, and Kym Anderson. 1998. "On the Need Organisation for Economic Co-operation and for More Economic Assessment of Quarantine/sps Development. Policies." Australian Journal of Agricultural and Schultz, Theodore W. 1980. "Nobel Lecture: The Resource Economics 42(4, December), forthcoming. Economics of Being Poor." Journal of Political Krueger, Anne O., Maurice Schiff, and Alberto Economy 88(4, August):639-51. Valdes. 1988. "Measuring the Impact of Sector- Strutt, Anna. Forthcoming. "Economic Growth, Trade Specific and Economywide Policies on Policy, and the Environment in Indonesia." Ph.D. Agricultural Incentives in LDCS." The World Bank diss. University of Adelaide. Economic Review 2(3, September):255-72. Strutt, Anna, and Kym Anderson. 1998. "Will Trade LeFrance, Jeffrey T. 1992. "Do Increased Commodity Liberalization Harm the Environment? The Case Prices Lead to More or Less Soil Degradation?" of Indonesia to 2020." Paper prepared for a World Australian Journal of Agricultural Economics Bank conference on trade, global policy, and the 36(1, April):57-82. environment, Environment Department, World Lindert, Peter H. 1991. "Historical Patterns of Bank, Washington, D.C., 22-23 April. Processed. Agricultural Policy." In P. Timmer, ed., Tyers, Rod, and Kym Anderson. 1992. Disarray in Agriculture and the State: Growth, Employment, World Food Markets: A Quantitative Assessment. and Poverty. Ithaca, N.Y.: Cornell University Cambridge, U.K.: Cambridge University Press. Press. World Bank. 1997. "Subsidy Policies and the Lutz, Ernst. 1992. "Agricultural Trade Liberalization, Environment." In Expanding the Measure of Price Changes, and Environmental Effects." Wealth: Indicators of Environmentally Sustainable Environmental and Resource Economics 2(1):79-89. Development, ch. 4. Washington, D.C.: World Martin, Will, and L. Alan Winters, eds. 1996. The Bank. Uruguay Round and the Developing Countries. World Resources Institute. 1996. World Resources Cambridge, U.K.: Cambridge University Press. 1996-97. New York: Oxford University Press. McCalla, Alex F. 1969. "Protectionism in Zilberman, David, Mahdu Khanna, and Leslie Lipper. International Agricultural Trade, 1850-1968." 1997. "Economics of New Technologies for Agricultural History 43(3, July):329-44. Sustainable Agriculture." Australian Journal of Radetzki, Marian. 1992. "Economic Growth and Agricultural and Resource Economics 41(1, Environment." In Patrick Low, ed., International March):63-80. 7 Property Regimes in Economic Development: Lessons and Policy Implications Daniel W Bromley This chapter is concerned with property regimes and all projects concerned with land and related natural property relations as they pertain to development projects resources. The centrality of property relations arises undertaken in the interest of improving the manage- not because property relations connect people to land ment of natural resources. Development projects rep- and other physical objects. Rather, property relations resent new streams of income and associated benefits are central in development because they connect peo- to a local community and sometimes to the nation- ple to one anotlher with respect to land and related state in which the community exists. But most devel- natural resources. Notice that the emphasis is not on opment projects bestow the bulk of their beneficial people and objects, but on people in relation to (and harmful) effects on those in closest proximity to objects or circumstances. Property relations are sim- the development intervention. Sometimes the local ply socially constructed contractual arrangements community will seek these new circumstances. among a group of people with respect to objects and Sometimes the development project will appear with- circumstances of value to them. Property relations are out local initiative. created by human communities to mediate individual The major difference in the long-run performance and collective behaviors regarding objects and cir- of these two types of projects can be profound indeed. cumstances of value to members of the community. One thing we know with almost striking certitude is Development projects often are less successful that when local citizens play a role in the conception than they would otherwise be precisely because of a development project, the chances for success are they ignore ex:isting property relations or, just as usually better than when projects simply appear from frequently, because they misunderstand existing prop- outside and are seen either as gifts or as impositions. erty relations. Although not central to the issues raised here, the This chapter addresses property relations in the point bears mention because it addresses a fundamen- context of developmental efforts. Some of these tal issue in the domain of "ownership"-whether we developmental efforts seek to mitigate natural think of ownership in relation to land and related nat- resource degradation. Other such efforts are con- ural resources or in relation to alterations in the social cemed primarily with increasing agricultural produc- and economic circumstances among local people aris- tion and other income-earning possibilities. And of ing from development projects. course some projects seek to do both. In the discus- Perceptions of ownership arise in the domain of sion, I summarize what we have learned over the property relations in and around the community in years in the domain of property relations in the devel- which development projects appear. These property oping world. I call attention to what has worked and relations will dominate the probability of success of what has not worked with respect to project design 83 84 Property Regimes in Economic Development: Lessons and Policy Implications and implementation and draw implications for policy Property Regimes Are Part of the Larger reform and institational design. In essence, I empha- Institutional Structure size the necessary components of a development pro- gram focused on land use and land management in the Property relations over land and related natural developing world. I begin with a section on the resources must be understood as part of the larger lessons learned before elaborating why these particu- institutional structure of a society (Bromley 1989). In lar findings are-or seem to be-true. other words, property regimes are a social construct and therefore cannot be seen as apart from the society Property Regimes: Lessons in which we address development problems. It is curi- for Development Planning ous that one must make this point, but some develop- ment planners clearly imagine that property regimes In this chapter, the term property relations and prop- stand apart from local culture. They are often likely to erty regimes are used interchangeably. Both terms are suppose, as well, that only one possible property much preferred to the more common term property structure exists-or ought to exist-throughout the rights because they focus on the full panoply of prop- world. Their favorite candidates are (1) the structure erty structures in a particular community (see box into which they (development planners) were social- 7.1). To talk of piroperty rights is to jump immediate- ized or (2) the structure that was implicit (or explicit) ly to a more legalistic and judicial idea as opposed to in the particular discipline in which they were trained. the social dimension of property regimes and proper- As social scientists, we are acutely aware that the ty relations. To deal with property rights is to enter the institutional setup of a Muslim society is very differ- domain of contention and disputes, whereas to deal ent from one that is Hindu or Christian. We readily with property relations is to address structures of con- accommodate those differences in our analysis of tractual relations among economic agents. development projects. A rural health program must be Box 7.1. Property Rights A property right is the capacity of the holder to com- for the implications accruing to others. We say that pel the authority system to come to his or her defense. the individual in such settings has privilege to use the To have a property right is to have the capacity to resource but also has no right to prevent others from require the authorities to defend your interests in an using it. An open-access resource is a free-for-all in object or a circumastance. which the rule of capture drives all users to take as Common property. A group of co-owners as a much as possible, as quickly as possible. There are no "corporate" entity has a right to exclude nonowners, property rights in open-access regimes. and that right is recognized by an authority system in State property. The nation-state maintains control which the common property regime is situated (say, of the area and will usually manage it through the the nation-state). Individual co-owners are bound by activities of a government agency-forestry service, both rights and dluties regarding behavior in respect parks service, and so forth. Individuals in the nation- of the asset. state may have the right to enter and use the resource, Open access. An open-access resource is one in with a clear duty to observe certain strictures on use. which there are no property rights for anyone. Under Private property. Individuals (or groups) have a open-access resource settings, the first individual or right to undertake socially acceptable uses and a duty group of individuals to appropriate resources to refrain from socially unacceptable uses. All others becomes, by default, their "owner." But ownership ("nonowners") have a duty to respect the boundaries under this situation arises from capture, not from of the property regime and have a right to expect that prior legitimization by the state. Under open access, only socially acceptable uses will occur. individuals are free to use a resource without regard Daniel W. Bromley 85 structured differently if it is intended for Pakistan than preallocative function of any economic system. And, if it is being implemented in Paraguay. A livestock by being preallocative, property relations must be project must look very different in northern India than understood as social constructs whose nature and in eastern Africa. A program to improve primary existence are precisely antecedent to what we ordi- schools should be quite different in Yemen than it is in narily regard as "economic" behavior. the Philippines. That is, we have learned that devel- It is indeed encouraging to see institutional opment programs and projects must be crafted to dif- arrangements receiving more attention in economics. ferent cultures and institutional structures. However, intellectual incoherence is inevitable if this Despite this cultural sensitivity to programs and recognition of the importance of institutions is projects, there is often a degree of institutional singu- allowed to proceed under the patently false notion that larity when it comes to property regimes. For some institutions are constraints on the otherwise "effi- reason, economists often find it difficult to admit that cient" workings of an economy. Institutions-proper- property regimes can and do show as much variabili- ty relations-are simply structural attributes of an ty across cultures as do other social arrangements. economy that provide agents with domains of choice The reason for this difficulty, no doubt, is that proper- within which they may act. They are, at once, con- ty relations are incorporated into economic models straints and liberation. A property right for Alpha (a and economic analysis such that there is but one way liberation) is a "duty" for Beta (a constraint). Property in which to organize behavior. Private (individual- relations exist for just this reason-to liberate Alpha ized) property is implicit in a theory of economic and to constrain Beta. The social problem, therefore, behavior that reifies individual autonomy predicated is how it shall be decided that Alpha deserves libera- on rational self-interest. Simply put, no other proper- tion and therefore that Beta warrants constraint. That ty regime is consistent with a theory of behavior question cannot be decided from within economic founded on methodological individualism. Meth- models that are themselves predicated on that institu- odological individualism is the philosophical position tional setup. Nor can that choice be decided on that believes the single agent to be the suficient unit grounds of "efficiency" for the simple reason that of analysis. Contemporary economics is built on the judgments of efficiency are themselves artifacts of the precepts of methodological individualism. prices and income distribution (and hence structures But the institutional arrangements of a society- of "effective" demand) that are the products of a par- including property relations-must be understood to ticular institutional structure (Bromley 1989, 1990). reflect prior values and expectations regarding future We are left with the realization that property rela- opportunities. Just as there is no "right" culture, there tions-property structures-must be seen as logically is no "right" property regime. Rather, there are prop- and socially prior to the economizing behavior we erty regimes responding to, indeed reflecting, mani- attempt to analyze and explain in our economic mod- fold interests and priorities. As Dahlman (1980: 85) els. The first lesson, therefore, is that we must under- observes: stand property regimes as particularistic structures In the process of defining property rights, the that gain their rationale because they are seen to economic system must make two interrelated address fundamental questions about which members decisions ... The first is to decide on the distri- of society deserve the protection from the state that bution of wealth; who shall have the rights to property relations entail. Property relations are mani- ownership of the scarce economic resources festations of whose interests count in a particular even before, as it were, trading and contracting social setting. begin. The second refers to the allocative func- tion of property rights; they confer incentives on Property Regimes Require External Legitimacy the decisionmakers within the economic system ... one set of decisions must be treated as The previous discussion provides a logical precursor endogenous for the system and constitute the to a second lesson about property relations. One of the exogenous conditions for each trading agent in very clearest lessons from the past several decades of the resulting set of trades; the second set of deci- economic assistance to the developing world is that sions is made in the context of the making of customary (and traditional) property regimes have these trades. fallen prey to the priorities and imperatives of influ- Property relations-property regimes-reflect this ential forces at the center of the nation-state. Against 86 Property Regimes in Economic Development: Lessons and Policy Implications these pressures-which can manifest themselves as The second lesson, therefore, is that any structure large-scale logging of valuable timber, the encourage- of property relations requires a commitment from the ment of extraction of minerals, the subsidization of recognized system of authority that enforcement will commercial fishing fleets, the promotion of massive be collectively assured instead of privately required resettlement projects, or the development of commer- (Bromley 1991). After all, when individuals must cial agriculture in pastoral or forest regions-tradi- enforce their own property rights, the concept of a tional property regimes are nothing. Communities of property "right" becomes a contradiction in terms. resource users cannot be expected to resist the eco- nomic and political power from the national capital in Ambiguous Rights Regimes Have Ambiguous order to stand up to such pressures for resource Efficiency and Distributional Consequences extraction. Sometimes, pressure on local property regimes is The record of development projects around the world more subtle. In particular, perhaps crop agriculture is is somewhat mixed with regard to the distributional given preference over pastoral activities. In such implications of ambiguous property rights structures. instances, the political and economic vulnerability of In some places we know that the rich will exploit the pastoralists means that the property regimes central to institutional vacuum that characterizes open-access pastoralism will be undermined. If a development resources and even some common property resources project is aimed at improving range and forest condi- (Blaikie, Harriss, and Pain 1992 and Jodha 1986). Yet tions for pastoral peoples, then the property regimes we also know that some ambiguity in property regimes central to this particular economic activity must be can work to the advantage of persons who do not con- recognized by the authority system (the nation-state) trol private property resources, which are rarely, if as a precondition of the project. It is not enough that ever, ambiguous. That is, some natural resources over pastoralists have traditional grazing "rights" in a par- which rights are ambiguous are available to individu- ticular area. If the property regime central to pastoral- als who would be excluded if the resources were more ism is to survive incursion by outsiders, then pas- valuable and therefore vulnerable to expropriation by toralists cannot be expected to mobilize their own the powerful elements in a community. The very mea- defense of the assets central to their survival. After all, gerness of the bounty from such assets means that the holders of private property are not expected to defend better-off elements in society do not bother with them their own claims to ownership; the authority system (Cordell and McKean 1992). does that for them. Although we know that in most However, when development projects increase the urban places in the developing world, the inattention income flow from natural resources over which rights and unreliability of official enforcement structures had previously been ambiguous, powerful individuals (local police) mean that protection of private property at the local level will usually figure out a way to (homes and businesses) has necessarily been "pri- expropriate at least a part-if not the majority-of the vatized" by those who can afford it. The rest, largely new income streams. Hence, although ambiguity can without assets worth serious protection, take their often work to the advantage of the dispossessed, it is chances. Indeed, to have a property right is to know unlikely to do so when development projects are that the authority system will come to your defense introduced into this kind of institutional environment. against the predatory behavior of others interested in Therefore, the third lesson suggests that development your situation or circumstances. projects introduced into settings in which property But successful development projects and programs relations are highly ambiguous will most probably require more than willingness of the authority system have economic and social impacts quite different to defend the property regimes in a particular location. from those postulated in the proposals and feasibility Members of the community will also have to have analyses that led to the project's acceptance and developed local structures of authority to manage nat- implementation in the first place. ural resources in a sustainable manner. Very often development projects come to those areas in which Property Relations Must Be Specified Prior resource degradation is rather advanced. And just as to the Implementation of Development Projects often, part of the reason for that degradation is the breakdown of authority over the resource-by internal A development project is a "relational contract" involv- conflict or by an iintrusion of outsiders into the area. ing three entities: (1) the donor agency, (2) the nation- Daniel W Bromley 87 state (the co-sponsor) into which the project is to be ing property regimes) need not entail every possible placed, and (3) the local individuals whose lives we detail of future circumstances associated with the hope will be altered for the better. We must recognize development project. But if certain general principles that each of these entities will always have its own are not spelled out, problems will certainly surface as agenda and that only in the rarest of circumstances will the benefits of the development intervention begin to those agendas be entirely coincident and mutually appear. To prevent this, it is good to think in terms of enforcing. (This is offered not as an assertion of impro- a general framework out of which greater detail will priety but as an empirical observation of great import to emerge. For instance, if a reforestation project is the ultimate success of development efforts.) But dis- planned, it is important to have in place a general pro- parate agendas suggest disparate perceptions of ends gram of protection and management before the trees and means-of objectives and of instruments. are planted. Such a framework needs to specify who Although the authority structure of donors and is responsible for monitoring use of the new plantings, nation-states assures that the implementation of a what uses are acceptable during which times of the project will appear to be unified and in harmony, we year (and when they may start), and how breakdowns encounter no such hegemony at the local level. in compliance will be handled. Indeed, we place projects in areas that-as a general For a project to rehabilitate an abandoned irriga- rule-warrant outside assistance precisely on account tion system, a "'constitution" should be worked out of the incoherence and contentious nature of local that specifies the order in which irrigators spread out institutional arrangements-of which property rela- along the major and minor distribution arteries shall tions are of paramount importance. Those conditions receive water ill any particular cycle of water avail- of institutional incoherence are, after all, often related ability. It will also be necessary to make sure that to situations of chronic poverty, serious natural rules of maintenance are spelled out, in a general way, resource degradation, and the inability of local people before the first water flows through the system. to undertake and to sustain adequate material provi- Of course we know that this is easier said than sioning. That is the "development problem." done. Or, more correctly, it is more easily done than Institutional dissonance and ambiguous property enforced. This brings us to the issue of the design and relations almost certainly favor the already advan- implementation of sustainable interventions (in con- taged when the benefits of a development project trast to sustainable development). Issues of sustain- begin to appear. And the record of development inter- able interventions take us, inevitably, back to a dis- ventions over the past several decades suggests that cussion of the larger institutional context within new income streams arising from development projects which the need for development projects arises. have a difficult time finding their way into the hands and pockets of those most in need of help. The puz- Resource Degradation Is the Result of Problems zlement about persistent poverty in the wake of That Precede P,roperty Regimes decades (and billions of dollars) of "development" projects is really no puzzle at all. The poor remain There is a genre of literature that seeks to blame cer- poor because the institutional arrangements rendered tain property regimes for widespread degradation of them poor prior to the development intervention, and natural resources. Most often, resource degradation is there are durable pressures-and nontrivial individu- laid at the feet of so-called "common property." als-to make sure that the mere advent of a "develop- Although there are indeed common property regimes ment project" does not somehow upset the institution- in which resource degradation is prevalent, the more al arrangements that created the current structure of probable institutional setup is one of "open access" in economic advantage in the first instance. which no property rights exist. But situations of open The fourth lesson, therefore, is that every develop- access-either in a narrow legalistic sense or in an ment intervention must be preceded by a concerted operational sense-are themselves manifestations of effort to ensure that the institutional arrangements larger institutional problems. have been modified so as to make sure the benefits go Hence, a more realistic assessment and diagnosis to the intended beneficiaries. This modification of the would indicate that resource degradation arises for "working rules" of the local economy must occur many reasons--only one of which may relate to the before the new benefit stream begins to materialize. particular property regime in place (Deacon 1994, This respecification of the working rules (includ- 1995). Much of the soil erosion from agricultural 88 Property Regimes in Economic Development: Lessons and Policy Implications lands occurs on 1private land; this is as true in the Ecological Variability Demands Flexible developing world as it is in the Corn Belt of the Institutions and Actions United States. It is curious indeed that private proper- ty is not blamed for soil erosion in Nebraska, while The instrumental nature of institutional arrangements, common property is blamed for soil erosion in including property regimes, is best demonstrated by Namibia. When Nebraska farmers allow 15 tons of thinking about the intersection of the realms of ecolog- soil per hectare annually to wash into the Niobrara ical variability and the holding of social and economic River, the blame is said to lie with the wrong technol- capital. In temperate climates, fixed assets such as land ogy (moldboard plowing), with a rate of time prefer- hold great economic potential and social status. Small ence that is too high, or with ignorance of the future wonder that we find in such places elaborate institu- yield-reducing effects of such erosion over the long tional and technological structures to define and control run. But when Namibian farmers allow 15 tons per access to, and control over, land. Most residents of tem- hectare of soil annually to wash into the Konkiep perate climates regard it as quite normal to view a fixed River (or to blow into the South Atlantic), the blame land base as both an economic asset and a source of goes to the fact that they are farming under a "primi- social status. These same individuals have been known tive and quaint" common property regime. This asym- to offer a bemused smile when told that African pas- metry reveals moire about the ideological predisposi- toralists gain social status from holding not land but tion of the commentator or analyst than it does about cattle. But of course this quaint practice is a product of the causal structure in each setting. their "primitive stage of development," it is said. These particular resource management "out- However, the fact that cattle are privately owned comes" are largely explained by existing institutional and land is not says less about stages of "develop- structures, of which property regimes are but a part. ment" than it does about the ecological reality within The outcomes are explained, as well, by existing cus- which Africans must make a living. This institutional toms and norms--and economic imperatives-that structure reflects an adaptive response to the reality of are often much rnore important than the particular provisioning where soils are poor, rainfall is fickle, property regime in which both farmers operate. and irrigation is not feasible. Pastoralism is a response The lesson we draw from this is that before offer- to these circumstances, and the institutions of ing development advice, we must build structurally resource management are a constructed overlay that coherent models that gain their legitimacy from their allows herding to function in such settings. It is not capacity to explain observed phenomena as opposed surprising that this structure of flexibility does, at to being merely consistent with observed phenomena. times, bump up against an alternative property rights For example, constructing a simple model showing structure-private property-constructed for rigidity. that "common property" will result in more erosion But institutional flexibility is a necessary attribute of than "private property" explains very little. More seri- certain economic systems (Cousins 1995, Behnke ously, such models-when employed in the service of 1994, and Behnke and Scoones 1992). prediction-quickly become the basis for normative Conflicts between pastoralists and sedentary agri- prescriptions without the benefit of empirical evi- culture in the arid climes must be understood as a con- dence or conceptual logic. Once someone has built a flict over property regimes. According to van den model "proving" that common property is more con- Brink, Bromley, and Chavas (1995: 374): ducive to erosion than private property, it is not a long In agriculture as well as livestock production, reach to predict that all common property regimes property rights emerge to secure income streams will have erosion in excess of the erosion from eco- generated by production activities. The nature of logically comparable private property regimes. And the income stream, then, may affect the type of from there, the normative element slips in to claim property right that is likely to be established. that all common property regimes should be convert- The crucial difference between sedentary farm- ed to private property in order to reduce resource ing and nomadic livestock production is that degradation in general and soil erosion in particular. they differ in their ability to react ex post to tem- The lesson here is that we must be careful to avoid poral uncertainty; in other words, they differ in building flawed causal models that become the basis flexibility. for suggesting false solutions to land management Unfortunately, property rights essential for live- problems. stock production in the Sahel have been eroded by a Daniel W. Bromley 89 long history of conflicts. More recently, a number of robust as opposed to scrawny plants, carry no policy state interventions have expropriated pastoralist prop- message whatsoever. erty rights crucial to their economic systems and The general "condition" of a specific natural clearly favored farmers over pastoralists in the alloca- resource in a particular place is a socially constructed tion of private property rights. These changes have concept. The policy relevance enters when different created general uncertainty over property rights to constructions become associated with the same eco- natural resources, thereby inducing a de facto open- logical reality. access situation. The resulting tragedy of open access, But the lesson we have learned over the past sev- induced by public policy, has substantially increased eral decades of development interventions in the the (transaction) costs of running the pastoralist econ- domain of land and related natural resources is that omy and adversely affected the pastoralists' ability to the only coherent social construction is that which overcome periods of drought (van den Brink, derives from those whose existence is most directly Bromley, and Chavas 1995). connected to the resources. Of course, technical The lesson for sustainable resource management is experts can help to ensure that local people grasp the that ecological settings that exhibit great variability prevailing scientific explanations (or conjectures) for require property regimes that allow quick and low- particular ecological outcomes. But given the large cost adaptations to new circumstances. Livestock as number of scientific "truths" that no longer command an asset, because of their mobility, provide this flexi- even minimal assent, such "technology transfer" must bility in a way that the immobility of land cannot pos- be quite cautious and circumspect. sibly equal. Small wonder that the institutions over livestock and land differ so profoundly between the Land Titles Are Not Necessary for Efficient temperate climates of the middle latitudes and the arid Investments in Productivity and semiarid reaches of Africa and Asia. I now wish to discuss the apparently self-evident Resource Degradation Is Contextual proposition that a "secure title" to land is a necessary condition for investment in, or the "wise" manage- Once we recognize that the social and economic ment of, land and related natural resources (Feder and meaning of various assets differs across ecological Noronha 1987, Mendelsohn 1994, and Southgate and social circumstance, then it follows automatically 1988). Interestingly, this "truth" is less true than many that the social and economic meaning of what is a imagine (Place and Hazell 1993). In fact, the contrary "resource" also differs. The social and economic proposition-that investment in land is a necessary meaning of a resource is something that brings value; condition to secure "title" to that land-is equally a resource is an "input" into something from which probable. Rather than investment requiring security, benefits flow. But this means that "resource degrada- security requires investment (Sjaastad and Bromley tion" cannot be understood without reference to the 1997). prior notion of what, exactly, is the set of resources to The confusion in this domain has to do with the which a particular society is predisposed to pay close precise meaning of "security" and "title." It seems to attention. be the experience in some parts of the world that the Overgrazing is a term that clearly derives its oper- titling and registration of land were the immediate ational content from a larger social and economic precursor to the dispossession of individuals who context. Unless one treasures particular desert plants imagined that this step would ensure their longevity for their own sake, their relative paucity in particular on the land. Instead, such title became a means where- locations is without normative significance in a devel- by moneylenders and others with some measure of opment context. Of course more and better plants political influence were able to acquire what the poor would allow more and better livestock, that would once thought was "their" land. Rather than enhancing then allow-and the chain is getting more tenuous as security, title had quite the opposite effect. But, of we proceed-more and richer herders, that would course, title served to establish security for those who allow more economic surplus to flow to urban areas, managed, by whatever means, to gain control of land that would allow "lift-off' to some state-defined that had heretofore been controlled by chiefs and "development." Absent that teleological sequence, other local notables. So the issue is not one of mere more or fewer plants in some comer of the Sahel, or title and security, but of security for whom. 90 Property Regimes in Economic Development: Lessons and Policy Implications To those trained in the modem scheme of things, it tive problem, such arrangements may still require an seems self-evident that individuals will only invest in oversight role for some higher authority if there is a some asset if their control over that asset is recog- national or international interest in the particular nized and secure. However, to return to the very first resource (Arya and Samra 1995). point about property regimes being part of a larger Still, an important lesson learned over the years is institutional structure, in some settings, Alpha may that the chances of sustainability are enhanced as a wish to invest in land for the benefit of Beta. When direct function of the extent to which local people this is the case, it is not Alpha's future that must be acquire "ownership" in the future benefit stream asso- secured, but Beta's. ciated with a particular natural resource. The lesson he.re is that we must be careful about the direction of causality when discussing fundamen- Implications for Development Policy tal economic behavior among disparate peoples in widely scattered corners of the developing world. To I have spelled out nine lessons for persons working in suppose that investment is stifled because bankers the general area of land rights and resource manage- cannot see a clear title against which a loan might be ment. In essence, these lessons pertain to (1) the offered is to impose a particular constellation of social embeddedness of property regimes, (2) the con- assumptions on a setting that has an entirely different cept of a "resource" and resource "degradation," (3) structural makeup. the social instrumentality of institutions such as prop- erty, and (4) those institutional arrangements that Mobilizing Local Interests Can Improve seem to enhance the success and sustainability of the Chances of Program Success development interventions. We must keep in mind that development endeavors The importance of ownership was suggested at the are both threatening to certain interests and also very outset. Many development programs and projects rewarding to others. Property regimes play the central require that established patterns of human interac- role in directing benefits and costs to different groups tion-and traditional resource uses-be modified. We at the local level. This means that the outcomes of know that efforts to impose such changes from out- development activities depend fundamentally on side are not sustainable, and this means that the bene- institutional arrangements in general and on property fit streams intended from such interventions will not relations in particular. materialize over the long term. This should not be surprising. After all, the institu- In natural resource projects, there is growing tional structure is largely responsible for the condi- recognition that "'community-based conservation" is tions that lead to the need for development assistance an important innovation in creating the conditions for in the first instance. Ameliorative activities in the way sustaining both natural resources and the local com- of development programs and projects cannot possi- mitment to sustainability (Western, Wright, and Strum bly succeed without careful analysis of these institu- 1994). The essence of community-based conservation tional precursors of current problems and without a is to reconstitute the incentives at the local level so clear program of correction as a precursor to project that those closest to the resource are given a greater implementation. stake in its long-run viability (Bromley 1994). In effect, those closest to the resource are given an References "ownership" interest in its future value. That approach generally gets the incentives right. The word "processed" describes informally repro- Of course, this is necessary but not sufficient to duced works that may not be commonly available ensure improved behaviors with respect to resources. through libraries. For instance, it is possible that an ownership interest Arya, S. L., and J. S. Samra. 1995. "Participatory in the future viability of particular natural resources Process and Watershed Management: A Study of simply allows for its more efficient degradation. After the Shiwalik Foothill Villages in Northern India." all, we have seen that an ownership interest in topsoil Asia-Pacific Journal of Rural Development 5(2, is no guarantee that it will not be washed away. So December):35-57. although creating a sense of ownership among locals Behnke, R. H. 1994. "Natural Resource Management in particular resources can solve much of the incen- in Pastoral Africa." Development Policy Review Daniel W. Bromley 91 12(1):5-27. Press. Behnke, R. H., and Ian Scoones. 1992. "Rethinking Deacon, Robert T. 1994. "Deforestation and the Rule Range Ecology: Implications for Rangeland of Law in a Cross-Section of Countries." Land Management in Africa." Environment Department Economics 70(4, November):414-30. Working Paper 53. World Bank, Washington, D.C. . 1995. "'Assessing the Relationship between Processed. Government Policy and Deforestation." Journal of Blaikie, Piers, John Harriss, and Adam Pain. 1992. Environmental Economics and Monagement "The Management and Use of Common- Property 28(1):1-18. Resources in Tamil Nadu, India." In Daniel W. Feder, Gershon, and Raymond Noronha. 1987. "Land Bromley and others, eds., Making the Commons Rights Systems and Agricultural Development in Work: Theory, Practice, and Policy, ch. 11. San Sub-Saharan Africa." The World Bank Research Francisco: ics Press. Observer 2(2, July):143-69. Bromley, Daniel W. 1989. Economic Interests and Jodha, N. S. 1986. "Common Property Resources and Institutions: The Conceptual Foundations of Rural Poor in Dry Regions of India." Economic Public Policy. Oxford: Blackwell Publishers. and Political Weekly, 5 July, 21(27):1169-81. 1990. "The Ideology of Efficiency: Searching Mendelsohn, Robert. 1994. "Property Rights and for a Theory of Policy Analysis." Journal of Tropical Deforestation." Oxford Economic Papers, Environmental Economics and Management supplement, 46(5, October):750-56. 19(1):86-107. Place, Frank, and Peter Hazell. 1993. "Productivity . 1991. Environment and Economy: Property Effects of Indigenous Land Tenure Systems in Rights and Public Policy. Oxford: Blackwell Sub-Saharaii Africa." American Journal of Publishers. Agricultural Economics 75(February): 10-19. .1994. "Economic Dimensions of Community- Sjaastad, Espen, and Daniel W. Bromley. 1997. Based Conservation." In David Western, R. "Indigenous Land Rights in Sub-Saharan Africa: Michael Wright, and Shirley C. Strum, eds., Natural Appropriation, Security, and Investment Demand." Connections: Perspectives in Community-Based World Development 25(4):549-62. Conservation, ch. 19. Washington, D.C.: Island Press. Southgate, Douglas. 1988. "The Econornics of Land Cordell, John, and Margaret A. McKean. 1992. "Sea Degradation in the Third World." Environment Tenure in Bahia, Brazil." In Daniel W. Bromley and Department Working Paper 2. World Bank, others, eds., Making the Commons Work: Theory, Washington, D.C. Processed. Practice, and Policy, ch. 8. San Francisco: ics Press. van den Brink, Rogier, Daniel W. Bromley, and Jean- Cousins, Ben. 1995. A Role for Common Property Paul Chavas. 1995. "The Economics of Cain and Institutions in Land Redistribution Programmes in Abel: Agro-Pastoral Property Rights in the Sahel." South Africa. Gatekeeper Series 53. London: Journal of Development Studies 31(3):373-99. International Institute for Environment and Westem, David, R. Michael Wright, and Shirley C. Development. Strum, eds. 1994. Natural Connections: Dahlman, Carl. 1980. The Open-Field System and Perspectives in Community-Based Conservation. Beyond. Cambridge, U.K.: Cambridge University Washington, D.C.: Island Press. 8 Equity and Environmental "Modifiers" for Rural Development Policies Peter Hazell Agriculture is different from other sectors in many opment experts tended to be interested in agricultural developing countries. In addition to being an impor- growth, and the lessons that emerged from their expe- tant production sector, it is the principal means of rience can be summarized as the 5 "I"s for agricultur- livelihood for most of the poor, it plays a key role in al growth (or for getting agriculture "moving," as ensuring food security, and, because a nation's farm- Mosher aptly stated it): ers are the custodians of many of its natural resources, * Innovation, the need for strong national agricultural it has a direct bearing on how these resources are research and extension systems (public and private) managed. For these reasons, agricultural development to generate and disseminate productivity-enhancing should simultaneously contribute to four principal technologies goals: growth, poverty reduction, food security, and * Infrastructure, the need for good rural infrastruc- sustainable natural resources management. ture, particularly roads and transport systems Experience suggests that there can be a high * Inputs, the need for efficient systems for delivering degree of complementarity between these four goals agricultural services, especially farm inputs, agro- if agricultural development (a) is broadly based and processing, irrigation water, and credit involves small and medium-size farms, (b) is market- * Institutions, the need for efficient markets that pro- driven, (c) is participatory and decentralized, and (d) vide farmers with unfettered access to domestic is driven by productivity-enhancing technological and intemational markets and the need for effec- change that does not degrade the resource base. Such tive public institutions to provide key public ser- growth reduces food prices and increases the effective vices demand for nonfood goods and services, particularly * Incentives, the need for conducive govemment in rural towns and market centers. By reducing pover- policies that do not penalize agriculture. ty and promoting economic diversification in rural It later became evident that governments also need areas, it also relieves pressures on the natural resource to promote the role of the private sector in agricultur- base. al production, marketing, and service provision. The requirements for broadly based agricultural In the 1970s and 1980s, attention focused on using development are reasonably well understood and agricultural development to reduce poverty and food should not be forgotten in our contemporary quest for insecurity as well as to contribute to growth. The "sustainability." Because they are so important, I lessons that emerged from that area can be summa- briefly review them before discussing sustainability rized as follows. In addition to the 5 "I"s, the follow- concerns. ing six "equity modifiers" are needed for agricultural Back in the 1950s and 1960s, agricultural devel- growth: 92 Peter Hazell 93 * Broadly based agricultural development. There are sive farming systems. This misuse is related to poor few economies of scale in agriculture, hence tar- management of inputs by public agencies (some irri- geting family farms is attractive on efficiency and gation departrnents), inappropriate polices that equity grounds. Small and medium-size farms encourage farmers to overuse inputs (fertilizer and need to be given priority in agricultural research pesticide subsidies, underpricing of water), and exter- and extension and in marketing, credit, and input nality problems that lead farmers to undervalue envi- supplies. ronmental costs and benefits (water pollution, * Participation. All rural stakeholders (not just the destruction of beneficial insects). rich and powerful) need to participate in the prior- The second type of environmental problems arises itization of public investments from which they are in extensive farming systems and is associated with expected to benefit or help finance. rapid population growth, poverty, and growth rates in * Land reforms. Market-assisted redistribution pro- agricultural productivity that is insufficient to meet grams, in particular, may be needed where too increasing need for food livelihood. Farmers in these much land is concentrated among large farms. areas typically use low levels of modem inputs (and * Investments in human capital. Investments in rural have limited opportunities to misuse inputs), and, education, clean water, health, family planning, because yield grows too slowly, they reduce fallows and nutrition programs are needed to improve the and expand cultivation into marginal and environ- productivity of poor people and increase their mentally fragile areas. Poverty and insufficient agri- opportunities for gainful employment. cultural intensilication are the fundamental problems, - Women. Because women play a key role in farm- but resource degradation is often worsened by inse- ing and ancillary activities, targeted programs are cure property rights, extemality problems, high costs warranted in agricultural extension and education of collective action for managing common properties as well as credit and small nonfarm business assis- or undertaking resource-improving investments, and tance programs. inappropriate govermment policies. * The rural nonfarm economy. The nonfarm econo- Recognizing these differences, I propose the fol- my in rural areas should be actively encouraged lowing eight environmental modifiers for "sustain- because it is an important source of income and able" agricultural development, these have yet to be employment, especially for the poor, and because fully worked out: it benefits from powerful income and employment * Give greater priority to backward regions in agri- multipliers when agriculture prospers. In many cultural development, even though many of these countries, these potential multiplier effects are may be resource-poor. Given rapid population constrained by investment codes and related legis- growth and Limited nonfarm opportunities, agricul- lation that discriminate against small, rural non- tural growth often is the only viable means of farm firms. meeting the food and livelihood needs of growing The new priority for environmental sustainability populations in these areas for the next few decades that has emerged in the 1990s does not negate the without excessive outmigration that adds to need for agriculture to continue contributing to already overloaded urban slums. Failure to do so growth, poverty alleviation, and increased food secu- will lead to worsening poverty and further degra- rity; it is just that agriculture is now required to do this dation of hillsides, forests, and soils. This will in ways that do not degrade the environment. To require providing additional resources for agricul- achieve this, we will still need the five "I"s and the six tural development rather than diverting resources equity modifiers. There are no shortcuts here. But from favorably endowed agricultural regions. some new environmental modifiers for sustainable * Give greater attention in agricultural research to agricultural development are now required. These sustainability features of recommended technolo- modifiers have yet to be worked out and tested gies, to broader aspects of natural resource man- through development experience, a process that in agement at the watershed and landscape levels, many ways is still at the research and design stage. and to the problems of resource-poor areas. Two fundamentally different types of environmen- * Ensure that farmers have secure property rights tal problems are associated with agriculture. The first over their resources. This does not necessarily type is associated with the misuse of modern inputs imply that government should invest in ambitious (irrigation water, fertilizers, and pesticides) in inten- land registration programs. In many cases (such as 94 Equity and Environmental "Modifiers" for Rural Development Policies Sub-Saharan Africa), indigenous tenure systems (irrigation and forestry department). Devolve man- still work surprisingly well and are better able to agement decisions to resource users, or groups of meet equity needs and recognize multiple-user users, wherever possible. This also requires the rights than fully privatized property rights systems. transfer of secure property or use rights. * Privatize common property resources or, where * Correct price distortions that encourage excessive this is not desirable (because of externality bene- use of modern inputs in intensive agriculture fits or for equity reasons), strengthen community (remove subsidies on fertilizers and pesticides, management systems. charge full costs for irrigation water and electrici- * Resolve externality problems through optimal ty). It may still be necessary to subsidize fertilizers taxes on polluters and degraders, regulation, in backward regions where current use is low and empowerment of local organizations, or appropri- soil fertility is being mined. ate changes in property rights. Free market prices * Establish resource monitoring systems to track are not always best; externalities may require opti- changes in the condition of key resources, educate mal tax or subsidy interventions. farmers about the environmental effects of their * Improve the performance of relevant public insti- actions, and delineate and protect sites of particu- tutions that manage and regulate natural resources lar environmental value. Part II Institutional and Social Perspectives In part L, we take mainly a policy perspective and Getting local groups and organizations to become argue that a sound policy environment is essential for self-managing organizations can extend over several sustainable agricultural development because, among years and does not happen without investment in other things, it creates an incentive framework that capacity building. Many programs have assumed that induces farmers to innovate. Here in part II, we take a geographic community is synonymous with commu- more institutional and social focus and make the point nity of interest. But communities are rarely homoge- that creation of appropriate institutional frameworks neous entities; nevertheless, the need to solve urgent and consideration of the social dimensions are equal- priority problems may bring different class and power ly important because they provide the structure or fab- groups together. The problem with existing social ric within which policies are applied and within which organization is that it is generally invisible and often farmers work. excludes women and the poor from the most impor- Limited success of central, top-down approaches to tant production and decisionmaking networks and development has led to recognition of the need for par- associations. It is important to understand the existing ticipatory community-driven development (Narayan, social system so as to draw on its strengths and embed chapter 9). Community-driven development is a new organizational forms in the existing social milieu. process in which community groups initiate, organize, Local elites often take leadership roles, and although and take action to achieve common interests and this is not necessarily bad, it can result in the hijack- goals. No single model is appropriate for all places ing of resources unless transparency and accountabil- and times. But successful community-driven devel- ity are somehow enforced. opment generally is characterized by five factors: The most important criteria determining appropri- local organizational capacity or the existence of ateness of technology are local management capacity viable community groups, an appropriate fit of tech- and the influence that the introduction of a technolo- nology to community capacity, effective agency out- gy has on social cohesion and benefit flows. Thus reach strategies, client-responsive agencies, and technology needs to be viewed in a holistic way to enabling policies. determine whether it fits into the existing social sys- Local organizational capacity is embedded in the tem and the direction of change that its introduction norms that enable groups of people to trust one anoth- will set into motion. In natural resource management er, work together, organize, solve problems, mobilize and agricultural projects, technological considerations resources, resolve conflicts, and network with others are important when considering what to plant, how to to achieve agreed goals. Viable community groups are harvest, and how to monitor. key to the success of community-driven development. Outreach mechanisms can be classified broadly 95 96 Part II. Institutional and Social Perspectives into two basic approaches, the extension approach and Community-driven development on a large scale the empowermen.t approach. The two often overlap, requires enabling policies and political support to and an effective outreach strategy contains elements protect agencies so they can initiate the reform of both. The extension approach is characterized more process and give it time to take root. It is much easier by information Idissemination, demand promotion, to support in the context of decentralization strategies. and delivery of particular services or inputs. The Although individuals play key roles in starting or pro- empowerment approach has the main goal of empow- tecting the reform process while it is still in the early ering local groups or creating local organizational stages, change can be sustained only if the rules and capacity for self-management through involvement in regulations at different levels provide incentives to decisionmaking. The empowerment approach uses work and organize. staff more intensively and usually requires greater For designing participatory community-based investment in local capacity building. In the extension development projects, the following 10 lessons from approach, the agency's desired outcomes are the use worldwide experience should be reflected in the of inputs and services delivered and increased effec- design of new projects, subject to testing with pilot tiveness and efficiency of programs. In the empower- activities: (1) develop clearly stated objectives logi- ment approach, the key outcomes are that households cally linked to strategies, outputs, indicators of suc- or groups at the community level have organized for cess, and physical or capacity-building outcomes, (2) self-management and show increased capacity for identify the key actors and their capacity and interests coordinated action, including management of local (through social assessment), (3) assess demand (as services or resources. measured, for example, by commitment before con- Client-centered agencies are needed to ensure that struction and substantial financial investment up-front demand is met and community self-management is in capital cost), (4) craft a self-selection process (with adequately supported by community fieldworkers. To a significant financial or organizational contribution become client centered, agencies must be convinced up-front), (5) structure subsidies that do not distort to change their rules and incentive structures so that demand (for example, by asking the poor to contribute the benefits from client responsiveness are more substantial amounts through their labor, (6) restruc- important to the agency than the costs. Experience ture the release of funds to support demand (such as demonstrates that trying to support community-based by fiscal decentralization), (7) plan for learning and approaches without fundamentally changing institu- for a plurality of models (the project must be seen as tional structures, rules, and incentives does not work. likely to adjust and change over time as local priori- Experience also shows that strong political and gov- ties change and as local organizations mature), (8) emmnent support at the highest levels generally is invest in outreach mechanisms and social organiza- needed for public agencies to become client centered. tion (investments are needed, and there are no short- A critical question is how to bring about govern- cuts to strengthening local social organization for col- ment commiitment to change. World Bank staff have lective action), (9) use participatory monitoring and used two main strategies to generate interest and com- evaluation (listening to people's suggestions is mitment: pilot projects as instruments to demonstrate empowering and encourages innovation and responsi- alternative strategies and a variety of participatory bility), and (10) redefine procurement rules, where techniques, including field visits to other regions and necessary (so that they do not hinder community ini- countries, to generate interest and a commitment to tiative). new ideas and action strategies. Indicators of success Ashby and others (chapter 10) illustrate these par- are also important because they send important sig- ticipatory principles by examining the experience nals to staff about program priorities. If community involving local organizations in watershed manage- involvement, number of women or the poor being ment. Their discussion is oriented toward providing reached, and number of functioning systems are not guidelines rather than a theoretical analysis. Also, it is reflected in indicators of success, there is little incen- based on experience of working with local organiza- tive for staff to change their way of doing things to tions from the perspective of development practition- reach these goals. Almost all projects that adjust as ers. This experience is evolving and accumulating they are implemented invest in numerous small stud- rapidly, and many practitioners are promoting a learn- ies that provide feedback on how different approach- ing process in which their approach to local-level es are working. watershed management is adapted and reoriented Part II. Institu,tional and Social Perspectives 97 swiftly based on analysis of what has gone before. other instances, entire structures are leveled as soon as Thus, although this discussion focuses on some of the the outside agencies imposing them leave the village. lessons learned and principles that can be derived In contrast, innovative management practices are from this experience, Ashby and others believe that rapidly developed and adopted when local organiza- the last word on what works and what does not work tions involve iarmers in experimenting with and is by no means decided. Indeed, the dynamic nature of adapting the principles of known conservation tech- the process of involving local organizations in water- niques to their own needs and constraints and in com- shed management is one of its most important charac- bining this information with local knowledge of the teristics and one that requires a focus on managing the landscape, soils, and vegetation. process rather than on defining a blueprint for how to Few local organizations at the watershed scale go about it. have arisen completely spontaneously without some Watershed management, especially in low-income intervention from outside agencies to catalyze their countries, usually involves a combination of conser- formation and action. It appears to be a characteristic vation and development of natural resources. of local organization on a watershed scale that out- Watershed inhabitants depend on these resources for siders play a significant role in enabling stakeholders their livelihoods, and one of the basic challenges is to to arrive at a joint plan of action that takes into establish and maintain management practices that rec- account transboundary effects not readily perceived oncile their economic needs with the long-run conser- or measured by stakeholders other than those directly vation of soil, water, and vegetation, especially where affected. resource degradation has significant off-site effects, By providing critical information about trans- such as the silting up of watercourses or reservoirs of boundary effects, external agencies can enhance the interest to influential stakeholders downstream. On- capacity of local organizations to mobilize local coop- site and off-site degradation of watersheds is wide- eration in watershed management. Another important spread, and the management practices for conserva- point of connection between local organizations and tion and regeneration are technically well understood: external agencies in watershed management is the land use planning to designate certain areas for con- need for external intervention in identifying the rele- servation, grazing, annual or perennial crops, and nat- vant stakeholders (which often have conflicting ural or managed forest; use of contour barriers on objectives) and bringing them "to the table," whether steep slopes; use of ground cover; use of agroforestry this refers to a negotiating table or to participation in and agrosilvopastoral systems; creation of buffer collective labor, monitoring, or enforcement of sanc- zones around watercourses and springs; and minimum tions. The involvement of local organizations in tillage, green manuring, and other soil improvement watershed management does not necessarily ensure practices that regenerate soils and permit the intensi- equitable participation of all relevant stakeholders in fication of agriculture where suitable. However tech- a resource management initiative (Ashby and others, nically desirable these practices, their adoption by chapter 10; a point also made by Narayan, chapter 9). resource-poor fanners is very uneven, in many cases Case studies show that women, the landless, marginal because the short-term economic returns to investing ethnic groups, and the laboring poor are all likely to in these practices exceed the costs for the individual be unrepresented in participatory watershed manage- farmer. As a result, top-down efforts to enforce ment. Leadership in local organizations is generally mandatory use of these practices or to promote them captured by the higher status, more well-to-do mem- using temporary subsidies have by and large failed to bers of dominant elites, and the priorities of local achieve lasting adoption. organizations tend to reflect their interests. Watershed Involving local organizations in designing techni- development projects should deliberately seek to cal innovations for soil and water conservation and in overcome this bias by setting up catchment commit- planning where to locate them in the landscape of a tees with representation of underprivileged groups. watershed is now recognized as a key element of suc- Ashby and others (chapter 10) highlight these impor- cessful adoption. There are many examples of tant principles with information from a case study enforced use of soil and water conservation practices tracing the evolution over time of local-level water- leading to accelerating erosion once the local people shed management in the Rio Cabuyal in Colombia. stop maintaining the structures, because they have no Participation is also one of the keys to successful commitment to keeping the structures in place. In technology generation among low-income farmers 98 Part 11. Institutional and Social Perspectives (Farrington and Thiele, chapter 11). In many areas of clients during the identification, testing, and dissemi- rainfed agriculture, where agroecological conditions nation of technologies. Significant organizational and are complex, diverse, and risk-prone and where stan- management challenges must be overcome if multi- dards of physical (communications, transport) and agency approaches and participation are to become social (education, health) infrastructure are low, spe- part of the mainstream of government research and cific efforts will have to be made to elicit farmers' extension strategy (Farrington and Thiele, chapter perceptions of needs and opportunities. In these areas 11). the private sector undertakes little research and devel- NGOS providing development-related services tend opment activity even though this is needed the most. to follow intensive face-to-face approaches with Attention has therefore turned to ways of improving small, local groups, often over a period of years. This the effectiveness of the kinds of research institutions empowering approach to participation allows NGOS to likely to dominate these areas in the coming decades, take the moral high ground and to dismiss efforts that namely those in the public sector. A radical view is do not reach these standards. However, from a wider that public sector research services are so ineffective policy perspective, govemment organizations cannot that they should be shut down and replaced with a mandate themselves to concentrate resources in a few low-cost but efficient capacity to "borrow" ideas. But villages as NGOS do, and they have to make the best of even then, capacity is still needed to adapt and test the difficult job of achieving whatever level of partic- "imported" technologies under local conditions. ipation is feasible through a thin spread of resources The public sector has generally performed poorly over a wide area. This leads government organiza- in difficult areas :for several reasons: (a) inadequate tions to seek functional participation to perform tech- orientation toward clients and participation by them in nology generation and dissemination more efficiently. the design and implementation of research and (b) Clearly, at times there are gray areas between func- insufficient interaction with local organizations hav- tional and empowering forms of participation, which ing close contact with farmers (nongovernmental represent different ends of a spectrum rather than organizations, membership organizations, develop- uniquely distinguishable categories. Nevertheless, the ment projects). Local organizations can create a distinction is important in policy contexts, not least "demand pull" on the public sector by helping to because each approach tends to be associated with articulate farmers' requirements. very different kinds of organizations. Multiagency approaches-not replacement by the Farrington and Thiele (chapter 11) present three private sector or complete abolition-are one way of case studies to illustrate innovative, participatory overcoming the limitations of the public sector. technology generation. The first case study describes Innovative multiagency approaches are now being scaling up participatory approaches in an unstructured tested in various ways. They seek synergistic interac- context over three years (1994-97) in the Udaipur dis- tions between different types of organizations, in trict in Rajasthan, India. These have sought to identi- which all work together to analyze particular prob- fy how NGOS and government research and extension lems, contribute jointly to their solution, review services might work more closely together. Some suc- progress in an iterative fashion, and make course cor- cesses are noted, but a number of difficulties remain. rections by mutual agreement as necessary. The principal constraints are deeply rooted in philoso- Field-based efforts to establish multiagency phies and operational procedures on both sides: gov- approaches to technology generation and dissemina- emient organization programs remain largely cen- tion in agriculture and natural resources have prolifer- trally determined. Problems on the NGO side include a ated in the past five years. These have involved varying reluctance of some to think in terms of strategic combinations of government organizations, non- development beyond the confines of a few villages in governmental organizations (NGOS), farmers organi- which they are working. This case study finds that the zations, and, at times, the private commercial sector. efforts to scale up participatory approaches will, for Although some success has been achieved in intro- the future, have to be organized on a more structured ducing participatory approaches at the diagnostic basis. stage of the research cycle (for instance, participatory A second case study describes scaling up partici- rural appraisal is now widely used, although often patory approaches in a structured context and uses an badly), much less attention has been paid to ensuring example from microwatershed rehabilitation. continued interaction between researchers and their Microwatersheds of 500-1,500 hectares containing Part II. Institutional and Social Perspectives 99 one or more villages have become the focus of policy they themselves experiment with new technologies, attention in several South Asian countries as a natural they learn in the best way they can. Being able to val- unit for implementing efforts to intensify agricultural idate and modify technologies for their own use, as and natural resource productivity. Improved natural well as to create entirely new technologies, makes the resource management potentially allows a quantum farmer much less dependent on outside sources of leap in agricultural production technology and in pro- information, which is very important in situations ductivity. where infrastructure and information services are A third case study describes changing the organi- lacking. zation and management of government services to Principle 2 suggests the use of rapid, recognizable scale up participation. It reports on the evolution in success of selected experiments to motivate farmers the eastern lowlands of Bolivia of a model for to innovate, rather than artificial incentives or subsi- strengthening links between research, extension, and dies. Principle 3 stresses the use of technologies that farmers. Central to the model is the highly innovative rely primarily on inexpensive, locally available Research-Extension Liaison Unit (RELU) explicitly resources. Principle 4 recomimends beginning the charged with linking the state agricultural research process with a very limited number of technologies, center with a broad range of intermediate institutions, so that the program is focused and can achieve the including NGOS, that carry out extension. When the maximum possible success right from the start, and RELU was created, its job was seen principally as that involving even the poorest farmers in the process. of transferring technology placed on the shelf by the Finally, principle 5 recommends training village lead- existing commodity-based structure of Centro de ers as extensionists and supporting them as they teach Investigaci6n en Agricultura Tropical (CIAT), but additional farmers, thereby creating and nurturing a applied research and testing soon became the princi- community-based multiplier effect. Now called pal activity for some RELU staff. farmer-to-farmer extension, this principle is effective Participatory principles are also key to successful and keeps costs down. To be selected as an extension- extension with long-run impact (Bunch, chapter 12). ist, fanners must have introduced innovations suc- One such approach-participatory people-centered cessfully on their own farm (which provides credibil- agricultural development-has been developed and ity!). tested extensively in Central America by World In many instances where people-centered agricul- Neighbors and has spread to NGOS in many nations tural development has been used successfully, yields around the world. of major crops have tripled and new crops have been Government extension services tend to neglect introduced. But success has not been uniform. farmers in resource-poor areas. One way to overcome According to one survey, in the 25 to 30 percent of this is to enlist farmers in solving their problems of villages where the response was best, yields contin- low productivity. A participatory process teaches vil- ued to increase after program termination, the number lagers how to experiment and teach one another of organizations increased, land prices increased dra- through the most efficient teaching process: learning matically, incomes increased, and outmigration either by doing. diminished drastically or was reversed. In 40 to 50 People-centered agricultural development consists percent of the average villages, yields more or less of a series of principles with much room for adapta- maintained or increased only marginally, whereas in tion to local circumstances, farmer needs, and the the 20 to 35 pelcent of the villages where impact was institutional context. Principle 1 is to motivate and poorest, yields dlecreased, although they continued to teach farmers to experiment with new technologies on be better than the yields at program initiation. a small scale, thereby reducing the risk of adoption One of the rnajor lessons of people-centered agri- and providing a means for them to continue to devel- cultural development is that sustainability does not op, adopt, and adapt new technologies in a permanent reside in technologies, which have a half-life of per- scientific process of innovation. This principle is fre- haps five yeatrs. Markets change, input prices quently referred to as participatory technology devel- increase, new technological opportunities appear, opment. Farmers who have very little or no formal varieties degenerate, pests spread, and competition education are not accustomed to learning through the becomes stiffer. The hope for sustainability of agri- written word or presentations; they learn much more cultural development is in the nature of an ongoing efficiently through their own experience. Thus when social process of widespread experimentation, sharing 100 Part II. Institutional and Social Perspectives of information, and innovation and group problem- extensive livestock grazing or was not farmed at all solving. In people-centered agricultural development, due to violence. Only 25 percent of the land suitable it is more important how things are taught than what for crop production was actually devoted to this use, particular technologies are taught. Thereby the process while the rest was left to pasture. This suggests that creates a sustainable agricultural development process. there are indeed large tracts of unused or underused It also, by its very nature, empowers people in ways land that could be subjected to land reform in order to useful for rural development efforts in general. increase agricultural productivity-a notion in line Giving farmers access to land and security of with available empirical evidence. tenure is essential to creating the right incentives for Because centralized land reform has not had much sustainable rural development. Theoretical reasons success, negotiated, market-based land reform is and empirical evidence suggest that land reform may being attempted. It is characterized by three main provide benefits in terms of both equity and efficien- principles. First, potential beneficiaries (with assets cy (Deininger, chapter 13). First, a large body of below a certain minimum level) can negotiate inde- research has demonstrated the existence of a robustly pendently with landowners and, once a deal has been negative relationship between farm size and produc- struck, are eligible to receive a .grant of up to 70 per- tivity due to the supervision cost associated with cent of the land's purchase price (subject to an upper employing hired labor. This implies that redistribution limit). Second, the government's role is limited to that of land from large to small farms can increase pro- of regulatory oversight and grant disbursement. Third, ductivity. Second, in many situations, landownership a decentralized institutional structure is needed to is associated with improved access to credit markets, ensure that land reforn is demand driven and coordi- providing benefits as an insurance substitute to nated with other government programs. Despite smooth consumption intertemporally. By enabling the favorable preconditions, and the government's poor to undertake indivisible productive investments expressed determination to distribute 1 million (or preventing them from depleting their base of hectares within four years, this new type of land assets), this coulcl lead to higher aggregate growth. reform has had a disappointingly slow start. Third, even in aggregate cross-country regression, the Brazil and South Africa also have, under different distribution of productive assets-more than the dis- conditions, recently initiated programs of negotiated tribution of incorne-seems to have an impact on land reform. With an institutional background very aggregate growth. Finally, land reform is expected to similar to that of Colombia (presence of land reform increase the environmental sustainability of agricul- legislation and a central land reform institute dating tural production and to reduce rural violence. from the early 1960s), negotiated land reform in Notwithstanding this apparent potential, actual Brazil has been initiated by individual states. The pur- experience with land reform has been disappointing in pose of the Brazilian interventions is to establish all but a few exceptional cases, such as Korea, Japan, cheaper, more agile policy alternatives to centralized Taiwan (China), and to some extent Kenya. Despite- land reform in an environment where the issue of land or because of-this, land reform remains a hotly reform is high on the political agenda and potential debated issue in a number of countries (Brazil, beneficiaries have at least some idea of what to do Colombia, El Salvador, Guatemala, South Africa, with the land. By contrast, negotiated land reform in Zimbabwe). South Africa has been adopted in the context of the There is a new type of negotiated land reform in national reconstruction program, in an environment in which land transfers are based on voluntary negotia- which productive small-scale agriculture was eradi- tion and agreement between buyers and sellers and in cated almost a century ago. There, much greater effort which the government's role is restricted to the provi- is required to establish the decentralized infrastructure sion of a land purchase grant to eligible beneficiaries. necessary to implement land reform, to provide com- This is being tried in Brazil, Colombia, and South plementary services such as marketing and technical Africa and is Deininger's main focus in chapter 13. assistance, and to increase beneficiaries' agricultural As Heath and Binswanger in chapter 3 indicate, and entrepreneurial capacity. small farmers in Colombia were often driven off their Although it is too early to judge whether negotiat- traditional lands and forced to eke out a living in mar- ed land reform can rise to the challenges that admin- ginal and environmentally fragile areas, and much of istrative land reform has failed to solve, the experi- the best agricultural land continued to be devoted to ence unfolding in Brazil, Colombia, and South Africa Part 11. Instittaional and Social Perspectives 101 (as described in chapter 13) can at least provide the those over other administrative functions. Thus the basis for a first assessment. Comparison of the most complete cases of decentralization are those approaches taken in different countries suggests that where local people collect revenues and decide local- (a) land reform through negotiation can only succeed ly how they will be spent. But not all functions of gov- if measures are taken to make the market for land emment can usefully be decentralized, and nonlocal sales and rental more transparent and fluid, (b) pro- groups may be in a better position than local ones to ductive projects are a core element of market-assisted appreciate long-term or large-scale issues and to act as land reforn (designed to establish economically disinterested arbiters of local disputes that cannot be viable and productive projects at a socially justifiable solved locally. This nonlocal perspective is vital in cost rather than to transfer assets), (c) the only way to conservation, which fundamentally concerns avoid- achieve effective coordination of the various entities ing or managing interspecies, intergenerational, involved in this process is through demand-driven and interregional, and international conflicts of interest. decentralized implementation, and (d) the long-run Empowerment of local groups should therefore be success of land reform is likely to depend critically on balanced by a continuing role for central government getting the private sector involved in implementation to deal with market failures and to ensure both social and the ability to use the land purchase grant to equity and environmental protection. "crowd in" private money. After stressing participation and decentralization Several of the chapters in part II (particularly as necessary elements of sustainable rural develop- chapters 9-12) stress the importance of participation ment throughout part II, we would be remiss if we for successful and sustainable rural development. failed to stress more explicitly the need to integrate Some also touch on the need to pursue decentralized gender. For several decades, food, agricultural, and approaches, which are important in and of them- natural resource management policies have been selves. Yet a certain degree of decentralization also designed without acknowledging that rural men and represents a precondition or facilitating role for par- women may have different preferences, face different ticipation. constraints, and respond differently to incentives. Decentralization is one aspect of the broader issue Indeed, until Ester Boserup's path-breaking work in of good governance and is, along with participation, a 1970, women were largely "invisible" within agricul- key element in sustainable rural development. In ture, and the predominant assumption was that the chapter 14, Caldecott and Lutz focus on whether or male household head made most, if not all, farm allo- not decentralization promotes the conservation of bio- cation and production decisions. logical diversity and, if so, under what conditions. Quisumbing and others, in chapter 15, show how Many past attempts to conserve biodiversity failed. neglecting the gender dimension in the design of One reason was an overly centralized approach, often development policy has led to failures in project involving top-down planning by technicians and implementation, failure to adopt new agricultural bureaucrats without concern for the opinions or well- technologies, oJr the adoption of new technology with being of the people affected by their decisions. In negative unanticipated impacts. Therefore, policy- many other countries too, the viability of centrally makers would do well to take gender issues into planned fortress reserves was undermrined by their account when formulating food, agricultural, and nat- cost and also by the democratic deficit built into them. ural resource management policies. Based on a review Nevertheless, complete decentralization can also be of empirical evidence, reducing inequalities between counterproductive to conservation. Under those con- men and women in human and physical capital can ditions, local people, perhaps in cooperation with out- lead to major efficiency and productivity increases in side entrepreneurs, may simply degrade and deplete agriculture. Moreover, addressing inequalities in the the resources faster and more efficiently. So central underlying distribution of property rights between governments clearly continue to have a role. The chal- men and women may be essential not only for lenge is to find the appropriate degree of decentral- increased agricultural productivity but also for envi- ization. ronmental and social sustainability. The second-round The most potent form of decentralization involves effects are also important: when women control the fiscal decentralization, that is, the transfer of authori- increases in income, those increases have a greater ty over spending decisions. These particular powers impact both on household food security and on invest- are retained by elite groups far more diligently than ments in child health and schooling. 102 Part II. Institutional and Social Perspectives Although participatory approaches to project design and-development approach recognizes the determi- have been recognized as a way of ensuring community nants of these asymmetries and their consequences for involvement, commitment, and ownership of develop- individuals, households, communities, and economic ment projects, they are not synonymous with incorpo- development. In essence, the women-in-development rating gender analysis. Participatory approaches to approach recognizes an outcome, whereas a gender- project design by themselves do not guarantee that and-development approach recognizes both the out- women will participate in either assessment of their come and its consequences as well as the process lead- needs or design of the proposed project. This means ing to those outcomes. The understanding gained from that a special effort needs to be made to have gender the gender-and-development approach facilitates properly addressed, such as by means of gender analy- understanding not only how a planned project or poli- sis (Quisumbing and others, chapter 15). This goes cy will affect men and women but also how it will beyond a women-in-development approach, which affect the underlying processes that condition the allo- identifies the differences in resource access, ownership, cation of rights, resources, and responsibilities within and control between men and women. Rather a gender- communities and households. 9 Participatory Rural Developgment Deepa Narayan In many countries, limited government success in In order to induce community-driven develop- managing natural resources, providing basic infra- ment on a large scale, it is important first to under- structure, and ensuring primary social services has led stand the dynamics at the household, group, or com- to the search for alternative institutional options. In munity level. Based on this understanding, it is recent years, a shift has occurred away from supply- important to define what needs to happen at succes- driven toward demand-driven approaches and from sively higher and more distant levels to support com- central command-and-control to local management or munity action. coproduction of resources and services in an effort to increase effectiveness, efficiency, and equity. Key Features of Successful Community-Driven One of these institutional options is participatory Development community-driven development, a process in which community groups initiate, organize, and take action Successful community-driven development is charac- to achieve common interests and goals. Substantial terized by five main factors: local organizational experience, both about what works and what does not, capacity or the existence of viable community groups, shows that no single model is appropriate for all the appropriate fit of technology to community capac- places and times. Although community groups have ity, effective agency outreach strategies, client- organized themselves from time immemorial to meet responsive agencies, and enabling policies. their needs, for the government or other outsiders to induce community-driven development on a large Local Organizational Capacity scale requires agencies to invest in local organization- al capacity and support community control in deci- Local organizational capacity or local-level social sionmaking. This reorientation, in turn, requires agen- capital is embeclded in the norms that enable groups cies to shift their internal rules, incentives, and skills of people to trusit one another, to work together, and to to focus on matching what the agency supplies with organize to solve problems, mobilize resources, what community groups want, are willing to pay for, resolve conflicts, and network with others to achieve and can manage. agreed goals. Participatory decisionmaking processes Experience also shows that community-driven help build confidence and capacity. When people development does not automatically include margin- cooperate and work together, they can overcome alized groups, the poor, women, or ethnic minorities problems of coordination, risk, and limits to their unless their inclusion is specifically highlighted as a information and individual skills. Viable community goal both at the agency and community levels. groups are key to the success of community-driven 103 104 Participatory Rural Development development. Development projects can be instru- why nongovernmental organizations (NGOs) that are mental in facilitating the emergence of local groups not limited to activities in a particular sector can start that organize themrnselves to solve their own problems with activities that produce quick results and then and network with others, including support agencies, move gradually to activities for which there is a time to mobilize resources and design solutions. However, lag. The Agha Khan Rural Support Program in north- for collective action to succeed, certain conditions western Pakistan starts by first building one small have to be met and some common mistakes avoided. productive infrastructure that is identified as the com- munity's first priority. The work is undertaken in part- Quick, local benefits. Collective action is easier to nership with the community (World Bank 1996). stimulate when benefits are quick, visible, and local, The FUNDASAL (Fundaci6n Salvadorefia de when they accrue only to those who participate, and Vivienda Minima) housing project for the poor in El when they are felt to be proportionate to costs. If ben- Salvador builds on the principle of incremental devel- efits from conservation measures such as forestry, opment. Houses are constructed in stages through fisheries, or wildlife protection accrue only at the dis- community self-help; a high priority is assigned to trict or national level, local community groups have securing tenure and services, completing the basic little incentive to participate. For this reason, cleanup structure of the house quickly, and letting families of rivers and lakes is difficult to induce through col- expand their houses over time. Each house is con- lective action. The long gestation period before bene- structed by 20 to 25 families working together; once fits accrue in forestry and tree-planting projects is a completed, the house is assigned by lottery to a par- disincentive to investment, particularly when com- ticipating family. The group then collectively starts bined with insecurity over tenure. Projects that pro- constructing the next house (Bamberger, Gonzalez- vide access to fodder, nontimber forest products, and Polio, and Sae-Hau 1982). firewood while timber is actually maturing are gener- ally more successful than those that rely solely on Common interest. When people experience prob- promised timber benefits. lems that they consider serious and cannot solve However, when quick visible results are induced through their own individual efforts, they are more artificially by agency-initiated shortcuts, they may be likely to come together to act collectively. Many pro- counterproductive to collective action. For example, grams have assumed that geographic community is in an attempt to motivate the formation of pastoral synonymous with "community of interest." This is not associations, the Mali Pastoral Associations Project necessarily the case and is a common source of prob- changed its strategy and constructed water points lems in mobilizing collective action. The community prior to mobilization; an evaluation of the project woodlots movement in India, for example, largely found that the change had the opposite effect because failed in its early days because it was assumed that the it violated the principle of local contribution and community was the appropriate unit of "common involvement in decisionmaking (Shanmugaratnam interest" to manage woodlots (Cernea 1989). and others 1992). Communities are rarely homogeneous entities. In irri- Another shortcut with negative long-run conse- gation associations, the interests of those at the tail quences is to lower or eliminate a community's finan- end invariably clash with those at the head end; for- cial contributions to induce quick action. Once elimi- mal village councils, such as the panchayats of India, nated, contributions are difficult to introduce later in may have different interests than the poor in the com- the same area; the result is lack of local ownership and munity. financial viability. In a rural water supply project in However, the need to solve urgent high-priority Nigeria, required community contributions were dra- problems may bring different class and power groups matically lowered to induce quick action and acceler- together. In South India, Robert Wade (1994) has doc- ate the pace of implementation. This lowered the umented how an entire village manages community- stake that community groups had in the success of the based irrigation systems and has developed a moni- boreholes (Boerma 1993). toring system to discourage water theft. Nonetheless, A more useful approach to inducing quick benefits as groups grow or spread over large areas, the influ- is to sequence tasks and break them up into small ence of social cohesion begins to break down, and it "doable" bits. Success in small tasks generates mutu- becomes more difficult to control and monitor the al trust and satisfaction. This is also one of the reasons behavior of individuals. For this reason, as groups Deepa Narayan 105 become larger, either they formalize regulations and organized arouncl irrigation systems. Rather than cre- delegate decisionmaking to smaller working groups ate new organizations, existing groups were encour- or they join in a federated structure that leaves deci- aged to register themselves as official farmers irriga- sionmaking at the local level. tion associations. In a Livestock Development Project in Mauritania, pastoral associations based partly on Authority over resources. In the past, agencies traditional organization and partly on new organiza- have been reluctant to shift the control and authority tions were created on a pilot basis, using environmen- over resource management to community groups. The tal and socioeconomic data. The average association assumption has been that community groups will not covered 50 villages, 2,500 square kilometers, and a make rational decisions and will misuse financial population of 14,000 people. These groups proved too resources. The experience has been to the contrary. In large to work effectively and were quickly subdivided the Philippines, when farmers groups were given (Shanmugaratnam and others 1992). authority to set and collect irrigation fees, farmers actually raised fees and improved their collection, Group leadership and capacity. The capacity of dramatically increasing the returns to the National groups to organ'ize themselves to undertake coordi- Irrigation Agency (NIA). In Orangi, an urban neigh- nated action is iraportant to their success. Local elites borhood in Karachi, Pakistan, the cost of sewers often take leade:rship roles, and although this is not declined as the incentive for government engineers to necessarily bad, it can result in the hijacking of overdesign sewers was removed. In several social resources unless transparency and accountability are funds (the Village Infrastructure Project in Indonesia somehow enforced. For example, in community- and the social fund in Malawi and Eritrea), as com- based rural water projects in Indonesia and Pakistan, munity groups were given the authority to manage success in community organizing was closely linked financial resources, they carefully monitored all to the presence of strong leaders interested in chang- expenses and often reduced the officially sanctioned ing the water supply situation. However, in labor costs paid to themselves to extend the money Zimbabwe, when local leaders implemented the poli- available for physical construction. To encourage cy regarding communal grazing schemes, they often accountability, more and more projects are supporting grabbed the external resources for themselves or the public posting of all project agreements with com- obstructed the formation of grazing groups (Scoones munities, including financial information, unit costs, and Matose 1993). Charismatic leadership may be labor costs, and so forth, to create a well-informed necessary to seed collective action, but managerial public and thus maximize the protection of project leadership is called for once the period of creation and resources. innovation is complete. When collective action is induced from the out- Social embeddedness. Experience demonstrates side, the focus :is often on creating the structure of the importance of nurturing institutions at the local committees, without matching the task to the capaci- level that have their roots in the local social context ty, knowledge, or technical skills of groups. Groups (Evans 1995). The Balochistan Primary Education also fail because too much is expected of them too Project in Pakistan shows how quickly community soon without any supportive training in management action spreads when it is embedded in the local social or specific skills. Getting local groups and organiza- organization, when tasks are clearly defined, and tions to become self-managing organizations can when time is spent developing the local organization. extend over several years and does not happen with- The problem with existing social organization is out investment in capacity building. that it is generally invisible and often excludes women and the poor from the most important produc- Group ownership and enforcement. To function tion and decisionmaking networks and associations. It effectively, community groups have to set the rules is important to understand the existing social system that define membership requirements, responsibilities so as to draw on its strengths and embed new organi- and benefits, accountability, punishments, and the res- zational forms in the existing social milieu. olution of disputes. This may be done formally, with In Nepal, for example, when a government policy written texts, fees, and fines, or informally through was promulgated to create farmers associations, assis- practice. Often, rules evolve in the course of the work. tant overseers found many informal groups of farmers Entry rules define who belongs and who does not 106 Participatory Rural Development as well as the obligations of members. Membership although most project documents stated that a goal can be based on a variety of factors: ownership of was to reach women (Narayan 1994). land, participation in farmers groups, gender, or age. Nor do community-based programs necessarily To limit the benefits to those who do the work, suc- reach the poor. Work with pastoral associations cessful groups often impose membership or user fees. reveals that as many as 80 percent of the herd in Mali Allocational rules define responsibilities, contribu- and 40 percent in Niger belong to absentee herd own- tions, and benefits. Unless members and managing ers. The capture of benefits and resources by the rich committees know what their responsibility is and how and powerful is ever present when the poor are bound they will benefit, they cannot be expected to perform to the rich in client/patron relationships and the their functions. A 1987 evaluation of the Orissa Social resource under consideration is scarce and therefore Forestry Project in India found that 82 percent of the has great economic value (Shanmugaratnam and oth- villagers, all of whom were supposedly members, did ers 1992). not know how produce from the village woodlot Thus, in Pakistan, village elites play a positive role would be distributed. Most did not expect to receive in the northem provinces in initiating simple drinking any share from the final output and looked on the water supply schemes, but in the southern provinces, community woodlot as another type of reserve forest. particularly Sind, elites capture irrigation water, often Not surprisingly, they had no interest in contributing blatantly and disproportionately (Bymes 1992). In to its establishment or maintenance. Kenya, where community leaders are sympathetic to Mechanisms to ensure accountability are also the plight of the poor, communities have instituted a important. The fact that group members know that an sliding scale of fees for the poor. In Tanzania, sliding effective monitoring mechanism is in place can deter fees based on poverty appear universal in community- violators and encourage others to report violations. managed systems. In other countries and settings, Monitoring can focus on payment of tariffs, extraction however, the poor are excluded when communities and use of resources, and performance of group man- impose user charges to achieve financial viability. agement committees, agency staff, and contractors. In The poor are most easily reached when program small groups that live in physical proximity, social services and structures are targeted specifically to pressure through peer monitoring is an important low- their needs. In the FUNDASAL Project in El Salvador, cost and effective technique. The Grameen Bank peer the poor constructed each other's houses, and this monitoring system is an important part of why the mutual help eliminated the need for a 10 percent group lending system works. If an individual defaults downpayment. Families were required to work in on a loan, the whole group is accountable. This cre- construction teams for 30 weekends and to participate ates incentives among members to monitor and sup- in meetings and training in organizational techniques port one another. (Bamberger, Gonzalez-Polio, and Sae-Hau 1982). The greater the resolution of disputes at the local level, the less burdened and expensive the overall sys- The Appropriate Fit of Technology tem. When conflicts cannot be solved quickly, group- to Community Capacity based schemes fall apart. In Zimbabwe, many grazing schemes have not been able to resolve basic boundary The most important criteria determining appropriate- conflicts or differences of opinion about technical ness of technology are the level of local management soundness. The presence of physical fences to mark capacity and the influence that the introduction of a boundaries has made no difference, and fences have technology has on social cohesion and benefit flows. been removed, ignored, or not maintained. Thus technology needs to be viewed in a holistic way to determine its fit with the existing social system Access by women and the poor In most societies, and the direction of change that its introduction will there are important differences between men and set into motion. women as regards their roles, needs, networks, skills, In the drinking water sector, poor maintenance led and knowledge. It cannot be assumed that what is to the search for village-level operations and mainte- appropriate for men is appropriate for women at the nance technologies, including hand pumps. Although community level. A study of rural water supply cost is important, the least-cost option is not neces- projects, for example, showed that only 17 percent of sarily the best. Experience has established that people the projects involved women in decisionmaking, desire house connections and are often more willing Deepa Narayan 107 to pay a relatively high charge for piped water than a Effective Outreach small fee for access to communally managed hand pumps, which are less convenient and reliable. Outreach mechanisms can be broadly classified into In natural resource management and agricultural two basic approaches, the extension approach and the projects, technological considerations are important empowerment approach. In reality, the two approach- when considering what to plant, how to harvest, and es often overlap:; effective outreach should combine how to monitor. Experience shows that community aspects of both. The extension approach is character- response is greatest when projects adopt approaches ized by information dissemination, demand promo- that provide a continuous flow of benefits, especially tion, and delivery of particular services or inputs. In where indigenous people or women are dependent on the empowerment approach, the main goal is to nontimber products, fodder or thatch grass, agricul- empower local groups or create local organizational tural intercrops, and seed. capacity for self-management through involvement in In Nepal, where indigenous farmers organiza- decisionmaking. The key challenge is to match the tions manage the majority of irrigated land, agency- outreach system with the desired outcomes to be led and donor-financed construction of permanent achieved at the community level. The empowerment headworks based on efficiency concerns has been approach is more staff-intensive and usually requires less than satisfactory (Ostrom 1993). The Kamala greater investment in local capacity building. In an Irrigation Project was meant to serve 25,000 irrigation project in Pakistan where success depended hectares but has never reached that goal. The system on autonomously functioning water users groups, the includes large, permanent, concrete headworks and a project provided for only 20 days of work per group fully lined canal. No fees have ever been imposed or per community organizer. This resulted in community collected, and agency staff spend most of their time organizers rushing from group to group, creating operating and maintaining the huge concrete head- groups without even allowing time for them to assess works and very little time maintaining the rest of the their own needs iByrnes 1992). The groups collapsed system. No farmer organization was created at con- soon after project completion. struction nor were farmers involved in decisionmak- The institutional forms within both systems vary. ing. Because the headworks increased the control Outreach can be done by restructuring agencies to add that nearby farmers had over water, these farmers client outreach personnel (NIA in the Philippines); by had little incentive to bargain with the tail-end farm- contracting outreach services to NGOS (Kenya Water ers. As a result, fanners located at a distance from and Health Organization in rural water supply), the the headworks have broken through the branch private sector (in agriculture in Malaysia), other gov- canals to obtain water, and armed conflict for water emient agencies, or other bilateral agencies; or by often occurs between the tail-enders and head- using existing lozal groups. enders. A recent study of several community-based pro- By contrast, in the nearby Pithuwas Irrigation grams in Mexico found that a key issue in reaching the Project, which is not dominated by permanent head- poor was investment in social organization or group works, the need to pool labor every year to construct formation (Carrasco, Esmail, and Piriou-Sall 1997). In the intakes and maintain the canal creates incentives FONAES, the national Solidarity Fund for Social for head-end farmers to collaborate and negotiate Enterprises, which provides assistance for productive water distribution. In sharp contrast to the Kamala activities to households below the poverty line, agency project, in the Pithuwas project, farmers have man- staff encourage the formation of groups of 15 people aged to irrigate 1,300 hectares using a system who then collectively develop detailed plans including designed to serve 600 hectares. Although no farmer financial details. FONAES staff also help groups to reg- groups were created at construction, a branch com- ister legally. The work is concentrated in a micro- mittee was later formed; this model has since spread region and then spreads outward, with members of through the entire system (Laitos and others 1986). groups training new groups and raising awareness If the tasks to be performed by the community- about the program in nearby cormmunities. level actors are complex and on a recurrent basis, Because community-level workers are often the such as long-term operations and maintenance, then first contact through which cornmunity people learn the investment needed in capacity building is gener- about agency programs, they are of critical impor- ally higher. tance. Unfortunately, the design of community out- 108 Participatory Rural Development reach systems is usually not given careful thought, ers in decisionmaking is limited by the structure of the often with disastrous consequences. It does not help extension approach, which emphasizes technical that community-level workers themselves have little instruction (Scoones and others 1993). voice within their own agencies.- Merely adding more In contrast, in Mali the role of agricultural exten- community workers without encouraging higher-level sion agents working on natural resources has been to staff to support community-level workers makes no enhance the capacity of village groups to manage nat- difference. ural resources. In the 1992 Natural Resources Management Project, extension agents supported Nature of the task. The extension approach is use- communities in carrying out their own needs analysis, ful in increasing the use of inputs not highly depen- in developing a plan of action, and in liaising with dent on coordinated action by individuals (such as external support agencies. This is a good example of agricultural inputs, seeds, fertilizer, nutritional sup- an extension approach that is empowering. plements). The dissemination of information is often made more efficient if groups are organized and ready Skills and characteristics of field agents. to receive infonnation or other inputs on predeter- Extension agents have technical skills, are informa- mined days. This includes, for example, the training tion specialists, and are often outsiders, typically with and visit (T&V) system, use of contact farmer groups, loyalty to the agency. Empowerment agents, in con- use of visiting health teams to monitor the growth of trast, must be highly trusted if they are to motivate babies on fixed clays in a month (being done success- community groups to collective action and must have fully in India and Indonesia), and use of the veterinary some technical information and skills. In India and department to vaccinate livestock (in Pakistan). It is Korea, the difference in the performance and charac- also efficient when communities or individuals are teristics of irrigation patrollers, the lowest-level work- only required to perform one-shot tasks such as col- er involved in canal operations, provides a graphic lecting money, contributing labor for construction example (Wade 1994). The Indian system, with its over a very short period of time, immunizing children, hierarchical organizational system based on central- or eliminating a particular pest. ized control and public works orientation, minimizes The empowerment approach is essential when identification between patrollers and local farmers community groups need to be involved in decision- and maximizes orientation and accountability to the making and take responsibility for long-term manage- Irrigation Department. Patrollers are full-time ment or when the tasks involve coordination at differ- employees of the agency and, hence, are only margin- ent levels. ally involved in farming themselves. They are select- ed by the agency engineer and become permanent Role of field agents. In the extension approach, employees after a period of probation. They are not field agents are often message focused. They channel posted near their own village and must move within information, provide technical expertise, and deliver six years to another area. The relationship is one of inputs, seeds, fertilizer, health or nutrition education, institutionalized mistrust and control by superiors. and so forth. The extension approach is weak in sup- The incentives, skills, and rules in the Korean sys- porting local institutional development for self- tem are almost the inverse and maximize the identifi- management. In the empowerment approach, field cation of patrollers with the existing social network of agents are first facilitators, catalysts, and organizers other farmers and village chiefs. The patrollers are for empowerment. Technical information is provided paid part-time; the rest of the time they are farmers. by specialists or the field agent to help local groups They are selected by the village chiefs and approved make informed decisions after weighing the costs and by the irrigation hierarchy; they must be renominated benefits of various options. every year by the village chiefs. They must have land To manage natural resources, in the Department of and reside in their area of work, and they are not post- Agricultural, Technical, and Extension Services ed from one place to another. In the Korean system, if (AGRITEX) system of extension in Zimbabwe the the irrigation system does not work well, the irrigation extension worker promotes a fixed and limited tech- patroller himself suffers, together with other farmers nical package through a range of T&V group with whom he has an ongoing relationship. approaches, demonstration plots, master farmer cer- Another important example is from the Tamil tificate training, and field days. Participation of farm- Nadu Nutrition Project, where the process of selecting Deepa Narayan 109 community-level health workers is an important con- approach, the primary activity initially is building tributor to success. Preference is given to poor, mar- confidence by tapping into people's knowledge and ried women from within the village, who are of sched- involving people in tasks; specialized information is uled caste with primary-level education and two introduced only as needed. Mass media are used to healthy children. They are hired as part-time workers. publicize the availability and rules of the program and Because the women have deep roots within the com- to establish transparency. munity and are poor but have thriving children them- selves, they are credible nutrition workers to famnilies Accountability. Within the extension approach, with malnourished children. The health workers have fieldworkers typically have their heads tumed toward a limited number of specific tasks and work with a the agency, which rewards, punishes, promotes, or local group of women to whom they are accountable ignores. In the empowerment approach, agents are (Heaver 1989). accountable to community groups even when they are By contrast, in the Integrated Child Development paid by agencies. To increase accountability, some Services Project in India, the anganwadi (child devel- programs require communities to certify that staff opment) workers are without clear priorities and are have completed work satisfactorily; in others, agency overloaded with a wide range of tasks including feed- staff are replaced by local people who are sometimes ing preschool children, conducting home visits, and also paid directly by communities. paying attention to pregnant and lactating mothers. Studies have found that almost half of the women are Outcomes desired at the community level. In the not recruited from the village in which they work and extension approach, the agency's desired outcomes are often of high caste. In the Indian context, it is very are that the inputs and services delivered are used and unlikely that these women will try to reach needy that programs are more effective and efficient. In the children from scheduled castes (Subbarao 1989 and empowerment approach, the key outcomes are that Heaver 1989). households or groups at the community level have Hiring women becomes particularly important in organized for self-management and show increased reaching women. In Pakistan and Yemen (the 1993 capacity for coordinated action including manage- Balochistan Primary Education Program and the 1992 ment of local services or resources. Basic Education Project), female teachers had to be hired before parents were willing to send female chil- Client-Responsive Agencies dren to school. In Nigeria, hiring female agricultural extension workers tripled the number of female farm- Client-centered agencies are needed to ensure that ers who were in contact with extension workers. demand is met and community self-management is adequately supported by community fieldworkers. To Control over decisions. In the typical extension become client centered, agencies must be convinced approach, control of and authority over decisions are to change their rules and incentive structures so that retained by the agency, and the field agent is the mes- the benefits from client responsiveness are more senger. In the empowerment approach, the agency important to the agency than the costs. Experience sets the parameters or starting conditions of partner- demonstrates that trying to support community-based ship, but control over management details is left with approaches without fundamentally changing institu- the community. tional structures, rules, and incentives does not work. Some social funds use the empowerment approach Key design featu-es of client-responsive agencies are in which fieldworkers offer a menu of options from summarized in table 9.1. which the community chooses those that best fit its Experience shows that strong political and govem- needs and financial and management capacity. The ment support for making public agencies client cen- agency defines the rules for partnership (for example, tered is needed at the highest levels. Such support has the amount of financial assistance available). occurred, for example, with the Programa Integral para el Desarrollo Rural (PIDER) in Mexico, NIA in the The role of information. In the extension approach, Philippines, public works programs in Indonesia, where the primary purpose is dissemination of infor- local government in Paraguay, the Ministry of mation, social marketing and the use of mass media Forestry in West Bengal in India, and in Zimbabwe. are of great importance. In the empowerment The performance of the irrigation sector in Mexico is 110 Participatory Rural Development Table 9.1. Reinventing Agencies to Support Community-Driven Development Lessons from the community Agency characteristics Agency mechanisms Group must have a felt need for Agency responds to demand and Use demand assessment, institute service; there is "commonality knows what people want outreach mechanisms, and employ of interest" social analysis to identify and understand key actors, their power, interests, and needs Group controls and has authority over Agency "lets go" of control over Define objectives and indicators of resources, decisions, and rule implementation details and spells success to support achievement of making out framework of rules for local control and reorient staff and interaction and negotiation with performance criteria communities Local group has the needed capacity Agency puts local empowerment Institute the capacity-building and skills and can mobilize and capacity building high on process, invest time and resources financial resources for long-term its agenda in training, devise new funding survival mechanisms that reach communities quickly, focus on strategies for groups to achieve financial self-sufficiency, and phase in payment of outreach workers by community groups Every group or community is unique Agency plans for diversity and Use a learning process approach, encourages local adaptation employ short planning horizons, modify implementation plans with feedback from monitoring and evaluation In many cases, the poor and marginal, Agency focuses on reaching the poor Reflect in objectives, institutional including women and indigenous and marginal mechanisms, targeting strategies, groups, are lefl. out and indicators of success the focus on poor and marginal groups a notable example of rapid and successful institution- Almost all projects that adjust as they implement al reform. invest in numerous small studies that provide feed- The critical question is how to bring about govern- back on how different approaches are working. The ment commitment to change. World Bank staff have Women-in-Development (WID) Project in The used two main strategies to generate interest and com- Gambia created a monitoring and evaluation unit mitment: (a) pilot projects as instruments to demon- during project preparation that honed its skills by car- strate alternative strategies and (b) a variety of partic- rying out several small studies. The Indonesia Rural ipatory techniques, including field visits to other Water Project has developed a national system of regions and countries, to generate interest in and com- monitoring key process and output indicators focus- mitment to new ideas and strategies for action. ing on changes at the community level. Indicators of success send important signals to The most graphic examples of the importance of staff about program priorities. If community involve- defining the parameters while supporting the evolu- ment, the number of women or the poor reached, and tion of local rules have emerged from the water and the number of systems functioning are not reflected in natural resource management sectors. Projects with indicators of success, there is little incentive for staff standardized forms of participation and universal to change their way of doing things to reach these applicability of implementation rules formulated by goals. the agency have not worked and, despite stated objec- Deepa Narayan 111 tives, have become agency-led rather than user-led if the rules and regulations at different levels provide programs. Programs that have created the overall incentives to work and organize. Three broad cate- framework of rules of interaction between users gories of rules are important: legislation, titles, and groups and agencies, while letting users groups work rights; control and use of funds and local fees; and out the details of how to manage resources, have other rules of the game. Also important is gaining emerged with numerous variations in local rules that political support or patronage in the right quarters. are managed, monitored, and enforced by local people. Legislative issues that govern the functioning of Elinor Ostrom (1990) highlights this fact in com- agencies and the relationship of agencies with other paring irrigation systems across and within countries. agencies and communities should be identified during For example, a study of a farmer-led Karjahi irriga- sectoral reviews. Key issues to examine include the tion system in Nepal found a diversity of rules even mandate of agencies, redefining the role of govern- within one small self-organized system, yet the sys- ment agencies as facilitators and regulators, civil ser- tem functioned well (Hilton 1990). In contrast, vice reform, new funding mechanisms, new systems Frances Cleaver (1990) in a national sample survey of of accountability for performance, legal status of more than 400 hand pumps in Zimbabwe found that community groups, simplification of legal registration applying microscopic rules (such as how deep the requirements, individual, group. or community own- community was required to dig) deterred community ership, and use and tenure rights over assets (particu- participation, local initiative, and problem solving. larly water, land, and other natural resources). Similarly, uniform application of rules in grazing Agencies and communities have little incentive to schemes, tree plantations, and woodlot management mobilize resources if they do not realize any immedi- has proven counterproductive. ate benefits. At the agency level, when fees collected The importance of social organization is increas- by agencies revert to central treasuries with no link- ingly recognized, and hence most agencies involve age to the agency's subsequent share of resources, institutional and social science (noneconomic) exper- agencies have little incentive to pay attention to tariff tise early in program development. Understanding the payment or collection. Dramatic improvements can social organization includes understanding local lead- occur when agencies are required to be financially ership and power systems; who has control and access autonomous and have access to user fees. to resources; who are the key social actors; what are It should not be assumed that having a centralized the local patterns of landownership, tenure, and use agency manage the collection of user fees is always rights; and what are the number, structure, and func- the most effective way to proceed. One of the key tioning of existing informal and formnal groups. In the problems that lhad to be addressed in the Bank- experience of NIA in the Philippines, baseline socio- supported NIA participatory irrigation projects in the cultural profiles and process documentation played Philippines was poor maintenance of irrigation important roles in pointing the direction for institu- canals. At the outset, it was assumed that the best way tional change. In Mexico, the National Water to solve the problem was to impose targeted increases Authority has an in-house group of senior social sci- both in the annual funding levels of operations and entists and communication specialists who design the maintenance expenses that NIA had to incur and in the strategy for community outreach, applied research, total amount of irrigation fees it had to collect annu- and communications. This work is then subcontracted ally. However, this undermined farmer-managed irri- to the private sector. gation, including fee collection, which was more effi- cient than NIA fee collection. Eventually NIA changed Enabling Policies its rules and gave control to farmers. Farmer-managed irrigation schemes worked so well that NIA subsidies Community-driven development on a large scale were eliminated in two years (NIACONSULT 1993). requires enabling policies and political support to pro- Many rules and regulations may need to be tect agencies so that they can initiate the reform changed, from required qualifications of community process and give it time to take root. It is much easier workers, teachers, and health educators to procure- to support reform in the context of decentralization ment rules. In the Balochistan Primary Education strategies. Although individuals do play key roles in Project, the quick spread of community-managed starting or protecting the reform process while it is schools for girls, required several policy changes. still in the early stages, change can be sustained only Some of the key policy changes to ensure recruitment 112 Participatory Rural Development of and support for female teachers from within the vil- minister of water resources, and the director general lage included lowering the minimum age to 14 years of the water directorate. In Mexico, the solidaridad and raising the maximum age to 40 years, legalizing a program, which devolves fiscal power to the munici- mobile teacher training program, and "sanctioning" palities, had the blessing of the president. the school and teacher post through the Education The Gambia WID Project survived the preparation Department Community Coordination Unit. Without process because of the strong interest of the president. these changes, no amount of community participation At later stages, when it appeared that the project and organization would have helped spread primary might not proceed, the president flew to Washington, education to girls quickly. In changing the rules, D.C., to meet with then-president of the World Bank, attention must be paid to protecting accountability Barber Conable, to persuade him of the importance of and performance. the project to his country. Since then, the project has When change is initiated on a large scale, there are been designated one of the Bank's most successful, usually losers and winners. Change is a political and strategies developed to reach women in agricul- process that rearranges power structures and alters ture have been incorporated in mainstream agricultur- people's access to and use of resources. Most of the al projects (Schmidt 1993). large-scale community-based approaches, especially Godmothers play important roles as well. In those that have worked through government and insti- Indonesia, wives of governors often are involved in tuted rapid, radical change, have had strong political getting service delivery agencies to take on gender support or patronage. issues in a serious way. In Colombia, the community- Judith Tendler (1993) highlights the positive and based day care program movement was actively sup- important roles played by "godfathers." Behind many ported by the wife of the president. stories of successful implementation has stood a "good governor" or a state secretary or a reform- Designing Participatory Community-Based nminded senior civil servant with access to and support Development Programs of top political leaders. The good governors have pro- vided protection from the pressure to hire mediocre The lessons from the worldwide experience in com- staff or to fire excellent staff on political grounds, to munity-driven development must be reflected in the make technically undesirable choices, and to delay design of new projects. This section discusses 10 fea- the transfer of funds from the central government. tures that are particularly important in the design of They have pressured agencies to produce results and participatory community-driven development proj- chosen one or two components of a project as "signa- ects. In situations of many unknowns, pilot activities ture activities" with which to leave their mark. Thus using the new learning and innovation loans are the in Sergipe, the governor fashioned the project around most appropriate. rural water supply; in Piaui, the governor supported the land purchase component; and in Ceara, the small- Set Objectives: Clarity, Priority, Links to Outputs scale riverine irrigation component won the gover- nor's support. Good governors have also mobilized A surprising number of projects do not have clearly their own resources, sometimes as much as three stated objectives logically linked to strategies, out- times the amount provided in the Bank loan, to sup- puts, indicators of success, and physical or capacity- port their favorite component within the time period building outcomes. Goal-oriented planning tools such of their tenure in office. as the logical framework analysis are aids in achiev- In Nepal, when the government decided in 1987 to ing clarity. These tools help to clarify and prioritize promote farmer associations in irrigation management objectives and articulate underlying assumptions, through radical policy changes, the Bank responded activities, and indicators of success. Brainstorming with a $20 million irrigation line of credit with tech- meetings and discussions can also be useful. nical assistance to support the project through the Establishing the priority of objectives is particular- United Nations Development Programme. The policy ly important in community development projects changes devolved power from a large central bureau- because as the pressure to produce tangible results cracy to lower levels, including farmers. These builds, shortcuts are taken, resulting in "build first, changes were made possible by support at the highest listen-dialogue-and-organize later." This is most like- levels, including the king, the minister of finance, the ly to happen when there are no godfathers, when tech- Deepa Narayan 113 nical agencies are in charge, or when success is based including NGO or community development specialists, on construction completed and inputs distributed have been particularly helpful in identifying and rather than on services functioning, being operated sketching the workings of existing social organization and maintained, and being used effectively. from which much can be learned about what does and Project experience establishes that women and does not work in a particular context. A range of par- the poor are not automatically reached through com- ticipatory and nonparticipatory techniques and tools munity-driven projects. If this is a goal, it must be are now available (Jacobs 1998, Narayan and reflected in the objectives, strategies, and indicators Srinivasan 1995, and Rietbergen-McCracken and of success. In such projects, targeting strategies that Narayan 1998). have low transaction costs become very important. Gender and poverty analysis tools are particularly Assess Demand useful. Early in project preparation, staff have found it Many problems have arisen because programs have useful to be open-minded about overriding objectives. been developed on needs assumed by planners rather These become increasingly clear as project prepara- than effective demand of community groups. A tion proceeds. They may change radically, as hap- review of 121 rural water supply projects shows that pened in the Matruh Natural Resource Management demand, as measured by commitment before con- Project, where objectives shifted from livestock to struction and substantial financial investment up-front natural resource management. To avoid designing in capital cost, contributes significantly toward both complex projects with multiple and competing objec- project effectiveness and overa]Ll beneficiary partici- tives, a hierarchy of objectives that can change in pation in decisionmaking (Narayan 1994). importance over time can be developed, thus making Demand can be assessed using a variety of the project manageable. methodologies: participatory techniques (community self-diagnosis, ranking of priority problems), benefi- Identify the Key Actors, Their Capacity, and Interests ciary assessment or "listening to the people" tech- niques (informal interviews, community meetings, Identifying the key actors at both the community and participant observation), or contingency valuation agency levels is critical. It should not be assumed that techniques that gauge what people are willing to pay the community is the logical unit of interaction with for different levels of service. In assessing demand, support agencies. Households or groups within a data should be disaggregated by gender, wealth level, community, individuals within a household, children, and other relevant characteristics. women, or those in particular occupations (such as farmers, informal sector workers, and the landless) Craft a Self-Selection Process may prove more appropriate. Social assessment is an important tool for identify- If collective action and some degree of self-manage- ing key actors and their interests, the existing social ment are the goal, designing a self-selecting process is organization, and aspects that need to be strengthened important. This includes community groups taking the or changed. At the community level, data from social initiative to become part of a project rather than agen- analysis are helpful in identifying local laws, rules, cies initiating the selection of communities. Self- and regulations governing interaction and access to selection is not a key strategy if delivery and use of resources. At the agency level, social assessment inputs to a targeted group is the prime objective. becomes an institutional tool for identifying the The single most important self-selection strategy is incentives, interest, and capacity of various service to institute a significant financial or organizational delivery agencies to support community-driven devel- contribution up-front before any project outputs are opment. Based on this analysis, either other agencies delivered. Many projects now require community are drawn in or fundamental reform is initiated. This groups to enter contracts and sign documents that may encompass redefining agency functions, restruc- detail mutual responsibility. These are useful only to turing funding mechanisms, or adding a new cadre of the extent that people understand what they are sign- staff to the organization. ing and when enforcement mechanisms exist on both Task teams find it useful to include national and sides (community and agency) to keep the other party international expertise in teams. Social scientists, accountable. 114 Participatory Rural Development As a gauge of interest, agencies may stipulate orga- Restructure Fund Release to Support Demand nizational tasks that community groups must perform before receiving project assistance. These may include A large-scale project may be demand oriented, but if mobilization of local resources and cash, formation of the flow of funds is not responsive, it is impossible to committees, demonstration of consultation with every- maintain a demand focus. Depending on the national one in the community, and submission of a proposal. and sectoral context, projects have used different Some projects require legal registration, an important strategies. factor if laws req[uire legally constituted bodies before One approach is to decentralize the programming governments or banks can transfer assets to the group of funds closer to where they will be used. In or involve the groups in decisionmaking. Indonesia rural water supply, programming of funds has been devolved to different levels. The higher the Structure Subsidies That Do Not Distort Demand cost, the further the project proposal has to travel for approval. However, authorities cannot tamper with To ensure inclusion of the poor and to retain a demand the basic design of the water system. In Ethiopia's orientation, it is important to structure subsidies so social fund, regional offices have been opened. that they do not violate the principle of demand. Another strategy is to create new funding mecha- Project experience illustrates that even the poor are nisms inside or outside the formal government struc- willing to contribute substantial amounts by offering ture. The structure and management of social funds their labor or by accessing credit, where available, if falls in this category. Some have been established the project meets a need and the service provider is with independent boards (managed with a majority of perceived to be reliable and trustworthy. NGO and private sector representatives) outside a par- Five approaches to structuring subsidies are common: ticular ministry; others are semiautonomous, but with * The subsidy is indirect and is invested in strength- high-level protection, often under the wing of the ening community organization or capacity through president or prime minister. Funding may also be outreach and training (Grameen Bank outreach channeled through intermediaries, NGOS, banks, and systems). other multilaterals, particularly the United Nations • Upper limits are set on the base amount; if higher Children's Fund (UNICEF) and International Fund for levels of service are desired, the community or Agricultural Development (World Bank 1997 and user must pay (Indonesia rural water supply proj- Narayan and Ebbe 1997). ect for low-income communities). * Where initial investments are high and beyond the Plan for Learning and Plurality of Models capacity of the community, organizational tasks completed by community groups are used to gauge Community development, especially the end of the commitment rather than requiring significant com- continuum that aims for community empowerment, is munity contribution to capital investment (Sri by definition dependent on hundreds of discrete deci- Lanka rural water systems in areas where hydro- sions made by individuals in communities. The geological systems are difficult; Senegal livestock review of experiences across sectors reveals that no centers). single form of project is successful in all settings; * Agencies defer or spread out large capital outlays projects have worked in some areas at certain times until the community has proven its management and not at others. ability and interest in the service (Pakistan Hence, it is important for the basic planning Balochistan ]Primary Education Project; the gov- assumption to be one of learning embedded in local emnment constructs a school only after a communi- knowledge systems. The project must be seen as like- ty has managed a community-based girls school ly to evolve, adjust, and change over time as local pri- for three years; the government trains the teacher, orities change and as local organizations mature. This, provides school supplies, and pays the teacher in turn, means that project components may change after the first three months). over time in order to maintain a fit with community * Agencies invest in the "trunk" or main network needs. The new adaptable lending instrument actively structures, and communities invest in secondary supports such a learning process approach. and tertiary distribution networks (irrigation, sew- It is also important not to search for one universal erage, piped water systems). answer or model for all times and all places, but to Deepa Narayan 115 plan for a multiplicity of implementation models. This and evaluation. Learning is more effective and effi- is particularly important for projects that require com- cient when feedback is listened to and when changes munities to perform complex long-range tasks. are put in place. Listening to people's suggestions is An increasing number of infrastructure (Tanzania, empowering and encourages innovation and responsi- Brazil, Indonesia) and natural resource management bility. Seeing information systems actually being used (Burkina Faso, Mali, Egypt) projects are adopting rewards the filling out of forms, which otherwise can structured learning as a specific strategy. These large- degenerate into meaningless busy-work. scale projects are based on a learning process The Gambia project used a very simple method to approach rather than on a blueprint design. The focus involve village women in identifying priority indica- is on a framework of principles and processes rather tors for monitoring and evaluation. A consultant was than on implementation blueprints. Planning is on a hired to develop a list of indicators. These were taken yearly basis, with plans for subsequent years depen- to the villages, and well-publicized consultative meet- dent on previous performance. This allows for the ings were held with women's groups who identified evolution of plans based on experience and fine- the key indicators from their perspective. These indi- tuning with time. Procurement of most materials is cators were then monitored by the project monitoring done on a yearly basis, budgets are indicative and and evaluation unit. flexible, and the project is conceptualized as a series of subprojects that are not implemented in a uniform Redefine Procurement Rules way. Procurement rules can hinder community initiative. Invest in Outreach Mechanisms When appropriate, these should support local-level and Social Organization procurement procedures. International bidding and bulk purchase are hardly appropriate when projects The level of investment needed in community social consist of many subprojects, whose nature and timing organization and intermediation is a function of the will be determined by community needs and readi- tasks expected of communities and the existing ness. strength of community organizations. In general, Three strategies have been found to be useful. strengthening indigenous or existing organizations is First, maximumna unit costs are set, as in Bolivia's easier than starting from scratch, although in some social fund. As long as the costs of local procurement cases this may be justified. If existing groups are do not exceed the maximum unit costs, local procure- dominated by the powerful and the elite, they may or ment is allowed, with a system put into place for may not reach women and the poor. Hence, many ensuring completion of quality work. Second, a cost projects that specifically target women or the poor threshold can be set, below which communities pro- invest in special strategies to reach them. These may cure goods locally based on market principles and include forming special organizations of the poor above which procurement follows standard local bid- such as the Grameen Bank, the Self-Employed ding or international bidding as appropriate. Third, as Women's Association of India, and women farmers in India, commimity projects are clustered to allow for groups in Nigeria and The Gambia. economies of scale and to attract a larger pool of local Perhaps the most important lesson that emerges is contractors and NGOS. that there are no shortcuts to strengthening local social organization for collective action. All shortcuts References for speedy implementation that circumvent local involvement in decisionmaking backfire sooner or The word "processed" describes informally repro- later. duced works that may not be commonly available through libraries. Use Participatory Monitoring and Evaluation Bamberger, Michael, Edgardo Gonzalez-Polio, and Umnuay Sae-Hau. 1982. Evaluation of Sites and There is a great temptation to freeze project designs Services Prcjects: The Evidence from El Salvador once implementation starts, even though much is Staff Working Paper 549. Washington, D.C.: written in project staff appraisal reports about learn- World Bank. ing, conducting studies, and performing monitoring Boerma, Pauline. 1993. "Lessons Learnt from the 116 Participatory Rural Development Implementation of the RUSAFIYA Project in a ington, D.C.: World Bank. Number of Selected Communities in Nigeria." Narayan, Deepa, and Katrinka Ebbe. 1997. Design of United Nations Development Programme-World Social Funds, Participation, Demand Orientation, Bank Water and Sanitation Program, World Bank, and Local Organizational Capacity. Discussion Washington, D.C. Processed. Paper 375. Washington, D.C.: World Bank. Byrnes, K. J. 1992. Water User's Associations in Narayan, Deepa, and Lyra Srinivasan. 1995. World Bank-Assisted Irrigation Projects in Participatory Development Tool Kit. Washington, Pakistan. Tectmical Paper 173. Washington, D.C.: D.C.: World Bank. World Bank. NIACONSULT. 1993. "An Evaluation of the Impact of Carrasco, Tanya, Talib Esmail, and Suzanne Piriou- Farmer's Participation on the National Irrigation Sall. 1997. "Decentralization and the Promotion of System's Performance." Report commissioned by Productive Projects: Case Studies of Community- the Participation Learning Group, Social Develop- Based Natural Resource Management and ment Department, World Bank, Washington, D.C. Income-Generating Projects in Oaxaca and Processed. Hidalgo, Mexico." Report prepared for the Ostrom, Elinor. 1990. "Crafting Irrigation Decentralization, Fiscal Systems, and Rural Institutions: Social Capital and Development." Development Group, World Bank, Washington, Paper prepared for the workshop on political theo- D.C. Processed. ry and policy analysis, Indiana University, Cemea, Michael M\. 1989. User Groups as Producers Bloomington. Processed. in Participatory Afforestation Strategies. Discussion . 1993. "The Evolution of Norms, Rules, and Paper 70. Washington, D.C.: World Bank. Rights." Paper prepared for the workshop on prop- Cleaver, Frances. 1990. "Community Maintenance of erty rights and the performance of natural resource Hand Pumps." Department of Rural and Urban systems, Royal Swedish Academy of Science, Planning, University of Zimbabwe, Herare. Stockholm, 2-4 September. Processed. Processed. Rietbergen-McCracken, Jennifer, and Deepa Narayan. Evans, Peter. 1995. "Government Action, Social 1998. "Participation and Social Assessment: Tools Capital, and Development: Creating Synergy and Techniques." Poverty Group, Social Develop- across the Public-Private Divide." University of ment Department, World Bank, Washington, D.C. California, Berkeley. Processed. Processed. Heaver, Richard. 1989. Improving Family Planning, Schmidt, Mary. 1993. "Case Study in Popular Health, and NVutrition in India: Experience from Participation: WID in Gambia." Report prepared Some World Bank-Assisted Programs. Discussion for the Learning Group on Participation, Social Paper 59. Washington, D.C.: World Bank. Development Department, World Bank, Wash- Hilton, Rita. 1990. "Cost Recovery and Local ington, D.C. Processed. Resource Mobilization: An Examination of Scoones, Ian, Jeanette Clark, Frank Matose, Colin Incentives in Irrigation Systems in Nepal." Report Phiri, Ole Hofstad, Isaac Makoni, and Sara prepared for the Decentralization, Finance, and Mvududu. 1993. "Future Directions for Forestry Management Project, U.S. Agency for Extension." In P. N. Bradley and K. McNamara, Intemational Development, Washington, D.C. eds., Living with Trees: Policies for Forestry Processed. Management in Zimbabwe. Technical Paper 210. Jacobs, Sue. 1998. A Manual for Participatory Washington, D.C.: World Bank. Methodsfor SocialAssessment. Washington, D.C.: Scoones, Ian, and Frank Matose. 1993. "Local World Bank. Woodland Management: Constraints and Laitos, Robby, and others. 1986. Rapid Appraisal of Opportunities for Sustainable Resource Use." In P. Nepal Irrigation Systems. Water Management N. Bradley and K. McNamara, eds., Living with Synthesis Report 43. Fort Collins: Colorado State Trees: Policies for Forestry Management in University. Zimbabwe. Technical Paper 210. Washington, Narayan, Deepa. 1994. The Contribution of People's D.C.: World Bank. Participation: Evidence from 121 Rural Water Shanmugaratnam, Nadarajah, Trond Vedeld, Anne Supply Projects. Environmentally Sustainable Mossige, and Mette Bovin. 1992. Resource Development Occasional Paper Series 1. Wash- Management and Pastoral Institution Building in Deepa Narayan 117 the West African Sahel. Discussion Paper 175. Incentive Problem: Organizational Determinants Washington, D.C.: World Bank. of a 'High-Quality Civil Service,' India and Subbarao, K. 1989. Improving Nutrition in India: Korea." Institute of Development Studies, Sussex Policies and Programs and Their Impact. Discussion University. Processed. Paper 49. Washington, D.C.: World Bank. World Bank. 1996. The Aga Khan Rural Support Tendler, Judith. 1993. New Lessons from Old Program, A Third Evaluation. Operations Projects: The Workings of Rural Development in Evaluation Study. Washington, D.C.: World Bank. Northeast Brazil. Operations Evaluation Study. . 1997. "Portfolio Improvement Program Washington, D.C.: World Bank. Review of Social Funds Portfolio." Washington, Wade, Robert. 1994. "Public Bureaucracy and the D.C. Processed. 10 Involving Local Organizations in Watershed Management Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg Involving local organizations in watershed manage- off-site stakeholder groups in the negotiation of com- ment is widely recognized as advantageous for three mon property resource use. In this respect, local orga- reasons. First, local organizations can often be very nizations may have an important role to play in mak- effective in generating and securing compliance with ing micropolicy within the watershed as well as in rules for the use of common property such as water, defining enabling policies that may be formulated and common grazing land, or forest and with management enforced by the state. They also have a role in pro- of buffer zones around conservation areas, all of moting technological innovation and the adoption of which may be important features of watershed man- conservation practices, whether these require collec- agement. This is in contrast to nonlocal agencies, tive or individual action. And they may mobilize and especially in low-income countries, which have diffi- deploy significant resources, both in cash and in kind, culty imposing sanctions on undesirable management that are needed to implement sustainable management practices or providing incentives that are lasting. of watershed resources. Second, organizations that involve local stakeholders This chapter examines experience with involving in the development of management practices and the local organizations in watershed management. The selection of technologies aimed at improving water- discussion seeks to provide guidelines, rather than a shed resource conservation often promote innovation theoretical analysis, based on experience of working by identifying locally appropriate technologies and with local organizations from the perspective of securing their adoption more effectively than external development practitioners. This experience is evolv- agencies. A third reason why central government ing and accumulating rapidly, and many practitioners agencies find it attractive to involve local organiza- are promoting a learning process in which their tions in watershed management is that devolution of approach to local-level watershed management is responsibility can both externalize some costs of swiftly adapted and reoriented based on analysis of enforcing conservation from the state to local com- what has gone before. Although this discussion focus- munities and reduce costs overall by creating condi- es on some of the lessons learned and principles that tions in which nonlocal agencies become more effi- can be derived from experience, the last word on what cient and effective through collaboration with local works and what does not work is by no means decid- organizations. ed. Indeed, the dynamic nature of the process of In watershed management, therefore, local organi- involving local organizations in watershed manage- zations regulate resource use, provide a forum for ment is one of its most important characteristics, and resolving conflicts among local stakeholders, and pro- one that requires a focus on managing the process vide a channel for the representation of on-site and rather than on defining a blueprint for how to go about 118 Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg 119 it. To highlight this important principle, we illustrate when local organizations are involved. The difficulty some of the dynamics of process management with of establishing definitive usufruct rights and obliga- information from a case study that traces the evolu- tions over resources with multiple uses is one reason tion over time of local-level watershed management. for skepticism about the efficacy of local organiza- tions for wate:rshed management (Uphoff 1986). The Social Ecology of Watersheds Multiple uses for any given watershed resource create multiple stakeholders, often with competing priori- In order to analyze the role of local organizations in ties. Typically, use by one stakeholder will generate watershed management, it is necessary to define both negative effects or externalities for others, as in the the physical and the social terms used. The perspec- case of upstream users who clear forests for arable tive of social ecology can be useful because it focus- cropping, whiclh causes soil erosion that silts down- es on the interaction between a physical landscape, stream reservoirs and may damage the hydrological including changes in the natural resource base, and cycle so that seasonal flow is affected. For this reason, the adaptive strategies people use to obtain liveli- conflict among stakeholders over the rights and con- hoods and to organize socially within that landscape. ditions for use of a given resource is a common fea- This interaction between the possibilities for resource ture of watershed management. use and ways of organizing is central to understanding A useful approach to assessing the potential role of the circumstances under which local organizations local organizations in watershed management is to turn out to be effective managers of watershed distinguish resources by their physical characteristics, resources. which affect the incentives and strategies users pursue Watersheds have been defined as "physically and which for this reason require different institution- defined subsets of rural society" and watershed man- al arrangements. Two physical distinctions are agement as "a question of social relations and coordi- whether a resource is stationary (grazing areas, nation among individuals with vested common inter- groundwater basins) or mobile (flowing streams) and ests in the watershed" (White 1995: 3) This definition whether the resource has storage capacity, either nat- recognizes that a watershed is a geohydrological unit ural or man-made (reservoirs or tanks). These two comprised of all land and water within the confines of physical characteristics affect the ease with which a drainage divide as well as an ecosystem where peo- users can obtain or generate information about the ple, livestock, vegetation, land, and water interact resource and thus their ability to coordinate resource (Jensen 1996). management (Schlager, Blomquist, and Tang 1994). In the watershed landscape, the flow of water cre- For example, in common property resources with ates a set of interlinked environmental effects from mobile flows that fluctuate unpredictably, such as upstream to downstream and involves multiple stream flow in a watershed, it is very difficult for resources (water, soil, vegetation) with multiple uses: users to assess the effects of use by one stakeholder on for example, water for irrigation, rural domestic use, the amount or quality of water available to another or downstream urban industrial use; forest for fire- stakeholder or the benefits to either user from a con- wood, forage, construction, timber, or ground cover; servation intervention. The extent to which local orga- and land for buffer zones, arable cropping, or grazing. nizations can monitor change in a mobile resource in Watersheds often include a mix of privately owned a watershed and can obtain information about cause- and common property resources. The latter may be and-effect relatdonships in the management of a open access or may have self-organized or externally mobile resource will determine how easy it is for imposed rules that govern access and use. The trans- them to play a role in that management. boundary effects typical of resource flows in water- When cause and effect canrnot be determined read- sheds, whether they originate from common or pri- ily, conflict over usufruct rights is more likely to vate property, mean that proper management requires occur, and it is easier to cheat on regulations about collective, rather than solely individual, action. The use. Even more problematic is when a mobile need for cooperation, negotiation, and collective resource cuts across multiple locations, as in the case agreements to manage transboundary effects is a basis of upstream and downstream locations in a watershed: for involving local organizations. coordinating rules of use across widely dispersed However, cooperation in watershed management locations increases the costs of management to local is not necessarily easy to initiate or sustain, even people (Schlager, Blomquist, and Tang 1994). 120 Involving Local Organizations in Watershed Management The existence or creation of storage capacity in a shed resources and form the tapestry of local institu- common property resource helps users to establish tional arrangements in a watershed. Many other types some control over use of the resource. It is then easi- of local organizations are operative in watersheds: er for local organizations to devise rules and sanctions exchange labor groups, which may be sporadic or long governing use, and, because users have a better term, credit groups that manage rotating funds or com- chance of capturing the benefits of conserving and munity banks, farmer-to-farmer agricultural research developing a resource with storage capacity, they are or extension groups, marketing associations and coop- more willing to work together to maintain storage and eratives, religious groupings, as well as groups preserve the resource without the intervention of involved with local government. Although many of external authorities. Locally defined institutional these are not involved directly in the governance of arrangements for managing irrigation vary depending resource management, they are usually stakeholders on whether or not the resource has storage capacity with interests in one or more watershed resources. (Schlager, Blomquist, and Tang 1994), but no empiri- The presence of such organizations, and the degree cal test of this relationship has been published with to which participation in them is widespread among reference to management of watershed resources. the local inhabitants of a watershed, is important In summary, the physical characteristics of a given because the social dynamics of cooperation on which resource targeted for regulation in a watershed are local organization for natural resource management likely to affect the type of institutional arrangements often depends are fundamentally driven by the extent that are workable and the role of local organizations in to which the people involved trust one another and are these arrangements. The role of local organizations in therefore able to predict how other users of a common watershed management will be strengthened by property resource are likely to behave. This interper- improving their access to information about the sonal trust is generated and sustained by repeated face- effects of alternative management interventions on to-face participation in small groups, and the more mobile resources or by creating storage capacity in people interact in this way, the greater the level of trust resources that are a focus of conflict. and cooperation that is possible (Glance and Huberman 1994). Institutions, and specifically the Local Organizations and Social Capital rules, norms, and conventions that local people devel- op to govern voluntary cooperation for resource man- With reference to natural resource management, local agement, depend on each individual's assurance that organizations are often understood to be self-organiz- other people can be expected to cooperate and adhere ing groups or groups of users voluntarily cooperating to these rules. Field studies in villages show that peo- in the regulation of common property resources. ple in small communities are commonly motivated by There are numerous examples of these and extensive the idea of fairness, as well as zeal-the opposite of evidence that, all over the world, these self-organizing free-ridership--driven by the desire to obtain approval groups have been weakened and undermined by the from the community (Bardhan 1993). All local organi- growth of governrent intervention. As a result, natural zations are a channel for positive sanctions, group resources degrade). However, very few of these approval, and status acquired for altruistic behavior groups have organized spontaneously for watershed that spills over into natural resource management. management. In most documented cases from devel- Organizations are a structured sets of roles, which oping countries, external agencies have played an themselves are defined by bundles of norms and rules important role in catalyzing the formation of local about what social behavior a particular role involves watershed management organizations. (Uphoff 1992). Local organizations may define roles The special nature of local organizations is that in natural resource management for individuals or they provide a basis for collective action, which groups, such as monitoring the state of forests, irriga- includes building consensus about problems and need- tion channels, or burning for land clearance. Such ed solutions, seeking and disseminating information roles involve applying sanctions for noncompliance about these, coordinating action, designating rights with the locally defined rules for forest, water, or land and responsibilities, as well as ensuring accountability management. The wider universe of local organiza- (Cernea 1993 and Uphoff 1992). Self-organizing tions found in watersheds have a number of functions groups for common property management are only that, although they may not be concerned directly one subset of the local organizations that use water- with watershed resource management, have signifi- Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg 121 cance for the efficacy of local organizations that do not the primary motivation. Participation in the coop- have this responsibility. These other functions include erative resource conservation activities was most mobilizing resources, such as labor or credit, lobbying strongly predicted by membership in preexisting, self- external agencies, managing information, and mediat- organized precooperatives (averaging eight members), ing and resolving conflicts (Pretty 1995). which in turn was highly correlated with membership Organizational density, measured by the number of in labor exchange groups. White (1995: 1691) con- local organizations and the degree to which participa- cludes that tion and overlapping membership is widespread in a Participatiorn in these groups apparently pro- population (Cemea 1993), can be used as an indicator vides the recessary experience of sharing of social capital or the capacity to work together. The information, building trust, constructing rules, concept of social capital refers to reciprocity based on monitoring relations, and sanctioning. This interpersonal trust, which is the foundation of cooper- experience and the assurance of solidarity and ative behavior. A high level of social capital, or orga- reciprocity allow members to share risk, lever- nization, that builds experience in working together is age resources, extend payback periods, test an important ingredient in the capacity of local people innovations, and make leading contributions to to manage their affairs successfully without depen- collective actions even when they know that dence on outside agencies. For example, a study of 25 these ventures might not be rewarded. agricultural projects several years after completion An example from Kenya illustrates how the found that those which had made an effort to build degree of self-organization and decisionmaking (or local organizations, such as water user associations, social capital formation) that local organizations had been able to sustain high rates of return; projects achieve in planning and implementing small catch- that had not done this had rates of return 5-10 times ment resource conservation is correlated with their lower than anticipated (Cernea 1987). capacity to sustain locally managed conservation. In Social capital is fostered by local organizations 1988 the Kenyan Ministry of Agriculture adopted the that reinforce norms of reciprocity (favors done now catchment approach, which concentrates soil conser- will be returned later), create channels of communica- vation efforts in a specified catchment (usually tion that enable people to verify the trustworthiness of 200-500 hectar'es) for a limited time, during which others, provide a local institutional memory about all farms are planned and conservation practices laid experiences of successful collective action, and out. It also involves the formation of a local catch- implement sanctions against those who act oppor- ment conservation committee consisting of 8-15 tunistically. Social capital can grow, when local orga- people, sometiines all newly elected by the inhabi- nizations are active, and can be depleted when they tants of a catchinent and sometimes based on existing are not (Sirianni and Friedland 1995). traditional organizations such as groups of elders or Thus one of the key functions of local organiza- cooperatives. The catchment conimittee helps other tions in watershed management is building the social farmers to plan and carry out soil and water conser- capital required for people to build consensus about vation activities, after receiving training from the and enforce the agreed use of watershed resources for ministry. In some cases, catchment committees have the diverse, multiple, and often conflicting possibili- continued their activities after the ministry staff have ties described in the previous section. A study of the moved on to new catchments, expanding their work development of collective action groups to control to include marketing and road maintenance. A com- transboundary erosion in 22 small, multiowner water- parison was made of six catchments to assess the sheds in Maissade, Haiti, illustrates this phenomenon. impact of the approach. The study found that when The watersheds averaged 9 hectares in size and the catchment committees are freely elected and ranged between 2 and 34 hectares. After two years of actively involved in planning the conservation mea- activity, 590 check dams were built by the groups, sures, they tencd to be active after ministry staff have each of which retained an average of 39 tons of soil moved on and to develop a higher degree of self- per hectare per year, providing financial benefits to organization, lormulating their own management farmers from increased and diversified production. rules. These active committees are also associated After examining a number of factors in order to with independent replication of the new practices by explain what motivated individuals to cooperate, the neighboring communities without payment or sub- study concluded that short-term economic gain was sidy (Pretty, Thtompson, and Kiara 1995). 122 Involving Local Organizations in Watershed Management Local Organizations and Technical Innovation tive management practices are developed and adopted in Watershed Management rapidly. One example is the Empresa de Pesquisa Agropecuaria e Difusao de Tecnologia (EPAGRI) Watershed management, especially in low-income microwatershed program in Santa Catarina, Brazil. countries, usually involves a combination of conser- Catchment committees were formed in microwater- vation and development of natural resources. sheds of about 150 families, and these promoted col- Watershed inhabitants depend on these resources for lective action in conjunction with the agronomists of their livelihoods, and one of the basic challenges is to the state extension and research service. Extensive establish and maintain management practices that rec- testing with farmers of more than 60 species of green oncile their economic needs with the long-ran conser- manure and cover crops, no-till agriculture, and recy- vation of soil, water, and vegetation, especially where cling as manure the effluent from piggeries stabilized there are significant off-site effects of resource degra- erosion, improved yields, and improved water quality. dation, such as the silting up of watercourses or reser- Similarly, the impact study of six villages with catch- voirs of interest to influential downstream stakehold- ment committees in Kenya referred to earlier found ers. On-site and off-site degradation of watersheds is that the more active, self-organizing committees were widespread, and the management practices for con- associated with a higher percentage of farms adopting servation and regeneration are technically well under- conservation practices and that farmers associated this stood: land use planning to designate certain areas for with higher yields, increased availability of fodder, conservation, grazing, annual or perennial crops, and and groundwater recharge (Pretty, Thompson, and natural or managed forest; use of contour baniers on Kiara 1995). steep slopes; use of ground covers; creation of agro- Local organizations can initiate institutional forestry and agrosilvopastoral systems; creation of change by creating new rules for watershed resource buffer zones around watercourses and springs; and management that catalyze or induce technical change use of minimum tillage, green manuring, and other of various kinds. An example is the Agha Khan Rural soil improvement practices that regenerate soils and Support Program in India. This program forms a vil- permit the intensification of agriculture where suit- lage institution to take on planning, conflict resolu- able. However technically desirable these practices, tion, extension, monitoring, and evaluation activities their adoption by resource-poor farmers is very in small catchments. Collective watershed manage- uneven, in many cases because the short-term eco- ment carried out through village institutions achieved nomic returns to investing in these practices exceed widespread adoption of new practices in an area cov- the costs for the individual farmer. As a result, top- ering more than 3,000 hectares from 1988 to 1991. down efforts to enforce mandatory use of these prac- This not only increased crop and livestock productiv- tices or to promote them using temporary subsidies ity but also stimulated the collective use of farm have by and large failed to achieve lasting adoption. equipment as well as innovation in plant protection Involving local organizations in designing techni- practices, postharvest processing, and pooled market- cal innovations for soil and water conservation and in ing of produce and credit (Shah 1994). planning where to locate them in the landscape of a An important feature of their contribution to watershed is now recognized as a key element of suc- improved watershed management is that local institu- cessful adoption. There are many examples of tions can institutionalize local capacity for innovation enforced use of soil and water conservation practices through collectively organized experimentation with that have accelerated erosion once the structures stop new practices and can provide a conduit for receiving being maintained., because the local people have no and exchanging information about technical innova- commitment to keeping the structures in place. In tions, which brings down the individual's cost of other instances, elntire structures have been leveled as experimenting. The Agha Khan project is an example soon as outside agencies imposing them leave the vil- of how progressing toward collective action to imple- lage (Hinchcliffe and others 1995). In contrast, when ment watershed development reduces the costs of local organizations involve farmers in experimenting implementation: for example, comparing two peri- with and adapting the principles of known conserva- ods-1988-89 and 1990-91-costs of planning tion techniques to their own needs and constraints and watershed management have declined from Rs325 to in combining this information with local knowledge Rs25 per acre and the cost of arranging community of the landscape, soils, and vegetation, then innova- plowing decreased from Rs125 to Rs13 per acre Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg 123 (Shah 1994: table 1). A detailed case study of 22 par- effects of variations in flow and to gain consensus on ticipatory watershed development projects found that what are reasonable limits to use. Moreover, regula- one of the main results of involving local organiza- tion of use may have to be coordinated among geo- tions in changing farmers' soil and water conservation graphically disperse users in upstream and down- practices was the capacity building that enabled a stream locations, which involves a high transaction process of technical innovation to be sustained cost for local organizations built on face-to-face inter- (Hinchcliffe and others 1995). action. External agencies have a critical role to play in In summary, local institutions can play an impor- helping local organizations obtain and manage infor- tant role in promoting technical change in watershed mation about cross-scale or transboundary effects. management. Collective organization can reduce the Some authors have interpreted the need for expert costs of acquiring information and experimenting knowledge from outside agencies to mean that the with and learning about potential changes that require role of local organizations will therefore be dimin- action on a collective rather than an individual basis ished (Schlager, Blomquist, and Tang 1994 and and can build local capacity to innovate. Self-organiz- Uphoff 1986). An alternative interpretation is that by ing groups may also reduce the costs of conservation providing critical information about transboundary to the individual by making it more likely that envi- effects, external agencies can enhance the capacity of ronmentally damaging opportunistic behavior will be local organizations to mobilize local cooperation in sanctioned, so that the individual who innovates does watershed management (Knapp and others 1997). not bear all the costs of getting his neighbors to do so. This approach d'raws on research on the dynamics of Finally, institutional change-new rules and regula- cooperative behavior that shows that when a large tions about watershed management that are created by group containing several factions (or stakeholders) local organizations-can induce technical change by begins cooperating, it does so through progressive creating conditions that make it more economically transitions, with cooperation spreading from small feasible to undertake new land use systems and pro- group to small group. The subgroup with the greatest duction or marketing innovations that increase the tendency to cooperate at first will probably be the one benefits of conservation management. with the lowest average costs for cooperation or the longest time horizon; their decision triggers coopera- The Interface between Local Organizations tion in another group with the next lowest costs and so and External Agencies on until there is a "cascade of further cooperation" as information spreads about the benefits of cooperation Local organizations are referred to throughout the from small group to small group, reducing uncertain- preceding discussion as self-organizing groups volun- ty within them (Glance and Huberman 1994: 80). The tarily cooperating in the regulation of watershed relationship of external agencies to self-organizing, resources. Few groups organized at the watershed voluntary cooperation for watershed management is level have arisen completely spontaneously without one of facilitating the access of stakeholder groups to some intervention from outside agencies to catalyze information, supporting their skills in processing their formation and action. At the watershed level, information, and promoting the cascade-like sharing outsiders appear to play a significant role in enabling of information among these groups in a watershed. stakeholders to arrive at a joint plan of action that The Indo-German Watershed Program in takes into account transboundary effects not readily Maharashtra, India, provides an illustration of how perceived or measured by stakeholders other than technical infonnation about resource flows that are those who are directly affected. not readily perceived by local stakeholders can, when This may be especially the case when mobile as provided by the outside agency to local people, fun- opposed to stationary resource flows are the focus of damentally alter the resource conservation strategy. watershed management or when multiple, competing The program sought to enhance percolation over the uses exist for a given resource, such as forest, that are whole microwatershed so that it acts as a large under- not easily perceived by stakeholders in different loca- ground reservoir, creating underground flows that tions. When, for example, watershed management is reach the mid-slopes some two months earlier than concerned with stream flow, a mobile resource that the lower slopes and providing additional water to the can often fluctuate unpredictably, it is difficult for generally poorer farmers located higher up the slope. users to inventory, monitor, and assess causes and At the outset, farners preferred to build expensive 124 Involving Local Organizations in Watershed Management check dams on streams to improve irrigation and to The Landcare experience in Australia provides invest first in improving the lower slopes (where the some insight into the roles of local and nonlocal orga- better-off villagers live). Specialized survey and land nizations in a large-scale process. Landcare began in use planning teams provided technical information to 1989 when an important farmers union and conserva- a village watershed committee and integrated that tion lobby obtained political support for a 10-year information with local knowledge of the watershed funding program of community-based, conservation landscape. In a planning meeting, all landowners in extension groups. These groups involve a broad cross- the watershed were involved in the development of section of rural people with a stake in catchment plan- detailed action plans, which the village committee ning, as a framework for individual property plans: was responsible for implementing (Farrington and farmers, schools, scientists, state agencies, and Lobo 1997). This blending of local knowledge, mobi- agribusiness, for example. The Landcare groups fos- lized through the village watershed committees, with ter an ethic of land stewardship, fueled by extensive externally supplied technical information consider- voluntary participation in innovative environmental ably relaxed the rigor with which technical informa- monitoring (land literacy). Users employ techniques tion is usually applied in the planning process in order and information that were largely the domain of spe- to define a type and location for conservation inter- cialists a few years ago, such as geographic informa- ventions that farmers found acceptable. An important tion system (GIs) and aerial surveys. Nationally fund- function of the local organizations involved in water- ed facilitators and coordinators work with Landcare shed management in this program was to provide a groups, and one of their main functions is to assist forum for blending local knowledge with externally groups with process management-how to manage supplied technical information and to mobilize local participatory planning, conflict resolution, and infor- people together with outsiders in a process that mation sharing, for example. This has, in turn, enables them to adapt technical standards for planning changed the role of external agencies from being reg- and monitoring resource use to meet local needs. ulators of resource use to assisting the Landcare The process of sharing information reduces each groups in the process of learning and capacity build- group's uncertainty about whether other groups will ing (Campbell 1994). cooperate to implement conservation, clarifies expec- Another important point of interface between local tations, and achieves consensus about the action to be organizations and external agencies in watershed planned and implemented by local organizations. In management is the need for external intervention in many experiences of participatory watershed develop- identifying the relevant stakeholders and bringing ment, the appearance of local organizations like them to the table, whether this is literally a negotiat- catchment committees is catalyzed by participatory ing table or participation in collective labor, monitor- rapid appraisals, which use techniques of group ing, or enforcement of sanctions. The involvement of analysis to facilitate the collective diagnosis by water- local organizations in watershed management does shed stakeholders of the status of physical and social not ensure equitable participation of all relevant resources and the causal relationships leading to stakeholders in a resource management initiative. degradation (for an overview of these techniques, see Case studies show that women, the landless, marginal Pretty 1995). It is critical that this group analysis ethnic groups, and the laboring poor are all unlikely to result in a plan of action formulated by local people be represented in participatory watershed manage- (in contrast to a diagnosis used by the outside agency ment (Hinchcliffe and others 1995). to formulate the plan of action). Otherwise the level of Leadership in local organizations is generally cap- voluntary cooperation and local commitment to the tured by the higher status, more well-to-do members plan is likely to be weak and of short duration. of dominant elites, and the priorities of local organi- In most cases, local organizations, such as catch- zations tend to reflect the interests of these elites. ment committees, with an active role in managing Some watershed development projects deliberately watershed development operate at a relatively small seek to overcome this bias by setting up catchment scale, in catchments of a few hundred hectares at committees with representation of underpriviledged most. There are as yet few reported experiences where groups (Jensen and others 1996); the effectiveness of local organizations coordinate actions dealing with this approach has yet to be evaluated. In the Indo- transboundary effects in watersheds of several thou- German Watershed Development Program, the village sand hectares. committee is nominated by the village assembly Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg 125 (gram sabha)-the gathering of all persons in a vil- Techniques available to improve the quality and lage boundary with voting rights. The program guide- quantity of downstream water included practices such lines stipulate that the committee should include rep- as enclosing areas along riverbanks and around resentatives of all social groups and that at least 30 springs where natural vegetation could reestablish percent should be women. It is important to ensure itself and planting contour barriers to control soil ero- that stakeholder groups represent the different inter- sion. However, there was little incentive for upland ests in use of a given watershed resource, even if these farmers to adopt any of these practices, because they groups do not already exist or have representation bore the costs but not the benefits, which accrued to within existing local organizations, because marginal- downstream users (Ashby and others 1997). While ized stakeholders may undermine a conservation plan external agencies concerned with natural resource carried out by others, in which they have no vested management saw bad soil management practices as a interest. major problem, farmers were more concerned with the poor availability of credit, which made fertilizer The Process of Local Organization difficult to obtain. for Watershed Management: A Case Study Stakeholders had no common perception of the problems (Ravnborg and Ashby 1996). Inhabitants of This section presents a case study of the process of the lower and mid-altitude watershed communities organizing at the local level for watershed develop- were concerned about the seasonal scarcity of piped ment in order to give a concrete example of how many water, the disappearance of streams, and the decreas- of the general principles discussed earlier translate ing flow of water in the river. They attributed this to into action. The case also highlights the importance of deforestation by the inhabitants in the upper-water- applying these principles in a learning process shed communities. The agencies responsible to down- approach, in which process management constitutes stream users were concerned about water quality in an important role of local organizations. the Cauca Valley, whereas the upper-watershed com- munities considered their own major problems to be Organizational and Resource Management Problems the lack of schools, electricity, all-weather roads, and in the Rio Cabuyal health services. Different watershed resources were managed by The Rio Ovejas is fed by five main tributaries into the independent organizations with different objectives. RLo Cauca, which is an important source of water and The piped water supply drawn from the headwaters of hydroelectric power for the city of Cali downstream the tributary watersheds was managed by a local orga- in the Cauca Valley in southern Colombia. The water nization, the junta del acueducto, while an external quality of the Rio Ovejas was poor as a result of the agency was responsible for the conservation of forest heavy siltation from upstream clearing and deforesta- reserves around the aqueduct intakes. There were no tion and pollution by agrochemicals, waste products formal channels for any joint decisionmaking or coor- from artesanal coffee postharvest processing, and the dination among the local and nonlocal entities. Still runoff of organic fertilizer, which is used heavily by worse, several other external agencies active within small farmers in the tributary watersheds. The erosion the catchment area had competing agricultural or con- in the upland watershed is a major cause of annual servation objectives. There were no incentives for flooding problems in the lower-lying areas. these agencies to coordinate with one another or with Tihe Rio Cabuyal catchment, situated in the Andes the local comrmunities. Credit, technical assistance, at an altitude of 1,200-2,200 meters, was selected as and commercialization outlets were promoted by agri- the pilot area within the Rfo Ovejas because of its size cultural development programs, providing upland and representativity with respect to the soil, agrocli- farmers with incentives to clear and burn secondary matic, and demographic characteristics of the larger bush fallow and forests to plant cassava, notorious watershed. Small-scale farming and day laboring are locally for its effects on soil degradation. the principal sources of income. Coffee, cassava, maize, and beans, and in the upper watershed also The Process of Organizational Change fruits, are the principal crops. Livestock production is of minor importance. The lower and mid-altitude Organization to tackle these problems was initiated by parts have relatively good access to markets. outside agencies interested in watershed develop- 126 Involving Local Organizations in Watershed Management ment. An interinstitutional coordinating committee program in which each had a defined role and contri- was formed with representatives from the public sec- bution. It was agreed that conservation activities tor, nongovernmental organizations (NGOs), and required complementary agricultural, production, and grower associations to collate all available informa- commercialization activities, mirroring community tion and make a joint diagnosis of the problems. The plans to compensate upland farmers. Each agency committee, in consultation with local extension began to identify its activities as complementary with agents and community leaders familiar with the Rio those of others. A watershed-based organization, Ovejas watershed, selected the Rfo Cabuyal as a pilot which called itself the Consorcio Interinstitucional catchment area for diagnosis and subsequent inter- para la Agricultura Sostenible en Laderas (CIPASLA), vention. The boundaries of the 7,000-hectare water- was formed of 12 participating agencies at this work- shed were set to include all persons able to appropri- shop and was charged with formalizing agreements ate water from the Rio Cabuyal, which involved 1,000 among them and raising funds for the joint program. families in 22 communities as local stakeholders. An important objective of the organizers of At the same tirne, in early 1993, community lead- CIPASLA was to create a forum in which local inhabi- ers began to visit different areas of the watershed to tants of the watershed could define, monitor, and motivate local people to collaborate in improving enforce their own regulatory rules and determine the watershed management. They focused their campaign benefits going to different stakeholders. As a result, a on the decreasing availability of water, which they watershed users association was formally constituted associated with the clearing and deforestation around as FEBESURCA (Federaci6n de Beneficiarios de la the springs and watercourses feeding the reservoirs at Subcuenca de Cabuyal). The local leaders, active in the head of the watershed. They rapidly organized a promoting the association, defined the basis for mem- series of comisiones or task forces to visit communi- bership as representation of all the different local ties in the upper reaches. This was the first time many organizations and interest groups, such as the junta had gone to these communities. They found much del aqueducto, schools, cooperatives, women's greater poverty than in the lower altitudes: impassable groups, and village government. Membership reflected cart tracks, an absence of schools, low-quality hous- the perception of the local leadership that FEBESURCA ing, poor straggly crops, and an extensive wasteland would be a pressure group composed of people like of bracken-covered fallow left after cropping. themselves who could be expected to mobilize around Farmers in the upper watershed described their prob- their agenda to protect water sources in the upper lems and concerns and explained how the clearing watershed, which feed the tanks that store water and and cutting of trees around the reservoirs provided not supply piped water to the lower areas. only cropland but also charcoal, almost their only By mid-1993, CIPASLA had developed a substantial source of cash income. budget from outside grants, participating organiza- The community leaders' strategy to promote con- tions, and participating communities. FEBESURCA had servation of the rernaining forest around the reservoirs its own share of this budget and provided a forum for was to seek support from external organizations, via local organizations to define an agenda to present to the coordinating committee, for programs to improve the external organizations in the form of community- the quality of life in the upper watershed as an explic- based projects. The projects that were executed most it reward for the commitment of upland farmers to rapidly were those that met the demand of mid-altitude cease cutting and clearing. They also began to fornu- farmers to protect their water supply from the upper late plans for lower-watershed farmers to teach the watershed. upland farmers how to improve their crops and to dis- cuss ways in which farmer cooperatives could pro- Regulating Resource Use mote the marketing of produce grown in the upper watershed. Local community negotiation to define, monitor, and At the same time, external organizations in the enforce conservation in the uplands began to replace coordinating committee began to develop a common externally imposed (but generally flouted) regula- set of objectives and activities to achieve them, con- tions. The formation of CIPASLA enabled the local solidating their efforts to catch up with the evolving community to initiate new arrangements for conser- agenda of the comnunity leaders. At a planning work- vation zones in the upper reaches of the watershed. shop in March 1993, participants identified a joint Previously, extemal agencies had been unable to Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg 127 implement a long-standing program to enclose buffer watershed. FEBESURCA had been led by a local interest zones to regenerate forest around watercourses. The group concerned with water and reforestation issues, local junta del acueducto and the regional watershed which represented the interests of the mid-altitude authority, together with community leaders, swiftly farmers. At first, FEBESURCA had not yet evolved as an began to work with the upper-watershed inhabitants, effective mechanism for channeling demand from a whose land yielded the springs and streams feeding wider constituency, and CIPASLA'S organizational the water reservoirs. Within a short period, FEBESUR- structure had not yet encompassed some significant CA leaders persuaded the watershed authority to relax stakeholders whose needs revolved around land use its technical recommendations for enclosing and pro- rights and management practices, as opposed to water. tecting the forest and for regenerating natural forest By 1994, it was decided that each of the 22 commu- cover around these water sources. This agency pro- nities should elect a representative to FEBESURCA. vided technical advice, delegating monitoring and FEBESURCA conducted a stakeholder analysis, sanctioning to FEBESURCA. This new arrangement facilitated by outsiders, to identify the conflicting enabled community leaders to negotiate with the interests involved. Participants identified the reasons upper-watershed farmers mutually acceptable limits for burning as well as solutions, including the accep- on use of these areas for agricultural purposes. tance of some burning for land preparation where FEBESURCA leaders organized community labor drawn there is a shortage of land and labor. A set of norms, from upper- and lower-watershed communities to specifying how and when burnings should be con- plant trees and enclose areas to be protected by mutu- ducted, was developed, and some communities formed al agreement. Over a period of 18 months, 135 groups to ensure compliance. Subsequently, farmers hectares were enclosed, and 150,000 trees were plant- generally seemed to be following the recommenda- ed. The community rapidly mobilized a total of 3,714 tions, such as making firebreaks before buming. person-days for this purpose. They explained their These norms were more successful in controlling motivation for working together as making an invest- burning than previous measures because they were ment in the future of their children, for whom the developed by farmers themselves and acknowledged availability of water will be critical. Also, the junta burning as the only feasible means of land preparation del acueducto announced that it would turn off access in some circumstances, rather than simply condemn- to piped water for families who failed to contribute ing all burning. Their development illustrates the role labor. of local organizations in providing a forum to analyze Subsequently, GIS analysis showed that small trib- and negotiate interests (Ravnborg and Ashby 1996). utaries throughout the watershed were as important as the upper-watershed springs and streams for supply- Technical Innovation ing groundwater and stream flow to the lower slopes. This information led to a reevaluation of the rehabili- CIPASLA badly needed technological innovations that tation strategy, and CIPASLA began to promote conser- would promote conservation while enhancing produc- vation in small catchments and downstream tribu- tivity. In 1993 FEBESURCA formed local agricultural taries as well. research committees (CIALS) in the upper-watershed communities in return for their commitment to protect Conflict Resolution the forest and watercourses. CIALs are committees of four or more farmers, elected by their community, In 1994 an outbreak of forest fires partly destroyed a who test adaptive technologies in the local environ- buffer zone created by FEBESURCA to protect water ment, combinin,g local knowledge and exotic tech- sources. The burning was sparked by slash-and-burn nologies, on topics chosen by the community (Ashby land clearance by people colonizing the upper water- and others 1995). Based on the CIAL's adaptive test- shed and highlighted the failure to identify the semi- ing, new crops for upland farmers were identified, and landless as stakeholders in the watershed or as an producer groups formed, funded by CIPASLA. These interest group with representation in the watershed groups channeled specialist training and market users association (Ravnborg and Ashby 1996). The advice from technical agencies to upland farmers. burning made it clear that securing the buffer zones CIPASLA also provided the impetus for the develop- around watercourses would have to involve the poor ment of small-scale dairying, which began to stimu- and the landless inhabitants of the upper part of the late changes in land use. Cut-and-carry pastures were 128 Involving Local Organizations in Watershed Management planted, replacing annual crops, and a milk producers Development, London. Processed. cooperative was established. FEBESURCA leaders used Glance, Natalie S., and Bernardo A. Huberman. 1994. their political clout to improve roads in the upper "The Dynamics of Social Dilemmas." Scientific watershed to facilitate transport and marketing. American (March):76-81. Middlemen began to appear in the upper watershed to Hinchcliffe, Fiona, Irene Guijit, Jules N. Pretty, and purchase the product, providing a steady weekly cash Parmesh Shaw. 1995. New Horizons: The income for the farmers. The introduction of commer- Economic, Social, and Environmental Impacts of cial production was linked to adoption of the long- Participatory Watershed Development. Gatekeeper rejected contour barriers, live fences, tree plantations, Series 50. London: International Institute for and buffer zones, just as the community leaders had Environment and Development. visualized. Jensen, Jens R. 1996. "Watershed Development: Concept and Issues." In J. R. Jensen, S. L. Seth, T. References Sawhney, and R. Kumar, eds., Watershed Development: Emerging Issues and Framework for The word "processed" describes informally repro- Action Plan for Strengthening a Learning Process duced works that may not be commonly available at All Levels, pp. 43-63. Proceedings of DANIDA'S through libraries. international workshop on watershed develop- Ashby, Jacqueline, T. Gracia, M. P. Guerrero, C. A. ment, DANIDA, Hubli and Bangalore, India, Quir6s, and J. I. Roa. 1995. Institutionalizing December 1995. WDCU Publication 1. New Delhi: Farmer Participation in Adaptive Technology DANIDA. Testing with the "cLAL. " Agricultural Administration Jensen, J. R., S. L. Seth, T. Sawhney, and R. Kumar, (Research and Extension) Network Paper 57. eds. 1996. Watershed Development: Emerging London: Overseas Development Institute. Issues and Framework for Action Plan for . 1997. "Supporting Farmer Experimentation Strengthening a Learning Process at All Levels. with Local Agricultural Research Committees." In Proceedings of DANIDA'S international workshop Laurens van Veldhuizen, Ann Waters-Bayer, on watershed development, DANIDA, Hubli and Ricardo Ramfrez, Debra A. Johnson, and John Bangalore, India, December 1995. WDCU Tompson, eds., Farmers'Research in Practice, pp. Publication 1. New Delhi: DANIDA. 245-63. London: Intermediate Technology Knapp, Edwin B., Jacqueline A. Ashby, Helle Munk Publications. Ravnborg, and William C. Bell. 1997. "A Bardhan, Pranab. 1993. "Analytics of the Institutions Landscape That Unites: Community-Led of Informal Cooperation in Rural Development." Management of Andean Watershed Resources." World Development 21(4):633-39. Paper presented at Global Challenges in Campbell, A. 1994. "Landcare in Australia: Spawning Ecosystem Management in Watershed Contexts, a New Models of Inquiry and Learning for special conference of the fifty-second annual con- Sustainability." Paper presented to the internation- ference of the International Soil and Water al symposium for farming systems research and Conservation Society, Toronto, 22-26 July. rural development, Montpellier, France, 21-25 Processed. November. Processed. Pretty, Jules N. 1995. "Participatory Learning for Cermea, Michael M. 1987. "Farmer Organizations and Sustainable Agriculture." World Development Institution Building for Sustainable Develop- 23(8):1247-63. ment." Regional Development Dialogue 8(2): Pretty, Jules N., J. Thompson, and J. K. Kiara. 1995. 1-24. "Agricultural Regeneration in Kenya: The . 1993. "Culture and Organization: The Social Catchment Approach to Soil and Water Sustainability of Induced Development." Conservation." Ambio 24(1):7-15. Sustainable Development 1(2):18-29. Ravnborg, Helle Munk, and Jacqueline Ashby. 1996. Farrington, John, and C. Lobo. 1997. "Scaling up Organizing for Local-Level Watershed Manage- Participatory Watershed Development in India: ment: Lessons from Rio Cabuyal Watershed, Lessons fromn the Indo-German Watershed Colombia. AGREN Paper 65. London: Agricultural Development Programme." DFID Natural Resource Research and Extension Network. Perspectives 17. U.K. Department for International Schlager, Edella, W. Blomquist, and S. Yan Tang. Jacqueline A. Ashby, Edwin B. Knapp, and Helle Munk Ravnborg 129 1994. "Mobile Flows, Storage, and Self-Organized Sociological Association, Washington, D.C., 20 Institutions for Governing Common-Pool August. Processed. Resources." Land Economics 70(3):294-317. Uphoff, Norma:n Thomas. 1986. Local Institutional Shah, Parmesh. 1994. "Participatory Watershed Development: An Analytical Sourcebook with Management in India: The Experience of the Aga Cases. Rural Development Committee. Westport, Khan Rural Support Programme." In Ian Scoones Conn.: Kumarian Press. and John Thompson, eds., Beyond Farmer First: . 1992. Learning from Gal Oya: Possibilities Rural People's Knowledge, Agricultural Research, for Participatory Development and Post- and Extension Practice, pp. 117-24. London: Newtonian Social Science. Ithaca, N.Y.: Cornell Intermediate Technology Publications. University Press. Sirianni, Carmen, and L. Friedland. 1995. "Social White, Thomas A. 1995. "The Emergence and Capital and Civic Innovation: Learning and Evolution of Collective Action: Lessons from Capacity Building from the 1960s to the 1990s." Watershed Management in Haiti." World Paper prepared for the meetings of the American Development 23(10):1683-98. 11 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers John Farrington and Graham Thiele Time-series studies of public sector agricultural offered in urban areas or in commercial agriculture research have generally demonstrated high rates of that migration is unlikely to reduce the overall volume return (Evenson and Pray 1991). Methodological of rural poverty by much in the coming decades. refinement has removed some of the earlier upward More and more private commercial companies are biases, but estimates remain substantially positive penetrating into rainfed areas with, for example, (Farrington, Thirtle, and Henderson 1997). However, seeds, agrochemicals, and farm machinery. This tends many such studies either focus on areas of commer- to occur where rainfall is reliable, because fragment- cial or semicommercial agriculture alone, where data ed markets and poor infrastructure elsewhere make are good, or use aggregate national-level data and so transaction costs unattractively high. It has been lump together commercial areas with areas where the argued that these areas urgently need the dynamism of returns to agricultural research are likely to be lower. private sector research and development, but, ironi- In areas where commercial agriculture is strong, cally, they are the least likely to receive it (for agroecological conditions tend to be homogeneous (as research and development on seeds, see Gisselquist in irrigated areas), and it is relatively easy to replicate and Srivastava 1997). In addition, it is unlikely that farmers' conditions on research stations. Furthermore, the private sector will perform in a socially optimal farners themselves tend to be articulate and self-con- way in areas where externalities are high, such as soil fident in conveying their requirements to researchers, and water conservation, or in areas where some thus strengthening the client focus of research and resources are not privately owned but "common extension. By contrast, in many areas of rainfed agri- pool." culture, where agroecological conditions are complex, Attention has therefore turned to ways of improv- diverse, and risk-prone and where standards of physi- ing the effectiveness of the kinds of research institu- cal (communications, transport) and social (educa- tions likely to remain dominant in these areas in the tion, health) infrastructure are low, specific efforts coming decades, namely those in the public sector. A will have to be made to elicit farmers' perceptions of radical view is that public sector research services in needs and opportunities. Our concern in this chapter is small countries have been so ineffective that they with just such areas, where almost 1 billion people should be shut down and replaced by a low-cost but live in poverty worldwide. efficient capacity to "borrow" ideas, for instance In many countries of Sub-Saharan Africa and through the Internet (Zijp, personal communication). South Asia, population levels continue to rise in these Although we empathize with some of the principles areas. Moreover, the number of rural poor remains so underlying this notion, we see it as impractical large in relation t:o the employment opportunities because of the substantial knowledge needed of the 130 John Farrington and Graham Thiele 131 country's agricultural conditions, the skills needed in in the technology development and transfer chain, interpreting what is often highly condensed documen- as does a study by the Overseas Development tation of new technology, and, most fundamentally, Institute (Carmey forthcoming). the capacity needed to conduct adaptive testing of * The 50-country study by ISNAR on agriculture and "imported" technologies under local conditions. For environmental research in small countries better or worse, this requires a continuing research (Eyzaguirre 1996), whose central analytical con- capability. However, as will become clear in the cept is a portfolio approach to research planning remainder of this chapter, we recognize the limita- and management in which multiple agencies (uni- tions of the public sector and see multiagency versities, government institutes, parastatals, pri- approaches-not replacement by the private sector or vate companies, nongovernmental organizations, complete abolition of public institutions-as one way and donor projects) work together to address spe- of making good some of these limitations. cific problems according to the comparative A number of major international studies have iden- advantage of each tified these limitations and suggested ways forward. * The major workshop on future options for agricul- Common threads among these studies are: tural extension held by the World Bank in June * That the public sector has generally not performed 1996 (Zijp 1996), which placed strong emphasis well in difficult areas on the desirability of multiagency approaches, * That a major reason for poor performance is inad- because the public sector, especially given its equate orientation toward clients and inadequate declining level of resources in many countries, is participation by them in the design and implemen- unlikely to be able to meet all the demands made tation of research on it. * That improvements require changes in the perfor- The proposals for multiagency approaches deriv- mance assessment criteria and reward systems for ing from this and other work are now being tested in public sector staff various ways, particularly in externally assisted proj- 9 That improvements also require additional ects. Some of these projects simply involve contractu- resources to permit more work in farmers' fields al relations in which nongovernmental organizations and less work on-station; however, levels of (NGOS), for instance, are contracted to provide exten- resources available to the public sector are at best sion services in a specific area (see, for instance, the static and in many cases declining experiment of Chile's Instituto Nacional de * That stronger interaction with local organizations Desarrollo Agropecuario described by Berdegue having close contact with farmers (nongovern- 1994). Others involve the establishment of research mental organizations, membership organizations, grant (endowment) funds that NGOS and others can development projects) can create a "demand pull" draw on in order to commission work from the public on the public sector by helping to articulate farm- sector (in some cases also the private sector) that ers' requirements. An extension of this argument is meets clients' needs. that these demands can be strengthened, and fund- A third category of innovative multiagency ing problems of the public sector alleviated, where approaches, which forms the basis of this chapter, such organizations can either use their own funds goes beyond narrowly contractual relations and seeks or draw on endowment funds, established by synergistic interactions between different types of donors or governments, to commission research. organizations. Here, all work together to analyze par- Such studies include: ticular problems, contribute jointly to their solution, * The nine-country study by the International review progress in an iterative fashion, and make Service for National Agricultural Research (ISNAR) course corrections by mutual agreement as necessary. on on-farm client-oriented research (Merrill-Sands Although some success has been achieved in intro- and Kaimowitz 1990), whose discussion of mech- ducing participatory approaches at the diagnostic anisms for linling research and extension could, as stage of the research cycle (for instance, participatory Farrington and Bebbington (1993) argue, be adapt- rural appraisal is now widely used, although often ed to embrace organizations outside the public sec- badly), much less attention has been paid to ensuring tor and to strengthen public-private links continued interaction between researchers and their * The ISNAR study led by Eponou (1996), which clients during the identification, testing, and dissemi- examines the role of formal farmers organizations nation of technologies. 132 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers Initiatives in the area of agricultural technology organizations to seek functional participation in an need to be set in the context of wider perspectives on effort to generate and disseminate technology more scaling up. The major review by Edwards and Hulme efficiently. Clearly, at times there are gray areas (1992) of NGOS' efforts to introduce participatory between functional and empowering forms of partici- approaches suggests four types of scaling up: pation, which represent different ends of a spectrum * Working with government to spread NGO methods rather than uniquely distinguishable categories. and change policy Nevertheless, as has been argued elsewhere * Conducting lolbbying and advocacy (Farrington 1997), the distinction is important in pol- * Expanding NGOS' own operations icy contexts, not least because each approach tends to * Strengthening the network of local membership be associated with very different kinds of organiza- organizations with which NGOS work. tions. Our concern here is largely with the more func- For reasons set out here, the premise underlying tional types of participation. this chapter is that public sector resources have to be Much of the literature on participation reports brought in to permit scaling up in this context. Our efforts to experiment with different participatory discussions therefore relate to the first type of scaling methods, often involving the staff of government up. The chapter focuses on a specific aspect of the organizations, NGOS, universities, or special projects question of scaling up participatory approaches where in discussions with individual farmers or, occasional- only limited success has been achieved to date: that is, ly, village groups, although the basis on which these the organizational and management measures that are groups are formed is rarely clear (see, for instance, the necessary if participation is to become part of the International Institute for Environment and mainstream of government research and extension Development, London, series notes on the participa- strategy. tory learning approach; Chambers, Pacey, and Thrupp 1989). Although interaction of this kind is clearly fea- Participation sible and useful, it does not necessarily form a basis for wide-scale implementation. A wide body of evi- Nongovernmental organizations providing development- dence now suggests that it is unrealistic to expect low- related services (as distinct from membership organi- resource rural people in difficult areas who have long zations such as farmers associations) tend to work been economically and politically marginalized to more commonly in health and education than in agri- have the confidence or skills to engage directly with culture and natural resources. Where they do work in "outsiders" such as government agencies (Carroll agriculture, they often apply the same Freirean princi- 1992). Many see an important role for NGOS in help- ples of participation (Freire 1972) as they do in other ing to build such skills and confidence and in helping sectors. Intensive face-to-face approaches with small, to support the emergence of diverse kinds of local local groups, often over a period of years, facilitate: organizations, so that interaction with government * The emergence of a "critical consciousness" of organization staff becomes more cost-effective than if wider social, economic, and political conditions it were conducted solely on an individual basis. facing rural communities that may bear on agricul- Many of the NGOS involved in agriculture are char- tural and natural resource constraints acterized by: * The identification of how communities can respond * Strong interest in low external-input agriculture to these constraints using their own resources * Ability to identify farmers' aspirations and the * The possibilities of drawing on external sources of needs and opportunities to which technologies support where their own resources are inadequate. need to be adapted This empowering approach to participation allows * Ability to identify indigenous knowledge and NGOS to take the moral high ground and to dismiss practice and so help in negotiating how they might efforts that do not reach these standards. However, complement modern technologies from a wider policy perspective, government organi- * Awareness of the wider contexts of livelihood in zations cannot mandate themselves to concentrate which initiatives toward agricultural change need resources in a few villages as do NGOS, and they have to be located. to make the best of the difficult job of achieving what- However, many observers raise valid questions over ever level of participation is feasible using resources their technical abilities, over possibly unjustified pre- spread thinly over a wide area. This leads government dispositions toward eco-based techniques, and exces- John Farrington and Graham Thiele 133 sive accountability to the rural poor, especially among India. Its population of some 1.5 million comprises a organizations that rely heavily on external funding. high proportion of scheduled tribes and castes, annu- Farrington and Bebbington (1993) air these poten- al rainfall is typically 800-1,000 millimeters, but tial strengths and weaknesses as a preamble to a major highly variable, irrigation is very limited, soils are empirical study covering 18 countries and more than poor, and the terrain is undulating (Alsop and others 70 case studies of NGOS' work in agricultural and nat- forthcoming). Conflicts among castes at the village ural resources development. They identify occasional level are common. The district has an atypically high cases in which NGOS and government organizations number of NGOS. many of whom seek to enhance rural succeeded in working together in a fully collaborative livelihoods through long-term empowerment mode, but many more in which the interaction was approaches. tangential, or through service-provision contracts, or Efforts over three years (1994--97) in Udaipur in which NGOS were working completely indepen- sought to identify how NGOs and government research dently of government, having tried and failed to and extension services might work more closely secure agreement or having decided that working with together. These efforts were initiated by the government was practically or philosophically impos- Government of Rajasthan and the Ford Foundation. sible. Their study identified the preconditions for suc- They comprised support to: cessful interaction among government organizations, * A central government Farm Science Center. This NGOS, and farmers groups in implementing participa- support permitted the center (which was located at tory approaches but only very few cases in which it a prominent Local NGO) to set up training courses was under way and none in which it had been expand- and experimentation in response to NGOS' require- ed to a substantial scale. ments, to invite representatives of some 35 NGOS Field-based efforts to establish wide-scale multi- and 10 government departments concerned with agency approaches to the generation and dissemina- aspects of extension or natural resource develop- tion of technology in agriculture and natural resources ment plus thte State Agricultural University and have proliferated in the past five years. These have related research stations to a quarterly forum meet- involved varying combinations of government organi- ing (which served as a basis for exchanging infor- zations, NGOS, farmers organizations, and, at times, mation, leading to joint field visits and some joint the private commercial sector. This chapter aims to planning of activities), and to create a documenta- draw out their main features, particularly with respect tion center for use by government organizations to the differing modalities for scaling up participatory and NGOS in the district and an agricultural approaches. It presents three contrasting case studies research fund to permit NGOS to commission and then draws lessons across the three.' The focus research that was not otherwise being conducted throughout is on organizational and management con- by government organizations. figurations for participatory approaches that elicit the * An external organization. This organization was views of rural people on their needs and their reac- mandated to assist in developing techniques for tions to the technologies on offer. monitoring the processes of multiorganizational Essentially, this involves collaborative approaches interaction, supporting individual organizations in which the respective organizations share objectives, (primarily NGOS) in adapting and introducing these make joint decisions on actions and course corrections, techniques, and in developing the confidence to and pool some resources to do so. Although one type of deal with senior government organization staff. It organization (for example, a government organization) was also charged with conducting a number of may help to finance others (NGOS, farmers groups), col- studies on issues relating to collaboration, with laboration toward shared objectives is essentially a developing a newsletter to document recent devel- higher order of interaction than simple contracting opments in NGo-government organization interac- arrangements, which are not considered here. tion, with generating confidence among organiza- tions to use its correspondence column as a means Case 1: Scaling up Participatory Approaches of drawing the attention of senior officials to inad- in an Unstructured Context equacies in collaborative efforts, and with engag- ing the interest of a local agency (the Farm Science Udaipur district in Rajasthan, India, is typical of the Center) in continuing with its publication. approximately 75 districts located in semiarid central * A number oJfNGOS. This support allowed NGOS to 134 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers collaborate with the extension services of the reluctance of some to think in terns of strategic Department of Agriculture in testing new crop development beyond the confines of a few villages in varieties in farmers' fields and funded a course in which they are working and failure to adopt a busi- participatory methods for the Department of nesslike approach in dealing with well-disposed Agriculture, N13OS, and others. senior government organization staff. NGO meetings A number of successes have been noted. In sever- with government organization staff are often charac- al cases, NGOS and government organizations have terized more by each NGO'S efforts to articulate its initiated joint projects. Training schedules are philosophy than to agree on concrete plans of action responding more closely to NGOS' and farmers' to which government organization staff can respond. requirements. Government organizations now know Overall, three years of work have laid some impor- more about the programs of other government organi- tant foundations, generated some collaborative actions zations than NGOs know about those of other NGOS. (and consolidated others) but led to improvements (as The flow of information across the NGO-govemMent yet unquantified) in agricultural productivity in only a organization divicle has improved. The newsletter on few hundred hectares. While the Government of recent developments has grown into a widely read Rajasthan and others are keen to carry this work for- vehicle for information and commentary. A number of ward into other areas, there is a growing consensus organizations see the value of process monitoring that, in order to be more cost-effective, efforts to scale techniques, both in making course corrections and in up participatory approaches will, for the future, have strategic planning. And, after a long struggle, senior to be organized on a more structured basis. government staff have accepted, and are prepared in a limited way to fhnd, the concept of paraextension Case 2: Scaling up Participatory Approaches workers put forward by an innovative middle-level in a Structured Context: An Example government organization official and several NGOS for from Microwatershed Rehabilitation implementation in areas chronically understaffed by government agents. Participatory varietal selection Microwatersheds of 500-1,500 hectares containing has been adapted to local conditions from a neighbor- one or more villages have become the focus of policy ing state and is being led at the village level by farm- attention in several South Asian countries as a natural ers' representatives. unit for implementing efforts to intensify agricultural However, a number of difficulties remain. From and natural resource productivity (Farrington and the government organization side, incentive and Lobo 1997). In India alone, they are attracting some reward systems in the university are such that $500 million a year of government investment, with researchers (even from the zonal station set up under additional funds from numerous donor-supported a World Bank scheme and mandated to service the projects. The biophysical concepts are well proven. area's requirements) have rarely participated in meet- Improved management of the common pool forest ings, let alone joinied in collaborative efforts. Lacking and pasture lands in the upper slopes: technical support, NGOS initially could not put up * Increases the production of biomass potentially clear, researchable proposals under the Agricultural useful in livestock production Research Fund (although this has now been addressed * Reduces erosion and runoff to some degree by allocating some of the fund to con- * Increases the infiltration of water so that water tracting independent technical support for the prepa- tables beneath the agricultural land on the lower ration of proposals). slopes rise, thereby allowing more water to be lift- The principal constraints, however, are deeply ed to counteract droughts within the season, reduc- rooted in philosophies and operational procedures on ing risk, increasing the productivity of existing both sides. Government organization programs crops, and in some cases permitting new crops to remain largely centrally determined, with perfor- be grown. mance targets set in relation to, for instance, the num- Improving natural resource management in this way ber of demonstrations held. There remains little room therefore potentially allows a quantum leap in agri- for flexibility, and the more enlightened senior gov- cultural production technology and in productivity. ernment organization staff have difficulty overcoming Pilot experiences typically indicate a doubling in crop prejudice, reluctance, and, in some cases, lethargy, productivity, increased employment opportunities, lower down. Problems on the NGO side include the diversification of the village economy, and, where John farrington and Graham Thiele 135 marketing opportunities are good, a several-fold Watershed Development Programme (IGWDP) began increase in milk production. However, the rehabilita- in 1989, but numerous informa]L preparations, includ- tion of resources has proven difficult to achieve sus- ing experimental watershed sites and contacts with tainably in anything other than small areas: tradition- government and other organizations, began in the al systems of common pool resource management mid-1980s. Currently with funds of some DM12 mil- have broken down under population pressure in many lion, the prograin covers 92,000 hectares in 20 dis- areas, and new joint action requires substantial face- tricts of Maharashtra, involving 50 NGOS working in to-face effort to create the necessary coherence and 74 watersheds. confidence at the village level and to define rights and responsibilities in an equitable fashion. The setting oJ appropriate criteria for the selection An additional difficulty in India is that several of watersheds, villages, and local-level NGO partners government departments are mandated to deal with and the design of local-level collaborative mecha- aspects of watershed development, and one-the nisms. Criteria for the selection of villages and water- Forest Department-has long had powers of "polic- sheds include, on the technical side: ing" in order to prevent depletion of the resource. * Assured irrigation on no more than 20 percent of Despite the recognized ineffectiveness of this net cultivated area approach, the department is reluctant in many areas to * Notable erosion, land degradation, resource deple- move to new, joint-management arrangements with tion, or water scarcity problems NGOS and village groups, in part because these limit * Villages located in the upper part of drainage sys- rent-seeking on the part of local officials. tems * Watersheds around 1,000 hectares in size and aver- Setting the Preconditions for Scaling up age rainfall around 1,000 millimeters a year at the Local Level * Village boundaries that correspond closely with those of the wvatershed Several factors argue for taking a structured approach * Cropping systems that do not include long-dura- to microwatershed rehabilitation: tion crops with high water requirements, such as T The need for a degree of consensus on rights and sugarcane. responsibilities within the village, or among the In terms of socioeconomic characteristics, villages affected, including agreement on mea- * Villages should be poorer than average with a high sures to protect the resource once recovery is proportion of scheduled tribes and scheduled under way (such as restrictions on livestock graz- castes. ing or the cutting of trees and fodder), on the * There should be no wide disparities in the size of pumping of water resources to irrigate crops, on landholding. the types of crops grown, and so forth * Villages should preferably have shown a concern * The need to bring together both villagers and tech- for resource conservation and should have a nical specialists in order to identify desirable soil known histoiy of coming together for common and water conservation and revegetation measures, causes. the levels and division of necessary financial As a condition for selection, villages must commit resources and material inputs, and the means by themselves to: which both sides can monitor progress * Ban the felling of trees * The need to progress along a toposequence from * Ban free grazing and undertake "social fencing" ridge to valley if the physical measures imple- for the protection of vegetation mented such as contour strips and tree planting are * Reduce any excess population of nondescript live- to be effective stock * The need for a "learning" approach, in which vil- * Keep water-intensive crops to existing levels, lages and outside agencies initially tackle a small preferably ban them segment of the watershed, decide on the basis of * Ban deep tubewells this experience whether they wish to proceed to * Contribute voluntary labor to a value of 16 percent the full watershed, and, if they do, incorporate of the unskilled labor costs of the project, with lessons from this experience. landless and poor single-parent households being The preparatory phase of the Indo-German exempt 136 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers * Start a maintenance fund for watershed develop- essary and acceptable. In addition, government maps ment showing individual landholdings and features such as * Take the steps necessary for achieving and main- streams were found to be inconsistent with both con- taining a sustainable production system tour maps and with reality on the ground. * Constitute a village watershed committee to main- The net area approach, subsequently developed in tain the common assets created. the IGWDP relies heavily on consultation with farmers Criteria for the selection of NGOS to support village in their own fields. The type and location of interven- organizations include: tions agreed with farmers are marked both on the * Their reputation and history, the extent to which ground (with lime) and on landholding maps. Fields they have achieved rapport with organizations of are assessed for slope, soil depth, soil texture, and ero- the people and of government, their perspective on sion status, and these data allow the costs of rehabili- watershed development, and their technical and tation to be derived. The village watershed commit- managerial capability tees and NGOS are presented with the maps generated * The length of time they have been active in the in this way and a copy of the principal spreadsheet in area the local language, so that they can discuss it prior to * Demonstrated willingness (in the event of weak submission. The net area planning approach allows familiarity with watershed management) to under- villagers to voice their preferences and priorities in take exposure visits elsewhere, to send village the design of watershed rehabilitation and to monitor youth and others on specific training programs, progress on the ground. However, it also allows tech- and to prepare and implement a demonstration nical inputs concerning, for example, tree species, project of at least 100 hectares spacing distance, planting arrangements, and bunding * Willingness to accompany village organizations from the Forestry Department. Also important, it pro- through a capacity-building program and to meet duces a plan that incorporates agreed work rate and the qualifying criteria before undertaking full funding norms and is readily accessible (and if implementation. accepted, fundable) by the government agency con- What is notable about the IGWDP's approach to cemed (in this case, NABARD). Using funds allocated NGOS is that its early experience of consultation with to them, the supporting NGO and village watershed the larger NGOS already experienced in microwater- committees in each watershed are required to hire a shed rehabilitation suggested that their commitment civil engineer for the preparation of full proposals, to long-term Freirean empowerment-type approaches and engineers are trained in the net planning to group formation and joint action and the inclination approach. of some toward purely indigenous development were A key feature of the IGWDP iS that it allows for a likely to predispose them against effective collabora- capacity-building phase of up to one year in which a tion with government agencies, against improved small segment of the watershed (typically 100 technologies coming from outside the project, and hectares) is rehabilitated. This allows villagers to against rapid scaling up. The IGWDP therefore decided develop the requisite skills, to determine whether they not to work with these organizations. feel comfortable with the approach, to suggest amendments as necessary, and to build up group- The design of village-level mechanisms for partic- based capacity for decisionmaking and joint action in ipatory planning, learning, and implementation. The the practical context in which these skills will ulti- initial planning approach used by the program was mately be applied on a larger scale. based on gross area planning in which cost norms pro- A technical support NGO-the Watershed Organiza- vided by NABARi]-National Bank for Agricultural tion Trust (WOTR)-created by the IGWDP provides and Rural Development-for specific types of inter- funds to village organizations and local NGOS for the vention and specific types of land were routinely capacity-building phase (currently to a ceiling of applied. A major shortcoming of this approach was Rs500,000-approximately $15,000-per watershed, that it relied on contour maps, which are inadequate to including administrative costs). capture features such as the extent to which individual During the capacity-building phase, WOTR pro- fields have been leveled. Existing farm-specific vides villagers and the supporting NGO with training improvements, such as leveling and bunding, are a in technical skills corresponding to the individual crucial determinant of what further measures are nec- components of watershed development, namely: John Farrington and Graham Thiele 137 * Soil and land management their involvement in rehabilitation and agreement to * Water management subsequent protection and access arrangements were * Crop management essential. The approach to obtaining agreement to * Afforestation undertake joint forest management and the experience * Pasture and fodder development with its implementation provide valuable lessons for * Livestock management adapting the approach to other areas. * Rural energy management * Other farm and nonfarm activities Establishment of channels for drawing on techni- * Community development. cal expertise following rehabilitation. Watershed These include, for instance, skills in surveying, rehabilitation increases the amount of fodder avail- staking, and nursery raising. Villagers and NGOS are able and so may permit the upgrading of livestock and also trained in interpersonal relations, social mobi- the intensification of production. Similarly, higher lization, and the management of village-based organi- water tables allow water to be pumped in order to zations. counteract the threat of drought, extend the season, or cultivate for a second season. The availability of Design of a sustainable mechanism for screening water reduces risk, allows higher-yielding or more and,funding individual proposals submittedfor water- valuable crops to be grown, and encourages higher shed rehabilitation. The IGWDP has established a proj- levels of input use. All of this implies a potential ect sanctioning committee headed by NABARD, which demand for new crops and management practices. has a strong reputation for probity and technical capa- Linkages with technical agencies in the period fol- bility. It is comprised of four representatives of NGOS, lowing rehabilitation are facilitated by WOTR, which the program coordinator, three representatives of the has a central role in the program's philosophy of cre- Government of Maharashtra, and one representative ating self-sustaining local organizations. It provides of the national government. Proposals have to be NGOs and village organizations with support and structured on agreed financial and technical norms, training in awareness creation, social mobilization, although there is some flexibility for modifying these and the planning, implementation, and monitoring of to accommodate new technologies. Once a proposal watershed development projects. WOTR also provides for rehabilitation has been approved, funds are autho- technical and managerial training support and puts rized from NABARD for both capacity-building and NGOS and community-based organizations in contact full-implementation phases, and NABARD iS involved with line departments of the Government of in monitoring progress. The central role played by a Maharashtra. WOTR has 29 staff covering the disci- respected national organization in assessing and plines of social mobilization, women's issues, agron- channeling finance to donor-supported projects is a omy, civil engineering, and computer applications. cornerstone of replicability: rupees can be channeled WOTR'S training is tailored to specific settings, using a through this mechanism once foreign funds have combination of structured workshops and less struc- dried up. tured techniques such as village meetings and expo- sure visits. The WOTR has proposed exploring farm- Mobilization of administrative and political sup- based and other income-generating opportunities to port from the early stages. The IGWDP focused on take advantage of the additional resources created by obtaining the political support of members of the leg- watershed development and providing extension islative assembly of Maharashtra, initially by inviting advice on environmentally sustainable and economi- them to see successfully rehabilitated pilot water- cally viable dryland farming systems. sheds. It then drew on this support in order to obtain a cabinet resolution implementing joint forest manage- Creating a Wider Institutional Structure ment arrangements in the state. These arrangements, and Expansion Pathways in turn, required the Forestry Department to work with village-based membership organizations and For several years prior to implementation of the pro- NGOS in designing and implementing rehabilitation gram, its architects were concerned that participatory measures. Given the high proportion of common watershed development should be replicable over property resources falling under Department of wide areas. Preparatory work was based on the Forestry control in the upper slopes of watersheds, premise that stakeholders need to be engaged at inter- 138 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers national, national, provincial, and local levels. village watershed committee, which, in matters relat- At the international level, the program receives ing to forest department lands in the village, works funds from two organizations under the German with the forest protection committee. During the Ministry of Economic Cooperation, both of which capacity-building phase, funds are channeled via have an interest in seeing the program succeed: WOTR into the NGO'S bank account, and the NGO iS German Technical Cooperation (Deutsche then responsible for contracting a civil engineer Gesellschaft fur technische Zusammenarbeit, GTZ) (diploma level) to help in drafting the watershed provides funds to WOTR for a 12-18-month capacity- development plan, together with the villagers them- building phase, and the German Development Bank selves, based on net area techniques. The engineer is (Kreditanstalt fur Wiederaufbau, KfW) provides trained in participatory net-based planning by staff of funds to NABARD, which disburses them to local-level WOTR. Toward the end of the capacity-building phase, agencies for the fuill-scale implementation phase (four the draft proposal is considered by the village water- years). shed committee and submitted to NABARD. At the national level, the principal stakeholders are If approved, funds for the full-implementation the Ministry of Finance (via NABARD) and the phase are channeled into a bank account operated Ministry of Agriculture. The Ministry of Finance is jointly by the NGO and the village watershed commit- ultimately responsible for the disbursement of funds, tee. WOTR provides ongoing support during the full- but the Ministry of Agriculture is keen to see devel- implementation phase, and NABARD and the program opment of watersheds on the ground and is not with- coordinator are responsible for monitoring and super- out influence. Channeling funds through NABARD has vision. Management costs go directly to the NGO, several advantages: whereas project funds go to the joint account of the • NABARD brings an interest by central government village watershed committee and the NGO. The expec- in the performance of the program. tation is that the role of NGOS will diminish over time * NABARD has an interest in raising the repayment as that of local-level membership organizations rates it has achieved historically in rainfed farming becomes stronger. Once the rehabilitation works are areas and so can be expected to commit itself to complete, half of the 16 percent contribution made by the success of the program. the village to the cost of unskilled labor is returned to • Individual NGOS and village watershed committees the village watershed committee to form the core of a can receive foreign funds channeled through maintenance fund. NABARD without having to obtain foreign Two principal expansion pathways are envisaged: exchange registration. * "Nodes" of approximately 1,000 hectares of * Several dozen NABARD staff have technical qualifi- watershed are used as a central demonstration that cations in subjects broadly related to agriculture neighboring villages come to see and as a potential and natural resource management. They feel com- training area for new villages forming a village fortable discussing technical issues with officials watershed committee; some village watershed of, for example, the forestry or agriculture depart- committees have already begun to register them- ments, and, in turn, command the respect of tech- selves as NGOS and to obtain the benefit of the nical staff in these departments. funding support available for NGOS while at the * Procedures developed with and through NABARD same time serving as a vehicle for a type of farmer- for the disbursal of foreign funds in this way lend to-farmer extension. themselves to any subsequent disbursal of * The intention is that the essential features of the Government of India funds. Maharashtra model be replicated in other states. At the state level, the principal stakeholders are These features include a cabinet resolution and the departments of agriculture, soil and water conser- various departmental orders analogous to those vation, and forestry. Ministers overseeing these passed in Maharashtra, the role of NABARD in dis- departments successfully promoted a cabinet resolu- bursing funds for agreed proposals, and the role of tion in 1992 in support of the program. This has been NGOS in supporting community-based organiza- a key move in facilitating supportive action by line tions and of WOTR in supporting both the commit- department staff. ments made by villagers and the fusion of local At the local level, during the capacity-building knowledge and technical norms in the net area phase, the village assembly (gram sabha) nominates a planning approach. Officials from other states John Farrington and Graham Thiele 139 (such as Gujurat and Andhra Pradesh) have come ever commanded, and some were already conducting to observe the approach and are expected to take a adaptive trials (Thiele, Davies, and Farrington 1988). franchise on it, allowing it to remain the intellec- An RELU was created in 1988 to link with interme- tual property of the IGWDP. diate organizations and facilitate extension work. The new model fitted well with the then-burgeoning liter- Case 3: Changing the Organization ature on increasing farmer participation (Farrington and Management of Government Services and Martin 1987), on intermediate organizations to Scale up Participation pulling down research (Roling 1988), and on multiple centers of innovation (Biggs 1990). It was internally This case study reports on the evolution in the eastem coherent and intellectually attractive but at the time lowlands of Bolivia of a model for strengthening links completely untested. RELU staff faced the challenge of among research, extension, and farmers (Thiele, making the model work. Wadsworth, and Velez 1998). Central to the model is the highly innovative Research-Extension Liaison Changing Functions Unit (RELU) explicitly charged with linking the state agricultural research center (Centro de Investigaci6n Different core iunctions have predominated within en Agricultura Tropical-IAT) with a broad range of the RELU without entirely displacing earlier ones. intermediate institutions, including NGOS, that carry Each function led to the creation of a new set of link- out extension. The dominant function of the unit has age mechanisms. The dominant function changed evolved over time and with each change new mecha- partly as a result of external pressures (principally nisms for linking with intermediate organizations from donors) and partly as difficulties with the previ- have been tried and adapted. As a result, although liai- ous function became apparent. son positions or units have been created elsewhere When the RE]_U was created, its job was seen prin- (Ekpere and Idowu 1989), nowhere else, as far as we cipally as that of transferring technology placed on are aware, has such a diverse yet coherent range of the shelf by the existing CIAT commodity-based struc- linkage mechanisms been developed. ture. RELU staff were based at the experimental station and spent part of their time working alongside CIAT Origins of the RELU researchers, whcm they were supposed to backstop, to find out about research advances. They advised exten- The Santa Cruz Department in lowland Bolivia is sionists on technology available in CIAT, gave talks to characterized by commercial farming, ranching, and farmers, and established demonstration plots. smallholder subsectors. Population density in the Verification trials were begun in 1991 and soon smallholder areas is low so that many farners can became the principal activity for some RELU staff, only be reached by long journeys across poor roads. with a consequent reduction in the importance of the From the mid-1970s to mid-1980s, the govern- bridging function. This move to adaptive research was ment research organization supported a conventional due to a number of factors: public sector extension service. Plagued by poor com- * Some of the material presented initially-for munications, long distances, weak infrastructure, and example, in agroforestry-was not ready for high absenteeism, it had little impact. Subsequently, extension, and further testing was essential. with advice and support from the then Overseas * Researchers (lid not always have technologies that Development Administration (United Kingdom) and were appropriate to farmers' needs. the World Bank, CIAT managers came to see the * A culture of action militated against sifting potential for building on existing links with interme- through research reports to find information on diary organizations. technology. Links were identified with four main categories of However, this shift into adaptive research and the intermediary organizations: NGOS, producers associa- strengthening of links with intermediate organizations tions (which mainly represented large producers), were accompanied by a weakening of links with CIAT development projects of the Regional Development researchers. RELU staff moved from their base in the Corporation, and suppliers of agricultural inputs. experimental station to the city office and, in some These organizations had a total of 129 full-time exten- cases, tested technologies obtained from outside sion staff, far more than CIAT's extension offices had Santa Cruz. 140 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers Initially, it had been supposed that the intermediate Linkage Mechanisms organizations were relatively competent at extension work. Subsequently, deficiencies were revealed, and a Rapid rural appraisals have proved to be a highly strong demand for training on the part of intermediate effective type of linkage and have helped to raise the organizations was identified. In 1993, following quality of interaction and coordination. United Kingdom-supported training to its own staff, Zonal meetings are probably the most powerful of the RELU began training extensionists. Modules were all the linkage mechanisms developed. They bring developed for t[raining in communication skills, together extensionists based in the zone, researchers including the preparation of talks, extension materi- from local experimental stations, and RELU staff. They als, extension methodology, and participatory rural are organized by the zonal coordinator who refers any appraisal. problems that cannot be resolved locally to subject In 1995 CIAT was reorganized using a matrix man- matter specialists, commodity researchers, or other agement model to promote a stronger farming sys- experts. During a typical meeting: tems perspective. Farming systems teams, including * A representative of each group explains the work both researchers and liaison RELU staff, were set up at proposed for the coming month to facilitate the the local experimental stations. Intermediate organi- coordination of activities. zations with technical backup from the farming sys- * Field problems are discussed and solutions shared. tems teams were encouraged to carry out their own * A talk or demonstration is given to address prob- verification trials. The RELU synthesized and diffused lems identified by participants. the results of this work. Recommendation workshops are organized each year in which researchers, RELU staff, and members of RELU Staff intermediate organizations discuss and update recom- mendations for a specific commodity, and these are Three types of staff have worked within the RELU, published as a booklet. This process has proved par- although their roles and duties have shifted over time: ticularly effective at increasing the interactions v Subject matter specialists are the largest group and between commodity researchers and extensionists. are responsible for groups of commodities. They Verification trials are carried out jointly by subject transform research findings into usable and acces- matter specialists, intermediate organizations, and sible extension recommendations, transmit this farmers with technical support from researchers. information ito intermediate organizations, and These tools play a useful role, despite inadequate feed research back to CIAT researchers. prior research on some technologies, differing degrees * Zonal coordinators were established in response to of involvement by extensionists, and variable feed- NGO pressure that RELU staff should work on an back to researchers. area and "systems" basis. They help to integrate Funds are allocated for collaborative activities the extension activities of all institutions working with extension organizations. Within the agroforestry in the area and ensure that extensionists receive the program, for example, small amounts of money have technical information they need from CIAT. In the- been made available to allow these organizations to ory, zonal coordinators draw on the skills of sub- estabLish small nurseries. Secondment of researchers ject matter specialists and researchers, when into the RELU unit has been used occasionally to pro- required, but in practice some zonal coordinators mote tighter links with the commodity programs. have moved into validation, where subject matter A manual explaining the functions of RELU staff, specialists have not been able to provide support, researchers, and extensionists in intermediate organi- or have replaced subject matter specialists on zations and procedures for technology transfer was training. Similarly, some subject matter specialists published and has been widely used. Loose-leaf tech- have concentrated their activities in a limited area nical sheets on CIAT technologies organized in a dis- and become zonalized. tinctive technical loose-ring binder were produced * Communicators, working closely with the two and have been distributed to 27 intermediate organi- other types of staff, operate an information center zations and 250 extensionists. Talks and courses on for intermediary organizations, prepare and dis- technologies for extensionists have been given. tribute publications targeted at extensionists, and However, these have tended to respond to immediate play a lead role in training. needs and have not met the need for longer sustained John larrington and Graham Thiele 141 training courses in more complex technologies. An comparison, and it is too early in most cases to con- information center for extensionists was set up. clude much about adoption. But the examples in box However, because only limited practical information 11.1, taken together with other inforrnation, allow cer- is available, the center is full of research reports and tain tentative conclusions. other documents of more value to university students CIAT has made more information available to than to extensionists. A database of extensionists and extension orga.nizations and in some instances intermediate organizations was established, but high (silage), this has led to technological change. turnover of extensionists has meant that this mecha- However, improved delivery of information about, nism has never become fully operational. and follow-up oni, promising technologies (granulosis virus) is still needed, and quality needs to be Did the RELU Approach Work? improved. The demand for research has increased (agroforestry), bujt this has not occurred with all areas Quantitative evaluation is difficult. The RELU has of research (silage) and is still not strong enough. been involved in so many technological areas that it is A number of factors contributed to successful hard to assess its impact, there is no clear control for implementation of the model. Box 11.1. Participatory Approaches Scaled up through the RELU Agroforestry research Silage Before the RELU was established, agroforestry An important constraint facing small dairy farmers in research was mainly carried out on station with little Santa Cruz is the lack of grazing during the dry win- farmer participation. The agroforestry subject matter ter months. Silage is used widely in highland areas of specialist chose technologies developed by Bolivia, but CIAT researchers did not regard it as researchers that seemed most promising for small appropriate for Santa Cruz. The subject matter spe- farmers in frontier areas who were diversifying into cialist in pastures, without the support of researchers, livestock and for whom sustaining pasture productivi- carried out verification trials with intermediary orga- ty and fencing were priorities. These included forage nizations on the use of silage. Economic analysis alleys, with rows of leguminous forage bushes and showed the technology to be profitable, and farmers timber trees undersown with pasture, living fence were very interested. The subject matter specialist posts, mainly for fencing pasture, and windbreaks. helped local organizations to carry out demonstrations Verification trials were set up to test these options, of the technology. By 1995-96, more than 100 dairy with the NGOS, subject matter specialist, and farmers farmers had adopted silage use. working together. Trial design was a single unreplicat- ed block compared with farmer practice. Granulosis virus to control soybean pest There was a lot of interest on the part of interme- In Brazil, a bioinsecticide based on granulosis virus to diate organizations in collaborating with the RELU in control the soybean looper (Anticarsia gemmitalis) is these validation trials. As of 1995-96 around 60 farm- widely used. cix.F researchers tested this technology ers were participating. However, it later turned out and showed it to be effective. Results were written up, that RELU staff had underestimated the complexity of a but no measures were taken to diffuse the technology. difficult technology: what was intended to be verifica- The subject matter specialist produced technical tion was in fact the first real testing. Nevertheless, sheets on the granulosis virus for extensionists and testing of prototype technologies by the RELU cut out arranged for on-farm demonstrations with the produc- research cycles and pulled agroforestry research down er association for soybeans. Initial results were to the farm level. As a result of the RELU's work, the encouraging, but the subject matter specialist, and agroforestry program itself was modified, and several others, left to complete a master's degree with researchers assumed responsibility for testing some project financing. The unit became overstretched, and technology that had been in verification. the promotion of granulosis virus stopped. 142 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers * The model was implemented following a process The chapter has reviewed partnerships in three of consultation with intermediate users. contexts: case study 1 reviewed partnerships in a pure, * CIAT'S smallness (it serves only the Santa Cruz relatively unstructured form; case study 3 reviewed Department) allowed flexible implementation and partnerships in a context in which restructuring of the modification of linkage mechanisms, where appro- public sector and of public-private relations evolved priate. over time; and case study 2 attempted to capture the * CIAT'S reputation as a research institution made it a major productivity gains offered by microwatershed- credible part;ner for intermediate organizations, based approaches. Microwatershed-based approaches particularly where it had a "new" technology (such are inherently more challenging for two reasons: first, as agroforestry) to recommend. because of the joint action needed to manage common The following factors limited the success of the pool resources and connect them with agriculture and, model: second, because of the need to integrate and sequence * RELU staff had little experience with on-farm test- the inputs from several government departments. ing, systems approaches, and farmer participation These three studies inform policy regarding the focus. Insufficient training was carried out in this institutional preconditions for sustainable scaling up area. of participatory approaches in several ways. First, it is * The RELU tended to be more reactive than pro- difficult to reconcile high degrees of participation in active. Mechanisms exist to prioritize and feed decisionmaking on technology options by local orga- information back to the research agenda, but they nizations (especially NGOS working with low-resource are informal and not transparent. farmers) with the government structure and perfor- * Operational costs were restricted during periods of mance norms designed to deliver services and inputs. financial insecurity. Where, as in Udaipur (case 1), an inclusive, participa- * Staff had to be trained and the model developed tory, and open-ended strategy involving larger and while day-to-day activities went on. This reduced well-established NGOS is pursued, it has proved diffi- the time and resources available for testing and cult to move from talking shop to taking action. The disseminating technology (see the granulosis virus history of relations among NGOS, their differing case). philosophies, lack of felt need to scale up, continuing (sometimes justified) doubt over what government Conclusions organizations can offer and whether trying to turn them into client-oriented organizations is worth the This chapter has argued that, especially under unreli- effort all have meant that the introduction of practical able rainfed conditions where the private commercial improvements to technology at the farm level by sector is likely to remain weak, there continues to be improving the articulation of farmers' needs has been an important role for public sector research and dis- slow. Although isolated and productive coalitions of semination. However, in such areas, it has been diffi- interests among certain NGOS and certain individuals cult to maintain an effective public sector capacity on or groups within government organizations have the ground, and research and extension have general- emerged, it would take (in addition to a more focused ly been only weakly oriented toward clients' needs. approach by NGOS) major reforms in the structure and Eschewing more radical solutions such as total performance of government organizations to improve abolition of public sector research capability and its service delivery radically at the farm level. NGOS are replacement by electronic media, we argue that the predisposed toward group-based approaches, even clients of research must place a stronger demand on where agricultural innovation is usually individualis- research if it is to be effective. Participatory approach- tic. There is some justification for this, given the need es are easy enough to implement on a small scale, but to strengthen self-confidence among villagers if they specific institutional arrangements have to be made to are to have any prospect of influencing government permit wide-scale implementation. These, we argue, services. Nevertheless, as relations with the public can best be based on partnerships between public and sector improve and incomes rise, much group activity private (especially nongovernmental) organizations is likely to become individualized. and farmers themselves, both to bring additional Second, improved management of common pool resources to bear and to help in reorienting the public resources will, however, continue to demand joint sector. action. Although a well-managed connection between John Farrington and Graham Thiele 143 common pool and private agricultural resources in Overall, there can be little doubt that cases 2 and 3 semiarid areas offers the prospect of substantial pro- offer prospects for scaling up participatory approach- ductivity gains and lower risk, it also imposes consid- es to technological change in agriculture and natural erable demands on organizations such as NGOS aiming resource management. Quantifying the changes on to support groups and help design joint action. the ground that are attributable to these new institu- Furthermore, as case study 2 indicates, watersheds tional arrangements is now of high priority. What is cannot be rehabilitated in an unstructured fashion: vil- striking is the degree of sophistication of organiza- lagers' views have to be sought and external technol- tional and management arrangements (case 2) and the ogy options considered, selected, and built into bank- critical revision of arrangements (case 3) needed to able projects that set out physical and financial norms, respond to earlier experiences or meet new needs. the rights and responsibilities of insiders and out- However, the differing local circumstances, and the siders, time schedules, sequences of activities, and so detailed local knowledge underpinning changes in on. Microwatershed rehabilitation promises signifi- organization and management arrangements, are such cant and potentially sustainable gains in the produc- that it is the principles identified, rather than the actu- tivity of agriculture and natural resources. However, it al experiences, that are transferable to other settings. also poses considerable challenges, both in the struc- turing and sequencing of interventions and in the Note requirements of local groups for support in joint action. Certainly, the demands are significantly 1. In cases 1 and 2, "structured" implies building greater than those made by participatory approaches on previous experience to enhance efficiency by in agriculture alone, whether individual or group agreeing in advance to preconditions for successful based. In this setting, the project in case 2 decided to collaboration, the respective roles of different organi- work with small, locally focused NGOS that had the zations, shared objectives, and, where possible, necessary value-based philosophy and had (or were intended concrete outputs. prepared to acquire) the necessary skills in manage- ment, leadership formation, conflict resolution, and References basic technical matters but did not have the entrenched predispositions of some larger NGOS. An The word "processed" describes informally repro- additional facet of this case was the scope and quality duced works that may not be commonly available of arrangements for drawing the technical and finan- through libraries, cial support of the relevant government organizations Alsop, Ruth, John Farrington, Elon Gilbert, and Rajiv into microwatershed rehabilitation and ensuring that Khandelwal. Forthcoming. "Coalitions of Interest: such arrangements, where necessary with adaptations, Partnerships fior Processes of Agricultural Change." could be transferred to other locations. New Delhi. Processed. Third, case 3 records the substantial changes made Berdegue Julio. 1994. "El sistema privatizado de in government organization and management in order extensi6n agricola en Chile: 17 afios de experien- to facilitate more participatory approaches in an area cia." Paper presented to the international sympo- where public sector research and extension faced sium for farming systems research and rural devel- chronic obstacles. A new multiagency approach opment, Montpellier, France, 21-25 November. emerged in which research interacted with different Processed. kinds of "intermediate" organizations working with Biggs, S. 1990. "A Multiple Source of Innovation farmers on an area-specific basis. In many ways, this Model of Agricultural Research and Technology represents the type of public sector response that Promotion." WVorld Development 18(11): 1481-99. appeared to be desirable in case 1: large farmers Carney, Diana. Forthcoming. Research and Farmers' already made vocal demands on research services; the Organizations: Prospects for Partnership. London: intention was that NGOS should do the same for low- Overseas Development Institute. resource farmers and that the public sector should be Carroll, T. 1992. Intermediary ANGOs: The Supporting organized and managed in ways designed to allow Link in Grassroots Development. West Hartford, this. A range of organizational arrangements and link- Conn.: Kumaxian Press. age mechanisms were used over the several years of Chambers, R., A. Pacey, and L. A. Thrupp, eds. 1989. experience gained to date in this case. Farmer First: Farmer Innovation and Agricultural 144 Innovative Approaches to Technology Generation and Dissemination for Low-Income Farmers Research. London: Intermediate Technology Institute. Publications. Farrington, John, and A. Martin. 1987. Farmer Edwards, M., and D. Hulme, eds. 1992. Making a Participation in Agricultural Research: A Review Difference: NGOs and Development in a Changing of Concepts and Practices. ODI Occasional Paper World. London: Earthscan Publications. 9. London: Overseas Development Institute. Ekpere, J., and I. Idowu. 1989. Managing the Links Farrington, John, C. Thirtle, and S. Henderson. 1997. between Research and Technology Transfer: The "Methodologies for Monitoring and Evaluating Case of the Agricultural Extension-Research Agricultural and Natural Resources Research." Liaison Service in Nigeria. The Hague: Agricultural Systems 55(2):273-300. International Service for National Agricultural Freire, P. 1972. Pedagogy of the Oppressed. Research. Harmondsworth, U.K.: Penguin. Eponou, T. 1996. Partners in Technology Generation Gisselquist, David, and Jitendra Srivastava. 1997. and Transfer: Linkages between Research and Easing Barriers to Movement of Plant Varieties for Farmers' Organizations in Three Selected African Agricultural Development. World Bank Discussion Countries. Research Report 9. The Hague: Paper 367. Washington, D.C.: World Bank. International Service for National Agricultural Merrill-Sands, D., and David Kaimowitz. 1990. The Research. Technology Triangle: Linking Technology Transfer Evenson, R. E., and C. E. Pray. 1991. Research and in Developing Countries. London: Westview Productivity in Asian Agriculture. Ithaca, N.Y: Press. Cornell University Press. Roling, Niels. 1988. "Extension Science." In Wye Eyzaguirre, P. 1996. Agriculture and Environmental Studies in Agricultural and Rural Development. Research in Small Countries: Innovative Cambridge, U.K.: Cambridge University Press. Approaches to Strategic Planning. Chichester, Thiele, Graham, P. Davies, and John Farrington. U.K.: John Wiley and Sons. 1988. Strength in Diversity: Innovation in Farrington, John. 1997. "Farmers' Participation in Agricultural Technology Development in Eastern Agricultural Research and Extension: Lessons Bolivia. ODI Network Paper 1. London: Overseas from the Last Decade." Biotechnology and Development Institute. Development Monitor, 30 March, pp. 12-15. Thiele, Graham, Jonathan Wadsworth, and Roy Velez. Farrington, John, and A. J. Bebbington, with D. Lewis 1998. "Creating Linkages: Lessons from and K. Wellard. 1993. Reluctant Partners? Agricultural Research and Extension Liaison in Nongovernmental Organizations, the State, and Lowland Bolivia." European Journal ofAgricultural Sustainable Agricultural Development. London: Education and Extension 4(4):213-24. Routledge. Zijp, W. 1996. "Report on a Workshop on Alternative Farrington, John, and C. Lobo. 1997. Scaling up Mechanisms for Funding and Delivering Participatory Watershed Development in India: Extension Held at the World Bank, Washington, Lessons from the Indo-German Watershed D.C., 18-19 June 1996." Reported in Newsletter of Development Programme. Natural Resource the Agricultural Research and Extension Network Perspectives 17. London: Overseas Development 34:8B9. London: Overseas Development Institute. 12 People- Centered Agricultural Development: Principles of' Extension for Achieving Long-Term Inmpact Roland Bunch During the past 27 years, a series of principles for flexible than most methodologies. Different organi- making agricultural extension work effective have zations in different countries have modified consider- been developed in Central America and spread to ably the specifics of the approach, without reducing mostly nongovernmental organizations (NGOS) in substantially the effectiveness of the basic ingredients. some 22 nations around the world. In a number of Therefore, we generally speak of people-centered countries, these principles have enabled programs to agricultural development as a series of principles. As triple families' basic grain yields at a total program long as these basic principles are employed, there is cost of less than $400 per family. Furthermore, many room for a large amount of adaptation to local cir- of these families have continued to increase their cumstances, farmaer needs, and institutional impera- yields after the program intervention: a recent in- tives, without significantly decreasing the effective- depth study showed that 15 years after program phase- ness of the overall approach. out, yields have once again doubled (Bunch and The basic principles are as follows: L6pez 1994). * Motivate and teach farmers to experiment with This chapter describes the inner synergism of new technologies on a small scale, thereby reduc- people-centered agricultural development, as well as ing the risk of adoption and providing a means its relationship to the situation of resource-poor farm- for them to continue to develop, adopt, and adapt ers in developing nations, the sustainability of agri- new technologies in a permanent scientific cultural development, and the empowerment of process of innovation. This principle is frequent- resource-poor people. The chapter also briefly ly referred to as "participatory technology devel- describes the history and spread of the principles of opment." people-centered agricultural development as defined * Use rapid, recognizable success in these experi- in Bunch (1982), their proven effectiveness (including ments to motivate farmers to innovate rather than the results of major evaluations), modifications made artificial incentives or subsidies. in their application, cases in which they have been * Use technologies that rely primarily on inexpen- scaled up, and lessons learned. sive, locally available resources. * Begin the process with a veiy limited number of The Principles of People-Centered Agricultural technologies, so that the program is focused, Development achieves the maximum possible success right from the start, and allows even the poorest farm- Although sometimes called a methodology, people- ers to become involved in the process. centered agricultural development is much more * Train village leaders as extensionists and support 145 146 People-Centered Agricultural Development: Principles of Extension for Achieving Long-Term Impact them as they teach additional farmers, thereby People-Centered Agricultural Development creating and nurturing a community-based multi- and the Situation of Resource-Poor Farmers plier effect. This principle is called "farmer-to- farmer extension" in many nations of Asia. All the major systems of agricultural extension used A good number of organizations already use one or in developing nations were transplanted from the more of these principles. Numerous farmer-to-farmer industrial countries. These systems were not used in extension programs in Southeast Asia train villagers industrial nations when they were going through their as extensionists, while a growing movement in South own processes of agricultural development but rather America uses participatory technology development, evolved well after they were economically developed. a name that emphasizes the development of technolo- Therefore, in many ways these systems are not appro- gy by villager farmers through small-scale experi- priate to the situation of resource-poor farmers in the mentation. Nevertheless, long experience in a diversi- developing world. ty of cultures has shown that the five principles, when In contrast, people-centered agricultural develop- used together, reinforce one another. That is, a syner- ment principles grew out of the experience of a group gy exists between the various principles. Training vil- of NGOS in Central America. They take into account lager extensionists, for instance, becomes much more the social, economic, and ecological conditions that complicated, and recognizable success much less exist in developing nations. common, when a program begins with multiple tech- The appropriateness of people-centered agricultur- nologies. It also becomes virtually impossible for al development for developing nations is further evi- farmers to experiment when the number of technolo- denced by the fact that these principles have been dis- gies is unlimited, especially when several of them are covered independently by a whole series of people quite expensive. Thus the principles achieve consider- and organizations around the world, from Polan Lacki ably more impact when they are applied as a group (1993) in Chile to PATECORE in Burkina Faso (see figure 12.1). (Atampugre 1993) and the Food and Agriculture Figure 12.1. The Synergy between the Principles of People-Centered Agricultural Development < ~~~~~~~~~~~~~~~Begin with a Use appropriate lmtdnme technologiesoftcnlge Use rapid, < > Motivate and recognizable teach farmers success to experiment Train village leaders as extensionists Roland Bunch 147 Organization's farmer field schools in Southeast Asia. own experience, Thus, when they themselves experi- Figure 12.2 illustrates this relationship between the ment with new technologies, they learn in the best principles of people-centered agricultural develop- way they can. ment and the situation of resource-poor farmers. On In the case of the second arrow, villager experi- the left is the list of principles and on the right are the mentation has a special value because many of the characteristics most often found in small rural com- technologies "transferred" to developing nations munities of the developing world. In such situations, come from industrial nations. Thus it is extremely we find not only obstacles and problems that a system important for villager farmers not only to validate of agricultural development must overcome, such as these technologies under their own particular condi- lack of formal education and economic infrastructure, tions but also to experiment with how they could be but also very important resources that a system of agri- modified, where necessary, to become more appropri- cultural development must take advantage of if it is to ate to the different circumstances in which they are be as effective and as efficient as possible. These being applied. include an abundance of well-motivated human In the third case, being able to validate and modi- resources and an ecology that in many cases is highly fy technologies for their own use, as well as to create favorable to very high rates of agricultural productivity. entirely new technologies, makes the farmer much The arrows indicate causal relationships. For less dependent on outside sources of information. example, the first principle-to motivate and teach This empowers the farmer. And in situations where farmers to experiment-has a causal relationship with infrastructure and information services are lacking, four characteristics of rural areas in developing the ability to adapt technologies and invent new ones nations. In the first case, people who have very little becomes extrernely important. or no formal education are not accustomed to learning The last of these causal relationships is perhaps the through the written word or by listening to presenta- most important of the four. If the world is to feed its tions. They leam much more efficiently through their rapidly growing population without destroying the Figure 12.2. The Relationship between People-Centered Development and the Situation of Resource-Poor Farmers in the Developing World Principles Characteristics of small rural communities 1. Motivate and teach farmers Lack olf formal education to experiment Ecological difference from countries with aL temperate climate 2. Use rapid recognizable success Lack oiF economic resources 3. Use appropriate technologies Abundant, well-motivated human resources 4. Start with a limited technology Lack oif power 5. Train villagers as extensionists Need for reorganization Lack of services and infrastructure 148 People-Centered Agricultural Development: Principles of Extension for Achieving Long-Term Impact planet's environment, we have to find a large number the people-centered agricultural development process of agroecological technologies that are appropriate to motivates them to continue experimenting and to the social, economic, and ecological circumstances of share what they know. developing nations. If we fail to involve in this mon- The third activity, that of villagers' sustaining the umental effort the millions of well-motivated villager growth in agricultural productivity wholly or largely leaders in the hundreds of thousands of villages in the by themselves, is a much more complicated issue. developing world, the threats of poverty and famine First we must ask ourselves what factors must exist in may well become acute. a community for the farmers to, at the very least, maintain high levels of productivity. Personnel from The Sustainability of Agricultural Productivity COSECHA (Asociaci6n de Consejeros para una Agricultura Sostenible, Ecol6gica y Humana) have In many developing countries extension services are asked this question of groups of agronomists in some overstaffed and underpaid. They lack means of trans- two dozen nations. The resulting list is virtually the port and often lack cutting-edge messages. Even if same anywhere the question is asked and almost they were clearly ahead of innovative farmers, exten- always includes the following: sionists are seldom working with more than one out of * The motivation to continue the development every four farmers, and even then farm visits may be process. so infrequent as to be largely ineffective (Saravia * Self-confidence and a respect for their own knowl- 1983). The result of this lack of resources is that gov- edge and culture. People who are convinced they ernments often target areas of high agricultural poten- are ignorant or incompetent all too often become tial, with the result that problems of poverty and incompetent (Schumacher 1973: 192). inequality become even more severe (information * The ability to organize and manage experiments. presented at an International Food Policy Research New pests attack crops, seeds degenerate, input Institute workshop on land degradation in the devel- prices rise, old markets dry up, and new ones oping world, Annapolis, Md., 4-6 April 1995). appear. The only way farmers can maintain pro- A way of meeting this challenge is for govern- ductivity and profitability in a modem, rapidly ments to enlist farmers in solving their problems of changing environment is constantly to be trying low productivity. Widespread experience on three out new technologies. continents shows that farmers can fulfill three major * Medium- to long-term use rights over a certain roles (and many minor ones) in furthering their own minimum of natural resources that are in satisfac- agricultural development: tory condition. Without a minimum of certain • Establish and manage experiments in order to resources-land and water-no one can produce modify those technologies already known and enough food to live well. develop new ones * Access to or ownership of adequate financial * Spread knowledge of useful technologies from one resources. This need not be very much. Most cur- farmer to another rent loan programs handle much more money than * Carry on, by themselves if necessary, the process- small farmers really need. But farmers do need at es of agricultural investigation and extension, once least some extra capital to risk in their experimen- they have learned them, thereby continuing to tation and invest in improvements. Most of this increase their yields. capital will usually result from their own higher If governments are to feed their growing populations, productivity. at a cost within reason, they must involve farmers in * A certain basic knowledge of biological and agro- these three activities. nomic processes. This knowledge is necessary in The people-centered agricultural development order to understand experimental results and process teaches vi'llagers how to experiment and teach decide what possibilities of improvement will be one another through the most efficient teaching most promising for future experimentation. process we know: learning by doing. Repeatedly, year * A diversified agriculture. Knowledge of a series of after year, farmers manage experiments. Some of crops, animals, and trees reduces risk and provides them also become extensionists, teaching farmers in a basis for future innovations. other villages what they themselves have learned. * The ability and motivation to share information And the success of the innovations they learn through about agricultural technologies with other farmers. Roland Bunch 149 No one farmer can experiment enough to continue The list appears to be long and difficult to achieve. improving his or her productivity. The only way However, in a l eople-centered program, very few of whole villages or areas can solve their problems these attributes require any special effort; most of and move ahead is for each farmer to be learning them are reinforced each growing season by the prin- from the experiments of dozens of other farmers. ciples used to design the program. Figure 12.3 com- • Organization-building capacity. With constant pares the principles of people-centered agricultural innovation, new needs and new opportunities will development with the factors of sustainability. Each present themselves. These will often best be seized arrow indicates a causal relationship between compe- or solved not through some preexisting structure, tent use of the principle and the strengthening of one but by new organizations, either permanent or tem- of the factors needed for sustainability. porary, that people will create if and when they are To take the first principle, for example, by experi- needed. menting, farmers gain the ability to manage experi- Catalysts in this process will be: ments through the time-honored method of learning * Contacts with outside sources of information and by doing. When farmers experiment, they gain a good support deal of basic agricultural knowledge. When farmers * Administrative capabilities, meaning the ability to know how to experiment, and are motivated to do so, plan strategically, to handle money and accounting they gain the ability to continue to diversify their agri- procedures, and to manage group dynamics culture. And when they are capable of constantly * Minimal rural infrastructure and access to markets acquiring information in this manner, they will, on a * A high rate of literacy among farmers. Although the sustainable basis, have something valuable to share process has worked in areas where functional adult with others. literacy is as low as 20 percent, higher literacy rates In conclusion, people-centered agricultural devel- make the process more efficient opment principles go a long way toward strengthen- Figure 12.3. The Relationship between People-Centered Agricultural Development and the Factors That Make the Agricultural Development Process Sustainable Principles Factors needed for sustainability Motivate and teach farmers to Motivation to continue the process experiment Self-confidence and self-respect Use rapid, recognizable success Ability to organize and manage experiments Rights over natural resources Use appropriate technologies Adequate financial resources Begin with a limited number of Basic agricultural knowledge technologies Diversified agriculture Train village leaders as Ability and motivation to share information extensionists Organization-building capacity 150 People-Centered Agricultural Development: Principles of Extension for Achieving Long-Term Impact ing precisely those factors that can make the develop- areas of life. Villagers' increased economic resources ment process self-sustaining at the village level. Thus, do not have to be used in agricultural activities; they it is the extension methodology-how we teach-that can be used in any walk of life. Nor will their organi- helps a program attain sustainability, much more than zational skills be used to build agricultural organizations the technology-what we teach. exclusively; they can be used to establish organiza- tions to further any ends the people choose. People-Centered Agricultural Development, Thus, while the process creates a sustainable Sustainability, and Empowerment agricultural development process, it also, by its very nature, empowers people in ways useful for all sorts It is also worthwhile to compare the factors needed for of development efforts. That is, we are not talking of sustainability with the factors identified by social sci- an effort that only makes agricultural development entists as sources of power in any society (figure 12.4). sustainable. Rather, we are talking about a process One can see immediately that virtually all the sources that provides people with a series of skills and atti- of power are inherent to one or several of the factors of tudes that will enable them to improve their own sustainability. That is, empowerment is inherent to any well-being in many aspects of life, from organiza- process that permits villagers to sustain the process of tional management to health improvement, and from agricultural development. If we wish agricultural employment-creation activities to home betterment. development to be an ongoing process-to be sustain- In a way, the people-centered agricultural devel- able-it must include empowerment. Unempowered opment process uses agricultural improvement, a farmers are simply incapable of carrying on the very widely present "felt need," to get people process. involved in a process that can start them on the road But villagers do not merely gain some sort of agri- toward, and provide them with many of the skills and cultural self-confidence. They gain self-confidence in attitudes necessary for them to continue, an overall general, which gives them more competence in many process of integrated rural development. Figure 12.4. The Relationship between the Factors of Sustainability of the Agricultural Development Process and Villager Empowerment Factors needed for sustainability Elements of empowerment Motivation to continue the process Self-confidence Self-confidence and self-respect Power of coercion Ability to organize and manage experiments Money Rights over natural resources Adequate financial resources Position Basic agricultural knowledge Prestige, influence Diversified agriculture Ability and motivation to share Knowledge information Organization-building capacity Organization Roland Bunch 151 People-Centered Agricultural Development Instead of having to work with each individual group History and Present Dissemination of farmers, the more expensive professional personnel become managers of farmer extensionists who do Most of the principles now included in people- most of the teaching. Therefore, each professional centered agricultural development were being tried agronomist trains, supports, and backs up a group of 6 out by one or another institution in the highlands of to 10 farmer extensionists. In this manner, each agron- Guatemala during the 1960s. In 1972, World omist can reach, intensively but indirectly, some 150 to Neighbors, with OXFAM /U.K.'s support, established 250 farmers. Even though the farmer extensionists are the San Martin Jilotepeque Integrated Development paid a small stipend (after the first year or two), the Program. This was the first program to employ all the reduction in costs per farmer reached is dramatic. principles of people-centered agricultural develop- World Neighbors and COSECHA in Central America ment in a single program. The results, partially have a standard of being able sustainably to triple the described here, greatly exceeded expectations. productivity of farmers' basic grain production (from Extemal evaluations confirmed the success, in agri- levels of between 400 and 800 kilograms per hectare) cultural productivity, health improvement, organiza- for a total program cost of less than $400 (in 1996 dol- tional quality, and so forth. Development lars) per farm family, including administration, salaries, Altematives, Inc., after studying 41 exemplary pro- transportation, and so forth. That is, if a program costs grams in seven nations around the world, gave its $50,000 each year for eight years, for a total of number one rating to the San Martin Program. $400,000, then about 1,000 farmers should have sus- Impressed by the results of this and a later sister pro- tainably tripled their productivity of basic grains. gram in San Jos6 Poaquil, Guatemala, World Neighbors This $400 per family cost includes leadership train- decided to apply people-centered development princi- ing, organizational strengthening, training in the basic ples to programs in Bolivia, Haiti, Honduras, Indonesia, principles of biology and agriculture, and, in most Kenya, Mexico, Peru, the Philippines, and Togo. In cases, agricultural diversification. Some of World every case except Kenya, these programs produced such Neighbors' programs have not attained this goal. Their significant results that they were heavily visited and first programs in Mexico (in Oaxaca and Tlaxcala considered worthy of emulation. states) were more costly. However, the San Martin Little by little, a series of other institutions have Program in Guatemala was a good deal more efficient adopted people-centered agricultural development than this, and the Guinope Program in Honduras cost principles for their own programs. Organizations as almost exactly what the goal would indicate. The diverse as the World Wildlife Fund, Trees for World Neighbors' Cantarranas Program cost about People/Germany, the Mennonite Central Committee, $450 per family and tripled productivity, while a pro- World Vision, Catholic Relief Services, the Aga Khan gram in Honduras, run by students from a college in Foundation, Cooperative for American Relief Oregon, United States, spent less than $300 per family Everywhere (CARE), Campesino a Campesino and tripled productivity. (Farmer to Farmer), German Technical Cooperation In any case, ihese figures are a good deal lower than (GTZ), Peace Corps, and Rodale International have those of virtually any other methodology. Many pro- applied the principles to programs they manage or gram proposals describe efforts that cost several thou- fund in one or more nations around the world. sand dollars per family benefited, while some cost well over $10,000 per family (25 times less efficient than The Effectiveness of People-Centered Agricultural the $400 goal). And most of the programs resulting Development Principles from these proposals never reach all the people they hoped to, nor do many of those people ever manage to The people-centered agricultural development triple their productivity sustainably. A fair number of approach not only has proven itself capable of achiev- agricultural de-velopment programs spend more than ing high rates of farmer adoption and increased pro- the $400 per farnily on artificial incentives alone. ductivity but, even more important, has done so at much less cost than other extension systems. Nature and SusVainability of the Impact Inasmuch as farmers become a major factor in the spread of both agricultural technologies and the devel- A key question is how much of the increases in pro- opment process, costs are considerably reduced. ductivity are sustained after the program's termina- 152 People-Centered Agricultural Development: Principles of Extension for Achieving Long-Term Impact tion. Are yields maintained after the outside inter- increased, and outmigration either diminished drasti- vention ends? A recent study of three of the earlier cally (in San Martin) or was reversed. In the Guinope programs (in San Martin Jilotepeque, Guatemala, and program villages, net inmigration occurred between in Guinope and Cantarranas, Honduras) showed that 1990 and 1994, in spite of heavy net outmigration yields not only maintained themselves, but continued, during the 1970s-a phenomenon without precedent in general, to increase significantly (Bunch and L6pez in Honduras, as far as we know. 1994). In San Martin, for instance, average maize One of the major conclusions of the study was that yields in the four villages studied were 400 kilograms sustainability does not reside in technologies. per hectare before the program (1972) and approxi- Whereas the technologies taught in these programs mately 2,000 kilograms per hectare at program termi- had a half-life of perhaps five years (after five years, nation (1979). But 15 years later (1994), with virtually half the technologies had been abandoned), the agri- no additional intervention in agricultural development cultural development process and increases in yields by any institution, maize yields averaged close to continued. Thus sustainability is not to be found in the 4,500 kilograms per hectare. That is, yields once correct choice of technologies. Markets change, input again doubled. In the same 15 years, bean yields (the prices increase, new technological opportunities second most important staple food in the area) appear, varieties degenerate, pests spread, and compe- increased 75 percent. The villagers themselves had tition becomes stiffer. Most technologies disappear, successfully carried on the process of agricultural sooner or later, in this fast-changing modern world. development. The hope for sustainability of agricultural develop- The picture was not, however, uniform. In the 25 ment is not in the nature of the technology, but in the to 30 percent of villages where the response was best, nature of the process. The study confirms that the yields continued to increase after program termina- social process-one of widespread experimentation tion. In the 40 to 50 percent of the average villages, and sharing of information, innovation, and group yields more or less maintained or increased only mar- problem-solving-is sustainable, year after year. ginally, whereas in the 20 to 35 percent of the villages Although formal studies have not yet been done, pre- where impact was poorest, yields decreased, although liminary empirical evidence indicates that similar they continued to be better than the yields at program phenomena have occurred in many other programs initiation. using people-centered agricultural development prin- Other signs of innovation were roughly propor- ciples around the world. tional. In the villages of poor impact, there was virtu- ally no ongoing innovation, whereas in the best vil- Obstacles to Achieving Such Impact lages, literally hundreds of experiments were made after program termination, and often more than 10 People-centered agricultural development principles successful innovations per community were being are not capable of overcoming all obstacles. They do used. require well-motivated extension personnel and a In San Martin, whole new systems of agricultural good understanding of how the principles can be production appeared. One of the four villages studied applied in each particular situation. Although some developed an intensive cattle-raising system in which World Wildlife Fund-supported programs in they were raising some seven head per hectare. Mesoamerica have been highly successful, such as the Another village became a major producer of fruit, Defensores de la Naturaleza Program in Guatemala while two others developed an intensive coffee- and the Calakmul and Linea Bi6sfera programs in producing system. And one of the villages even devel- southeastern Mexico, other programs using people- oped a small-scale (less than 1 hectare per family) sus- centered principles, because of a lack of motivation tainable forest management system, earning good on the part of personnel or extremely poor, top-down incomes by planting trees in open spaces (without administration, have failed. using a nursery) and cutting (on average) the number We are often asked under what farming conditions of trees they could harvest on a sustainable basis each the people-centered agricultural development princi- year. ples would not work. We suspect they would not work Other impacts were also studied. In the high- with mechanized, highly capitalized farmers, because impact communities, the number of organizations these farmers would probably not be willing to teach increased, land prices increased dramatically, incomes one another. They also might not work in areas for Roland Bunch 153 which no known group of two or three technologies Faso, technological offerings were limited by the would bring significant increases in yields, such as the severity of the environment and the paucity of inno- semiarid areas of the African Sudan and the higher vations known to be successful there), have shown reaches of the Peruvian Andes. remarkable levels of impact. The people-centered agricultural development A second major modification has been that of the process is also very difficult, or even impossible, in an Campesino a C'ampesino Program in Nicaragua. In institutional framework that discourages decision- this case, taking advantage of the villager dynamism making at a level close to the village extensionist that coursed through the country in the 1980s, the pro- (Pretty and Chambers 1994). Both the principles and gram initiated a process of villagers teaching villagers the technologies used must be adapted according to that spread a srmall number of beneficial agricultural differences in farming systems, cultures, land tenure, innovations across a much wider area much more environmental factors, accessibility of villages and rapidly than any other program. using people-centered markets, and so forth. Institutions that fail to provide agricultural development principles. They did this the space for such adaptations will rarely be success- almost totally with volunteer villager labor, with an ful in using the people-centered agricultural develop- absolute minimum of organizational and financial ment approach. support, and with the involvement of a minimum of Successful application of the methodology is also professionals. I'hus a program working in some 35 made much more difficult where previous develop- municipios-probably in excess of 300 villages-did ment programs were highly paternalistic, giving away not have, until recently, more than four professional all sorts of inputs, food, and other artificial incentives. agronomists on its entire agricultural extension staff Nevertheless, even in these sorts of situations, signif- nationwide. icant successes have been achieved. Some observers claim that this incredibly rapid expansion with apparent far-reaching success has Important Modifications in the Application resulted in low percentages of villagers actually inno- of People-Centered Agricultural Development vating within rnost villages. Other observers feel that Principles such rapid expansion with minimal supervision could only occur in a nation in which major change has As people-centered development principles have been occurred and villagers are fairly well-organized and applied around the world, they have been modified highly motivated, as is the case in Nicaragua present- considerably. This modification, by and large, has ly. Unfortunately, we in COSECHA have not been able been a very healthy one, not only allowing the process to follow the progress of Campesino a Campesino to take into account local variations and differing cul- closely enough to know whether these observations tural or economic factors but also helping to under- are justified or not. stand the limits to modifying their application without Some organizations have tried using criteria for reducing their effectiveness. selecting villager extensionists. These criteria often What are the major modifications and what have include having a grade-school education, having we learned from these experiences? First of all, we established leadership or prestige in the community, have learned that limiting the initial number of inno- and having practiced the suggested innovations suc- vations taught is easily the most important of the prin- cessfully on their own farm. The first of these criteria ciples, while at the same time being the one about has proven counterproductive in most cases, and the which program personnel, at first, are usually the least second is largely irrelevant. The third criterion, how- enthusiastic. In fact, there seems to be a remarkable ever, has proven to be of tremendous value, because inverse correlation between a program's cost-efficien- anyone who produces three times what his or her cy (measured in the number of families whose pro- neighbors produce will have all the credibility he or ductivity in basic grains can be tripled for each she needs. $10,000 spent) and the number of innovations intro- A fourth innovation has been that of trying to duced in a village in the first year of the program. arrive at scientifically valid conclusions at a 1 or 5 Even programs that did not consciously use people- percent level of significance from the farmers' numer- centered agricultural development principles, but ous and varied experiments. One methodology, whose early innovations were limited by other factors presently being tried in World Neighbors' South (for example, in the PATECORE Program in Burkina American programs, has been to convince each 154 People-Centered Agricultural Development: Principles of Extension for Achieving Long-Term Impact farmer to repeat each experiment several times experimental basis. In contrast, the Souza Tobacco (Beingolea Ochoa n.d.). Statistical analysis, using Company-which limits its work to tobacco produc- either hand calculations or modified stability analysis, tion, including the introduction of cover crops- has produced statistically valid results (Hildebrand employs a large corps of farmer extensionists. Both 1993). However, some observers suspect that farmers programs, although very large, have had excellent will not continue to repeat their experiments, because levels of impact. Of the large-scale programs working they have little personal motivation for doing so. with small- to medium-scale farmers, EPAGRI'S is per- Another approach has been to use the data from a haps the most successful in Latin America. large number of farmers who are doing a similar Perhaps the major innovation that EPAGRI has used experiment without repetitions on any one farm. to make people-centered agricultural development Again, the calculations can be done by hand or com- principles work in a large-scale program is to organize puter, and results have been found to be more reliable agronomists into regional teams. This procedure has and predictive than those from experiment stations been used precisely to reduce the verticality or (Rzewnicki and others 1988). This procedure could be bureaucratic nature of the organization. Although extremely important as more and more farmers begin there is a team leader, personnel feel a sense of team- to experiment. It will help farmers and scientists to work, of everyone working together to achieve a com- enter into a mutually profitable dialogue and will mon goal. This allows closer personal ties among make it possible for the tremendous amounts of valu- extensionists and administrators, more openness to able information that farmers produce to enter into the discuss issues openly and learn from one another, more formal repositories of agricultural knowledge. more freedom for decisionmaking at the village level, An additional variant is to incorporate people-cen- more upward flow of information from the field, and tered agricultural development principles into other less organizational distance between farmers and kinds of rural development programs. Many of these statewide administrators. Each one of these factors principles have been used in health programs, family tends to increase the effectiveness of large-scale agri- planning programs, vocational education programs, cultural extension programs (Pretty 1995). and even a participatory post-earthquake housing pro- gram. Indications are that, where applicable, they Lessons Learned have increased the effectiveness of these programs, also. Through the years, and having visited more than 250 agricultural development projects around the world, The Experience of Large-Scale People-Centered we in COSECHA have learned the following lessons: Agricultural Development Programs * The people-centered agricultural development approach can be very effective in a wide number In two cases, people-centered agricultural develop- of cultures, environments, and political settings. ment principles have been tried out on a large scale. * Limiting the initial number of technological inno- The excellent work in Nicaragua of Campesino a vations to one or two simple technologies that will Campesino, assisted by the collaboration of a good provide a dramatic impact and get the social number of NGOS, whose collaboration Campesino a process moving is probably the most important Campesino encourages, has already been mentioned. principle. The second case is that of EPAGRI and the Souza * Sustainability of the process of agricultural Tobacco Company, which collaborate with each other improvement is not to be found in the technologies in the State of Santa Catarina in southem Brazil. Well introduced, but in the social process of active, over 100,000 farmers have made major changes in farmer-run innovation and dissemination of ideas. their agricultural practices (including use of cover * The achievement of this sustainable social process crops, zero tillage, and improved pig raising) that requires a series of conditions that include villager have at least tripled yields. Many farmers are now empowerment as a necessary ingredient. harvesting maize at a rate of 7 tons per hectare and * The people-centered agricultural development more. process requires well-motivated personnel, a good In this case, EPAGRI uses people-centered agricul- deal of decisionmaking freedom at the grassroots, tural development principles, except that it has only and an ample understanding of the principles and recently begun using farmer extensionists, and on an how they can and should be applied in each partic- Roland Bunch 155 ular situation. Hildebrand, Peter E. 1993. "Steps in the Analysis and The self-confidence, organizational abilities, high- Interpretation of On-Farm Research-Extension er incomes, and overall empowerment achieved Data Based on Modified Stability Analysis: A are applicable to all participatory development Training Guide." Staff Paper Series. Food and efforts. This style of agricultural development Resource Economics Department, University of work can help to stimulate a much broader process Florida, Gainesville. Processed. of integrated rural development. Lacki, Polan. 1993. Development of the Small Farm: From Dependency to Self-Reliance. 2d ed. References Santiago: Food and Agriculture Organization of the United Nations. The word "processed" describes informally repro- Pretty, Jules N. 1995. Regenerating Agriculture: duced works that may not be commonly available Policies and Practice for Sustainability and Self- through libraries. Reliance. London: Earthscan Publications. Atampugre, Nicholas. 1993. Behind the Lines of Pretty, Jules N., and Robert Chambers. 1994. "Towards Stone: The Social Impact of a Soil and Water a Learning Paradigm, New Professionalism and Conservation Project in the Sahel. Oxford: OXFAM. Institutions for a Sustainable Agriculture." In Ian Beingolea Ochoa, Julio. n.d. "Planning and Scoones and John Thompson, eds., Beyond Conducting Experimental Field Trials in Peasant Farmer First: Rural People's Knowledge, Communities." Pamphlet published by Office for Agricultural Research, and Extension Practice. the Andean Area, World Neighbors, Santiago, London: Intemational Institute for Environment Chile. Processed. and Development. Bunch, Roland. 1982. Two Ears of Corn: A Guide to Rzewnicki, Phil E., and others. 1988. "On-Farm People-Centered Agricultural Improvement. Experiment Designs and Implications for Locating Oklahoma City: World Neighbors. Research Sites." American Journal of Alternative Bunch, Roland, and Gabino L6pez. 1994. "Soil Agriculture 3(4):168-73. Recuperation in Central America, Measuring the Saravia, Antonio. 1983. Un enfoque de sistemas para Impact Three to Forty Years after Intervention." el desarrollo agricola. San Jose: Instituto Paper presented at the intemational policy work- Interamericano de Cooperaci6n para la Agricultura. shop of the Intemational Institute for Environment Schumacher, E. F. 1973. Small Is Beautiful: and Development, Bangalore, India, 28 November- Economics as If People Mattered. New York. 2 December. Processed. Harper and Row. 13 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa Klaus Deininger Theoretical reasons and empirical evidence suggest for land reform that allows small farmers to realize that land reform may provide benefits in terms of both their productive potential at a cost that is comparable equity and efficiency. First, a large body of research to other types of government interventions (educa- has demonstrated the existence of a robustly negative tion) would be very desirable. relationship between farm size and productivity due This chapter describes a new type of negotiated to the supervision costs associated with employing land reform in which land transfers are based on vol- hired labor. This implies that redistributing land from untary negotiation and agreement between buyers and large to small farms can increase productivity (see sellers and in which the government's role is restrict- Binswanger, Deininger, and Feder 1995 for refer- ed to providing a land purchase grant to eligible ben- ences). Second, in many situations, landownership is eficiaries. The main focus is on Colombia, where leg- associated with improved access to credit markets, islation mandating negotiated land reform was passed providing benefits as an insurance substitute to in 1994. The chapter describes the difficulties encoun- smooth consumption intertemporally. By enabling the tered in implementing the new model, reviews solu- poor to undertake indivisible productive investments tions that have emerged, and, by comparing the initial (or preventing them from depleting their base of experience to that of Brazil and South Africa-two assets), this could lead to higher aggregate growth other countries that recently began implementing a (see Bardhan, Bowles, and Gintis forthcoming for negotiated approach to land reform-attempts to draw references). Finally, even in aggregate cross-country some lessons that might be of use to other countries regression, the distribution of productive assets- where the distribution of land is very unequal and more than the distribution of income-seems to have agrarian reform is rapidly moving up on the political an impact on aggregate growth (Birdsall and Londoiio agenda. 1997 and Deininger and Squire forthcoming). The chapter is structured as follows. We begin by Notwithstanding this apparent potential, actual discussing some of the reasons underlying the change experience with land reform has been disappointing in of the government's land reform policy and then elab- all but a few exceptional cases such as Korea, Japan, orate on the principles of market-assisted land reform Taiwan (China), and to some extent Kenya. Despite- and some of the difficulties experienced in making it or because of-this, land reform remains a hotly operational. We then discuss the mechanisms devel- debated issue in a number of countries (Brazil, oped in five pilot municipios and, by comparing them Colombia, El Salvador, Guatemala, South Africa, with prevailing practices, describe some of the details Zimbabwe) some of which are spending considerable associated with implementation as well as farm mod- amounts of resources for this purpose. A mechanism els that have been developed in these municipios. This 156 Klaus Deininger 157 is followed by a section comparing the mechanisms nation to support, or at least live with, exceptionally used in Colombia to those adopted in Brazil and high levels of rural violence that increasingly consti- South Africa and briefly discussing some of the impli- tute a drag on the whole economy (estimates in the cations for monitoring the new approach. Colombian press put the losses associated with rural violence at about 15 percent of gross domestic prod- The Rationale for Market-Assisted Land Reform uct). The government sees the reduction in rural vio- lence as an important goal of land reform. In Colombia, land reform has been a long-standing Third, structural adjustment made the lack of concern to correct the inequitable distribution of land, adaptability in the large-farm sector particularly bla- to increase the productivity and environmental sus- tant. Large mechanized farms that cultivated mainly tainability of agricultural production, and to reduce traditional crops with minimal inputs of labor were widespread rural violence. Maldistribution of land in highly indebted and therefore unable to adjust to the rural areas, dating back to the encomiendas given out new environment and take advantage of the opportu- following the Spanish conquest, has been reinforced nities for exports of nontraditional crops. Unable to and exacerbated in more recent times by a number of respond to the loss of agricultural protection in a pro- factors: ductive way, thte large-farm sector resorted to large- * Tax incentives for agriculture that imply that rich scale lobbying. Establishment of a dynamic small- individuals acquired land in order to offset taxes farm sector would, it was hoped, enable Colombia to on nonagricultural enterprises capitalize on its agroecological diversity and signifi- * Legal impediments to the smooth functioning of cantly increase its exports of traditional and nontradi- the land rental and sales markets. Share tenancy tional crops. has been either directly outlawed or, when this was None of these concerns is new. In the 1950s a lifted, discouraged by the fact that tenants receive World Bank mission identified the maldistribution of property rights to whatever land improvements productive resources, especially land, as one of the they have made, making it in principle impossible root causes of economic stagnation. In 1961 the gov- to terminate their leases ernment established the National Land Reform * Credit and interest rate subsidies plus dispropor- Institute (Instituto Nacional Colombiano de Reforma tionate protection of the livestock subsector that Agraria-INCO]RA) to bring about a more equitable provide incentives for agricultural cultivation with distribution of assets in the rural economy. However, very low labor intensity (World Bank 1996). for most of its history, this institute followed a cen- * The use of land to launder money acquired by drug tralist and paternalistic approach to land reform, until lords. 1994, when a law was passed to create the basis for a These factors have implications for factor use, market-oriented and more beneficiary-driven employment generation, and welfare in rural areas. approach. We will return to the details of this law after First, while small farmers were often driven off their briefly examining the rationale that motivated the traditional lands and forced to eke out a living in mar- government to shift to the market-oriented or negoti- ginal and environmentally fragile areas, much of the ated approach. best agricultural land continued to be devoted to extensive livestock grazing or was not farmed at all Why through the Market? due to violence (Heath and Binswanger 1996). Only 25 percent of the land suitable for crop production From an institutional point of view, the main reasons was devoted to this use, while the rest was left to pas- for the shift from an interventionist to a market- ture. This suggests that there are indeed large tracts of assisted model of land reform were the limited suc- unused or underused land that could be subjected to cess of centralized land reform and the elimination land reform in order to increase agricultural produc- of the traditional source of finance for INCORA. Even tivity-a notion in line with available empirical evi- though considlerable amounts of resources were dence.' spent on land reform (INCORA's average annual bud- Second, rural employment growth since the 1950s get in the late l 980s was about $140 million), almost has been dismally low, significantly below the stan- 35 years of operations produced little visible effect. dard even of Latin American countries (Misi6n Social In the early 1 990s the administrative costs of trans- 1990). This appears to have increased peasants' incli- ferring land were very high, amounting to about 50 158 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa percent of the total land reform budget or about purchase price (subject to an upper limit).3 $15,000 per beneficiary. INCORA appeared to be more * INCORA's role is limited to that of regulatory over- effective in regularizing spontaneous settlement on sight and grant disbursement. the frontier than in converting the landless into suc- * A decentralized institutional structure ensures that cessful agricultural entrepreneurs in areas transferred land reform is driven by demand and coordinated from large owners. Many beneficiaries of the INCORA with other govemment programs. This structure program abandoned full-time agriculture and rented revolves around the Consejo Municipal de out most or all of their land, in many cases to the old Desarrollo Rural (CMDR), which, through a sub- landlord. In the aggregate, almost 35 years of state-led committee for land reform issues (Comite land reform hardly affected overall land distribution; Municipal de Reforma Agraria, CMDR), is intended between the 1960s and 1990, the Gini coefficient of to create the institutional structure to integrate var- the operational land distribution fell only 3 percentage ious entities and make land reform an integral part points, from 0.87 to 0.84. of the municipal development strategy. Traditionally, INCORA was financed by a share of Despite favorable preconditions, and the govem- duties on agricultural imports. The elimination of ment's expressed determination to distribute 1 million these duties in the course of structural adjustment hectares within four years, the land reform program meant that other sources of funding-ones subject to had a disappointingly slow start.4 In the remainder of greater public scrutiny-had to be found. INCORA'S this section, we consider some of the obstacles that agreement to adopt a less patemalistic approach freed have contributed to this lack of progress and review it from the burden of many of its erstwhile functions, regulatory measures that have been taken in the mean- including selection of beneficiaries, adjudication of time to address and, if possible, eliminate some of land, provision of technical assistance, and account- them. ing, enabling it to focus more effectively on regulato- ry oversight. In fact, this reorientation was a condition What Is to Be Financed? of the continued existence of the institution.2 In addition to these institutional factors, structural It is well accepted that enabling the poor to establish adjustment enhanced the scope for market-assisted viable agricultural enterprises without being burdened land reform by increasing both the supply of and the by excessive debt requires a grant (Binswanger and demand for land. On the supply side, adjustment Elgin 1988 and Carter and Mesbah 1993). Based on reduced -the profitability of large-scale farming by historical experience, the Colombian land reform law eliminating agricultural protection while increasing establishes that the maximum possible grant is 70 the opportunities for investment in other parts of the percent of the price of land for a viable family farm economy. This made selling a more attractive propo- (defined as one that generates at least three minimum sition for landowners. However, the dynamic response salaries).5 The remaining 30 percent of the purchase to structural adjustment and the simultaneous decline price for land, plus any additional start-up investment, in annual crop production widened the inequality has to be obtained from other sources-either the between regions and worsened rural poverty in a farm household's own resources or a regular loan number of areas. This was accompanied by higher from a financial institution. It was hoped that by shar- levels of rural violence and demand for land. All of ing the risk, financial intermediaries would provide this suggests that achieving a sustainable reduction of additional assurance of the economic viability of land rural poverty will require more far-reaching measures, reform enterprises. possibly including a redistribution of assets. A shortcoming of this arrangement is that grant resources are restricted to the purchase of land, even Principles of Negotiated Land Reform though it is recognized that establishment of a farm enterprise as envisaged in the law will require funds The legal framework for negotiated land reform is to establish on-farm infrastructure and improvements characterized by three main principles: like drainage, sheds, and perennials as well as work- * Potential beneficiaries (with assets below a certain ing capital like animals, tools, and machinery. This minimum level) can negotiate independently with restriction has two undesirable consequences. landowners and, once a deal has been struck, are First, the explicit prohibition of using the grant for eligible for a grant of up to 70 percent of the land's anything except land purchases may drive up the price Klaus Deininger 159 of land. Instead of providing an incentive for benefi- about 15 hectares. Although the government's desire ciaries to bargain and obtain land at the lowest possi- to prevent the land reform program from contributing ble price, it creates incentives for collusion between to minifundizaci6n (the creation of commercially sellers and buyers to overstate land prices, divide the nonviable garden plots or periurban squatter surplus between them, and let the government foot the schemes) is understandable and legitimate, the impo- bill. This has indeed happened; landlords in many sition of a minimum farm size is unlikely to achieve instances have overstated the price of land, and-by this goal. covering the complete land value with the 70 percent First, such a requirement concentrates a limited grant-have obtained a subsidy element of 100 per- amount of resources among very few beneficiaries, cent.6 Consequently, in 1996 the price of land creating an agrarian bourgeoisie much better-off than acquired through direct intervention by INCORA was the average farmer in any given municipio. It also lower than the price of land acquired by beneficiaries neglects the considerable human capital and other through negotiated land reform in the open market, assets as well as experience with financial and mar- leading to widespread dissatisfaction and calls for a keting institutions that are necessary to operate a suc- return to the interventionist paradigm.' cessful 15-hectare farm.9 This tendency is exacerbated Second, even in the absence of such creative solu- by the provision of a land purchase grant only for the tions, the law created an incentive structure strongly transfer of land without complementary investments. biased toward the transfer of well-developed agricul- Conversations with farmers and land reform benefi- tural land that was already located close to infrastruc- ciaries suggest that they prefer the provision of ture and endowed with the necessary complementary smaller plots together with funds for complementary investment. This makes it more difficult to target the investments over the provision of large plots without program to the truly poor in outlying areas, an issue such investments. This preference is supported by the that is of particular importance because land reform experience of many past reform beneficiaries who has considerable scope to target poor individuals who were forced to rent out the larger part of their hold- are usually beyond the reach of more conventional ings owing to a lack of complementary resources.'° government programs. The inability to use the grant Second, a centrally imposed minimum size for investments such as drainage, irrigation, or trees requirement neglects the potential of the rural poor- that are attached to the land is clearly contrary to the especially those in proximity to urban areas-to goal of developing underused lands in inaccessible derive income from a variety of sources. It leaves lit- areas.8 It reduces land reform to a mere redistribution tle room for beneficiaries to expand their holdings of existing assets rather than the creation of new ones, gradually by renting or purchasing additional land and even if the aggregate cost per beneficiary were the fails to account for the considerable heterogeneity of same as under a program targeting underused areas. In conditions prevailing in different parts of the country. the more likely event that undeveloped land could be Decentralized administration of land reform is more acquired more cheaply and beneficiaries' own contri- likely to accommodate this heterogeneity by permit- bution would reduce the cost of establishing on- and ting local institutions to determiine matters such as the off-farm infrastructure, such an approach would be optimum farm size given local conditions. Local insti- associated with direct inefficiencies and a reduction in tutions should likewise decide how grant elements the number of beneficiaries who could be served by should be distri'buted, striking a balance between pro- any given budget. viding them to comparatively few beneficiaries or to so many beneficiaries that available resources are What Is the Goal of Land Reforn? spread too thinly for farms to be viable. To have land reform contribute to the creation of Who Are the Main Actors? viable agricultural enterprises, rather than a rural pro- letariat with plots too small to generate a sustainable The Colombian law clearly recognizes that negotiated livelihood, the government established that the land land reform can, work only if it is built on local initia- transferred under the land reform program would tive and if beneficiaries and their representatives have to provide beneficiaries with an income of at coordinate the functions previously performed by a least three minimum salaries. In practice (assuming centralized institution. The institutional structure average land quality), this works out to a farm size of established is exemplary from a conceptual point of 160 Making MIarket-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa view, envisaging (a) decentralized decisionmaking Measures Taken to Overcome Initial Constraints characterized by maximum local participation expressed through the preeminent role of the local To overcome these constraints, the following changes councils (CMDR and CMRA); (b) private sector involve- were made in the legal framework."' It was clarified ment in bringing together potential buyers, provision that the goal of market-assisted land reform is the of complementary credit, and technical assistance to establishment of viable productive projects (proyectos continue during the first two years of production on productivos) rather than the mere transfer of land. A the land received; and (c) the limitation of INCORA to mechanism was devised to facilitate the use of grant preventing the misuse of funds, ensuring that regula- funds to finance nonland investments, thus overcom- tory requirements are met, and coordinating the dif- ing the bias inherent in previous legal provisions. ferent government agencies involved. Although the concept of a minimum target income The structure, however, functioned poorly for land reform beneficiaries was maintained, this tar- because poor dissemination of the law-among both get income was reduced to two, rather than three, min- beneficiaries and :[NCORA staff-prevented the effec- imum salaries. The income-generating capacity of the tive participation of beneficiaries. By default then, productive project no longer is based on municipio- INCORA interpreted its mandate so broadly as to run level averages but has to be demonstrated by the ben- the program without private sector participation and eficiary in an economically viable farm plan, which in many cases without the participation of the institu- can include income from nonagricultural sources. tions that were expected to represent beneficiary Responsibility for the approval of land transfers interests at the local level. This provided little incen- was decentralized from INCORA'S national headquar- tive for the private sector to assume the functions ters to the regional offices, while mechanisms for envisaged in the law. Private real estate agents, who increased accountability to prevent abuse of this were supposed to act as information brokers in mak- authority were introduced. It was clarified that a func- ing potential buyers and sellers aware of market tioning municipal council and a municipal land demands, found it difficult to compete with compara- reform plan were preconditions for the introduction of ble INCORA services that were offered free of charge. land reform in any given municipality. In addition, INCORA'S continued involvement in the decisionmaking process-including its ability to The Practice of Market-Assisted Land Reform manipulate political levers-put the institution in a position to virtually guarantee a "successful" out- The experience gathered during the first two years of come. the law's implementation tells us little about the "Success" was defined more in terms of transfer- mechanisms required for the success of a program of ring land and exhausting available budgets than in this nature. To generate such insight in a rapid and establishing viable rural enterprises. Consequently, cost-effective way, the government has identified five even approved projects were deprived of economic pilot municipios, selected to reflect the heterogeneity analysis and reliable sources of technical assistance. of the country, where new methods of truly market- It is thus no surprise that the financial sector, still assisted land reform will be introduced.'2 Comparing conscious of a long history of forced lending to land the experiences of these pilots with INCORA land reform, remained unconvinced and cautious. reform practices illustrates a number of important Financing of the land as well as any complementary issues (see table 13.1). credit therefore remained a near-monopoly of the Two key insights have emerged from these com- government-owned Caja Agraria where political parisons: the central importance of productive proj- demands to extend credit to land reform beneficiaries ects and the great benefit that can be gained from a clashed with the desire to restructure this enterprise, structured and well-defined process at the local level leading to prolonged indecision and deadlock. that leads to the elaboration of a municipal land "Beneficiaries" meanwhile were sitting on the land reform plan, especially in situations where preexisting without complementary credit, without a clear idea of capacity is limited. The next section describes the how to organize production or marketing, and with main elements of such locally devised plans and the typical clientelist mentality ("INCORA will pro- briefly summarizes the experience of their implemen- vide") that market-assisted land reform was meant to tation. We then touch on the arrangements for moni- eliminate. toring and evaluating the pilot. Klaus Deininger 161 Table 13.1. Mechanisms to Implement Market-Assisted Land Reform in Colombia, Brazil, and South Africa Colombia &razil Task INCORA Market assisted INCRA Market assisted South Africa, pilots Land Selection by INCORA By beneficiaries; with Purchase or expropriation; Negotiated by community; Community initiative selection based on political means to increase average cost of $11,600 willing seller (including pressure; cost of transparency and per family; mainly banks)-willing buyer; $18,000-$22,000 provide technical legalization of expected cost of $3,000 per family assistance occupied lands per beneficiary Land financing 70 percent of land value Administration of grant Agrarian reform bonds Loan to approved Maximum grant of (up to $22,000) as grant resources and provision for unimproved land and beneficiaries from a R15,000 for planning (20 percent cash; of additional credit for cash for improvements commercial bank and land purchase 50 percent bonds); land purchase and and crops; beneficiaries (considerable subsidy 30 percent through working capital by in theory expected to pay element) Caja Agraria credit commercial bank back, but not enforced (lengthy delays) Beneficiary Point scheme for social Comprehensive Through INcRA based on Self-selection of Self-selection of selection need and agricultural registration; preselection examination of beneficiaries; clearance beneficiaries subject experience; in practice based on social criteria, agricultural knowledge; of price and title by to maximum income ad hoc selection based and final selection based in practice almost all State Land Institute; criterion (less than on individual farms on productive projects are regularized decentralized approval; Rl,500 per month) squatters occupied lands ineligible Farming Perceived to be necessary Key issue for selection; No specific arrangement; Up to 8 percent of Provincial plan as a project only to obtain a different forms of project value is precondition, but few definition bank loan technical assistance available for technical specific guidelines available; farm models assistance in project provided and no farm available at municipal preparation and models elaborated level implementation; farm models elaborated at state level Other Credit for land and working Independent financial Credit of up to $1,150 Access to PROCERA credit Responsibility of financing capital, the big bottleneck institutions to provide (average $610) for fond like other land reform beneficiaries causing implementation integrated credit for the and housing and $7,500 beneficiaries delays whole project (average $4,500) for working capital; 70 percent subsidy element; minimal cost recovery Off-farm Complex set of Identified and costed Provided by INcRA ($3,200 Grant of $4,000 per Services and to some investments interinstitutional in municipal land in 1994, now up to beneficiary, disbursed extent infrastructure coordination with littie reform plan $8,000), almost all directly to the provincial results up to now for roads community responsibility; coordination with the center still weak 162 Making Market-Assisted Land Reforn Work: Initial Experience from Colombia, Brazil, and South Africa which is essentially a means test based on assets. The The Municipal Land Reform Plan whole process provides an opportunity to integrate land reform into a broader program of capacity build- Three key functions of the municipal land reform plan ing and social assistance at the municipal level. In a are to provide a more systematic way of identifying demand-driven and decentralized process, this could potential beneficiaries and land, to provide farm mod- resolve at least some aspects of the potential conflict els that can be used to elaborate productive projects between the dual objectives of equity and efficiency, by beneficiaries, and to clarify and establish linkages objectives that are to some extent unavoidable if land to other institutions. The preparation of this plan, its reform is to make a long-term sustainable contribu- final approval, and the monitoring of its implementa- tion to poverty reduction. tion are the responsibilities of the municipal land Enabling potential beneficiaries to register at pub- reform committee. The existence of a land reform lic offices and police stations in outlying villages has plan of this type is emerging as a precondition for had a considerable impact on the number of registra- going ahead with land reform in a given municipio. tions. Contrary to the procedures followed earlier, the information supplied is checked for consistency. The Identification of Potential Beneficiaries names of rejected and accepted aspirantes (with rea- sons for rejection) are posted publicly. The publicity Even though a needs-based qualification system- of the selection process seems not only to have which assigned points to potential beneficiaries based increased accountability but also to have led local on their socioeconomic needs-existed under INCO- authorities to understand the tradeoffs between differ- RA, in practice the resulting decisions were almost ent programs, as well as the scope and limitations of always overridden by considerations other than maxi- land reform. In all of the municipios involved in the mizing equity or productivity. In order to prevent pilot, alternative programs have been initiated to what it saw as the unrealistic expectations and waste (temporarily or permanently) take care of the specific of resources that would result from potential benefi- needs of groups who will not benefit from land reform ciaries negotiating directly with landowners interested in the immediate future. These include chicken hatch- in selling, INCORA continued to select beneficiaries on eries and other microenterprises for female household a case-by-case basis once a given farm had been put heads, construction of rural roads under seasonal up for sale and central approval for the release of the food-for-work schemes, as well as relocation of farm- necessary land purchase funds had been obtained. In ers from environmentally fragile areas and reforesta- these cases, to be able to disburse funds quickly, sales tion of critical zones. of farms were often quite secretive. The selection The information obtained through the beneficiary committees that were established included workers of questionnaire is then aggregated at the municipio the existing farm who were generally careful not to level and used to provide a profile of potential benefi- admit too many contenders from outside.'3 ciaries that contains, among other things, a descrip- By contrast, the pilot approach is based on the tion of their capacity and human capital endowments, principle that establishing clear rules of the game, their current employment status, agricultural experi- ensuring the participation of beneficiaries in all stages ence, asset position, and access to land (as sharecrop- of the process, and increasing the flow of information pers or workers), their organization in groups (with are the best ways to generate "realistic expectations" some form of organizational structure), their current and create a transparent and competitive market for and past involvement in credit and financial markets, land. The process starts with a systematic information and access to other forms of government assistance. campaign throughout the municipio, which should Based on its profile of aspirantes, the local coun- result in a fairly complete inscription of all potential cil, in close collaboration with the responsible institu- land reform beneficiaries (aspirantes) in a registry to tions, develops a training program for potential bene- be maintained by INCORA. A brief questionnaire pro- ficiaries who passed the first stage of prequalification. vides basic information on beneficiaries' educational The rationale is that, without a basic familiarity with level, their agricultural experience (if any), their the economic requirements of specific types of pro- sources of income, and their access to other types of duction and with their own capacity to meet these government services such as education or health. This requirements, beneficiaries will be unable to judge the information is then used to conduct a prequalification, productive capacity of a given farm or meaningfully Klaus Deininger 163 negotiate with landlords. Most municipios seem to information not only among potential buyers but have decided to offer training to about twice the num- also among any sellers of land, specifically infor- ber of those tentatively qualified in order to allow for mation on thie mechanisms of market-assisted land attrition and to ensure that potential beneficiaries do reform and the modalities of payment under this indeed compete in trying to put together the most program viable farm projects. * Increasing sellers' awareness of the scope and potential for alternative forms, such as land rental, Selection of Land that could temporarily or permanently provide potential beneficiaries with access to land, imme- The availability of large amounts of unused or under- diately increasing efficiency and perhaps serving used land in large holdings implies that in a reason- as a springboard for the landless to acquire the ably fluid land market there would be plenty of sup- information and agricultural experience necessary ply to enable potential beneficiaries to choose the to put together a productive project most suitable lands and negotiate a competitive * Encouraging more effective collection of existing price."4 In practice, rationing was largely by nonprice municipal land taxes-a strategy in line with the means, including outright corruption, and the main central government's desire to enlarge the revenue challenge for landowners willing to sell was to obtain base of local governments and to reduce gradually clearance from the regional and central INCORA the need for central transfers. offices regarding the availability of funds to purchase Because there is little justification for a land their land, which was seen as the precondition for pro- reform progranm that would push up prices for land, ceeding with negotiations. thus transferring resources to rich landlords rather In the pilot municipios, this has been replaced by a than to poor beneficiaries, municipios are eligible for more systematic procedure. The first step is to deter- land reform grants only if they can demonstrate that mine ecologically suitable zones and, based on cadas- existing supply is between double and triple the tral information, to establish an inventory of the land amount of land to be transacted so as to ensure a truly according to size classification that could be used to competitive land market. identify target areas for agrarian reform. Areas where land reform would result in environmental hazards, Farm Project Development where soil fertility is insufficient, or where the exist- ing ownership structure is already characterized by The most fundamental challenge in the transition small- to medium-size holdings are thus eliminated a from a centrally administered and paternalistic land priori. This gives beneficiaries a better idea of the reform into a market-driven one is to change partici- potential supply of land for land reform, helps to set pants' mentality. The outcome of land reform will rely realistic goals, and puts into perspective the potential heavily on the initiatives and entrepreneurship of ben- contribution of land reform for solving the social eficiaries who rnust themselves identify opportunities problems of a given municipio. By relating effective for profitable economic development. This represents supply and demand, the systematic approach to land a profound reorientation on the part of farmers who selection that is required to establish a municipal are accustomed to waiting passively to be settled on land reform plan makes evident the potential of land plots preselected by government officials. Without the reform to push up land prices within the municipio. change in mentality that this reorientation implies, the This in turn identifies municipios in which outmigra- potential of market-assisted land reform to alleviate tion (or use of very small plots of land) is the only poverty and increase productivity is vastly dimin- viable option for executing land reform in a cost- ished. effective manner. Thus, in contrast to the approach followed by Experience from the pilots indicates that land that INCORA, in which agricultural production and produc- had traditionally been registered with INCORA was tivity were given little if any consideration, the artic- often of marginal quality and hardly suitable for land ulation of viable and economically feasible farm plans reform, while some of the best land continued to lie takes a pivotal role in the pilot. Under the traditional idle or underused. Specific measures to resolve this approach, beneficiaries generally elaborated their issue have included: "productive projects" after obtaining access to the * More effective and systematic dissemination of land, with little systematic guidance and no discretion 164 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa in the use of the technical assistance funds adminis- diversified crop production, and an important live- tered by INCORA. Farm plans were regarded as a stock component in virtually each of the cases. tedious necessity ior obtaining access to complemen- The requirements of land for the average land tary credit rather than as a legitimate justification for reform beneficiary have more or less halved, from receiving the award of public funds."5 Without a clear about 15 hectares under the INcORA-based market- understanding of the economic potential of the farms assisted model to about 8 hectares on average. The to be established, expected returns, and alternative average land purchase grant has shrunk from Col$22.5 offers of comparable land, beneficiaries' ability to million to Col$8.3 million (with an exchange rate of engage in substantive bargaining was greatly dimin- slightly less than 1,000 pesos to the dollar). If projec- ished. It was only natural that INCORA generally took tions about employment generation are correct, the the lead in "negotiating" with the landlord. This nego- land purchase grant needed per additional employ- tiation ordinarily amounted to a mere formality ment created is about $3,000, a figure that compares revolving around the price set by an official appraiser. favorably with other types of government interven- Without any training or capacity building before tions. accessing the land, it is not surprising that beneficia- A brief summary illustrates not only the diversity of ries showed little awareness of the importance of sup- models that are suggested for different agroecological port services or marketing. Even at its best, land conditions but also the high labor intensity of the farm reform remained a. largely closed affair that would not plans adopted, the importance of complementary generate significant externalities beyond the proper- investments, and the high working capital requirements ties concerned. in most of the models (see table 13.2).'7 The models are Compared with this, the process introduced in the characterized by high financial and economic rates of pilot municipios attempts to provide beneficiaries with return (between 60 and 90 percent), suggesting that a clear idea of productive opportunities consistent with even if some difficulties are encountered during imple- their abilities before they formulate productive proj- mentation, they are likely to be economically worth- ects that form the basis for "shopping" for land. To this while. Beyond these figures, the farm models have end, agricultural professionals, either from govern- proven to be of great importance in the debate about ment agencies or from local NGOS, are employed to land reform, especially at the municipal level. establish crop budgets for a range of options actually Information concerning productive projects, practiced in the municipio and to conduct training demand for land by different groups in the population, courses and hold meetings to disseminate them. current land values, available supply of land at vari- Aggregation of these into farm plans involving more ous prices, the social and productive infrastructure intensive land use and sustainable income for benefi- required to make these productive, and estimated ciaries provides the basis for the formulation of "pro- costs of different components of land reform (land, ductive projects." Once potential beneficiaries under- complementary on-farm investments, and off-farm stand them, they can adapt the model budgets to spe- investments) are all incorporated in the municipal cific farms that are available on the market and use land reforn plan. This, together with information on them in subsequent price negotiations. the contributions expected from different participants The shift in emphasis away from maximizing the (beneficiaries, central government, local institutions), amount of land and minimizing additional capital allows local authorities to put together a more com- investment and toward fully using available family prehensive and coordinated program of land reform labor has resulted in productive projects that are quite that is in line with the specific needs and opportunities different from those promoted before.'6 Three differ- (including the fiscal capacity) of the municipio. This ences are particularly noteworthy. All of the projects information identifies the need for governmental and include a "garden plot" component, setting aside nongovernmental institutions (NGOs, banks, local gov- about 1 hectare for domestic consumption needs ernments, providers of technical assistance, universi- (including chickens, one pig, and a cow) and intensive ties) that are supposed to play key roles in the dissem- cultivation of vegetables or fruits, with the surplus to ination, capacity building, execution, and follow-up of be sold in the market. Emphasis is on providing full the land reform process. This serves to alert municipal employment of the family's labor force throughout councils to cases in which, due to very low levels of the year, meaning that farm plans focus on high-value local capacity and beneficiary organization, broad par- crops rather than traditional bulk commodities, more ticipation of beneficiaries cannot be taken for granted. Klaus Deininger 165 In such circumstances, municipal councils may con- Financing and Linkages to Other Institutions sider specific measures to improve dissemination and capacity and to increase competition among the insti- Given the scant attention traditionally devoted to ben- tutions involved in land reform execution. eficiaries' economic viability and the high transaction Table 13.2. Main Economic Parameters of Productive Projects Developed for Four Pilot Municipios, by Model Number, Colombia Rivera Montelibano Puerto Wilches Fuente de Oro Indicator (1) (2) (3) (1) (2) (1) (2) (3) (4) (1) (2) (3) Farm Farm size (hectares) 6.0 5.0 4.0 10.0 7.0 12.5 8.5 6.5 6.5 6.0 8.0 6.0 Labor days used 917 660 537 224 380 1,543 492 1,170 1,385 1,187 1,064 4 Total investment 15.2 14.2 16.7 10.9 13.3 11.3 12.4 15.4 9.2 13.3 13.0 13.3 Percentage for land 9.2 8.2 9.7 5.6 7.0 4.8 6.4 10.4 4.2 9.0 8.8 9.0 Percentage for company 6.0 6.0 7.0 5.3 6.3 6.5 6.0 5.0 5.0 4.3 4.2 4.3 Grant 9.4 8.7 10.5 5.7 8.0 6.4 7.6 10.8 6.4 8.8 8.6 8.9 Own contribution 1.8 1.8 0.9 1.6 1.9 5.1 3.0 2.3 3.3 1.3 1.4 1.3 Bank credit 4.0 3.7 5.3 3.6 3.4 -0.2 1.8 2.3 -0.5 3.2 3.0 3.1 Grant/ employment 2.3 2.9 4.3 5.6 4.6 0.9 3.4 2.0 1.0 1.6 1.8 1.2 Working capital 11.6 4.4 8.0 13.0 13.3 7.6 6.6 4.0 6.6 10.6 12.2 16.0 Output Total 28.6 19.6 30.4 18.3 14.7 14.7 14.8 13.5 13.0 22.5 20.3 32.3 Per person 6.9 6.5 12.5 18.0 8.5 2.1 6.6 2.5 2.1 4.2 4.2 4.4 Land allocation (hectares) House plot 1 1 0.5 0.5 0.5 1.5 1.5 1.5 1.5 1 1 1 Tomatoes, beans 2 Tobacco, 2 1.5 maize Maize, rice, 3 6 sorghum Plantain 0.5 4 2 Rice, vegetables, 5 maize Manioc 2 Oil palm 3 5 Citrus 1 Papaya Rubber 5 Improved 3 2 2 6 1.5 6 6 1 1 2 pasture Note: Costs are in millions of 1997 pesos (with 980 pesos to the dollar). Source: Municipal land reform plans for the four municipios. 166 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa costs associated with repossession of the land that indeed developing their productive potential, and to served as collateral for land reform loans, it is not sur- deduct loan repayments at the source, rather than rely- prising that commercial banks are extremely hesitant ing on unrealistic expectations of foreclosure. to advance credit to land reform beneficiaries. The resulting de facto monopoly of Caja Agraria in pro- Monitoring and Evaluation viding credit opened the door to political maneuver- ing and-because this institution is itself undergoing Although use of a grant-based mechanism that relies restructuring-lengthy delays and bureaucratic on market transactions to redistribute productive requirements in the award of credit. As a result, many assets is an innovative approach, it involves a num- beneficiaries who have already accessed land under ber of largely unproven procedures and, without the INCORA mechanism continue to wait for the sale to close monitoring and supervision, may produce be finalized. In the meantime, the inability to obtain unexpected and unintended outcomes. It is therefore even working capital credit severely impedes their essential to have an information system that continu- productive performance (and motivation). ously monitors each successive stage of the land Under the pilot, a number of cooperative banks reform's implementation in addition to assessing its already active in rural areas have agreed to lend to land long-term impacts. This makes it possible to quickly reform beneficiaries and thus create competition with identify, evaluate, and rectify unforeseen deviations Caja Agraria. The preferred arrangement bears simi- from the program's overall objectives, as well as larity to contract farming. The banks work closely with determinants of successful implementation, lending the providers of technical assistance (ensuring that the the process a responsiveness so sorely lacking in farm business established by beneficiaries will indeed INCORA practices. This is particularly important in generate the desired revenues) and help farmers mar- view of the focus on decentralized implementation. ket produce. This enables them to supervise the use of Table 13.3 relates key components of monitoring and credit more closely, to ensure that enterprises are impact evaluation to the issues discussed in the Table 13.3. Key Variables for Land Reform Planning, Monitoring, and Impact Assessment Unit Municipal land reform plan Monitoring Impact assessment Beneficiaries Beneficiary profile (capacity, Grant per beneficiary and Increase in income welfare) employment Consumption smoothing Training requirements Group formation (assets) Additional employment Credit market access generated Social services Targeting efficiency Reduction of violence Improvements in access to land Projects Demand and supply of land Characteristics of farms Agricultural productivity Characteristics of transferred Environmental sustainability productive projects Implementation of projects Complementary Repayment performance investments needed (planned and actual) C'ost by component (who pays?) Institutions Institutional capacity Effectiveness in dissemination Strengthened local (technical assistance, and capacity building govemment finance, marketing) Efficiency of land transfer Fiscal sustainability process Private sector participation Klaus Deininger 167 municipal land reform plan. Key questions to be national reconstruction program, in an environment answered in this context are: in which productive small-scale agriculture was erad- * What instruments are best suited to reach particu- icated almost a century ago. Here, much greater lar target groups, maximize the net benefits of effort is required to establish the decentralized infra- land reform (or minimize the cost of its execu- structure necessary to implement land reform, to pro- tion), and are compatible with rapid attainment of vide complementary services such as marketing and the government's quantitative goals? How do they technical assistance, and to increase beneficiaries' perform in terms of beneficiaries' adherence to the agricultural and entrepreneurial capacity. project plans they have established (and the asso- ciated economic benefits)? Brazil * What are the direct and indirect impacts on agri- cultural productivity and poverty? How does this With a land distribution among the most unequal in compare with other instruments at the govern- the world, Brazil's situation is similar to Colombia's ment's disposal, such as construction of infra- in a number of respects. There is a very large and structure or support to education? And under what vocal political demand for land reform; a recent study conditions can land reform constitute a sustain- by the Food ancl Agriculture Organization estimates able and worthwhile use of public funds? that 2.5 million families are potential candidates for Monitoring describes what is happening on the land reform. A land reform institute (INCRA) was ground, compares it to what had been set out in the established in 1969, distributing 10 million hectares municipal land reform plan, and uses the information to 200,000 families and colonizing about 14 million to update the parameters contained in the municipal hectares for about 75,000 beneficiary families since land reform plan. In addition, the information pro- then. Land reform has recently acquired considerable vided can be used to (a) assess whether the mecha- political importance; a federal minister for agrarian nisms used reach target populations and areas (poor reform was appointed in 1996, and the land reform rural dwellers and underused lands), (b) provide an budget tripled from $0.4 billion in 1994 to $1.3 billion ex ante estimate of the expected benefits of land in 1995 with a further increase to $2.6 billion pro- reform, both for productivity and for poverty reduc- posed in 1997. tion, and (c) provide information on whether imple- Although the large majority of these federal funds mentation is progressing as expected. Impact assess- will be spent according to the old process, which-at ment, in turn, is concerned with the ultimate impact, a cost of about S30,000 per beneficiary-is not only both direct and indirect, of the program on household costly but also lengthy and bureaucratically cumber- well-being, agricultural productivity, environmental some, state governments in the Northeast have been sustainability, and institutional strengthening. moving ahead and setting up a decentralized market- assisted pilot scheme."8 This scheme is intended to Comparison with Market-Assisted Land Reform speed up the land reform process, reduce costs by in Other Countries more than half, and provide the basis for a model that could eventually be adopted nationwide. Brazil and South Africa have also, under different The main diflferences from the old mechanism are conditions, recently initiated programs of negotiated the following (compare table 13.1): land reform. With an institutional background very * Rather than relying on a lengthy process of expro- similar to that of Colombia (presence of land reform priation, land is selected by community groups on legislation and a central land reform institute dating a willing seller-willing buyer basis. This is expect- from the early 1960s), negotiated land reform in ed to reduce the price of land from $11,000 to Brazil has been initiated by individual states. The $3,000, mainly by avoiding the need to pay for purpose of the Brazilian interventions is to establish expensive land improvements that are of little use cheaper, more agile policy alternatives to centralized for small-scale agriculture. land reform in an environment where the issue of * Landlords are compensated with cash instead of land reform is high on the political agenda and poten- highly discounted government bonds. This pro- tial beneficiaries have at least some idea of what to vides a strong incentive for landowners-includ- do with the land. By contrast, negotiated land reform ing many banks who hold title to large tracts of in South Africa has been adopted in the context of the land as collateral for nonperforming loans-to sell 168 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa land to land reform beneficiaries. assistance can be financed, on a voluntary basis, * Government's role is reduced to providing assur- through the community grant-is appropriate. ance that there are no problems with the land titles However, evaluation of the initial projects should and ensuring that the price negotiated between com- examine whether this excludes certain target munity groups and landlords is within acceptable groups or whether more stringent requirements boundaries. Projects are approved at the state level. may improve the quality of the projects estab- * Technical assistance is provided on a strictly lished, as well as beneficiaries' repayment capaci- demand-driven basis; beneficiaries can use part of ty and the overall benefits of land reform. the community grant made available under a * Since the working capital credit provided to bene- World Bank loan to contract private providers; ficiaries is subsidized (with a subsidy element of CONTAG (Federation of Rural Workers) partici- 70 percent), other sources of working capital, as pates in the state councils and assists with infor- well as marketing channels, need to be available to mation dissemination and land purchase negotia- ensure the longer-term economic viability of land tions. reform beneficiaries. This may well be one of the The only commonality between pilots and the most critical aspects of land reform. nationwide land reform process is the fact that benefi- Given the political importance of land reform and ciaries under the new process have access to a subsi- the limited knowledge of both the most appropriate dized loan under a special program (PROCERA) for mechanisms to implement this reform as well as the land reform beneficiaries. magnitude of the productivity and poverty impact, the Although the broad principles are similar to those government has established the Central Institute for in Colombia, the process is considerably more flexi- Agrarian Studies to (a) encourage discussion between ble and agile. The main points of difference are the all parts of civil society, academics, and politicians on following: land reform issues, (b) carry out a thorough and care- * Because grant financing is provided for comple- ful monitoring and evaluation of the land reform mentary and community-level infrastructure rather process either directly or through local institutions, than land itself, beneficiaries who expect to repay and (c) make the data collected in this process avail- their land purchase loan have an incentive to able to national and international researchers, thus reduce the price of land as much as possible. This acting as a hub in a broader network of countries and creates an incentive to focus on lands that are cur- researchers interested in negotiated land reform. rently underused, thus reducing the expected pur- chase price and directing land reform to areas South Africa where the social gains from the intervention are maximized. While South Africa shares with Brazil and Colombia * The process of beneficiary selection is less formal a highly unequal land distribution, policies that and bureaucratic than in Colombia, relying on an neglected the land rights of the nonwhite population infrastructure of existing community associations have systematically exacerbated these inequalities. to conduct information campaigns and ensure The Native Lands Act of 1912 prohibited blacks from broad representativeness. Community control establishing new farming operations, sharecropping, facilitates greater flexibility in project execution or renting land outside the reserves, which made up and should be particularly adequate if, through only 7.7 percent of the country's area. Inside the close supervision and ex post accountability, a reserves, an artificial form of "traditional" tenure with minority of politically vocal and well-represented maximum holding sizes and restrictions on land trans- farmers can be prevented from appropriating most actions was imposed. Subsequent policies of "black of the benefits. spot removal" transferred the large majority of black * As long as potential beneficiaries already have farmers who legitimately owned land outside the experience with the technology to be used on the reserves into the homelands where tenure restrictions, new farms, as tenants for instance, repayment high population density, and lack of capital and mar- incentives are strong.9 And as long as beneficiary ket access made commercial agriculture virtually organizations exist that can assist farmers in impossible. Labor laws that discriminated against preparing projects and in the initial stages of estab- blacks in favor of whites and generous capital subsi- lishment, the solution chosen-where technical dies contributed to successive evictions of large parts Klaus Deininger 169 of the black population from white farms, where they * Redistribution. As the main component of the gov- had been employed as labor tenants and farm workers. emient's lanid reform policy, redistribution aims Although the Native Lands Act was repealed in to complement the market by providing land for 1993, the momentous task of reversing these policies productive and residential purposes to a large and their consequences was left to the government number of rural blacks who were dispossessed that entered power following the 1994 elections. In during aparthteid and who are interested in obtain- attempting to do so, this government had to contend ing land. It aims to do so by providing a one-time not only with the extremely unequal land distribution settlement/land acquisition grant of R15,000 (the average amount held per person was 1.3 hectares ($3,300), equivalent to the national housing sub- by blacks compared with 1,570 hectares by whites) sidy available in urban areas to eligible beneficia- but also with the lack of any local government struc- ries, defined as anyone with a monthly salary ture, widespread absence of administrative capacity, a below R1,500. The choice of land redistribution highly indebted large-farm sector, and fear that redis- rather than expropriation (which, as in Colombia, tribution would wreak havoc on agricultural produc- can still be used as an instrument of last resort) tivity and jeopardize national food security. The gov- was based on the need to maintain public confi- ernment decided to adopt a land reform policy that dence in the land market and more generally to would redress the injustices of apartheid, foster affirm the government's respect for individual national reconciliation and stability, underpin eco- property rights. It also reflects the recognition that nomic growth, improve household welfare, and alle- expropriation. in other countries has failed to pro- viate poverty (Government of South Africa 1996). vide rapid access to land for a large number of peo- The three central components of this policy are as fol- ple, instead degenerating into lengthy political lows: maneuvering and rent-seeking. The number of * Restitution. Legal processes have been put in place potential land reform beneficiaries is considerable; to compensate (in cash or in kind) individuals who estimates indicate that there are about 200,000 were victims of forced removals after 1913. All labor tenants, 1 million farm workers, and as many restitution cases are dealt with through the Land as 7 million-8 million blacks in the reserves. Claims Court and Commission, established in Capacity building is particularly important 1994, to which claims have to be submitted within because the large majority of the rural population has three years (that is, by the end of 1997). Even if the never seen a successful and productive small farm, legal process can be completed in a speedy man- and many land reform beneficiaries themselves seem ner, the inability of the vast majority of the popu- to believe that efficient agricultural production is pos- lation to furnish written evidence makes this sible only on large farms. The South African govern- option feasible for only a small part of the popula- ment is well aware of these issues and has set up a tion. number of pilots in different provinces to accumulate * Land tenure reform. This component seeks to experience and improve the execution of land reform. improve tenure security of all South Africans by These projects are gathering momentum and have recognizing individual as well as communal own- provided a number of valuable lessons. ership rights to land, giving people the right to The almost complete fungibility of the land pur- make decisions about their own tenure system, chase grant represents an important advantage over adjudicating disputes, reforming tenancy laws, and the Brazilian and Colombian models. It prevents indi- attempting to end discrimination against women in viduals without comparative advantage in farming land allocation and holding. It is intended to create from becoming land reform beneficiaries just to the administrative infrastructure that will provide secure the government subsidy, while eliminating the hitherto disadvantaged groups with access to land possibility that land reform will inflate land prices. under a wide array of arrangements that are in line Although the government wishes to execute land with agroecological endowments and community reform with maximum local participation, the lack of characteristics. It is hoped that this will provide the an institutional structure has up to now made effective regulatory environment for a land rental market by decentralization difficult.20 This has, at times, resulted transferring land to more productive users and in confusion regarding the criteria and responsibility redressing the inefficiencies of the apartheid sys- for making decisions across alternative projects.2" tem. Although land reform in South Africa is the 170 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa responsibility of the Department of Land Affairs, entities involved in this process is through demand- provincial governments determine how and where driven and decentralized implementation; and (d) the complementary services and infrastructure are provid- long-run success of land reform is likely to depend on ed. This would be a difficult arrangement in the best getting the private sector involved in implementation of circumstances and threatens to fragment the land and the ability to use the land purchase grant to reform program in an environment in which adminis- "crowd in" private money. trative capacity is limited, few well-established proce- dures and mechanisms exist, and ethnic and political Making land markets more transparent and fluid. differences between provincial and central authorities Selection of land should be demand driven and effect- create a feeling of animosity and distrust.22 These ed through beneficiaries. However, information on shortcomings can be addressed by making the process land prices as well as beneficiaries' ability to assess more demand driven and incorporating current realistically the value of a piece of land or the poten- landowners more systematically in the selection of tial productive returns to its more intensive cultivation beneficiaries, elaboration of farm plans, and use of are often limited. Three mutually reinforcing strate- technical assistance. gies to deal with this constraint are (a) the provision The applicable regulations envisage that a land use of technical assistance at the community level, includ- plan should lay out the farm model to be adopted by ing assessment of the adequacy of the land price at the beneficiaries. Most of these plans seem to have been point of transaction; (b) the cofinancing of the land elaborated by authorities without the significant par- purchase through a private bank that, because it shares ticipation of beneficiaries, limiting their sense of in the risk of default, will have an incentive to assess ownership and of responsibility for the formulation the economic feasibility of the proposed farming proj- and outcomes of projects. Although the lack of bene- ect; (c) the creation of a "market information system" ficiary capacity makes this a difficult process, a better in the form of a public announcement of land prices understanding of the target group's specific needs and for plots transacted in the market, both with and with- aspirations, and of the productive opportunities in out use of a land purchase grant. It should be clear that given local areas, could have very beneficial effects. negotiated land reform is intended to be a comple- The experience from land reform in Latin America ment, rather than a substitute, for other forms of gain- illustrates that, in the absence of such a plan and asso- ing access to land. This understanding is fostered ciated productive projects, there is a danger that ben- when it becomes clear that under a regime of negoti- eficiaries will revert to survival farming and that land ated land reform, land rental is a means for beneficia- reform will adopt a "populist-welfarist" paradigm ries to accumulate experience and start-up capital, not that, by negating the impact of land reform on pro- something that augurs subsequent government expro- ductivity, will threaten the poverty-reducing potential priation. Moreover, whatever restrictions on land of the intervention (Cross and others 1996).23 rental persist should be eliminated. Up to now none of the models considered here has contemplated in depth Conclusion: Comparing Different Approaches the potential benefits of interregional migration and ways to encourage such migration to reduce the cost Although it is too early to judge whether negotiated of a land reform program and at the same time land reform can rise to the challenges that administra- enhance its impact on productivity. tive land reform has failed to solve, the unfolding experience in the three countries described can at least Focusing on productive projects. Productive proj- provide the basis for a first assessment. Comparison ects are likely to be the key of market-assisted land of the approaches taken in different countries suggests reform (a) because productivity is the only objective that (a) land reform through negotiation can only suc- criterion to put an upper bound on the price that can ceed if measures are taken to make the market for land be expected to be paid, (b) because it provides a nec- sales and rental more transparent and fluid; (b) pro- essary basis for financial intermediaries to evaluate ductive projects are a core element of market-assisted and eventually support such projects, and (c) because land reform, which is designed to establish economi- it requires beneficiaries to familiarize themselves with cally viable and productive projects at a socially jus- the realities they are likely to confront as independent tifiable cost rather than to transfer assets; (c) the only farmers, and in the process provides them with greater way to achieve effective coordination of the various clarity on their own aspirations as well as the poten- Klaus Deininger 171 tial-and the limitations-of land reform to con- ful private financiers but also to develop a better tribute to the attainment of these goals. understanding of the requirements and the alterna- tives available for individuals who for various rea- Decentralizing implementation. Experience with sons are not the most promising candidates for land centralized land reform has revealed it as costly and reform. If, by doing so, the process of negotiated land inefficient, which is incompatible with the very prin- reform contributes to changing beneficiaries' atti- ciple of decentralized negotiation. In Colombia, a tudes and manages to transform them from passive two-year attempt to execute negotiated land reform in objects who expect government to deliver a turnkey a centralized manner has convinced all but the most solution into subjects of the process who are able to hard-headed INCORA staff of the virtues of a decen- convert a one-time subsidy into a permanent tralized participatory approach. "Absence of capaci- improvement in their livelihood, it will have more ty" is unlikely to be an excuse for choosing a more than achieved its purpose. centralized approach: in most cases peasants were much faster to figure out how to work with the system Notes than was INCORA'S bureaucracy. Avoidance of such a costly leaming process in Brazil owes much to the 1. Balcazar (1990) summarizes the existing litera- presence of community organizations that were ready ture on three points, namely (a) small farms farm to apply to land reform the experience gained in more intensively than large farns, as measured by decentralized poverty alleviation projects. Even in value of output per unit of area, (b) between 1973-76 South Africa, where the absence of local institutions and 1988 average physical yields on small farms at the start of the program made a relatively central- increased about 82 percent (this seems to be an ized mode of implementation necessary, the rapid unweighted average across different types of crops) emergence of local governments has already facilitat- whereas those oni large farms remained stagnant, (c) ed significant steps toward greater decentralization. regional as well as commodity-specific studies do not find any systematic relationship between farm size Maximizing private sector involvement. Two and adoption of new technology or improvements in lessons have become evident about the financing of productivity. land reform. First, selective subsidization of a specif- 2. Although one of the main arguments for keeping ic part of the land reform package to the exclusion of INCORA was that its field presence would speed up others (as was originally the case in Colombia) is implementation of the program, this has not material- likely to be counterproductive, and a flat grant that ized. Three years after passage of the law that man- can be used for all types of expenditures (as in South dated market-assi-sted land reform, the level of execu- Africa) is clearly preferable. Second, ensuring access tion is far behind schedule and INCORA still has a staff to credit for working capital and other requirements above the level that many outside observers consider as a part of the land reform package seems to be necessary to execute effectively a truly decentralized absolutely essential for beneficiaries to develop and demand-driven market-assisted land reform pro- financially sustainable operations. Given the high gram. Restructuring (or complete closure) of INCORA transaction costs of providing credit in rural areas without a proven alternative would clearly have been and the greater need for monitoring when most difficult from both an operational and a political point clients lack previous exposure to credit, a unified of view. However, land reform would probably have source of financing that caters to beneficiaries' needs been implemented faster if the government had imme- for land purchase as well as working capital is most diately initiated decentralized pilots of the type desirable. Full support by the commercial private described in this chapter. Instead, the government sector, private and cooperative banks in particular, on waited for the reigulatory framework of the new law whose continued participation the success of market- (which ironically was entrusted to INCORA) to be assisted land reform is likely to depend most, relies almost completed, only to discover that considerable heavily on the formulation of economically viable redrafting was required to make the regulations com- and technically feasible productive projects.24 Getting patible with executing land reform through the mar- participants-government bureaucrats and NGOs as ket. well as potential beneficiaries-to face the realities 3. Modalities of financing are quite flexible-sell- of the market is essential, not only to convince doubt- ers are supposed to obtain 50 percent of the land's 172 Making Market-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa purchase price in cash (a 30 percent contribution by be enforceable and may even be counterproductive in the beneficiary and 20 percent from the government that it discourages successful reform beneficiaries grant) and the remainder in government bonds, which from intensifying production on part of their land and are traded at a considerable discount in the secondary renting out the other part, while preventing unsuc- market. A main boon for potential sellers is that any cessful beneficiaries from exiting. sales of land under the land reform are exempt from 11. Because changing the law itself was consid- capital gains tax. ered to be too cumbersome and lengthy, the changes 4. Following passage of the law in 1994, it was were introduced through specific regulations govern- envisaged that INCORA would spend about 50 percent ing implementation of the law. of available funds for market-assisted land reform and 12. These municipios are San Benito Abad in 50 percent according to the old model (direct inter- Sucre, Rivera in Huila, Fuente de Oro in Meta, vention) in 1995. Because the regulations to comple- Montelibano in C6rdoba, and Puerto Wilches in ment the law (for which INCORA had the main respon- Santander. sibility) proceeded very slowly, INCORA first attempted 13. Although one would expect that the transfer of to shift resources from the market to the administered large and relatively extensively cultivated farms component and, after this proved impossible, had to would provide an opportunity for accommodating disburse almost $20 million during a very short peri- additional workers, the opposite often happened in od at the end of the year to avoid losing the funds. practice, due to the unwillingness of existing workers Given the time pressure, it is not surprising that the to reduce their share and the associated incentives to processes used were not necessarily in line with the exaggerate the amount of land needed to establish a principles of market-assisted land reform. productive unit. This tendency to expel laborers is 5. To arrive at the maximum grant size per benefi- well known from the theory of cooperatives. ciary, the total land reform budget was divided by the 14. This is supported by anecdotal evidence-in average number of past beneficiaries. all of the places visited, potential sellers of land were 6. Although these practices were encouraged by the most ardent. There is a need to quantify the mag- the cumbersome mechanisms involved with obtaining nitudes involved-that is, both the degree of underuse credit to finance the 30 percent of land value not cov- and the amount of land that hypothetically would be ered by the government's purchase grant, this is hard- available for a productivity-enhancing land transfer. ly an excuse for them. 15. This perception is illustrated by the fact that 7. Even though all land reform transactions are to official records for the land transferred in 1996 indi- be conducted through the mechanism described, cate that in all cases where no credit from Caja INCORA maintains a budget for direct intervention in Agraria was needed (because the seller advanced the cases of social or other emergency. 30 percent complementary credit to the buyer), a farm 8. As a result, it is possible to use the grant to plan or productive project was "not required." finance coffee trees that are already planted but not 16. Indeed, examination of the farm models rec- that have yet to be established. ommended by INCORA for some of the land reform 9. The concentration of benefits may be expedient settlements indicated that-despite their considerable from a political point of view, especially if the bene- physical extent-these models failed to provide year- fits can be appro]priated by the about 10-15 percent of round full employment for the beneficiary and his or the peasantry that is politically well organized and her family. collaborates closely with INCORA. With respect to 17. Even model 1 in Montelibano-a model of farm size, even in the European Union, the average very low labor intensity-is considerably more labor farm size in 1990 (including the former German intensive than the average INCORA project in this Democratic Republic) was, with 14.8 hectares, slight- municipio. ly below the Colombian "minimum size" of 15 18. The steps involved in the process of expropri- hectares. While average farm sizes in Greece and Italy ation, which is applicable to land that is used to less were only 4 and 5.6 hectares, respectively, even in the than 80 percent, are as follows. First, there is a visit by Netherlands and Germany, average farm size was 16 an INCRA mission to assess the value of land and hectares (Eurostat 1995). improvements, followed by expropriation through 10. The categorical prohibition on rental of land presidential decree and confirmation of expropriation reform land (included in the 1994 law) is unlikely to through a federal court (emissdo de posse) in a Klaus Deininger 173 process that takes about a year. Once this is accom- mainly as a means to enhance their own power, which plished, landowners are compensated with titulos da depends on the number of people they are able to reforna agraria (with a real interest rate of 6 percent; accommodate rather than the productive success of bearing a discount of 25-40 percent in the market); the programs that are established. anecdotal evidence of excessive compensation 23. "NGOS and pilot structures are ... beginning to abounds. The necessary infrastructure investment is be afraid that they would not be able to do more than then included in the subsequent year's INCRA budget provide land redistribution beneficiaries with the min- (it takes one to two years for the infrastructure to be imum of land anid secure tenure and that this alone established), and INCRA announces that the land is would amount to just dumping them-equivalent to available, selecting beneficiaries based on agricultur- the apartheid practice of dumping removals victims in al skills. In practice, all cases are limited to regular- the middle of the veld with no shelter and no way to ization of existing squatter settlements. Once selected, make a living" (Cross and others 1996: 166). beneficiaries are eligible for credit from PROCERA 24. There seems to be broad consensus in the liter- (maximum, $7,500; average, $4,500), a special pro- ature on land issues that the full benefits from gram for land reform beneficiaries that is adminis- landownership materialize only if landowners have tered through INCRA (with an effective subsidy of access to financial markets (Carter and Zegarra 1995). about 70 percent). 19. Because it is likely that this will require some References monitoring (and thus increase costs), repayment will depend on the structure of the bank's incentives, in The word "processed" describes informally repro- particular who will bear the cost of a possible default duced works that may not be commonly available on the loan and the nature of the compensation the through libraries. bank receives. Balcazar, Alvarc. 1990. "Tanao de finca, dinamica 20. Considerable progress along these lines has tecnologia y rendimnientos agricolas." Coyuntura been made, for example by contracting out monitor- Agropecuaria 7(3):107-25. ing and evaluation of the land reform pilots to local Bardhan, Pranab, Samuel Bowles, and Herbert Gintis. (traditionally black) universities, a step that, it is Forthcoming. "Wealth Inequality, Wealth hoped, will not only increase the quality of evaluation Constraints, and Economic Performance." In but also enhance the capacity of these institutions, Anthony Atkinson and Francois Bourguignon, which has traditionally been quite weak. eds., Handbook on Income Distribution. 21. The green paper states that (a) most critical and Amsterdam: Elsevier North-Holland. desperate needs will be attended to first, (b) priority Binswanger, Hans P., Klaus Deininger, and Gershon will be given to projects where institutional capacity Feder. 1995. "Power, Distortions, Revolt, and is sufficient to allow quick and effective implementa- Reform in Agricultural Land Relations." In J. tion, (c) sustainability and viability of the project need Behrman and T. N. Srinivasan, eds., Handbook of to be ensured. and (d) government will spread its Development Economics. Vol. 3. Amsterdam: efforts geographically and across different types of Elsevier Science B.V. beneficiaries. Binswanger, Hans P., and M. Elgin. 1988. "What Are 22. Based on a case study of a provincial land the Prospects for Land Reform?" In A. Maunder reform pilot in Kwazulu Natal, Cross and others and A. Valdes, eds., Agriculture and Governments (1996) report four main shortcomnings, namely (a) a in an Interdewendent World. Proceedings of the complete lack of interinstitutional coordination (con- twentieth International Conference of Agricultural ceming access to water resources), (b) underestima- Economists, Buenos Aires. Aldershot: Dartmouth. tion of the time and energy required for legal issues Birdsall, Nancy, and Juan Luis Londonio. 1997. such as adjudication between conflicting land claims, "Asset Inequality Matters: An Assessment of the (c) political tensions between the African National World Bank's Approach to Poverty Reduction." Congress and Inkafha Freedom Party (at the national American Economic Review 87(2):32-37. and provincial level, respectively) which prevent the Carter, Michael R., and Dina Mesbah. 1993. "Can effective delivery of services other than land to the Land Market Reform Mitigate the Exclusionary beneficiaries, and (d) a worrying tendency of local Aspects of Rapid Agro-Export Growth?" World chiefs to interfere because they perceive land reform Development 21(7):1085-100. 174 Making MWarket-Assisted Land Reform Work: Initial Experience from Colombia, Brazil, and South Africa Carter, Michael R., and Eduardo Zegarra. 1995. der EU = SPEL-Data for EU Agriculture = "Reshaping Class Competitiveness and the Dounees SPEL pour l'Agriculture de l'EU. Trajectory of Agrarian Growth with Well- Luxembourg: Office des Publications Officielles Sequenced Policy Reform." Agricultural des Communautes Europeenne. Economics Staff Paper Series 379. University of Heath, John, and Hans P. Binswanger. 1996. "Natural Wisconsin-Madison. Processed. Resource Degradation Effects of Poverty and Cross, Catherine, T. Mngadi, S. Sibanda, and V. Jama. Population Growth Are Largely Policy-Induced." 1996. "Making a Living under Land Reform: Environment and Development Economics Weighing up the Chance in KwaZulu-Natal." In 1(1):65-83. M. Lipton, F. Ellis, and M. Lipton, eds., Land, Misi6n Social. 1990. El desarrollo agropecuario en Labour, and Livelihoods in Rural South Africa. Colombia. Vol. 1. Bogota: Ministerio de Agricultura, Vol. 2: KwaZulu-Natal and Northern Province, Departamento Nacional de Planeaci6n. chap. 6. Natal: Indicator Press, University of South Africa, Government of. 1996. "Our Land. Natal. Green Paper on South African Land Policy." Head Deininger, Klaus, and Lyn Squire. Forthcoming. Office, Department of Land Affairs, Pretoria. "New Ways of Looking at Old Issues: Inequality Processed. and Growth."' Journal of Development Economics. World Bank. 1996. Review of Colombia's Agricultural Eurostat. 1995. SPEL-Daten fur die Landwirtschaft and Rural Development Strategy. Washington, D.C. 14 Decentralization and Biodi-versity Conservation Julian Caldecott and Ernst Lutz This chapter adopts a particular perspective within a and Michael Wells, and Zimbabwe by Brian Child much broader study of decentralization (Binswanger, (see Lutz and Caldecott 1996). In addition, Ajit Shah, and Parker 1994), which is itself only one Banerjee and Emnst Lutz (1996) reviewed 32 World aspect of the greater issue of good governance. Here, Bank and Global Environment Facility projects with we focus on one aspect of decentralization: that is, conservation goals. They focused on project-related whether or not it promotes the conservation of bio- variables and their contribution to the effectiveness of logical diversity and, if so, under what conditions? We habitat conservation. know that conservation is part of sustainable develop- Part of the background to all this work is the ment and good governance (WCED 1987, IUCN, UNEP, knowledge that many past attempts to conserve bio- and wwF 1991, UNCED 1992, WRI, IUCN, and UNEP diversity failed. ODne reason was an overly centralized 1992, and UNEP 1995), that participation by local approach, often involving top-down planning by tech- communities is needed for effective conservation, and nicians and bureaucrats without concern for the opin- that this may have to be supported by a significant ions or well-being of the people affected by their deci- degree of decentralization (Bhatnagar and Williams sions. Examples are common and include efforts of 1992, Wells, Brandon, and Hannah 1992, and World the former Soviet Union, whose system of parks and Bank 1994). But questions remain concerning the reserves helped conserve much of the country's bio- kind and scale of decentralization, its linkage with diversity, but whose centralized decisionmaking was participation, and the role of other elements such as seriously distorted by inefficiency and corruption. For incentives, enforcement, and institutional capacity in the system to function, government needed to have achieving a socially optimal degree of conservation. enough power to be able to ignore local aspirations We now need to discover how exactly and by what and local conditions, and when that power failed, mechanisms these various factors are linked. many nature reserves were immediately exploited by To seek answers to these questions, case studies local groups thal could conceive of no reason not to were undertaken in 10 countries: Colombia by do so. In many other countries too, the viability of Eduardo Uribe, Costa Rica by Julian Caldecott and centrally plannecl "fortress reserves" has been under- Annie Lovejoy, India by Shekhar Singh, Indonesia by mined by their cost and also by the democratic deficit Julian Caldecott, Kenya by Joyce Poole and Richard built into them (Bonner 1993). Complete decentral- Leakey, Nepal by Uday Sharma and Michael Wells, ization, however, can also be counterproductive to Nigeria by Julian Caldecott and Andrew Babatunde conservation. Urnder those conditions, local people, Morakinyo, the Philippines by Maria Dulce Cacha perhaps in cooperation with outside entrepreneurs, and Julian Caldecott, Russia by Margaret Williams may simply degrade and deplete the resources faster 175 176 Decentralization and Biodiversity Conservation and more efficiently. So there clearly remains a con- needed to advance those interests. Finally, there is the tinuing role for central governments. The challenge is process of communication and negotiation by which to find the degree of decentralization that is most the various groups agree to the terms of transactions appropriate. among them. These transactions represent the These issues will perhaps become clearer if con- exchange of rights and obligations or the payment of servation is taken to mean improving the allocation compensation for impacts caused by one group ful- and use of environmental resources, including the filling its own set of interests on the opportunity for various components of biodiversity (distinct popula- others to fulfill theirs. tions, species, ecosystems, and so forth) and the range It seems reasonable to facilitate these processes, of ecological services and goods they provide. These thus reducing the average cost of transactions for the resources are consumed at various levels of society, benefit of society as a whole. One way to do so is to which can be classified broadly as local (comprising assist in the peaceful redistribution of power, while the community and bioregional levels), national, and also establishing mechanisms for bargaining based on global. Many different costs and benefits are associat- the new power structure. In overly centralized soci- ed with managing resources in a particular way at eties, this may mean promoting decentralization as each level of society, and managing them to maximize one component of a broader reforn process. This benefits at one level may impose costs on other levels. could involve, for example, giving local groups the Thus what is internal to one level may be external to authority to manage environments and to tax and another. spend in order to do so. But not all functions of gov- To illustrate the effect of this, assume that an ermnent can usefully be decentralized (Prud'homme action at the local level creates 10 units of benefit at 1995), and nonlocal groups may be in a better position that level, while causing negative impacts (or exter- than local ones to appreciate long-term or large-scale nalities) worth eight units at the national level and two issues and to act as disinterested arbiters of local dis- at the global level. These impacts can take many putes that cannot be solved locally. This nonlocal per- forms, for example the loss of species, watersheds, spective is vital in conservation, which fundamentally fisheries, or carbon stores. One way for nonlocal soci- concerns avoiding or managing interspecies, inter- ety to avoid such impacts is to prevent local people generational, interregional, and international conflicts from acting in their own interests, such as by impos- of interest. Empowerment of local groups should ing a fortress reserve on them. But a fairer alternative therefore be balanced by a continuing role for central would be to pay the local level 10 or more units of government to deal with market failures and to ensure resources to refrain from undertaking the action con- both social equity and environmental protection. cerned. This transfer payment may take the form of a Local activities occur in the context of national rural development project offering alternative benefits policies, laws, and institutions, but often out of sight and is the rationale behind many of the integrated of the agents of government (Caldecott and Fameso conservation-development projects that have been 1991). They can directly affect the components of undertaken in recent years (McNeely 1988, biodiversity, in which the national and global levels of Ruitenbeek 1990, 1992, Brandon and Wells 1992, society are now increasingly interested. These levels Wells 1992, and Wells and Brandon 1993), albeit with wish to protect viable and representative natural mixed success. ecosystems, particularly terrestrial ones in the humid Thus each level of society has its own interests in tropics, where most of the world's estimated 10 mil- how biodiversity resources are used, and payments lion-100 million species exist (UNEP 1995). Tropical from one to another can in principle be used to com- species are especially vulnerable to major and wide- pensate losing parties so that all sides are satisfied. spread habitat change, such as that brought about by Decentralization can be seen as a process by which clear-felling, fire, or agriculture. Thus the single best property rights and bargaining powers are redistrib- way to maintain richness of global species overall is uted among the levels of society, and there are at least to create and manage an adequate system of protected three distinct aspects to this. First, there is the educa- areas in the tropics. Other approaches complement tional (or self-discovery) process by which each this aim or can serve others, including the use of off- group identifies itself, its priorities, and hence its site techniques such as captive breeding and the con- interests.' Second, there is the empowerment process servation of wild populations under logging, hunting, by which each group obtains the bargaining power fishing, and other kinds of harvesting to the extent this Ji lian Caldecott and Ernst Lutz 177 can be done sustainably. Once conservation is adopted as a policy priority, Given the externalities involved, certain resource a special challenge is created for good governance, management options may exist for which the global which is uniquely concerned with managing actual or social net benefit is maximized and where, with potential conflicts between the needs of people and appropriate compensation, all actors at various levels those plants and animals, between the needs of differ- can be equally or better off. A complication emerges, ent generations of people, and between the needs of however, if people at the local level are unaware of widely separated. populations of people. This chapter the value of wild species, for example, or of the con- aims to explore some of the implications of entering sequences of changing land use in water catchments, this new territory. In doing so, it draws insights from either for themselves or for society as a whole. A field experience that can help to make sense of the rational response to this at the national and global lev- complex interaction among all the factors involved, els, in addition to or instead of compensation or other thereby helping to clarify the relationship between incentives, would be to invest in education at the local decentralization and biodiversity conservation. level. This should reduce the cost of any transfer pay- ments that may still be required, while taking advan- Processes of Decentralization tage of the fact that education, once paid for, contin- ues to yield benefits without further cost (Tilak 1989, The country stu(dies reveal certain common features Haddad and others 1990, and Herz, Subbarao, and of decentralization processes. Until the mid-1970s in Raney 1991). The high long-term cost-effectiveness Kenya, Nepal, and Zimbabwe or the late 1980s and of education helps to explain why education activities early 1990s in C(osta Rica and Colombia, countries are now so often linked to conservation projects (for had a centralizedl but fragmented approach to envi- example, GEF 1995). ronmental management and rural development (table Table 14.1. Changes in National Conservation Institutions in Select Countries Country Fragmented and centralized Integrated and centralized Mediating body Colombia 1968-93: National Institute of 1993: Ministry of the 1993: reformed Natural Renewable Resources Environment autonomous regional corporations Costa Rica Pre-1989: National Parks Service, 1989-95: National Parks 1987: National Parks Forestry Directorate, Wildlife Service, National System Foundation Service of Conservation Areas 1989: National (SINAC) Biodiversity Institute (INBio) Kenya Pre-1976: National Parks Board, 1976-90: Wildlife 1990: Kenya Wildlife Game Department, Forestry Conservation and Service (parastatal) Department Management Department Nepal Pre-1973: royal hunting preserves 1973: Department of National 1986: King Mahendra Parks and Wildlife Trust for Nature Conservation Conservation Philippines Pre-1992: Protected Areas and 1992: National Integrated 1992: Foundation for Wildlife Bureau, Tourism Authority, Protected Areas System the Philippine Environment National Power Corporation, (NIPAS), Department of Department of Agriculture Environment and Natural Resources Zimbabwe 1960-75: Department of National 1975: Department of National 1989: CAMPFIRE Association Parks and Wildlife Management Parks and Wildlife Management and intensive conservation areas 178 Decentralization and Biodiversity Conservation 14.1). Control over resources was divided horizontal- A Role for Mediating Organizations ly and vertically among line ministries and other agencies, which competed among themselves for These experiences confirm that decentralization is power and funds. It is unclear what prompted changes inevitably and fundamentally a political process to this system in each country, but we know that because it redistributes power. Because of this, an attempts were made to rationalize and consolidate important factor in decentralizing a country is the pres- authority over protected area systems, creating more ence of a body able to mediate among various interest integrated but still centralized arrangements. groups and to promote smooth and effective transfers As these new institutions were created, it became of power. These are often officially sponsored NGOS, possible to reform them by amending the laws under such as the CAMPFIRE Association in Zimbabwe, the which they were established. Thus Nepal's National National Biodiversity Institute (INBio) in Costa Rica, Parks and Wildlife Conservation Act of 1973 was the King Mahendra Trust for Nature Conservation in amended in 1989 to authorize multiple-use conserva- Nepal, and the autonomous regional corporations in tion areas and the involvement of nongovernmental Colombia. All were created by government and have organizations (NGos) and again in 1993 to authorize close formal and especially informal links with gov- buffer zones and local revenue-sharing arrangements. emient, but they also have a clear, independent role Meanwhile, Zimbabwe's Parks and Wild Life Act of and much operational and financial autonomy in ful- 1975 was amended in 1982 to empower the district filling it. Other attempts at making such arrangements councils over wildlife. In some cases, reform was are represented by the parastatal Kenya Wildlife driven by events in the field, rather than by the nation- Service and the national parks boards in Nigeria, but al legislature. Thus in the early 1990s, the Guanacaste these have tended to be vulnerable to political interfer- conservation area in Costa Rica and the first CAMP- ence. FIRE districts in Zimbabwe went beyond what was In the absence of an effective mediating body, other then current law to develop local arrangements for NGOS may adopt a less formal but analogous role. This conservation.2 In these cases, leadership by inspired may involve a local NGO developing a partnership with individuals encouraged local groups to take risks that official agencies in its area of operations. An example were rewarded later when laws were changed to is that in Nizhny Novgorod, Russia, between Dront authorize what had already been implemented on the Eco-Center (the NGO), the regional unit of the Ministry ground. But in other cases, local environmental of Environmental Protection, and the managers of the activism provoked reactionary efforts to restore cen- federally funded Kerzhenski reserve. Here, the three tral control. An example is Ogoniland in the Niger groups collaborate to devise and implement regional delta region of Nigeria, where local action against conservation initiatives. International conservation environmental damage caused by the oil industry was NGOS such as the World Wide Fund for Nature also harshly suppressed by the Nigerian military. often position themselves as intermediaries between In Indonesia, those field projects that have been official agencies and the public in the countries where effective in influencing central government seem to they work. Some specialize in integrating themselves have done so by cautiously going beyond what was as closely as possible with governments and can be the previous official interpretation of certain laws. effective in providing technical advice and policy Examples include projects at Arfak in Irian Jaya and guidance to them. Conservation NGOS with close links Kayan-Mentarang in East Kalimantan, where com- to government are evident, for example, in Indonesia munity mapping of resource boundaries and local par- and Nigeria, but before 1993 were kept more remote ticipation in resource use helped to create subtle but from government in Colombia and still remain so in important changes in official attitudes to local Kenya. Establishing a new and effective mediating empowerment. In Costa Rica and Zimbabwe, reforms body in a country means that existing NGOS have to that gave local people far greater control over and reposition themselves accordingly, and adjustments of opportunity to profit from the management of wild this kind (which are seldom smooth) have been seen in species and natural ecosystems prompted political Costa Rica, Zimbabwe, and elsewhere. reaction. Aggressive questioning of CAMPFIRE began in Zimbabwe in about 1992, and SINAC reforms were Role of Local and International NGOs suspended by a conservative regime in Costa Rica from 1990 to 1994.3 Local and international NGOs have several important Julian Caldecott and Ernst Lutz 179 roles to play in conservation and rural development assets on which nature tourism is based or to creating projects (Wells, Brandon, and Hannah 1992 and GEF incentives to encourage conservation among the local 1993). They tend to be the seed bed of new concepts population. that are later adapted and used by official donors and Ecosystems within nature reserves contain governments. This role is mainly due to their ability to resources whose use can contribute to local revenues attract innovative individuals, to promote their ideas under certain circumstances. Local revenues vary through advocacy, and to try out their concepts according to the nature of the resource concerned and through small and highly visible pilot projects. other factors such as the presence of markets. Selling Because their approach tends to be flexible, consulta- trophies to hunters (as in Zimbabwe), or biodiversity tive, and sensitive to local needs, NGOS can also act as prospecting permits to pharmaceutical companies (as buffers between large projects and small communi- in Costa Rica), or lodging and food to tourists (as in ties, helping to scale down project activities to a more Nepal or Indonesia) all require technologies and atti- appropriate level. Local NGOs are therefore often used tudes that may not have existed previously in each as contractors to deliver community services on country. If arrangements allow revenues to be cap- behalf of large projects. Finally, NGOS can reduce the tured by local people, then this will tend to promote shortfall between the local empowerment planned by decentralization and, if other conditions are favorable, a project and that actually achieved, as evidenced by conservation. Bat the opposite effects are seen if most Banerjee and Lutz (1996). This is because they are benefits are obtained by nonlocal groups, such as cor- sensitive to the betrayal of ideals, will campaign in porations exploiting timber, fish, minerals, or tourism public if they see this happening, and therefore have without the consent of local groups. There are many an important role to play in keeping elite institutions such cases in the country studies, including hotels in honest. the Royal Chitwan national park in Nepal and the Amboseli national park in Kenya and plantation Fiscal Decentralization developments and logging concessions in and around the Okomu reserve and Cross River national park in The most potent form of decentralization is the trans- Nigeria. fer of authority over spending decisions, and elite Nature reserves have potential value, which can be groups retain these particular powers far more dili- turned into revenue if the technology and markets to gently than those over other administrative functions. do so exist. But if those revenues are large, they may Thus the most complete cases of decentralization are attract the interest of powerful groups or individuals. those where local people collect revenues and decide Thus as wildlife resources in Zimbabwe came to gen- locally how they will be spent. This was the key to the erate local revenues for the CAMPFIRE districts, pres- strength of the CAMPFIRE districts in Zimbabwe and sures grew to use them in ways that relieved central the Guanacaste conservation area in Costa Rica, and government of its spending responsibilities. This had the same kind of autonomy is seen to some extent in an effect like one of expropriating local funds, and the joint forest management areas in India, buffer such a response can be expected when central gov- zone areas in Nepal, autonomous regions in ernments with limited resources see some groups of Colombia, newly empowered municipalities in the citizens becoming wealthier than others. Unless the Philippines, and certain project areas in Indonesia. central governrent has a special reason not to inter- Arrangements for making decisions on how to vene, or the newly enriched groups have enough polit- spend local funds vary, and some are more transparent ical influence to defend their wealth, at least part of it than others. Zimbabwe's CAMPFIRE communities give will be taken away from them. Thus at some stage, a an example of full participation in a process of debate political response will be needed to ensure that and decision, with allocations to public works and pri- enough benefits are retained by local people to sustain vate pockets being made openly and in cash. At the the decentralized decisionmaking unit. other extreme is the Masai Mara reserve in Kenya, which is under the control of the district council but is Some Reasons for Decentralization affected by a serious lack of accountability and trans- parency in decisionmaking. There, the large sums of The fact that many countries are tending to decentral- money collected from entrance fees contribute little ize might be explained in two ways. First, rural peo- either to maintaining the wildlife and biodiversity ple may be becoming better organized, better educat- 180 Decentralization and Biodiversity Conservation ed, and more aware of economic and ecological real- Incentives and Enforcement ities and hence more assertive in protecting their interests. Alternatively, the rapid degradation of Willingness of local people to respect boundaries and ecosystems in most countries may simply be making rules protecting nature reserves should be strength- those ecosystems less valuable and less interesting as ened by obtaining their prior agreement to them. The objects of control by powerful groups and individu- incentive for government to seek this agreement lies als. Virgin rain forests and unfished seas can be in the need to spend fewer resources on policing exploited profitably by corporate investors, but where there is strong local support for conservation.4 hacked forests and exhausted fisheries are not worth As shown in the Arfak mountains in Irian Jaya, such investment and may be passed by the elite to Indonesia, for example, setting reserve boundaries by local control until they recover or until new tech- local consultation can help people to accept them as nologies or markets change their value. Both process- permanent and inviolable and are therefore much es are probably at work and will interact in a complex cheaper to maintain than boundaries imposed by cen- and patchy way depending on many factors in each tral authority alone. Similarly, from Zimbabwe we country. The balance emerging in each place and time know that local participation in wildlife management reflects the level of government commitment to decisions and associated revenues can reverse long- decentralize, but this is not enough on its own, and standing hostility to wildlife (see also Pye-Smith and appropriate laws, policies, human capacity and insti- Feyerabend 1994). tution building, and time are all needed as well. An apparent paradox here is that local threats to These other requirements flow from the many ways nature reserves can also arise from private motives to in which local autonomy can be threatened and the exploit resources. Examples can be drawn from the many kinds of responses needed to overcome those country study on Russia, which documents the effects threats. of a sudden collapse of central power after a long period of highly centralized govemance. Violations of Links to Conservation Effectiveness reserves by or with the connivance of local people and local authorities are now common and include poach- Some threats to communities and their environments ing endangered species (Siberian tigers in Lazovsky can be solved locally, an example being in the and Sikhote Alinski), constructing roads (in Masoka area of Zimbabwe where secure land tenure Samarskaya Luka) and buildings (in Pri-Oksky allowed local people to evict nonlocal squatters. But Terrasny), grazing livestock (in Daurski and Altaiski), well-armed poaching gangs provide a challenge that fishing (in Magadanski and Kostomuksha), and clear- local people alone may be unable to meet, as hap- cut logging in several reserves in European Russia. In pened in several of Kenya's national parks and these cases, local groups have taken power without reserves. This also applies to planning failures and being able to limit resource exploitation, either cases where impacts on the environment were not because local people are not aware of the value of anticipated, such as impacts caused by road projects nature reserves to themselves or because they have no in Colombia's Salamanca national park and by water tradition of holding those in power accountable, or diversions in Kenya's Amboseli and Tsavo West both. Similar problems have arisen elsewhere, where national parks. A condition of local autonomy is local elites have received new opportunities to profit therefore the protection of local environments by the arrival of mining or logging companies. In through proper central planning and properly these cases, much damage can be done before local enforced laws on assessing and managing environ- awareness of environmental consequences gives birth mental impacts. This is one reason why conservation to local opposition to the elite's decisions. can be seen as a joint venture between local and non- To offset such risks requires appropriate incentives local levels of society. It suggests that as local actors and policing to complement community-level motiva- become stronger, there is a need to strengthen the tions and action. Enforcement may be welcomed by willingness and capacity of nonlocal institutions to local people if it is carried out justly and impartially, support decentralization. This highlights the role of if they have a role in defining its scope and approach, central government in promoting actions, such as or if it clearly benefits them directly or indirectly. An conserving biodiversity, that local people may not be example of the latter can be found in northeastern able or willing to perform unaided. Kenya, where antipoaching work by the Kenya Jiulian Caldecott and Ernst Lutz 181 Wildlife Service is welcomed locally because it tends vide block grants and other forms of assistance such to protect local people against armed Somali intrud- as help in mornitoring and applying environmental ers. For government, a key issue guiding investment impact assessment and spatial planning laws. in policing is the likely cost-effectiveness of different kinds of effort, and two factors contribute strongly to Scientific Expertise this (Caldecott 1996). First, there is its style, which is influenced by factors such as vigilance, consistency, A final issue is whether policies and decisions are sci- public relations effort, staff quality, and legal process. entifically sound. Conservation biologists, for exam- Second, there is the level of local compliance, which ple, are trained to make judgments on the minimum is influenced strongly by the style of policing, the size of viable reserves, the location of their manage- extent of local involvement in setting rules and aware- ment zones, ancd the likely consequences for wildlife ness of why they are needed, and the economic incen- of the hunting laws applied in their buffer zone. tive structure applied to local people's use of nature Unless they themselves possess such skills, local reserves and wild species. These factors all interact groups will neecl support from specialists if they are to with one another, and proper policing has a positive make such judgments or if they are to set up the mon- effect on compliance, while improper policing is inef- itoring systems for ecosystems and populations with- fective or worse. out which no management system that is more than Proper policing requires that those who undertake moderately intensive can succeed for long. it are seen by local people to act on behalf of an Conservation requires paying just as much attention authority that has a legitimate claim on their loyalty. to ecology as to the social, managerial, and financial This can occur at the community level of society, and aspects of proper policing and to the economic and such social pressure is the main means by which psychological factors that operate through incentive endogenous cultural rules are enforced in any society. structures to regulate the use of resources and thereby Examples include hunting taboos and sacred forests, increase the likelihood of conservation success. both of which are common among tribal peoples (Caldecott 1988 and Kiss 1990). Exogenous rules, Concluding Ob:servations however, such as those protecting biodiversity, may have no equivalent basis in people's upbringing and The diverse experiences reported in the country stud- may be hard to enforce at the community level ies and project analyses show that decentralization because of kinship ties and friendships among the and conservation are complex processes that interact people responsible. Finally, most rural communities with one another in many ways. From the historical lack an adequate tax base with which to finance more reviews, we can conclude that centralized, top-down than the most rudimentary volunteer protection conservation has seldom been effective, except where efforts, which will only be effective in exceptional large budgets are available for enforcement and the cases. society concerned is willing to accept a rather undem- Conservation enforcement services have the best ocratic conservation process. Looking at the more morale and are best accepted locally when they work recent experience of countries in giving new respon- on behalf of a level of society beyond that of the com- sibilities to local government units and NGOS, we can munity but not as remote as that of the nation, which see that this creates both opportunities and potential we call the bioregional level. This is the case in the problems. To take advantage of the former while Guanacaste conservation area in Costa Rica, where avoiding the latter, a cluster of arrangements must be wardens work on behalf of a regional committee of made as a whol'e if conservation is to work well in a local stakeholders. The main drawback here is likely decentralized setting. Of these, seven merit special to be one of resources, which are limited in rmost rural attention: bioregions.5 Without adequate funds held locally, and * Local participation, especially in a way that allows considering that threats can occur through external local people to understand and endorse the bound- planning failures, the national level of society will aries and management plans of nature reserves and need to accept considerable responsibility for conser- that promotes clear tenure over land and other vation policing. The latter is therefore often best done resources in and around the reserves on behalf of local authorities with the assistance and * Capacity building, especially to increase skills and supervision of the national level, which should pro- accountabiliiky among local government units and 182 Decentralization and Biodiversity Conservation NGOS, so they can work together to promote con- those decisions. servation and rural development Weak environmental awareness, caution, self- * Incentive structures, especially those that allow regulation, tenure, and accountability are thus typical local communities to keep income from the sus- of conservation problems. They can usually be tainable use of nature reserves and other biodiver- found-alone or in combination, obviously or dis- sity assets guised-wherever the components of biodiversity are * Conditional subsidies, especially where divergent being eroded by human action. Decentralization can costs and benefits of conservation are experienced directly help to solve problems of self-regulation, by local and nonlocal groups, making it necessary tenure, and accountability but has a more indirect role for global and national society to bridge the gap in improving environmental awareness and caution. with livelihood investments or grants This awareness depends on public understanding of * Appropriate enforcement, especially against pow- how ecosystems work, and decentralization can help erful local or central interests, and always in the to increase local responsibility for environmental context of education and public relations activities management, thus making this process more relevant * Stakeholder forums and ecoregional executives, and interesting to local people. Greater interest then which need decisionmaking and fiscal authority to tends to promote willingness to learn about ecology. fulfill their three main roles of avoiding conflict Self-regulation, tenure, and accountability are through dialogue, authorizing conservation action, strongly interconnected. Some reforms that a country and requesting help from nonlocal society to meet can make will help to ensure that people who decide local development priorities how to use environments are directly affected by the - Enabling policies, laws, and institutions to provide consequences of those decisions. By shortening the a clear and supportive framework for conservation feedback loops between decisions and their effects, on behalf of national government, creating incen- such reforms reward a cautious approach to decision- tives at the local level to harmonize development making. Moreover, changes that give authority specif- with conservation and thereby reducing the need ically to people living within the managed environ- for enforcement. ment encourage decisions that are responsive to local The country studies also illustrate many of the key conditions. If, in addition, other local stakeholders are issues in sustainable development, even though they encouraged and enabled to question those decisions, report only a small sample of recent global experi- then responsibility is promoted and a strong force for ence. A review of this material and other evidence in good governance created. this volume and elsewhere (Barzetti 1993, IIED and All such changes tend to improve environmental ODA 1994, Western, Wright, and Strum 1994, White management and are often consistent with key ele- and others 1994, UNEP 1995, and Caldecott 1996) ments in the process of conservation (see Janzen leads us to conclude that conservation problems can 1991, 1992 and WRI, IUCN, and UNEP 1992). The first usually be traced to one or more of the following of these is that the components of biodiversity must underlying causes. First, people who depend on an be saved in order to preserve the option to use them. ecosystem may not be aware of the connection The second is that people must study and learn about between the ecosystem's well-being and their own. what those components are and how they might be Second, they may be incautious in changing the used. The third is that they must be used sustainably ecosystem to make it more productive, without realiz- in order to meet the economic, intellectual, and other ing that this can do more harm than good to the inter- needs of society. The final element is that they and ests of the majority. Third, the presence of external their uses must be taught about, so they will be valued effects may be connected with local actions. Fourth, by people. From this point of view, then, decentral- local people may have no accepted rules governing ization can be important in allowing biodiversity to be the use of the ecosystem, often because a traditional perceived as a local resource for local husbandry, management system has broken down or has been which in turn can motivate local people to preserve replaced by a central and ineffective one. Fifth, they the biodiversity in their own environment. may not have the authority or ability to manage the Reforms of this kind yield benefits that are not ecosystem exclusively in their own interests. Finally, limited to local people, because they can also reduce key decisions changing the use of the ecosystem may the cost to a government of meeting its own conser- be made without involving all the people affected by vation aims (Caldecott, Bashir, and Mohamed 1995). Julian Caldecott and Ernst Lutz 183 They do so partly by encouraging local understanding that is no longer sheltered by a national govern- of how to make better use of ecosystems (thus reduc- ment may become vulnerable to groups wishing to ing the need for inducements to do so) and by prompt- exploit it, for example as a source of raw materials ing people both to insist on and assist in proper envi- or as a dumping ground for wastes. Where nation- ronmental policing (thus reducing its recurrent cost). al governments are no longer able to control such They also promote dialogue and trust and may reduce threats, localities must be helped to communicate the cost of negotiating settlement of claims arising and collaborate to prevent them from being singled from impacts within and between local and other lev- out and overwhelmed one by one. els of society. If combined with strengthened resource * Finally, uncertainties abound in any political tenure, participation can help local people to capture process, including decentralization. Thus there is some of the economic benefits of conservation, fur- always a risk that the need to protect nature ther rewarding local policing effort at little cost to reserves may be forgotten for a time. government. Finally, better environmental manage- The last observation draws attention to the fact that ment helps governments to avoid the cost of rehabili- the process of decentralization in many countries is tating societies that have been blighted by environ- not driven by public interest in biodiversity conserva- mental damage. By promoting local participation, tion, but rather by a desire for better access to the decentralization can have an important role in all this, fruits of economic development through democratic but it is not a panacea and must always be seen, participation. Conservation will benefit from this only judged, and planned in context. This leads us to five to the extent thaot ecosystems and the biodiversity they main observations: contain are seen as resources to sustain development, * Precipitate and unplanned decentralization can in other words as valuable resources that people (or neutralize national and global influence, while giv- different levels of society) may wish to control for ing powers to local societies that may lack ade- their own particular benefit. If there is no such per- quate skills and accountability to use those powers ception, then conservation benefits will accrue from properly. This means that the social context should decentralizatiorn only accidentally and, if biodiversity be analyzed carefully before changes are attempt- continues to be perceived as valueless by newly ed. empowered local groups, only temporarily. * Redistributing power may affect those who were Because co:nservation requires pernanent solu- formerly doing well, prompting them to resist tions to problems of species extinction and environ- change. Thus mediating bodies will be needed to mental degradation, it must involve changing percep- smooth the transfer of power, and these must be tions and values among people who control the fate of trusted enough by all sides to be able to reassure ecosystems. In decentralized circumstances, this them. Such bodies may have to slow down the means local people. Because their main motive for process while studies and consultations occur and seeking more power is likely to be to enhance their while people seek alternative livelihoods and economic position, the link between conservation and adjust their expectations of the society in which local enrichment (or poverty avoidance) must be they live. These bodies can take many forms, but made very explicit. This is a major theme in modem they must be genuinely independent of the main conservation projects, which have to show that real parties and competent to act, as well as be trusted benefits can come from conservation. To do this, the and trustworthy. definition of wealth may have to be expanded to * A locality or bioregion can empower itself by uni- include its biological and cultural dimensions as well lateral action, but this can provoke efforts to as its more conventional economic aspect. People reassert control by groups that feel threatened. may also have to be helped to perceive the value of These groups might seek to reclaim bioregional long-term and future benefits as higher than that of revenues or to reverse events through political, temporary get-rich-quick schemes. administrative, or military intervention in the These new and demanding conditions for success- name of national unity. To prevent this, other ful conservation projects imply the need to identify changes in law and policy may be needed to pro- and sustain those features of the project area that tect the newly empowered bioregion and to permit allow wealth creation and ensure poverty avoidance. it to sustain itself. An aim is to neutralize forces that may damage local * There is the risk that a bioregion or community ecosystems and erode biodiversity, while promoting 184 Decentralization and Biodiversity Conservation the local capture of economic benefits and the more Areas and Economic Development in Latin productive use of natural and artificial ecosystems. America and the Caribbean." World Conservation These all depend on adequate levels of environmental Union, Gland, and Inter-American Development awareness and security of resource tenure. If these are Bank, Washington, D.C. Processed. achieved in a project area, they will create more Bhatnagar, Bhuvan, and Aubrey C. Williams. 1992. favorable conditions for other forms of investment, Participatory Development and the World Bank: both by strengthening local institutions and by Potential Directions for Change. Discussion Paper improving local knowledge and thinking skills. The 183. Washington, D.C.: World Bank. interplay among all of these factors means that appro- Binswanger, Hans, Anwar Shah, and A. N. Parker, priate decentralization of relevant functions is as comps. 1994. "Decentralization, Fiscal Systems, much a necessary (albeit insufficient) condition for and Rural Development, Revised Proposal/ conservation as conservation is for sustainable devel- Request for Research Support Budget Funding." opment. Rural Development Department, World Bank, Washington, D.C. Processed. Notes Bonner, R. 1993. At the Hand of Man: Peril and Hope for Africa's Wildlife. London: Simon and Schuster. 1. Or conscientizacdo (Freire 1984), that is, learn- Brandon, Katrina E., and Michael Wells. 1992. ing to perceive and act against oppressive elements of "Planning for People and Parks: Design Dilemrnas." reality (which in our view include harmful environ- World Development 20(April):557-70. mental externalities). Caldecott, Julian 0. 1988. Hunting and Wildlife 2. Under the Communal Areas Management Management in Sarawak. Gland: World Con- Programme for Indigenous Resources (CAMPFIRE), servation Union. local people own the rights to manage and profit from . 1996. Designing Conservation Projects. wildlife resources in their communal lands. This gives Cambridge, U.K.: Cambridge University Press. them a strong incentive to keep wildlife populations Caldecott, Julian O., S. Bashir, and S. Mohamed. as productive assets, rather than allow them to be 1995. Issues in Sustainably Financing Coastal overharvested or lost through habitat destruction. Conservation. Proceedings of the International 3. Sistema Nacional de Areas de Conservaci6n Coral Reef Initiative South Asia regional work- (SINAC), Costa Rica's National System of Conservation shop, the Maldives, 29 November-3 December. Areas, incorporates a number of reserves, each of Chatham, U.K.: Natural Resources Institute; which is managed by a locally appointed and locally London: Overseas Development Administration. accountable committee with a very high degree of Caldecott, Julian O., and T. F. Fameso. 1991. "TFAP autonomy in all areas of decisionmaking. Nigeria: Findings of the Preliminary Conservation 4. The term policing refers here to all forms of and Environment Study Mission." Commission of publicly authorized surveillance and intervention that the European Communities, Brussels. Processed. aim to prevent harm. Freire, P. 1984. Pedagogy of the Oppressed. New 5. Although not to the usual extent at Guanacaste, York: Continuum. which by 1992 had access to an externally financed GEF (Global Environment Facility). 1993. "The Social $11 million endowment fund. Challenge of Biodiversity Conservation," ed. by S. H. Shelton and written by G. Castilleja, P. J. Poole, References and C. C. Geister. Working Paper 1. Washington, D.C. Processed. The word "processed" describes informally repro- . 1995. "Quarterly Operational Report, August duced works that may not be commonly available 1995." Washington, D.C. Processed. through libraries. Haddad, Wadi D., Martin Carnoy, Rosemary Rinaldi, Banerjee, Ajit, and Ernst Lutz. 1996. "Analysis of and Omporn Regel. 1990. Education and World Bank and GEF Projects." In Ernst Lutz and Development: Evidence for New Priorities. Julian 0. Caldecott, eds., Decentralization and Discussion Paper 95. Washington, D.C.: World Biodiversity Conservation. Washington, D.C.: Bank. World Bank. Herz, Barbara, K. Subbarao, and Laura Raney. 1991. Barzetti, V., ed. 1993. "Parks and Progress: Protected Letting Girls Learn: Promising Approaches in Julian Caldecott and Ernst Lutz 185 Primary and Secondary Education. Discussion Income Distribution: Past Evidence and Further Paper 133. Washington, D.C.: World Bank. Analysis. Discussion Paper 46. Washington, D.C.: IIED and ODA (International Institute for Environment World Bank. and Development and Overseas Development UNCED (United Nations Conference on Environment Administration). 1994. Whose Eden? An Overview and Development). 1992. Agenda 21. Rio de of Community Approaches to Wildlife Management. Janeiro: United Nations Conference on London. Environment and Development; Nairobi: United IUCN, UNEP, and WWF (World Conservation Union, Nations Environment Programme; New York: United Nations Environment Programme, and United Nations Secretariat. World Wide Fund for Nature). 1991. Caring for UNEP (United Nations Environment Programme). the Earth: A Strategy for Sustainable Living. 1995. Global Biodiversity Assessment. Cambridge, Gland. U.K.: Cambridge University Press. Janzen, D. H. 1991. "How to Save Tropical WCED (World Commission on Environment and Biodiversity." American Entomologist 37(3, Development). 1987. Our Common Future. fall):159-71. Oxford: Oxford University Press. . 1992. "A South-North Perspective on Wells, Michael. 1992. "Biodiversity Conservation, Science in the Management, Use, and Economic Affluence, and Poverty: Mismatched Costs and Development of Biodiversity." In 0. T. Sandlund, Benefits and Efforts to Remedy Them." Ambio K. Hindar, and A. H. D. Brown, eds., Conservation 21(3):237-43. of Biodiversityfor Sustainable Development. Oslo: Wells, Michael P., and Katrina E. Brandon. 1993. Scandinavian University Press. "The Principles and Practice of Buffer Zones and Kiss, Agnes, ed. 1990. Living with Wildlife: Wildlife Local Participation in Biodiversity Conservation." Resource Management with Local Participation in Ambio 22(2-3):157-62. Africa. Technical Paper 130. Washington, D.C.: Wells, Michael, Katrina E. Brandon, and Lee Hannah. World Bank. 1992. People and Parks: An Analysis of Projects Lutz, Ernst, and Julian 0. Caldecott, eds. 1996. Linking Protected Area Management with Local Decentralization and Biodiversity Conservation. Communities. Washington, D.C.: World Bank. Washington, D.C.: World Bank. Western, D., M. Wright, and S. Strum. 1994. Natural McNeely, Jeffrey A. 1988. Economics and Biological Connections: Perspectives in Community-Based Diversity. Gland: World Conservation Union. Conservation. Washington, D.C.: Island Press. Prud'homme, R. 1995. "The Dangers of White, A. T., L. Z. Hale, Y. Renard, and L. Cortesi, Decentralization." The World Bank Research eds. 1994. Collaborative and Community-Based Observer 10(2):201-20. Management of Coral Reefs: Lessons from Pye-Smith, C., and G. B. Feyerabend. 1994. The Experience. West Hartford, Conn.: Kumarian Press. Wealth of Communities: Stories of Success in World Bank. 1994. "The World Bank and Local Environmental Management. London: Participation." Operations Policy Department, Earthscan. World Bank, Washington, D.C. Processed. Ruitenbeek, H. J. 1990. Economic Analysis of WRI, IUCN, and UNEP (World Resources Institute, Conservation Initiatives: Examples from West World Cons,ervation Union, and United Nations Africa. Godalming, U.K.: World Wide Fund for Environmernt Programme). 1992. Global Nature. Biodiversity Strategy: Guidelines for Action to 1992. "The Rainforest Supply Price: A Tool for Save, Study, and Use Earth's Biotic Wealth Evaluating Rainforest Conservation Expenditures." Sustainably and Equitably. Washington, D.C.: Ecological Economics 6(July):57-78. World Resources Institute; Gland: World Tilak, Jandhyala B. G. 1989. Education and Its Conservation Union; Nairobi: United Nations Relation to Economic Growth, Poverty, and Environment Programme. 15 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick For several decacles, food, agricultural, and natural cultural technologies, or the adoption of new technol- resource management policies have been designed ogy but with negative unanticipated impacts (see without acknowledging that rural men and women Alderman and others 1994 for a review). At the same may have different preferences, face different con- time, gender roles in agriculture are not immutable straints, and respond differently to incentives. Indeed, and may be affected by economic incentives: the gen- until Boserup's (1]970) path-breaking work, women der division of labor by task may change, as expedi- were largely "invisible" within agriculture, and the ent (see Tisch and Paris 1993 on the Philippines), and predominant assumption was that the male heads of women farmers in Africa may increase their produc- household made nmost, if not all, farm allocation and tion of cash crops, traditionally a male domain, as production decisions. Subsequent work (for example, these become more profitable (Saito, Mekonnen, and Dixon 1982 and Safilios-Rothschild 1985) has shed Spurling 1994). Project experience suggests that light on women's contribution within different farming knowledge of the relative bargaining power of men systems, whether as farm managers, laborers, or man- and women could help to predict the outcome of agers of natural resources. More recently, a growing development projects and promote a design that literature has questioned the assumption that house- would not lead to a deterioration of women's access to holds make decisions as one (see Alderman and others and control over resources (von Braun and Webb 1994 and Haddad, Hoddinott, and Alderman 1997 for 1989). reviews). This body of work shows how policymakers This chapter argues that policymakers would do may not be able to predict accurately individuals' well to take gender issues into account when formu- responses to policies due to different preferences as lating food, agricultural, and natural resource man- well as asymmetries in the distribution of rights, agement policies. Based on a review of empirical evi- resources, and responsibilities within households. dence, it shows that reducing inequalities between In many societies, gender is an important determinant men and women in human and physical capital can of the distribution of rights, resources, and responsi- lead to major efficiency and productivity increases in bilities among individuals, households, and communi- agriculture. Moreover, addressing inequalities in the ties (Moser 1989). Because this distribution is socially underlying distribution of property rights between determined, rather than innate or biological, it is also men and women may be essential not only for subject to policy intervention. Project experience increased agricultural productivity but also for envi- shows how neglecting the gender dimension in devel- ronmental and social sustainability. The second-round opment policy design has led to failures in project effects are also important: this chapter shows that implementation, failures in the adoption of new agri- increases in income, when controlled by women, have 186 Agnes R. Quisumbing, Lynn R. Brown, Lawrence Halddad, and Ruth Meinzen-Dick 187 a greater impact both on household food security and Credit is particularly important during the planting on investments in child health and schooling. Finally, season, when seeds and cash inputs have to be pur- paying attention to gender issues in the design of poli- chased, and in the lean season before the harvest, as cies and projects is crucial to designing projects that stocks for consumption are depleted. Credit may also are both successful and sustainable. be essential to smoothing consumption in case of crop failure or drought. Gender-Specific Constraints in Agriculture Collateral requirements, high transaction costs, limited education and mobility, social and cultural Despite their importance in agriculture, women usual- barriers, and the nature of women's business limit ly have lower levels of physical and human capital women's ability to obtain credit. Property that is than men. In many developing countries, women have acceptable as collateral, especially land, is usually lower school enrollment rates, literacy rates, and registered in men's names, and the types of valuables schooling attainment. In the early 1980s, average lit- women have, such as jewelry, are often deemed unac- eracy rates for men in developing countries were over ceptable by fotmal financial institutions. The trans- 50 percent, while more than two-thirds of women action costs involved in obtaining credit-transporta- were still illiterate (Seager and Olson 1986). This dis- tion costs, paperwork, time spent waiting-may be parity continues to be larger in rural areas, where edu- higher for wornen due to higher opportunity costs cational attainment is lower, and persists despite high from forgone activities. Indeed, in rural Kenya, the private rates of return to women's schooling distance to a bank is a significant determinant of (Quisumbing 1995) and high social returns to women's probability of obtaining credit but does not women's education (Subbarao and Raney 1996). affect men's borrowing behavior (Saito, Mekonnen, Laws governing women's rights to land and other and Spurling 1994). Women's lower educational lev- property-whether legal or customary-typically els relative to men, their lack of familiarity with loan provide them weaker rights to land. Many religious procedures, ancL social and cultural barriers may con- laws also discriminate against women in land rights strain their mobility and interaction with predomi- (see Agarwal 1994 for examples for South Asia).2 nantly male credit officers or moneylenders. Women's use rights under customary law in many Moreover, women may be credit constrained because African countries are allocated from their husbands their role as primary caregivers and the health risks and natal families based on their position within a kin- associated with childbearing lead to intermittent ship group and, in particular, on their relationship to a employment, which makes them risky clients for male relative (father, brother, husband). These rights banks. Women also tend to be involved in the produc- entitle women to farm the land, often in exchange for tion of relatively low-return crops that are not includ- labor on their husband's and other family plots. ed in formal sector lending programs. Exclusion from However, these indigenous customs have been modi- local groups, such as fanners groups, may prevent fied by Western colonization, which introduced pri- women from receiving both credit and extension vate ownership by individual registration of land and advice, particularly if the extension worker plays an often discriminated against women in titling. important role in credit delivery. Moreover, extension Furthermore, because women obtain land rights usu- workers tend to favor landowners (Agarwal 1994), ally through a male relative, there is no guarantee that who are rarely women, thereby giving men preferen- they will retain these rights in the case of death or tial access to information. divorce. Moreover, a farm household survey in Kenya In spite of women's important role as farm man- and Nigeria (Saito, Mekonnen, and Spurling 1994) agers and agricultural workers, whether as family or finds that more male than female heads of households, hired laborers, they have seldom been beneficiaries of and more male than female farmers, are able to exer- extension. Traditional extension systems based on cise their land rights fully. Absence of formal land single-commodity extension often fail to consider the rights and smaller land sizes cultivated by women crops and activities in which women are involved. may be critical because land is usually needed as col- Community or rural extension, in contrast, may cover lateral in credit markets. the broad spectrum of women's activities but runs the Access to credit, both formal and informal, has risk of treating specific tasks superficially. important implications for the ability to attain a stable Furthermore, extension systems in many developing standard of living and undertake productive activities. countries are overloaded; in the late 1980s, ratios in 188 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development Africa, Asia, the Near East, and Latin America range African farming systems, plots controlled by women from one agent to 2,000 or 3,000 farmers (Saito and have lower yields than plots farmed by men. Studies Spurling 1992). In contrast, in Europe and North analyzing the technical efficiency of male and female America, one extension agent serves 300-400 farm- farmers have suggested that this is due to lower levels ers. In addition, women are underrepresented among of input use on women's plots and not to inherent extension agents. Even in regions with a long tradition managerial differences between male and female of female farming, such as Africa, in the early 1990s, farmers. For example, studies from Kenya, Korea, only 11 percent of extension staff and 7 percent of Nigeria, and Thailand, summarized in table 15.1, field extension staff were women. Moreover, although show that female farmers are equally efficient as male female extension workers may be trained in agricul- farmers, once individual characteristics and input lev- ture, they are often mandated to give advice only on els are controlled for (Moock 1976, Saito, Mekonnen, home economics subjects. This may constrain the and Spurling 1994, and Jamison and Lau 1982). delivery of agricultural extension messages to female There may be significant gains from increasing farmers, who may be restricted from interacting with women's level of physical and human capital, as sug- male extension agents and who prefer to interact with gested by simulations using coefficients from the pro- female agents. duction functions estimated in the studies cited. Lastly, time is a scarce resource for rural women. Simulations performed using Moock's (1976) coeffi- Many studies have documented the long hours that cients for female farmers suggest that if female maize women spend in productive activities compared with farmers were given sample mean characteristics and men (Brown and Haddad 1995 and McGuire and input levels, their yield would increase 7 percent. If Popkin 1990). In Africa, women spend as much as they were given men's mean input levels and other two hours a day caring for children, three hours characteristics, yields would increase 9 percent. preparing food, and two hours fetching fuel and water. Giving all women at least a year of primary education In rural Asia, food-processing activities take up two to would raise yields 24 percent, which reflects the gains three hours daily, and in rural Bangladesh, women to providing primary education in a setting where may spend as long as six hours a day fetching water. women have very low educational levels. Simulations Where women's nutritional status is already compro- with the Saito, Mekonnen, and Spurling (1994) coef- mised or at risk, demanding physical labor may have ficients suggest a 22 percent increase in women's significant negative effects on their nutritional status, yields on plots of maize, beans, and cowpeas if as suggested by Higgins and Alderman (1997), based women farmers were given the characteristics and on a Ghanaian study. The extent of rural women's input levels of male farmers. Increasing land area and multiple responsibilities often imposes time and ener- fertilizer use to the levels of male farmers would gy constraints on their participation in programs increase women's yields 10.5 and 1.6 percent, respec- designed to increase their incomes and on their will- tively. However, these studies do not take into account ingness to adopt new technologies. the relationship between input application and farmer characteristics. Since more-educated farmers are Productivity and Efficiency more likely to adopt new technologies and apply modem inputs, the contribution of such inputs may be Women's limited control over productive resources overstated and that of education underemphasized. has important consequences for agricultural produc- This may underestimate the consequences of under- tivity. Econometric evidence on gender differences in investing in women's education in rural areas. agricultural productivity points to the importance of More recent literature suggests that the asymmetry investing in women by increasing their human capital, of roles and obligations within the household, partic- through education and extension, and by increasing ularly in African farming systems, may underlie their access to physical and financial inputs women's lower yields. A study in Burkina Faso (Udry (Quisumbing 1995, 1996). 1996) finds that plots controlled by women have sig- nificantly lower yields than similar plots within the Technical Efficiency of Male and Female Farmers household that are planted with the same crop in the same year, but controlled by men. The yield differen- Numerous studies have found that, where male and tials are due to lower input intensity on female-managed female farmers manage separate plots, as in many plots: much less male labor per hectare is devoted to Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick 189 plots controlled by women than to similar plots con- If the household had common objectives and pref- trolled by men. Child labor and unpaid exchange erences, factors of production would be allocated effi- labor are also applied more intensively to plots con- ciently to the various productive activities of the trolled by men. Lastly, virtually all fertilizer is con- household. The consistently higher input intensities centrated in plots controlled by men, even though the on men's plots suggest misallocation of household marginal product of fertilizer diminishes. These dif- resources, which is traced to stronger incentives for ferences in input intensity between male- and female- individuals to achieve high output on their own plots, managed plots persist even after land quality, mea- as well as imperfect labor allocation processes within surement error, or risk management behavior are the household. Estimates show that reallocating cur- taken into account. rently used factors of production across plots could Table 15.1. Production Function Studies with Estimates of Male-Female Differences in Technical Efficiency Dependent variable Study Sample Sex offarmer variable Coefficient definition Kenya Moock (1976) All farmers: Female dummy x log of area 0.090 Log of maize males = 101; planted in maize output per acre females = 51 Female dummy x log of -0.280* plant population per acre Female dummy x log of labor 0.108*** input per acre Female dummy x primary 0.167*** schooling dummy Female dummy x log of -0.028 extension contact index Bindlish and All farmers: Female head dummy -0.022 Log (ln) of crop Evenson (1993) male heads = 434; production in 1990 female heads = 241 Saito, Mekonnen, All farmers: Male farmer dummy -0.017 Log of total value and Spurling males = 306; of crop production (1994) females = 147 at farmer level Nigeria Saito, Mekonnen, All farmers: Male farmer dummy -0.130 Log of total value and Spurling number of fanners = 226; of production at (1994) males = 210; farmer level females = 15 Korea, Rep. of Jamison and Mechanical farms = Male head dummy 0.95*** Log of value of Lau (1982) 1,363 (90.2 percent agricultural crop male heads) output Nonmechanical farms = Male head dummy 0.059 Log of value of 541 (87.6 percent agricultural crop male heads) output Thailand Jamison and Lau (1982) Chemical farms = 91 Male head dummy 0.076 Log of output (97.7 percent male (kilograms) heads) Nonchemical farms = Male head dummy 0.269 Log of output 184 (99.5 male heads) (kilograms) * Significant at 10 percent. ** Significant at 5 percent *** Significant at 1 percent. 190 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development increase the value of household output by 10 to 20 tions by Burger and Gunning suggest that female percent (Alderman and others 1995). decisionmakers are unresponsive to increases in early adoption by male farmers. This is consistent with Technological Adoption by Male and Female Farmers other findings that previous awareness and adoption of the technology, particularly by farmers of the same In some countries, especially where new technologies sex, also increase the probability of adoption. The sig- are associated with "male" crops or activities, women nificance of gender-specific copying effects high- may be less likely to adopt new crops or technologies. lights the need not only for female extension agents to Female farm managers are less likely to grow coffee work with female farmers, but also for female contact in Kenya and rear livestock in Tanzania but are equal- farmers to be chosen. The importance of cooperatives ly likely to have cattle in Kenya and to grow coffee or and extension in many of these studies emphasizes the cocoa in C6te d'Ivoire (Appleton and others 1991). need to provide support infrastructure to rural areas. Female farmers in Zambia are less likely to use oxen Most of the studies reviewed also suggest that in cultivation (Jha, Hojjati, and Vosti 1991), because farmers with larger areas cultivated and higher values plowing is traditionally a male activity. However, of farm tools are more likely to adopt new technolo- women may more readily adopt technologies related gy. To the extent that women farmers may have less to the tasks they perform, particularly if the extension education, have less access to land, and own fewer agent is female. Evidence from Kenya suggests that tools, they may be less likely to adopt new technolo- female farmers are equally likely to apply technical gies. A relatively unexplored area is the extent to advice from extension agents and are even more like- which weaker property rights to land and trees dimin- ly to adopt relatively complex practices such as top ish the incentives to adopt higher-valued, permanent dressing, chemical use, and stalk borer control crops and environmentally sustainable farming prac- (Bindlish and Evenson 1993). Another study from tices. Kenya (Saito, Mekonnen, and Spurling 1994) indi- cates that female farmers are more likely to adopt Property Rights and Natural Resources improved seed and use agrochemicals. In Nigeria, although male farmers are more likely to use insecti- The distribution of property rights to land and natural cide, male and female farmers are equally likely to resources underlies men's and women's differential use fertilizer. control over productive resources. Not only does Despite the mixed evidence on technological addressing gender asymmetries in control over adoption by gender, most of the studies find that resources improve productivity, equity, and human more-educated farmers are more likely to adopt new development, it also is critical to environmental sus- technologies. In Kenya, more-educated females, in tainability. One of the basic arguments for paying particular, are more likely to adopt coffee. Raising the attention to property rights is that secure tenure educational level of female farmers by giving them encourages investment, which leads to more produc- universal primary education has higher marginal tive and sustainable use of the resource (see Besley effects on the probabilities of adopting coffee than 1995 and Place and Hazell 1993). But if women are raising the educational level of all farmers, due to the blocked from certain avenues of investment (for generally lower levels of female education in Africa example, tree planting), or if they know that particu- (Burger and Gunning, personal communication). lar investments and increases in productivity will lead Simulations suggest that a 10 percent increase in the to the loss of their access to land, their insecurity of percentage of women having completed primary tenure can be a barrier to productivity and their incen- schooling leads to a 6 percent increase in early adop- tives to sustain the resource over time. This is often tion and a 14 percent increase in late adoption. the case when agriculture becomes commercialized, Increasing female education also has a higher impact especially where food production is under women's on adoption than increasing land sizes-a 10 percent control, but the income from cash crops goes to men increase in arable land increases early adoption only (for example, see Chambers and Morris 1973 and Dey by 2 percent and late adoption by 6 percent. 1981). Providing universal primary education may also Official policies toward granting rights from the stimulate early adoption by female farmers, whom state range from favoring men (as in many irrigation other women are more likely to imitate. Other simula- systems; see Zwarteveen 1997) to giving preference Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick 191 to women or women's groups. Even where official resources to exploit (for example, grazing lands policies are gender-neutral in govemment-allocated require cattle, nmarine fisheries require boats, ground- rights, women may have difficulty acquiring those water requires wells and pumps). Creating a level rights (see Lastarria-Cormhiel 1997). Limitations to playing field for women may require addressing these access include money, where legal or illegal adminis- hurdles as much as the formal rules and laws relating trative payments are made, little knowledge of the to resource tenure. public institutions, and distance, which may constrain These issues speak principally to formal rights of women more due to their lower educational attain- control, such as exclusion, sale, and availability of ment and limited mobility. Limited access also results land to others. 'Women's access to resources for their from traditional expectations of women's place and own productive and reproductive activities is more behavior that keep women in the private domain (see prevalent than t:heir control. The pattems of access are Freudenberger 1994). more complex and more nuanced. It is in the informal The reasons for divergence between de jure and de use rights where we are more likely to see the flexi- facto equity are particularly important for program bility and subtleties that characterize actual practice. design. For example, acquiring resource rights But the process of formalizing property rights though through labor contributions ("sweat equity") appears privatization and titling of resources reduces the com- to be a more equitable route than cash purchase or plex bundle of iights into a single unitary right. When inheritance for resource-poor households and individ- customary use rights are not recognized, many uals. But this does not always hold in practice. women and marginal users lose out. The effect of gen- Women may also be constrained by time, owing to der relations on formal and informal property rights their additional domestic responsibilities, from con- suggests that policymakers need to look beyond legal tributing sweat equity to irrigation, trees, and land rights and to consider removing gender-based con- development projects. straints to other services and rights, which combine to There is a tendency to idealize community limit women's access to property (Meinzen-Dick and resource allocation as being very equitable. Indeed, others 1997). communal tenure systems often provide for all house- There is also a need to explore altematives to free- holds to have some land (although women's rights hold tenure that allow more flexible use patterns, may be subordinate to men's; see Lastarria-Cormhiel which can benefit women as well as men. To do this 1997). The poor depend heavily on common property requires good examples of tenure arrangements that resources such as pasture or forests that are available accommodate multiple users. Because many of these to members of a community (Jodha 1992), and many are customary rather than statutory arrangements, communities have norms that no one should be denied there is a need for more written documentation that access to basic drinking water. However, community could be disseminated to policymakers and others norms regarding the appropriate status for women involved in shaping tenure arrangements. Especially may be the greatest barriers to women's control over important in this regard are examples where the resources, especially independent rights to the resource or the products of that resource have entered resource. Clearing trees is often a way to establish the market, rather than only subsistence products with land rights under communal land tenure systems in more limited demand. Africa, but tree clearing is almost exclusively a male Looking specifically at women's uses of natural task in most societies and thus precludes women from resources allows us to see a wider variety of uses and establishing a land right. In other cases, planting trees hence gives a inore accurate accounting of productiv- to prevent soil erosion or improve output requires ity. Hedgerows, wetlands, forests, and other niches clear land rights. Even use of the commons may are sources of food, fuel, thatching, mats, or medici- require access to other complementary means of pro- nal plants that, are especially important to women duction, such as animals to use grazing lands, favor- (Rocheleau and Edmunds 1997). Property regimes ing those with more resources (see Brouwer 1995). and resource management systems that maximize out- Looking at the complementary inputs required to put of a single commodity may appear to give the obtain rights sheds light on the barriers to equitable highest retums under conventional analysis, but when access. Land titling requires political connections and we look at the full spectrum of uses, other property know-how, as much as cash. Even com-mon property regimes may have a higher value of output.3 For or open-access resources may require some private example, looking only at marketed logs as the output 192 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development of a forest misses the total value of fodder, fruits, domestic responsibilities prevented them from sup- "minor forest products," and kindling. Measuring plying the labor needed when required, lowering their only the grain output of an irrigation system misses profit share and causing many women to withdraw. the vegetable gardens, livestock watering, domestic By 1995 the number of male farmers exceeded the water supply, and many household enterprises for number of women (Ahenkora and others 1995). which women use water from the system. Failure to However, groups made up of all men or all women do recognize the value of those resources, or the custom- not necessarily imply homogeneity and the ability of ary rights of their users, can reduce both the produc- all to participate. A credit program in Mali was target- tivity and equity of management systems. ed to women using group formation as a "collateral" In many cases, collective action by women has mechanism through peer group pressure to repay. An been instrumental in securing rights for women, either evaluation indicated that women with preschool chil- as a group or individually. Where women are blocked dren were less able to use the credit to generate a pos- from holding land individually, they may be able to itive return than women without preschool children obtain a parcel for a women's group to use for a col- (De Groote and others 1996). Homogeneity of group lective garden or nursery, for example. Women's irri- members in terms of the activity or its goal may be gation groups in Bangladesh have invested in water more important than gender per se (see, for example, pumps, enabling them to use or sell groundwater (van Baland and Platteau 1996). This requires empirical Koppen and Mahmud 1995). In addition, collective determination and the lessons from successful and action may lead to a change in the rules, permitting unsuccessful cases to be shared more widely. individual women to obtain stronger rights over the resource (see Agarwal 1994 for good examples from Gains to Increasing Women's Incomes South Asia). Although there is a large number of both nongovernmental organizations (NGOs) and local self- The previous sections have shown how disparities in help groups, an important question in each area is the men's and women's human and physical capital and presence (or absence), shape, objectives, and mem- access to natural resources have costs in terms of for- bership of women's organizations as vehicles for gone increases in agricultural output and the potential learning more about women's roles and needs and as for sustainable natural resource management. a potential base for women to assert their needs. Constraints to increasing women's incomes may also The other side of this issue concems the integra- have important consequences for food security, indi- tion of women as rights-holders and decisionmakers vidual welfare, and investment in the next genera- into traditionally male-dominated institutions for col- tion's human capital. A growing body of evidence lective resource management. Policies that devolve suggests that women spend a higher proportion of authority for resource management from the state to their incomes on expenditures related to household local institutions make it all the more critical that the food and nutrition security. A number of studies con- local institutions function effectively. But how does ducted in the 1980s suggest that women and men the gender composition of local institutions affect spend income under their control in systematically their strength and effectiveness? Male emigration or different ways (Guyer 1980 and Dwyer and Bruce diversification out of agriculture makes this increas- 1988). Women typically spend a high proportion of ingly important in many parts of the world. A related their income on food and health care for children as question is whether women are better off integrating well as other goods for general household consump- into existing male-dominated groups or setting up tion. In contrast, men retain discretionary control over their own groups for resource management (nurseries, a higher proportion of their own incomes for personal social forestry action)? The different roles and respon- expenditures. These studies have been confirmed by sibilities of women can prejudice their ability to inte- more recent studies that provide quantitative mea- grate successfully in mixed groups. A food producer sures of the different effects of men's and women's cooperative set up in 1992 in Gbefi, Ghana, with grant incomes. Many of these studies have already attempt- funding conditioned on 50 percent of the members ed to control for unobservable household-level and being female initially had a membership level of 59 community factors and the endogeneity of women's percent women. Members were required to provide labor income.4 labor to the cooperative fields, receiving a profit share Evidence from Africa, Asia, and Latin America in proportion to their labor input. However, women's shows that women's income has a greater effect on Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick 193 household food security and preschooler nutrition three programs. Indeed, for one of the programs run than men's income (see table 15.2). In southwestern by the Grameen Bank, total household consumption Kenya, for a given level of household income, increases by TkI8 for every TklOO lent to a woman female-controlled income share has a positive and and by Tkl 1 fcr the same amount lent to a man (Pitt significant effect on household consumption, while and Khandker 1995). Doss's (1997) study finds that, men's income has a negative effect (Kennedy 1991). in rural Ghana, a larger share of assets controlled by Similarly, in Taiwan (China), women's income share women increases the budget shares allocated to food has a significant and positive effect on household and improves child education outcomes; it decreases budget shares allocated to staples and education and a the budget shares of alcohol, recreation, and tobacco. negative effect on budget shares allocated to alcohol Consistent with the earlier literature, Wang (1996) and cigarettes (Thomas and Chen 1994). In finds that in Zambia, mother's income increases Guatemala, the average yearly profits from nontradi- weight-for-age more for infant girls than for infant tional agricultural export crops would increase house- boys, while father's income increases weight- and hold food expenditures by twice as much if they were height-for-age more for boys than for girls. controlled by women rather than their husbands (Katz Why do men and women tend to spend income dif- 1992). Finally, the effect of women's unearned ferently? Societal and cultural norms may assign income on child survival probabilities in Brazil is women the role of "gatekeepers," who ensure that almost 20 times higher than that of men's unearned household memabers, especially children, receive an income (Thomas 1994). adequate share of available food. Alternatively, More recent work summarized in table 15.3 has women may prefer to spend more on children's daily attempted to control for the possibility that govern- needs because they spend more time with them. ment programs are placed as a result of community Women may also face different constraints than men. characteristics rather than random assignment (Pitt To minimize the competing demands on their time, and Khandker 1996) as well as to derive alternative for example, women may spend more on food measures of nonlabor income based, for example, on because they purchase more expensive calories that asset ownership (Doss 1997). Pitt and Khandker's take less time to prepare. Finally, women and men (1996) study of three NGO credit programs in may have different income flows and thus different Bangladesh finds that in the overwhelming majority transaction costs. In other words, because women's of cases, the effects of women's credit compared with income tends to come more frequently and in smaller credit provided to men are as expected-that is, amounts, it may be more readily spent on household greater female labor supply, greater accumulation of daily subsistence needs than lumpier seasonal women's nonland assets, greater use of contracep- income, which tends to come to men and is likely to tives, and lower fertility. In addition, a strong result on be spent on more expensive items (Hamilton, Popkin, household total consumption is found in two out of and Spicer 1984). A recent study tested this hypothe- Table 15.2. Effects of Men's and Women's Income or Assets on Household Welfare Ratio of effect of Effect of Effect of women 's income or women's income men's income assets to that of Country and study Effect on or assets or assets men's income or assets Africa Kenya, Kennedy Household calorie level Positive n.a. n.a. (1991) Asia Taiwan (China), Thomas Household budget share Negative Negative 1.3 and Chen (1994) for alcohol Latin America Brazil, Thomas (1994) Child weight-for-height Positive Positive 4.2 Brazil, Thomas (1994) Child survival Positive Positive 18.2 Guatemala, Katz (1992) Food expenditures Positive Positive 2 n.a. Not applicable. Table 15.3. Summary of Recent Studies of the Effects of Female Income, Assets, and Credit on Selected Welfare Outcomes Country Study (data year) Estimation and test Conclusion Comments Rogers, Swindale, Honduras Ordinary least squares: "Controlling for income level, the percent No instrumenting and Ohri- (1993-94) effect of percent of income of income earned by the child's mother has a of the share of Vachaspati earned by mother on positive effect on nutritional outcome ... the women's income (1996) preschooler standardized positive effect ... decreases as children get height-for-age older." (p. 150) Pitt and Khandker Bangladesh Weighted exogenous "The set of female credit variables is Estimation and sampling (1995, 1996) (1991-92) sampling maximum statistically significant in seven of eight cases of households take care likelihood fixed effects: at the 0.05 percent level. By contrast the set of to identify the average testing for differential male credit variables is significant in three out impact of the credit impact of male and female of eight cases ... the hypothesis that male and program;the key is to borrowing from NGOS on female credit parameters are jointly equal for sample households in eight outcomes-boy's and each of these three (NGO) programs is program villages that are girl's schooling, women's rejected in only four cases: women's labor ineligible due to some and men's labor supply, supply, women's nonland assets, contraception, exogenous constraint (such total household expenditure, and fertility." (p. 41) as landholding restrictions contraception use, fertility, in the absence of a land and value of women's Household consumption increases by Tkl8 market) nonland assets for every TklOO lent to a woman and Tkl 1 for the same amount lent to a man (Pitt and Khandker 1995) Ward-Batts United Ordinary least squares and Follows up on work of Lundberg, Pollak, and Total expenditure not (1996) Kingdom Tobit: Are budget shares on Wales (1997) instrumented; changes in (1973-76 and 23 goods affected by dummy other policies not fully 1980-82) variable for policy shift in Strong negative impact on men's clothing, accounted for U.K. child benefit allocation tobacco, housing, strong positive impacts (value added tax) (from men and women to on children's clothing, fuel, and food at home women)? Doss (1997) Ghana Ordinary least squares For urban households, female asset share has Total expenditure not (1991-92) and logit: budget share a significant impact on seven out of eight instrumented; assets = equations savings budget share categories: food (+), alcohol (-), accounts and education (+), recreation (-), tobacco (-), businesses controlled by household (-), and miscellaneous (-); for household; assignation rural households, female farmland share has of ownership of assets a significant impact on food (+), alcohol (-), assumes that recreation (-), tobacco (-), and completeness knowledge and of vaccinations (+). For rural households responsibility for female asset share significantly affects four business equal control out of eight budget share categories: food (+), alcohol (-), recreation (-), tobacco (-), and child education outcomes (+) Wang (1996) Zambia Fixed-effects instrumental Mother's income increases weight-for- Because of fixed-effects (1991-93) variables age more for infant girls than for infant methods used, can only boys; father's income increases boys look at the differential weight- and height-for-age more for boys effect of mother's and than for girls; the difference between these father's income on child differential effects is only significant anthropometric outcomes; for weight-for-age at the 10 percent level; need to use panel IV for height-for-age, the differential methods that allow for effect of male and female income is instrumentation of time- significantly different invariant variables (for example, the Hausman- Taylor estimator) Gammage (1997) Nigeria Univariate logit Association of women's income contributions Needs multivatate analysis (1996) with decisions on reproductive health and to control for other fertility: subsets of these associations are observable factors and significant and positive, while none is for unobservable factors negative and significant 196 The Impor-tance of Gender Issues for Environmentally and Socially Sustainable Rural Development sis by controlling for the flows of income earned by City, the United Nations declared that 1976-85 would men and women in Niger (Hopkins, Levin, and be the U.N. Decade for Women: Equality, Haddad 1994). The findings indicate that the timing Development, and Peace. The 1995 Fourth World of female income flows has a significant effect on Conference on Women in Beijing set forth a platform both total household expenditures and food expendi- of action for member states to undertake to improve tures in a given season, while the timing of male women's status and welfare. Although much has been income flows has no effect. This suggests that women learned over the years, much remains to be done. have less access than men to resources that tend to Not surprisingly, an early approach to reducing the even out consumption, such as credit and savings. asymmetries focused on targeting projects and pro- Thus, both the timing of overall household income grams to women. Unfortunately, in most cases it was and the flow of income by gender influence seasonal no more successful at improving women's status than food expenditures. previous policies targeted to the household, and there- What are the policy and programmatic implica- by often implicitly to men (Penia, Webb, and Haddad tions of the differential impacts of women's income? 1996). There are two key reasons for this: women are The implications ,for the design of transfer programs not isolated economic actors, and women per se are are straightforward: when the benefits to targeting not a homogeneous group. There are numerous exam- women (in terms of the intervention's objective) out- ples of project failure with respect to the former. For weigh the costs, then targeting should be undertaken. example, when irrigation was introduced in The More generally and more importantly, these results Gambia in the early 1980s to raise yields, commer- stress the impact of women's economic status within cialize rice production, and increase women's share of the household on a number of important policy objec- household income, community initiatives ended up tives: fertility reduction, food purchases, nutrition, reducing women's income. Previously, women were and education to name a few. The previous sections the rice growers. Yield increases transformed rice suggested that excluding women from productive from a private crop under the control of women to a activity is a waste of talent that will result in lost eco- communal crop under the control of men (von Braun nomic productivity today. The results in this section and Webb 1989 and Dey 1981). Moreover, more suggest that the second-round impacts of improving recent attempts by donors to safeguard women's economic opportunities for women could be profound access to land were frustrated by managers of the in terms of improving economic productivity tomor- project, who sided with male household heads in dis- row-through lowered fertility, improved child sur- putes over access to land (Dey 1992). vival, and increased human capital. Neither are women, as a group, homogeneous. Heterogeneity is demonstrated by differences in class Toward Gender-Sensitive Project Design and caste, ownership of land versus landlessness, rental versus ownership, life cycle stage, marriage The previous sections have documented the striking order (where there is polygyny), position as a female differences between men and women in the owner- head of household or as part of a joint or male-headed ship of, access to, and control of assets and why that household, and household composition, all of which matters for a wide range of outcomes such as agricul- are sources of variance that may be greater than their tural productivity, food security, environmental sus- common interests as women (Meinzen-Dick and oth- tainability, and human welfare. If policymakers and ers 1997). Class and power relationships crosscut project designers continue to implement programs gender. In Mexico, privatization of land dominated by that do not take account of gender roles, responsibili- the well-to-do has marginalized poor men as well as ties, and resource constraints, neither sustainable eco- poor women, with men migrating out (Goldring nomic growth nor successful development is likely. 1996). Ethnicity may be important, as in Ecuador, Recognition of these differences, however, is not where indigenous and mestizo women have differen- new. The Percy Amendment to the U.S. Foreign tial access to productive resources such as land Assistance Act, passed by the U.S. Congress in 1973, (Ahlers 1995). mandated the integration of women into national Many researchers and policymakers who are unfa- economies, recognizing that doing so would improve miliar with gender issues use a "shortcut" approach to development efforts as well as the status of women. In reaching women by identifying female-headed house- 1975, at the World Conference on Women in Mexico holds to whom to provide benefits or promote equity. Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick 197 Although these households may be perceived to have of 121 were highly participatory in their design, and more autonomy, they are by no means more homoge- of those, only 10 successfully reached women neous than women in general. They are differentiated (Narayan 1995). Respect needs to be given to local not only by the factors identified above but also by the norms of behavior to ensure that women are given process by which they became female headed: appropriate opportunities to identify their needs and divorce, widowhood, or male outmigration. For those participate in project discussions. This may involve de jure female heads of household, inheritance pat- holding separate meetings with men and women, terns-patrilineal or matrilineal-are a significant structuring worrmen's meetings to fit with their domes- determinant of the degree of resource access and own- tic work schedules, and identifying who the implicit ership. female leaders are. Even where planned projects are Successful project or policy formulation is unlikely thought to be relevant to either men or women, they without gender analysis.5 This goes beyond a women- are likely to be less successful if spouses are not in-development approach. In the latter, the differences involved in pre.project community discussions. In an in resource access, ownership, and control between oilseeds processing project in Zimbabwe, men only men and women are identified; in the former, the permitted their spouses to become fully committed to determinants of these asymmetries and their conse- project meetings and activities after the project had quences for individuals, households, communities, been discussed with them and they could appreciate and economic development are recognized (Murphy the benefits to thleir wives and families (Mehra 1996). 1995). In essence, the women-in-development Successful gender analysis at this early stage will approach recognizes an outcome, whereas a gender- ensure not only good technical design but also good and-development approach recognizes both the out- socioeconomic design of projects: a design that takes come and its consequences as well as the process account of local production systems, resource bases leading to those outcomes. The understanding gained and access, distribution of labor, distribution of poten- from the gender-and-development approach facili- tial project benefits, and the bargaining power of men tates understanding not only how a planned project or and women both within and across different classes policy will affect men and women but also how it will and castes. affect the underlying processes that condition the allo- The results of gender analysis should then be oper- cation of rights, resources, and responsibilities within ationalized. In the first instance, having completed the communities and households. gender analysis stage, a planner should be in a posi- We recognize that gender analysis requires tion to determiine the gender "design" or balance of resources. However, the costs of failing to understand the project. At one end of the spectrum, projects may these asymmetries can be high in terms of wasted be designed solely for women, with only female par- resources or negative effects on household welfare ticipants. At the other end of the spectrum, projects and resource use (von Braun, Puetz, and Webb 1989). may be completely mainstreamed, with a gender- Additionally, recognizing the process by which gen- neutral project design giving equal opportunities for der asymmetries are determined broadens the array of participation to men and women. Somewhere in- policy instruments available to tackle them. For between these two options lie hybrid approaches- example, changing legal frameworks or incentive having a "women-only" component within a larger structures can create an enabling environment to project or having separate budgeting and reporting of encourage project or policy success. objectives by gender within a mainstream project. Although participatory approaches to project There is no "first-best" recommendation because each design have been recognized as a way of ensuring has its advantages and disadvantages and the appro- community involvement, commitment, and owner- priate choice depends on the community's character- ship of development projects, they are not synony- istics and the p:roject's goals. mous with incorporating gender analysis. As shown Mainstreaming, for example, has the advantage of here, community resource allocation is not necessari- being able to draw on all project resources and is thus ly equitable. Neither do participatory approaches to more likely to be sustainable. It also provides scope project design guarantee that women will participate for capacity tuilding. Unless the organization or in either their needs assessment or the proposed proj- implementing agency is sensitive to women's needs, ect design (Mosse 1995). A World Bank study finds however, the project itself can be gender insensitive, that only 20 rural water supply projects in a portfolio such that women may be marginalized. Moreover, it 198 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development may be difficult to monitor performance with regard cessful in meeting women's needs in situations where to gender. Mainstreamed projects are more likely to there is no constituency for more gender-integrated be successful if a high level of awareness of gender projects. Striving to make the project visible but sen- issues is reflected in project design. A strong commit- sitive to community norms can work to build a con- ment from local institutions, community leaders, and stituency for the future. This can be the first step to project directors is also essential if women are to reap achieving gender integration, first through a project benefits in proportion to their participation. subcomponent, after which it is fully mainstreamed. Another optioni is to have an identifiable budget Having decided on the appropriate gender balance and reporting structure within a mainstream project. of the project, it is important to ensure the ongoing In this model, attention to gender issues is more visi- participation of women. During gender analysis, the ble, gender-specific indicators are easier to monitor, "critical path" should be identified-those aspects of and the project itself can raise awareness of the need the project that must be achieved to guarantee overall to pay attention to gender issues. Although separate project success with respect to women's participation budgeting can help to maintain a high profile for and accrual of benefits. These elements should form a women in the project, greater care needs to be taken central core of project monitoring activities. A system to prevent tokenism: it may simply be used as an of indicators should be devised at the beginning of the excuse to allocate a small budget for women's partic- project that will adequately track the performance of ipation. the project. This is particularly important where there A separate wornen's project component allows the is a lack of constituency for gender issues and where needs of women to be met more specifically than in a supervision is provided from outside the community. mainstream project. This enables the project to be vis- Care should be exercised so that women can make ible and fine-tuned to women's needs, while being the required and envisaged project commitments. For linked to the mainstream. It may also enable project example, where "sweat equity" represents the com- implementers to tap other resources and access tech- munity commitment to a project, it is necessary to nical expertise. It also has a higher chance of guaran- ensure that the timing of such provision fits into teeing women access to project benefits. However, women's domestic schedules; otherwise project tim- such projects are often a low priority of policymakers ing may need to be modified. Where user groups are and implementing organizations. They are typically formed to manage natural resources, appropriate rep- underbudgeted and small in scale and, without suffi- resentation by women should be assured. Even where cient commitment from project leaders, face the dan- projects are targeted solely to women, such as repro- ger of being marginalized. They may also divert atten- ductive health programs, they are less likely to be suc- tion from gender issues in the broader project. Lastly, cessful if program delivery does not assign value to having a tendency to focus on welfare issues, they women's time and recognize their significant eco- may neglect women's role as agricultural producers nomic roles and responsibilities (Mehra 1996). and managers. However, they are a good method of demonstrating to local communities the efficacy of Concluding Remarks addressing women's issues and can build a con- stituency for mainstreaming gender issues in the We have reviewed a literature suggesting that gender future. asymmetries and failure to acknowledge them have Finally, organizations may choose to implement a short-run and long-run costs. In the short run, gains in free-standing women-in-development project. Such a agricultural productivity and household food security project addresses women's specific needs, is visible, are forgone. In the longer run, sustainability of agri- and is easily monitored. However, it is often difficult cultural production is compromised by less-than-opti- to get a constituency for the project from both com- mal natural resource management, as are investments munity leaders and government. Because free-stand- in schooling and fertility control. ing projects are typically small in scale, they can be If the gains to recognizing and mitigating gender invisible to local and national government ministries asymmetries are so compelling, why do they persist? and thus may not raise awareness of gender issues for First, gender analysis is not yet a widely used tool for future projects. They often focus on welfare rather policy analysts and project designers and imple- than broader issues of empowerment and resource menters, partly due to the confusion between the access. Nevertheless, these projects can be highly suc- women-in-development goals and the gender-and- Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick 199 development goals. Second, gender analysis can be References costly in terms of data collection (more disaggregat- ed, though better targeted) and training. Third, there is The word "prozessed" describes informally repro- confusion as to whether targeting women necessarily duced works that may not be commonly available reduces asymmetries. Fourth, although the social and through libraries. economic environment can be altered to reduce or Agarwal, Bina. 1994. A Field of One's Own: Gender eliminate these asymmetries, it is a difficult process and Land Rights in South Asia. Cambridge, U.K.: that may involve changing the legal environment. Cambridge University Press. Fifth, many institutions beyond the household (such Ahenkora, Salome, Elizabeth Akpalu, Joanna Kerr, as community groups, extension agents, credit Kofi Marfo, and Francis Ulzen-Appiah. 1995. groups, project managers) need to see the value to all "Gender and the Implementation of Structural their members of using the gender lens. This involves Adjustment in Africa: Examining the Micro- operationalizing gender analysis within ministries, Meso-Macro Linkages. The Ghana Study." Report NGOS, and projects. prepared for the Canadian International The challenge for research is to improve estimates Development Agency for the Structural and documentation of the costs of gender asymme- Adjustment t.nd Gender in Africa (sAGA) Initiative. tries and to develop better operational tools for gender North-South Institute, Ottawa. Processed. analysis. These tools are needed to enhance the abili- Ahlers, R. 1995. Untitled electronic mail communica- ty of development actors to assess, a priori, the costs tion submitted to the gender-prop e-mail confer- and benefits of gender analysis for themselves. ence, International Food Policy Research Institute, Washington, D.C., 13 October. Processed. Notes Alderman, Harold, John Hoddinott, Lawrence Haddad, and Christopher Udry. 1995. Gender 1. A classic example is Jones's (1986) study of Differentials in Farm Productivity: Implications Cameroon. Rice is considered a male crop, so any for Household Efficiency and Agricultural Policy. income generated from it was controlled by men, Food Consumption and Nutrition Division even if the crop was produced by women. Discussion Paper 7. Washington, D.C.: Consequently, few women entered into improved rice International Food Policy Research Institute. cultivation. Instead, they continued to grown Alderman, Harold, John Hoddinott, Lawrence sorghum, which they controlled, despite its lower Haddad, and Stephen Vosti. 1994. "Strengthening returns. Agricultural amd Natural Resources Policy through 2. In Hindu law, women do not have the right to Intrahousehold Analysis: An Introduction." American own, acquire, or dispose of property, although they Journal ofAgricultural Economics 76(5):1208-12. may have the right to some personal property. In Appleton, Simon, David L. Bevan, Kees Burger, Paul Islamic law, although women's landownership rights Collier, Jan Willem Gunning, Lawrence Haddad, are explicit, the share given to female heirs is half that and John Hoddinott. 1991. "Public Services and given to males. Household Allocation in Africa: Does Gender 3. This is related to the broader question of Matter?" Center for African Studies, Oxford accounting for women's productivity, including both University, Oxford. Processed. home and market production. See Beneria (1992) and Baland, Jean-Marie, and Jean-Philippe Platteau. 1996. Quisumbing (1996) for reviews of these issues. Halting Degradation of Natural Resources. Is 4. Econometric approaches have involved using There a Role for Rural Communities? Oxford: fixed-effects methods to control for unobserved Clarendon Press and Food and Agriculture household and community-level factors. The endo- Organization of the United Nations. geneity of women's income has been addressed by Beneria, Lourdes. 1992. "Accounting for Women's instrumenting women's income or using nonlabor Work: The Progress of Two Decades." World income instead of total income. Developmenir 20(11):1547-60. 5. Good sources for gender analysis include Besley, Timothy. 1995. "Property Rights and Feldstein and Jiggins (1994), Thomas-Slayter, Esser, Investment Incentives: Theory and Evidence from and Shields (1993), and Fong, Wakeman, and Ghana." Journal of Political Economy 103(5): Bhushan (1996). 903-37. 200 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development Bindlish, Vishva, and Robert Evenson. 1993. Evaluation Gender Analysis in Agriculture. West Hartford, of the Performance of T&V Extension in Kenya. Conn.: Kumarian Press. Technical Paper 208. Washington, D.C.: World Fong, Monica S., Wendy Wakeman, and Anjana Bank. Bhushan. 1996. Toolkit on Gender in Water and Boserup, Ester. -[970. Women's Role in Economic Sanitation. Survey Toolkit Series 2. Washington, Development. New York: St. Martin's Press. D.C.: World Bank, Poverty and Social Policy Brouwer, R. 1995. "Common Goods and Private Department. Profits: Traditional and Modem Communal Land Freudenberger, M. S. 1994. "Tenure and Natural Management in Portugal." Human Organization Resources in The Gambia: Summary of Research 54(3):283-94. Findings and Policy Options." Land Tenure Center, Brown, Lynn R., and Lawrence Haddad. 1995. "Time University of Wisconsin, Madison. Processed. Allocation Patterns and Time Burdens: A Gammage, Sara. 1997. "Women's Role in Decision Gendered Analysis of Seven Countries." Making at the Household Level: A Case Study in International Food Policy Research Institute, Nigeria." Intemational Center for Research on Washington, D.C. Processed. Women, Washington, D.C. Processed. Chambers, Robert-, and Jon Morris, eds. 1973. Mvvea: Goldring, L. 1996. "Irrigation, Privatization, and An Irrigated Rice Settlement in Kenya. Munich: Migration." Paper submitted to gender-prop e-mail Weltforumr Verlag. conference, Intemational Food Policy Research De Groote, Hugo, Eileen Kennedy, Ellen Institute, Washington, D.C., 11 October. Payongayong, and Lawrence Haddad. 1996. Processed. "Credit with Education for Women in Mali: Self- Guyer, Jane. 1980. "Household Budgets and Women's Selection and Impacts on Women's Income, with Incomes." African Studies Center Working Paper Implications for Adult Female and Preschooler 28. African Studies Center, Boston University, Nutrition." Report for the U.S. Agency for Boston. Processed. Intemational Development. International Food Haddad, Lawrence, John Hoddinott, and Harold Policy Research Institute, Washington, D.C. Alderman. 1997. Intrahousehold Resource Processed. Allocation: Methods, Models, and Policy. Dey, Jennie. 1981. "Gambian Women: Unequal Washington, D.C.: International Food Policy Partners in Rice Development Projects." Journal Research Institute; Baltimore, Md.: Johns Hopkins of Development Studies 17(3, April): 109-22. University Press. 1992. "Gender Asymmetries in Hamilton, S., Barry Popkin, and D. Spicer. 1984. Intrahousehold Allocation of Land and Labor: Women and Nutrition in Third World Countries. Some Policy Implications." Paper presented at the Praeger Special Studies. New York: Begin and conference on intrahousehold resource allocation: Garvey. policy issues and research methods, Intemational Higgins, Paul A., and Harold Alderman. 1997. "Labor Food Policy Research Institute and World Bank, and Women's Nutrition: The Impact of Work Washington, D.C., February. Processed. Effort and Fertility on Nutritional Status in Dixon, Ruth B. 1982. "Women in Agriculture: Ghana." Journal of Human Resources 22(3): Counting the Labor Force in Developing 577-95. Countries." Population and Development Review Hopkins, Jane, Carol Levin, and Lawrence Haddad. 1994. 8(3, September):539-65. "Women's Income and Household Expenditure Doss, Cheryl. 1997. "The Effects of Women's Patterns: Gender or Flow? Evidence from Niger." Bargaining Power in Household Health and American Journal of Agricultural Economics Education Outcomes: Evidence from Ghana." 76(5):1219-25. Department of Economics, Williams College, Jamison, Dean, and Lawrence Lau. 1982. Farmer Williamstown, Mass. Processed. Education and Farm Efficiency. Baltimore, Md.: Dwyer, D., and .J. Bruce. 1988. A Home Divided: Johns Hopkins University Press for the World Women and Income in the Third World. Palo Alto, Bank. Calif.: Stanford University Press. Jha, Dayanatha, Behjat Hojjati, and Stephen Vosti. Feldstein, Hilary Sims, and Janice Jiggins, eds. 1994. 1991. "The Use of Improved Agricultural Tools for the Field: Methodologies Handbook for Technology in Eastern Province." In Rafael Celis, Agnes R. Quisumbing, Lynn R. Brown, Lawrence Haddad, and Ruth Meinzen-Dick 201 John T. Milimo, and Sudhir Wanmali, eds., Agricultural Economics 58(5, December):831-35. Adopting Improved Farm Technology: A Study of Moser, Caroline. 1989. "Gender Planning in the Third Smallholder Farmers in Eastern Province, World: Meeting Practical and Strategic Gender Zambia. Washington, D.C.: International Food Needs." World Development 17(11, November): Policy Research Institute. 1799-825. Jodha, N. S. 1992. Common Property Resources: A Mosse, David. 1995. "Authority, Gender, and Missing Dimension of Development Strategies. Knowledge: Theoretical Reflections on Discussion Paper 169. Washington, D.C.: World Participatory Rural Appraisal." Economic and Bank. Political Weekly, 18 March. Jones, Christine. 1986. "Intrahousehold Bargaining in Murphy, Josette L. 1995. Gender Issues in World Response to the Introduction of New Crops: A Bank Lending. Washington, D.C.: World Bank, Case Study from Northern Cameroon." In Joyce L. Operations Evaluation Department. Moock, ed., Understanding Africa's Rural Narayan, Deepa. 1995. The Contribution of People's Households and Farming Systems. Boulder, Colo.: Participation: Evidence from 121 Rural Water Westview Press. Supply Projects. Environmentally Sustainable Katz, Elizabeth. 1992. "Intrahousehold Resource Development Occasional Paper 1. Washington, Allocation in the Guatemalan Central Highlands: D.C.: World Bank. The Impact of Nontraditional Agricultural Pefia, Christine, Patrick Webb, and Lawrence Haddad. Exports." Ph.D. diss. University of Wisconsin, 1996. Women's Advancement through Agricultural Madison. Change: A Review of Donor Experience. Food Kennedy, Eileen. 1991. "Income Sources of the Rural Consumption and Nutrition Division Discussion Poor in Southwestern Kenya." In J. von Braun and Paper 10. Wlashington D.C.: International Food R. Pandya-Lorch, eds., "Income Sources of Policy Research Institute. Malnourished People in Rural Areas: Microlevel Pitt, Mark M., ard Shahidur R. Khandker. 1995. "The Information and Policy Implications." Working Impact of Group-Based Credit Programs on Poor Paper on Commercialization of Agriculture and Households in Bangladesh: Does the Gender of Nutrition 5. International Food Policy Research Participants Matter?" Department of Economics, Institute, Washington, D.C. Processed. Brown University, Providence, R.I. Processed. Lastarria-Cornhiel, Susana. 1997. "Impact of . 1996. Hrousehold and Intrahousehold Impact Privatization on Gender and Property Rights in of the Grameen Bank and Similar Targeted Credit Africa." World Development 25(8):1317-33. Programs in Bangladesh. Discussion Paper 320. Lundberg, Shelly J., Robert A. Pollak, and Terence J. Washington, D.C.: World Bank. Wales. 1997. "Do Husbands and Wives Pool Their Place, Frank, and Peter Hazell. 1993. "Productivity Resources? Evidence from the United Kingdom Effects of Indigenous Land Tenure Systems in Child Benefit." Journal of Human Resources Sub-Saharar Africa." American Journal of 32(3):462-80. Agricultural Economics 75(1):10-19. McGuire, Judith S., and Barry M. Popkin. 1990. Quisumbing, Agnes R. 1995. Gender Differences in Helping Women Improve Nutrition in the Agricultural Productivity: A Survey of Empirical Developing World: Beating the Zero Sum Game. Evidence. Food Consumption and Nutrition Technical Paper 114. Washington, D.C.: World Division Discussion Paper 5. Washington, D.C.: Bank. International Food Policy Research Institute. Mehra, Rekha, ed. 1996. Taking Women into Account: . 1996. "Male-Female Differences in Lessons Learned from NGO Project Experiences. Agricultural Productivity: Methodological Issues Washington, D.C.: International Center for and Empirical Evidence." World Development Research on Women. 24(10):1579--95. Meinzen-Dick, Ruth S., Lynn R. Brown, Hilary S. Rocheleau, Diane, and Davids Edmunds. 1997. Feldstein, and Agnes R. Quisumbing. 1997. "Women, Men, and Trees: Gender, Power, and "Gender, Property Rights, and Natural Resources." Property in Forest and Agrarian Landscapes." World Development 25(8):1299-302. World Development 25(8):1351-71. Moock, Peter. 1976. "The Efficiency of Women as Rogers, Beatrice L., Ann Swindale, and P. Ohri- Farm Managers: Kenya." American Journal of Vachaspati. 1996. "Determinants of Household 202 The Importance of Gender Issues for Environmentally and Socially Sustainable Rural Development Food Security in Honduras." Department of Tisch, S., and Thelma Paris. 1993. "Labor Nutrition, Tufts University, Bedford, Mass. Substitution in Philippine Rice Farming Systems: Processed. An Analysis of Gender." International Rice Safilios-Rothschild, Constantina. 1985. "The Research Institute, Manila. Processed. Persistence of Women's Invisibility in Agriculture: Udry, Christopher. 1996. "Gender, Agricultural Theoretical and Policy Lessons from Lesotho and Production, and the Theory of the Household." Sierra Leone." Economic Development and Journal of Political Economy 104(5):1010-46. Cultural Change 33(2, January):299-317. van Koppen, Barbara, and Simeen Mahmud. 1995. Saito, Katrine, Hailu Mekonnen, and Daphne "Women and Water-Pumps in Bangladesh: The Spurling. 1994. Raising the Productivity of Women Impact of Participation in Irrigation Groups on Farmers in Sub-Saharan Africa. Discussion Paper Women's Status." Department of Irrigation and 230. Washington, D.C.: World Bank. Soil and Water Conservation, Wageningen Saito, Katrine, and Daphne Spurling. 1992. Agricultural University, Wageningen, The Developing Agricultural Extension for Women Netherlands. Processed. Farmers. Discussion Paper 156. Washington, von Braun, Joachim, Detlev Puetz, and Patrick Webb. D.C.: World Bank. 1989. Irrigation Technology and Commercialization Seager, Juri, and Ann Olson. 1986. Women in the of Rice in The Gambia: Effects on Income and World: An International Atlas. New York: Simon Nutrition. Research Report 75. Washington, D.C.: and Schuster. International Food Policy Research Institute. Subbarao, Kalanidhi, and Laura Raney. 1996. "Social von Braun, Joachim, and Patrick Webb. 1989. "The Gains from Female Education: A Cross-National Impact of New Crop Technology on the Study." Economic Development and Cultural Agricultural Division of Labor in a West African Change 44(1):105-28. Setting." Economic Development and Cultural Thomas, Duncan. 1994. "Like Father, Like Son, or Change 37(20):513-34. Like Mother, Like Daughter: Parental Education Wang, M. 1996. "Gender Differences in and Child Health." Journal of Human Resources Intrahousehold Resource Allocations: An 29(4):950-88. Empirical Analysis of Child Health in Zambia." Thomas, Duncan, and Chien-Liang Chen. 1994. Department of Economics, George Washington "Income Shares and Shares of Income." Labor and University, Washington, D.C. Processed. Population Working Paper 94-08. Rand Ward-Batts, Jennifer. 1996. "Out of the Wallet and Corporation, Santa Monica, Calif. Processed. into the Purse: Does Income Control Really Affect Thomas-Slayter, Barbara P., A. Lee Esser, and M. Intrahousehold Resource Allocation?" Department Dale Shields. 1993. "Tools of Gender Analysis. A of Economics, University of Washington, Seattle. Guide to Field Methods for Bringing Gender into Processed. Sustainable Resource Management." ECOGEN Zwarteveen, M. 1997. "Water: From Basic Need to Research Project, International Development Commodity: A Discussion on Gender and Water Program, Clark University, Worcester, Mass. Rights in the Context of Irrigation." World Processed. Development 25(8):1335-49. Part III Technical Issues and Perspectives In parts I and II we stressed the need to put in place consider "moved soil" as "lost soil," even though appropriate policy and institutional frameworks that much of it may be deposited on the same or other agri- are conducive to sustainable rural development. One cultural land. Soil degradation is predicted to have of the ways in which these frameworks function is by severe effects on agricultural productivity, yet evi- creating the right incentives for farmers to innovate in dence on the magnitude of these effects is hard to their production technologies. Here in part III we find. In fact, in many cases declines in productivity focus mainly on technical issues because sustainable are claimed with little or no supporting evidence at rural development must also be technically sound. We all. cover the following issues: soil conservation, main- Degradation can be slowed or arrested by a large streaming of biodiversity, integrated pest manage- range of options, including cultural practices such as ment and other strategies to reduce pesticide use, eco- contour plowing and minimum tillage, vegetative nomic and environmental aspects of irrigation and practices such as grass strips, strip cropping, and veg- drainage, livestock-environment interactions, inap- etative barriers, and mechanical measures such as ter- propriate deforestation, and agroforestry. races and cutoff drains. Adoption of any of these tech- Chapter 16 by Lutz and others seeks to shed niques can be costly, either directly in investment empirical light on the question of soil conservation. requirements or indirectly in production forgone, and Soil degradation represents a reduction in the land's some measures are better suited to some conditions actual or potential uses. Many cultivation practices than to others. The critical question facing farmers, tend to degrade soil over time. For example, cultiva- and society as a whole, is whether the benefits of a tion can expose soil to water and wind erosion, repeat- given conservation measure or set of measures are ed tillage can weaken soil structure, crop production sufficient to make the costs worth bearing. can remove nutrients, and use of machinery can com- The problem of soil degradation and conservation pact the soil. Soil degradation, in turn, affects produc- can be examined from two perspectives. The first is tivity. As soil is degraded, crop yields decline or input that of society as a whole. From this perspective, all levels (and hence costs) rise in an effort to keep or the costs and benefits of a given activity must be con- restore productivity. sidered, including off-site effects. From the second Despite long-standing concern about these prob- perspective, that of the farmers themselves, only the lems, surprisingly little hard evidence exists on their costs and benefits that actually accrue to the agent magnitude. The degradation figures quoted in the lit- making the decisions about resource use are consid- erature are often extrapolated from very limited data ered. and may exaggerate the problem because they often Lutz and others, in chapter 16, examine the returns 203 204 Part III. Technical Issues to investment in conservation measures mainly from the soil's physical structure. the farmers' point of view for two reasons. First, deci- Whether conservation measures are profitable for sions about land use are ultimately made by the farm- the farmers is an empirical and site-specific issue. ers themselves in light of their own objectives, pro- Returns to conservation depend on the specific duction possibilities, and constraints. Understanding agroecological conditions faced, on the technologies the incentives facing individual farmers is necessary, used, and on the prices of inputs used and outputs pro- therefore, if patterns of resource use are to be under- duced. Generally, the farmers' decision to invest in stood and appropriate responses to problems formu- conservation is based on normal considerations of lated. Second, land use problems generally depend benefit and cost: farmers tend to adopt conservation heavily on site-specific biophysical characteristics, measures when it is in their interest to do so, unless which can vary significantly even within small areas. some constraint is present. Cases in which returns to A farm-level approach also places the emphasis firm- conservation are low or negative correlate well with ly on the effects that degradation has on farm produc- low adoption rates. Profitability of conservation prac- tivity. In developing countries, where substantial tices is a necessary, but not always a sufficient, condi- numbers of people depend directly on agricultural tion for their adoption. Factors other than strict cost- production, the effect of degradation on yields is often benefit considerations also play a role. Institutional critical. This is not to belittle the importance, in some and motivational issues must be considered together situations, of off-farm effects, such as siltation of with the results of the cost-benefit analysis. reservoirs and waterways. Advocates of soil conservation often argue that In making their decisions about land use, farm subsidies are indispensable to induce farmers to adopt households need to consider both the agroecological conservation measures. But such statements assume and the economic characteristics of the environment that conservation is inherently desirable whether or in which they operate. The household's problem can not there is concrete evidence that the benefits out- be summarized as one of maximizing the present weigh the costs. Lutz and others, in chapter 16, show value of the stream of expected net returns to agricul- that this may be far from the case; frequently, the ben- tural production. With regard to the adoption of con- efits of specific conservation techniques (such as servation measures, the issue is whether returns under mechanical structures) do not justify their costs. the optimal path of the new, more conserving system Unless there are important off-site effects or the price are sufficiently greater than returns under the path of signals received by farmers are significantly distorted, the current, more degrading system to justify the cost subsidies to induce adoption will not enhance eco- of switching. nomic efficiency. Whatever their justification, the use The estimated productivity losses of soil erosion of subsidies encounters several difficulties. First, the and degradation vary considerably, as do the estimat- divergence between social and private returns to con- ed effects of conservation practices on yields. In some servation must be established, so that intervention can instances, yields are expected to recover once conser- be directed where it will be most effective. Second, it vation measures are established-partly because soil is difficult to design appropriate incentive structures regenerates once the processes of degradation are so that social objectives are met. Third, subsidization halted, partly because fertilizers are used more effi- schemes often create perverse incentives for farmers. ciently, and partly because improved cultivation prac- One important way for governments to help is to tices are sometimes introduced together with conser- make sure that constraints such as insecure tenure do vation. Elsewhere, conservation measures might slow not prevent farmers from adopting conservation mea- but not halt the decline in yield. sures. Also, governments already conduct research on Introducing conservation measures often has an soil conservation and provide, through extension ser- adverse effect on production because some of the area vices, some assistance to farmers who undertake con- cultivated is turned over to use as diversion ditches, servation work. However, research in experiment sta- terraces, or hedges. Physical structures, in particular, tions tends to favor technical efficiency (including usually reduce the area available for cultivation more structural measures such as terraces) over cost-effec- than 10 percent. But conservation measures can also tiveness. Further, government extension work is often have beneficial impacts. Besides reducing soil loss ineffective. In many cases, nongovernmental organi- and the rate of decline in yield, they can encourage the zations NGOS, such as Vecinos Mundiales in Central retention of moisture and stimulate improvements in America, have proven to be more effective than gov- Part III. Technical Issues 205 emnment at presenting the range of options to farmers potentially dangerous loss of biodiversity. But wild and delivering related technical assistance. Given the species are essential for agricultural improvement wide variety of conditions that farmers face, govern- because they are the source of new economic plants ment extension services should also provide, explain, and animals and. provide important services such as and demonstrate to farmers the corresponding variety pollination and pest control. of options available rather than, as has often happened Possible remedial measures can be adopted to in the past, pushing broadly for the adoption of spe- address the loss of biodiversity associated with agri- cific techniques. And it may be innovative as well as cultural development: (a) minimizing habitat frag- effective for governments to decentralize decision- mentation by providing wildlife corridors along making and channel budgetary resources for soil con- "bridges" of natural habitat, (b) shifting to integrated servation to the local level. This would allow commu- pest management (IPM) strategies, such as rotating nities to participate and contract assistance from those crops and relying on biocontrol agents to check crop from whom the greatest contributions can be expected. and livestock pests, and (c) eliminating fiscal and reg- Chapter 14 discussed biodiversity conservation ulatory measures that promote homogeneity in crop with a focus on protected areas. But much of the and livestock production. world's biodiversity exists in human-managed or A balanced conservation strategy includes in situ -modified systems. That is the focus of Srivastava and conservation (for example, maintaining animal and others in chapter 17. Biodiversity has two important plant genetic resources in places where they occur dimensions: the genetic variation within species and naturally) as well as ex situ conservation (for exam- populations and the preservation of habitat. The sig- ple, maintaining them in seed or field gene banks). In nificance of variation within a species, although poor- situ conservation of crops and livestock can be sup- ly appreciated, is critical, particularly for agriculture. ported by (a) emphasizing the safeguarding of plants The continued productivity of existing crops and live- and animals for the future improvement of agriculture stock hinges in large part on harnessing the genetic as a new dimension of existing wilderness parks and variation found within each species. Habitat conser- biological reserves, (b) creating world heritage sites vation seeks to safeguard natural habitats for wild for genes for agricultural development, (c) integrating species and populations and to manage habitats that agrobiodiversity in ecotourism where appropriate, (d) have been modified for human use, such as farmland. helping to find markets for lesser-known crops and Agrobiodiversity is found in habitats that have been local varieties under the motto "use it or lose it," and modified for crop and livestock production. (e) helping to find ways for livestock owners to gen- Agrobiodiversity includes all plants and animals that erate more revenues from threatened breeds. contribute directly or indirectly to the raising of crops Better in situ conservation can also be supported and livestock. by rapid agrobiodiversity assessment teams. Many Although it is conceptually useful to differentiate development organizations, including the World agrobiodiversity from the larger array of species and habi- Bank, have procedures for screening projects for their tats, the boundaries between biodiversity and agrobio- environmental impacts before they are approved. For diversity are not clear-cut. All of biodiversity is poten- the most part, such screening focuses on off-site tially of use to agriculture, particularly with the impacts of agricultural development and includes an emerging biotechnologies. Agriculture is highly assessment as to whether the project is likely to lead dynamic, and the interface between wild and domes- to the loss of fo:rest or other "natural" habitats. But ticated plants and animals is constantly shifting. This surveys of agrcbiodiversity may also be needed fact underscores the importance of conserving as before a development project is implemented. much biodiversity as feasible for agricultural devel- An assessment of biodiversity before a project is opment in the future. approved and implemented would encompass bio- How agriculture can be intensified without damag- diversity in relatively undisturbed habitats and agro- ing biodiversity is a critical question for rural devel- biodiversity. Rapid assessment methodologies have opment. Environmentally inappropriate intensifica- been tried and tested more fully with the first catego- tion of agriculture has led to eutrophication of lakes ry. Much work remains to be done in designing and estuaries, loss of soil micro-organisms, accelerat- methodologies for adequately assessing agrobiodiver- ed soil erosion, contamination of groundwater, and sity. Rapid assessments of agrobiodiversity should draining of wetlands. All of these activities trigger a document the diversity of crops cultivated by local 206 Part III. Technical Issues farmers, the varietal diversity of each crop, the num- one-has increased the use of agrochemicals. ber of livestock species raised, the number of breeds Intensive monoculture systems using high-yielding per livestock species, and the effect of the land use cereal varieties have resulted in an environment con- practice on soil biodiversity as an indicator of soil ducive to the buildup and infestation of pests, and the "health" (Srivastava and others, chapter 17). consequent use of pesticides has disrupted the natural The new vision for agricultural research adopts a balance between pest and predator. The rapid growth holistic approach that is more sensitive to environ- in pesticide use can also be attributed to the misinfor- mental concerns, while still addressing the need to mation and risk aversion of both policymakers and boost the yields and incomes of rural producers and farmers as well as to the profit motive, including sales caretakers of the land. It includes, but is not restricted techniques of the chemical industry. Further, indis- to, (a) integrated pest management, (b) a participatory criminate and injudicious use of agrochemnicals often approach with farmers, (c) better use of farmer knowl- has been encouraged by inappropriate or inadequate edge, (d) greater support for research, development, and pesticide policies in developing countries. Finally, dissemination of lesser-known crops and animals, (e) bilateral and multilateral aid for pesticide use, provid- support for research on new crops and livestock, (f) ed through grants or subsidized prices, has permitted greater sensitivity to the value of a mosaic of land governments to distribute pesticides to farmers at low uses, (g) greater diversity of habitats within land use or no cost. systems, (h) greater reliance on recycling of organic Indiscriminate use of pesticides in the production matter, (i) shift of the research focus from individual of various crops has often impaired health due to traits to lifetime and herd productivity characteristics, direct or indirect exposure to hazardous chemicals, (j) determination of the critical number of breeds for contaminated ground and surface waters through conservation purposes, and (k) an effort to learn more runoff and seepage, transmitted pesticide residues about the genetic components of adaptation in live- through the food chain, increased the resistance of stock. pest populations to pesticides, thereby reducing their The notion of a new research paradigm has impli- efficacy and causing pest outbreaks, and reduced the cations for institutional development and the explo- population of beneficial insects ("good bugs" like par- ration of new ways of doing business. Innovative asites and predators), thereby reducing the effective- institutional arrangements would include more effec- ness of pest control strategies that attempt to mini- tive partnerships among agricultural research centers, mize pesticide use. NGOS, growers associations, private companies A reappraisal of the role of pesticides in agricul- involved in the imanufacture and sale of agricultural ture has been under way for some time. But only technologies, universities and agricultural extension recently has the topic shifted from technological solu- agencies, and development lending institutions. To tions (for example, better protection of workers and some degree, all of these partnerships are being consumers, classification and phaseout of the most explored and tested. dangerous chemicals) to the agricultural policy agen- Generating the gains in agricultural productivity das of government and donors. This has been associ- necessary to secure food availability and livelihoods ated with a wider shift in focus from agricultural pro- in the developing world over the coming decades duction to environmentally sustainable management requires an approach in which the intensification of of production systems in which IPM is a critical ele- agricultural systems is consistent with the conserva- ment. Integrated pest management is a key compo- tion of the natural resource base. This approach nent of integrated farming practices that are based on requires less reliance on the intensive use of external an understanding of ecology and the interaction inputs and greater dependence on management skills between crops or animals and their pests, as well as an and location-specific knowledge of agroecosystems. understanding of the environments in which pests Integrated pest management constitutes one such operate. Yet, despite notable successes in some coun- approach and is critical to sustainable rural develop- tries, widespread implementation of IPM remains an ment (Schillhorn and others, chapter 18; Pingali and elusive goal in most parts of the world. Gerpacio, chapter 19). Host-plant resistance-the plant's ability to resist Agricultural intensification-the movement from high levels of pest infestation-has been the corner- an extensive to an intensive production system or stone of the scientific strategy for developing sustain- from a subsistence production system to a commercial able pest control systems for cereal crops. Significant Part III, Technical Issues 207 advances have been made in the last three decades in insecticides. the development and dissemination of crop varieties IPM is knowledge-based; it p:resents by far the most with resistance to the major cereal pests. Much of the difficult challenge to traditional, small-scale farmers advance has come through the use of conventional in developing countries who are making the transition breeding approaches, although substantial gains in the to scientific farming. IPM requires farmers to grasp development of resistance could come through the use complex sets of data that are often anything but self- of modern biotechnology tools. The introduction of evident, unitary, and standardized or amenable to varieties with host-plant resistance has dramatically trial-and-error learning. The institutional and eco- reduced the need for insecticides in rice and maize noniic structure in the rural sector of developing and the need for fungicides in wheat. economies also requires some policy intervention to Given the extent of breeding efforts in producing reconcile long-term societal goals with short-term pest-resistant crop cultivars, and evidence of insignif- individual objectives in pest control. Promoting sus- icant productivity benefits of pesticide use, why are tainable pest management within an IPM framework high levels of agrochemicals still used in crop pro- requires improved research and extension linkages, duction? First, the dissemination of crop varieties effective farmer training methods, community action, resistant to pest pressures has not been accompanied and an undistorted price structure. by extension messages on the reduced need for pes- The IPM concept is holistic: it requires farmers to take ticides. Second, farmers' pest control decisions, sci- a systems view of the farm enterprise and to under- entists' research priorities, and policymakers' pre- stand the interlinkages among various components of scriptions, which are all based largely on perceived the system. Therefore, farmer training in sustainable pest-related losses in yield (in turn often not related to pest management is an essential component of a strat- actual losses) have led to the promotion of pesticide egy to achieve minimum use of insecticides. The use. Consequently, high and injudicious applications eventual goal is to build farmers' capacity to identify of broad-spectrum pesticides have continued as and solve problems based on a thorough understand- before, causing the breakdown in varietal resistance. ing of field ecology. The experience of the Food and New varieties generated to replace cultivars with Agriculture Organization's farmer field schools has resistance breakdown have subsequently been over- shown that trained farmers use significantly lower come through further biotype changes in pest popula- levels of pesticicles than untrained farmers. They are tions. This breeding treadmill could only be overcome also more likely to experiment with other components through dramatic changes in crop management prac- of sustainable production systems, such as improved tices, especially in the use of pesticides. fertilizer management and more efficient water man- Under low levels of pest infestation, natural con- agement. trol is the economically dominant strategy for pest As yet, several unresolved research and policy management. (This is true for insects and diseases for issues are related to intensive farmner training in IPM. rice, wheat, and maize.) Natural control relies on The most imporLant one is related to the costs and predator populations to control pest infestations under benefits of farmer training. The costs of training mil- normal circumstances, when pest-resistant varieties lions of farmers are enormous and need to be justified are used. Pesticides may have to be used as a last clearly in terms of farm-level benefits, both reduced resort in the rare instance of high infestations. Natural expenditure on insecticides and reduced social costs control does not imply doing nothing; rather, it is of pesticide use. Attention also should be focused on based on the premise of in-depth farmer knowledge of the opportunities for reducing the overall costs of the pest-predator ecology and frequent monitoring of training. There are essentially two options, which are field conditions by the farmer. In this regard, natural not mutually exclusive, for reducing the costs of train- control can be considered the ultimate goal of an IPM ing. The first is to train a core group of farmers with- program, and farmers who are well versed in IPM tech- in a geopolitical init, such as a municipality, and then niques would converge toward it (Pingali and to rely on farmer-to-farmer training for disseminating Gerpacio, chapter 19). This implies a paradigm shift the IPM message to a wider group of farmers. The sec- in the traditional IPM strategies from when best to ond is to condense the complex message into simple apply to when not to apply. Therefore, continued rules that are easy for the farmer to implement. An investments in IPM training would be essential for the example, from rice, of such a rule is "Do not spray successful adoption of reduced or minimum use of insecticides against leaf-feeding insects for the first 208 Part III. Technical Issues 40 days of crop growth." The rule is based on detailed use, to account for negative externalities and short- pest ecology studies showing that the predominant term subsidies to account for positive externalities in insect pests during the first 40 days of crop growth are the use of IPM, (d) orientation of research and tech- leaf-feeding insects and that even very high levels of nology policies to generate a steady supply of relevant infestations of these insects rarely lead to any loss in pest management information and technologies, yield (Pingali and Gerpacio, chapter 19). including adequate budget allocations for research, The "no spray for 40 days" rule is an example of a extension, and training, and (e) signing of and adher- simple message, distilled from in-depth scientific ence to international agreements and conventions investigation, that can be transmitted easily to farm- (Schillhom and others, chapter 18). ers. A small set of such rules, which are mutually con- Over the past three decades, the dramatic spread of sistent, would go a long way toward improving insect irrigated agriculture was mainly responsible for pest management in tropical crop production. Simple increases in food production that kept pace with rules ought not to be seen as a substitute for farmer growth in population. This was achieved in part training but rather as a strong complement to a train- through the large-scale development of new water ing program. Simple messages can be transmitted resources and construction of new irrigation capacity rapidly even while investments are made in trying to (Chakravorty, chapter 20). The International reach all farmers with intensive training. Irrigation Management Institute has estimated that to The success of pest management programs meet future demand, irrigated agriculture may need to depends on, among other things, collective organiza- deliver output increases of more than 3.5 percent tion against infestations of migratory pests. The annually-a daunting task by any measure of perfor- actions of individual farmers in the management of mance. This growth must come primarily from their pest problems could have detrimental effects on increases in irrigation efficiency and not from addi- the community as a whole. In this regard, manage- tions to new irrigation capacity. Moreover, as is ment of pests could be viewed as a common property becoming increasingly common in the industrial problem and dealt with through effective collective countries and even in many developing countries, ris- action. In Asian agriculture, where farms are uniform- ing demands for residential use of water and rising ly small and farmers are nearly homogeneous, collec- environmental concerns are beginning to put serious tive action for pest control seems quite attractive. limitations on new water development projects. Synchronization of planting is the single most impor- Producing more food with less water-placing a tant community action strategy in pest management. smaller financial burden on the taxpayer and reducing Even with a well-established IPM program, pesti- the environmental costs-will be the major challenge cides may have to be kept as a technology of last for irrigation in the twenty-first century. resort. Essentially, the idea of pesticide use in IPM is Despite major investments in irrigation during the to spray only when imperative, using the smallest past four decades-the World Bank alone has invest- amount possible to do the job. However, agrochemi- ed more than $31 billion and leveraged an additional cals continue to have a significant impact on produc- $53 billion from partner countries and cofinanciers- tivity in the management of weeds; herbicides will irrigation projects around the world are in trouble for continue to be the preferred alternative in the foresee- a multitude of reasons. Ex post project benefits have able future, even when the health costs of herbicides often been far lower than projected ex ante returns are explicitly taken into account. and full-cost pricing of water is rare. Most projects A number of policy and regulatory instruments are only recoup some fraction of their operation and available to governments to encourage environmen- maintenance (o&M) costs. Low tax collections and poor tally sound and economically rational pest manage- o&M of project structures have further led to high ment practices. The more important of these instru- rates of water loss through seepage and percolation, ments are (a) development of a system that increases and inadequate investments in drainage have rendered the awareness of policymakers, consumers, and pro- large tracts of prime agricultural land unusable ducers of the hazards of pesticide use, (b) develop- because of salinity and waterlogging. The low price of ment of a regulatory framework to ensure appropriate irrigation water, which is often unrelated to water use, and safe production, distribution, and use of pesti- has led farmers to withdraw too much, giving rise to cides, (c) introduction of appropriate economic incen- widespread salinity and waterlogging. Investments in tives, including taxes and special levies on pesticide the irrigation sector in the developing countries have Part III. Technical Issues 209 relied too much on bureaucratic federal and state native uses such as to meet urban needs. However, in agencies that have paid little attention to the econom- many developing countries, the transition from the ic pricing of irrigation services, the creation of reli- short to the long run is of profound importance and is able delivery systems, and the participation of users in beset by serious problems relating to the redistribu- the o&M of projects. Adverse health and environmen- tion of rents accruing to beneficiaries. Most irrigation tal effects of water use have received little attention, systems in developing countries are heavily subsi- partly because of a lack of coordination and interac- dized by the taKpayer, with substantial rents from tion between water and public health agencies. water accruing to large landowners and the financial- These problems call for a new approach that is ly better-off farmers who are usually located in the more comprehensive and integrates the elements of upper reaches of the distribution system. irrigation policy, which include the economic and Many governments, policymakers, and environ- environmental effects and the associated intersectoral mental groups are beginning to realize that a more linkages (Chakravorty, chapter 20). This approach efficient irrigation system means more water for must be built on pricing of resources that reflects everyone. This has induced changes in the way water opportunity costs whenever possible and creation of tights have been allocated, and restrictions to trading delivery infrastructure that promotes efficient alloca- in water have often been removed. In developing tion and, if feasible, trading of the resource among countries, govemnments are now motivated and will- users. ing to undertake major institutional reforms as a pre- Optimal resource allocation is best done by means condition to initiating new multilaterally funded irri- of a smoothly functioning market system. In the case gation projects. of water, these conditions are not met on several New "holistic" thinking is also reflected in the cur- counts. For example, the mobility of water makes it rent emphasis on "basinwide" management promoted difficult to measure and hold, thereby making the by agencies such as the International Irrigation establishment of property rights difficult. Economies Management Institute (Chakravorty, chapter 20). of scale in water generation and conveyance often These agencies argue that water efficiency must be create conditions for a natural monopoly. The joint measured not within a project but over an entire basin product nature of water, which enables multiple users so as to include the reuse of drainage water from seep- to access a given stock of water sequentially, and the age and percolation. This suggests that policies such externalities created as a result of the degradation of as the adoption of sprinkler irrigation within an irri- quality through use make the creation and functioning gation project may improve the efficiency of water of water markets extremely difficult. For these rea- use within the project but reduce the availability of sons, the water sector in both industrial and develop- water elsewhere in the basin. Similarly, salt and ing countries has been characterized by public invest- chemical pollution of water through the application of ment in the construction of irrigation projects, state inputs may adversely affect agriculture downstream ownership and operation of facilities, and regulated of the project. These issues can only be handled if the distribution and allocation of water. basin is adopted as the relevant unit of analysis and The introduction of markets that allow trading the relevant externalities are internalized. among users would lead to a more efficient allocation Livestock can damage global natural resources in of water. Water markets function in different parts of a number of ways, but they can also contribute to the world, often without help from outside parties. environmental balance (Steinfeld and others, chapter However, they only work under very selective condi- 21). About 34 million square kilometers, or 26 percent tions: (a) the presence of effective water users associ- of the world's land area, are used for grazing live- ations or water authorities, (b) the removal of legal stock. In addition, 3 million square kilometers, or and institutional regulations that prohibit or limit about 21 percent of the world's arable area, are used water trading, and (c) the creation of infrastructure for cereal production for livestock feed. Livestock (canals) and control systems for the transfer and mea- produce 13 billion tons of waste per year. A large part surement of water. of this is recycled, but where animal concentrations In the long run, properly functioning water mar- are high, waste poses an enormous environmental kets may allocate scarce supplies to their most valued hazard. Livestock grazing can affect the water balance use. They may also improve the efficiency of water in in certain areas. Water is needed to produce fodder the agricultural sector and thus release water for alter- and feed concentrate, to provide drinking water for 210 Part III. Technical Issues animals, and to drain surplus waste and chemicals. Middle Eastern countries, and, in the semiarid Livestock interact directly and indirectly with biodi- zones, overseeding or planting of adapted fodder versity. may regenerate the vegetation. The introduction of Policies need to be designed to correct the envi- a multispecies grazing pattern will often encourage ronmental effects of livestock production. These poli- better use of vegetation and may have positive cies should address the underlying causes of environ- effects on plant and animal biodiversity. mental degradation and must be flexible, site-specific, Livestock, mainly through their input function and well targeted. Instruments to enhance positive and within a mixed crop-livestock system, enhance the mitigate negative environmental impacts include pric- main natural resource-land. Animal manure and ing, regulations, and institutional development. The traction make the land more productive than would collective purpose of these instruments is to establish be the case in their absence. Thus, all technologies feedback mechanisms to ensure that the use of live- that reduce nutrient losses from manure, and stock is consistent with overall social objectives. improve the efficiency of their application, Within an enabling policy framework, a wide enhance land productivity. range of technologies are available, while others still * Technologies that save natural resources by allow- need to be developed. Technologies can be grouped ing farmers to get more revenue from the same into four different sets (Steinfeld and others, chapter iesource or to get the same from less. The live- 21). Although there is some overlap, these categories stock sector possesses and continues to develop help to provide a good picture of livestock-resource technologies that increase the efficiency of natural interactions: resource use. In particular, these technologies tar- Technologies that reduce the environmental dam- get feed conversion because feed typically age by alleviating the direct pressure on natural accounts for 60-70 percent of production costs. resources or by reducing the pollution load Better feed conversion saves land used for its pro- through modifying the chemical or physical char- duction, while reducing the animals' waste load. acteristics of products. In the arid and semiarid Technologies also provide solutions to saving and grazing areas, careful water development can help sparing other natural resources such as water and bio- to prevent environmental damage. Investments in diversity. market technology may also reduce environmental * Technologies that turn waste into products by clos- pressure by encouraging greater off-take. In addi- ing cycles. Historically, livestock were kept tion, new and more benign methods are available because they used resources for which there was to control diseases. Environmental damage of no alternative use. This explains why efficiency intensive systems can be reduced significantly by per animal was not, and in many low-input sys- focusing on emissions from manure by, for exam- tems still is not, a major concern. The conversion ple, improving collection and storage. The main of organic waste into livestock products, although focus should be on reducing nitrogen losses, most associated with livestock waste, reduces the envi- of which are in the form of ammonia from manure. ronmental hazards associated with crop and Nutrient losses during and after application of agroindustrial waste. Also, livestock consume manure on soils can be significantly reduced by food wastes and increasingly so, as urban agglom- injecting or applying manure into the subsoil. eration and changing eating habits offer a window Better timing of application in response to crop of opportunity for the collection of food waste requirements avoids further losses and enhances from catering units to be recycled as feed. Large the nutritive value of manure. In tanneries, dairies, amounts of straw, otherwise burned on the fields or and slaughterhouses, anaerobic systems can purify slowly decomposing with little nutritional benefit wastewater and reduce by half the biological oxy- to the crops, may be turned into quality feed, for gen demand (BOD), while more sophisticated example through urea treatment. The cycles of anaerobic systems reach 90 percent BOD purifica- matter can also be closed by using livestock waste tion. as feed, energy, or fertilizer. Technologies that enhance natural resources by Novel concepts are being developed to integrate making them more productive or richer. For graz- crop and livestock production in a farming area rather ing systems in arid zones, "deferred grazing," than on a mixed farm (Steinfeld and others, chapter which has been a traditional practice in many 21). This method of area-wide crop-livestock integra- Part III. Technical Issues 211 tion allows individual enterprises to operate separate- sector, will depend on technology to substitute for ly while linking the flows of energy and organic and natural resources. This trend toward knowledge- mineral matter through markets and regulations. This intensive systems is already widely observed. allows for highest efficiencies at the enterprise level, Expansion ofr livestock production has often been while maximizing social benefits. Thus, the motto for named as one of the main causes of tropical defor- the more densely populated parts of the developing estation. Contemporary concern about deforestation world is to intensify, but not concentrate, animal pro- focuses on tropical countries because that is where the duction. A policy framework is needed to organize majority of forest cover is being removed (Kaimowitz crop and livestock production in such a way. and others, chapter 22). In the period 1980-90, 137.3 Grazing systems will remain a source of exten- million hectares of tropical forests were cleared, about sively produced animal products. To some extent, 7.2 percent of the total that existed in 1980. Any these systems can intensify production by incorporat- claims regarding the magnitude of global deforesta- ing new technologies, especially in the higher-poten- tion must be taken with caution, however, because tial areas (subhumid and highland areas). Often this there are serious problems with the data and defini- can be facilitated by stronger organizations, local tions used. empowerment, and regulation of access to resources. Deforestation often implies: In the semiarid and subhumid tropics, where there is * The loss of -livelihoods for forest-dependent peo- potential for mixed farming, policies need to facilitate ple, many of whom may not wish or be able to find the transition of grazing systems into mixed-farming other sources of employment systems by integrating crops and livestock (manure * Decreasing stocks of fuelwood and nontimber for- management, animal draft, residue feeding, and fod- est products as well as of industrial timber der crops). * Greater soil erosion and river siltation Mixed-farming systems will see continued intensi- * Substantial loss of species and genes, in view of fication and important growth. Smallholder and fami- the high level of endemic biodiversity in tropical ly mixed farming will remain predominant for some forests time to come, with livestock based on crop by-prod- * Substantial emission of carbon dioxide, which ucts and surplus (Steinfeld and others, chapter 21). contributes to global warming Important productivity gains can be achieved by * Other types of local and regional climate change. enhancing the flow of nutrients and energy between The concern is not only with deforestation but also the two components. Livestock's role, in addition to with forest degradation, which can be defined as a production, is to enhance and substitute natural decrease of density or increase of disturbance in for- resources. The environmental and economic stability est classes. In the long run, this may be just as impor- of this system makes it the prime focus for the con- tant as deforestation itself. tinuing transfer and development of technology. The process of deforestation must be analyzed at Industrial systems in areas of high animal densities two levels: agents and causes. The agents of defor- will face the challenge of coping with higher produc- estation are the people who physically (or through tion costs as a result of more stringent regulations and decisions over their labor forces) convert forests to pollution levies. This will remove, in some cases, the nonforest uses: small farmers, plantation and estate competitive edge that industrial production has over owners, forest concessionaires, infrastructure con- land-based production. Potentially, this will also raise struction agencies, and so forth. In Latin America, the prices for livestock products, reduce demand, and most deforestation is associated with medium- and provide incentives for intensification. large-scale operations (resettlement schemes, large- As a result of the interaction between livestock scale cattle ranching, hydroelectric dams) and is char- production systems and natural resources, coupled acterized by tranisitions from closed forest to nonforest with factors such as market access, there are develop- land uses. In Africa, deforestation is related largely to ment opportunities as well as threats to sustainability. the expansion of small-scale farming and rural popu- A comprehensive perspective is needed to ensure an lation pressure (the growing number of smallholders), enabling policy framework in which to introduce associated with conversion from closed forest cover effective technologies (Steinfeld and others, chapter to short fallow farming. In Asia, deforestation is asso- 21). Technology remains a key component because ciated with boch relatively large operations (as in future development, including that of the livestock Latin America) and rural population pressure (as in 212 Part III. Technical Issues Africa). There are no simple, universal single-cause * Areas that have large numbers of forest-dependent explanations. Different agents coexist and closely people who show no inclination to abandon their interact in many countries, and their relative impor- existing livelihood strategies tance varies over time and between regions. Given * Areas that, by virtue of their rich potential and such complex relationships, there are no clear guilty comparative advantage for timber production, or innocent parties, and no one should expect neat, make this the most profitable use of land, even simple solutions. after the area has been logged for the first time Individuals and businesses deforest because it is * Fragile areas where the ecological cost of conver- their most profitable alternative. To get them not to sion resulting from "downstream" effects out- deforest in situations where forest clearing is inappropri- weighs any economic gain from nonforest land ate, deforestation must be made less profitable or other uses alternatives (either based on retaining forests or com- * Periurban areas where forests play a key role in pletely outside forest areas) must be made more prof- conserving aquifers, providing fuelwood, and sup- itable. This is, in fact, the main thesis of Kaimowitz porting recreational and tourism activities. and others (chapter 22). It is important to distinguish inappropriate from The "causes" of deforestation refer to the multiple appropriate deforestation for three reasons. First, it is factors that shape agents' actions and, in particular, often assumed (at least implicitly) that all tropical their decision to deforest. These include market forces deforestation is inappropriate, and this is not neces- (international price fluctuations of agroexport com- sarily the case. Second, to the extent that inappropri- modities), economic policies (currency devaluation), ate deforestation can be defined, the geographical legal or regulatory measures (a change in land tenure areas (and sometimes socioeconomic groups) can be laws), institutional factors (the decision to deploy specified that should be the targets of policy designed more forest rangers to a particular area), and political to reduce inappropriate deforestation. Third, by clearly decisions (a change in the way forest concessions are defining the areas appropriate for conversion, pres- allocated), among others. A key question is how cer- sure can conceivably be eased on forests where con- tain causes can be manipulated to influence the version is inappropriate. behavior of agents, so as to lessen the rate of inappro- Kaimowitz and others (chapter 22) offer a rudi- priate deforestation (Kaimowitz and others, chapter mentary conceptual framework for distinguishing 22). inappropriate from appropriate deforestation. It takes Some deforestation is inappropriate for two rea- into account three cross-cutting categories of valua- sons. First, deforestation generally causes negative tion: biophysical, economic, and political. Deci- externalities that generate costs to society that are not sionmakers must take all of them into account simul- reflected in existing prices and has long-term conse- taneously. quences that individual producers rarely consider. Deforestation can be made less profitable by Second, the relative importance of the negative exter- * Reducing the demand or prices for products pro- nalities tends to grow over time if an increasing pro- duced from newly cleared land portion of deforestation occurs in areas that have only * Increasing the unit costs and riskiness of activities marginal value for agriculture but sequester large associated with deforestation amounts of carbon, have fragile soils, or are high in * Eliminating speculative gains in land markets. biodiversity. Alternatives to deforestation can be made more Any decision regarding which deforestation is profitable by appropriate or not is ultimately political in nature and * Increasing the income stream to be obtained by cannot be justified on purely technical grounds. maintaining forests Nevertheless, forest clearing is more likely to be inap- * Reducing the costs of maintaining forests propriate when it involves the following types of * Increasing the opportunity costs of labor and capi- areas: tal that might otherwise be used in activities asso- * Areas that have little value for agriculture by ciated with deforestation (Kaimowitz and others, virtue of the quality of soil or gradient of land chapter 22). * Areas that have large amounts of biodiversity, par- All the policies analyzed are evaluated on the basis ticularly endemic biodiversity, that is poorly repre- of the following six criteria: effectiveness, ability to sented in existing protected areas be targeted, equity, political viability, direct cost, and Part III. Technical Issues 213 indirect cost. There is no perfect or generalizable pol- gal, difficult, and often unprofitable activities of tem- icy for reducing inappropriate deforestation. There are porary agriculture in forestlands. However, if the new no "first-best" options. Each national situation is dif- land use is profitable (growing cocoa, coffee, cinna- ferent, much uncertainty remains about key cause- mon, rubber, or fruit trees, or even timber trees like and-effect relations, and there are usually tradeoffs eucalyptus, teak, or gmelina) and if the potential cap- among effectiveness, ability to be targeted, political ital gains from "capturing" some real estate from the viability, and direct and indirect costs of policies. government forests are high, it might be very difficult Most of the policy instruments discussed are very to slow the rate of forest conversion by such people. blunt instruments in regards to stopping deforesta- Thus the principal reforms that could reduce inap- tion-governments will be forced to choose from a propriate tropical deforestation are likely to be a com- mix of measures specifically crafted to local condi- bination of the following government policies: tions. * Eliminate subsidies to agricultural and pastoral Logging will usually lead to increased forest con- industries that encourage deforestation version by an influx of migrants, if the following con- * Eliminate legal and institutional incentives or ditions apply simultaneously: requirements to clear forests as a basis for gaining * Roads are constructed that open up new areas of recognized land tenure forests. * Reform forest industry concessions and licenses to * The use of nonforest land is much more profitable provide incentives for long-term sustainable man- than the retention of forests (in part due to policy agement distortions). * Develop innovative institutional arrangements for * Forest boundaries are poorly enforced by govern- devolving more decisionmaking authority and ment agencies (such as the forest service or the responsibility to those whose livelihoods and qual- national parks service), and given the institutional ity of life are directly linked to the extent and qual- or legal context, people expect that the land they ity of tropical forests occupy, claim, or "stake out" will eventually be * Encourage voluntary market differentiation by recognized, even legalized, by the government, for consumers that discriminates positively toward example, by creating an open-access "frontier." products that have been sustainably produced from * A large pool of unemployed or landless people, or forests with very low incomes and prospects, constitute * Facilitate the recognition of and compensation for potential migrants (whether moving spontaneously environmental services provided by forests and or sponsored by the state or private agoindustries). ensure that transfer payments are received by the The pace of colonization might be related to the persons making the decisions at the forest fron- difference between current incomes of potential tiers. migrants and the amount they expect to earn by Most of the major environmental problems in colonizing forest areas. developing counitries are due not to the pursuit of eco- This assessment, if correct, suggests that the nomic development, but rather to incorrect economic answer to the forest conversion issue is not necessari- policies: poorly defined property rights, underpricing ly to stop logging per se, or to stop logging in all new of resources, state allocations and subsidies, and areas, or to ban all new road construction in forest neglect of nonrmarketed social benefits. Instead of areas, but rather to reform those policies and institu- devising new policies to stop further resource and tions that at present make forest colonization seem environmental deterioration while promoting devel- more attractive than the potential migrants' current opment, one should first try to eliminate those (legal, activities. This might include the pull factors (to social, political, and institutional) factors that cause or reduce the profitability of illegally clearing forests or exacerbate the problems (a point made repeatedly in of speculating in land that was supposed to be kept as part I of this vo:Lume). forest) or the push factors (to increase the limited One response to deforestation has been to encour- livelihood options outside of forests). The evidence age agroforestry: a land use system in which trees, from the rapid economic growth of Asian economies shrubs, palms, and bamboos are cultivated on the before 1997 is that as employment and income same land as agricultural crops or livestock for eco- prospects outside the agricultural sector improved, nomic and environmental reasons. The system fewer people wanted to undertake the dangerous, ille- includes managing natural regrowth, seeding, plant- 214 Part III. Technical Issues ing, and maintaining trees as border plantings, and is profitable, smallholders will and should adopt interplanting them in agricultural crops, in woodlots, agroforestry incrementally and gradually because of in home gardens, or in other systems. Agroforestry's management and resource constraints; poorer farn- special characteristics are that it includes a large num- ers, in particular, are often hampered by limited land, ber of species, configurations, and management inten- labor, and capital resources and their need to ensure sities, has longer gestation than most agricultural food security and reduce risks. crops, and produces outputs with multiple uses. Local scarcity of wood products is, as might be Numerous projects have been established in the past expected, a key motivator in adopting nontraditional two decades throughout Central America and the agroforestry. Projects must begin by assessing the Caribbean to promote communal and individual tree scarcity of wood as well as the existence of local plantings and agroforestry systems (Current and oth- markets for products. Taungya, perennial intercrops, ers, chapter 23). The main emphasis used to be tree trees on contours, and tree lines have proved to be the planting for fuelwood. But many of those projects easiest systems to introduce. Results are mixed for were unsuccessful. Subsequently, the emphasis shift- alley cropping, home gardens, windbreaks, green ed from fuelwood production to the concept of mul- manuring, dispersed trees in cropland, and tree-pas- tipurpose tree species in agroforestry systems. The ture systems. Farmers are willing to invest in rehabil- establishment of multipurpose tree species on farm- itating their land where systems also produce prod- land can provide a wide range of benefits. Conflicts ucts or income, and they prefer less-intensive sys- do arise between the cultivation of trees and agricul- tems. tural crops, but some agroforestry systems allow The demonstration effect of fast-growing tree farmers to integrate trees into their farming systems, species on farms, and of benefits on demonstration with only small drawbacks for crop production or plots, has helped to expand agroforestry activities, even an increase in farm productivity overall. reducing the costs of extension and increasing its The shift in emphasis toward multipurpose tree effectiveness. Rather than offering standard designs, species was accompanied by a corresponding realiza- programs serve farmers best if they offer a broad tion that a project is more likely to succeed if project selection of species and systems from which to planners consult local communities about their per- choose those most suitable to their household's needs ceived needs and design projects to meet those needs and resources. Involving local people as paratechni- instead of imposing schemes that they may not con- cians is often a successful, low-cost approach to pro- sider a priority. moting technology. In addition to the estimated financial return, farm- Financial incentives and subsidies should be kept ers attach considerable importance to how an agro- to a minimum. Agroforestry technologies promoted forestry system fits into the overall farm production by extension should be financially profitable, and thus system and the existing land, labor, and capital con- adoptable, for the farmer without subsidies. The pos- straints (Current and others, chapter 23). Many agro- sible exceptions are time-limited financial incentives forestry systems are profitable to farmers under a for early adopters of unfamiliar technologies. In-kind, considerable range of economic conditions, and vari- material inputs encourage farmers to experiment with ous types of (low-intensity) traditional agroforestry and adopt agroforestry, but the experiences with food- are practiced in many areas. Even when agroforestry for-work incentives are mixed. 16 The Costs and Benefits of Soil Conservation in Central Amnerica and the Caribbean Ernst Lutz, Stefano Pagiola, and Carlos Reiche Soil degradation can be defined as a reduction in the Aggregate measures such as these, however, often land's actual or potential uses (Blaikie and Brookfield have a weak empirical basis. Studies that directly 1987). Many cultivation practices tend to degrade soil measure erosion rates and the factors that influence over time. For example, cultivation can expose soil to them are few and have generally been scattered and water and wind erosion, repeated tillage can weaken unsystematic. Even less effort has been devoted to soil structure, crop production can remove nutrients, studying other forms of land degradation, such as and use of machinery can compact the soil. Central nutrient depletion, damage to physical and chemical America's often mountainous environment and heavy properties of soil, or reductions in its capacity to rainfall make much of the region particularly vulner- retain moisture. Table 16.1 presents an overview of able to degradation-a problem exacerbated by popu- available estimates of erosion rates in Central lation pressures that have opened up new areas only American countries. The figures presented in the table marginally suited to agriculture. Soil degradation, in were obtained in a variety of ways and are therefore turn, affects productivity. As soil is degraded, crop not always strictly comparable, but they do give some yields decline or input levels (and hence costs) rise in idea of the great diversity of erosion rates present an effort to restore productivity. within the region. Despite long-standing concern about these prob- Predictions abound of the catastrophic effects that lems, surprisingly little hard evidence exists on their soil degradation will have on agricultural productivi- magnitude. The degradation figures quoted in the lit- ty. Evidence on the magnitude of these effects, by erature are often extrapolated from very limited data contrast, is hard to find. In fact, in many cases claims and may exaggerate the problem because they often of declines in productivity are made with no support- consider "moved soil" as "lost soil," even though ing evidence at all (see Biot, Lambert, and Perkin much of it may be deposited on other agricultural 1992 for some African examples). Leonard (1987), land. For instance, in a recent assessment of the extent for example, simply asserts that a pattern of extensive of human-induced soil degradation, the International land use leading to soil loss or decline in fertility is Soil References and Information Centre estimated apparent in the Caribbean areas of Central America. that 56 percent of the land in Central America has Speaking of the highland areas, he points to increas- experienced moderate degradation (implying that pro- ing reports of localized desertification in areas of ductivity has been substantially reduced) and that 41 western Honduras and Costa Rica. He also mentions percent has experienced strong degradation (implying that cotton yields are reportedly declining where that agricultural use has become impossible; severe erosion has been experienced. But nowhere Oldeman, Hakkeling, and Sombroek 1990). does he indicate the size or rate of fertility loss. More 215 216 The Costs and Benefits of Soil Conservation in Central America and the Caribbean Table 16.1. Empirical Evidence on Soil Erosion in Central America and the Caribbean Average rate of erosion Metric tons Country and Rainfall Slope Farming per hectare region or area (millimeters) (percent) system per year Millimeters Dominican Republic Taveras - - - 275 Northcentral - 36 Various 24-69 Southwest - 30 Various 2-1,254 - El Salvador Metapan 1,600 - Corn, beans 49 Haiti Camp-Perrin 2,000 30 Hedges 4-45 Papaye 1,200 25 Grass hedge 8 Honduras Tatumbla, Morazarn 2,000 45 Corn, beans 42 3 900-1,500 15-40 - 18-30 Nicaragua Cristo Rey 1,700 30-40 Cotton 40 Panama Cuenca del Canal 1,200 35 Rice 153 1,200 35 Corn 137 1,200 35 Rice 118 Coc1e 1,937 - Rice, corn 340 17 Chiriquf 1,500-2,800 - Cassava, beans 35 5 1,500-2,800 - Pasture 77 11 1,500-2,800 - Coffee 183 27 - Not available. Note: Figures are rounded. Source: Case studies in Lutz, Pagiola, and Reiche 1994. generally, the assumption that fertility must be declin- these soils is itself productive, although less so than ing rapidly is usually left implicit from statements the topsoil. The Chiriquf region in Panama provides about high rates of erosion. another example of this. Conversely, areas with shal- But erosion rates, even where they are significant, low soils or unfavorable subsoils, such as the may have very little effect on productivity under cer- Turrubares area in Costa Rica, can be very sensitive to tain conditions. Erosion rates in the Tierra Blanca area even limited rates of erosion. The same is true of other of Costa Rica's Cartago Province, for example, are forms of soil degradation. The impact of nutrient loss extremely high, but the effect on productivity is minor on productivity, for example, depends on the initial because soils in that region are deep (up to 1 meter in stock of nutrients and on their rate of regeneration. some places) and have high organic matter throughout Given the different effects of soil degradation on the soil profile. Moreover, the subsoil that underlies productivity, a specific soil conservation technique- Ernst Lutz, Stefaino Pagiola, and Carlos Reiche 217 particularly an expensive one-may not necessarily farm level is the most apt to incorporate site-specific be worthwhile from the perspective of a farm house- effects. hold or society. Degradation can be slowed or arrest- A farm-level approach also places emphasis firm- ed by a large range of options, including cultural prac- ly on the effects of degradation on farm productivity. tices such as contour plowing and minimum tillage, In developing countries, where substantial numbers of vegetative practices such as grass strips, strip crop- people still depend directly on agricultural produc- ping, and vegetative barriers, and mechanical mea- tion, the effect of degradation on yields is often criti- sures such as terraces and cutoff drains. Adopting any cal. This is not to belittle the importance, in some sit- of these techniques can be costly, either directly in uations, of off-farm effects of soil degradation, such investment requirements or indirectly in production as siltation of reservoirs and waterways. But even forgone, and some measures are better suited to some where such off-farm effects are the primary concern, conditions than to others. The critical question facing considering them first at the farm level is appropriate farmers and society as a whole is whether the benefits because conservation measures have to be imple- of a given conservation measure or set of measures mented on farms.1 are sufficient to make the costs worth bearing. In making their land use decisions, farm house- holds need to consider both the agroecological and the Conceptual Issues economic characteristics of the environment in which they operate. In addition, they often face numerous The problem of soil degradation and conservation can constraints, such as tenure problems, liquidity con- be examined from two perspectives. The first is that straints, and the need to meet consumption require- of society as a whole. From this perspective, all the ments and to compensate for missing or incomplete costs and benefits of a given activity must be consid- markets. Moreover, many farm decisions are made in ered. If agricultural production leads to siltation of the context of considerable risk and uncertainty. A reservoirs, for example, this represents a real cost to complete analysis of land use decisions, therefore, society that should be considered together with the requires that one look at the issue in the context of value of the output obtained and any effects on fertil- overall decisionmaking of the household (Singh, ity. In addition, valuation of the resources used and Squire, and Strauss 1986 and Reardon and Vosti obtained from agricultural production should be 1992). adjusted for any distortions resulting from policy The farm household's problem can be formulated interventions or market failures, in order to measure as one of maximizing the utility of consumption over their true opportunity cost. From the second perspec- time, subject to a budget constraint imposed by the tive, that of the farmers themselves, only the costs and returns from agriculture over time and from nonfarm benefits that actually accrue to the agent making the activities and subject to any other constraints it might decisions about resource use are considered. These face. Singh, Squire, and Strauss (1986) show that if costs and benefits are valued at the prices these agents markets exist for all goods and services, the problem actually face, with no attempt to adjust for distortions. of maximization is separable, in the sense that pro- This chapter examines the returns to investment in duction decisions are made independently of con- conservation measures mainly from the farmers' point sumption decisions. Even when production decisions of view for two reasons. First, decisions about land are not separable, however, they can be analyzed use are ultimately made by the farmers themselves independently as long as the "prices" of goods for and not by social planners or government agencies. which markets are missing are interpreted as shadow Farmers decide how to use their land in light of their prices that reflect the farm household's perception of own objectives, production possibilities, and con- the severity of the constraints it faces (de Janvry, straints and not on the basis of any theory of the social Fafchamps, andL Sadoulet 1991]). good. Understanding the incentives facing individual The household's problem, then, can be summa- farmers is necessary, therefore, if patterns of resource rized as one of maximizing the present value of the use are to be understood and appropriate responses to stream of expected net returns to agricultural produc- problems formulated. Second, land use problems gen- tion (Pagiola 1993). In practice, data are generally not erally depend heavily on site-specific biophysical available to estimate complex maximization models. characteristics, which can vary significantly even But for empirical analysis, the model can be reformu- within small areas (Pagiola 1993). Analysis at the lated to fit a cost-benefit analysis framework. The 218 The Costs and Benefits of Soil Conservation in Central America and the Caribbean Table 16.2. Sites of Country Studies Sites Biophysical environment Degradation problem Conservation measures proposed Costa Rica Barva area, Important coffee-producing Soil loss affecting Diversion ditches Province of Heredia region; relatively deep soil, nutrients available to but vulnerable to erosion coffee because of topography Tierra Blanca- Important vegetable-producing Deep soils so decline Diversion ditches recommended, San Juan Chicod, area; deep volcanic soils in yield is not significant; but interfere with prevalent Province of Cartago erosion washing away cultivation practices seed and fertilizer and exposing rocks Turrubares, Central Previously used for pasture, Very high rates of Diversion ditches or terraces Pacific region now converted to production erosion; thin soils of cocoa yam for export vulnerable to erosion Dominican Republic El Naranjal Subsistence agriculture; High rates of erosion Diversion ditches at 10-meter intervals, subwatershed, steep slopes, soils of live barriers, and cropping on the Peravia Province moderate natural fertility contour Guatemala Patzit6, Department Small-farmer area; strongly Heavily affected Terraces with a protected embankment of Quich6 undulating topography; soils of by soil erosion medium depth and fertility Haiti Maissade Hilly area; generally less Erosion Ramp pay (indigenous technique: Watershed, degraded and more productive crop residue placed along the Central Plateau than most other hilly contour, held in place by stakes), region regions of Haiti hedgerows along the contour, and contour rock walls Honduras Tatumbla, Subsistence agriculture Susceptible to water Diversion ditches Department of predominant; thin topsoil, with erosion, especially protected by live barriers Francisco Morazdn low levels of organic material in the high areas and of many nutrients Yorito, Small-scale subsistence Cleared plots vulnerable Diversion ditches with live barriers Department of Yoro agriculture, still largely forested; to erosion shallow, easily erodible soils, of medium to low natural fertility Nicaragua Santa Lucia valley, Subtropical foothills, moderately High risk of erosion Manually constructed watershed of deep soils; one of the most due to steep slopes, diversion ditches with stone Malacatoya River productive areas in the country scarce vegetation cover, barriers and intense precipitation; deforestation on upper slopes Panama Cocle Subsistence agriculture Rapid yield decline on Combination of erosion using slash-and-burn techniques, cleared plots; deforestation prevention measures (planting with plots cultivated one year on the contour, live and every five; shallow soils, dead barriers, diversion ditches) generally low in organic matter and improved cultivation practices and nutrients, on steep slopes Ernst Lutz, Stefano Pagiola, and Carlos Reiche 219 household's choice can be thought of as selecting unlikely to affect adoption. Of course, if adopting a between two or more alternative cropping systems. new production system is unprofitable for the farm For concreteness, one might think of a household household, the question of whether other constraints deciding whether to replace its traditional cultivation might prevent its adoption does not arise. system, in which conservation measures are limited to practices such as contour plowing, with a more con- Methodology serving system, which might include, for example, terraces or reduced-tillage techniques. Each system is Cost-benefit analysis techniques lend themselves well characterized by a distinct production function and to the evaluation of soil conservation measures soil growth function and, therefore, generates a dif- because they provide a coherent framework for inte- ferent optimal path. From the household's perspec- grating information on the biophysical and economic tive, the problem is whether returns under the optimal environments facing farmers. Variants of these tech- path of the new, more conserving system are suffi- niques have been used to examine a number of soil ciently greater than returns under the optimal path of conservation cases: for example, in the Dominican the current, more degrading system to justify the cost Republic (Veloz and others 1985), in India (Magrath of switching. 1989), and in Kenya (Pagiola 1992). Other simple Basically, it is in the farm household's financial techniques, such as calculating the value of lost nutri- interest to adopt the new, more conserving system if ents (Repetto and Cruz 1991), can only provide rough the net present value (NPV) of the incremental returns indicators of the severity of the problem. They cannot from switching is positive (NPV > 0). This formulation provide guidance in selecting the best response. is equivalent to a standard cost-benefit analysis for- The basic principles of the analysis are straight- mulation and lends itself particularly well to empirical forward. First, the effects of continued erosion (or analysis because data are often available in a suitable other types of soil degradation) on productivity are form. Observing practices in use allows time paths of estimated for the time horizon of interest. These are yield and use of inputs to be constructed; these are then used to estimate returns at each point in time. then used to project costs and revenues over time. The Second, the calculations are repeated under the condi- method can also be used if the only data available are tions that would be experienced if a specific conser- on total costs and revenues in each period, and it lends vation measure were adopted. The returns to the itself well to the incorporation of lumpy investments investment in this measure are then obtained by tak- and other discontinuities in cropping practices ing the difference between the streams of discounted (Walker 1982 and Taylor and others 1986).2 costs and benefits in the cases with and without con- The discussion so far has assumed that the only servation. It rnust be stressed that this method esti- constraints on behavior are those imposed by the mates the returns to the specific conservation mea- properties of the biophysical system. The NPV > 0 cri- sures being examined, not to conservation per se. A terion is thus a necessary, but not a sufficient, criteri- finding that certain conservation practices are not on for adoption of a new production system, because profitable does not imply that all conservation mea- other factors might prevent adoption of a new system sures are not profitable-often, numerous measures even if the NPV estimate is positive. In principle, these designed to reduce degradation rates are already being other constraints can be built into the optimization practiced. framework. The effect of insecure tenure might be As was argued in the previous section, when the included, for example, by limiting the length of the analysis is carried out at the farm level using prices time horizon. In practice, however, it generally proves actually facing farmers, a positive NPV estimate for a easier to compute the profitability of a system assum- given conservation measure can be interpreted as ing that no constraints hold and then verifying showing that adoption of the measure is profitable for whether other constraints are binding. The cost-bene- the farmer. Farmers should, in principle, be willing to fit calculations themselves often provide insight into adopt such a measure voluntarily. But, as with all whether particular constraints are likely to prove cost-benefit analysis, there is no guarantee that other, binding. The length of time it takes for an investment unexamined options would rnot be preferable. When to be repaid, for example, indicates whether tenure several options are known to exist, the analysis can be problems are likely to pose problems. If the invest- repeated for each in turn, and the most profitable ment is repaid very rapidly, insecurity of tenure is among them found. 220 The Costs and Benefits of Soil Conservation in Central America and the Caribbean For this article and the larger work from which it of the biophysical environment; building and validat- is drawn, the availability of data dictated the choice ing a complete and realistic model are complex both of the sites studied (see table 16.2) and of the endeavors. Even calibrating existing models is far aspects of the problem analyzed-erosion and from easy. mechanical methods of conservation.3 Research Obtaining the required economic data was gener- efforts have focused almost exclusively on problems ally less problematic. Crop production budgets, used arising from erosion, to the neglect of other forms of to estimate returns, were generally widely available, soil degradation, and most conservation projects in although rarely at the degree of disaggregation needed. the past have tended to emphasize mechanical conser- Fortunately, preliminary budgets built from available vation structures. Consequently, the case studies do secondary data were easy to confirm, supplement, and not present a comprehensive overview of soil conser- correct during fieldwork. The most important task vation problems and practices in the region; they do, was to ensure that the crop production budgets accu- however, illustrate the wide diversity of conditions rately reflected practices and prices in the area. Inputs encountered, help explain farmers' behavior, and indi- provided by the households themselves, such as fam- cate appropriate policy responses. ily labor, were priced at their cost in the nearest mar- The country studies, with the exception of Haiti, ket. Output and input prices used in the analysis were were conducted by local practitioners. In most cases, meant to represent long-run real price trends for out- teams were composed of economists, agronomists, puts and inputs. Assessing the discount rate is crucial, and soil scientists from relevant government agencies. given the intertemporal nature of the problem, but This collaborative and participatory approach to the beset by controversy. Here, because the analysis research proved successful in drawing on local examines the profitability of conservation from the knowledge and expertise. It also developed local ana- farm household's viewpoint, the appropriate discount lytical capacity. rate to use should be the farmers' cost of borrowing or The data needecl on the nature and rate of degrada- their rate of time preference. Little empirical evidence tion caused by current practices, on the effects of exists on either, however (Holden, Shiferaw, and Wik degradation on future productivity, and on the effects 1998). Therefore, and to facilitate comparability of of conservation practices were very scarce. Different results across study sites, a common real discount rate methods for estimating the required relationships of 20 percent was used in each case study. In addition, were chosen, depending on the nature of the available the internal rate of return (IRR) was computed in each data. Econometric techniques were sometimes case. If the appropriate discount rate, assuming it is employed to estimate the effect on yield of certain known, is smaller than the IRR, the proposed conser- observed conditions (such as the presence or absence vation measures would be profitable. of certain conservation measures). Disentangling the impact of soil degradation is very difficult (Capalbo Effects of Degradation on Productivity and Antle 1988). For our purposes, however, estimat- ing a time trend oif yields with and without a given The estimated productivity losses vary considerably conservation measure was usually sufficient, although across the case studies; table 16.3 presents findings even this limited objective encounters problems such for some of the crops analyzed. In several cases, the as bias in sample selection when nonconserved and data point to rapid rates of decline in yield. In the conserved fields are compared. In addition, many of Maissade watershed of Haiti, for example, yields of the case studies had to rely on farner recall data and corn and sorghum would decline as much as 60 per- were able to control for other sources of yield varia- cent over a decade in the absence of conservation tion, such as weather, only to a limited extent. In other measures. In the Tatumbla region in Honduras, coin cases, simple models of the physical environment- yields would decline almost 50 percent in 10 years if such as the Universal Soil Loss Equation and, in no conservation measures were used. Elsewhere, esti- Haiti, the Soil Changes under Agroforestry model- mated declines would be minor. Coffee yields in the were employed, using a mixture of experimental and Barva region of Costa Rica, for example, would observational data. This modeling approach is more decline just over 10 percent in 10 years, and there is flexible because it allows parameter values to be reason to believe that this rate of decline is overesti- drawn from a variety of data sources. But it requires mated. In Costa Rica's Tierra Blanca region, declines detailed knowledge (both qualitative and quantitative) in potato yield caused by erosion would be compen- Ernst Lutz, Stefano Pagiola, and Carlos Reiche 221 Table 16.3. Estimated Inpact of Soil Degradation on Productivity of Select Crops in the Areas Studied (percentage of initial yield) Years Projected Country and area Crop 10 20 30 40 50 shutdown year Costa Rica Barva Coffee 89 78 67 56 46 20 Turrubares Coco yam 0 0 0 0 0 4 Dominican Republic El Naranjal Pigeon peas 58 16 0 0 0 Peanuts 100 100 100 100 100 16a Beans 77 53 30 0 0 Guatemala Patzite Corn 0 0 0 0 0 lOb Haiti Maissade Corn, 41 22 10 1 0 25 sorghum Honduras Tatumbla Corn 53 39 39 39 39 8 Yorito Corn 82 65 47 41 41 11 Note: Projected year for production shutdown is in the absence of conservation measures. a. The 16-year shutdown period applies to the pigeon peas-beans-peanuts intercrop system. Because the three crops are cultivated together, peanut cultivation is assumed to cease when the yields of the other crops make pro- duction uneconomic. b. Corn can be produced in years one through nine, but decline in yield is so rapid that it reaches zero in year 10. Source: Case studies in Lutz, Pagiola, and Reiche 1994. sated easily by small increments in fertilizer use; Blanca the production of potatoes would remain prof- indeed, potato production has been increasing steadi- itable more or less indefinitely even without conserva- ly despite high rates of erosion. The effects of degra- tion. dation can also vary significantly across crops, even Not all the damage caused by soil degradation in the same area, as shown by the data from El takes the fomi of yield losses. In Tierra Blanca, for Naranjal in the Dominican Republic. example, the effects of degradation on agricultural If no conservation measures were adopted, returns production are reflected primarily in higher costs aris- to agricultural production would gradually decline in ing from the need to apply higher rates of fertilizer, each of the cases studied. Eventually, production from the lower efficiency of fertilizer (because some is would become uneconomic and cease, although washed away), and from the need to remove stones exactly when this will happen will vary, depending on that accumulate on fields as soil is eroded. In Panama's the rate of decline in yield, the cost of production, and Cocle Province, agricultural production could only be the price of the output. (Because farmers are likely to sustained for a very short time on a given plot if no adjust their production practices as yields decline, the conservation mneasures were used. The costs of degra- time before production becomes unprofitable is likely dation, therefore, are reflected primarily in the need to to be overestimated.) The very high rates of decline clear new plots of land at frequent intervals. experienced in Turrubares mean that the production of These examples, together with the diverse effects coco yam would shut down in four years if no conser- on yield, reinf orce the need for site-specific informa- vation measures were adopted; by contrast, in Tierra tion for understanding degradation problems and 222 The Costs and Benefits of Soil Conservation in Central America and the Caribbean devising effective ways of helping farmers respond to by about 14 percent, while terrace construction in the them. Note, however, that these case studies are by no Patzite region of Guatemala led to a 15 percent reduc- means a random sample of degradation conditions in tion. Further, terracing often entails the movement of the region; they are drawn from sites for which data earth, which can bring unproductive soil to the surface. were available, and therefore primarily from areas In Tierra Blanca, diversion ditches had the additional where degradation problems were serious enough to disadvantage of interfering with the prevailing produc- warrant data collection. Consequently, they probably tion practices, which rely heavily on mechanical represent high-case scenarios of the degree and rate of equipment. Such drawbacks heavily influence the ulti- degradation in the region. mate profitability of these conservation measures. The estimated effects on yields of conservation Because some of the productivity estimates are practices are likewise varied. In some instances, based on weak or incomplete data, extensive sensitiv- yields would recover once conservation measures ity analyses were incorporated into each of the case were established-partly because soil regenerates studies. The results are robust to changes in the esti- once the processes of degradation are halted, partly mated effects on yield in several cases, but in others because fertilizers are used more efficiently, and part- results are affected significantly by changes in ly because improved cultivation practices are some- assumed rates of decline in yield. In such instances, times introduced together with conservation. In the the premium to additional research would be high. In Tatumbla area of Honduras, for example, corn yields the Santa Lucia case study in Nicaragua, data were would increase about 145 kilograms annually if diver- insufficient to estimate the effects of degradation on sion ditches were built and improved planting prac- productivity. Simulation analysis was used, therefore, tices were adopted, up to a maximum set by the local to examine returns to the proposed conservation mea- agroecological conditions and the technology sures (manually constructed diversion ditches with employed by the farmers. Elsewhere, conservation stone barriers) under a range of assumptions about the measures might slow but not halt the decline in yield. effect of degradation and conservation on yield. The In the Turrubares area of Costa Rica, for example, results of the simulations show that the proposed con- diversion ditches would halve the rate of decline in servation measures are likely to be profitable only if yield; the much more expensive terraces, in contrast, the yield benefits of conservation are substantial. would reduce the rate to one-tenth. Again, the diversi- ty of conditions is evident. Farm-Level Returns to Soil Conservation As well as reducing soil loss and hence the rate of Measures decline in yield, conservation measures can affect yields by encouraging the retention of moisture and Effects on yield are not the only factors to consider in stimulating improvements in the soil's physical struc- analyzing the costs and benefits of investing in a ture (English, Tiffen, and Mortimore 1994). In Haiti's given conservation measure. The cost of constructing Maissade area, land treated with conservation struc- and maintaining conservation measures must also be tures was found to produce an average of 51 percent considered. The cost of constructing and maintaining more corn and 28 percent more sorghum than plots conservation measures must also be considered. Table without conservation structures in 1988 (a year of 16.4 summarizes the results of a full economic analy- poorly timed rainfall) and an average of 22 percent sis of each of the case studies where data were suffi- more corn and 32 percent more sorghum in 1989 (a cient to allow adequate assessment. more normal year). In dry areas, therefore, soil con- The most profitable conservation measure studied, servation can often reduce the risk of crop failure by in terms of rate of return, was found in Maissade, improving moisture retention. Haiti. This indigenous technique, known as ramp pay, Although introducing conservation measures can consists of crop stubble laid out along the contour, bring long-term benefits, it also can often result in supported by stakes, and covered with soil. It is cheap adverse effects on production because some of the area to construct and very effective in halting erosion. cultivated was turned over to use as diversion ditches, Moreover, without conservation measures, yield terraces, or hedges. Physical structures, in particular, would decline particularly rapidly in that area. High usually reduced the available area for cultivation by rates of return were also estimated for conservation more than 10 percent. Construction of cutoff drains in measures in Turrubares, Costa Rica, where highly Tierra Blanca, for example, reduced the effective area profitable export crop production was threatened by Ernst Lutz, Stefano Pagiola, and Carlos Reiche 223 Table 16.4. Estimated Returns to Investments in Conservation in the Case Study Areas Internal rate Number of Conservation Net present value of return years to break Country and area measure Crop (U.S. dollars) (percent) even Costa Rica Barva Diversion ditches Coffee -920 < 0 > 100 Tierra Blanca Diversion ditches Potatoes -3,440 < 0 > 100 Turrubares Diversion ditches Coco yam 1,110 84.2 2 Terraces Coco yam 4,140 60.2 3 Dominican Republic El Naranjal Diversion ditches Pigeon peas, -132 16.9 > 100 peanuts, beans Guatemala Patzite Terraces Corn -156 16.5 > 100 Haiti Maissade Ramp pay Corn, sorghum 1,180 Undefined 0 Rock walls Corn, sorghum 956 Undefined 1 Honduras Tatumbla Diversion ditches Corn 909 56.5 4 Yorito Diversion ditches Corn 83 21.9 18 Panama Cocl6 Terraces Rice, corn, 34 27.2 8 yucca, beans Note: Net present value is computed over 50 years, using a 20 percent real discount rate. rapid rates of yield decline, and in the Tatumbla area In Maissade in Haiti, Turrubares in Costa Rica, of Honduras, where yield decline would also have and Patzit6 in Guatemala, data were available to been very rapid if no conservation measures had been examine the returns to different forms of conserva- taken. tion. In Maissade, ramp pay is clearly superior to The least profitable conservation measures studied rock walls, wHich are more expensive and lack the were found in Barva and Tierra Blanca, Costa Rica. agronomic advantages of ramp pay. In Turrubares, the The Tierra Blanca case is particularly interesting, choice is less clear: terraces slow erosion much more because rates of erosion are very high. But the effectively than diversion ditches, but they are also region's deep volcanic soils mean that degradation more expensive to construct and entail a greater has very little effect on productivity. In fact, produc- reduction in effective cultivated area. The tradeoff tion would be higher without the proposed conserva- that must be mnade between effectiveness and cost is tion measures-diversion ditches-because their con- fairly easy in this case because the greater effective- struction would reduce the effective cultivated area ness of terraces more than compensates for their addi- and, by interfering with current production practices, tional cost. Buit this is not always true. In Patzite, for would increase the costs of production. It is not sur- example, a conabination of diversion ditches and live prising that farmers in the area had little interest in barriers appears to be substantially more profitable adopting these conservation measures. than terraces, even if they are much less effective. 224 The Costs and Benefits of Soil Conservation in Central America and the Caribbean This case appears to be more representative of condi- Obstacles to Adopting Conservation Measures tions encountered in Central America: in analyses of 20 conservation techniques in Mexico, for example, Profitability of conservation practices is a necessary McIntire (1994) finds that cultivation and cropping but not always a sufficient condition for their adop- practices, including vegetative barriers, are superior tion. Factors other than strict cost-benefit considera- to structural measures. Only when crop production is tions also play a role (Murray 1994). Some of these very profitable but extremely vulnerable to degrada- factors are reflected in the cost-benefit analysis to the tion (as in the case of Turrubares) are expensive con- extent that they affect the prices facing farmers. The servation measures likely to be justified. effect of imperfect factor markets, for example, would Unfortunately, data were insufficient to examine be reflected in higher prices for inputs, which would differences in returns within the study areas. Evidence affect the profitability of production activities. Most from Kenya (Pagiola 1994) suggests that returns to often, however, institutional issues must be consid- conservation can vary considerably even within nar- ered together with the results of the cost-benefit rowly defined agroecological zones. Farmers on dif- analysis. The analysis carried out for the case studies ferent slopes, for example, experience different rates does not always provide conclusive evidence on of erosion. They also face different costs of conserva- these, but it does provide some insights. tion; the optimal spacing of terraces and diversion It has often been argued that insecure property ditches, for example, is a function of slope. Whether rights dissuade farmers from undertaking long-term these differences are significant in any given instance investments, such as investments in soil conservation, is an empirical matter. because they may not themselves be able to reap the In each case, adoption rates correlate well with the benefits (Ervin 1986 and Wachter 1992). To respond estimated profitability of conservation. The profitabil- to this problem, numerous efforts have been made to ity of ramp pay is confirmed by its widespread adop- make tenure more secure by providing farmers with tion in Maissade. Conservation measures were also legal title to their land. The U.S. Agency for adopted at high rates in the Tatumbla region of International Development (USAID), for example, has Honduras and the Turrubares region of Costa Rica; funded titling projects in several countries, including not surprisingly, adoption rates were very low in El Salvador and Honduras. But equating land titles Tierra Blanca. Adoption rates were also low in Yorito, with secure tenure and thus with increased investment Honduras; there, the studies estimated the conserva- is too simplistic. Unless numerous improvements are tion measures to be marginally profitable, but the esti- made to the legal system and governmental institu- mates were based on particularly weak data and were tions, land titles often prove to be too costly to obtain fairly sensitive to changes in assumptions. Thus it or enforce for most farmers, and unless access to cred- may be perfectly rational for farmers not to adopt the it is improved for farmers holding titles, the desired proposed conservation measures. In some cases-in effect on investment may not materialize (L6pez Tierra Blanca, foir instance-degradation simply is 1996). not a significant problem for productivity. In others, The time required for investments in conservation the costs of the proposed conservation measures are measures to break even provides an important indica- too high relative to their benefits. The case of Patzite tor of the severity of tenure insecurity. Farmers with in Guatemala illustrates this best: although degrada- insecure tenure may doubt that they will be able to tion is relatively rapid and, if left untreated, will make profit from conservation measures if the benefits will production uneconomic within a decade, the proposed be reaped in the distant future. Table 16.4 shows that terraces are very expensive and take a lot of the land in most of the case studies, profitable conservation out of cultivation. Again, this is not to say that all con- measures had relatively short payback periods. Where servation measures are unprofitable. Visits to Tierra long payback periods were forecast, the measures Blanca show, for example, that although farmers have were either unprofitable or only marginally profitable not adopted diversion ditches, they do plant along and thus unlikely to be adopted even in the absence of contours and, on steeper slopes, construct temporary tenure problems. Other evidence from the case studies bunds on their fields. (The effects of these measures also suggests that tenure insecurity is not as signifi- are implicit in the estimates of degradation and of cant a problem as is sometimes thought. About 80 per- impact on productivity for the "without conservation" cent of the farmers in the Tatumbla area in Honduras case.) own land by occupation, that is, they do not have legal Ernst Lutz, Stefono Pagiola, and Carlos Reiche 225 title, yet most have adopted the recommended conser- of cases studied and weakness of the available data, vation measures. In the Patzite region in Guatemala, broad lessons must be drawn with care. It seems safe the proportion of farmers without title is similar; only to say, however, that except when high-value crops 10 percent of farmers have title to their land. Although are planted on very fragile soils (like the coco yam in erosion is a significant problem, adoption of conser- Turrubares), expensive mechanical structures are vation measures in this area has been relatively slow. unlikely to be profitable for the farmers. Conservation At first sight, this might appear to be evidence for the measures are particularly likely to be profitable either importance of titling. But negative profitability of the when they are cheap and simple or when they allow recommended conservation measures is more likely the adoption of improved practices. to account for low adoption rates than insecure tenure Generally, the farmers' decision to invest in con- or lack of land titles. servation is based on normal considerations of benefit Another oft-cited obstacle to adoption is the lack and cost: farmers tend to adopt conservation measures of capital markets. If credit markets fail, adoption of when it is in their interest to do so, unless some con- conservation will be limited by the farmers' ability to straint is present. Cases in which retums to conserva- self-finance the required investments (Pender 1992). tion are estimated to be low or negative correlate well The research carried out for this project did not bring with low adoption rates. to light any direct evidence on the functioning of cap- A full examination of the role of government poli- ital markets in the areas studied. The estimated rates cy in conservation would require a broader analysis of return for investments in conservation measures than that undertaken here; in particular, off-site effects shown in table 16.4 give some indication of the max- of degradation would have to be explicitly included imum rates that could be supported before the invest- and allowance made for distortions in observed price ments would become unprofitable. Several are signals resulting from govemment policies or market encouragingly high.4 failures. Nevertheless, several important points emerge from the present analysis. Conclusions and Policy Implications Advocates of soil conservation often argue that subsidies are indispensable to induce farmers to adopt Whether conservation measures are profitable for the conservation measures. But such statements assume farmers is an empirical and site-specific issue. that conservation is inherently desirable whether or Returns to conservation depend on the specific agro- not there is concrete evidence that the benefits out- ecological conditions faced, on the technologies used, weigh the costs. The results presented in this article and on the prices of inputs used and outputs produced. show that this may be far from the case; frequently, Hard data on the extent of soil degradation and its the benefits of specific conservation techniques (such effects on productivity remain extremely scarce as mechanical structures) do not justify their costs. despite several decades of soil conservation efforts Unless there are important off-site effects or the price (Lal 1988). More systematic research is needed on signals received by farmers are significantly distorted, soil degradation and its consequences, and there is subsidies to induce adoption would therefore not considerable scope for collaboration on such research, bring increased economic efficiency. because all countries within Central America include When off-site effects are present, the rationale for a large number of agroecological regions, and many intervention is potent, because the farmers' estimation agroecological regions are found in more than one of retums to conservation will pay inadequate atten- country. Regional organizations such as CATIE tion to its social benefits. In the Santa Lucfa Milpas (Centro Agron6mico Tropical de Investigaci6n y Altas watershed in Guatemala, for example, a USAID Ensenianza, Tropical Agricultural Research and project uses subsidies (so-called pago social) to Higher Education Center) have an obvious coordinat- induce farmers to build terraces and thus reduce ing role to play. The payoff is likely to be high, flooding in the historic town of Antigua. In the same because the approach to soil conservation is more tar- watershed, far:rners who do not receive subsidies gen- geted, with efforts concentrated where they are need- erally use less costly conservation methods such as ed most. vegetative barTiers and live fences. Although these The results of the case studies carried out in the measures are profitable to the farmers, they may not region show that conservation is profitable in some control floods. cases, but not in others. In view of the small number The effect of price distortions is more difficult to 226 The Costs and Benefits of Soil Conservation in Central America and the Caribbean establish. The many factors that affect the profitabili- initially very high, even though the evidence suggests ty of a given conservation measure and their compli- that the measures were unprofitable from the farmers' cated interactions make it hard to predict whether a perspective. In 1985, more than 90 percent of the distortion encourages or discourages conservation area's farms practiced soil conservation. Five years (Pagiola 1996). Recent evidence suggests that typical later, however, only half of these farms continued to policy distortions in developing countries tend to do so. Subsidies can only convince farmers to modify encourage degradation (Panayotou 1993), but more their behavior as long as they continue to be paid. In work is needed to substantiate this point. The best contrast, MARENA'S successor seems to have stimulat- way of dealing with policy distortions or market fail- ed considerable use of conservation techniques with- ures is to attempt to eradicate the distortions them- out offering subsidies; in fact, the cost of participation selves; subsidization should be resorted to only in the was quite high because conservation was tied to rare instances that such direct action is virtually access to irrigation. Although sufficient data were not impossible. available to analyze the new practices fully, they Whatever their justification, the use of subsidies appear to be highly profitable. encounters several difficulties. First, the divergence Another risk in designing subsidization schemes is between social and private returns to conservation that of creating perverse incentives for farmers. In must be established, so that intervention can be direct- Costa Rica, for example, a reforestation credit system ed where it will be most effective. However, subsidies unintentionally encouraged farmers to deforest their are often used where no off-site effects are present, land so that they might qualify for the credit. The wasting scarce budgetary resources in areas where expectation that subsidies are forthcoming to fund they are not justified by any social benefits. In Costa conservation efforts may also encourage farmers to Rica, for exampile, the soil conservation service delay conservation, even when such measures are pri- (SENACSA) subsidizes half the cost of establishing con- vately profitable, in the hope that the government will servation measures on the fields of small farmers, bear part of their cost. Even when they are justified, irrespective of location. Subsidies are also provided in then, subsidies must be used with great care. cases such as Turrubares, where individual farmers Governments should also ensure that constraints already have sufficient incentive to conserve purely such as insecure tenure do not prevent farmers from on productivity grounds. Conversely, subsidies are adopting conservation measures. But such efforts not always provided in cases where off-site effects are require substantiating research if they are to be effec- present. More commonly, subsidies are provided to tive. Too often the existence of tenure problems and construct, but not maintain, the conservation mea- the effectiveness of titling as a solution are simply sures, so farmers sometimes allow them to decay. In taken as given. Nicaragua, for example, terraces were built on fields Governments already do some research on soil in the Lake Xolotlan watershed above Managua in an conservation and provide, through extension services, effort to reduce flooding in the city and sedimentation some assistance to farmers who undertake conserva- in its reservoirs. Built at no cost to the farmers, these tion work. However, research in experiment stations terraces interfered with cultivation practices and did has tended to favor technical efficiency (including not result in private net benefits to the farmers; most structural measures such as terraces) over cost-effec- were soon destroyed. Similar experiences have tiveness. Further, government extension work is often occurred in the Tierra Blanca area of Costa Rica. ineffective. In many cases, nongovernmental organi- The second problem in using subsidies, then, is the zations, such as Vecinos Mundiales in Central difficulty of designing appropriate incentive struc- America (L6pez and Pfo Camey 1994) have proven to tures so that social objectives are met. The case of the be more effective at presenting the range of options Lake Xolotlan watershed illustrates a situation in to farmers and delivering related technical assistance. which subsidies are insufficient to overcome the Given the wide variety of conditions that farmers divergence between private and social returns to con- face, government extension services should also pro- servation. The El Naranjal watershed in the vide, explain, and demonstrate to farmers the variety Dominican Republic provides another example. Here, of options available rather than, as has often happened the USAID-funded Management of Natural Resources in the past, pushing specific techniques. And it may be Project (MARENA) provided subsidized credit to par- innovative as well as effective for governments to ticipating farmers. Consequently, adoption rates were decentralize decisionmaking and channel budgetary Ernst Lutz, Stejfano Pagiola, and Carlos Reiche 227 resources for soil conservation to the local level. This degradation problem-the inappropriate use of com- would allow communities to participate and contract mon property lands-are outside the scope of this assistance from those from whom the greatest contri- research. For a discussion of off-farm effects, see butions can be expected. Magrath and Arens (1989); for common property Research is not likely to produce a "breakthrough issues, see Bromley (1992). technology" that will solve all conservation problemns. 2. Combined investments from households in a Improvements are likely to be more marginal. But, village or watershed area are sometimes required to alone or in combination, improved techniques can manage land degradation problems effectively. For have a significant impact on productivity. analysis of such problems in the same area as the Modifications in the ramp pay technique used in Haiti Haiti case study described below, see White and are an example. Here, the traditional practice of gath- Runge (1992). ering crop stubble along the contour was improved by 3. In addition to the sites listed in table 16.2, more exact placement and by covering the structure research was carried out at other sites in several of the with upslope soil, thus discouraging rat infestations countries listed and at several sites in El Salvador. and encouraging surface flow infiltration. These Data on these sites-in particular, on the effects of changes made the practice much more effective in degradation on yields-are insufficient to allow a full halting degradation and more acceptable to farmers. analysis of the returns to conservation measures. Similar improvements in techniques arising from 4. Even when rates of return to investment in con- research have been successful in West Africa (Reij servation are hiigh, however, conservation might not 1992). be undertaken if even higher rates of return can be The conflict between conservation and production obtained by investing in off-farm income opportuni- noted in many of the case studies often affects the ties. Southgate (1994), for example, argues that high returns to conservation very significantly. Attempts to returns to urban employment in Ecuador encourage develop practices that reduce or eliminate this con- farmers to depreciate their land assets and then move flict-overlap technologies, in the terminology of to urban areas. Similarly, Schneider and others (1993) Reardon and Vosti (1992)-should be especially argue that perceptions of limitless land resources in encouraged. And to make the research truly useful, it the Amazon prompt farmers to "mine" their soils and should be carried out primarily on the farm and in then move on. close consultation with farmers. References Notes The word "processed" describes informally repro- The data used in the case studies were collected and duced works that may not be commonly available analyzed by Mauricio Cuesta and Hector Manuel through libraries. Melo Abreu (Costa Rica); Jose Abel Hemrndez Biot, Yvan, Robert Lambert, and Scott Perkin. 1992. (Dominican Republic); Jose Bueno Alferez, Jose What's the Problem? An Essay on Land Roberto Hernandez Navas, and Rafael Lazo Degradation, Science, and Development in Sub- Mel6ndez (El Salvador); Luis Eduardo Barrientos, Saharan Africa. Discussion Paper 222. East Saul Adolfo Lima, and Pedro Antonio Rosado Anglia: University of East Anglia, School of (Guatemala); Jon L. Jickling and Thomas A. White Development Studies. (Haiti); Jose Wilfredo Andino, Carlos Awad Ramirez, Blaikie, Piers, and Harold Brookfield. 1987. Gabino L6pez Vargas, Gilberto Palma, Augustino Pfo "Defining and Debating the Problem." In Piers Camey, and Antonio Valdez (Honduras); Danilo Blaikie and Harold Brookfield, eds., Land Antonio Montalban and Miguel Obando Espinoza Degradation and Society. London: Methuen. (Nicaragua); and Cesar Isaza, Julio Santamarfa, and Bromley, Daniel W., ed. 1992. Making the Commons Tomas Vdsquez (Panama). The authors are indebted Work: Theory, Practice, and Policy. San Francisco: to Hans Binswanger, John English, John McIntire, ics Press. Augusta Molnar, Alfredo Sfeir-Younis, Stephen Vosti, Capalbo, John M., and Susan M. Antle. 1988. and several members of the Editorial Committee for Agriculturnl Productivitv: Measurement and valuable comments and suggestions. Explanation. Washington, D.C.: Resources for the 1. Off-farm effects and another important land Future. 228 The Costs and Benefits of Soil Conservation in Central America and the Caribbean De Janvry, Alain, Marcel Fafchamps, and Elisabeth Accounting Approach." Environment Department Sadoulet. 1991. "Peasant Household Behavior Working Paper 18. World Bank, Washington, D.C. with Missing Markets: Some Paradoxes Processed. Explained." The Economic Journal 101:1400-17. McIntire, John. 1994. "A Review of the Soil English, John, Mary Tiffen, and Michael Mortimore. Conservation Sector in Mexico." In Ernst Lutz, 1994. Land Resource Management in the Machakos Stefano Pagiola, and Carlos Reiche, eds., District, Kenya: 1930-1990. Environment Paper 5. Economic and Institutional Analyses of Soil Washington, D.C.: World Bank. Conservation Projects in Central America and the Ervin, David E. 1986. "Constraints to Practicing Soil Caribbean. Environment Paper 8. Washington, Conservation: Land Tenure Relationships." In D.C.: World Bank. Stephen B. Lovejoy and Ted L. Napier, eds., Murray, Gerald. 1994. "Technoeconomic, Organiza- Conserving Soil: Insights from Socioeconomic tional, and Ideational Factors as Determinants of Research. Ankeny, Iowa: Soil Conservation Soil Conservation in the Dominican Republic." In Society of America. Ernst Lutz, Stefano Pagiola, and Carlos Reiche, Holden, S. T., B. Shiferaw, and M. Wik. 1998. eds., Economic and Institutional Analyses of Soil "Poverty, Market Imperfections, and Time Conservation Projects in Central America and the Preferences: Of Relevance for Environmental Caribbean. Environment Paper 8. Washington, Policy?" Environment and Development 3(1): D.C.: World Bank. 83-104. Oldeman, L. R., R. T. A. Hakkeling, and W. G. Lal, Rattan C. 1988. "Preface." In Rattan C. Lal, ed., Sombroek. 1990. World Map of the Status of Soil Erosion Research Methods. Ankeny, Iowa: Human-Induced Soil Degradation: An Explanatory Soil and Water Conservation Society. Note, 2d ed., rev. Wageningen: International Soil Leonard, H. Jeffrey. 1987. Natural Resources and Reference and Information Centre. Development in Central America: A Regional Pagiola, Stefano. 1993. "Soil Conservation and the Environmental Profile. New Brunswick, N.J.: Sustainability of Agricultural Production." Ph.D. Transaction Books for the International Institute diss. Food Research Institute, Stanford University, for Environment and Development. Palo Alto, Calif. L6pez, Gabino Vargas, and Augustin Pfo Camey. . 1994. "Soil Conservation in a Semi-Arid 1994. "Practical Experiences and Lessons Learned Region of Kenya: Rates of Return and Adoption by Vecinos Mundiales from Soil Conservation by Farmers." In T. L. Napier, S. M. Camboni, and Work in Rural Communities of Honduras." In S. A. El-Swaify, eds., Adopting Conservation on Ernst Lutz, Stefano Pagiola, and Carlos Reiche, the Farm: An International Perspective of the eds., Economic and Institutional Analyses of Soil Socioeconomics of Soil and Water Conservation. Conservation Projects in Central America and the Ankeny, Iowa: Soil and Water Conservation Caribbean. Environment Paper 8. Washington, Society. D.C.: World Bank. . 1996. "Price Policy and Returns to Soil L6pez, R. 1996. "]Land Titles and Farm Productivity Conservation in Semi-Arid Kenya." Evironmental in Honduras." Department of Agriculture and and Resource Economics 8:225-71. Resource Economics, University of Maryland, Panayotou, Theodore. 1993. Green Markets: The College Park, Md. Processed. Economics of Sustainable Development. ICEG Lutz, Ernst, Stefano Pagiola, and Carlos Reiche, eds. Sector Studies 7. San Francisco: ics Press. 1994. Economic and Institutional Analyses of Soil PCEO (Proyecto de Control de Erosi6n de Occidente). Conservation Projects in Central America and the 1981. "Internal Report." Irena, Nicaragua. Caribbean. Environment Paper 8. Washington, Processed. D.C.: World Bank. Pender, John L. 1992. "Credit Rationing and Farmers' Magrath, William B. 1989. "Economic Analysis of Irrigation Investments in Rural South India: Soil Conservation Technologies." Divisional Theory and Evidence." Ph.D. diss. Food Research Working Paper 1989-4. Environment Department, Institute, Stanford University, Palo Alto, Calif. World Bank, Washington, D.C. Processed. Reardon, Thomas, and Stephen A. Vosti. 1992. Magrath, William B., and Peter Arens. 1989. "The "Issues in the Analysis of the Effects of Policy on Costs of Soil Erosion on Java: A Natural Resource Conservation and Productivity at the Household Ernst Lutz, Stefano Pagiola, and Carlos Reiche 229 Level in Developing Countries." Quarterly Perspective of the Socioeconomics of Soil and Journal of International Agriculture 31(4, Water Conservation. Ankeny, Iowa: Soil and October-December):380-96. Water Conservation Society. Reij, Chris. 1992. "Building on Traditions: The Taylor, Daniel B., Douglas L. Young, David J. Walker, Improvement of Indigenous swc Techniques in the and Edgar L. Michalson. 1986. "Farm-Level West African Sahel." Paper presented at the inter- Economics of Soil Conservation in the Palouse national symposium on soil and water conserva- Area of the Northwest: Comment." American tion, Soil and Water Conservation Society, Journal of Agricultural Economics 68(2, Honolulu, Hawaii, 19-21 October. Processed. May):364-65. Repetto, Robert, and Wilfrido Cruz. 1991. Accounts Veloz, Alberto. Douglas Southgate, Fred Hitzhusen, Overdue: Natural Resource Depreciation in Costa and Robert Macgregor. 1985. "The Econornics of Rica. Washington, D.C.: World Resources Erosion Control in a Subtropical Watershed: A Institute. Dominican Case." Land Economics 61(1, May): Schneider, Robert, Gunars Platais, David Rosenblatt, 145-55. and Maryle Webb. 1993. "Sustainability, Yield Wachter, Daniel. 1992. "Land Titling for Land Loss, and Inmediatismo: Choice of Technique at Conservation in Developing Countries." the Frontier." LATEN Dissemination Note 1. Latin Divisional WvVorking Paper 1992-28. Environment America Technical Department and Environment Department, World Bank, Washington, D.C. Division, World Bank, Washington, D.C. Processed. Processed. Walker, Davidt J. 1982. "A Damage Function to Singh, Inderjit, Lyn Squire, and John Strauss, eds. Evaluate Erosion Control Economics." American 1986. Agricultural Household Models: Extensions, Journal of Agricultural Economics 64(4, Applications, and Policy. Baltimore, Md.: Johns November): 145-55. Hopkins University Press for the World Bank. White, Thomas A., and C. Ford Runge. 1992. Southgate, Douglas. 1994. "The Rationality of Land "Common Property and Collective Action: Degradation in Latin America: Some Lessons from Cooperative Watershed Management in Haiti." the Ecuadorian Andes." In T. L. Napier, S. M. Working Paper P92-3. Center for International Camboni, and S. A. El-Swaify, eds., Adopting Food and Agricultural Policy, University of Conservation on the Farm: An International Minnesota, St. Paul. Processed. 17 Toward a Strategy for Mainstreaming Biodiversity in Agricultural Development Jitendra Srivastava, Nigel Smith, and Douglas A. Forno The conservation of biodiversity has emerged as a agriculture and biodiversity are intimately connected major priority. It is no longer just the concern of bird and that one cannot survive without the other. watchers and a handful of field botanists. Citizens Continued progress in raising and sustaining agricul- and politicians alike have rallied to the cause of sav- tural yields hinges on better protecting and harness- ing biodiversity, which is harnessed by cultures in ing the planet's biological riches. various ways to produce food and other products. Agriculture is often seen as the "enemy" of bio- Much of the planet's remaining biodiversity will be diversity, rather than as part of it. This perception lost unless future needs can be met from areas arises because the raising of livestock and crops already cultivated or grazed. Demand for food and inevitably alters vast expanses of the earth's surface. other agricultural products is likely to increase sig- Population growth and other factors encourage defor- nificantly over the next few decades. As societies on estation, pastoralists are squeezed into ever-diminish- every continent become more urbanized and income ing spaces and sometimes overgraze the land, and levels rise, more livestock products are consumed, high-input, modem farming practices frequently pol- and greater demands are placed on landscapes to pro- lute the water and soil with chemicals. All of these duce feed for cattle, pigs, and other domestic ani- activities trigger a potentially dangerous loss of bio- mals. This process is likely to exert further pressure diversity. However, some land use systems and agri- on biodiversity unless a concerted effort is made to cultural practices enhance biodiversity within man- adopt more environmentally sound agricultural prac- aged landscapes. For example, the judicious use of tices. livestock waste as organic manure enhances the species diversity of macrofauna (Bohac and Agriculture: Friend or Foe of Biodiversity? Pokarzhevsky 1987). Also, inappropriate agricultural practices can be modified to mitigate adverse impacts Some dramatic changes will be needed in the ways on the environment. that people raise crops and livestock if much bio- diversity is to survive the next 50 years. How agri- Agricultural Intensification: A Bane or Blessing culture is transformed and intensified in a sustainable for Biodiversity? manner will be the key to how many species and how much genetic variation are still around in the next To many, agricultural intensification means more century. A focus on conserving biodiversity in "pro- purchased inputs, such as fertilizers, pesticides, her- tected areas" alone will not work (box 17.1). The bicides, and machinery. Eutrophication of lakes and main purpose of this chapter is to make the case that estuaries, loss of soil micro-organisms, accelerated 230 Jitendra Srivastava, Nigel Smith, and Douglas A. Forno 231 soil erosion, contamination of groundwater, and The Core Issue draining of wetlands-to mention a few of the adverse impacts of some agricultural practices-attest How agriculture can be intensified while enhancing to the dangers inherent in intensifying agriculture biodiversity is the critical question tackled in this without regard to the long-term consequences for the chapter. Our goal is to identify some of the critical natural resource base. dimensions to this issue, to illuminate their multiple This chapter does not purport to provide all the facets, and to suggest policies that mitigate the answers. It is only one step in that direction. But sus- adverse impacts of agriculture on the environment. tainable agricultural intensification would include Our concern here is not only to highlight ways in approaches such as: which agricultural practices can be tailored so that * More rational use of nutrients, space, and energy they are more environmentally friendly, however. We in all land use systems are especially concerned with incorporating greater * Greater recycling of nutrients biodiversity wilhin agricultural production systems. * Better use of biological resources to raise and New approaches to agricultural research and devel- maintain yields of crops and livestock, for exam- opment are being tried in various places around the ple, by improving the use of genetic resources and world, and virtually all of them emphasize harnessing biocontrol agents of crop pests and managing biological resources better than in the * Greater appreciation for, and use of, indigenous past. Instead of excessive reliance on an arsenal of knowledge, especially concerning numerous potent chemicals to improve soil fertility and thwart neglected crops that could help improve liveli- the attacks of in,sects and disease-causing organisms, hoods and the environment agricultural research is geared increasingly to the * More effective measures for soil and water conser- manipulation of genes and release of predators of vation crop pests, among other biological assets. When * The deployment of "environmental corridors" in crops and livestock are bred so that they can thrive landscapes that have been transformed by agricul- under the incessant onslaught of challenges to pro- ture and livestock raising. ductivity, agricultural production systems become more resilient. Agricultural intensification does not automatically Box 17.1. A Holistic Approach to Biodiversity trigger greater lharm to the environment. On the con- Conservation trary, it can save and enhance biodiversity. Benign policies and practices that enhance agricultural pro- Protection of a sample of natural habitats is neither ductivity as well as biodiversity conservation are sufficient nor desirable to conserve biodiversity for possible. This chapter pinpoints ways in which this two simple reasons: (a) most of the world's biodiver- has already been accomplished in certain areas and sity exists in human-managed or -modified systems, suggests measuires that might accelerate the wider and (b) land use patterns and sociopolitical factors in adoption of sound practices. Such information will be areas adjacent to parks and reserves have major useful to personts engaged not only in designing and implications for the integrity of biological diversity in implementing agricultural development projects but "protected" areas (Pimentel and others 1992). This also in establishing priorities for agricultural relationship has clearly been demonstrated by the fate research. Individuals involved primarily in the con- of 62 bird species in an 86-hectare woodland in West servation of biodiversity as it is most widely under- Java. After several square kilometers of surrounding stood-the safeguarding of wildlife and "nat-ural" woodland were destroyed, 20 bird species disap- habitats-mighit also find use in the discussion on the peared, four declined almost to extinction, and five complementarities between agriculture and environ- more declined noticeably (Diamond, Bishop, and Van mental conservation. Balen 1987). The remaining species appeared to be unaffected. This example highlights the need for Biodiversity and Agrobiodiversity Defined regional conservation (Ricklefs 1987) and the need for biodiversity conservation in both protected areas Before explorin1g the complex issues surrounding and agricultural ecosystems. agricultural development and biodiversity, it is useful and____agricultural ________ecosystems._______ to define what we mean by biodiversity and agrobio- 232 Toward a Strategy for Mainstreaming Biodiversity in Agricultural Development Figure 17.1. Agrobiodiversity as a Subset of Biodiversity 4 ~~Biodiversity Crop species and varieties Livestock species and breeds grobiodiversi Soil micro-organisms in cultivated areas Biocontrol agents for crop and livestock pests Wild populations with spontaneous \ ~~~~~~~~~~~introgression or breeding \ \ Ne~~ar relatives with spontaneous/ diversity, the subset of biological resources that sup- ing biodiversity in habitats transformed by humans, port agricultural production (figure 17.1). such as farmland, rangeland, and forests exploited for Biodiversity has three main dimensions: the genet- lumber and nontimber forest products. How farmers ic variation within species and populations, the num- and livestock owners manipulate natural resources is ber of species, and the preservation of habitat (Srivastava, Smith, and Fomo 1996). The significance Box 17.2. Biodiversity and the Tropics of variation within a species is not widely appreciated but is critical, particularly for agriculture. The contin- The conservation and management of biodiversity are ued productivity of existing crops and livestock particularly critical in lower latitudes because they hinges in large part on harnessing the genetic varia- contain many more species than are typically found in tion found within each species. temperate areas. Agrobiodiversity is also exceptional- The second dimension of biodiversity is fairly ly high in the humid tropics. Tropical rain forests straightforward: it is an index of species richness or cover only 7 percent of the earth's surface but more stheinumberwa: of distc pant andexof acims ricnagven, X than half of the world's biota (Wilson 1988). A couple environment. Thus tropical rain forests are especially of examples suffice to illustrate the extraordinary lev- rich in species, and their fate has major implications els of biodiversity found in tropical rain forests and to for many crops important for subsistence and cash underscore their importance for conservation and the income in the tropics and subtropics (box 17.2). In need for careful management of their biological rich- order to protect species and genetically distinct popu- es. A 13.7 square kilometer portion of the La Selva each ~ spce,i.sncsayt aeur reserve in Costa Rica contains almost 1,500 plant lations of e cies, species, more than are found in all of the United their environments. Kndm(yr 98.TeMlynpnnuai The issue of habitat conservation has two parts Kingdom (Myers 1988). The Malayan peninsula is safeguarding natural habitats for wild species and only about a third as large as the United Kingdom, but populations and managing habitats that have been it contains five-and-a-half times as many plant modified for human use, such as farmland. The sec- species (Whitmore 1985). Tropical forests contain ond item is poorly understood among the general pub- wild populations of hundreds of crops, such as lic and many policymakers. Economic growth and mango, coffee, and rubber, as well as candidates for poverty alleviation hinge in large measure on manag- new crops and livestock. Jitendra Srivastava, Nigel Smith, and Douglas A. Forno 233 therefore critical to the question of conserving bio- cent of the land is in nominally protected parks or diversity. reserves, and some 70 percent of the total land surface It is in habitats that have been modified for crop is in agriculture or managed forests (Pimentel and and livestock production that one finds agrobiodiver- others 1992). Considerable biodiversity is often found sity, which includes all plants and animals that con- in areas managed for agricultural production or tribute directly or indirectly to the raising of crops and extractive products. livestock. How and why rural people conserve, enhance, and Although it is conceptually useful to differentiate use biodiversity have rarely been taken into account agrobiodiversity from the larger array of species and when designing management interventions and devis- habitats, it is worth emphasizing that the boundaries ing policy for agricultural development and natural between biodiversity and agrobiodiversity are not resource management. But the active participation of clear-cut. All of biodiversity is potentially of use to farmers, ranchers, and pastoralists-and especially agriculture, particularly with the emerging biotech- resource-poor operators-is essential in the design nologies. Agriculture is highly dynamic, and the inter- and implementation of biodiversity and agricultural face between domesticated plants and animals and development projects (Thrupp, Cabarle, and Zazueta wild species is constantly shifting. A wild plant or ani- 1994 and Wilcox and Duin 1995). mal of little or no current market value could eventu- ally provide significant employment and income in The Role of the Business Community the future. The fact that it is not easy to draw a firm line between biodiversity and agrobiodiversity only One of the leitmotifs of this chapter is that market underscores the importance of conserving as much forces can be harnessed to ferret out and promote biodiversity as feasible to maintain its future value for crops or old varieties that are in danger of slipping agricultural development, among other reasons. into extinction. The same idea applies to ancient, but Agriculture has a direct stake in safeguarding the dwindling, livestock breeds and some unconventional wider biodiversity. Wild species are essential for agri- livestock species, such as iguanas. If markets can be cultural improvement because they are a source of found for "forgotten" crops and livestock breeds, they new economic plants and animals and provide impor- are more likely to survive. Gene banks and embryo tant services such as pollination and pest control. stores cannot save all the varieties and breeds that are Moreover, advances in biotechnology are pushing no longer commercially viable. Markets are constant- back the limits to exchanging genes between unrelat- ly changing, and new opportunities are emerging for ed organisms. tapping some of these dwindling pockets of agro- Given the ultimate importance of biodiversity in biodiversity to generate income for locals and enrich its broadest sense to agriculture, a strategy for main- the diets of people around the world. streaming biodiversity in agricultural development Many of the innovative approaches to marketing needs to address the off-site impacts of land use sys- the untapped wealth of biodiversity in Latin America tems. Steps in this direction are outlined in the call for are likely to come from relatively small-scale enter- a new agricultural research and development para- prises. One example is Kapok Intemational, based in digm. Work is already under way to address a range of Chagrin Falls, Ohio. Kapok International has a issues related to off-site impacts, including the reduc- Brazilian subsidliary in Manaus, Brazil, and currently tion or elimination of agricultural pollutants in markets some Armazonian fish. It soon expects to groundwater and in runoff and the placement of begin marketing unusual Amazonian fruits for the greater emphasis on integrated pest management juice and candy industry in North America (box 17.3). (IPM) strategies. With assistance from the International Finance To many, biodiversity conservation is essentially Corporation, an innovative trust fund has been estab- synonymous with protecting showcase mammals, lished to spur greater private sector investments in birds, and wild and spectacular landscapes. Although improving the use and management of the wealth of efforts to safeguard the habitats of wildlife warrant biological resources in Latin America. The support, this conventional approach to conservation Biodiversity Enterprise Fund for Latin America is of biodiversity is insufficient. Most of the earth's sur- designed to help companies start up or expand busi- face has now been modified in various ways by ness activities linked to sustainable use of the conti- human activities. On a global scale, less than 5 per- nent's biodivers.ity, which is both underused and, in 234 Toward a Strategy for Mainstreaming Biodiversity in Agricultural Development Box 17.3. Business Opportunities in Agrobiodiversity: The Case of Kapok International Turning the tide on the rapid destruction of species- One product that Kapok International expects to rich rain forests can only occur if the economic needs market in the near future is cupua u cocoa. of the local people are addressed. To that end, Kapok Pronounced coo-poo-a-su, this relative of cacao ren- International is seeking to develop new markets for ders a fine chocolate-like drink with a hint of citrus. some intriguing Amazonian fruits that are little CupuaCu makes a creamy-tasting cocoa, but without known outside the region. Based in Chagrin Falls, the milk, a major plus for the millions of people who Ohio, Kapok International is exploring the potential suffer from lactose intolerance. Fine cupua u choco- of some nontimber forest products and some promis- late bars may soon grace supermarket shelves. ing crops that carL be grown or collected on a sustain- Several drinks containing cupua u pulp have able basis. In this manner, economic value will be appeared recently in several U.S. markets. added to forests, and locals can improve their stan- A lack of infrastructure in most parts of the dards of living. Amazon is a major impediment to exploring the full Market opportunities in the United States abound potential of the region's numerous tropical fruits. for "cause--oriented" foods. Consumers in the United Poorly maintained roads impede the timely delivery States, as well as many other countries, are increas- of fruits to processing plants. Unsanitary conditions ingly interested in novel foods, especially if they and inadequate refrigeration at the region's few agro- come from exotic locations such as the Amazon. And industrial plants also hinder the export of frozen pulps some consumers are willing to pay a premium for and purees to demanding markets in North America extractive and agricultural products that have been and Europe. With supportive agricultural policies and obtained without damaging the environment or the an improved investment climate, such barriers can be cultures of local people, especially indigenous overcome to the benefit of consumers, local people, groups. and biodiversity. Box 17.4. Biodiversity and Sustainability of Agricultural Systems As with any ecosystem, the functioning and sustain- value. Insects provide a virtually untapped source of ability of agricultural systems depend greatly on bio- food (Defoliart 1989), dyes (Metcalf, Flint, and logical diversity. Through an experimental study, Metcalf 1962), and pharmaceutical products (Eisner Tilman, Wedin, and Knops (1996) demonstrate that 1990). The elimination or addition of even one more diverse plant communities use and retain nutri- species can have profound effects. Pimentel and oth- ents more efficiently, thereby attaining greater pro- ers (1992) cite a remarkable example to illustrate this. ductivity and reducing the loss of nutrients leaching Until recently, oil palm trees in Malaysia were polli- from the ecosystem. Various soil microbes also effec- nated manually-an inefficient and expensive way of tively recycle nutrients (Lal 1991). Earthworms, performing the task. Ten years ago, the government insects, and fungi all play a vital role in the function- introduced a tiny weevil from West Africa's forests ing and sustai:nability of agricultural systems. that is associated with palm pollination. The pollina- Biological diversity also enhances natural pest control tion of palm trees in Malaysia is now accomplished mechanisms in agroecosystems. In fact, every single entirely by the weevil, with annual savings of $140 species that exists in agroecosystems has intrinsic million (Greathead 1983). some cases, overexploited. Compared with traditional and in other regions (see box 17.4). investments in agriculture, fisheries, and forestry, the The private sector has critical roles to play at vari- $20 million to $30 million fund for biodiversity-based ous steps of the process of making better use of agro- enterprises is a small but promising start. As experi- biodiversity. In the case of planting material, multina- ence accumulates as to what works and what does not tional seed companies dominate the production and with such specialized investment funds, the expansion sale of major cereals, but opportunities exist for of such efforts should be considered in Latin America small-scale private seed companies to exploit the Jitendra Srivastava, Nigel Smith, and Douglas A. Forno 235 growing demand for unconventional crops, particular- ed in maintaining animal and plant genetic resources ly in areas where the crops have recently been intro- in places where they occur naturally (in situ), while duced. Although the private sector is still a relatively others suggest that genetic variation is more useful, minor source of new varieties in most developing and often safer, if it is maintained in seed or field gene countries, this picture is gradually changing banks (ex situ). Both strategies are needed to sustain (Dalrymple and Srivastava 1994). Policies at the the productivity of agricultural systems. national and intemational levels are needed to facili- tate this encouraging trend. Promising in situ approaches. The following approaches to in situ conservation of crops and live- Major Recommendations stock warrant particular support: * Emphasize a new dimension to existing wilderness This section brings together major policy recommen- parks and biological reserves by safeguarding dations within a framework that both highlights the plants and animals for the future improvement of multiple dimensions of agriculture-biodiversity inter- agriculture. Some existing parks and reserves con- actions and suggests some ways to proceed. We tain wild populations or near-relatives of crops and explore fresh approaches on in situ conservation of livestock; highlighting such attributes would plant and animal genetic resources, underscore the strengthen the case for improving their protection value of work on the systematics of plants and ani- and management. A major problem here is that mals to help us better understand variation within many areas under consideration for protection species and relationships between species, and make a have not been adequately inventoried from the case for assembling interdisciplinary teams to assess botanical standpoint. agrobiodiversity before agricultural development * Create world heritage sites for genes for agricul- projects are implemented. Major policy recommenda- tural development. Most parks and reserves have tions span issues such as promoting quality pricing, been set up to protect spectacular animals or rare extending credit to traditional varieties as well as and endangered plants and habitats. Although such modern cultivars, and reducing agricultural subsidies, efforts deserve continued support, the agrobio- which often exacerbate environmental problems. diversity aspect of environmental conservation has Dimensions to the emerging agricultural research par- been largely overlooked. Many areas are "hot adigm that stresses sustainability are also examined. spots" for agrobiodiversity anid warrant better con- A strategy is needed to help policymakers recon- servation and management in their own right. cile the task of preserving biodiversity while at the * Integrate agrobiodiversity in ecotourism where same time increasing agricultural productivity. In the appropriate. A virtually untapped market exists for first section, we explored some of the important rela- showing tourists who visit nature reserves and tionships between agriculture and biodiversity and major archaeological sites "traditional" villages argued that biodiversity conservation and agricultural and farmlands in the surrounding area. development can be complementary activities. Here * Help find markets for lesser-known crops and we explore some tentative steps toward an overall local varieties under the motto "Use it or lose it." strategy for mainstreaming biodiversity in agricultur- If a highly localized crop that is on the decline sud- al development. denly ignites interest in a distant market, growers will be motivated to look around for highly pro- AgendaforAction ductive varieties. This search for planting material might save varieties that would otherwise be aban- Various policies and practices can be devised to pro- doned. mote biodiversity conservation in managed land- * Help find ways for livestock owners to generate scapes and to enhance the greater use of bioiversity in more revenues from threatened breeds. If such agriculture (table 17.1). The major points for policy breeds can "pay" for themselves, they are less like- consideration are summarized here. ly to be replaced by the more widespread breeds or to lose their identity in continuous crossbreeding. Toward a balanced conservation strategy. Too often conservation is posed as an either/or proposi- Greater involvement of the private sector. Too tion: some argue that resources should be concentrat- often, biodiversity conservation is seen as a public Table 17.1. Toward a Strategy for Mainstreaming Biodiversity in Agricultural Development Objective Proposed action Responsible organizations Commitment to enhancing biodiversity in agriculture Increase government commitment Conduct substantive dialogue with client countries Multilateral development banks, bilateral to mainstreamning biodiversity on the need to conserve and manage aid agencies, NGOS in agriculture and rural development biodiversity resources Increase farmer commitment to Support efforts to establish property rights so that National governments incorporate greater diversity farmers are more willing to invest in long-term in farming systems strategies to incorporate more biodiversity, such as intercropping with perennials; identify other incentives that encourage farmers to promote agrobiodiversity in their cropping systems Generate greater understanding of Prepare training manuals and packets Multilateral development banks, bilateral aid how to implement policies and sound for human resource development agencies, international agricultural research practices to enhance biodiversity in centers, NGOS, agricultural extension agencies agricultural development Diversify farming operations with Support more research in lesser- National agricultural research programs viable alternative crops known or more localized crops Collect sufficient data on the Develop methodologies and performance Multilateral development banks, bilateral potential impact of agricultural indicators for assessing agrobiodiversity aid agencies, national agricultural research development projects on in areas targeted for agricultural development programs, NGOS, farmers biodiversity and agrobiodiversity New agricultural research and extension paradigm Increase participation of farmers Conduct substantive dialogue with developing- Farmers, multilateral development banks, in the design and implementation country institutions on the need to incorporate, international agricultural research centers, NGOS of agricultural development projects not just consult, local farmers in agricultural to enhance agrobiodiversity research and development projects Increase the deployment of Offer greater support for developing polygenic International agricultural research centers, genetically diverse (heterozygous) resistance, multilines, high-yielding, national agricultural research programs populations of crop varieties and open-pollinated varieties instead of hybrids animal breeds on managed landscapes Improve productivity of fanning Conduct more research on undervalued crops and Farmers, international agricultural research and livestock raising in marginal livestock that are better adapted to difficult centers, national agricultural research programs environments environments than the "mega" crops and livestock Create greater awareness of Develop and disseminate "sound practices" Multilateral development banks, bilateral aid appropriate practices for tailored to task managers in developing-country agencies, NGOS, national agricultural research enhancing biodiversity in institutions and international development agencies programs, agricultural extension agricultural development Create rich landscape mosaics with Develop an understanding of factors that motivate Research: universities, national agricultural a variety of habitats for wildlife farmers to retain or destroy buffer strips and research programs; implementation: as well as crop and livestock woodlots or to plant a wider variety of crops; agricultural extension production suggest policies to promote landscape mosaics Develop understanding of how to Prepare training manuals and packets for Multilateral development banks, bilateral aid agencies, implement policies and sound human resource development intemational agricultural research centers, practices to enhance biodiversity NGOS, agricultural extension agencies in agricultural development Approaches to conservation of agrobiodiversity '9-~ Achieve appropriate balance Develop methodology lor estimating International agricultural research centers (especially between in situ and ex the value of in sitL conservation of International Plant Genetic Resources Institute), situ approaches by crop genetic resources as a basis for universities, Food and Agriculture Organization allocating limited funds Create incentives for in situ Identify and implement a new system of intellectual Food and Agriculture Organization conservation of crop and livestock property rights applicable to traditional varieties genetic resources and breeds to benefit local people Create incentives to maintain Explore the possibility of recognizing intellectual Food and Agriculture Organization materials in ex situ collections property rights for some genetic materials in through recognition of their collections with royalties to be paid to gene tangible economic value banks and groups that provide the germ plasm 238 Toward a Strategy for Mainstreaming Biodiversity in Agricultural Development sector responsibility. Experience has shown that and animal materials collected, particularly in the reliance on parks and reserves alone to conserve bio- tropics. This shortage stems in part from the pull of diversity does not work because local populations glamorous fields, such as microbiology and biotech- often do not receive any benefits from them. Indeed, nology. Development organizations and donors could they may even suffer damages from wildlife-damag- help fill this need by providing: ing crops or livestock or be denied access to resources * Support to expand the room for botanical and zoo- they once depended on for part of their subsistence or logical collections in museums, universities, and income. other research-oriented institutions in developing To counteract the reliance on formal protected countries areas, community empowerment and management of * Scholarships for students from the developing natural resources have become a new rallying cry world to undertake taxonomic training in both among donors and organizations that promote conser- developing and industrial countries. vation and rural development. The decentralization of control over natural resources can bring many bene- Rapid Agrobiodiversity Assessment Teams fits, but "community" management alone may not work in all cases. Many development organizations, including the Both protected areas and community or local ini- World Bank, have procedures for screening projects tiatives to manage and conserve biodiversity warrant for their environmental impacts before approval. For continued support, but they are not enough. A third, the most part, such screening focuses on the off-site parallel approach is called for, particularly in the case impacts of agricultural development and assesses of agrobiodiversity: greater involvement of the private whether the project is likely to lead to the loss of for- sector. Where possible, market forces can and should est or other "natural" habitats. But surveys of agro- be hamessed to conserve and better use biodiversity. biodiversity are also needed before a development project is implemented. Tapping farmer knowledge. As a matter of policy, An assessment of biodiversity before a project is all agricultural research programs and development approved and implemented would encompass two projects should include components that seek to main topics: biodiversity in relatively undisturbed incorporate farmers in the design and implementation habitats and agrobiodiversity. For both situations, bio- of research and development efforts. The benefits of diversity indicators would need to be established and such collaboration far outweigh the "costs" of taking their significance for decisionmaking assessed. Rapid the time and trouble to improve the channels of com- assessment methodologies have been tried and tested munication between the research establishment and more fully with the first category. Much work remains the intended beneficiaries: farmers and livestock man- to be done in designing methodologies for adequately agers. assessing agrobiodiversity. Indigenous knowledge has made specific contribu- tions to sustainable agriculture: Biodiversity in relatively undisturbed habitats. The * Crop varieties adapted to harsh growing conditions implications of the agricultural development project * Cropping pattems that minimize the buildup of for the expected shift in land use patterns would need diseases to be analyzed, paying special attention to the likeli- * Breeds of livestock that tolerate poor feed or resist hood that the project might exert greater pressure on diseases and pests that afflict imported breeds remaining relatively intact habitats. Using geographic * Intimate knowledge of botanical resources in for- information system techniques, one could begin est and other habitats that could provide leads for delineating habitats with minimal disturbance where plant domestication or other uses. much of native biodiversity is concentrated. For the different "natural" habitats, an inventory of Greater support for systematics. The task of con- the plant and animal species would be called for, not- ducting in situ conservation and agrobiodiversity sur- ing any species with restricted distributions (this is veys prior to agricultural development will be made already part of the environmental assessment process) much easier if competent specialists are available to and identifying any wild populations of domesticated identify and sort materials. Currently, not nearly plants and animals that could be of special interest to enough taxonomists are available to analyze the plant breeders. Efforts should be made to establish whether Jitendra Srivastava, Nigel Smith, and Douglas A. Forno 239 spontaneous introgression is occurring between for producing "clean" produce but also for offering weedy populations and crops, because such exchange a diverse array of fruit and leaf types will help to of genes often improves the hardiness of crops. generate heterogeneity on agricultural landscapes. Agrobiodiversity. A survey is warranted to docu- Quality pricing is especially important for fruits, ment: vegetables, coffee, and cacao because a discrimi- * The diversity of crops cultivated by local farmers nating pricing structure would signal farmers that * Varietal diversity of each crop some of the lesser-known varieties also have mar- * The number of livestock species raised ket value anid that it is worth their trouble to * The number of breeds per livestock species process and handle produce with care. * Assessment of land use practices on soil biodiver- * Open credit to nonrecommended varieties and sity as an indicator of soil "health." livestock. For the most part, farmers can only obtain credit for government-approved varieties or Performance indicators. Performance indicators "super" breeds. Such discrimination accelerates can serve as "warning bells" when agricultural prac- the process of abandoning traditional varieties and tices threaten to impair the integrity of both "natural" breeds and contributes to genetic erosion. and cultural habitats. Dangers signs include: ' Continue to encouruge the reductio of farming * Natural habitat loss subsidies. This process is well under way in many * Habitat fragmentation countries, in part because of international trade * Species loss even when natural habitat remains agreements. ][t should be encouraged because agri- intact cultural subsidies often trigger biodiversity loss. * Decline in biodiversity of crop species on-farm Such losses occur because farmers often use more * Decline in biodiversity within species. purchased inputs, such as environmentally damag- The following is a sample of remedial measures ing pesticides and fertilizers or irrigation water, that can be adopted to address the loss of biodiversity than they would if they had to pay market prices. associated with agricultural development: * Minimize habitat fragmentation by providing A new agricultural research and development wildlife corridors along "bridges" of natural habi- model. The old research model emphasized maximiz- tat ing output, almost at any cost. Also, research tended * Shift to integrated pest management strategies, to focus on commodities rather than on production such as crop rotation and reliance on biocontrol systems as a whole. The interrelatedness of the parts agents, to check crop and livestock pests of farming systems was neither appreciated nor * Eliminate fiscal and regulatory measures that pro- understood, with sometimes adverse affects on the mote homogeneity in crop and livestock produc- landscape and peoples' livelihoods. The new vision tion for agricultural research adopts a more holistic * Support research on traditional varieties that can approach that is sensitive to environmental concerns, achieve high yield while still addressing the need to boost yields and * Support research on modern varieties that are incomes of rural producers and caretakers of the land. genetically resistant to pests and diseases. The evolving agricultural research paradigm More work is needed, however, to operationalize includes, but is not restricted to: an indicator matrix, particularly with regard to * Adopt integrcated pest management. IPM strategies methodologies for measuring indicators. include the release of biocontrol agents, deploy- ment of genetically resistant cultivars and breeds, Adjustment of the policy environment. A wide more judicious use of pesticides and herbicides, range of policy areas are involved, such as credit, alteration of cropping patterns to thwart the build- trade, intellectual property rights, and land tenure. up of pests arid diseases, and placement of greater The following policy levers would help conserve and emphasis on crop rotation, where economically improve the use of agrobiodiversity: feasible, to retard soil degradation and reduce pest * Promote quality pricing. If farmers receive premi- pressure. um prices for unusual, but attractive, varieties of * Use a participatory approach with farmers. Two crops, they will be encouraged to grow them. A types of on-fann research are typically found: price grading system that rewards farmers not only demonstration plots on farmers' land and experi- 240 Toward a Strategy for Mainstreaming Biodiversity in Agricultural Development mental work that involves farmers and other stake- ty of soil micro-organisms, which are so important holders in the design of models from the "ground in nutrient recycling. up." Much more of the second type of on-farm * Focus research on lifetime and herd productivity research is needed that involves farmers, pastoral- characteristics. Deterministic simulation models ists, and other "clients" of agricultural research and live animal experimentation can be used in and development from the inception of the study some cases to achieve this goal. design. In this manner, research would be more * Determine the critical number of breeds for con- demand driven. servation purposes. Analysis of genetic variation * Improve the use of indigenous knowledge. How can be used to highlight the genetic "spacing" and why local people use natural resources can between breeds and to identify those breeds that provide important information for more appropri- are significantly different or unique from others. ate agricultural research and development efforts. * Learn more about genetic components of adapta- * Support research, development, and dissemination tion in livestock. A better understanding of traits of lesser-known crops and animals. Neglected tra- such as resistance to ticks and use of body reserves ditional varieties and breeds, many of which are would aid breeding efforts and would likely under- particularly well suited to difficult environments, score the importance of safeguarding so-called would be included in this broad research effort. "minor" breeds. But sustained support for research on the major food and industrial crops as well as livestock Institutional development. The notion of a new remains essential. research paradigm has implications for developing * Support research on new crops and livestock. new institutional arrangements and exploring new Scope exists for new crops and livestock to fill ways of doing business. Innovative institutional specialty market and environmental niches. In arrangements would include more effective partner- some cases, natural vegetation communities could ships between: be managed for the production of new domesticat- * Agricultural research centers and NGOS ed animals. A deeper comnitment to research on * Agricultural research centers and growers' associ- crop and livestock candidates would thus under- ations score the value of conserving biodiversity and nat- * Agricultural research centers and private compa- ural habitats. nies involved in the manufacture and sale of agri- * Be sensitive to the value of a mosaic of land uses. cultural technologies Even land uses that are "desirable" from the biodi- * Universities and agricultural extension agencies versity viewpoint can be promoted too far. * Development lending institutions and all of these Biodiversity in managed landscapes is often best entities. served by promoting a mixture of land uses that To some degree, all these connections are being provides varied habitats for wildlife adapted to explored and tested. Results are mixed. Initial suspi- altered areas. cion between some national agricultural research cen- * Achieve a greater diversity of habitats within land ters and NGOS has gradually subsided, while in other use systems. Biodiversity within a land use system, cases misgivings about interinstitutional collaboration such as intensive cereal cropping, can be achieved have been borne out by experience. Links between by allowing for a variety of habitats, such as ripar- agricultural extension agencies and generators of ian buffer strips, shelter belts, windbreaks, strip technology are, with few exceptions, weak. To facili- cropping, and wetlands. Diversity of habitats on tate future collaborative ventures in the management the landscape creates more niches for wildlife, of agrobiodiversity, organizations' roles and responsi- some of which are beneficial in controlling crop bilities will have to be established more clearly at the pests. More diverse habitats, including managed outset. ones, also promote the more efficient use of nutri- ents and create microclimates that buffer crops References from inclement weather. * Recycle organic matter. Measures such as incor- The word "processed" describes informally repro- porating livestock or green manure-no-till or duced works that may not be commonly available minimum-till -farming-help to sustain the diversi- through libraries. Jitendra Srivastava, Nigel Smith, and Douglas A. Forno 241 Bohac, J., and Pokarzhevsky. 1987. "Effect of Tropical Forests." The Environmentalist 8(autumn): Manure and NPK on Soil Macrofauna in 1-20. Chemozem Soil." In J. Szegi, ed., Soil Biology and Pimentel, D., U. Stachow, D. A. Takacs, H. W. Conservation of Biosphere, 2 vols. Proceedings of Brubaker, A. R. Dumas, J. J. Meaney, J. A. S. the ninth international symposium. Budapest: O'Neil, D. E. Onsi, and D. B. Corzilius. 1992. Akademiai Kiado. "Conserving Biological Diversity in Agricultural! Dalrymple, D. G., and J. P. Srivastava. 1994. Forestry Systems." Bioscience 42(5):354-62. "Transfer of Plant Cultivars: Seeds, Sectors, and Ricklefs, R. E. 1987. "Communily Diversity: Relative Society." In J. R. Anderson, ed., Agricultural Role of Local and Regional Processes." Science Technology: Policy Issues for the International 235:167-71. Community, pp. 180-207. Wallingford, U.K.: Srivastava, Jitendra, Nigel J. H. Smith, and Douglas C.A.B. International. Forno. 1996. Biodiversity and Agriculture: Defoliart, G. R. 1989. "The Human Use of Insects as Implications for Conservation and Development. Food and as Animal Feed." Bulletin of the Technical Paper 321. Washington, D.C.: World Entomological Society of America 35(1):22-35. Bank. Diamond, J. M., K. D. Bishop, and S. Van Balen. Thrupp, L. A., B. Cabarle, and A. Zazueta. 1994. 1987. "Bird Survival in an Isolated Javan "Participatory Methods in Planning and Political Woodland: Island or Mirror?" Conservation Processes: Linking the Grassroots and Policies for Biology 1(2):132-42. Sustainable Development." Agriculture and Eisner, T. 1990. "Prospective for Nature's Chemical Human Values 11(2-3):77-84. Riches." Issues in Science and Technology Tilman, D., D. Wedin, and J. Knops. 1996. (National Academy of Sciences) 6(2):31-34. "Productivity and Sustainability Influenced by Greathead, D. J. 1983. "The Multi-million Dollar Biodiversity in Grassland Ecosystems." Nature Weevil That Pollinates Oil Palm." Antenna (Royal 379(6567):718-20. Entomological Society of London) 7(3):105-07. Whitmore, T. C. 1985. Tropical Rain Forests of the Lal, Rattan. 1991. "Soil Conservation and Far East. Oxford: Clarendon Press. Biodiversity." In D. L. Hawksworth, ed., The Wilcox, B. A., and K. N. Duin. 1995. "Indigenous Biodiversity of Microorganisms and Invertebrates: Cultural and Biological Diversity: Overlapping Its Role in Sustainable Agriculture, pp. 89-104. Values of Latin American Ecoregions." Cultural Wallingford, U.K.: C.A.B. International. Survival Quarterly 18(winter):49-53. Metcalf, C. L., W. P. Flint, and R. L. Metcalf. 1962. Wilson, E. 0. 1988. "The Current State of Biological Destructive and Useful Insects. New York: Diversity." In E. 0. Wilson and F. M. Peter, eds., McGraw-Hill. Biodiversity, pp. 3-18. Washington, D.C.: Myers, N. 1988. "Threatened Biotas: 'Hotspots' in National Academy Press. 18 Integrated Pest Management: Strategies and Policies for Effective Implementation Tjaart W. Schillhorn van Veen, Douglas A. Forno, Steen Joffe, Dina L. Umali-Deininger, and Sanjiva Cooke Generating the gains in agricultural productivity nec- including yield, profits, risk, sustainability, safety, and essary to secure food availability and livelihoods in pest population dynamics. IPM is a key component of the developing world over the coming decades integrated farming practices that are based on an requires an approach in which the intensification of understanding of ecology and the interaction between agricultural systems is consistent with conservation of crops or animals and their pests as well as an under- the natural resource base. This approach requires less standing of the environments in which pests operate. reliance on the intensive use of external inputs and An essential aspect of IPM is its integration of tcCll- greater dependence on management skills and loca- nical and social knowledge. This integration requires tion-specific knowledge of agroecosystems. Inte- a sound understanding of key pest constraints and bio- grated pest management (IPM) constitutes one such logical and farm management systems. These are approach. often highly location-specific, and farmer participa- The IPM concept is far from new. Farmers used tion and networking are essential in the design of integrated pest control long before scientists coined modern science-based IPM schemes. The major limit- the term. In traditional farming systems, pest manage- ing factor, however, is insufficient backup of farmers ment is inseparable from sound farm management. It with adequate and applicable IPM-related technolo- involves farm practices that promote good plant and gies. Another problem is a lack of appropriate farm- animal health and keep pest losses in check without focused research. the use of manufactured chemicals. Changes in farm- Development agencies, notably bilateral programs ing systems during the past half century lost sight of sponsored by Germany, the Netherlands, and the this approach, and chemical control methods became United States, as well as those sponsored by the Food the pillar in the control of pests and diseases in mod- and Agriculture Organization, have been advocating em agriculture. During the past decade, however, and supporting efforts to implement IPM for more than growing concerns about the risk and negative effects 20 years. The World Bank is now a significant IPM of chemical methods have spurred agriculturists, supporter and has invested some $80 million in imple- environmentalists, and economists to explore pest mentation projects since 1988. However, despite these management strategies that have fewer side effects on efforts and notable successes in some countries, wide- public health and the environment. The most well- spread implementation of IPM remains an elusive goal known among these strategies is IPM. in most parts of the world. IPM can best be described as a farmer's knowledge If IPM is to become widespread, farmers must have and use of an optimal mix of pest control tools and the appropriate incentives, relevant knowledge, and tactics, taking into account a variety of factors, practical techniques to make use of nonchemical- 242 Tjaart W. Schillhorn van Veen, Douglas A. Forno, Steen Joffe, Dina L. Umali-D)eininger, and Sanjiva Cooke 243 based approaches. To apply IPM, farmers need to choices and may especially benefit from cost-effec- accept a practice that is usually more management- tive IPM approaches. and labor-intensive than the use of chemical agents. Because agricultural pests are rarely limited to sin- Hence fanners will need to see a demonstrable eco- gle farms but occur regionally or beyond, there are nomic payoff. Ultimately, the choice of pest manage- limits to farmer-centered approaches. Most pest con- ment technology will be influenced by the costs, ben- trol programs also need a regional or national base as efits, and availability of competing alternatives, as a guide, as exemplified by national and regional quar- well as by any rules or other social norms governing antine regulations, by the highly successful eradica- their use. tion of screwworm disease in the United States and Governments influence the prospects for wide- more recently in northern Africa, and by the beneficial spread implementation of IPM through the incentive effects of nationally or regional]ly harmonized pesti- structures and regulations affecting the choice of pes- cide use. ticides or alternative approaches. Government influ- ence is also strong through support for research, Pest Management, Farmers, and the Chemical extension, education, and training initiatives. This Industry dimension of IPM implementation, including the influence that donor organizations can have on gov- The profit motive of the chemical industry has been ermnent polices, is one of the most important but partly responsible for the detrimental environmental least-documented aspects of the challenge facing the and public healt h effects of pesticides. Initial increas- international community in implementing the recom- es in pesticide use after World War II were fueled by mendations of the United Nations Conference on extra supplies and industrial capacity. In later years, Environment and Development (UNCED) on IPM. the industry has been responding to the demand, albeit Consequently, this chapter reviews how govern- somewhat manipulated, of its clients. Farmers ments can support and encourage the implementation embraced pesticides because of labor savings, increased of IPM and how the World Bank and other develop- crop security, higher quality and homogeneity of prod- ment agencies can assist this process. The first section uct, and persuasive messages from research, exten- considers the factors that affect farmers' choice of sion, and industry advertising activities. Consumers pest management technology. The second section increasingly demanded visibly blemish-free agricul- reviews issues related to development strategies and tural and horticultural products; the industry respond- government policies and shows how many countries ed by developing an array of pest control chemicals.' have created a policy environment that favors the use Consumers, however, are becoming increasingly of chemical pesticides. The third section assesses the interested in food safety and are demanding whole- policy instruments that governments can use to create some products and production practices that are not a more level playing field for alternative pest man- detrimental to thle environment. agement practices and to promote IPM where appro- Just as public attitudes have changed, so too has priate. The fourth section discusses national IPM poli- the pesticide industry. Insecticide use in agriculture, cy and the need for more farmer-centered research, initially the major culprit in public health concerns, extension, and training. has been stagnant or declining in the past two decades in industrial countries (National Research Council Farmers and Pest Management 1989), although herbicide use has increased. "Cure- all" drugs and agrochemicals are very slowly being Farmers have available a range of pest management replaced by target-specific chemicals, albeit at higher techniques, from natural controls based primarily on costs. Current research in the agrochemical industry is cultural, physical, and mechanical techniques to the increasingly directed at highly specific interventions use of biological control agents or chemical pesti- and at disease- or weed-resistant genetics (as well as cides. In simple terms, the choice of pest management herbicide-resistant crops that could increase future technique is a function of costs (purchased inputs, herbicide use). Following stricter regulation, as other variable costs such as labor, and fixed costs such demanded by society, considerable improvements as the cost of sprayers or information) and returns have been made in the safe handling of pesticides, (labor-saving prevention of crop loss in monetary or although this has been applied more widely in the subsistence terms). Poor farmers are limited in their industrial world than in the developing world. 244 Integrated Pest Management: Strategies and Policies for Effective Implementation Economic Threshold in Pest Management Decisions a varied and ecologically well-balanced mnix of crops (and livestock and wildlife) helps to limit the losses. The concepts cf action thresholds and economic The past half century has seen a worldwide trend thresholds are often used, albeit in a simplified way, toward more intensive and often mechanized agricul- to explain decisionmaking in pest management. Thus tural systems that increase productivity and food sup- Stern (1973: 261) defines the action threshold level as plies. "the population density at which control action will result in little likelihood of the pest population Agricultural Development Strategies: exceeding the economic threshold level." The eco- Changing Incentivesfor Pesticide Use nomic threshold level is the "break-even point at which the value of loss in yield quantity or quality is Production intensification, however, has been associ- equal to the cost of a control method that successful- ated with a trend toward monoculture and chemical- ly eliminates pest damage and yield loss." The action based approaches to pest management. One side and economic thresholds overlap, but they are not effect has been to create conditions conducive to the necessarily the same; farmers' actions may derive rapid growth of pest populations and to increase the from a lack of specific knowledge (for example, "see- risk that such populations will reach epidemic propor- ing many bugs" rather than assessing potential dam- tions. Formal agricultural research and extension sys- age) or from advertising campaigns that promote use tems have responded with the development of high- at levels where costs may be higher than benefits. Too yielding plant (and livestock) varieties, combined often farmers, because they have only limited under- with technology packages of pesticide- or drug-based standing of pest ecology, assume that all bugs are approaches to counteract losses. Although an effective harmful. and economic course of action in the short term, the In some cases it may indeed pay to take action use and continuing development of cheap and quick- very early against emerging pest populations, thereby acting pesticides and veterinary pharmaceuticals may preventing later buildup. In other cases it may not pay exacerbate the long-term pest problem by eliminating to take any action at all; many farmers may underes- the population of pest predators ("good bugs," such as timate the recuperative ability of their plants and ani- ladybirds, earthworms, and birds) that help to keep mals. Cotton plants, for example, can compensate populations of undesirable bugs in check. Pest preda- well for early defoliation by leafworms; yet Sudanese tors thus improve the natural defense systems of cotton growers still spray early, inadvertently elimi- plants and animals. nating beneficial pests and thus creating subsequent Less emphasis has been placed on maintaining problems with aphids and whiteflies. genetic diversity, including the maintenance of bene- Uncertainty about the degree of crop losses caused ficial organisms, and on breeding for disease and pest by pests creates output risks for farmers. The manner resistance in plants and animals. A widely pursued in which farmers respond to these risks (or their development policy of industry-led growth has been degree of risk aversion) in turn influences pest man- accompanied by the taxation of agriculture (keeping agement activity. If they perceive the level of pest food prices low), which has been partly compensated infestation and damage to be uncertain, farmers may through input subsidies (including pesticides). Rising apply pesticides at a lower threshold level to reduce labor costs associated with the commercialization of the risk of crop losses. This decision contributes to agriculture have discouraged further the use of more more frequent applications and greater total volume labor-intensive traditional pest management practices. applied. Uncertainty about the effectiveness of a par- ticular technique could also lead to increased pest Governments, Donors, and Excessive Use management activity to reduce risk at the margin of Chemical Pesticides (Feder 1979). The increased dependence on pesticides fostered by Policy Issues in Pest Management agricultural development strategies has led over time to the entrenchment of a chemical mode of pest con- In traditional farning systems a dynamic equilibrium trol as the dominant paradigm in pest management. (or endemic stability) exists between many pests and Today, many developing-country governments have beneficial organisms. A degree of loss is accepted but incentive frameworks that favor pesticide use over the Tjaart W Schilihorn van Veen, Douglas A. Forno, Steen Joffe, Dina L. Umali-Deininger and Sanjiva Cooke 245 adoption of more environmentally benign iPM vided through grants or subsidized prices, permits approaches. governments to distribute pesticides to farmers at low Pesticide use tends to be promoted through both or no cost. Moreover, lack of information on alterna- direct and indirect government and industry measures tive, more environmentally benign approaches to pest (table 18.1) and through the mechanism of prices management often biases extension service toward (Waibel 1994 and Farah 1994). Direct government pesticide use. price subsidies reduce the farmer's costs of using pes- Governments indirectly subsidize the cost of pesti- ticides and encourage increased use. Weak or inap- cides through overvalued exchange rates that make propriate legislation and regulations and their inade- pesticide imports more available, through public sec- quate enforcement contribute further to improper use. tor investments in pesticide research and development Bilateral and multilateral aid for pesticide use, pro- that are not recouped from pesticide users, and Table 18.1. Direct and Indirect Measures Fostering the Use of Pesticides Direct measures Indirect measures Price factors Nonprice factors Price factors Nonprice factors Government Public sector policies, such Indirect public Public sector policies, price subsidies as direct involvement of sector subsidies such as taxation the public sector in marketing Public sector Weak or inappropriate Overvaluation of exchange Technology policy restricting the production and pesticide legislation rates that facilitate entry of more environmentally distribution pesticide imports benign technologies (import ban, lack of intellectual property protection) Cost subsidies to Weak (or lack of) enforcement Pesticide bias in government Weak (or lack of) private pesticide of pesticide-use regulations: research investments environmental policy manufacturers conflicts of interest Pesticide sales Public sector extension service Foreign and domestic direct- Farm input trade guarantee tax exemption promotion of pesticides investment incentives, including for the manufacrure of pesticides Subsidized credit Government pest management Minimum wage policy that Inadequate curricula of and insurance tied policies, such as support discriminates against more agricultural extension to pesticide use for promotion or pesticide labor-intensive and research (pest eradication; environmentally benign specific legislation regarding production techniques prophylactic treatments) Preferential Prochemical bias of industry High import duties import duties information and extension on alternative technologies and exchange rates advice for pesticide imports Plant protection Strong pesticide lobby outbreak budget of interest groups Donor price subsidies Donor pesticide grant or aid at subsidized costs Source: Adapted from Waibel 1994. 246 Integrated Pest Management: Strategies and Policies for Effective Implementation through special foreign and domestic investment organisms. These hazards give rise to negative exter- incentives (for example, duty-free imports of capital nalities that impose costs on society and the environ- goods and tax incentives) that extend to pesticide ment. But these costs are rarely reflected in the price manufacturing. Policies that create barriers to the that users pay, and other forms of compensation are entry of new technologies (for example, lengthy, cum- not applied. In this situation there is divergence bersome, and costly registration procedures for seeds between the private costs to the user and costs to soci- of improved disease-resistant plant varieties or lack of ety, the true "social costs." Thus the level of pesticide intellectual property protection) also limit farmers' use is not optimal from society's perspective, and pest management technology options. Banks may intervention is warranted. also encourage pesticide use by linking credit to cer- A number of policy and regulatory instruments are tain agronomic practices, including the use of pesti- available to governments to encourage environmental- cides (Thrupp 1990). ly sound and economically rational pest management National governments sometimes also use pesti- practices. The more important of these instruments cide regulation to protect their own pesticide industry. are outlined here (their implementation is discussed In some cases such protection is provided to outdated later): manufacturing processes or to unsafe manufacturing * Development of a system that increases the aware- plants or products, which prevents new and safer ness of policymakers, consumers, and producers of products from entering the market. Pesticide indus- the hazards of pesticide use tries often use governments to expand the market for * Development of a regulatory framework to ensure their products by encouraging them to promote tech- appropriate and safe production, distribution, and nology packages that combine seed, fertilizer, and use of pesticides pesticide; by supporting applied research using pesti- * Reorientation of agricultural and environmental cides; or by supporting extension services that policies to introduce appropriate economic incen- endorse pesticide use in general. tives, including taxes and special levies on pesti- cide use, to account for negative externalities, and Integrated Pest rManagement short-term subsidies, to account for positive exter- and Public Policy Tool Box nalities in the use of IPM - Orientation of research and technology policies to A reappraisal of the role of pesticides in agriculture generate a steady supply of relevant pest manage- has been under way since the publication of Rachel ment information and technologies, including ade- Carson's Silent Spring 40 years ago (Carson 1962 and quate budget allocations for research, extension, Farvar and Milton 1972). But only in the last few and training years has the topic shifted from technological solu- * International agreements and conventions. tions (for examp]Le, better protection of workers and consumers, classification and phaseout of the most Information Programs and Education dangerous chemicals) to the mainstream of govern- ment and donor agricultural policy agendas. This has Many of the policies that encourage excessive pesti- been associated with a wider shift in focus from agri- cide use were developed at a time when little infor- cultural production per se to environmentally sustain- mation was available on the health hazards of pesti- able management of production systems, as reflected cide use and on sustainability. Greater awareness of in the UNCED Agenda 21 agreements. these issues among policymakers, producers, and con- Government intervention to influence farmers' sumers, which can be achieved by improving the flow choice of technology between pesticide- and non- of information, encourages reducing pesticide use by pesticide-based approaches can be justified by the helping to make decisions that better reflect true environmental and public health implications of pes- social costs. Informed consumer preferences can ticide use. These include harmful effects on public become a powerful factor influencing pest manage- health, ranging from acute poisoning and a high risk ment practice. Informed consumers can make their of cancer among pesticide-handling farmworkers to preferences clear through pressure on public authori- contamination of food with pesticide residues (see ties to regulate the development, handling, and use of chapter 19). There are also wider environmental pesticides; through the market, by paying a premium effects on water, soil, and a variety of nontargeted for pesticide-free produce; and directly through dia- Tjaart W. Schillhorn van Veen, Douglas A. Forno, Steen Joffe, Dina L. Umali-Deininger, and Sanjiva Cooke 247 logue with producer groups. Public investment in education and communications media to provide Box 18.1. Pesticide Deregulation and unbiased information on the benefits and risks of pes- Development of Resistance irn China and India ticide use is necessary as part of a general effort to level the playing field between pesticide dependency Dramatic declines in cotton production in China in and IPM. the past few years have occurred as major pests, such as the cotton bollworm (Heliothis armigera), have Regulation become resistant to many insecticides. The economic losses have been enormous. In some counties of the Establishment of a pesticide regulatory system is the Yellow River valley, yields have declined 50-60 per- primary means used to ascertain pesticide quality, to cent through a combination of resistance and drought. minimize exposure, and to determine a proper balance In Hebei, Henan, and Shandong, overall yields have between the risks and benefits of using agrochemi- been reduced 310-50 percent. In addition, the number cals. The major objectives are to (a) protect pesticide of sprays used has risen from between 8 and 10 to users and handlers, (b) protect public health, (c) pre- between 15 and 25, and the seasonal cost of sprays vent negative ecological effects, (d) prevent the build- has doubled or tripled over a short period. In the up of pesticide resistance in target and nontarget pop- worst-affected areas, cotton production has become ulations, and (e) protect beneficial pests. Appropriate completely uneconomic. A report by the Intemational regulatory frameworks are necessary for pesticide Organization of Pesticide Resistance Management regulation; this is an area where deregulation is mis- (IOPRM 1993) finds that the single most important fac- placed (see box 18.1). tor behind the development of resistance was the Regulatory systems operate through a variety of unregulated pesticide market. compliance provisions and mechanisms (table 18.2). A similar picture has emerged in India. Here Regulatory authorities may restrict, tax, or ban the deregulation has led to an increase in the number of production, import, sale, or use of certain pesticides. smaller companies selling (subsidized) pesticides Use criteria may be determined by geographical with no effective oversight of the product quality or zones (for example, sensitive habitats), according to of the advice given to farmers. Throughout India, different classes of user (for example, institutions or there are more than 86,000 dealers. Under these cir- individual farmers), or by commodity-production cumstances of uncontrolled marketing of pesticides, it system and pesticide resistance management. is likely to be very difficult, if not impossible, to Introduction of such a regulatory framework requires implement an 1PM policy (Gillham and others 1995). an understanding of the feasibility of various regula- tory options under prevailing conditions in a particu- lar country. Economic and Legal Disincentives A pesticide registration scheme requires manufac- turers to demonstrate, by providing relevant data, the One approach to internalizing the negative externalities safety and efficacy of specific pesticides for particu- associated with pesticide use is called the polluter-pays lar uses when used according to the registered principle. Acccrding to this principle, polluters should instructions. Often allied to the registration process is be made to pay the costs of the damage they inflict on a requirement for accurate labeling, including infor- society. In the case of pesticides, strict implementa- mation about safe use of the product and potential tion of polluter payment is complicated by the diffi- hazards and associated risks of individual pesticides. culty and the cost of clearly identifying who the pol- This requirement usually carries over to advertising luters are, to what degree each user contributed to the and extension messages. Monitoring and evaluation damage, and the need to estimate the optimal level of specific products may also be required. Through a of pesticide use and the associated optimal tax. licensing and certification scheme, pesticide handlers Using the polluter-pays principle to control pesticide may be required to demonstrate adherence to set use is being seriously considered in the European standards and completion of relevant training. The Union and elsewhere (Bonny and Charles 1993). registration of pesticides in the United States and Implementation of the polluter-pays principle has Europe is often used as a baseline for suitability of included the imposition of licensing fees, user fees, use in other countries. and a sales tax. These measures can be tailored to the 248 Integrated Pest Management: Strategies and Policies for Effective Implementation Table 18.2. Common Pesticide Regulatory Activities Type of regulation Enforcement process Target grouip Pesticide registration Quality control of new Manufacturers, formulators, agrochemicals; risk assessment and distributors of new pesticides Pesticide reregistration Continuous risk assessment Manufacturers, formulators, (of new formulations of existing agrochemicals and distributors or other use) Bans on unsafe pesticides Banning of sales and Manufacturers, formulators, manufacturing distributors, and end users (no indemnification for the farmer) Food safety Development of quality Producers, processors, standards; monitoring of food market chains, and researchers quality; prosecution of violators Pesticides in groundwater Monitoring of application and Producers and manufacturers groundwater residue; prosecution of violators Storage and disposal of pesticides Certification and inspection; Producers, distributors, and prosecution of misuse pesticide applicators User or farmworker safety Certification and training Producers, applicators, and farmworkers Resistance prevention Voluntary or mandatory Producers, applicators, participation and distributors characteristics of individual pesticides. Such levies adopters who are paving the way for more widespread are usually further supported by penalties for non- adoption through their demonstration effect; for exam- compliance, negligence, and failure to meet relevant ple, an initial matching grant to farmer groups may ini- standards. Overall, taxation of pesticide use offers a tiate a collective pest scouting scheme or other means more flexible tool than the all-or-nothing pesticide of enhancing pest monitoring or farmer networks. registration requirement (that is, the pesticide is either approved for use or not), but taxation requires Research and Extension more sound (and probably costlier) enforcement. A bias in favor of pesticides in the funding of research, Economic Incentives extension, and training in many countries is a response to distorted economic incentives that artificially boost Positive incentives could also be used to promote the returns to pesticide-related research. This bias also environmentally benign pest management practices. reflects the economic strength of pesticide industries, These incentives include support for alternative prac- both domestic and foreign, that have been able to pro- tices, either through temporary grants and tax breaks mote and support the use of proprietary packages. One or through research and extension. For example, to means by which governments can encourage a social- encourage the adoption of environmentally friendly ly nearly optimal mix of research and related down- technologies, governments could set up short-term, stream activities is the judicious use of the economic time-bound programs that provide matching grants to and regulatory provisions outlined above. This would farmers who adopt IPM practices. This subsidy may be have the effect of bringing the costs of pesticide use economically justified because it rewards early closer to the social costs, thus stimulating demand for Tjaart W. Schillhorn van Veen, Douglas A. Forno, Steen Joffe, Dina L. Umali-Deininger, and Sanjiva Cooke 249 Table 18.3. Examples of Successful Integrated Pest Management Applications Crop Country Impact Soybean Brazil Pesticide use on soybeans reduced 80 percent in seven years Rice India (Orissa) Insecticide use on rice reduced Rice Indonesia Pesticide use on rice reduced more than 60 percent Cotton Nicaragua Pesticide use reduced one-third during the 1970s Various Cuba Pesticide use reduced 80 percent during the 1980s Source: Postel 1987. alternative techniques including IPM (table 18.3 shows Implementation Strategies for IPM some examples of a successful switch from pesticides to IPM). It would also encourage industry to invest in Experiences from Africa, Asia, Latin America and the the supply of environmentally benign technologies. Caribbean, and the United States (California) are yielding valuable lessons about implementation of International Agreements and Conventions IPM-based programs in developing countries (Teng and Heong 1988, Wiebers 1993, Kiss and Meerman Since the early 1980s the United Nations system has 1991, and Murray and Nathan 1994). These experi- taken the lead in the establishment of international ences suggest that there are two key strategic ele- agreements covering pesticides. The Food and ments in implementing IPM on1 a wide scale: (a) cre- Agriculture Organization, for example, has issued a ation of a level playing field for different modes of series of detailed guidelines on the procurement, pest management, which means eliminating policies trade, and handling of pesticides under the umbrella that promote environmentally unsustainable pest of the International Code of Conduct on the management techniques and strengthening regulatory Distribution and Use of Pesticides (FAo 1990). These institutions and (b) implementation of positive mea- guidelines cover all aspects, including regulatory pro- sures to promote IPM through support for public visions and industry practice. Acceptance of the code awareness, research, extension, and training, with an has been slow, often because implementation is in the emphasis on decentralized, farmer-centered initia- hands of agricultural plant protection departments, tives. many of which are also involved in the distribution of The first of these measures is necessary to create pesticides. an enabling environment for IPM but is rarely suffi- A relatively recent addition to internationally cient to promote implementation. The second mea- accepted practice is the prior informed consent provi- sure is also necessary but fails to have a significant sion that is being jointly implemented by FAO and impact unless the enabling environment is in place. United Nations Environment Programme. Under the The political commitment and management needs in informed consent procedures, listed products (those terms of policy and institutional change are complex that have been banned or severely restricted in a num- in that they cross sectoral boundaries and need to ber of countries) can be exported to another country cover both raacro- and micro-level initiatives. only if that country has full information and gives Careful sequencing and targeting of specific mea- explicit consent. The original list of 12 chemicals is sures are also needed to minimize the economic being expanded to 39. Currently, 120 countries are costs associated with policy shifts, in particular signatories to the agreement. where this wilil create or strengthen a constituency Another important international standard, estab- opposed to change. lished by the World Health Organization, groups pes- ticides according to their oral toxicity in mammals. National Integrated Pest Management Policy These data provide a reference for relevant standards including the maximum residue levels of allowable Establishment of a national IPM policy framework pesticides in the Codex Alimentarius agreement, provides a useful first step in implementing an IPM which governs the quality of agricultural produce. strategy at the national level. Several countries have 250 Integrated Pest Management: Strategies and Policies for Effective Implementation already moved to do this, such as Cuba (Rosset and Box 18.2. Integrated Pest Management Benjamin 1994), Indonesia (Kenmore 1991), the in Indonesia Netherlands (Vijftigschild 1996), and the United States (U.S. Department of Agriculture, In the early 1980s, extensive areas of rice production Agricultural Research Service 1993). The advan- in Indonesia were infested with the brown plant-hop- tages stem from overseeing the disparate package of per Nilaparvata lugens. It was no longer economical- measures needed to implement IPM, both upstream ly possible to control the pest, despite ongoing resis- policy elements and on-farm implementation, with- tance breeding by the International Rice Research in a single, coherent decision framework. The same Institute and intensive pesticide applications by farm- decision framework can provide the basis for con- ers. sultations with all relevant stakeholders to secure The Indonesian national IPM policy was broad institutional support for what may involve a announced in 1986 by presidential instruction and significant shift (see the example of Indonesia; box supported by the Food and Agriculture Organization 18.2). Finally, government's explicit statement of of the United Nations (FAO; see Kenmore 1991 and intent relative to IPM provides a strong signal to all World Bank data). This followed research evidence relevant actors. that insecticides were not increasing rice yield signif- A broad consultation process using workshops icantly and, by destroying natural enemy populations, and seminars can be used to define targets and pri- were destabilizing the system of rice production. orities and to define appropriate institutional roles Many insecticides were banned; subsidies were and responsibilities for different elements of imple- reduced or removed; crop protection and associated mentation, thus helping to bring the various public, research and extension institutions were strength- private, and government organization stakeholders ened. A major iPM training effort began. By mid-1992 in IPM together within a concerted and open pro- about 1,000 pest observers, 3,000 extension staff, and gram. This process also helps to raise awareness and at least 150,000 farmers had been trained to observe understanding by both consumers and producers and understand the local ecology of the planthopper about the health and environmental risks associated and its natural enemies and to follow simple threshold with pesticides and to create a broad constituency in rules. favor of implementing IPM and mobilizing public During 1987-90, the volume of pesticides used funds to support implementation. on rice fell more than 50 percent, while yields increased about 15 percent. Farmers' incremental net Information, Analysis, and Priorities profits are estimated at $18 per IPM-trained farmer per season. This compares with an estimated average Policies that support IPM will increase aggregate training cost of about $4 per farmer. The federal gov- social welfare, but they may also negatively affect the emnment has benefited from a $120 million a year (85 productivity and livelihoods of some groups in the percent) reduction in pesticide subsidies and has short term. Thus Zilberman and others (1991) find begun to realign these funds into further IPM initia- that bans on pesticide imports, when no provisions tives. In 1993 the government of Indonesia, the have been made for effective substitutes, result in World Bank, and the U.S. Agency for International lower production levels, higher prices, a substantial Development committed $53 million to a project that loss of discretionary income for consumers, and a will build on and extend the use of IPM throughout redistribution of income toward agricultural produc- the country. ers who do not use pesticides. To achieve a smooth This Indonesia project is an often-quoted exam- transition to environmentally sound pest management ple, but India's irrigated rice ecosystem is also well practices and to maintain progress toward wider agri- suited to natural pest control. Complete IPM cultural sector growth targets, some targeting and approaches in other production systems and ecosys- phasing of policy changes are necessary. tems may require more information and different A first step is to determine the status and availabil- approaches. However, the field school and farmer ity of suitable IPM technologies and the prospects for training and empowerment concept has positive farmer adoption. This determination will be specific to effects beyond pest control. the system of production and requires field data based on a multidisciplinary assessment. For example, in Tjaart W Schillhorn van Veen, Douglas A. Forno, Steen Joffe, Dina L. Umali-Deininger, and Sanjiva Cooke 251 some high-input systems, direct measures to limit pes- Tobin 1994). Several factors explain this situation. A ticide use, through either the pricing mechanisms or comprehensive and rigorous network of registration, regulation, may be the single most important way to legislation, standards, training, monitoring, and infor- influence technology choice toward IPM. In theory, mation systems appears to be more expensive than agricultural service providers, both public and private, alternative public sector programs with a shorter-term will adjust the supply of information and technologies and more visible direct impact.. These costs are raised to meet farmers' requirements for alternative tech- by limitations of infrastructure and communications. niques. In practice, a variety of market imperfections At the same time, the financial impact of IPM policies, and technical limitations and constraints on the flow of such as lower subsidies, taxes, and licensing and reg- information could slow the pace of any adjustment. istration fees, falls most heavily on producers and A concerted effort to support farmers' transition to marketers of pesticides and initially on farmers IPM through on-farm adaptive research, demonstration, engaged in the production of cash crops, who use the and training can be a powerful and effective adjunct to most pesticides. These are both powerful interest economic and regulatory measures. This has been groups. effective in irrigated rice production systems in Given such constraints, there is a need to: Indonesia and other Asian countries (see box 18.2). * Develop a streamlined regulatory system with low Although such an intensive training effort can be cost- administrative costs (registration procedures, for ly, a growing body of evidence suggests that it yields example, can often use data from other countries high economic returns (Pincus 1994). with comparable or higher standards) Although measures to ensure undistorted pricing for * Ensure that the institutions charged with imple- pest management technologies continue to be impor- menting economic and regulatory measures have tant, a critical constraint to IPM implementation is like- adequate authority to push through and enforce ly to be the availability of effective and competitive relevant measures alternative (nonchemical) techniques. For example, * Seek to recover implementation costs with revenue there remains a shortage of low-cost IPM technologies (for example, through registration fees) that are relevant to the mixed-farming systems preva- * Combine implementation of policies with public lent in much of Sub-Saharan Africa (Kiss and information and education on the benefits Meerman 1991). In the absence of viable alternatives, * Encourage a transparent and open process that any measures that would raise the costs of pesticides includes consultation with industry and other inter- may be perceived as placing unwarranted limitations est groups for possible resolution of conflicting on growth of productivity and livelihoods, with the interests. long-run social benefits being heavily discounted by both producers and policymakers. A relatively sub- Farmer-Centered Research, Extension, and Training stantial investment in the generation of new knowl- edge and practical IPM techniques is therefore essen- The involvement of farmers in helping to generate tial, along with improved infrastructure and services. locally speciiic techniques suitable for particular The aim is to ensure that as farming systems evolve, farming systems appears to be an important factor farmers continue to have access to relevant pest man- determining the success of efforts to implement IPM, a agement technologies that make minimal use of finding that is likely to apply equally to other sustain- chemical pesticides. able production technologies and resource manage- ment approaches. The field implementation of IPM Economic and Regulatory Measures requires that farmers, extension workers, and local crop protection technicians have a practical under- Many developing countries face significant con- standing of th,e ecology and life cycle of major pests straints to implementing the economic and regulatory and their natural enemies and that this knowledge be provisions required to support IPM. Surveys indicate translated into appropriate decisionmaking tools (for that 25 percent of developing countries lack the legis- example, action thresholds) and practical control tac- lation and 80 percent lack the resources to implement tics. or enforce such legislation (FAo 1993). This is espe- Training of farmers, pest observers, agricultural cially the case in Africa, with 76 percent of countries development officers, and extension agents should lacking pesticide-control statutes (FAO 1993 and therefore represent a major component of any imple- 252 Integrated Pest Management: Strategies and Policies for Effective Implementation mentation initiative. This training should take place centers of the Consultative Group on International close to the field; much of the information will be Agricultural Research and other international research locally specific and generated in collaboration with groups. farmers through adaptive research on farmers' fields. Such a decentralized initiative, involving the estab- Note lishment of farmner field schools, has been very suc- cessful in a number of Asian countries (Asian 1. The global market for agrochemicals is current- Development Bank 1993). ly estimated at $21 billion; nearly 27 percent of this is A related feature of successful initiatives has been in North America, 30 percent in Western Europe, and the role of pilot projects as the platform for demon- 14 percent in Japan. The market growth in OECD strating and validating IPM techniques before more (Organisation for Economic Co-operation and widespread extension and training. This training is Development) countries has flattened out, but use in backed by a continuous flow of information from pest the developing world is expected to grow consider- scouts, farmers, and others to decisionmakers and is ably. In 1990 developing countries accounted for 20 consequently knowledge-, information-, and research- percent of world agrochemical use, but this percent- intensive. Such information can be transmitted daily age is expected to double by 2000. However, use in (weather reports, for example) or weekly (by radio, these countries is limited largely to a few crops: television, or newspapers). Support for IPM imple- cacao, cotton, coffee, flowers, fruits (mainly banana), mentation is inconsistent with a top-down technolo- maize, rice, tobacco, and vegetables. See OECD gy-transfer approach. (1995). An alternative mode, which has been used in Indonesia, is the so-called facilitation model (box References 18.3). Rather than transferring ready-made packages of technology and blanket recommendations, this The word "processed" describes informally repro- approach seeks to enhance the capacity of farmers to duced works that may not be commonly available learn, to make use of their experience, and to adapt through libraries. technologies as required (Roling and van de Fliert Asian Development Bank. 1993. IPM Guidelines. 1994). The use of such a different approach has impli- Manila. cations for investment design, the design and man- Bonny, S., and R. Charles. 1993. "Perspectives d' agement of institutions, training of staff, and links evolution de l'emploi des engrais et des phy- between research and extension, including those with tosanitaires dans 1' agriculture fran,aise." Cahiers d'economie et sociologie rurale 26(l):29-62. Box 18.3. Lessons Learned in Indonesia Carson, Rachel. 1962. Silent Spring. Cambridge, Mass.: Riverside Press. .de Fliert, Pontius, and Rling FAO (Food and Agriculture Organization of the United According~ to va Nations). 1990. International Code of Conduct on (1995), expenence in Indonesia has taught the fol- lowing lessons: the Distribution and Use of Pesticides. Amended lowing lessons: vrin oe * Introduction of IPM cannot be compared to the version. Rome. transfer of technology from expert to nonexpert; . 1993. Safe and Efficient Use of Pesticides in transfer,itrequire,sohfacilitation of a learning Africa. Rome. rather, it requires the Farah, Jumanah. 1994. Pesticide Policies in Developing process. Countries: Do They Encourage Excessive Use? * The critical issue in IPM training for rice is under- cuntio P er3 E age ExCei Use? standing the ecology of the rice field. Discussion Paper 238. Washington, D.C.: World * Trinig IM intrutor als reuirs a earing Bank. p Training ipm instructors also requires a learming Farvar, M. Taghi, and John P. Milton, eds. 1972. The process and is related directly to daily farming Careless Technology: Ecology and International experiences. *Experencesion workerswillinevitablybinvolveDevelopment. Garden City, N.J.: Natural History . Extension workers will inevitably be involved in Press. IPM training. * Farmers are considered the best primary trainers in Feder, Gershon. 1979. "Pesticides, Information, and Pest Management under Uncertainty." American ipm for other farmers. Journal ofAgricultural Economics 61(1):97-103. Tjaart W. Schillhorn van Veen, Douglas A. Forno, Steen Joffe, Dina L. Umali-Deininger, and Sanjiva Cooke 253 Gillham, Fred E. M., Thomas M. Bell, Tijen Arin, Back, One Siep Forward: Cuba National Policy for Graham A. Matthews, Claude Le Rumeur, and A. Alternative Agriculture. Gatekeeper Series 46. Brian Hearn. 1995. Cotton Production Prospectsfor London: International Institute of Environment the Next Decade. Technical Paper 287. Washington, and Development. D.C.: World Bank. Stem, V. M. 1973. "Economic Thresholds." Annual IOPRM (International Organization for Pesticide Review of Entomology 18:259-80. Resistance Management). 1993. "The Management Teng, P. S., and K. L. Heong. 1988. "Pesticide of Resistant Cotton Pests in China." Paris. Management and Integrated Pest Management in Processed. Southeast Asia." Consortium for International Kenmore, Peter E. 1991. "Indonesia's Integrated Pest Crop Protection, Beltsville, Md. Processed. Management: Policy, Production, and Environ- Thrupp, Lori Ann. 1990. "Inappropriate Incentives for ment." Paper presented at U.S. Agency for Pesticide Use: Agricultural Credit Requirements in International Development environment and agri- Developing Countries." Agriculture and Human culture officers conference, Colombo, Sri Lanka, Values 7(3-4):62-69. 11 September. Processed. Tobin, R. J. 1994. Bilateral Donor Agencies and the Kiss, Agnes, and Frans Meerman. 1991. Integrated Environment: Pest and Pesticide Management. Pest Management andAfrican Agriculture. Technical Arlington, Va.: Winrock International Environmental Paper 142. Washington, D.C.: World Bank. Alliance. Murray, D. L., and E. L. Nathan. 1994. "From U.S. Department of Agriculture, Agricultural Pesticides to Integrated Pest Management: Pros- Research Service. 1993. "USDA Programs Related pects for an Alternative Development Paradigm in to Integrated Pest Management." USDA Program Central American Agriculture." Paper presented at Aid 1506. Bleltsville, Md. Processed. the eighteenth international congress of the Latin van de Fliert, Elizabeth, J. Pontius, and Niels R6ling. American Studies Association, Atlanta, Georgia, 1995. "Searching for Strategies to Replicate a 10-12 March. Processed. Successful Extension Approach: Training of IPM National Research Council. 1989. Alternative Trainers in Indonesia." European Journal of Agriculture. Washington, D.C.: National Academy Agricultural Education and Extension 1(1):41-63. Press. Vijftigschild, R. A. N. 1996. "The Relationship OECD (Organisation of Economic Co-operation and between Pesticide Use and Environmental Burden Development). 1995. "No. 6 Guidelines for Aid for Arable Farms in the Netherlands." In H. Waibel Agencies on Pest and Pesticide Management." In and J. C. Zadocks, eds., Institutional Constraints Guidelines on Environment and Aid. Paris. to IPM. Plesticide Policy Project. Hanover, Pincus, J. 1994. "Farmer Field School Survey: Impact Germany: ]nstitut fur Garten-bauokonomie. of IPM Training on Farmers." Pest Control Waibel, H. 1994. "Towards an Economic Framework Behavior Internal Report. National IPM Program, of Pesticide Policy Studies." In S. Agne, G. Fleischer, Jakarta. Processed. and H. Waibel, eds., Proceedings of the Workshop on Postel, Sandra 1987. Defusing the Toxics Threat: Pesticide Po,licies. Gottingen, Germany: Institut fuir Controlling Pesticides and Industrial Waste. World- Agrarokonomie der Universitat Gottingen. watch Paper 79. Washington, D.C.: Worldwatch Wiebers, Uxve-Carsten. 1993. Integrated Pest Institute. Management and Pesticide Regulation in Develop- Roling, Niels, and Elizabeth van de Fliert. 1994. ing Asia. Technical Paper 211. Washington, D.C.: "Transforming Extension for Sustainable Agriculture: World Bank. The Case of Integrated Pest Management in Rice in Zilberman, Dlavid, A. Schmitz, G. Casterline, E. Indonesia." Agriculture and Human Values 11 Lichtenberg, and J. B. Siebert. 1991. "The (2-3):97-108. Economics of Pesticide Use and Regulation." Rosset, Peter, and Medea Benjamin. 1994. Two Steps Science 253(5019):518-22. 19 Toward Reduced Pesticide Use for Cereal Crops in Asia Prabhu L. Pin gali and Roberta V Gerpacio Intensive crop monoculture systems, popularized by for potato, and Forget, Goodman, and de Villiers the green revolution, created an environment that was 1993). conducive to pest growth. Although pest-related loss- Technological change is dramatically reducing the es in yield were small in percentage terms, they were need for pesticide use in cereal crop production sys- highly visible and led policymakers to fear major out- tems. Today, with improved host-plant resistance, breaks (Rola and Pingali 1993). This desire to avoid improved crop management, and advances in inte- large-scale pest outbreaks resulted in policies that grated pest management, along with a reformed poli- made pesticides easily accessible and affordable at the cy environment that is beginning to discourage the farm level (Pingali and Rola 1995). Pesticides were use of pesticides, the link between pesticides and seen both as a guarantee against crop failure and as a cereal crop productivity has become less clear. means of realizing the full potential yield of the crops Certainly in the case of insecticides for rice and fungi- (Waibel 1986). Although frequent and indiscriminate cides for wheat, recent evidence indicates that the pro- use of pesticides is the norm in developing countries ductivity benefits of applying chemicals are marginal today, farm-level promotion of pesticide use is at best. In the case of maize, chemical use under sub- accompanied by little, if any, training in judicious use, sistence production systems has always been low, but farmer safety, and adverse ecological and health con- commercialization has led to increased pesticide use. sequences. However in recent decades, maize varieties with Indiscriminate pesticide use in the production of improved resistance to insects and diseases have been various crops has been documented to result in the developed. Herbicides are the one uncertainty for the following: (a) health impairment due to direct or indi- future of chemical use. Adequate nonchemical con- rect exposure to hazardous chemicals, (b) contamina- trols for weeds are not yet available, and herbicide use tion of ground and surface waters through runoff and is increasing dramatically as the opportunity cost of seepage, (c) the transmittal of pesticide residues labor rises across the developing world. through the food chain, (d) increased resistance of This chapter describes the opportunities currently pest populations to pesticides, thereby reducing their available for achieving a dramatic reduction in the use efficacy and causing pest outbreaks, and (e) the reduc- of agrochemicals in the production of cereal crops in tion of the population of beneficial insects like para- Asia. The first section summarizes past trends and sites and predators, thereby reducing the effectiveness future prospects for pesticide use in Asia for the three of pest control strategies that attempt to minimize the major cereals: rice, wheat, and maize. The second sec- use of pesticides (see Pingali and Roger 1995 for rice, tion identifies the factors that contribute to the rise in Cormwell 1995 for tobacco, Crissman and Cole 1994 farm-level demand for pesticides and the factors that 254 Prabhu L. Pin gali and Roberta V Gerpacio 255 could reduce that demand. The third section high- egory I and category II chemicals, as classified by the lights the advances that have been made in the gener- World Health Organization, many of which have been ation and dissemination of cereal crops that are less banned for agricultural use in mnost industrial coun- dependent on chemicals because they are resistant to tries. These insecticides include organochlorine endo- insects and diseases. The final section discusses the sulfan; organophosphates such as methyl parathion, integrated management approaches that are necessary monocrotophos, and chlorpyrifos; carbamates such as for maximizing and sustaining the productivity gains BPMC, carbaryl, and carbofuran; and pyrethroids such offered by resistant varieties. as cypermethrin and deltamethrin. They continue to be widely used in tropical Asia due either to ignorance The Pesticide Market and Use or to substantial:Ly lower costs than the safer altema- in Asian Cereal Production tives. There is considerable variation from country to From $900 million in 1960, the global agricultural country and froml crop to crop in the type and quanti- pesticide market was valued at $27.5 billion in 1993 ty of pesticides used. Herbicides dominate North and at $30.3 billion in 1995, indicating an annual American and European domestic markets, but insec- growth rate of almost 11 percent. Global pesticide ticides are more common elsewhere in the world, sales in 1993 were equivalent to almost 2 million met- especially in the Asia-Pacific region. The share of ric tons of active ingredients, of which 497,000 metric East Asia in global pesticide consumption is 14 per- tons-or 25 percent-were consumed in Asian agri- cent but accounts for 54 percent of Asia's pesticide culture (FAO 1997). Cereal crop production across the use. Wood Mackenzie Consultants, Ltd. (1993 as world is reported to be the second largest consumer of cited in IRRI 1995) reports that, in 1993, Japan alone agricultural pesticides, next only to fruit and veg- accounted for 62 percent of all pesticides consumed in etable production. In 1995 rice and maize production Asia's rice production: 46 percent of the insecticides, consumed about 23 percent of all agrochemicals, 71 percent of the herbicides, and 72 percent of the while the production of fruits and vegetables con- fungicides used on rice in the region (table 19.1). The sumed about 26 percent (IFPRI 1996). Global rice pro- next large users of rice pesticides in Asia were the duction, for instance, consumed about $3.2 billion Republic of Korea, China, and India, at 12, 11, and 7 worth of agricultural pesticides (12 percent of the percent, respectively. 1993 pesticide sales), of which $2.6 billion (or 81 per- Aggregate figures, though, tend to mask the inten- cent) were spent in Asia. sity of pesticide use. Malaysia used just 9 percent of Before the 1960s, pesticides were mostly con- Asia's total pesticide consumption but was the sumed in industrial nations. With agricultural mod- region's most intensive user of pesticides, applying emization in developing countries in the mid-1960s, 23.42 kilograms of active ingiredients per hectare of developing-country exports became an important part arable land. In contrast, India consumed 14 percent of of the global pesticide industry. In the 1960s and all agricultural pesticides in Asia but applied only 1970s, France, Germany, Italy, Japan, Switzerland, about 0.43 kilogram of active ingredients per hectare United Kingdom, and the United States accounted for of arable land. 75 percent of world chemical production and provid- In Asian cereal crop systems, insecticides are used ed approximately 85 percent of all agrochemicals predominantly in rice and commercial maize produc- used in Asia (Boardman 1986 and United Nations tion; almost none is used in wheat production. 1990, as cited in Pingali and Rola 1995). Fungicides are generally not important for cereal A large portion of the pesticides exported to the crops, except for rice production in the East Asian developing countries are products no longer used in countries. The use of fungicides for rice may increase the industrial countries on experimental and health in the rest of Asia if the current push toward higher grounds. Weir and Shapiro (1981), for instance, point rice yields is pursued (Pingali, Hossain, and Gerpacio out that at least 25 percent of U.S. pesticide exports 1997). In the case of wheat, the plant's ability to resist are products that are banned, are heavily restricted, or high levels of disease pressure has made fungicide use have never been registered for use in the United almost obsolete in Asian wheat production systems. States. Warburton, Palis, and Pingali (1995) also However, the use of herbicides is increasing for all report that the vast majority of insecticides used in three cereals, as rising agricultural wages, due to developing Asian countries are highly hazardous cat- increasing off-farm employment opportunities, 256 Toward Reduced Pesticide Use for Cereal Crops in Asia Table 19.1. Share in Consumption of Pesticides Used in Rice Production in Select Asian Countries, 1993 (computed based on rice pesticide sales) Country Insecticides Herbicides Fungicides All pesticides Bangladesh 1.03 0.54 1.38 0.95 China 16.55 8.67 7.59 11.22 India 12.41 2.49 4.14 6.58 Indonesia 4.14 0.65 0.69 1.94 Japan 45.60 71.40 72.41 61.81 Korea, Rep. of 13.24 11.92 11.31 12.17 Myanmar 1.03 0.54 0.69 0.76 Philippines 1.55 1.19 0.41 1.10 Thailand 2.48 1.95 0.69 1.79 Vietnam 1.96 0.65 0.69 1.14 Source: For basic data, IRRI 1995. reduce the cost-effectiveness of hand weeding Agricultural Modernization and the Demand (Naylor 1997). for Agrochemicals Although Asia's rice production remains the major consumer of agricultural pesticides, the composition In the developing world, agricultural modernization, of its pesticide market has been changing in recent which includes agricultural intensification and years. The proportion of insecticides used in Asian increased market orientation, led to rapid growth in rice production, for example, decreased from about 45 the demand for agrochemicals. The rapid spread of percent (of all pesticides) in 1984 to 37 percent in the green revolution for cereal crops, especially in 1993 (table 19.2). Total insecticide use, for rice pro- Asia during the late 1960s and the 1970s, led to an duction, actually decreased 10 percent from 1990 to upward shift in the demand for modern agricultural 1993 (table 19.3). In the same time period, the pro- inputs, including pesticides. Although the demand for portion of herbicides in total chemical use increased pesticides as a whole grew rapidly, the trends for from 27 to 35 percent. Herbicide use grew at a rate of chemical subgroups, such as insecticides, herbicides, about 4 percent a year during 1987-92. and fungicides, varied substantially by country or In 1984, rice production in China used 64 percent region and by crop. Pingali and Rola (1995) provide a of all insecticides used, 8 percent of herbicides, and stylized representation of the factors leading to 28 percent of fungicides; by 1993, insecticide use had regional differences in pesticide use. The following dropped to 54 percent, while herbicide use had risen summarizes their discussion. to 27 percent of all pesticides used. India and the Agricultural intensification-the movement from Philippines also demonstrated the same trend in rice an extensive to an intensive production system or production. In contrast, rice economies with cheap from a subsistence production system to a commercial agricultural labor, such as Myanmar and Vietnam, still one-has been extensively documented to promote show increasing use of insecticides, but decreasing the use of agrochemicals. Table 19.4 shows the differ- use of herbicides, in rice production (table 19.2). ential effects of increasing land scarcity and increas- Indiscriminate and injudicious use of agrochemi- ing market orientation on the demand for chemicals. cals is the direct result of inappropriate or inadequate Consider first a sparsely populated subsistence soci- pesticide policies in developing countries. ety with limited access to markets. Agriculture in such Enlightened policy environments along with the intro- societies is characterized by extensive land use and duction of crop varieties resistant to insects and dis- almost complete reliance on nontraded inputs, such as eases, as well as the promotion of integrated pest farmyard manure. Pest pressure is low in such culti- management, generally result in less use of pesticides, vation systems and is kept that way through a variety more use of safer chemicals, and improved farmer of management practices, such as crop rotations and safety and enviromnental sustainability. the use of traditional cultivars with known resistance Prabhu L. Pingali and Roberta V Gerpacio 257 Table 19.2. Proportion of Insecticides, Herbicides, and Fungicides Used in Rice Production in Select Asian Countries, 1984-93 (percent) Country and year Insecticides Herbicides Fungicides Bangladesh 1984 61.3 19.4 19.4 1988 42.6 27.7 29.8 1993 40.0 20.0 40.0 China 1984 64.3 7.8 27.9 1988 77.9 4.2 17.9 1993 54.2 27.1 18.6 India 1984 79.1 9.4 11.5 1988 77.1 12.1 10.8 1993 69.4 13.3 17.3 Japan 1984 29.8 37.4 32.8 1988 32.1 40.7 27.1 1993 27.1 40.6 32.3 Philippines 1984 59.5 24.3 16.2 1988 37.5 37.5 25.0 1993 51.7 37.9 10.3 Myanmar 1984 36.4 45.5 18.2 1988 33.3 33.3 33.3 1993 50.0 25.0 25.0 Vietnam 1984 50.0 33.3 16.7 1988 30.8 38.5 30.8 1993 63.3 20.0 16.7 All Asia 1984 44.5 26.8 27.6 1988 42.1 31.4 25.7 1993 36.8 35.1 27.6 Source: For basic data, nuu 1995. to chronic pests. handled by family labor, increased insect pressure is In subsistence societies, increasing land scarcity no longer amenable to traditional management prac- due to population growth leads to agricultural intensi- tices, and small amounts of insecticides begin to be fication-that is, increased intensity of land use used for cereal crops, even in subsistence societies. (Boserup 1965 and Pingali and Binswanger 1987). Low to moderate amounts of fungicides also tend to Where the opportunity costs of family labor are low, be used for crops such as cotton, tobacco, and horti- food production continues to rely predominantly on cultural products, especially fruits and vegetables (see nontraded inputs. Pest pressure increases with intensi- table 19.5). fication due to increased spatial and temporal carry- The contrasting scenario is one of a sparsely pop- over of pests. Although increased weed pressure is ulated area that has excellent market access. In this 258 Toward Reduced Pesticide Use for Cereal Crops in Asia Table 19.3. Percentage Change and Annual Growth Rate of Sales in Insecticides, Herbicides, and Fungicides Used in Rice Production in Select Asian Countries, 1980-93 (percent) Country and Percentage change Annual growth rate chemical 1980-84 1985-89 1990-93 1981-86 1987-92 Bangladesh Insecticides - -37.5 -66.7 - -6.3 Herbicides - 0.0 -66.7 - -6.1 Fungicides - 33.3 -47.4 - 6.8 China Insecticides 109.1 17.9 -8.6 12.0 4.4 Herbicides 40.0 -37.5 - 9.0 Fungicides 31.6 0.0 22.2 1.2 2.1 India Insecticides 175.0 13.8 -28.6 8.5 4.4 Herbicides 62.5 53.3 -17.9 9.3 9.4 Fungicides 14.3 33.3 25.0 0.4 11.9 Japan Insecticides 62.3 51.9 1.4 14.6 -1.1 Herbicides 6.8 53.4 14.6 11.6 1.0 Fungicides 3.1 16.6 34.6 11.0 1.9 Philippines Insecticides 37.5 -40.0 -21.0 7.4 0.0 Herbicides 125.0 40.0 -21.4 17.6 2.5 Fungicides 0.0 83.3 -78.6 1.1 12.0 Myanmar Insecticides 4.0 167.9 36.2 7.3 15.5 Herbicides 0.0 83.3 155.8 16.0 31.6 Fungicides 0.0 191.7 -6.8 8.7 9.9 Vietnam Insecticides 20.0 -16.7 58.3 3.3 10.4 Herbicides - 20.0 -50.0 - 0.0 Fungicides - 150.0 -58.3 - 13.6 All Asia Insecticides 77.9 36.4 -10.4 11.5 3.3 Herbicides 23.0 46.0 24.4 13.0 3.7 Fungicides 12.8 35.5 10.2 9.4 4.2 - No data available. Source: For basic data, iRn 1995. Table 19.4. Intensification, Market Orientation, and Demand for Agrochemicals Land/labor ratio Low market orientation High market orientation High Low demand for all chemicals Herbicides predominate Low Insecticides predominate High demand for all chemicals Source: Pingali and Rola 1995. Prabhu L Pin gali and Roberta V Gerpacio 259 Table 19.5. Agrochemical Use, by Stage of Agriculture and Cropping System Otherfield crops Horticulture Farm characteristics Cereals (rice) (cotton, tobacco) (fruits, vegetables) Land-abundant All chemicals., none All chemicals, All chemicals, subsistence none to low none to low Land-scarce Insecticides, Insecticides, fungicides, Fungicides, subsistence low to moderate low to moderate low to moderate Land-abundant Herbicides, Herbicides, fungicides, Fungicides, herbicides, market-oriented moderate to high moderate to high moderate to high Land-scarce, Insecticides, herbicides, All chemicals, high All chemicals, high market-oriented high Source: Rola and Pingali 1995. case, agricultural intensification is high due to high a large number of pest outbreaks have been associat- land values, but unlike the subsistence case, the ed with injudicious pesticide applications rather than opportunity cost of labor is also high, hence the high with the use of modem high-yielding cultivars, high levels of traded input use. Although increasing insect cropping intensity, or high use of chemical fertilizers pressures can be controlled through appropriate crop (Heinrichs and Mochida 1984, Kenmore and others rotations and seasonal fallows, weeds are the domi- 1984, Joshi and. others 1992, and Schoenly and others nant constraint. High levels of herbicide use are the 1996). Outbreaks of secondary pests of rice, notably norm. High fungicide use can also be observed for the brown plan-hopper, that were previously of minor horticultural crops, especially fruits and vegetables. significance began to occur in regions adopting mod- Where market access and land scarcity are both high, em varieties and using agrochemicals (Pingali and high opportunity costs of land and labor result in agri- Gerpacio 1997). High and injudicious applications of cultural intensification with high use of traded inputs. pesticides disrapted the rich diversity of pest and The demand for all chemicals is high in such soci- predator populations, where the species richness and eties. abundance of predator populations are often greater Modem cereal crop production systems further than those of pest populations. aggravate pesticide use. Intensive monoculture sys- The rapid growth in pesticide use can also be tems using high-yielding cereal varieties result in an attributed to the misinformation and risk aversion of environment conducive to the buildup and infestation both policymakers and farmers. In an extensive of pests, and the consequent use of pesticides disrupts review of rice case studies from across Asia, Rola and the natural balance between pests and predators. The Pingali (1993) provide documentation that policy- risk of insect- and disease-related losses grew because makers' perception of yield losses is higher than farm- early modem varieties were often highly susceptible ers' perception of yield losses, which in turn is sub- to local pests and because crops were no longer het- stantially higher than actual yield losses. Both farm- erogeneous. Regular prophylactic application of pes- ers' and policymakers' perceptions of pest-related ticides promoted by extension services and supported yield losses are anchored in exceptionally high losses by government subsidies became a standard part of during major infestations, even when the probability the early green revolution package. of such infestations is low. In cases of minor infesta- In addition, because early formulations of pesti- tions, physical damage to the plant is easily visible but cides were often nonselective, pesticides proved does not always result in a loss of yield. Heong, equally lethal to beneficial pests that preyed on crop Escalada, and Mai (1994) have shown that leaf dam- pests. In many cases, pesticide use actually lowered age, even as much as 50 percent, has insignificant yields, as crop pests, freed of their natural predators, effects on rice yields, yet 80 percent of insecticide multiplied without constraints (Pagiola 1995 and Rola applications for rice in Asia are targeted toward leaf- and Pingali 1993). In the case of rice in tropical Asia, feeding insects. 260 Toward Reduced Pesticide Use for Cereal Crops in Asia Although in aggregate terms cereal crops account crops and, in the absence of alternative pest control for the bulk of pesticide use in Asia, on a per hectare strategies, increasing share of the agrochemical mar- basis their use is quite small when compared with the lev- ket (Pingali and Rola 1995). els applied on fruits, vegetables, and other high-value With recent advances in developing the cereal commodities. Pesticide application in high-value crops crop's ability to tolerate insects and diseases, and in is related to consumer demand for aesthetically integrated pest management, the outlook for future appealing agricultural products. Because these high- reductions in pesticide use is very promising. Modern value agricultural products enjoy a substantial price science has reduced the dependence on insecticides premium for unblemished physical appearance, risk- for rice and maize and the dependence on fungicides averse farmers tend to apply pesticides, beyond the for wheat. The dependence on herbicides for cost- technical optimum, in order to capture this price dif- effective management of weeds continues to be prob- ferential. Pagiola (1995), for example, reports that in lematic, with little scientific advance either in genetic Bangladesh where 70 percent of pesticides are used or in crop management. on rice, the amounts used per unit of area and the total area affected are both relatively small. On vegetables, Breeding for Host-Plant Resistance the use of insecticides follows a pattern almost dia- as a Technological Alternative to Pesticides metrically opposed to that found in rice. Whereas rice is sprayed only two to three times a season, vegetables Host-plant resistance-the plant's ability to resist such as eggplant and country beans are often sprayed high levels of pest infestation-has been the corner- several times a week. A survey of eggplant producers stone of the scientific strategy for developing sustain- in Jessore indicates a range of application from 17 to able pest control systems for cereal crops. Significant 150 times for one crop cycle (Kabir and others 1994, advances have been made in the last three decades in as cited in Pagiola 1995). Other vegetables have rates the development and dissemination of crop varieties of use that are lower, but still generally higher than with resistance to the major cereal pests (table 19.6). rice. Cauliflower and cabbage, for instance, are com- Much of the advance has occurred through the use of monly sprayed three to four times. In the case of conventional breeding approaches, although substan- wheat, maize, and rice, which are the most important tial future gains in resistance development could cereal crops in Asia, pesticides do not enhance physi- come through the use of modern biotechnology tools. cal quality in any way, and there is no price differen- Host-plant resistance ought to be seen as an essential tial to capture (Rola and Pingali 1993). Given the pos- building block in an integrated pest management itive income elasticity of demand for fruits and veg- strategy. Attempts to promote resistant varieties with- etables, the long-term prognosis for developing out a commensurate change in pesticide management economies is one of increasing areas under these practices have generally failed due to the early break- Table 19.6. The Evolution of Host-Plant Resistance for the Major Cereals, 1960s-90s Period Rice Wheat Maize 1960s Striped stemborer Stem rust European corn borer 1970s Brown planthopper, Septoria tritici blotch Earworms, tropical borers, green leafhopper, southwestern corn borer rice whorl maggot 1980s Yellow stemborer, Leaf rust Fall armyworm wahite-backed brown planthopper, thrips 1990s Bacillus thuringiensis Spot blotch, fusarium Bacillus thuringiensis scab, stripe rust Prabhu L. Piigali and Roberta V Gerpacio 261 down in resistance. achieved on resistance to the second-priority diseases, such as barley yellow dwarf, tan spot, and septoria Conventional Breeding for Host-Plant Resistance nodorum blotch (Dubin and Rajaram 1996). In the case of maize, most of the released hybrids Although, technically tedious and time-consuming, and composites have high resistance to foliar dis- conventional breeding techniques have led to signifi- eases, and breeding for resistance to prevalent insect cant progress in generating varieties resistant to major pests and diseases has made considerable progress insect pests and diseases. Early work at the (Commonwealth Agricultural Bureaux 1980). More International Rice Research Institute (IRRI) in the recently, Kumar and Mihm (1995) have found that Philippines reported that most of the important CIMMYT has developed potentially useful single cross sources of resistance to major diseases and pests had hybrids and varieties (for example, CML 139 x CML been incorporated into modem rice varieties (IRRI 135, CML 139 x Ki3, CML 67 x CML 135) that, 1972). The emphasis on releasing rice varieties with because of improved resistance, suffer less damage improved host-plant resistance to major insect and and percentage loss in grain yield to fall armyworn, disease pressures has continued through subsequent southwestern corn borer, and sugarcane borer than decades. Lines considered promising enough have earlier varieties. In 1993, it was reported that screen- been named as varieties, while others have proven to ing of CIMMYT bank accessions and elite germ plasm be good parents in new crosses. Named varieties and has continued to find new or better sources of maize promising breeding lines provide resistance to impor- resistance to stemborers, armyworms, maize weevils, tant rice insect pests such as green leafhoppers, brown spider mites, rootworms, and corn earworms (Mihm leafhoppers, stemborers, and gall midge and to dis- and others 1994, as cited in Edmeades and Deutsch eases such as blast, bacterial leaf blight, bacterial leaf 1994). streak, tungro, and grassy stunt (IRRI 1985). For decades, plant breeders have ensured that each For wheat and maize, the Centro Intemacional de new elite cultivar and breeding line possesses a wide Mejoramiento de Mafz y Trigo (CIMMYT, Intemational range of genes for resistance to pests and diseases. Maize and Wheat Improvement Center) based in However, breeding programs are challenged by the Mexico has an active program that develops maize fact that genetic variation is not confined to the host and wheat germ plasm with desirable levels of resis- plant but is also a feature of thne pest or parasite. With tance to insect pests and diseases. CIMMYT combines certain types of resistance, selection for a resistant "shuttle breeding" and "hot spot" multilocation test- population of the host is followed closely by natural ing within Mexico and abroad to obtain multiple dis- selection of the pest or parasite for those variants that ease resistance for the different agroecological zones can overcome the resistance. The risk is greater when in mega-environments where wheat and maize are large areas are planted year-long to a single, homoge- grown (Dubin and Rajaram 1996). Insect problems neous cultivar in association with indiscriminate use are not as important in wheat as in the other two cere- of pesticides. 'Pest and pathogen populations change als. "Shuttle breeding" refers to a method in which over time, and genetic resistance breaks down, while breeding generations from the same crosses are alter- some pest ancl disease problems remain intractable. nately selected in environmentally contrasting loca- To keep ahead of evolving and changing pest popula- tions as a way of combining desirable characteristics. tions, breeding activities constantly have to aim for "Hot spots" are locations where significant variability increased genetic diversity against pests and cultivars for a pathogen exists and where screening plant gen- with more durable resistance. Recent advances in erations exposes the genotypes to as broad a range of biotechnology can help to improve the durability of virulence genes and combinations as possible. This host-plant resistance. screening, together with multilocation testing, increases the probability of developing durable resis- Contributions of Biotechnology to Breeding tance. for Host-Plani, Resistance In wheat, considerable genetic progress has been achieved in developing host-plant resistance, with Biotechnology adds a new dimension to the research regard to leaf rust, septoria leaf blotch, fusarium scab, on improved host-plant resistance by enhancing con- and bacterial leaf streak in the high-rainfall areas of ventional breeding efforts through the use of wide tropical highlands. Some progress has also been hybridization, molecular marker technologies, and 262 Toward Reduced Pesticide Use for Cereal Crops in Asia genetic engineering. Biotechnology enhances the effi- plant breeders indirectly to select genes of interest. ciency of conventional breeding efforts by reducing Instead of selecting progeny plants having the gene of the length of time involved and the inherent trial-and- interest by checking for the action of the gene, scien- error nature of conventional breeding. Biotechnology tists use marker-aided selection to identify the pres- may also make resistance more durable by incorpo- ence of the linked marker in the plant's DNA. Such rating alien and novel genes into the cultivated plant markers are especially useful in breeding for resis- species. In the short to medium term, the contribu- tance to nonendemic pests and diseases or for genes tions of biotechnology are likely to be most signifi- with overlapping effects that can contribute to com- cant in improving the durability of pest resistance plex and more durable resistance. rather than in improving yield and grain quality. The In the case of rice, more than 20 single genes for following paragraphs discuss the prospects of bio- disease and insect resistance have been located rela- technology for the major cereal crops. tive to the restriction fragment length polymorphism Wide hybridization allows breeders to use genetic (RFLP) markers. Among them are several genes for diversity drawn from the wider gene pool of a crop's resistance to rice blast fungus and major genes for entire genus. A crucial tissue culture technique, resistance to bacterial blight and three insects (brown embryo rescue, enables breeders to cross cultivated planthopper, white-backed brown planthopper, and species with wild species or distant relatives and to gall midge). A current project in wheat based on the stabilize select breeding lines faster in the laboratory use of marker-assisted selection aims to develop resis- than in the field. Pieces of the wild genome are inte- tance to barley yellow dwarf, a viral disease that grated into the chromosomes of the elite parents, use- affects wheat as well as rice and barley. ful genes are transferred from different wild species Genetic engineering refers to a process that into cultivated varieties, and the progeny are readily includes the identification and incorporation of eco- stabilized. nomically valued genes into a specific crop (with Embryo rescue has produced hybrids between cul- genes originating from any living organism, from a tivated rice and 12 wild species. Genes for resistance virus, or even from chemical synthesis), the delivery to brown planthopper, white-backed brown planthop- systems that introduce the desired gene into the recip- per, bacterial blight, and blast have already been ient cells, and the expression of the new genetic transferred from wild species into elite rice breeding information in the recipient cells. Plant genetic engi- lines (IRRI 1993a). neering techniques provide breeders with new oppor- Diversity for resistance to or tolerance of several tunities to improve the efficiency of production and stresses is now available within the wide hybridiza- increase the utility of agricultural crops. Modern vari- tion wheats under prevailing conditions in Mexico. eties could be further improved by inserting foreign Resistance to Karnal bunt, septoria tritici blotch, and genes into plant cells and regenerating viable, stable, spot blotch from wide hybridization wheats is being and fertile plants that possess useful traits such as dis- used to improve bread wheat based on disease screen- ease and insect resistance, better-quality grain, and ing data obtained over several years. Resistant wide resistance to certain herbicides. Perhaps the biggest hybridization genetic stocks of wheat have now been advantage of these techniques is that they have the crossed onto elite cultivars, and the resulting lines are potential to shorten the time required to develop expressing diversity for resistance to these three dis- improved traits in wheat plants. eases. This germ plasm is being incorporated into The first group of useful foreign genes to be put advanced bread wheat lines to diversify the genetic into rice included soybean trypsin inhibitor for insect base of resistance to or tolerance of other stresses as resistance and two barley genes, chitinase and ribo- well (Pingali and Rajaram 1997). some inhibitor protein (RIP), for resistance to the fun- Molecular marker technology increases the effi- gal diseases sheath blight and blast (IRRI 1993b). ciency of combining specific desirable genes into When introduced into rice, novel genes such as the improved breeding lines by facilitating the selection Bacillus thuringiensis (Bt) gene for insect resistance of genes. A genetic map is used to show the relative and coat protein genes for tungro resistance should position of genes o.n a chromosome and the distance impart high levels of resistance. Coat protein-induced between genes. The position of useful and important virus resistance is directed against the rice tungro genes on the chromosomes is roughly identified or virus (Beachy and others 1989), the most important tagged by the use of molecular markers, allowing disease problem for rice (Herdt and Riely 1987). A Prabhu L. Pingali and Roberta V Gerpacio 263 coat protein-mediated resistance to rice stripe virus weeds, especially in direct-seeded and minimum- (an insect-transmitted virus) has been introduced in tillage systems. 'Widespread use of such plants could, Japonica rice (Hayakawa and others 1992). The Bt however, increase the risk of cross-pollination of the genes are targeted against stemborers (Joos and gene into weed and wild species, thus aggravating, Morrill 1989), which are the most important insect rather than reducing, the problem of weed competition. pest problem for rice (Herdt and Riely 1987). Insect- To date, however, there is no evidence that such cross- resistant rice generated by introducing modified endo- pollination has occurred. A herbicide-resistant plant toxin gene of Bt has been shown to produce Bt protein could also encourage high levels of herbicide use and and to enhance insect resistance in Japonica rice thereby contributr to increased environmental pollu- plants (Fujimoto and others 1993). tion. However, a herbicide-resistant plant could Genes for resistance to viral, bacterial, fungal, and encourage the switch to smaller doses of safer herbi- other diseases have been identified and transferred cides. Moreover, a host of intellectual property issues successfully into many crops (Ahl Goy and Duesing have to be dealt with in the transfer of genes and 1995 and Persley 1990). The first genetically engi- advanced biotechnology tools from advanced-country neered Indica rice based on protoplast system (Datta laboratories to developing-country research programs and others 1990) now provides resistance to sheath where they woul(d be used to develop improved cereal blight in Indica rice (Lin and others 1995). Several crops. chitinase genes in combination with Bt genes have been introduced in rice to enhance protection against Pesticide Productivity, Yield Variability, and Host- pathogenic fungi, bacteria, and insects. The identifica- Plant Resistance., Evidence from Experimental Trials tion of many other genes that may enhance the genetic and Farmers' Fields yield potential of rice or confer resistance to biotic and abiotic stresses is still at an early stage. After their iden- On-farm experimnents and examination of farmers' tification, these genes have to be isolated and con- yields do not suggest a yield or profitability response to structed in a manner suitable for their introduction and pesticide applications on resistant crop varieties. The expression into the rice genome. introduction of varieties with host-plant resistance has In the case of maize, genetically engineered plants dramatically reduced the need for insecticides in rice for resistance to the com borer have been produced and maize and the need for fungicides in wheat. using the Bt gene. Although the transformation of Evidence for the three major cereals indicates that the wheat plants is still in its infancy, there are significant extent of crop ]oss due to insects and diseases has hopes of developing additional transgenic plants with dropped over the last two decades and that the extent of fungal resistance and abiotic stress tolerance. yield loss due to the failure to apply chemicals has Although progress in biotechnology is rapid, only in declined significantly. a few cases have agronomically viable plants been Rola and Pingali (1993) summarize the data on rice developed. Although some progress has been made, crop losses due to insects in the Philippines and indi- gene mapping techniques remain slow and in need of cate that estimated crop losses have decreased since the refinement to facilitate marker-aided gene selection (N. late 1960s and early 1970s. Pathak and Dyck (1973) Huang, iRRi, personal communication). The procedure recorded a crop loss of 22.5 percent during 1968-72, is expensive and requires highly trained personnel. but more recently, both Litsinger (1991) and Waibel Traditional plant breeding problems are resolved, but (1986) recorded losses of only 8.6 and 8.9 percent, challenges facing plant improvement programs respectively. The drop in insect-related crop losses can remain essentially the same. The germ plasm still be attributed to the widespread use of resistant cultivars needs to be screened for desirable genes, which then in farmers' fields after the mid-1970s. Both Litsinger have to be transferred to the crop, and, finally, the per- and Waibel observe no significant differences in yield formance of the improved plant has to be tested in the between insecticide-treated and -untreated plots in field. In addition, assessing the usefulness of genetic more than half the trials. In otlher words, half the time engineering techniques requires understanding the pos- farmers do not need to spray against rice pests, espe- sible negative consequences. One example is the incor- cially on varieties with notable host-plant resistance. poration of a gene for herbicide resistance into cereal Pingali and Gerpacio (1997) report that the rice yield crops. A herbicide-resistant plant could sustain crop response, on facmers' fields, to insecticide application productivity in the face of increasing competition from has been negligible with the adoption of resistant vari- 264 Toward Reduced reiLL/ Cereal Crops in Asia eties. Prophylactic application of insecticides on resis- Pingali 1993) and the associated loss in farmer produc- tant varieties tends to increase, rather than decrease, the tivity (Antle and Pingali 1994). Although such studies variability in crop yields by disrupting the balance have not been conducted for wheat and maize in devel- between pest and predator and by inducing the break- oping countries, similar results could be anticipated. down in resistance. Agrochemicals continue to have a significant Host-plant resistance has similarly been successful impact on productivity in the management of weeds, in controlling the major corn insect pests, such as the where herbicides will continue to be the preferred alter- corn borer, corn earworm, and fall armyworm. The use native in the foreseeable future. In intensive cereal crop of insecticides on insect-resistant corn varieties has gen- production systems, the use of herbicides is more cost- erally not been economically worthwhile. Diseases, effective than the use of human labor, because the sea- such as rusts of the stem and leaf in wheat, that used to sonality of weeding often creates labor scarcity so that cause significant loss in yields have been controlled in wage rates for weeding are higher than those for other the last two decacles through improved resistance. Sayre crop management operations. The use of herbicides has and others (1997) report that there is no significant dif- also gotten a boost from the general rise in farm wages ference in wheat. yields between plots sprayed with due to overall economic growth and the growth in non- fungicides and those not sprayed when resistant vari- farm employment opportunities, particularly in Asia. eties are used. As cheap herbicides become available and as farm The widespread availability of insect- and disease- wages rise, cereal crop farmers increasingly substitute resistant varieties for the major cereals has reduced the herbicides for human labor, and the savings in labor productivity benefits and the profitability of applying cost more than compensate for the additional cost of insecticides and fingicides. Discussion on productivity herbicides. Pingali and Marquez (1997) document benefits of pesticides revolves around the presumption the productivity benefits of herbicide use, even when that pesticides are a risk-reducing input-that is, they the health costs of herbicides are taken explicitly reduce the variability of yield. One needs to question into account. this presumption, both for insecticides and for fungi- There are few genetic and management alternatives cides (Pingali and Rola 1995). In the case of insecti- to herbicides, and those that exist are generally not very cides, Rola and Pingali (1992, 1993) provide empirical cost-effective. The use of more competitive cereal vari- evidence for rice production showing that insecticide eties can avert the effect of weed competition and the applications increase variability of yield. They estimate consequent use of more herbicides, but there appears to be an insecticide response function and evaluate the a tradeoff between yield and the plant's ability to com- returns to four alternative pest management strategies pete with weeds (Moody 1991). Research on varietal for the Philippines, using both farm panel data and data improvement for weed management in cereals is still at from on-fann experiments conducted over several a very early stage (Khush 1996). Among the manage- growing seasons. IResponse function estimates indicate ment options for weed control that minimize labor use very modest effects of insecticides on the mean and are the manipulation of water in the paddy, in the case variance of yield distribution. Comparing the returns to of rice (Pingali, Hossain, and Gerpacio 1997), and, in prophylactic control, economic threshold levels, farm- the case of wheat and maize, the use of ridge tillage ers' practice, and natural control, Rola and Pingali find, systems (Sayre 1996) and the use of cover crops and for lowland tropical rice systems, natural control to be intercropping. However, none of these options has the economically dominant pest management strategy. proven to be economically as attractive as the use of Natural control, in association with varietal resistance, herbicides, and the challenge for the research and poli- is consistently more profitable in an average year than cy community is to find cost-effective mechanisms for prophylactic treatrnent and economic thresholds. The reducing herbicide use in cereal crop production. dominance of natural control is found to hold even The development of genetic resistant to Striga, the when farmers' risk aversion is taken into account. It most important weed in Africa, is an exception to the lack becomes even greater when the health costs of expo- of success in using genetic means for controlling weeds. sure to insecticides are taken explicitly into account, Approximately 50 million hectares devoted to cereal because the positive production benefits of applying production in Africa are in areas infested with Striga. insecticides are overwhelmed by the higher health Although three species of Striga are dominant, maxi- costs. The value of crop loss to pests is invariably lower mum loss is caused by Striga hernonthica. Both toler- than the cost of pesticide-related illness (Rola and ance (plants that have lower yield loss than susceptible Prabhu L. Pingali and Roberta V Gerpacio 265 plants) and resistance (plants that prevent attachment of farmer. In this regard, natural control can be consid- the parasite) are known to exist in maize and other ered the ultimate goal of an integrated pest manage- crops. In a recent study (Kim. Adetimirin, and ment (IPM) program, and farmers who are well versed Akintunde 1997), Striga reduced maize yields by 25 in IPM techniques would converge toward it. This percent in tolerant hybrids and 50 percent in suscepti- implies a paradigm shift in the traditional IPM strate- ble hybrids over a range of nitrogen levels. Genetics gies from when best to apply to when not to apply. seem to be a viable option to be pursued in the pesti- Therefore, continued investments in IPM training are cide-free management of Striga. essential for the successful adoption of reduced or minimum use of insecticides. Policies for Sustaining a Minimum-Pesticide Strategy Integrated Pest Management Given the extent of breeding efforts in producing Although it is generally recognized that the sustain- pest-resistant crop cultivars and evidence of insignifi- able use of host-plant resistance requires a more inte- cant productivity benefits of pesticide use, why are grated approach to pest management, farm-level prac- high levels of agrochemical use so prevalent in crop tices of pesticide management have not changed sig- production? First, the dissemination of crop varieties nificantly over the past 20 to 30 years. Farmers have resistant to pest pressures is not accompanied by been slow to adopt IPM practices, although they read- extension messages on the reduced need for pesticides. ily accept seeds and other tangible crop production Second, farmers' pest control decisions, scientists' technologies. However, IPM is knowledge-based and research priorities, and policymakers' prescriptions, presents by far the most difficult challenge to tradi- which are all based largely on perceived pest-related tional, small-scale farmers in developing countries as yield losses (in tum often unrelated to actual yield they make the transition to scientific farming (IRRI losses), promote the use of pesticides (Rola and 1994 and Goodell 1984). IPM requires farmers to Pingali 1993). Consequently, high and injudicious grasp complex sets of data that are often anything but applications of broad-spectrum pesticides continue as self-evident, unitary, and standardized or amenable to before, causing the breakdown in varietal resistance. trial-and-error learning. The institutional and eco- New varieties generated to replace cultivars with nomic structure in the rural sector of developing resistance breakdown are subsequently overcome economies also requires some policy intervention to through further biotype changes in pest populations. reconcile long-term societal goals with short-term This "breeding treadmill" can only be overcome individual objectives in pest control. Promoting sus- through dramatic changes in crop management prac- tainable pest management within an IPM framework tices, especially in the use of pesticides. However, requires improved research-extension linkages, effec- whenever govemment policies, such as subsidies and tive farmer training methods, community action, and credit programs, keep pesticide prices artificially low, an undistorted price structure. the incentives are also low for farmers to invest in The iPM concept is holistic; it requires farmers to take knowledge that improves their perceptions and to a systems view of the farm enterprise and to under- adopt alternative pest control strategies with a broad stand the interlinkages among the various components ecological approach. of the system. Disseminating such a holistic message Under low levels of pest infestation, natural con- requires an extension system that has traditionally trol is the economically dominant strategy for manag- been geared toward promoting component technolo- ing pests (Rola and Pingali 1993). This is true for gies, such as improved seeds and fertilizer, to adopt insects and diseases for rice, wheat, and maize. new skills. Pest and pesticide problems are intrinsi- Natural control relies on predator populations to con- cally local in nature. National policy ought to nourish trol pest infestations under normal circumstances, the rural community's capacity to handle pest prob- when pest-resistant varieties are used. Pesticides may lems effectively, profitably, and equitably. To be suc- have to be used as a last resort in the rare instance of cessful, the IPM concept needs to be adapted to partic- high pest infestations. Natural control does not imply ular local situations. Such adaptation has to be done "do nothing"; rather, it is based on the premise of in- with close collaboration of researchers, extension per- depth farmer knowledge of the pest-predator ecology sonnel, and farmers. Most IPM success stories have and frequent monitoring of field conditions by the been preceded by research done in farmer fields with 266 Toward Reducea . _ the farmer actively participating in all stages of the rule is based on detailed pest ecology studies, which research process (Escalada and Heong 1993). have shown that the predominant insect pests during The wider dissemination of locally validated the first 40 days of crop growth are leaf-feeding research results requires a decentralized extension insects and that even very high levels of infestations system in which the formulation of the extension mes- of these insects rarely lead to any loss in yield sage is determined at the subprovincial or municipal (Heong 1990). Leaf-feeding insects are very visible, level. In such a system, the extension worker acts as a and farmers tend to attach great importance to con- local-level researcher adapting research and technolo- trolling them (Heong, Escalada, and Lazaro 1995). gy to suit local agroecological conditions rather than Controlling leaf-feeding insects, however, comes at a as an agent transmitting messages that are barely cost that substantially exceeds the value of yield understood and often inappropriate to the situation at saved, if any. Early-season insecticide applications hand. Such a decentralized extension setup will only tend to wipe out the leaf-feeding insects as well as the come about with a paradigm shift from the current beneficial predator populations that are building up emphasis on the top-down transfer of information. in the paddy. Rice paddies receiving one or two Farmer training in sustainable pest management is insecticide applications within the first 40 days tend an essential component of a strategy toward minimum to be susceptible to secondary pest infestations, espe- use of insecticides. The eventual goal is to build cially the brown planthopper, which build up unchal- farmer capacity in identifying and solving problems lenged due to the lack of natural controls. Controlling based on a thorough understanding of field ecology. the growth of secondary pests requires further appli- The experience of the Food and Agriculture cations, thus spiraling the use of insecticides. By not Organization's farmer field schools has shown that applying insecticides early in the crop season, farm- trained farmers use significantly lower levels of pes- ers can reduce the need for them later on due to the ticides than untrained farmers (Kenmore 1991). abundance of predator populations in the paddy Trained farmers are also more likely to experiment fields. with other components of sustainable production sys- The "no spray for 40 days" rule is an example of tems, such as improved fertilizer management and a simple message, distilled from in-depth scientific more efficient water management. investigation, that can be transmitted easily to farm- There are as yet several unresolved research and ers. A set of such rules, which are mutually consis- policy issues related to intensive farmer training in tent, could go a long way toward improving the man- IPM. The most imrportant one concerns the costs and agement of insect pests in tropical cereal production. benefits of farmer training. The costs of training the Simple rules ought to be seen not as a substitute for 120 million farmers in Asia are enormous and need to farmer training but rather as a strong complement to be justified in terms of farm-level benefits, both a training program. Simple messages can be trans- reduced expenditures on insecticides and reduced mitted rapidly even while investments are being social costs of pesticide use. Attention also should be made to reach all tropical rice farmers with intensive focused on opportunities for reducing the overall training. costs of training. There are essentially two options, which are not mutually exclusive, for reducing train- Community Action ing costs. The first is to train a core group of farmers within a geopolitical unit, such as a municipality, and The success of pest management programs depends then rely on farmer-to-farmer training for disseminat- on, among other things, collective organization ing the IPM message to a wider group of farmers. against infestations of migratory pests. Actions of There are definite scale economies to the farmer-to- individual farmers in the management of their pest farmer training approach if the quality of the message problems could have detrimental effects on the com- transmitted does not deteriorate as it gets passed munity as a whole. In this regard, management of down the line. pests can be viewed as a common property problem The second option is to condense the complex that is best dealt with through effective collective message into several simple rules that are easy for the action. In Asian agriculture, where farms are uni- farmer to implement. An example, from rice, of such formly small and farmers are nearly homogeneous, a rule is: Do not spray insecticides against leaf-feed- collective action for pest control seems quite attrac- ing insects for the first 40 days of crop growth. The tive. Synchronization of planting is the single most Prabhu L Pingali and Roberta V Gerpacio 267 important community action strategy in pest manage- ciated with indiscriminate pesticide use. Antle and ment. Pingali (1994) have shown that a 100 percent tax on Asynchrony is significantly related to the buildup insecticides for rice could actually improve produc- and field-to-field carryover of pest populations. tivity by improving farmer health through a reduction Because synchronous planting can prevent the build in exposure and an increase in labor productivity. For up of pests and thereby reduce the damage to crops, industrial-country agriculture, Carlson and Wetzstein it could greatly improve yields. The extent of syn- (1993) and Zilberman and others (1991) argue that chrony would be based on the minimum period in pesticide taxation has the potential to eliminate many which pests can complete one life cycle. Rice fields of the problems associated with pesticide use. in a contiguous area have to be planted within three to four weeks in order to capture the benefits of syn- References chrony (Heinrichs and Mochida 1984). Several factors may constrain synchronization of The word "processed" describes informally repro- planting schedules in a particular location and for duced works that may not be commonly available particular crops. Loevinsohn (1985, 1987) investi- through libraries. gates the causes, extent, and effects of synchronous Ahl Goy, P., and J. Duesing. 1995. "From Pots to rice cultivation in Nueva Ecija, Philippines. These Plots: Genetically Modified Plants on Trial." include irregularities in water distribution and Bio/Technology 13(5):454-58. drainage or variation in the arrival of irrigation water Antle, J. M., and Prabhu L. Pingali. 1994. "Pesticides, at farms, unavailability of a tractor or custom hire Productivity, and Farmer Health: A Philippine work during land preparation, unavailability of labor Case Study." American Journal of Agricultural during transplanting, and lack of access to credit. Economics 76(3):418-30. On the whole, treating pests as a common proper- Beachy, R. N., B. L. Subba Rao, K. Gough, P. Shen, ty resource (or a public liability) would entail collec- and M. Kaniewska, with H. H. Hibino. 1989. tive action by constituents and support from the public "Characterization of the Viruses Associated with sector. Recognizing both the positive and the nega- the Tungro Disease." Paper presented at the third tive externalities of pest control agents, group action annual meeting of the Rockefeller Foundation could be much more effective than individual action. International Program on Rice Biotechnology, The challenge to governments is to create an envi- Colombia, Mo., 8-10 March. Processed. ronment that promotes these strategies in ways that Boardman, R. 1986. Pesticides in World Agriculture: achieve growth, equity, and environmental sustain- The Politics of International Regulation. New ability. York: St. Martin's Press. Boserup, Ester. 1965. Conditions of Agricultural Pesticide Pricing Policies and IPM Growth. Chicago: Aldine Publishing Company. Carlson, Gerald, and M. E. Wetzstein. 1993. Even with a well-established IPM program, pesticides "Pesticides and Pest Manageiment." In G. Carlson, may have to be kept as a technology-of-last-resort. D. Zilberman, and J. A. Miranowski, eds., Essentially, the idea of using pesticides in IPM is to Agricultural and Environmental Resource spray only when imperative, using the smallest Economics. New York: Oxford University Press. amount possible to do the job. To make IPM more Commonwealth Agricultural Bureaux. 1980. attractive, pesticides ought not to be subsidized. In Perspectives in World Agriculture. London: Unwin fact, farmers will bother to learn and to apply IPM Brothers; Old Woking, U.K.: The Gresham Press. techniques only when the cost savings associated Cormwell, J. 1995. "Risk Assessment and Health with insecticide application make it worth their Effects of Pesticides Used in Tobacco Farming in while. The highly acclaimed IPM training in Malaysia." Health Policy and Planning (United Indonesia was preceded by a comprehensive reform Kingdom) 10(December):43-1-37. of pesticide regulation and the removal of all subsi- Crissman, C., and D. Cole. 1994. "Economics of dies for rice insecticides (Rola and Pingali 1993). In Pesticide Use and Farm Worker Health in addition to the removal of subsidies, a sustainable Ecuadorian Potato Production." American Journal pest management program may warrant the taxation ofAgricultural Economics 76(August):593-97. of agrochemicals to account for the social costs asso- Datta, S. K., A. l'eterhaus, K. Datta, and I. Potrykus. 268 Toward Reduced Pesticide Use for Cereal Crops in Asia 1990. "Genetically Engineered Fertile Indica Rice Ecology 7(2-3):201-18. Recovered from Protoplasts." BiolTechnology Heong, K. L. 1990. "Feeding Rates of the Rice 8(8):736-40. Leaffolder, Chaphalocrocis medinalis (Lepidoptera: Dubin, H. J., and S. Rajaram. 1996. "Breeding Pyralidae), on Different Plant Stages." Journal of Disease-Resistant Wheats for Tropical Highlands Agricultural Entomology 7(2):81-90. and Lowlands." Annual Review of Phytopathology Heong, K. L., M. M. Escalada, and A. A. Lazaro. 34:503-26. 1995. "Misuse of Pesticides among Rice Farmers Edmeades, G. E., and J. A. -Deutsch, eds. 1994. Stress in Leyte, Philippines." In Prabhu L. Pingali and P. Tolerance Breeding: Maize That Resists Insects, A. Roger, eds., Impact of Pesticides on Fanner Drought, Low Nitrogen, and Acid Soils. Maize Health and the Rice Environment. Norwell, Mass.: Program Special Report. Mexico D.F. : Centro Kluwer Academic Publishers; Los Bafnos, Internacional de Mejoramiento de Mafz y Trigo. Philippines: International Rice Research Institute. Escalada, M. N'[., and K. L. Heong. 1993. "Com- Heong, K. L., M. M. Escalada, and Mai Vo. 1994. "An munication and Implementation of Change in Crop Analysis of Insecticide Use in Rice: Case Studies Protection." In D. J. Chadwick and J. Marsh, eds., in the Philippines and Vietnam." In Rice 1Pm Crop Protection and Sustainable Agriculture. Ciba Network Workshop Report on Message Design for Foundation symposium on world food production a Campaign to Encourage Farmers' Participation by means of sustainable agriculture: the role of in Experimenting with Stopping Early Insecticide crop protection, held in collaboration with the Spraying in Vietnam. Network workshop, Ho Chi Centre for Research on Sustainable Agricultural Minh City, Vietnam, 25-28 May. Los Bafios, and Rural Development, Madras, India, 30 Philippines: International Rice Research Institute. November-2 December 1992. Chichester, U.K.: Herdt, R. W., and F. Z. Riely. 1987. "International John Wiley and Sons. Rice Research Priorities: Implications for FAO (Food and Agriculture Organization of the United Biotechnology Initiatives." In Proceedings of the Nations). 1997. "Statistics on Pesticide Use." Rockefeller Workshop on Allocating Resources for Internet copy. Statistical Analysis Service, Developing-Country Agricultural Research, 6-10 Statistics Division, Rome. Processed. July, Bellagio, Italy. New York: Rockefeller Forget, G., T. Goodman, and A. de Villiers. 1993. Foundation. Impact of Pesticide Use on Health in Developing IFPRI (International Food Policy Research Institute). Countries. Proceedings of a symposium held in 1996. Pest Management: The Expected Trends and Ottawa, Canada, 17-20 September 1990. Ottawa: Impact on Agriculture and Natural Resources to International Development and Research Centre. 2020. 2020 Vision Food, Agriculture, and the Fujimoto, H., H. Itoh, M. Yamamoto, J. Kyozuka, and Environment Discussion Paper. Washington, D.C. K. Shimamoto. 1993. "Insect Resistant Rice IRRI (International Rice Research Institute). 1972. Generated by Introduction of a Modified Delta- Rice, Science, and Man. Papers presented at the Endotoxin Gene of Bacillus thuringiensis." tenth anniversary celebration of the International Bio/lechnology 11(10, October):1151-61. Rice Research Institute, 20-21 April. Los Bafios, Goodell, G. 1984. "Challenges to International Pest Philippines. Management Research and Extension in the Third . 1985. Twenty-fifth Anniversary Celebration. World: Do We Really Want IPM to Work?" Bulletin of Papers presented at the twenty-fifth anniversary the Entomological Society of America 30(3):18-25. celebration of the International Rice Research Hayakawa, T., Y Zhu, K. Itoh, Y. Kimura, T. Izawa, Institute, 6-8 June. Los Banios, Philippines. K. Shimamoto, and S. Toriyama. 1992. . 1993a. IRRI Program Report for 1992. Los "Genetically Engineered Rice Resistant to Rice Bafios, Philippines. Stripe Virus, an Insect-Transmitted Virus." . 1993b. Rice in Crucial Environments: IRRJ Proceedings of the National Academy of Sciences 1992-1993 Corporate Report. Los Bafios, Philippines. 89(20):9865-69. . 1994. IRRI Program Report for 1993. Los Heinrichs, E. A., and 0. Mochida. 1984. "From Bafios, Philippines. Secondary to Major Pest Status: The Case of . 1995. World Rice Statistics 1993-94. Los Insecticide-Induced Rice Brown Planthopper, Bafios, Philippines. Nilaparvata lugens, Resurgence." Protection Joos, H., and W. Morrill. 1989. "Control of Insect Prabhui L. Pingali and Roberta V Gerpacio 269 Pests on Rice Using Bacillus thuringiensis Springer-Verlag. Genes." Paper presented at the third annual meet- Loevinsohn, M. 1985. "Agricultural Intensification ing of the Rockefeller Foundation International and Rice Pest Ecology: Lessons and Implications." Program on Rice Biotechnology, Colombia, Mo., Paper presentedl at the International Rice Research 8-10 March. Processed. conference, International Rice Research Institute, Joshi, R. C., B. M. Shepard, P. E. Kenmore, and R. Los Bafios, Ph[lippines, 1-5 June. Processed. Lydia. 1992. "Insecticide-Induced Resurgence of -. 1987. 'Insecticide UJse and Increased Brown Planthopper (BPH) on IR62." International Mortality in Rural Central Luzon, Philippines." Rice Research Newsletter 17(3, June):9-10. Lancet 1(8546):1359-62. Kabir, K. H., F. M. A. Rouf, M. E. Baksh, N. A. Mihm, J. A., J. Deutsch, D. Jewell, D. Hoisington, and Mondal, M. A. Hossain, M. A. Satter, and M. S. M. D. G. de Leon. 1994. "Improving Maize with Asaduzzaman. 1994. "Insecticide Usage Pattern Resistance to Major Insect Pests." In G. E. on Vegetables at Farmers' Level of Jessore Edmeades and J. A. Deutsch, eds., Stress Region: A Survey Report." Research report. Tolerance Breeding: Maize TFhat Resists Insects, Jessore Regional Agricultural Research Stations, Drought, Low Nitrogen, and Acid Soils. Maize Jessore, Bangladesh. Processed. Program Special Report. Mexico D.F.: Centro Kenmore, P. E. 1991. Indonesia's Integrated Pest Internacional le Mejoramiento de Mafz y Trigo. Management: A Model for Asia. Intercountry IPc Moody, K. 1991. "Weed Management in Rice." In D. Rice Programme. Manila, Philippines: Food and Pimentel, ed., Handbook of Pest Management in Agriculture Organization of the United Nations. Agriculture, 2d ed., pp. 301-28. Boca Raton, Fla.: Kenmore, P. E., F. 0. Carifno, C. A. P6rez, V. A. Dyck, CRC Press. and A. P. Gutierrez. 1984. "Population Regulation Naylor, R. 1997. Herbicides in Asian Rice: of the Rice Brown Planthopper (Nilaparvata Transitions in Weed Management. Palo Alto, Luigens) within Rice Fields in the Philippines." Calif.: Stanford University, Institute for Journial of Plant Protection in the Tropics 1(1): International Studies; Los Bafios, Philippines: 19-37. International I'ice Research Institute. Khush, G. 1996. "Rice Varietal Improvement for Pagiola, Stefano. 1995. Environmental and Natural Weed Management." in R. Naylor, ed., Herbicide Resource Degradation in Intensive Agriculture in in Asian Rice: Transition in Weed Management. Bangladesh. Environmental Economic Paper 15. Palo Alto, Calif.: Stanford University, Institute of Washington, D.C.: World Bank. International Studies; Los Banios, Philippines: Pathak, M. D., and V. A. Dyck. 1973. "Developing an International Rice Research Institute. Integrated Method of Rice Insect Pest Control." Kim, Soon-Kwon, V. 0. Adetimirin, and A. Y. PANS 19(4):534-44. Akintunde. 1997. "Nitrogen Effects on Striga Persley, G. J. 1990. Beyond Mendel's Garden: Hermonthica Infestation, Grain Yield, and Biotechnology in the Service of WorldAgriculture. Agronornic Traits of Tolerant and Susceptible Wallingford, U.K.: C.A.B. International. Maize Hybrids." Crop Science 37(3):711-16. Pingali, Prabhu L., and Hans P. Binswanger. 1987. Kumar, H., and J. A. Mihm. 1995. "Antibiosis and "Population Density and Agricultural Tolerance to Fall Armyworm, Spodoptera Intensification: A Study of the Evolution of Frugiperda (J. E. Smith), Southwestern Corn Technologies in Tropical Agriculture." In G. Borer, Diatraea Grandiosella Dyar, and Johnson and IR. Lee, eds., Population Growth and Sugarcane Borer, Diatraea Saccharalis Fabricius, Econonmic Development. Washington, D.C.: in Selected Maize Hybrids and Varieties." National Research Council. Maydica 40(3):245-51. Pingali, Prabhu L., and R. V. Gerpacio. 1997. "Living Lin, W., C. S. Anuratha, K. Datta, I. Potykus, S. with Reduced Insecticide Use in Tropical Rice." Muthukrishnan, and S. K. Datta. 1995. "Genetic Food Policy 22(2):107-18. Engineering of Rice for Resistance to Sheath Pingali, Prabhu L., M. Hossain, and R. V. Gerpacio. Blight." BiolTechnology 13(7):686-91. 1997. Asiamn Rice Bowls: The Returning Crisis? Litsinger, J. A. 1991. "Crop Loss Assessment in Wallingford, U.K.: C.A.B. International. Rice." In E. A. Heinrichs and others, eds., Rice Pingali, Prabhu L., and C. B. Marquez. 1997. Insects: Management Strategies. New York: "Herbicides and Rice Farmer Health: A Philippine 270 Toward Reduced Pesticide Use for Cereal Crops in Asia Study." In R. Naylor, ed., Herbicides in Asian Maiz y Trigo. Rice: Transitions in Weed Management. Palo Alto, Sayre, K. D., R. P. Singh, J. Huerta-Espino, and S. Calif.: Stanford University, Institute for Rajaram. 1997. "Genetic Progress in Reducing International Studies; Los Banos, Philippines: Losses to Leaf Rust in CIMMYT-Derived Mexican International Rice Research Institute. Wheat Cultivars." CIMMYT Wheat Program, Centro Pingali, Prabhu L., and S. Rajaram. 1997. Intemacional de Mejoramiento de Maiz y Trigo, "Technological Opportunities for Sustaining Mexico D.F. Processed. Wheat Productivity Growth." Paper presented at Schoenly, K., J. E. Cohen, K. L. Heong, G. Arida, A. the Illinois World Food and Sustainable T. Barrion, and J. A. Litsinger. 1996. "Quantifying Agriculture Program conference on meeting the the Impact of Insecticides on Food Web Structure demand for food in the twenty-first century: chal- of Rice-Arthropod Populations in Philippine lenges and opportunities for Illinois agriculture, Farmers' Irrigated Fields: A Case Study." In G. Urbana-Champaign, Ill., 27 May. Processed. Polis and K. Winemiller, eds., Food Webs: Pingali, Prabhu L., and P. A. Roger, eds. 1995. Impact Integration of Patterns and Dynamics, chap. 32. of Pesticides on Farmer Health and the Rice New York: Chapman and Hall. Environment. Norwell, Mass.: Kluwer Academic United Nations. 1990. Commodity Trade Statistics. Publishers; Los Baiios, Philippines: International Statistical Paper Series D, vol. 38. no. 1-19. New York. Rice Research Institute. Waibel, H. 1986. The Economics of Integrated Pest Pingali, Prabhu L., and A. C. Rola. 1995. "Public Control in Irrigated Rice: A Case Study from the Regulatory Roles in Developing Markets: The Philippines. Crop Protection Monograph. Berlin: Case of Pesticides." In Prabhu L. Pingali and P. A. Springer-Verlag. Roger, eds., Impact ofPesticides on Fanner Health Warburton, H., F. G. Palis, and Prabhu L. Pingali. and the Rice Environment. Norwell, Mass.: 1995. "Farmer Perceptions, Knowledge, and Kluwer Academic Publishers; Los Banios, Pesticide Use Practices." In Prabhu L. Pingali and Philippines: lIntemational Rice Research Institute. P. A. Roger, eds., Impact of Pesticides on Farmer Rola, A. C., and Prabhu L. Pingali. 1992. "Choice of Health and the Rice Environment. Norwell, Mass.: Crop Protection Technologies under Risk: An Kluwer Academic Publishers; Los Bafios, Expected Utility Maximization Framework." Philippines: International Rice Research Institute. Philippine Journal of Crop Science 17(1):45-54. Weir, D., and M. Shapiro. 1981. The Circle of Poison: . 1993. Pesticides, Rice Productivity, and Pesticide and People in a Hungry World. San Farmers' Health: An Economic Assessment. Francisco: Third World Publications for the Washington, D.C.: World Resources Institute; Los Institute for Food and Development Policy. Bafios, Philippines: International Rice Research Wood Mackenzie Consultants Ltd. 1993. "Global Institute. Rice Pesticide Market." A report submitted to the Sayre, K. D. 1996. "The Role of Crop Management Plant Protection Department of Ciba-Geigy Ltd., Research at CIMMYT in Addressing Bread Wheat Basil, Switzerland. Processed. Yield Potential Issues." In M. P. Reynolds, S. Zilberman, D., A. Schmitz, G. Casterline, E. Rajaram, ancd A. Macnab, eds., Increasing Yield Lichtenberg, and J. B. Siebert. 1991. "The Potential in VVheat: Breaking the Barriers. Mexico Economics of Pesticide Use and Regulation." D.F.: Centro Intemacional de Mejoramiento de Science 253(August 2):518-22. 20 The Economic and Environmental Impacts of Irrigation and Dlrainage in Developing Countries Ujjayant Chakravorty The world's population is expected to grow from competition with the diversion of water for irrigation approximately 6 billion today to nearly 10 billion by in the United States and in many other industrial coun- the middle of the next century (United Nations tries. The same pattern is likely to repeat itself in the Population Fund 1993). Most of this growth will take developing countries with the rise in environmental place in the developing world, increasing its share of consciousness and the concornitant degradation of the earth's population to 90 percent. Over the past environmental resources. three decades, the dramatic spread of irrigated agricul- Thus the irrigation sector in developing countries ture, which enabled widespread adoption of high- will be expected to deliver higher agricultural output yielding production technology, was mainly responsi- with reduced supplies of water, because there is little ble for enabling increases in food production to keep hope that new investment in irrigation will match the pace with the growth in population. This was global trend of diverting water from agriculture to achieved through the large-scale development of new urban and environmental uses. The capital cost of new water resources and the construction of new irrigation irrigation deve:Lopment rose 70-116 percent per capacity. hectare in the 19 80s alone, according to figures avail- Both rising population and rising incomes due to able from the World Bank (Dinar 1997), and new economic development are expected to increase sub- investment in irrigation has declined substantially in stantially the demand for food in the coming years. countries such as China and India. Producing more The International Irrigation Management Institute food with less water-placing a smaller financial bur- (iIMI 1992) has estimated that to meet this demand, den on the taxpayer and reducing the environmental irrigated agriculture will need to increase output more costs-will be the major challenge for irrigation in the than 3.5 percent annually-a daunting task by any twenty-first century (Lenton 1994). measure of performance. However, the major problem However, in spite of its significant contribution to is that this growth must come primarily from more agricultural procluction and food security, the current efficient irrigation rather than from new capacity. state of irrigaticn is not encouraging. Despite major Moreover, as is becoming more common in the indus- investments in irrigation during the past four trial countries and even in many developing countries, decades-the World Bank alone has invested more higher residential demand for water and rising envi- than $31 billion and leveraged an additional $53 bil- ronmental concerns are beginning to restrict new water lion from partner countries and cofinanciers-irriga- development projects. Environmental demands for tion projects around the world are in serious trouble water to preserve instream uses, wetlands, freshwater for a multitude of reasons: ex post project benefits fisheries, and other recreational purposes are in direct have often been far lower than projected ex ante 271 272 The Economic and Environmental Impacts of Irrigation and Drainage in Developing Countries returns, and full-cost pricing of water is rare, with The Economics and Pricing of Irrigation Water: most projects recouping only a fraction of their opera- Critical Issues tion and maintenance (o&M) costs. The low price of irrigation water, which is often unrelated to water use, Economic principles dictate that under certain condi- has led farmers to withdraw too much, giving rise to tions regarding the characteristics of the resource and widespread salinity and waterlogging. For example, of the market in which it is allocated, resource alloca- Biswas (1991) suggests that more than a quarter of tion is best done by means of a smoothly functioning global irrigated acreage has been damaged by saliniza- market system (Young and Haveman 1985). In the tion. case of water, these conditions are not met on several On a broader scale, investments in the irrigation counts. For example, water is mobile, which makes it sector in developing countries have relied too much on difficult to measure and hold, thereby making difficult bureaucratic federal and state agencies that have paid the establishment of property rights. Economies of little attention to the economic pricing of irrigation scale in water generation and conveyance often create services, the creation of reliable delivery systems, and conditions for a natural monopoly. The joint product the participation of users in the operation and mainte- nature of water, which enables multiple users to nance of projects. Adverse health and environmental access a given stock of water sequentially, and the effects of water use have received little attention, part- externalities created as a result of the degradation of ly because of the lack of coordination and interaction quality through use make the creation and functioning between water and public health agencies. of water markets extremely difficult. For these rea- These problefns call for a new approach that is sons, the water sector both in industrial and in devel- more comprehensive and that integrates the different oping countries has been characterized by public elements of irrigation policy, which include the eco- investment in the construction of irrigation projects, nomic and environmental effects and the associated state ownership and operation of facilities, and regu- intersectoral linkages. This approach must be built on lated distribution and allocation of water. pricing of resources that reflects opportunity costs Irrigation water projects in developing countries whenever possible and the creation of delivery infra- face several important issues that arise from the exis- structure that pro.motes the efficient allocation and, if tence of insufficient or outmoded institutional mecha- feasible, trading of the resource among users. nisms and economic policies. Many of the policies This chapter reviews current thinking and practice that allocate water among users were introduced at a on the economics of irrigation and drainage in devel- time when there was no perceived scarcity of water oping countries. In particular, it examines recent expe- and the primary goal was to provide cheap and abun- riences with the reform of irrigation water policy and dant water supplies to encourage people to occupy attempts to draw broad conclusions and policy recom- and farm new settlement areas. These rules are most- mendations for which approaches work and which do ly some variation of the first-come, first-served prin- not and under what conditions. The chapter is ciple that allowed settlements to occur at low enforce- addressed to policymakers and practitioners and aims ment and transaction costs. These distributional to complement their specific expertise on resource mechanisms are still in use in most irrigation projects, management issues with a more global perspective on where water allocation is based on the seniority of issues concerning the economics and management of water rights, which prevents the trading and transfer agricultural water resources. of water among users. Low tax collections and poor The second section discusses critical issues related to o&M of project structures have further led to high the economics of irrigation in the developing countries, rates of water loss through seepage and percolation, followed by a section highlighting problems related to and inadequate investments in drainage have rendered rainfed agriculture, which is still the primary means of large tracts of prime agricultural land unusable food production in the developing world. Next, a section because of salinity and waterlogging. deals with the environmental problems that have arisen as a result of inadequate drainage and have resulted in Conveyance Losses large irrigated areas losing their productive potential. This is followed by a section highlighting a few impor- One major problem in irrigation management is the tant factors that contribute to the success or failure of irri- continued use of traditional design principles that gation projects. A final section concludes. implicitly assume that the opportunity cost of water is Ujjayant Chakravorty 273 zero. These techniques were appropriate in an era officials taking bribes for supplying more water to when water was plentiful. For example, the design certain villages and with the bribery permeating every capacity at the source and the allocation of water to level of the bureaucratic and political apparatus. The each farmer are often based on exogenously deter- system has a well-developed scheme of rewards and mined crop water requirements, not on the shadow punishment: bureaucrats who participate in corruption price of water. This implies the overuse of available are transferred to "lucrative" posts, while those who water resources to irrigate a project area that is small- do not are shunted to undesirable or remote locations. er than economically optimal (Chakravorty and In a simulation model of this phenomenon, Roumasset 1994). Frequently, politicians, in an Chakravorty, Hochman, and Zilberman (1995) find attempt to maximize the number of beneficiaries, that under plausible conditions, better management of exert pressure to extend the project area beyond what canal water may expand the service area of the project is specified in the design (Repetto 1986). Over time, by as much as 200-300 percent, with similar increas- large transmission losses and excessive withdrawals es in agricultural output. This can be done, for exam- of water by large and influential farmers (who typi- ple, by water users associations that collect water cally are located in the upstream reaches of the project charges as well as operate and imaintain the distribu- canals) shrink the irrigated area covered by the proj- tion system. Reducing water losses through better ect. maintenance allows the irrigated area in their model Wade (1984) provides a fascinating description of to expand. However, if the demand elasticity for agri- this phenomenon for an irrigation project in South cultural production is low, then the gains in service India. He points out that canal performance can be area and output may be smaller. Rigorous empirical substantially improved by better operation and main- evidence of the impact of better canal management is tenance and that output can be expanded significantly not available, however, and is an important problem with as little as 50-100 percent more outlays for for further research. maintenance and without expensive upgrading of facilities. Wade describes efficiency in an irrigation Salinity and Waterlogging project as akin to stretching a membrane. Higher lev- els of canal performance require the water supply to The complex relationship between transmission loss- be stretched through better management. Poor opera- es and the buildup of salinity and waterlogging is tion and maintenance, in contrast, lead to a contrac- revealed in an ongoing study conducted by the tion of the system, limiting water allocation to Government of P'akistan and the IIMI (Lenton 1994). upstream locations. The study finds that because the delivery of water in Wade further compares two water delivery sched- the lower reaches of the distribution system was unre- ules in northern and southern India that have very liable, farmers leaned heavily on tubewell systems for different impacts on irrigation efficiency and water irrigation. Because the tubewell water was of poor distribution. In northwestern India, irrigation is char- quality, intensive use led to high salinity in the project acterized by elaborate rotational delivery schedules area. This discovery was a complete surprise to the that require all farmers to share the available water. project researchers, who had initially assumed that This is enforced by allowing each field to withdraw salinity was the direct result of a rising water table. In all the water available in the channel for a fixed num- this project, researchers found no evidence of water- ber of minutes each week. This system ensures that logging, but they did find a high incidence of salinity. both head- and tail-enders receive their allocations. In They also observed that poor maintenance of distri- contrast, in southern India, paddy irrigation is charac- bution canals, unauthorized withdrawals of water, and terized by continuous flow. Water flows out of all out- capacity constraints led to spatial inequity between lets at the same time, and the Irrigation Department is farmers at the head and tail ends. Farmers at the tail supposed to ensure that the head-enders do not take end received much less water than those at the head more than their allocated amount and leave water in end and faced significantly more variable water sup- the canal for the tail-enders. This system also supplies plies. According to the author, these findings repre- the full water requirement to certain blocks, while sent a radical departure from the traditional approach depriving lands not in the service area of any water at to the problem of salinity, which focuses on waterlog- all. Wade provides a detailed account of the opportu- ging and overwatering. The new approach to control- nity for corruption that this system promotes, with ling salinity in the project focuses on the adoption of 274 The Econiomic and Environmental Impacts of Irrigation and Drainage in Developing Countries crops with a higher tolerance for salinity and better ment charges a flat rate for electricity use, which is management of water distribution, which helps to one reason why there has been a rapid growth in pri- reduce the high reliance on tubewell irrigation. vate tubewells run on electricity. The flat rate induces farmers to overirrigate and to select larger-capacity Maintenance of Plublic Irrigation Facilities pumps so that they can irrigate quickly, which is use- ful in an environment of random and fluctuating One of the most serious issues facing policymakers power supply. and international agencies like the World Bank and A major study of a sample of World Bank-funded the regional banks is the need to maintain project projects in irrigation and drainage between 1961 and structures. Service in public irrigation facilities tends 1985 reveals some interesting results. Jones (1994) to deteriorate rapidly a few years after commissioning, finds that estimates of economic return from projects mostly because of poor maintenance. Systems are fre- average 15 percent, lower than the 21 percent predict- quently operated so that the availability of water has ed at the project appraisal stage. The study finds that little to do with the seasonal pattem of water demand two of the biggest factors affecting returns from irri- by farmers (Repetto 1986). The warabandi rotation gation are those over which the project has no con- schedule, long popular in South Asia as a means of trol-macroeconomic distortions and commodity delivering canal water to farmers on a rotation sched- prices. However, among the factors that can be con- ule, is sometimes unreliable, with farmers never trolled is the scale of the project-the benefit-cost knowing when they will receive their allocation. This ratio seems to be greater in bigger projects. The study is mainly because of poor maintenance in upstream does not find a direct relationship between benefits regions and the natural instability of the canal bed, from an irrigation project and o&M, mainly because which in turn affects the water level and the amount water charges are usually paid directly into the central of water withdrawn from each individual outlet. This treasury, although some countries have enacted legis- uncertainty leads farmers to underinvest in important lation requiring collections to be used directly for production inputs, such as fertilizer and seeds. o&M (Radosevich 1986). However, other studies, In the public tubewell irrigation system in the state such as the Asian Development Bank's assessment of of Uttar Pradesh in India, steady decreases are irrigation user charges, suggests a direct link between observed in both the hours of operation and the area cost recovery and the financial autonomy of the orga- irrigated per tubewell. So although the number of nization undertaking o&M. In projects where o&M tubewells has increased over the years, the total activities are financed out of water fees, cost recovery acreage served has declined. These systems were put is achieved up to the costs of operation and mainte- in place to provide irrigation for low annual cropping nance. intensity and limited water application. However, the surface distribution system has failed to distribute the Water Trading and Markets limited water over the target area. Incomplete distri- bution canals have meant that farmers farther away Economists generally believe that the introduction of from the source do not receive enough water. markets and trading among users lead to efficient Inadequate operation and maintenance have led to fre- allocation. Water markets function in different parts of quent breakdowns, delayed repairs, and interrupted the world, often without help from outside parties. service (Cunningham 1992). Despite the poor perfor- However, they only work under very selective condi- mance of public irrigation systems, the government tions and do not seem to work when any of those con- has continued to develop public groundwater ditions are not satisfied. For example, in Alicante, schemes, mainly because such schemes have relative- Spain, irrigation systems in the huerta (irrigation ser- ly short construction periods and can reach beneficia- vice area) consist mostly of small farmers cultivating ries who either cannot afford private tubewells or up to 1 hectare of land or less. Water supply is scarce, inhabit inaccessible regions. irregular, and sufficient only for about a quarter of the Groundwater irrigation is characterized by high land (Reidinger 1994). The huerta has been in use for operating costs relative to capital costs, unlike surface hundreds of years and receives water from a reservoir, water irrigation. The government also provides subsi- two interbasin canals, and a local groundwater source. dized capital costs by providing farmers with cheap The water users association, which runs the irrigation electricity and diesel. In Uttar Pradesh, the govern- system, owns only part of the water and buys the rest Ujjayant Chakravorty 275 on a regular basis. It allocates water by holding auc- * Removal of legal and institutional regulations that tions every week in which tickets to a certain amount prohibit or limit water trading. For example, water and duration of water flow are traded. This system rights could le made distinct from land rights and allows for scarce water in the huerta to be allocated to should be specified in volumetric terms. its most valued use. * Creation of infrastructure (such as canals) and con- In the northern Indian states of Punjab and trol systems for the transfer and measurement of Haryana, a limited water market, locally called wara- water. However, the capital, operation, and main- bandi, exists in which farmers trade water turns. This tenance costs of these systems must be lower than happens mainly because the operating system of the the incremental benefits from water trading. irrigation canals operates on a different schedule than the water scheduling needs of the farmers, resulting in Effect on Income Distribution long delays between irrigation (Reidinger 1974). In California, where surface water rights are based In the long run, properly functioning water markets on a first-come, first-served queuing system, users are may allocate scarce supplies to their most valued use. allowed to divert water for "beneficial" use but are They may also improve the efficiency of water in the not allowed to sell the water or to trade it among agricultural sector and thus release water for alterna- themselves. These systems generally allow users to tive uses, such as to meet urban needs and protect draw as much water as they need, often paying a aquatic habitat and coastal areas. However, in many diversion fee that is much lower than the true cost of developing countries, the transition from the short run water delivery. The California drought of 1986-92 to the long run is of profound importance and is beset reduced the state's water supplies by nearly half. The with serious problems related to the redistribution of water districts responded to this situation through sev- rents accruing to beneficiaries. For example, as point- eral measures, including creating a water bank that ed out by Repetto (1996), Wade (1984), and others, bought and sold water. The limited water trading most irrigation systems in developing countries are induced by the creation of the water bank was highly heavily subsidized by the taxpayer, with substantial successful and enabled the transfer of water from sur- rents from water accruing to large landowners and plus to deficit plots. The reallocation of water to higher- financially better-off farmers, who are usually located valued crops and fallowing of lands that were being in the upper reaches of the distribution system. Any used to cultivate water-intensive crops (such as alfal- policy reforms that introduce water markets and fa) resulted in only a 20 percent reduction in gross improve conveyance structures that facilitate water revenue, using only half the amount of water as trading will move water prices closer to opportunity before. Although drought conditions no longer exist cost. As Chakravorty, Hochman, and Zilberman in California, there is a move to continue encouraging (1995) show, these reforms would allow the improved water trading in the state, partly because lawmakers project to serve more beneficiaries but would also sig- are under pressure to allocate more water to environ- nificantly reduce the rents accruing to beneficiaries mental uses. Sophisticated trading mechanisms have located closer to the upstream regions. Reform pro- been introduced, such as an electronic bulletin board posals must take this into account, possibly by subsi- for the exchange of information among water suppli- dizing upstream farmers or making them shareholders ers and demanders in the huge Westlands Water of the water utility, so that they receive a share of the District in the Central Valley, an experiment that is aggregate rents from the project. now being emulated by 23 other districts (Chak- Another problem is the possible capitalization of ravorty and Zilberman 1997). infrastructure improvements into land values and the These experiences in limited water trading pro- displacement of small and marginal farmers by more duce several important clues to the conditions that resourceful farmers. For example, if upstream farmers are need to be created for trading to occur: allowed to sell their water, downstream farmers will * The presence of effective water users associations be subject to even smaller supplies. In that case, poli- or water authorities. Such an organization can be cies need to be designed to ensure that downstream effective in resolving disputes over water trades, farmers can obtain rights over use of the return flows including their third party impacts. These associa- or are compensated with some portion of the proceeds tions can also help organize trades and provide of the trade. The large inequities prevalent in most information on them. developing-country irrigation systems make it imper- 276 The Economic and Environmental Impacts of Irrigation and Drainage in Developing Countries ative that income distribution issues of irrigation sumers outside the project, for example, in the form of reform be studied thoroughly before they are imple- lower food prices and increased food security or mented. through spillover effects in the regional economy. If these externalities were incorporated in the marginal- Marginal-Cost Pricing cost calculations, it might well be optimal to charge farmers below marginal cost. There are other prob- Pricing of irrigation services has probably been the lems with marginal-cost pricing. Farming activities most troublesome issue to deal with for both intema- vary from season to season and across locations with- tional agencies involved in irrigation as well as in the same project and climate. To charge marginal- national and local govermments. In most developing cost prices each time would require not only perfect countries, the revenue collected is not even enough to information on the parameters of demand and cost but cover o&M costs, let alone capital expenditures. Water also the use of complex pricing formulas. This may pricing techniques vary considerably between coun- mean high administrative and political costs that tries and often between projects and regions within could well offset the gains from marginal-cost pricing the same country. The pricing principles applied in (Sampath 1992). However, as the scarcity value of most projects are based on financial criteria such as water rises with time, and the cost of computing time- recovery of o&M costs rather than on any economic and space-dependent prices decreases with the advent concept such as marginal-cost pricing. Not only do of low-cost microcomputers, it will become econom- O&M costs vary across projects, but also there is con- ically feasible to use complex pricing policies for irri- siderable uncertainty regarding their magnitude, and gation water as is the practice for other spatially dis- irrigation authorities in developing countries have tributed commodities such as electricity and natural been known to inflate their estimates to secure funds gas. for increased construction activity. In many countries, Implementation of marginal-cost pricing implies the inability of the irrigation agency to charge for that farmers farther from the water source must be water services reduces its claim on government charged higher prices to account for conveyance costs resources. For example, in Egypt, employee salary than those located closer to the source. This also scales for revenue-generating sectors are higher than implies that farmers located farther away will with- for nonrevenue-generating sectors such as irrigation draw less water than farmers in the head reaches of (Perry 1996). the distribution system. A pricing policy that discrim- Pricing schemes for irrigation water vary consider- inates between farmers at different locations may be ably from place to place. The most commonly imple- difficult to implement. It has been suggested that if mented schemes are not based on the quantity of conveyance losses are reduced by lining canals and water withdrawn but on a flat rate that is not related to maintaining control structures, the cost differences in actual use, often leading farmers to use water until its water delivery between locations will also be reduced, marginal value is zero. In many countries, water rates so that a spatially uniform pricing policy may be a are levied on a per hectare basis and depend on the reasonable approximation to marginal-cost pricing type and extent of crop irrigated. Often land taxes are (see Chakravorty, Hochman, and Zilberman 1995). levied to tax the increased benefits created by As Sampath (1992) points out, many water econo- improvements in irrigation. Betterment levies are also mists preach the virtues of marginal-cost pricing with- used to tax the higher land value capitalization from out realizing that social and political objectives may irrigation. In somne countries, such as Peru, volumetric support a deviation from pricing water at marginal fees are charged based on minutes of flow, while in cost. Pricing models developed by economists need to Israel, where water is distributed through closed incorporate these multiple objectives. Solving these pipes, block prices are in effect, which consist of a second-best models may yield water prices that are basic low price for an initial volume and higher prices closer to those observed in practice. The difference for additional blocks (Sampath 1992). between these second-best prices and the realized Various surveys have shown that marginal-cost prices may provide a more accurate picture of the sub- pricing principles are rarely applied in water projects optimality of water prices. However, although recog- in developing countries (see Small 1987). A reason nizing diverse social and political objectives (such as often cited for departures from marginal-cost pricing equitable distribution of benefits among beneficiaries) is the existence of externalities from irrigation to con- is important and provides important clues to current Ujjayant Chakravorty 277 pricing practices, it may be difficult to justify those mented a complete package of measures to ensure the objectives because the pricing policies they promote efficient use of water. For example, each farmer's cause a range of problems that were not originally allocation of water is based on the area cultivated, anticipated, such as adverse environmental impacts crop mix, and water requirements for each crop. and rent-seeking. What may be needed is analysis Wasteful practices come with penalties in the form of showing that these noneconomic objectives within the reduced acreage or fines for excessive consumption of irrigation sector may actually cause economic and water. Water charges are based on a progressive block environmental impacts that result in serious efficien- rate structure with a basic rate for 80 percent of the cy and equity effects in the economy as a whole. allocation and hiigher rates for additional amounts. The liMI has examined irrigation cost recovery and Water supplied (luring the peak irrigation season car- water pricing policies in USAID (U.S. Agency for ries a 40 percent premium. All the water supplied is International Development) projects in Egypt and the metered, and the government subsidizes the installa- impact of alternative pricing mechanisms on the agri- tion of more efficient irrigation technology and other culture sector (Perry 1996). The study evaluates three control devices. As a result, more than 90 percent of policies: (a) a flat rate, independent of type of crop or the irrigated area is under sprinkler irrigation, and the cropping intensity, (b) a crop-based charge linked to rest is under drip technology (Tahal Consulting water consumption, and (c) a volumetric charge. The Engineers 1991). results show that although a flat rate charge has no impact on crop selection, full recovery of allocated Water Manageinent in Rainfed Areas costs to agriculture would reduce farm incomes by about 4.5 percent. More important, the study finds In spite of phenomenal increases in acreage under irri- that a crude crop-based charge in which water charges gation, rainfed agriculture continues to be a major are proportional to farm demand is a good approxi- source of food production in many developing coun- mation for volumetric pricing because it induces tries. An estimated 70 percent of the cropped land in farmers to grow water-efficient crops. This finding Asia is rainfed. These production systems are com- suggests that volumetric pricing, which requires water plex and diverse and are constantly affected by the measurement, may not be cost-effective under condi- uncertainties of monsoon, drought, and flood cycles. tions of relatively plentiful supply. Rainfed farming systems are also characterized by Further modeling to simulate the impact of a 15 per- wide variations in soil quality, ranging from the dry cent reduction in water supplies found that service and sterile soils of arid zones to the deep, humus-rich, charges equal to 25 percent of farm incomes would be water-soaked soils of tropical rain forests (FAO 1982). required to reach the targeted reduction in demand. Both the delicient and the excess supply of rain However, under uniform rationing of water, production contribute to thc instability of production. The high falls only 4 percent. If only tail-end farmers are given degree of uncerlainty regarding availability leads to a reduced supplies of water, the losses increase to 7 per- less-than-optimal investment in agricultural inputs, cent. These figures suggest the importance of pricing especially in the new green revolution package based policies that allow for optimal water conservation on high-yielding and fertilizer-responsive seed vari- through crop and technology choice at the farm level. eties that have benefited farmers who have access to No discussion of water pricing mechanisms is irrigated lands. Thus farmers who crop rainfed, often complete without referring to the case of Israel, which marginal, lands generally tend to be from the eco- not only is located in one of the most arid and water- nomically poorer sections of society, and their inabil- scarce regions of the world but also has one of the ity to invest in modem agricultural technologies highest water use efficiencies in irrigated agriculture. serves to widen the gap between rich and poor in the Between 1951 and 1985, total output from irrigated rural areas of developing countries. land grew tenfold, with only a fivefold increase in irri- In certain regions, such as South and Southeast gated acreage. However, aggregate water use Asia, rainfall is highly variable. About 80 percent of increased only threefold, and water use per hectare rain falls in a period of four to six months, and the rest actually dropped about 40 percent (Bhatia, Cestti, and of the year is mDstly dry. Thus cropping periods have Winpenny 1995). Crop production per cubic meter of to be adjusted to make optimal use of the rainy sea- water almost tripled during this period. The Water son, but if the rain is delayed, or there is a prolonged Commission of the Ministry of Agriculture has imple- dry spell, the cropping schedule is often severely dis- 278 The Economic and Environmental Impacts of Irrigation and Drainage in Developing Countries turbed. Empirical data suggest a close relation ging and salinity are irrigation without drainage, between rainfall and crop productivity (Saleh and excessive irrigation, and large leakages from the canal Bhuiyan 1995). system (such as a delivery efficiency of 35-40 percent Rainfed agriculture is affected by three major from canal head to root zone). Many of these practices types of problems. These can be classified as soil- are intricately linked to water pricing and manage- related (undulating and sloping lands, poor fertility), ment policies. For example, pricing irrigation water socioeconomic (fragmented holdings, poor economic too low leads to excessive irrigation as well as lower condition of farmers), and climatic (water deficit). A incentives to adopt modern irrigation technologies number of relevant technologies can be targeted (Caswell and Zilberman 1986), resulting in low water toward rainfed environments. According to Katyal efficiencies and relatively large water losses. and Venkateswarlu (1993), these include: The problem is particularly severe in Pakistan * Land treatments that promote maximum preserva- where semiarid climatic conditions encourage the tion of rainwater where it falls, that is, in situ accumulation of salts in the root zone. In addition, moisture preservation. For example, off-season irrigation water supplies add 1.2 metric tons of salt tillage increases the use of rainwater. per hectare to the root zone every year. Unregulated * Collection of water during high-intensity rains and pumping of groundwater further aggravates the situa- recycling it during periods of water scarcity, which tion by mobilizing salts dissolved in the groundwater could lead to stable production from drylands. The aquifer. All this has contributed to more than a third of runoff can be collected in ponds and used for life- the gross command area being waterlogged and 14 saving irrigation of the rainy-season crop or irriga- percent of it being saline. The twin problems of water- tion before sowing or after the season. logging and salinity are especially severe in Sindh * Efficient use ojfharvested water. Experiments have Province, where the water table is less than 8 feet been conducted to determine the optimal timing of deep in more than a third of the cropped area and supplemental irrigation for the rainfed crop. For more than half the area is moderately saline. High example, moisture stress during the flowering of groundwater tables inhibit root growth and therefore rainy-season crops causes considerable damage. reduce crop yields. Salinity retards plant growth, and Supplemental irrigation at this stage may be high- its impact in reducing Pakistan's agricultural output is ly beneficial. estimated to be on the order of nearly 25 percent. These technological fixes must be undertaken in Donor agencies and the Pakistan government relation to other economic and institutional reforms agree that the lack of an effective drainage system is and must be done for the catchment as a whole. Thus, the principal threat to the sustainability of agriculture land and water conservation measures need to be in the Indus basin. The steps to be taken require that adopted along with crop diversification, use of rele- the drainable surplus must be removed from the vant agroecological information, and access to credit source but also that internal drainage must be suffi- and inputs by rainfed fanners. Here too, farmer par- cient and the Indus Basin Irrigation System must be ticipation and agency commitment are critical. For provided a drainage outlet to the Arabian Sea. In the example, there may be asymmetry in the distribution past, half-hearted attempts to program drainage failed of benefits because farmers owning land in the catch- to address some of the key institutional and economic ment area may need to be compensated by farmers in factors affecting water use and drainage practices. the command area who are the direct beneficiaries of Important components were neglected, such as o&M, water conservation. This, in turn, requires substantial public participation, drainage research, and the appli- cooperation and negotiation between upstream and cation of research results to the project. Recently, the downstream farmers (Pandey 1991). Pakistan government agreed to initiate a comprehen- sive approach to river basin management that will Enviromnental Impacts of Irrigation and Drainage include the reform of water markets and individual property rights, the restructuring of government agen- There are severe environmental impacts of irrigation cies, and the formation of water users associations. projects, including waterlogging and salinity, and This integrated approach to drainage management their resulting adverse impacts on health include the is expected to contribute not only to better drainage incidence of vector-bome diseases such as schistoso- but also to a wide variety of direct and indirect bene- miasis and malaria. The principal causes of waterlog- fits. In particular, it is expected to benefit tenants, Ujjayant Chakravorty 279 smallholders, and tail-enders by improving the effi- ed to induce water users to use water efficiently. ciency of service delivery, improving the availability The World Bank recently undertook a review of 64 of irrigation water and drainage, and lowering costs. active irrigation and drainage projects in 28 countries The improvements in drainage structures and o&M (World Bank 1996). The review highlights the main will lead to less salinity and waterlogging and higher issues and problems with the irrigation portfolio and output of crops. Preliminary estimates suggest that suggests how te improve its qttality. The Bank report farmers can expect to increase their incomes as much identifies several key problems in its irrigation port- as 30 percent as a result of the anticipated drainage folio: improvements, although the exact magnitude will * Excessive optimism regarding the commitment of depend on a wide variety of factors, including macro- the host government during project preparation economic factors, the initial level of salinity and and appraisal waterlogging, and agroecological conditions. Thus * Miscalculation at the appraisal stage of what may farmers will benefit not only because the steady be achieved in terms of hydraulic efficiencies and decline in yields will stop but also because output will cropping intensities, which reduces the projects' increase gradually and the associated water manage- cash returns and lowers thei:r economic outcome as ment measures will increase the availability of water, well as their economic rate of return especially for tail-enders. * Performance of economic analysis of projects In many regions, the diversion of water for irriga- based on a single probability of water availability tion has resulted in the reduction of available water through rain-fall or other sources. supplies to sensitive ecosystems such as wetlands that The report concludes that improving irrigation support vital plant and animal populations. In the Aral lending depends on several factors, some of which are Sea region, the diversion of water from lakes to sup- external to the projects being implemented. For exam- port irrigated rice and cotton cropping in remote ple, the slump in commodity prices in the 1980s led to desert areas of Kazakhistan and Uzbekistan has a drop in rice prices and to lower returns for the entire resulted in a drop in water level in the lake from 53 to portfolio of rice-based irrigation projects. Several 39 meters above sea level and a tripling of salinity measures are now being considered for improving the levels as a result of the intrusion of salt water from the rate of project success in meeting targeted develop- Aral Sea. This has decimated the fish population in ment goals: the lake and caused severe climatic and other changes * Elimination of the strict targets on professional in the region's microenvironment (Dinar 1997). input, overaLl cost, and speed in project prepara- tion, which are causing a decline in the quality of Success Stories and Lessons for Policy Reform projects at the entry level. The two years typically given for a project to receive funding from the Several important lessons can be gleaned from the Bank are too short to complete a series of tasks cumulative experience with irrigation over the past including land acquisition, field investigations, four decades. Some critical ones are as follows. It is development of stakeholder participation, and universally acknowledged that the present system of detailed engineering design of irrigation projects. water pricing (or the lack of it) cannot continue for- This has led to poor project quality and a general ever and that some form of water pricing reform must preference for projects that require less time and be implemented. It is also clear that the straightfor- budget to prepare. ward introduction of water markets will not work. The * Better understanding of the technical, economic, introduction of market mechanisms must be preceded social, and organizational aspects of irrigation by the creation of conditions that ensure the proper projects. Arnong other things, for example, the functioning of markets. For example, investments Bank has called for improved understanding of the must be made in infrastructure such as conveyance interaction between project design and water man- facilities capable of transporting water between loca- agement, sc that rehabilitation and maintenance tions and in control structures for volumetric mea- improvements lead directly to an improvement in surement of water flows. Legal impediments to water service. It also has called for designs that not only trading must be removed, although they do not seem provide improved service but also are simpler to to hamper trading in water when other conditions are operate. favorable. Transferable water rights need to be creat- A few irrigation projects have shown exemplary 280 The Economic and Environmental Impacts of Irrigation and Drainage in Developing Countries performance mainly because of strong government beginning to be motivated and willing to undertake commitment, user participation in the management of major institutional reforms as a precondition to initi- structures, and, in many cases, institutional reform ating new multilaterally funded irrigation projects. and commitment of the irrigation authority to project However, major impediments remain. In many implementation. For example, modernization of the countries, the physical and institutional structures Office du Niger, one of the oldest and largest schemes must be improved before sound economic and man- in Sub-Saharan Africa, resulted in the tripling of aver- agement policies can be implemented. User participa- age paddy yields to about 5 tons per hectare, which tion in operation and maintenance is now an essential compares well with yields obtained under the green component of any project, but the factors essential to revolution in Asia. Similarly, the Mexican water forming effective water users associations are still authority the, Comisi6n Nacional de Aguas, complet- being understood. The cultural and historical setting ed two successful projects that involved decentraliza- also plays a key role in determining beneficiary par- tion of its o&M activities in irrigation districts through ticipation. the promotion of water users organizations. This Another important insight obtained from irrigation resulted in an increase in the cost recovery component projects is the importance of adopting an integrated of the o&M budget from a low 18 percent to about 80 approach to project rehabilitation. For example, if a percent nationwide. In Colombia, legal reforms gave project aims to improve drainage in a particular irri- water users associations a key role in the management gation project, it needs to consider existing water of irrigation districts and required full recovery of institutions, prevailing prices of water, the issue of o&M costs and pairtial recovery of investment costs. surface and groundwater allocation, and the adoption As a result of changes created by the new law, o&M of technology by farmers. Simply creating better recovery targets were met in three of the four districts drainage structures may not work in the long run. The financed by the irrigation project. same is true in rainfed agriculture. A watershed-based approach that considers the impacts of management Concluding Remarks policies on both upstream and downstream beneficia- ries has a better chance of working than policies that Several key lessons emerge from this discussion. do not consider linkages across sectors or beneficiary From all accounts, the task of improving irrigation groups. management has become more complex. This is partly This new "holistic" thinking is reflected in the cur- because as populations have increased, so have rent emphasis on "basinwide" management promoted demands for urban water use and for better environmen- by agencies such as the International Irrigation tal quality. The low marginal product of water in agri- Management Institute (Seckler 1996). They argue that culture and the relatively high value of water in urban water efficiency must be measured not within a proj- and environmental uses make it likely that more and ect but over an entire basin so as to include the reuse more water will be siphoned away from agriculture in of drainage water from seepage and percolation. This the future. This again implies that less water will be suggests that policies such as the adoption of sprinkler available to meet the agricultural needs of an expand- irrigation within an irrigation project may improve the ing population. This situation is beginning to induce a efficiency of water use within the project but also structural change in the nature of irrigated agriculture. reduce the availability of water elsewhere in the basin. In spite of the presence of rent-seeking interests, Similarly, salt and chemical pollution of water restrictive property rights regimes, and a general through application of inputs may adversely affect belief that water, [like air, should not be bought and agriculture downstream of the project. These issues sold, fundamental changes are already beginning to can only be handled if the basin is adopted as the rel- happen. evant unit of analysis and the relevant externalities are Many governments, policymakers, and environ- internalized. These issues become more complicated mental groups, such as in California, are beginning to if the basin spills across political boundaries, as in the realize that a more efficient irrigation system means case of the Imperial Irrigation District (IID) in more water for everyone. This has induced changes in Califomnia. A study by the California Department of the way water rights have been allocated, and, in Water Resources calculated that the IID could save many cases, restrictions on trading in water have been nearly half a million acre-feet of water through canal removed. In developing countries, governments are lining and other measures, which could be used to Ujjayant Chakravorty 281 augment municipal water supplies. However, these Policy." Paper presented at the workshop on sus- "savings" are also claimed by Mexico, which argues tainable management of water in agriculture: that its aquifers receive recharge through seepage issues and policies, Organisation for Economic from IID canals (Cummings and Nercissiantz 1992). Co-operation and Development, Athens, November. Rural Development Department, World Bank, References Washington, D.C. Processed. FAO (Food an(I Agriculture Organization of the United The word "processed" describes informally repro- Nations). [L982. Rainfed Farming Systems in Asia duced works that may not be commonly available and the Pacific: Cropping Patterns in Agro- through libraries. Ecological Zones of Southeast Asia. Bangkok. Bhatia, Ramnesh, Rita Cestti, and James Winpenny. IIMI (International Irrigation Management Institute). 1995. Water Conservation and Reallocation: Best- 1992. "Irnproving the Performance of Irrigated Practice Cases in Proving Economic Efficiency Agriculture: IIMi's Strategy for the 1990s." and Environmental Quality. A World Colombo. Processed. Bank-Overseas Development Institute Joint Jones, W. I. 1994. "A Review of World Bank Study. Washington, D.C.: World Bank. Irrigation Experience." In Guy Le Moigne, K. Biswas, Asit K. 1991. "Inaugural Address." Presented William Easter, Walter J. Ochs, and Sandra before the seventh world congress on water Giltner, eds., Water Policy and Water Markets: resources, International Water Resources Selected Papers and Proceedings from the World Association, Rabat, Morocco, May. Processed. Bank's Ninth Annual Irrigation and Drainage Caswell, Margriet, and David Zilberman. 1986. "The Seminar, Ainnapolis, Maryland, December 8-10, Effects of Well Depth and Land Quality on the 1992. Technical Paper 249. Washington, D.C.: Choice of Irrigation Technology." American World Bank. Journal ofAgricultural Economics 68:798-811. Katyal, J. C., and B. Venkateswarlu. 1993. Chakravorty, Ujjayant, Eithan Hochman, and David "Management of Water Resources in Rainfed Zilbennan. 1995. "A Spatial Model of Optimal Areas of India: The Critical Issues." icID Bulletin Water Conveyance." Journal of Environmental 42(2):1-9. Economics and Management 29:25-41. Lenton, Roberto. 1994. "Research and Development Chakravorty, Ujjayant, and James Roumasset. 1994. for Sustainable Irrigation Management." Water "Incorporating Econornic Analysis in Irrigation Resources Development 10(4):417-24. and Management." Journal of Water Resources Pandey, Sushil. 1991. "The Economics of Water Planning and Management 120(6):819-35. Harvesting and Supplementary Irrigation in the Chakravorty, Ujjayant, and David Zilberman. 1997. Semi-Arid Tropics of India." Agricultural Systems "California's Water Experience: Lessons of the 36:207-20. Drought." In Down to Earth. New Delhi: Center Perry, Chris. 1996. Alternative Approaches to Cost for Science and Environment. Sharing for Water Service to Agriculture in Egypt. Cummings, Ronald, and Vahram Nercissiantz. 1992. Research Report 2. Colombo: International "The Use of Water Pricing as a Means for Irrigation Management Institute. Enhancing Water Use Efficiency in Irrigation: Radosevich, G. 1986. "Legal and Institutional Case Studies in Mexico and the United States." Aspects of Irrigation Water Management." In K. Natural Resources Journal 32:731-55. C. Nobe and R. K. Sampath, eds., Irrigation Cunningham, John. 1992. "Public and Private Management in Developing Countries: Current Groundwater Development in India: The Case of Issues and Approaches. Boulder, Colo.: Westview Uttar Pradesh." In Guy Le Moigne, Shawki Press. Barghouti, and Lisa Garbus, eds., Developing and Reidinger, Richard. 1974. "Instituting Rationing of Improving Irrigation and Drainage Systems: Canal Water in Northern India: Conflict between Selected Papers from World Bank Seminars. Traditional Patterns and Modern Needs." Technical Paper 178. Washington, D.C.: World Economic Development and Cultural Change Bank. 23(1):79-104. Dinar, Ariel. 1997. "Irrigated Agriculture and the . 1994. "Observations on Water Markets for Environment: Problems and Issues in Water Irrigation Systems." In Guy Le Moigne, K. 282 The Economic and Environmental Impacts of Irrigation and Drainage in Developing Countries William Easter, Walter J. Ochs, and Sandra Small, Leslie. 1987. Irrigation Service Fees in Asia. Giltner, eds., Water Policy and Water AMiarkets: ODI/IIMI Irrigation Management Network Paper Selected Papers and Proceedings from the World 87/1C. London: Overseas Development Institute. Bank's Ninth Annual Irrigation and Drainage Tahal Consulting Engineers, Ltd. 1991. "Israel Water Seminar, Annapolis, Maryland, December 8-10, Sector Review." Paper presented at the intema- 1992. Technical Paper 249. Washington, D.C.: tional workshop on comprehensive water World Bank. resources management policy, World Bank, Repetto, Robert. [l986. Skimming the Water: Rent- Washington, D.C., June. Processed. Seeking and the Performance of Public Irrigation United Nations Population Fund. 1993. State of the Systems. Research Report 4. Washington, D.C.: World's Population. New York. World Resources Institute. Wade, Robert. 1984. "Ilrigation Reform in Conditions Saleh, A. F. M., and S. I. Bhuiyan. 1995. "Crop and of Populist Anarchy: An Indian Case." Journal of Rain Water Management Strategies for Increasing Development Economics 14:285-303. Productivity of Rainfed Lowland Rice Systems." World Bank. 1996. "Portfolio Improvement Program: Agricultural Systems 49:259-76. Irrigation Subsector." Agriculture and Natural Sampath, Rajan. 1992. "Issues in Irrigation Pricing in Resources Department, Washington, D.C. Developing Countries." World Development Processed. 20(7):967-77. Young, Robert, and Robert Haveman. 1985. Seckler, David. 1996. The New Era of Water "Economics of Water Resources: A Survey." In Resources Management: From "Dry" to "Wet" Allen Kneese and James Sweeny, eds., Handbook Water Savings, Research Report 1. Colombo: of Natural Resource and Energy Economics. Vol. International Irrigation Management Institute. 2, pp. 465-529. Amsterdam: North-Holland. 21 Livestock and the Environmient: Issues and Options Henning Steinfeld, Cornelis de Haan, and Harvey Blctckburn Livestock interact with the environment in many The Dynamics of Livestock Production Systems ways. In the first section, this chapter provides the overall context and discusses the main issues. In the For a large part, livestock interact with the environ- second section, it suggests policy and project inter- ment within the confines of a production system. ventions that can address those issues, including the These production systems are evolutionary responses tradeoffs between productivity and the environment. to population pressure, resource endowment, and marketing opportunities. Sere and Steinfeld (1996) The Context and Issues distinguish three main production systems: grazing, mixed-farming,, and industrial systems. Livestock can damage global natural resources in a Grazing systems are mainly based on native grass- number of ways, but there are also many examples of land, with no or only limited integration with crops. environmental balance and positive contributions. These systems often do not involve imported inputs Livestock interact on a global scale with land, water, and generally have a low caloric output per hectare air, and plant and animal biodiversity. About 34 mil- (Jahnke 1982). In arid grazing systems, the potential lion square kilometers, or 26 percent of the world's for increasing productivity is extremely limited, and land area, are used for grazing livestock. In addition, livestock resources are best managed by allowing for 3 million kilometers, or about 21 percent of the variability-or "at disequilibrium" (Behnke, Scoones, world's arable area, are used for cereal production for and Kerven 1993)-and by adapting animal pressure livestock feed. Livestock produce 13 billion tons of continuously to the highly variable rainfall and avail- waste a year. A large part of this is recycled, but ability of feed. The number of stock is allowed to where animal concentrations are high, waste poses an fluctuate in response to unfavorable "bust" or drought enormous environmental hazard. Livestock and live- conditions and favorable "boom" conditions. Where stock waste cause gaseous emissions that have infrastructure and water development or crop important local and global impact on the environ- encroachment interfere with resources and patterns of ment. Livestock grazing can affect the water balance movement, ecologically sound systems are often dis- in certain areas. Water is needed to produce fodder rupted, overgrazing and land degradation occurs, and and feed concentrate, to provide drinking water for the livelihoocl of pastoral communities is invariably animals, and to drain surplus waste and chemicals. threatened. In semiarid zones, there is some scope for Livestock interact directly and indirectly with biodi- intensification and transition into mixed-farming sys- versity; although biodiversity often is compromised, tems, because these areas have some potential for there are also examples of mutual benefits. cropping. Parts of southern Asia have developed 283 284 Livestock and the Environment: Issues and Options along this path, and the sudano belt in Africa is fol- the importance of roughage as feed is decreasing at lowing the same course. Some of the humid and sub- the expense of cereals and agroindustrial by-products. humid zones have been precluded from livestock There is an important species shift toward monogas- development because of problems of disease (African tric animals, mainly poultry and pigs. Ruminant meat sleeping sickness, which is transmitted by tsetse flies accounted for 54 percent of total meat production in in Africa) or accessibility. With increasing influx of the developing countries in 1970, decreased to 38 per- population and better methods of controlling disease, cent in 1990, and is projected to decrease further to 29 the threat of African sleeping sickness will become percent in 2010 (Alexandratos 1995). This species manageable, and these as well as other grazing sys- shift reflects the use of, and better conversion rates tems (subhumid savannas of Brazil, for example) for, concentrate feeds by monogastric animals. have excellent potential and may develop into mixed- Livestock production is becoming separated from farming systems. its land base, is urbanized, and is beginning to assume Historically, the integration of livestock and crop features of industrial production. In recent years, activities into mixed-fanning systems represents the industrial livestock production globally grew at twice main avenue for intensification. As by-products (crop the rate (4.3 percent) of production in mixed-farming residues, manure) of one enterprise serve as inputs systems (2.2 percent) and more than six times the rate into the other, this system is, in principle, environ- in grazing systems (0.7 percent; Sere and Steinfeld mentally friendly. In many places of the world, 1996). This trend continues to accelerate. In agroeco- closed-cycle smallholder farms have developed and logical terms, livestock production is growing more continue to be a predominant feature. rapidly in humid and subhumid zones than in arid When land scarcity emerges as a pressure point for tropical zones or highlands, essentially following pat- farming systems, farmers respond by specializing in terns of human population growth. crop production, by following a path of intensifying the crop-livestock interaction (Christiaensen, Tollens, Grazing and Overgrazing and Ezedinma 1995), or by abandoning land-based production and moving into industrial livestock pro- About 34 million kilometers, or 26 percent of the duction where market demand is strong. world's land area, are used for grazing livestock. Industrial production systems are detached from Grazing animals can improve soil cover by dispersing immediate land in terms of feed supply and waste dis- seeds with their hoofs and through their manure, posal. Where the demand for animal products increas- while controlling shrub growth, breaking up soil es rapidly, land-based systems fail to respond, leading crusts, and removing biomass that otherwise might to animal concentrations that are out of balance with provide fuel for uncontrolled bush fires. These the capacity of the land to absorb waste and supply impacts stimulate grass tillering, improve seed germi- feed. High concentrations of animals close to human nation, and thus improve land and vegetation. agglomerations bear enormous pollution problems However, heavy grazing causes soil compaction and and associated human health risks. Because of water erosion and decreases soil fertility, organic matter pollution and health hazards, most of these systems content, and water infiltration and storage. are moved out of city boundaries as soon as infra- Overgrazing in hilly environments can accelerate ero- structure permits. The decision of Singapore to aban- sion. don livestock production altogether is a case in point. The arid rangelands have generally been related to The opportunities that arise from strong market widespread degradation. Dregne, Kassas, and demand for meat and milk conflict with the limited Rozanov (1992) and World Resources Institute (1992) potential to expand the conventional resource base on argue that the majority of the world's rangelands are which livestock traditionally relied: that is, grazing moderately or severely desertified. The United land and feed resources of low or no alternative use. Nations Environment Programme (UNEP; see Increasingly the world livestock sector resorts to Oldeman, Hakkeling, and Sombroek 1990) estimates extemal inputs, notably high-quality feed but also that, since 1945, about 680 million hectares, or 20 more productive breeds, better animal health, and percent of the world's grazing lands, show significant general husbandry inputs. Grazing systems offer only soil degradation. limited potential for intensification, and livestock pro- However, if irreversibility and declining produc- duction is becoming increasingly crop-based. Thus tivity are taken as the main characteristics of degrada- Henning Steinifeld, Cornelis de Haan, and Harvey Blackburn 285 tion (Nelson 1990), then the actual situation of the be temporary. The continuous "disequilibrium" pro- world's arid grasslands and rangelands is better than tects land, especially in arid areas with annual vegeta- generally reported. For example, in the Sahel, satellite tion, because the low grazing pressure after a drought imageries (Tucker, Dregne, and Newcomb 1991) facilitates recuperation, and, although with a time lag, show vegetation at the same northern limits as before overall grazing pressure is adjusted to the amount of the big droughts of the 1970s and 1980s. They point feed available (the theoretical bases for range man- to a fluctuating pattern, or a "contracting and expand- agement under those conditions have recently been ing Sahara," rather than a continuously expanding well described by Behnke, Scoones, and Kerven 1993 desert. This resilience is confirmed by our own analy- and Scoones 1994). Flexibility and mobility are there- sis showing that animal production per head and per fore key requirements to achieve sustainable use of hectare has increased, rather than declined, over the rangeland in these areas. past three decades (see box 21.1). This has occurred in Where this mobility is impaired and customary spite of a large increase in the number of livestock and practice impe(led, degradation often occurs. The a decrease in the area of rangelands. nationalization of arid rangelands that was introduced In the light of this evidence, arid rangelands with by many governments in the postcolonial period in less than a 90-day growing season are now seen as Africa and Asia undermined the intricate fabric of containing dynamic and highly resilient ecosystems, customary practice by replacing an ecologically well- especially under traditional management of continu- balanced system of communal land use with a free-for-all ous adjustments to highly variable rainfall, both in open-access system (see box 21.2). Policies to "settle" time and in space. Arid vegetation is extremely the pastoralists, to promote ranches, and to regulate resilient, and most of the changes observed are the stocking rates, which were major principles of arid result of severe dry spells and are therefore likely to rangeland development of the 1960s and 1970s, reduced the critical mobility and flexibility in that system. In the Middle East and Central Asia, state farms are being privatized and cut up, thereby imped- gBox 21.1. Productmvty Trends in the SahLelinpstrlt'molty ing pastoralists' mobility. An analysis of livestock production in five Sahelian countries over a 30-year period, carried out as part of water pointsf The devopmen of ' . . ~~~water points fcor human and livestock use often opens this study, shows a 93 percent increase in the meat prosucdy prhecwar an a percent increase in the. up arid lands and contributes to land degradation (see produced per hectare and a 47 percent increase in the box 21.3). Around the water points, animal trampling meat produced per head. At the same time, there was a 22 percent increase in the animal population (from acauss are surfaces, end degad land typically, .. . . ~~~~~~~accounts for 5 to 10 percent of total area. Normally, 14.5 to 17.6 iillion total livestock units). soil fertility is quite high, but the impact of hoofs, This increase in productivity occurs in both cattle including soil compaction, and heavy grazing impede and small ruminants. Part of the increased productiv- ity might result from the progressive move of the live- noral aterS ofhvegetan. Long-term di stoc poulaton o moe hmid rea andthe Senegal and Sudan (Thomas and Middelton 1994) do stock population to more humid areas and the not find a significant expansion of the sacrifice areas increased use of crop residues. The long-term trend of individual water points. More important, water points to sustained increases in productivity and a sta- development in ard areas might upset entire ecosys- ble resource base. dvlpetlLaL ra Lh pe nr css tems by changing the relationship between traditional 30 wet- and dry-season grazing areas and converting tra- 25 _ Kg meat/TLU ditional dry-season grazing into year-around grazing. 20 _ In the semiarid zones, those with more than a 90- 15 day growing season, land degradation caused by graz- 10 ing livestock is much more serious than at the fringes 5 of the desert. Data from a transect in Mali show that _ _ _ v _ _ Kg meatUHa. land degradation in areas with 600-800 millimeters of t . 0. cJ 4 't . , rainfall was significantly greater than in areas with c° a' a' a' a' m' ot350-450 millimeters. In areas with higher rainfall, the percentage of barren soil increased from 0 to 10 per- 286 Livestock and the Environment: Issues and Options Box 21.2. The Deterioration of Common Property Box 21.3. Positive Effects of Intensification: Resources in India The Machakos Case In a study of 75 villages in seven rather arid states of Human pressure and intensification can also work India, Jodha (1992) finds that the common property positively. English, Tiffen, and Mortimore (1992) resource area declined by as much as 30 to 50 percent show that in the semiarid Machakos district in Kenya between the 1950s and 1982 and that cattle declined the natural resource base improved despite a 500 per- 20 percent in favo:r of small ruminants. The tradition- cent growth in population during the past 60 years, an al communal management of the common property increase in cropland from 35 percent of the higher- resource had broken down: potential area in 1948 to 81 percent in 1992, and an * In 1950, 70 villages had formal or informal rules increase in the livestock population per farm from 5.4 for managing the common property resource; in animal units in 1940 to 16.2 in 1980. Dynamic mar- 1982 only eight villages maintained those rules. ket development made farming profitable, and off- * In 1950, 55 villages levied formal or informal farm employment generated enough capital for taxes to maintain the common property resource; investments in soil and water conservation. none did so in 1982. Horticulture and smallholder dairy production are the * In 1950, 65 villages had formal or informal obliga- main activities generating the cash for resource con- tions to maintain their common property resource; servation. The famine predicted in the 1930s for the in 1982 only 13 had them. Machakos district never occurred. As traditional common property resource manage- ment shifted to open access, common property y in North Africa and the Middle East, ld resources degraded significantly, as shown by the 75 especiayp ea- pretdecline in the number of trees in the common ing to continuous growth of the livestock population, perct decline in the number of ntcwhich is no longer constrained by the poor productiv- property resource and the decline in the number of itwfpsuei o-analrgos grazing days (Jodha 1992). ~~ity of pasture in low-rainfall regions. grazing days (Jodha 1992). Water Data on the effect of grazing on water infil- cent over the period 1950-90, whereas in arid areas, tration are site-specific. However, light-to-moderate there was no significant change (Mainguet and others grazing can improve or maintain infiltration, whereas 1992). A similar situation is reported in parts of North heavy grazing almost always reduces it. Long-term Africa and the Middle East (Sidahmed 1996), observations on the Edwards plateau in Texas have although quantitative data are not available. shown that both heavy grazing and no grazing cause lower infiltration and higher erosion rates than mod- Land. This degradation is linked to growing popu- erate grazing (Grazingland Management Systems lation. Crop encroachment, fuelwood collection, and 1996). Water pollution by agrochemicals can be overgrazing are the interlocking factors causing land caused by the control of livestock disease vectors, degradation in semiarid zones. Crop encroachment such as ticks (the carrier of many diseases, such as not only exposes the soil directly to the erosive effects anaplasmosis and east coast fever) and tsetse flies (the of winds and downpours but also progressively ham- carrier of African sleeping sickness in man and ani- pers the movement of animals because it obliterates mals), and of weeds. Ticks have been controlled tra- the passage between wet- and dry-season grazing ditionally by cattle dips or sprays using organochlo- areas. Fuel and fertilizer subsidies, especially in the rines. Several donor-supported projects promoted this Middle East and North Africa, often exacerbate the technology in the 1970s and early 1980s. Fauna has conversion into marginal cropland of sites with higher been damaged both by inappropriate dosing, leading potential within arid rangelands. Drought emergency to increased resistance of ticks, as well as by inappro- programs, which hand out subsidized concentrate priate drainage, causing the liquid in dips to be dis- feed, probably also contribute to range degradation, charged frequently in open water after use. allowing too many animals to be maintained on the range and preventing an ecologically normal regener- Plant and animal biodiversity. Arid vegetation is ation of range vegetation after the drought. These feed extremely resilient. Most of the changes observed are subsidies have now become a structural phenomenon, a result of dry spells and therefore are most likely Henning Steinfeld, Cornelis de Haan, and Harvey Blackburn 287 temporary. The decrease in woody and perennial and contain approximately 50 percent of the world's species observed in recent decades in the Sahel biodiversity. Since 1950, more than 200 million (Boudet and others 1987) is most likely the result of hectares of tropical rain forests have been lost, as a droughts. Resting an area brings the original flora result of ranching, crop cultivation, and forest back again (Hiernaux 1996), which indicates that the exploitation. Ranching-induced deforestation is one loss of these species is not irreversible. In effect, an of the main causes of loss of some unique plant and extensive review of grazing and production data of animal species in the tropical rain forests of South and 236 sites worldwide, including many sites in the arid Central America, the world's richest source of biodi- zone, shows that there is no difference in biomass pro- versity. Since 1950, the area in Central America under duction, species composition, and root development pasture has increased from 3.5 million to 9.5 million in response to long-term grazing in the field hectares, and cattle populations have more than dou- (Milchunas and Lauenrot 1993). Grazing by animals bled from 4.2 million to 9.6 million head (Kaimowitz can even improve plant biodiversity. 1995). In Asia and Africa, deforestation is the result Complementarity can be observed between mainly of crop expansion. wildlife and livestock. There is increasing evidence Land speculation, titling procedures, and govern- that, in low-rainfall areas, the combination of wildlife ment-provided financial incentives have been the and livestock can result in greater biodiversity and main reasons for ranch-induced deforestation higher income for pastoralists and ranchers. Usually, (Kaimowitz 1995). Land speculation is heavily influ- livestock-wildlife combinations do not require signif- enced by road construction because land prices are icant reductions in livestock stocking rates. For exam- closely correlated with the distance to an all-weather ple, a reduction of only 20 percent of the cattle stock- road. In many countries, land titling procedures still ing rate would create the "niche" for most wildlife require that land be under agricultural use before a species to prosper (D. Western, 1996, personal com- title can be given. Such procedures, of course, encour- munication). Byrne, Staubo, and Grootenhuis (1996) age deforestation. Incentive policies, in the late 1960s estimate a larger dietary overlap, and thus the need for and 1970s, which provided subsidized interest rates a greater reduction, but agree that there is substantial with lenient reimbursement conditions, and beef complementarity. There is also fairly general agree- export subsidies played an important role in the ment that wildlife ranching on the basis of meat alone expansion of ranching. They have been phased out, is not financially viable. Game meat is now attractive and this has reduced investment in large ranches by because it occupies a niche market. Combining the two absentee owners. This is reflected in a decline in activities, with adequate sharing of the benefits of com- deforested area. In Central America in the 1980s, rain mon goods, is therefore considered the best strategy. forest disappeared at an annual rate of 430,000 In humid areas, clearing forest and savanna to hectares a year. This decline slowed to 320,000 establish pastures causes soil nutrients to leach out hectares annually over the period 1990-94, although rapidly. Weeds displace grasses, and artificial pastures this should be seen in context: by 1990, only about 19 can only be sustained for a period of up to 10 years. million hectares of rain forest remained. More than 50 percent of the pasture areas in the Amazon area have now been abandoned in a degrad- Crop-Livestock Interactions: Intensification ed state. Natural regeneration of forests is quite diffi- and Involution cult, especially when areas are large. Under good management and using modern technology, however, The integration of crop and livestock still represents the establishment of pasture and the introduction of the main avenue for intensification of food produc- cattle can be the second-best option for maintaining tion. Mixed farming provides farmers with an oppor- soil fertility. For example, the Centro de Investigaci6n tunity to diversify risk from single-crop or livestock en Agricultura Tropical (CIAT 1994) developed a sta- production, to use labor more efficiently, to have a ble land management system, using grass and nitro- source of cash, and to add value to low-value or sur- gen fixing, that also proved attractive to smallholders. plus feed. To varying extents, mixed-farming systems allow the waste products of one enterprise (crop by- Livestock and Deforestation products, manure) to be used as inputs to the other enterprise (as feed or fertilizer). Mixed farming is, in Tropical rain forests cover about 720 million hectares principle, beneficial for land quality because it helps 288 Livestock and the Environment: Issues and Options to maintain soil fertility. In addition, the use of rota- approximately 28 percent of the world's arable land, tions between various crops and forage legumes equivalent to 52 percent of total cropping land in replenishes soil nutrients and reduces soil erosion developing countries. In southeastern and eastern (Thomas and Barton 1995). Asia, animal draft is decreasing in importance as rapid Adding manure increases the soil's nutrient reten- mechanization takes place. In contrast, many areas in tion capacity (or cation exchange capacity), improves Africa are developing into mixed farming, with draft the physical condition by increasing the water-hold- animals playing a key role in providing tillage for ing capacity, and improves the structure. This is a cru- expanded crop areas and higher production. As a cial contribution because, in many systems, it is the renewable form of energy, animal draft substitutes for only avenue available to farmers for improving the fossil energy and the use of other natural resources. soil's organic matter. It is also substantial in econom- Past policies have sometimes limited the synergis- ic terms. Approximately 20 million tons, or 22 per- tic effect of crops and livestock in nutrient-deficient cent, of total nitrogen fertilization out of 94 million situations. Imposing high import duties to protect tons in world agriculture (FAo 1997) and 11 million domestic cereal production pushed cropping into mar- tons, or 38 percent, of phosphate are of animal origin, ginal areas and upset the equilibrium between crops representing about $1.5 billion worth of commercial and livestock. Subsidization of inorganic fertilizer fertilizer. Not only does animal manure replenish soil reduced the incentives for nutrient cycling using live- fertility, but it also helps to maintain or create a better stock. climate for soil micro-flora and fauna. In many places of the world, subsistence farms However, mixing crops and livestock neither gen- with crop, livestock, and household closely interlinked erates new nutrients (with the exception of nitrogen have developed and continue to be a predominant fea- fixation by leguminous plants) nor reduces nutrient ture. With human population pressure increasing fur- surpluses. But livestock, even with low levels of tech- ther, the need for intensification brings livestock more nology, allow for (a) the spatial and temporal alloca- and more into cropping areas and integrates them in tion of nutrients from areas with lower returns from nutrient and energy cycles (see box 21.4). Almost cropping to those with higher returns, (b) the acceler- throughout the world, the family-based and diversi- ation of the turnover of nutrients in the production fied mixed-farming system has come under pressure cycle, and (c) the reduction of nutrient losses within either from novel technologies and market forces or the cycle compared to agricultural production without from resource degradation and poverty. Two major livestock. features emerge. One scenario leads to specialization, The key issue is the nutrient balance. Most mixed- where market forces and technological requirements farmning systems of the developing world have a neg- force mixed-fanning systems to grow in unit size and ative nutrient balance. Deficits are partially covered to specialize. With specialization, there are fewer by a flow of nutrients from grazing areas to cropland. opportunities for on-farm integration of crops and As population pressure changes the ratio of cropland livestock. Another significant trend is what has been to grazing land, and when other sources are not avail- described as "involution" or a collapse of the mixed- able, fertility gaps tend to widen. Reported deficits farming system. In virtually all tropical highland range from about 15 kilograms of nitrogen per hectare areas the relatively high densities of human popula- a year in Mali to more than 100 kilograms of nitrogen tion are traditionally sustained by rather complex, per hectare a year in the highlands of Ethiopia (de mixed-farming systems. Population pressures may Wit, Westra, and Nell 1996). The result is that crop decrease farn sizes to a point where associated land yields continue to decline. Resource degradation, pressures are no longer compensated by commensu- poverty, and population pressure carry a high risk of rate increases in land productivity, resulting in disin- conflict, as the recent events in Rwanda have proven. tegration of the system. Livestock, often large rumi- However, positive trends-intensifying and diversify- nants, can no longer be maintained on the farm. This ing production-a] so occur. A key factor facilitating results in a greater deficit of nutrients and energy and such positive trends is access to markets. leads to natural resource degradation and loss of Animal draft is still a main component of many investment. There is mounting evidence that human smallholder farms in developing countries, substitut- population pressures, poverty, and resource degrada- ing for human labor and drudgery. An estimated 250 tion, aggravated by lack of access to markets and million working animals provide the draft power to employment opportunities, are cause-and-effect fac- Henning Steinfeld, Cornelis de HYaan, and Harvey Blackbum 289 Box 21.4. The Feed-Food Controversy One-third of the world's grain harvest ends up in the large ruminant populations to produce without being digestive tracts of livestock, where the conversion of fed high-quality feed. Second, the quality of different vegetable protein into animal protein is incurring loss- proteins needs to be considered. Proteins of animal ori- es of 60 to 90 percent of edible protein. This fact con- gin have a much higher digestibility and nutritive value cerns large parts of the public. than most vegetalble proteins. Among the 996 million tons of concentrates used However, the question of feed-food competition is in 1994 (FAO 1996), all cereals, roots and tubers, puls- somewhat ill-posed. Rather than looking at what es, and some feedstuffs of animal origin (milk pow- potential food goes to livestock, one should look at der) can be classified as edible. Edible feedstuffs pro- the resource requirements. For example, a highly vide livestock with 74 million tons of protein. intensive dairy production system may rely on alfalfa In the same year and by contrast, livestock produced fodder produced under irrigation and with heavy use 199 million tons of meat, 532 tons of milk, and 45 mil- of extemal inputs including fossil fuel, fertilizer, and lion tons of eggs, altogether yielding 53 million tons of pesticides. Alfalfa is not edible, but the land and other protein. Leaving aside the differences in nutritional resources could be used for food production. value, the world's livestock sector consumes more edi- Apparently the faod versus feed debate has to do with ble protein than it produces. The input-output ratio is the value that food has beyond its commodity price- 1.41 and continues to grow because livestock produc- a less tangible religious and cultural dimension. tion increasingly relies on grains, and this reliance is A recent FAO study (1996) shows that the increas- only partially offset by gains in the efficiency of feed ing use of feed grains has not had an adverse effect on conversion. However, such input-output considerations the provision of cereals for human consumption. In overlook two important aspects: first, ruminants have times of food shortages, such as in 1974-75, adjust- the capacity to produce protein without being fed pro- ments are made in the use of feed, and food con- tein because of rumen flora activity. This allows for sumption of cereals remains largely unaffected. Box 21.5. Reducing Methane Emissions from Digestive Fermentation through Strategic Supplementation in South Asian Countries Urea and other supplemental nutrients are mixed percent, while increasing milk production 35 per- with molasses to make them palatable to livestock. In cent. addition, molasses provide the energy needed to real- Efficient digestion in the rumen requires a diet that ize the improved microbial growth that can result contains essential nutrients for the fermentative from enhanced ammonia levels. These multinutrient micro-organisms. Lack of these nutrients lowers the blocks have been used in many countries including animal's productivity and raises methane emissions Bangladesh, India, Indonesia, and Pakistan (Habib per unit of product. For animals on low-quality feed, and others 1991, Hendratno, Nolan, and Leng 1991, the primary limitation to efficient digestion is the low Leng 1991, and Saadullah 1991). Typical results concentration of ammonia in the rumen. Supplying have been increases of 20-30 percent in milk yield, ammonia can therefore greatly enhance digestive effi- increases of 80-200 percent in growth rate, and ciency and the use of available feed energy. Ammonia increases in reproductive efficiency. In addition, can be supplied by urea, chicken manure, or soluble methane emissions per unit of product have declined protein that degrades in the rumen. Urea is broken as much as 40 percent. Bowman, Croucher, and down in the rumen to form ammonia, and adding urea Picard (1992) estimate that strategic supplementation to the diet is the most effective method of boosting of dairy animals would reduce methane emissions 25 rumen ammonia levels. 290 Livestock and the Environment: Issues and Options tors of the involution (Himalayan hills, African high- all feeds in the industrial countries and 12 percent in lands, Andean countries, Java). the developing world. This involution of previously well-integrated mixed The current expansion of cropland is globally very fanning is another livestock environment "hot spot." small-0.1 percent annually, which compares with Here, it is not the interaction between livestock and nat- growth in crop production of 1.9 percent a year (FAO ural resources that creates a degradation problem but 1996). This suggests that the bulk of the increased rather the socioeconomic context that leads to a dimin- demand for crops is being met by the intensification ishing interaction that eventually ceases altogether. of cropland rather than expansion. Crops differ in the degree of depletion of soil Concentrate Feed Production moisture and water resources, in their relative demands on soil nutrients, and in their typical need for Cereals are the major component of livestock concen- applications of pesticides. In general, cereal crops, trate feed: 32 percent of the world's cereal production and in particular maize, have the potential to cause is consumed by livestock, and this has become an greater environmental damage than other crops. This issue of debate (see box 21.5). This cereal is produced is due to heavy use of fertilizers and pesticides, high on roughly 21 percent of the total land devoted to demand for water, and poor ground cover, in the early cereal. All totaled, cereals, oilseeds, and roots and stages of plant development. On the contrary, poten- tubers used to feed livestock added up to an average tial impacts are generally lowest for legume crops, of 750 million tons over the 1990-92 period. An addi- such as soybeans and pulses. Environmental risks due tional 252 million tons (about 24 percent) of all con- to nitrate and phosphate losses are greatest in maize centrates were processing by-products (brans and and wheat, while the risks of soil nutrient depletion oilcakes) for which there is little alternative use (see are greatest in cassava and sweet potato. figure 21.1). Averages based on 1990-92 data indicate that Waste from Processing global cereal production was 1.8 billion tons. Of this, 600 million tons were used for livestock feed. Maize Like waste from animal production, the processing is the most important feed grain, accounting for 55 of animal products results in environmental damage percent of total feed grain used, followed by barley where it is concentrated and unregulated. This is the and wheat. Soybean is the most important oil meal case in urban and periurban environments, particu- because it supplies more than half the requirements larly in developing countries. Slaughtering requires for these high-protein feeds. large amounts of hot water and steam for steriliza- By the mid-1980s, concentrate feed accounted for tion and cleaning. Therefore, the main polluting about a quarter of all feeds for livestock; and this pro- component is wastewater (Verheijen and others portion was growing at about 0.2 percentage point 1996). annually. Concentrates comprise about 40 percent of The concentration of organic compounds in Figure 21.1. Global Production of Livestock Concentrate Feed, 1965-92 (millions of tons) 1,000 800 600 200___ _ 0 1965-67 1975-77 1985-87 1990-92 MN Cereals EBrans U Oilseeds/meals * Roots/tubers Henning Steinfeld, Cornelis de Haan, and Harvey Blackburn 291 wastewater leads to oxygen demand, usually emission from fossil fuels, the same grasslands and expressed as biological oxygen demand (BOD). forests recapture part of these emissions. In effect, Wastewater contains fat, oil, protein, carbohydrates, technologically improved savannas accumulate 50 and other biodegradable compounds. Degradation of tons (or the annual emission of about 15 cars) more these organic substances requires oxygen. In addi- emissions per hectare than the traditional unimproved tion, wastewater usually contains insoluble organic savannas (Fisher and others 1995). and inorganic particles called suspended solids. In Leaving asi(le the respiration of domesticated ani- most developing countries, tannery effluent is sew- mals, livestock can be associated with about 15 to 20 ered, discharged into inland surface waters, or irri- percent of total carbon dioxide emissions. When the gated on the land (Verheijen and others 1996). High carbon sequestration capacity of natural and improved concentrations of salt and hydrogen sulfide present grassland is considered, the contribution to the net in tannery wastewater greatly affect the quality of increase in carbon dioxide concentrations in the water. Suspended matter such as lime, hair, and atmosphere is probably much smaller. There are, fleshings make the surface water turbid and settle to however, notable differences between the systems. the bottom, thereby affecting fish. Chromium tan- Grazing systerns, with little fossil fuel inputs and ning is toxic to fish and other aquatic life. When often considerable sequestration, are likely to have a mineral tannery wastewater is applied on the land, positive carbon balance, whereas this balance is like- soil productivity is adversely affected and some part ly to turn negative as intensity and commercial inputs of the land may become completely infertile. Due to increase, such as in industrial systems. infiltration, groundwaters are also adversely affect- ed. Discharge of untreated tannery effluent into sew- Methane. M[ethane gas is much more aggressive ers causes deposition of calcium carbonate and (24 times) than carbon dioxide in causing global cli- blocks sewers. mate change. It is the product of animal production and manure mamagement, rice cultivation, production Gaseouis Emissions of Livestock and Waste and distribution of oil and gas (pipelines), coal min- ing, and landfills. Livestock production and manure Livestock and livestock waste produce gases. Some management contribute about 16 percent of total are local, such as ammonia. Others, such as carbon annual production of 550 million tons. Methane gas is dioxide (CO2), methane (CH4), ozone (03), and produced mainly as a by-product of the feed digestion nitrous oxide (N20), affect the world's atmosphere, of ruminants (box 21.5). On average, about 6 percent by causing global warming or global climate change. of the feed energy is lost in methane (U.S. Environmental Protection Agency 1995). Methane Carbon dioxide. There are three main sources of emissions are the direct result of the capacity of rumi- livestock-related carbon dioxide emissions. First, all nants to digest large amounts of fibrous grasses and domesticated animals emit carbon dioxide as part of other feeds that cannot be used for human consump- their basic metabolic function or respiration, estimat- tion. Pigs and poultry cannot digest these fibrous ed at a total of 2.8 billion metric tons annually. feeds, which is why emissions from these animals are Second, carbon dioxide emissions result from bio- relatively low. Methane emissions per unit of product mass burning, part of which can be attributed to the are highest when the quality of feed and level of pro- clearing of land and setting of bush fires to create pas- duction are love, conditions that prevail in the arid and ture and enhance pasture growth. Third, carbon diox- humid tropics and subtropics. This offers the opportu- ide is released in relation to livestock-related con- nity for a win-win situation in which improved feed- sumption of fossil fuel for heating, manufacturing ing leads to more efficient and profitable production machinery, and producing feed. and lower methane emissions. Carbon dioxide is the least aggressive of the green- Of methane emanating from animal production, 20 house gases but is emitted in large quantities. percent comes from manure stored under anaerobic Annually, savanna and deforestation fires pump conditions (U.S. Environmental Protection Agency approximately 1.6 billion tons into the atmosphere, 1995). Here the picture is reversed: high levels of between 20 and 25 percent of the total anthropogenic methane emissions from manure management, as well CO2 emission of 7.1 billion tons (Houghton, as from large production units, are usually associated Callander, and Vemey 1995). However, unlike the with high levels of productivity and intensity. 292 Livestock and the Environment: Issues and Options Despite growing livestock populations, global reducing environmental damage associated with live- methane emissions from livestock remain static for stock is to establish feedback mechanisms that ensure two reasons. First, increases in productivity lower that livestock's use is consistent with overall social emission levels per animal and per unit of product. objectives. Advances in feed resources and nutrition and breed There is ample evidence that economywide poli- development are significant contributing factors. cies, including fiscal and monetary policies or struc- Second, monogastric production is growing at a much tural adjustment programs, have a strong impact on faster pace than ruminant production. About 80 per- the natural resource base. These complicated interac- cent of the total growth of the livestock sector is tions are poorly understood, and generalizations are attributed to pigs and poultry, which emit compara- difficult to make. However, where market failures tively small amounts of methane. occur, or policy distortions and poverty cause envi- Any reduction in methane production, however, is ronmental damage, broad policy reforms that pro- likely to result in an increased emission of other mote efficiency or reduce poverty can generally be gases, notably carbon dioxide and nitrous oxide, regarded as beneficial to the environment. because fossil fuels and fertilizer will be required in The objective is not necessarily to modify the the intensification process. original broader policies that have conventional eco- nomic or social goals; this would be an unrealistic Nitrous oxide. This is the most aggressive green- proposition, particularly in developing countries. house gas (320 times C02). It is produced in animal Rather, complementary measures with specific or manure, which contributes about 400,000 tons of localized foci may need to be designed to mitigate the nitrogen a year or 7 percent of total global emissions. negative effects or enhance the positive effects of the orig- Indirectly, livestock are associated with N20 emis- inal policies on the environment (Munasinghe 1994). sions from grasslands and, through their concentrate The challenge is to design policies that correspond to feed requirements, with emissions from arable land the intended social and economic objectives but still and nitrogen-fertilizer use. comply with environmental sustainability. What Is to Be Done? Institutions Policies need to be designed to correct the environ- Institutions are required to develop environmental mental effects of livestock production, which are not policies and laws and to assist line agencies in incor- reflected in product and input prices. These policies porating environmental concerns into economic should address the causes of environmental degrada- development planning and budgeting. There must also tion and must be flexible, site-specific, and well tar- be institutional capacity to monitor compliance and geted. Instruments to enhance positive and mitigate enforce regulations. Institutions are essential to man- negative environmental impacts include pricing, reg- aging access to common property resources, such as ulations, and institutional development. The collec- grazing land or forest, or ensuring security of tenure. tive purpose of these instruments is to establish feed- For most traditional grazing systems under pres- back mechanisms thiat ensure that livestock's use is sure by changing property rights, institutions need to consistent with overall social objectives. be improved by decentralizing decisionmaking. Within an enabling policy framework, a wide Local empowerment is required in all situations, but range of available technologies can be employed, especially where governmental institutions are weak. while others still need to be developed. Their purpose The need to transfer authority and responsibility for is not simply to reduce the environmental damage by resource management to the lowest level at which it reducing the polluting load. They should also be can be exercised effectively is increasingly recog- designed to enhance and spare natural resources nized. Often, consultation and direct participation of through the use of livestock and to turn waste into the community in decisionmaking enable local useful products. knowledge to be harnessed and responsibility for identifying problems and finding solutions to be Policy Design taken at the local level. Institutions need to establish stewardship for environmental resources and to regu- The collective role of all policy instruments aimed at late access. This, in particular, refers to extensive sys- Henning Steinfeld, Cornelis de Haan, and Harvey Blackburn 293 tems where a large part of the environmental damage can be traced to a regulatory vacuum left by eroding traditional institutions. . The perception that pastoralists maintain unproduc- Community-based wildlife management is tive animals in their herds for prestige rather than eco- becoming generally accepted as an essential compo- no.ic reasons is still widespread. This, in the eyes of nent of sustainable wildlife management, particularly . . o ' ~~~many, iS one of the main reasons for overstocking and livestock-wildlife interactions. There has been con- land degradation. However, almost all studies on pas- siderable progress in decentralizing wildlife manage- land agropast all sththere as ment, especially in East Africa, and in sharing the toral and agropastoral systems show that there are benefits from wildlife with the involved population veryL few unproducive amals in traditional herds (Ottichilo 1996). Key issues involve maintaining the (ILCA 1994). Aimmals are sold when they have their strength~~~~~~~ an coeso oftetaiinaoilfbi opttmum market weight. Unproductive animals are strength and cohesion of the traditional social fabrc sometimes found in "investment herds" owned by and achieving an equitable distribution of revenues traders or civil servants who, in the absence of reli- from wildlife among the local population. able and remunerative banking systems in Sub- Property rights instruments have the potential to Saharan Africa. invest in livestock. grant resource users a tangible interest in the envi- ronmental consequences of their actions (Young 1996). Property rights need to be enforced by gov- reflected in their market price. Examples include con- ernments and, in a subsidiary fashion, by traditional centrate feed, fossil fuel, inorganic fertilizer, livestock institutions. Lack of security of tenure has often been products, land, mechanization, and genetic material. identified as a prime cause of land degradation. In the These should be priced higher by abolishing subsidies Brazilian Amazon, for example, security of tenure or, in some situations, imposing taxes. This, in addi- has been conditional on land clearance, with the tion to the quantitative effect of reduced consumption, result that significant areas of rain forest have been will induce a more efficient use of natural resources, cleared solely as a means to obtain access to wealth. with both economic and environmental gains. It will Land tenure arrangements are a key factor in facili- also favor a more even spatial distribution of livestock tating long-term investment in sustainable resource production and promote land-based systems. use, for both grazing and mixed systems. Corresponding]y, subsidies or tax relief can be pro- Government institutions become critical in estab- vided where natural resources are saved or used wise- lishing and enforcing a regulatory framework for the ly, such as by using renewable energy (methane) or protection of areas or the control of wastes, such as protecting biodiversity. Essentially, the recommenda- for industrial production systems or animal product tion is for full-cost pricing of the provision of all processing units. To strengthen the pollution control goods and services used by livestock producers and, and enforcement mechanisms in the developing also, those sectors that compete with the livestock world will be an important task in the future. sector for resources and markets. The establishment of financial institutions may There are many examples of incentive policies that alleviate pressure on the environment where signifi- reduce grazing pressures. Levying grazing fees for cant numbers of animals are kept primarily for communal areas is common. Fees can also be pro- investment purposes, such as by absentee herd own- gressive, with larger herds paying more per head ers in Africa or Latin America (see box 21.6). (Narjisse 1996). Full cost recovery, especially for Positive interest rates on safe deposits are providing water and animal health services, is another key ele- better altematives for investment if the primary pur- ment of good pastoral policies. Water has often been pose for keeping livestock is not production. a free good supplied publicly. Full cost recovery would reduce the number of large boreholes and Incentive Policies therefore reduce local degradation around these water points. Appropriate benefit-sharing systems for the Incentive policies rely on market forces. The more a protection of biodiversity, on a local and international production system is exposed to the market, the more scale, will nee(d to be devised. For intensive systems, susceptible it will be to price changes. In particular, levies on the discharge of waste and tradable manure intensive production depends on inputs that contain a quota schemes have been established in many indus- high component of natural resources, which are rarely trial countries iBrandjes and others 1996). 294 Livestock and the Environment: Issues and Options Regulations resources or by reducing the pollution load through modifying the chemical or physical char- Regulations can be used to promote more sustainable acteristics of work products use of resources by prescribing quantitative limits (of * Technologies that enhance natural resources by emissions or number of animals or use of inputs), making them more productive or richer technical methods, and access to resources. A further * Technologies that save natural resources by mak- regulatory instrunment is zonation, to promote a more ing it possible to generate more revenue from the even spatial distribution of animal production. For same resource or the same revenue from less grazing systems in the humid and subhumid zones, * Technologies that turn waste into products by clos- the establishment of protected areas for the preserva- ing cycles. tion of biodiversity is often the only appropriate mea- sure. For intensive mixed-farming systems and indus- Reduction of Environmental Damage trial systems, the control approach focuses on regulating manure management. Enforcement of regulations In the arid and semiarid grazing areas, careful water within reasonable social costs remains a major chal- development can help to prevent environmental dam- lenge. age. Investments in market technology may also reduce environmental pressure by encouraging Training and Extension greater off-take. In addition, new and more benign methods are now available to control diseases. Knowledge transfer, in many instances, is the key to Vaccines are available for the tick-borne diseases keeping a balance between livestock and the spread by a number of subspecies, although not yet environment. A considerable array of technologies are for east coast fever. Long-duration tick sprays now available to address environmental issues in mixed- exist that can be applied on individual animals, with farming and industrial systems. Livestock extension negligible environmental effect. A number of environ- services in the developing world have traditionally mentally benign control methods, using extremely focused on animal health services at the expense of low concentrations of easily degradable insecticides, production issues, and there has been an almost com- also exist for tsetse control. plete neglect of the environmental aspects of livestock Environmental damage of intensive systems can production. The introduction of environmentally be significantly reduced by focusing on emissions more benign practices often requires complex social from manure by, for example, improving collection and institutional changes, which front-line extension- and storage techniques. The main focus should be on ists are poorly equipped to handle. Alternative tech- reducing nitrogen losses, most of which are in the nology is badly needed in grazing systems, even form of ammonia from manure. Most losses can be though no ecologically and socially viable alterna- avoided by manure collection and covered storage tives to the traditional pastoral systems of arid zones facilities. Minimal amounts of ammonia are emitted have yet been identified. Although the current effi- when manure is collected under solid floors, and 80 to ciency of those traditional systems makes this a for- 90 percent reductions can be achieved by covering midable task, population pressure in these areas storage tanks (Voorburg 1994). A further reduction of makes it imperative as well. odor and ammonia loads can also be achieved through "bio-filters" or "bio-washers" that absorb odors and Technologies ammonia from polluted air. This is done by oxidizing ammonia into NO2 and NO3. Up to half of the ammo- Technological change occurs as a response to changes nia can be eliminated through air washing systems, in the availability of inputs and in demand. Over the which are, however, costly in investment and opera- past decades the livestock sector has responded main- tion (Chiumenti, da Borso, and Donanton 1994). ly by increasing efficiency and by undertaking the Nutrient losses during and after application of major structural shifts outlined earlier. Four sets of manure on soils can be significantly reduced by technologies, which overlap somewhat, can work to injecting or applying manure into the subsoil the benefit of the environment: (Brandjes and others 1996). Better timing of applica- * Technologies that reduce the environmental dam- tion in response to crop requirements avoids further age by alleviating the direct pressure on natural losses and enhances the nutritive value of manure. Henning Steinfeld, Cornelis de Haan, and Harvey Blackburn 295 Nitrification inhibitors can be added to slurry to reduce leaching from the soil under wet conditions. Box 21.7. Alternatives to Cereal Feeding In tanneries, dairies, and slaughterhouses, anaero- bic systems can purify wastewater and reduce by half Countries in the humid and subhumid tropics are the biological oxygen demand, while more sophisti- cereal-deficit. ]Livestock production, in particular cated anaerobic systems reach 90 percent BOD purifi- monogastric production, is thus faced with high cation (Verheijen and others 1996). Wastewater treat- prices for feed concentrates. This has spurred the ment first separates solids from the liquid, followed development of sugarcane-based feeding systems in a by biological treatment under anaerobic conditions number of tropical countries (Colombia, Cuba, (lagoons). Nutrients, such as phosphorus, are then Vienam, Philippines; see Preston and Leng 1994). removed by chemical or physical processes such as Sugarcane is one of the highest yielders of biomass adsorption, stripping, or coagulation. The same per unit of time and area. The juice can be used to process removes the remaining BOD as well as feed monogastrics, while the tops can be used to feed pathogens. ruminants. As a perennial crop, sugarcane production has very low rates of erosion and can be produced Resource-Enhancing Technologies with low external input. In the past, the association of sugarcane and livestock production has been prob- For grazing systems in arid zones, "deferred graz- lematic because sugarcane was traditionally produced ing," a traditional practice in many Middle Eastern on large plantations, geographically separated from countries, may regenerate vegetation, while in the livestock prodtuction. Recent developments in the semiarid zones, this may be accomplished by over- diversified use of sugarcane may lead to more vil- seeding or planting adapted fodder. The introduction lage-based, intensive systems of monogastric produc- of a multispecies grazing pattern will often encour- tion in the humid tropics. age better use of the vegetation and may have posi- tive effects on plant and animal biodiversity. Such technologies may be part of overall "management" Resource-Sparing Technologies approaches, which should explicitly acknowledge the efficiency of the current mobile systems and their The livestock sector continues to develop technolo- nonequilibrium status. For the humid tropics, peren- gies that increase the efficiency of natural resource nial grasses and legumes have been developed that use. In particular, these technologies target feed con- maintain soil fertility better than any other crop. version because feed typically accounts for 60 to 70 Biodiversity may also be enhanced through careful percent of production costs. Better feed conversion management of wildlife-livestock interactions in pas- saves land, while reducing the animals' waste load toral systems as well as through prevention of bush (see box 21.7). encroachment. A wide array of technologies improve feed con- Livestock, mainly through their input function version. They are mainly used in industrial systems in within a mixed crop-livestock system, enhance the industrial countries, but their use is quickly picking up main natural resource: land. Animal manure and trac- in developing countries as well. The most important tion make the land more productive than would be the ones are: case in their absence. Thus, all technologies that * Introducing multiphase feeding, whereby feed reduce nutrient losses from manure and improve the composition is better suited to the needs of animal efficiency of their application enhance land productiv- classes, which means that [ess waste is produced ity. For example, stall feeding doubles the effective and less nitrogen and phosphates appear in waste availability of nitrogen and phosphorus. Fodder shrubs and in the environment and trees reduce soil erosion and improve soil fertility. * Improving the accuracy of determining nitrogen In Indonesia, agroforestry systems were successfully and phosphate requirements, followed by improv- introduced with three strata, including grass, fodder ing the balance of feeds with these essential nutri- shrubs, and tree crops such as oil and coconut palms or ents. In this respect, the balance of essential amino cashew nuts. Raising the productivity of land through acids, the building blocks of feed proteins, has crop-livestock interaction protects other land from improved in the rations of pigs and poultry. For being brought into cultivation. example, a combination of more balanced feed, 296 Livestock and the Environment: Issues and Options improved digestibility, and inclusion of synthetic products, although associated with livestock waste, amino acids substantially reduces the protein con- reduces the environmental hazards associated with tent in feed and hence reduces the excretion of crop and agroindustrial waste. Also, more and more nitrogen and phosphorous by 20 to 40 percent (van food wastes are consumed by livestock, as urban der Zijpp 1991) agglomeration and changing eating habits offer a win- * Improving digestibility, by adding an enzyme dow of opportunity for the collection of food waste (called phytase) that catalyzes the digestion of from catering units to be recycled as feed. Large phosphates contained in feed and reduces waste amounts of straw, otherwise burned on the fields or loads on the environment left to decompose slowly with little nutritional benefit * Promoting feeding systems that reduce intake and to the crops, may be turned into quality feed, for stopping buffet-style ad libitum feeding, popular example through urea treatment (Li-Biagen, 1996, in the 1980s personal communication). * In mixed-farming systems of lower intensities, The cycles of matter can also be closed by using adding strategic supplementation for specific livestock waste as feed, energy, or fertilizer. classes of animals such as lactating cows or grow- Recycling manure-by feeding it to animals, live- ing animals stock, and fish (Muller 1980)-is practiced only on a * Improving the efficiency of feed conversion with limited scale. Reluctance to use manure as feed origi- enhanced genetic potential, better health and envi- nates mainly from the fear of health risks but is also ronmental conditions, and improved management due to the low nutritive value of manure, with the - Improving efficiencies in feed conversion for the exception of poultry manure. Poultry manure is incor- livestock sector as a whole by shifting to mono- porated in the diets of intensive systems, particularly gastrics, especially poultry and fish. This trend is as ruminant feed. In Asia, especially, manure is fed to likely to continue and will be particularly strong in fish and pigs (China). The low quality of manure as the developing countries. feed and high processing costs limit its competitive- Biodiversity is best preserved by intensifying live- ness vis-a-vis traditional commercial feeds. Sanchez stock production while intensifying crop production. (1995) provides a recent overview of the possibilities. Both will reduce the land requirements for given vol- Technology also exists to use the energy content of umes of final product, alleviate pressures on habitats manure. Biogas plants of all sizes and levels of and biodiversity, and limit the use of land and water sophistication not only recover the energy contained resources. in manure but also eliminate most of the animal and human health problems associated with contamina- Waste Technologies tion of waste by micro-organisms (see box 21.8). Other methods of controlling the waste load are the Historically, the raison d'etre for keeping livestock purifying and drying of manure. was its use of resources for which there was no alter- Promising approaches exist to reduce emissions native use. This explains why efficiency per animal from manure lagoons by recovering methane and was not, and in many low-input systems still is not, a using it for energy. Large confined animal operations major concern. allow such techniques to be profitable. This methane The conversion of organic waste into livestock can be used for on-farm energy to generate electricity, and the slurry effluent can be used as animal feed, as Box 21.8. Biogas in China aquaculture supplements, or as crop fertilizer. The controlled bacterial decomposition of the volatile In China, more than 5.3 million rural biogas systems solids in manure reduces the potential for contamina- are in place, producing 1.25 million cubic meters tion from runoff, significantly reduces pathogen lev- annually. Biogas is used for household heating and els, removes most noxious odors, and retains the cooking, poultry hatching, tea roasting, and grain and organic nitrogen content of the manure. fruit storage. The slurry is used as fertilizer, for fish Methane can be recovered in covered lagoons fanrning, and for pig feed, which shows good results where manure solids are washed out of the livestock in semi-intensive production systems (Henglian housing facilities with large quantities of water, and 1995). the resulting sluny flows into an anaerobic primary lagoon. Such anaerobic conditions result in signifi- Henning Steinfeld, Cornelis de hraan, and Harvey Blackburn 297 cant methane emissions, particularly in warm cli- land, water, and biodiversity. mates, but the methane can be recovered by placing Mixed-farming systems will see continued inten- an impermeable floating cover over the lagoon and sification and growth. Smallholder and family mixed applying negative pressure. Alternatively, digesters farming will remain predominant for some time to can be used. Large-scale digesters are engineered ves- come, with livestock based on crop by-products and sels into which a mixture of manure and water is surplus. Important productivity gains can be placed and retained for about 20 days. The digester is achieved by further enhancing the flow of nutrients heated to about 60°C, after which the gas is drawn off and energy between the two components. Live- and used for energy. Large dairy and pig farms with stock's role, in addition to production, is to enhance high energy requirements find these systems to be and substitute natural resources. The environmental cost-effective. Small-scale digesters do not include and economic stability of this system makes it the heating and are appropriate for warm climates only. prime focus for the continuing transfer and develop- ment of technology. Where involution occurs, in The Needfor Action areas of extreme population pressure, resource degradation, and. poverty, this must be fought Novel concepts are being developed to integrate crop through accelerated uptake of technology (feed and livestock production in a farming area rather than resources, animal. traction, development of small on a mixed farm. This method of area-wide integra- ruminants, and so forth). Under more favorable tion of crops and livestock allows individual enter- agroecological and market conditions, new forms of prises to operate separately while linking the flow of industrial production will have to be established. energy and organic and mineral matter through mar- These industrial systems will have to be based on the kets and regulations. This allows for highest efficien- resource endowments of a region, if nutrient bal- cies at the enterprise level while maximizing social ances are to be maintained and the environment's benefits. A policy framework is needed to organize ability to absorb waste respected. We are therefore crop and livestock production in this way. projecting industrial systems integrated into a wider concept of land use, particularly for pig and poultry Strategies for Livestock Production Systems production. This trend is already under way in some industrial countries. This would blend resource-sav- Putting the environment into the forefront does not ing technologies with the absorptive capacities of mean that only environmental objectives count. On the surrounding land. New organizational arrange- the contrary, environmental goals can only be tackled ments will have to be found to allow specialized effectively if and when sound economic objectives units to capitalize on economies of scale. The strate- are met. gy is thus to transform mixed-farming systems into Grazing systems will remain a source of exten- specialized and commercial enterprises (through sively produced animal products. To some extent, infrastructure and institutional development, animal these systems can intensify production by incorporat- production and health technologies, and processing) ing new technologies, especially in areas with high- in rural areas where land pressure is on the increase er potential (subhumid and highland areas). Often and where the market allows. this can be facilitated by stronger organizations, Industrial systems in areas of high animal densities local empowerment, and regulation of access to will have to cope with higher production costs as a resources. Where there is potential for mixed farm- result of more stringent regulations and pollution ing, policies need to facilitate the transition of graz- levies. In some cases, this will remove the competi- ing systems into mixed-farming systems in the semi- tive edge that industrial production has over land- arid and subhumid tropics by integrating crops and based production. Potentially, this will also raise the livestock (manure management, animal draft, prices for livestock products and reduce demand. residue feeding, fodder crops, and so forth). Higher prices will provide incentives for the land- Grazing systems can also intensify by diversify- based systems to intensify. Scales of industrial pro- ing and opening up other uses. In grazing systems, duction will grow further because of economies of livestock's role, in addition to providing a livelihood scale for waste treatment. This system's purpose is to pastoral people and market production, should be mainly to satisfy the soaring demand in many parts of to protect the natural resource base, in particular the developing world, which is often over and above 298 Livestock and the Environment: Issues and Options the supply capacity of land-based production and at interaction between livestock production systems and maximum efficiency of resource use. natural resources, coupled with factors such as market As resources become more scarce, livestock pro- access, there are development opportunities as well as ducers must continue to search for technologies that threats to sustainability. They are sketched in figure increase the efficiency of resource use if the rapidly 21.2 and lead to the identification of areas for strate- increasing demand is to be met without putting addi- gic intervention. tional strain on natural resources. The challenge is to To minimize environmental damage, governments obtain more efficiency without concentrating animal should, in very general terms, intervene as suggested production in a given area. Limiting the number of in figure 21.2. Strategic interventions need to focus on livestock while maintaining market mechanisms the phases of transition from one state to another, through, for example, tradable emission quotas seems where entirely new sets of technology are introduced to be an appropriate choice. Ideally, the advanced in an effort to intensify (where the agroecological and resource-saving technologies and the absorptive market potential allows), and on the fundamental capacities of extended rural areas should be married. pressures of poverty, population growth, and weak Thus the motto for most of the industrial world and institutions. the more densely populated parts of the developing It is evident that questions relating to livestock and world is to intensify, but not concentrate, animal pro- the environment cannot be solved in environmental duction. Such an approach would promote the spread terms only. A comprehensive perspective is needed to of processing into these areas and bolster economic ensure an enabling policy framework in which effec- development. tive technologies can be introduced. Technology In a schematic way, we can thus identify pathways remains the key component because future develop- for livestock production systems. Intensification, spe- ment, including that of the livestock sector, will cialization, and organization are the processes that depend on technology to substitute for natural characterize the different phases. As a result of the resources. This trend to knowledge-intensive systems Figure 21.2 Development Pathways of Livestock Systems Organization Specialization Intensification Specialized <~~~~~> 9 ~~~~~~crop systems Speialze ~~~~ +~~~~~~~~~~ ~~Area-wide ( Mixed systems 3 > ~~~~crop-livestock < u / f ~~~~~~~~~~~~~~~~~~integration Idstil\ livestockv/ Vlivestock systm ..Degradation . .. ..Involution ......... / Nutrient surplus Population pressure, economic growth Henning Steinfeld, Cornelis de Haan, and Harney Blackburn 299 is widely observed: smart technologies, supported by Protection Agency, Global Change Division, astute policies, can help to meet future demands while Washington, D.C. Processed. maintaining the integrity of the natural resource base. Brandjes, P. J., J. de Wit, H. G. van der Meer, and H. Better information on which to base decisionmaking Van Keulen. 1996. "Environmental Impact of is urgently required. Animal Manure Management." Consultant's report for the multidonor study on livestock and Conclusions the environment, Food and Agriculture Organiza- tion, World Bank, and U.S. Agency for Improved management is essential if the world's nat- International Development, Rome. Processed. ural resources are to continue to provide the basis for Byrne, P. V., C. Staubo, and J. G. Grootenhuis. 1996. life support and human well-being. Only with "The Economics of Living with Wildlife in improved management can the dual objectives of sus- Kenya." In J. IBojo, ed., The Economics of Wildlife. tainable agricultural production be fulfilled-to feed AFTES Working Paper 19. World Bank, the world's growing population, while sustaining its Washington, I).C. Processed. natural resource base. Livestock production is the Chiumenti, R., F. da Borso, and L. Donanton. 1994. largest user of land and is about to become the most "Dust and Ammonia Reduction in Exhaust Air important agricultural activity in terms of economic from Piggeries Using a Mechanical Wet output. Left to uncontrolled growth, the environment Separation System." In Proceedings of the Seventh will suffer, compromising human welfare. However, Technical Consultation of the European this is unlikely to happen. The opportunities not only Cooperative Research Network on Animal Waste to mitigate environmental damage but also to tap the Utilization. REUR Technical Series 34. Rome: immense development potential offered by livestock Food and Agriculture Organization. are large: awareness, political will, and readiness to Christiaensen, L., E. Tollens, and C. Ezedinma. 1995. act are growing among all those involved and ensure "Development Patterns under Population Pressure: that the problems will be tackled effectively. Agricultural Development and the Cassava- Livestock Interaction in Smallholder Farming References Systems in Sub-Saharan Africa." Agricultural Systems 48(1):51-72. The word "processed" describes informally repro- CIAT (Centro (le Investigacion en Agricultura duced works that may not be commonly available Tropical). 1994. Annual Report 1993. Cali, through libraries. Colombia. Alexandratos, N., ed. 1995. World Agriculture: de Wit, J., P. r. Westra, and A. J. Nell. 1996. Towards 2010. Rome: Food and Agriculture "Environmental Assessment of Landless Ruminant Organization of the United Nations; Chichester, Production Systems." Consultant's report for the U.K.: John Wiley and Sons. multidonor study on livestock and the environ- Behnke, R. H., Ian Scoones, and C. Kerven. 1993. ment, Food and Agriculture Organization, World Range Ecology in Disequilibrium. London: Bank, and U.S. Agency for International Overseas Development Institute and the Development, Rome. Processed. International Institute for Environment and Dregne, H., M. Kassas, and B. Rozanov. 1992. Development Commonwealth Secretariat. "Status of Desertification and Implementation of Boudet, G., M. Carriere, P. Christy, and H. Guerin. the United Nations Plan to Combat 1987. "Paturages et elevage au Sud de la Desertification." Desertification Control Bulletin Mauritanie. Etude integree sur les paturages, leur 20:6-18. conservation et leur restauration." Institut English, J., M. Tiffen, and M. Mortimore. 1992. d'Elevage et de Medicine Veterinaire des Pays "Land Resource Management in the Machakos Tropicaux (IEMVT), Maison Alfort, France. District, Kenya 1930-1990." Environment Paper Processed. 5. World Bank, Washington, D.C. Processed. Bowman, R. L., J. C. Croucher, and M. T. Picard. FAo (Food and Agriculture Organization of the United 1992. "Reducing Methane Emissions from Nations). 1996. "Cereals: Feed for Animals or Ruminant Livestock: India Prefeasibility Study." Food for People?" Fact sheet distributed at the A.T. International for the U.S. Environmental World Food Summit, Rome, 13-17 November. 300 Livestock and the Environment: Issues and Options Processed. Livestock Development in Tropical Africa. Kiel, .1997. "Current World Fertilizer Situation and Germany: Kieler Wissenschaftsverlag Vauk. Outlook 1994/95-2000/2001." Food and Agriculture Jodha, N. S. 1992. Common Property Resources: A Organization, United Nations Industrial Missing Dimension in Development Strategies. Development Otganization, and World Bank Discussion Paper 169. Washington, D.C.: World Working Group on Fertilizers, Rome. Processed. Bank. Fisher, M., I. M. Rao, M. A. Ayarza, C. E. Lascano, Kaimowitz, David. 1995. Livestock and Deforestation and J. I. Sanz. 1995. "Carbon Storage by in Central America. EPTD Discussion Paper 9. Introduced Deep-Rooted Grasses in the South Washington, D.C.: International Food Policy American Savannas." Nature 371(6494):236-38. Research Institute; Coronado, Costa Rica: Instituto Grazingland Management Systems. 1996. Interamericano de Cooperacion Agricola. "Environmental Impact Assessment of Livestock Leng, R. A. 1991. "Improving Ruminant Production Production in Grassland and Mixed-Farming and Reducing Methane Emissions from Ruminants Systems in Temperate Zones and Grassland and by Strategic Supplementation." EPA/400/1-91/004. Mixed-Farming Systems in Humid and Subhumid U.S. Environmental Protection Agency, Washington, Tropic Zones." Consultant's report for the multi- D.C. Processed. donor study on livestock and the environment, Mainguet, M., L. Guyot, M-C Chemin, and A. Food and Agriculture Organization, World Bank, Braunstein. 1992. "Modifications to the Sahelian and U.S. Agency for International Development, and Sudanian Ecosystems in Western Mali." In H. Rome. Processed. E. Dregne, ed., Degradation and Restoration of Habib, G. S., S. Basit Ali Shah, S. Wahidullah, G. Arid Lands. Lubbock: Texas Tech University. Jabbar, and Ghufranullah. 1991. "Importance of Milchunas, D. G., and W. K. Lauenrot. 1993. Urea Molasses Blocks on Animal Production "Quantitative Effect of Grazing on Vegetations (Pakistan)." International symposium on nuclear and Soils over a Global Range of Environments." and related techniques in animal production and Ecological Monographs 63(4):237-366. health, International Atomic Energy Agency, Muller, Z. 0. 1980. Feedfrom Animal Wastes: State of Vienna, Austria, 15-19 April. Processed. Knowledge. Animal Production and Health Paper Hendratno, C., J. V. Nolan, and R. A. Leng. 1991. 18. Rome: Food and Agriculture Organization of "Importance of Urea Molasses Multinutrient the United Nations. Blocks for Ruminant Production in Indonesia." Munasinghe, Mohan. 1994. "Economic and Policy International symposium on nuclear and related Issues in Natural Habitats and Protected Areas." In techniques in animal production and health, Mohan Munasinghe and Jeffrey McNeely, eds., International Atomic Energy Agency, Vienna, Protected Area Economics and Policy: Linking Austria, 15-19 April. Processed. Conservation and Sustainable Development. Henglian, X. 1995. "The Production of Rural Family Washington, D.C.: World Bank and the World Biogas in China." In Proceedings of the Workshop Conservation Union. of Peri-Urban Livestock Wastes in China, Beijing, Narjisse, H. 1996. "The Range Livestock Industry in China. 19-22 September 1994. Rome: Food and Developing Countries: Current Assessments and Agriculture Organization of the United Nations. Prospects." In Proceedings of the Fifth International Hiernaux, P. 1996. The Crisis of Sahelian Pastoralism: Rangeland Congress, Salt Lake City. 23-28 July Ecological or Economic. Pastoral Development 1995. Vol. 2: Invited Presentations. Denver: Network Paper 39a. London: Overseas Devel- Society for Range Management. opment Institute. Nelson, R. 1990. Dryland Management: The Houghton, J. T., B. A. Callander, and S. K. Verney, "Desertification Problem." Technical Paper 116. eds. 1995. Intergovernmental Panel on Climate Washington, D.C.: World Bank. Change and an Evaluation of the IPCC IS 92 Oldeman, L. R., R. T. A. Hakkeling, and W. G. Emission Scenarios. Cambridge, U.K.: Cambridge Sombroek. 1990. "World Map of the Status of University Press. Human-Induced Soil Degradation. An Explanatory ILCA (International Livestock Centre for Africa). Note." International Soil Reference and Information 1994. Annual Report 1993. Addis Ababa. Centre/United Nations Environment Programme, Jahnke, H. E. 1982. Livestock Production Systems and Wageningen, The Netherlands. Processed. Henning Steinfeld, Comnelis de Haan, and Harvey Blackburn 301 Ottichilo, W. K. 1996. "Wildlife-Livestock and U.S. Agency for International Development, Interactions in Kenya." Paper prepared for the Rome. Processed. World Bank/Food and Agriculture Organization Thomas, D. S. G., and N. J. Middleton. 1994. workshop, balancing livestock and the environ- Desertification: Exploding the Myth. New York: ment, Washington, D.C., an event associated with John Wiley. the fourth conference on environmentally sustain- Tucker, C. J., H. E. Dregne, and W. W. Newcomb. able development, World Bank, Washington, D.C. 1991. "Expansion and Contraction of the Sahara Processed. Desert from 1980-1990." Science 253:299-301. Preston, T. R., and R. A. Leng. 1994. "Agricultural U.S. Environmental Protection Agency. 1995. "Study Technology Transfer: Perspectives and Case of Livestock-Environment Interactions: Global Studies Involving Livestock." In J. Anderson, ed., Impact Domain. Methane Emissions." Consultant's Agricultural Technology: Policy Issues for the report for the multidonor study on livestock International Community. Oxon, U.K.: C.A.B. and the environment. Food and Agriculture Intemational. Organization, World Bank, and U.S. Agency for Saadullah, M. 1991. "Importance of Urea Molasses International Development, Rome. Processed. Block Lick and Bypass Protein on Animal van der Zijpp, A. J. 1991. "Sustainable Animal Production (Bangladesh)." International sympo- Production: A Response to Environmental sium on nuclear and related techniques in animal Problems." In J. Boyazoglu and J. Renaud, eds., production and health, International Atomic The Livestock Production Sector in Eastern Energy Agency, Vienna, Austria, 15-19 April. Europe as Affected by Current Changes, pp. Processed. 191-97. EAAP Publication 57. Wageningen: Sanchez, M. 1995. "Use of Animal Excreta as Feed." European Association for Animal Production. In Proceedings of the Workshop of Peri-Urban Verheijen, L. A. H. M., D. Wiersema, L. W. Hulshoff Livestock Wastes in China, Beijing, China, 19-22 Pol, and J. De Wit. 1996. "Management of Waste September 1994. Rome: Food and Agriculture from Animal Product Processing." Consultant's Organization of the United Nations. report for the multidonor study on livestock and Scoones, Ian. 1994. Living with Uncertainty: New the environment, Food and Agriculture Directions in Pastoral Development in Africa. Organization, World Bank, and U.S. Agency for London: Intermediate Technology. International Development, Rome. Processed. S6re, Carlos, and Henning Steinfeld. 1996. World Voorburg, J. H. 1994. "Farmers Options to Reduce Livestock Production Systems: Current Status, Odor and Ammonia Emissions from Animal Issues, and Trends. Animal Production and Health Buildings and Storage." In Proceedings of Seventh Paper 127. Rome: Food and Agriculture Technical Consultation of the European Organization of the United Nations. Cooperative Research Network on Animal Waste Sidahmed, A. E. 1996. "The Rangelands of the Utilization. REUR Technical Series 34. Rome: Food Arid/Semiarid Areas: Challenges and Hopes for and Agriculture Organization. the 2000's." Keynote address to the symposium on World Resources Institute. 1992. World Resources range management, international conference on 1992/1993. New York: Oxford University Press. desert development in the Arab gulf states, KISR, Young, M. D. 1996. "Maintaining Harmony: Kuwait, March. Processed. Equitable and Efficient Means to Minimize the Thomas, D., and D. Barton. 1995. "Interactions Adverse Impact of Livestock on the Environment." between Livestock Production Systems and the Paper presented at the workshop on blancing live- Environment: Impact Domain. Crop-Livestock stock and the environment, an event associated Interactions." Consultant's report for the multi- with the fourth conference on environmentally sus- donor study on livestock and the environment, tainable development, World Bank, Washington, Food and Agriculture Organization, World Bank, D.C. Processed. 22 Public Policies to Reduce Inappropriate Tropical Deforestation David Kaimowitz, Neil Byron, and William Sunderlin This chapter seeks to explain the ways in which, and * Substantial emission of carbon dioxide, which the extent to which, public policy measures might be contributes to global warming used to reduce inappropriate tropical deforestation. * Other types of local and regional climate change. In this introductory section, we define the terms used We follow the Food and Agriculture Organization (deforestation, agents of deforestation, and inappro- of the United Nations (FAO) in defining deforestation priate deforestation) and present our central argu- as "the sum of all ... transitions from natural forest ment and the structure of our chapter. classes (continuous and fragmented) to all other In the last 8,000 years, world forest cover classes" (FAO 1996: 22). This definition focuses declined more than 40 percent, from 62 million to 33 attention on the loss of natural forests and is framed million square kilometers (Bryant, Nielsen, and in the context of specific definitions of forest and Tingley 1997). Contemporary concern about defor- nonforest classes of land use. Analysis of transitions estation focuses on tropical countries because that is among these classes can shed light on the cause-and- where the majority of forest cover is being removed. effect relations underlying deforestation and help to In the period 1980-90, an estimated 137.3 million determine the optimal uses of land. Based on this def- hectares of tropical forests were cleared, about 7.2 inition, the amount of deforestation in the period percent of the total that existed in 1980 (FAO 1996).' 1980-90 was estimated at 8.2 percent in Asia, 6.1 Any claims regarding the magnitude of global defor- percent in Latin America, and 4.8 percent in Africa estation must be made with caution, however, (FAO 1996). because there are serious problems with the data and Forest degradation is a crucial adjunct to defor- definitions used. estation. Forest degradation is a "decrease of density Regardless of the exact magnitude and location of or increase of disturbance in forest classes" (FAO tropical deforestation, it is clear that deforestation 1996: 21). Degradation potentially implies a major often implies: loss of forest products and ecological services, even * The loss of livelihoods for forest-dependent peo- when there is little deforestation as such (outright ple, many of whom may not wish to or be able to conversion of forest to another land use). In what fol- find other sources of employment lows, we focus principally on deforestation, but con- * Decreasing stocks of fuelwood and nontimber for- cerns about degradation may be just as important in est products as well as industrial timber the long run. * Greater soil erosion and river siltation The process of deforestation must be analyzed at * Substantial species and gene loss, in view of trop- two levels: agents and causes. The agents of defor- ical forests' h.igh level of endemic biodiversity estation are the people who physically (or through 302 David Kaimowitz, Neil Byron, and William Sunderlin 303 decisions over their labor forces) convert forests to change in land tenure laws), institutional factors (the nonforest uses: small farmers, plantation and estate decision to deploy more forest rangers to a particular owners, forest concessionaires, infrastructure con- area), and political decisions (a change in the way struction agencies, and so forth. FAO (1996) attributes forest concessioins are allocated), among others. This most deforestation in Latin America to medium- and chapter discusses how certain causes can be manipu- large-scale operations (resettlement schemes, large- lated to influence the behavior of agents, so as to scale cattle ranching, hydroelectric dams) and notes lessen the rate of inappropriate deforestation. that deforestation in that region is characterized by Two main reasons lead us to believe that some transitions from closed forest to nonforest land uses. deforestation is inappropriate and that inappropriate In Africa, deforestation is related largely to the deforestation is a significant problem. First, defor- expansion of small-scale farming and rural popula- estation general:Ly causes negative externalities that tion pressure (the growing number of smallholders) generate costs to society not reflected in existing associated with the conversion of closed forest cover prices and has long-term consequences that individ- to short fallow farming. Deforestation in Asia is asso- ual producers rarely consider. Second, the relative ciated with both relatively large operations (as in importance of the negative extermalities tends to Latin America) and rural population pressure (as in grow over time if an increasing proportion of defor- Africa) and involves conversion of closed forest to estation occurs in areas that have only marginal value long and short fallow, plantations, and nonforest for agriculture but that sequester large amounts of uses. In the larger countries, such as China, India, and carbon, have fragile soils, or are high in biodiversity. Indonesia, different types of processes dominate Any decision regarding which deforestation is depending on the specific context of the region appropriate or not is ultimately political in nature and involved. There are no simple, universal single-cause cannot be justified on purely technical grounds. explanations. Nevertheless, many forestry policy experts agree that The relative contribution that different agents forest clearing is more likely to be inappropriate make to deforestation is difficult to determine for two when it involves the following types of areas: reasons. First, little reliable information exists on this * Areas that currently have little value for agricul- subject. Second, even if it did, in many situations var- ture by virtue of the quality of their soil or gradi- ious agents operate in the same location and cannot ent of the land be analytically separated. Agents are often interde- * Areas that have large amounts of biodiversity, pendent (for example, cattle ranchers who supply particularly endemic biodiversity, that is poorly smallholders with chain saws), and different activi- represented in existing protected areas2 ties can take place sequentially in the same location * Areas that have large numbers of forest-dependent (small farmers frequently occupy abandoned timber people who show no inclination to abandon their concessions, and many farmers engage in salvage existing livelihood strategies logging prior to clearing land for farmiing). * Areas that, by virtue of their rich potential and One reason for the intensity and confusion in the comparative advantage for timber production, deforestation debate is that proponents of different make this the most profitable use of land, even views tend to emphasize the role of loggers, small after the area is logged for the first time farmers, or large ranchers and to portray these agents * Fragile areas where the ecological cost of conver- as being universally dominant, when in fact the dif- sion resulting from "downstream" effects out- ferent agents coexist and closely interact in many weighs any economic gain from nonforest land countries, and their relative importance varies over uses time and between regions. Given such complex rela- * Periurban areas where forests play a key role in tionships, there are no clear guilty or innocent par- conserving aquifers, providing fuelwood, and ties, and no one should expect neat, simple solutions. supporting recreational and tourism activities. The "causes" of deforestation refer to the multiple It is important to distinguish inappropriate from factors that shape agents' actions and, in particular, appropriate deforestation for three reasons. First, it is their decision to deforest. These causes include mar- often assumed (at least implicitly) that all tropical ket forces (international price fluctuations of agricul- deforestation is inappropriate, and this is not neces- tural export commodities), economic policies (cur- sarily the case. Second, to the extent that inappropri- rency devaluation), legal or regulatory measures (a ate deforestation can be defined, the geographical 304 Public Policies to Reduce Inappropriate Tropical Deforestation areas (and sometimes socioeconomic groups) can be should not be forested is a tall challenge, precisely specified that should be the targets of policy designed because it requires giving attention to three (not to reduce inappropriate deforestation. Third, by always mutually consistent) criteria and because clearly defining the areas appropriate for conversion, much remaining tropical forest cover falls into the pressure can conceivably be eased on forests where "ambiguous" zone described in table 22.1. The three conversion is inappropriate. categories are not discrete but rather the end points Table 22.1 offers a rudimentary conceptual and the midpoint of a gradient that ranges from clear- framework for distinguishing inappropriate from ly unsuitable to extremely suitable. appropriate deforestation. It takes into account three The chapter's basic thesis is as follows. cross-cutting categories of valuation: biophysical, eco- Individuals and businesses deforest because it is their nomic, and political. No one category alone provides most profitable alternative. To get them not to defor- a sufficient basis for cogent decisionmaking on est in situations where forest clearing is inappropriate, whether a forest should be protected or cleared. deforestation must be made less profitable, or other Decisionmakers must take all of them into account alternatives (either based on retaining forests or com- simultaneously. pletely outside forest areas) must be made more prof- The horizontal correspondences in this table are itable. idealized and do not necessarily correspond to the Deforestation can be made less profitable by "real world." For example, it can be assumed that * Reducing the demand or prices for products pro- over time, because an ever larger share of remaining duced from newly cleared land tropical forests are in remote or hilly areas, these * Increasing the unit costs and riskiness of activities forests will be less appropriate for conversion associated with deforestation because of their relative inadequacy for agriculture or * Eliminating speculative gains in land markets. their high biodiversity value.3 These same forests, Alternatives to deforestation can be made more however, will likely be targeted by a growing number profitable by of stakeholders (precisely because they are scarce), - Increasing the income stream to be obtained from and the chance of "win-lose" and "lose-lose" out- maintaining forests comes will increase as well. Identifying which forests * Reducing the costs of maintaining forests Table 22.1. Conceptual Framework for Distinguishing "Appropriate" and "Inappropriate" Deforestation (criteria influencing appropriateness) Form of deforestation Biophysical Economic Political Inappropriate High adverse local or High utility to local people No potential downstream effects of as forest and low or stakeholders conversion, high biodiversity, temporary agricultural ultimately gain and high biomass (or other) potential from the conversion ("lose-lose") Ambiguous Intermediate status between Intermediate status Some stakeholders the characteristics listed between the win, others lose above and below characteristics listed from the conversion above and below ("win-lose") Appropriate Low adverse local or Low utility to local people All potential downstream effects of as forest and high and stakeholders gain conversion, low biodiversity, lasting agricultural from the and low biomass (or other) potential conversion Note: One might rightly ask, "If the situation is 'lose-lose,' why would someone in this situation deforest?" Here we mean "lose-lose" in two possible senses. First, the agent(s) deforesting expected a gain but had to move (and therefore "lost") because the soil could not support agriculture. Second, stakeholders appear to gain (because they obtain a source of livelihood through deforestation) yet lose in the relative sense (because they could have obtained a better livelihood through means other than dieforestation). David Kaimowitz, Neil Byron, and William Sunderlin 305 * Increasing the opportunity costs of labor and cap- policies involved are mostly national in scope, and gov- ital that might otherwise be used in activities asso- emments can irmplement some of them easily and cheap- ciated with deforestation. ly (at least with regard to their direct costs). Because The chapter's organization follows this typology the policies' direct effects are proportional in magni- of policy options for reducing inappropriate defor- tude to the quantity of products sold, larger producers estation. The first section focuses on policies affect- bear most of whatever negative effects the policies ing the prices of and demand for agricultural and may have on producers' incomes. forestry products. The second section discusses poli- Nevertheless, these policies are blunt instruments, cies that make production associated with deforesta- which are difficult to target and thus likely to have tion more or less costly and risky. Section three perverse side effects. They rely entirely on the mar- examines policies that affect land speculation. ket to distinguish where and how supply should be Section four looks at different ways to increase the reduced, and this is unlikely to meet environmental profitability of maintaining forests. Section five objectives efficiently. In some cases, lower prices addresses the issue of the opportunity costs of capital may lead supply to fall more rapidly on lands that are and labor. All the policies analyzed are evaluated on economically marginal for agriculture, but this does the basis of the following six criteria: effectiveness, not hold in other situations, and land that is econom- ability to be targeted, equity, political viability, direct ically marginal for agriculture is not necessarily the cost, and indirect cost. In each section, a summary most important land to maintain under forest cover to table evaluates the policy reforms discussed with meet the objectives of biodiversity, watershed protec- respect to these criteria. tion, or timber production.4 Often, it is preferable to We conclude that there is no perfect or generaliz- find ways to meet the demand for agricultural products able policy for reducing inappropriate deforestation. by using policies designed to promote more intensive There are no "first-best" options. Each national situ- agricultural p:roduction or concentrate production ation is different, much uncertainty remains about where it is most appropriate rather than by reducing key cause-and-effect relations, and tradeoffs usually aggregate demand. In fact, depressing the demand for exist between the criteria for evaluating the policies. tropical agricultural products lowers the incentives Some policies are effective, but difficult to target, for technological changes that might lead such prod- costly, or not politically viable. Other policies are ucts to be produced more intensively, on less land. easier to target, but more costly or not politically When deforestation is driven largely by land specula- viable, and so forth. tion or production of subsistence or nonmarketed goods, these policies will have little impact. Besides, Policies to Reduce Prices and Demand it is difficult to win political support for policies that for Tropical Agricultural and Forestry Products restrict the demand for or production of tropical products, particularly export products, and that Demand for tropical products is a function of the size restrict the incomes of primary producers in tropical and incomes of the consuming population, relative developing countries. prices, quantitative trade restrictions, and consumer Where higlh agricultural prices result from direct preferences. Policies that affect any of these variables or indirect government subsidies, as is often the case, can also affect demand. This section first analyses the for example, with sugarcane production, the argu- effect on forest clearing of policies that affect the ment for reducing demand by eliminating those sub- demand for agricultural goods, then looks at policies sidies is somewhat stronger. Guaranteed price sup- that influence the demand for forest products. port schemes have, in the past, increased the expected High prices and demand for tropical agricultural profitability of deforestation both by increasing the commodities can stimulate the clearing of forests to expected sale price of outputs and by reducing the produce these commodities, as illustrated by the "ham- commercial risks associatecl with such ventures. burger connection" in Central America or the cassava Market promotion and marketing boards have similar fodder export schemes in Thailand. Therefore, under effects: they both reduce market risk and volatility certain circumstances, policies that reduce the and increase long-run average returns to producers. demand for tropical agricultural products could limit Under these circumstances, eliminating subsidies or the conversion of forests to croplands and pastures. marketing assistance for products from newly cleared This type of approach has certain advantages. The forestlands may be effective, low cost, and socially 306 Public Policies to Reduce Inappropriate Tropical Deforestation progressive, although in some countries it may be between economic growth and deforestation is for politically unacceptable because of the perceived countries to promote patterns of economic develop- impacts on farmers' incomes. In particular, the fact ment that are less dependent on primary agricultural that most price subsidies accrue to large farmers, who and forestry commodities. The alternative is to devel- produce most of the (subsidized) agricultural prod- op a well-managed, ecologically and economically ucts, may make them politically difficult to elimi- sustainable forest sector. In either case, it is essential nate. to adopt policies that limit inappropriate forest con- Ceteris paribus, higher population implies more version. demand for agricultural products. Depending on the Governments influence the relative prices of, and scope of the markets, the level of local, national, or hence demand for, tropical products through various international population may be relevant. However, measures, including exchange rate policies, tariffs, income level and distribution, consumer preferences, nontariff trade barriers, pricing of government utili- and other factors that affect relative prices generally ties, and price controls, among others. The trend in influence demand even more, at least in the short or tropical countries in recent years has been toward medium term. Moreover, population growth can also trade liberalization and more free market policies. influence forest clearing through numerous other The European and U.S. governments, in contrast, causal relations, which often have conflicting effects continue to use tariff and nontariff trade barriers to (Bilsborrow and Geores 1994). After a considerable regulate imports of agricultural products to meet lag, policies that lead to lower population growth commercial, political, and environmental objectives. could potentially limit deforestation by lowering the Real currency devaluations, designed to increase demand for agricultural products, but the magnitude the relative price of tradable goods and services com- of that effect is unlikely to be large. Because popula- pared with nontradables, generally increase real pro- tion also affects many other important policy vari- ducer prices of tropical agricultural products, making ables, governments are unlikely to design population it more profitable to convert forestland to agriculture policies principally with regard to their impact on (Capistrano 1990 and Wiebelt 1994). This problem is deforestation. most likely to arise where export producers are Per capita income and economic growth greatly important agents of deforestation, such as in certain influence the demand for tropical products, although, countries of Central America (beef), South America again, this is only one way they affect deforestation. (soybeans), and West Africa (cocoa; Kaimowitz, Several authors have speculated about the existence Thiele, and Pacheco 1996 and Kaimowitz 1997). of an "environmental Kuznets curve," whereby at Because exchange rate policies are broad and far- lower levels of per capita income economic growth reaching, under most circumstances concerns over leads to greater deforestation as a result of increased inappropriate deforestation cannot be expected to be demand for tropical products, but after income levels a major factor in their design, nor should they be. reach a certain threshold, deforestation declines However, in situations where a large portion of antic- because the economies become less dependent on ipated export growth generated by devaluations agricultural and forest products, primary production comes from the expansion of agricultural production becomes more intensive, new employment opportu- to areas currently under forests, the issue cannot be nities arise, and demand for forest preservation grows avoided. Policymakers should explicitly consider the (Capistrano 1990). tradeoffs between the possible benefits of short-term The empirical evidence on this issue is weak, con- increases in exports and the costs associated with for- tradictory, and based mostly on cross-sectional data est conversion. Alternatively, they might try to devise and on the few "tiger economies of East Asia," which simultaneous parallel measures to minimize the pre- may not reflect how the majority of tropical countries dictable adverse impacts. are likely to evolve in the future. What evidence there A similar argument applies to agricultural price is, however, suggests that given existing levels of per policies focusing on production for domestic con- capita income in most tropical countries, without a sumption, such as price controls, import restrictions, change in the pattern of development, economic guaranteed minimum prices for food crops, or public growth is likely ito lead to greater forest clearing in marketing of foodstuffs. In most cases, the issue of the medium tenm (Stern, Common, and Barbier deforestation will appropriately receive little atten- 1996). This implies that one way to avoid a tradeoff tion when designing these policies. In some David Kaimowitz, Neil Byron, and William Sunderlin 307 instances, however, policy changes can be expected chase cattle. For example, the highest rate of popula- to shift relative prices in a way that will promote tion increase in the Philippine uplands was in the rapid agricultural expansion in forested areas. In municipalities with logging concessions (Garrity, Ghana, for example, Munasinghe and Cruz (1994) Kumar, and Guiang 1993). But in other situations, have shown that policies that improve agricultural they do not affect forest conversion.5 Moreover, poli- prices promote extensive, rather than intensive, agri- cies that depress the demand for timber can be a dis- cultural growth and that the environmental costs may incentive for sustainable forest management, where not justify the benefits. In such circumstances, poli- that is a possibi]ity, even though in the short term they cymakers should explicitly consider whether short- limit some of the negative effects associated with cur- term agricultural growth is worth the long-term con- rent logging practices. The question, therefore, is not sequences and implement damage control measures to reduce the aggregate demand for tropical timber, alongside the necessary macroeconomic reforms. but rather to reduce the demand for the tropical timber The conversion of forests to pastures in Central that is closely and visibly associated with inappropri- America illustrates many of these points. Many ate tropical deforestation and degradation, while, if authors have cited this as a case where strong demand possible, stimulating the demand for tropical timber for beef imports in the United States in the 1970s and that can be proven not to be linked to deforestation. high international beef prices led cattle ranching to We return to this issue below. expand rapidly at the expense of forests (Myers As with agricultural products, larger populations, 1981). During the 1980s, however, Central American economic growth, exchange rate devaluation, and beef and dairy producer prices fell sharply as a result more free market pricing policies generally increase of increased protectionism and reduced per capita both the domestic and export demand for timber and beef consumption in the United States, domestic hence logging. Large devaluations, in particular, have price controls, overvalued exchange rates, and inter- been associated with major increases in tropical tim- national donations of powdered milk. This, along ber exports from Bolivia, Central Africa, and with other factors, reduced the pace of forest conver- Southeast Asia (Anderson, Constantino, and Kishor sion to pasture, but by no means stopped it. 1995 and Tchotngui and others 1995). Ironically, when demand and prices declined, live- There has been much controversy over the use of stock production fell more rapidly in traditional cat- log export taxes or bans to reduce the logging of nat- tle regions than along the agricultural frontier, ural forests, among other objectives (Barbier and oth- because producers on the frontier had few alterna- ers 1994).6 These policies are relatively easy to tives to investing in cattle, could simultaneously enforce (although there always tends to be some slip- speculate in land markets, and obtained high initial page) and generally reduce the amount of logging in yields from pasture in recently deforested areas the short run, but at high economic cost. Their medi- (Kaimowitz 1997). Thus, as with the cocoa example um- and long-term impacts are less certain, but many (Ruf and Siswoputranto 1995), it cannot be generally experts believe that bans discourage efficient log- assumed that the "new" producers at the forest mar- ging, wood processing, and investment in sustainable gin will be the first to become nonviable as demand forest management and encourage the domestic con- contracts-although they are physically "at the mar- sumption of timber. This sometimes leads to even gins," they are economically less marginal than the greater logging than would have occurred without the producers in "older" areas. ban (as occurred in Indonesia). Similar arguments apply to tropical timber, but Bans and taxation of tropical timber imports, cur- there are important differences. Because industrial rently being discussed in certain industrial countries, logging in most tropical countries is selective and are similar, with two notable differences. First, does not involve direct conversion of large forest because not all countries that import timber are like- areas to other uses, the effect on forest conversion of ly to adopt suc:h a policy, its effect on the domestic policies that affect the demand for forest products is prices of timber in exporting countries will probably much more indirect. Logging companies frequently be smaller than with an export ban (Buongiomo and facilitate forest conversion to other uses by building Manurung 1992). Second, importing countries roads that assist farmers subsequently to enter and appear to be rmore prepared to distinguish timber clear land or by providing investment capital to farm- taken from sustainably managed forests from timber ers or ranchers that can be used to clear land or pur- taken from forests that are not sustainably managed. 308 Public Policies to Reduce Inappropriate Tropical Deforestation This could provide an incentive to invest in sustain- macroeconomic, agricultural, and trade policies have able forest management that does not exist with tra- on forest clearing, actively promote patterns of eco- ditional log export bans. It could still prove difficult, nomic growth that rely less on the liquidation of nat- however, to define sustainable forest management ural capital, and take a more long-term view in and monitor whether imported timber comes from assessing policy tradeoffs. forests that meet those standards. Finally, brief mention should be made of the case Policies That Increase the Costs and Risks of commercial firewood collection, which is some- of Production Associated with Deforestation times directly responsible for clearing forests in drier periurban areas. Higher urban incomes generally lead In this section, we examine policy measures that can to rapid substitution of fuelwood with other energy reduce inappropriate deforestation by increasing the sources and hence can actually reduce the pressure on costs and risks of production associated with defor- forests (and vice versa.) Attempts to promote the sub- estation. Often these policy measures eliminate gov- stitution of fuelwood by providing subsidies that emient subsidies or other distortions generated by lower the price of energy sources such as kerosene, previous policies and thus actually save govemments butane, or electricity, however, have had little suc- money, rather than involving additional expenses. First cess in reducing the consumption of fuelwood and we review policies related to agricultural production, have been expensive for governments (Mercer and such as agricultural input subsidies, colonization Soussan 1992). schemes, and research and extension. Then we look at In summary, as shown in table 22.2, any policies policies affecting logging in natural forests. Finally, designed to restrict forest conversion by reducing the we examine the implications of road construction and demand for primary products are blunt instruments, maintenance and of zoning and protected areas, which which make them alternatives of last resort. In some affect both agriculture and logging. circumstances, however, they may be the only realis- Basically, government actions that encourage tic alternative for achieving rapid reductions in inap- agricultural intensification tend to slow deforestation propriate deforestation. Policymakers must become at the forest-agriculture frontier, while actions that more conscious of the potential impact that their encourage extensification of agriculture (bringing Table 22.2. Demand-Related Policies Political Policies Effectiveness Targetability Direct costs Indirect costs Equity viability Population control Limnited None Moderate Low Unknown Moderate' Limits on economic growth Moderate None Low High Unknown Low Appreciated exchange rates Moderate5 None Variablec High Mixed Variable' Price controls on Moderate Limited Moderate High Mixed Low tropical products Export bans and taxes Moderated Limited Low Variablee Favorable Variable on logging Import restrictions on Moderated Moderate Low' Unknown Favorable Moderate primary products (coffee, cocoa, timber, or beef) a. Population control is not politically viable in some highly religious countries. b. Appreciated exchange rates will be more effective where forest conversion is for export production. c. Appreciated exchange rates have higher direct costs for governments and are less politically viable when countries have balance of payments or debt problems. d. Policies are more effective in the short run. e. Indirect costs depend on the size of the sector and are higher in the medium and long run. f. Targetability of import restrictions is higher if certification is involved, but direct costs are also. David Kaimowitz, Neil Byron, and William Sunderlin 309 new land into production) are likely to promote ing land colonization and resettlement schemes to deforestation. We define intensification as increasing reduce inappropriate deforestation would probably the amount of labor, capital, and information applied be moderately effective, be readily targetable, have per hectare of land. Whether the expansion of market no cost to the national treasury, and result in substan- demands and higher product prices lead to more or tial savings in budgetary outlays. Its political viabili- less pressure on the forest-agriculture frontier will ty is unclear. We suggest that subsidies or programs depend on the farmers' choice of technology (that is, that encourage extensive expansion of agriculture how much to intensify and how much to expand an would promote deforestation, while subsidies that area) in response to perceived opportunities. favor intensification would tend not to encourage Assuming that farmers respond rationally to the deforestation. Thus it would be appropriate to distin- expected net returns from either means of increasing guish among agricultural subsidies, before conclud- their output, any public policies that affect their input ing that their abolition would help to slow deforesta- prices will affect their decisions. The effects of gov- tion. For example, removal of fertilizer subsidies in emnment subsidies for agricultural inputs such as fer- parts of Sub-Saharan Africa may well accelerate tilizer, water, seeds, pesticides, credit, fuel, and trans- deforestation. ]!t would be a mistake, therefore, to port on agricultural intensification and extensification believe that eliminating subsidies alone is sufficient are far from uniform. Subsidies for supplying irriga- to solve the problem of inappropriate deforestation. tion water generally facilitate intensification, because Brazil and Cerntral America, for example, greatly it is not feasible to supply irrigation water to remote reduced the tax incentives and subsidized credit for areas at the forest frontier. Fuel and transport subsi- cattle ranching in recent years, and in both cases dies tend to favor extensification by making it more inappropriate cleforestation declined, but remained profitable to farm in remote areas. Livestock credit high (Oz6rio (le Almeida and Campari 1995 and also tends to facilitate extensification, because tropi- Kaimowitz 1997). cal livestock production is typically an extensive Although agricultural research and extension activity. Fertilizers, seeds, and agricultural credit activities have rarely been aimed explicitly at pro- subsidies can facilitate either intensification or exten- moting deforestation, nevertheless there are instances sification. It is generally accepted that the agricultur- where farmers adopted the recommended new tech- al green revolution, whose success was at least par- nologies not only on their existing fields but also on tially due to the subsidized availability of the inputs new fields that were formerly forest (even protected required for intensification (credit, fertilizer, pesti- forests and national parks). As Angelsen (1995) cides, transport), helped to save considerable areas of notes, if a new agricultural technology is worth forests in Southeast Asia by making the cultivation of adopting on ex-isting farmland, it may also be viable unirrigated rice on forest hillsides much less prof- on nearby lands that were previously left in forests itable. and may make it profitable to farm in places where Governments of many countries have deliberately previously it was not. Thus, for example, improve- and explicitly encouraged forest conversion in the ments in soybean technology in Bolivia and Brazil past through official resettlement and colonization have been an important factor in promoting the clear- programs, and some still do, for example the Trans ing of forest areas to expand soybean production. Amazonia Highway and the Rondonia colonization New pasture varieties may also have made it prof- scheme in Brazil and the transmigration program in itable to clear forest for pasture in Latin America in Indonesia. Because so many governments have used places where otherwise it would not have been. One these policies, it might seem at first glance that argu- possible policy reform would be to ensure that the ing for their removal would be hopeless. Moreover, content of agricultural extension services is not many governments in Africa, Asia, and Latin biased to encourage deforestation. However, such America have used "national defense" arguments to measures are likely to be only moderately effective, justify their colonization schemes, and these argu- they are difficult to target, and, although they have ments carry great political weight. Nevertheless, the few cost or equity consequences, they are unlikely to current political climate favoring the elimination of find a strong political constituency. public subsidies and less government intervention Frequently recommended policy reforms that te'nds to favor these reforms. would reduce the profitability of industrial logging Removing credit and input subsidies and eliminat- (and so presumably reduce the amount of inappropri- 310 Public Policies to Reduce Inappropriate Tropical Deforestation ate deforestation due to such logging) include than logs for export. Accelerated industrialization has increasing stumpage charges and other fees and been accomplished in Indonesia and Malaysia, reducing other direct and indirect subsidies to timber through a suite of policy measures linked to trade and processors. Timber stumpage charges and other tim- tariff policies, tax incentives, log export restrictions, ber fees are often extremely low, and governments "cheap logs," worker training programs, and so forth frequently justify this as a way to attract new invest- (de Los Angeles and Idrus 1990). However, most ment or to capture social benefits like infrastructure independent studies suggest that the short-run eco- development and employment creation. For example, nomic cost to Indonesia has been extremely high the Bangladesh Forest Industries Development (Lindsay 1989 and Douglas 1995). Even Malaysia's Corporation made only notional payments (a few successful and widely acclaimed policies for acceler- cents per ton) for areas, logs, and bamboo for more ated industrialization and "downstream, value added than 20 years (FAo 1993). The Hindustan Paper processing" have recently been assessed by a World Corporation in Kerala (India) paid $0.50 per ton of Bank team as economically inefficient and too ambi- eucalyptus pulpwood compared to the Forest tious. By making logging more profitable than it Department's direct production costs of almost $25 would otherwise be, the policy on subsidies con- per ton (FAO 1993). Naturally, the industry operations tributed to the expansion of area logged annually. expanded considerably, but when these stumpage Further, if the prices of wood were set at artificially charges were eventually increased to realistic levels, low levels, distorting wood processing decisions and the company's activities contracted, leaving excess leading to excessive physical wastage (Douglas capacity, unemployment, demands for more planta- 1995), then greater areas would have to have been tions, and yet more logging of protected natural logged to achieve the same final output of timber forests. Subsidizing log input prices is a very ineffi- products. cient way of creating additional employment and dis- Road construction plays a crucial role in defor- courages intensive recovery of finished products in estation and has been the topic of much analysis and mills.7 The systems used for (noncompetitive) alloca- debate. It is often argued that new road construction tion of logging concessions and licenses have attract- presents the greatest threat to remaining forests by ed almost as much criticism as the log-pricing issue, providing access to forested areas (Bryant, Nielsen, because of the effects on mismanagement of forests and Tingley 1997 and Dudley, Jeanrenaud, and and accelerated deforestation as well as the revenues Sullivan 1995). Roads are often essential both to forgone by the state. The issuance of 500 timber con- market farm outputs and to deliver inputs to the farm cessions in Thailand by 1968, covering half of the and can greatly increase the profitability of agricul- country, opened up vast new areas for encroachment tural and wood production in frontier areas. In many and official agricultural development (FAo 1993). countries one of the most common forms of assis- Substantial increases in the fees and charges set tance requested by farmers is the extension and by governments for the exploitation of natural forests upgrading of road transport systems. might reduce logging in inappropriate areas and Sometimes, however, roads do not create opportu- eliminate some of the least efficient operators from nities for deforestation; rather, the existence of those the industry, who may also be among the most opportunities leads to the construction of roads. If destructive. But their effectiveness for reducing inap- prospective farmers perceive their best opportunity to propriate deforestation could be mixed if they also be clearing some forest and converting the land to deterred sustainable forest management, so they another use, they can and will ask governments to would need to be carefully targeted. There would be construct roads to facilitate that. In this case, roads no direct costs; rather they would increase govern- facilitate deforestation, but they may not be its prin- ment revenues if they did not lead to massive clo- cipal cause. Thus, in analyzing the impact of public sures and reductions in volumes more than propor- road construction on forests, it is important to take tional to the price increases. Although the equity into consideration the reasons why specific roads implications could be very favorable, the political have been built. In some cases, the prime impetus is viability is uncertain at best. to provide access to timber for logging, in others it is As one form of indirect subsidy to local industry, for farming, and in still others (such as the Trans countries have operated a two-price scheme, where Amazonia or Trans Borneo highways) it is for mili- logs for local processing are considerably cheaper tary or "national development" reasons. David Kaimowitz, Neil Byron, and William Sunderlin 311 The road construction and transport policy reforms Many countries have claimed national forests, parks, that might be required to reduce deforestation by and reserves but have been unwilling or unable to increasing the costs of deforesting activities do not enforce such claims effectively. If governments were necessarily involve less road building, but rather a to enforce reserve boundaries more effectively and change in the nature, type, and location of roads built. were consistent in not recognizing "squatters" within In particular, existing road systems would have to be reserves, this could substantially increase the risks intensified, rather than expanding new roads into pre- and reduce the expected profitability of deforestation. viously remote forested areas. Such reforms could be Although such measures could be targeted, they may effective, targeted, and budget-neutral and could have be expensive, ineffective, and inequitable in develop- low indirect costs. However, because increased road ing countries vwith many poor landless people. Forest intensity could lead to capital appreciation of existing resource degradation has been made worse by restrict- farmlands, the effects may be regressive. In addition, ing peoples' rights to use land and neglecting their tra- because of the strong political demand for roads by ditional uses and their capacities to preserve forests. A landholders in remote areas and many governments' new approach that actively engages local people in interests in promoting colonization in those areas, decisions and management, and ensures that they are such reforms may not be politically palatable. not disadvantaged by forest conservation or produc- Although the topics of land use zoning and the tion activities, is currently being applied and evaluat- declaration and enforcement of protected areas are ed worldwide. vast and warrant many volumes (Wells, Brandon, and It is not difficult to conceive of a suite of policies Hannah 1992), we confine ourselves to summary that would retard or discourage deforestation of pro- comments. The rate of deforestation can be substan- tected forest areas by increasing the risks and conse- tially reduced where protected areas are effectively quences to those who deforest. However, there seem managed and have the support of surrounding popula- to be few bases; for optimism so long as the prevailing tions. However, where effective management and political tendency is to ignore deforestation and popular support do not exist, declaration of parks or encourage the expansion of alternative land uses. As national forests may actually accelerate deforestation. summarized in table 22.3, most such policy reforms Table 22.3. Policies That Increase the Costs and Risks of Production Activities Associated with Deforestation Political Policies Effectiveness Targetability Direct costs i'ndirect costs Equity viability Reduced subsidies for Moderate Moderate Saves money Low Favorable Very low certain agricultural inputs to high linked with expansion of farm areas Reduced support for Moderate High Saves money Low Favorable Very low colonization and to mixed settlement schemes Reduced technical and Low Low None None Unclear Low to advisory support for moderate activities on newly cleared lands Reduced road and Moderate Moderate Saves money Low Favorable Variable transportation subsidies to mixed Removal of subsidies to Moderate High Saves money Some Favorable Variable logging and forest industries to high possible doing destructive exploitation decline in exports Removal of tax and credit High Moderate Saves money Low Favorable Low subsidies for agriculture on newly cleared lands 312 Public Policies to Reduce Inappropriate Tropical Deforestation appear to be moderately effective, equitable, and tunities for capital gains from land sales. This could readily targeted and would reduce government expen- potentially be achieved through a combination of ditures. However, such subsidies currently exist land tenure, taxation, infrastructure, credit, and because many governments apparently see the expan- macroeconomic policies. sion of agriculture and (eventually) higher incomes Some Latin American governments have changed for farmers as desirable goals, irrespective of the con- their land titling and agrarian reform policies in an sequences for deforestation. attempt to decouple forest clearing from increased tenure security. Formally, they no longer require pub- Policies That Discourage Forest Clearing lic lands to be cleared for claimants to obtain or to Establish Property Rights maintain property rights. They now deny the right to request land titles to farmers who have inappropri- Many people clear forests not only to reap the bene- ately deforested the land after the new policies have fits of selling standing timber or producing crops and gone into effect (Sunderlin and Rodriguez 1996). livestock on the cleared land but also to gain proper- Land titling programs for humid tropical areas have ty rights over that land or to strengthen their existing attempted to provide secure property rights to exist- rights (Oz6rio de Almeida and Campari 1995, ing settlers, without obliging them to clear land for Angelsen 1995, and Mendelsohn 1994). This allows that purpose. Many Latin America countries require them to earn an implicit rent from the land and per- people who receive property rights over previously haps later sell the acquired land at a substantial prof- public lands in tropical forest areas to maintain a cer- it. In many Latin American countries, one condition tain percentage of that land in forest. Brazil, for for prospective landholders to obtain legal recogni- example, for many years required 50 percent of each tion of property rights over formerly public lands is rural property within Legal Amazonia to remain for- to demonstrate that they are "productively" occupy- est and now requires 60 percent (Burford 1991). ing those lands. They do so by clearing forest and The impact of these policies has ranged from planting pasture or crops. Deforested land is also bet- mildly effective to negative, with the main problems ter protected from expropriation by agrarian reform being weak implementation, forest clearing provoked agencies or invasion by squatters (Utting 1992). In by uncertainty surrounding the policy changes, and Southeast Asia and West Africa, farmers frequently the infusion of new migrants attracted by the chance clear forests to plant perennial crops such as cocoa or to obtain land titles. Governments find it somewhat rubber, partly to establish or improve their property expensive to implement these policies and hard to rights over that land (Angelsen 1995). Even when target them based on specific local conditions. Where land is private, forests often belong to governments. governments are unwilling or unable to enforce legal This gives landholders an additional incentive to property rights, land claimants keep clearing forests clear the forest to obtain broader property rights to ward off other potential claimants. Although there (Dorner and Thiesenhusen 1992) or greater legal are typically few strong objections at the national or recognition of de facto claims. international level, local elites usually oppose any In much of Africa, communal land tenure regimes attempts to restrict their ability to gain property rights and ethnically segmented land markets traditionally over deforested lands. As a result, the policies may be have limited the relevance of land speculation as a less equitable in practice than they are in theory. factor in deforestation, although perennial crops are Throughout the tropical world, there is increasing often planted to establish long-term usufruct rights interest in establishing or supporting existing com- (Dorner and Thiesenhusen 1992). This situation has mon property regimes for land and forests to protect slowly begun to change, however, as integration into indigenous and other forest-dependent people, to the global market economy and demographic pres- encourage sustainable forest management, and to limit sures have increased, and governments have sought the scope of land markets. The literature on the topic to establish private property regimes. rarely stresses the third aspect, but it is potentially To discourage the clearing of forests in order to quite important. gain or improve property rights, governments can use Common property regimes can effectively limit measures that break the link between forest clearing deforestation related to land speculation in certain and tenure, can tax nonforested (cleared) land more places. They are relatively cheap to establish legally heavily than forested land, or can reduce the oppor- where they already exist informally, and govern- David Kaimowitz, Neil Byron, and William Sunderlin 313 ments do not have to compensate other land public services gradually improve in agricultural fron- claimants. However, they are difficult to create where tier areas, real local land prices will rise and will informal common property regimes did not exist in encourage land speculation. There is ample evidence the past (Richards 1997). Policies that promote com- that similar processes are at work in Indonesia, the mon property regimes are equitable in the sense of Philippines, and Thailand. supporting groups that have traditionally been poor Certain government subsidies can increase land and marginalized, but they often involve ensuring prices and thus stimulate deforestation for speculative large amounts of resources to small numbers of indi- purposes where property rights can be secured or viduals. improved by clearing land (Binswanger 1994). Road Theoretically, land taxes could be used to discour- construction, in addition to facilitating access to age forest clearing by taxing cleared lands at a high- forested areas, raises land values and offers opportu- er rate than forested land. Land taxes and capital nities for capital gains (Angelsen 1995). Credit subsi- gains taxes could also potentially discourage land dies or tax incentives for agriculture or livestock that speculation by raising the cost of holding land as a are preferentially available to landholders also hedge against inflation or as a source of capital gains, increase the value of holding land. Govemment pur- charging landowners for the costs of infrastructure chases of land for protected areas, refugee resettle- improvement, and lowering the price of land (Strasma ment, infrastnicture construction, or other purposes and others 1987). However, experience to date has not often inflate local land markets. Removing these sub- been encouraging. Because few countries have used sidies can greaLtly reduce inappropriate deforestation, punitive land and capital gains taxes to discourage especially where high subsidies existed in the past. land speculation, it is hard to assess their practical Governments can actually save money by doing this, effectiveness. Most govemments have found it diffi- and there are iFew indirect costs. Because most such cult to administer such taxes effectively, because of subsidies presumably go to wealthy families and cor- the large amounts of information required and the porations, except perhaps for road construction, elim- high potential for evasion, or have avoided them, inating them tends to favor equity. Important interest because of potential political opposition (Skinner groups suppo:rt these subsidies, but others often 1991). oppose them. Because the purchase of land is a common hedge In summary, eliminating subsidies that promote against inflation, macroeconomic policies that reduce land speculation and strengthening either private prop- inflation may also limit the incentives for land specu- erty or common property regimes can sometimes be lation. This seems to be particularly relevant for cer- effective tools against inappropriate deforestation (see tain Latin American countries, such as Brazil, where table 22.4). Most other proposed policies that might high inflation rates were previously common but were limit forest clearing to gain secure tenure still have not not associated with general economic collapse. It is been tested fu]ly. Land taxation, in particular, merits also true, however, that as long as infrastructure greater consideration, given the notable gap between investments continue, regional markets grow, and its theoretical potential and practical difficulties. Most Table 22.4. Policies That Discourage Deforestation to Establish Property Rights Political Policies Effectiveness Targetability Direct costs Irdirect costs Equity viability Land titling policies Negative Moderate Moderate Low Unknown Moderate to low Support for common Moderate Moderate Low Low Mixed Variable property regimes to moderate Land and capital gains taxes Unknown Low High Low Unknown Low Fewer credit, tax, Moderate Moderate Saves money Low Favorable to Variable and road subsidies mixed Anti-inflationary Unknown Low Low Variable Mixed Moderate macro policies to high 314 Public Policies to Reduce Inappropriate Tropical Deforestation policies discussed in this section are easy to target to Thus although some market assessments have specific regions but are much more difficult to adapt to proposed that market forces will increase the poten- the conditions of individual properties. Generally, the tial timber revenues from retaining natural forests for more effective (and less purely symbolic) these poli- high-value products, we have seen little basis for cies are likely to be., the less politically viable they are, relying on that. The effect is likely to be too little, too although eliminating certain credit subsidies and tax late. Further, if timber concessionaires harbor doubts incentives is apparently an exception. about their long-term security of tenure, they are like- ly to continue to behave in a short-term exploitative Policies That Increase the Profitability way; even if they agree that tropical timber values of Maintaining Forests might soar in the future, they doubt that they will benefit from such price rises. Having considered many policy instruments that In the very short term, higher timber prices and might reduce the profitability of removing tropical more roads may encourage unsustainable practices forests, we turn now to possible interventions that but in the long term may be necessary for sustainable could increase the profitability of retaining them and practices to be financially viable. Short-term licenses managing them on a long-term sustainable basis. We do not encourage long-term stewardship of forests, consider three aspects in this regard: increased rev- but even 25-year licenses are less than many cutting enues from timber, increased revenues from nontim- cycles, destroying any motivation to nurture ber forest products, and the possibility of payments regrowth. "Cleaning operations" every 5 to 10 years for global environmental services. can be observed. Even with very long-term licenses, To the extent that excessive logging (both in area concessionaires in many tropical countries may and in degree of damage caused) contributes to defor- behave as if their tenure were insecure, because of estation, some recent efforts to reduce deforestation extreme uncertainty about political changes. In many have focused on policies to encourage forest man- countries, logging rights are not transferable, reduc- agement and logging practices on concessions and ing the incentive to maintain a high residual value of state forests that are less destructive and more sus- the forest enterprise "as a going concern." tainable (Bryant, Nielsen, and Tingley 1997 and Transferability of licenses might indirectly impose Dudley, Jeanrenaud, and Sullivan 1995). Analysts commercial penalties for owners who have not (Douglas 1995) have suggested that market forces looked after the forest under their control and might could provide a significant incentive for more sus- provide financial rewards (as higher transfer prices) tainable forest management, as the expected future for those who have not high-graded, who have pro- values of remaining commercial forests increase tected advanced growth, and who have installed good greatly and as accessible stocks of tropical forests are infrastructure. Systems like the "evergreen licenses" depleted. However, recent assessments (Byron and devised in British Columbia (Canada) with "claw- Ruiz Perez 1996 and Sayer and Byron 1996) of how back" provisions-providing strong commercial forestry practices may change in response to chang- incentives for long-term sustainable forest manage- ing markets for forest products conclude that "fron- ment-may warrant consideration and application in tier logging" of relatively remote areas in the tropics tropical developing countries. may become less important in the future. Once the In 1990 the International Tropical Timber forests in more accessible areas (close to roads, Organization (an agency of UNCTAD) reached agree- rivers, and ports) have been exploited, any cost ment between producer and consumer countries that advantage that these forests might have had will be by the year 2000, all tropical timber entering interna- rapidly eroded. Remaining natural forests will tional trade would come from sustainably managed become less able to compete with the outputs of the forests. Other recent initiatives (such as those of the rapidly expanding plantation sector in the tropics and Forest Stewardship Council) advocate policies that subtropics, except in specialty markets. In contrast to might make sustainable timber management more rising costs and declining quality of logs from natural profitable (through independent inspection and certi- forests, the volume and quality of plantation material fication of secure and longer-term concessions, will continue to improve, and technological advances accompanied by ecolabeling of products) to achieve in plantation silviculture and wood processing will premium prices or preferential market access for tim- continue to lower unit production costs. ber from certified, well-managed forests. Since 1990 David Kaimowitz, Neil Byron, and William Sunderlin 315 nongovernmental organizations (NGO) and, to a less- reserves in the Brazilian Amazon) to local NGO ini- er extent, governments have urged consumers to pur- tiatives to promote new tree species of economic chase only certified or ecolabeled tropical timber. potential in the highlands of northern Sumatra. They wish to exclude or minimize the sale of timber Numerous substantial policy and development initia- from poorly managed forests, while simultaneously tives are pursuing this proposition. Yet debate is rag- obtaining premium prices for certified timber. ing about the extent to which efforts to promote the The certification approach has generated much extraction, use, and marketing of nontimber forest debate: Who certifies? On what basis? Who controls products represent a real option for slowing defor- the chain of custody between the forest and the final estation and imnproving the livelihoods of forest peo- consumer to guarantee validity? How many con- ple. Dove (1993) has argued that incorporating non- sumers would pay what sort of premium for certified timber forest products into international markets will timber? Would this system be voluntary or compul- achieve neither socioeconomic nor environmental sory (and if compulsory, is it legal under the rules of objectives and may actually be counterproductive in the World Trade Organization)? In spite of the many both senses-at best it diverts attention from what he unresolved issues, polices by both governments and sees as the real causes of deforestation and degrada- NGOS to encourage and respect certification might tion and their political resolution. give some positive commercial rewards for sustain- The potential for these policy measures to slow able forest management. Although it is generally deforestation depends on whether one accepts that politically acceptable, it remains unclear how effec- forest people are destroying tropical forests because tive (or even relevant) certification will be in reduc- the stream of income from forests is too low (or ing inappropriate deforestation on a broad scale, because they simply do not understand the magnitude where conversion to other land uses poses the prima- of these benefits). Dove believes that forest destruc- ry threat to forests. tion is caused by loggers, migrants, and predatory To summarize, the policy interventions that might outsiders-that forest people are the victims rather reduce deforestation by increasing expected long- than the perpetrators of deforestation. Thus he argues term timber values rely on the identification and that efforts to halt forest destruction should focus on development of niche markets for tropical timbers- the economic behavior of the former, not of the for- small quantities of very high-value wood carefully est people. He asserts that increasing commercializa- removed from forests. Certification and ecolabeling tion of nontimber forest products will foster predato- are a special case of this approach that attempts to ry exploitation and destroy traditional management dissociate "sustainable forest management" from systems that rmight have helped to protect forests. exploitative logging that causes deforestation. The Recent research by the Center for International feasibility, effectiveness, and costs of this approach Forestry Research-cIFoR (Byron and Ruiz-Perez remain to be determined. If one of the major disin- 1996)-has shown how the distribution of property centives to sustainable forest management is insecu- rights and the ability (or otherwise) of local people to rity of tenure or fear of expropriation, removal of claim and enforce such rights are crucial determi- such disincentives would appear an attractive, even nants of the outcome of commercialization of non- necessary, policy option. But there is no evidence timber forest products. For a strong "community" of that this is a sufficient condition for slowing com- users and collectors of nontimber forest products mercial deforestation. with organization and political empowerment, com- Numerous international NGOS (such as Cultural mercialization of nontimber forest products can Survival Enterprises and Conservation International) enhance welfare and forest conservation. Conversely, have argued that finding new products in tropical the existence of a strong political elite whose mem- forests, finding new markets for new or existing bers may seek to expropriate the value of nontimber products, and marketing nontimber forest products forest products as soon as they realize that significant more effectively will make tropical forests much too potential rents exist may lead to serious social and valuable to destroy. Both government institutions and environmental exploitation. NGOS have incorporated such efforts into their poli- Policy proposals and development actions aimed cies and programmatic agendas, ranging from the at slowing deforestation by increasing the stream of World Bank pilot project for Brazil (with a signifi- revenues fromn nontimber forest products for forest cant component aimed at supporting four extractive peoples could be effective, targeted, equitable, low 316 Public Policies to Reduce Inappropriate Tropical Deforestation cost, and moderately politically acceptable to gov- ing material left in forests and increase the amount of emments of countries whose tropical forests are future carbon sequestration by healthy, rapidly growing inhabited, but only in specific institutional and legal immature forests. As yet, there is no legal obligation for contexts. power utilities (in the United States) to expend share- It has also been proposed that having beneficiaries holder funds in Central America or Malaysia to offset pay for the environmental services provided by trop- their U.S. emissions, nor are there tax credits at pres- ical forests could provide a market-based stimulus ent.8 Most studies have looked at the demand side for retaining forests and slowing deforestation. The based on the theoretical costs if companies have to proposed policy reform is to pay those who retain pay carbon taxes or if they have to pay to reduce their forests for the environmental services their actions emissions, which most do not, except in Denmark provide to the wider public. Analogous to the and Norway. upstream-downstream analyses in which beneficia- The Framework Convention on Climate Change ries compensate the providers of environmental has debated proposals for "international tradable car- externalities in a watershed or on a national scale, bon emissions permits" that appear to be an efficient there have been numerous proposals for capturing the and low-transaction costs approach to determining "option and existence values" of tropical forests, as who would emit and sequester carbon dioxide and perceived by affluent residents of industrial coun- where. Yet countless practical difficulties remain. In tries, as a way of internalizing some of the global the context of their efficacy in preventing tropical externalities associated with tropical deforestation. deforestation, it appears unlikely that payments The U.N. Convention on Biological Diversity calls (whether for carbon, biodiversity, or watershed pro- for further research on the equitable distribution of tection) would have much impact unless they actual- benefits and costs of forest conservation, which is ly went to those making the deforestation decisions seen as a prerequisite to devising suitable interna- on the ground, rather than to national governments. tional or intranational compensatory schemes. To Another recent notion is that industrial countries, date, this has focused on two main areas: biodiversi- or the private sector, or NGOS should offer developing ty prospecting and carbon sequestration. countries with tropical forests some "payment for The agreement between Merck Pharmaceuticals carbon storage." Apart from the practical difficulties and the Biodiversity Institute (InBio) of Costa Rica of paying local decisionmakers not to do something, for the commercial exploration and assessment of this would have to be considered not only for forests that country's biodiversity has attracted great public- but also for oil and coal. Basically, it would seem to ity. It appears to be an example of using commercial, violate the "polluter pays" principle, which has generally market-based transactions to make retaining tropical guided international environmental policy and financial forests more valuable than destroying them. Despite transfers. But these proposals also await the results of the publicity about the potential of "biodiversity research documenting and measuring the physical prospecting" to reward forest managers and indige- relationships, that is, measuring watershed outputs nous forest people for retaining forest biodiversity, it and soil erosion losses, quantifying biodiversity, or has recently been argued that the realistic potential of measuring the physical extent of carbon sequestra- such payments is very small, apart from a very few, tion, which are still far from complete. very small, niche markets (Simpson 1997). In summary, policies to encourage the retention of Similarly, the U.N. Framework Convention on tropical forests by capturing payments for the global Climate Change has authorized the so-called "activi- environmental services provided remain of question- ties jointly implemented" whereby carbon dioxide able effectiveness, targetability, and cost. So many (CO2) emitters (such as power utilities) in industrial unknowns remain that the whole area is decidedly countries pay for carbon sequestration in developing ambiguous (see table 22.5.) countries, as offsets for their CO2 emissions. To date, these offsets have taken the form of payments for the Policies That Increase the Opportunity Costs establishment of new forests (such as the Dutch of Labor and Capital Forests Absorbing Carbon Emissions-FACE- Foundation) or the payment to logging companies for We have argued that any policies that reduce the rel- the adoption of reduced-impact logging techniques, ative profitability of clearing forests, as compared to which can reduce the amount of damaged and decay- retaining them, tend to slow deforestation rates. David Kaimowitz, Neil Byron, and William Sunderlin 317 Table 22.5. Policies That Increase the Profitability of Maintaining Forests Political Policies Effectiveness Targetability Direct costs lIdirect costs Equity viability Niche marketing Moderate To be Undeter- Undeter- Moderate High in for tropical timbers, determined mined mined Europe, including certification doubtful in and ecolabeling exporting countries Abolition of disincentives Moderate High Low? Low; Possibly Variable for long-term sustainable inequitable management Greater security of Moderate High Low to Low Favorable Variable tenure for forest people to high moderate Development of markets Moderate High Low !Low Favorable High for nontimber forest products to low Creation of a system for Unknown Unclear Low Low to Undeter- Unclear transfer payments for moderate mined to date biodiversity conservation, carbon storage, and watershed management Because labor and capital are major factor inputs in clear forests for marginal agriculture, which is often forest conversion, but much less so in retention, temporary and unsustainable. The forest populations increasing the opportunity costs of labor and capital have been willing to clear land, to log illegally, and could be expected to reduce deforestation rates. to hunt and collect nontimber forest products to earn Policies that increase the opportunity costs of labor cash incomes. If the prices they receive are low, they and capital include general policies that promote harvest more in order to earn the required incomes! growth as well as specific policies that promote In cities and towns, high-technology industries, employment. In parts of East and South Asia (Korea tourism, and the service sector are booming as pros- and Malaysia), the opportunity costs of labor and perous enclaves. This rapid economic growth in urban, capital have clearly increased, and deforestation nonland-based sectors is already having profound seems to have slowed and even been reversed. In impacts on forests and deforestation rates. Although other places where alternative returns to labor and rapid growth in agriculture, timber industries, and capital remain low (Cambodia and Laos), deforesta- plantation estates may contribute to deforestation, tion rates remain high. This suggests possible con- employment growth in sectors like tourism, textiles, nections between these costs and the extent of inap- and manufacturing has increased labor opportunity propriate deforestation. costs to such an extent that it has reduced deforesta- Even within Southeast and East Asian countries, tion. However, where the deforesters have been left there appears to be socioeconomic stratification and out of that growth, and no more attractive livelihoods differentiation: remote, poor, rural forest villages are have emerged, deforestation has continued. The evi- being left out of the economic miracle. Most have dence to date is circumstantial and anecdotal, but the experienced very low income growth, if any; people policy implications could be that the promotion of sus- who have little formal education, limited access to or tainable, nonrural employment options may reduce the control over valuable resources, and poor transport types of deforestation caused by poor people trying to and communications infrastructure struggle to main- earn a meager living from deforesting. Similarly, tain traditional living standards. One of the options where deforestation is driven by the interests of large for those remaining in the villages and forests is to capital, the implication could be that with the emer- 318 Public Policies to Reduce Inappropriate Tropical Deforestation gence of more lucrative options (through the develop- instruments with regard to stopping deforestation- ment of capital markets), these parties might lose inter- governments will be forced to choose from a mix of est in deforestation (see table 22.6). measures specifically crafted to local conditions. This diagnosis of who is clearing forests, where, There is no magic policy reform that will slow all and why, and the observation that where more attrac- types of tropical deforestation. Most panaceas that tive alternative employment options have emerged, have been advocated have been revealed as merely people (both large-scale and peasants) have voluntar- partial solutions applicable under limited conditions. ily moved out of forest-degrading activities suggest a It is crucial to resolve the interrelations between possible policy measure. Economic growth, with rea- deforestation, logging, road construction, and con- sonable equity of access to new industries and version to alternative land uses. Logging will lead to resources, may make certain types of deforestation in increased forest conversion by an influx of mnigrants, particular contexts sufficiently unremunerative as to if the following conditions apply simultaneously: slow it voluntarily. The pursuit of equitable and sus- * Roads are constructed that open up new areas of tainable economic development is a stated goal of forests. most developing countries, so it would be fortuitous * The nonforest land use is much more profitable indeed if that process inherently encouraged forest than the retention of forests (in part due to the pol- conservation. Unfortunately, there is little evidence icy distortions described). that the solution is so simple. In terms of the criteria * Forest boundaries are poorly enforced by govern- used here, policies to promote industrialization and ment agencies (the forest service or national parks employment generation in nonland-based sectors are service) and an institutional or legal context exists already on the political agenda; although they might in which people can expect that the land they be effective in reducing deforestation, and equitable occupy, claim, or "stake out" will eventually be if they could be achieved, relatively few countries recognized or even legalized by the government, have achieved this. Indeed, promoting sustainable for example, by creating an open-access frontier. economic growth and employment creation has been * A large pool of people who are unemployed or more elusive than policies to slow deforestation landless or who have very low incomes and directly but, in the long run, might be the eventual prospects constitute potential migrants (whether solution. moving spontaneously or sponsored by the state or private agoindustries). We hypothesize that the Conclusions pace of colonization might be related to the dif- ference between current incomes of potential There is no perfect or generalizable policy for reduc- migrants and the amount they expect to earn by ing inappropriate deforestation. There are no "first- colonizing forest areas (hence the importance of best" options. Each national situation is different, the opportunity costs of labor). much uncertainty remains about key cause-and-effect All of the instances, of which we are aware, of relations, and there are usually tradeoffs among rapid forest clearance by "squatters" appear to be effectiveness, ability to be targeted, political viabili- consistent with this hypothesis. Conversely, where ty, and direct and indirect costs of policies. Most of these conditions do not apply (which admittedly is the policy instruments discussed here are very blunt still rare in tropical developing countries), forests can Table 22.6. Policies That Increase the Opportunity Costs of Labor and Capital Used in Deforestation Political Policies Effectiveness Targetability Direct costs Indirect costs Equity viability Increased urban Moderate Low Ambiguous Low; high Ambiguous High employment and potential wages indirect gains Improved capital markets Low Low Low Low Unclear High David Kaimowitz, Neil Byron, and William Sunderlin 319 be logged but subsequently remain under manage- while collecting more of the potential economic ment and not be cleared or seriously degraded (in rents to remove the "gold-rush" mentality that has peninsular Malaysia). characterized the industry for much of the past 40 This assessment, if correct, suggests that the years. answer to the forest conversion issue is not necessar- * Develop innovative institutional arrangements for ily to stop logging per se, or to stop logging in all devolving more decisionmaking authority and new areas, or to ban all new road construction in for- responsibility to those whose livelihoods and quality est areas, but rather to reform those policies and insti- of life are linked directly to the extent and quality of tutions that at present make forest colonization seem tropical forests. In association with this, both pro- more attractive than the potential migrants' current duction and conservation activities (for the full activities. This might include the pull factors (to range of consumptive and nonconsumptive goods reduce the profitability of illegally clearing forests or and services from forests) can be fostered, or at of speculating in land that was supposed to be kept as least not discouraged, from both government forest) or the push factors (to increase the limited forests and private property outside of govern- livelihood options outside of forests). The evidence ment forest reserves. from the rapid economic growth of Asian tiger * Encourage voluntary market differentiation by economies is that as employment and income consumers that discriminates positively toward prospects outside the agricultural sector improve, those products that have been sustainably pro- fewer people want to undertake the dangerous, ille- duced from forests and penalizes those products gal, difficult, and often unprofitable activities of tem- that have contributed to deforestation. This could porary agriculture in forestlands. However, if the new be applied not only to tropical timber and to non- land use is very profitable (growing cocoa, coffee, timber forest products like Brazil nuts, but also to cinnamon, rubber or fruit trees, or even timber trees other products such as chocolate, coffee, and like eucalyptus, teak, or gmelina) and if the potential leather. capital gains from "capturing" some real estate from * Facilitate the recognition and compensation for the govemment forests are high, it might be very dif- environmental services provided by forests-at ficult to slow the rate of forest conversion by such global, national, and local scales-through mech- people. The case for intensification of agriculture as a anisms linked directly to the amount of environ- primary instrument for reducing deforestation has been mental services generated (whether carbon cogently summarized by Waggoner. (1995).9 sequestration or biodiversity conservation or Thus the principal govemment policy reforms that watershed protection) and ensure that transfer could reduce inappropriate tropical deforestation are payments are received by those making the deci- likely to be: sions at the forest frontiers. i Eliminate subsidies to agricultural and pastoral Most of the major environmental problems in industries that encourage deforestation and, if developing countries are due not to the pursuit of necessary, offset these by schemes targeted at economic development but rather to incorrect eco- intensification of existing agricultural areas. This nomic policies, such as poorly defined property also applies to road construction; that is, intensifi- rights, underpricing of resources, state allocations cation of existing rural road networks rather than and subsidies, and neglect of nonmarketed social expansion into hinterlands to open up new areas benefits. Instead of trying to devise new policies to (unless they are specifically zoned for conversion stop further resource and environmental deterioration from forest to agriculture). while promoting real development, we should first * Eliminate existing legal and institutional incen- eliminate those (legal, social, political, and institu- tives or requirements to clear forests as a basis for tional) factors that cause or exacerbate the problems. gaining recognized land tenure and to reward "squatting" or colonization of protected forest Notes reserves. * Reform forest industry concessions and licenses 1. These figures are derived from the data in FAO to provide incentives for long-term sustainable (1996: table 4.2.13). They are based on the F3 defnition management, through increasing the duration and of "forest," which includes closed, open, long fallow, security of licenses and making them tradable, and fragmented forests. 320 Public Policies to Reduce Inappropriate Tropical Deforestation 2. The official target of the International Union for do either. the Conservation of Nature (IUCN) is for at least 10 per- 9. Waggoner (1995) concludes that if during the next cent of each major forest ecological zone to be given 60-70 years the world farmer could reach the average protected area status. However, Iremonger and others yield of today's average U.S. corn grower (8 tons of (1997) conclude that of 138 regional and ecoregional grain equivalent per hectare), then 10 billion peo- zones worldwide, only in insular Southeast Asia has this ple-twice the world's current population-would target been achieved. need only half the current area of cropland, while eat- 3. This is not to imply that the areas of highest bio- ing at today's average calorie levels in the United diversity value are necessarily remote and hilly sites States. undesirable for agriculture (or vice versa). Indeed, con- flicts arise because the high biodiversity areas frequent- References ly are the remaining forest areas of high agricultural potential. The word "processed" describes informally repro- 4. Ruf and Siswoputranto (1995) analyze the eco- duced works that may not be commonly available nomics of growing cocoa on newly cleared forestlands, through libraries. relative to existing farmlands. They find that newly Anderson, R. J., L. F. Constantino, and N. M. Kishor. cleared lands have a consistent price advantage of 1995. "Stabilization, Structural Adjustment, and $300-$400 per hectare (the "natural fertility subsidy"). Bolivia's Forestry Exports." Environment This means that if world prices fall, the supply from Division Dissemination Note 13. Latin America growers on existing farmlands contracts first, so clear- Technical Department, World Bank, Washington, ing new forest becomes the only viable option as prices D.C. Processed. fall. Angelsen, Arild. 1995. "The Emergence of Private 5. There are numerous examples, covering vast areas Property Rights in Traditional Agriculture: around the world, where forests have been logged but Theories and a Study from Sumatra." Paper pre- are still reasonably productive and ecologically stable. sented at the fifth common property conference: Thus it is overly simplistic, and incorrect, to equate reinventing the commons, International Asso- commercial logging with deforestation. Industrial log- ciation for the Study of Common Property, Bodo, ging is most likely to lead to deforestation when associ- Norway, 24-28 May. Processed. ated with spontaneous small-scale settlement and con- Barbier, Edward B., J. C. Burgess, J. Bishop, and B. version or with large-scale agroindustry conversion. Aylward. 1994. The Economics of the Tropical 6. In fact, log export bans are rarely intended as Timber Trade. London: Earthscan Publishers. conservation measures, but rather as part of a pack- Bilsborrow, Richard, and M. Geores. 1994. age to encourage accelerated domestic processing, "Population, Land-Use, and the Environment in with employment generation, value added, and, it is Developing Countries: What Can We Learn from hoped, higher net foreign exchange earnings. Cross-National Data?" In D. Pearce and K. Brown, 7. Manasan (1989) has observed that although eds., The Causes of Tropical Deforestation: The Southeast Asian governments justified their direct Economic and Statistical Analysis of Factors and indirect subsidies by the expectation that the Giving Rise to the Loss of Tropical Forests, pp. expansion of forest industries would generate jobs, 106-33. London: University College London Press. only the Philippines offered a direct financial incen- Binswanger, Hans. 1994. "Brazilian Policies That tive to employment, via tax rebates. Yet even this was Encourage Deforestation in the Amazon." World ineffective because it was neutralized by other tax Development 19(7):821-29. concessions, which promoted capital-for-labor sub- Bryant, D., D. Nielsen, and L. Tingley. 1997. The Last stitution. Overall, relatively few jobs have been cre- Frontier Forests: Ecosystems and Economies on the ated despite the substantial subsidies, and the impact Edge. Washington, D.C.: World Resources Institute. on forests has been severe (Douglas 1995 and Buongiorno, J., and Togu Manurung. 1992. "Predicted Lindsay 1989). Effects of an Import Tax in the European Com- 8. Although spending $1 million on offsets and munity on Intemational Trade in Tropical Timbers." tree planting in tropical developing countries is Journal of World Forest Resource Management cheaper than spending $10 million on controlling 6:117-37. emissions at home, there are no legal obligations to Burford, Nicolette. 1991. "Land Scarcity in Amazonia: David Kaimowitz, Ne,l Byron, and William Sunderlin 321 Migrant Farmers' Access to Land and Forest Pacheco. 1]996. "The Effects of Structural Conservation in the State of Rond6nia." M.Sc. the- Adjustment on Deforestation and Forest Degra- sis, Oxford Forestry Institute, Oxford. dation in Lowland Bolivia." Center for International Byron, R. N., and M. Ruiz Perez. 1996. "What Future Forestry Research, Bogor, Indonesia. Processed. for the Tropical Moist Forests 25 Years Hence." Lindsay, H. R. 1989. "Indonesian Forestry Policy: An Commonwealth Forestry Review 75(2):124-29. Economic Analysis." Bulletin of Indonesian Capistrano, Ann D. 1990. "Macro-Economic Economic Studies 25(1):99-115. Influences on Tropical Forest Depletion: A Cross- Manasan, R. C3. 1989. "A Review of Investment Country Analysis." Ph.D. diss. University of Incentives in ASEAN Countries." Working Paper Florida, Gainesville. 88-27. Phillipines Institute of Development de los Angeles, M., and R. Md Idrus. 1990. "Investment Studies, Manila. Processed. Incentives of Southeast Asian Countries in the Mendelsohn, R. 1994. "Property Rights and Tropical Wood-based Sector." In Proceedings of the Deforestation." Oxford Economic Papers 46 Malaysian Timber Industry Board Asia-Pacific (October):750-56. Conference, Kuala Lumpur, 24-26 September1990. Mercer, D. E., and J. Soussan. 1992. "Fuelwood Kuala Lumpur: Malaysian Timber Industry Board. Problems and Solutions." In N. P. Sharma, ed., Domer, P., and W. C. Thiesenhusen. 1992. Land Tenure Managing the World's Forests, Looking for and Deforestation: Interactions and Environmental Balance between Conservation and Development, Implications. Discussion Paper 34. Geneva: United pp. 177-214. Dubuque, Iowa: Kendall/Hunt Pub- Nations Research Institute for Social Development. lishing Company. Douglas, J. J. 1995. "The Economics of Long-Term Munasinghe, Mohan, and Wilfrido Cruz. 1994. Management of Indonesia's Natural Forest." World "Economy-wide Policies and the Environment." Bank, Jakarta. Processed. Environment Paper 10. Environment Department, Dove, M. R. 1993. "A Revisionist View of Tropical World Bank, Washington, D.C. Processed. Deforestation and Development." Environmental Myers, Norman. 1981. "The Hamburger Connection: Conservation 20:1-17. How Central America's Forests Became North Dudley, N., J. P. Jeanrenaud, and F. Sullivan. 1995. America's liamburgers." Ambio 10(1):3-8. Bad Harvest: The Timber Trade and the Degradation Oz6rio de Almeida, A. L., and J. S. Campari. 1995. of the World's Forests. London: Earthscan. Sustainable Settlement in the Brazilian Amazon. FAO (Food and Agriculture Organization of the New York: Oxford University Press. United Nations). 1993. Forestry Policies of Richards, M. 1997. "What Future for Forest Selected Countries in Asia and the Pacific. FAO Management Based on Common Property Resource Forestry Paper 115. Rome. Institutions in Latin America?" Development and -. 1996. Forest Resources Assessment 1990: Change 28(1):95-118. Survey of Tropical Forest Cover and Study of Ruf, F., and P. S. Siswoputranto, eds. 1995. Cocoa Change Processes. FAO Forestry Paper 130. Cycles: The Economics of Cocoa Supply. Rome. Cambridge, U.K.: Woodhead Publishing. Garrity, D. P., D. M. Kumar, and E. S. Guiang. 1993. Sayer, J. A., and R. N. Byron. 1996. "Technological "The Philippines." In Sustainable Agriculture and Advance and the Conservation of Forest the Environment in the Humid Tropics, pp. Resources." Journal of World Ecology and Sus- 549-624. Washington, D.C.: National Academy tainable Development 3(3):43-53. Press. Simpson, R. D. 1997. "Biodiversity Prospecting: Iremonger, S., V. Kapos, J. Rhind, and R. Luxmore. Shopping the Wild Is Not the Key to Conservation." 1997. "A Global Overview of Forest Conservation." Resources 126:12-15. Paper prepared for the world forestry conference, Skinner, J. 1991. "If Agricultural Land Taxation Is So Turkey, October. Processed. Efficient, Why Is It So Rarely Used?" The World Kaimowitz, David. 1997. Livestock and Deforestation Bank Economic Review 5(1):113-33. in Central America in the 1980s and 1990s: A Stern, D. I., M. S. Common, and Edward B. Barbier. Policy Perspective. Bogor, Indonesia: Center for 1996. "Economic Growth and Environmental International Forestry Research. Degradation: The Environmental Kuznets Curve Kaimowitz, David, Graham Thiele, and Pablo and Sustainable Development." World Devel- 322 Public Policies to Reduce Inappropriate Tropical Deforestation opment 24(7): 1141-160. Nations Research Institute for Social Devel- Strasma, J. D., J. Alm, E. Shearer, and A. Waldstein. opment. 1987. "Impact of Agricultural Land Revenue Waggoner, Paul E. 1995. "How Much Land Can Ten Systems on Agricultural Land Usage." Associates Billion People Spare for Nature? Does Technology in Rural Development, Burlington, Vt. Processed. Make a Difference?" Technology in Society Sunderlin, W., and J. Rodriguez. 1996. Cattle, 17(1):17-34. Broadleaf Forests, and the Agricultural Wells, Michael, Katrina Brandon, and Lee Hannah. Modernization Law of Honduras, The Case of 1992. People and Parks: Linking Protected Area Olancho. CIFOR Occasional Paper 7. Bogor: Management with Local Communities. Washing- Center for International Forestry Research. ton, D.C.: World Bank, World Wide Fund for Tchoungui, R., and others. 1995. "Structural Nature, and U.S. Agency for International Adjustment and Sustainable Development in Development. Cameroon." 'Working Paper 83. Overseas Wiebelt, M. 1994. "Protecting Brazil's Tropical Development Institute, London. Processed. Forest, A CGE Analysis of Macroeconomic, Utting, P. 1992. Trees, People, and Power: Social Sectoral, and Regional Policies." Kiel Working Dimensions of Deforestation and Forest Paper 638. Kiel: Kiel Institute of World Protection in Central America. Geneva: United Economics. 23 Costs, Benefits, and Farmer Adoption of Agroforestry Dean Current, Ernst Lutz, and Sara Scherr In Central America and the Caribbean, where the are used for fencing; coffee and cocoa plants are cul- empirical part of our work was done, the forest cover tivated for shade; and fallow (shifting cultivation) in each of the eight countries studied declined systems are maintained that sequentially rotate trees between 10 and 24 percent during the 1980s (table with crops. 23.1). In the two countries with the highest popula- Agroforestry systems introduce the deliberate use tion density (El Salvador and Haiti), only 6.2 and 1.3 of woody perennials (trees, shrubs, palms, bamboos) percent of forest cover, respectively, remained by on the same land management unit as agricultural 1990. crops, pastures, and animals. This may consist of spa- Deforestation means increasing scarcity of tial arrangement in the same place at the same time or wood-a principal source of fuel and construction a sequence over time; it may include woodlots, materials in rural areas-and contributes to environ- taungya, and rnanaged woody fallows for forest tree mental degradation of fragile agricultural lands. One production. For farm areas devoted to perennial response of governments, backed by nongovernmen- crops, systems include deliberate perennial inter- tal organizations (NGOs) and international donors, has crops and horne gardens. The main systems intro- been to encourage agroforestry: a land use system in duced for annual cropland are trees planted in lines which trees, shrubs, palms, and bamboos are culti- on boundaries, alley cropping of annual crops vated on the same land as agricultural crops or live- between tree rows, windbreaks, barrier hedges along stock for economic and environmental reasons: to the contours of fields, and trees planted on terrace ris- replenish wood stocks while upgrading land through ers. diminishing erosion, water loss, and other natural phe- Although agroforestry activities has attracted nomena. The system includes managing natural solid support and are likely to receive further sub- regrowth; seeding, planting, and maintaining trees as stantial investnent from public agencies, NGOS, and border plantings; and interplanting them in agricul- international donors, they have generally been poor- tural crops, in woodlots, in home gardens, or other ly documented. Monitoring of projects, where it systems. Agroforestry reinforces traditional refor- exists, has been weak or focused narrowly on admin- estation efforts by reducing the pressure on remain- istrative concerns (Scherr and Muller 1991 and ing natural forests. Current, Rivas, and G6mez 1994). Traditional agroforestry practices already play an Most research and analysis has dealt with the important role in many farning systems-fruit trees physical and biological aspects of agroforestry sys- are planted in home gardens, pastures, and fields and tems, neglecting systematic analysis of agroforestry's among permanent tree crops; live shrubs and bushes economic contributions at the farm level (Swinkels 323 324 Costs, Benefits, and Farmer Adoption ofAgroforestry Table 23.1. Key Indicators in Case Study Countries Population Reforestation as density Rate of GNP per Percentage Rate of a percentage (population population capita Forest cover Forest cover change in deforestation, of area per square growth (1992 U.S. in 1990 per capita forest cover 1981-90 deforested Country kilometer) (percent) dollars) (percent) (hectares) 1981-90 (percent)' per year' Costa Rica 62.7 2.8 1,960 28.8 0.5 -23.8 2.9 7.5 Dominican Republic 149.0 2.1 1,050 22.5 0.2 -24.2 2.8 1.1 El Salvador 257.1 1.4 1,170 6.2 0.0b -16.8 2.2 16.1 Guatemala 89.0 2.9 980 39.3 0.5 -15.6 1.7 3.1 Haiti - - - 1.3 0.Ob -10.3 4.8 73.3 Honduras 48.2 3.3 580 41.2 0.9 -19.4 2.1 0.4 Nicaragua 30.0 2.7 340 50.8 1.6 -16.8 1.9 1.5 Panama 32.5 2.1 2,420 41.1 1.3 -17.0 1.9 0.8 Total 77.3 - - 36.9 - -18.1 - 2.3 Latin America and the Caribbean 24.2 1.9 2,180 56.2 2.3 - 7.1 0.8 5.0 -Not available. a. Deforestation is defined as the change of land use from forest to other land use or depletion of forest crown cover to less than 10 percent. b. Extremely small values. Source: World Bank 1994, BID 1993, and FAO 1993. and Scherr 1991 and Scherr and Muller 1991), even Central America and the Caribbean (see figure 23.1) though the absence of a profit motive is a root cause and drew together the lessons learned for the design of many project failures (Tschinkel 1987 and Murray and execution of future efforts. Consultants from the 1991). The appropriateness of particular agroforestry region and staff at the Tropical Agricultural Research systems to subregional economic conditions is rarely and Higher Education Center (CATIE) in Costa Rica assessed (Scherr 1992). Nor has there been much carried out eight country studies on the basis of an ear- comparison of different extension approaches and lier pilot study (Current and Lutz 1992) under the their influence on farmer adoption of agroforestry overall direction of the World Bank, with methodolog- (Vergara and MacDicken 1990). ical support from the International Food Policy Important questions remain about which strate- Research Institute. Each study included a country gies to pursue for agroforestry development. Which overview and case studies of two or three agroforestry agroforestry practices, under different environmental projects, analyzing in detail the technical interventions and economic conditions, are profitable for farmers? and their adoption, their economic performance at the Under what conditions are farmers willing to invest farm level, the project's approach to extension, and in agroforestry systems oriented to environmental policies that influenced implementation of the project rehabilitation? Do certain conditions justify subsidies and its adoption by farmers. National workshops and a and nonmarket incentives? What level and type of regional workshop were held for farmers, extension- field extension assistance are needed to disseminate ists, project managers, and decisionmakers, with rep- new agroforestry species and information and to resentatives from farmer organizations, NGOS, national encourage adoption? What are the principal con- agencies, and international donors to review prelimi- straints at the farm level to further adoption of agro- nary findings. To our knowledge, this was the first forestry? What kinds of policy action would encour- large-scale study of farm-level profitability of agro- age adoption? forestry in the tropics. To answer these questions, this study analyzed the The specific objectives of the study were to: experiences of 21 projects in eight countries of * Document earlier agroforestry efforts and their Dean Current, Ernst Lutz, and Sara Scherr 325 Figure 23.1. Approximate Location of Case Studies 1)OMInkan , § wdf ~Hondutas Salv3dor w 1 a*a Central -* -Costa America lessons for tree planting on farms Caribbean to promote communal and individual tree • Determine costs and benefits of agroforestry sys- plantings and agroforestry systems. Agroforestry tems from the farmers' perspective development i:n the region generally reflected-and * Determine the advantages and disadvantages for in some cases led-international trends in thinking farmers of different types of agroforestry systems and action about agroforestry. In the late 1970s and under different site and resource conditions early 1980s when the emphasis in forestry was on * Identify the effects of institutional factors and fuelwood production, agroforestry systems were used government policies on the adoption of agro- on farms for planting trees for fuelwood (Eckholm forestry systems. 1975 and King 1979). Fuelwood projects were imple- This chapter summarizes the principal findings of mented worldwvide (U.S. Comptroller General 1982 the country studies, comparing them where possible and USAID/ROCAP 1979), and research was undertak- with the experience of other regions. Beginning with en on fuelwood species (National Academy of a brief background history of agroforestry in the Sciences 1980). But the response to fuelwood proj- region, the article describes the investigative meth- ects in rural communities was discouraging: farmers ods used, outlines the characteristics of the projects were not enthusiastic about contributing the sizable reviewed, and discusses the findings on household investment in labor-for planting, felling, sectioning, benefits, the pros and cons of different extension and splitting--needed to produce a product with a approaches, and the effects of different policies. relatively low market value. Furthermore, the fact that fuelwood has traditionally been collected as a Evolution of Agroforestry in Central America free good (even on private lands) lessened its attrac- and the Caribbean tion as a product for farmers. In El Salvador, with its high population density and low forest cover, a recent Numerous projects have been established in the past national study showed that only 33 percent of the two decades throughout Central America and the rural population considered the acquisition of fuel- 326 Costs, Benefits, and Farmer Adoption ofAgroforestry wood a problem (Current and Juarez 1992). species was accompanied by a corresponding realiza- Subsequently, the emphasis shifted from fuel- tion that a project is more likely to succeed if project wood production to the use of multipurpose tree planners consult local communities about their per- species in agroforestry systems. The establishment of ceived needs and design projects to meet those needs multipurpose tree species on farmland can provide a instead of imposing schemes that communities may wide range of benefits. For farm families, they may not consider a priority.' In line with this thinking, help to meet household consumption needs, provide research focused on traditional agroforestry systems a source of income, or improve land husbandry to (Budowski 1987), and many forestry agencies have ensure future production. As well as fuelwood, mul- been changing their emphasis from "guarding" the tipurpose tree species supply green manure, wood for forest to helping communities produce tree plantings construction, fence posts, fuel, raw materials for local that, in turn, may help to protect the remaining forest. or industrial processing, and cash income and savings This stronger community role in the design of projects from any of these. Conflicts do arise between the cul- and technology has parallels in Africa and Asia tivation of trees and agricultural crops, but many (Kerkhof 1990, Budd and others 1990, French and agroforestry systems allow farmers to integrate trees van den Beldt 1994, and Raintree 1991). into their farming systems, with only small draw- backs for crop production or even an increase in farm Methods of the Study productivity overall (Gregersen, Draper, and Elz 1989 and Raintree 1991). The study compared the economic and technical per- Social benefits from establishing trees on farms formance of the most prevalent agroforestry systems can be significant and are often cited to justify tree used in the 21 projects. planting projects supported by national governments and donors. Trees can provide protection from soil Analysis of Household Benefits erosion generated by water and wind, protecting future productive potential and preserving important The analysis identified the role-both intended and watersheds that feed hydroelectric projects and supply actual-of trees in the household's livelihood strate- water for population centers; they provide wood prod- gy: to satisfy household needs, to generate cash ucts for the farm, thus relieving the pressure on natu- income or savings, or to supply insurance. It also ral forest areas (Current and Lutz 1992); they provide assessed the effects of the agroforestry practice on raw materials for rural industries that generate labor, land, and other household resources (including employment for rural communities (Arriola and household risks) and used cost-benefit and related Herrera 1991 and Campos, Rodriguez, and Ugalde analyses to assess the profitability of the agroforestry 1991); and they provide environmental and social practice relative to farmers' alternative systems. This benefits, such as bird habitats, water retention capaci- was done using sensitivity analysis to determine the ty, or shade for dwellings. range of conditions (such as input or output prices, Regionally, the most important applied research productivity levels) under which an agroforestry initiative has been the Tree Crop Production technology was likely to be profitable (Scherr 1995). (Madelenia) Project financed by the U.S. Agency for The data on inputs and outputs and costs and International Development (USAID) Regional Office prices required to develop the multiyear agroforestry for Central America and Panama and executed by enterprise budgets were generated in various ways, CATIE (Belaunde and Rivas 1993). Other internation- among them recall data and plot measures from single- ally important research programs include those of the visit farm surveys, multivisit studies of case study Forestry Fuelwood Research and Development farms, and results of on-farm research trials in the Project in Asia (French and van den Beldt 1994), the project area. Costs and benefits were valued at the International Center for Research in Agroforestry prices that these agents actually face, with no attempt (Steppler and Nair 1987), and the Nitrogen Fixing to adjust for economic distortions. Farmers were Tree Association (Nitrogen Fixing Tree Association interviewed about the means and the opportunity 1993). Numerous national and international NGOS costs of using land, labor, and capital resources in have also undertaken applied research and extension agroforestry. They also provided qualitative assess- activities (Cook and Grut 1989 and Nair 1990). ments and anecdotal information on social and envi- The shift in emphasis toward multipurpose tree ronmental costs and benefits; quantitative informa- Dean Currentt, Ernst Lutz, and Sara Scherr 327 Table 23.2. Farming and Agroforestry Systems in the Case Study Projects Agroforestry Principal tree Project and country systems promoted species used Costa Rica Community Agricultural Center: Hojancha A,E,H 5,21,31 PRODAF Forestry /Agroforestry Development Project A,E,H,L 5,11,17 COOPELDOS-windbreaks for coffee and dairy products F 9,14,15,19 Traditional Agroforestry Systems of the Humid Tropics A,C,L 13,17,20 El Salvador Agroforestry Support to Low-Income Rural Communities D,E,G,H,I 18,20,24,31 Community and Family Nursery Program D,E,G,H 8,18,31 Guatemala Agroforestry Program of DIGEBOS, CARE, Peace Corps, and DIGEBOS/CATIE Madelefia 3A,C,D,G,L 3,9,17,18,20,22,25 La Maquina Settlement-Guatemala E,G,H 6,18 Honduras Development of the Broadleaved Forest (BLFDP) A,C,E,G 20,23,29 COHASA Rural Development B,E,H 20,24 Sierra de Omoa Management Unit A,I 13,20 Land Use Planning (LUPE) C,D,E,H 13,20 Nicaragua Control of Erosion in the West and Reforestation F 18,24,30 Support to Farm Forestry and Rehabilitation of Windbreaks E,F,G,H 18,20,24,30 Demonstration Farm "La Esperanza" C,E,L 7,8,16,20,27 Panama Agroforestry for Community Development (INRENARE/CARE) B,E,F,G,H 2,18,24,26 Food Production and Community Development-Aquaculture (MIDA/WFP) H 24,26,31 Dominican Republic ENDA-Caribe: Desarrollo Rural Integrado de Zambrana C,G 2,7 Fondo Rotativo Agroforestal y Comercial (FORESTA) C,E,H 2,17,18,20 Haiti Agroforestry Outreach Project-Pan American Development Foundation D,E,H 4,8,10,18,24 Maissade Watershed Management Project-Save the Children E 1,8,12,28 Code Agroforestry systems 5 Bombacopsis quinatum (Eucalyptus) A Perennial crop with trees (Pochote-C. Amer.) 19 Eucalyptus saligna (Eucalyptus) B Tree intercropped with agriculture 6 Caesalpinia velutina 20 Gliricidia sepium (Mother of cocoa) C Alley cropping (Aripin-C. Amer.) 21 Gmelina arborea (Gmelina) D Contour plantings 7 Calliandra calothyrsus (Calliandra) 22 Grevillea robusta (Silk oak) E Border plantings and live fences 8 Cassia siamecz (Yellow cassia) 23 Inga vera (Guaba) F Windbreak 9 Casuarina cuninghamiana 24 Leucaena leucocephala (Leucaena) G Taungya (Casuarina) 25 Melia azederach (Chinaberry) H Woodlot 10 Catalpa longissima (Catalpa) 26 Pinus caribaea (Caribbean pine) I Fa'low I I Cedrela odorata (Spanish cedar) 27 Pitehcolobiutn saman (Raintree) J Home garden 12 Colubrina arborescens 28 Simaruba glauca K Fodder trees (Coffee colubrina) (Aceituno-C. Amer.) L Trees in pasture (silvopastoral) 13 Cordia alliodora (Onion cordia) 29 Swietenia macrophyla 14 Croton niveus (Colpachi -C. Amer.) (Honduras mahogany) Code Principal tree species (common name) 15 Cupressus lusitanica 30 Tecoma stans 1 Acacia auriculiformis (Mexican cypress) (Yellow trumpet flower) (Japanese acacia) 16 Enterolobium cyclocarpum 31 Tectona grandis (Teak) 2 Acacia mangium (Mangium) (Earpod tree) 3 Alnus acuminata (Alder) 17 Erythrina sp. (Mountain immortelle) 4 Azadirachta indica (Neem) 18 Eucalyptus camaldulensis Table 23.3. Project Extension Approaches Extension tools Information and training Marketing, Incentives Alumber of Number of processing, Field trips, Cash Country Number of families technical Extension Food for or Demonstration Local training Material incentive and project communities participating staff visits work infrastructure plots Radio promoters courses inputs or credit Costa Rica Traditional 3 280 None 0 0 0 0 0 0 0 0 0 Community 53 300 2 0 0 0 * 0 * 0 0 o Agricultural Center PRODAF 69 144 5 *0 0 0 0 * 0 COOPELDOS 19 119 1.5 0 0 o 0 0 0 0 0 0 Dominican Republic ENDA-Caribe 66 800 5 0 0 0 0 0 0 0 0 0 FORESTA - 89 5 0 0 0 0 0 0 0 0 S Guatemala La Maquina - - - 0 0 0 0 0 0 0 0 0 DIGEBOS/CARE 350 10,000 86 0 0 * 0 0 0 0 0 0 Madeleia 3 22 1,309 22 0 0 0 0 0 0 0 0 0 Haiti Agroforestry Outreach 6' - 200 a 0 0 0 0 0 0 0 0 NGOS Maissade Watershed Associations 2,400 12-15 *0 0 0 0 0 0 0 0 Honduras Broadleaf forest 10 656 32 0 0 0 0 0 0 0 0 0 COHASA 6 1,500 5 (-) 0 0 0 o o o * * LUPE 44 5,427 193 (45) 0 0 0 0 0 0 0 0 0 Omoa 7 109 3 0 0 0 0 0 0 0 0 o Nicaragua Erosion - 3,000 30 + 600b 0 0 0 0 0 0 0 0 0 Windbreaks 16 500 20 0 0 0 0 0 0 0 0 0 La Esperanza 11 765 3 * 0 0 0 0 * 0 0 0 Panama INRENARE/CARE 15 (10) 157 (80) (10) * 0 0 0 0 0 0 0 0 MIDA/WFP 11 districts 4,380 (600) (57) 0 0 0 0 0 0 0 0 0 El Salvador Agroforestry support 100 3,302 20 0 0 0 0 0 0 0 0 0 Nursery program - 22,470 - 0 0 0 0 0 0 0 0 0 - Not available. Note: Numbers in parentheses represent the first phases of the projects. Numbers outside the parentheses represent the totals to 1992. a. Associations. b. Workers were hired only during the establishment phase. Source: World Bank data. 330 Costs, Benefits, and Farmer Adoption ofAgroforestry tion was generally not available. the focus on smallholders (average farm size was less The study deliberately emphasized household than 20 hectares in most projects and less than 5 benefits because, since farmers make their decisions hectares in five projects). Together, the projects cost on land use in light of their own objectives, produc- approximately $150 million and represented more tion possibilities, and constraints, understanding their than 50,000 farm households; they covered a variety incentives is essential to understanding patterns of of agroforestry systems and site and socioeconomic resource use and formulating appropriate responses to conditions (see table 23.2). problems. Ultimately, the success of any agroforestry Table 23.3 summarizes key elements of the exten- system depends on how fanners, not project planners, sion strategies used by the projects. Such strategies view its costs and benefits. Furthermore, the fann- can generally be divided into four broad categories: level approach places the focus firmly on the effects on using existing forestry or agricultural extension net- farm productivity, which are particularly important in works; training and providing logistical support to this region. (This is not to minimize the importance, in extension agents hired specifically for the project; some situations, of off-farm effects.) integrating project extension agents with national extension agencies; and supporting extension staff Analysis of Technology Adoption organized and managed by NGOS. The projects stud- ied represented all four of these categories. The most Researchers collected information on indigenous common methods used to provide information and agroforestry practices, farmiing systems, and land use training to the farmers were visits by extension problems as well as the farmers' criteria for selecting agents or paratechnicians, field trips and training agroforestry interventions. Characteristics of tech- courses, and demonstration plots. Six projects pro- nology adoption (the systems selected and the extent, vided marketing assistance. Most projects used mate- durability, and determinants of adoption) were rial incentives (free seedlings or other inputs), and assessed through informal farmer surveys. Formal many others used cash incentives, credit, or food-for- and informal surveys of farmers, focus groups, and work incentives. (Food for work has been used for investigations by project staff identified specific communal and privately owned nurseries, as well as problems, and researchers assessed the experience of for tree planting on both communal and private lands. the project with the selection of tree species, tree- Plant survival rates have been much higher in com- crop interactions, and methods of establishing trees. munity nurseries where trees are planted on private land and for private tree plantings.) Policy Assessment Findings: Incentives for Adopting Agroforestry, The study identified policy influences on agro- Adoption Patterns, and Implications forestry adoption, project implementation, and the for Extension profitability of agroforestry from several sources. Country overviews drew on public documents, key The success or failure of the effort to promote agro- informants, and unpublished literature. Project case forestry depends ultimately on how many farmers studies used interviews with farmers and project staff take up the system and persevere with it. Perceived as well as project reports to gain policy insights and profitability for the household is a critical influence to document constraints caused by specific legal, on a farmer's initial decision to include agroforestry institutional, and policy problems. National and in the farm's activities; the systems adopted and the regional workshops provided a forum for project scale and pace of adoption will then be determined by executives and policymakers to develop specific rec- practical matters, such as the resources available to farm- ommendations for policy action. ers and the specific constraints they face. All of these have important implications for the design of exten- Characteristics of the Projects Reviewed sion efforts. The 21 projects analyzed were selected on the basis Incentives for Adopting Agroforestry of the age of the project (older projects being pre- ferred), the availability of economic data, the will- The study assessed the profitability of agroforestry at ingness of project staff to cooperate in the study, and the household level in light of financial benefits and Dean Current, Ernst Lutz, and Sara Scherr 331 Table 23.4. Averages of Indicators by Agroforestry System Number of Cost- Average Agroforestry systems Net present value benefit Return Payback person-days system studied 10 percent 20 percent ratio to labor period per year Trees with 5 2,863 1,300 1.8 5.03 3.4 165 agricultural crops Alley cropping 9 1,335 847 2.1 6.98 1.9 56 Contour planting 4 1,426 761 1.6 3.75 2.0 116 Perennial with trees 4 2,867 1,405 1.8 4.19 4.0 139 Home garden 4 - 372 2.2 Taungya 8 6,797 2,868 2.5 15.98 4.9 53 Woodlot 10 764 -33 1.0 3.41 9.2 12 -Not available. Note: This table should be used for general comparison only. The systems analyzed represent only a small sample of each system and cannot be considered a random representative sample. Net present values and return to labor are influ- enced by localized market conditions and do not fairly represent differences between planting systems. The payback peri- ods and average man-days required per year can more reliably be compared. (Current, Lutz, and Scherr 1995 present this information by each individual system analyzed.) Source: Current, Lutz, and Scherr 1995. costs, risk management, and the nature and accessi- With most situations experiencing reductions bility of markets for agroforestry products. in access to off-farm supplies, growing demand, declining site productivity, and Financial results. Financial analyses were under- increased exposure to risk, it should not be sur- taken for the major agroforestry systems used in the prising that tree-planting activity does increase region-trees intercropped with agricultural crops, as agriculture and land use become more inten- alley cropping, contour plantings, perennial crops sive. Though it would be incorrect to assume with interplanted trees, home gardens, taungya, and that this alwa;ys happens, there is a general pro- woodlots (see Current, Lutz, and Scherr 1995, for gression towards more planted trees as agricul- detailed results).2 Table 23.4 shows averages of the ture and pressures on land intensify, and exist- financial indicators calculated for these systems and ing tree stocks diminish, within most systems. net present values for each technology-that is, all A system's financial value may derive from the costs and benefits at the plot level estimated at 10 and value of the tree products it provides (woodlots, 20 percent real discount rates and compared to the net taungya, timber species in coffee and cacao); the pro- present values for one or more alternative practices. tection and subsequent increase in agricultural pro- This approach was adopted because the systems duction it affords (windbreaks, shade trees in peren- often differed considerably. The net present values nial crops); its contribution to soil improvement were found to be positive at a 20 percent discount through nitrogen fixation, organic matter, and rate in all but one (woodlots) of the seven systems improved soil structures; or a combination of these. analyzed. The outputs most commonly produced in the eight Local scarcities of timber, fuelwood, and poles, countries and seven systems analyzed were fuel- which differ not only from country to country but wood, charcoal, and roundwood products for local from region to region within countries, are critical to markets and household use. In Costa Rica and the the economics and adoption of agroforestry. The Dominican Republic, sawlogs were major products. higher prices and correspondingly higher profits that Prices for tree products were highly variable often accompany scarcity can be an important incen- across the region and within each country. Fuelwood tive for farmers to turn to agroforestry.3 This finding prices provide an example of this variability and the is consistent with those from qualitative assessments effect it may have on a system's profitability: accord- and empirical case studies from other parts of the ing to our estimates, fuelwood prices ranged from $2 world. Summarizing results for South Asia and East per cubic meter in Panama to almost $6 in Africa, Arnold (1995: 272-73) concludes that Guatemala. In Guatemala City, fuelwood prices were 332 Costs, Benefits, and Farner Adoption ofAgroforestry more than twice the price of the same fuelwood in with two to five years for the other systems. Thus the other parts of the country. In contrast, despite a rela- long gestation period commonly assumed for tree- tively low forest cover around Le6n, Nicaragua, a based systems does not seem to be a problem for cubic meter of fuelwood cost only $2.14, one of the most of the agroforestry practices used in this region. lowest prices in the region. This apparent contradic- Of course, absolute profitability is less important tion can probably best be explained by differences in to a farmer's decisionmaking than the profitability the employment situation, the amount of fuelwood relative to the available alternatives. Comparing the marketing, and relative access to the tree resource. financial returns for agroforestry with those for alter- One of the social benefits of windbreaks in Nicaragua native farming practices, the case studies found that is that they serve as a source of fuelwood, providing agroforestry systems produced significantly higher gainful (though illegal) employment for the surround- returns per hectare-at least 25 percent-in approxi- ing communities and creating a competitive market mately 40 percent of the cases and at least 10 percent that maintains prices at a low level. greater returns in more than half the cases. In more Financial analyses for all of the seven practices stud- than 75 percent of the cases, the returns to labor were ied showed average net present values of more than higher for systems including agroforestry than the $300 per hectare (the exception was woodlots). local agricultural wage. Findings on profitable returns from taungya and from Financial results were generally most sensitive to intercropping of multipurpose tree species in perennial yields and product prices, followed in importance by crops confirmed findings of earlier regional case studies labor; the results were least sensitive to the prices of of the same region. The on-farn profitability of agro- purchased inputs such as seedlings, agricultural seed, forestry in Central America appears to be consistently and agrochemicals. Including trees in agricultural high, under a wide range of conditions and for different systems makes tree, crop, and livestock systems less systems. Perhaps the reason for this high profitability is sensitive to changes in yields and prices and that farmers modify systems to ensure profitability improves overall profitability. under their own farm and local conditions. Returns for perennial intercrops of coffee and Risk management. Adopting agroforestry systems black pepper were almost all high; low prices for can both reduce and increase the risks to the security plantain and cacao reduced the returns to those sys- of a household's livelihood. Income diversification is tems in Costa Rica. On cropland, net present values a strongly positive benefit that reduces the risks for for alley cropping and trees on contours were high cropping systems, whether annual or perennial, and except in Nicaragua. Analysis of the alley cropping for livestock farming. But risks are associated with systems used conservative assumptions about the the cultivation of trees, as with any other crop. medium- and long-term effects on yield of green Unusual drought conditions or poor planting materi- manuring and nitrogen fixation; high returns were als can mean failure to establish the trees; uncertain often due to the added value of the wood product. market prospects-most projects were weak in Except in Costa Rica, windbreaks brought low developing or improving farmers' access to tree prod- returns because of the high costs of establishment uct markets-increase risk; badly managed trees, or and the suppression of yields in adjacent crops. In the overly competitive tree species, can destroy associat- case of government-sponsored windbreaks in Nic- ed crops; and where permit procedures and govern- aragua, they were valued only for their fuelwood, ment regulations are restrictive, there is a risk of not although more valuable products had been harvested being able to harvest trees-a major obstacle to ini- from them, and this valuation underestimated the tial adoption. These risks, and perceived risks, reduce value of the windbreak. In Costa Rica, increased the rates of adoption and expansion, even where yields of coffee and milk production alone made the average financial returns are good. establishment of windbreaks profitable in three years. Typically the benefits received from the trees off- (Windbreaks protect pasture grass from desiccation set reported losses in agricultural production, but data and reduce the stress on cows.) on the impacts of trees on agricultural crops are Cost-benefit ratios were above 1.6 for six of the scarce. Positive effects of trees on crops were report- seven agroforestry systems, with woodlots once more ed for agroforestry systems that produced organic the exception. Payback periods were quite long for mulch and green manure or wind protection. This the woodlot systems, averaging nine years compared was true of trees (Alnus) planted on contours in Dean Current, Ernst Lutz, and Sara Scherr 333 Guatemala, Leucaena intercropped with corn in Hon- market has had :o adapt to the normal vicissitudes of duras, and windbreaks in Costa Rica. In Honduras, supply and demand; as new nurseries have developed the financial analyses reflected a decline in agricul- outside Hojanchta, the demand and prices for certain tural production in the alternative systems without species have declined, so that some nurseries have trees. had to sell plants at a loss, while others have special- ized their product to suit industrial growers who have Markets. Most farmers in the study initially adopt- special requirements for their plants. ed agroforestry systems to meet household subsis- Studies by Cook and Grut (1989) and Warner tence needs, a result consistent with the findings of (1993) for Africa and the collection edited by household surveys in six countries in Asia in 1991, Raintree and Francisco (1994) for Asia and by Arnold reported by the Forestry/Fuelwood Research and (1995) for Africa and Asia show the importance of Development Project (French and van den Beldt local markets in providing incentives for agro- 1994). In 12 of the projects, a significant part of total forestry. Nevertheless, they still conclude that house- output continued to be reserved for use on the farm. hold use is the driving force for most adoption. Once subsistence needs were satisfied, farmers Arnold (1995: 278) further argues that distinctions became interested in market opportunities; in seven between the two are blurred: projects, local markets were important, and for 12 In most of the situations studied self-sufficien- projects, considerable output was directed to region- cy, in particular tree products, proved to be the al or urban markets (usually through middlemen who primary objective. However, with the growing purchased locally). dependence of farm households on income to The market for farm-grown tree products varies meet at least part of their needs, as forest prod- within the region. In El Salvador and on the Pacific ucts such as fuelwood, fodder, and fruits coast of Guatemala, markets have developed for rural become progressively commoditized, and the construction; poles and roundwood have also found a marketplace provides opportunities to substi- market in Guatemala for tobacco-drying racks. In La tute purchased inputs such as fertilizer for Maquina, Guatemala, when the roundwood market inputs previously supplied by growing trees, for construction became saturated, the market for the distinction between production for subsis- tobacco racks picked up the slack. But farmers in tence or sale has progressively less meaning. Costa Rica who had been led to believe that they Not only will producers sell what is surplus to would be able to sell products from thinnings (wood their subsistence needs, but they will sell a three to five years old) encountered a limited market commodity needed in the household if the that quickly became saturated, leaving many without opportunity cost of doing so is advantageous. an outlet. The market for sawlogs is strong in Costa The role of local and regional woodfuel markets Rica but requires wood 11 to 15 years old. Similar in encouraging farmers to adopt agroforestry appears problems with saturation in certain markets have also to be more important in much of Central America and arisen in Asia. Asia than in Af-rica, possibly because farming sys- Local organizations are working to develop new tems are more highly commercialized in Central local, national, and international markets-notably in America and the Caribbean than in much of Africa, Guanacaste, Costa Rica, where tree farmers have and the region is more integrated economically. Thus formed a regional association of local organizations farmers may be more inclined to specialize in partic- and cooperatives that is studying markets for a ular tree species and tree products; the multiproduct regional forest industry based on their new and exist- feature of agrofiarestry, which appears to be so high- ing plantations. On the Pacific coast of Guatemala, ly valued in Africa and Asia (Raintree and Francisco tree farmers recently formed a similar organization to 1994), is generally less valued in Central America. explore markets for their tree products. In Hojancha, Costa Rica, early work to improve Adoption Patterns nursery practices for some of the most commonly planted species engendered a nursery industry that Potential profitability may be a key incentive in now supplies plants for individual and industrial tree farmers' decisions to adopt agroforestry, but the via- planting throughout the country and provides a bility of a system for a particular household or area source of employment and income for Hojancha. The will be determined by the resources available to them 334 Costs, Benefits, and Fanner Adoption ofAgroforestry and the constraints they face. taungya or block plantings, followed by trees in some kind of linear arrangement (border plantings, contour Management factors. Land, labor, and capital plantings, live fences). Evidence of recent adoption resources are as influential as profit considerations in was mixed for alley cropping, home gardens, wind- farmers' decisions about agroforestry (Dewees 1993 breaks, green manuring, and tree-pasture systems. and Scherr 1992). The resources available to the fann The choice of particular tree species for use in agro- family limit the type of agroforestry systems that can forestry technologies was dictated by considerations be adopted, because different systems require differ- such as familiarity to farmers, growth performance, ent resources. Home gardens and alley cropping, for market value, ease of propagation and management, instance, may be very labor-intensive, whereas tim- multiple-use options, interaction with crops (see ber trees or woodlots require little labor but much table 23.5), availability, project recommendations, land. and environmental effects. The study identified 32 An advantage of agroforestry is that tree manage- different forest tree species promoted by at least one ment operations can often be scheduled for periods of project, along with a large number of well-known or slack labor demand; indeed, the case studies suggest- exotic fruit trees. The most frequently used species ed that this is a significant incentive for adoption. In were Eucalyptus camaldulensis, Gliricida spp., and El Salvador, fuelwood prices follow the agricultural Leucaena leucocephala. Tectonis grandis (teak), calendar: during slack agricultural periods, farmers Acacia mangium, and Cassia sp. were also men- dedicate more time to extracting and processing fuel- tioned by at least four projects (see table 23.2 for wood, and the increase in supply is reflected in a sea- common names). Projects had tended in the past to sonal decline in fuelwood prices (Current and Juarez limit the options they provided to introduced species. 1992). One constraint on the adoption of alley crop- More recently, there was a shift in the opposite direc- ping is that it needs labor for pruning, which fre- tion, with some projects offering only native species. quently conflicts with major agricultural operations. A mix of species appeared to be desirable, from the Intensive agroforestry systems, such as alley crop- farmers' point of view. ping, ale less attractive to farmers with available land for fallowing. Scale and pace of adoption. Most empirical stud- Management of agroforestry ranges in scope and ies have found that farmers tend to adopt agroforestry sophistication from simple gathering to active man- gradually, and in the projects studied here, significant agement of naturally growing resources, domestica- adoption commonly required 5 to 10 years. This tion and establishment of trees in selected sites, would seem an especially wise course of action for expansion in the scale of tree planting, and intensive smallholders on resource-poor farms, given the need management (Scherr 1992). The level of intensity to ensure food security and reduce risks. Continuous, that is economic for farmers depends on factors such small-scale adoption seems a more appropriate as scarcity of forest resources, level and composition objective than crash planting programs for small- of demand, returns to alternative uses of land, labor, holders. Studies documenting cases of rapid adoption and capital resources, and the costs of production rel- (for example, of eucalyptus in India) have found ative to the costs of available substitutes for tree poorer economic returns. Several theoretical cost- products and services (bricks for building poles, benefit modeling exercises also show that enterprises kerosene for fuelwood). The Panama project demon- are more profitable when costs and risks of tree strates that a fallow system for recuperating soils and establishment are spread over several years, obtain- the extraction of forest products are a very profitable ing early benefits from multiproduct systems and a traditional alternative where land is available for that more regular flow of final products over time type of management. The study found the highest (Swinkels and Scherr 1991; selected cases). intensity of agroforestry management-for example, Farm size clearly influences the pace and scale of trimming branches to limit shading of agricultural adoption: even within the limited range of resource- crops-in areas of greatest land shortage, such as El poor farms targeted by the project, researchers found Salvador and the Guatemalan highlands. that adoption was greater on larger farms. Farmers selected and adapted agroforestry practices and man- Agroforestry practices and tree species. The most agement to suit their particular labor, land, and capi- prevalent agroforestry systems found on farms were tal resources. Female participation was significant in Dean Current, Ernst Lutz, and Sara Scherr 335 Table 23.5. Interaction of Agroforestry Systems with Other Crops Agroforestry system Positive tree-crop interactions Negative tree-crop interactions Perennial crop with trees Complementary Competes for water and nutrients; shade Trees intercropped Complementary Competes for water and nutrients; shade with agricultural crops Alley cropping Complementary Competes for water and nutrients; shade Contour planting Increases production by providing Competes for space, water, and nutrients protection from soil erosion and contributing organic material Windbreak Increases production by providing Competes for space, water, and nutrients protection from winds Taungya Crop production helps to cover Shade and competition eliminate crop production the cost of plantation establishment in three to four years Woodlot Isolates trees from crops, avoiding Competes for crop production area; poor matching potential negative tree-crop interactions of site-species and management has led to soil erosion Home garden Generally does not compete for crop No tree-crop interface, so little or no interaction production area; provides for tree product and fruit production on the same area Note: Choice of species and management may determine whether interaction is positive or negative and may limit or enhance the magnitude of the impact (positive or negative). In a complementary interaction, trees shade crops so that inputs improve tree growth, and trees provide organic material and nutrients to crops. In a competetive interaction, agricultural production decreases as a result of competition for sunlight, water, and nutrients, and tree harvesting may iamage perennials. Source: Authors' data. several projects. Information from four projects in El farmers are reluctant to plant trees because they fear Salvador and Guatemala indicates that the average possible expropriation of the land or future inability to number of trees planted ranged between 440 on farms harvest trees without permits. Most projects have with an average size of 2.2 hectares (or 200 trees per developed mechianisms to overcome this, such as con- hectare) and 1,160 on farms with an average of 10.4 tracts and written agreements giving the farmer hectares (or 112 trees per hectare). Although the explicit rights to harvest trees, but such agreements farmers with more land planted more trees in total, are often not applicable nationally. Farmers without farmers with less land tended to plant more trees per clear land rights-renters, squatters, or residents of hectare and dedicated a greater percentage of their land that is de jure under forest reserve or protec- total land area to trees. Although farm size may affect tion-face substantial obstacles to agroforestry. adoption, it was not a limiting factor. For example, Lack of formal land tenure generally is not a bind- extension monitoring data from El Salvador show ing constraint to adoption (Cook and Grut 1989 and that 40 percent of the farmers participating in a com- Godoy 1992), except in certain situations in Africa, munity nursery program owned less than 1 hectare of where constraints to women on tree planting or diffi- land. culties in protecting trees established in fields used by pastoralists during the fallow period may be a deter- Effects of land tenure and land use regulation. rent. By contrast, tree ownership and disposal rights, Harvesting laws and regulations, created to protect particularly the restrictions and insecurities created by forests, are a significant barrier to the adoption of state claims of control over harvesting, have been par- agroforestry in many places. In several countries, ticularly problematic in dry areas in West Africa 336 Costs, Benefits, and Farmer Adoption ofAgroforestry (Shepherd 1992) as well as in Central America. Terms Information on available options. Projects should of sharecropping arrangements also appear to be a provide farmers with information on resource constraint in parts of Asia (Arnold 1995). requirements and performance of a variety of agro- forestry designs and species rather than with standard Implications for Extension designs. This would require promotion programs to broaden the basket of choice of species and systems The forces that influence the spread of agroforestry, to provide detailed information and performance and the patterns of adoption that have emerged, under different types of management. Farmers could should be a central concern in designing or modify- then select those that best meet household needs and ing projects, particularly the amount of extension resources and modify the basic design to suit their assistance offered and the form it should take. own preferences. Sometimes tree planting on farms has been promoted irrespective of the profitability of the agroforestry Demonstration, technical assistance, and train- system or how it fits into management of the house- ing. Agroforestry development is clearly occurring in hold and farm resource. Of course, there is room for parts of the region as a natural response to economic extension and supply-driven innovation to act as a incentives and subsistence needs. Qualitative infor- catalyst by promulgating new species or agroforestry mation from farmers, project staff, and informed systems, for instance through demonstration plots or observers suggests that the demonstrated benefits of farms. But extension and technical assistance need to fast-growing tree species and systems often have be highly responsive to farmers' concerns and be able been a strong impetus to the expansion of agro- to provide a range of options and to outline their pros forestry activities (see box 23.1). Nonetheless, even and cons. They should not push tree planting as where agroforestry systems are demonstrably prof- something that is intrinsically good. itable, some technical assistance is needed to facili- Box 23.1. The La Miquina Project in Guatemala: A Case of Participatory Field Research with Demand-Driven Diffusion of Technology La Maquina is a farming area of approximately 345 aging multipurpose tree species. Between 1986 and square kilometers on the Pacific coastal plain of 1991, about 238,000 seedlings of multipurpose tree Guatemala. Activities were initiated there in 1983 species were provided to 550 farmers. under the Lena Project, a regional project with core La Maquina's success is principally attributable to professional staff located at CATIE in Costa Rica, oper- three aspects of the project: the demonstration effect, the ating in participating countries through a national coor- quality of the product, and the evolution of incentives. dinator working with a counterpart from the forestry * The demonstration effect meant that dissemination agency. Under their direction, two or three profession- was driven initially by demand, followed later by als assisted with t]he activities, along with support and directed extension activities to support the effort. field personnel provided by the forestry agency. * As the product of a research project, planting stock The principal activity was to establish silvicultural was reliable and technical assistance in planting experiments to evaluate the performance of species in and management helped to guarantee quality. different systems. Rather than being restricted to exper- Incentives were modified as the project developed. imental fields, research plots were established on farm- To launch tree planting, free seedlings were given to ers' lands by agreement with the owners, and the local farmers and were often also transported to the farm. population was involved in the research effort. In 1987 Once the community became interested, farmers the work was expanded to include plots on demonstra- arrived at the nursery to request seedlings and tion farms managed by the farmers under various agro- received technical assistance from the staff. Later, forestry systems. Detailed information was collected seedlings were provided in exchange for an equiva- on the inputs and returns for both agricultural and lent number of plastic bags. Most recently, the project forestry/agroforestry plots. and a new effort in an adjacent community began pro- A nursery was also established in the area, managed moting family nurseries. Thus the project gradually by two resident technicians who also provided techni- shifted the costs and responsibilities for tree produc- cal assistance to farmers interested in planting or man- tion to the farmers. Dean Current, Ernst Lutz, and Sara Scherr 337 tate adoption and to provide information about tree the perceived risks of new planting and the need for management for unfamiliar species and configura- out-of-pocket cash expenditures. Food-for-work tions. The process of diffusion is slow, in part incentives, by contrast, were very problematic, com- because farmers take a long time to evaluate peren- monly leading to dependency, failure to maintain sys- nial systems and in part because they are frequently tems once they were established and incentives were unfamiliar with product markets at the farm or group no longer given, and reluctance to undertake subse- level. Technical assistance can accelerate diffusion. quent agroforestry efforts without payment. This can be phased out after the technology has been transferred, although, as the La Maquina example Credit for agroforestry. Evidence from the study indicates, continued applied research to produce the suggests that formal lines of credit may not be suit- best tree species and to discover which conditions able for promoting smallholder agroforestry. Poor suit them best remains very important even after farmers seldom use formal production credit, because public extension on agroforestry has been phased of the risks involved, and when they do, they may out. reserve it for othier priorities. Credit is mainly needed Training and employing local people as paratech- to cover the costs of hired labor for planting relative- nicians was judged to be a successful, low-cost ly large acreages-an activity not relevant for small- means of promoting technology in nine of the proj- holders. Where credit is made available, integrated or ects. Paratechnicians could generally turn to profes- nontargeted systems (rather than separate lines of sional staff for technical expertise. Use of paratech- credit for specific crops or trees) would benefit farm- nicians develops local human resources and the ers most. capacity to continue diffusing agroforestry beyond the project period. Social and Environmental Benefits Provision of planting materials. Of the variety of Quantitative data on the social (as distinguished from strategies for providing germ plasm-direct seeding private) benefits provided by agroforestry systems are of native and introduced species, family nurseries, sparse. But there is some information and anecdotal community nurseries, and specialized centralized evidence that these projects are making social and nurseries producing containerized seedlings in "root environmental contributions that could justify further trainers" for easier transport-community and family promotion for that purpose. Agroforestry systems are nurseries proved superior to centralized nurseries, replacing less stable systems in watersheds, protect- when backed by good technical assistance and logis- ing the remaining forest resource by providing an tical support. Farmers in many projects indicated a alternative supply of tree products, protecting cities willingness to purchase seedlings at market prices. from the effects of airborne dust, and providing sources of employment and income generation for Financial and material incentives. The case stud- rural communities, thus helping to slow outmigration. ies suggested that financial incentives and subsidies These benefits, of course, generally do not accrue to may not, in fact, be needed for adoption, at least for the farmer and therefore tend not to figure prominent- the size and duration found. The important objective ly in adoption decisions. As in Central America, Cook for extension is to put forward agroforestry technolo- and Grut (1989) found for Africa that environmental gies that will be profitable for the farmer without benefits, although appreciated by farmers, are not suf- subsidies. There may be an argument for providing ficient to spur adoption, except in areas of intensive time-limited financial incentives to early adopters of production where yields are declining. unfamiliar technologies, to be reduced and eventual- Some specific examples of environmental and ly eliminated once farmers are aware of the benefits. social benefits were noted in Central America: Small, in-kind material inputs (seedlings, water- * Cacao plantings with tree shade introduced in ing cans, planting bags, and fertilizer), which were fragile watersheds in Choloma, Honduras, are generally provided with extension services, proved replacing shifting agricultural systems with peren- widely effective in persuading farmers to try agro- nial tree crops that protect and build soil. In forestry. The advantage of these incentives was not Choluteca, llonduras, communal and individual so much that they had an effect, which was often plantings combined with soil conservation mea- quite modest, on profitability but that they reduced sures are providing a similar benefit. In addition, 338 Costs, Benefits, and Farmer Adoption ofAgroforestry in most areas where projects have been operating, Ugalde 1991). The increased employment oppor- the burning of agricultural fields and pastures has tunity is an advantage for Hojancha, which is a declined. region whose outmigration adds people to already * In Honduras, El Salvador, and Guatemala, tree overburdened urban slums. Another social benefit planting is providing an alternative source of tree derives from the fact that agroforestry activities products and thus reducing the higher social costs, are often undertaken in periods when the demand such as loss of biodiversity and increased down- for labor is low. Recent efforts to process small stream flooding, entailed in "mining" a natural diameters and explore markets for tree resources forest. Data from three projects demonstrated that will provide further possibilities for employment in Honduras, farmers who depended on the natural and income generation. forest for up to 100 percent of their tree products In order to get a better empirical handle on social are now meeting the bulk of that demand from and environmental impacts, much more data will have their own tree plantings. In El Salvador, a com- to be collected than are currently available. More mon- munity close to the capital city, which used to pro- itoring of projects is needed in general, and that moni- duce charcoal for sale from natural vegetation, is toring should include data collection to allow social now using tree plantings for that purpose. In and environmental impact hypotheses to be tested. Guatemala, tree plantings on the Pacific coast are being used as posts for an expanding tobacco Conclusions industry instead of the remaining mangrove for- est, which would have been the most probable The attitude of the individual farmer to agroforestry source of that material. is crucial to the success of agroforestry projects. • In Costa Rica, the Monteverde Conservation Farmers' perception of the role that the system will League has for several years promoted the estab- play in their farm's production system and of its lishment of windbreaks, which have improved the costs, benefits, and profitability will determine the productivity of dairy and coffee for participating extent and durability of adoption and should guide farmers. A recent study by researchers from Duke project strategies on extension and institutional and University has demonstrated that those wind- policy issues. breaks are acting as biological corridors connect- ing remnant forest patches. In addition, because The Profitability of Agroforestry and Its Role their productivity has increased as a result of their in the Farming System participation with the Monteverde Conservation League, participating farmers are more receptive Agroforestry, in many ways, is comparable to other to protecting and improving their natural forest parts of the farming system. Its special characteris- and expanding the width of the windbreaks to tics are that it includes a large number of species, serve environmental purposes. configurations, and management intensities, has * In Nicaragua, extensive windbreaks were estab- longer gestation than most agricultural crops, and lished to protect Le6n from dust carried into the produces outputs with multiple uses. city by winds. Reportedly, the windbreaks have In addition to the estimated financial return, farm- reduced the dust and, with it, the health costs for ers attach considerable importance to how an agro- the population. forestry system fits into the overall farm production * Finally, as this study demonstrates, these projects system and the existing land, labor, and capital con- improved rural welfare by providing needed tree straints. Even more important for a farmer's decision products and generating income and employment. than the absolute profitability of agroforestry systems A documented example is the effort of the Centro may be the returns relative to alternative options. The Agrfcola Cantonal in Hojancha, Costa Rica. most profitable agroforestry systems sometimes have According to a report prepared in 1990, the com- the most market risks. Some marginally profitable mercial nurseries in Hojancha that developed systems are widely used to meet specific household from the reforestation efforts generated 30,000 subsistence needs. Once those are met, there is an person-days of employment in 1989, and seed col- interest in market opportunities. The existence or lection from plantations provided an additional development of markets is crucial for expansion (see 4,900 person-days (Campos, Rodriguez, and similar findings in Alvarez-Brylla, Lazos-Chavero, Dean Current, Ernst Lutz, and Sara Scherr 339 and Garcia-Barrios 1989 for Mexico; Mussak and agroforestry systems may not yet be economic. For Laarman 1989 for Ecuador; French and van den smallholders, in contrast, intensive systems may be Beldt 1994 for Asia). more interesting, but food security and risk issues are Many agroforestry systems are profitable to farm- more critical. ers under a considerable range of economic condi- tions, and various types of (low-intensity) traditional Extension Strategies agroforestry are indeed practiced in many areas. Many agroforestry systems are profitable at real dis- The demonstration effect of fast-growing tree species count rates of 20 percent or higher. Intercropping, if on farms, and of benefits on demonstration plots, has managed properly, tends to be superior to woodlots. helped to expancd agroforestry activities, reducing the Soil-improving interventions could be profitable costs of extension and increasing its effectiveness. even with conservative assumptions of environmen- Rather than offering standard designs, programs tal effects. The performance of windbreaks tends to would serve farrners best if they would offer a broad be poor, perhaps because design and local participa- selection of species and systems from which to tion are unsatisfactory. Perennial intercrops diversify choose those most suitable to the household's needs income, add value per unit of land, improve cash and resources. flow, and cause only limited loss of the main product. Technical assistance is needed to facilitate adoption and to provide information about managing unfamiliar Agroforestry Adoption species and configurations. An endogenous process of agroforestry is c'Learly developing in parts of the region Local scarcity of wood products is, as might be in response to economic incentives and subsistence expected, a key motivator in adopting nontraditional needs. But this process is slow, and its acceleration agroforestry. Projects must begin by assessing the would require access to limited, but continuous, farm- scarcity of wood, as well as the existence of local or group-level technical assistance. Where new mar- markets for products, when locating agroforestry kets for products are being accessed or developed, efforts. extension services may also provide critical informa- Taungya, perennial intercrops, trees on contours, tion and assistance in marketing. and tree lines have proved to be the easiest systems Involving local people as paratechnicians is often to introduce. Results are mixed for alley cropping, a successful, low-cost approach to promoting tech- home gardens, windbreaks, green manuring, dis- nology. Their participation makes projects more sus- persed trees in cropland, and tree-pasture systems. tainable by developing local human resources and Farmers are willing to invest in rehabilitating their thus the capacity to diffuse agroforestry beyond the land where systems also produce products or income, project period. and they prefer less-intensive systems. Financial incentives and subsidies should be kept Even when agroforestry is profitable, smallhold- to a minimum. Agroforestry technologies promoted ers will and should adopt agroforestry incrementally by extension should be financially profitable, and and gradually because of management and resource thus adoptable, for the farmer without subsidies, with constraints. Poorer farmers, in particular, are often the possible exception of time-limited financial hampered by limited land, labor, and capital incentives for early adopters of unfamiliar technolo- resources and their need to ensure food security and gies. In-kind, material inputs encourage farmers to reduce risks. Programs to promote rapid, large-scale experiment with and adopt agroforestry, but the expe- adoption of agroforestry may put smallholders at risk riences with food-for-work incentives are generally or bias adoption and benefits heavily toward higher- not promising. income farmers; continuous, small-scale adoption is a more appropriate objective for smallholder pro- Institutional and Policy Implications grams. Preferred agroforestry systems offer short- term and intermittent benefits that permit farmers to There has been considerable regional (Alfaro and oth- self-finance investments. ers 1994) and international (FAO 1993 and Oram and Adoption patterns differ between small and large Scherr 1993) discussion recently about the types of pol- farms. For medium and large farms, fallow and icy reforms necessary to promote agroforestry as well extensive grazing are still important, and intensive as forestry. In general, the issues raised concerning 340 Costs, Benefits, and Farmer Adoption of Agroforestry macro- and intersectoral policy linkages, trade policies, (Nicaragua), Manuel G6mez (Panama), Abel institutional development, and conservation policy are Hernandez (Dominican Republic), and John Jickling relevant across-the-board. For Central America and the and T. Andrew White (Haiti). The case studies are Caribbean, particular attention should be paid to presented in Current, Lutz, and Scherr (1995). improving the institutional structure for tree product Thanks also go to the many individuals interested in markets (information, monitoring, grading, and stan- agroforestry issues who provided information or who dards); modifying regulations that restrict markets for participated in the national workshops or in the farm-produced products; providing public support ser- regional meeting in El Salvador. The comments by vices for decentralized NGO extension and paraexten- Chona Cruz, Peter Dewees, John McIntire, Augusta sion efforts; and incorporating agroforestry into plan- Molnar, and Stefano Pagiola on earlier drafts are ning efforts. much appreciated. Institutional responsibilities for agroforestry 1. Although many trees have multiple potential extension and support need to be defined. Because uses, management for one use may preclude other agroforestry falls between the ministries of forestry uses. One cannot produce high-quality wood in alley and agriculture, the institutional "home" for agro- cropping systems. Fodder banks and living fences forestry activities has been uncertain. Nongovernmental generally do not produce timber. organizations have taken a leading role in providing 2. These results include information on three information and support but have sometimes undertak- windbreak systems for which averages could not be en isolated and uncoordinated efforts. Semi- calculated because the systems were not fully com- autonomous projects coordinated closely with host- parable, as well as a number of other, less frequently government agencies have been effective. used systems. They also show that certain systems Lack of land titles does not in and of itself appear are more prevalent and important in some countries to be a significant constraint to agroforestry adoption than in others owing to climatic, ecosystem, or cul- in most areas. The important point is how secure tural factors. farmers feel about their property rights with or with- 3. Financial profitability was determined through out an official title. De facto property rights generally a cost-benefit analysis of the agroforestry systems provide farmers with enough security. However, in reviewed in their entirety, not the incremental net general, no tree planting is taking place on rented benefit of the agroforestry system over an alternative land. system of land use. The profitability of the agro- forestry systems was compared with that of an alter- Future Research native. More precise assessment is needed of the effects of Glossary specific policies and extension approaches. For that purpose, empirical studies are needed on a large This glossary is based on Rocheleau, Weber, and number of farms, with specific agroforestry practices Field-Juma (1988). sampled across a range of environments, farm size, Agroforestry. The deliberate use of woody perennials and market access. Monitoring of on-farm environ- (trees, shrubs, palms, bamboos) on the same land mental effects would be particularly interesting, giv- management unit as agricultural crops, pastures, en the dearth of information in that area. and animals. This may consist of a mixed spatial arrangement in the same place at the same time or Notes a sequence over time. Alley cropping. Cultivation of annual crops between Funding from United Nations Development rows of trees or hedgerows. It is sometimes called Programme and ;institutional support from Hans hedgerow intercropping. Binswanger and Michael Baxter are gratefully Boundary planting. Lines of multipurpose trees or acknowledged. The authors are indebted to the con- shrubs planted along borderlines and boundaries sultants who prepared the country case studies: Dean dividing properties or land uses. Current and Ricardo Hernandez (Costa Rica), Carlos Contour planting. Rows of trees or shrubs planted Reiche (El Salvador), Otto Samoyoa (Guatemala), along the contour of the field. Carlos Rodrfguez (Honduras), Rodolfo Vieto Home garden. A complex collection of woody and Dean Current, Ernst Lutz, and Sara Scherr 341 herbaceous plants deliberately grown in small Garcia-Barrios. 1989. "Home Gardens of a Humid plots in or near home compounds, often associat- Tropical Region in Southeast Mexico: An Example ed with the production of small domestic animals. of an Agroforestry Cropping System in a Recently Live fenceposts. Use of living trees, rather than dead Established Community." Agroforestry Systems posts, as fenceposts. In Central America, these are 8(2):133-56. usually used to support barbed wire fencing. The Arnold, J. E. M. 1995. "Retrospect and Prospect." In fenceposts may also be managed for fuelwood or J. E. M. Arnold and Peter A. Dewees, eds., Tree poles. Management in Farmer Strategies: Responses to Multipurpose trees or shrubs. A woody perennial Agricultural Intensification, pp. 272-87. Oxford: grown to provide more than one product or ser- Oxford University Press. vice. Arriola, Francisco, and Rudy Herrera. 1991. Casa Taungya. A system in which new forest plantations Cabrican, Quetzaltenango, Guatemala. Agricultura are established together with food and cash crops, sostenible en las laderas Centroamericanas: which continue to be intercropped until shaded Oportunidades de colaboraci6n interinstitutional. out by the maturing plantation. Coronado, Costa Rica: Instituto Interamericano de Trees intercropped with annual crops. Closely spaced Cooperaci6n Agricola. trees planted or maintained in cropland for their Belaunde, Elvira, and A. C. Rivas. 1993. Responding products or their positive effects on associated to Practice and Affecting Policy: The National crop production (for example, nitrogen fixation or and Regional Impact of Local Experience in the microclimate improvement). Madelenia IProject in Central America. Rural Trees intercropped with perennial crops. Closely Development Forestry Network Paper 16a. spaced trees planted or maintained in perennial London: Overseas Development Institute. crops for their products or their positive effect on BID (Banco Interamericana de Desarrollo). 1993. associated crop production (for example, shade or "Progreso econ6mico y social en America nitrogen fixation). Latina." Informe 1993. Special theme: Inversion Windbreak. Strips of trees or shrubs planted to pro- en recursos humanos. Washington, D.C. tect fields, homes, canals, or other areas from Budd, W. William, Irene Duchhart, L. H. Hardesty, wind and blowing soil or sand. and Frederick Steiner, eds. 1990. Planning for Woodlot. Stands of trees planted and managed to pro- Agroforestry. Amsterdam: Elsevier. duce various tree products, associated plants, or Budowski, Gerardo. 1987. "The Development of services. Agroforestry in Central America." In Howard A. Woody fallow. Fallow is land resting from cropping, Steppler and P. K. Ramachandran Nair, eds., which may be grazed or left unused, often colo- Agroforestry: A Decade of Development, pp. nized by natural vegetation. For woody fallows, 69-88, Nairobi: International Center for Research the fallow is left uncultivated for sufficient time in Agroforestry. that woody plants come to predominate; farmers Campos, Oscar, Emel Rodriguez, and Luis Ugalde. may enrich, manage, or harvest from this woody 1991. "Desarrollo agropecuario sostenible en la fallow. regi6n de lF[ojancha, Guanacaste, Costa Rica." In Agricultura sostenible en las laderas centroamer- References icanas: Oportunidades de colaboracion interin- stitucional. Memorias del taller bajo patrocinio The word "processed" describes informally repro- de Cooperaci6n Suiza para el Desarrollo, organi- duced works that may not be commonly available zado por el Centro Internacional de Agricultura through libraries. Tropical, Instituto Interamericano de Cooperacion Alfaro, Marielos, Ronnie de Camino, Maria Ileana para la Agrfcultura, Centro Agron6mico Tropical de Mora, and Peter Oram, eds. 1994. Taller regional: Investigaci6in y Ensenanza, Centro Internacional de Necesidades y prioridades de investigaci6n en Mejoramiento de Mafz y Ti-go, pp. 245-77. Costa politicas forestales y agroforestales para Rica: Instituto Interamericano de Cooperaci6n Latinoamerica. San Jose, Costa Rica: Instituto Agricola. Interamericano de Cooperaci6n Agricola. Cook Cynthia, and Michael Grut. 1989. Agroforestry Alvarez-Brylla, M. E., E. Lazos-Chavero, and J. R. in Sub-Saharan Africa: A Farmer's Perspective. 342 Costs, Benefits, and Farmer Adoption of Agroforestry Technical Paper 112. Washington, D.C.: World Gregersen, Hans, Sydney Draper, and Dieter Elz. Bank. 1989. People and Trees: The Role of Social Current, Dean, and Modesto Juarez. 1992. The Present Forestry in Sustainable Development. Washington, and Future Status of Production and Consumption D.C.: Economic Development Institute and the of Fuelwood in El Salvador. Tropical Agricultural World Bank. Research and Higher Education Center-U.S. Kerkhof, P. 1990. Agroforestry in Africa: A Survey of Agency for International Development-El Project Experience. London. U.K.: The Panos Salvador. Turralba, Costa Rica: Tropical Institute. Agricultural Research and Higher Education King, K. F. S. 1979. "Concepts of Agroforestry." In Center. M. T. Chandler and D. Spurgeon, eds., Current, Dean, and Ernst Lutz. 1992. "A Preliminary International Cooperation in Agroforestry. Economic and Institutional Evaluation of Selected Proceedings of an International Conference, pp. Agroforestry Projects in Central America." 1-14. Nairobi, Kenya: International Center for Divisional Working Paper 1992-38. Environment Research in Agroforestry. Department, World Bank, Washington, D.C. Murray, Gerald F. 1991. "The Tree Gardens of Haiti: Processed. Agroforestry among Caribbean Peasants." In Current, Dean, Ernst Lutz, and Sara Scherr. 1995. David Challinor and Margaret Hardy Frondorf, Costs, Benefits, and Farmer Adoption of Agro- eds., Social Forestry: Communal and Private forestry: Project Experience in Central America Management Strategies Compared, pp. 35-44. and the Caribbean. Washington, D.C.: World Washington, D.C.: U.S. Agency for International Bank. Development; Baltimore, Md.: Johns Hopkins Current, Dean, Carlos Rivas, and Manuel G6mez. University Press. 1994. El sistema estandarizado de registros para Mussak, M. F., and J. G. Laarman. 1989. "Farner's actividades de extensi6n forestal MIRAExT: Una Production of Timber Trees in the Cacao-Coffee herramienta para seguimiento y evaluaci6n. Region of Coastal Ecuador." Agroforestry Systems Turralba, Costa Rica: Tropical Agricultural 9(2):155-70. Research and Higher Education Center. Nair, P. K. R. 1990. The Prospects of Agroforestry in Dewees, Peter A. 1993. Trees, Land, and Labor. the Tropics. Technical Paper 131. Washington, Environment Paper 4. Washington, D.C.: World D.C.: World Bank. Bank. National Academy of Sciences. 1980. Firewood Eckholm, Eric. 1975. The Other Energy Crisis: Crops: Shrub and Tree Species for Energy Firewood. Worldwatch Paper 1. Washington, Production. Vol. 1. Washington, D.C.: National D.C.: Worldwatch Institute. Academy Press. FAO (Food and Agriculture Organization of the Nitrogen Fixing Tree Association. 1993. Nitrogen United Nations). 1993. Forestry Policies of Fixing Tree Research Report. Vol. 11. Hawaii. Selected Countries in Asia and the Pacific. Oram, Peter, and Sara J. Scherr, eds. 1993. Report of Forestry Paper 115. Rome. the African Regional Workshop on Forestry and French, James H., and Rick J. van Den Beldt, eds. 1994. Agroforestry Policy Research. Nairobi, Kenya, 20 Forestry/Fuelwood Research and Development November-4 December. Washington, D.C.: Project. Final Report. Bangkok, Thailand: Winrock International Food Policy Research Institute. International. Raintree, John B. 1991. Socioeconomic Attributes of Godoy, Ricardo A. 1992. "Determinants of Smallholder Trees and Tree Planting Practices. Rome: Food and Commercial Tree Cultivation." World Development Agriculture Organization of the United Nations. 20(5):713-25. Raintree, John B., and Hermina A. Francisco, eds. G6mez, Manuel, and Thomas McKenzie. 1991. 1994. Marketing of Multipurpose Tree Products in "Resumen anual de operaciones econ6micas en Asia. Proceedings of an International Workshop fincas demostrativas, del Proyecto Madelenia, en Held in Baguio City, Philippines, 6-9 December, Am6rica Central." Internal technical report for the 1993. Bangkok, Thailand: Winrock International. Madelenia Project. Turrialba, Costa Rica: Tropical Rocheleau, Dianne, Fred Weber, and Alison Field- Agricultural Research and Higher Education Center. Juma. 1988. Agroforestry in Dryland Africa. Processed. Nairobi, Kenya: International Council for Dean Current, Ernst Lutz, and Sara Scherr 343 Research in Agroforestry. Farmers in Upland Watersheds: Experience in Scherr, Sara J. 1992. "Not Out of the Woods Yet: Central America." International Tree Crops Challenges for Economics Research on Agro- Journal 4(4):249-68. forestry." American Journal of Agricultural Eco- U.S. Comptroller General. 1982. Changes Needed in nomics 74(3):802-08. U.S. Assistance to Deter Deforestation in . 1995. "Economic Analysis of Agroforestry Developing Countries. Report to the Congress of Systems from the Farmers' Perspective." In Dean the United States. Washington, D.C.: General Current, Ernst Lutz, and Sara Scherr, eds., Farmer Accounting Office. Adoption and Economic Benefits of Agroforestry: USAID/ROCAP (U.S. Agency for International Project Experience in Central America and the Developmerlt/Regional Office for Central American Caribbean. Washington, D.C.: World Bank. Programs). 1979. "Fuelwood and Alternative Scherr, Sara J., and Eva Muller. 1991. "Technology Energy Sources." Washington, D.C. Processed. Impact Evaluation in Agroforestry Projects." Vergara, Napoleon T., and Kenneth G. MacDicken. Agroforestry Systems 13(3):235-58. 1990. "Extension and Agroforestry Technology Shepherd, Gill, ed. 1992. Forest Policies. London: Delivery to Farmers." In Kenneth G. MacDicken Overseas Development Institute. and Napoleon T. Vergara, eds., Agroforestry: Steppler, Howard A., and P. K. Ramachandran Nair, Classification and Management, pp. 354-75. New eds., Agroforestry: A Decade of Development. York: John Wiley and Sons. Nairobi: International Center for Research in Wamer, Katherine. 1993. Patterns of Farmer Tree Agroforestry. Growing in Eastern Africa: A Socioeconomic Swinkels, Rob A., and Sara J. Scherr, comps. 1991. Analysis. Nairobi, Kenya: Oxford Forestry Economic Analysis of Agroforestry Technologies: Institute and International Center for Research in An Annotated Bibliography. Nairobi, Kenya: Agroforestry. International Center for Research in Agroforestry. World Bank. 1994. World Development Report 1994. Tschinkel, Henry. 1987. "Tree Planting by Small New York: Oxford University Press. Part IV Conclusions Meeting the food and fiber needs of growing popula- of sustainabiliy under which capital can be changed tions in developing countries while conserving the from one category into another, while keeping the resource base is and will continue to be a great chal- total constant. Yet we also recognize that there may lenge for everyone involved in rural development. be critical levels of each type of capital (critical This volume brings together recent findings and cur- ecosystems) beyond which capital should not be sub- rent thinking intended to help meet the challenge. stituted. Past agricultural growth has been associated with Sustainable rural development must appropriately a variety of environmental and resource problems integrate environmental concerns (Hazell and Lutz, including excessive deforestation and forest degrada- chapter 2). Such integration includes correcting price tion, water depletion, waterlogging and salinization, distortions that encourage excessive use of modem fish stock depletion, soil degradation, health effects, inputs in intensive agriculture, ensuring that farmers and biodiversity losses. Comprehensive, empirically have secure property rights over their resources, strong estimates for the costs of resource degradation strengthening community management systems and associated with agriculture in developing countries empowering local organizations, improving the per- do not seem to exist; in their absence, Hazell and formance of relevant public institutions that manage Lutz (chapter 2) present indicative estimates from and regulate natural resources by devolving manage- various sources. ment decision.s to resource users, or groups of users, A key message in this volume is that environmen- wherever pos sible, and giving greater attention to tal problems do not have to be the inevitable outcome sustainability features of recommended technologies of agricultural growth. Problems are usually associat- (Hazell, chapter 8). ed with inappropriate incentive systems (broadly It is evident that environmentally and socially sus- defined), inappropriate institutional settings, insuffi- tainable rural development depends on appropriate cient investment in resource-poor areas, inadequate policy frameworks, appropriate institutional poverty and social concerns, and political economy approaches, as well as appropriate technologies. We systems in which the rich and powerful extract rents. have structured our conclusions into these three parts, Although no precise definition is offered, sustain- which mirror the three main parts of the volume. able agricultural development is conceptualized as agricultural income traditionally measured in nation- Appropriate Policy Frameworks a] income accounts corrected for changes in the value of the stock of capital, which includes natural, man- It essential thiat an appropriate policy framework be made, and human capital. We also adopt a definition in place that is conducive to sustainable rural devel- 345 346 Part IV Conclusions opment. In too many countries, there are still too than their deterioration. But Boserup effects are far many obvious distortions in the incentive frame- from automatic; they are an outcome of investment works. Most of these should be removed simply on decisions made by farmers, who require a positive the basis of sound economics and efficiency.' In incentive regime and access to markets and to soil many cases, these reforms would have direct or indi- and water resources. Adverse policy regimes force rect positive impacts on the environment; that is, tak- impoverished peasants to mine, rather than augment, ing environmental concerns into account would often the land resources. This mining is especially damag- add to the benefits of reforms that should have been ing if they only have access to marginal land or to undertaken on the basis of economics. But many humid tropical forest frontiers. Therefore, the reforms are not carTied out for political economy rea- removal of adverse policies should be given top pri- sons, where perhaps small but important groups or ority. individuals derive benefits from the existing distor- Adverse policies often reduce economic efficien- tions that they are not willing to give up voluntarily. cy, increase poverty, and degrade natural resources Government leaders either may themselves benefit (Heath and Binswanger, chapter 3). Faulty policies from the distortions or may feel that they cannot have tended to favor the modernization of large-scale afford to lose the support of a particular stakeholder farming at the expense of more efficient and employ- group that is benefiting from the current policies. ment-intensive family farms. Reforming adverse Resource and environmental problems are quite policies would be a win-win-win situation because different for the array of agroecological conditions the same policy changes could lead to more growth, that one encounters. For example, in irrigated agri- less poverty, and more sustainable natural resource culture the main problem is the inappropriate use of management. inputs (irrigation water, seeds, fertilizers, pesticides, Although we emphasize here the importance of and machinery), often in response to a faulty incen- appropriate agricultural policies including overall tive framework, while in rainfed agricultural areas, incentive frameworks (including macro and trade the main problem may be resource degradation due to policies), we also recognize the importance of proper insufficient yield growth in relation to population infrastructure (related to market access), education growth (Hazell and Lutz, chapter 2). (in particular, the education of girls), health, and Sustainable agriculture seeks the integrated use of macroeconomic stability. In some contexts-for a wide range of pest, nutrient, soil, and water man- example, where farmers are poorly integrated into agement technologies. It also aims for increased markets or where there are significant market imper- diversity of enterprises within farms so that by-prod- fections and farmers are facing significant risks-it ucts or wastes from one component or enterprise may also be necessary to assist in additional ways, become inputs to another (Pretty, chapter 4), thereby such as by interlinked contracts (Holden and reducing negative impacts on the environment. Binswanger, chapter 5). The diversity of situations implies that sustainable New policies to stimulate sustainable rural devel- agriculture cannot be a fixed model that is to be opment are much needed, and, to succeed, these poli- imposed; rather, it should be a process of learning cies have to build on an understanding of the deci- (Pretty, chapter 4; Bunch, chapter 12). What needs to sionmaking environment and behavioral responses of be made sustainable, therefore, is the social process small farmers. Small farmers largely are rational of innovation itself, which implies an enhanced and respond to changes in the set of constraints and capacity to adapt to unexpected changes and emerg- opportunities they face. But inefficiencies may still ing uncertainties. For such innovation to occur, it is accrue due to economies of scale, poverty and sub- necessary to have an enabling policy environment. sistence constraints, unequal access to markets, and Chapter 3 (Heath and Binswanger) and chapter 5 various policies disfavoring small farmers. These (Holden and Binswanger) elaborate on this point, inefficiencies may also lead to underinvestments in supported by empirical evidence that conducive pol- conservation. There may thus be grounds for inter- icy and institutional environments are a key condi- ventions from the viewpoints of efficiency and equi- tion for inducing innovations. ty as well as sustainability. Given the complexity of The Boserup hypothesis states that, under the small-farm problems, it is necessary to start with right conditions, higher population and market access pilot projects and to adopt a flexible project design can lead to improvements in natural resources rather that is accompanied by systematic monitoring and Part IV Conclusions 347 evaluation to allow for maximum learning and con- ment projects i:ncrease the flow of income from nat- sequent adjustment and upscaling. ural resources over which rights previously may have Part of improving the overall incentive framework been ambiguous, powerful individuals at the local usually means reducing trade distortions. Many level will usually devise a way to expropriate at least countries began to reduce such distortions during the part-if not thne majority-of these new income past decade. In general, trade liberalization leads to streams. Hence, although ambiguity can often work overall net gains, but some negative environmental to the advantage of the dispossessed, it is inlikely to effects are possible (Anderson, chapter 6). Where do so when development projects are introduced into they occur, environmental policy or regulatory this kind of institutional environment. Therefore, actions may be needed. To support these, one should development projects introduced into settings in make more efforts to quantify the impacts that are which there is great ambiguity about property rela- likely to be associated with trade and other econom- tions will most probably have economic and social ic reforms. impacts quite different from those postulated in the Unilaterally removing the policies that keep proposals and Lfeasibility analyses that led to the proj- domestic food prices artificially low in developing ect's acceptance and implementation in the first place. countries would boost incentives to invest in agricul- Therefore, every development intervention must be tural research. But at the higher price levels, the preceded by a concerted effort to ensure that the insti- cost-in terms of degrading the farm sector's natural tutional arrangements have been modified so as to resource base-to those countries of failing to reduce ensure that the benefits go to the intended beneficia- distortions in the markets for farm chemicals and irri- ries. This modification of the "working rules" of the gation water would be greater than when product local economy must occur before the new benefit prices, and hence input intensities, were lower. This stream begins to materialize (Bromley, chapter 7). suggests the need to introduce simultaneous reform Eliminating the privileges of rural elites that are in the pricing of outputs and inputs. embedded in adverse policies is an extremely diffi- The outcomes of policy reforms and development cult task. These policies did not come about by acci- activities depend on institutional arrangements in dent, but as the consequence of a historical evolution general and property relations in particular (Bromley, that involved bargaining among politically strong chapter 7). Property regimes are central in develop- groups looking out for their interests at each point in ment because they connect people to one another time (Heath and Binswanger, chapter 3). In practice, with respect to land and related natural resources. policy is the net result of the actions of different Development projects are sometimes less successful interest groups pulling in complementary and oppos- than they would otherwise be precisely because the ing directions, making the outcome uncertain (Pretty, existing property relations have been ignored or mis- chapter 4). Analyzing the joint consequences of the understood. policies, using sector studies, for example, may aid Bromley offers nine main lessons for develop- the reform effort. But it is unlikely to be sufficient. ment practitioners: (1) property regimes are part of Steps must be taken to provide peasants with greater the larger institutional structure, (2) all property input into policymaking (Heath and Binswanger, regimes require external legitimacy, (3) ambiguous chapter 3). Unless more political actors with enlight- rights regimes have ambiguous efficiency and distri- ened hearts and minds see it as in their interest to butional consequences, (4) property relations must be bring poor peasants more strongly into the policy specified prior to the implementation of development process, reforms of the policy environment and of projects, (5) resource degradation is the result of land may continue to progress slowly. problems that precede property regimes, (6) ecologi- How can development agencies encourage cal variability demands flexible institutions and reforms when political economy constraints are actions, (7) resource degradation is contextual, (8) blocking them? First, they should seek unity among land titles are not necessary for efficient investments themselves as to what reforms are to be undertaken. in productivity (security of tenure is much more Second, they should take a firm approach, supported important), and (9) mobilizing local interests can by credible analyses of current policies that show improve a program's chances of success. who are the g;ainers from the current policies and who An elaboration of lesson number 4 may be impor- are the losers (producers, consumers, treasury, envi- tant for development practitioners. When develop- ronment). New lending or grants should be provided 348 Part IV Conclusions after progress has been made in specific areas. Unless another, work together, organize, solve problems, a government demonstrates its commitment to mobilize resources, resolve conflicts, and network improving economic development, outsiders should with others to achieve agreed goals. Viable commu- not be "enablers" of inefficiency. nity groups are key to the success of community- In addition to building a basis for seeking unity driven development. Getting local groups and orga- among donors and holding discussions with govern- nizations to become self-managing organizations can ments, applied, empirical analyses may inform and extend over several years and does not happen with- activate public opinion, strengthen reform-oriented out investment in capacity building. Local elites constituencies, and create more local ownership often take leadership roles, and, although not neces- through workshops, seminars, and meetings with civil sarily bad, this can result in the hijacking of resources society and nongovernmental organizations (NGOS). unless transparency and accountability are somehow enforced. Appropriate Institutional Approaches The appropriateness of technology is determined by the local management capacity and how the intro- The approach to rural development must be partici- duction of a technology influences social cohesion patory and decentralized. A sound policy environ- and benefit flows. Technology needs to be viewed in ment is essential for sustainable agricultural develop- a holistic way to determine its fit with the existing ment, among other things, because it creates the social system and the direction of change that will be incentive framework that induces farmers to inno- set into motion by its introduction. vate. Appropriate institutional frameworks and con- Outreach mechanisms can be classified broadly sideration of the social dimensions are equally into two basic approaches: the extension approach important because they provide the structure or fab- and the empowerment approach. The two often over- ric within which policies are applied and form an lap, and an effective outreach strategy contains ele- important part of the context within which farmers ments of both. The extension approach is character- work. ized by dissemination of information, promotion of Projects must be participatory and decentralized demand, and delivery of particular services or inputs. because building local ownership is necessary for The empowerment approach is characterized by the sustainable efforts (Narayan, chapter 9). It has some- empowerment of local groups or the creation of local times been argued that participation is time-consum- organizational capacity for self-management through ing and that decentralized approaches are difficult involvement in decisionmaking. because of the generally lirnited institutional capaci- Client-centered agencies are needed to ensure that ty at local or regional levels. But experience shows demand is met and community self-management is that these objections are not well founded if the adequately supported by community fieldworkers. To intention is to further locally driven assistance that become client-centered, agencies must be convinced produces sustainable impacts over the long term. to change their rules and incentive structures. Interest Limited development success of central, top- and commitment to change can be generated by pilot down approaches has led to recognition of the need projects that demonstrate alternative strategies and for participatory, community-driven development use a variety of participatory techniques, including (Narayan, chapter 9). Community-driven develop- field visits to other regions and countries. ment is a process in which community groups initi- Community-driven development on a large scale ate, organize, and take action to achieve common requires enabling policies as well as political support interests and goals. No single model is appropriate to protect agencies so that they can initiate the reform for all places and times. But successful community- process and give it time to take root. Although indi- driven development generally has five main charac- viduals play key roles in starting or protecting the teristics: local organizational capacity or the exis- reform process in the early stages, change can be sus- tence of viable community groups, the appropriate fit tained only if the rules and regulations at different of technology to community capacity, effective levels provide appropriate incentives to work and agency outreach strategies, client-responsive agen- organize. cies, and enabling policies. To increase the probability of success with partic- Local organizational capacity is embedded in the ipatory community-based development projects, the norms that enable groups of people to trust one following 10 lessons from worldwide experience Part IV Conclusions 349 should be reflected in the project design, subject to involve local stakeholders in the development of testing with pilot activities: (1) start with clearly stat- management practices and the selection of technolo- ed objectives logically linked to strategies, outputs, gies aimed at conserving watershed resources often indicators of success, and physical or capacity-build- promote innovation by identifying locally appropri- ing outcomes; (2) identify the key actors and their ate technologies and securing their adoption more capacity and interests (using social assessment, for effectively than extemal agencies. Third, devolving example); (3) assess demand (as measured by com- responsibility to local organizations can both exter- mitment before construction-for example, with a nalize some costs of enforcing conservation from the financial investment up-front in capital cost); (4) state to local communities and also reduce costs craft a self-selection process (such as by requiring a overall, by creating conditions in which nonlocal financial or organizational contribution up-front); (5) agencies become more efficient and effective through structure subsidies that do not distort demand; (6) their collaboration with local organizations. restructure fund release to support demand (such as Although many lessons can be learned and princi- by implementing fiscal decentralization); (7) plan for ples derived from experience, the last word on what learning and a plurality of models (the project must works, and what does not work, is by no means be seen as likely to adjust and change over time as decided (Ashby and others, chapter 10). Indeed, the local priorities change and as local organizations dynamic nature of the process of involving local mature); (8) invest in outreach mechanisms and social organizations in watershed management is one of its organization (investments are needed, and there are most important characteristics and requires focusing no shortcuts to strengthening local social organiza- on managing t:he process rather than on defining a tion for collective action); (9) use participatory mon- blueprint for how to go about it. An interesting case itoring and evaluation (listening to people's sugges- study in chapter 10 illustrates this. tions is empowering and encourages innovation and Cooperation in watershed management is not nec- responsibility); and (10) redefine procurement rules, essarily easy to initiate or sustain, even when local where necessary (so that they do not hinder commu- organizations are involved. Multiple uses for any nity initiative). given watershed resource create multiple stakehold- Conserving or improving natural resources often ers, often with, competing priorities. For this reason, requires collective action by groups of users, but conflict among stakeholders over the rights and con- there is evidence that self-organizing groups have ditions for use of a given resource is a common fea- been weakened and undermined by the growth of ture of watershled management. Thus, one of the key government intervention, with negative conse- functions of local organizations in watershed man- quences for resource management. Involving local agement is to build the social capital required for organizations in rural development contains princi- people to build consensus about and enforce the ples of both participation (Narayan, chapter 9) and agreed use of watershed resources for diverse, multi- decentralization (Caldecott and Lutz, chapter 14). ple, and often conflicting purposes. Ashby and others (chapter 10) explain these princi- Few local organizations at the watershed scale ples at the watershed level and illustrate them with have arisen without some intervention from outside practical experiences and a case study. agencies to catalyze their formation and action. At There are three main advantages in involving the local level, outsiders usually play a significant local organizations in watershed management (Ashby role in enabling stakeholders to arrive at a joint plan and others, chapter 10). First, local organizations can of action that takes into account transboundary often be very effective in generating and securing effects not readily perceived or measured by those compliance with rules for the use of common proper- who are not directly affected. Another important ty such as water, common grazing land, or forest and point of interface between local organizations and the management of buffer zones around conservation external agencies in watershed management is the areas, all of which may be important features of need for external intervention in identifying the rele- watershed management. This is in contrast to the dif- vant stakeholiders and bringing them to the table, ficulties experienced by nonlocal agencies, especial- whether this is literally a negotiating table or partici- ly in low-income countries, in imposing sanctions on pation in collective labor, monitoring, or enforcement undesirable management practices or in providing of sanctions. incentives that are lasting. Second, organizations that The involvement of local organizations in water- 350 Part IV Conclusions shed management does not necessarily ensure equi- Participatory principles are also key to successful table participation of all relevant stakeholders in a extension with long-run impact (Bunch, chapter 12). resource management initiative (chapters 9 and 10). One such approach called participatory people-cen- Case studies show that women, the landless, margin- tered agricultural development (PCAD) has been al ethnic groups, and the laboring poor are unlikely to developed and tested extensively in Central America be represented in participatory watershed manage- by World Neighbors and has spread to mostly non- ment. Leadership in local organizations is generally governmental organizations in many nations around captured by the higher status, more well-to-do mem- the world. bers of dominant elites, and the priorities of local Government extension services have tended to organizations tend to reflect their interests. Therefore, neglect farmers in resource-poor areas. One way to watershed development projects should deliberately overcome this is to enlist the farmers' own efforts in seek to overcome this bias by setting up catchment solving their problems of low productivity. A partici- committees with representation of underprivileged patory process teaches villagers how to experiment groups. and teach each other through the most efficient teach- Participation is also one of the keys to successful ing process: learning by doing. generation of technology among low-income farmers PCAD consists of a series of principles with much (Farrington, chapter 11). Participatory approaches are room for adaptation to local circumstances, farmer easy enough to implement on a small scale, but spe- needs, and the institutional context. They include cific institutional arrangements have to be made to motivating and teaching farmers to experiment with permit wide-scale implementation. These can best be new technologies on a small scale, thereby reducing based on partnerships between public agencies, pri- the risk of adoption and providing them with a means vate organizations (especially NGOS), and farmers to continue developing, adopting, and adapting new themselves, both to bring additional resources to bear technologies in a permanent process of innovation. and to reorient the public sector. Chapter 11 argues This principle is frequently referred to as participato- that, especially under unreliable rainfed conditions ry technology development. Another key principle is where the private commercial sector is likely to the importance of training village leaders as exten- remain weak, public sector research and dissemina- sionists and supporting them as they teach additional tion continue to have an important role. However, in farmers, thereby creating and nurturing a community- such areas, it has been difficult to maintain an effec- based multiplier effect. This principle is called tive public sector capacity on the ground, and farmer-to-farmer extension. research and extension have generally been oriented In many instances, where PCAD has been used suc- only weakly toward clients' needs. cessfully, yields of major crops have tripled, and new Multiagency approaches-not replacement by the ones have been introduced. But success has not been private sector or complete abolition-are one way of uniform. One survey showed that, in the 25 to 30 per- overcoming the limitations of the public sector. cent of villages where the response was best, yields Innovative multiagency approaches are now being continued to increase after program termination, the tested in various ways. They seek synergistic interac- number of organizations increased, land prices tions in which different types of organizations work increased dramatically, incomes increased, and out- together to analyze problems, contribute jointly to migration either diminished drastically or was their solution, review progress in an iterative fashion, reversed. In 40 to 50 percent of the average villages, and make course corrections by mutual agreement as yields more or less maintained or increased only mar- necessary. ginally after project completion, whereas in the 20 to Although some success has been achieved in 35 percent of the villages where impact was poorest, introducing participatory approaches at the diagnos- yields decreased, although they continued to be bet- tic stage of the research cycle (for instance, partici- ter than the yields at program initiation. patory rural appraisal is now widely used, although One of the major conclusions of PCAD is that sus- often badly), much less attention has been paid to tainability does not reside in technologies, which ensuring continued interaction between researchers have a half-life of perhaps five years. Markets and their clients during subsequent identification, change, input prices increase, new technological testing, and dissemination of technologies (Farrington opportunities appear, varieties degenerate, pests and Thiele, chapter 11). spread, and competition becomes stiffer. The hope Part IV Conclusions 351 for sustainability of agricultural development is in to eligible beneficiaries (Deininger, chapter 13). the nature of an ongoing social process of widespread As Heath and Binswanger (chapter 3) indicate, experimentation and sharing of information, of inno- small farmers in Colombia were often driven off their vation, and of group problem-solving. In PCAD, it is traditional lands to eke out a living in marginal and more important how things are taught than what par- environmentally fragile areas, and much of the best ticular technologies are taught. Thereby the PCAD agricultural land. continued to be devoted to extensive process creates a sustainable process of agricultural livestock grazing or was not farned at all due to vio- development. By its very nature, it also empowers lence. Only 25 percent of the land suitable for crop people in ways useful for rural development efforts in production was actually devoted to this use, while the general. rest was left to pasture. This suggests that there are One central condition for moving toward more indeed large tracts of unused or underused land that sustainable rural development is for farmers to have could be subjected to land reform in order to increase access to land and to have security of tenure. agricultural productivity-a notion in line with avail- Therefore, even though land reform is politically sen- able empirical evidence. sitive and difficult to implement, it is very important Centralized land reform was attempted in that these issues be tackled. Theoretical reasons and Colombia but did not have much success. Now nego- empirical evidence suggest that land reform may pro- tiated, market-based land reform is being tried. It is vide benefits both for equity and for efficiency. First, characterized by three main principles. Potential ben- a large body of research has demonstrated the exis- eficiaries (with assets below a certain minimum tence of a robustly negative relationship between level) can negotiate independently with landowners farm size and productivity due to the supervision cost and, once a deal has been struck, are eligible for a associated with employing hired labor. This implies grant of up to 70 percent of the land purchase price that redistributing land from large to small farms can (subject to an urpper limit). The government's role is increase productivity. Second, in many situations, limited to that of regulatory oversight and grant dis- landownership is associated with improved access to bursement. A decentralized institutional structure credit markets, providing benefits as an insurance ensures that land reform is driven by demand and is substitute to smooth consumption intertemporally. coordinated with other government programs. By enabling the poor to undertake indivisible pro- Despite favorable preconditions, and the govern- ductive investments (or preventing them from deplet- ment's expressed determination to distribute 1 mil- ing their base of assets), this could lead to higher lion hectares within four years, this new type of land aggregate growth. Third, even in aggregate cross- reform has had a disappointingly slow start. country regression, the distribution of productive Brazil and South Africa have, under different con- assets-more than the distribution of income-seems ditions, recently initiated programs of negotiated to have an impact on aggregate growth. Finally, land land reform. With an institutional background very reform is expected to increase the environmental sus- similar to that of Colombia (presence of land reform tainability of agricultural production and to reduce legislation and a central land reform institute dating rural violence (Deininger, chapter 13). from the early 1960s), negotiated land reform in Notwithstanding this apparent potential, actual Brazil has been initiated by individual states. The experience with land reform has been disappointing purpose of the Brazilian interventions is to establish in all but a few exceptional cases such as Japan, cheaper, more agile policy alternatives to centralized Korea, Taiwan (China), and to some extent Kenya. land reform in an environment where the issue of Despite-or because of-this, land reform remains a land reform is high on the political agenda and poten- hotly debated issue in a number of countries (Brazil, tial beneficiaries have at least some idea of what to Colombia, El Salvador, Guatemala, South Africa, and do with the land. By contrast, negotiated land reform Zimbabwe). in South Africa has been adopted in the context of a A new type of negotiated land reform is now national reconstruction program in an environment in being tried in a number of countries (Brazil, which productive small-scale agriculture was eradi- Colombia, South Africa) where land transfers are cated almost a century ago. Here, even greater effort based on voluntary negotiation and agreement is required to establish the decentralized infrastruc- between buyers and sellers and the government's role ture necessary to implement land reform, to provide is restricted to the provision of a land purchase grant complementary services such as marketing and tech- 352 Part IV Conclusions nical assistance, and to increase beneficiaries' agri- groups, making it necessary for global and national cultural and entrepreneurial capacity. society to bridge the gap with livelihood investments Several of the chapters in part II (particularly or grants, (5) appropriate enforcement, especially chapters 9-12) stress the importance of participation against powerful local or central interests and always to successful, sustainable rural development. Some in the context of education and public relations activ- also touch on the need to pursue decentralized ities, (6) stakeholder forums and ecoregional execu- approaches, which are important for many reasons, tives, which need decisionmaking and fiscal authori- including their role in facilitating participation. ty to fulfill their three main roles of avoiding conflict Decentralization is one aspect of the greater issue through dialogue, authorizing conservation action, of good governance, and one question is whether it and requesting help from nonlocal society to meet promotes or hinders biodiversity conservation and local development priorities, and (7) enabling poli- under what conditions (Caldecott and Lutz, chapter cies, laws, and institutions to provide a clear and sup- 14). portive framework for conservation on behalf of Many past attempts to conserve biodiversity national government, thus creating incentives at the failed. One reason was an overly centralized approach, local level to harmonize development with conserva- often involving top-down planning by technicians tion and reducing the need for enforcement action and bureaucrats without concern for the opinions or (Caldecott and Lutz, chapter 14). well-being of the people affected by their decisions. In addition to participation and decentralization, But because of the externality issues involved, com- appropriate attention to gender issues is also needed plete decentralization can also be counterproductive to achieve sustainable rural development. For sever- to conservation. Under complete decentralization, al decades, food, agricultural, and natural resource local people, perhaps in cooperation with outside management policies have been designed without entrepreneurs, may simply degrade and deplete the acknowledging that rural men and women may have resources faster and more efficiently. So central gov- different preferences, face different constraints, and ernments clearly continue to have a role. The chal- respond differently to incentives. Indeed, women lenge is to find the degree of decentralization that is tend to be "invisible" within agriculture, and the pre- most appropriate. dominant assumption is that the male head of house- Decentralization and conservation are complex hold makes most, if not all, farm allocation and pro- processes that interact with one another in many duction decisions. ways. Decentralizing by giving new responsibilities Quisumbing and others (chapter 15) show how to local government units and NGOS creates both neglecting the gender dimension in development pol- opportunities and potential problems. To take advan- icy design has led to failures in project implementa- tage of the former while avoiding the latter, a cluster tion, failure to adopt new agricultural technologies, of arrangements must be made as a whole if conser- or adoption of the new technology but with negative vation is to work well in a (moderately) decentralized unanticipated impacts. Therefore, policymakers setting. Of these, seven institutional and incentive would do well to take gender issues into account elements merit special attention: (1) local participa- when formulating food, agricultural, and natural tion, especially in a way that allows local people to resource management policies. A review of empirical understand and endorse the boundaries and manage- evidence shows that reducing inequalities between ment plans of nature reserves and that promotes clear men and women in human and physical capital can tenure over land and other resources in and around lead to major efficiency and productivity increases in the reserves, (2) capacity building, especially to agriculture.2 Moreover, addressing inequalities in the increase skills and accountability, among local gov- underlying distribution of property rights between ernment units and nongovernmental organizations to men and women may be essential not only for enable them to work together to promote conserva- increased agricultural productivity but also for envi- tion and rural development, (3) incentive structures, ronmental and social sustainability. The second- especially those that allow local communities to keep round effects are also important. When controlled by income from the sustainable use of nature reserves women, income increases have a greater impact both and other biodiversity assets, (4) conditional subsi- on household food security and on investments in dies, especially where divergent costs and benefits of child health and schooling. conservation are experienced by local and nonlocal Although participatory approaches to project Part I'V Conclusions 353 design (such as discussed in chapter 9) have been rec- to reduce pesticide use, economic and environmental ognized as a way of ensuring community involve- aspects of irrigation and drainage, livestock-environ- ment, commitment, and ownership of development ment interactions, inappropriaLte deforestation, and projects, they are not synonymous with incorporating agroforestry. gender analysis. Participatory approaches to project Soil is one of the basic inputs into the agricultural design by themselves do not guarantee that women production system. Soil degradation, in turn, affects will participate in either assessment of their needs or productivity. As soil is degraded, crop yields decline design of the proposed project. This means that a spe- or input levels (and hence costs) rise in an effort to cial effort needs to be made to have gender properly keep or restore productivity. Despite long-standing addressed, such as by gender analysis. This goes concern about these problems, surprisingly little hard beyond a women-in-development approach, which evidence exists on their magnitude (Lutz and others, identifies the differences in access to, ownership of, chapter 16). and control of resources between men and women. A Degradation can be slowed or arrested by a large gender-and-development approach recognizes the range of options, including cultural practices such as determinants of these asymmetries and their conse- contour plowirg and minimum tillage, vegetative quences for individuals, households, communities, practices such as grass strips, strip cropping, and veg- and economic development. In essence, the women- etative barriers, and mechanical measures such as ter- in-development approach recognizes an outcome, races and cutoff drains. Adoption of any of these whereas a gender-and-development approach recog- techniques can be costly, either directly in investment nizes both the outcome and its consequences as well requirements or indirectly in production forgone, and as the process leading to those outcomes. The under- some measures are better suited to some conditions standing gained from the gender-and-development than to others. The critical question facing farmers, approach facilitates understanding not only how a and society as a whole, is whether the benefits of a planned project or policy will affect men and women given conservation measure or set of measures are but also how it will affect the underlying processes sufficient to make the costs worth bearing. that condition the allocation of rights, resources, and Decisions about land use are ultimately made by responsibilities within communities and households. the farmers themselves in light of their own objec- Projects can integrate women's concerns by main- tives, production possibilities, and constraints. strearning them or by given them an identifiable bud- Understanding the incentives facing individual farm- get and reporting structure within a project. The ers is necessary, therefore, if patterns of resource use appropriate choice depends on the community's char- are to be understood and appropriate responses to acteristics and the project's goals. problems formulated. Farmers seek to maximize the present value of the stream of expected net returns to Appropriate Technologies agricultural production. With regard to adoption of conservation rmeasures, the issue is whether returns Sustainable rural development must be technically under the optimal path of the new, more conserving sound and innovative and must consider environmen- system are sufficiently greater than returns under the tal and social impacts. In order to move closer toward path of the cuirent, more degrading system to justify environmentally and socially sustainable rural devel- the cost of switching. opment, much technical innovation is needed. Estimated productivity losses of soil erosion and Farmers have been innovating under difficult condi- degradation vary considerably, as do the estimated tions and often under poor incentive frameworks. A effects on yields of conservation practices. Returns to question is how they can be better supported in their conservation (lepend on the specific agroecological efforts. As discussed in parts I and II of this volume, conditions faced, on the technologies used, and on the main support must come through conducive poli- the prices of inputs used and outputs produced. cy and institutional frameworks, including proper Generally, the farmers' decision to invest in conser- infrastructure. It must also come through appropriate vation is based on normal considerations of benefit research and empowering extension. and cost: farrmers tend to adopt conservation mea- Our technical discussion covered the following sures when it is in their interest to do so, unless some issues: soil conservation, mainstreaming of biodiver- constraint is present. Cases in which returns to con- sity, integrated pest management and other strategies servation are estimated to be low or negative corre- 354 Part IV Conclusions late well with low rates of adoption. Profitability of ural habitats for wild species and populations and to conservation practices is a necessary but not always manage habitats that have been modified for human a sufficient condition for their adoption. Factors other use, such as farmland. Agrobiodiversity is found in than strict cost-benefit considerations also play a habitats that have been modified for crop and live- role. Institutional and motivational issues also must stock production. Agrobiodiversity includes all be considered together with the results of the cost- plants and animals that contribute directly or indi- benefit analysis. rectly to the raising of crops and livestock. Advocates of soil conservation often argue that Although it is conceptually useful to differentiate subsidies are indispensable to induce farmers to agrobiodiversity from the larger array of species adopt conservation measures. But such statements and habitats, the boundaries between biodiversity and assume that conservation is inherently desirable agrobiodiversity are not clear-cut. The interface whether or not there is concrete evidence that the between wild and domesticated plants and animals is benefits outweigh the costs. Lutz and others (chapter constantly shifting. This fact underscores the impor- 16) show that this may be far from the case; fre- tance of conserving as much biodiversity as feasible quently, the benefits of specific conservation tech- for agricultural development in the future. niques (such as mechanical structures) do not justify How agriculture can be intensified without dam- their costs. Unless there are important off-site effects aging biodiversity is a critical question for rural or the price signals received by farmers are signifi- development. Environmentally inappropriate intensi- cantly distorted, subsidies to induce adoption will not fication of agriculture has led to eutrophication of bring increased economic efficiency. lakes and estuaries, loss of soil micro-organisms, One important way for governments to help is to accelerated soil erosion, contamination of ground- make sure that constraints such as insecure tenure do water, and draining of wetlands. All of these activities not prevent farmers from adopting conservation mea- trigger a potentially dangerous loss of biodiversity. sures. Also, governments already conduct some A balanced conservation strategy includes in situ research on soil conservation and provide, through conservation (maintaining animal and plant genetic extension services, some assistance to farmers who resources in places where they occur naturally) as undertake conservation work. However, research in well as ex situ conservation (maintaining genetic experiment stations has tended to favor technical resources in seed or field gene banks). In situ conser- efficiency (including structural measures such as ter- vation of crops and livestock can be supported by (a) races) over cost-effectiveness. Further, government emphasizing the safeguarding of plants and animals extension work is often ineffective. In many cases, for the future improvement of agriculture as a new nongovernmental organizations, such as Vecinos dimension of existing wilderness parks and biologi- Mundiales in Central America, have proven to be cal reserves, (b) creating world heritage sites for more effective than government at stimulating tech- genes for agricultural development, (c) integrating nological development (see chapter 12). agrobiodiversity in ecotourism where appropriate, Some of the beneficial effects of conservation (d) finding markets for lesser-known crops and local may come through better moisture conservation and varieties under the motto "use it or lose it," and (e) better conservation of organic substances, both of finding ways for livestock owners to generate more which may also benefit soil biodiversity. Chapter 14 revenues from threatened breeds. Better in situ con- discusses biodiversity conservation with a focus on servation can also be supported by rapid agrobiodi- protected areas. But much of the world's biodiversity versity assessment teams (Srivastava and others, exists in human-managed or -modified systems chapter 17). (Srivastava and others, chapter 17). Biodiversity has The new vision for agricultural research adopts a two important dimensions: the genetic variation holistic approach that is more sensitive to environ- within species and populations and the preservation mental concerns, while still addressing the need to of habitat. The significance of variation within a boost yields and incomes of rural producers and care- species is not widely appreciated but is critical, par- takers of the land. It includes, but is not restricted to, ticularly for agriculture. The continued productivity (a) integrated pest management (IPM), (b) a participa- of existing crops and livestock hinges in large part on tory approach with farmers, (c) better use of farmer harnessing the genetic variation found within each knowledge, (d) greater support for research, develop- species. Habitat conservation seeks to safeguard nat- ment, and dissemination of lesser-known crops and Part IV Conclusions 355 animals, (e) support for research on new crops and ture has been under way for some time. This has been livestock, (f) greater sensitivity to the value of a associated with a wider shift in focus from agricul- mosaic of land uses, (g) greater diversity of habitats tural production per se to environmentally sustain- within land use systems, (h) greater reliance on recy- able management of production systems in which IPM cling of organic matter, (i) shift of the research focus is a critical elernent. Integrated pest management is a from individual traits to lifetime and herd productiv- key componentl of integrated farming practices that ity characteristics, (j) determination of the critical are based on an understanding of ecology and the number of breeds for conservation purposes, and (k) interaction between crops or animals and their pests, an effort to learn about genetic components of adap- as well as an understanding of the environments in tation in livestock. which pests operate. The notion of a new research paradigm has impli- Significant advances have been made in the last cations for institutional development and the explo- three decades in the development and dissemination ration of new ways of doing business. Innovative of crop varieties with resistance to the major cereal institutional arrangements would include more effec- pests. Much of the advance has come through the use tive partnerships among agricultural research centers, of conventional breeding approaches, although sub- NGOS, growers association, private companies stantial gains in the development of resistance could involved in the manufacture and sale of agricultural come through the use of modern biotechnology tools. technologies, universities and agricultural extension The introduction of varieties with host-plant resis- agencies, and development lending institutions. To tance has dramatically reduced the need for insecti- some degree, all of these partnerships are being cides in rice and maize and the need for fungicides in explored and tested. wheat. Generating the gains in agricultural productivity Given the extent of breeding efforts in producing necessary to secure food availability and livelihoods pest-resistant crop cultivars and evidence of insignif- in the developing world over the coming decades icant productivity benefits of pesticide use, why are requires an approach in which the intensification of high levels of agrochemicals still used in crop pro- agricultural systems is consistent with the conserva- duction? First, the dissemination of crop varieties tion of the natural resource base. This approach resistant to pest pressures has not been accompanied requires less reliance on the intensive use of external by extension messages on the reduced need for inputs and greater dependence on management skills pesticides. Second, farmers' pest control decisions, and location-specific knowledge of agroecosystems. scientists' research priorities, and policymakers' pre- Integrated pest management constitutes one such scriptions, which are all based largely on perceived approach and is critical to sustainable rural develop- pest-related losses in yield (in turn often not related ment (Schillhorn and others, chapter 18; Pingali and to actual losses) have led to the promotion of pesti- Gerpacio, chapter 19). cide use. Consequently, high and injudicious applica- Agricultural intensification-the movement from tions of broad-.spectrum pesticides have continued as an extensive to an intensive production system or before, causing the breakdown in varietal resistance. from a subsistence production system to a commer- New varieties generated to replace cultivars with cial one-generally has increased the use of agro- resistance breakdown have subsequently been over- chemicals. Indiscriminate use of pesticides in the come through further biotype changes in pest popula- production of various crops has often impaired health tions. The breeding treadmill could only be overcome due to direct or indirect exposure to hazardous chem- through dramatic changes in crop management prac- icals; contaminated ground and surface waters tices, especialLy in the use of pesticides. through runoff and seepage; transmitted pesticide Under low levels of pest infestation, natural con- residues through the food chain; increased resistance trol is the economically dominant strategy for pest of pest populations to pesticides, thereby reducing management. (This is true for insects and diseases for their efficacy and causing pest outbreaks; and rice, wheat, and maize.) Natural control relies on reduced the population of beneficial insects ("good predator populations to control pest infestations bugs" like parasites and predators), thereby reducing under normal circumstances, when pest-resistant the effectiveness of pest control strategies that varieties are used. Pesticides may have to be used as attempt to minimize pesticide use. a last resort in the rare instance of high infestations. A reappraisal of the role of pesticides in agricul- Natural control does not imply doing nothing; 356 Part 1V Conclusions rather, it requires in-depth farmer knowledge of the the first 40 days of crop growth." The rule is based on pest-predator ecology and frequent monitoring of detailed pest ecology studies showing that the pre- field conditions by the farmer. In this regard, natural dominant insect pests during the first 40 days of crop control can be considered the ultimate goal of an IPM growth are leaf-feeding insects and that even very program, and farmers who are well versed in IPM high levels of infestations rarely lead to any loss in techniques would converge toward it (Pingali and yield (Pingali and Gerpacio, chapter 19). Gerpacio, chapter 19). This implies a paradigm shift Even with a well-established IPM program, pesti- in the traditional IPM strategies from when best to cides may have to be kept as a technology of last apply pesticides to when not to apply. resort. Essentially, the idea of pesticide use in IPM is IPM is knowledge-based; it presents by far the to spray only when imperative, using the smallest most difficult challenge to traditional, small-scale amount possible to do the job. The one area where farmers in the developing countries who are making agrochemicals continue to have a significant impact the transition to scientific farming. IPM requires farm- on productivity is in the management of weeds; her- ers to grasp complex sets of data that are often any- bicides will continue to be the preferred alternative in thing but self-evident, unitary, standardized, or the foreseeable future, even when the health costs of amenable to trial-and-error learning. The institutional herbicides are explicitly taken into account. and economic structure in the rural sector of devel- A number of policy and regulatory instruments oping economies also requires some policy interven- are available to governments to encourage environ- tion to reconcile long-term societal goals with short- mentally sound and economically rational pest man- term individual objectives in pest control. Promoting agement practices. The more important of these are sustainable pest management within an IPM frame- (a) development of a system that increases the aware- work requires improved research and extension link- ness of policymakers, consumers, and producers of ages, effective farmer training methods, community the hazards of pesticide use, (b) development of a action, and an undistorted price structure. regulatory framework to ensure appropriate and safe The IPM concept is holistic; it requires farmers to production, distribution, and use of pesticides, (c) take a systems view of the farm enterprise and to introduction of appropriate economic incentives, understand the interlinkages among various compo- including taxes and special levies on pesticide use, to nents of the system. Therefore, farner training in sus- account for negative externalities, and short-term sub- tainable pest management is an essential component sidies, to account for positive externalities in the use of a strategy to achieve minimum use of insecticides. of IPM, (d) orientation of research and technology The eventual goal is to build farmer capacity to iden- policies to generate a steady supply of relevant pest tify and solve problems based on a thorough under- management information and technologies, including standing of field ecology. The experience of the Food adequate budget allocations for research, extension, and Agriculture Organization's farmer field schools and training, and (e) signing of and adherence to has shown that trained farmers use significantly international agreements and conventions (Schillhorn lower levels of pesticides than untrained farmers. and others, chapter 18). They are also more likely to experiment with other Over the past three decades, the dramatic spread components of sustainable production systems, such of irrigated agriculture was mainly responsible for as improved fertilizer management and more effi- increases in food production that kept pace with cient water management. growth in population. This was achieved in part Training millions of farmers will be costly. In that through the large-scale development of new water regard, there are essentially two options, which are resources and construction of new irrigation capacity not mutually exclusive, for reducing training costs. (Chakravorty, chapter 19). The International The first is to train a core group of farmers within a Irrigation Management Institute has estimated that to geopolitical unit, such as a municipality, and then to meet future demand, irrigated agriculture may need rely on farmer-to-farmer training for disseminating to deliver output increases of more than 3.5 percent the IPM message to a wider group of farmers. The sec- annually-a daunting task by any measure of perfor- ond is to condense the complex message into several mance. The major challenge is that this growth must simple rules that a;re easy for the farner to imple- come primarily from increases in irrigation efficien- ment. An example, from rice, of such a rule is "Do cy and not from additions to new irrigation capacity. not spray insecticides against leaf-feeding insects for Moreover, as is becoming increasingly common in Part IV Conclusions 357 the industrial countries and even in many developing Partly because of budgetary pressures, some countries, rising demand for residential use of water developing countries are more willing to undertake and rising environmental concerns are beginning to institutional reforms in an effort to improve the pric- put serious limitations on new water development ing and trading of water. New "holistic" thinking is projects. Producing more food with less water-plac- also reflected in an emerging emphasis on "basin- ing a smaller financial burden on the taxpayer and wide" management promoted by agencies such as the reducing environmental costs-will be the major International Irrigation M[anagement Institute challenge for irrigation in the twenty-first century. (Chakravorty, chapter 20). These agencies argue that Irrigation projects around the world are in trouble water efficiency must be measured not within a proj- for a number of reasons: ex post project benefits have ect but over an entire basin so as to include the reuse often been far lower than projected ex ante returns; of drainage water from seepage and percolation. This full-cost pricing of water is rare, with most projects suggests that policies such as the adoption of sprin- recouping only a fraction of their operation and kler irrigation within an irrigation project may maintenance (o&M) costs; low tax collections and poor improve the efficiency of water use within the project o&M of project structures have led to high rates of but reduce the availability of water elsewhere in the water loss through seepage and percolation; and basin. Similarly, salt and chemical pollution of water inadequate investments in drainage have rendered through application of inputs may adversely affect large tracts of prime agricultural land unusable agriculture downstream of the project. These issues because of salinity and waterlogging. The low price can only be handled if the basin is adopted as the rel- of irrigation water, which is often unrelated to use, evant unit of analysis and the relevant externalities has led farmers to withdraw too much, giving rise to are internalized. salinity and waterlogging. Investments in the irriga- Livestock have often been seen as being a signif- tion sector in developing countries have relied too icant cause of environmental degradation. Indeed, much on bureaucratic federal and state agencies that they can damage the global natural resources in a have paid little attention to the economic pricing of number of ways, but they can also contribute to envi- irrigation services, the creation of reliable delivery ronmental balance (Steinfeld and others, chapter 21). systems, and the participation of users in the opera- About 34 million square kilometers, or 26 percent of tion and maintenance of projects. Adverse health and the world's land area, are used for grazing livestock. environmental effects of water use have received lit- In addition, 3 mnillion square kilometers, or about 21 tle attention, partly because of a lack of coordination percent of the world's arable area, are used for pro- and interaction between water and public health ducing cereal for livestock feed. Livestock, in turn, agencies. generate 13 billion tons of waste per year. A large These problems call for a new approach that is part of this is recycled, but where animal concentra- more comprehensive and integrates the elements of tions are high, waste poses an enormous environ- irrigation policy, which include the economic and mental hazard. environmental effects and the associated intersectoral Livestock grazing can affect the water balance in linkages (Chakravorty, chapter 20). This approach certain areas. 'Water is needed to produce fodder and must be built on pricing of resources that better feed concentrate, to supply drinking water for ani- reflects opportunity costs whenever possible and cre- mals, and to drain surplus waste and chemicals. ation of delivery infrastructure that promotes the effi- Livestock interact directly and indirectly with biodi- cient allocation and, if feasible, trading of the versity, and livestock and livestock waste cause resource among users. gaseous emissions with important local and global The introduction of markets that allow for trading impact on the environment. of water among users would lead to a more efficient Policies need to be designed to correct the nega- allocation of the resource. But water markets only tive environmental effects of livestock production. work under very selective conditions: (a) the pres- These policies should address the underlying causes ence of effective water users associations or water of environmental degradation and must be flexible, authorities, (b) the removal of legal and institutional site-specific, and well targeted. Instruments to regulations that prohibit or limit water trading, and enhance positive and mitigate negative environmen- (c) the creation of infrastructure (canals) and control tal impacts include pricing, regulations, and institu- systems for the transfer and measurement of water. tional development. The collective purpose of these 358 Part IV Conclusions instruments is to establish feedback mechanisms to the majority of forest cover is being removed ensure that the use of livestock is consistent with (Kaimowitz and others, chapter 22). In the period overall social objectives. 1980-90, 137.3 million hectares of tropical forests Within an enabling policy framework, a wide were cleared, about 7.2 percent of the total that exist- range of available technologies can be employed, ed in 1980. Any claims regarding the magnitude of while others still need to be developed. Technologies global deforestation must be taken with caution, can be grouped into four sets (Steinfeld and others, however, because there are serious problems with the chapter 21). Although there is some overlap among data and definitions used. categories, they help to provide a good picture of Deforestation often implies: livestock-resource interactions: * The loss of livelihoods for forest-dependent peo- * Technologies that reduce the environmental dam- ple, many of whom may not wish to or be able to age by alleviating the direct pressure on natural find other sources of employment resources or by reducing the pollution load * Decreasing stocks of fuelwood and nontimber for- through modifying the chemical or physical char- est products as well as of industrial timber acteristics of products * Greater soil erosion and river siltation * Technologies that enhance natural resources by * Substantial loss of species and genes, in view of making them more productive or richer the high level of endemic biodiversity in tropical * Technologies that save natural resources by allow- forests ing farmers to generate more revenue from the * Substantial emission of carbon dioxide, which same resource or to generate the same from less contributes to global warming * Technologies that turn waste into products by * Other types of local and regional climate change. closing cycles (a point also made by Pretty, chap- The concern is not only with deforestation but ter 4). also with forest degradation, which can be defined as Novel concepts are being developed to integrate a decrease of density or increase of disturbance in crop and livestock production in a farming area forest classes. In the long run, this may be just as rather than on a mixed farm (Steinfeld and others, important as deforestation itself. chapter 21). This method of area-wide crop-livestock The process of deforestation must be analyzed at integration allows individual enterprises to operate two levels: agents and causes. The agents of defor- separately while linking the flows of energy and estation are the people who physically (or through organic and mineral matter through markets and reg- decisions over their labor forces) convert forests to ulations. This allows for highest efficiencies at the nonforest uses: small farmers, plantation and estate enterprise level, while maximizing social benefits. owners, forest concessionaires, infrastructure con- Thus the motto for the more densely populated parts struction agencies, and so forth. In Latin America, of the developing world is to intensify, but not con- most deforestation is attributed to medium- and centrate, animal production. large-scale operations (resettlement schemes, large- As a result of the interaction between livestock scale cattle ranching, hydroelectric dams), and defor- production systems and natural resources, coupled estation is characterized by transitions from closed with factors such as market access, there are devel- forest to nonforest land uses. In Africa, deforestation opment opportunities as well as threats to sustain- is largely related to the expansion of small-scale ability. A comprehensive perspective is needed to farming and rural population pressure (growing num- ensure an enabling policy framework in which to ber of smallholders) associated with the conversion introduce effective technologies (Steinfeld and oth- from closed forest cover to short fallow farming. In ers, chapter 21). Technology remains a key compo- Asia, deforestation is associated with both relatively nent because future development, including that of large operations (as in Latin America) and rural pop- the livestock sector, will depend on technology to ulation pressure (as in Africa). There are no simple, substitute for natural resources. This trend to knowl- universal single-cause explanations. Different agents edge-intensive systems has been widely observed. coexist and closely interact in many countries, and Expansion of livestock production has often been their relative importance varies over time and between named as one of the main causes of tropical defor- regions. Given such complex relationships, there are estation. Contemporary concern about deforestation no clear guilty or innocent parties, and no one should focuses on tropical countries because that is where expect neat, simple solutions. Part IV Conclusions 359 Individuals and businesses deforest because it is products that have been sustainably produced their most profitable alternative. To get them not to from forests deforest in situations where forest clearing is inap- * Facilitate recognition of and compensation for propriate, deforestation must be made less profitable, environmental services provided by forests and or other alternatives (either based on retaining forests ensure that transfer payments are received by the or completely outside forest areas) must be made persons making decisions at the forest frontiers. more profitable. This is the main thesis of Kaimowitz One response to deforestation has been to encour- and others (chapter 22). A key question is how certain age agroforestry: a land use system in which trees, causes can be manipulated to influence the behavior shrubs, palms, and bamboos are cultivated on the of agents, so as to lessen the rate of inappropriate same land as agricultural crops or livestock for eco- deforestation. nomic and environmental reasons. The system There is no perfect or generalizable policy for includes manag.ing natural regirowth, seeding, plant- reducing inappropriate deforestation (Kaimowitz and ing, and maintaining trees as border plantings, and others, chapter 22). Each national situation is differ- interplanting them in agricultural crops, in woodlots, ent, much uncertainty remains about key cause-and- in home gardens, or in other systems. Agroforestry's effect relations, and there are usually tradeoffs special characteristics are that it includes a large among effectiveness, ability to be targeted, political number of species, configurations, and management viability, and direct and indirect costs of policies. intensities, has longer gestation than most agricultur- Most of the policy instruments are very blunt instru- al crops, and produces outputs with multiple uses. ments for stopping deforestation, and governments In addition to the estimated financial return, farm- will be forced to choose from a mix of measures, ers attach considerable importance to how an agro- specifically crafted to local conditions. forestry system fits into the overall farm production The answer to the issue of forest conversion is not system and the existing land, labor, and capital con- necessarily to stop logging per se, or to stop logging straints (Current and others, chapter 23). Many agro- in all new areas, or to ban all new road construction forestry systems are profitable to farmers under a in forest areas, but rather to reform those policies and considerable range of economic conditions, and vari- institutions that make forest colonization seem more ous types of (low-intensity) traditional agroforestry attractive than the potential migrants' current activi- are practiced in many areas. Even when agroforestry ties (Kaimowitz and others, chapter 22). These might is profitable, smallholders will and should adopt include the pull factors (to reduce the profitability of agroforestry incrementally and gradually because of illegally clearing forests or of speculating in land that management anid resource constraints; poorer farm- was supposed to be kept as forest) or the push factors ers, in particular, are often hampered by limited land, (to increase the livelihood options outside of forests). labor, and capital resources and by their need to Thus the principal reforms that could reduce inap- ensure food security and reduce risks. propriate tropical deforestation are likely to be a Local scarcity of wood products is, as might be combination of the following government policies: expected, a key motivator in adopting nontraditional * Eliminate subsidies to agricultural and pastoral agroforestry. Projects must begin by assessing the industries that encourage deforestation scarcity of wood and nontimber products, as well as * Eliminate legal and institutional incentives or the existence of local markets for products. Taungya, requirements to clear forests as a basis for gaining perennial intercrops, trees on contours, and tree lines recognized land tenure have proved to be the easiest systems to introduce. * Reform forest industry concessions and licenses Results are mixed for alley cropping, home gardens, to provide incentives for long-term sustainable windbreaks, graen manuring, dispersed trees in crop- management land, and tree-pasture systems. Farmers are willing to * Develop innovative institutional arrangements invest in rehabilitating their land where systems also for devolving more decisionmaking authority produce products or income, and they prefer less- and responsibility to those whose livelihoods and intensive systems. quality of life are linked directly to the extent The demonstration effect of fast-growing tree and quality of tropical forests species on farms, and of benefits on demonstration * Encourage voluntary market differentiation by plots, has helped to expand agroforestry activities, consumers that discriminates positively toward reducing the costs of extension and increasing its 360 Part IV Conclusions effectiveness. Rather than offering standard designs, reduced subsidization. Yet despite these improve- programs serve farmers best if they offer a broad ments, many countries still implement a variety of selection of species and systems from which to policies that directly or indirectly favor the use of pes- choose those most suitable to their household's needs ticides and impede the adoption of ecologically and resources. Involving local people as paratechni- sound pest control techniques. Negative side effects cians is often a successful, low-cost approach to pro- of pesticides with medium and high toxicity include moting technology. groundwater contamination, chronic or acute health Financial incentives and subsidies should be kept problems for farmers and farmworkers, losses of to a minimum. Agroforestry technologies promoted nontarget crops and other species, pesticide resis- by extension should be financially profitable, and tance, and pesticide residues in food. Subsidization of thus adoptable, for the farmer without subsidies. The water, mainly in the form of a lack of cost recovery, possible exceptions are time-limited financial incen- also remains a problem with serious long-term con- tives for early aclopters of unfamiliar technologies. sequences. Only a handful of countries have made In-kind, material inputs encourage farmers to experi- strides in increasing cost recovery and moving ment with and adopt agroforestry, but the experiences toward full-cost pricing, which would help to with food-for-work incentives are mixed. improve management, reduce waste, reduce overuse, Implementing many of the needed changes to and thereby reduce the problems of misallocation, move toward more sustainable rural development waterlogging, and salinity. Finally, over the past will not be easy. As discussed in this volume, it will decade or two, many parastatals and marketing require significant efforts by farmers and many oth- monopolies have been abolished or reformed. Yet ers on many fronts. It will also require a greater com- many still remain. Such intervention often results in miitment to the goal of sustainable agricultural devel- welfare losses for producers or consumers or both, opment by many national governments (which often marketing operations may not be efficient, and there have an urban bias) as well as by development assis- is potential for misuse of funds. tance agencies. 2. In many developing countries, women have lower rates of school enrollment, literacy, and Notes schooling attainment. This disparity continues to be larger in rural areas, where low educational attain- 1. Examples of necessary reforms can be given for ment persists despite high private rates of return many subject areas and for many countries. With to women's schooling and high social returns to regard to pesticide subsidies, many countries have women's education. Index Action threshold of pest management, 244 development Africa: animal draft in, 288; deforestation agents/ Agricultural extensification. See Extensive farming causes in, 211, 303, 358; gender issues in, 188-89, systems 190; 1980-90 deforestation in, 12, 302; pesticide Agricultural intensification: biodiversity conservation regulation in, 251; poverty-degradation linkage in, approach to, 231, 233, 235, 236-37 table, 354, 355; 51; resource-conserving technologies in, 37; tran- environmental impact of, 7, 9, 11, 12, 13, 93, 205, sitioning property regimes in, 62, 312 239; and high/low market orientation, 256-57, 258 African National Congress, 173 n22 table, 259; Eand indiscriminate pesticide use, 11, Agency outreach: client-centered features of, 96, 206, 230-31, 244, 254, 256-57, 355; positive 109-11, 110 table, 348; community people's role effects of, in Machakos, 286 box; reduced defor- in, 107-8, 122-24, 126-27, 349; empowennent estation with, 308-9, 319; transition from extensive versus extension approach of, 96, 107, 348; by system to, 16. See also Irrigated agriculture enabling policies for, 96, 111-12; field agents's Agricultural productivity: agroforestry's integration role in, 108-9; and godfather/godmother role, 112; with, 214; Boserup on, 22-23, 58; canal water to mediate decentralization, 178, 183; social assess- management for, 273; in Colombia 1950-87, 25, ment tools of, 111, 113; watershed management role 26 table; of Colombia's small farmers, 27-28, 30, of, 97, 122-24, 349; watershed management role 171 nl; farm-level perspective on, 204, 217, 219; of, in Rio Cabuyal, 125-27. See also Government future demands on, 9, 35; gender issues tied to, organizations; Nongovernmental organizations 101, 188-90, 189 table, 352; genetic variation's Agha Khan Rural Support Program, 42, 63, 104, ties to, 205, 230, 232, 234 box, 354; in irrigated 122-23 agriculture systems, 3, 11, 271, 356-57; in Israel, Agricultural development: for backward regions, 93; 277; land redistribution's impact on, 100, 156; decentralized/participatory approach to, ix-x, 1-2, land tenure tied to, 6-7, 190, 293, 351; with peo- 348; five views on future of, 4, 35-36; focus of, in ple-centered development, 145, 151-52; in rainfed 1950s and 1960s, 7, 92; focus of, in 1970s and systems, 277-78; relation of farm size to, 57, 351; 1980s, 7, 92-93; four principal goals of, 7, 92; in resource-poor areas, 16; salinization effects on, property relations central to, 6-7, 83, 86, 87, 347; 278; with single versus multiple outputs, 191-92; sustainability failures in, 40; sustainability modi- soil conseriation effects on, 204, 222-23; soil fiers for, 7, 9-10, 93-94. See also Community-dri- degradation's impact on, 203, 204, 215-16, ven development projects; People-centered agri- 220-22, 221 table, 225, 353; soil factors of, 13. cultural development; Sustainable agricultural See also Agricultural intensification; Productive 361 362 Index projects Alicante (Spain), 274-75 Agricultural research: aid community's concerns Allen, Jennifer, 13 about, 5-6, 71, 79; holistic approach of, 206, 231, Alley cropping, 40, 331, 331 table, 332, 334, 335 239-40, 354-55; on host-plant resistance, 261-63; table monoculture approach of, 244, 259; multiagency Altaiski (Russia), 180 approaches to, 98. 131, 350; in multipurpose tree Amazonian product marketing, 233, 234 box plantings, 326; needed in rainfed areas, 14, 98, Amboseli National Park (Kenya), 179, 180 130; predicted public funding of, 76; rates of Andean countries: cattle sector in, 26; peasant farm- return on, 18-19, 80, 130; reform/investment ing in, 24, 33 n2 needs of, 18-19; regulatory influence on, 248-49; Anderson, Jock R., 23 on soil conservation, 226-27; trade policy's stimu- Anderson, Kym, 1, 5, 6, 79 lation of, 72, 79. See also Participatory technology Animal draft, 288 generation Antle, J. M., 267 AGRITEX (Department of Agricultural, Technical, Aral sea region, 279 and Extension Services, Zimbabwe), 108 Arfak project (Irian Jaya, Indonesia), 178, 180 Agrobiodiversity: assessments of, prior to projects, Arid rangelands: characteristics/assets of, 88-89, 285, 205-6, 238; biodiversity's interface with, 232 fig., 286-87; technologies for, 295; UNEP on degrada- 233, 354; business community's marketing of, tion of, 284 233-35, 234 box; defined, 205; ecotourism's inte- Arnold, J. E. M., 331, 333 gration of, 235; with mixed land uses, 240; policies ASEAN (Association of Southeast Asian Nation), 77 to conserve, 239; of tropical rain forests, 232 box Ashby, Jacqueline A., 2, 96-97 Agrochemicals. See Pesticides Asia: deforestation agents/causes in, 211-12, 303, Agroforestry: country case studies of, 324-25, 325 358; insecticide usage in, 255, 256, 256 table, 257 fig.; definition/characteristics of, 213-14, 323, table; 1980-90 deforestation in, 12, 302; rainfed 359; environmental/social benefits of, 337-38; share of land in, 277; technologies in, 37 extension strategies for, 328-29 table, 330, Asian Development Bank, 274 336-37, 339; farmer training in, 214, 337, 339; Asociaci6n de Consejeros para una Agricultura fuelwood projects of, 214, 325-26; incentives to Sostenible, Ecol6gica y Humana (COSECHA), 41, promote, 214, 326, 330, 333, 337, 360; institution- 148, 151, 154 al responsibility for, 340; land tenure constraints Association of Southeast Asian Nation (ASEAN), 77 on, 335-36, 340; management factors in, 334, 339; Asymmetric information, 5, 51 multipurpose tree projects of, 214, 326; participa- Australian Landcare program, 124 tory research in, 141 box; profitability issues of, 214, 323-24, 326-27, 338, 359; risk factors in, Babatunde, Morakinyo, 175 332-33; smallllarge farmer adoption of, 334-35, Bacillus thuringiensis (Bt) gene, 262-63 339; tree-crop interactions of, 334, 335 table. See Balcazar, Alvaro, 171 nl also Agroforestry costs-benefits analysis Balochistan Primary Education Project (Pakistan), Agroforestry costs-benefits analysis: country case 105, 109, 111-12 studies of, 324-25, 325 fig.; of household benefits, Banerjee, Ajit, 175, 179 326, 330; of individual agroforestry systems, 331 Bangladesh, 75, 192, 193, 260 table, 331-32, 339, 340 nn2-3, 359; of manage- Bangladesh Forest Industries Development ment operations, 334, 339; of pace/scale of adop- Corporation, 310 tion, 334-35; prior research's neglect of, 323-24; Barbier, Edward B., 23 project characteristics of, 327 table, 330; and pro- Barnum, H. N., 58 ject extension strategies, 328-29 table, 330, Barva (Costa Rica), 223 336-37, 339; of tree impact on crops, 332-33; of Basic Education Project (Yemen), 109 tree products, 331-32, 338 Basinwide management of water, 209, 280-81, 357 Agroforestry Outreach Project (Haiti), 40 Bates, J. M., 80 Aid community, research funding concerns of, 5-6, Bean yields, in San Martin, 152 71, 79 Bebbington, A. J., 131, 133 Alderman, Harold, 188 Beghin, John, 78, 80 Index 363 Beneficiaries: farm size requirement for, 159, 160, Black spot removal policies (South Africa), 168-69 164, 172 n9; financing arrangements for, 158-59, Bolivia: people-centered development in, 150; RELU 160, 165-66, 171 n3; INCORA's selection of, 162, in, 99, 139-41 172 n13; INCRA's selection of, 168, 172-73 n18; Boserup, Ester, 1031, 186 of land-restricted grants, 158-59, 172 n5; market Boserup effects: incentive structure tied to, 4, 5, 23, access needs of, 171, 173 n24; of municipal land 346; in Machakos versus Ethiopia, 23; of popula- reform, 162-63, 166, 166 table; productive pro- tion growth/market access, 22-23; of productivi- jects of, 163-64, 172 n1S; rental restriction on, ty/market access, 58 159, 172 nlO; of South African land reform, 169, Botswana, 23 170, 173 n23; transparent land markets for, 170. Braverman, Avishay, 51, 52 See also Farmers; Small farmers Brazil: colonization schemes in, 309; decentralized Berry, R. A., 57 land reform in, 171; female expenditures in, 193; Binswanger, Hans P., 1, 3-4, 5, 52, 55, 100, 351 negotiated land reform in, 100-101, 161 table, Biodiversity: agricultural system's dependence on, 167-68, 173 n19, 351; nontimber forest products 205, 230, 232, 234 box; of arid versus humid of, 315; people-centered development in, 154 rangeland, 287; assessments of, prior to projects, British Department for International Development 205-6, 238-39; business community's marketing (fonnerly Overseas Development Administration), of, 233-35, 234 box, 238; genetic variation dimen- 40, 139 sion of, 205, 232, 354; habitat conservation dimen- Bromley, Daniel W., 1, 6, 7, 88 sion of, 205, 232, 354; and inappropriate defor- Brown planthopper (Nilaparvata lugens), 250 box, estation, 303-4, 320 nn2-3; species richness 259 dimension of, 232; weak local awareness of, 177, Brundtland Commission, 36 182. See also Agrobiodiversity; Biodiversity con- Bt (Bacillus thuringiensis) gene, 262-63 servation Bullock stocks, 55, 56 Biodiversity conservation: agriculture's incorporation Bunch, Roland, 2, 41-42, 145 of, 231, 233, 235, 236-37 table, 354, 355; with Burkina Faso: gender/efficiency study in, 188-89; biodiversity prospecting, 316; business communi- PATECORE Program in, 146, 153 ty's marketing of, 233-35, 234 box, 238; causes of Business-as-usual optimists, 4, 35 problems in, 182; central government's role in, Butz, Earl, 38 101, 175-76, 352; conventional versus holistic Byerlee, Derek, 79 approach to, 231 box, 233, 235, 238; costs/benefits Byrne, P. V., 287 of, 176, 182; country case studies of, 175, 177 table, 177-78, 184 nn2-3; decentralized approach Cacha, Maria Dulce, 175 to, 2, 101, 352; enforcement of, 180-81, 182, 352; Cairns Group, 80-81 ex situ strategies for, 205; global net benefit of, Caja Agraria (Colombia), 33 n3, 160, 165, 166, 172 176-77; incentive structures for, 180-81, 182, nlS 352; in situ strategies for, 205-6, 235, 238, 354; Calakmul program (Mexico), 152 local benefits of, 183-84; local participation/ Calamai, P., 14 agreement on, 180, 181-82, 352; past failures in, Caldecott, Julian, 2, 101, 175 101, 175; scientific expertise required for, 181; California Department of Water Resources, 280-81 sustainability's reliance on, 2, 182-84. See also California water trading, 275 Biodiversity; Integrated pest management Cameroon study, 199 nl Biodiversity Enterprise Fund for Latin America, Campesino a Campesino (Farmer to Farmer), 151, 233-34 153, 154 Biogas systems, 296, 296 box CAMPFIRE (Zimbabwe). See Communal Areas Biological oxygen demand (BOD), 210 Management Programme for Indigenous Resources Bioregion: conservation enforcement by, 181, 184 n5; Cantarranas program (Honduras), 41, 42 table, 151 risks to, when decentralized, 183 Capital gains taKes, to discourage deforestation, 313, Biotechnology research, 35, 261-63 313 table Bishop, Joshua T., 13, 23 Capital stock: assessing depreciation of, 13; categories Biswas, Asit K., 272 of, 3, 345; substitutions versus transformations of, 364 Index 10, 19 nl; sustainability issues for, 9-10, 345 Centro Agron6mico Tropical de Investigaci6n y Carbon dioxide emissions: from deforestation fires, Ensenianza. See Tropical Agricultural Research 291; from livestock, 291; transfer payments for and Higher Education Center storage of, 3]L6, 317 table, 320 n8 Centro de Investigaci6n en Agricultura Tropical CARE (Cooperative for American Relief (Bolivia, CIAT), 99, 139, 140, 141 Everywhere), 151 Centro Internacional de Mejoramiento de Mafz y Caribbean: agroforestry case studies in, 214, 325 fig.; Trigo (CIMMYT), 261 cattle sector in, 26; 1981-90 deforestation in, 323, Cereal crops: and feed-food controversy, 36, 289 box; 324 table; soil conservation costs-benefits in, 218 high-yield production of, 259; livestock consump- table; soil erosion rates in, 216 table. See also tion of, 36, 283, 290, 357; 1990-92 global produc- Agroforestry costs-benefits analysis; Soil conser- tion of, 290; 1961-91 in developing countries, 75; vation costs-benefits analysis pesticide use on, 254, 255, 260; pest-resistance of, Carlson, Gerald, 267 206-7, 260 table, 261-63; soil depletion effects of, Carrying capacity: and soil degradation, 23, 33 nl; 290. See also Maize; Rice; Wheat and unsustainability assumptions, 22 CGIAR. See Consultative Group for International Carson, Rachel, 246 Agricultural Research Carter, M. R., 57 Chakravorty, Ujjayant, 2, 273, 275 Case studies: of agroforestry, 324-25, 325 fig., 327 Chalfant, J. A., 59-60 table, 328-29 table; of biodiversity conservation, Chavas, Jean-Paul, 88 175, 177 table, 177-78, 184 nn2-3; of microwa- Chayanov, A. V., 58 tershed rehabilitation, 134-37; of participatory Chemical industry. See Pesticide industry technology generation, 98-99, 133-37; of RELU Cheung, S. N. S., 56 in Bolivia, 139-41; of soil degradation, 218 table, Child, Brian, 175 220; of watershed management, 125-28. See also China: Asian share of pesticides in, 255, 256, 256 Agroforestry costs-benefits analysis; Soil conser- table; biogas systems in, 296, 296 box; deregulat- vation costs-benefits analysis ed pesticides in, 247; salinization problems in, 12; Cassia sp., 334 sustainable project in, 44 Catchment resource conservation (Kenya), 121 Chiriquf region (Panama), 216 Catholic Relief Services, 151 CIALS (local agricultural research conmmittees), 127 CATIE (Centro Agron6mico Tropical de Investigaci6n CIAT. See Centro de Investigaci6n en Agricultura y Ensefianza). See Tropical Agricultural Research Tropical and Higher Education Center CIFOR (Center for International Forestry Research), Cattle. See Livestock sector 315 Center for International Forestry Research (CIFOR), CIMMYT (Centro Internacional de Mejoramniento de 315 Mafz y Trigo), 261 Central America: agroforestry benefits in, 337-38; CIPASLA (Consorcio Interinstitucional para la agroforestry case studies in, 214, 324-25, 325 fig.; Agricultura Sostenible en Laderas), 126-27 1981-90 deforestation in, 323, 324 table; people- Cleaver, Frances, 111 centered development in, 99, 145, 146, 151; ranch- Cline, W. R., 57 ing-induced deforestation in, 287, 307; soil con- CMDR (Consejo Municipal de Desarrollo Rural), 158 servation programs in, 41, 42 table; soil erosion Cocaine revenues, 28-29 rates in, 215, 216 table; value of tree products in, Cocil Province (Panama), 221 331-32; yield of select crops in, 220-22, 221 Codex Alimentarius agreement, 249 table. See also Agroforestry costs-benefits analy- Coffee yields, in Costa Rica, 220, 221 table sis; Soil conservation costs-benefits analysis Cole, M. A., 80 Central government: biodiversity conservation role Collective action: based on common interest, 104-5; of, 101, 175-76, 352; conservation enforcement institutional design of, 18; IPM's dependence on, by, 180-81; failed land reform policies of, 100, 208, 266-67; leaders of, 65; local organizations' 157-58 provision of, 120, 122; reduced costs with, Central Institute for Agrarian Studies (Brazil), 168 122-23; required by transboundary effects, 119; Centro Agrfcola Cantonal (Costa Rica), 338 resource management through, 17-18; self-selec- Index 365 tion strategies for, 113-14. See also Local organi- management zations Consorcio Interinstitucional para la Agricultura Colombia: agricultural labor market in, 25, 25 fig., Sostenible en Laderas (CIPASLA), 126-27 29-30; cattle sector of, 26, 100; colonization poli- Consultative Group for International Agricultural cies of, 27; decentralized conservation in, 177 Research (CGIAR), x, 19, 252 table, 177-78; employment/wage differentials in, Consumption: production's ties to, 4-5, 51; and risk 29; failed migration policy in, 24-25; farm size in, diffusion mechanisms, 55-56; in separable versus 172 n9; farm-size/yields in, 27-28; fiscal decen- nonseparable models, 58 tralization in, 179; godmother-supported policy in, CONTAG (Federation of Rural Workers, Brazil), 168 112; irrigation project success in, 280; land prices Contour planting, 331, 331 table, 332, 335 table in, 28-29; land tenure changes in, 27, 27 table, Contract farming, and plantation production, 57 157; large-farmer bias in, 25-26, 28, 30, 33 n3, Cook, Cynthia, 333, 337 157; negotiated land reform in, 100-101, 158-59, Cooperative for American Relief Everywhere 161 table, 351; 1950-87 agricultural productivity (CARE), 151 in, 25, 26 table; resource degradation in, 24-25; Corn: pest-resistance of, 264; yields of, in Haiti, 220, rural violence disincentive in, 26, 30, 157. See also 221 table, 222 Instituto Nacional Colombiano de Reforma COSECHA. See Asociaci6n de Consejeros para una Agraria; Municipal land reform program; Agricultura Sostenible, Ecol6gica y Humana Negotiated land reform Costa Rica: agroforestry benefits in, 338; coffee Colonization policies: in Colombia, 27; and defor- yields in, 220, 221 table; decentralized conserva- estation, 213, 309, 318; reduced support of, 309, tion in, 177 cable, 177-78; fiscal decentralization 311 table, 319 in, 179; soil conservation measures in, 218 table, Comisi6n Nacional de Aguas (Mexican water author- 222-23, 223 table; soil degradation in, 13, 216; ity), 280 subsidization schemes in, 226; tree products of, Common property: deteriorated resources of, in India, 331, 333; value of windbreaks in, 332 286 box; to limit deforestation, 312-13, 313 table; Costs-benefits analysis of agroforestry. See local organizations' regulation of, 118, 120; priva- Agroforestry costs-benefits analysis tization of, 94; as resource degradation cause, 87, Costs-benefits analysis of soil conservation. See Soil 88; rights to, 16-17, 84 box; and watershed's conservation costs-benefits analysis transboundary effects, 119, 137 Cotton bollworm (Heliothis armigera), 247 Communal Areas Management Programme for Cotton crop, in China, 244 Indigenous Resources (CAMPFIRE, Zimbabwe), Covariate risk, 5; and credit/insurance markets, 54; in 178, 179, 184 n2 low population density areas, 51 Community-driven development projects: client-cen- Credit markets: for agroforestry, 337; failed policy tered agencies of, 96, 109-11, 110 table; enabling intervention in, 62-63; and gender issues, 187, policies for, 96, 111-12, 348; five characteristics 193; high covariance and, 54; inherited endow- of, 95-96, 103, 348; gender analysis in, 197-98; ments element of, 54; large-farmer bias of, 28; in godfather/godmother role in, 112; learning low-risk/high-risk villages, 54; moral hazard and, process approach to, 114-15; local organizational 54; for nonrecommended livestock, 239; and pes- capacity of, 95, 103-6, 113-14, 121; monitor- ticide use, 246; in surplus-producing areas, 64; and ing/evaluation of, 115; objectives for, 112-13; use of group lending, 63 outreach mechanisms of, 95-96, 107-9; the Crop insurance markets: as commercially unviable, poor's access to, 106, 113; social dimension of, 53-54; failed policy intervention in, 62-63 105, 111, 113; subsidy/demand element of, 114; Cross, Catherine, 173 n22 technology element of, 95, 106-7, 348; ten princi- Cross-compliances. See Interlinkage mechanisms ples for design of, 96, 112-15, 348-49; World Cross River National Park (Nigeria), 179 Bank strategies for, 96, 110. See also Local com- Cruz, Wilfrido, 307 munity; Local organizations Cuban IPM polizy, 250 Conable, Barber, 112 Cupuagu cocoa, 234 Consejo Municipal de Desarrollo Rural (CMDR), 158 Current, Dean, 2 Conservation management. See Natural resource Currie, Lauchlin, 24 366 Index Dahlman, Carl, 85 Direct use values, defined, 10 Daurski (Russia), 180 Diversion ditches, 204, 222, 223, 224 Decentralization: biodiversity case studies of, 175, Domestic farm prices: artificially low versus higher, 177 table, 177-78, 184 nn2-3; capacity-building's 5-6, 347; developing countries' reform of, 78-79, importance to, 181-82, 352; central government's 80; Uruguay Round's impact on, 77 role in, 101, 175-76, 352; in community-driven Dominican Republic: crop yields in, 221 table; soil development, 96, 111-12, 348; concerns/risks in, conservation costs-benefits in, 218 table; soil ero- 183; of fiscal decisionmaking, 101, 105, 179; of sion rates in, 216 table; tree products of, 331 IPM extension, 266; of livestock management, Dove, M. R., 315 292-93; local benefits from, 182, 183; mediating Drainage management, integrated approach to, body's role in, 178-79, 183; in negotiated land 278-79, 280, 357 reform, 100, 158, 159-60, 170-71; political/eco- Dregne, H., 284 nomic reasons for, 179-80, 182-83; redistribu- Dront Eco-Center (Russia), 178 tion/bargaining processes of, 176, 178, 184 nl; to Drug trafficking, 28-29 reduce deforestation, 213; in soil conservation, Dyck, V. A., 263 205, 226-27; in sustainable rural development, ix-x, 1-2, 38, 41, 43, 46, 348 East Asia, 317; animal draft in, 288; global share of Defensores de la Naturaleza Program (Guatemala), pesticides in, 255; intensification-agroforestry 152 connection in, 331 Deforestation: agents of, 211-12, 302-3, 358-59; Ecolabeling of timber, 315, 317 table agricultural extensification and, 308-9; agro- Economic threshold of pest management, 244 forestry response to, 213-14, 359; as appropriate Ecotourism, and agrobiodiversity, 235 versus inappropriate, 212, 303-4, 304 table, 320 Ecuador, 196, 227 n4 nn2-3; causes of, 212, 303, 358-59; in Central Education: and adoption of technologies, 149, 188, America and Caribbean, 41, 323, 324 table; in 190; and conservation enforcement, 182, 352; Colombia, 25; common property's limitation on, female levels of, 186, 187, 188, 360 n2; to reduce 312-13, 313 table; definitions/implications of, transfer costs, 149, 177 302, 358; economic discentives for, 212-13, Edwards, M., 132 304-5; labor/capital factors and, 58, 317; livestock Edwards plateau (Texas), 286 production and, 211, 287, 358; logging/coloniza- Egypt, 12, 276, 277 tion and, 213, 307, 309, 318-19, 320 n5; from El Naranjal watershed (Dominican Republic), 226 1980-90, 12, 211, 302, 319 nl, 358; relocating El Salvador: agroforestry projects in, 334, 335, 338; agricultural production and, 72-74, 75, 76; road fuelwood projects in, 325-26, 334; FUNDASAL construction and, 213, 310, 318; valuation cate- project in, 104, 106; land titling projects in, 224; gories for, 41. See also Agroforestry; Tropical for- soil erosion rates in, 216 table; tree products mar- est conservation policies ket in, 333 Deininger, Klaus., 2, 100 Empowerment approach: in decentralization process, De Janvry, Alain, 58-59 176; extension approach versus, 96, 107, 348; De Moor, A., 14 facilitation focus of, 108; goals of, 109; for local Department of Primary Industry (Australia), 42 decisionmaking/accountability, 108, 109; NGOs' Developing countries: expanding rural labor market use of, 98, 132; in people-centered development, in, 75-76; extensionist/farmer ratio in, 188; mar- 150, 150 fig. ginal-cost pricing in, 276-77; pesticide exporta- Empresa de Pesquisa Agropecuaria e Difuso de tion to, 255; pesticide regulation in, 251; pesticide Tecnologia (EPAGRI, Brazil), 122, 154 use in, 244-46, 245 table, 252 nl; population Enforcement of conservation: at bioregional level, growth in, 271; raising domestic prices in, 78-79, 181, 184 n5; defined, 184 n4; education context of, 80; trade liberalization effects in, 71-72, 80 182, 352; of endogenous versus exogenous rules, Development Alternatives, Inc., 151 181; local compliance with, 180-81 Development Assistance Committee of the English, J., 286 Organisation for Economic Co-operation and Environmental effects: agricultural productivity and, Development, 40 ix, 1, 9, 93; of agroforestry systems, 337-38; and Index 367 costs-benefits issues, 10-11, 337; in extensive reform and, 72. See also Environmental effects; farming systems, 7, 11-12, 93; in intensive farm- Off-site impacts ing systems, 7, 11, 93; of irrigation projects, 11, 208-9, 271, 278, 279; labor market transitions and, Facilitation model, in Indonesia, 252, 252 box. See 75-76; land usage changes and, 75; of livestock- also Participatory technology generation crop interaction, 287-90, 295; of livestock sector, Factor productivity, in family-based versus large- 209-10, 211, 284-87,358; of livestock waste, 283, scale sector, 27, 33 n4 291-92, 357; of pesticide misuse, 2, 246, 254, 355, Fafchamps, M., 58-59 360 nl; policy reinstrumentation and, 76-77; real Family-farm units: as "equity modifier," 93; factor income growth and, 72; relocating agricultural productivity of, in Colombia, 27, 33 n4; increased production and, 72-74, 74 fig.; trade liberalization optimal size of, 57; Law 160 on, 31-32. See also and, 5, 6, 71, 72, 78, 347; Uruguay Round and, Farm household models; Small farmers 77-78. See also Deforestation; Natural resource FAO. See United Nations Food and Agriculture degradation; Off-site impacts; Soil degradation Organization Environmental Kuznets curve, 306 Farmers: agroforestry adoption by, 326, 330, 334-35; Environmental pessimists, 4, 35-36 as agroforestry paratechnicians, 214, 337, 339; EPAGRI (Empresa de Pesquisa Agropecuaria e collective organizing of, 17-18; degradation Difuso de Tecnologia, Brazil), 122, 154 incentives to, 14, 43; and fuelwood projects, Eponou, T., 131 325-26; indigenous knowledge of, 238, 240; infor- Erosion. See Soil erosion mation needs of, 15-16; innovation capacities of, Escalada, M. M., 259 x, 1, 4, 41-12, 99, 111, 122, 127-28, 145; IPM Ethiopia: carrying capacity studies in, 23, 33 nl; con- incentives for, 242-43, 250-51; IPM's challenges servation investment in, 23-24, 61, 64; income to, 265, 356; pest management options for, 244, and discount rates in, 57 245-46; as risk-averse versus risk-neutral, 59-60; Eucalyptus camaldulensis, 334 small-scale experimentation by, 145, 146, 147, European Commission, 40 148, 153-54; soil conservation issues for, 204, European Union, 77, 172 n9, 247 217, 219; technology adaptation by, 40-41, 42 European Union Council Regulations, 76 table, 127-28, 145, 147, 148-49; trained as exten- Existence values, defined, 10 sionists, 99, 145-46, 151, 153; trained in IPM, Experimentation by farmers: in people-centered 207-8, 251--52, 266, 356. See also Beneficiaries; development, 145, 146, 147, 148; statistical analy- Large farmers; Resource-poor farmers; Small sis of, 153-54. See also Innovations farmers; Training of farmers Export bans/taxes, to limit deforestation, 307-8, 308 Farmer-to-farmer extension: Brazil's large-scale use table of, 154; multiplier effect of, 99, 146, 350; reduced Ex situ conservation strategies, 205 costs with, 151, 207, 266, 356; and selection crite- Extension approach: in agroforestry projects, 328-29 ria, 153; in Southeast Asia, 146 table, 330, 336-37, 339; decisionmaking/account- Farmer to Farmer program. See Campesino a ability in, 109; as empowering, in Mali, 108; Campesino empowerment approach versus, 96, 107, 348; gen- Farm household models: to analyze structural adjust- der issues of, 187-88, 190; to increase inputs, 108, ment effects, 59; with missing markets, 58-59; 187-88, 190; to induce conservation adoption, poverty/sustainability relationship in, 60; as sepa- 204-5, 354; to livestock production, 294; technical rable versus nonseparable, 58; yields/conservation instruction focus of, 108, 226, 354 relationship in, 61, 64. See also Family-farm units; Extensionists. See Farmer-to-farmer extension Small farmers Extensive farming systems: environmental problems Farm Science Center (Udaipur district), 133 in, 7, 11-12, 93, 308-9; subsidization of, 309; Farrington, John, 2, 98, 131, 133 transition to intensive systems from, 16. See also FEBESURCA. See Federaci6n de Beneficiarios de la Rainfed agriculture areas Subcuenca de Cabuyal Extemalities: addressed by incentives, 11, 15, 94; Feder, Gershon, 57, 62 local versus regional costs of, 10; and market price Federaci6n de Beneficiarios de la Subcuenca de distortions, 14-15; Pareto-relevant, 63; trade Cabuyal (FEBESURCA), 126, 127, 128 368 Index Federation of Rural Workers (CONTAG, Brazil), 168 and-development approach to, 197-98; incentive Female marriage contracts, to diffuse risk, 55 systems and, 186; income expenditures as, Fertilizers: agricultural PSE and, 74 fig.; natural 192-93, 193 table, 194-95 table, 196, 199 n4, 352; methods replaced by, 37; new modernists on, 36; property rights as, 186, 187, 190-92, 352; techni- production relocation, and use of, 74-75; in rain- cal efficiency as, 188-90, 189 table; unresolved by fed areas, 12; soil degradation effects on, 221; sub- participatory approach, 197, 353. See also Gender- sidization of, .14 and-development approach; Women; Women-in- Field agents of outreach programs, 108-9 development approach Finkelshtain, Israel, 59-60 General Agreement on Tariffs and Trade (GATT), 71, Fiscal decentralization, 101, 105, 179 77, 80. See also Uruguay Round Fish stock depletion, causes of, 12 Genetic diversity: deemphasis on, 244; tied to agri- FONAES (Solidarity Fund for Social Enterprises), cultural productivity, 205, 230, 232, 234 box, 354 107 Genetic engineering, 262-63 Food and Agriculture Organization of the United German Development Bank (Kreditanstalt fur Nations. See United Nations Food and Agriculture Wiederaufbau, KfW), 138 Organization German Technical Cooperation (Deutsche Food-for-work programs, 63, 64-65, 360 Gesellschaft fur technische Zusammenarbeit, Ford Foundation, 133 GTZ), 138, 151 Forest degradation, 12, 211, 302, 358. See also Germany, average farm size in, 172 n9 Deforestation Gerpacio, Roberta V., 2, 263-64 Forestry Department (India), 135, 137 Ghana, 193, 307 Forestry/Fuelwood Research and Development Gichuki, Francis, 23 Project (Asia), 333 Gliricida ssp., 334 Forestry Fuelwood Research and Development Global Trade Analysis Project (GTAP), 77-78 Project (Asia), 326 Government organizations: functional participation Fourth World Conference on Women (Beijing, 1995), approach of, 98, 132; IPM implementation role of, 196 243; NGO interaction with, in Udaipur district, Francisco, Hermina A., 333 133-34; pesticide incentives of, 244-46, 245 table; Freire, P., 132 research/dissemination limitations of, 98, 130, Fruits/vegetables: fungicide usage on, 257, 259, 259 131, 142, 350; research funding by, 76 table; insecticide usage on, 260 Grameen Bank, 63, 106, 115, 193 Fuelwood: agroforestry's production of, 214, 325-26; Grants: advantages of South African, 169; eligibility variable prices for, 331-32 requirements for, 163; incentives created by, 159, Fuente de Oro municipio (Meta), 172 n12 168, 172 n6, 173 n19; maximum size of, 100, 158, Fundaci6 Salvadorefia de Vivienda Minima (FUN- 172 n5; municipal program's reduction of, 164; for DASAL), 104, 106 nonland investment, 160, 168; restricted to land Fungicides: for horticulture, 257, 259, 259 table; purchases, 158-59, 172 n8 Japan's Asian share of, 255; reduced productivity Granulosis virus technology, 141 box benefits of, 264; reduced use of, 254; for rice pro- Grazing production systems: in arid rangelands, duction, 255, 257 table 284-85; with cropping potential, 211, 283-84; deforestation impact of, 287; feed conversion in, The Gambia, 196; WID Project in, 110, 112, 115 296; grazing fees for, 293; with mixed-farm poten- GATT. See General Agreement on Tariffs and Trade tial, 283-84, 297; positive carbon balance in, 291; Gbefi (Ghana), 192 regulatory control of, 294; resource-conserving Gender-and-development approach: limited use of, strategies for, 295, 297; in semiarid rangelands, 198-99; using mainstreamed project, 197-98; 285-86, 287; water infiltration effects of, 286 using separate women's project, 198; women-in- Greece, average farm size in, 172 n9 development approach versus, 102, 197, 353 Gremios (commodity associations), 32 Gender issues: class/caste variances and, 196; devel- Grepperud, Sverre, 33 nl, 60-61 opment policy's neglect of, 101-2, 106, 113, 186, Grootenhuis, J. G., 287 352; of extension systems, 187-88, 190; gender- Groundwater irrigation systems, 274, 278 Index 369 Group Farming for Rice Program (India), 44 Holden, Stein T., 1, 4, 5, 57, 59, 60, 61, 63, 64 Group Farming Initiative of the Kerala State (India), Home gardens (agroforestry system), 331, 331 table, 44 334, 335 table Grut, Michael, 333, 337 Honduras: agroforestry benefits in, 337-38; GTAP (Global Trade Analysis Project), 77-78 Cantarranas program in, 41, 42 table, 151; corn GTZ (Deutsche Gesellschaft fur technische yields in, 220., 221 table, 222; Guinope program in, Zusammenarbeit), 138, 151 41, 42 table, 151; land titling projects in, 224; peo- Guanacaste conservation area (Costa Rica), 178, 179, ple-centered (levelopment in, 150; reversed outmi- 181, 184 n5 gration in, 152; soil conservation costs-benefits in, Guatemala: agroforestry projects in, 334, 335, 338; 218 table; soil erosion rates in, 216 table corn yields in, 221 table; Defensores de la Horticulture. See Fruits/vegetables Naturaleza Program in, 152; female expenditures Host-plant resistance: for cereals, 1960s-90s, 260 in, 193; fuelwood prices in, 331-32; San Martin table; conventional breeding for, 261; essential to Jilotepeque program in, 41, 42 table, 151, 152; soil IMP, 260-61: with genetic engineering, 262-63; to conservation measures in, 218 table, 223; tree herbicides, 263; with molecular marker technolo- products market in, 333 gy, 262; and reduced pesticide use, 206-7, 254, Guinope program (Honduras), 41, 42 table, 151, 152 263-64, 355; with wide hybridization, 262 Gujarat (India), 42 Hot spot testing, 261 Hulme, D., 132 Habitat conservation: agricultural benefits from, 240; Humid rangelands: degradation effects on, 287; tech- defined, 205; in situ approach to, 235; of natural nologies for, 295 and modified habitats, 231 box, 232-33, 354 Haiti: corn/sorghum yields in, 220, 221 table, 222; ICRISAT village-level studies: on conservation people-centered development in, 150; ramp pay investment decisions, 61; on intertemporal mar- measure in, 222, 223, 223 table, 227; soil conser- kets, 54; on profit rate of small farmers, 57; on risk vation costs-benefits in, 218 table; soil erosion diffusion mechanisms, 55-56 rates in, 216 table IDD (Imperial Irrigation District, California), 280-81 Hampton, S., 59, 60 IGWDP. See Indo-German Watershed Programme Haryana (India), 275 IIMI. See International Irrigation Management Hazell, Peter, 1, 3, 7, 345 Institute Health effects, of pesticide misuse, 13, 206, 246, 254, Impact assessment, defined, 167 264, 355, 360 nl Imperial Irrigation District (IID, California), 280-81 Heath, John, 1, 3-4, 100, 351 Import restrictions, to limit deforestation, 306, 308 Hedges (conservation measure), 204, 222 table Heliothis armigera (cotton bollworm), 247 INBio (National Biodiversity Institute, Costa Rica), Heong, K. L., 259 178, 316 Herbicides: continued use of, 243, 254, 255-56, 264, Incentives. See Subsidies/incentives 356; genetic alternatives to, 264-65; Japan's Asian Income expenditures: gendered welfare effects of, share of, 255; management alternatives to, 264; 192-93, 193 table, 194-95 table, 199 n4, 352; tied market orientation, and demand for, 258 table, to income flows, 193, 196 259; for rice production, 257 table. See also Income growth: and demand for tropical products, Pesticides 306; from drainage improvements, 279; and popu- Hicks, Sir John, 9 lation/education effects, 72 Higgins, Paul A., 188 INCORA. See Instituto Nacional Colombiano de High-input agriculture, 36, 37; sustainable agricul- Reforma Agraria ture's transition from, 38, 44, 45. See also Inputs, INCRA (Brazilian land reform institute), 161 table, modem external 167-68, 172-73 nl8 Hindustan Paper Corporation (Kerala, India), 310 Incremental de velopment principle, 104 Hochman, Eithan, 273, 275 India: Asian share of pesticides in, 255, 256 table; Hoff, Karla, 52 common property issues in, 17, 286 box; conser- Hojancha (Costa Rica), 333, 338 vation investment studies in, 61; contract farming 370 Index effects in, 57; deregulated pesticides in, 247; fiscal 92; production relocation, and use of, 74-75; decentralization in, 179; Indo-German Watershed reduction of, in 1990s, 37, 44-45; taxation option Program in, 123-24; irrigation field agents in, for, 74. See also Fertilizers; Herbicides; 108-9; irrigation systems in, 273, 274; microwa- Insecticides; Pesticides; Resource-conserving tershed rehabilitation in, 134-35; Orissa Social technologies Forestry Project in, 106; risk aversion studies in, Insecticides: decreased use of, 243, 254; exportation 56; sustainability policy in, 44 of hazardous, 255; Japan's Asian share of, 255; Indica rice, 263 market orientation, and demand for, 257, 258 Indirect use values, defined, 10 table; reduced productivity benefits of, 264; for Indo-German Watershed Programme (IGWDP): rice crops, 255, 257 table, 263-64; on vegetables capacity-building phase of, 136-37, 138; external versus rice, 260 agency's role in, 123-24; full-implementation In situ conservation strategies, 205-6, 235, 238, 354 phase of, 138; funds/coverage of, 135; joint mnan- Instituto Nacional Colombiano de Reforma Agraria agement initiative of, 137; multi-level stakehold- (INCORA): establishment of, 157; financing ers of, 125, 138; planning approach of, 136; selec- arrangements under, 158-59, 165-66; interven- tion criteria of, 135-36 tionist budget of, 172 n7; land redistribution poli- Indonesia: colonization schemes in, 309; contract cies of, 27, 30-31, 157-58, 170-71; and Law 160, farming effects in, 57; decentralized conservation 31, 171 n2, 171-72 n4; municipal land reform ver- in, 178; facilitation model in, 252, 252 box; fiscal sus, 160, 161 table; productive project limitations decentralization in, 179; industrialization policies of, 163-64, 172 nnl5-16; selection of beneficia- in, 310; IPM application in, 250, 250 box, 267; ries by, 157-58, 162, 172 n13 people-centered development in, 150; pesticide Integrated Child Development Services Project subsidies removed in, 14, 75, 250 box, 267; rubber (India), 109 production efficiency in, 58; Rural Water Project Integrated pest management (IPM): collective action in, 110; soil loss in, 13; sustainability policy in, 44; approach of, 208, 266-67; consumer preference Uruguay Round and, 78 for, 246-47; country applications of, 249 table, Indus Basin Irrigation System (Pakistan), 278 250 box; economic incentives for, 208, 242-43, Industrial countries: extensionist/farmer ratio in, 188; 246, 248, 267; farmer training in, 207-8, 251-52, pesticide production/exportation by, 255; policy- 266, 356; with host-plant resistance, 206-7, supported degradation in, 43; rural-to-services 260-61, 355; implementation strategies for, 239, transition in, 75; sustainable projects in, 39 table; 249-51, 356; national policy framework for, 208, trade liberalization and, 71, 72 246, 249-50; natural control goal of, 207, 265, Industrial livestock production systems, 211, 284, 356; with participatory technology generation, 291, 294, 297 208,242, 243, 246, 248-49,251-52, 265-66,356; Industrial-world-to-the-rescue group, 4, 36 regulatory system of, 208, 246, 247, 248 table, Inkafha Freedom Party, 173 n22 251; to remediate biodiversity losses, 205, 233; Innovations: after program termination, 152; in slow adoption of, 265 Campesino program, 153; enabling policy envi- Intensive farming systems. See Agricultural intensifi- ronment for, 4, 346; by farmer extensionists, 99, cation 145; by Gujarat farmers, 42; by Honduran farmers, Interlinkage mechanisms: disadvantage of using, 66; 41-42; by Karjahi farmers, 111; by local water- efficiency/equity issues and, 56; to induce technol- shed communities, 97, 118, 122-23, 349; in ogy adoption, 43-44,51-52,63,64; market imper- l950s-60s, 92; by Rio Cabuyal farmers, 127-28; fections and, 63, 65-66, 346 sustainable productivity through, x, 1, 346. See Internal differentiation, in village economy models, also Experimentation by farmers; Resource-con- 52, 52 fig. serving technologies Internal rates of discount, experimental study of, 54 Inputs, modem external: extension approach for, 108, International Center for Research in Agroforestry, 326 187-88, 190; gender dimension of, 186, 188-90, International Code of Conduct on the Distribution and 189 table, 199 nl; irrigated agriculture's misuse of, Use of Pesticides (FAO 1990), 249 3, 9, 11, 93, 346; modernist development on, 36, International Crop Research Institute for the Semi- 40; natural controls versus, 37, 207; in 1950s-60s, Arid Tropics (ICRISAT). See ICRISAT village- Index 371 level studies nl; property rights over, 17, 119, 137; urban/envi- International Finance Corporation, 233 ronmental diversion of, 271, 280. See also Water International Food Policy Research Institute, 18-19 pricing; Water trading/markets International Institute for Environment and ISNAR (International Service for National Development, 38 Agricultural Research), 131 International Irrigation Management Institute (IIMI), Israel, water pricing schemes in, 276, 277 208, 209, 271, 273, 277, 280, 356, 357 Italy, average farm size in, 172 n9 International Organization of Pesticide Resistance IUCN (International Union for the Conservation of Management (IOPRM), 247 Nature), 320 n2 International Rice Research Institute (IRRI), 11, 250 box, 261 Japan, pesticides in, 255, 256 table International Service for National Agricultural Japonica rice, 2153 Research (ISNAR), 131 Jodha, N. S., 286 International Soil References and Information Centre, Jones, Christine, 199 nl 215 Jones, W. I., 274 International Tropical Timber Organization, 314 International Union for the Conservation of Nature Kaimowitz, David, 2, 212 (IUCN), 320 n2 Kalfayan, J., 57 Intertemporal markets: failed policy intervention in, Kamala Irrigation Project (Nepal), 107 62-63; high covariance and, 54; ICRISAT Kapok International (Ohio), 233, 234 box research studies of, 54; imperfection/absence of, Kassas, M., 284 53-54; inherited endowments element of, 54. See Kayan-Mentarang project (East Kalimantan), 178 also Credit markets; Crop insurance markets Kazakhistan, 279 IOPRM (International Organization of Pesticide Kenya: alley cropping in, 40; compliance with Resistance Management), 247 enforcement in, 180-81; decentralized conserva- IPM. See Integrated pest management tion in, 177 table, 177-78; female expenditures in, Iremonger, S., 320 n2 193; gender issues in, 187, 189 table, 190; land IRRI. See International Rice Research Institute titling program in, 62; local organizational capaci- Irrigated agriculture: basinwide approach to, 209, ty in, 121, .122; people-centered development in, 280-81, 357; economic problems in, 208-9, 150; returns to conservation in, 224 271-72, 357; and environmental effects, 2, 11, Kerr, J., 61 208-9, 271, 278, 279; future output demands on, Kerzhenski reserve (Russia), 178 14, 208, 271, 356-57; income distribution issues KfW (Kreditanstalt fur Wiederaufbau), 138 for, 275-76; integrated drainage improvements in, Khandker, Shahidur R., 193 278-79, 280; intensification effects in, 11; in King Mahendra Trust for Nature Conservation Israel, 277; local innovation's importance to, 111; (Nepal), 178 maintenance/operation of, 273, 274; overuse of Knops, J., 234 modern inputs in, 3, 9, 11, 93, 346; pricing issues Korea: Asian share of pesticides in, 255, 256 table; for, 276-77; and property rights issue, 17, 119, field agents' role in, 108-9; gender/efficiency 137; returns to research in, 130; subsidization of, studies in, 189 table 209, 275, 360 nl; successful projects of, 39 table, Kostomuksha (Russia), 180 279-80; water allocation techniques of, 272-73; Kumar, H., 261 World Bank studies on, 274, 279-80. See also Kwazulu Natal (South Africa), 173 n22 Irrigation water; Water pricing; Water trading/ markets Labor market (Colombia): agriculture's share of, 25, Irrigation water: allocated by irrigation projects, 25 fig.. See also Rural labor market (Colombia) 272-73; allocated by water trading/markets, 209, Lacki, Polan, 146 274-75, 357; basinwide management of, 209, Lake Xolotlan (Nicaragua), 226 280-81, 357; Israeli allocation of, 277; livestock Landcare (Australia), 124 grazing effects on, 209-10; losses/salinity of, Land Claims Court and Commission (South Africa), 273-74; overuse/misuse of, 2, 12, 14, 17, 309, 360 169 372 Index Land invasion, in Colombia, 27, 32 Livestock Development Project (Mauritania), 105 Land reform: biased against the poor, 26, 27-28, 33 Livestock feed: concentrates as, 290, 290 fig.; and n3, 157, 159, 172 n8; flawed approaches to, 28; feed-food controversy, 36, 289 box; land used to INCORA's policies of, 27, 30-31, 157-58, produce, 209, 357; subsidies for, and range degra- 163-64, 172 nl4; Law 160 on, 31-32, 33 n6, 159, dation, 286; sugarcane-based, 295 box; technolo- 172 nnlO-ll; market-assisted model of, 100, 156, gies for conversion of, 210, 295-96 157; potential equity/efficiency benefits of, 30, Livestock production policies: collective purpose of, 100, 156, 162, 351. See also Negotiated land 292, 357-58; of decentralized decisionmaking, reform 292-93; of full-cost pricing, 293; regulatory Land taxes, to discourage deforestation, 313, 313 instruments of, 294; technology component of, table 210, 294-98, 358; of training/extension, 294 Land tenure: as agroforestry adoption issue, 335-36, Livestock production systems: with area-wide crop 340; Colombian changes in, 27, 27 table, 157; integration, 210-11, 297, 358; development path- communal system of, and women, 191; as conser- ways of, 298, 298 fig.; grazing type of, 211, vation investment issue, 219, 224-25, 314, 317 283-84; industrial type of, 211, 284; institutional table; decoupled from deforestation, 312, 313 regulation of, 292-93; mixed-farming type of, 211, table; evolved effectiveness of, 17, 94; interlink- 284; resource-conserving strategies for, 297-98; age and, 56; Law 160 on, 32; South African reform zonation of, 294. See also Grazing production sys- of, 169; sustainable productivity's ties to, 6-7, tems; Industrial livestock production systems; 190, 293, 351 Mixed-farming systems Land titles: agricultural-use precondition for, 287; Livestock sector: cereal consumption by, 36, 283, decoupled from deforestation, 312, 313 table; 290, 357; crop integration with, 287-90, 295; failed institutionalization of, 62; as gender issue, deforestation impact of, 211, 287, 358; drug traffic 187, 191; versus security of tenure, 6-7, 89-90, and, 29; favored by Colombian policy, 25, 26, 100, 224-25 157; feed-food controversy of, 36, 289 box; as Laos, 40-41 flexible asset, 6, 88-89; in situ conservation Large farmers: Colombian policy favoring, 25-26, 30, approach to, 235; land share for grazing of, 209, 33 n3, 157; commodity associations of, 32; non- 283, 284, 357; monogastric species shift in, 284; farming benefits to, 28; small-farmer yields ver- processing-wastes of, 290-91; production systems sus, 27-28; structural adjustment effects on, 157, of, 283-84; research/development model for, 240; 158; subsidized products of, 305-6 resource-conserving technologies for, 210, La Selva reserve (Costa Rica), 232 box 294-98, 358; in Sahelian countries, 285 box; soil Latin America: deforestation agents/causes in, 211, degradation from, 284-86; trade liberalization 303, 358; land titling reforn in, 312; marketing and, 75; unproductive animals of, 293 box; water biodiversity of, 233-34, 234 box; 1980-90 defor- infiltration effects from, 209-10, 286; wildlife's estation in, 12, 302; people-centered development complementarity with, 287. See also Livestock in, 154; resource-conserving technologies in, 37 feed; Livestock production policies; Livestock Law la of 1968 (Colombia), 27 production systems; Livestock waste Law 30 of 1988 (Colombia), 27 Livestock waste: efficient recycling of, 296; gaseous Law 101 of 1993 (Colombia), 28, 33 n3 emissions of, 289 box, 291-92; global amount of, Law 160 of 1994 (Colombia), 31-32, 33n6, 159, 172 209, 283, 357; to improve soil fertility, 288; reduc- nn1O-11 ing emissions from, 210, 294, 296-97; reducing Law 200 of 1936 (Colombia), 27 nutrient losses from, 294-95 Lazovsky (Russia), 180 Local community/organizations: agencies responsive Leakey, Richard, 175 to, 96, 109-11, 110 table, 348; assessing Leonard, H. Jeffrey, 215 demands/needs of, 113; assessing investment in, Leucaena leucocephala, 334 115; balanced by nonlocal perspective, 101, Ley de Aparcerfa of 1975 (Colombia), 27 175-76, 352; collective action by, 17-18, 120, Lfnea Bi6sfera program (Mexico), 152 122; conservation benefits for, 104, 183-84; con- Literacy rates, 149, 186. See also Education servation decentralized to, 178, 184 n2, 292-93; Litsinger, J. A., 263 enforcement role of, 180-81; entry/allocational Index 373 rules of, 105-6; extension versus empowerment pollination in, 234 box; pesticide usage in, 13, 255 approach to, 96, 107, 108; fiscal authority/ Mali, 13, 108, 192, 285 accountability of, 105, 182; institutional design Mali Natural Resources Management Project, 108 principles for, 18; multipurpose tree-planting role Mali Pastoral Associations Project, 104 of, 214, 326, 340 nl; negotiated land reform role Malthusian effects of population growth, 5, 35-36, of, 159-60; nontimber revenues for, 315-16; 51, 57 power structures within, 97, 103, 105, 106, Management of Natural Resources Project (MARE- 124-25, 348, 350; resource degradation by, NA, Dominican Republic), 226 175-76, 180, 182; self-selection process of, Manasan, R. G., 320 n7 113-14; social dimension of, 105, 111, 113, 121; Manure. See Livestock waste technology's fit with, 95, 106-7. See also Maps, contour /ersus landholding, 136 Community-driven development projects; Local MARENA (Management of Natural Resources organizations in watershed management; People- Project, Dorninican Republic), 226 centered agricultural development Market access: Boserup effects of, 4, 22-23, 58; and Local elites, 4, 105, 106, 347, 348; as watershed man- deforestation, 58; landownership tied to, 100, 156, agement participants, 97, 124-25 171, 173 n24 Local organizations in watershed management: Market failures, versus market imperfections, 50. See advantages of, 118, 349; external agencies' inter- also Market imperfections face with, 123-24, 126-27, 349; Landcare experi- Market imperfections: deterninants of, 52-53; failed ence of, 124; of Rio Cabuyal pilot catchment, policy decisions on, 61-63; farm household model 125-27; roles/functions within, 119, 120-21; with, 58-59; and farm size/productivity relation- social capacity goals of, 121, 239; storage capaci- ship, 57; impact of, on small farmers, 4-5, 50, 65; ty's importance to, 120; technical innovation suc- interlinkage response to, 51-52, 63, 65-66, 346; in cess of, 97, 118, 122-23; types of, 120 low population density areas, 51; market failures Loevinsohn, M., 267 versus, 50; and risk diffusion mechanisms, 55-56; Logging: export taxes/bans on, 307, 320 n6; and for- soil degradation models with, 61; and transmission est conversion, 213, 307, 318-19, 320 n5; promot- of poverty, 54-55; types of, 5, 50. See also ing sustainable practices of, 314-135; reducing Interlinkage mechanisms profitability of, 309-10 Marshall, Alfred, 56 L6pez. Gabino, 41-42 Masai Mara reserve (Kenya), 179 Lovejoy, Annie, 175 Masoka area (Zimbabwe), 180 Low-input agriculture, and sustainability, 36, 37 Matruh Natural Resource Management Project, 113 Lutz, Ernst, 1, 2, 3, 101, 175, 179, 203-4, 345, 354 Mauritania Livestock Development Project, 105 Meat production, species shift in, 284 McConnel, K. E., 60 Mekonnen, Hailu, 188 Machakos district (Kenya), 4, 23, 286 box Mennonite Central Committee, 151 McIntire, John, 52, 224 Merck Pharmaceuticals, 316 McNamara, Robert, ix Methane gas emissions, 202, 289 box, 291, 296-97 Magadanski (Russia), 180 Methodological individualism, of contemporary eco- Maharashtra watershed program. See Indo-German nomics, 85 Watershed Programme Mexican National Water Authority, 111 Mahmud, Simeen, 192 Mexico: cultivation/cropping practices in, 224; FON- Maissade (Haiti), 121, 222, 223, 223 table, 224 AES outreach programs in, 107; gender/class Mai Vo, 259 issues in, 196; godfather-supported policy in, 112; Maize: in farm household models, 59, 60; genetic irrigation project success in, 280; National Water engineering of, 263; insect pests of, 261; 1990s Authority in, 111; people-centered development pesticide use on, 255; pest-resistance of, 254, 260 in, 150, 152; PIDER program in, 109 table. 261; profitable yields of, 43, 152, 154; as Microwatershed rehabilitation: biophysical successes share of feed grain, 290 of, 134-3:5; capacity-building phase of, 136-37; Malayan peninsula, 232 box expanded structure/pathways for, 137-39; joint Malaysia: industrialization policies in, 310; palm tree action in, 142-43; Maharastra case study of, 374 Index 98-99, 134-37; screening/funding mechanism for, (Nepal), 178 137; selection criteria for, 135, 136; structured Native Lands Act of 1912 (South Africa), 168 approach to, 135, 143, 143 nl; technical expertise Natural control pest management: as goal of IPM, following, 137 265; versus pesticides, 37, 207, 264, 355-56; prof- Middle East, 286, 295 itability of, in Philippines, 264 Migration: rural-rural, 29; rural-urban, 24-25, 30, 152 Natural habitats: as global share of land, 233; in situ Mihm, J. A., 261 conservation of, 235; protection of, as insufficient, Ministry of Agriculture (Colombia), 32 231 box, 238-39 Ministry of Agriculture (India), 138 Natural resource degradation: as contextual, 6, 89; Ministry of Agriculture (Kenya), 121 costs-benefits analysis of, 10-11; environmental Ministry of Economic Cooperation (Germany), 138 effects of, 11-13; flawed causal models of, 88; Ministry of Environmental Protection (Russia), 178 high discount rates and, 58; as institutional prob- Ministry of Finance (India), 138 lem, 6, 87-88; by local community, 175-76, 180, Missing markets: farm household model with, 58-59; 182; output price distortions and, 14-15, 74, 75; soil conservation models with, 61 policy support of, 1, 3, 4, 14, 43, 346; population Mixed-farming systems: animal draft component of, growth and, 5, 51, 57; poverty-reducing solution 288; grazing system's transition to, 283-84; inten- to, 14, 16, 63, 93, 292; production relocation and, sification/growth in, 211, 284, 287, 297; involu- 75; technologies' role in, 18; urban migration's tion trend in, 288-99, 297; nutrient balance issue failed solution to, 24-25, 30; Uruguay Round and, of, 288; regulatory control of, 294; resource-con- 78. See also Deforestation; Environmental effects; serving strategies for, 297; specialization trend in, Soil degradation 288 Natural resource management: through collective Molecular marker technology, 262 action, 17-18, 192; inadequate research on, 18-19; Montelibano municipio (C6rdoba), 172 n12 information requirements for, 15-16, 94; learning Monteverde Conservation League (Costa Rica), 338 process approach of, 115; local organizations' role Moock, Peter, 188 in, 118, 120; local ownership essential to, 7, 83, Morocco, 59 90, 148, 348; market imperfections' impact on, Mortimore, Michael, 23, 286 4-5, 50, 65; nonprice intervention factor of, 15; Multiagency research approach. See Participatory prices/subsidies factor of, 14-15; property rela- technology generation tions central to, 6-7, 16-17, 93-94; Rio Cabuyal Munasinghe, Mohan, 307 case study of, 125-28; using weak versus strong Municipal land reform program: eligibility for grants sustainability, 10; women's role in, 190-92. See under, 163; farm size requirement of, 164; financ- also Agroforestry; Biodiversity conservation; ing arrangements of, 166; INCORA's practices Integrated pest management; Soil conservation; versus, 160, 161 table; information system of, Tropical forest conservation policies; Watershed 164-65; land selection system of, 163; mecha- management nisms to implement, 161 table; monitoring/evalua- Nature reserves, 179, 180 tion of, 166 table, 166-67; productive projects of, Negotiated land reform: in Brazil versus Colombia, 164, 165 table; selection of beneficiaries under, 168; in Brazil versus South Africa, 167, 351-52; 162-63, 168. See also Productive projects conceptual versus functional structure of, 159-60; Myanmar, 256, 256 table decentralized approach to, 159-60, 170-71; defined, 100, 156; farm-size requirement of, 159, NABARD. See National Bank for Agricultural and 160, 172 n9; financing arrangements of, 158-59, Rural Development 160, 165-66, 171 n3; INCORA's role in, 158, 160, Nakajima, C., 58 171 n2, 171-72 n4; INCRA's mechanisms of, 161 Narayan, Deepa, 1 table, 167-68, 172-73 n18; land selection system National Bank for Agricultural and Rural of, 163, 170; local participation required for, Development (NABARD), 136, 137, 138 159-60; maximum grant size in, 158, 172 n5; National Biodiversity Institute (INBio, Costa Rica), mechanisms to implement, 161 table; private sec- 178, 316 tor's role in, 160, 171; productive projects of, 160, National Parks and Wildlife Conservation Act of 1973 164, 165 table, 170; reasons for shift to, 157-58; Index 375 rental prohibition of, 159, 172 nIO; revised goals 226 of, 160, 172 nll; selection of beneficiaries under, Ogoniland (Nigeria), 178 162-63, 168, 172 nI3, 351; in South Africa, Okomu reserve (Nigeria), 179 168-70, 173 nn2O-23; three principles of, 100, Open-access property: and resource degradation, 6, 158, 351. See also Municipal land reform pro- 84 box; rights to, 84 box gram; Productive projects Option values, clefined, 10 Nepal, 105, 107; decentralized conservation in, 177 Orangi (Pakistan), 105 table, 177-78; fiscal decentralization in, 179; god- Organisation for Economic Co-operation and father-supported policy in, 112; Karjahi irrigation Development (OECD), 40, 252 nl system in, 111 Orissa Social Forestry Project (India), 106 Netherlands: average farm size in, 172 n9; national Ostrom, Elinor, 18, 111 IPM policy of, 250 Output price distortions: and pesticide use, 74, 75, 80; Net present value (NPV) criteria, 219 and resource degradation, 14-15 New modernists, 4, 36, 40 Overseas Development Administration (now British NGOs. See Nongovernmental organizations Department for International Development), 40, NIA program, 109, 111 139 Nicaragua: agroforestry benefits in, 338; Campesino a Ownership. See Property regimes Campesino Program in, 151, 153, 154; fuelwood OXFAM, 151 prices in, 331-32; soil conservation costs-benefits in, 218 table; soil erosion rates in, 216 table; sub- Pacayas (Honduras), 41 sidization schemes in, 226; value of windbreaks in, Pagiola, Stefano, 60, 260 332 Pakistan, 106, f07; Balochistan Primary Education Nigeria, gender issues in, 109, 187, 189 table, 190 Project in, 105, 109, 111-12; drainage manage- Nilaparvata lugens (brown planthopper), 250 box, ment in, 278-79; salinity/waterlogging study in, 259 273-74; water scarcity problems in, 12; women Nitrogen Fixing Tree Association, 326 extension workers in, 109 Nitrous oxide emissions, of manure, 292 Palis, F. G., 25:5 Nizhny Novgorod (Russia), 178 Palm pollination, manual versus natural, 234 box Nongovernmental organizations (NGOs): agro- Panama: agroforestry management in, 334; fuelwood forestry support from, 323, 340; decentralization prices in, 331-; soil degradation in, 216 table, 221 role of, 178-79; group-based approach of, 142; Pareto-relevant externality, interlinkage mechanisms IGWDP and, 136-37, 138; nontimber forest prod- and, 63 ucts agenda of, 315; participatory approach of, 16, Parks and Wild Life Act of 1975 (Zimbabwe), 178 98, 132; post-rehabilitation role of, 137; soil con- Participation: case studies incorporating, 98-99; servation role of, 226, 354; strengths and weak- decentralization's role in, 101, 352; empowering nesses of, 132-33; task sequencing by, 104; in versus functional use of, 98, 132; principles of, in Udaipur district, 133-34 community-based development, 96; and unre- Nonseparability of production/consumption, 4-5, 51, solved gender asymmetries, 197, 353; used in 58 technology generation, 97-98. See also Nontimber forest products, marketing of, 315-16, 317 Community-driven development; People-centered table agricultural development North African livestock feed subsidies, 286 Participatory technology generation: among low- Nueva Ecija (Philippines), 267 income farimers, 97-98; in Bolivia, 139-41; case Nurseries, community versus centralized, 337 studies of, 98-99; empowering versus functional approach to, 98, 132; field-based proliferation of, OECD (Organisation for Economic Co-operation and 133; four types of scaling up, 132; indigenous Development), 40, 252 nl knowledge through, 238, 240; IPM's reliance on, Office du Niger (Sub-Saharan Africa), 280 242, 243, 251-52, 265-66, 356; ISNAR studies Off-site impacts: addressed by new research, 233; of on, 131; in Maharashtra, 134-37; multiagency natural resource degradation, 13, 97, 122, 203, approach to, 98, 131, 350; with RELU linkages, 217, 225, 227 nl; subsidization justified by, 225, 139-41, 141 box, 143; in resource-poor areas, 376 Index 99-100, 145; structured context of, 134-37, 143, misuse of, 14, 206, 245 table, 245-46, 360 nl; use 143 nl; two on-farm types of, 239-40; in Udaipur of, by stage of agriculture, 259 table. See also district, 133-34; unstructured context of, 133-34; Herbicides; Insecticides; Integrated pest manage- in watershed management, 96-97; wide-scale ment; Pest management implementation of, 137-39, 142 Pest management: action/economic thresholds of, Pastoralism. See Livestock sector 244; consumer influence on, 246-47, 260; costs- PATECORE Program (Burkina Faso), 146, 153 benefits criteria of, 243; in developing countries, Pathak, M. D., 263 244-46, 245 table; justified government interven- Patzite region (Guatemala), 222, 223, 224, 225 tion into, 246; monoculture approach to, 244, 259; PCAD. See People-centered agricultural development national/regional base for, 243; synchronization Peace Corps, 151 strategy of, 266-67; traditional approach to, 242, Pender, John L., 57, 61 244; training's impact on, 266; using natural con- Pennsylvania, 43 trol versus pesticides, 37, 207, 264, 355-56. See People-centered agricultural development (PCAD): also Integrated pest management empowerment through, 150, 150 fig., 154; large- Pest-resistance. See Host-plant resistance scale application of, 154; as learning-by-doing Pests: corn insect, 264; maize insect, 261; rice insect, process, 148, 152; lower costs of, 151; obstacles 261; wheat insect, 261 to, 152-53; origins of, 146-47; productivity gains Philippines: Asian share of pesticides in, 256 table; with, 145, 151-52; program modification of, decentralized conservation in, 177 table, 177-78; 153-54; rural characteristics tied to, 147 fig.; sus- employment incentive in, 320 n7; fiscal decentral- tainability factors of, 99-100, 148-50, 149 fig., ization in, 105, 179; NIA program in, 109; people- 350-51; synergistic principles of, 99, 145-46, 146 centered development in, 150; pest management fig., 350; uneven success of, 99, 152, 350. See also in, 264; reduced crop losses in, 263; synchronous Farmer-to-farmer extension rice cultivation in, 267 Percy Amendment to U.S. Foreign Assistance Act PIDER (Programa Integral para el Desarrollo Rural, (1973), 196 Mexico), 109 Perennial with trees (agroforestry system), 331, 331 Pilot manicipios (Colombia): INCORA's practices table, 332, 335 table versus, 160, 161 table; names/locations of, 172 Peru, 150, 276 n12. See also Municipal land reform program Pesticide industry: governmental protection of, 246; Pimentel, D., 234 box growth rate of, 255; increased herbicide use in, Pingali, Prabhu L., 2, 255, 256, 259, 263-64, 267 243, 254, 356; profit motive of, 243; regulation of, Pithuwas Irrigation Project (Nepal), 107 247, 248 table Pitt, Mark M., 193 Pesticides: agricultural PSE and, 74 fig.; for Asian Policing. See Enforcement of conservation rice crops, 255-56, 256 table, 257 table; for cereal Polluter-pays principle, 43, 247-48; violation of, 316 crops, 254, 255, 260; disincentives for use of, 43, Poole, Joyce, 175 247-48, 264, 267; environmental effects of, 2, Poor, the: FONAES program for, 107; global statistics 246, 254, 355, 360 nl; exportation of, 255; global on, ix, 35; land reform biased against, 26, 27-28, market for, 252 nl, 255; health effects of, 13, 206, 33 n3, 157, 159, 172 n8; neglected by community- 246, 254, 264, 355, 360 nl; Indonesia's reduced based programs, 106, 113; projects targeted to, use of, 44; intensive agriculture's use of, 11, 206, 115; structuring subsidies for, 114. See also 230-31, 244, 254, 256-57, 355; international stan- Poverty; Resource-poor farmers; Small farmers; dards for, 208, 249; IPM's regulation/registration Village economies of, 247, 248 table; lethal to pest predators, 244, Population growth: Boserup effects of, 4, 5, 22-23, 259; market orientation, and demand for, 256-57, 346; and collective action, 18; environmental pes- 258 table, 259; motivations for use of, 243, 259, simists on, 35-36; income growth and, 72; 260, 355; natural control versus, 37, 207, 264, Malthusian scenario of, 5, 51, 57; productivity 355-56; new modernists on, 36; output price dis- demands of, 9, 271, 306 tortions and, 74, 75, 80; production relocation and, Posner, J. L., 33 n2 74-75, 76; reduced productivity benefits of, 254, Potato yields, 220-21 264; social versus private costs of, 246; subsidized Poultry manure, 296 Index 377 Poverty: in Colombia, 24; cost-sharing schemes to 191; legitimation of, 6, 85-86, 293; pastoralist ver- address, 64; food-for-work policy on, 63, 64-65; sus sedentary, 88-89; preallocative/allocative global statistics on, ix, 35; and high discount rates, functions of, 85; and reduced externalities, 15; 57-58; resource degradation's ties to, 14, 16, 63, rights to, 84 box; with single versus multiple out- 93, 292; and risk diffusion insufficiency, 56; in puts, 191-92; specified prior to projects, 6, 87, South Asia, 130: in Sub-Saharan Africa, 130; and 347; types of 84 box. See also Common property; sustainability, in farm household models, 60; Land tenure; Land title; Private property; Usufruct transmitted via imperfect markets, 54-55; urban rights migration's failed solution to, 24-25, 30. See also Protein, animal xersus vegetable, 289 box Poor, the; Resource-poor farmers; Small farmers; Public sector reszarch: predicted funding of, 76; pro- Village economies posals for improvements in, 131; rates of return Pretty, Jules N., 1, 4 on, 18-19, 130; research/dissemination limitations Price controls, to limnit deforestation, 305-6, 308 of, 98, 130, 131, 142, 350; on soil conservation, table, 320 n4 226-27 Price supports (Uruguay Round): as "blue-box" Puerto Wilches municipio (Santander), 172 n12 instruments, 76; Cairns Group's action on, 80; Punjab (India), 275 environmental subsidies in lieu of, 76. See also Subsidies/incentives Quarantine import restrictions, 76-77, 81 Pri-Oksky Terrasny (Russia), 180 Quisumbing, Agnes, 2, 101, 352 Private property, 85; failed institutionalization of, 62; rights to, 84 box; and watershed's transboundary Rainfed agriculture areas: fertilizer usage in, 12; rain- effects, 119 fall variations in, 277-78; research needs of, 14, Private sector: biodiversity marketing role of, 233-35, 98, 130; resource degradation in, 3, 11-12, 16, 234 box, 238; negotiated land reform role of, 160, 346; sustainable agriculture's transition from, 38; 171; research by, in rainfed areas, 130 sustainable projects in, 39 table; technological PROCERA (Brazilian program), 168 fixes for, 278 Producer subsidy equivalent (PSE), and use of pollu- Raintree, John B., 333 tive inputs, 74 fig. Ramp pay (soil conservation measure), 222, 223, 223 Production relocation of agriculture: and environmen- table, 224, 227 tal concerns, 72-73; and labor/capital effects, Rayner, Anthony J., 80 75-76; and land degradation, 75; and livestock Reardon, Thomas, 227 sector effects, 75; and pollutive inputs, 74-75; RELU. See Research-Extension Liaison Unit research's amplification of, 79; trade reform caus- Repetto, Roberc, 13, 275 ing, 73 figs., 73-74 Research. See Agricultural research Productive projects: economic indicators of, 164, 165 Research-Extension Liaison Unit (RELU), 99; adap- table; employment goal of, 164, 172 n16; tive research by, 139-40; factors for success of, INCORA's approach to, 163-64, 172 nnlS-16; as 142; linkage mechanisms of, 140-41; origins of, key to negotiated reform, 170; rnunicipio's 139; participation scaled up through, 141 box, 143; approach to, 164; revised goal of, 160, 172 nll staff of, 140 Programa Integral para el Desarrollo Rural (PIDER, Resource-conserving technologies: in Central Mexico), 109 America, 40-41; education, and use of, 149, 188, Property regimes: with ambiguous distributional con- 190; extemal agencies' input on, 123-24; farmers' sequences, 6, 86, 87, 347; central to sustainable adaptation of, 40-41, 42 table, 127-28, 145, 147, development, 7, 83, 90, 148, 348; as conservation 148-49; gender dimension of, 186, 188-90, 189 investment issue, 6-7, 16-17, 89-90, 93-94, 219, table, 199 nl; in Gujarat, 42; high discount rates 224-25; cultural/social dimension of, 6, 84-85; and, 58; interlinkage to induce adoption of, 43-44, deforestation to establish, 312; discrirninatory reli- 51-52, 63; for livestock sector, 210, 294-98, 358; gious laws on, 187, 199 n2; failed policies on, 62; local capacity issue of, 95, 106-7, 122-23, 149, gender inequalities in, 186, 187, 190-92, 352; 348; in people-centered development, 145, 146, income growth and, 72; Law 160 on, 31-32, 33 n6, 153, 154; rates of return on, 80; in Rio Cabuyal, 159, 172 nnlO-1l; legal/customary divergence in, 127-28; sustainability and, 37, 38, 99-100, 152, 378 Index 154, 350-51; in Thailand, 42-43; trade policy's Saito, Katrine, 188 expansion of, 72, 79; yields, and use of, 61, 64 Salamanca National Park (Colombia), 180 Resource-poor farmers: government extension's Salinization: in Aral Sea region, 279; conveyance neglect of, 99; people-centered development for, losses and, 273-74; induced by subsidies, 12; in 99-100, 145, 146-48; productivity growth for, 16; intensive farming systems, 11; principal causes of, risky investment in, 16 278 Restitution policy, of South African land reform, 169 Samarskaya Luka (Russia), 180 Rice: genetic engineering of, 262-63; insect pests of, Sandmo, A., 59, 60 261; integrated pest management of, 250 box, 264; Sanitary and phytosanitary (SPS) measures (Uruguay intensification effects on, 11; molecular marker Round package), 77, 81 selection on, 262; pesticide use on, 255-56, 256 San Martin Jilotepeque Integrated Development table, 257 table, 260, pest-resistance of, 260 table, (Guatemala), 41, 42 table, 151, 152 261, 263-64; sustainable policy for, 44; synchro- Santa Catarina (Brazil), 122, 154 nous cultivation of, 266-67 Santa Cruz Department (Bolivia), 139 Rio Cabuyal catchment, 125-28 Santa Lucia (Nicaragua), 222 Rio Cauca, 125 Santa Lucia Milpas Altas watershed (Guatemala), 225 Rio Ovejas, 125 Sayre, K. D., 264 Risk diffusion mechanisms: heterogeneity of produc- Schillhorn van Veen, Tjaart, 2 tion as, 59; insufficiency of, in imperfect markets, Schneider, Robert, 227 n4 55-56, 59; surplus production as, 60 Schultz, T. W., 51, 56 Rivera municipio (Huila), 172 nl2 Selassie, Haile, 24 Road construction: and deforestation, 213, 310, 318; Self-Employed Women's Association (India), 115 reduced subsidies for, 311, 319 Semiarid rangelands: intensification effects in, 286 Rodale International, 151 box; land degradation in, 285-86; technologies Rola, A. C., 256, 259, 263, 264 for, 295 Rondonia colonization scheme (Brazil), 309 SENACSA (Costa Rican soil conservation service), Rosenzweig, M. R., 52, 55 226 Royal Chitwan National Park (Nepal), 179 Sensibility sustainability, and capital transformations, Rozanov, B., 284 10 Ruf, F., 320 n4 Serageldin, Ismail, 10 Ruminants, 284, 289 box, 296 Sere, Carlos, 283 Rural Development: From Vision to Action. A Sector Shanmugaratnam, N., 63 Strategy, ix Shapiro, M., 255 Rural development projects: ambiguity effects on, 86, Sharecropping: Colombian reductions in, 27, 27 table; 87, 347; culturallsocial dimension of, 84-85, 89, Law 160 on, 32 105, 111, 113; gender issues of, 101-2, 106, 113, Sharma, Uday, 175 186, 196-97, 352; property relations central to, 6, Shiferaw, B., 57, 61, 64 83, 86, 87, 347; three entities/agendas of, 86-87; Shuttle breeding, for host-plant resistance, 261 as transfer payment, 176. See also Agricultural Sikhote Alinski (Russia), 180 development Silage use, in Santa Cruz, 141 box Rural elites. See Local elites Silent Spring (Carson), 246 Rural labor market (Colombia): declining farm sector SINAC (Sistema Nacional de Areas de Conservaci6n, in, 30; INCORA policy's impact on, 162, 172 n13; Costa Rica), 178, 184 n3 low growth rate in, 157; male-female differential Sindh Province (Pakistan), 278 in, 29; 1993 earning sources of, 30 Singh, Inderjit, 217 Rural Water Project (Indonesia), 110 Singh, Shekhar, 175 Russia, 178, 180 Sistema Nacional de Areas de Conservaci6n (SINAC, Rwanda, 40, 288 Costa Rica), 178, 184 n3 Siswoputranto, P. S., 320 n4 Sadoulet, E., 58-59 Small farmers: discount rates of, 57-58; farm size Sahelian countries, livestock sector in, 285 box, 287 requirement for, 159, 164, 172 n9; food-for-work Index 379 programs for, 63, 64-65; incomes per capita of, table. See also Soil erosion 57-58; inefficiencies of, 51, 56, 63, 346; interlink- Soil erosion: average rates of, in Central America, 216 age policies for, 63-64, 65; labor-rich versus table; on-site/off-site costs of, 13; productivity labor-poor, 58; land reform biased against, 26, effects of, 204, 215-16; unsystematic measure- 27-28, 33 n3, 157, 159, 172 n8; market imperfec- ment of, 215, 225; Uruguay Round and, 78 tion effects on, 4-5, 50, 65; production/consump- Soil moisture, 204, 222 tion nonseparability of, 4-5, 51, 58; productivity Solidarity Fund for Social Enterprises (FONAES), gains of, in Colombia, 27-28, 30, 33 n4, 171 nl; 107 profitability studies of, 57, 326, 330; risk diffusion Sorghum yields, 220, 221 table, 222 mechanisms of, 55-56, 59-60; and segmentation South African land reform: beneficiaries of, 169, 170, phenomenon, 28, 33 n5. See also Family-farm 173 n23; "black spot removal" policies of, 168-69; units; Farm household models; Resource-poor Brazilian versus, 167, 351-52; decentralization farmers; Village economies steps in, 171; mechanisms to implement, 161 Smapath, Rajan, 276 table; negotiated form of, 100-101; policy compo- Soil Changes under Agroforestry (model), 220 nents of, 169, 173 nn2O-21; shortcomings of, 170, Soil conservation: Central American programs of, 41, 173 n22 42 table; decentralized approach to, 205, 226-27; South Asia: agroforestry in, 331; deforestation rever- farm-level perspective on, 204, 217, 219; govern- sals in, 317; livestock nutrient supplementation in, mental/NGO roles in, 204-5, 354; government 289 box; microwatershed rehabilitation in, research on, 226-27; locally-induced adoption of, 134-35; rainfall variations in, 277-78; rural pover- 122; manure's contribution to, 288; with multipur- ty levels in., 130; warabandi rotation system in, pose tree plantings, 326; nonseparability models 274, 275 of, 60-61; off-farm investments versus, 227 n4; Southeast Asia., 146-47; animal draft in, 288; defor- off-site effects of, 225, 226; physical structures estation in, 312; employment incentive in, 320 n7; used in, 204, 222, 223, 224; productivity effects of, farmer-to-farmer extension in, 146; rainfall varia- 204, 222-23; subsidization of, 204, 225-26, 354; tions in, 277-78 techniques of, 203, 216-17, 353; unprofitability Southgate, Douglas, 227 n4 of, in Patzite, 224. See also Soil conservation South India, 273 costs-benefits analysis South Queensland (Australia), 42 Soil conservation costs-benefits analysis: adoption Souza Tobacco Company (Brazil), 154 rates tied to, 204, 224, 225, 353-54; of conserva- Soviet Union, former, 175 tion measures' profitability, 222-24, 223 table; SPS (sanitary and phytosanitary measures of Uruguay country case studies of, 218 table, 220, 227 n3; of Round package), 77, 81 effects on yields, 220-22, 221 table; and lack of Spurling, Daphine, 188 capital markets, 225; and land tenure issues, 219, Squire, Lyn, 58, 217 224-25; maximization model reformulated to, Sri Lanka, 13, 57 217, 219; with positive NPV, 219; principles/ Srinivasan, T. N., 57 methodology of, 219-20; using simulations, 222; Srivastava, Jitendra, 2, 205 variability of, within sites, 224 State property, rights to, 84 box Soil degradation: agricultural productivity effects of, Staubo, C., 287 203, 204, 215-16, 225, 353; from cereal crops, Steinfeld, Henning, 2, 283 290; conservation techniques to arrest, 203, Stern, V. M., 244 216-17, 353; country case studies of, 218 table, Stiglitz, Joseph E., 51, 52, 56 220, 227 n3; defined, 215; excess carrying capaci- Strauss, John, 217 ty and, 23, 33 nl; fertilizer inefficiencies from, Striga hermonthica (African weed), 264-65 221; from livestock sector, 284-86; nonseparabili- Strong sustainability, and capital substitutions, 10 ty models of, 60-61; on-site/off-site costs of, 13, Structural adjustment effects: on Colombia's large- 203; on-site/off-site effects of, 217, 225, 227 ni; farm sector; 157; and farm household models, 59; participatory/local research on, 220; policy distor- on INCORA's funding sources, 158 tion's impact on, 226; unsystematic measurement Strutt, Anna, 77-78 of, 215, 225; yield losses from, 204, 220-22, 221 Stumpage charges, 310 380 Index Sub-Saharan Africa, 130 ity of, in U.S., 38, 43 Subsidies/incentives: annual cost of, 12; Bangladesh/Indonesian reform of, 75; for Taiwan, 193 Bosrupian development, 4, 5, 23, 346; of client- Tamil Nadu Nutrition Project, 108-9 centered agencies, 96, 109, 111; for conservation Tanzania, 23 investment, 180-81, 182, 204, 352, 354; created Tatumbla area (Honduras), 222, 224-25 by grant resources, 159, 168; 172 n6, 173 nl9; Taungya (agroforestry system), 214, 331, 331 table, externality costs-benefits and, 11, 15, 94; gender 332, 334, 335 table roles and, 186; intensification/extensification Taylor, J. E., 59, 60 effects of, 309; with interlinked contracts, 63-64; Technologies. See Resource-conserving technologies for irrigation systems, 209, 275, 360 nl; justified Tectonis grandis, 334 by off-site effects, 225, 226; of Kamala versus Tenancy. See Land tenure Pithuwas projects, 107; large-farmer bias of, Tendler, Judith, 112 25-26, 28, 30, 33 n3, 157; for livestock feed, 286; Terraces (conservation measure), 204, 222, 223 local ownership element of, 7, 90; for logging Thai-German Highland Development Project, 42-43, employment, 310, 320 n7; in Machakos versus 43 table Ethiopia, 23-24; in 1950s-60s, 92; nonprice inter- Thailand: contract farming effects in, 57; gender/effi- vention and, 15; for pesticide use, 14, 206, ciency studies in, 189 table; land titling program 244-46, 245 table, 360 nl; price distortions and, in, 62; Thai-German Highland Development 14-15; to promote agroforestry, 214, 330, 337, Project in, 42-43, 43 table 339, 360; to promote IPM, 208, 242-43, 246, 248, Thampapillai, Jesuthason, 23 267; to promote land-based livestock systems, Thiele, Graham, 2, 98 293; removal of, for sustainability, 14, 94, 239, Ticks, 286, 294 250 box, 346, 360 nl; removal of, to reduce defor- Tierra Blanca (Costa Rica), 216, 220-21, 222, 223, estation, 213, 287, 305-6, 313; to retain demand, 224 96, 114; risk of perverse, 226; for soil conserva- Tiffen, Mary, 23, 286 tion, 204, 225-26, 354; in Uruguay Round, 76 Tilman, D., 234 Sugarcane-based livestock feed, 295 box Timber stumpage charges, 310 Sumatra, 57-58, 315 Togo, 150 Sustainable agricultural development: "adding value" Trade liberalization policies: agricultural production characteristic of, 38; biodiversity conservation's effects of, 72-76, 73 figs.; and environmentalist importance to, 2, 182-84, 205-6, 231, 238-39; concerns, 5, 6, 71, 72, 78, 347; implications of, for and capital substitutions, 10, 345; and carrying developing countries, 80; in nonagricultural sec- capacity, 22-23; complex concept of, 9, 36-37, 45, tors, 77; policy reinstrumentation responses to, 345; countries with projects of, 38, 39 table, 44; 76-77; real income effects of, 72; research invest- current extent/impact of, 37-38; decentralized/par- ment effects of, 72, 79; and resource degradation, ticipatory approach to, ix-i, 1-2, 38, 41, 43, 46, 75, 78; and use of pollutive inputs, 74-75 348; for deficit-producing areas, 64-65; defined, Training and visit (T&V) system, 108 3, 9-10; environmental costs/benefits of, 10-11; Training of farmers: as agroforestry paratechnicians, environmental modifiers for, 7, 9-10, 93-94; 214, 337, 339; as extensionists, 99, 145-46, 151, failed projects of, 40; goals of, 7, 37, 44; high- 153, 350, 356; in IPM, 207-8, 251-52, 266, 356; input/low-input arguments for, 36, 37; interlinkage by municipios, 162-63; reducing cost of, 151, 207, mechanisms for, 63-64, 65-66, 346; land tenure 266, 356; using simple rules, 207-8, 266, 356 tied to, 6-7, 190, 293, 351; local ownership ele- Transaction costs: in low population density areas, 51; ment of, 7, 83, 90, 148, 348; local versus regional in village economy models, 52, 52 fig. assessment of, 10; people-centered principles of, Trans Amazonia Highway, 309 148-50, 149 fig., 150 fig.; policy framework for, 1, Transboundary effects in watersheds, 119, 123, 137 4, 45 table, 45-46, 50-51, 345-48; resource-con- Traxler, Greg, 79 serving technology and, 2, 37, 38, 99-100, 152, Tree Crop Production (Madelefia) Project, 326 154, 350-5 1; for surplus-producing areas, 64; tran- Tree products. See Tropical products sition from high-input to, 38, 44, 45; unprofitabil- Trees for People (Germany), 151 Index 381 Trees with agricultural crops (agroforestry system), estation agents, 303; deforestation definition by, 331, 331 table, 335 table 12, 302; farmer field schools of, 207, 266, 356; Tropical Agricultural Research and Higher Education feed grain study by, 289 box; IPM support from, Center (CATIE), 225, 324, 326 44, 207, 242, 250 box, 356; pesticide guidelines Tropical forest: in "ambiguous" zone, 304, 304 table; of, 249 certified/ecolabeled timber of, 315; as global bio- United Nations Framework Convention on Climate diversity share, 287; and land tenure issues, 314; Change, 316 and land titling reform, 312, 313 table; logging of, United States: Amazonian products marketed in, 233, 213, 307, 318-19, 320 n5; 1980-90 deforestation 234 box; national IPM policy of, 250; pesticide of, 12, 211, 302, 319 nl, 358; species richness of, exports fromrr, 255; sustainable agriculture in, 38, 232, 232 box. See also Deforestation; Tropical for- 43 est conservation policies U.S. Agency for International Development Tropical forest conservation policies: of appreciated (USAID), 224, 225, 250 box, 277, 326 exchange rates, 306, 308 table; on colonization U.S. Foreign Assistance Act, Percy Amendment to, schemes, 309, 311 table, 319; of common proper- 196 ty regimes, 312-13, 313 table; of export Universal Soil Loss Equation (model), 220 bans/taxes, 307-8, 308 table; of import restric- Urea nutrient supplementation, for livestock, 289 box tions, 306, 308 table; of land/capital gains taxes, Uribe, Eduardo, 175 313, 313 table; of land tenure security, 314, 317 Uruguay Round: and agricultural production reloca- table, 359; of land titling reform, 312, 313 table; of tion, 71, 73--74; "blue box" price supports of, 76, nonbiased extension, 309, 311 table; on nontimber 80; economic welfare effects of, 78; environmen- forest products, 315, 317 table, 319, 359; of pay- tal subsidies in, 76; GTAP modeling of, 77-78; ments for environmental services, 316, 317 table, nonagricultural agreements of, 77; SPS measures 319, 359; of population control, 306, 308 table; of of, 77. See also Trade liberalization policies price controls, 305-6, 308 table, 320 n4; of pro- USAID. See U.S. Agency for International tected area management, 311; on road construc- Development tion, 310-11, 311 table, 319; on stumpage charges, Usufruct rights: gender issues and, 191; needed for 310; on subsidization schemes, 308-9, 311, 311 sustainability, 148; over multiple use resources, table, 313, 313 table, 319, 359; on timber values, 119; with perennial crops, 312; religious laws on, 314-15, 317 table, 319; on urban employment/ 187, 199 n2 wages, 317-18, 318 table Uttar Pradesh (India), 274 Tropical products: bans/taxes on, 307-8; of Central Uzbekistan, 279 American agroforestry projects, 331-32, 338; import restrictions on, 306; income/population Van den Brink, Rogier, 88 growth and, 306; market opportunities for, 333, Van Koppen, Barbara, 192 340; policies to depress demand for, 308 table; Vecinos Mundiales (Central America), 204-5, 226, price controls on, 305-6 354 Tsavo West National Park (Kenya), 180 Vegetables. See Fruits/vegetables Tubewell irrigation systems, 273, 274 Vietnam, 256, 256 table Turrubares (Costa Rica), 216, 221, 222-23, 223 table, Village economies: credit and crop insurance markets 224 of, 53-54; market imperfections in, 52-53; trans- Tyers and Anderson's model (1992), 73 action costs and differentiation in, 52, 52 fig., 53 fig. Udaipur district (Rajasthan, India), 98, 133-34 Village Infrastructure Project (Indonesia), 105 Udry, C., 52 Violence, as agricultural disincentive, 26, 30, 157 United Nations Conference on Environment and La Violencia of 1958, 30 Development (UNCED), 243 Vosti, Stephen A., 227 United Nations Environment Programme (UNEP), 249, 284 Wade, Robertl, 104, 273, 275 United Nations Food and Agriculture Organization Waggoner, Paul E., 320 n9 (FAO): on Brazilian land reform, 167; on defor- Waibel, H., 263 382 Index Wang, M., 193 Wide hybridization, 262 Warabandi rotation system, 274, 275 Wik, M., 57, 60 Warburton, H., 255 Wildlife: allocated revenues from, 179; decentralized Wastewater: anaerobic purification of, 210, 295; from management of, 293 livestock processing, 290-91 Wildlife Service (Kenya), 178, 180-81 Water. See Irrigation water Williams, Margaret, 175 Water Commission of the Ministry of Agriculture Windbreaks (agroforestry system), 332, 334, 338, 340 (Israel), 277 n2 Waterlogging, 12, 273, 278 Women: agroforestry participation by, 334-35; in col- Water points, and soil degradation, 285 lective action groups, 192; communal tenure and, Water pricing: Egyptian mechanisms of, 277; at full 191; development policy's neglect of, 10 1-2, 106, cost, 360 nl; Israeli mechanisms of, 277; at mar- 113, 186, 352; education level of, 187, 360 n2; as ginal cost, 276-77; outmoded policies of, 272-73, "equity modifier," 93; extension system's relations 357; and poor irrigation practices, 278; water trad- with, 187-88, 190; as extension workers, 109; in ing/markets and, 275 female marriage contracts, 55; in The Gambia pro- Watershed management: appraisal process in, 124; ject, 115, 196; gender-and-development approach and competing multiple stakeholders, 119, 138, to, 102, 197-98; and household welfare effects, 349; conservation/regeneration elements of, 97, 192-93, 193 table, 195-96 table; income flows 122; defined, 119; external agencies' role in, 97, of, 193, 196; in Indo-German Watershed 122-24; inequitable local participation in, 97, Development Program, 125; rural unemployment 124-25; of mobile resources, 119, 123; participa- of, 29; technical efficiency studies of, 188-90, 189 tory principles of, 96-97; Rio Cabuyal case study table; time resources of, 188, 198; unrepresented of, 125-28; of storage capacity resources, 120. See in watershed management, 97, 350; women-in- also Local organizations in watershed manage- development projects for, 102, 197, 198. See also ment; Microwatershed rehabilitation Gender; Gender-and-development approach; Watershed Organization Trust (WOTR), 136-37, 138 Women-in-development approach Watersheds: defined, 119; multiple resources/uses in, Women-in-development approach: in The Gambia, 119-20; transboundary effects in, 119, 123, 137. 110, 112, 115, 196; gender-and-development See also Irrigated agriculture; Local organizations approach versus, 102, 197, 353; pros and cons of, in watershed management; Watershed management 198 Water trading/markets, 209; examples of, 274-75; Women-in-Development (WID) Project (The income distribution effects of, 275-76; precondi- Gambia), 110, 112, 115, 196 tions for, 275, 279, 280, 357 Woodlots (agroforestry system), 331, 331 table, 332, Weak sustainability, and capital substitutions, 10 334, 335 table Wedin, D., 234 Wood Mackenzie Consultants, Ltd., 255 Weed management. See Herbicides World Bank: agricultural extension workshop of, Weir, D., 255 131; community-driven development strategies Welfare effects: of female income expenditures, of, ix-x, 96, 110; on failed projects, 40; IPM sup- 192-93, 193 table, 194-95 table, 199 n4, 352; of port from, 242, 250 box; irrigation investments trade liberalization policies, 78 by, 208, 271; irrigation study by, 274, 279-80; on Wells, Michael, 175 Malaysia's industrialization, 310; nontimber for- West Africa, 17, 312, 335-36 est products project of, 315; research funding West Java, 231 by, 19 Wetlands Water District (Central Valley, California), World Conference on Women (Mexico, 1975), 196 275 World Health Organization, 249, 255 Wetzstein, M. E., 267 World Neighbors, 41, 150, 153-54 Wheat: diseases of, 262; fungicide use on, 254, 255; World Resources Institute, 13, 284 genetic engineering of, 263; insect pests of, 261; World Trade Organization (WTO), 6, 71, 77, 80-81. molecular marker selection on, 262; pest-resis- See also Uruguay Round tance of, 260 table, 261, 262; wide hybridization World Vision, 151 of, 262 World Wide Fund for Nature, 178 Index 383 World Wildlife Fund, 151, 152 Zambia, 57-58, 59, 193 WOTR (Watershed Organization Trust), 136-37, 138 Zilberman, David, 250, 267, 273, 275 WTO. See World Trade Organization Zimbabwe, 23, 105; AGRITEX system in, 108; decentralized conservation in, 177 table, 177-78, Yemen Basic Education Project, 109 180, 184 n2; fiscal decentralization in, 179; irriga- Yields: and conservation investment, 61, 64; host- tion system in, 111 plant resistance and, 263-64; losses in, and pesti- Zimbabwe Department of Agricultural, Technical, cide use, 259, 355; of small versus large farmers, and Extension Services (AGRITEX), 108 27-28; soil degradation's impact on, 204, 220-22, Zonal coordinators of RELU, 140 221 table; variability in, 220 Zonation, of livestock production systems, 294 DISTRIBUTORS OF cOie-BookImportcentre The Middle East Observer INDONESIA P.O. Box 2825 41, Sherif Street Pt. Indira Limited W ORLD BANK Beijing Cairo Jalan Borobudur20 Tel: (20 2) 393-9732 P.O. Box 181 PUBLICATIONS Chinese Corporation for Promotion of Fax: (20 2) 393-9732 Jakarta 10320 Prices and credit terms varyfrom Humanities Tel: (62 21) 390-4290 country to country. Consult your 52 You Fang Hu Tong, FINLAND Fax: (62 21) 390-4289 local distributor before placing an B Nei Da Jie Akateeminen KirIakauppa Be'ijing P.O. Box 128 IRAN order. Tel: (86 10)66072494 FIN-0OO1 Helsinki Ketab Sara Co. Publishers ARGENTINA Fax: (86 10) 660 72 494 Tel: (358 0) 121 4418 Khaled Eslamboli Ave., 6th Street ARGEiciNT del LibroINA ernacional Fax: (358 0) 121-4435 Delafrooz Alley No. 8 Oficina del Libro b nternacional COLOMBIA E-mail: akatilaus@stockmann.fi P.O. Box 15745-733 Av. Cordobs 1877 Infoenlace Ltda. URL: http://www.akateeminen.com/ Tehran 15117 1120 Buenos Aires Carrera 6 No. 51-21 Tel: (98 21) 8717819; 8716104 Tel: (54 1) 815-8354 Apartado Aereo 34270 FRANCE Fax: (98 21) 8712479 Fax: (54 1) 815-8156 Santatride Bugotri, D.C. Editions Eska E-mail: ketab-sara@necdanet.ir E-mail: olilibro@satlink.como Tel: (57 1) 285-2798 5, avenue de l'Opera AUSTRAUA, FIJI, PAPUA NEW GUINEA, Fax: (57 1) 285-2798 75001 Paris Kowkab Publishers ASTRLOMNISA,NFIJ, VANUATNW UINA Tel: (33 1) 42-86-56-00 P.O. Box 19575-511 ANDSAMOA COTE D'IVOIRE Fax: (33 1)42-60-45-35 Tehran D.A. Information Serices Center d'Edition et de Diffusion Atricaines Tel: (98 21) 258-3723 648 Whitehonf e Road (CEDA) GERMANY Fax: (98 21) 258-3723 Mitchamr 3132 04 B.P. 541 UNO-Verlag Victoa Abidjan 04 Poppelsdorfer Allee 55 IRELAND Telr (i1) 3 9210 7777 Tel: (225) 24 6510; 24 6511 53115 Bonn Government Supplies Agency Fax: (61) 3 9210 7788 Fax: (225)25 0567 Tel: (49 228) 949020 Olfig an tSolithair E-mail: service@dadireftcom.au Fax: (49 228) 217492 4-5 Harcourt Road UEL: http://wwwdadirect.com.au CYPRUS URL httpY/wwwuno-verlag.de Dublin 2 URL: http://wwwdedirect.com.aaCenter for Applied Research E-mail: snoverlsg@aol.com Tel: (353 1) 661-3111 AUSTRIA Cyprus College Fax: (353 1) 475-2670 Gerold and Co. 6, Diogenes Street, Engomi GHANA Weihburggasse 26 RO. Box 2006 Epp Books SerVices ISRAEL A-1011 Wien Nicosia P.O. Box 44 Yozmot Literature Ltd. Tel: (43 i) 512-47-31 -0 TTeI: (357 2) 59-0730 TUC P.O. Box56055 Fax: (43 1) 512-47-31-29 Fax: (357 2) 66-2051 Accra 3 Yohanan Hasandlar Street URL: http:/1www.gerod.co/at.on2ine Tel: 223 21 778843 Tel Aviv 61560 URL: htt:/lwww.grold.colLonhineCZECH REPUBLIC Fax: 223 21 779099 Tel: (972 3) 5285-397 BANGLADESH ~~~~USIS, NIS Prode(na Fax: (972 3) 5285-397 BANGLADESH Haveikova 22 GREECEFa:(73)58-9 Micro Industries DevelopmentHaekv22GEC Assistance Society (MIDAS) 130 00 Prague 3 Papasotiriou S.A. R.O.Y Intemational Aouse 5s Road 16 Tel: (420 2) 2423 1486 35, Stournara Str. PO Box 13056 Dhanouse RoAd 16 Fax: (420 2) 2423 1114 106 82 Athens Tel Aviv 61130 Dhaka 1209 URL: http://wwwnis.cz/ Tel: (30 1) 364-1826 Tel: (972 3) 649 9469 Tel: (880 2) 326427 Fax: (30 1) 364-8254 Fax: (972 3) 648 6039 Fax: (880 2) 611188 DENMARK E-mail: royil@netvision.net.il SamfundsLItteratur HAITI ;JELGIUM Rosenoerns Alle 11 Culture Diffusion Palestinian Authority/Middle East Jean De Lannoy DK-1970 Frederiksberg C 5, Rue Capois Index Information Services Avn du RoiL202 Tel: (45 35) 351942 C.P. 257 PO.B..19502 Jerusalem A06. Brussels 2Fax: (45 35) 357822 Port-au-Prince Tel: (972 2) 6271219 Tel: (322) 538-5169 URL: http://wwwsl.cbs.dk Tel: (509) 23 9260 Fax: (9722) 6271634 Fax: (32 2) 538-0841 Fax: (509) 23 4858 ECUADOR ITALY, LIBERIA BRAZIL Libri Mundi HONG KONG, CHINA; MACAO Licosa Commissionaria Sansoni SPA Publicacies Tecnicas Intemacionais Ltda Libreria Intemacional Asia 2000 Ltd. Via Duca Di Calabria, 1/1 PublicabesXto Teceicas minernac,onas LIa P RO. Box 17-01-3029 Sales & Circulation Department Casella Postale 552 Rua Peixoto Gomide, 209 Juan Leon Mera 851 302 Seabird House 50125 Firenze Tel: (55 11) 259-6644 Quito 22-28 Wyndham Street, Central Tel: (39 55) 645-415 Tel: (S5 11) 259-6644 Tel: (593 2) 521-606; (593 2) 544-185 Hong Kong, China Fax: (39 55) 645-415 Fax: (55 11) 258-6990 Fax: (593 2) 504-209 Tel: (852) 2530-1409 E-mail: licosa@ftbcc.it E-mail: postmaster@pti.uol.br E-mall: librimul @librimundi.com.ec Fax: (852) 2526-1107 URL: http://www.fthbcc.it/licosa URL: http://www.uol.br E-mail: librimu2@librimundi.com.ec E-mail: sales@asia2000.com.hk CANADA URL: http://www.asia2000.com.hk JAMAICA CAnADA Publishing Co. Ltd. CODEU Ian Randle Publishers Ltd. Renouf Publishing Co. Ltd. Ruiz de Castilla 763, Edif. Expocolor HUNGARY 206 Old Hope Road, Kingston 6 5369 Canotek Road Primer piso, Of. #2 Euro Info Service Tel: 876-927-2085 Ottawa, Ontario KIJ 9 Quito Margitszgeti Europa Haz Fax: 876-977-0243 Tel: (613) 745-2665 TeVFax: (5932)507-383;253-091 H-1138 Budapest E-mail: irpl@colis.com E-mail: orderdept@renoutbooks.com E-mail: codeu@impsat.net.ec Tel: (36 1) 350 80 24, 350 80 25 URL: http:l/ www.renoufbooks.com Fax: (36 1) 350 90 32 JAPAN EGYPT, ARAB REPUBLIC OF E-mail: euroinfo@mail.matav.hu Eastem Book Service CHINA Al Ahram Distribution Agency 3-13 Hongo 3-chome, Bunkyo-ku China Financiai a Economic Publishing Al Galaa Street INDIA Tokyo 113 China Financial & Economic Publishing Cairo Allied Publishers Ltd. Tel: (81 3) 3818-0861 House Tel: (20 2) 578-6083 751 Mount Road Fax: (81 3) 3818-0864 8, Da Fe Si Dong Jie Fax: (20 2) 578-6833 Madras - 600 002 E-mail: orders@svt-ebs.co.jp Tel (86 10) 6333-8257 Tel: (91 44) 852-3938 URL: http.//www.bekkoame.orip/-svt-ebs Fax: (86 10) 6401-7365 Fax: (91 44) 852-0649 KENYA Oasis Official Bucharest SWITZERLAND Africa Book Service (E.A.) Ltd. P.O. Box 3627 Tel: (40 1) 613 9645 Librairie Payot Service Institutionnel Quaran House, Mfangano Street Wellington Fax: (40 1) 312 4000 Cotes-de-Montbenon 30 PO, Box45245 Tel: (644)499 1551 1002 Lausanne Nairobi Fax: (64 4) 499 1972 RUSSIAN FEDERATION Tel: (41 21) 341-3229 Tel: (254 2) 223 641 E-mail: oasis@actrix.gen.nz Isdatelstvo Fax: (41 21) 341-3235 Fax: (254 2) 330 272 URL: http://www.oasisbooks.co.nz/ 9a, Kolpachniy Pereulok Moscow 101831 ADECO Van Diermen EditionsTechniques KOREA, REPUBLIC OF NIGERIA Tel: (7 095) 917 87 49 Ch. de Lacuez 41 Dayang Books Trading Co. University Press Limited Fax: (7 095) 917 92 59 CH1807 Blonay International Division Three Crowns Building Jericho Tel: (41 21) 943 2673 783-20, Pangba Bon-Dong, Socho-ku Private Mail Bag 5095 SINGAPORE; TAIWAN, CHINA Fax: (41 21) 943 3605 Seoul Ibadan MYANMAR; BRUNEI Tel: (82 2) 536-9555 Tel: (234 22) 41-1356 Hemisphere Publication Services THAILAND Fax: (82 2) 536-0025 Fax: (234 22) 41-2056 41 Kallang Pudding Road #04-03 Central Books Distribution E-mail: seamap@chollian.net Golden Wheel Building 306 Silom Road NORWAY Singapore 349316 Bangkok 10500 Eulyoo Publishing Co., Ltd. SWETS Norge AS Tel: (65) 741-5166 Tel: (66 2) 235-5400 46-1, Susong-Dong Book Deparment, Postboks 6512 Etterstad Fax: (65) 742-9356 Fax: (66 2) 237-8321 Jongro-Gu N-0606 Oslo E-mail: ashgate@asianconnect.com Seoul Tel: (47 22) 97-4500 TRINIDAD & TOBAGO Tel: (82 2) 734-3515 Fax: (47 22) 97-4545 SLOVENIA AND THE CARRIBBEAN Fax: (82 2) 732-9154 Gospodarski Vestnik Publishing Group Systematics Studies Ltd. PAKISTAN Dunajska cesta 5 St. Augustine Shopping Center LEBANON Mirza Book Agency 1000 Ljubljana Eastem Main Road, St. Augustine Librairie du Liban 65, Shahrah-e-Quaid-e-Azam Tel: (386 61)133 83 47; 13212 30 Trinidad& Tobago, West Indies RO. Box 11-9232 Lahore 54000 Fax: (386 61) 133 80 30 Tel: (868) 845-8466 Beirut Tel: (92 42) 735 3601 E-mail: repansekj@gvestnik.si Fax: (868) 645-8467 Tel: (961 9) 217 944 Fax: (92 42) 576 3714 E-mail: tobeftrnidad.net Fax: (9619) 217 434 SOUTH AFRICA, BOTSWANA Oxford University Press For single titles: UGANDA MALAYSIA 5 Bangalore Town Oxford University Press Southeem Africa Gustro Ltd. University of Malaya Cooperative Sharae Faisal Vasco Boulevard, Goodwood PO Box 9997, Madhvani Building Bookshop, Limited PO Box 13033 P.O. Box 12119, Ni City7463 Plot 16/4 Jinja Rd. P.O. Box 1127 Karachi-75350 Cape Town Kampala Jalan Pantai Baru Tel: (92 21) 446307 Tel: (27 21) 595 4400 Tel: (256 41) 251 467 59700 Kuala Lumpur Fax: (92 21) 4547640 Fax: (27 21) 595 4430 Fax: (256 41) 251 468 Tel: (60 3) 756-5000 E-mail: ouppak@TheOffice.net E-mail: oxfordQoup.co.za E-mail: gus@swiftuganda.com Fax: (60 3) 755-4424 E-mail: umkoop@tm.net.my Pak Book Corporation For subscription orders: UNITED KINGDOM Aziz Chambers 21, Queen's Road International Subscription Service Microinfo Ltd. MEXICO Lahore RO. Box 41095 RO. Box 3, Omega Park, Alton, INFOTEC Tel: (92 42) 636 3222; 636 0885 Craighall Hampshire GU34 2PG Av. San Fernando No. 37 Fax: (92 42) 636 2328 Johannesburg 2024 England Col. Toriello Guerra E-mail: pbc@brain.net.pk Tel: (27 11) 880-1448 Tel: (44 1420) 86848 14050 Mexico, D.F. Fax: (27 11) 880-6248 Fax: (44 1420) 89889 Tel: (52 5) 624-2800 PERU E-mail: iss@is.co.za E-rmail: wbank@microinfo.co.uk Fax: (52 5) 624-2822 Editorial Desarrollo SA URL: http://www.microinfo.co.uk E-mail: infotec@rtn.net.mx Apartado 3824, Ica 242 OF. 106 SPAIN URL: http://rtn.net.mx Lima 1 Mundi-Prensa Libros, S.A. The Stationery Office Tel: (51 14) 285380 Castello 37 51 Nine Elms Lane Mundi-Prensa Mexico S.A. de C.V. Fax: (51 14) 286628 28001 Madrid London SW8 5DR clRio Panuco, 141-Colonia Cuauhtemoc Tel: (34 91) 4 363700 Tel: (44 171) 873-8400 06500 Mexico, D.F PHILIPPINES Fax: (34 91) 5 753998 Fax: (44 171) 873-8242 Tel: (52 5) 533-5658 Intemational Booksource Center Inc. E-mail: libreria@mundiprensa.es URL: http://www.theso.co.ukl Fax: (52 5) 514-6799 1127-AAntipolo St, Barangay, Venezuela URL: http://wmw.mundiprensa.com/ Makati City VENEZUELA NEPAL Tel: (63 2) 896 6501; 6505; 6507 Mundi-Prensa Barcelona Tecni-Ciencia Libros, S.A. Everest Media Intemational Services (P) Ltd. Fax: (63 2) 896 1741 Consell de Cent, 391 Centro Cuidad Comercial Tamanco GPO Box 5443 08009 Barcelona Nivel C2, Caracas Kathmandu POLAND Tel: (34 3) 488-3492 Tel: (58 2) 959 5547; 5035; 0016 Tel: (977 1) 472 152 International Publishing Service Fax: (34 3) 487-7659 Fax: (58 2) 959 5636 Fax: (977 1) 224 431 Ul. Piekna 31/37 E-mail: barcelona@mundiprensa.es 00-677 Warzawa ZAMBIA NETHERLANDS Tel: (48 2) 628-6089 SRI LANKA, THE MALDIVES University Bookshop, University of Zambia De Lindeboomlinternationale Publicaties b.v- Fax: (48 2) 621-7255 Lake House Bookshop Great East Road Campus P.O. Box 202, 7480AE Haaksbergen E-mail: books%ips@ikp.atm.com.pl 100, Sir Chittampalam GardinerMawatha RO. Box32379 Tel: (31 53) 574-0004 URL: http://www.ipscg.waw.pl/ips/export/ Colombo 2 Lusaka Fax: (31 53) 572-9296 Tel: (94 1) 32105 Tel: (2601) 252 576 E-mail: lindeboo@worldonline.nl PORTUGAL Fax: (94 1) 432104 Fax: (260 1) 253 952 URL: http://www.worldonline.nl/-lindeboo Livraria Portugal E-mail: LHL@sr.lanka.net Apart ado 2681, Rua Do Carmo 70-74 ZIMBABWE NEW ZEALAND 1200 Lisbon SWEDEN Academic and Baobab Books (Pvt.) Ltd. EBSCO NZ Ltd. Tel: (1) 347-4982 Wennergren-Williams AB 4 Conald Road, Graniteside Private Mail Bag 99914 Fax: (1) 347-0264 P. 0. Box1305 PRO. Box 567 New Market S-171 25 Solna Harare Auckland ROMANIA Tel: (46 8) 705-97-50 Tel: 263 4 755035 Tel: (64 9) 524-8119 Compani De Librarii Bucuresti S.A. Fax: (46 8) 27-00-71 Fax: 263 4 781913 Fax: (64 9) 524-8067 Str. Lipscani no. 26, sector 3 E-mail: mailfwwi.se THE WORLD BlANK 1818 H Street, N.W. Washington, D.C. 20433 USA Telephone: 202.477.1234 Facsimile: 202.477.6391 Telex: MCI 64145 WORLDBANK MCI 248423 WORLDBANK World Wide Web: http://www.worldbank.org E-mail: books@worldbank.org 14 42 49 9 780821 342497 ISBN 0-82134249-