a: WORLD BANK ENVIRONMENT PAPER NUMBER 3 417t Environmental Economics and Sustainable Development Mohan Munasinghe 9/ _ :0f L e Xf W - E= t 00Sitt aQ½ RECENT WORLD BANK ENVIRONMENT PAPERS No. 1 Cleaver, Munasinghe, Dyson, Egli, Peuker, and Wencelius, editors, Conservation of West and Central African Rainforests/Conservation de laforet dense en Afrique centrale et de l'Ouest No. 2 Pezzey, Sustainable Development Concepts: An Economic Analysis WORLD BANK ENVIRONMENT PAPER NUMBER 3 Environmental Economics and Sustainable Development Mohan Munasinghe The World Bank Washington, D.C. Copyright G 1993 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 July 1993 Environment Papers are published to commnunicate the latest results of the Bank's environmental work to the development community with the least possible delay. The typescript of this paper therefore has not been prepared in accordance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. The World Bank does, not guarantee the accuracy of the data included in this publication and accepts no responsibility whatsoever for any consequence of their use. Any maps that accompany the text have been prepared solely for the convenience of readers; the designations and presentation of material in them do not imply the expression of any opinion whatsoever on the part of the World Bank, its affiliates, or its Board or member countries concerning the legal status of any country, territory, city, or area or of the authorities thereof or concerning the delimitatiorL of its boundaries or its national affiliation. 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 permission 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, 27 Congress Street, Salem, Massachusetts 01970, U.S.A. The complete backlist of publications from the World Bank is shown in the annual Index of Publications, which contains an alphabetical title list (with full ordering information) and indexes of subjects, authors, and countries and regions. The latest edition is available free of charge from the Distribution Unit, Office of the Publisher, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, The World BanLk, 66, avenue d'Iena, 75116 Paris, France. Mohan Munasinghe is chief, Policy and Research Division, in the Environment Department of the World Bank. Library of Congress Cataloging-in-Publication Data Munasinghe, Mohan, 1945- Environmental economics and sustainable development / Mohan Munasinghe. p. cm. - (World Bank environment paper ; no.3) Includes bibliographical references. ISBN 0-8213-2352-0 1. Environmental policy-Economic aspects. 2. Economic development-Environmental aspects. I. Title. II. Series. HC79.E5M865 1993 363.7-dc2O 92-42952 CIP Foreword The decade of the 1980s has witnessed a fundamental development may be identified; based on economic, change in the way governments and development ecological, and socio-cultural criteria. Reconciling agencies think about environment and development. these concepts and operationalizing them will be a The two are no longer regarded as mutually exclu- formidable task, which is only now getting under- sive. It is now recognized that a healthy enviromnent way. In the meantime those maldng decisions have is essential to sustainable development and a healthy to find ways of introducing such concerns into their economy. Moreover, economists and planners are analysis in a practical way. beginning to recognize that economic development The objectives of this paper recognize these which erodes natural capital is often not successful. needs. It seeks to help analysts, practitioners and In fact, development strategies and programs which policymakers in the field by reviewing the latest do not take adequate account of the state of critical thinking on the concepts and techniques of sustain- resources-forests, soils, grasslands, freshwater, able development and the valuation of environmental coastal areas and fisheries-may degrade the re- impacts, so that they can be explicitly taken into source base upon which future growth is dependent. account in development decisionmaking. Since its creation, the Vice Presidency for Envi- The application of environmental economic ronmentally Sustainable Development (ESD) has principles is examined, in order to not only extend placed the highest priority on the analysis of these conventional cost-benefit analysis but also make bet- important issues. Within ESD, the Environment De- ter use of other methods such as multi-criteria deci- partment's work, in particular, has focused on the sionmaking. This analytical discussion is further links between environment and development, and illustrated by a careful selection of the most recent the implications of these linkages for development empirical work. These case studies cover a range of policy in general. The objective of the Environment projects and circumstances to be encountered in a Paper Series is to make the results of our work variety of developing countries. available to the general public. Increasing environmental awareness and con- cerns over sustainability have broadened the range of issues that need to be examined in the assessment Ismail Serageldin of the potential impacts of proposed projects and Vice President prograrns. Three different concepts of sustainable Environmentally Sustainable Development Acknowledgement The author is pardcularly grateful to Noreen Beg. Special thanks are owed also to Emst Lutz. In addfition, the author is grateful to Shakeb Afsah, Edward Barbier, Robin Bates, Wilfred Beckerman, Jan Bojo, Wilfrido Cruz, Partha Dasgupta, John Dixao, John English, Gunnar Kohlin, Randy Kramer, Karl-Goran Maler, Anil Markandya, Peter Meier, David Pearce, Annika Persson, Narendra Sharma, Adelaida Schwab, Tom Teitenberg, Hirofumi Uzawa, Jeremy Warford, Diale Whittington, and Carlos Young, for assistance. This work was supported in part by grants from the Governments of Norway and Sweden. Contents Abstract vi PART I: ANALYTICAL FRAMEWORK 1 1. Introduction I Conceptual Basis of Sustainable Development 1 Focus of the Paper 4 2. Linking Economics and the Environment 5 The Role of Environmental Economics 5 Recent Developments 7 Conventional Project Evaluation 7 National Income Accounts and Macroeconomic Performance 11 Economywide Policies and the Environment 13 3. Framework for Environmental-Economic Decisionmaking 19 Valuation of Environmental Costs and Benefits 19 Multi-Objective Decisionmaking (or Multi-Criteria Analysis) 28 The Discount Rate 33 Risk and Uncertainty 35 PART II: CASE STUDIES 38 4. Short Case Studies of Environmental Valuation 38 Change in Productivity Method 38 Loss of Earnings Method 39 Travel Cost and Contingent Valuation Methods 40 Contingent Valuation Method to Estimate Option, Existence, and Bequest Values 43 5. Two World Bank Valuation Case Studies 46 Vatuation of Biophysical Resources in Madagascar 46 Improving Decisionmaking in the Sri Lanka Power Sector 51 6. Conclusions 60 Bibliography 61 Annex 1. Using Shadow Prices 69 Annex 2. Summaries Of Environmental Valuation Case Studies 72 Change in Productivity Method 72 Loss of Earnings Method 78 Travel Cost and Contingent Valuation Methods 80 Annex 3. Estimating International Travel Costs 93 Annex 4. Using a Simple Biodiversity Index 95 Annex 5. Recent Work on Policies and Environmental Implications 97 Abstract One essential step towards achieving sustainable cation. To facilitate this, an extended bibliography is development is the economically efficient manage- provided. ment of natural resources. This paper explains the First, a number of shorterdeveloping country case key role of environmental economics in facilitating studies which cover a wide range of practical valuation the more effective incorporaLtion of environmental methods, are reviewed. Next, two longer case studies concerns into development clecisionmaking. Tradi- describe the current state-of-the-art in applicalion of a tionally, the economic analysis of projects and poli- combination of valuation techniques in developing cies (including the techniques of shadow pricing), countries. The Madagascar example focuses or the use has been developed to help a country make more of methods such as contingentvaluation andtravel cost, efficient use of scarce resources. However, "exter- to value forest resources and biodiversity. The Sri nalities", mainly those arising from adverse environ- Lanka case study examines the scope for application mental consequences, often have been neglected in of multicriteria techniques (in addition to economic the past. It is also important to recognize the social valuation), to improve decisionmaking in the energy and ecological objectives that are a part of sustainable sector. development, and to reconcile these concepts and We may conclude generally, that furtheir appli- operationalize them within the economic framework. cation to practical problems in developing countries This paper reviews concepts and techniques for is required (rather than further theoretical develop- valuation of environmental impacts that enable such ment), of the environmental valuation concepts and environmental considerations to be explicitly con- techniques presented in the paper. Such case study sidered in the conventional cost-benefit calculus work can be most effective when carried out as part used in economic decisionmaking. Key related as- of project preparation. A major purpose in this en- pects including environmental impacts of economy- deavor is at least to indicate orders of magnitude, if wide policies (both macroeconomic and sectoral), it is not possible to provide more accurate numbers. discount rate issues, and mulli-criteria analysis are Some alternatives can be ruled out, and gross envi- reviewed. The process of internalizing environ- ronmental errors avoided in this fashion. Also, one mental externalities may be ifacilitated by making can often identify the key environmental indicators even rough qualitative assessments early on in the to which the decision is sensitive and focus atlention project evaluation cycle. The advantages of such an on them. approach include: the early exclusion of options that The evidence presented suggests that the valu- are not sound environmentally; more effective in- ation techniques for determining "use values" may depth consideration of environmentally preferable be applied successfully in appropriate cases. How- alternatives; and better opportunities for redesigning ever, examples involving the estimation of "non-use projects and policies to achieve, sustainable develop- values" are virtually non-existent in the developing ment goals. world, and rather scarce even in the industrialized There are an increasing number of attempts to nations. both improve and make use of economic techniques The use of multiobjective decision methods also to value environmental assets in developing coun- needs to be expanded, (as illustrated in the Sri Lanka tries. This paper seeks to help practitioners in the case study), given the difficulties in using purely field, whose main concern is to keep up with the monetary methods of cost benefit analysis in many advances most relevant to their own areas of appli- contexts. PART I: ANALYTICAL FRAMEWORK 1. Introduction The state of the environment is a major worldwide provides many other generalized services ranging concern today. Pollution in particular is perceived as from simple amenities to irreplaceable life support a serious threat in the industrialized countries, where functions (e.g., stabilization of the global climate or the quality of life had hitherto been measured mainly filtering out of harmful ultraviolet rays by the strato- in terms of growth in material output. Meanwhile, spheric ozone layer). One complication is that these environmental degradation has become a serious functions tend to interact negatively-for example, impediment to economic development and the al- overloading the waste absorbing capability will gen- leviation of poverty in the developing world. erally reduce the supply of other productive environ- Mankind's relationship with the environment mental inputs as well as general life-support has gone through several stages, starting with functions. primitive times in which human beings lived in a In this context, several authors have pointed out state of symbiosis with nature, followed by a pe- that human activities are part of an open dynamic riod of increasing mastery over nature up to the socioeconomic subsystem which is embedded in the industrial age, and culminating in the rapid mate- global ecosphere (examples of early writings are: rial-intensive growth pattern of the twentieth cen- Odum 1973, Odum 1975; while some recent ones tury which adversely affected natural resources in include Costanza 1991, and Folke and Janssen many ways. The initial reaction to such environ- 1992). The rapid growth of the socioeconiomic sub- mental damage was a reactive approach charac- system in modem times has begun to overload some terized by increased clean-up activities. Most of the capabilities of the ecosystem (locally as well recently, mankind's attitude towards the environ- as globally). Many environmentalists argue that un- ment has evolved to encompass the more proactive bounded (and especially material-intensive) eco- design of projects and policies that will help antici- nomic growth would not be sustainable in the long pate and minimize environmental degradation. In run, given that the ecosphere is finite (see, for exam- this context, the world is currently exploring the ple: Goodland, Daly and El-Serafy 1991). concept of sustainable development, an approach that will permit continuing improvements in the pre- Conceptual Basis of Sustainable sent quality of life at a lower intensity of resource Development use, thereby leaving behind for future generations an undiminished or even enhanced stock of natural re- In addition to its environmental roots, the concept of sources and other assets.1 sustainable development which emerged in the The environmental assets that we seek to protect 1980s, draws also on the experience of several dec- provide three main types of services to human soci- ades of development efforts. In the 1950s and 1960s, ety, and the consequences of their degradation must the focus of economic progress was on growth and be incorporated into the decisionmaking process. increases in output, based mainly on the concepts of First, it has been known for centuries that the natural economic efficiency. By the early 1970s, the large and resource base provides essential raw materials and growing numbers of poor in the developing world, and inputs which support human activities. Second, the the inadequacy of "trickle-down" benefits to these environment serves as a sink to absorb and recycle groups, ledto greater efforts to directly improve income (often at little or no cost to society) the waste prod- distribution. The development paradigm shifted to- ucts of economic activity. This function has received wards equitable growth, where social (distributional) much greater attention in modem times, especially objectives were recognized as distinct from and as where overloading of the sink capacity has occurred. important as economic efficiency (see left side of Fig- Finally, there has been increasing recognition, par- ure 1.1). ticularly in the last two decades, that the environment 2 Part I: Analytical Framework Figure .1 Tradeoffs among the Three Main Objectives of Sustainable Development - Income redistrib. - Environmental - Employment Assessment - Targetted - Valuation Assistance - Internalisation SOCIAL QBJECTIVE - Popular Participation ECOLOGICAL OBJECTIVE P-OVERTY/EQUITY - Consultation NATURAL RESOURCEiS - Pluralism Early attempts were made to weight benefits three major points of view: economic, social and (and costs) of development projects according to the ecological, as shown in Figure 1.1. income level of the beneficiaries, thereby incorporat- Attempts are continuing (as described later in ing social equity concerns directly into economic this volume) to integrate environmental concerns decisions. However, the lack of success with such into conventional economic decisionmaking- approaches resulted in a more pragmatic procedure, mainly by valuing environmental assets and impacts whereby the economic and social objectives were of development efforts and using them in procedures kept separate, but reconciled ihrough the pursuit of ranging from project level cost-benefit analysis to more efficient production coupled with targeted pov- environmentally adjusted national accounts at the erty-alleviating initiatives (e.g., employment genera- macroeconomic level (see the right side of Figure tion, direct subsidies, etc.) to assist low income 1.1). However, to the extent that some functions of groups. the environment cannot be valued in monetary Protection of the enviromnent has now become terms (e.g., biodiversity), other techniques like the third major objective of development. By the multi-criteria analysis need to be used to trade off early 1980s, a large body of evidence had accumu- non-comparable objectives. lated that environmental degradation was a major Finally, the interaction between the social and barrier to development. The concept of sustainable ecological objectives are probably the least well development has therefore evolved to encompass understood (see the base of Figure 1.1). The grow- Part !: Analytical Framework 3 ing importance of topics like public participation in point out that it may be more relevant to examine the decisiomnaking, consultation of affected groups, and maintenance of the set of opportunities, as opposed pluralism, are manifestations of this awareness. to the preservation of the value of the asset base. This is because preferences and technology are not held Three Approaches to Sustainable Development constant through successive generations, so preserv- ing a constant value of the asset base could be redun- It is possible to identify three different concepts of dant. By concentrating on the size of the opportunity sustainable development that reflect the ideas pre- set, the importance of the conservation of biodiver- sented above-i.e., the economic, the ecological sity becomes more evident, both in terms of ecologi- and the socio-cultural (Munasinghe and McNealy cal and economic approaches to sustainability. The 1992). preservation of biodiversity allows the system to The economic approach to sustainability is retain resilience by protecting it from external based on the Hicks-Lindahl concept of the maximum shocks, in the same manner that preservation of the flow of income that could be generated while at least capital stock protects assets for future consunption. maintaining the stock of assets (or capital) which The difference between the two concepts lies in the yield these benefits (Solow 1986, Maler 1990). There consequences of a loss in ecological resilience. Un- is an underlying concept of optimality and economic der the Hicksian income measure, a society that efficiency applied to the use of scarce resources. consumes its fixed capital without replacement is not Problems of interpretation arise in identifying the sustainable. Using an ecological approach, loss of kinds of capital to be maintained (e.g., manufactured, resilience implies a reduction in the self-organization natural, and human capital) and their substitutability, of the system, but not necessarily a loss in produc- as well as in valuing these assets, particularly eco- tivity. This depends to a certain extent on the capacity logical resources. The issues of uncertainty, irre- of human societies to adapt and continue functioning versibility and catastrophic collapse pose in the face of stress and shocks. The linkage between additional difficulties (Pearce and Turner 1990). socio-cultural and ecological sustainability is thus The ecological view of sustainable develop- demonstrated through the organizational similarities ment focuses on the stability of biological and physi- between human societies and ecological systems. cal systems. Of particular importance is the viability In a separate paper, Perrings (1992) points out of subsystems that are critical to the global stability that sustainable development is not necessarily syn- of the overall ecosystem (Perrings 1991). Protection onymous with the maintenance of the status quo. of biological diversity is a key aspect. Furthermore, Biodiversity conservation does not require the pres- "natural" systems may be interpreted to include all ervation of all species, nor the maintenance of current aspects of the biosphere, including man-made envi- environmental conditions. An evolutionary system ronments like cities. The emphasis is on preserving requires that a level of biodiversity be maintained the resilience and dynamic ability of such systems to that will guarantee the resilience of the ecosystems adapt to change, rather than conservation of some on which human consumption and production, and "ideal" static state. therefore human welfare, depend. Sustainable devel- The socio-culiural concept of sustainability opment demands compensation for the opportunities seeks to maintain the stability of social and cultural foregone by future generations, because today's eco- systems, including the reduction of destructive con- nomic activity changes the level or composition of flicts (UNEP et al. 1991). Both intragenerational biodiversity in a way that will affect the flow of vital equity (especially elimination of poverty), and inter- future ecological services, and narrow the options generational equity (involving the rights of future available to unborn generations. This holds true even generations) are important aspects of this approach. if positive rates of economic growth indicate an Preservation of cultural diversity across the globe, and increase in the instrumental (or use) values of op- the better use of knowledge concerning sustainable tions currently available. practices embedded in less dominant cultures, should Reconciling these various concepts and opera- be pursued. Modem society would need to encourage tionalizing them as a means to achieve sustainable andharness pluralism and grass-roots participation into development is a formidable task. The diversity of a more effective decisionmaking framework for so- short-term needs and concerns, as well as long- cially sustainable development. term goals throughout the world, suggests that In comparing the concepts of ecological and there is no universally "right" or "wrong" sustain- economic sustainability, Githinji and Perrings (1992) able development. 4 Part 1: Analytical Framework One practical approach that may be more useful cause difficulties in estimating the sustainable price to policy-makers and the public is the concept of of a resource. The institutional difficulties in the maximization of net benefits of economic and social implementation of inter-generational cornpensa- development, subject to maintaining the services tion schemes would be significant. Despite such from, and stock of natural resources over time. This difficulties, the development of methodologies to implies that renewable resources, especially if they reflect sustainability constraints in shadow pricing is are scarce, should be utilized at rates less than or important, if economic decisions are to consider the equal to the natural rate of regeneration. The effi- intergenerational effects of the depletion of natural ciency with which non-renewable resources are used resource stocks (see also the following section on should be optimized subject to substitutability be- discount rates). tween these resources and t,chnological progress. Waste should be generated at rates less than or equal Focus of the Paper to the assimilative capacity of the environment, and efforts should be made to protect intra- and intergen- This volume reviews how environmental economics erational equity. Finally, the iimplementation of sus- could facilitate the efficient use of natural resources tainable development will require a pluralistic and (both mineral and biological), as well as mamnade consultative social framework that, among other capital and human resources-an objective wvhich is things, facilitates the exchange of information be- a vital step towards sustainable development. Part A tween dominant and hitherto disregarded groups in describes the framework for analysis. Special atten- order to identify less material and pollution inten- tion is paid to the key role of environmental econom- sive paths for human progress. ics in helping value environmental and natural resources more precisely and in intemalizing the Sustainability Constraint costs and benefits of using such resources into the conventional calculus of economic decisionmaking. In conventional econonmic analysis, biases exist against More generally, the identification of sustainable de- the adequate valuation of natural capital, and the costs velopment options requires: of natural capital depletion. Often, there is an inbuilt * Good understanding of the physical, lack of consideration for the rights of future genera- biological and social impacts of human tions in the traditional decisicinmaking process. The activities. original wvork of Hotelling which established the * Better estimates of the economic value of principle of a depletion premnium or user cost for damage to the environment that help to exhaustible resources, was a key step in beginning to improve the design of policies and pro- place a value on current resource use based on future jects and lead to enviromnentally sound benefits foregone (for a good review, see Dasgupta investment decisions. and Heal 1979). More generally, if it is accepted that * Development of policy tools and strength- rents from natural capital depletion (both renewable ening of human resources and institutions and exhaustible) should be shared with future genera- to implement viable strategies and man- tions (as encompassed in the concept of sustainable age natural resources on a sustainable development), and that a cautious approach must be basis. adopted with regard to natural resource stock deple- Part B of the volume contains a number of tion, then shadow pricing ought to reflect a sustain- shorter case studies grouped and presented according ability constraint (Pearce et al. 1991). One approach to the various techniques of environmental valuation is to ensure that sufficient assets remain to ensure a discussed earlier. Wo longer case studies that illustrate non-decreasing flow of future per capita welfare or the combined use of several valuation techniques are consumption (Pezzey 1992). described next. Finally, the conclusions are presented The emphasis on a sustainable supply ensures followed by the bibliography and annexes. that natural capital is not regarded as a free good, and therefore preempts the biases mentioned earlier. If an Note appropriate rule was applied, differences in the rela- tive scarcity of resources would be reflected in the 1. This broad definition is based on the World Cbniuis- sustainable price. However, certain practical consid- sion on Environment and Development, 1987. For a recent erations need to be addressed in determining such a review of altemative definitions of sustainable development. rule.rLations need knowleaddress detrmanig curechan see Pezzey, 1992. rule. Lack of knowledge of the demand curve can 2. Linking Economics and the Environment The Role of Environmental Economics project evaluation/cost-benefit analysis (CBA), sec- toral/subnational studies, multisectoral macroe- Environmental economics plays a key role in identi- conomic analysis, and international economic fying options for efficient natural resource manage- analysis (finance, trade, etc.), at the different hierar- ment that facilitate sustainable development. It is an chic levels. essential bridge between the traditional techniques Unfortunately, the analysis of the environment of decisionmaking and the emerging more envi- cannot be carried out readily using the above socio- ronmentally sensitive approach. Enviromnental economic structuring. As shown in Figure 2.2, one economics helps us incorporate ecological con- convenient breakdown recognizes environmental is- cerns into the conventional framework of human sues that are related to: (1) global and transnational society, as shown in Figures 2.1 and 2.2. concems (for example, climate change, ozone layer Figure 2.1 indicates the hierarchical nature of depletion); (2) natural habitats (for example, forests decisionmaking in modem society. The global and and other ecosystems); (3) land (for example, agri- transnational level consists of sovereign nation cultural); (4) water resources (for example, river states. In the next level are individual countries, each basins, aquifers, watersheds); and (5) urban-indus- having a multisectoral macroeconomic structure. trial areas (for example, large cities, airsheds). In Various economic sectors (such as energy, industry, each case, a holistic environmental analysis would agriculture, transport, etc.) exist within each country. seek to study a physical or ecological system in its Finally, each sector consists of different subsectors, entirety. Complications arise because such natural projects and local schemes. systems tend to cut across the decisionmaking struc- The usual decisionmaking process that corre- ture of human society mentioned earlier. For exam- sponds to this structure, relies on techno-engineering, ple, a forest ecosystem (like the Amazon) could fmancial and economic analyses of projects and poli- affect the global climate, span several countries, and cies. In particular, we note that conventional economic also interact with many different economic sectors analysis has been well developed over the past several within each country. decades, and uses a variety of techniques including Figure 2.1 The Socio-Economic Structure Figure 2.2 The Environmental System SOCIOECONOMC EN\lRONMENTAL STRUCTURE SYSTEM INTER- GLOBAL NATIONAL TRANSNATIONAL _ CO Z _1 _NATURAL < H~~~~~~~ ~~~~~~~~~~~~~~~~~~~ABITATS < NATIONAL2 o i) MACROECON. 0 z oiz LL 2 < Z LAND 0 o < SECTDRAL 0 ZO -..U Z UBSECTORAL E _ . 8 <3 FZ wRNECT URBAN-ND _ 6 Part f: Analytical Framework The causes of environmental degradation arise of environmental economnics has also been added- from human activity (ignoring natural disasters and the latter plays a key bridging role, by helping to map other events of nonhuman origin), and therefore, we the EA results onto the framnework of conventional begin in the upper part of Figure 2.1. The physical economic analysis. Once again, a variety of environ- (including biological and social) effects of socioeco- mental economic techniques facilitate this process of nomic decisions on the environment must then be incorporating environmental issues into traditional traced through to the bottom half of the diagram, and decisionmaking. These methods include economic the techniques of environmental assessment (EA) valuation of environmental impacts (at the local/pro- have been developed to facilitate this difficult analy- ject level), integrated resource management (at the sis. For example, deforestation of a primary moist sector/regional level), environmental Lmacroe- tropical forest may be caused by hydroelectric dams conomic analysis and environmental accounting (at (energy sector policy), roads (transport sector pol- the economywide, multisector level), and icy), slash and burn farming (agriculture sector pol- global/transnational environmental economic analy- icy), mining of mninerals (industrial sector policy), sis (at the international level). The analytical tech- land clearing encouraged by land-tax incentives (fis- niques mentioned above overlap considerably, and cal policy), and so on. Disentangling and prioritizing therefore this conceptual categorization should not these multiple causes (right side) and their impacts be interpreted too rigidly. (left side) will involve a complex environmental Once the foregoing steps are completed, pro- assessment exercise. jects and policies must be redesigned to reduce their The various elements in Figures 2.1 and 2.2 are environmental impacts and to shift the development combined in Figure 2.3, while the new analytical tool process towards a more sustainable path. Clearly, the Figure 2.3 Incorporating Environmental Concerns Into Decisionmaking ENVIRONMENTAL ANALYTICAL SOCIO ECONOMIC SYSTEM TOOLS AND METHODS STRUCTURE GLOBAL INTER- TRANSNATIONAL NATIONAL z 0io 0- I~~ ~~~~~~F cn I LU < NATU RAL Z O < Z HABITATS L Z Z < Z- SECTORAL WATER I-~~~~co~ o < (3 < -REGIONAL WATER0 L zw< Z wLI L)0> Z a > .< P-w_r C- ,URBAN-INDUST. UJ> a. C.< - SUBSECTORAL AND AIR -PROJECT Part !: Analytical Framework 7 formulation and implementation of such policies is efficiency is a necessary but not sufficient condition itself a difficult task. In the deforestation example for sustainable development. described earlier, the decisionmakers who wish to In 1989, the World Bank issued its guidelines protect this single ecosystem are likely to face prob- for environmental irnpact assessment of projects, as lems in coordinating policies in a large number of the natural culmination of many years of increasing disparate and (usually) non-cooperating ministries concern about the environmental dimension of de- and line institutions (such as energy, transport, agri- velopment.' Thus enviromnental analysis has been culture, industry, finance, and forestry). elevated to the same level of importance as the other traditional aspects of project evaluation: eco- Recent Developments nomic, financial, social, and technical analyses. The valuation of environmental impacts takes on Although the consolidation of environmental eco- added urgency in this context, since it would per- nomic theory and its application to empirical issues mit environmental concerns to be incorporated ef- spread throughout the developed world in the 1970s, fectively into the normal decisionmaking process. the incorporation of environmental issues into devel- opment planning is a relatively recent phenomenon. Conventional Project Evaluation A review of the literature in the field reveals the paucity of writings touching upon the environment. The successful completion of a development project Academic development economics barely acknow- usually involves several well-defined steps. The sys- ledged the field until recently (Dasgupta and Maler tematic approach used by the World Bank in a 1990). Much of the literature on the environmental typical project cycle includes identification, prepa- economics of developing countries has emerged ration, appraisal, negotiations and financing, im- since the mid-1980s. This is largely as a result of the plementation and supervision, and evaluation. increasing emphasis being placed on enviromnental issues by major aid-giving institutions. The Project Cycle The Brundtland Report recognized the role of economics in sustainable development-both in as- Project identification involves preliminary selection sessing costs of environmental degradation in devel- (by the borrowing country and the World Bank) of oping countries, and in designing relevant incentives potential projects that appear to be feasible and con- to limit such degradation (World Comiission on form to national and sectoral development goals. In Environment and Development 1987). The recent the preparation phase which may last one year or worldwide concern regarding the environment has more, the borrower studies the engineering-techni- strengthened the emphasis placed on enviromnental cal, envirorunental, institutional, economic and fi- sustainability as an important criterion for sound natu- nancial aspects of a proposed project. Project ral resource management. Anotherfacet of this concern appraisal consists of a comprehensive and systematic is reflected in the increased attention paid to intergen- review of all aspects of the project, culminating in an erational equity and the role of discount rates in eco- appraisal report that discusses comprehensively the nomic calculations. national and sectoral strategies as well as the engi- Decisionmakers in most developing countries neering/technical, environmental, institutional, eco- now accept that poor managementof the environ- nonric, and financial issues. ment has become a significant barrier to develop- The appraisal report is the basis for justifying ment. Since sustainable development includes other the investment, as well as the negotiations at which long standing, high priority objectives like economic the borrower and financier (the World Bank) discuss growth and poverty eradication, scarce resources the measures required to ensure the success of the (both natural and man-made) must be used as effi- project, and the conditions for funding. The resulting ciently as possible. The conventional approach relied agreements are included in loan agreements which on the systematic analyses of economic decisions (in together with the appraisal report are considered and particular, investment and pricing policies) that in- accepted by the Bank's Board of Executive Directors fluenced the production of goods and services. How- and the borrowing governmenL The borrower is re- ever, as discussed earlier, this focus on economnic sponsible for implementing the project according to conditions mutually agreed on with the Bank. Super- 8 Part I: Analytical Framework vision of the implementation process is carried out ciency border prices (see Annex 1 for more details of by the Bank through periodic field visits and progress shadow pricing). However, for the financial analysis reports from the borrower Evaluation is the final of projects, B, C and r may be defined in financial stage of the project cycle, following disbursement of terms. the loan. Project performance audits are carried out If projects are to be compared or ranked, the one by an independent Bank department, and where ap- with the highest (and positive) NPV would be the propriate, involve review of previous project docu- preferred one, that is, if NPVi > NPV[I (where ments and field visits. This analysis yields valuable NPVi = net present value for project i), then project experience that helps imprcve the work at all stages I is preferred to project II, provided also thait the scale of the project cycle. of the alternatives is roughly the same. More accu- rately, the scale and scope of each of the projects Economic Analysis and Cost-Benefit Criteria under review must be altered so that, at the margin, the last increment of investment yields net benefits Cost-benefit analysis (CBA ) is the key element in the that are equal (and greater than zero) for all the appraisal stage of the project cycle. CBA seeks to projects. Complexities may arise in the analysis of assess project costs and benefits using a common interdependent projects. yardstick. Benefits are defined in relation to how a The intemal rate of return (IRR) is also used as project improves human welfare. Costs of scarce re- a project criterion. It may be defined by: sources used up by the project are measured by their opportunity costs-the benefit foregone by not using T these inputs in the best alternative application. I (Bt - Ct)/(1 + IRR)' = In addition to this economic test, as previously = o mentioned, a number of other aspects (including technical, environmental, institutional, and financial Thus, the IRR is the discount rate which reduces criteria) also need to be ccnsidered in project ap- the NPV to zero. The project is acceptable if praisal. Thefinancial and economic analyses of pro- IRR > r, which in most normal cases implies jects are different. The former uses market or NPV > 0 (ignoring cases in which multiple roots financial costs to estimate the financial viability and could occur, which may happen if the annual net profits of the project enterprise or firm. By contrast, benefit stream changes sign several times). Problems the economic analysis seeks to capture economic of interpretation occur if altemative projects have effects on the whole econolmy, using shadow prices widely differing lifetimes, so that the discount rate that reflect opportunity costs. Extemalities are val- plays a critical role. ued wherever possible (as described below). Some Another frequently used criterion is the benefit- of the criteria commonly used in the cost-benefit test cost ratio (BCR): of a project are described next, with the emphasis on economic rather than financial evaluation. [ T 1 F T The most basic criterion for accepting a project BCR = I B,/ (1 + r) ') compares costs and benefits to ensure that the net = J / =I present value (NPV) of benefits is positive: If BCR > 1, then NPV > O and the project is T acceptable. NPV = E (Bi - Ct )/(1 + r)' Each of these criteria has its strengths and weak- ,=O nesses, but NPV is probably the most useful. The NPV test may be used to derive the least-cost rule. In where B, and Ct are the benefits and costs in year t, certain cases, the benefits of two altemative projects r is the discount rate, and T is the time horizon. may be equal (that is, they both serve the same need Both benefits and costs are defined as the differ- or demand). Then the comparison of alternatives is ence between what would occur with and without the simplified. Thus: project being implemented. As described later, in economic testing B, C, andr are defined in economic T terms and appropriately shaiow priced using effi- NPV, - NPV11 = y, [ C11t - Cl,I I/ (1 + r) t *=0 Part I: Analytical Framework 9 since the benefit streams cancel out. Therefore, if national development goal might be embodied in an acceptableobjectivefunctionsuchas aggregatecon- T T sumption. Usually, one important task of the analyst E CIIt/ (1 + r) ' > I C 1t/ (1 + r) t; is to maximize this consumption subject to con- 1=0 ,=O straints-including limits on the availability of re- sources (like capital, labor and environmental this implies that NPVI > NPV11. assets),structural distortions in theeconomy, and so In other words the project which has the lower on. Then, the shadow price of a given scarce eco- present value of costs is preferred. This is called the nomic resource represents the change in value of the least-cost alternative (when benefits are equal). objective function, caused by a marginal change in However, even after selecting the least-cost alterna- the availability of that resource.3 While the general tive, it would still be necessary to ensure that this equilibrium approach is conceptually important, it project would provide a positive NPV. is too cumbersome and data-intensive to use in many cases. In practice, partial equilibrium tech- Shadow Pricing niques may be used, that evaluate the impact of the change in the availability of a given resource on a In the economist's idealized world of perfect compe- few key sectors or areas, rather than throughout the tition, the interaction of atomistic profit-maximizing economy (see also, the following sections on envi- producers and utility-maximizing consumers gives ronmental accounting and environmental impacts rise to a situation that is called Pareto-optimal.2 In of economywide policies.) tis state, prices reflect the true marginal social costs, Two basic types of shadow prices exist. These scarce resources are efficiently allocated and, for a involve whether or not society is indifferent to in- given income distribution, no one person can be made come distribution considerations. To illustrate this better off without making someone else worse off point, consider the simple national goal of maximiz- (Bator 1957). ing the present value of aggregate consumption over However, conditions are likely to be far from a given time horizon. If the consumption of different ideal in the real world. Distortions due to monopoly individuals is added directly regardless of their in- practices, external economies and diseconomies come levels, then the shadow prices derived from (such as environmental impacts which are not inter- such a model are termed efficiency prices because nalized in the private market), interventions in the they reflect the pure efficiency of resource allocation. market process through taxes, import duties and sub- Alternatively, when increasing the consumption of sidies, all result in market (or financial) prices for the lower income groups becomes an important ob- goods and services which may diverge substantially jective, this consideration is given a greater weight from their shadow prices or true economic values. in evaluating aggregate consumption. Then, the re- Furthermore, the reliance on strict efficiency crite- sultant shadow prices are called social prices. ria for determining economic welfare implies the The goal of shadow pricing is, therefore, either passive acceptance of the existing (skewed) in- efficiency- or socially-oriented. In brief, efficiency come distribution-this may be socially and politi- shadow prices try to establish the actual economic cally unacceptable, especially if there are large values of inputs and outputs, while socially oriented income disparities. Such considerations necessi- shadow prices take account of the fact that the in- tate the use of appropriate shadow prices (instead come distribution between different societal groups of market prices) of project inputs and outputs, to or regions may be distorted in terms of overall na- determine the optimal investment decisions and poli- tional objectives. This may call for special adjust- cies, especially in the developing countries where ments; for example, giving greater weight to benefits market distortions are more prevalent than in the and costs accruing to the poor relative to the rich. In industrialized countries. practice, such fornal weighting schemes are seldom Given the complex nature of modem societies, used inproject evaluation-instead, income distribu- it is often conceptually and empirically useful to tional and other social issues are addressed through attempt to capture all the key economic relationships direct targeting of beneficiaries and similar ad hoc in a comprehensive "General Equilibrium" model of approaches. In our analysis, we will place primary the macroeconomy. In such a model, the overall emphasis on efficiency shadow pricing. 10 Part I: Analytical Framework Nonpriced inputs and outputs must be shadow- excessive environmental damage, until rmore accu- priced to reflect their economic opportunity costs rate data and valuation studies can be carried out. In (see Annex 1 for details of border shadow prices and such cases, the initial emphasis is on cost effective- conversion factors). Major categories of such non- ness (i.e., achieving pollution targets at the lowest priced inputs and outputs are common property re- cost), rather than valuing the benefits of control sources and externalities (especially those arising measures.For example, quantity controls on air pol- from environmnental impacts). Access to common lution that limit the aggregate emission level may be property resources is not restricted, and therefore combined with an initial allocation of emissionrights exploitation tends to occur on a first come, first among existing and potential individual polluters served basis, often resulting in overuse (beyond the (which collectively do not exceed the total emission sustainable level). In particular, public goods are a limit). This is analogous to defining property rights class of environmental resources (e.g., a beautiful to an open access resource-in this case, the airshed view) that are freely accessible and indivisible (i.e., over a particular region. Next, it would be logical to enjoyment by one individual does not preclude en- encourage schemes like marketable pollution per- joyment by others). These properties lead to "free- mits (which may be competitively traded among riding"-a situation in which one consumer, either polluters), to achieve an economically efficient re- knowingly or unknowingly, uses the resource at a distribution of "pollution rights" within the overall price less than the efficient cost of making that re- emission limit. However, specific minimum quantity source available, and in the process takes advantage controls may not be an efficient long-term solution, of the greater contributions of others (this is the case if no attempt is made to compare the mnarginal costs when wastewater dischargte taxes are paid by con- of compliance with the real benefits provided (i.e., sumers using a transnational groundwater source in marginal damages avoided--especially as environ- one nation, and the environmental benefits of cleaner mental conditions improve over time). water are shared by consumners in another country In practice, it is often prudent to us>^ ia variable who draw from the samne aquifer, but do not pay such mix of both price and quantity controls to protect the taxes.)4 environment (Pearce and Turner 1990). A mixed Externalities are defined as beneficial or ad- system allows the various policy instruments to be verse effects imposed on others for which the origi- flexibly adjusted depending on marginal cleanup nator of these effects cannot charge or be charged (as costs. In this way, an optimal outcome cam be ap- the case may be).5 If a (damaging)externality can be proached even without full information concerning economically valued or shadlow priced, then a charge control costs (Baumol and Oates 1988). or tax may be levied on the perpetrator, to compen- sate for and limit the damage. This is the so-called Numeraire "Pigouvian" or "price-control" approach to environ- mental regulation. The basic concepts and tech- To derive a consistent set of economic shadow prices niques for economic valuation of environmental for goods and services, a common yardstick or nu- impacts underlying this approach are discussed later meraire to measure value is necessary. The choice of in Chapter 3. the numeraire, like the choice of a currency unit, Unfortunately, many externalities are not only should not influence the economic criteria for deci- difficult to measure in physical terms but even more sionmaking, provided the same consistent frame- difficult to convert into monetary equivalents (that work and assumptions are used in the analysis. is, to measure the "willingness to pay" of the parties The same nominal unit of currency maLy have a affected by the externalities). Quite often therefore, different value depending on the economic circum- the so-called "quantity-control" approach is taken, stances in which it is used. For example, a rupee- by imposing regulations and standards, expressed worth of a certain good purchased in a duty free shop in physical measurements only (e.g., safe mini- is likely to be more than the physical quantity of the mum standards for pollution), that try to eliminate same good obtained for one rupee from a retail store, the perceived external damages. Especially when after import duties and taxes have been levied. There- environmental pollution is severe and obvious, fore, it is possible to distinguish intuitively between setting standards could sente as a useful first step the border-priced rupee, which is used in intema- to raise consciousness about the problem and limit tional markets free of import tariffs, and a domestic- Part I: Analytical Framework 11 priced rupee, which is used in the domestic market of productive activity, is not included in subject to various distortions. A more sophisticated terms of current costs or depreciation of example of the value differences of a currency unit natural wealth. Thus, resource-based in various uses arises in countries where investment goods are underpriced in the market-the for future economic growth is considered inade- lower the value added, the larger is the quate. In these instances, a rupee-worth of savings extent of underpricing of the final product that could be invested to increase the level of future (Dasgupta and Maler 1991). It follows that consumption, may be considered more valuable than countries that export primary products do a rupee devoted to current consumption. so by subsidizing them, often with dispro- Amost appropriate numeraireinmany instances portionately large adverse impacts on the is a unit of uncommitted public income at border poorest members of society (who are less shadow prices (Little and Mirrlees 1974). Essen- able to protect themselves)-the small cul- tially, this unit is the same as freely disposable for- tivator, the forest dweller, the landless eign exchange available to the government, but peasant, and so on. Currently, there are no expressed in terms of units of local currency con- estimates of such hidden costs or "subsi- verted at the official exchange rate. Annex 1 contains dies", but if there were, the GDP of many a discussion of this particular yardstick of value. The countries could well be significantly border-priced numeraire is particularly relevant for lower. In addition, natural resource deple- the foreign exchange-scarce developing countries. It tion raises intergenerational equity issues represents the set of opportunities available to a to the extent that the productive assets country to purchase goods and services on the inter- available to future generations are unfairly national market. diminished (see the discussion on discount rate in the next section). National Income Accounts 3. Abatement or cleanup activities (for exam- and Macroeconomic Performance ple, those that result in expenditures in- curred to restore the environment) often In order to accurately recognize and include environ- serve to inflate national income, but the mental concerns in macro-econornic analyses, stand- offsetting environmental damages are not ard income accounting techniques must be included. In the case of private firms, de- re-examined. Gross domestic product (GDP) is the fensive environmental expenditures (that commonly used growth measure, based on transac- is, measures to reduce or avoid environ- tions in markets. GDP is the basis on which many mental damage) are dedected from fmal aspects of macroeconomnic policy are determined. value added. However, if such cleanup However, its shortcomings include neglect of income costs are undertaken by the government or distributional concerns, non-market activities, and by households, they are added to national even more crucial-environmental degradation. output.The resulting GDP estimate is in- In terms of the envirom-nent, there are three correct because: (a) harmful outputs like weaknesses in the current national accounting frame- pollution are ignored; and (b) beneficial work (Lutz and Munasinghe 1991): inputs related to enviromnental needs are 1. National accounts may not represent wel- implicitly under valued.6 fare accurately, because the balance sheets To overcome these deficiencies in presently do not fully include environmental and used accounting techniques, it is necessary to de- natural resources, and therefore, important velop a System of National Accounts (SNA) that is changes in the status of such resources are capable of yielding an Environmentally-adjusted net neglected. Domestic Product (EDP) and an Environmentally- 2. The true costs of using natural resources in adjusted net Domestic Income (EDI). National level human activity are not recorded in conven- decisionmakers and macroeconomic planners (typi- tional national accounts. The depletion or cally, in a Ministry of Finance or National Plamning) degradation of natural capital (such as the routinely rely on the conventional SNA to formulate stock of water, soil, air, minerals, and wil- economic policies. Thus, a supplementary environ- derness areas), which occurs in the course mentally-adjusted SNA and corresponding perform- 12 Part I: Analytical Framework ance indicators would encourage such policymakers adjusted net Domestic Product 1 (EDP 1) was calcu- to reassess the macroeconomic situation in light of lated, which incorporates the "economic" depletion environmental concerns and to trace the links be- costs of natural resource use, (but does not account tween economywide policies and natural resource for the degradation of environriental quality and management (Muzondo et ia. 1990). corresponding losses of non-marketed environ- The World Bank has worked closely with the mental services that is reflected in EDP2). Next, UN for the past decade, to better incorporate envi- EDP2 was estimated after subtracting the costs of romnental concerns into the present round of revi- degradation of environmental quality from EDPI. sions of the SNA framework. As an interim measure, EDP2 was estimated to range from 90 to 97 percent the "UNSO Framework" (Bartelmus, Stahmer, and of NDR. The final results showed that consumption van Tongeren 1989, UNSO 1990) was created. This exceeded net environmentally adjusted domestic framework is referred to as the System for Environ- production in most years. However, lack of physical mentally-adjusted EconomicAccounts (SEAA). Its ob- data made it extremely difficult to obtain accurate jective is to integrate environmental data sets with estimates. Significant fluctuations in commodity existingnationalaccountsinforrnation,whilemaintain- prices also reflect the difficulties for governments in ing SNA concepts and principles insofar as possible. attempting to maintain sustainable development Environmental costs, benefits, and natural re- policies. source assets, as well as ex]penditures for environ- Additionally, contrary to findings, Papua New mental protection, are presented as satellite Guinea is not necessarily depleting its capital base, accounts in a manner consistent with the account- as the capital gain from erosion of external debt ing framework of the SNA. (caused by inflation reducing the value of ihe debt) In brief, the SEAA seeks to maintain the essen- is about 4% of GNP in real terms. The substitutability tial integrity of the existing SNA; but at the same of capital is therefore an issue to be considered in the time, encourages (through the satellite accounts) the definition of "income". collection and compilation of relevant infornation Aside from these World Bank-supported stud- on natural resources and the environment. An impor- ies, few examples exist of the application of environ- tant element of the SEAA is its ability to utilize mental accounting in developing countries (and even information that may be generated by other measure- less in the developed world). The UN Economic ments such as physical resource accounting at the Commission for Latin America and the Envirorunent regional or sectoral level. The satellite accounts con- (ECLAC) and the United Nations Environment Pro- stitute an important step towards the eventual goal of gramme (UNEP) performed two case studies in Latin computing an EDP and an EDI. America that apply enviromnental accounting meth- The World Bank, together with the UN Statisti- odologies to limited areas within countries. The Ar- cal Office (UNSO), has recently completed case gentina study valued a forest ecosysitem by studies in Mexico (van Tongeren et al. 1991), and estimating the costs of improving productive func- Papua New Guinea (Bartelmus et al. 1992), to deter- tions and of maintaining ecosystem functions. Re- mine how such accounts can be prepared. The Papua sults gained were employed in modelling alternative New Guinea study demonstrates the feasibility of management and exploitation scenarios. The Mexico applying the SEAA framework in a country with study calculated adjustments to the gross product, relatively weak institutional capacities and limited due to a biological corridor, using market valuation data availability, (a scenario that would exist in many of replacement cost in the agricultural and ,forestry resource-rich developing countries). Depreciation of sectors, and constructed physical balance sheets for produced assets was calculated to be between 9 and individual resources (CIDE 1992). The H[icksian 11 percent of Gross Domestic Product (GDP), result- concept of income was utilized to provide a revised ing in a conventional Net Domestic Product (NDP) measure of the region's income. of between 89 and 91 percent cif GDP. Environmental Gilbert performed a case study of Botswana impacts were assessed for the agriculture, forestry, using an environmental accounting franiework mining and energy sectors. within a larger modelling and information system. The authors estimated that these impacts The framework consisted of stock accounts (clescrib- amounted to 2.1 percent of NDPon average, for the ing natural resource stocks in physical units); re- 1986-1990 time period. First, the Environmentally- source user accounts (describing stocks in physical Part 1: Analytical Framework 13 and monetary units); and socio-economic accounts oriented reforms with the complementary goal of (which focused more on demographics, environ- environmental protection. Indeed, a key message of mental policy, and the use of resources). However, the 1992 World Development Report is the need to full implementation of the framework has not been identify and exploit such "win-win" policy reform possible because of severe constraints on informa- opportunities. For example, addressing problems of tion (CTDIE 1992). land tenure as well as access to financial and social An early application of environmental account- services not only yield econoric gains but also are ing in a developing country was performed by essential for promoting environmental stewardship. Repetto et al. (1989) for the World Resources Insti- Similarly, refomis to improve the efficiency of indus- tute. The study collected data on petroleum, timber, trial or energy related activities will reduce both eco- and soil resources. The approach used is based upon nomic waste and environmental pollution (World Bank physical stock and flow accounts of natural re- 1992). sources, and the valuation of these stocks. It has been Many instances of excessive pollution or re- suggested that the valuation method used in forestry source over-exploitation are due to market failures overestimates the true resource depreciation (Peskin and policy distortions exacerbated by unemploy- with Lutz 1990) but has nevertheless proved ex- ment, landlessness, and poverty (Munasinghe et al. tremely useful as an indicator of the magnitudes 1993). Therefore, broad policy reforms, which usu- involved in adjustments to GDP through environ- ally promote efficiency or reduce poverty, also mental accounting (CIDIE 1992). Alater study by the should be generally beneficial for the environment. Tropical Science Center and the World Resources However, some of these reforms may have negative Institute (1991) utilized the same valuation method- environmental effects, depending on pre-existing ology for forestry, but focused on providing detailed (and often localized) constraints or distortions- methods for the technical estimates of deforestation, such as inadequately defined property or resource soil erosion, and coastal fishery over-exploitation in rights. The challenge is to trace the complicatedpaths Costa Rica. An example from Brazil is provided in by which such policy changes ultimately affect in- Ser6a da Motta and Young (1991). Lutz (1993) is a centives for efficient resource use at the firm or recent comprehensive volume on environmental ac- household level. The objective is not necessarily to counting. directly modify the original broader policies (which have other conventional, economic, or property re- Economywide Policies and the Environment lated goals), but rather to design more specific or localized complementary measures. The latter would Economywide policies (both macroeconomic and help mitigate negative effects or enhance the positive sectoral) play a significant role in the rate of deple- impacts of the original policies on the environment tion of natural resources and the level of environ- Such complementary actions would include both mental degradation. Fiscal and monetary policies, market-based approaches (like Pigouvian taxes on structural adjustment programs, and stabilization environmental externalities or allocation of limited measures all have an effect on the natural resource pollution rights coupled with marketable pennits), as base. Unfortunately, interactions between the econ- well as non-market methods (such as command-and- omy and the environment are complex and not well control techniques). This process of articulating a understood. The ideal approach is a "general equilib- range of policies becomes more difficult when rium" analysis that traces both the economnic and economywide reform programs address very broad environmental effects of economywide policy re- macroeconomic distortions. forms. However, such comprehensive methods are seldom feasible in developing countries where data General Economywide Policies and skills are more scarce. "Partial"approaches that help to identify the most important impacts of At the outbreak of the economic crisis of the early economywide policies may be more practical. 1980s, many developing countries that had been No simple generalizations are possible as to the running substantial budget and trade deficits and likely environmental effects of broad policy meas- financing these by increasing external debt were ures. Nevertheless, opportunities have been missed forced to adopt emergency stabilization programs. oftenfor combining poverty reduction- or efficiency- 14 Part I: Analytical Franework These programs may have had unforeseen environ- fish exports), then the degree of ownership would mental consequences. influence how production and resource slocks were One important environmental impact of the cri- managed. Reactions might range from more invest- sis was related to poverty and unemployment. The ment in and maintenance of assets (if environmental stabilization efforts often necessitated currency de- costs were internalized by owner-users) to rapid de- valuations, controls on capital, and interest rate in- pletion (when the users had no stake in the resource creases. When income levels dropped, tax revenues stock). The possibility of the latter result is empha- decreased accordingly. As unemployment increased, sized in another study (Capistrano and Kiker 1990). governments fell back upon expansionary financing They propose that increasing the competitiveness of policies, which led to increases in consumer prices. world exports would also increase the opportunity The effect of such policies on the poorest population cost of keeping timber unharvested. This could lead groups often drove them onto marginal lands, resulting to forest depletion that significantly exceeds natural in soil erosion or desertification. Fuel price increases regenerative capacity. Another recent study (Kahn and and lowered incomes also contributed to deforestation McDonald 1991) used empirical evidence to suggest and reductions in soil fertility, as the poor were forced that a correlative link exists between debt and defores- to use fuelwood and animal dung for heating, lighting, tation. They propose that debt burdens cause myopic and cooking. behavior that often results in overdepletion of forest Aside from the contractionary aspects of short- resources-through deforestation rates that may not be term stabilization measures, many macroeconomic optimal in the long run, but are necessary to meet short policies also have potentially important effects on term needs. resource use and the environment. Unfortunately, no The important influence of the macroeconomic easy generalizations as to the directions of these context for agriculture has already been shown by the effects are possible; they can be either beneficial or classic studies of Johnson (1973) and Schuh (1974). negative, depending on specific conditions. For ex- More recently Krueger, Schiff, and VWldes (1991) ample, real currency devaluations have the effect of have compiled detailed country examples suggesting increasing international comnpetitiveness, and raising that economywide factors may in fact be more imn- production of internationally tradable goods (for ex- portant than sectoral policies in agriculture. These ample, forestry and agricultural products). If the studies point out that when a broad assessment agricultural response occurs through crop substitu- perspective is adopted, direct output price inter- tion, environmental impact would depend on ventions by government turn out to have less effect whether the crop being prornoted tended to be envi- on agricultural incentives than indirect, economy- romnentally benign (such as tea, cocoa, rubber) or wide factors, such as foreign exchange rates and environmentally damaging i(such as tobacco, sugar- industrial protection policies. cane, and corn). Environmental impacts would also The impact of economywide policies are impor- depend on whether increased production led to farm- tant for the environment, although studies are only ing on new land (which could result in increased starting to quantify the importance of their role. For deforestation) or to more efficient use of existing example, Hyde et al. (1991) cite studies in Brazil and farmland. Another possibility is that overvaluationof in the Philippines that demonstrate how economic the exchange rate (and the resulting negative terms policy spillovers constitute an important source of of trade, decreased competitiveness of products and deforestation. Agricultural subsidies in Brazil prob- lower farmgate prices), may well push small cultiva- ably contribute to half of forest destruction in Ama- tors onto more environmenatally fragile marginal zonia (Mahar 1988, Binswanger 1989, Lewandroski lands, in an attempt to absorb the effects of the price and McClain 1990). Ageneral equilibrium simulation changes. for the Philippines suggests that foreign exchange In a recent review of the links between growth, rate changes, although motivated by general balance trade policy and the environnent, (Lopez 1991), the of payments concems, have major implications for author argues that the output from a natural resource the demand for wood products, therefore influencing such as a forest or fishery (where production depends logging rates (Boyd et al. 1990). The case of fielwood critically on the stock) also will be affected by other may be as interesting, since fuelwood shortages have factors such as property rights. Thus, if trade policy been identified as the major forestry problem in increased the value of output (for example, timber or many developing countries. It is possible that na- Part 1: Analytical Framework 15 tional price policies on fuel and investment policies tion, the reduction of government subsidies, or en- on altemative energy sources may be as important as ergy pricing reforms. Each such policy would have any sector specific program for addressing the fuel- different implications for various environmental ar- wood problem. eas of concern. For examnple, devaluation would Arecent study of WorldBankcountry andsector significantly affect timber and crop prices and there- work has sought to present an overview of how fore influence deforestation trends (indicated by the economywide policy reforms to promote develop- arrows in the top half of Figure 2.4). This approach ment have numerous unanticipated environmental is explored more systematically in Table 2.1. effects (Cruz and Munasinghe 1992). Ongoing work The first column of Table 2.1 contains a few on this project includes a series of country case among the many issues addressed by macroe- studies. conomic and sectoral decisionmakers. The The linkages among economywide policies and economywide policies in the second colunm of the the environment may be examined from two major table are usually designed to address these issues, perspectives - economic and environmental. First, with the corresponding direct economic objectives or consider the viewpoint of macroeconomic and sec- development impacts in the third column. Examples toral planners or decision-makers, especially those of second-order, and often unanticipated environ- involved in national economic planning in the Min- mental impacts are listed in the fourth and last col- istry of Finance or Planning, or in key sectoral line umn. Industrial protection reform provides one Ministries. Inicreasingly, they would wish to know example of how both positive and negative environ- the likely impact of a specific policy on a range of mental effects (last row, last column of Table 2.1), environmental issues. Iypical policies might in- might be associated with such an economywide pol- clude local currency devaluation, price liberaliza- icy. Therefore, to properly evaluate broad policy Figure 2.4 Interaction Among Economic Policies and Environmental Issues National Economic Plan: Environmental Action Plan: Economywide Policies Environmental Issues Devaluation * ___ - * \*0 Deforestation * 0 Low Energy Prices * 0 * 0 Urban Air Pollution * 0 16 Part I: Analytical Framework Table 2.1 Some Typical Examples of Direct and Indirect (Environmental) Impacts of Policies Policy Direct Indirect Policy Issue Reform Objectives/Effects (Environmental E^ffects) 1. Trade Flexible Promote industrial Export promotion may deficits. exchange rates. competitiveness, exports; lead to more deforestation reduce imports. for export, but it could also lead to substitution of tree crops for annual crops. In addition, industrial job creation may reduce pressures on land resources. 2. Food security and Agricultural Increase crop yields and May reduce spontaneous unemployment. intensification in settled acreage; absorb more rural rnigration to ecologically ]ands and resettlement labor. fragil areas. However, programs for new areas. there is potential for over- use of fertilizers and chemicals. 3. Industrialization Iteduce tariffs and special Promote competition and More openness may lead protection, associated with investment incentives. industrial efficiency. industry to adopt more inefficient production. energy-efficient or less pollution-prone technologies. However, it may also lead to in- rnigration of hazardous industries. reforms, their direct and indirect effects need to be nomic policies, including foreign exchange and ag- identified, and any trade-offs between their conven- ricultural price policy reforms. tional development contribution and their environ- This second approach is illustrated in Table 2.2, mental effects need to be assessed. where the first column lists some typical environ- Retuming to Figure 2.4, let us consider the mental issues that are incorporated in environmental second (or converse) viewpoint of environmental action plans. These issues are catogorized according policyrnakers-e.g., those in ihe Environment Min- to varying economic characteristics in the second istry who are charged with preparing the national column. Column 3 contains a sampling of broad enviromnental action plan (NEAP). They may be policy reforms that could affect the particular envi- quite concerned to find out which of a bewilderingly ronmental problem. For example, water availability wide range of economic poLicies (current or pro- and quality will depend primarily on domestic price posed) would have substantial imnpacts on a specific policies while energy generation will often be very environmental issue that has high priority in the sensitive to international fuel prices. Thus, rational- NEAP (e.g., urban air pollution or deforestation). If izing domestic water charges and subsidies will be air pollution is the major concem, then the NEAP crucial in promoting more sustainable water use, should include a detailed review of policies such as whereas trade and exchange rate policies (that influ- energy pricing (indicated by the arrows in the bottom ence fuel prices) will be relevant for improving en- part of Figure 2.4). On the other hand, addressing ergy efficiency. deforestation issues would require particular atten- Ideally, both the above approaches (eccinomnic tion to be paid to a somewhat different set of eco- and environmental) should be developed consis- Part l: Analytical Framework 17 Table 2.2 Some Typical Environmental Issues, Their Characteristics, and Sensitivity to Policy Reforms Resource/Environmental Management Issue Sectoral Economic Characteristics Relevant Policy Reforms 1. Agricultural expansion and * many small, competitive * reduction of taxes and subsidies deforestation decisionmakers are involved * exchange rate and trade reforms * outputs, inputs are mostly * poverty and income internationally traded distribution policies * government implements * property rights reforms substantial production subsidies and trade intervention 2. Water depletion and degradation * supply side is dominated by * intersectoral pricing govermnent or monopolies; bulk * reduction of subsidies and of resource use goes to large introduction of charges for commercial enterprises and resource degradation irrigation systems * resource is not internationally traded but sectoral use and productivity for main user groups substantially differ * prices are highly regulated 3. Energy use and air pollution * as with water, supply side * exchange rate reforms dominated by government and * reduction of cross-subsidies monopolies * privatization programs for * inputs (coal, oil) are generally generating and distribution traded; output broadly linked to all activities production activities * sectoral investment and pricing highly centralized tently within the same comprehensive framework of tegrating the domestic economy into the world econ- national decisionmaking. However, in practice they omy (Fischer and Thomas 1990). tend to be quite distinct and disjointed, often because Adjustment lending started in the early 1980s to of historical constraints. Overcoming such problems address the problem of maintaining growth in coun- may require far-reaching reforms in decisionmaking tries severely affected by the worldwide economic processes and institutions, even before new policies crisis. The crisis was triggered by the second round are formulated and implemented. of energy price hikes, the collapse of export markets, and increased international interest rates (World StructuralAdjustment Bank 1988). These external conditions, combined with lack of competitiveness of local industries, lag- Structural adjustment programs address a well-de- ging employment generation, and persistent budget fined set of reform areas. These will generally in- deficits, led to unsustainable current account deficits. clude establishing an appropriate macroeconomic framework for growth, introducing a set of support- Thus, trade-oriented reforms, including tanff ing sectoral policies and investment efforts, and in- reductions and devaluation, became key components 18 Part I: Analytical Framiework of adjustment lending. Because the deficit is identi- known about the intervening or mediating role of cafy equal to the use of foreign savings, adjustment factors such as property rights and population implications for managing external debt have also re- growth. ceived much attention. In turn, this led to concern for In the early to mnid-1980s, environmental aspects stimulating domestic savings., increasing the role of were relatively neglected in adjustment lending op- taxation in resource mobilization, decreasing the extent erations. There were few, if any, explicit loan com- of government expenditures, and making investments ponents or conditionalities that were more productive. "environmental." Thus, the 1989 study emphasized The benefits and costs of such programs are the potential complementarities between the adjust- highly country-specific. In a recent review of the ment reform components and environmental objec- Latin American experience, Birdsall and Wheeler tives. For example, trade liberalization was (1991) conclude that there is no evidence to show considered potentially environmentally beneficial. that open developing country economies are more Increased agricultural export prices would imply prone to pollution. They argue that the inflow of higher land values and therefore more investment in foreign technology and capital would tend to bring land. Of course, assumptions on land tenure security in better pollution standards. At the same time, it is were needed to complete the argument. the pollution-intensive heavy industries sector that The above study has been updated recently has generally benefitted froni protective industrial (Warford et al. 1992). The current review concludes and trade policies. Nevertheless, some concern per- that environmental aspects have been increasing in sists that encouraging foreign investment and priva- importance in adjustment programs. Recent pro- tization might lead to the growth of "pollution grams, from 1988-92, includemore thanjust anomi- havens," given the weakness of environmental regu- nal treatment of environment. For example. about lations in most developing countries. Trade liberali- 60% of the 58 countries involved in the review had zation also could encourage the growth of adjustment programs whose loan components or energy-intensive and/or highly polluting industry. conditionalities were environmentally motivated. However, pollution caused by industrialization During FY79-FY87, the comparable figure would only could be offset by afforestaticin (although this does have been 37%. In addition, environmental policy re- not necessarily compensate residents of polluted forms have been the subject of some sectoral adjust- areas), and limited by appropriate taxation policies ment programs. that encouraged the use of pollution abatement In recent years, non-govermnental organiza- technologies. tions and academic institutions have contributed to Next, we examine to what extent those who the debate on the environmental effects of stmictural design structural adjustment lending programs have adjustment and stabilization programs-among the progressed in recognizing their environmental impli- most recent efforts being the studies undertaken by cations and in incorporating policy measures that the World Wildlife Fund (WWF) and the World could reduce negative impacts or enhance environ- Resources Institute (WRI). mentally beneficial effects. An initial assessment of The WWF-sponsored report (Reed, ed., 1992) the environmental relevance of structural adjustment included three case studies, in C6te d'Ivoire, Thai- reforns was undertaken by Seblastian and Alicbusan land, and Mexico. In CBte d'Ivoire, no direct linkage (1989). This study looked at how the adjustment was established between macroeconomic policies lending reforms may have affected environmental and the environment. Underlying institutional fac- factors. The reforms included (I) relative price tors and market and policy failure at the sectoral level changes in agricultural outputs, inputs, energy, ex- had a greater impact on natural resource use than port taxes; (2) trade and industry policy reformn; (3) macroeconomic instruments. Again, in the case of changes in public expenditure programs; and (4) Mexico, no clear linkages between macroeconomic institutional reforms by sector. At the time of com- policy and the environment could be established pletion of this paper, most of the links between (although, at the micro level, inappropriate pricing adjustment-related policies and the environment fo- for water and energy has resulted in adverse environ- cused on inter-sectoral effects, e.g., between in- mental consequences). In Thailand, evidence sug- creased agricultural export prices and incentives for gested that, while structural adjustment programs agricultural land managemrent. Also, little was may have reduced resource depletion and environ- Part I: Analytical Framework 19 mental degradation per unit of output, overall envi- include industrial promotion and investment incen- ronmental quality declined because of the increase tives, tax incidence, import liberalization and trade, of output in the aggregate. Because the underlying and energy pricing. These are undertaken to improve market failures with regard to the natural resource resource allocation, but they also have important sector were not adequately addressed, and social and implications for macroeconomic stability and environmental costs were not fully internalized, the growth (Fischer and Thomas 1990). The first is pri- trade-off (between economic growth and loss of en- marily a microeconomic goal while the second is an vironmental quality) may not have been economi- adjustment goal. cally efficient or enviromnentally sustainable. For example, taxation and government expen- The WRI (Cruz and Repetto 1992) performed a ditures are the prime microeconornic mechanisms study on the environmental effects of structural ad- for resource allocation. However, they also comprise justrnent and stabilization programs in the Philip- the basic elements of fiscal policy. In turn, fiscal pines. They found that structural adjustment policies policy has a critical macroeconornic impact because had not adequately examined adverse envirounental it directly determines the fiscal deficit and therefore impacts of macroeconomic policies. This resulted in affects the current account deficit and, after a lag, increased emissions, concentrated pollution and con- investment levels.7 Tax reform issues are particularly gestion, increased pressure on open-access re- relevant from an environmental management per- sources, and encouraged overexploitation of spective because they can have a wide range of non-renewable resources. They suggest several poli- potential impacts on resource use. The choice of the cies (for example, the taxation of resource rents, tax or tariff base, for example, can lead to substantial energy taxes, and the elimination of industrial incen- changes in the level of pollution-related activities. tives), that could limit environmental damage while Such environmental implications, however, are not at the same time promoting economic objectives considered in conventional assessments that focus on such as fiscal balance, poverty alleviation, and eco- fiscal effectiveness. Other sectoral reforms, such as nomnic efficiency. those dealing with energy pricing and industrial ex- Rationalization of public expenditure is an inte- ports, also affect macroeconomic stability and have gral part of many Structural Adjustment Loans thus played a regular role in adjustment programs (SALs), and usually emerges from recommendations (Yagci et al. 1985, World Bank 1989). on spending priorities made in Public Invest- Beyond their role in contributing to a stable ment/Expenditure Reviews (PI/ERs). macroeconomic environment, sectoral policies also Such reviews provide a consistent framework have economy-wide relevance in terms of promoting within which the borrowing country and the Bank growth from the supply side. These include sectoral could discuss how spending decisions might be investment and pricing policies as well as sectoral best tailored to meet the priority objectives of regulation and institutional development. As noted macroeconomic management. Environmental con- by Fischer and Thomas (1990), the traditional ap- siderations would be incorporated better into deci- proach to development was through investment in sions affecting intra sectoral resource allocation, agriculture, industry, infrastructure, and human re- as well as specific projects investment, by more sources. However, the contribution of these sec- effective use of the PI/ER process. toral investments also depended largely on the macroeconomic policy context and on the presence Sectoral Policies of enabling institutions. Practical examples of the effects sectoral poli- While adjustment is inherently a macroeconomic cies have on the environment have been described in effort, involving macroeconomic policy reform, it several studies. Some countries subsidize urban con- also requires specific sectoral reforms (Cruz and sumers by placing price ceilings on food. In such Munasinghe 1992). The key macroeconomic vari- cases, the environmental consequences will be the ables in adjustment programs are the investment- same as for currency overvaluation, as both result in savings gap, the fiscal deficit, the trade deficit, the lowered incentives to increase production of interna- exchange rate, and the rate of inflation. The microe- tionally tradable crops. conomic or sectoral reforms relevant to adjustment 20 Part I: Analytical Framework 'iWo classic studies by Mahar (1989) and Bin- Notes swanger (1989) highlighted how distorted public policies have exacerbated environmental problems 1. For details, see World Bank, Operationail Directive in the Brazilian Amazon region over several decades. 4.01. Mahar showed that poor and landless peasants have 2. Further details concerning the sumnmanized material Ml this section may be found in Dasgupta. P., S. M[arglin and contributed to deforestation, basically responding tO A.K. Sen, 1972. Guidelines for Project Evaltuati on, UNEDO, incentives such as highways that opened up jungle New York Little, I.M.D. and Mirrlees, J. 1974. ProjectAp- areas, govemment land grants, access to public in- praisal and Planning for Developing Countries, New York, frastructure, and cash subsidlies. Some of the blame Basic Books; Munasinghe, M., 1990b. Energy Analysis and is attributable directly to govemment settlement pro- Policy, Butterworths Press, London; Ray A., 1984. Cost BenefitAnalysis, Baltimore: Johns Hopkins University Press; jects. Large scale livestock as well as iron ore smelt- and Squire, L. and H. Van der Tak. 1975. Economic Analysis ing (in the Carajas area) have contributed to land of Projects, Baltimore, Johns Hopkins University Press. degradation. Binswanger, focussing on the agricul- 3. In the more technical context of a mathematical tural sector, argued that tax and land distribution programming macroeconomic model, the optimal values of policies, and the provision of credit to farmers, not the dual variables (that correspond to the binding resoure availability constraints in the primal problem) have dimen- only encouraged deforestation but also worsened sions of prce, and could be interpreted as shadow prices income distribution (by favoring larger landowners). (LAenberger 1973, Sassone 1977). More recently, Lutz and Young (1992) have 4.AseminalreferenceonthefreeriderprincipleisP.A. stressed that rational agricultuural policies that elimi- Samuelson 1954. "The Pure Theory of Public Expenditures", nated subsidies for farm inputs (like fertilizers and Review of Economics and Statistics, 36, 370-89. pesticides) could provide both economic and envi- 5. For a review of the basic theory of extemailities, see ponnesticides)icouldSpr ider both al. (1993)argue that R. Coase, "The Problem of Social Cost", Journal qfLaw and ronmental gains. Schneider et al. (1993) argue that Economics, October 1960; and D.M.G. Newbery, "Exteunali- unsustainable destruction of ithe Amazon is the result ties: The Theory of Envirounental Policy" in G.A. Hughes of farmers, ranchers and loggers responding quite and G.M. Heal (eds), 1980,Public Policy and the TaxSystem. predictably and rationally to distorted incentives 6. Dasgupta and Maler (1991) suggest that in order to arising from poor govemmeint policies and political avoid double-counting, expenditures that enhance the re- instability. source base, (such as planting fcrests), should not be included instability, ~~~~~~~~~~~in national incomne comnputations, since they are already re- Some recent examples of World Bank sectoral flected in the value of changes in the resource stockcs. work that attempted to address issues of environ- 7. Stanley Fischer and Vmod Thomas, "Policies for mental concem through policy reform are detailed in Economic Development" in American Journal of Agricul- Annex 5 (Cruz and Munasinlghe 1992). tural Economics,Vol. 72, No. 3, August 1990. 3. Framework For Environmental-Economic Decisionmaking As discussed earlier, a comprehensive framework for Basic Concepts of Economic Value cost-benefit analysis of projects in developing coun- tries was in place by the 1970s, including sophisti- Conceptually, the total economic value (TEV) of a cated techniques of shadow pricing to compensate resource consists of its (i) use value (UV) and (ii) for economic distortions, as well as social weights to non-use value (NUV). Use values may be broken adjust for income distributional concerns. With the down further into the direct use value (DUV), the emphasis on market-oriented reforms and economic indirect use value (IUV) and the option value (OV) liberalization in the 1980s, some of the enthusiasm (potential use value). One needs to be careful not to has waned for applying such a complex system rig- double count both the value of indirect supporting orously (Little and Mirrlees, 1990). functions and the value of the resulting direct use.1 Nevertheless, the rise of environmental con- One major category of non-use value is existence cerns, and growing evidence of associated externali- value (EV).2 We may write: ties and market failures have revived interest in economic evaluation of projects and policies. In TEV = UV + NUV view of the increasing risk of environmental degra- or TEV = [DUV + IUV + OV] + [NUV] dation, even modest attempts to assess environ- mental impacts and risks would be well justified. Figure 3.1 shows this disaggregation of TEV in Ideally, environmental cost and benefits should be schematic form. Below each valuation concept, a quantified economically and integrated into tradi- short description of its meaning and a few typical tional cost benefit analysis. In cases where valuation examples of the environmental resources underlying of environmental impacts is problematic, other tech- the perceived value, are provided: niques such as multicriteria analysis may be used to * direct use value is determined by the con- incorporate the environmental concerns into invest- tribution an environmental asset makes to ment decisions and policies. current production or consumption; * indirect use value includes the benefits Valuation of Environmental Costs derived basically from functional services and Benefits that the enviromnent provides to support current production and consumption (e.g., Incorporating environmental concerns into eco- ecological functions like natural filtration nomic decisions involves two basic steps. First, the of polluted water or recycling of nutri- "with"and "without"project scenarios are compared ents); to identify the physical impacts (broadly defined to * option value is basically the premium that include also ecological and social effects) of a given consumers are willing to pay for an unu- economic activity. Engineers, biologists, social sci- tilized asset, simply to avoid the risk of not entists, and other experts are required to determine having it available in the future (a more such impacts. An important issue, outside the scope detailed version is provided later in this of this paper, is that such physical impacts are com- section); and plex and often poorly understood. * existence value arises from the satisfac- Economically valuing such physical impacts tion of merely knowing that the asset constitutes the second step in the environmental-eco- exists, although the valuer has no inten- nomic analysis, as described below. tion of using it. Option values and non-use values are shaded in the figure, to caution the analyst concerning some of the ambiguities associated with defining these con- cepts-as shown in the examples, they can spring Figure 3.1 Categories of Economic Values Attributed to Environmental Assets (with Examples from a Thopical Forest) Total Economic Value . I,,I Use Values Non-Use Values I I . ltII. l I Direct Use Indirect Use Option Existe'ce Other Non-Use Values Values Values Vle Valu's Output that Functional Future direct Value from can be benefits and indirect knowledge of consumed use values continued directly existence o Food o Ecological o Biodiversity o Habitats o Biomass functions o Conserved o Endangered o Recreation o Flood control habitats species o Health o Storm protection e ADecreasing "tangibility" of value to individuals Source: Adapted h-o Peare (1992). Part I: Analytical Framework 23 from similar or identical resources. Economic theory on the level of services attributed to nonuse values. clearly defines TWV (see next section), but there The question of strictly partitioning use and non-use tends to be considerable overlap and ambiguity in the values therefore becomes more complex. breakdown categories, especially with regard to non- For the practitioner, what is important is not use values. Therefore, these categories are useful necessarily the precise conceptual basis of economic mainly as an indicative guide. Further, the distinc- value, but rather the various empirical techniques tions often become irrelevant in practical estimation that permit us to estimate a monetary value for many since the objective is to measure TUV rather than its enviromnental assets and impacts. However, there is components (see Randall 1991 for details). uncertainty in the results derived from some of these Non-use values tend to be linked to more altru- techniques even in developed market econornies, and istic motives (Schechter and Freeman, 1992). The therefore their use in developing countries should be differing formns of altruism include intergenerational tempered by caution and sound judgement. altruism, or the bequest motive; interpersonal altru- ism or the gift motive; stewardship (which has rather Overview of Practical Valuation Techniques more ethical, as opposed to utilitarian origins); and q-Altruism, which states that the resource has an A variety of valuation techniques may be used to intrinsic right to exist. This final definition is clearly quantify the above concepts of value. The basic well outside the conventional economic theory, in- concept of economic valuation underlying all these corporating the notion that the welfare function techniques is the willingness to pay (WTP) of should be derived from something more than purely individuals for an environmental service or re- humanutility. Quiggin(I991)has put forward amore source.3 Willingness to pay itself is based on the controversial point-that the more radical forms of area under the demand curve, as illustrated below altruism could not be reconciled within a conven- by a simplified case. tional BCA framework. In Figure 3.2, the curve D(So) indicates the There has been considerable debate recently demand for an environmental resource (e.g., the concerning the applicability of non-use values. number of visits made per month, to a freshwater Smith (1992) states that further research is needed in source like a river). Xo is the original demand level order to better describe the public good services at the price p (e.g., the cost of making a trip, underlying non-use values. Consider the indirect including the value of time spent for travelling). utility function The total WTP or value of the services provided by the environmental resource is measured by the area U = V(Y,P,Q) OHEA which consists of two main components: (1) the area OBEA or (pXo), which represents the where Y is the individual's income, P is a vector of total cost; and (2) the area BEH, which is called prices for private goods available on markets (or the consumer surplus or net benefit (i.e., the net through well-defined conditions of access to assets value over and above actual expenses). The point such as recreation sites), and Q is a vector of services H represents the choke price, at which demand (both use and non-use) derived from a given environ- falls to zero. mental resource. The distinction between use and Next, we extend this example to examine what non-use value is made by introducing constraints that could happen if the quality of the environmental linkQ to one or more of the goods rationed by prices. service is improved (e.g., by clean-up of the fresh- Smith points out that for Q to include services based water source). The normal response would be an on non-use values, it must be capable of changing increase in demand, represented in Figure 3.2, by without an individual undertaking any actions. The a shift in the demand curve from D(So) to D(SI). factors altering Q must satisfy the requirements of a The new level of demand is Xi (assuming the same public good-they must be non-exclusive and non- price p), yielding the corresponding total WTP appropriable. If attention is paid to the source of equal to area OGFI, and the new net benefit meas- services underlying nonuse values, it may emerge ured by area BF. Thus, the quality improvement that the components of Q based on use values depend will result in an incremental increase in the value 24 Part 1: Analytical Framework Figure 3.2 Benefit Increase Due to the Improvement in Quality of an Environmental Asset Value p 1. E D(S1) Demand oftheenviromnentalresource-givenbytheshaded mental asset could be defined in terms of thle differ- areaEFIH.4 ence between the values of two expenditure (or cost) The theoretically correct demand function to functions. The latter are the ninimum amounts re- use in estimating value, is the compensated or Hick- quired to achieve a given level of utility-for a sian one which indicates how demand varies with household (or output-for a firm) before and after price while keeping the user's utility level constant varying the quality of, price of, and/or access to, the Equivalently, the change in value of an environ- Part I: Analytical Framework 25 environmental resource in question, while keeping then the market valuation will be a partial meas- all other aspects constant (see also Annex 3). ure, and shadow price corrections may need to be Problems of measurement may arise because made. The foregoing comments apply to change of the commonly estimated demand function is the production. Marshallian one-which indicates how demand var- Often, the result of the impact cannot be directly ies with the price of the environmental good, while related to a market activity. In some of these cases, keeping the user's income level constant. In practice, the WTP could be estimated at conventional market it has been shown that the Marshallian and Hicksian value by using a closely related proxy. Care should estimates of WTP are in good agreement for a variety be exercised on the following points: a) the rele- of conditions, and in a few cases the Hicksian func- vant attributes affected by the environmental im- tion may be derived, once the Marshallian demand pact might, in the case of the proxy measure, be function has been estimated (Willig 1976, Kolstad mixed with other attributes, thereby affecting the and Braden 1991). value of proxy; and b) if the proxy attributes are What people are willing to accept (WTA) in the identical to the ones lost by the impact, then the way of compensation for environmental damage, is value given by the proxy is only a lower bound for another measure of economic value that is related to the true WTP. This approach applies to the follow- WTP. WTA and WTP could diverge as discussed ing techniques: loss of earnings, defensive expen- below (Cropper and Oates 1992). In practice either diture, replacement cost, and shadow project. or both measures are used for valuation, and there- In certain cases the WTP can be estimated fore references to WTP in the case studies may be through derivation of a demand function for the broadly interpreted to include WTA also, unless environmental asset through analysis of actual be- otherwise stated. havior. Examples of this approach (also called surro- As shown in Table 3.1 valuation methods can be gate market techniques) include travel cost, wage categorized, on the one hand, according to which differential, and property valuation. The WTP can type of market they rely on, and on the other hand, also be elicited through a controlled experiment or by considering how they make use of actual or po- direct interviews, using the artificial market tech- tential behavior of economnic agents.5 nique or contingent valuation. Next, we provide some general comments about While empirical evidence indicates that ques- valuation, before separately discussing each of the tions about willingness to accept compensation yield techniques in Table 3.1 in greater detail. higher answers than questions about willingness to Under specific conditions, such as when the pay to retain the same amenity, there is no explana- environmental impact leads to a marginal change in tion for this discrepancy on the basis of economic the supply of a good or service that is bought on a theory. It has been suggested that WTA questions competitive market, the WTP can be estimated di- need more time to be properly understood and as- rectly in terms of changes valued at prevailing mar- sininlated, and that the gap between WTA and WTP ket prices. If the market is not fully competitive, narrows with successive iterations. However, it has Table 3.1 Taxonomy of Relevant Valuation Techniques Conventional market Implicit market Constructed market Based on actual behavior Effect on Production Travel Cost Artificial Market Effect on Health Wage Differences Defensive Cost Property Values Preventive Cost Proxy Marketed Goods Based on potential Replacement Cost Contingent Valuation behavior Shadow Project Other 26 Part I: Analytical Framework also been suggested that people are willing to spend with other measures of the demand for water actual income or wealth less readily than "opportu- (Markandya 1991). nity" income or wealth-money they do not yet have Whittington et al. (1992) carried out a revealing but may obtain (Knetsch and Sinden, 1984). It may study in Nigeria. The study was designed a) to meas- also be the case that individuals are more cautious ure the demand for different types of drinking water when weighing the net benefits of changing assets programs in Nigeria in a way that would allow pre- than when no change is made. Generally, WTP is dictions of household participation patterns with dif- considered to be more consistent and credible a ferent designs and pricing schemes; b) to evaluate the measure than WTA. However, when significant dis- feasibility of using CVM in rural villages where crepancies exist between the two measures, then the respondents have limited education and monetary higher values of WTA may be more appropriate when resources; and c) to examine the consequences of valuing losses in environmental benefits. This in turn giving respondents time to evaluate contingent valu- would lead to higher values for conservation projects ation questions before taking their bids. The authors (Markandya 1991). found that respondents who were allowed time to WTPis often used even when the situation being think about WTP questions made significantly lower valued involves a loss of amenity-(the question is bids the second time around. This was tnre for both then presented in the form of what one is willing to public water systems and private connections. The pay to prevent a loss (Markandya, 1991). Hanemamn authors concluded that the differences were not a (1991:) showed that the rneasure to be employed result of strategic behavior because of the open- depends on the problem under consideration. He ended interview process used, and multivariate demonstrated that the difference between WTA and analyses of WTP bids. They noted that these find- WTP varies directly with the income elasticity of ings are probably not easily transferable to devel- demand for S (the desired good or service) and oped country studies, given the increased inversely with the elasticity of substitution between consumer awareness in the industrialized world. S and private goods. Therefore, WTP and WTA are We conclude this section by noting that environ- equal if the income elasticity of demand for S is zero, mental impacts also may be classified as indicated in or if S is a perfect substitute for a private good. Table 3.2. Conversely, as the elasticity of substitution between In the table, on-site and off-site are relative S and private goods approaches zero, then the differ- terms, since they depend on how widely the project ence between WTA and WTP could become large. site or "boundary" is defined. For example, inunda- In developing countries, the ability to pay be- tion of land near a new dam is definitely on-site, comes a concern. Especially in low income areas, whereas distant effects like extent of siltation far money values placed on environmental goods and downstream of the dam are off-site. In-between im- services are traditionally low. One way of addressing pacts could be classified in either category, depend- this concern within cost benefit calculations is to use ing on the project boundary assumed in the; analysis. income weights. However, the use of income weights There is a related time dimension. For example, is somewhat constrained by the data problems in off-site effects also tend to occur in the more distant getting income or consumption levels of the con- future, than on-site impacts (although this is not cerned groups. In such cases, the decision criterion always the case). Clearly, there is a systematic in- has to evaluate the income distribution impacts sepa- crease in the difficulty of valuation as the analyst rately, i.e., based on the social concept of sustainable progresses from on-site, short-run, market priced development discussed in Chapter 1 (see also the impacts to off-site, long-run, non-market priced ef- Madagascar case study in Chapter 5, and Markandya fects. In the developing world, given data, time and 1991). An alternative approach would be to measure resource constraints, it would be advisable to avoid benefits of environmental imaprovements in terms of attempting to value off-site, non-market priced envi- willingness to accept income for loss of amenity, as ronmental impacts. opposed to willingness to pay for an environmental improvement. For example, in water supply projects, WTP can be zero, whereas WTA is more consistent Part I: Analytical Framework 27 Table 3.2 Ease of Valuation of Environmental Impacts Availability of Direct Market Prices to Value ImIpact Impact Location Inpact Time Yes No On-Site Short-Run Quite Easy to Value Sometimes Possible to Value Long-Run Highly Possible to Value Possible to Value in Special Cases ............................................................................................................................... ....................................................................... ............................... Off-Site Short-Run Often Possible to Value Difficult to Value Long-Run Sometimes Possible to Value Rarely Possible to Value Source: Adapted from Hamilton, Dixon and Miller (1989). Actual Behavior in Conventional Markets This technique seeks to avoid ethical controver- sies associated with valuing a single specific life, The techniques summarized below rely mainly on attempting instead to place a value on changes in the directly observable effects or actions valued at mar- statistical probability of ill-health or death (some- ket prices (see north-west corner of Table 3.1). what like the actuarial values used by life insurance Effect on production. An investment decision firms). Moreover, governments and public health often has environmental impacts, which in turn affect authorities routinely set priorities and allocate health the quantity, quality or production costs of a range of expenditures which affect human well-being. This outputs-that may be valued readily in economic in turn provides a baseline for determining implicit terms.There are examples of this in two case studies values placed by society on various health risks, both that follow in chapter 4. In the study on soil conser- for prevention and cure of illness. vation in Lesotho, the increased production from Defensive or preventive costs. Often, costs conserved land is estimnated. In the valuation of 1 may have been voluntarily incurred by communities hectare of Peruvian rainforest the values of different or individuals to mitigate or undo the damage caused production schemes are compared. Furher examples by an adverse environmental impact. For example, include impacts on tropical wetlands (Barbier et al. if the drinking water is polluted, extra filtration 1991) and the effects of sedimentation on coral diver- and/or purifying chemicals may need to be used. sity and ultimately on fish production (Hodgson and Then, such additional defensive or preventive expen- Dixon 1988). ditures (ex-post) could be taken as a minimum esti- Effect on health. This approach is based on mate of the benefits of mitigation. The assumption health effects caused by pollution and environmental is that the benefits of avoided environmental degra- degradation. One practical measure that is relevant dation at least exceed the costs of avoidance. The is the value of human output lost due to ill health or advantage of the technique is that defensive or pre- (in some cases) premature death. The loss of poten- ventive outlays (already made) are easier to deter- tial net earnings (called the human capital technique) mine than the value of the original environmental is one proxy for foregone output, to which the costs damage. One weakness is that the defensive actions of health care or prevention may be added (as a form are sometimes decided upon quite arbitrarily with of replacement/defensive expenditure)-see the little reliance on market forces, so that the costs bear Zimbabwe water supply and health study in chapter little relation to the potential environmental benefit. 4. The above measure assumes that earnings reflect Recently, Harrington et al. (1989) evaluated the the value of marginal product and that medical treat- economic damages of a waterbome disease outbreak, ment costs are well defined. The method also en- emphasizing that the valuation of averting behavior counters difficulties when the cause-effect link requires the establishment of a relationship between between environmental quality and ill-health is un- observable defensive expenditures, and non-observ- clear, or the sickness is chronic (i.e., of long dura- able willingness to pay. tion). 28 Part 1: Analytical Framework Potential Behavior in Conventional Markets Implicit Markets This section summarizes techniques in which Often, relevant market data is not available in di- potential or future actions could be valued in conven- rectly usable form, to value environmental resources. tional markets to provide a measure of environmental In many such cases, analysis of indirect market data degradation, provided there is a high degree of cer- (e.g., using statistical and econometric, multiple re- tainty that such actions will be undertaken. gression methods) permits the valuation to be carried Replacement cost and shadow project. If an out implicitly. A variety of such surrogate market- environmental resource that has been impaired is based methods-including travel cost, the "hedonic" likely to be replaced in the future by another asset methods (property value and wage differential), and that provides equivalent services, then the costs of proxy goods-as well as their applicability under replacement may be used as a proxy for the environ- different circumstances, are described below. mental damage. This is an ex-ante measure similar Travel cost. This method seeks to determine the in spirit to the (ex-post) defensive costs approach. It demand for a recreational site (e.g., number of visits may be argued that the benefits from the environ- per year to a lake or game park), as a function of mental resource should be at least as valuable as the variables like consumer income, price, and various replacement expenses.This approach is especially socio-economic characteristics. The price is usually relevant if there is a sustainability constraint that the sum of observedcost elements like (a) entry price requires certain asset stocks to be maintained intact to the site; (b) costs of travelling to the site; and (c) (see the discussion in Chapter 1). foregone earnings or opportunity cost of time spent. The replacement cost approach has been applied The consumer surplus associated with the estimated to deterioration of groundwater resources in the Phil- demand curve provides a measure of the value of the ippines, by determining the cost of developing alter- recreational site in question. More sophisticated ver- native water sources (Munasinghe 1990c). The same sions include comparisons (using regression analy- technique of estimating potential ex-post mitigation sis) across sites, where environmental quality is also expenditures represented by the increased costs of included as a variable that affects demand (for a health care, is used in the Manicaland case study that detailed survey, see Mendelsohn 1987 and Bockstael follows. et al. 1991). The use of shadow projecst reflects an institu- Until a few years ago, most applications of this tonal judgement on the value of replaced environ- technique were to be found in the market economies, mental assets, and has been discussed in the context but quite recently, several examples have emerged of project-level sustainability. Such a shadow pro- involving developing world applications. Three of ject is usually designed specifically to offset the the case studies sunmmarized in chapter 4 use the environmental damage caused by another project. travel cost method. In one case study, the travel cost The cost of the shadow project is a measure of the for domestic trips to a forest reserve in Costa Rica is value of environmental assets that are thereby re- used. In another, a study on the value of elephants in stored. The original project and shadow project to- Kenya, the travel cost of tourists from Europe and gether form a sustainable package which helps to North America is used as one estimate of consumer maintain undiminished, some vital stock of environ- surplus. Finally, Lumpinee Park in Thailand was mental resources. For example, if the original project used also as an example of travel cost analysis in a was a dam which inundated some forest land, then developing country setting. In chapter 5, the Mada- the shadow project might involve the replanting of gascar case study uses a sophisticated international an equivalent area of forest, elsewhere. Often the travel cost model. equivalency criterion is hard to satisfy exacly-in the Property value. In areas where relatively com- above example, the two tracts of forest may have the petitive markets exist for land, it is possible to de- same volume of biomass, but could differ widely in compose real estate prices into components terms of biodiversity. attributable to different characteristics like house size, lot size, proximnity to schools, shops, parks, etc. (Cropper and Oates 1992). To value an environ- Part I: Analytical Framework 29 Figure 3.3 The Trade-Off Curve in Multi-Objective Decisionmaking family of equipreference curves cn~~~~~~~~~~~~~~~~N N~~~~~~~~~~~ (I) N- a0 .- N. O .> trade-off curve N 0 Objective Z2 (Biodiversity Loss) mental variable like air or water quality, the method the benefits of the projects, to determine the bias of seeks to determine that component of the property the hedonic method, if any. The IDB also used the value attributable to the relevant environmental vari- hedonic price method to estimate the benefits of a able. Thus, the marginal WTP for improved local water quality improvement project in Ecuador. How- environmental quality (air, land or water) is reflected ever, the model used seems to undervalue the bene- in the increased price of housing in cleaner neighbor- fits of the sewerage connections, and the valuation hoods. of the waste treatment plant resulted in very uncer- This method has limited applicability in devel- tain benefits (the estumation model did not obtain a oping countnes because it requires a well functioning significant coefficient on the dummy variable used housing market, as well as sophisticated information to measure water quality, although the sign is right). and tools of statistical analysis. Jimenez (1983) used Additionally, the benefits of a cleaner river on a this technique to explain changes in housing prices regional or national basis were not accounted for. The in a Manila slum area, upgraded partly due to water author concludes that the contingent valuation and sanitation service improvements. Markandya method (see below) may have been more appropriate (1991) examines Inter-American DevelopmentBank to use in this case. (IDB) projects that utilized the hedonic price method Wage differences. As in the case of property to estimate the benefits of installing a sewerage pro- values, the wage differential method attempts to re- gram in El Salvador, and a drainage program in late changes in an economic price variable (i.e., the Brazil. He points out that, since hedonic pricing can wage rate), to environmental conditions. The under- lead to an overestimation of benefits, it would be lying assumption is that there is some component of helpful if more than one method were used to value the wage that is determined by the environmental 30 Part 1: Analyticai Framework Figure 3.4 Exclusionary Screening MORE Tradeoff curve CMAX Feasible set of options U) 0 BMAX LESS BIODIVERSITY LOSS MORE pollution or hazard associ,ated with the job or work vide an example involving marketed and non-mar- site. The technique is relevant when competitive keted fish substitutes. labor markes exist, where wages (that reflect the marginal product of labor) iequilibrate the supply and ConstructedMarkets demand for labor. (See earlier discussion on shadow pricing and wage rates). One consideration is that In cases, where market information cannot be used this method relies on private valuation of health risks, directly or indirectly, market-like behavior needs to rather than social ones. In this context, the level of be deduced through construction or simulation. The information concerning occupational hazards must methods summarized below depend on direct ques- be high in order for private individuals to make tions, surveys or marketing experiments. meaningful tradeoffs between health risks and remu- Contingent valuation. When relevant market neration. Finally, the effects of all factors other than behavior is not observable, the contingent valuation environment (for example, skill level, job responsibil- method (CVM) puts direct questions to individuals ity, etc.) that might influence wages must be eliminat to determine how much they might be willing-to-pay to isolate the impacts of environment (WTP) for an environmental resource, or h1ow much Proxy marketed goods. This method is useful compensation they would be willing-to-accept when an environmental good or service has no read- (WTA) if they were deprived of the same resource. ily determined market value, but a close substitute The contingent valuation method is more effective exists which does have a competitively determined when the respondents are familiar with the environ- price. In such a case, the mLarket price of the substi- mental good or service (e.g., water quality) and have tute may be used as a proxy for the value of the adequate information on which to base their prefer- environmental resource. Barbier et al, (1991) pro- ences. CVM is likely to be far less reliable when the Part I: Analytical Framework 31 object of the valuation exercise is a more abstract other types of bias. (For details, see The Energy aspect like existence value. Journal, 1988). Munasinghe (1990a) provides sev- Generally, declared WTA tends to be signifi- eral early examples of the application of CVM to candly greater than the corresponding WTP. This value the quality of electricity services in developing may be partly attributable to "strategic bias" where countries. respondents feel that they would be better off inflat- A review by Pearce and Markandya (1989) ing the amounts they would receive rather than the compared valuation estimates obtainedfrom market- sums to be paid out, if the hypothetical questions based techniques and CVM, using results from seven posed were somehow to become a reality in the studies carried out in industrial nations. They found future. "Status quo bias" could also serve to increase that the corresponding estimates overlapped within WTA or WTP estimates-in this case, individuals an accuracy range of plus or minus 60 percent. The who have become accustomed to a certain level of conclusion is that CVM, cautiously and rigorously enviromnental quality simnply may be signalling their applied, could provide rough estimates of value that instinctive resistance to change. Providing more in- would be helpful in economic decisionmaking, espe- formation about options as well as opportunities to cially when other valuation methods were unavailable. reconsider choices, could help to reduce this form of Recently, Strand and Taraldset (1991) tested for bias. In the case of poorer individuals, WTP may be multiple bias in a contingent valuation study in Nor- limited by the ability-to-pay, whereas WTA is not. way. They followed a two-step procedure, compar- Generally, the CVM questionnaires have to be care- ing valuation of air pollution alone; and secondly fully designed, implemented and interpreted to over- deriving an implicit valuation of the air pollution come the above mentioned difficulties, as well as issue from the ranking of a broader set of six envi- Figure 3.5 Trade-Offs Among the Three Basic Objectives of Sustainable Development ECONOMIC x ECs OLOGICAL 0 x SOCIAL 32 Part 1: Analytical Flrarnework ronmental issues (problems due to oil spills at sea; ment of a group of relevant "experts"is sought to freshwater pollution; noise pollution; air pollution; estimate the desired values. garbage and waste accumulation; and pressure on recreational areas). They found that expressed valu- Multi-Objective Decisionmaking ations of air pollution reductions were three to four (or Multi-Criteria Analysis) times as high as the true values, and believe that this is due to the upward bias of what they term "mental Valuation techniques seek to estimate environmental accounting"-whereby individuals tend to focus too impacts in economic terms. Often, projects or poli- much of their total valuation of a much larger set of cies and their impacts are embedded in a system of environmental goods on the particular object to be broader (national) objectives, e.g., defined by social valued. If biases had not existed, the authors estimate and ecological concerns of the type discussed in the that from total WTP for a]ll six issues the amount introduction. To the extent that the impacts of pro- allocated for air pollution would be only 50-60% of jects and policies on these broader objectives may the figure stated when they were asked about air be valued economically, all such effects can be pollution alone. incorporated into the conventional decisionmak- Despite the limitations of the method, in certain ing framework of cost-benefit analysis. However, circumstances, contingent valuation may be the only some social and biophysical impacts cannot be easily available technique for estimating benefits. It can quantified in monetary terms, and multi-objective and has been applied to common property re- decisionmaking offers a complementary approach, sources, amenity resources with scenic, ecological which may facilitate the optimal choice arnong in- or other characteristics, and to other situations vestment options or policies available. where market information is not available Multi-objective decisionmaking, or multi-crite- (Mitchell and Carson 1989). ria analysis (MCA), differs from CBAin three major Examples of contingenit valuation surveys in areas (van Pelt, 1991). While CBA focuses on effi- developing countries are given in the case studies. ciency (although incorporation of income distribu- The use of the contingent valuation method for esti- tion objectives may be attempted), MCA (toes not mating the value of viewirng elephants in Kenya impose limnits on the fonns of criteria, allowing for shows that it is possible to achieve an understanding consideration of social and other forms of equity. of the order of magnitude of the benefits with only Secondly, while CBA requires that effects be meas- limited efforts. The study on willingness to pay for ured in quantitative terms, to allow for the applica- water services in southern Haiti tests the methodol- tionof prices, multi-objective decisionmaking canbe ogy for different biases and reliability of estimated broken down into three groups: one that requires values. The willingness to pay for improved sanita- quantitative data, a second that uses only qualitative tion in Kumnasi, Ghana, is another example of CVM data, and a third that handles both simultaneously. successfully employed in a developing country. Finally, multi-objective decisionmaking does not re- Artificial market. Such markets could be con- quire the use of prices, although they might be used structed for experimental purposes, to determine to arrive at a score. CBA uses prices which may consumer willingness to pay for a good or service. sometimes be adjusted according to equity weight- For example, a home water purification kit might be ing. Multi-objective analysis uses weighting involv- marketed at various price levels or access to a game ing relative priorities of different groups as opposed reserve might be offered on the basis of different ad- to pricing. Van Pelt concludes that the choice be- mission fees, thereby facilitating the estimation of the tween the two methodologies is a matter of the trade- value placed by individuals on water purity or on the off between methodological and empirical use of a recreational facility, respectively. considerations. If efficiency is the only criterion and Other. A variety of essentially opinion- and prices are available to value efficiency attributes, survey-based techniques for determining values of CBA is preferable. However, in many cases, a pau- environmental impacts have been used. They are city of data, and the need to incorporate social and related to the CVM approach but are more ad-hoc. biophysical impacts makes the use of multi-objective One of these is the Delphi melhod, where the judge- analysis a more practicable and realistic option. Part I: Analytical Framework 33 Multi-objective decisiornaking calls for desir- one objective against the other-typical equiprefer- able objectives to be specified. These often exhibit a ence curves are illustrated in Figure 3.3. The pre- hierarchical structure. The highest level represents ferred alternative is that which results in the greatest the broad overall objectives (for example, improving utility-which occurs (for continuous decision vari- the quality of life), often vaguely stated and, hence, ables as shown here) at the point of tangency D of not very operational. Some of these, however, can be the best equi-preference curve, with the trade-off broken down into more operational lower level ob- curve. jectives (e.g., increase income) so that the extent to Since the equi-preference curves are usually not which the latter are met may be practically assessed. known other practical techniques have been devel- Sometimes only proxies are available (e.g., if the oped to narrow down the set of feasible choices on objective is to enhance recreation opportunities, the the trade-off curve. One approach uses limits on attribute number of recreation days can be used). objectives or "exclusionary screening". For exam- Although value judgements may be required in ple, in Figure 3.4, the decisionmaker may face an choosing the proper attribute (especially if proxies upper bound on costs CMAX (i.e., a budgetary con- are involved) in contrast to the single-criterion meth- straint). Similarly, ecological experts might set a odologies used in economic cost-benefit analysis, maximum value of biodiversity loss BMAX (e.g., a measurement does not have to be in monetary terms. level beyond which the ecosystem suffers cata- More explicit recognition is given to the fact that a strophic coUapse). These two constraints help to variety of concerns may be associated with planning define a more restricted portion of the trade-off curve decisions. (darker line), thereby narrowing and simplifying the An intuitive understanding of the fundamentals choices available. of multi-objective decisionmaking can be provided Of the several multi-criteria methods that have by a two-dimensional graphical exposition such as in been developed, the practical method which is most Figure 3.3. Assume that a scheme has two non-com- suitable to determine the best alternative available mensurable and conflicting objectives, Zi and Z2. depends upon the nature of the decision situation.6 For example, Zi could be the additional project cost For instance, interactive involvement of the decision- required to protect biodiversity, and Z2, some index maker has proved useful in the case of problems char- indicating the loss of biodiversity. Assume further acterized by a large number of decision variables and that alternative projects or solutions to the problem complex causal interrelationships. Some objectives can (A, B and C) have been identified. Clearly, point B be dealt with through direct optimization, while others is superior (or dominates) to A in terms of both Zi require the satisfaction of a certain standard (e.g., level and Z2 because B exhibits lower costs as well as of biological oxygen demand or BOD not below 5 biodiversity loss relative to A. Thus, alternative A milogranis per liter). may be discarded. However, we cannot make such a More generally, five main forms of multi-criteria simple choice between solutions B and C since the evaluation methods have been identified (Pearce and former is better than the latter with respect to objec- Tumer, OECD, 1990). These are the aggregation, lexi- tive Zi but worse with respect to Z2. In general, more cographic, graphical, consensus-maximizing, and con- points (or solutions) such as B and C may be identi- cordance approaches. fied to define the set of all non-dominated feasible Aggregation techniques add up scores over a solution points that form an optimal trade-off curve range of criteria to identify the optimal solution. It or curve of best options. includes methods such as the Environmental Valu- For an unconstrained problem, further ranking ation System, which estimates the net enviromnental of altematives cannot be conducted without the in- impact of a project in temis of a single composite troduction of valuejudgements. Specific information index score. Four key environmental factors are as- has to be elicited from the decisionmaker to deter- sessed: ecology, pollution, aesthetics and human in- mine the most preferred solution. In its most com- terest. Criteria are weighted by a panel of experts plete form such infonnation may be summarized by using the Delphi method. Value transfornation func- a family of equi-preference curves that indicate the tions are applied to all project impacts, which could way in which the decisionmaker or society trades off result in overly subjective indexing, and the applica- 34 Part !: Analytical Framework tion of scores to certain indicators of pollution (e.g., representation of decision problems in the sense that emission levels) can be questioned. Bearing these several objectives can be accounted for. However, a qualifications in mind, this method could be utilized key question concerns whose preferences are to be for example, in determining whether to follow a more considered. The model only aids a single decision- conventional (fossil-fuel based) energy policy in the maker (or a homogeneous group). Various interested medium term, or whether to pursue an alternative groups will often assign different priorities, to the strategy which placed greater emphasis on energy respective objectives, and normally it may not be conservation and renewable energy sources. possible to determine a single best solution via the The lexicographic approach ranks criteria from multi-objective model. Also, the mathematical the most to the least important, thereby not allowing framework imposes constraints upon the ability to for trade-offs among criteria. The method is best effectively represent the planning problem. Nilon lin- suited to cases where priorities, are clear. An example ear, stochastic, and dynamic formulations can assist is a toxic waste facility, where acceptable/non-ac- in better defining the problem but impose costs in ceptable risk levels can be relatively easily defined. terms of complexity in formulation and solving the Graphical approaches employ a method that model (Cocklin 1989). plots altematives on graphs with reference to bench- Nevertheless, in constructing the model tie ana- mark positions (which would be the 'best' scores of lyst communicates information about the nature of any particular alternative). Weighting is used to as- the problem. He specifies why factors are important sign relative importance. The method does not con- and how they interact. Liebman (1976) observes that sider possible collinearity between criteria, a "modelling is thinking made public" and considers problem that often emerges in multi-criteria analysis. the transfer of knowledge as perhaps the most impor- It also requires that data be in ratio or interval form. tant contribution of modelling. With respect to the A version of this method has been utilized in the second point of criticism (i.e., diverse preferences), U.S.A., to value the externalilies that are to be re- Liebman suggests that there is value to be gained in flected in bidding programs for permits to release constructing models from differing perspectives and sulfur dioxide. comparing the results. Consensus maximization aggregates individual In addition to facilitating specific tradeoff deci- preferences to form group consensus. The Borda- sions at the project level, the multi-objective ap- Kendall method ranks criteria in terms of importance proach could also help in selecting strategic paths. in order to reveal an overall index of imnportance. The broadest representation of multi-criteria trade- This method could be helpful, for example, in at- offs arising from sustainable development efforts are tempting to estimate the siting of a hydropower pro- shown in Figure 3.5, where the axes are labelled ject, by taking into account preferences across according to the three basic objectives (economic, different interest groups (farmers, government plan- social and ecological) discussed in the Introduction ners, conservation groups, etc.) (see Figure 1.1). Conceptually, three broad indicators Concordance methods assume non-compensa- may be envisaged-one for each major objective, tory values between criteria, thereby negating the with movements along any given axis indicating an possibility of establishing trade-offs. Comparisons improvement in that indicator. In practice, each such are carried out on a pairwise basis, and a mathemati- indicator would itself consist of many disaggregate cal concordance or discordance index is used to yield sub-indices. Some limited mapping from one axis to indexes of dominant and non-dominant solutions. A the other may be possible (e.g., to the extent that score foreach alternative is establishedby examining environmental assets could be valued, then ecologi- the overall comparative performance of an altema- cal concems may be incorporated within the eco- tive against the whole range of criteria. A possible nomic indicator). application of this could be for a wetlands drainage The non-inferior (or trade-off) surface XYZ project that would result in irreversible loss of biodi- shown in Figure 3.5 is the best that could be achieved versity. given the current availability of resources and lakowl- The major accomplishment of multi-objective edge. In the real world, societies would operate at decision models is that they allow for more accurate points such as A that are well within this surface- Part /: Analytical Framework 35 due to constraints and inefficiencies. Therefore, the rium between consumption and investment), equals first step would be to identify actions that would both the marginal rates of time preference and return move us closer towards the non-inferior surface (e.g., on capital. In practice, government policy distor- along the path AB). Such so-called "WIN-WIN" tions and market failures lead to divergences be- activities that simultaneously improve all three indi- tween the rates of time preference and return on cators would have the highest priority. One clear capital. Furthermore, the social rate of time prefer- example mnight be to reduce the pollution and degra- ence may be less than the individual time preference dation of freshwater resources in developing coun- rate, because long-lasting societies are likely to have tries. First, maldng higher quality water available a bigger stake in the more distant future than rela- would increase economic output (e.g., by improving tively short-lived individuals. health productivity). Second, it would certainly im- The rate of capital productivity is often very prove the ecological status. Finally, the social in- high in developing countries, because of the scarcity dicator would rise also, since water quality of capital. In the poorer countries, the rate of time improvements tend to benefit mainly the poor. Once preference also is elevated in many cases, because of the shift from A to B is achieved, movements along the urgency of satisfying immediate food needs the surface (such as B C) would be the next step-this rather than ensuring long-term food security (Pearce process is more difficult to determine, since it would and Turner 1990). involve trade-offs among the different objectives. A study was conducted in rural India which compared estimated rates of time preference of heads The Discount Rate of rural households. Results were derived from hy- pothetical questions, and games involving real and Discounting is the process by which costs and significant awards (Pender and Walker 1992), and benefits that occur in different time periods may be indicated the existence of a marked positive time compared. Economists typically use a forward-look- preference. Wealth was inversely and significantly ing approach in which past (or sunk) costs and bene- related to the discount rate in many of the experimen- fits are ignored, while a discount rate is applied to tal games and hypothetical questions. Aproportional future costs and benefits to yield their present values. 10% rise in net wealth was accompanied by a 3-7% T he use of standard criteria for cost-benefit analysis fall in the discount rate. (CBA), such as the net present value (NPV) and In both the games and the questions, the dis- internal rate of return (IRR), have been explained in count rate declined with increases in the time frame more detail earlier. (See the section on cost-benefit and the magnitude of reward (Pender 1992). This criteria). suggests that the neoclassical discounted utility The issue of choosing an appropriate discount model, which imnplies that the rate of time preference rate has been discussed in the context of general CBA is independent of time frames and the amounts of for many years.7 The long term perspective required commodities, does not always hold true. The lack of for sustainable development suggests that the dis- properly developed capital mnarkets in many devel- count rate might play a critical role in intertemporal oping economies often causes investment decisions decisions concerning the use of enviromnental re- to be linked to consumption decisions, and to depend sources. on the nature of preferences. For policy purposes, the Tvo concepts help shape the discount (or inter- importance of understanding intertemporal prefer- est) rate in a market economy. First, there is the rate ences is evident. of time preference which determiines how individu- Higher discount rates may discriminate against als compare present-day with future consumption. future generations. This is because projects with Second, there is the rate of return on investment (or social costs occurring in the long term and net social opportunity cost of capital), which determines how benefits occurring in the near term, will be favored an investment (made by foregoing today's consump- by higher discount rates. Projects with benefits ac- tion) would yield a stream of future consumption cruing in the long run will be less likely to be under- (net of replacement). In an ideally functioning mar- taken under high discount rates. It is therefore a ket, the interest rate (determines the point of equilib- logical conclusion that future generations will suffer 36 Part !: Analytical Framework from market discount rates detennined by high rates nario in which growth and consumption are falling of current generation time preference and/or produc- (e.g., catastrophic global warming in 100 years), then tivity of capital. g could become negative and consequently CTP may Based on the foregoing, some environmentalists be small or even negative. In this case, with CTP as have argued that discount rates should be lowered to the discount rate, future costs and benefits would facilitate environmentally sound projects meeting loom much larger in present value terms than if the the CBA criteria. However, this would lead to more conventional opportunity cost of capital (say 8%) investment projects of all types, thereby possibly was used, thereby giving a larger weight to long- threatening the environmentally fragile resource term, intergenerational concerns. The key point is bases. Krautkraemer (1988) argued that lowering that it may be mnisleading to choose discount rates discount rates can in fact worsen environmental deg- without assuming some consistent future scenario.9 radation-by reducing the cost of capital and thereby Thus an optimistic future would be associated with lowering the cost of production such that more is higher discount rates than a gloomy one--which is consumed in the near term relative to the case where consistent since the risk of future catastrophies should discount rates were higher (for a more recent expo- encourage greater concern for the future. sition c2f this point see Norgaard 1991). Pearce (1991) andvan Pelt (1991) make thecase Many environmentalists believe that a zero dis- that the discount rate traditionally used for project count rate should be employed to protect future gen- appraisal at international institutions may be too erations. However, employing a zero discount rate is high. Pearce points out that if rates above 1-2% are inequitable, since it would imply a policy of total used, then global warming is not taken into consid- current sacrifice, which runs counter to the proposed eration as a serious concern, and future generations aim of eliminating discrinination between time pe- would be left to bear the cost. The only means of riods-especially when the present contained wide- achieving such a low rate would be to cisregard spread poverty (Pearce 1991). utility discounting as unethical, to abandon opportu- There is some basis, in traditional discount rate nity cost discounting, and to place specific restric- analysis, to argue in favour of using declining (and tions on the nature of the income-utility function. He even negative) discount rates for evaluating costs and concludes that use of the opportunity cost rate alone benefits over very long (or multi-generational) time does not appear justifiable on grounds of intergen- periods, when welfare and returns on investment erational equity, and therefore that an appropriate may be falling. Consider the consumer rate of time range of estimates would be 2-5%. Van Pelt argues preference which has components: CTP = a + ,g that, in many parts of the world, high discount rates Here, a represents the preference of an individual cannot be justified on the basis of CBA theory. Ap- for comsumption today rather than in the future-it plication of CBA principles would suggest that in may be based on the myopic notion of "pure" pref- many African countries as well as India, the rate of erence, as well as the risk perception that future discount should be as low as 5%. He suggests that consumptionmaynever be realized. , is the elasticity the rate of discount should be determined on a coun- of marginal welfare and g is the growth rate of try-specific basis, and be regularly updated, as is the consumption. The second term (0g) reflects the fact case with other shadow prices. that the declining marginal welfare of consumption Norgaard makes the case that manipulating dis- combined with increases in expected future con- count rates to reflect sustainability concerns results in sumption will combine to make future consumption an inefficient use of capital. Instead he suggests direct less valuable than present-day consumption, i.e., income transfers to compensate for environmerntal deg- since we are likely to be richer in the future, today's radation. He utilizes a general equilibrium model to consumption is more highly valued. demonstrate that income transfers to future genera- The general consensus is that a is close to zero, tions, through the efficient allocation of resources, usually 0-3%, and j may be in the range of 1 to 2. result in new levels of savings and investment, a shift Thus if g is large (i.e., high expected economic in the types of investments, and a different interest (or growth rates), then CIP could be quite large too.8 discount) rate. The rate of interest may increase or On the other hand, if we consider a long-range sce- Part 1: Analytical Framework 37 decrease, but this is irrelevant, since it merely serves work of policy andprojectdecisions; (c) In the case as an equilibrating price. of projects leading to irreversible damage, CBA should Since the discount rate may be an inappropriate be adapted to the extent possible, to include a measure- tool to facilitate such intergenerational transfers, a ment of the foregone benefits of preservation in the better altemnative might be to impose a sustainability computation of costs; and (d) where valuation of envi- constraint, whereby current well-being is maximized ronmental impacts is difficult, and large irreversible without reducing the welfare of future generations damage might occur, restrictions might be set to limit below that of the current generation. The aim would the enviromnental degradation within acceptable bio- be to ensure that the overall stock of capital (broadly logical or physical norms. defined) is preserved or enhanced for future genera- tions (see also the discussion of sustainability con- Risk and Uncertainty straints in Chapter 1). In practice, this would entail monitoring and measurement of capital stocks (man- Risk and uncertainty are an inherent part of economic made, human and natural) and an overarching invest- decisions. Risk represents the likelihood of occur- ment policy that sought to ensure that compensating rence of an undesirable event like an oil spill. In the investments offset depreciation of existing assets case of uncertainty, the future outcome is basically (Pearce 1991). Apart from the previously detailed unknown. Therefore, the risk of an event may be attempts to include depreciation of natural resource estimated by its probability of occurrence, whereas stock in national income accounting, little has been no such quantification is possible for uncertainty accomplished in this area. Where such a sophisti- since the future is undefined. The risk probability cated approach may be impractical, simple rules that and severity of damage could be used to determine limit specific environmental impacts (e.g., ground- an expected value of potential costs, which then water pollution standards) may be a useful first step would be used in the CBA. to protect the rights of future generations. However, the use of a single number (or ex- In the case of projects leading to irreversible pected value of risk) does not indicate the degree of damage (such as destruction of natural habitats, and variability or the range of values that might be ex- so on), the benefits of preservation may be incorpo- pected. Additionally, it does not allow for individual rated into standard cost-benefit methodology using perceptions of risk. If the future cannot be perceived the Krutilla-Fisher approach (Markandya and Pearce clearly, then the speed of advance should be tailored 1988). Benefits of preservation will grow over time to the distance over which the clarity of vision is as the supply of scarce environmental resources de- acceptable. Global warming is an illustrative exam- creases, demand (fueled by population growth) in- ple. In the past, the greenhouse effect of C02 emis- creases, and existence value possibly increases. The sions was not known or recognized as a risk. At the Krutilla-Fisher approach incorporates these increas- present time, there is still considerable uncertainty ing benefits of preservation by including preserva- about the future impacts of global warning, but tion benefits foregone within project costs. The given the large magnitude of potential consequences, benefits are shown to increase through time by the caution is warranted. As more understanding of the use of a rate of annual growth. While this approach phenomenon is gained, the uncertainty may be trans- has the same effect on the overall CBA as lowering formed into estimates of future risk probability. discount rates, it avoids the problem of distorted The traditional and simple way of incorporating resource allocations caused by arbitrarily manipu- risk and uncertainty considerations in project level lating discount rates. CBA has been through sensitivity analysis.Using In summary, the following conclusions may optimistic and pessimistic values for different vari- be reached, within the context of environmental ables can indicate which variables will have the most cost-benefit analysis: (a) The normal range for pronounced effects on benefits and costs. Although opportunity costs of capital (e.g., 6-12 percent) sensitivity analysis need not reflect the probability of may be used as the discount rate; (b) Efforts should occurrence of the upper or lower values, it is useful be made to ensure that compensating investments for determining which variables are most important offset capital stock degradation within a frame- to the success or failure of a project (Dixon et al. 38 Part I: Analytical Framnework 1988). More sophisticated approaches to analyze risk cases. The issue is the practical redundancy of OV and uncertainty are available. (See, for example, (which may be captured in other measures), rather Braden and Kolstad 1991).In particular, deterninis- than its conceptual validity. This view is reflected by tic a point estimate of value could be quite mislead- Freeman (1993):"I think it is time to expunge option ing, whereas a range of values helps identify more value from the list of possible benefits associated robust options. Various criteria such as mini-max and with environmental protection". minimum-regret may be used (Friedman 1986). Environmental policy formulation is compli- The issue of uncertainty plays an important role cated by the presence of numerous forms of uncer- in environmental valuation and policy formulation. tainty. As an illustration, Bromley (1989) identified Option values and quasi-option values are based on six different aspects of uncertainty in the case of air the existence of uncertainty. Option value (OV) is pollution resulting from acid deposition. They are (i) essentially the premium that consumers are willing identification of the sources of particular pollutants; to pay to avoid the risk of not having something (ii) ultimate destination of particular emissions; (iii) available in the future (see the study on valuation of actual physical impacts at the point of destination; elephants in Kenya.) While option value has been (iv) human valuation of the realized impacts at the described in various ways, one useful technical defini- point of destination of the emissions; (v) the extent tion is the difference between tlhe ex-ante and ex-post to which a particular policy response will have an welfare associated with the use of an environmental impact on the abovementioned factors; and (vi) the asset (Smith 1983). The sign of option value depends actual cost level and the incidence of those costs that upon the presence of supply and/or demand uncer- are the result of policy choice. tainty, and on whether the consumer is risk averse or Bromley suggests that the way in which policy- risk loving (Pearce and Turner 1 990). makers address these uncertainties depends on their Quasi-option value (QOV) is the value of pre- perception of the existing entitlement structure. The serving options for future use in the expectation that interests of the future are only protected by an enti- knowledge will grow over time. If a development tlement structure that imposes a duty on current takes place that causes irreversible environmental dam- generations to consider the rights of future genera- age, the opportunity to expand knowledge through tions. He terms them "missing markets," noting that scientific study of flora and fauna is losL Uncertainty "future generations are unable to enter bids to protect about the benefits of preservation tobe derived through their interests". In the absence of such a structure, future knowledge expansion (which is independent of decisionmakers may tend to follow a policy that development) leads to a positive QOV. This suggests ignores costs to future generations, and minimis that the development should ba postponed until in- costs to current generations at the expense of the creased knowledge facilitates a more informed deci- future. If the entitlement structure is adjusted, the sion. If information growth is contingent upon the policymaker can then examine three policy instru- development taking place, which is unlikely in an ments to ensure that future generations are not made environmental context, then QOV is positive when the worse off: mandated pollution abatement; full com- uncertainty regards the benefits of preservation, and pensation for future damnages (e.g., by taxation); and negative when the uncertainty is; about the benefits of an annuity that will compensate the future for costs the development.10 imposed in the present. In the face of uncertainty, the Recently, the applicability of option value has first option would appear to be the most efficient. come into question (Markandya 1991). If calcula- Other important sources of uncertainty linked tions are performed in terms of the option price with environmental issues include uncertainty about (valuing what a person would pay for future benefits land tenure, which leads to deforestation and unsus- today) then the option value may be redundant. The tainable agricultural practices, and uncertainty about option value basically calls for an individual to add resource rights, which can accelerate the rate of option value to expected future benefits to bring total depletion of a nonrenewable resource. Policyrnakers value up to the option price. As most CVM methods can address these issues by instituting land reforms, carry out valuations in terms of the option price, the and by designing appropriate taxation policies that use of option value may be impractical in many Part I: Analytical Framework 39 return rents to public sources rather than to private attribute (i.e., cleanness of water) are weak complements agents. (Maler 1974). 5. In a recent paper, Maler (1992) classifies valuation methods into two broad groups: (1) Surveys of Willingness Notes to Pay (like contingent valuation), and (2) Production Func- tion based. He divides category (2) into two sub-categories: 1. For a discussion and example of tiis, see Aylward, B. (2a) output measurable in markets (corresponding to the and E. Barbier, "Valuing Environmental Functions in Devel- secondcolumninTable3.1-conventionalmarkets); and(2b) oping Countries", Biodiversity and Conservation, 1992, vol. output not measurable (other methods in Columns 2 and 3 of 1, pp. 34-50. Table 3.1). 2. The various terms in the equation for TEV may be 6. For an introductory overview relevant to natural grouped in somewhat different ways, for convenience. See resource analysis, see Munasinghe, M. 1992. Water Supply for example Walsh, R. G., J. B. Loomis, and R. A. Gillman, and Environmental Management, Boulder, CO: Westview 1984, "Valuing Option, Existence and Bequest Demands for Press. An extensive survey including references to about 150 Wildemess." Land Economics, 60, (1). In order to measure applications has been done by C. Romnero and T. Rehman. willingness to pay for wildemess protection, they sought to 1987. pp. 61-89. A shorer, but more recent, survey is by F. separate (a future-oriented) preservation value from recrea- Petry. 1990. tional use value (in current use). Accordingly, these authors 7. For more details, see P. Dasgupta, et al 1972; A.C. defined preservation value (PV) as option value plus exist- Harberger, 1976; and Little and Mirrlees, 1974. ence value plus bequest value, i.e., PV = [ OV + EV + BV 1. 8. Assumning a consumption based numeraire for wel- 3. For an up-to-date exposition, see C.D. Kolstad and fare W, we may write W = W(c). Then B = - (c(d2W/dc2)/ JB.Braden, "Environmental Demand Theory", Chapter 2 in (dW/dc)) and g = (l/c)(dc/dt). Marginal welfare increases Braden and Kolstad (eds.) 1991, Measuring the Demandfor with consumption: (dW/dc) > 0; but at a declining rate: Environmental Quality, Elsevier, New York, 1991; and A. (d2W/dc2) < 0. Therefore P > 0, so that the sign of the term Randall, "Total and Non-Use Values", Chapter 10 in Braden (3g) is the same as sign of g. (See Dasgupta & Heal 1979, and Kolstad (eds.) 1991. chapter 10, for details). 4. Fromatheoreticallystrictviewpoint,the abovemeas- 9. For an example where the discount rate is endo- ure of value is correct in this case, only if the good in question genized and reflects future consumption, see Uzawa 1969. (i.e., visits to the water source) and the environmental quality 10. See Pearce and Turner 1990, and Fisher and Hane- mann 1987. PART II: CASE STUDIES 4. Short Case Studies of Environmental Valuation In this section shorter case studies from developing to hybrids). The calculation, which is valid for one countries are briefly outlined. For ease of reference, year only, does not take into consideratiorn the im- they are grouped according to the valuation tech- pacts on soil conservation. Calculations for maize niques described earlier. While the case studies are and sorghum show a real marginal internal financial not necessarily ideal and exhibit various shortcom- rate of return of minus 21 and minus 30 percent, ings, the main purpose of thie exposition is to illus- respectively for the two alternative options. trate the implementation of some of the techniques An economic cost-benefit analysis is per- of environmental impact valuation. Further details of formed, distinguishing between productivity gains these case studies are provided in Annex 2. These due to increased use of commercial inputs, and im- illustrations apply primarily to the direct and indirect pacts due to increased soil conservation. Unlike in use value categories shown in Figure 3.1. Because the earlier financial analysis, soil conservation developing country examples describing attempts to measures are included here, as the calculation cov- estimate option, existence, and bequest values are ers a five-year period. Costs are primarily labor rare, several studies applied to the industrialized and material input costs. This analysis is carried countries are also presented. out taking into account the perspective of the Le- sotho government and the donor agency, the Swed- Effect on Production Method ish International Development Authority ('SIDA). Primary benefit categories include increased pro- Cost-Benefit Analysis of Land Improvement duction of sorghum and maize (incremental yield in Lesotho] due to conservation relative to future decline in yield due to erosion), fruit, fuelwood, and fodder. The Farm Improvement with Soil Conservation Given various assumptions regarding the rate of (FISC) project was initiated in Southern Lesotho in growth of project implementation factors (see de- 1985, with the overriding aim of raising agricultural tails in Annex 2), the results of the base case are an production among smallholdler fanming households NPV of minus M 7.0 million at a discount rate of through soil conservation measures, subsidized in- 10 percent and minus M 5.6 million at 1 percent. puts, and afforestation. Table A 2.2 demonstrates that conservation crop In pursuit of this goal the project has rehabili- benefits are more significant at a lower discount tated old terrace structures, constructed new ones, rate, as they are slow to materialize. and added other structures for conservation. It has The qualitative interpretation is that the project promoted the use of hybrid maize and sorghum, the makes a loss that is significant in compariscin to the planting of fodder grasses, and tree planting. It has resources invested. In terms of the overriding target also promoted rotational grazing on communal ran- of the project, to raise agricultural production among geland. Conservation in this study is defined as the smallholder farming households, the project cannot promotion of optimum use of land in accordance be termed successful. However, the benefits of soil with its capability (so as to assure its maintenance conservation may have been underestimated in the and improvement). desire to improve crop yields (FISC is oriented to- A financial analysis of two cultivation options wards production rather than preservation). Given is carried out that compares a high commercial input demand uncertainty arising from lack of information alternative (implying the use of fertilizer and hybrid on future population, food habits, agricultural tech- seed) with the traditional alternative (implying no nology, and capacity to import food; and supply use of fertilizer and locally saved seeds as opposed uncertainty about the possibility of droughts leading Part II: Case Studies 41 to soil losses, it may be advisable to ascribe a positive predict the market outcome than one which focused option value to soil conservation efforts to be incor- on the returns to land. This deduction is confirmed porated into the stream of costs and benefits. Again, by behavior in the market, where forest extractivism if long-term protection of the land base is assigned a has tended to vanish wherever labor has had reason- higher priority relative to immediate producuvity in- able alternatives. creases, there would be a basis for more substantial subsidization of cover crops such as fodder grasses at Effect on Health Method the expense of traditional cropping. Aside from the debatable benefits of its produc- Economic Analysis of a Water Supply and Health tion-as opposed to conservation-oriented ap- Program in Zimbabwe4 proach, FISC serves as an important model in terms of its emphasis on consultation and community par- Fredriksson and Persson (1989) evaluate the Mani- ticipation and of its reliance on low-cost, labor-inten- caland Health, Water and Sanitation Program in Zim- sive field techniques. babwe using social cost benefit analysis. The objectives of the program are to improve living con- Valuation of an Amazonian RainforesP2 ditions in the communal areas of Manicaland through: (i) improving existing and constructing new Most financial appraisals of tropical forests have water supplies that ensure an acceptable quantity and focused exclusively on timber resources and have quality of water for domestic use and that are reliable ignored the market benefits of non-wood products, and accessible to the community; (ii) improving sani- thus providing a strong market incentive for destruc- tation conditions by constructing latrines and conse- tive logging and widespread forest clearing. In an quently preventing fly breeding and the transmission effort to illustrate the values of non-wood forest of disease; and (iii) providing health education to products, the authors present data concerning inven- improve hygienic practices and motivate behavioral tory, production, and current market value for all the changes. commercial tree species occurring in one hectare of The domestic price of labor is used as a nu- Amazonian forest. They arrive at a combined NPV meraire. The shadow price of foreign exchange was of US$6,820 for a fruit and latex production and estimated to be Z$1.75 per US$ in the analysis. selective cutting project, with logging contributing The shadow price of unskilled labor is set at zero just 7% of the total. This compares favorably with an in the dry season, due to unemployment in Zim- estimated NPV of US$3,184 for timber and pulp- babwe. In the peak season, there is a shortage of wood obtained from a 1 hectare plantation in Brazil- labor. As uncertainty exists because of possible de- ian Amazonia, and an NPV of US$2,960 for fully lays in the harvest, a sensitivity analysis is made in stocked cattle pastures in Brazil. the evaluation, where the shadow wage of unskilled In order to extrapolate the value of the project, labor is set to 100%, 75% and 50% of the market which is based on the value of 1 hectare, it would be wage of Z$0.46 per hour. The opportunity value of useful if attempts were made to incorporate elastici- children's labor is set to zero during the whole year. ties of demand for the products. The broader validity (This is a little surprising, given that children repre- of the results for remote areas is doubtful, since this sent a significant labor force in many developing study was carried out for an area that was quite close countries). For skilled labor, the shadow wage is set to a market town. at the market wage. Another methodological concern is the use of The social discount rate is assumed to be the rate returns per hectareas the unit of comparisonbetween of return in the private sector, which would mean a different uses. A recent World Bank study concludes real social discount of 4.86% since the commercial that an appropriate analysis would compare returns bank lending rate to the industrial sector at the end per productive unit, including land, labor and capital of 1987 was 14.75% and the inflation rate was valued at their opportunity cost.3 Since land in the 9.89%. A time horizon of 40 years was established Amazon is generally abundant relative to labor, an which was the duration time of benefits from the analysis based on the returns to labor would better project. 42 Part II: Case Studies Materials supplied by the project were valued at if a shadow wage of 100% and 100% disease reduc- market prices. Community input is valued at the tion is achieved. opportunity cost of unskilled labor. The constructor While most of the benefits come from disease is assumed to be paid the shadow price of skilled reduction, it must be pointed out that benefits are labor. probably underestimated through the use of a lower The change in consumer surplus is determiined bound. This is because the value of a saved life has by calculating the average price for water collected been underestimated, through the use of a lower in the wet season and dry season, both before and bound. Secondly, the paper does not take into ac- after the project. The price of water in the wet season count other benefits such as local industry that would is calculated using 75 peak period days with no energy benefit from improved water supply. T'hird, as in- savings but with a shadow wage greater than zero. In come distribution will probably improve as a result the remaining 105 days energy is saved and the shadow of the project, benefits could have been given a wage equals zero. Using this infonnation, the change greater weight. The authors conclude that the project in consumer surplus is then calculated at shadow wages is a success from an overall societal point of view. of 50%,75%, and 100%. Cost of illness estimnates consist of treatment Travel Cost and Contingent Valuation costs, costs of lost production and costs for extra Methods transportation. Since willingness to pay for healthier and longer lives was not estimated, the values used The Consumer Surplus From Visits to a Costa significantly underestimate the true values. For treat- Rican Rainforestr ment costs, the costs for private treatment are used as the opportunity cost. This study measures the value of ecotourism at a A life saved would be valued very highly by the tropical rainforest site in Costa Rica using the travel individual concerned and his/her family. From soci- cost method. By observing travel behavior, the ety's viewpoint, a lower bound or miniimum estimate authors reveal that Costa Rican visitors are willing of the gain from a life saved is derived by the authors to pay $35 per household to visit the site. The study using future production and consumption. The finds that visitation is highly correlated with educa- weighted average of the wage adjusted for both tion (and therefore probably income), and that house- unemployment and the income in communal lands is holds in areas with high population densities make used as an estimate of a child's future production, and more trips. future consumpticil is estimated as final household The paper only considers domestic visits, al- expenditure per capita. The net present value of the though foreign visitors to the site outnumbered do- difference between production and consumption is mestic visitors by four to one in 1988. Foreign the net output gained by saving the child. The same visitation is likely to be worth far more than domes- measure is utilized to de,termine net present output tic, as foreign visitors have higher travel costs, and a produced by an adult's life if he/she is saved today. greater value of travel time because of higher earn- Sensitivity analyses were conducted for social ings. Additionally, they provide foreign exchange. discount rates of 2%, 4.86%, 7.24% and 9%; disease Nevertheless, if we use the same value of $35 per visit reductions of 40%, 70%and 100%; and wet season for all visitors, this would result in an NPV of $1,250 shadow wages for unskilled labor of 50%, 75%, per hectare. This figure is one to two times the mag- and 100% of the market wage for casual workers nitude of the purchase price currently paid by the on commercial farms in Manicaland. In the base reserve for the acquisition of new lands. case, with an estimated social discount rate of 4.86%, It is unclear, however, whether the authors as- shadow wage 100% of the market wage, and a 100% sess the cost (in foregone earnings) of time spent at health improvement, the internal rate of return was site. It would have been useful to have a clearer greater than the social (fiscount rate and the net definition of their term "cost of travel tinme." present value of the project was strongly positive. At In addition, they use a linear demrand function a social discount rate of 7.24%, estimated in the alter- as opposed to the more popular semi-log functional native approach, the project is not found profitable even form, as visitation rates from many zones were zero. Part II: Case Studies 43 In a similar study, Willis and Garrod (1991) made the US$30 million, respectively, based on an estimnate of case for the superiority of the semi-log functional 250,000 to 300,000 adult safaris per year. (This is an form over the linear for the Clawson-Knetsch Zonal example of option value-the premium consumers Travel Cost Method. They also found that the zonal are willing to pay to avoid the risk of supply uncer- travel cost method probably overestimated the con- tainty.) sumer surplus for their sample of travel cost studies. Note that both methods produced anmual view- The individual travel cost method comes closer to ing values for elephants of around US$25 million. contingent valuation results. Although these estimates are rough, they are a useful As in the Costa Rican study, most travel cost guide to the order of magnitude of value. studies look at single-purpose, single destination trips. A more general methodological problem is how The Willingness to Payfor Water Services to deal with multiple destination trips. In most cases in Haiti7 of intemational tourism to developing countries the travel cost would need to be attributed to many The contingent valuation method was used in this activities at a number of sites. The problem then study to estimate consumers' willingness to pay for becomes to elicit the specific value given to a certain an improved water system in a village in southem site. Haiti. The project was executed by CARE. The re- search team devised tests in an attempt to correct The Value of Viewing Elephants on Safaris biases that could threaten the validity of the survey in Kenya6 results, such as strategic bias, starting point bias, and hypothetical bias. The results of the survey, utilizing The travel cost method was used to estimate a de- an ordered probit model as opposed to a linear model, mand function for safaris in Kenya. The analysis is demonstrated that willingness to pay for a new water based on the 80 percent of tourists who come to system (whether for a public standpost or for a pri- Kenya from Europe and North America. Surveys of vate connection) was positively correlated to in- tour operators and visitors provided data for the come, the cost of obtaining water from existing estimation of land costs, air fare, and travel time sources, and the education of household members. It costs. Travel time costs were weighted at 30 percent was negatively correlated with the individual's per- to reflect the fact that vacation time is valued at lower ception of the quality of water at the traditional than the gross wage rate. A weighted average con- source used before the construction of the improved sumers' surplus of US$725 is estimated. This gives water supply system. The sex of the respondent was a total consumer surplus for those on safari in the statistically significant in the model for public stand- range of US$182 million to US$218 rnillion annu- posts but not in the model for private connections. ally, depending on the assumed level of visitation. The mean of WTP bids for public standposts To identify the contribution elephants make to represented 1.7 percent of household income, while the value of a safari, tourists were asked to allocate the mean WTP bid for private connections was 2.1 the pleasure and enjoyment of their trip over various percent of household incomes. These bids are signifi- categories of experience. Elephants represented 12.6 cantly lower than the 5 percent rule-of-thumb often percent of total enjoyment. Therefore, the estimated used in rural water supply planning as an estimate of economic value of a safari yields a viewing value for maximum "ability to pay" for private connections. elephants of US$23 million to U$27 million per year. However, the bids are based on the assumption that In order to assess consumers'willingness to pay the public standposts are already in place. to maintain the elephant population at current levels The results of this study show that in a case through increased enforcement activity, a survey was involving a familiar (basic need) commodity,it is designed, utilizing the contingent valuation ap- possible to obtain reasonable, consistent answers in proach. Attempts were made to adjust for biases. The a contingent valuation survey conducted among a average value was 89 dollars while the median was very poor, illiterate population. Contingent valuation 100 dollars. This yields an annual viewing value of is likely to see greater future use in developing coun- US$22 million to US$27 million and US$25 to tries; for collecting information on individuals'will- 44 Part II: Case Studies ingness to pay for various public infrastructure pro- expected inverse relationship between travel costs jects, and (with more difficulty), also for environ- and number of visits. mental protection services (such as the treatment of The TC model used was of the standard form- industrial wastewater flowing through residential ar- visitation rates were assumed to be a function of total eas). travel cost, availability of substitute sites, and in- come. The total travel cost variable included both Willingness to Payfor Improved Sanitation in Ku- out-of-pocket travel costs plus the monetary value of masi, Ghana8 time spent traveling, using a representative wage rate. 187 people were interviewed and divided into Willingness to pay was estimated through a contin- 17 zones of origin within Bangkok. A regression gent valuation approach in Kumasi, Ghana (Whit- equation of visitation rates on travel cost was esti- tington et al., 1992). Options provided to consumers mated, and a demand curve derived for public use of were water closets with a piped sewerage system and the park. The area under the curve, (the consumers' ventilated pit latrines (KVIPs). KVIPs provided a surplus of park users), was estimated at 13.2 million cheaper option for sanitation, as they did not require baht a year. sewer connections and installation of water closets. The 187 persons interviewed for the TC study Households that already had connections to were also interviewed for the CVM study. Visitors water were asked their WTP for a water closet and a for recreational purposes indicated a slightly lower KVIP. Households with water closets were asked the WTP for yearly contributions to maintain the park amount they would pay for a sewer connection. than visitors for morning and evening exercise pur- Results showed that WTP for a WC or a KVIP was poses. However, the latter group's willingness to pay roughly the same for houses without WCs. House- per visit was less than that of recreational users. holds with bucket latrines were willing to pay the An additional 225 people were interviewed lowest for KVIPs; those using public latrines bid throughout Bangkok, including people who had significantly higher prices, up to 30-35% more, never used the park, in an attempt to estimate the which demonstrated dissatisfaction with the incon- broader "social" value of the park. When CVM re- venience and lack of privacy in the public systems. sults were adjusted by the appropriate age-corrected Overall mean bids of around $1.5 per month com- population figures, the W`IP measure of park users pared to average existing expenditures of about $0.5 was estimated at 13 million baht per year, and for the per month WTP was found to be less than costs of survey of Bangkok residents, 116.6 million baht per supply. year, (thus demonstrating the more realistic WTP of The study concludes that the required subsidy constant users of the park). The consumers' surplus for a WC system for Kumasi would amount to $60 and welfare gain associated with the continued exist- million. The required overall subsidy would be $4 ence of the park is clearly demonstrated. million for the KVIP system. Although benefits of improved health were not estimated, a subsidy Contingent Valuation Method to Estimate would probably be justifiable in terms of benefits Option, Existence, and Bequest Values to public health. Hardly any developing country examples are avail- Value of a Thai National Park9 able in this category, and a cautious approach is recormnended. For illustrative purposes, four U.S. In a 1980 study, TC and CVM methods are used to examples are provided below. These studies all used estimate the economic value to Bangkok residents, the contingent valuation approach to obtain actual of the Lumpinee Park in Thailand. The study is measures for option, existence, and/or bequest val- unusual in that the urban setting is not usually con- ues. The first case was quite possibly the earliest sidered suitable for a travel cost approach. However, survey that developed a methodology to attempt to even with short visits and minimal travel time and determine existence value. Both this first and the costs to use the park, the analysis still confirmed the second studies, while revealing the significance of existence values, examine them as an adjumct to Part II: Case Studies 45 theirmainfocus,whichisrecreationalusevalue.The Option Price and Existence Value of Wildlifell third case is perhaps the first study that undertook to examine total preservation value in depth, broken This study measures the option price (option value down into its three separate components of option, plus expected consumer surplus) and existence value existence, and bequest value. The fourth study is of grizzly bears and bighorn sheep in Wyoming- basically an attempt to measure the effects of infor- both of these species being endangered by human mation disclosure (endangered status and physical activity in the area. A mail survey was sent out, with appearance) on existence values for endangered spe- questions being directed towards hunters and non- cies. hunters. Hunters were asked their WTP for a "stamp" allowing them to hunt in new hunting areas in either Existence Value of Preserving Visibiliy'0 five or fifteen years for grizzly bear or bighorn sheep. Respondents were each confronted with only one The survey attempts to measure annual household time horizon. The probability of supply was variable. willingness to pay (WTP) to preserve visibility in the Nonhunters were asked to specify their WTP for the Grand Canyon-both WTP if visibility preservation existence of the animals or for the opportunity to were to be extended to the entire southwestern observe them in the future. parklands region and WTP to prevent plume blight As expected, the overall option price increased seen from Grand Canyon National Park. For the as the probability of supply increased. Contrary to purpose of the study, only the major source of air expectations, no systematic relationship could be pollution in the region, coal-fired power plants, was determined showing that bids based on certain de- the focus. Over 600 households in Denver, Los An- mand exceeded those based on uncertain demand. geles, Albuquerque, and Chicago participated in the Existence values and observer option prices were survey. One-third of the respondents were asked a significant. The mean bids for observer option prices pure user value question: How much would they were in the range of US$20 for both grizzly bear and be willing to pay in higher entrance fees per day bighorn sheep, regardless of the time element. This for visibility protection at the Grand Canyon or is on a par with option bids for hunters at high levels other parks? The other two-thirds of the respon- of supply certainty. Existence values are high for dents were asked how much they would be willing grizzly bear ($24 at five years, $15.20 at fifteen to pay in higher electric power bills to preserve years), but are significantly lower for bighorn sheep visibility in the parklands, a measurement of total ($7.40 and $6.90 respectively). preservation value (defined by the authors as the sum of existence plus user value). The authors Option, Existence, and Bequest Values interpreted existence value as the difference be- of Wilderness12 tween total preservation value and user value. The preservation value bids are substantially The key question posed here is the amount of wilder- higher than the user value bids, apparently signifying ness to be protected in Colorado. A sample of 218 that existence value is an important component of resident Colorado households participated in a mail total economic value. The authors are careful to point survey. Respondents were asked to report their will- out that visitation plans were not an overwhelming ingness to pay into a special fund to be used exclu- factor in determining preservation value bids, and sively for the purpose of protecting wilderness. This that knowledge acquired through previous visits was payment vehicle is recognizable to Colorado resi- also considered relatively unimportant in the deter- dents, being similar to the state income tax form's mination of bids. Moreover, preservation bids did not checkoff for nongame wildlife preservation. Respon- decline with distance, which seems to indicate that dents were asked to write down the maximum nonuse value was an important component in the amount of money they would be willing to pay respondents' bids. annually for protection of current wilderness, and for hypothetical increases in wilderness depicted on four maps. Once this budget allocation was completed, 46 Part II: Case Studies respondents were asked to allocate the highest endangered species and preservation altematives amount reported among ifour categories of value: plays an important part in determining the replicabil- recreational use, option, existence, and bequest de- ity and usefulness of existence valuation results. mands. Total preservation benefits are estimated as The experiment was conducted using isolated the residual after recreatien use benefits have been experimental and control groups of paid university- subtracted from total WTP for wildemess protection. level student subjects in the United States. CVM was Preservation values were estimated by developing an used to measure preservation bids for a humpback appropriate econometric model of willingness to pay whale preservation fund. The experimental group by survey households and by aggregating values was then provided with more information about the across households in the state. whales (through the screening of a film), and both Results indicate that as the quantity of wilder- groups were then questioned again. The experimen- ness increases, annual household preservation values tal group increased their bids by 32 percent from their increase at a decreasing rate, except for bequest original values, and the control group increased their value, which is linear. Option value had a strong bids by 20 percent. This may be attributable to the positive association with income. In-state wilderness fact that all respondents had more time to reconsider users had a much higher option value than nonusers, their bids, and perhaps demonstrates how prefer- indicating that recreational use is an important ele- ences are learned through the interview process it- ment in the determination of option value. Existence self, even in the absence of new information. value was positively related to the importance of Finally, all control and experimental, subjects preservation of natural scenery, ecosystems, and ge- were asked to fully allocate a lump-sum windfall netic strains. Existence vaJue increased with fre- gain of $30 arnong preservation funds for three ani- quency of wilderness trips undertaken. All income mal species, given four scenarios, containing differ- groups valued existence of wvildemess about equally. ent levels of information about physical appearance Interestingly, a wide range of workers (skilled and and endangered status. unskilled) would pay US$1.50 more for existence The effects of information disclosure on re- demands than would persons in other occupations. sponses was more evident here. Faced with zero Bequest value was not influenced by the number of information distinguishing species, the subjects' children living at home with respondents. This seems willingness to pay to preserve each species was to indicate that bequest value is correctly defined as nearly equal. Given infornation on physical appear- the satisfaction from interpersonal transfers of wil- ance, they allocated more to the monkey-like animal derness to indefinite future generations rather than as compared with the rabbit-like or rat-like animal, specifically to the children of the respondent. Retired reflecting a strong anthropomorphic tendency. Given persons were willing to pay US$6.15 more for be- information on endangered status, respondents allo- quest demand than were other respondents. All in- cated significantly more funds to the animal that was come groups valued bequest demands about equally. endangered but savable as compared with ubiquitous The authors conclude that, even without taking or extremely rare animals. When information was into account the preservation estimates of nonresi- provided on both physical appearance and endan- dents of the state, adding preservation value to the gered status, the endangered but savable species consumer surplus of recrealional value had a sub- received the highest allocation followed by the rare stantial effect on the benefit value for wildemess. and abundant species. These results suggest that information about endangered status may be rela- Existence Value of Endangered Species13 tively more important to respondents than informa- tion about physical characteristics in formulating This study tests the hypothesis that an individual's preservation bids. WTP to preserve a particular animal is significantly In conclusion, it appears that information dis- influenced by information provided about the ani- closure can influence perceived marginal efficiency mal's physical and behavioral characteristics and of investment in a preservation fund and thereby about its endangered status. Public awareness about Part II: Case Studies 47 result in changes of an individual's budget allocation 6. This case study is derived from Brown and Henry, strategy. 1989. 7. This case study is derived from Whittington, et al. 1990. Notes 8. This case study is derived from Whittington et al., 1992. 1. This case study is derived from J. Bojo 1991. 9. This case study is derived from Grandstaff and Dixon 2. This case study is derived from Peters, et. al. 1989. in Dixon and Hufschmidt (eds.). 1986. 3. World Bank (1992). 10. This case study is derived from Schulze, et al. 1983. 4. This case study is derived from Fredriksson and l1. This case study is derived from Brookshire, 1983. Persson, 1989. 12. This case study is derived from Walsh, et al. 1984. 5. This case study is derived from Tobias and Mendel- 13. This case study is derived from Samples, et al. 1986. sohn, 1991. 5. Two World Bank Valuation Case Studies The two longer case studies in this section were sonnel, and the development of roads, visitors'facili- carried out recently in the World Bank, and use ties, and other infrastructure. Another important set combinations of valuation methods to facilitate of costs that are often ignored are the opportunity decisionmaking. costs associated with the foregone uses of park land. Benefits include both use values and non-use Valuation of Biophysical Resources values. Most parks do not allow exploitation of forest in Madagascar' resources, and the primary uses are therefore for tourism and research. Tourism can generate consid- In the last two decades, environmental and resource erable revenues for the country from both entrance economnists in industrialized countries have devel- fees and travel expenditures. National parks also oped non-market techniques for the valuation of generate a number of non-use benefits, among which environmental costs and benefits. These methods existence value and option value are important. Ex- have been successfully applied in developed coun- istence value is defined as the willingness to pay tries although they have rarely been used in develop- (WTP) to preserve the park-by individuals who ing countries where the neaed for national park never plan to use it. Option value is defined as the creation is typically much higher. In this section, the willingness to pay to maintain the park for possible estimation of the non-market costs and benefits of use in the future. Other benefits may include reduced establishing a national park on nearby villages, and deforestation, watershed protection and climate valuation of the park as an international tourism regulation. This study focuses on the measurement destination are discussed. The results summarized of some of the more important and more difficult to below are from the first stage in the analysis to arrive measure economic inpacts, namely the impact of the at a rational decision concerning the proposed crea- park on local villagers and the benefits of the new tion of the Mantadia National Park in Madagascar. park to foreign tourists. Further details are available in Kramer, Munasinghe, Sharma et al. (1992). Opportunity Cost Analysis Madagascar is one of the economically poorest and ecologically richest countries in the world, and The opportunity cost approach uses standard eco- it has been designated by the international commu- nomic analysis based on market values, to detemiine nity as a prime area for biodiversity whose ecosys- the net economic benefits associated with altemative tems are also atgreatrisk. Maclagascaris onIUCN's uses of one or more resources. In this case, the list of megadiversity countries because of its ex- relevant opportunity costs would be those associated traordinarily high rates of species endemism. As a with the alternative uses by people living near and in result of the enormous biodiversity, the intema- the park. The creation of the park imposes a consid- tional donor community is providing large sums of erable economnic burden on the local population. The money in order to save as much of Madagascar's opportunity cost is estimated by determining recent biodiversity as possible. The government of Mada- land use and by projections of future land use gascar is also taking steps to control forest degradation changes in the absence of a national park. and to protect biodiversity. There are no human settlements within the Man- Three main methods for ithe valuation of envi- tadia National Park boundaries, but several villages ronmental impacts are applied here-the contingent lie in close proximity. The villagers depend on the valuation, the travel cost, and the opportunity cost forests in and around the park for forest products and approaches. agriculture. The form of shifting cultivation used for The creation of a national park generates many agriculture production in eastem Madagascar is criti- both indirect and direct costs and benefits. Costs cally important as a mechanism for deforestation. It arise from land acquisition (if the land had been is also the only means of livelihood known to many previously privately owned), the hiring of park per- of the inhabitants of the region. Furthermore, fuel- Part II: Case Studies 49 wood is collected from the forests, a wide range of The village survey had to be modified, to over- fish and animals are consumed, and a number of come constraints. For example, in devising the ques- different types of grass are harvested and used for tions for the village survey, initially it appeared from assorted purposes. villager responses that they were willing to pay to The opportunity costs associated with these eco- protect the forest. However, it transpired that this nomic activities were estimated using a survey of 351 response was not related to perceived non-use bene- households in 17 villages within a 7.5 km radius fits from thepark, butrather was theresultof a sense around the park boundary. The survey was adminis- of coercion arising from arrests of villagers by local tered by a local NGO well versed in rural survey authorities, for incursions into the park area. The techniques, and it was undertaken following a recon- questions were therefore reformulated in terms of naissance visit to the village, several focus group willingness-to-accept compensation for being de- interviews, conversations with people well ac- nied access to forestland within the park. quainted with the area, and a pretest covering 25 households. It was administered in Malagasy, the TravelCostMethod national language. The questionnaire focused upon (i) establishing Travel cost models use the amounts of time and the extent of the dependence of the local villagers on money visitors spend traveling to a site as the price forests nearby for obtaining a wide variety of forest proxies, together with participation rates and visitor products; (ii) establishing the extent to which the attributes, to estimate the recreational value of the villagers used the forest for shifting cultivation; and site. Recreation in Madagascar's national parks con- (iii) assessing local attitudes toward conservation of trasts sharply with the standard assumptions that the the forests. Questions related to socio-economic trip is a single-purpose, single destination day-trip to variables, land use, time allocation, and house- a site that affords some particular recreational expe- hold production activities were also asked. The rience or typical quality which can be substituted for final section used the contingent valuation those available at similar sites. Instead, recreators in method discussed below. Madagascar can be divided into two groups consum- A separate questionnaire was administered to ing distinct goods: (i) local visitors who make day the village leaders. It focused on issues pertaining to trips to national parks to view the local natural envi- general agricultural patterns, markets and prices of ronment; and (ii) intemational tourists who under- goods sold, village history and migration patterns, take lengthy trips to experience unusual natural forest related cultural issues, and details on shifting settings and cultures. This part of the study focuses cultivation practices. on the second aspect. The novel international travel cost method used here is summarized in Annex 3, and Contingent Valuation Method presented in full in Mercer and Kramer (1992). The method is based on the assumption that The contingent valuation method (CVM) uses sur- individuals travel to a single country where they vey techniques to establish the value of goods and engage in a variety of activities. Estimating the services which are not usually exchanged in markets. model requires specific data on how each household distributes its time across activities during the time In this study, the CVM was used in both the horizon of the model, and ideally it implies the village survey and the tourist survey. In the tourist collection of full trip itinerary data as well as travel survey, the CVM was used as an altemative method cost information for foreign visitors. The itinerary to the travel cost method for estimating the total value data includes the distribution of time between activi- of the park to the tourists. These questions were ties for each individual, the costs of pursuing such phrased in terms of how much more the foreign tourists activities, and the features of the various activities would have been willing to pay for their trip, if the new that lead to differences across individuals in their park hadbeen createdfor them to visit.These questions ability to undertake them. were also pretested and revised prior to the implemen- Based on the theoretical model, several ques- tation of the tourist survey. tionnaires were prepared and translated into French 50 Part II: Case Studies and administered to visitors in the small Perinet The survey indicates that fuelwood is the most Forest Reserve adjacent to the Mantadia National important forest product collected (see Table 5.2). Park. The questions consisted of questions on the The average household collects about 6,164 kgs, or cost of the current trip to iMadagascar, details of about US $38 worth of firewood per year. T'he total previous nature related tourist trips, the process for value of collected firewood is $13,289 per year. Total deciding on trip destinations, contingent valuation value of other collected forest products is US$ 818. questions for the willingness to pay for visits to the To estimate the opportunity cost to villagers of Mantadia National Park, and a series of socio-demo- establishing the Mantadia National park, income graphic and economic questions. In addition, a from agricultural and forestry activities was esti- "Madagascar Trip Diary" was developed which elic- mated for three different groups of villages. The ited detailed itinerary, cost, time, and quality infor- mean value of losses was $91 per household per year. mation for the current trip to Madagascar. The Aggregating over all households living in the vicinity questionnaires were tested in the US with a focus of the park and using a 10% discount rate and twenty group of previous visitors to Madagascar, and in year time horizon, the net present value of the oppor- Madagascar the questionnaires were revised follow- tunity costs was estimated to be $556,010. ing pretests with a small sarnple of visitors to the The contingent valuation study of the villages Perinet Reserve and discussions with local Malagasy indicates the villagers' perceptions of the forest. collaborators. Due to political unrest in Madagascar Forty percent seemed to think that forests do not help during the time of the survey, only 94 surveys were soil protection, although sixty-five percent agreed completed, including a 14% return rate on the travel that floods occur less frequently with forests. inter- diaries. Although this data setL is inadequate for full estingly, 91% of the respondents agreed that primary implementation of the international travel cost forests are "more fun" than secondary forests, which method, the estimates of mean willingness to pay to suggests a recreational value of the forests, but 77% visit the new park derived from it provide useful of the respondents did not think that preserving for- insights. ests in order to preserve ancestral graves was very important. Finally, 68% of the respondents thought Preliminary Empirical Results that it is advantageous to clear the forests as a form of pest management. The average household size in the surveyed villages The responses to the contingent valuation ques- was 4.6 persons. In 1988, average per capita income tions indicate that on average, a compensation of rice was $190 in Madagascar, and the villages in the equivalent in value to $108 per year per household survey may well have incomes lower than the aver- would make households as well off with the park as age. Several villages are very isolated, and many do without. Aggregating over the population in the park not have access to medical facilities, running tap area, this implies a necessary one time compensation water, electricity and primary schooling. Approxi- of approximately $673,078, assuming a 10%7o dis- mately 95% of the householcls own land, and the count rate and twenty year time horizon. average amount of land owned is 1.9 hectares per In the tourist survey, income for the visitors household (see Table 5.1). In the survey, 36% of the ranged from $3,000 to $300,000 with a mean of households own a watch, 33% own a radio and 97% $59,156 (see Table 5.3). The average tourist was 39 have a kerosene lamp to light their huts. The average years old and had completed 15 years of education. household produces 487 kg of paddy rice per year, Visitors came from 13 countries. Trips rangecd from or about US $128 worth of rice. Most households 3 to 100 days in length, with a mean of 27 days, and also engage in shifting cultivation. Eighty percent of with 1-8 days spent at Perinet (mean of 2 days). the households surveyed said that they would add to Expenditures for the trip ranged from $335 to $6,363 existing land for cultivation. Ninety-nine percent of with an average trip costing $2,874. The mean trans- these acknowledged that they planned to cut forests port cost to reach Madagascar was about $1,390 to add to their land. The average household planned while transport costs within the island averaged al- to cut 1.6 hectares of forested land in the coming year most $590. A further breakdown of tourist charac- to undertake shifting cultivation. Part 11: Case Studies 51 Table 5.1 Land Use Information for Villages Variable Number of Observations Range Mean Total quantity of farmiand 311 0 to 9 1.89 per household (hectares) Planned increase in culti- vated land per household 256 0 to 10 1.70 (hectares) Annual quantity of farm- land planted with rice per 289 0.04 to 5 1.04 household (hectares) Total annual rice yield per 296 2 to 3,600 487.00 household (kilograms) Total annual quantity of rice marketed per house- 249 0 to 990 41.80 hold (kilograms) Total annual value of rice 296 $0.5 to 1,101 $128.00 yield per household ($) Table 5.2 Value of Forest Products Collected by Viflagers Total Annual Value Mean Annual Value for all Villages per Household Forest Products Number of Observations ($US) ($US) Fuelwood 316 13,289 38.0 Crayfish 19 220 12.0 Crab 110 402 3.7 Tenreck 21 125 6.0 Frog 11 71 6.5 teristics by country of origin, showed considerable tions as a function of travel costs, socioeconomic diversity (see Table 5.4). characterisitics, and quality variables. The model Using data from the tourist survey supple- was used to project benefits to tourists that would mented by data from a separate survey of travel result from a 10% improvement in facilities at the experts, an econometric analysis was performed to park. The average increase in willingness-to-pay was apply the travel cost method. The model, known as estimated to be $24 per tourist, resulting in an annual a random utility model, examines the allocation of benefit to foreign tourists of $93,600, based on the trip choices to international nature tourism destina- current annual visitation rate at the Perinet Reserve. 52 Part H.: Case Studies Table 53 Summary Statistics for Complete Sample of Tourists Variable Number of Observations Range Mean Income 71 $3,040 to 296,400 $59,156 Education 86 10 to 18 years 15 years Age 87 16 to 71 years 38.5 years Number of Days in 83 3 to 100 days 26.6 days Madagascar Number of Days in Perinet 80 1 to 8 days 2 days Total Cost of Trip to 78 $335 to $6,363 $2,874 Madagascar Total Cost to Madagascar 47 $352 to 5,000 $1,388 Total Cost in Madagascar 43 $8 to 2,000 $588 Table 5.4 Summary Statistics for Tourist by Country of Origin Mean Mean Mean Number of Number of Mean Mean Percent of Expenditure Days in Days in Mean Age Education Income Country Sample ($US) Madagascar Perinet (years) (years) ($US) Britain 20.2 3,332 18 1.6 45 15.8 36,891 Italy 21.4 2,357 21.4 1.9 34 14.2 112,000 France 15.5 2,481 36 1.9 34 15 63,197 Germany 11.9 3,172 24.8 1.8 40 15 42,304 Switzerland 11.9 3,200 37.6 2.3 36 15.6 51,243 USA 4.8 3,097 18.5 2.75 49 16.5 53,515 Other 1. 6629 081 (7 countries) 14.3 2,726 26.6 2.91 40.8 4 33,997 At a 10% discount rate, this would generate a net annually. This amounts to $2.16 million as the pre- present value of $796,870 of benefits associated with sent value for the stream of benefits over 20 years- the park over twenty years. again assuming a 10% discount rate. Utilizing the contingent valuation method, in a discrete choice format, the mean bid for tourists to Conclusions view the new park (conditional on seeing the same number of lemurs) was $65. Assuming current visi- Several tentative conclusions can be drawn from the tation pattems continue, the total additional willing- early results of this study. Non-market valuation ness-to-pay to visit the new park would be $253,500 techniques can provide useful information for eco- Part II: Case Studies 53 Table 5.5 Summary of Economic Analysis of Mantadia National Park Estimates of Welfare Losses to Local Villagers from Establishment of Park Method Used Annual Mean Value per Household Aggregate Net Present Value Opportunity Cost $91 $673,078 Contingent Valuation $108 $566,070 Estimates of Welfare Gains to Foreign Tourists from Establishment of Park Method Used Annual Mean Value per Trip Aggregate Net Present Value Travel Cost $24 $796,870 Opportunity Cost $65 $2,160,000 nomic evaluation of national parks. A major strength At the same time, the early findings indicate of this study is the opportunity to compare valuation future issues. Reliance on WTP is fundamental to the techniques (Table 5.5). For the village component, economic approach, but tends to overemphasize the the estimated welfare estimates based on two entirely importance of value ascribed to richer foreign visi- differentmethods, opportunity cost analysis and con- tors. If conflicting claims to park access were to be tingent valuation method, were remarkably similar determined purely on this basis, residents (especially, ($91 and $108 per householdper year). The estimates the poor local villagers) are likely to be excluded. of tourist benefits based on the travel cost method Therefore, as indicated in the introduction, the so- and contingent valuation method were somewhat ciocultural concepts of sustainable development (es- more disparate ($24 versus $65 per trip) but it is pecially intragenerational equity and distributional noteworthy that the benefit estimates are of the same concerns) would need to be invoked to protect the order of magnitude. It is important to note that the basic rights of local residents-perhaps in the form higher contingent valuation estimate may reflect of a "safe minimum" degree of access to park facili- some non-use values, while the travel cost method is ties. for use value only. Further, research of this type would have impli- Improving Decisionmaking in the Sri Lanka cations for policy, investment decisions, resource Power Sector2 mobilization, and project design and management. The information can help govermments decide how The incorporation of environmental externalities to (a) allocate scarce capital resources among com- into decisionmaking is particularly important in the peting land use activities, and (b) choose and imple- power sector, where environmental concerns (rang- ment investments for natural resource conservation ing from greenhouse gas emissions of fossil-fueled and development. Results can also be used in deter- plants to the impacts of inundation at hydro plants) mining or influencing pricing, land use, and incen- have posed increasingly difficult constraints to pro- tive policies. At the local level, the findings can be ject implementation. It is also clear that in order for used to determine compensation for local villagers environmental concerns to play a real role in power for foregone access to forest areas designated as sector decisionmaking, one must address these issues national parks. In addition, the research findings can early-at the sectoral and regional planning stages, show the value of a park as a global environmental rather than only at the stage of project environmental asset to foreigners, thus influencing the external as- assessment. sistance for conservation programs at the local level. 54 Part II: Case Studies Unfortunately, as soon as one is dealing with power, increasing use of demand side management, power sector issues at this aggregate planning level, and .system efficiency improvements), that make the application of many of the project-level valuation decisionmaking quite difficult-even in the absence techniques discussed earlier becomes extremely dif- of the environmental concerns. The study is rela- ficult, for two main reasons. The first is the nature of tively unique in its focus on these kinds of planning the impacts themselves-the health effects of pollut- issues, as opposed to the more usual policy of assess- ants from coal fired generating stations, the potential ing environmental concerns only at the project level loss of biodiversity associated with large scale hydro after the strategic sectoral development decisions reservoirs, the impacts of greenhouse gas emis- have already been made. sions-all are exceptionally difficult to value. In- deed, attempts to do so would very likely focus Environmentallssues attention on the validity of the valuation techniques themselves, rather than the policy trade-offs that Sri Lanka (see map in Figure 5.1) is one of the more must be made. The second reason concerns the scale densely populated countries of the world, and land of analysis. Many of the techniques discussed in this availability is an important issue. In general, hydro paper are most appropriate at the micro-level: the use plants are in the wet zone areas where there is little of the contingent valuation approach is much more vacant land nearby for resettled inhabitants to relo- valid where respondents can be asked specific ques- cate, while land that is available at greater distances tions about impacts of a particular project to which is often seen by potential evacuees as undesirable they can relate-as the previous example of Mada- because of questions concerning the availability of gascar illustrates well. However, this may be very adequate water supply. A rough but effective way of difficult to apply in situations where one is dealing comparing the likely extent of potential land-related with a potentially large number of technology, site environmental impacts across projects is the area and mitigation options. inundated per KWh of capacity. This varies between It is in these kinds of situations that the tech- zero and as much as 150 hectares per K"ha The niques of multi-criteria analysis (MCA) may be ap- correlation between the installed capacity and the plied. Such techniques first gained prominence as amount of land to be inundated is poor; large projects practical evaluation tools in the 1970s, when the do not necessarily mean worse envirormnerntal im- intangible environmental externalities lying outside pacts and vice versa. conventional CBAmethodologies were increasingly The progressive loss of Sri Lanka's natural for- recognized. It also met one objective of modern ests over the past 50 years is well documented, and decisionmakers, who preferred to be presented with is one of the country's most important environmental a range of feasible alternatives as opposed to one concerns. Power sector projects will be scrutinized "best" solution. As explainecd earlier, MCA allows very carefully for their potential impact on what for the appraisal of alternatives with differing objec- natural forest areas remain, even if it is true that the tives and varied costs and benefits, which are often power sector per se has been a relatively minor assessed in differing units of mneasurement. contributor to the loss of forest lands. The main Such an approach was used by Meier and reason for deforestation in the past has been planned Munasinghe (1992) in a study of Sri Lanka. The agricultural development and settlement schemes, objective was to demonstrate how environmental chena cultivation, encroachment by unplanned set- externalities could be incorporated into power sys- tlement and cropping, illicit logging and luncon- tem planning in a systematic and efficient manner. trolled fuelwood and timber extraction. Sri Lanka presently depends largely on hydro power Relatively little is known about ambient air for electricity generation, but over the next decade quality in Sri Lanka. In most parts of the country the there seems little choice other than to begin building air quality is fairly good, a reflection of the limited large coal- or oil-fired stations, or to build hydro extent of industrialization except in Colombo, and plants whose economic returns and environmental the natural ventilation provided by strong monsoonal impacts are increasingly unfavorable. In addition, winds. In Colombo, however, the sharp increase in there are a wide range of other options (such as wind automobile and bus traffic over the past decade has Part II: Case Studies 55 led to strong indications of increasing deterioration evitable that sites will need to be found on the coast of air quality. Nevertheless, based upon what we do to accommodate such thennal plants. The economic know about pattems of energy utilization certain importance of preventing environmental degradation inferences can be drawn. It is fairly certain that at in the coastal zones is well established. Foreign present the power sector contributes only marginally tourism, an important source of foreign exchange, is to air pollution in Sri Lanka. However, this is ex- largely focused on the country's sandy beaches and pected to change significantly once the anticipated coastal estuaries and lagoons. The marine fishery coal burning power plants are added to the system industry provides employment to some 100,000 per- beginning in the late 1990's. sons, and is the largest source of animal protein for Acid rain is likely to become an increasingly Sri Lanka. The main environmental issue concerns important environmental issue in the Asia-Pacific the discharge of heated effluents into waters of the region given the fact that the energy plans in many coastal zone, where there exist numerous ecosystems countries, in particular India and China, call for rapid that are extremely sensitive to temperature increases, development of fossil energy systems. Acid rain is including 1) coral reefs, sea grass beds, benthic corn- largely a long-range phenomenon, and it is fairly munities, mangrove stands, rocky and other shores, obvious that the extent to which acid rain is or will 2) zooplankton and phytoplankton communities be experienced in Sri Lanka is as much a function of which are free floating, and 3) nursery grounds for emission trends of acid rain precursors in India as in fish and prawns. Sri Lanka itself. Global warming and transnational acid rain are Methodology conceptually different from local environmental im- pacts, since in the former case the impacts will occur Multi-criteria analysis has been developed expressly predominantly in other countries. If the main eco- for situations where decisions must be made taking nomnic objective is to maximize welfare in Sri Lanka, into consideration more than one objective which decision-makers in Sri Lanka would be unwilling to cannot be reduced to a single dimension. Its central incur additional costs if the benefits of such actions focus is the quantification, display and resolution of accrue mainly to other nations. In this case study, it trade-offs that must be made when objectives con- is assumed that Sri Lanka will be reimbursed by the flict. In the case of application to the power sector, international community for the incremental costs of there may well be strategies that have beneficial global warning mnitigation efforts, or that the Gov- impacts on both environmental and economic objec- emnment would have signed an international agree- tives-most energy efficient investments that are ment committing itself to undertake certain C02 economically justifiable also bring about a reduction emission reduction measures. in emissions and hence improve enviroDnental qual- Because Sri Lanka is a small island, which has ity as well as economic efficiency. But most options been isolated for relatively long periods, there are a require that trade-offs must be made: wind plants, for large number of endemic species. Among Asian example, potentially provide substantial environ- countries, Sri Lanka has the highest level of bio- mental benefits but are more expensive than other logical diversity. The NSF Committee on Research options. Priorities in Tropical Biology identifies Sri Lanka The overall methodology is illustrated in Figure as demanding special attention. Biological diver- 5.2, and involves the following steps: sity is under threat in Sri Lanka primarily from the * The definition of the options to be exam- progressive reduction in its natural forests and ined. other natural habitats, especially through the selec- * The selection and definition of the attrib- tive exploitation of tree species, particularly for utes, selected to reflect planning timber. Therefore, the power sector is likely to objectives. come under intense scrutiny from this perspective. * The explicit economic valuation of those As the generation mix shifts from one that is impacts for which valuation techniques predominantly hydro to one in which large baseload can be applied with confidence. The resul- fossil-fueled plants play an increasing role, it is in- 56 Part II: Case Studies Figure 5.1 Sri Lanka -~~~~~~~~~~ mm STltiil o GmmAo a~~ ~~ ~~~~~~~~~~~ t% S --2ttxit m - m uv 13hMVA= CZ_ I'_ -t132 KY tANS*M UNI SS )V KY aA Of fW BAY 132 KY C stRs OF J&V6L O STAMM MAA i \awa#t CtLCr A / au s~~~~~~~~~~~~~~~~Tmvn CANYON U Wn 7Sx - S., .......A..dh p Gd. \ HX X~~~~~~~~~~~~~~XMDW& OCEAN;. 2O *AI Part II: Case Studies 57 tant values are then added to the system The main set of policy options examined, be- costs to define the overall cost attribute. yond variations in the mix of hydro and thermal * The quantification of those attributes for plants, included (i) demand side management (using which explicit economic valuation is in- the illustrative example of compact fluorescent light- appropriate, but for which suitable ing); (ii) renewable energy options (using the illus- quantitative impact scales can be defined. trative technology of wind generation); (iii) * The translation of attribute value levels improvements in system efficiency (using more am- into value functions (known as "scaling"). bitious targets for transmission and distribution * The display of the trade-off space, to fa- losses than the base case assumption of 12% by cilitate understanding of the trade-offs to 1997); (iv) clean coal technology (using pressurized be made in decisionmaking. fluidized bed combustion (PFBC) in a combined * The definition of a candidate list of op- cycle mode as the illustrative technology); and (v) tions for further study: this also involves pollution control technology options (illustrated by the important step of eliminating from a variety of fuel switching and pollution control further consideration options that are options such as using imported low sulfur oil for clearly inferior. diesels, and fitting coal buming power plants with In some applications it may be appropriate to flue gas desulfurization (FGD) systems). add two further steps: the definition of weights for Attribute selection: Great care needs to be ex- each attribute, and the application of an amalgama- ercised in criteria or attribute selection-they should tion rule to provide a single overall ranking of op- reflect issues of national (as opposed to local project tions. However, the Sri Lanka case study did not level) significance, and ought to be limited in num- follow this approach. ber. There is little gain from a proliferation of attrib- utes. Increasing the number of attributes is not a Application to the Sri Lanka Power Sector substitute for assigning proper weights to environ- mental attributes in the decision process. On the Policy option definition: A variety of options were contrary, the more attributes considered the more selected for study,including a whole range of siting, complex the analysis, and the higher the probability pollution control mitigation and technology options. that the results will be hard to interpret and decision- Indeed, it is very important that as few a priori makers will not find the exercise useful. It often judgements as possible are made about the "feasibil- occurs that, in a desire to be comprehensive, there is ity" or "practicality" of options, because for the an inclusion of all possible impacts, making it more analysis to be useful, meaningful trade-offs must be difficult to demonstrate trade-offs, and possibly in- examined. For example, in the case of the Trinco- troducing biases through a reluctance to assign low malee coal fired power plant (on the north-east weighting to attributes. coast-see Figure 5.1), the environmental impact The following environmental criteria or attrib- assessment prepared in the mid 1980s considered utes were used in the study. To capture the potential only a very narrow range of options: all altematives impact on global warming, C02 emissions were de- studied involved sites on Trincomalee bay, with fined as the appropriate proxy. To be sure, the rela- once-through cooling to a shallow bay inlet. Other tionship between global C02 concentrations and the south coast sites had been eliminated earlier on actual physical impacts that may follow, such as sea grounds of high cost (because these sites could not level rise or changes in monsoonal rainfall pattens accommodate large coal transport vessels, resulting are still poorly understood, and in any event unlikely in higher transport costs). Yet the additional costs of to be captured by simple linear correlations. How- an evaporative cooling system, or of an outfall sys- ever, since Sri Lanka's contribution to worldwide tem that would discharge heated effluents to the emissions will remain extremely small, the assump- deeper parts of the Trincomalee Bay, proved to be tion of linearity of impacts (relative to global CO2 less than the incremental coal transport costs to a site emissions) is not unreasonable. on the south coast. To capture health impacts, use was made of the population-weighted increment in fine particulates 58 Part II: Case Studies Figure 5.2 ENVIRONMENTAL DEFINE SITING ISSUES TECHNOLOGY, ISSUES ~~AND MITIGATION SCENARIOS gANO I F) DEFINE omADDITIONAL,.*- SCENARIOS DEFINE APPLY ENVIRONMENTAL N YNM ES ATTRIBUTES ALUATIQONS MULTI-ATTRIBUTE __ _ __ _m ANALYSIS ECONOMIC ATTRIBUTE mNALYSX (SYSTEM NPV) ECONOMIC POLICY ANALYSIS ITRADE-OFF CURVE.S and NOx attributable to each source. To this end a high. It should be noted that what is captured in this simple Gaussian plume model was applied to all of attribute is the separate and purely political objective the major sites, calculated incremental ambient con- of employment creation, rather than the strictly eco- centrations for lkn square cells to within a 20km nomic benefits that would be captured by the use of radius, and multiplied by the population in each cell. shadow wage rates appropriate to reflect high unem- To capture other potential air pollution impacts, ployment in the construction cost estimates. All of such as acid rain, S02 and N0x emissions were used. these impacts were appropriately discoun1i and ex- As an illustrative social impact, the study used the pressed as a present value. Finally, to capture the creation of labor opportunities. Employment crea- potential biodiversity impacts, a probabilistic index tion is an important objective of national policy, and was derived (as discussed below). in Sri Lanka there has occurred frequent discussion Attribute quantificatdon: The problems of of the need for employment creation in the south quantification are well illustrated in the case of the where youth unemploymenlt rates are especially biodiversity attribute. At the planning level,, detailed Part 11: Case Studies 59 site specific information at the potential power plant Some Illustrative Results sites is unlikely to be available. Consequently the only quantification that appears possible is to derive With options and attribute definitions in hand, the a probabilistic index that gives the decision-maker case study then generated the multi-dimensional infornation about the likelihood that the detailed trade-off space. Using the ENVIROPLAN model, envirorunental impact statement will reveal the pres- the attribute values for each of the environmental ence of an endemic species, significantly impact attributes, and the cost attribute (for which average ecosystems of high biological diversity, or affect a incremental cost over a 20-year planning horizon habitat already in a marginal condition. was used) were calculated for every option, with There are a number of practical problems in results displayed as a series of two-dimensional deriving an appropriate index. The first is that the trade-off curves. In a final step, the list of candidate value of the area lost is a function of what remains plans for further study was then derived by examin- of the habitat. For example, the loss of the last hectare ing dominance relationships among all criteria si- of an ecosystem would be unacceptable, whereas the multaneously. loss of one hectare if 1,000 hectares remain would be Figure 5.3 illustrates a typical trade-off curve, much less. Second, ecosystems may require a mini- in this case for health impacts. The "best" solutions mum area for long term survival, which implies that are those that lie closest to the origin, and the so- the value function would need to tend to inflnity as called trade-off curve, defined by the set of "non-in- it approaches that minimum value. Perhaps even ferior" solutions, represents the set of options that are more importantly, the argument is sometimes made superior, regardless of the weights assigned to the that the value to be ascribed to the loss of habitat different objectives. For example, on this curve, the associated with some regulatory or govermmental option defined as "iresid", which calls for the use of decision depends on whether it remains secure. The low sulfur imported fuel oil at diesel plants is better details of the biodiversity index derived in the case on both the cost and the environmental objective than study are discussed further in Annex 4. the use of flue gas desulfurization systems (identified Someimpacts, however, resistdirectquantifica- as the point "FGD"). tion, even in terms of the sort of probabilistic scale A quite different trade-off curve was derived for derived for biodiversity. For example, the quantifi- biodiversity, and on Figure 5.4 is illustrated as the cation of potential damages to aquatic ecosystems trade-off between biodiversity index value and aver- from thermal discharges is extremely problematic, in age incremental cost. Most of the options have an large part because of the difficulties in extrapolating index value that falls in the range of 50-100: the no from one ecosystem to the other. The general effects hydro option has an essentially zero value, because of thermal discharges into coastal waters are of the thermal projects that replace hydro plants in this course well known. Discharges into the well-mixed, option tend to lie at sites of poor biodiversity value surface layer would usually have the general ten- (either close to load centers or on the coast). For dency to repel fish. On the other hand, if the dis- example,while wind plants would require rather charge is below the thermocline, thermal discharges large land area, the vegetation of the area on the south would have a generally beneficial effect, as the up- coast has relatively low biodiversity value, and there- welling effect caused by plume buoyancy brings fore the overall increase in biodiversity impact of this nutrients to the layers near the surface. However, option is small. Thus, the best options (or non-infe- attaching specific numerical estimates to the values rior curve) include the no hydro option, and run-of- of this general function is essentially impossible. river hydro options that require essentially zero What can be done as a generic calculation that can inundation. Note the extreme outlier at the top right be used to compare different sites is to begin with a hand corner, which is the Kukule hydro dam-it has definition of what is considered to constitute an a biodiversity loss index (B = 530) that is an order of acceptable environmental risk; for example, say a magnitude larger than for other options (B = 50 to temperature increase of no more than 1 °C at the 70). surface. The surface area over which this criterion is The case study drew several useful conclusions. exceeded is then calculated as a function of the The first four listed below are of a methodological cooling system design proposed. nature, and deal with the extent to which multi-attrib- ute methods are potentially effective in assisting 60 Part II: Case Studies Figure 5.3 frade-Off Curve of kWh Costs Versus Health Impacts Index 5.9 5.8 u +kukuleHD 5.7 no DSM 5.6 5.5 lowScoal+iresid 5.4 - u +FGD U noCoal E a) 'a 5.3 - owScoal 51 resid m Mawella>D 5.2 a ~~~T&D>11 5.1 5 \ +wind 4.9 + noHydro . ............. basecase(with DSM) 4.8+PC .+UO2001 4.7 _ I ; . I 2 3 4 5 6 7 NPV(health Impacts,2010,10%) decision-makers. The remaining ones deal with the Third, the case study indicated that certain op- substantive policy recommendations whose focus is tions were in fact clearly inferior, or clearly superior, to ensure that environmental considerations are ap- to all other options when one examines all impacts propriately incorporated in the planning process. simultaneously. For example, the high dam version First, the results of the case study indicate that of the Kukule hydro project can be safely excluded those impacts for which valuation techniques are from all further consideration as a result of poor relatively straightforward and well-established- performance on all attribute scales (including the such as valuing the opportunity costs of lost produc- economic one). On the other hand, implexmaentation tion from inundated land, or, estimating the benefits of certain demand side management measures domi- of establishing fisheries in a reservoir-tend to be nates all other options; i.e., they yield positive gains quite small in comparison to overall system costs, in terms of economic and enviromnental criteria. and their inclusion into the benefit-cost analysis does Fourth, the results indicate that it is possible to not materially change results. derive attribute scales that can be useful proxies for Second, even in the case where explicit valu- impacts that may be difficult to value. For example, ation may be difficult, such as in the case of mortality use of the population-weighted incremental ambient and morbidity effects of air pollution, implicit valu- air pollution scale as a proxy for health impacts ation based on analysis of ihe trade-off curve can permitted a number of important conclusions that are provide important guidance to decisionmakers. For independent of the specific economic value assigned example, the study determined that the value of to health effects. Thus, the study clearly demon- human life necessary to justify flue gas desulfuri- strated that if the health effects of pollutants associ- zation at potential sites for coal fired-power plants ated with fossil fuel combustion (particularly fine was on the order of $1.5 million. This is at least particulates and NOx) are to be considered, then the one (if not two) orders of magnitude greater than most effective strategy for reducing the overall popu- what would be needed to justify the installation of lation dose is to install tighter pollution controls at modem diagnostic equipment at the regional hos- oil-burning power plants located in or near urban pitals. Part I: Case Studies 61 Figure 5.4 Trade-Off Curve of kWh Costs Versus Biodiversity Index 5.8 - +kukule D 5.7 * no DSM 5.6 - 5.5 - 5.4 - +FGD - > ? c E 5.3 iowScoal 0) T5 2 iresid 5.2 - T&D>11 a 5.1 - U +wind 5- 4.9 _ oHydro Trade-off curve +ku 4.8 +UO2001 4.7 ¢ I l l l 0 100 200 300 400 500 600 Blodiverslty idex areas, rather than installing FGD systems at the more fluidized bed combustion (PFBC) technology for remote sites suitable for coal-burning power plants. coal power, (iv) replacement of some coal-fired Finally, with respect to the practical implica- power plants (on the South coast) by diesel units; and tions for planning, the study came to a series of (v) the need to re-examine cooling system options for specific recommendations on priority options, in- coal plants. cluding (i) the need to systematically examine de- mand side management options, especially Notes fluorescent lighting; (ii) the need to examine whether the present transmission & distribution loss reduc- 1. This case study is derived from Kramer, Munasinghe, tion target of 12% ought to be further reduced; (iii) Sharm, et al.. 1992. ... . ~~~~~~2. This case study is derived frmn Meier and Mus- the need to examine the possibilities of pressurized i 92.T 6. Conclusions One essential step towards achieving economically developing country is required rather than further efficient management of natural resources and for- theoretical development. Such case study work can mulating a practical strategy for sustainable develop- be most effectively carried out as part of project ment, is the effective incorporation of environmental preparation. A major purpose in this endeavor is not concerns into decisionmaking. Traditionally, the eco- to provide fine-tuned numbers but to indicate orders nomnic analysis of projects and policies (including the of magnitude. Some alternatives can be ruled out and techniques of shadow pricing), has been developed gross environmental errors avoided in ti.s fashion. to help a country make more efficient use of scarce Also, one can often identify the key environmental resources. "External effects," especially those indicators to which the decision is sensitive and focus arising from adverse environrnental consequences, attention on them. often have been neglected. Some modest evidence exists that the valuation This report has reviewed concepts and tech- techniques for determining use values may be ap- niques for valuation of environmental impacts that plied successfully in appropriate cases. However, enable such environmental considerations to be ex- examples involving the estimation of nonuse values plicitly considered in the conventional cost-benefit are virtually nonexistent in the developing world, calculus used in economic decisionmaking. Even and rather scarce even in the industrialized nations. rough qualitative assessments early on in the project The use of multiobjective decision methods also evaluation cycle may facilitate the process of inter- needs to be explored in greater depth, as an alterna- nalizing these environmental externalities. The ad- tive to more purely economic valuation methods. vantages include early exclusion of environmentally Nevertheless, developing counties are attempt- unsound altematives; more effective in-depth con- ing increasingly to both improve and make use of sideration of environmentally preferable altema- economic techniques to value environmenital assets. tives; and opportunities for redesigning projects and While the academic literature usually focuses mainly policies in order to achieve sustainable development on the development of the techniques, there are also goals. sector- or topic-related approaches.14 For practitio- Certain specific shortcomings and difficulties ners, the important concern is to keep up with and associated with the case studies were discussed ear- make use of the advances most relevant to itheir own lier. More generally, greater application of the envi- areas of application. To facilitate this, an extensive ronmental valuation concepts and techniques bibliography is included at the end of this chapter. presented in this paper to practical problems in a Bibliography Anderson, D., The Economics of Afforestation - A counting - Framework for an SNA Satellite Case Study in Africa, The World Bank Occa- System", Review of Income and Wealth, 1989. sional Paper No. 1/New Series, The Johns Hop- Bator, F.J., "General Equilibrium, Welfare and Allo- kins University Press, Baltimore, Md., 1987. cation", American Economic Review, March Attaviroj, P., "Soil Erosion and Land Degradation in 1957, pp. 22-59. the NorthemThai Uplands", in J.A. Dixon, D.E. Baumol, W.J. and W.E. Oates, The Theory of Envi- James and P.B. Sherman (Eds.), Dryland Man- ronmental Policy, (2nd ed.), Cambridge Univer- agement: Economic Case Studies, Earthscan sity Press, Cambridge, 1988. Publications Ltd., London, 1990. Binswanger, H., "Brazilian Policies that Encourage Aylward, B., The Economic Value of Ecosystems: 3 Deforestation in the Amazon", ENV Depart- - Biological Diversity, LEEC Gatekeeper Series ment Working Paper No. 16, World Bank, No. GK91-03, IIED/UCL London Environ- Washington, D.C., April 1989. mental Economics Centre, June 1991. Birdsall, N. and D. Wheeler, "Openness Reduces Aylward, B., and E.B. Barbier, "Valuing Environ- Industrial PollutioninLatin America: The Miss- mental Functions in Developing Countries", ing Pollution Haven Effect", paper presented at Biodiversity and Conservation, 1992, vol. 1, pp the Symposium on International Trade and the 34-50. Environment, International Economics Depart- Balik, M. and R. Mendelsohn, "Assessing the Eco- ment, World Bank, Washington D.C., Novem- nomic Value of Traditional Medicines from ber 1991. Tropical Rain Forests", Conservation Biology, Bishop, R.C., "Option Value: An Exposition and Vol. 6, No. 1, March 1991. Extension", Land Economics, Vol. 58, No.1, Barbier, E.B., The Economic Value of Ecosystems: I February 1982, pp. 1-15. - Tropical Wetlands, LEEC Gatekeeper Series Bishop, J. and J. Allen, The On-Site Costs of Soil No. GK89-02, August 1989. Erosion in Mali, Environment Department , The Economic Value of Ecosystems: 2 - Working Paper 21, the World Bank, Washing- Tropical Forests, L-EC Gatekeeper Series No. ton, D.C., 1989. GK91-01, January 1991. Bockstael, N.E., K.E. McConnell and I. Strand, , The Economics of Controlling Degrada- "Recreation", in Braden, J.B., and Kolstad, tion: Rehabilitating Gum Arabic Systems in Su- C.D., eds., op. cit., Chapter 8, p. 227-270. dan, LEEC Discussion Paper No. DP90-03, Bojo, J., The Economics of Land Degradation: The- IIED/UCL London Environmental Economics ory and Applications to Lesotho, Dissertation Centre, June 1990. for the Doctor's Degree in Economics, Stock- Barbier, E.B., W. M. Adams and K. Kimmage, Eco- hohm School of Economics, 1991, pp.259 - 350. nomic Valuation of Wetland Benefits: The Hade- , "Benefit-Cost Analysis of the Farm Im- jia-Jama'are Floodplain, Nigeria, LEEC provement with Soil Conservation Project in Discussion Paper No. DP91-02, April 1991. Maphutseng, Mohale's Hoek District, Leso- Barde, J.P., and D. W. Pearce (eds.) Valuing the tho", in J.A. Dixon, D.E. James and P.B. Sher- Environment: Six Case Studies, Earthscan Pub. man (eds.), 1990, op. cit. Ltd., London, June 1991. , Economic Analysis of Agricultural Devel- Bartelmus, P., E. Lutz, and S. Schweinfest, Inte- opment Projects. A Case Study from Lesotho, grated Environmental and Economic Account- EFI Research Report, Stockholm School of ing: A Case Study for Papua New Guinea, Economics, 1990. Envirornent Working Paper No. 54, World Bojo, J., K.G. Maler, and L. Unemo, Environment Bank, Washington D.C., July 1992. and Development: An Economic Approach, Bartelmus, P., C. Stahmer, and J. van Tongeren, Kluwer, Durdrecht, 1992. "Integrated Environmental and Economic Ac- 64 Bibliography Boyd, R., W. Cruz and WF. Hyde, "Unintended Costanza, R., (ed.), Ecological Economics: The Sci- Policy Spilovers on Upland Resources and the ence and Management of Sustainability, Co- Environment", Ohio University, unpublished lumbia University Press, New York, 1991. manuscript, 1990. Cropper, M.L. and W.E. Oates, "Environmental Eco- Braat, L.C. and J.B. Opschoor, "Risks in the Bot- nomics: A Survey", Journal of Economic Lit- swana Range-Cattle System", in J.A. Dixon, erature, Vol. XXX, June 1992, pp. 675-740. D.E. James and P.B. Sherman (eds.), 1990, op. Cruz, W. and M. Munasinghe, Economic Policy and cit the Environment: A Review of Bank Work, En- Braden, LB. and C.D. Kolstad (eds.), Measuring the vironment Department, World Bank, Vlashing- Demand for Environmental Quality, Elsevier, ton D.C., 1992. New York, 1991, Chapters 2 and 10. Cruz, W. and R. Repetto, The Environmental Effects Bromley, D. W., "Entitlements, Missing Markets, of Stabilization and StructuralAdjustment Pro- and Environmental Uncertainty", Journal of grams: The Philippines Case, World Resources Environrnental Economics and Management, Institute, Washington D.C., September 1992. vol. 17,1181 - 194, 1989. Dasgupta, P. and G. Heal, Economic Theory and Brookshire, D.S., L.S. Eubanks and A. Randall, "Es- Exhaustible Resources, Cambridge University timating Option Prices and Existence Values for Press, Cambridge, 1979. Wildlife Resources", Land Economics, Vol. 59, Dasgupta, P. and K.G. Maler, "The Environment and No. 1, February 1983. Emerging Development Issues", proceedings of Brown Jr., G. and W. Henry, Trhe Economic Value of the World Bank Annual Conference on Devel- Elephants, London Environmental Economics opment Economics 1990, Beijer Reprint Series Centre, Discussion Paper 89-12, 1989. No. 1, Beijer International Institute of Ecologi- Brown, K. and D.W. Pearce, 'The Economic Value cal Economics, Stockholm, 1991. of Non-Market Benefits, of Tropical Forests: Dasgupta, P., S. Marglin and A. K. Sen, Guidelines Carbon Storage", in J. Weiss, The Economics of for Project Evaluation, UNIDO, New ProjectAppraisal and theEnviromnent, Edward York, 1972. Elgar, London, 1993. Dixon, J., R. Carpenter, L. Fallon, P. Shermaan, and Capistrano, D. and C. F. Kiker, "Global Economic S. Manipomoke, Economic Analysis of t!he En- Ifluences on Tropical Closed Broadleaved vironmental Impacts of Development Projects, ForestDepletion, 1967- 1985", paperpresented published by Earthscan in association with the at the International Society for Ecological Eco- Asian Development Bank, London, 1988. nomics Conference, May 1990, World Bank, Dixon, J., and M.M. Hufschmidt (eds.), Economic Washington D.C. Valuation Techniques for the Envirounent, Coase, R., 'The Problem of Social Cost", Journal of Johns Hopkins University Press for the East- Law and Economics, October 1960. West Center,1986. Cocklin, C., "Mathematical Programming and Re- Dixon, J.A., D.E. James and PB. Sherman, The sources Planning I: The limitations of Tradi- Economics of Dryland Management, Earthscan, tional Optimization", Journal of Environmental London, 1989. Management, vol. 28, 1989, pp. 127-141. Dixon, J.A. and P.B. Sherman, Economics of Pro- Committee of International Development Institu- tected Areas, Island Press, Washington D.C., tions on the Environment (CIDIE), 'The Pre- 1990. sent State of Environmental and Resource Durojaiye,B.O., and A.E. Ikpi. 'TheMonetary Value Accounting and its Potential Application in De- of Recreational Facilities in a Developing Econ- veloping Countries", prepared for Environ- omy: A Case Study of Three Centers in Nige- mental and Natural Resource Accounting ria." Natural Resources Journal 28(2):315-328, Workshop, UNEPHeadquaters, Nairobi, 22-26 1988. February 1992. Energy Journal, Special Issue on Electricity Reli- ability, Vol. 9, December 1988. Bibliography 65 Evenson, R., "Genetic Resources:Assessing Eco- Institute and University of California, River- nomic Value", in J. Vincent et al., Valuing Envi- side, July 1992. ronmental Benefits in Developing Economies, Goodland, R., H. Daly and S. El-Serafy, Environ- Michigan State University, East Lansing, 1992. mentally Sustainable Economic Development Finney, C.E. and S. Westem, "An Economic Analysis Building on Brundtland, ENV 46, Environment of Environmental Protection and Management: Department, The World Bank, Washington, An Example from the Philippines", in The En- D.C., 1991. vironmentalist, Vol. 6, No. 1, 1986. Grandstaff, S., and J.A. Dixon, "Evaluation of Fischer, S. and V. Thomas, "Policies for Economic Lumpinee Park in Bangkok, Thailand." in John Development", American Journal of Agricul- A. Dixon and M.M. Hufschmidt (eds.), Eco- tural Economics, Vol. 72, No. 3, 1990. nomic Valuation Techniques for the Environ- Fisher, A. C., and M. Hanemann, "Quasi Option ment: A Case Study Workbook, Johns Hopkins Value: Some Misconceptions Dispelled", Jour- University Press, Baltimore, Md., 1986. nal of Environmental Economics and Manage- Hamilton, L. S., J. A. Dixon and G. 0. Miller, "Man- ment, 14, 1987, pp. 183 - 190. grove Forests: An Undervalued Resource of the Fleming,W.M.,"PhewaTalCatchmentManagement Land and of the Sea," in E. M. Borgese, N. Program: Benefits and Costs of Forestry and Ginsburg and J. Morgan (eds.), Ocean Yearbook Soil Conservation in Nepal", in L.S. Hamilton No. 8, Univ. of Chicago Press, Chicago, II., (ed.), Forest and Watershed Development and 1989. Conservation in Asia and the Pacific, Westview Hanemann, W. M., "Willingness to Pay and Willing- Press, Boulder, CO, 1983. ness to Accept:How Much Can They Differ?", Folke, Carl and AnnMari Jansson, The Emergence of American Economic Review, vol. 81, no.3, June an Ecological Economics Paradigm: Examples 1991. from Fisheries and Agriculture, Beijer Reprint Harberger, A.C., Project Evaluation: Collected Pa- Series No. 9, Beijer International Institute of pers, University of Chicago Press, 1976. Ecological Economics, Stockholm, 1992. Harrington, W., A.J. Krupnick and W.O. Spofford, Fredriksson, P. and A. Persson, The Manicaland Jr., 'The Economic Losses of Waterbome Dis- Health, Water, and Sanitation Program in Zim- ease Outbreak", Journal of Urban Economics, babwe -A Social Cost Benefit Analysis, Depart- Vol. 25, No. 1. 1989, pp. 116-137. ment of International Economics and Hartunian, N.S., C.N. Smart, and M.S. Thompson, Geography, Stockholm School of Economics, The Incidence and Economic Costs of Major 1989. Health Impairments, D.C. Heath, Lexington, Freeman, A.M., The Measurement of Environmental Mass, 1981. Resource Values, World Resources Institute, Hodgson, G., and J.A. Dixon, Logging Versus Fish- Washington D.C., 1993. eries andTourism in Palawan, Paper No.7, East Friedman, J.W., Game Theory with Applications to West Center, Honolulu, HI, 1988. Economics, Oxford, Oxford University Press, Holmberg, G., "An Economic Evaluation of Soil 1986. Conservation in Kitui District, Kenya", in J.A. Gigengack, A.R. et al., "Global Modelling of Dry- Dixon, D.E. James and P.B. Sherman (eds.), land Degradation", in J.A. Dixon, D.E. James 1990, op. cit. and P.B. Shennan (eds.), 1990, op. cit. Hyde, W.F. et al,"Forest Economics and Policy Gilbert, A., "Natural Resource Accounting: A Case Analysis: An Overview", World Bank Discus- Study of Botswana", in J.A. Dixon, D.E. James sion Paper No. 134, Washington D.C., 1991. and P.B. Sherman (eds.), 1990, op. cit. Jimenez, E., "The Magnitudes and Determinants of Githinji, M. and C. Perrings, "Social and Ecological Home Improvement in Self-Help Housing: Ma- Sustainability in the Use of Biotic Resources in nila's Tondo Project", Land Economics, Vol.59, sub-Saharan Africa: Rural Institutions and De- No. 1 (1983), pp. 70-83. cisionmaking in Kenya and Botswana", Beijer Johnson, D.G., World Agriculture in Disarray, Macmnillan Press Ltd, London, 1973. 66 Bibliography Kahn, J.R. and J.A. McDonald, 'Third World Debt University of Maryland, College Park, MD, De- and Tropical Deforestation", prepared by Oak cember 1991. Ridge National Laboratory, Tennessee, for the Luenberger, D., Introduction to Linear and Von-lin- U.S. Department of Energy, December 1991. ear Programming, Addison-Wesley, Reading Kim, S. and J.A. Dixon, "Economic Valuation of MA, 1973. Environmental Quality Aspects of Upland Ag- Lutz, E. (ed), Toward Improved Accounting for the ricultural Projects in Korea", in J.A. Dixon and Environment, World Bank, Washington D.C., M.M. Hufschmidt (eds.), 1986, op. cit. 1993. Kolstad, C.D. and J.B. Braden, "Environmental De- Lutz, E. and M. Munasinghe, "Accounting for the mandTheory", Chapter 2 inBraden and Kolstad Envirornment," Finance and Development, Vol. (eds.) 1991, Measuring rhe Demand for Envi- 28, March 1991, pp. 19-21. ronmental Quality, Elsevier, New York. Lutz, E. and M. Young, "Integration of Environ- Knetsch, J.L. and J.A. Sinden, "Willingness to Pay mental Concems into Agricultural Policies of and Compensation Demanded: Experimental Industrial and Developing Countries", World Evidence of an Unexpected Disparity in Meas- Development, Vol. 20, No. 2 (1992), pip. 241- ures of Value", Quarter,!y Journal of Econom- 253. ics, August 1984. MacRae, D., Jr., and D. Whittington, "Assessing Kramer R.A., M. Munasinglie, N. Sharma, E. Mer- Preferences in Cost-Benefit Analysis: Reflec- cer, and P. Shyamsundar, "Valuation of Impacts tions on Rural Water Supply Evaluation in of Conservation and Development Projects: A Haiti." Journal of Policy Analysis and Manage- Case Study in Madagascar", paper presented at ment 7(2):246-263,1988. the Environnental Valuation Workshop, World Maler, K.G., "Economic Theory and Environmental Bank, Washington D.C., December 1992. Degradation: ASurvey of Some Problenis",Re- Krautkraemer, J., 'The Rate of Discount and the vista de Analisis Economico, Vol. 5, No. 2, pp. Preservation of Natural Environments", Natu- 7 - 17, November 1990. ral Resources Modelling,, Vol. 2, winter 1988. Maler, K.G., "Production Function Approach in De- Krueger, A., M. Schiff, and A. Valdes, The Political veloping Countries", Beijer Reprint Series No. Economy of Agricultural Pricing Policies, The 6, Beijer Institute, Stockholm, 1992. World Bank, Washington D.C., 1991. Mahar, D., "Govermment Policies and Deforestation Lewandrowski, J.K. and E. McClain, "Resource Use in Brazil's Amazon Region", a World Bank and Deforestation in Brazil", USDA Economic publication in cooperation with the World Wild- Research Service, unppublished manuscript, life Fund and the Conservation Foundation, 1990. Washington, D.C., 1989. Liebman, J., "Some Simple-Minded Observations Margulis, S., Back of the Envelope Estinuates of on the Role of Optimization in Public Systems EnvironmentalDamageCosts inMexico, Policy Decisionmaking", Interfaces, Vol. 6, 1976, pp. Research Working Paper, World Bank, Wash- 102 - 108. ington, D.C., January 1992. Little, I.M.D. and J.A. Mirrlees, Project Appraisal Markandya, A., TheEconomicAppraisal of Projects: and Planning for Developing Countries, Basic The Environmnental Dimension, Inter-American Books, New York, 1974. Development Bank, Washington D.C., July Little, I. M. D. and J. A. Mirrlees, "Project Appraisal 1991. and Planning Twenty Years On", paper pre- Markandya, A. and D.W. Pearce, "Enviromnental sented at the Annual Coniference on Develop- Considerations and the Choice of the Discount ment Economics, World Bank, April 26 and 27, Rate in Developing Countries", ENV Depart- 1990. ment Working Paper No. 3, Washington D.C., Lopez, R., 'The Environment as a Factor of Produc- May 1988. tion: The Economic Growth and Trade Policy Markandya, A. and D.W. Pearce, "Social Costs of Linkage", mimeo, Economics Department, Tobacco Smoking", British Journal of Addic- tion, Vol. 84, No. 10, 1989, pp. 1139-1150. Bibliography 67 McConnell, K.E., and J.H. Ducci, "Valuing Environ- Munasinghe, M., and J. McNealy, "Key Concepts mental Quality in Developing Countries: Two and Terminology of Sustainable Development", Case Studies." Prepared for session on "Contin- UNU Conference on Sustainability, 22-25 June gent Valuation Surveys in Developing Coun- 1992, Washington, D.C. tries," AEA/AERE Annual Meetings, Atlanta, Munasinghe, M., W. Cruz and J. Warford, "Do Georgia, 1989. Macroeconomic Policies Hurt the Environ- Meier, P. and M. Munasinghe, Incorporating Envi- ment," Finance and Development, Sept. 1993. ronmental Concerns into Power Sector Deci- Muzondo, T.R., K.M. Miranda, and A.L. Borenberg, sionmaking: A Case Study of the Sri Lankan "Public Policy and the Environment", IMF Power Sector, ENV Department, World Bank, Working Paper, International Monetary Fund, Washington D.C., December 1992. Washington D.C., June 1990. Mendelsohn, M., "Modelling the Demand for Out- Newbery, D.M.G., "Externalities: The Theory of En- door Recreation", Water Resources Research, vironrental Policy" in G.A. Hughes and G.M. Vol. 23, No 5,1987, pp. 961-7. Heal (eds.), Public Policy and the Tax System, Mercer, E. and Kramer R., "An International Travel 1980. Cost Model: Estimating the Recreational Use Norgaard, R. B., "Sustainability as Intergenerational Value of a Proposed National Park in Madagas- Equity: The Challenge to Economic Thought car". Paper presented at the Association of En- and Practice", Internal Discussion Paper, Asia vironmental and Resource Economists Regional Series, Report No. IDP-97, World Contributed Paper Session: Recreation Demand Bank, June 1991. Models Allied Social Science Association An- Odum, E. P., Ecology: The Link Between the Natu- nual Meeting, New Orleans, January 5, 1992. ral and Social Sciences, Holt-Saunders, New Mitchell, R.C., and R.T. Carson, Using Surveys to York, second edition, 1975. Value Public Goods: The Contingent Valuation Odum, H. T., "Energy, Ecology, and Economics", Method, Resources for the Future, Washington, Amnbio, vol. 2, pp. 220-227, 1973. D.C., 1989. Pearce, D.W., "An Economic Approach to Saving the Mu, X. et al., "Modelling Village Water Demand Tropical Forest" in D. Helm (ed.), Economic Behaviour: A Discrete Choice Approach", Policy Towards the Environment, Blackwell, Water Resource Research, Vol. 26, No. 4. Oxford, 1991. Munasinghe, M., Electric Power Economics, Butter- Pearce, D. W., "Economic Valuation and the Natural worths Press, London, 1990a, chapters 2,7,8,9. World", Policy Research Working Paper Munasinghe, M., Energy Analysis and Policy, But- WPS988, World Bank, Washington D.C., Octo- terworths Press, London 1990b. ber 1992. Munasinghe, M., "Managing Water Resources to Pearce, D.W., E.B. Barbier, A. Markandya, S. Bar- Avoid Environmental Degradation", ENV rett, R.K. Tuner and T. Swanson, Blueprint 2: Working Paper No. 41, World Bank, Washing- Greening the World Economy, Earthscan Publi- ton, D.C. 1990c. cations Ltd., London, 1991. Munasinghe, M., Water Supply and Environmental Pearce, D.W. and S. Puroshothaman, "Protecting Management, Westview Press, Boulder, CO, Biological Diversity: The Economic Value of 1992. Medicinal Plants", in T. Swanson, Biodiversity Munasinghe, M., and E. Lutz, "Environmental-Eco- and Botany:The Values of Medicinal Plants, nomic Analysis of Projects and Policies for Sus- Cambridge University Press, Cambridge, 1993, tainable Development", ENV Working Paper forthcoming. No. 42, World Bank, Washington, D.C., Janu- Pearce, D.W., and A. Markandya, Environmental ary, 1991. Policy Benefits: Monetary Valuation, OECD, Munasinghe, M., "Biodiversity Protection Policy: Paris, 1989. Environmental Valuation and Distribution Is- Pearce, D.W. and R.K. Turner, The Use of Benefits sues", Ambio, Vol. 21, No. 3, May 1992. Estimates in Environmental Decisionmaking, 68 Bibliography report to the Environment Directorate, OECD, Quiggin, J., "A Research Program for a Critique of Paris, July 1990. Total Valuation", University of Maryladld Aus- Pearce, D.W., and R.K. Turner, Economics of Natu- tralian National University, mimeo, 1991. ral Resources and the Environment, Harvester Randall, A., "Total and Non-Use Values", Chapter 10 Wheatsheaf, 1990. in Braden and Koistad (eds.) 1991, Measuring Pearce, D.W. and J. Warford, World Without End: the Demand for Environmental Quaility, El- Environment, Economics, and Sustainable De- sevier, New York. velopment, Oxford University Press, Oxford, Randall, A. and J.R. Stoll, "Existence Value in aTotal 1993. Valuation Framework", in Managing Air Qual- Pender, J.L., "Experimental Measurement of Dis- ity and Scenic Resources at National Pt7ks and count Rates in Rural India", Stanford Univer- WildernessAreas,WestviewPress,BoulderCO, sity, manuscript, 1992. 1983. Pender, JLL and T.S. Walker, "Experimental Meas- Ray, A., Cost Benefit Analysis, Johns Hopkins Uni- urement of Tume Preference in Rural India", versity Press, Baltimore MD, 1984. Stanford University and ICRISAT, manuscript, Reed, D., (ed.), Structural Adjustwnt and tAe Envi- 1992. ronment, Westview Press, Boulder CO, 1992. Perrings, C., "Biotic Diversity, Sustainable Develop- Repetto, R., "Economic Policy Reform for Natural ment, and Natural Capital", paper presented at Resource Conservation", ENV Department Intemational Society for Ecological Econom- Working Paper No.4, World Bank, Washington ics, Stockholm, August 1992. D.C., May 1988. Perrings, C., Ecological Sustainability and Environ- 'Repetto, R, W. Magrath, M. Wells, C. Beer;and F. mental Control, Centre for Resource and Envi- Rossini, Wasting Assets: Natural Resources in ronmental Studies, Australian National the National Income Accounts, World Re- University, 1991. sources Institute, Washington D.C., 1989. Peskin, H.M. with E. Lutz, "A. Survey of Resource Repetto, R., W. Cruz, R. Solorzano, R. de Camino, and Environmental Accounting in Industrial- R. Woodward, L. Tosi, V. Watson, A. Vazquez. ized Countries", Environment Working Paper C. Vilalobos, and J. Jimenez, Accounts Over- No. 37, World Bank, Washington D.C., 1990. due: Natural Resource Depreciation in Costa Peters, C.M., A.H. Gentry, and R.O. Mendelsohn, Rica, Tropical Science Center, San Jose, and "Valuation of an Amazoniian Rainforest", Na- World Resources Institute, Washington D.C., ture, Vol. 339, 29 June 1989, pp. 655-656. 1991. Petry, F., "Who is Afraid of Chloices? A Proposal for Romero, C. and T. Rebman, "Natural Resource Man- Multi - Criteria Analysis as a Tool for Decision- agement and the Use of Multiple Criteria Deci- making Support in Development Planning", sionmaking Techniques: A Review", European Journal of International Development, Vol. 2, Journal of Agricultural Economics, Vol. 14, 1990. pp. 209-231. 1987, pp. 61 - 89. Pezzey, J., "Sustainable Development Concepts: An Ruitenbeek, R., "The Rainforest Supply Price: A Economic Analysis", World Bank Enviromnent Tool for Evaluating Rainforest Conservation Paper No. 2, World Bank, Washington D.C., Expenditures", Ecological Economics, V0Ol. 6, 1992. No. 1, July 1992. Phantumvanit, D., "A Case Study of Water Quality Samples, K.C., J.A. Dixon, and M.M. Gowen, "In- Management in Thailand", in Y.J. Ahmad (ed.), formation Disclosure and Endangered Species Analysing the Options, UNEP Studies, 5. Per- Valuation", Land Economics, Vol. 62, NVo. 3, gamon Press Ltd., Oxford, 1982. August 1986. Phelps, C.E., "Death and Taxes: An Opportunity for Samuelson, P.A., 'The Pure Theory of Public Expen- Substitution", Journal of Health Economics, ditures", Review of Economics and Statistics, Vol. 7, 1988, pp 1-24. vol. 36,370-89. Sassone, P.G., "Shadow Pricing in CBA: Mathemati- cal Programniing and Economic Theory", The Bibliography 69 Engineering Economist, vol.22, Spring 1977, Squire, L. and H. van der Tak, Economic Analysis of pp. 219-33. Projects, Johns Hopkins University Press, Bal- Schecter, M. and S. Freeman, "Some Reflections on timore MD, 1975. the Definition and Measurement of Non-Use Statistical Office of the United Nations (UNSO), Value", Natural Resources and Environmental "SNA Handbook on Integrated Environmental Research Center and Departmentof Economics, and Economic Accounting", New York, pre- University of Haifa, June 1992 (draft). limninary draft, October 1990. Schneider, R., G. Platais, D. Rosenblatt, and M. Stocking, M., The Cost ofSoil Erosion in Zimbabwe, Webb, "Sustainability,Yield Loss and Imedia- In Terms of the Loss of Three Major Nutrients, tismo," LATEN Dissemination Note No. 1,The Working Paper 2, Soil Conservation Pro- World Bank, Washington, D.C., April 1993. gramme, Land and Water Development Divi- Schuh, G.E., "The Exchange Rate and U.S. Agricul- sion, Food and Agriculture Organization, ture", American Journal of Agricultural Eco- Rome, December 1986. nomics, Vol. 56, no. 1, 1974. Strand,J. and A. Taraldset, The Valuation ofEnviron- Schulze, W.D., D.S. Brookshire, E.G. Walther, K. mental Goods in Norway: A Contingent Valu- Kelley MacFarland, M.A. Thayer, R.L. Whit- ation Study with Multiple Bias Testing, worth, S. Ben-David, W. MaimL and J. Molenar, Department of Economics, University of Oslo, "The Economic Benefits of Preserving Visibil- March 1991. ity in the National Parklands of the Southwest", Thomas, V., "Evaluating Pollution Control: The Case Natural Resources Journal, Vol. 23, January of Sao Paulo." Chapter 13 in The Economics of 1983. Urbanization and Urban Policies in Develop- Sebastian, I., and A. Alicbusan, "Sustainable Devel- ing Countries, G.S. Tolley and V. Thomas (eds.), opment: Issues in Adjustment Lending Poli- The World Bank, Washington, D.C., 1987. cies", ENVPR Working Paper No. 1989 - 6, Tobias, D. and R. Mendelsohn, "Valuing Ecotourism October 1989, World Bank. in aTropical Rain-Forest Reserve",Ambio, Vol. Ser6a da Motta, R. and C. E. F. Young, Natural 20, No. 2, April 1991. Resources and National Accounts: Sustainable Tumer, R.K., "Wetland Conservation: Economics Income from Mineral Extraction in Brazil, and Ethics," in Economics, Growth and Sustain- IPEA, Rio de Janeiro, 1991. able Environments: Essays in Memory of Rich- Smith, V.K., "Nonmarket Valuation of Environ- ard Lecomber, D. Collard, D.W. Pearce and D. mental Resources: An Interpretive Appraisal", Ulph, (eds.), 121-159. Macmillan, London, North Carolina State University, Raleigh, July 1988. 1992. UNEP, IUCN, andWWF. Caringfor the Earth. 1991. Smith, V.K., "Option Value: A Conceptual Over- Uzawa, H., "'Time Preference and the Penrose Effect view", Southern Economic Journal, vol. 49,3, in a Two-Class Model of Economic Growth," J. pp. 654-658, 1983. Political Economy, Vol. 77, 1969, pp. 628-652. Society for Promotion of Wastelands Development, van Pelt, M.J.F., Sustainable Development and Pro- "Economic and Social Change in a Small Rural ject Appraisal in Developing Couries, paper Community in the Degraded Lower Shivalik presented at the 31st European Congress of the Hill Range in North India", in J.A. Dixon, D.E. Regional Science Association, Lisbon, Portu- James and P.B. Sherman (eds.), 1990, op. cit. gal, 27-30 August, 1991. Society for Promotion of Wastelands Development, vanITngeren, J., S. Schweinfest, E, Lutz, M. Gomez "Dryland Management Options in Wastelands Luna, and F. Guillen Martin, "Integrated Envi- Development: Jawaja Block, Rajasthan (In- ronmental and Economic Accounting: A Case dia)", in J.A. Dixon, D.E. James and P.B. Sher- Study for Mexico", Environment Working Pa- man (eds.), 1990, op cit. per No. 50, World Bank, Washington D.C., De- Solow, R., "On the Intergenerational Allocation of cember 1991. Natural Resources", Scandinavian Journal of Vmcent, J.R., E. W. Crawford and J.P. Hoehn, eds., Economics, volume 88, no. 1, pp. 141-149. Valuing Enviromnental Benefits in Developing 70 Bibliography Countries, proceedings of a seminar held Feb- Services in Developing Countries: A Case Study ruary-May 1990, at Michigan State University, of the Use of Contingent Valuation Surveys in East Lansing, MI, Special Report 29, September Southern Haiti", Economic Development and 1991. Cultural Change, Vol. 38, No. 2, January 1990. Walsh, R. G., J.B. Loomis, and R.A. Gillman, "Valu- Whittington, D. et al., "Calculating the Value of Time ing Option, Existence and Bequest Demands for Spent Collecting Water:Some Estimates for Wilderness", Land Economics, Vol. 60, No. 1, Ukunda, Kenya", World Development, Vol. 18, February 1984. No. 2, 1990. Wang, G., W. Wu and F. Li, "Economic Evaluation Whittington, D., A. Okorafor, A. Okore, and A. of Land Levelling in Weibei Dry Upland, McPhail, "Strategy for Cost Recovery in the Shaanxi, China", in J.A. Dixon, D.E. James and Rural Water Sector: A Case Study of Nsukka P.B. Sherman (eds.), 1990, op. cit. District, Anambra State, Nigeria." Draft. Infra- Wang, G. and L. Han, "Economic Evaluation of structure and Urban Development Depnrtment, Dryland Peanut Growing with Perforated Plas- The World Bank, Washington, D.C., 1989. tic Mulch", in J.A. Dixon, D.E. James and P.B. Willis, K.G. and G.D. Garrod, "An Individual Sherman (eds.), 1990, op. cit. Travel-Cost Method of Evaluating Forest Rec- Warford, J. and A. Schwab, "Issues in Adjustment reation", Journal of Agricultural Economics, Lending Policies", Environment Policy and Re- Vol. 42, No. 1, pp. 33-42, January 1991 . search Division, WorldBank, WashingtonD.C., World Commission on Environment and Develop- draft, 1992. ment, Our Common Future, Oxford University Whittington, D., D.T. Lauria, A.M. Wright, K. Choe, Press, Oxford, 1987. J.A. Hughes, and V. Swarna, Household De- World Bank, World Deveopment Report 1992: De- mandfor Improved Sanitation Services: A Case velopment and the Environment, Oxford Uni- Study of Kumasi, Ghana, UNDP-World Bank versity Press, 1992. Water and Sanitation Program, Report No. 3, World Bank, Adjustment Lending: An Evaluation of Washington, D.C., May 1992. Ten Years of Experience, Policy and Research Whittington, D., V. K. Smith, A. Okorafor, A. Okore, Series No. 1, Washington D.C., 1989. J. Leng Liu, and A. McPhail, "Giving Respon- World Bank, "Brazil: An Analysis of Environmental dents Time to Think in Contingent Valuation Problems in the Amazon", Report No. 9104 - Studies:A Developing C'ountry Application", DR, Vol. 2, Annex 5, Washington D.C., 1992. Journal of Environmental Economics and Man- Yagci, F. et al, Structural Adjustment Lending -An agement, vol. 22, pp. 205-225, 1992. Evaluation of Program Design, Staff Working Whittington, D., J. Briscoe,, X. Mu, and W. Barron, Paper No. 735, the World Bank, Washington "Estimating the Willingness to Pay for Water D.C., 1985. ANNEX 1. USING SHADOW PRICES The estimation and use of shadow prices is facilitated other domestic or foreign users, the border-priced by dividing economic resources into tradeable and marginal social benefit (MSB) of this foregone do- nontradeable items. Tradeables and nontradeables mestic consumption or of reduced export eamings, are treated differently. The values of directly im- would be a more appropriate measure of social costs. ported or exported goods and services are already The socially optimal level of total consumption known in border prices, that is, their foreign ex- for the given input (Qopt) would lie at the point where change costs converted at the official exchange rate. the curves of MSC and MSB intersect. Price and Locally purchased items whose values are known nonprice distortions lead to nonoptimal levels of only in terms of domestic market prices, however, consumption Q Qopt characterized by differences must be converted to border prices by multiplying between MSB and MSC. More generally, a weighted the former prices by appropriate conversion factors average of MSC and MSB should be used if both (CFs). effects are present. The MSB would tend to dominate Border (Shadow) Price = Conversion Factor x in a short-run, supply constrained situation; the MSC Domestic (Market) Price would be more important in the longer run, when expansion of output is possible. The MSC of nontradeable goods and services BP = CF x DP from many sectois can be determined through appro- priate decomposition. For example, suppose one peso-worth of the output of the construction sector For those tradeables with infinite elasticities- (valued in domestic prices) is broken down succes- of world supply for imports, and of world demand sively into components. This would include capital, for exports-the cost, insurance, and freight (C.I.F.) labor, materials, and so on, which are valued at pesos border price for imports and the free-on-board Cl, C2,... Cn in border prices. Since the conversion (F.O.B.) border price for exports may be used (with factor of any good is defined as the ratio of the border a suitable adjustment for the marketing margin). If price to the domestic price, the construction conver- the relevant elasticities are finite, then the change in sion factor equals: import costs or export revenues, as well as any shifts in other domestic consumption or production levels n or in income transfers, should be considered. The CCF = , C free trade assumption is not required to justify the use = of border prices since domestic price distortions are adjusted by netting out all taxes, duties, and subsi- The standard conversion factor (SCF) may be dies. used with nontradeables that are not important A nontradeable is conventionally defined as a enough to merit individual attention or that lack commodity whose domestic supply price lies be- sufficient data. The SCF is equal to the official ex- tween the F.OB. export price and CI.F. inport price. change rate (OER) divided by the more familiar Items that are not traded at the margin because of shadow exchange rate (SER), appropriately defined. prohibitive trade barriers, such as bans or rigid quo- Using the SCF to convert domestic priced values into tas, are also included within this category. If the border price equivalents is conceptually the inverse increased demand for a given nontradeable good or of the traditional practice of multiplying foreign cur- service is met by the expansion of domestic supply rency costs by the SER (instead of the OER) to or imports, the associated border-priced marginal convert foreign exchange to the domestic price social cost (MSC) of this increased supply is the equivalent. The standard conversion factor may be relevant resource cost. If the incremental demand for approximated by the ratio of the official exchange the nontradeable results in decreased consumption of 72 Annex 1: Using Shaaow Prices rate to the free trade exchange rate (FTIER), when the where (m) and (u) are the foregone marginal output country is moving toward a freer trade regime: and overhead costs of labor in domestic prices, and a and c are corresponding conversion factors to con- OER eX + nM vert these values into border prices. SCF = ER If we are interested only in efficiency pricing, FTER eX (1 - t) + nM (1 + tin) then we may stop here. However, if social pricing is important, consider the effect of these changes on where X = F.O.B. value of exports, M = C.I.F. value consumption patterns. Suppose a worker receives a of imports, e = elasticity of domestic supply of ex- wage Wn in a new job and that the income foregone ports, n = elasticity of domestic demand for imports, is WO, both in domestic prices. Note that W'n may not tx = average tax rate on exports (negative for sub- necessarily be equal to the marginal product fore- sidy), and tm = average tax rate on imports. gone En It could be assumed, quite plausibly, that Illustrative of important tradeable inputs used in low-income workers consume the entire increase in many development projects are capital goods and income (Wn - Wo). Then this increase in consump- petroleum-based fuels. Soime countries may have tion will result in a resource cost to the economy of other fuels available, such as natural gas or coal b(Wn - Wo). The increased consumption also pro- deposits. If no clear-cut export market exists for these vides a benefit given by w(Wn - Wo), where (w) indigenous energy resouroes, then they cannot be represents the MSB, in border prices, of increasing treated like tradeables. If there is no altemative use domestic-priced private sector consumption by one for such fuels, an appropriate economic value is the unit. Therefore, MSC of the production or extraction of gas or coal plus a markup for the discounted value of future consumption foregone (or "user cost"). If another SWR = a.m + c.u + (b - w)(Wn - WO) high value use exists for these fuels, the opportunity costs of not using the resources in altemative uses should be considered as their economic value. The symbol b represents the MSC to the econ- Two important nontradeable primary factor in- omy, resulting from the use of the increased income. puts are labor and land, the next subjects for discus- For example, if all the new income is consumed, then sion. The foregone output of workers used in the b is the relevant consumption conversion factor or energy sector is the dominant component of the resource cost (in units of the numeraire) of making shadow wage rate (SWR). Consider a typical case of available to consumers one unit worth (in domestic unskilled labor in a labor surplus country-for ex- prices) of the marginal basket of (n) goods that they ample, rural workers employed for dam construc- would purchase. In this case tion. Complications arise in estimating the opportunity cost of labor, because the original rural n income earned may not reflect the marginal product b = Igi.CF1 of agricultural labor. Furthermore, for every new job created, more than one rural worker may give up former employment. Allowance must also be made where gi is the proportion or share of the i'th good in for seasonal activities such as harvesting, and over- the marginal consumption basket and CaF is the head costs like transport expenses. Based on the corresponding conversion factor. foregoing, the efficiency shadow wage rate (ESWR) The corresponding MSB of increased consump- is given by: tion may be decomposed further; w = dlv, where U/v is the value (in units of the numeraire) of a one-unit increase in domestic-priced consumption aoruiing to ESWR = a.m + c.u someone at the average level of consumption (ca). Therefore, v may be roughly thought of as the pre- mium attached to public savings, compared to "av- erage" private consumption. Under certain Annex 1: Using Shadow Prices 73 simplifying assumptions, b = I/v. If MU(c) denotes assets (see main text). Examples might be the flood- the marginal utility of consumption at some level c, ing of virgin jungle because of a hydroelectric dam then d = MU(c)/MU(ca). Assumng that the marginal that would involve the loss of valuable timber, or utility of consumption is diminishing, d would be spoilage of a recreational area that has commercial greater than unity for "poor" consumers with c < ca, potential. and vice versa. The shadow price of capital is usually reflected A simple form of marginal utility function could in the discount rate or accounting rate of interest be MU(c) = c-n. (ARI), which is defined as the rate of decline in the value of the numeraire over time. Although there has been much discussion concerning the choice of an Thus, d = MU(c)/MU(ca) = (ca/c)n . appropriate discount rate, in practice the opportunity cost of capital (OCC) may be used as a proxy for the ARL in the pure efficiency price regime. The OCC Making the further assumption that the distribu- is defined as the expected value of the annual stream tion parameter n = 1, gives d = ca/c = ia/i where isfi of consumption, in border prices net of replacement, is the ratio of net incomes, which may be used as a which is yielded by the investment of one unit of proxy for the corresponding consumption ratio. public income at the margin. The consumption term (b-w) in the expression A simple formula for the social-priced ARI, for SWR disappears if, at the margin, a) society is which also includes consumption effects, is given by indifferent as to the distribution of income (or con- sumption), so that everyone's consumption has equivalent value (d=1); and b) private consumption ARI = OCC [s + (1 - s)w/b] is considered to be as socially valuable as the uncom- mitted public savings (b=lhv). The appropriate shadow value placed on land where (s) is the fraction of the yield from the original depends on its location. Usually, the market price of investment that will be saved and reinvested. urban land is a useful indicator of its economic value Usually, the rigorous estimation of shadow in domestic prices, and the application of an appro- prices is a long and complex task. Therefore, the priate conversion factor (such as the SCF) to this analyst is best advised to use whatever shadow prices domestic price, will yield the border-priced cost of have already been calculated. Alternatively, the ana- urban land inputs. Rural land that can be used in lyst would estimate a few important items such as the agriculture may be valued at its opportunity costs- standard conversion factor, opportunity cost of capi- the net benefit of foregone agricultural output. The tal, and shadow wage rate. When the data is not marginal social cost of both urban and rural land precise enough, sensitivity studies may bemadeover should reflect the value of associated environmental a range of values of such key national parameters. ANNEX 2: SUMMARIES OF ENVIRONMENTAL VALUATION CASE STUDIES1 Effect on Production Method valid for one year only. In the economic analysis, soil conservation measures are considered as the calcu- Cost-Benefit Analysis of Land Improvement lation covers a five year period. in Lesotho2 Crop sampling was carried out in cooperation with the FISC staff during five seasons, 1986-1990. The Farm Improvement with Soil Conservation The results show that farmers using a "high-input" (FISC) project was initiated in 1985 in Mohale's management do receive higher yields on average, but Hoek district in southern Lesotho, and is gradually that very substantial variations of yields make this a being expanded. The project is now used as a model risky investment. for a national training program in soil conservation. Financial calculations for maize and sorghum Other soil conservation projects are already using show that the yield must increase by 125 percent and FISC as a model. The choice of the FISC project for 144 percent, respectively, in order to achieve a real study is further justified by its modem approach in rate of return of 10 percent. These target increases dealing with land degradation: production orienta- in yield are significantly higher than the average tion, labor-intensive techniques, and popular partici- values achieved in the areas under high-input man- pation. Furthermore, information for research could agement. The two options actually yield negative be obtained at low cost. The project area is fairly real marginal IRRs of 21 and 30 percent:, respec- typical for lowland Lesotho where most of the crop tively. production takes place. Wilh some adjustments the Maintained participation in the project appears calculations could be used for other areas in Lesotho limited after the initial boost when consenration ef- or even for other areas with similar geographic and forts result in in-kind payments. Most likely, project socio-economic features in other countries. sales are merely replacing alternative, less accessible The overriding aim of the FISC project is to sources of supply. There are no convincing signs of raise agricultural productioni. It has rehabilitated old a major transformation of the crop management re- terrace structures, constructed new ones and added gime. The long-ternn impact of physical conservation other structures for conservation; promoted hybrid works may be thne only net impact as far as the major maize, hybrid sorghum, and fodder grasses; and grain crops are concerned. Financial budgets for fruit planted thousands of tree seedlings. It has also pro- and fuelwood trees show more promising retums, how- moted rotational grazing on communal rangeland. ever, and have also met with greater interest among The project area covers almost 26,000 hectares and local people. reaches about 22,000 people. Possible explanations for the lack of farmer response to the promotion of high-inputrmanagement Finacial Analysis are discussed, including, among other things, land tenure, credit for agricultural investments, and risk The financial analysis was done from a household pertaining to agricultural investments. Most serious perspective, using market prices. Two management is the problem of risk. Crop yields are very unreliable options for cultivation of maize and sorghum have in Lesotho. The fanners are quite aware of this and been compared in financial terms. The "high-input" will (informally) calculate the chances of losing in- alternative implies the use of commercial fertilizer vested resources. Demands for yield increases for and hybrid seed. The "traditional" alternative implies maize and sorghum have been shown to be consid- no use of fertilizer and locally saved seeds instead of erable in order to reach an acceptable level of finan- hybrids. The less immediate impact of soil conserva- cial return (10 percent real rate). It is certainly not tion measures is left out of this calculation, which is Anne-x 2: Summaries of Environmental Evaluation 75 Case Studies irrational of the farmer to adopt a careful approach * Livestock products by promoting im- in the face of these risks. proved grazing management. Additional benefit items to consider are: Econoniic Analysis * Training of personnel, and the introduction of improved communal management, with The economic analysis considers the perspective of potential extra-project impacts the Lesotho Government and the donor agency, the * Off-site physical impacts, such as less sil- Swedish International Development Authority tation of dams, less maintenance costs for (SIDA). Overall project performance has been re- roads and bridges, improved water qual- corded for the period March 1985 to December 1990. ity, etc Current, firm plans for work until mid 1992 have * Secondary benefits for the community at been incorporated, and extrapolations have been large as a result of the increase in income made from past performance in relation to the avail- from agriculture. ability of future financial means. Quantifcation of costs and benefits. Not all The analysis distinguishes between productiv- costs are readily available in monetary terms. Exam- ity impacts due to increaseduse of commercial inputs ples are the temporary loss of soil from new, un- and to improved soil conservation. The distribution, grassed terraces, "loss" of land to terraces and other sale, and use of commercial fertilizer and hybrid seed structures, and increased maintenance for roads due in the project areas has been monitored. In the short to greater use. Labor cost for soil conservation is not term, distribution of these inputs has increased some- included since it was reported to be negligible. This what as they are used as in-kind payments for con- view is, however, controversial. Monetary cost data servation work on individually controlled land. were taken from project accounts and complemented Convincing signs are lacking, however, for a lasting by the executing company's data for costs paid by the impact in terms of commercial sales or their use. donor agency directly to the company. Farmers are known to save project distributed inputs The crop benefits due to hybrid seeds and fertil- for several years, and the level of use in the project izer use is assumed to be negligible. Only small areas is not significantly different from use in non- quantities of the inputs have been distributed through project areas. Therefore, the project cannot be cred- the project and the financial analysis showed their ited with a rise in productivity due to increased use use to be questionable. For those that wanted the of inputs. inputs they were available through other channels. Identification of costs and benefts. Financial To determine the benefits from soil conserva- costs are identified through project and donor ac- tion a number of factors need to be considered. First counts. There are also costs of soil conservation is the issue of whether soil loss actually affects crop works. These are borne by the farmers and have to production at all in this particular area. Second, if it be estimated separately. Financial costs have to be does, to what extent soil loss occurs, and third, how adjusted in several ways to arrive at real economic this rate affects productivity. costs, however. Potential benefits of the project in- There is reason to believe that soil loss immedi- clude increasing production of ately affects the average crop production area since * Maize, sorghum, and crop residues due to the average topsoil depth is estimated to be 25 centi- the use of fertilizer, hybrid seed and con- meters, a level at which the water retention capacity servation of soil and nutrients is reduced. Research in the area indicates that the * Fruits (peaches and apples) annual soil loss is roughly 15 tons per hectare (t/ha) * Fuelwood from pine and other tree on poorly managed soils and 5 tlha for the areas species under project-influenced conservation management. * Fodder grasses such as Eragrostis and Through comparisons with other studies on loss of Bana grass productivity due to soil loss, a 1 percent annual * Vegetables from communal gardens spon- decline in yield on nonconserved land was assumed. sored by the project The impact of soil conservation on crop produc- tion can be expressed as 76 Annex 2: Summaries of Environmental Evaluation Case Studies IQit = dYit x AYi x LACt x PI x CSi financial wage. The project payment of M 3.5 per day for unskilled labor is shadow priced at M 2.5 which is the generally accepted local wage for daily where IQ = incremental production (kg); laborers. dY = relative crop decline avoided due to conserva- Local financial prices were used for the crops. tion; Based on a comparison of the protein and energy AY= the base level of average (14-year) yieldfor the content of maize residues as compared to five sub- district in kilograms per hectare; stitutes, the approximate price of M 60 per ton of LAC = accumulated area under conservation man- maize crop residues was derived. The same figure agement (hectare); will be applied to sorghum residues. PI = project impact: the share of AH affected by the The value of fuelwood is based on ihe values project's actions (percent); and calorific content of its substitutes, displayed in CS = the share of maize an(d sorghum respectively of Table A2. 1. the cultivated land (percent); The actual substitution value will vary depend- t = time index (year 1 ... T); and ing on the household's situation. A rough, weighted i = crop index. average could be calculated as follows: (a) 5 percent will substitute dungr (0.25 * The assumption here is that PI (project impact) M 100.61) equals 1, as conservation activities are assumed to be (b) 5 percent will substitute brushwood (0.75 * nil in the absence of the project. 3 M 24.33) Up to 1992, more than 18,000 seedlings of apple Weighted average: M 43.40 and peach trees were to be delivered to farmers in the As for the period after 1992, a number of as- project area. The survival rate was estimated to be 50 sumptions have been made. The level of costs is percent. The number of remaining fuelwood trees by assurmed to remain at the average for 1985-1992, but 1992 is estimated atroughly 1]30,000. Thefodderbene- with all costs associated with expatriate (nondomes- fits are rather small due to the small areas planted and tic) services, such as consultant fees, back-stopping the opportunity cost of the land used by the executing company, external evaluation mis- Road and bridge maintenance in the area will sions, etc., taken out. This implies a level of 54 not be significantly affected by the improved land percent of the previous average. management and soil retention, and no major dams As for benefits, it is assumed that the project for hydropower or irrigation in the project areas will continues to add newly conserved crop land at a be affected. performance rate of 50 percent of the previous level. Valuation in economic prices. The Loti (pl. However, from the accumulated area under conser- Maloti) is fixed on a par with the internationally vation management reached the previous year, a convertible South African Rand. There is no black decline rate of 2 percent per year is applied. Another market for Rands or Maloti.. This indicates that the assumption is that there will be no lasting imapact on distorting impact of foreign exchange restrictions is not significant. The official exchange rates have therefore been used, based on International Mone- Table A2.1 Economic Value of Fuelwood tary Fund data. In 1990 the exchange rate was set to Substitutes per Cubic Meter roughly 3 Maloti per US$. Since there is a transfer of income to Lesotho of M/kg M about 20 percent of the import value, inclusive of MJ/kg Eq. kg (1987) (1990) duties, the value of importeid components is multi- plied by a factor of 0.8 when going from financial to Brushwood 16.0 550 .03 24.33 economiic prices.Dug19 68 0 001 Skilled and semi-skilled labor has access to the Dung 12.9 682 .10 100.61 large South African labor market and is priced at the Crop residues 14.0 629 .02 18.54 Annex 2: Summaries of Environmental Evaluation 77 Case Studies the level of use of fertilizer and hybrid seed. Further- Table A2.2 Discounted Share of Total Benefit more, it is assumed that distribution of fruit trees continues, but declines to a level of 50 percent of Benefit Shares project maximum. For fuelwood it is assumed that Item 1 percent 10 percent tree planting declines to a level of 50 percent of project average achievement in 1986-1991. For fod- Sorghm 30 25 der grasses, the assumption in the base case wil be Maize 28 23 that fodder growing stabilizes at 50 percent of the Fruit 29 38 level achieved in 1986-1991. Fuelwood 11 12 Results and Sensitivity Analyses Fodder 1 2 Total 100 100 Given these assumptions, the results of the base case are an NPV of minus M 7.0 million at a discount rate of 10 percent and minus M 5.6 million at 1 percent. The internal rate of return is minus 0.8 percent. reflecting the higher marginal utilities of the recipi- The qualitative interpretation is that the project ents as a group as opposed to the donors as a group. makes a loss that is significant in comparison to the This weight, however, should then be consistently resources invested. The present value of costs is only applied to alternative projects. M 9.9 million at 10 percent and M 33.6 million at 1 percent. In terms of the overriding target for the Discussion project as defined by the donor-to raise agricultural production among smallholder farming house- In the quantification of costs, the labor requirement holds-the project cannot be shown to be successful for maintenance of conservation structures was omit- when the benefits are related to the costs. The table ted. However, if a higher cost estimate is accepted, below summarizes the benefit categories, which pro- the size of conservation benefits change drastically. vide some overview of the relative importance of It can be shown that the break-even point for maize various benefit items. Each benefit category has been is 15 person-days, given a time horizon of 50 years discounted by 1 and 10 percent respectively. and a discount rate of 10 percent. Given a discount Table A2.2 shows that fruit is a significant bene- rate of 1 percent, the maximum labor input is raised fit item that deserves more thorough monitoring in to 34 days, before the net present value of conserva- the future. This, and conservation benefits, make up tion benefits approximates zero. Thus, it is possible the bulk of benefits, and deserve closer scrutiny in that this benefit item has been considerably overesti- terms of sensitivity testing. mated. Only empirical measurement in project areas The robustness of this base case result is tested could resolve this issue. using sensitivity analysis of alternative assumptions An additional point is option value. Lesotho is with regard to: discount rate, post-1992 project per- extremely dependent upon migrant labor remit- formance, fruit income, erosion impact on crop tances. Ahigh price scenario is designed to build into yields, future grain prices, and distributional the calculation the possibility of a substantial in- weights. crease in relative prices as a proxy for the somewhat The qualitative impression of the sensitivity unlikely, but not impossible, event that a mass of analysis is that, if the rate of 10 percent is used, the migrant workers are forced to return to Lesotho. base case result remains robust, although the size of While the nominal price level in fact may not be the deficit is changed. If the lower rate of 1 percent changed considerably due to the extent of the market, is considered acceptable as a standard, however, the the real price would increase for Basotho farmers as base case result could be qualitatively altered by incomes would fall. The value of the project is thus several factors, mnaldng the project perform better correlated to the size of the GNP of Lesotho. From than expected. With the lower rate, the project could both a macro and a micro perspective, the soil also be justified using a heavy distributional weight conservation program is an insurance against hard 78 Annex 2: Summaries of Environmental Evaluation Ccase Studies times.Thisleavesthedecisionmakerswithapartially This has given a strong incentive for destructive quantified problem: is the present value of the op- logging and widespread forest clearing. tions higher than the negative NPV of the stream of This valuation was based on a systematic bo- costs and benefits that have been valued? tanical inventory of 1 hectare of Peruvian rainforest However, soil consenration is not the only pos- along the Rio Nanay near the small village of sible insurance against declines in migrant labor Mishana, 30 kilometers southwest of the city of incomes. Lesotho needs food security, not necessar- Iquitos. Annual precipitation in the region averages ily more domestic production of grain. Establishing 3,700 millimeters; soils are predominantly infertile an economic capacity to buy grain on the world white sands. The inhabitants of Mishana are indige- market through other development projects nmay be nous people who make their living practicing shifting a more efficient alternative.. Lesotho has a compara- cultivation, fishing, and collecting a wide variety of tive advantage in inexpensive labor, not in good forest products to sell in the Iquitos market. agricultural land and a beneficial climate. Screen- ing available development project options for their Method and Data profitability is therefore a useful exercise. Further- more, the capacity to ensure food security is de- The inventory showed 50 families, 275 species and pendent on the size of the population, a neglected 842 trees of at least 10.0 centimeters in diameter. Of matter which needs urgenit attention in Lesotho. the total number of trees on the site, 72 species (26.2 While nonagricultural investments may be percent) and 350 individuals (41.6 percent) yield more efficient from a macroeconomic point of view, products with an actual market value in Iquitos. the majority of the population, and the poorest part, Annual production rates for fruit trees and palms live in the rural areas. If the primary value of a project were either measured from sub-samples or estimated like FISC is not so much to raise production imme- from interviews with collectors. Latex yields were diately but rather to protect to a significant degree the taken from the literature. The merchantable volume land base in the long term, more substantial subsidi- of each timber tree was calculated using published zation of cover crops such as fodder grasses at the regression equations relating diameter to commer- expense of traditional cropping could be justified. cial height. Even if fodder could not be commercially sold, the Average retail prices for forest fruits were col- grower would provide an insurance service while lected in monthly market surveys. The officially protecting the land for future potential uses. The controlled rubber prices were used. Four inde- economic viability of this option needs to be worked pendent sawmill operators were interviewed to de- out. termine the mill price of each timber species. The This study cannot conclusively provide an an- labor investment associated with fruit collection and swer to the future value of the FISC approach, but latex tapping was estimated in person days per year has given some reasons why massive, full-scale rep- based on interviews and direct observation of local lication involving a substantial number of expatriate collecting techniques. The harvest cost was based on personnel should be avoided. Continued efforts the minimum wage rate, US$2.50 per day.5]3ased on should be subject to close moritoring of their efficiency earlier studies the transport cost for fruit and latex in order to justify any further funding. The original was estimated at 30 percent of total market value report also contains a discussion of the income distribu- while extraction cost for timber was set at 40 percent tional impacts of theproject, which is not analyzedhere of total value. due to lack of space and direct relevance to the topic of environmental valuation. Results Valuation of an Amazonian Rainforest The market value of the fruit production in the sam- ple area was almost US$650 per year. Annual rubber Most financial appraisals of tropical forests have yields amount to about US$50. Deducting collection focused exclusively on timber resources and have and transportation costs gives net annual revenues from ignored the market benefits of non-wood products. fruits and latex of US$400 and US$22, res$ectively. Annex 2: Summaries of Environmental Evaluation 79 Case Studies The net present value (NPV) of this production, at 5 market benefits if the appropriate resources are ex- percent discount rate and assuming that 25 percent ploited and properly managed. of the frut crop is left in the forest for regeneration, is estimated at US$6,330. Effect on Health Method The hectare of forest also contains 93.8 cubic meters of merchantable timber. If liquidated in one Economic Analysis of a Water Supply and Health felling, this sawtimber would generate a net revenue Program in Zimbabwe6 of US$1,000 on delivery to the sawmill. A logging operation of this intensity, however, would damage The purpose of the paper is to evaluate the Manica- much of the residual stand and greatly reduce, if not land Health, Water and Sanitation program in Zim- eliminate, future revenues from fruit and latex trees. babwe utilizing social cost-benefit analysis. Data The net financial gains from timber extraction would was gathered on cost-benefit analysis, water and be reduced to zero if as few as 18 trees were damaged sanitation projects, and health statistics in Zim- by logging. babwe. Because the health statistics do not fully Periodic selective cutting would yield a maxi- reflect the mortality and morbidity rates, benefits of mum of about 30 cubic meters per hectare every 20 the proposed water and sanitation program will be years. With a weighted average price of US$17.21 underestimated. per cubic meter and deducting harvest and transport Only communal lands in rural Manicaland are costs, the net revenue is about US$310 at each cutting studied. No consideration is given to resettlement cycle. The net present value would be US$490. areas or commercial farm areas. The study considers The combined NPV of fruit, latex and selective only two major benefits-health improvements and cutting would be about US$6,820, with logging con- change in consumer surplus for water. tributing about 7 percent of the total. Timber man- agement appears to be a marginal financial option Background in this forest, especially considering the possible impact of logging on fruit and latex trees. The Manicaland Province, organized in 7 districts, is situated in the east of Zimbabwe. It is characterized Compaisons by a relatively high altitude and a diverse relief, implying varying patterns of rainfall, temperature, T-he NPV of this piece of rainforest compares well soils, and natural farmiing regions, and has the high- with other uses of rainforests. Using the same dis- est rainfall in the country. The province is agricultur- count rate, 5 percent, the NPV of the timber and ally rich and produces forestry, fruit, maize, pulpwood obtained from a I hectare plantation of groundnuts, sunflowers, tea, coffee, cotton, dairy, Gmelina arborea in Brazilian Amazonia is estimated and beef products. Approximately 90% of the pro- at US$3,184, or less than half that of the forest. vincial population of 1.2734 million (1987) lives in Similarly, gross revenues from fully stocked cattle rural areas and 66% lives in communal lands. Over pastures in Brazil are reported to be US$148 per 50% of the population is under 15 years of age. Over hectare per year. This gives a present value of 65% of the economically active population in Mani- US$2,960. Deducting the costs of weeding, fencing, caland is working in agriculture, either as subsistence and animal care would lower this figure significantly. farmers or as permnanent or seasonal laborers on Both these estimates are based on the optimistic commnercial farms. assumption that plantationforestry and grazing lands The objective of the program, within the are sustainable land-use practices in the tropics. framework of Swedish and Norwegian support to Tropical forests perform vital ecological serv- Zimbabwe's health sector, is to improve living ices, they are the repository for an incredible diver- conditions in the communal areas of Manicaland sity of germplasm, and their scientific value is through: immeasurable. The results from this study indicate * Improving existing and constructing that tropical forests can also generate substantial new water supplies that ensure an ac- ceptable quantity and quality of water 80 Annex 2: Summaries of Environmental Evaluation Case Studies for domestic use and that are reliable and shadow wage meets the market wage. The opportu- accessibleforthe community; nity cost of children's time is set at zero. * Improving sanitation conditions by con- Social rate of discount. Using Helmer's ap- structing latrines; proach (taking the rate of return in the private sector), * Giving health education to improve hygi- one would derive a real social discount rate of 4.86%. enic practices and instigate behavioral If the World Bank discount rate of 10% were used, changes. adjusted for foreign exchange rate changes, the real discount rate would be 7.24%. There is no official Shadow Pflces Used discount rate for government projects in Zimnbabwe, so the authors conducted a sensitivity analysis with Domestic prices were used as a numeraire. That is, the social discount rate ranging from 2% to 4%. local currency is expressed in Zimbabwean dollars Time horizon. Atime horizon of 40 years is used (Z$) and foreign currency is expressed in $US. in the analysis, given that benefits are expected to Foreign exchange rate. Harberger's formula is remain as a result of reinvestment. utilized to calculate the shadow price of foreign exchange. Since Zimbabwe ist a pricetaker, the sup- Calculation of Costs ply and demand elasticities of foreign exchange can be replaced by the import demrand and export supply The materials supplied by the project are valued elasticities. Assuming that the export supply elastic- at market prices. Community input, except for the ity is set to zero, and that no quantitative restrictions constructor's payment in case of latrines, is valued at exist, the Harberger 's formula reduces to the opportunity cost of labor. Since the shadow wage of unskilled labor has been set at zero during the dry P = n M (1 + T)R = (1 + T) R season (when construction takes place), the costs of n M community input except for the constructor "s pay- ments are zero. The constructor is paid the s,hadow where R' = the shadow price of foreign exchange and price of skilled labor: $Z50 for a double latrine, and R the official exchange rate; M = c.i.f. value of $Z250 for a multi-compartment latrine. imports in terms of foreign currency; T = import duties; and n = elasticity of demand for foreign Calculation of Benefits exchange. Given the assumption that the export supply Change in consumer surplus. The price of elasticity equals zero, the H'arberger's formula now water is calculated by mean kCals of energy used in approaches the standard UNIDO guidelines ap- walking to and carrying water, the energy cost per proach. Using 1987 data, the authors determine that $Z1, andthe time cost with shadow wages set alL 50%, one extra unit of foreign exchange can buy goods 75%, and 100%. The change in consumer surplus per worth 1.1799 units on the domestic market. Because year is equal to of the existence of capital restrictions in the form of quotas in Zimbabwe, import duties are increased [ (q2-ql) 1 substantially so thatthe demand for foreignexchange 180 q(Pw -(PW2) + 2 (PWI -Pv2) equals the supply. Therefore a shadow price of for- eign exchange of 1.75 is used. (q2 - ql) Shadow price of labor. In the dry season, there + 185 q, (Pdl - Pd2) + 2 (Pdl - PcO is underemployment in the agricultural sector, so the shadow wage for labor is set at zero. Given that uncertainty exists in the peak harvest season, when where: there is a shortage of labor, the authors conduct a q1 = quantity carried home in 1 day, before the project; sensitivity analysis, where the shadow wage is set at q2 = quantity carried home in 1 day, after the project; 100%, 75%, and 50% of the market wage of Z$0.46 Pwi = price/i in the wet season, before the project; per hour. For skilled laborers, it is assumed that the Pw2 = price/i in the wet season, after the project; Annex 2: Summaries of Environmental Evaluation 81 Case Studies Pdl =price/I inthedryseason,beforetheproject;pd2 Vc = -64.71/(1 + i)t + (1307.21 - 795.13)/(1 + i)t = price/I in the dry season, after the project. The change in consumer surplus per person, when full coverage is reached, is shown below. The corresponding net present value of output gained by society, for each adult saved, is: Shadow wage 50 percent 75 percent 100 percent VA = (275.00 - 129.41)/(1 + i)t; 0.99 1.18 1.37 with i = social rate of discount. Health improvements. Cost of illness estimates In Table A2.4 the values for different discount consist of treatment costs, costs of lost production, rates are shown as lower bound estimates. and costs for extra transportation. The authors point In the base case, with an estimated social dis- out that as willingness to pay is not taken into ac- count rate of 4.86%, shadow wage at 100% of the count, the values arrived at may severely underesti- market wage, and a 100% health improvement, the mate true values. internal rate of return was greater than the social For treatment costs, the costs of private treat- discount rate and the net present value of the project ment are used as the opportunity cost. A consultation was strongly positive. At a social discount rate of with a physician costs $Z10.80 and the cost of nurs- 7.24%, estimated in the alternative approach, the ing is $Z3.00 per hour. The transportation cost used project is not found profitable even if a shadow is $Z 0.50 per single joumey for all inpatients. The wage of 100% and 100% disease reduction is lower bound of the total value per year of reduced achieved. morbidity when full reduction is reached, is shown While most of the benefits come from disease in Table A2.3. reduction, it must be pointed out that benefits are The value of a life saved would be very high to probably underestimated because the value of a the individual concerned or his/her family. The saved life has been understated-through the use of authors attempt to establish a minimum value by a lower bound. Secondly, the paper does not take into adopting society's viewpoint to determine the net account other benefits such as local industry that output gained by saving a life. They use the human would benefit from improved water supply. Third, as capital approach-based on the stream of average income distribution will probably improve as a result income minus average consumption, discounted of the project, benefits could have been given a back to the time of avoided death. greater weight on social equity grounds. Therefore The average household income in communal lands is $Z550. If the household consists of 2 adults and 4.5 children on average, the income received by Table A2.3 Total value in $Z of health improve- each adult is $Z275. The authors estimate that each ments per year after the year 2005 at a social child consumes $Z64.71 per year, and that each adult discount rate of 4.86% for different values of living on conumunal lands consumes $Z129.41. disease reductions and shadow wage Assuming an urban migration rate of 20%, and an unemployment rate of 20%, the authors use the Shadow Disease reduction' weighted average of the wage adjusted for unem- wage ployment and the income in communal lands as an (percent) 40% 70% 100% estimate of the child's future production: $Z1307.21. Future consumption is estimated as final household 50 2,501,936 4,378,388 6,254,839 expenditure per capita, $Z795.13. The net present value of output gained by pre- 75 2,505,760 4,385,081 6,264,401 venting a child's death, taking into consideration 100 2,509,585 4,391,774 6,273,962 only production and consumption aspects, is: 82 Annex 2: Summaries of Environmental Evaluation Case Studies the authors conclude that the project is likely to be a Populations, densities, and illiteracy rates for each success from a societal point of view. cant6n were taken from a 1986 census. Distances were measured along the most likely roads between Travel Cost and Contingent Valuation the major population center of each cant6n and Methods MCFR. The travel cost was estimated at US$0.15 per kilometer. This includes out-of-pocket costs, a frac- The Consumer Surplus From Visits tion of fixed costs, and the value of travel time. to a Costa Rican Rainforest 7 The demand function for visits (VISYIATION RATE) was assumed to be linear and to depend on This study measures the value of domestic eco-tour- the travel cost (DISTANCE), the population density ism to the privately owned Monteverde Cloud Forest (DENSITY) and the illiteracy rate (ILLITERACY). Biological Reserve (MCFR). The MCFR straddles the continental divide in Costa Rica and consists of 10,000 hectares of rugged terrain, the vast majority (I) VISITATION RATE = ao + a1DISTANCE + of which is virgin rainforest. Tourism to the reserve a2DENSITY + a3ILLITERACY + e has increased markedly over the 18 years of its existence, both in terms of domestic and foreign visitation rates, despite the relatively remote locality where 'e' is an error tenn assumed to be independent and difficulty of accessing the site. and normally distributed. The model was eslimated using multiple regressions. The se3mi-log functional Data, Method and Results form could not be used on this data because the visitation rate from many cant6ns was zero. Two In 1988, 755 out of approximately 3,000 domestic specifications were estimated; with and without illit- visitors left their addresses at the reserve for the eracy rates. The results are presented in Table A2.5. opportunity to win wildlife photographs. The sample All coefficients have the expected sign. The showed a similar geographical distribution as a con- coefficient on price is negative and statistically sig- trol sample and is assumed to be representative of the nificant. Higher population densities result in more true domestic visitor population. trips, which is expected since people living in less Costa Rica is divided into 81 cant6ns. Each dense cant6ns probably have nearby rainforests to cant6n is treated as an observation (this is used visit. The higher the illiteracy rate the lower the instead of the concentric zones mentioned in Chapter visitation rate, which indicates that visitation is posi- 3). Visitation rates (number of visits per 100,000 tively correlated with education, and probably per- residents) were calculated for each cant6n by divid- manent income. ing observed numbers of trips ty census populations. The visitation rates predicted in table A2.5 are lower than actual rates, since they only predict the Table A2A Lower bound gain to society from visitation observed in the sample. Adjusting this for a saved life in $Z at different discount rates the whole sample (3000/755) yields an accurate per capita visitation rate. The linear demand functions Social discount rate estimated in Table A2.5 suggest that visitation would (percent) drop to zero only at distances of 328 kilometers and 347 kilometers, respectively, for the two regressions. 2 4.8 7.24 9 At the presumed US$0.15 per kilometer, this implies a maximum price per visit of US$49 and US$52, Child respectively. (< S yrs) 8,441 2,813 1,094 934 The consumer surplus for each cant6n is the (<5 yrs) 8,441 2,813 integral under the demand function (I) between the Adult actual price for this cant6n and the maximum jprice. (> 5 yrs) 1,307 1,131 1,011 489 The results are summed across all cant6ns, yielding an annual consumer surplus estimate of US$97,500 Annex 2: Summaries of Environmental Evaluation 83 Case Studies and US$116,200, respectively. Given that there are Value of Viewing Elephants on Safaris in Kenya8 about 3,000 Costa Rican visitors per year, the site is worth about US$35 per domestic visit. Assuming the Travel Cost Approach real value of this recreational flow remains the sarne over time and using a real interest rate of 4 percent, the The travel cost method can be used to estimate a present value of domestic recreation at this site is demand function for going on safari in Kenya. between US$2.4 million and US$2.9 million. The consumer surplus (CS) is the difference be- tween what people actually pay and themaximum Discussion they would be willing to pay. This net economic benefit from a safari does not show up in market This consumer surplus estimate of about observations but would be lost to the interna- US$100,000 per year does not include foreign visi- tional society if safaris were prohibited. tors. Foreign visitors outnumbered domestic visitors The analysis is based on a survey taken from by four to one in 1988. Foreign visitation is likely to samples of the approximately 80 percent of the be worth far more than domestic since foreign tour- tourists to Kenya who came either from North ists with higher incomes and lack of nearby substi- America or Europe. The other 20 percent of the tutes probably value the site more than domestic tourists were assumed to have the same average visitors. The present value estimate is probably too consumer surplus as the sample. low considering the fact that visitation has been In 1988therewere63,000visitorsfromNorth growing at 15 percent a year for the last five years. America and 350,000 from Europe. Normalizing Still, if the same figure, US$35, is used for all visitors for population differences gave 0.2316 and 0.9826 and also for the future, that would mean anet present visitors per 1,000 population for North America value of US$1,250 per hectare. The price that the and Europe, respectively. reserve currently pays to acquire new land is between The price of safari is defined as the sum of US$30 and US$100 per hectare. This suggests ex- land travel costs, air fare, and travel time costs. pansion of protected areas near this reserve is a These are summarized in Table A2.6. Land costs well-justified investment. were estimated by creating a quality weighted Finally, it should be noted that the recreational pr ice index from the tour operator s'surveys.9 The value of standing forests is but one of its potential air fare and travel time were estimated from visi- benefits. The total value of the forest includes bene- tors' surveys. Average annual income was fits from renewable harvests of many commodities, US$45,000 giving an estimated hourly wage of biological diversity, ecological services, and sites for US$22.50, which is weighted at 30 percent to scientific research. reflect that vacation time is less valued than gross wage rate. Table A2.5 Domestic Demand for Visits to Monteverde VISITATION RATE = 36.17 - 0.121 DISTANCE + 0.008 DENSITY (4.20) (2.77) (2.76) Adjusted R2 = 0.145 VISITATION RATE = 44.42 -0.107 DISTANCE +0.006 DENSITY -0.001 ILUTERACY (4.28) (2.40) (1.82) (1.40) Adjusted R2 = 0.156 2 Note: The t-statistics are in parentheses. The number of observations (cant6ns) are 81. The low R2 values are probably due to the lack of additional data. For example, information is unavailable on socioeconomic variables that mnight better explain differences in visitor travel behavior. 84 Annex 2: Summaries of Environmental Evaluation Case Studies We now have the minimum requirement to es- Results Fmm Tavel Cost Approach timate a demand curve-two observations of price- quantity. Expecting demand to be a linear function The weighted average consumer surplus is about we have: US$725. Based on discussions with tour operators and with personnel in the economic section of the U.S. Enbassy in Kenya, the number of adults going P = 4,023 - 1,674 Q on safari eachyear was estimated at between 250,000 and 300,000. This gives a total consumer swuplus for those on safari in the range of US$182-US$218 where P is the sum of land and air travel time costs, million annually, depending on the assumed level of and Q is holiday visits per :1,000 population. Note visitation. that we have yet to address the problem of the per- To identify the contribution elephants make to centage of visitors on holiday to Kenya that actually the value of a safari, tourists on safari were asked in goes on safari. the tourist survey to allocate the pleasure and enjoy- Given a linear demand curve, per person con- ment of their trip over four stipulated categories of sumer surplus is the choke pjice (the price at which experience. "Seeing, photographing, and learning demand is driven to zero) minus the actual price paid, about the wildlife" made up 50 percent of the pleas- divided by two. For North America, the consumer ure according to the answers. In a follow-up question surplus is given by: concerning only the enjoyment of the wildlife, the interviewees attributed 25 percent of their wvildlife pleasure to seeing African elephants. Applying the CS. = 0.5 * (4,023 minus 3,635) = US$194 share of 12.6 percent, attributed by the visitors to elephants, to the estimated economic value of a safari yields a viewing value for elephants of 23 to 27 For Europe it is million dollars per year. Contingent Valuation Appmach CfS. = 0.5 * (4,023 - 2,378) = US$822.50 The tourists' survey contains a series of contingent valuation questions. One of the questions (see box It seems reasonable that a safari, a once-in-a- A2.1) asks people to pay in the form of a special lifeftime adventure for most North Americans, most annual permit (or increased safari cost) of 100 dollars of whom had a vety satisfactory experience, would which would maintain the elephant population at be worth 5 percent more than it costs. It seems current levels through increased enforcement activ- plausible that a similar experience at less cost would ity. Sixty-five percent of the respondents said they be worth 35 percent more than the cost for a Euro- would pay 100 dollars. The average was 89 dollars pean. while the median was 100 dollars. Some respondents dislike translating important qualitative experiences into a dollar metric and re- Table A2.6 Land, Airfare, and Travel Tine Costs (all monetary units US$) Land travel Hourly Region cost Air fare Travel time wage Weight Time cost Total price North America 1,465 1,900 40hrs 22.50 0.30 270 3,635 Europe 957 1,300 18hrs 22.50 0.30 121 2,378 Annex 2: Summaries of Environmental Evaluation 85 Case Studies spond with a zero response. There were a substantial Although the estimates are rough, they are almost number of zero responses. However, to maintain a certainly a good guide to the order of magnitude of short questionnaire no follow-up questions were asked value. The viewing value of elephants is more likely todistinguish"protesting"respondentsfrom"genuine" 25 million dollars annually than 2.5 million or 250 zero respondents. Ib diminish the importance of the million dollars. It does not seem prudent for Kenya's zeros, the median value, 100 dollars, has been used 1988 Wildlife Management and Conservation instead of the average. budget to be under 200,000 dollars when tens of Respondents could have a strategic bias to give millions of dollars in viewing value of elephants large values if they thought the result would lead to alone are at stake. policy decisions they like but would not have to pay for. Respondents may also put in large values if they The Willingness to Payfor Water Services regard the question as a sort of referendum in which in Southern HaiO they vote, as it were, for a broader, perhaps moral issue. However, the largest response to this question In rural areas, many of those who are in the service was 500 dollars, less than 1 percent of the respon- area of new water supply systems have chosen to dent's income and about 3 percent of the cost of continue with their traditional practices. If rural his/her safari. There was therefore no "trimming" of water projects are to be both sustainable and replica- data. ble, an improved planning methodology is required Starting point bias was not tested for due to that includes a procedure for eliciting information on inadequate sample size. As to the credibility of the the value placed on different levels of service, and median value, 100 dollars, it seems modest inasmuch tariffs must be designed so that at least operation and as it is 3 percent of the total cost of a safari. If one maintenance costs (and preferably capital costs) can thinks introspectively about the value over and above be recovered. A key concept in such an improved the cost of a very satisfying moderately expensive planning methodology is that of "willingness to pay" experience, 100 dollars does not appear to be a (WTP). suspiciously high number and some think it some- Two basic theoretical approaches are available what low. for making reliable estimates of households' WTP. The first, the "indirect" approach, uses data on ob- Results Fum Content Valualion Method served water use behavior (such as quantities used, travel times to collection points, perceptions of water Combining the median value of willingness-to-pay quality) to assess the response of consumers to dif- of 100 dollars with the estimate of 250,000 to ferent characteristics of an improved water systemn. 300,000 adult safaris per year yields an annual view- The second, or "direct" approach, is simply to ask an ing value of elephants of between 25 to 30 million individual how much he or she would be willing to dollars. If the mean value of 89 dollars per person is pay for the improved water service. This survey accepted, the viewing value is decreased to between approach is termed the "contingent valuation 22 million and 27 million dollars. method" and is the focus of the case study. Note that both methods produce annual values of around 25 million dollars for viewing elephants. Box A2.1 Question 10 in survey: Special Fees and Permits Suppose that the current population of elephants can be maintained if additional foot, vehicle, and aerial pa- trols are operated on a sustained basis in the parks. If these patrols can be supported by a special 100 dollar annual permit (or included in each visitor's safari cost), are you willing to support this permit fee. [18] NO, I am not willing to pay $100 for this permit. [34] YES, I am willing to pay $100 for this permit. [ ] I am willing to pay a maximum of for this permit. 86 Annex 2: Summaries of Environmental Evaluation Cas e Studies The Study Area obtaining water from existing sources, and the education of household members, and negatively In August 1986 the research team conducted a correlated with the individual's perception of the contingent valuation survey and source observa- quality of water at the traditional source used tions in Laurent, a village in southern Haiti. At the before the construction of the improvedi water time the United States Agency for International supply system. The authors hypothesize that the Development was funding a rural water supply WTP bids of women respondents would be higher project designed to proviide services to about than those of men because women carry most of 160,000 individuals in 40 towns and villages. The the water, but alternative interpretations are cer- project was executed by CARE. The affiliation tainly possible. with CARE provided access to villages and justi- The research design attempted to test fied the presence of the tearm to the local popula- whether WTP bids are systematically related to tion. the variables suggested by economic thecry. Dif- Haiti, with two-thirds of the population at an ferent ways of posing the questions were tried. annual Per capita income of less than US$155 in The bidding game format worked better than 1980, provides a field setting similar to the situ- direct, open-ended questions. The bidding game ation in much of Africa and some parts of Asia. In was very familiar and easily understood because such poor areas an accur ate understanding of the it was similar to the ordinary kind of bargaining willingness of the population to pay for rural that goes on in local markets of rural Haiti. Tests water services is likely to be particularly impor- were also included for the existence and magni- tant for sound investment decisions. tude of several types of threats to the validity of The population of Laurent is about 1,500, the survey results, such as strategic bias, starting predominantly small farmers with a few people point bias, and hypothetical bias. having regular wage employment. Remittances Strategic bias may arise when respondents be- from relatives and friends are common. More lieve they may influence an investment or policy than 80 percent of the population is illiterate and decision by not answering the interviewer's ques- malnour:ishment among children is widespread. tions truthfully. Such strategic behavior may influ- There are seven sources of fresh water within ence answers in either of two ways. Suppose an approximately 2 kilometers of most of the popu- individual is asked how much he would be willing to lation: one protected well and six springs in dry pay to have a public standpost near his house. If river beds. The springs provide only modest he/she thinks the water agency or donor will provide amounts of water, and individuals often wait more the service if the responses from the village are than an hour to draw supplies. The average 3 positive, but that someone else will ultimately pay kilometers round trip to a water source can some- for the service, there will be an incentive to overstate times take several hours. The preferencefor clean his/her WTP. On the other hand, if the individual drinking water is strong, and people sometimes believes the water agency has already made the de- will walk considerable distances past alternative cision to install public standposts in the village and sources to collect drinking water that is consid- that the purpose of the survey is to determine the ered pure. amount people will pay for the service (in order to assess charges), the individual will have an incentive Research Design to understate the true WTP. An attempt to estimate the magnitude of the bias Economic theory suggests tthat an individual's was made by dividing the study population in two demandfor agoodisafunctionof thepriceofthe groups. One group was read an opening stalement good, prices of substitute and complementary intended to minimize strategic bias. The group was goods, the individual's income, and the individ- told that the international agency, CARE, had already ual's tastes. Maximum WTP for a new water sys- decided to build the new system and that people tem will vary from household to household and would neither have to pay CARE for the system nor should be. positively related to income, the cost of pay money at the public fountains. The second group Annex 2: Summaries of Environmental Evaluation 87 Case Studies was read another statement that was accurate but left ond consisted of a number of specific water-related more questions about the purpose of the study unan- questions. In the third section the enumerator read swered, especially concerning the role of the inter- one of the statements used to test for strategic bias view in designing a water fee. and showed the photographs of public standposts in The hypothesis was that if individuals acted other villages. The respondent was then asked to strategically, then bids from those who received the present bid per month for (a) public standposts (as- second statement would be lower than bids from the sunming no private connection) and (b) for a private first because the former would fear that a high bid connection (assuming public standposts were al- would result in a higher charge by the community ready installed). The fourth section was a series of water committee. questions on the health and education of family Starting-point bias exists if the initial price in members and the household's assets (such as radios a bidding-game affects the individual's final WTP. or kerosene larnps). The latter was used, along with This could, for example, be the case if the respondent observations about the quality of the house itself, as wanted to please the interviewer and interpreted the a substitute for expenditure questions, to forTn a initial price as a clue to the "correct" bid. To test for household wealth index. starting-point bias, three different versions of the The second part of the fieldwork consisted of questionnaire were randomly distributed, each with observing the quantities of water collected by indi- different initial prices in the bidding game. viduals at all the sources used by the population of Hypothetical bias may arise for two kinds of the village. The objective of these observations was reasons. First, the respondents may not understand to verify the information individuals provided in or correctly perceive the characteristics of the good household interviews on the sources they used and being described by the interviewer. This has been a the quantities of water collected. All sources were particular problem when the contingent valuation observed on the same day from sunrise to sunset. The method has been used to measure individuals' WTP analysis of the source observation data for Laurent for changes in environmental quality. For example, increased the confidence in the quality of the water- it may be difficult for people to perceive what a use data obtainedfrom the household interviews. Out change in sulfur dioxide or dissolved oxygen means of 119 observations of trips to water sources, the in terms of air or water quality. This bias is not likely interview responses were consistent with the source in the present case. The respondents were farniliar observation for 101 households (85 percent). with public water fountains and private water con- nections and photos of public standposts built in Analysis of Contngent Valuation Bids nearby villages were shown during the interview. The second source of hypothetical bias is the Fourteen percent of the households gave an answer possibility that the respondents do not take the ques- of "I don't know" in response to WTP question for tions seriously and will respond by giving whatever public standposts; there was a 25 percent non-re- answer first comes to mind. The test for this is the sponse rate for the WTP question for private connec- same as for the applicability of consumer demand tions. The mean for the bids for the standposts, 5.7 theory: were bids systematically related to the vari- gourdes per month, (US$1.14) seemed realistic. ables suggested by economnic theory? The test for strategic bias showed the antici- pated higher bids for those who had received the Field Procedure neutral statement, but the difference was not statisti- cally significant (t-statistics of 1.1 and 0.5, respec- Fieldwork in the village consisted of two parts: tively, for bids on standposts and private household surveys and source observation. The ma- connections). On the basis of this test, the hypothesis jority of households in Laurent were interviewed that respondents were not acting strategically when (170 questionnaires completed out of approximately they answered the WTP questions cannot be rejected. 225 households). The household interview consisted of four sections. The first dealt with basic occupa- The test for starting-point bias showed that the tional and demographic data on the family. The sec- bids did not vary systematically with the starting- 88 Annex 2: Summaries of Environmental Evaluation C'ase Studies point. The null hypothesis: that the three samples are Conclusions from the same population cannot be rejected, al- though the confidence intervals are wide. The results of this study suggest that it is possible to On the basis of these results, there was no reason do a contingent valuation survey among a very poor, to attempt to adjust the WTP bids for strategic or illiterate population and obtain reasonabile, consis- starting-point bias. The mean of WTP bids for the tent answers. The results strongly suggest that con- public standposts was 5.7 gourdes per household per tingent valuation surveys are a feasible method for month. Assumning an average annual income in estimating individuals' willingness to pay for im- Laurent of 4,000 gourdes i(US$800), the mean bid is proved water services in rural Haiti. It may also prove about 1.7 percent of household income and is signifi- to be a viable method for collecting information on cantly lower than the 5 percent rule of thumb often individuals' willingness to pay for a wide range of used in rural water supply planning for maximum public infrastructure projects and public services in "ability to pay" for public standposts. The mean of developing countries. WTP bids for private connections, 7.1 gourdes, was not much higher (2.1 percent of household income), Willingness to Payfor Improved Sanitation but these bids were based on the assumption that the in Kumasi, Ghanall public standposts were already in place. The variations in the bids for public standposts A case study was conducted in Kumasi, Ghana, to and private connections were modeled as a function determine household demand for two sanitation of the identified explanatory variables. The depend- technologies: water closets (WCs) with a piped sew- ent variable obtained from the bidding game is prob- erage system, and Kumasi ventilated improved pit ably not the maximum amount the household would latrines (KVIPs). The objective of the study is to be willing to pay but rather an interval within which demonstrate how information on household demand the true willingness to pay falls. Linear regression is for improved sanitation services can be gathered not an appropriate procedure for dealing with such utilizing a contingent valuation survey, an,d how this an ordinal dependent variable because the assump- information can be used to plan sanitation invest- tions regarding the specification of the error term in ments. the linear model will be violated. An ordered probit model was instead used to explain the variations in Study Design WTP bids. The results of the estimations can be seen in Observers were placed at a random sarnple of 30 Table A2.7. The coefficients for all the independent public latrines in neighborhoods in Kumasi, and at variables are in the direction expected. The t-statis- all 10 public latrines near the public marketplace, to tics indicate that the variables for household wealth, determine usage levels of public latrines (such usage household education, distance of the household from requires payment of a fee in some localities). Aseries the existing water source, and water quality are all of interviews was then carried out with personnel significant at the 0.05 level in both models. The sex involved in the management and operation of all 40 of the respondent was statistically significant in the public latrines in the sample. Information was gath- model for public standposts, but not in the model for ered on the cash flow position regarding the latrines, private connections. The results clearly indicate that and on the facilities themselves. Data was gathered the WTP bids are not random numbers but are sys- also on the amount of waste collected by Kumasi's tematically related to the variables suggested by eco- six desludging trucks. nomic theory. The major part of the field work was related The ordered probit model can be used to predict to the design and implementation of a large house- the number of households in a community which will hold survey, to collect information on existing use anew source if various prices were charged. Such sanitation practices and WTP for improved sanita- demand schedules are precisely the kind of informa- tion services. An initial household survey of 50 tion needed by planners and engineers to make sound households was conducted in July 1989. This was investment decisions. Annex 2: Sunmaries of Environmental Evaluation 89 Case Studies Table A2.7 Willingness-to-pay bids for public standpost and private connections Dependent variable: Probability that a household's WTP falls within a specified interval For a public standpost For a private connection Independent variables Coefficient 7-ratio Coefficient i-ratio Intercept 0.841 1.350 -0.896 -1.344 Household wealth in- 0.126 2.939 0.217 4.166 dex Household with for- 0.064 0.232 0.046 0.194 eign income (1 if yes) Occupation index -0.209 -0.848 -0.597 -2.541 (1 if farmer) Household education level .157 2.113 0.090 1.818 Distance from exist- 0.001 5.716 0.000 1.949 ing water source Quality index of ex- -0.072 -2.163 -0.099 -2.526 isting source Sex of respondent -0.104 -5.41 -0.045 -0.207 Log-likelihood -206.01 -173.56 Restricted log-likeli- -231.95 -202.48 hood Chi-square (free- 51.88 57.83 dom=7) Adjusted likelihood 0.142 0.177 ratio Degrees of freedom 137 120 followedbyapretestingquestionnaireforahundred socioeconomic characteristics, such as education, households. income, etc. A two-stage, stratified.sampling proce- The final survey questionnaire had four parts. dure was utilized to select a random sample of 1,633 The first part related to demographic characteristics. households. The household survey was caried out The second was concerned with the household's over a five-week period in October and November existing sanitation conditions, and the level of satis- 1989. The overall response rate to the survey was faction with these services. The third contained ques- very high-only 45 households refused to be inter- tions about WTP for improved sanitation facilities. viewed. Usable interviews were completed by 1,224 Fourth and finally, there were questions relating to respondents. 90 Annex 2: Summaries of Environmental Evaluation Case Studies Household preferences for improved technologies sewer connection for households currently with- out water. Each household was only asked about Surveyors asked the question: "If a WC (con- its WTP for services relevant to its specific cir- nected to a sewer system) and a KVIP each cost cumstances. the same amount per month, which one would you On average households without a WC were prefer?" Contrary to expectations that respon- willing to pay about the same for a WC or KVIP dents would stronglyprefer aWC,45% indicated ($1.43 vs. $1.47). Households with a WC were a preference for KVIPs, and only 54% preferred willing to pay $1.32 for a sewer connection. WCs. In an attempt to identify any underlying Households without water connections were will- patterns for expressed preferences, a logit model ing to pay $1.56 for a KVIP and $2.53 per month was estimated to examine the relationship be- for both a water connection and WC. Results tween household preferences and socioeconomic suggest that expenditures for water and sanita- and other characteristics of the household. The tion are not substitutable-the demand is addi- dependent variable was the household's prefer- tive. ence for either WCs or KVIPs assuming equal Three multivariate modeling approaches costs; higher values of the dependent variable were utilized to analyze the relationship of vari- indicated a preference for the KVIP over a WC. ous determinants in the WTP bids. These were Dependent variables included sex, age, education, Ordinary Least Squares (OLS), Stewart Maxi- income, and existing household sanitation facili- mum Likelihood, and Ordered Probit. Results ties. obtained from the multivariate analysis are ro- The unrestricted model used a list of 21 bust, and show that WTP information is system- independent variables to explain household pref- atically related to socioeconomic characteristics erences.Theoverallfit of thismodelwasnot good; of the household in all areas, and that r-esponses it could not be confirmed that parameter esti- obtained conformed with prior expectations and mates were not equal tci zero. Surprisingly, no consumer demand theory. linkage could be established between higher in- Household with higher incomes bid signifi- comes and/or better education and a greater pref- cantlymorefor improved services than thosewith erence for WCs. lower incomes. Owners bid more for improved The restricted model used four independent facilities than renters. Respondents with access to variables, all of which were significant at the 10% a private water connection bid more for KVIPs level in the unrestricted raodel. In this restricted than respondents without a private connection, model,thevariablerelatedtothe"time-to-think" suggesting a desire to continue improving basic effect was significant at the 5% level, which indi- services. Households in multistorybuildiings were cates that respondents who had time to consider willing to pay less for a KVIP than households their answers were more likely to prefer KVIPs. living in single-story buildings, possibly because The authors suggest that this later hesitancy may of the limited convenience of KVIPs in multistory be linked to a diminished enthusiasm for an un- dwellings. familiar system (WCs coinected to a sewer sys- tem) once respondents were able to reconsider Policy options their responses. The authors point out that information on WTP Household willingness to pay for improved sanitation can help in sanitation for improved sanitation planning. Costs of both KVIPs and WCs are higher than what households are willing topay for Respondents were asked about their WTP for five them, and therefore public subsidization would different types of services: KVIPs, WCs with sewer be necessary if improvements were to be imple- connections, sewer connections for households that mented. It is evident that conventional sewerage already had WCs, private water connections, and systems, and therefore WCs, are not affordable both a private water connection and a WC with a for the majority of the population. The authors Annex 2: Summaries of Environmental Evaluation 91 Case Studies suggest that if policy reforms were introduced in the Visitation rates or number of visits per 1,000 financial and housing markets, household demand people, (not the number of visitors, since many make for KVIPs might be significantly increased without repeat visits) were calculated for all 17 zones, and the need for subsidization. are demonstrated in Table A2.8. Information on the round-trip monetary travel cost and the average tine Applicability of CVM in Esfimaing Household for each trip was collected. Time spent traveling was Denund for Sewerage Services then changed into a monetary value using a repre- sentative wage rate and was then added to the travel The study identifies two areas in which additional cost. Aregression equation was estimated, regressing research is needed. Contingent valuation questions visitation rates against average total travel cost. Us- for alternative technologies were asked sequentially ing data on changes in admission fees, changes in in this survey. Research is required to develop cost- demand for visits with higher fees can be estimated effective ways of determining how households (Table A2.9). These results were then plotted on a would choose between multiple options presented linear demand curve, and the consumer surplus was simultaneously at varying prices. estimated at Baht 13.2 x 106. Capitalized at 10%, this Secondly, the study used the single household yielded a value of Baht 132 million, or $US 6 million, as the unit of analysis. Additional research shouldbe at the 1980 exchange rate. undertaken on how best to estimate collective will- Under the contingent valuation approach, users ingness to pay for improved facilities in situations were directly asked a hypothetical question: "If the where group housing is common-for example, in government could no longer allocate the budget to an apartment building. maintain the park but in fact was considering turning The study helps to broaden the horizons of it into other uses, what would be the maximum planners, engineers, and policy analysts, by placing amount you are willing to contribute per year to keep more of a focus on "demand-side" issues (for hous- the park?" Two sets of values were obtained from the ing, capital, and infrastructure), as opposed to purely survey-a user value, from interviews conducted at technological alternatives or government financing. the park, and values from respondents who were interviewed at their residences. Value of a ThaiNational Park2 The 187 persons interviewed for the TC study were also interviewed for the CVM study. The In an attempt to measure the value of Lumpinee Park amount per year that respondents were willing to pay in Thailand, two surveys were conducted, a demand appeared to be linked with the frequency of visits. function for the service of the park was estimated, Visitors for recreational purposes indicated a slightly and the value of the consumer's surplus associated lower WTP for yearly contributions to maintain the with the demand curve was computed to represent park than visitors for morning and evening exercise the value of the park in the year of the study (1980). purposes. However, the latter group's willingness to The Travel Cost model used was of the standard pay per visit was less than that of recreational users. form-visitation rates were assumed to be a function The authors decided to estimate the number of visi- of total travel cost, availability of substitute sites, and tors per year (2.1 million) as the number of visits, as income. The total travel cost variable included both the value of the park will depend to a large extent on out-of-pocket travel costs plus the monetary value of the proportion of the visitors using the park for time spent traveling, using a representative wage recreational and exercise purposes and their average rate. 187 people were interviewed and divided into visitation rates. 17 zones of origin within Bangkok. The initial survey User's willingness to pay, or what the authors was divided into weekday and weekendusers. It was call the user's hypothetical valuation, was estimated found that the park's use rate was heaviest on the at 13 million baht, and is therefore virtually identical weekend. An attempt was made to determine if week- to the travel cost results. day and weekend users were statistically different in An additional 225 people were interviewed terms of values, income, etc. It was found that this throughout Bangkok, including people who had was not the case. never used the park, in an attemipt to estimate the 92 Annex 2: Swnmaries of Environmental Evaluation Case Studies broader "social" (or what is now more commonly Table A2.8 Visitation Rate per 1,000 Population termed "non-use") value of the park. When CVM per Year for All Zones results were adjusted by the appropriate age-corrected population figures, the WTPmeasure of park users was Sample estimated at 13 million baht per year, and for the survey Popula- Vlisitation of Bangkokresidents, 116.6millionbabiperyear, (thus Zone tion Person Percent rate/1,000 demonstrating themorerealisticWTPof constantusers of the park). The consumer surplus and welfare gain associated with the continued existence of the park is 1 190,450 10 5.3 590 clearly significanL 2 235,647 27 14.4 1,288 Notes 3 77,112 4 2.1 583 4 131,542 23 12.3 1,965 1. Portions of this Annex are based on an original draft 5 380,416 30 16.0 886 prepared by Gunnar Kohlin. 2. This case study is derived from J. Bojo, 1991. 6 519,869 18 9.6 389 3. This is supported byobservations during fieldwork 7 523,831 37 19.8 794 in the area before project initiation. 4. This case study is derived from Peters, et al. 1989. 8 123,109 3 1.6 274 5. All prices in 1987 U.S. dollars using anexchangerate 9 479,659 4 2.1 94 of twenty infis to the dollar. 6. This case study is derived from Fredriksson and 10 201,334 4 2.1 223 Persson, 1989. 11 388,333 6 3.2 174 7. This case study is derived from Tobias and Mendel- sohn, 1991. 12 255,555 7 3.7 308 8. This case study is derived from Brown and Henry, 13 262,097 1 0.5 43 1989.13 22071 054 9. The survey contained 17 questions and was distrib- 14 140,249 4 2.1 320 uted at some lodges and given to tourists during parts of May 15 382,621 5 2.7 147 and June, 1988. There were 53 respondents. 10. This case study is derived fromn Whittington, et al., 16 204,434 1 0.5 55 11. This case study is derived from Whittington et al., 17 113,769 3 1.6 296 1992. 12. This case study is derived from Grandstaff and a Dixon, in Dixon and Hufschmidt (eds.), 1986. Total 4,610,027 187 99.6 a. Percentages do not total 100 due to rounding. Table A2.9 Visits in One Year at Various Admission Fees in Baht (B) Number of visits at various admission fees (B) Popula- Total cost Zone tion (Blvisit) 0 2 4 6 8 10 12 14 16 18 20 1 190,450 8.08 161,978 139,714 117,438 95,169 72,900 50,631 28,362 6,093 -- -- -- 2 235,647 3.72 260,483 232,929 205,375 177,821 150,268 122,714 95,160 67,606 40,052 12,499 -- 3 77,112 10.25 55,800 46,783 37,768 28,750 19,734 10,717 1,700 -- -- -- -- 4 131,542 5.04 135,254 119,873 104,492 89,111 73,730 58,349 42,968 27,587 12,206 -- -- 5 380,416 8.64 311,085 266,604 222,123 177,641 133,160 88,678 44,197 -- -- -- -- 6 519,869 10.00 383,788 323,000 262,213 201,426 140,639 79,851 19,064 -- -- -- -- 7 523,831 13.66 274,624 213,373 152.123 90,872 29,621 - -- -- -- -- -- 8 123,109 16.65 43,021 28,626 14,231 -- -- -- -- -- -- - - 9 479,659 14.18 236,884 180,798 124,713 68,627 12,541 -- - -- -- -- -- 10 201,334 15.50 83,893 60,352 36,810 13,268 - 11 388,333 20.35 51,701 6,294 -- - -- -- - -- - -- -- 12 255,555 19.52 46,424 16,543 - - 13 262,097 17.16 83,776 53,129 22,482 - - - -- - -- - -- 14 140,249 17.01 46,058 29,659 13,260 - -- - - - -- -- -- 15 382,621 18.43 93,890 49,151 4,412 - - -* - -- - -- -- 16 204,434 27.59 - - -- - - -- -- -- -- 17 113,769 21.70 6,167 - -- - - -- -- -- - Total visits per year 2,274,826 1,766,828 1,317,440 942,685 632,593 410,940 231,451 101,286 52,258 12,499 0 ANNEX 3. ESTIMATING INTERNATIONAL TRAVEL COSTS1 Most travel cost studies specify the problem in to Madagascar, relative to trips to other countries. terms of valuing single-purpose, single destination This is described in more detail below (for a rigorous day-trips to a site that provides aparticular recreation mathematical analysis, see Mercer and Kramer experience of typical quality that is interchangeable 1992). with those available at many similar sites. Recreation There are two types of goods in the model: in Madagascar's national parks is not consistent with recreation service flows and nonrecreation service these assumptions. Recreators in Madagascar can be flows. Recreational service flows are produces by divided into two groups consuming two distinct combining market travel services and timre to travel goods: a) local recreators who consume day-trips to to a country with nature tourism trip experiences in national parks to escape the city and view the local the country. Nonrecreation service flows include natural environment; and b) international nature market commodities combined with time. tourists who consume mu]ti-week trips to an exotic Nature tourism trip experiences in a country are land, to experience unusual landscapes, fauna, flora, produced by choosing a bundle of activities and and cultures. utilizing in-country market services and travel time Basic travel cost models can be applied directly to visit the location of the activities. The activities to estimate the demand of local recreators. By con- can be thought of either as single destinations within trast, the decision by foreigners to pursue nature the country, or groups of destinations. Activities in tourism in Madagascar and at a particular site such Madagascar, for example, might include travelling to as the proposed national park is much more complex the proposed Mantadia National Park to view the than the models presented in standard travel cost indiri lemurs and to a specific beach to swim and studies. Such foreign visitCrs fly great distances and sunbathe, or to a group of parks for birdwatching and utilize various modes of travel to visit numerous sites a group of beaches for snorkeling. The production and enjoy many different activities in Madagascar. function of these nonhomogeneous nature tourism Few foreign tourists plan trips solely to visit a par- trip experiences is therefore a function of the activi- ticular national park in Madagascar. More com- ties chosen, the travel time to reach the location of monly, they first decide to take a trip to Madagascar the activities, and the market goods used to travel to to view nature. Then, they decide on a specific itin- the location of the activities. erary that may include several parks, cultural sites, Once the site of the activity is reached, the etc. Therefore, modeling the demand for foreign activity is produced by combining market goods (e.g. travel to a national park in Madagascar for nature guide services, camnping equipment), environmental related recreation on the basis of the travel cost services of the site of the activity, and on-site time. method requires a significanit reformulation of tradi- Households considering international nature tourism tional models. are assumed to attempt to maximize utility. The The international travel cost method is most utility function of the individual is then solved sub- easily understood by dividing the decision process ject to the production function above and the usual into several stages. First, the household examines the full income constraint. costs associated with enjoying the various activities This model can be used to specify the total use available in each country. Using the knowledge value of international nature trips in temis of the given in the first stage, the household miniimizes the maximal amount an individual would be willing to total cost of nature tourism trip experiences defined pay for the trip and still enjoy a given level of by a certain bundle of activities. Finally, the house- satisfaction or utility. This implies that the value of a hold uses the information from stages one and two to trip is the difference between the minimurm income compare the benefits and (mininiized) costs of a trip required to reach the given initial level of utility at Annex 3: The International Travel Cost Model 95 choke price levels, minus the minimum income re- Thus, Y* is the minimum expenditure required quired to reach the initial utility level at the initial to achieve the same utility level U°, if choke prices prices. This is explained below. P*v, C*NM, and C*M prevailed for visits to Madagas- A specific application of the model is that it car and for activities within Madagascar (i.e., if these seeks to identify the tourists' willingness-to-pay for prices were raised from the base level up to values just the creation of a new national park (Mantadia). Mer- high enough to choke off demand for international cer and Kramer define a total use value (TUV) based travel and in-country activities-in Figure 3.2 in Chap- on two major components: ter 3, the point H is an example of a choke price). a) decision to visit Madagascar (versus other TUV may be broken down into components in countries); several different ways, but for their purposes, the b) decision to undertake specific recreational authors rewrite (1) as follows: activities in Madagascar. TUV = e(P*v, C*NM, C*M, ZO, U°) - With respect to item (b), a further disaggrega- e(Pov, CONM, cOM, Z°, UO) tion is made. Thus the model focuses on two activi- ties within Madagascar: = [e(P*v, C*NM, C*M, Z°, U°) - (b) (i) Recreation outside Mantadia park; e(P°v, C*NM, C*M, Z, U°)] (4) (b) (ii) Recreation in Mantadia park. We may write + [e(P°v, C*NU, C*M, Z°, U°) - e(Pov, CONM CTm, Z0, U0)] (5) TUV = Y -Y (1) + [e(P0v, CONM C*M, Z°, U°) - e(Pov, CONM, COM, Z0, UO) (6) The Y's are income levels which may be ex- Note that line (4) indicates the willingness-to- pressed in terms of corresponding expenditure func- pay for (or value of) a simple visit to Madagascar- tions e, as follows: but with no recreational activities undertaken (i.e., choke prices C*NM and C*m prevail). Line (5) relaxes the cost of recreation outside Mantadia (CON1) and Y° = e(P°V, C°NM, C°M, Z0, U°) (2) indicates the WTP both to visit Madagascar and to undertake activity outside Mantadia park. Finally, by reducing the cost of recreation in Mantadia (C0M), where P°v = base price of a visit to Madagascar line (6) indicates the value of a visit to Madagascar plus recreational activities both outside and within C°NM = base cost of recreation outside Mantadia Mantadia. park The above sequencing of components facilitates the estimation of value attached by tourists to the COM = base cost of recreation in Mantadia park creation of a new national park. Survey techniques may then be used to determiine actual cash costs and Z° = vector expressing base levels of other goods time expended by intemational tourists visiting and services Madagascar, and thereby estimate their WTP for the various components. U° = base utility level Note In other words, Y° is the minimum expenditure re- quired to achieve the utility level U0, given the base 1. This annex is drawn from a detailed paper by Mercer levels for all prices and costs. and Kramer (1992). Similarly Y = e(P*v, CNM, CM,S Z, U ) (3) ANNEX 4: USING A SIMPLE BIODIVERSITY INDEX This annex summarizes the development of a basic of low biodiversity, such as a sand dune) whereas the biodiversity index that was developed to facilitate the loss of one hectare if 1,000 ha remain would be much analysis of environmental impacts at hydroelectric less. dam sites in Sri Lanka. It may be considered a first This approach to valuation of biodiversity is step, constrained by scarcity of data, resources and subject to several caveats. First, ecosystems may time. require some minimum area for long term survival, Detailed site specific information at potential which implies that the value function would need power plant sites is unlikely to be available at the to tend to infinity as it approaches that miniimum long-range system planning stage. Thus, the only value. quantification of biodiversity impacts that appears Second, and perhaps more importantly, the ar- possible at this level of aggregation is a probabilistic gument is sometimes made that the value to be as- estimate that gives the decision-maker information cribed to the habitat associated with some regulatory about the likelihood that the detailed environmental or governmental decision depends upon whether impact study will reveal adverse effects on an en- what remains is secure. For example, the cost of the demic species, significant inmpacts on ecosystems of loss of 1 ha of habitat if 1,000 remains rnight be high biological diversity, or degradation of a habitat valued as negligible, if that remaining habitat is already in a marginal condition. It should be noted that endemicity and bio-diversity are not necessarily Table A4.1 Relative Biodiversity Value correlated: an endemic spec:ies may be encountered in an area of low biodiversity, and areas of high rank ecosystem w biodiversity may contain no endemic species. It is certainly true, however, that extinction of an endemic species would very likely constitute a "show-stop- 1 lowland wet evergreen 0.98 per" at the project stage; and it is also true that at least forest as far as Sri Lanka is concerned, its endemic species 2 lowland moist evergreen 0.98 are most likely to be encountered in areas of high forest biodiversity. 0.90 A biodiversity index wll have several constitu- l o rs ent elements. First is the nature of the impacted 4 upper montane forest 0.90 system itself. In Table A4.1, the main ecosystems 5 riverine forest 0.75 encountered in Sri Lanka are ranked and assigned a 6 dry mixed evergreen 0.50 value, wj, that captures the relative biodiversity value forest of different habitats. The scale is to be interpreted as a strict ratio scale (i.e. zero indicates zero amount of villus 0.40 the characteristic involved, and a habitat value of 0.1 8 mangroves OAO implies ten times the value of a habitat assigned the 9 thorn forest 0.313 value of 0.01). The second element concerns the relative valu- 10 grasslands 0.3) ation, because the value of the area lost is a function 11 rubber lands 0.24) of the proportion of the habitat that is lost. For 12 home gardens 0.20 example, the loss of the last hectare (ha) of an eco- h 10 system would be unacceptable, and hence assigned 13 salt marshes an infnite value (even if the habitat involved were 14 sand dunes 0.10 Annex 4: Using a Simple Biodiversity Index 97 Figure A4.1 Biodiversity Index Value and Reservoir Size in Sri Lanka 1.3- 1.2 _ *KUKU 022 1.1 1 _ GING05 0.s9 0.8 0.7 § 0.6 - a GING052 1 0.5 _ 0.4 _KELA 071 iD 3 MAGUO43 0.3- 0.2 Other smai plvis iunl-ot-river) 0.1 _ 0 075 0 20 40 60 Reservor area, kM 2 protected from encroachment. On the other hand, if Bi = y Ai w the remaining 1,000 ha is vulnerable to encroach- ment, then the loss of that 1 ha under consideration might be assigned a much higher value. where Aj is the ha of ecosystem of typej at site i; and Such reasoning is not logical, because there is a wj is relative biodiversity value of type j (as defined confusion between costs and benefits. On the cost inTable A4.1). side, it really does not matter whether the loss is In the case of Sri Lanka, the biodiversity index attributable to a power plant, or to agricultural devel- is strongly correlated with reservoir size (see Figure opment, or even to illicit felling: the loss is the same A4. 1). This is simply a consequence of the fact that in both cases. On the other hand, the benefits to all of the larger projects inundate relatively large society of these activities may be quite different. Yet amounts of natural forest of high biodiversity value: it is only land use planning at the local or regional this would not necessarily be the case elsewhere. level that can address the costs and benefits of alter- However, when one excludes the five very large native uses in a systematic way. projects with reservoir sizes in excess of 15 km2, the Thus the biodiversity index associated with site correlation is quite poor, reflecting the diversity of i, Bi, is simply defined as ecosystem types encountered at the smaller sites. ANNEX 5: RECENT WORK ON ECONOMIC POLICIES AND ENVIRONMENTAL IMPLICATIONS Environmental Area/Problem Policy Issue or Reform Country . _ - or .eforrn 3 [(Study) |Land degradation Water degradation Airand atmo- Energy and trans- Other environ- Witdin Economy- and deforestation and depletion spheric polludion portation mental Sector wide efficiency problems 2 AFRICA Burkina Faso Decliing forest Loss of wildlife Promote community land Environmental and biodiversity management, technical managemtent project, support and environmental PRS434,1991 monitoring _ Burundi Inefficient use of Reform price stmucture of Energy sector energy power, oil, and woodfuels rehabilitation project, PR5411,1991 Central African Deforestation Wildlife Support to government for Republic degradation forest and wildlife Naturl resource protection management project, P15225, 1990 Ghana Agricultural Polludon from Industrial pollution Mining Effect of agricultural Environmental intensification and domestic and externalides development on implication of land degradation industrial waste sustainable agriculture. accelerated growth, Sectoral implications of white cover, Jasdip accelerated growth strategy Singl, 1992 . _ Note: This list focuses on work outside the Envirounment DeparmtmenL Note on report codes: ER-Economic Report; PR-President's Report; AR-Appraisal Report; SR-Sector lvory Coast Deforestation and Trade policy effects on land Tadc liberalization and decline of fallow use change (using GIS) deforestation, conference paper, 1991. Ramon Lopez Forest protcction; taining Forestry sector project, Deforestation Inefficient wood and investment in wood PR4879.1990 _ __ processing industry enya Deforestation; Agricultural water Introduce environmental Impact of irrigation Forestry development agroforestry depletion regulation: tree planting; investments p-Jec, PRS3So0,990 roion flscal subsidies; tenure form Protected areas and wwddlife services Need for wildlife Promote community project, PR5644, 1992 I management wildlife nanagement Lesotho land degradation Soil conservation L management and and sod erosion programs, such as pasture conservation project, improvement, tree planting PR4053. 1988 Madagascar Soil emsion, Coastal fishery Improve land security |Enviment pmgram, deforestaion over-use through tidtling; promote AR8348, 1990 (first environmnental education phase of BAP) Malawi Soil esion, Agricultural water Pricing reforms; X |(Economic Report Env deforestation depletion; urban institution building Mgt. Strtegy) water pollution _ _ _ Maudtius Lack of land use Investment program and Emizomnmental planning institution support for land monitoring aDd manunagement development project.{i PR53811990 Econtomic development LAnd allocation Industrial water Loss of tablish environmental Devise a National witi envirmental pollution biodiversity institutions and use environmental policy management, SR7264, economic incentives for 1988 _ - _ pollution abatement -' ~~~~~~~~tI I|Mozambique SoDl salinidtion Water-borne Address resource Unanticipated effects of (Olivares) from Inigadon; diseases from undervaluation irrigation development rangmve irrigadon (degradation _ _ X Nigeria Deforestation and Water Institutional and policy Establish land security to Towards the il degradation contamination, reforms for specific promote conservation; development of an ' especially among resource areas recommended review of nviroamenral acdon rural and landless trade policy impacts on lan for Nigeria, gray poor resource-based products; over SR9002,1990 energy pricing reforms . r Rwanda Excess demand Woodfuel pricing and Policies goveming issues and options in for fuelwood taxation fuelwood and petroleum die energy sector, substitution SR8017. 1991 I I O enegal Soil erosion Promote research on agro- Second agricultural forestry and soil rseareh projet, conservation PR5064 1990 Uganda Urban water Institutional and physical Second water supply availabiity and improvements in delivery project, PR5214, 1990 waste water and treatment capacity; treatment more pollution control proposed Zambia Economic Dependence on Export promotion to reduce recovery program, mining dependence on mining IPR5483, 1991 Zimbabwe Land Overgrazing Introduce forestry Need for tenure Forest resources degadation, and decline of regulation reforms and re- management and forest loss wildlife distribudon |development project. PR5086, 1990 Economics of Widlife Decline of Devaluation and tourism wildlife and implicadons on wildlife biodiversity loss Agriculture sector Inefficient land Promote land equity and memnorajdum, SR9429, use conservation 1991 I W . =~~~~~~~~~~~~~~~~~~~ Environmental Area/Problem Policy Issue or Reform F Country_ (Study) Land degradation Water degradation Airand atmo- Energy and trans- lOther envimn- Within Economy- |and dcforcstation and depletion spheric pollution portation mcntal Sctor wide _____________randdefresatin 1ad dpicion spheic olltio jemcienc jprbem _ficiency problems jASIA Il __ _ ____ i_ _ _ _ _ __ _ _ _ _ ___ ___ ___ __ _ |Bangladesh Fishery I t 1 Feshwater Increase private sector role sector review, gray fishery in fishery development; cover SR8830, 1991 development improve credit access India Deforestation and Water resource Industry ndusty regulation; Role of population pressure India's environment- watershed degradation pollution soil conservation on limited land; protection of a strategy for World stabilization program material and energy Bank assistance, intensivc industry SR7676, 1989 Mahanshtra funl High incidence of Investment in rural water water supply and water-borne supply and sanitation environental disease systems sanitaton, AR9202, 1991 Industrial pollution Urban water Industrial air Investment, technical control project, polludon, water pollution assistance, and monitoring |PR5485, 1991 supiy aond at the state level l ~~~~~sanitadion Integrated watershed Deforestation and ' Industrial pollution Environmental Institutional and investment development (plains) land degradation; impact of energy support for watershed project, PR5241, soil erosion, projects management; community 1990; (hills) prject, watershed management PR15187, 1990 - degradadon i _ 5 .CZ.1 tII PI China Environmental Execssive use Urban air pollution Inefficient coal Institutional reforms Reforn of industrial Strategy paper, of fertilizer and use incentves; green cover SR degradadon of implications of economic re- . o 669, 1991 Nick marginal lands stmructring o Anderson srcui Technical assistance for Liaoning urban Urban water use llution abatement and infrastructure project, and pollidon ater prieing AR8250. 1991 o Efficiency and Unvestments and policy Investment and policy nvitonment l Urban air polluton Widespread coal ptions for reducing coal- options for reducing irpaac of eoal use, use related polludon household energy use gray cover SR89tS, 1991 Technical and institudonal spport for plantaton National afforestadon ap between anagement and forest project. AR8487. 1990 forest resources sector planning and use; poor quality of plantation efforts Coal pricing in Inefficient use of Reforms in subsidized coal China, WBDPI38, coal pricing system Yves Albouy . .. . . 0 wk Indonesia Contuainated Unsafe solid Investment, technical Second Jabotabek urban water waste disposal assistance and policy urban development reforms to reduce water- project, AR8339. 199 related pollution Third Jabotabek Household ommunity participation in urban development waste disposal aste disposal systems project. AR8397. 1990 Trade and Excessive louing Trade policy, gains from deforestadon, trde, and implications for conference paper, net welfare Cados Braga. 1991 Taxadon nd Urban air pollution Role of taxes and subsidies environment, in energy, transport, and Eskdland et al petrochemical industries Second forestry Deforestadon Technical assistance and insdtutions n concession monitoring conservation pmject, PR5320. 1990 Perdlizer restructuring project, Environmental Investment in industry PR5231, 1990 impact of together with environmental fertilizer use management program Yogyakarta upland Soil erosion aea development Technical support for soil roJect, PR5383, 1991 conservation and watershed ,_______________ __________.___| stabilization ;a Krea Urban air pollution LNG utilization LNG promotion Gas udilization study, to reduec air SR8142,1990 _ pollution ;__ Malaysia Reforstation ad Support for reforestation Forestry sector sustainable ased on so-l and glohb; g study, green cover, forestry its fi1991. Wiens n _ _ _ _ ' '' __ H Ii |NepalsOn erosion and Agricultural development SL rst loss policy effects, SAL incorporates resource2i management Papua New Guinea Forest loss Institution building The forestry sector a and forestry investments WAP'rev-ew, SR8031, 1990 _ Philippines Deforestation and Mangrov Resource rent taxation and Environment and soil emsion degradation and institutional reforms in Env. natural resource coastal fishery SECAL sector adjustment depletion rogrm. red cover. 1989 PR5452 Sri Lanka Mining coastal Urban water Excess fuelwood New regulation and energy Environmental land degradation pollution demand conservation Managenment Strategy, green cover SR9649, 1991 l ailand Forest lss and Urban air and Degradation of Under-pricing of dirty fuels; Growth-orientcd policies not Country economic soil erosion water pollution resources regulate land use sufficiently sensitive to memorandum, gray associated with distribution and cover ER9627. 1991 accelerated environmental implications igrwth 0 Co-ntry I Environmental Area/Problem Policy Issue or Reform Country __ (Study) Land degradation Water degradation Airand atmo- Energy and trans. Other environ- Within Economy- and dcforestation and depletion spheric pollution portation |mental Sector wide IAsia. I ~ efficiency problems a EEurope find Ccmrai IAsia. I Bulgaria l nefricient energy State enterprise and price Structural adjustment Use meforms loan project. SAL P5S88, 1991 Environment strategy Industrial pollution Material and Economic restructuring study, yeDlow cover, energy intensive implications for pollution and 1991 industry energy Czechoslovakia Inefficient energy Pricing and regulatory Structural adjustment use reform in energy sector; an project, PR5583, establish institutional and 1991 lpolicy framework for environmental protection Turkey Inefricient Reform power pricing and Turkish electricity electricity use provide technical authority rstructuring assistance for loan project. PR5422, environmental assessment 1991 . II Poland Water pollution Air pollution Promote decentralized Environment from industry mgt approach; management project, institutional support ' PR5256, 1990 Energy market Energy use Energy price reform development, SR8224, efficiency j;i99 __ __ _ I IL. i i . ~ ~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ - B. tSI Environmental Area/Problem Policy Issue or Reform Country (Study) Land degradation Water degradalion Airand atmn- Energy and trans- Other environ- Within Economy- - and deforestation and depiction spheric polluion portation mental Sector wide , . efficiency problems t; Middle East and North 1 g 8 Africa Algeria Water resource Assess water use in Agriculture: a new over-exploitation agriculture opportunity for growth, SR7419, 1990 Environmental Land degradaton Agricultural water Industrial polludon Population Hazardous waste Industrial expansion probiems and issues, depletion from promoted growth and management, and water programs need to be yellow cover in French, heavy industries pressure on treatment coordinated with pollution 1991, Nadia Saad limited arabic abatement land, tourism and site _ l _ _____ _ _ _ .degradation Egypt Need for land Agricultural water Improve efficiency of future Land reclamation reclamadon supply investments in land subsector review, reclamation SR8047, 1990 Environmental issues Water resource Potendal use of Tourism Need for Iegislation and paper, white cover, scarcity identified natural gas development institutional capacity for Nadia Saad, 1990 as most serious rmserves has not management; investments environmental ensured needed for water constraint protection of development sites Jordan inefficient energy Oil and energy price Energy sector study, use reforms SR7984, 1990 . ,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~- 0 Tunisia Overgrazing Vulnerability to Low efficiency of watcr use Address increase in herd Agricultural drought and size from fodder subsidy expenditure review, agricultural water during droughlt SR9511, 1991 depletion Small farmers' potential Inadequate land Inefficient use of and prospects: a conservation irrigation at farm Water pricing reforn Land tenure, reform land technical study. level markets to promote soil SR9323, 1991 conservation Yemen Massive tubewell Remove irrigation expansion and subsidies groundwater depletion _ C3 Environmental Area/Problem Policy Issue or Reformi Country __=__._.._.__. _.___ (Study) Land degradation Water degradation Airand atmo- Energy and trans- Other environ- Within Economy- and deforestation and depletion spheric pollution portation mental problems Sector wide :Z. efficiency Latin America and - Caribbean __ Argentina Energy price and Assess tax and subsidy Energy sector study, tax distortions schemes SR7993, 1990 Agricultural sciences Pest control and Increased government and institutional increased capacity for environmental dcvelopment project, chemical management 1991 ER residues; overfishing Environmnental issues Soil erosion, Urban water paper, SR8141, 1989 salinization. and pollution Biodiversity Institutional reorganization Undertake natural resource eforestation losses accounting Bolivia eforestation; Technology improvement Agricultural land degradation to reduce wind erosion technology and soil erosion and chemical usC development project, in the Altiplano 1991 ER |BRzjl Increased logging Trade lib., gains from trade, Tnde and and net welfare Environment conference paper,1991, Carlos Braga National Deforestation and Watkr pollution Airpolludion Migrtion and Strengthen government Environmental soil erosion deforestation in env. agencies, and Project, AR 8146 northern frontier reguladon Promote LNG Promote more efficient Hydrocarbon subsdtution and energy pricing transport and increased processing, ER 1991 efficiency of petroleum use Chile Deforestation Industrial pollution Overfishing Resource undervaluation Economic openness and Structural adjustment and rent-seeking contribution to cleaner and environment industrial technologies study. 1991. Scarsborough _ Colombia Industrial pollution Technical assistance to Industrial government and industrial restructuring and enterprises development project, ER 1991 Costa Deforestation; Resource undervaluation; Cattle industry subsidies Rica pasture land focus on regulatory promote forest clearing; Country Economic degradation approach inequitable land distribution Memorndum, 1989 ^ (El-Serafy) ._ _ Dominican Republic Limited energy Remove energy subsidies Issues and options resources and in the energy sector, distorions SR8234, 1991 _ _ __o _. e tig Ru r Deforstation. Improve land tenure tco Rural devclopment erosion encourage conservation. --~ poject, AR9437,199I Reduce poverty and absorb niral labor to reduce pressure on resources Lower Guayas flood Improve drainage Environmental monitoring. Land titling to encourage - control project. ER works in farms mangrove mnanagement, fann improvements 1991 and rivers institutional strengthening Municipal Waste disposal Investment in infrastructure evelopment and and sewage and technical assistance rban infrastructure infrastructure for environnental proect ER 1991 mnitoring Haid Deforestation and Eicessive Technical and financial Forestry and soil erosion consumption of support to TFAP envimnmnental fuelwood protection pomject, resources AR9307. 1991 Honduras |Dforcstation Vnd Poverty and Support for government ASALaddresses lan(I tenure Structural adjustment soti cmsion environmental resource mgt. efforts and forest mgt. issues. credit ptoject, degradation poverty reduction PR5453 Peru Deforestation Industrial pollution Overfishing Resource undervaluation Economic openness and Structural adjustment and rent-seeking contribution to cleaner study, 1991 industrial technologies Scarsborough Venezuela Deforestation Wasteful use of Decline of wildlif Institutional progress Increase fertilizer and Envirmnnental issues from catte and agricultural, including first environment pesticide prices to reduce in Venezuela, gray ilegal logging industrial, and ministry in LAC; stumpage water contamination cover SR8272, 1991 household water valuation; water charges |Mexico Industrial and T Trade and trucking Fiscal policy and transportation restrictions and impact on air environment project pollution pollution Transport and Poluio* r- Monitoring and tclecom. SECAL, trucking enforcement of standards gPR5254, 1.990. |Water and sanitation Urban water Price reformns, projcct {project, PR89173, 1990 supply aind investmncnt sanitation l lMand governgnntd Mining sector monitoring capacity restflcturing project, AR9428, 1991 |Allocation of Deregulate mining and Promote employment and Decentralization an mining rights and cstablish environmental acccss to social servies to regional .rjc|off-site costs standards; improve poor and indigenous development project institutional capacity at populations to reducc fordisadvantaged t99 local levels pressure on forcsts states, AR8786, 1991 l l l ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~Comprehensive Transport Air Quality Conserve Pollution from Ctransportaion policy to Management in thc remaniing humid transport sourtcs nprmote efficiency and Mexico Metropolitan forest pollution abatement a Area, SR10045, 1992 p a Me I.. Distributors of World Bank Publications ARCENTINA The Middle EastObeever KENYA SOUTH AFRICA, BOTSWANA Caelos Hirsch, SRL 41, Sherif Street Africa Book Servic (EA.) LtL For singie tikiw Galera Guem- Cairo Quaran Hous. Mfangaw Strzt Oxford Univerity Pri Florida 165, 4th Floor-Ofc. 455/465 P.O. Box 45245 Southern Afria 1333 Buenas Aires FINLAND Nairobi P.O. Box 1141 Akateeminen KIrjakauppa Cape Town 8C00 AUSTRAUA, PAPUA NEW CUINEA. P.O. Box 128 KOREA, REPUBLIC OF FIJI, SOLOMON ISLANDS, SI:00l10 Helsinki 10 Pan Korea Book Corpontion For oreeal'ies Ordm: VANUATU, AND WESTERN SAMOA P.O. Box 101, Kwangwhamun Inational Subcription Service D.A. Information Servios FRANCE Seoul P.O. Box 41096 648 Whitehors Road World Bank Publiations CrAighau Mitcham 3132 66, avenue d'14na MALAYSIA johannesburS2024 Victoria 75116 Paris Univermty of Malaya Cooperative Bookahop, Limited SPAIN AUSTRIA GERMANY P.O. Box 1127, Jalan Pantai Baru Mundi-Prensa Librs, S.A. Gerold and Co. UNO-Veriag 59700 Kuala Lumpur Castello37 Graben 31 Poppelsdorfer Alle 55 2001M Madrid A-IOI Wien D-5300 Bonn 1 MEXICO WNFOTFC Libria [nternaconal AEDOS BANGLADESH HONG KONG, MACAO Apartado Postil 224860 Consel de Cent. 39 Micn Industi Development Asia 2000 Ld. 14060 Tlalpan, Mexio OF. 08009 Barcelona Asistance Soiety MIDAS) 46.48 Wyndham Street HouxeS. Road 16 Winning Centre NETHERLANDS SRI LANKA AND THE MALDIVES Dhanmondi R/Are 2nd loor De LindeboomInO/ -Publikaties Lake House Bookshop Dhaka 1209 Central Hong Kong P.O. Box 202 P.O. Box 244 7480 AE Haalksbergen 100, Sir Chittmpilam A. Brach evs INDIA Gardiner Mawatha Pine View. Ist Floor Allied Publishers PrivateLtd. NEW ZEALAND Colonibo2 100 Agrabad Commercial Area 751 Mount Road EBSCO NZ Ltd. Chittgong 4100 Madras - 600 002 Private Mail Bag 99914 SWEDEN New Market Forsingle titrlo 76, KD.A. Avenue Brgnch ofta: Auckland Fritzes Fackboksfore,aet Kuina 9100 15 J.N. Heredia Marg Regeringsatn 12, Box 1656 Bailard iEstate NIGERIA 510327Stockholn BELGIUM Bombay - 400 038 University Pre Limited Jean De Lannoy Three Crowns Building Jericho For sm t : Av. du Roi 202 13/14 Asf Ali Road Private Mail Bag 5095 Wennergren-Wlliam AB 1060 Brussels New Delhi -110 002 [badan P. . Box 1305 S-17125 Sona CANADA 17 Chittaranan Avenue NORWAY La Dlffuseur Calcutta -700 072 Narvesen Information Center SWITZERLAND C.. 85, 1501B rue Amp8re Book Department For sixgle riwth Boucherville. Quebec layadeva Hostel Building P.O. Box 6125 Etterad Librarine Payot J48 5E6 5th Main Road, Gandhinagar N-0602 Osio 6 Case postale 3212 Bangalore - 560 009 CH 1002 Lausnne CHILE PAKISTAN Invertc GCT S.A. 3-5-1129 Kachiguda Mirza Book Agncy For nawlm oirds: Americo Vespucio Norte 1165 Cron Road 65. ShahrahQuaid-e-Azam lrairie Payot Santiago Hyderabad -500027 P.O. Box No. 729 Service de Abonnements Lahore 54000 Case poase 3312 CHINA Prathana Flab, 2nd Floor H 1002 Lausanne Chuia Ftnancial & Economie Near Thakom Baug, Navmngpum PERU Publishing House Ahmedabad -380 009 Editoril Desarrolo SA THAILAND 8. Da Fo Si Dong lie Apartado 324 Central DepartmentStore Beijing Patala House Lima 1 306 Silom Road 16-A Ashok Marg Bangkok COLOMBIA Lucknow - 226 001 PHILIPPINES Infoenace Ltda. International Book Center TRPDAD & TOBAGO, ANTIGUA Apartado Aereo 34270 Central Bazaar Road Suite 1703, GCtyland 10 BARBUDA, BARBADOS, Bogota D.E. 60 Baja Nagar Condominium Tower I DOMINICA GRENADA, GUYANA, Nagpur 440 010 Ayala Avenue. H.V. dela JAMAICA, MONTSERRAT, ST. COTE D'IVOIRE Cost Extntsion KITTS & NEVIS, ST. LUCA. Centre dEdition et de Diffusion INDONESIA Makati Metro Manila ST. VINCENT I GRENADINES Africaines (CEDA) Pt Indira limited Systematics Studies Unit 04 B.P. 541 Jalan Borobudur 20 POLAND #9 Watts Street Abidan 04 Plateau P.O. Box 181 International Publishing Service Curm-e Jakarta 1320 UL Pieknn 31/37 Trinidad West Indies CYPRUS 00.677 Warzawa Center of Applied Reserch IRELAND TURKEY Cyprus College Covernment Supplies Agency For suoscri;m ordes: Infotei 6. Diogenes Street Engorni 4-5 Harcourc Road INS Journals Narlabake Sok. No. US P.O. Box 2006 Dublin2 Ul. Okrezna 3 Cagaloglu Nicosia 02-916 Warzawa Istaibu ISRAEL DENMARK Yozmot Literature Ltd. PORTUGAL UNITD KINGDOM SamfundsUtteratur P.O. Box 56055 Livraria Portugal Microinfo Ltd. Rosnoerns AI# 11 Tel Aviv 61560 Rua Do Carmo 70-74 P.O. Box 3 DK-197 FrederiksbergC i200 Lisbon Alton, Hampshire GU342PG ITALY England DOMINICAN REPUBUC Liosa Commisionaria Sensoni SPA SAUDI ARABIA, QATAR Editors Taller. C porA. Via Duca Di Calabria. 1/1 jarir BookStore VENEZUELA Restauracion a Lsabel La Cat6llca 309 Casella Postale 552 P.O. Box 3196 ibria del Eate Aprado de Correos 2190 Z-1 50125 Firenze Riyadh 11471 Aptdo. 60.337 Santo Domingo Caracas 1060-A JAPAN SINGAPORE. TAIWAN EGYPT, ARAB REPUBUC OF Eastern Book Service MYANMARBRUNEI Al Ahram Hongo 3-Chome. Bunkyo-ku 113 Information Publcations Al Caa Street Tokyo Private. Ltd. Cairo Golden Wheel Building 41. Kallang Pudding #04-03 Singapore 1334 The World Bank Headquarters European Office Tokyo Office 1818 H Street, N.W. 66, avenue d'Frna Kokusai Building 3-chom Washington, D.C. 75116 Paris, France 1-1 Marunouchi 3-chome Chiyoda-ku, Tokyo 100, Japan Telephone: (202) 477-1234 Telephone: (1) 40.69.30.00 Facsimile: (202) 477-6391 Facsimile: (1) 40.69.30.66 Telephone: (3) 3214-5001 Telex: wul 64145 WORLDBAN\K Telex: 640651 Facsimile: (3) 3214-3657 RCA 248423 WORLDBK Telex: 26838 Cable Address: INTBAFRAD WASHINGTONDC Cover design by Walton Rosenquist ISBN 0-8213-2352-0