_lrPo lto WORLD B3ANK TECHNICAL PAPER NUMBER 180 Trends in Agricultural Diversification Regional Perspectives * V-4 Shawki Barghouti, Lisa Garbus, and Dina Umali, editors 1-- RECENT WORLD BANK TECHNICAL PAPERS No. 115 Le Moigne, Plusquellec, and Barghouti, Dam Safety and the Environment No. 116 Nelson, Dryland Management: The "Desertification" Problem No. 117 Barghouti, Timmer, and Siegel, Rural Diversification: Lessons from East Asia No. 118 Pritchard, Lending by the World Bank for Agricultural Research: A Review of the Years 1981 through 1987 No. 119 Asia Region Technical Department, Flood Control in Bangladesh: A Plan for Action No. 120 Plusquellec, The Gezira Irrigation Scheme in Sudan: Objectives, Design, and Performance No. 121 Listorti, Environmental Health Components for Water Supply, Sanitation, and Urban Projects No. 122 Dessing, Support for Microenterprises: Lessons for Sub-Saharan Africa No. 123 Barghouti and Le Moigne, Irrigation in Sub-Saharan Africa: The Development of Public and Private Systems No. 124 Zymelman, Science, Education, and Development in Sub-Saharan Africa No. 125 van de Walle and Foster, Fertility Decline in Africa: Assessment and Prospects No. 126 Davis, MacKnight, IMO Staff, and Others, Environmental Considerations for Port and Harbor Developments No. 127 Doolette and Magrath, editors, Watershed Development in Asia: Strategies and Technologies No. 128 Gastellu-Etchegorry, editor, Satellite Remote Sensing for Agricultural Projects No. 129 Berkoff, Irrigation Management on the Indo-Gangetic Plain No. 130 Agnes Kiss, editor, Living with Wildlife: Wildlife Resource Management with Local Participation in Africa No. 131 Nair, The Prospects for Agroforestry in the Tropics No. 132 Murphy, Casley, and Curry, Farmers' Estimations as a Source of Production Data: Methodological Guidelines for Cereals in Africa No. 133 Agriculture and Rural Development Department, ACIAR, AIDAB, and ISNAR, Agricultural Biotechnology: The Next "Green Revolution"? No. 134 de Haan and Bekure, Animal Health in Sub-Saharan Africa: Initial Experiences with Alternative Approaches No. 135 Walshe, Grindle, Nell, and Bachmann, Dairy Development in Sub-Saharan Africa: A Study of Issues and Options No. 136 Green, editor, Coconut Production: Present Status and Priorities for Research No. 137 Constant and Sheldrick, An Outlook for Fertilizer Demand, Supply, and Trade, 1988/89-1993/94 No. 138 Steel and Webster, Small Enterprises under Adjustment in Ghana No. 139 Environment Departinent, Environmental Assessment Sourcebook, vol. I: Policies, Procedures, and Cross-Sectoral Issues No. 140 Environment Department, Environmental Assessment Sourcebook, vol. H: Sectoral Guidelines No. 141 Riverson, Gaviria, and Thriscutt, Rural Roads in Sub-Saharan Africa: Lessons from World Bank Experience No. 142 Kiss and Meerman, Integrated Pest Management and African Agriculture No. 143 Grut, Gray, and Egli, Forest Pricing and Concession Policies: Managing the High Forest of West and Central Africa No. 144 The World Bank/FAO/UNIDO/Industry Fertilizer Working Group, World and Regional Supply and Demand Balances for Nitrogen, Phosphate, and Potash, 1989/90-1995/96 No. 145 Ivanek, Nulty, and Holcer, Manufacturing Telecommunications Equipment in Newly Industrializing Countries: The Effect of Technological Progress No. 146 Dejene and Olivares, Integrating Environmental Issues into a Strategy for Sustainable Agricultural Development: The Case of Mozambique (List continues on the inside back cover) WORLD BANK TECHNICAL PAPER NUMBER 180 Trends in Agricultural Diversification Regional Perspectives Shawki Barghouti, Lisa Garbus, and Dina Umali, editors The World Bank Washington, D.C. Copyright i 1992 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 August 1992 Technical Papers are published to communicate the results of the Bank's 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 condusions 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 induded 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 delimitation of its boundaries or its national affliation. 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 dassroom 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, Department F, The World Bank, 1818 H Street, N.W., Washington, D.C. 20433, U.S.A., or from Publications, The World Bank, 66, avenue d'I6na, 75116 Paris, France. ISSN: 0253-7494 Shawki Barghouti is Chief in the Agriculture Operations Division of the South Asia Region Country Department Ill of the World Bank. Lisa Garbus and Dina Umali are consultants to the Natural Resources Division of the Agriculture and Rural Development Department of the World Bank. Library of Congress Cataloging-in-Publication Data Trends in agricultural diversification: regional perspectives / [edited by] Shawki Barghouti, Lisa Garbus, and Dina Umali. p. cm.-(4World Bank technical paper, ISSN 0253-7494; no. 180) Includes bibliographical references. ISBN 0-8213-2191-9 1. Agricultural diversification-Developing countries. 2. Agricultural diversification-Asia, Southeastern. I. Barghouti, Shawki M. II. Garbus, Lisa, 1963- . m. Umali, Dina L. IV. Series. S482.T74 1992 338.1'09172'4-dc2O 92-24632 CIP FOREWORD In pursuing strategies to promote agricultural development, policymakers and development planners are increasingly recognizing agricultural diversification as an important alternative, not only to maintain or improve farm incomes, but also to increase agricultural productivity through more efficient resource use and the avoidance of further environmental degradation resulting from overintensive monocultural production. The success of agricultural diversification programs, however, is influenced by a wide variety of factors, including the agricultural environment, the availability of appropriate technologies, the structure and performance of markets, and the broad range of govermnent policies affecting agriculture. Such multiple determinants of success increase the difficulty of studying agricultural diversification, which is further complicated by the multiple perspectives (farm, regional, sectoral, and intersectoral) from which diversification can be viewed. Within the Bank, there has been increasing interest in agricultural diversification as a development option for various regions, including Southeast Asia, the Middle East and North Africa, Sub-Saharan Africa, Central America, and South America's Pacific Rim countries. In response, the Bank's Agriculture and Rural Development Department (AGR) requested and was awarded a special grant by the Bank's Research Committee to conduct a series of country studies on the current status of and future prospects for agricultural diversification in Southeast Asia, particularly Indonesia, Malaysia, the Philippines, and Thailand. AGR also commissioned studies on increasing the flexibility of agricultural systems in the Middle East and North Africa, Sub-Saharan Africa, Central America, and South America's Pacific Rim countries. This publication presents the results of these studies, highlighting Southeast Asia, where notable progress has been achieved in agricultural diversification. The papers provide detailed information about how each country or region has approached the policy and technical issues of diversification and the analytic tools employed to articulate them. In response to changes in the agricultural sector, the Bank is emphasizing lending for broadly based sectoral operations rather than for narrow commodity-based projects. It also supports the inventory and sustainable exploitation of regional resources and the introduction of advanced technical concepts in irrigation and marketing. In its operational work, the Bank is attempting to determine appropriate frameworks to evaluate diversification and is turning increasingly to multidisciplinary approaches in project design and appraisal. This volume represents the first step in understanding agricultural diversification in a regional context and should serve as a springboard to future research and implementation efforts. Michel Petit Director Agriculture and Rural Development Department I INTRODUCTION Agricultural diversification creates opportunities for achieving higher and more stable rural incomes through the more efficient use of resources and the exploitation of comparative advantage. It involves both the introduction of diverse commodities and technologies and the development of related job opportunities in rural agroindustry. It is an evolutionary, demand-driven development process that should be supported by appropriate policies, technologies, infrastructure, and services. Because farm-level diversification broadens income sources for rural households, it can also serve as a risk-minimizing strategy. As a development strategy, however, agricultural diversification is influenced by several factors, including technology availability, agricultural support systems capability, flexibility of irrigation systems, capital accessibility and availability, and the political will of governments. Particular constraints include limited technologies for alternative crops, irrigation water supply and management inefficiencies, poorly developed agricultural markets, weaknesses in research and extension, and unfavorable government policies. Interest in agricultural diversification as a development alternative has been increasing for various reasons. The diminishing returns to the Bank's lending portfolio in the rice-based systems of Southeast Asia have underscored the need to diversify. In the Middle East and North Africa, the focus is shifting from food security based on self-sufficiency to food security sustained through trade based on comparative advantage and more diversified agricultural production. Poor agricultural performance in Sub-Saharan Africa over the past two decades has stimulated interest in diversification as a possible venue for agricultural growth in that region. Central America and South America's Pacific Rim countries have been involved in export diversification over the past 50 years, with mixed results; however, the need for a diversified economy is well accepted. In response to these developments, the Bank's Agriculture and Rural Development Department commissioned studies on increasing the flexibility of agricultural systems in these regions. The studies focus on technical, economic, and institutional aspects, including. microlevel constraints to farmer adjustments and macrolevel policy responses. This publication presents the results of these studies, highlighting Southeast Asia, where notable progress has been achieved in agricultural diversification. The policy instruments and objectives to stimulate the process of diversification differ by country. A common problem is the limited flexibility of agricultural systems in adjusting to economic, political, and technological change, thereby constraining the ability of farmers to adjust to shifting opportunities to maximize their benefits and minimize risk and uncertainty. The papers in this volume stress these points. First, Petit and Barghouti analyze the forces that drive the diversification process, both past and future. Zandstra then focuses on related technological considerations. Broad examinations of agricultural diversification in the regions discussed above follow: Southeast Asia (Timmer), the Middle East and North Africa (Somel), Sub-Saharan Africa (Jaffee), and Central America and South America's Pacific Rim countries (Tabora). Southeast Asia is then highlighted, with Pingali examining diversification out of rice monoculture into multicrop enterprises and Miranda looking at irrigation management requirements to facilitate this movement. Next, agricultural diversification experiences in Indonesia (Kasryno, Manwan et al.); Malaysia (Mat and Chen); the Philippines (Adriano and Cedillo, Adrias); and Thailand (Siamwalla et al.) are presented. These studies explore diversification performance at the farm, regional, and sectoral levels and examine the future role of government in the diversification process. The editors thank the following persons for their assistance in the preparation of this publication: G. Edward Schuh, Avishay Braverman, Antony Cole, Dirk Leeuwrik, Sushma Ganguly, Carol Timmer, Wilfred Candler, William Cuddihy, and Peter Hazell. TABLE OF CONTENTS Overview Diversification: Challenges and Opportunities Michel Petit and Shawki Barghouti 1 Technological Considerations in Agricultural Diversification H. G. Zandstra 15 Regional Perspectives Agricultural Diversification in Asia: Lessons from the 1980s and Issues for the 1990s C. Peter Timmer 27 Agricultural Diversification in the Middle East and North Africa Kutlu Somel 39 Enhancing Agricultural Growth through Diversification in Sub-Saharan Africa Steven Jaffee 61 Central America and South America's Pacific Rim Countries: Experience with Export Diversification Panfilo C. Tabora, Jr. 93 Focus on Southeast Asia Diversifying Asian Rice-Farming Systems: A Deterministic Paradigm Prabhu L. Pingali 107 Irrigation Management for Crop Diversification Senen M. Miranda 127 Indonesia: Diversification as an Agricultural Policy Instrument Faisal Kasryno 135 - viii - Agricultural Diversification in Indonesia: The Potential Roles of Research, Extension, and Irrigation Ibrahim Manwan, Ahmad Dimyati, A.M. Fagi, and M. Oka Adnyana 151 Rationalizing Crop Diversification in Malaysia's Nongranary Irrigated Areas Jaafar Mat and Ng Chau Chen 167 Agricultural Diversification in the Philippines Marietta S. Adriano and Jocelyn L. Cedillo 181 The Diversified Crops Irrigation Engineering Project in the Philippines Alberto S. Adrias 191 Public Policies toward Agricultural Diversification in Thailand Amnmar Siamwalla, Direk PatamasiAwat, and Suthad Setboonsarng 199 DIVERSIFICATION: CHALLENGES AND OPPORTUNITIES Michel Petit and Shawki Barghouti Policymakers and agricultural planners are increasingly focusing on agricultural diversification as a strategy to promote agricultural development because it offers opportunities to reduce production and price risks, increase flexibility, increase agricultural incomes, and sustain productivity and growth. Although attention to the dialectical relationship between diversification and specialization is not new, concern about diversification has recently intensified because of major developments in the agricultural sector in many developing countries, particularly in Asia. This paper first reviews these developments and some of their implications. Then, it discusses the dimensions of diversification and its role in advancing agriculture in developing countries. Finally, policy reforms, essential components of future agricultural development strategies, are proposed. The paper's framework, in outlining the main issues and policies of agricultural diversification, is based on the need to understand how success in agricultural development often creates the need for a major reorganization of resources, both within the agricultural sector and between agriculture and the rest of the economy. This process is most likely to occur when success is based on the introduction of new technology in the production of agricultural commodities. Tlhe paper purposely avoids a "pick-the-winner" approach to diversification. It argues that creation of a profitable environment for new crops requires substantial technical research, innovative economic analysis, and policy adjustment. The paper's theme is the desirability of introducing greater flexibility into farming systems and of broadening the basis of the rural economy in general. Recent Developments in the Agricultural Sector The most notable development in the agricultural sector is its declining contribution to the national economy (table 1). Farm incomes are lower than urban incomes. Consequently, the nonagricultural sectors are drawing resources away from agriculture, and low prices and incomes are "pushing" farmers into urban jobs. Farmers are diversifying their incomes by seeking partial employment outside the farm sector. Growth in the rural nonfarm economy has been significant in most developing countries. Nonfarm activities now account for 20 to 30 percent of rural employment in Asia and Latin America and 10 to 20 percent in Africa (World Bank 1990b). Most successful countries use this structured transformation for more efficient resource allocation and better income distribution. The second important development involves the rapid progress in the development of advanced agriculturalproduction technologies. New technologies have enabled farmers to increase their output and provided them with new production options and more flexible farming systems. Much of the increase in world food production since 1960 has resulted from intensified production of rice, wheat, maize, poultry, and dairy products. This production growth has been achieved primarily through yield increases. Baanante et al. (1989) projected that between 1982/84 and 2000, yield increases will account for approximately 50 to 80 percent of production growth in Latin America, the Near East and North Africa, Sub-Saharan Michel Petit is director of the World Bank's Agriculture and Rural Development Department. Shawki Barghouti, formerly chief of that department's Agricultural Technology and Natural Resources Division, is now chief of the Agriculture Operations Division, South Asia Region Country Department 11, World Bank. Africa, and Asia; expansion of cultivated land TABLE 1. PERCENTAGE SHARE OF and increasing cropping ntensity will account AGRICULTURE IN GDP AND EMPLOYMENT for the remainder (figure 1). The basic IN SELECTED COUNTRES, 19 AND 19M technologies responsible for these output % SHARE OF % SHARE OF increases are improved plant varieties and AGRIC. N GDP LABOR DN AGRIC. animal breeds, agrochemicals, and irrigation COUNTRY 1965 1988 1965 1988 systems. High-yielding varieties of rice, for exam.ple, largely contributed to the doubling Souaheas: aia Malaysia 28 21 60 36 of yields in several Asian countries (figure 2). Thailand 32 17 82 59 Technological improvements in Philippines 26 23 57 49 irrigation systems have also expanded Indonesia 56 24 71 55 production opportunities. Whereas traditional Midde ast mid irrigation technologies (furrow, border, and Nonh Af,ica flood irrigation), which involve water delivery Egypt 29 20 55 42 Turkey 34 19 75 52 to plants through gravitation, usually resulted Sy& 25 22 52 25 in substantial water losses and limited Morocco 22 IS 61 40 uniformity in water distribution, modem Tunisia 20 17 49 32 irrigation technologies (particularly, sprinkler Jodna 18 8 57 264 and drip irrigation) are characterized by increased water use efficiency. Furthermore, they have opened greater opportunities to Source: World Bank 1982,1990b. cultivate soils with low water-holding capacity (sandy and rocky soils) and to farm low- quality lands and steep slopes. This transition has also enabled regions facing limited water supplies to shift from low-value crops with high water requirements (e.g., cereals) to high-value crops with lower water requirements, such as fruits, vegetables, and oilseed. It has also allowed the use of low-quality (e.g., highly saline) water in regions with high temperatures and high evaporation rates. These technological advances, however, have been a double-edged sword. The technological changes in both crop production and irrigation were accompanied by an increasing scarcity of suitable and productive agricultural resources; scarce water resources and a diminishing availability of suitable lands with fertile soils are evident in many developing countries. Concurrently, global changes in output, particularly of cereals, resulted in lower commodity prices (Timmer 1988). Under these conditions, farmers, especially in countries that have narrowed their food self-sufficiency gap, have had to search for new crops, new technology, and new production systems to sustain their farming enterprises. Thus, improved technologies have not always brought with them stability and improved incomes. In Southeast Asia and the Middle East, this has translated into the rapid growth in production of nontraditional agricultural commodities such as fruits, vegetables, roots and tubers, and poultry (table 2). The third important development is the increasing awareness of and concern about the impact of current production practices on the environment. Environmental degradation resulting from intensive production practices is a growing problem in many developing countries. For example, the long-term decline in rice yields can be attributed to the combined effects of increased pest pressure, the rapid depletion of soil micronutrients, changes in soil chemistry induced by intensive cropping, and increased reliance on low-quality irrigation water (Flinn and De Datta 1984; Pingali et al. 1990). Agricultural diversification thus offers one approach for stemming further environmental degradation through the establishment of multicommodity production systems that are not only economically profitable, but also environmentally sound. Last, the world market is characterized by shi,ting consumer demands. The world market has become more competitive, and consumption patterns, especially among high-income communities, are Figure 1. Projected Contribution to Production Growth of Yield, Land, and Cropping Intensity, 1982/84-2000 5X Cl7X l ^ ZOX :1~~~~~~~~~2% 15% 17% 20%12 23% 22% 26% 11% 39% 77% 63% 57% 69% 49% 93 developing Sub-Saharan Near East/ Asia (excl. People's latin countries Africa North Africa Republic of China) America [EI] Yield Arable land Cropping intensity Source: Baanante et al. 1989. Figure 2. Rice Yields per Hectare Harvested for Selected Asian Countries, Five-Year Moving Averages, 1953-88 6.0 f4.O - ~Japan 5.0 - _ ~~~South Korea Ta,i--w'an-~ 4.0- 3.0 <3 ~~~~~~~~~Indonesia 2.0 1.5 ~~~~~ , . ^ _ _ < ~~~~~~~~India I I I ~ ~ ~~I I I I 1955 1960 1965 1970 1975 1980 1985 Year Source: IRM 1990. - 4 - shifting away from more traditional staples toward higher demands for TABLE 2. PRODUCTION OF SELECTED COMMODITIES livestock products, fruits, and IN SELECTED SOUTHEAST ASIAN AND MIDDLE vegetables. The sharp increases in fruit, EASTERN COUNTRIES, 1970-88 (MILLION MT) nut, and vegetable exports of selected COMMODITY/COUNTRY 1970 1975 1980 1985 1988 Asian and Near Eastern countries, FrTdy presented in table 3, indicate the rapidly Maasia 084 o784 0.96 1508 16137 increasing demand for these products. Philippines 2.04 2.74 6.15 6.18 5.11 TFsiland 2.37 4.10 7.18 5.43 5.56 Egypt 1.45 2.06 2.29 2.99 3.52 JIrdan 0.06 0.03 0.09 0.12 0.17 Moroso 1.45 1.27 1.67 1.75 2.16 Diversification Experience in Syr 0.34 0.52 0.74 0.99 1.43 Southeast Asia and the Middle East Turkey 6.29 6.37 7.73 8.00 8.92 Vegerals Indonesia 2.40 2.59 2.46 3.51 3.53 Recent changes have motivated Mlaysia 0.45 0.51 0.61 0.53 0.54 farmers and agricultural policymakers to Thilan 1093 2.50 2175 2.46 2148 search for more flexible and Egypt 5.30 6.40 7.30 8.97 10.02 environmentally sound cropping patterns, Morocco 0.96 1131 1126 1145 1.55 Syfia ~~~0.72 2.11 3.11 2.20 2.54 more diversified economic activities, and ZSya 9.17 10.29 13.03 16.60 16.87 more responsive production systems that can benefit from new opportunities aa 13.95 15.53 16.38 16.84 18.27 created by changes in technology and the Palip s 1.33 1.79 3.61 25063 0275 market. The focus on diversification Thailand 3.82 7.62 16.94 19.50 22.53 EVP ~~~~0.67 0.86 1.39 1.71 2.02 extends the discussions on agricultural Jordn - - 0.01 1.03 0'.05 development to areas that have not Morocwo 0.28 0.20 0.54 0.84 0.80 received much attention in the last 20 Turiey 1.92 2149 3100 4110 4135 years, during which international efforts P. focused on increasing the production of ndoneia 0.22 0.40 0.72 1.24 1.44 Malaysia 0.43 26 ~~~2.57 4.14 5.03 staple crops. The main objective now is Phlppsa 0. o.o216 0.01 0.03 0.04 to respond to the increasing concern - - 0.02 0.09 0.16 about the flexibility of agricultural Suar Indonesia Q.71 103 i.25 1.7 1.86 production systems. This requires Malaysia - 0.05 0.05 0.08 0.09 assessment of available opportunities ifor Phii- 1.93 2.39 2.27 1.61 1.34 modernizing agricultural systems through ES 0.53 0.55 0.67 0.90 1.03 diversification. The potential rfor Morocco 0.16 0.26 0.36 0.49 0.69 diversifcation. he potenial ifor Syria 0.03 0.03 0.05 0.06 0.02 widespread diversification in production, Tur1ey 0.65 0.99 0.94 1.40 1.41 marketing, processing, and agroindustry Po,- exists, but the process will not occur Indoaisa 75100 110100 17600 231800 445.00 automatically: technological short- Psiippins 96.00 136.00 226.00 190.00 237.00 Thailan 90.00 260.00 383.00 470.00 556.00 comings, institutional rigidities, and Egypt 85.00 94.00 117.00 165.00 112.00 policy distortions are major obstacles. Jordan 8.00 21.00 26.0W 55.00 64.0 Moro=c 31.00 46.00 99.00 96.00 102.00 The diversification process m Syria 14.00 20.C0 49.00 91.00 73.00 Southeast Asia and the Middle East has Turkey 105.00 171.00 247.00 271.00 290.00 been spurred by several factors: (a) Pa8 3 declining returns from cereal-based Malaysia 70.00 113.00 154.00 252.00 200 farming operations; (b) the availability of hi i 369.00 323.00 412.00 397.00 570.00 . . . ~~~~~~ ~ ~~~~~~ ~~~~Thailazi 210.00 175.00 267.00 378.00 334.00 advanced irrigation technology; (c) the development and adoption of improved a In h of neic Wm. Souce; U.S. Departme of Agricssture 1990. high-value crops; (d) increased domestic and regional demand for fruits, -5- vegetables, and livestock products, (e) the growth of private agribusiness involved in TABLE 3. FRUIT, NUT, AND VEGETABLE processing and marketing; (f) the shift of EXPORTS IN SELECTED SOUTHEASF ASIAN AND labor out of agriculture as employment NEAR EASTERN COUNTRIES, 1970-88 (MILLION USS) opportunities in other sectors became COULTRY 1970 1975 1980 1985 1988 increasingly available and rewarding; and (g) reduced government intervention in farmers' Malaysia 19.05 30.79 39.52 39.46 38.74 decisionmaking and the removal of Philippine. 29.58 113.08 213.51 230.73 266.75 distortionary policies that favor selected crops Thailand 76.89 275.32 900.41 814.18 1172.48 and limited commodities (Timmer 1988- Egypt 50.51 82.29 99.99 135.58 198.77 Jordan 9.59 24.39 49.13 31.02 27.69 Tsakok 1989; Barghouti 1990; Barghouti et Morocco 149.03 194.70 40.02 205.64 282.48 al. 1990; World Bank 1990a; Somel 1992). Syria 6.54 11.04 34.19 5.86 37.41 Agricultural policy reforms have been Turkey 137.63 277.48 786.93 744.16 1200.13 an important factor in the promotion of agricultural diversification. In particular, Source: U.S.Deparunent of Agriculture 1990. preferential treatment for cereal production, such as subsidized water and fertilizers, has frequently been removed or at least reduced. Public services in research, extension, credit, and marketing have expanded to cover production and processing of cereal and noncereal crops. For Southeast Asian and Middle Eastern countries, the main appeal of diversification is that the process identifies viable areas for investment in agriculture and rural development that may reduce the cost of price stabilization programs and transfer some of the burden of price instability to farmers and away from the budget. Because the large budgetary expenditures involved in the stabilization of agricultural prices are usually unsustainable in the long run, the promotion of more flexible agricultural systems can alleviate much of the political and economic need for such large budgetary expenditures (Iimmer 1988). During the 1970s, farmers in several Asian countries significantly increased total production of rice. In response to stable prices for rice and subsidized imports during that decade, rice output grew at an annual rate of 3.7 percent in Southeast Asia (IRRI 1990). The adoption of high-yielding varieties combined with the expansion of irrigation facilities and higher application rates of fertilizer led to much of the growth in production. Surpluses developed in the mid-1980s, forcing down rice prices in domestic and world markets. In response, farmers in some countries began, to varying degrees, to shift to the production of other commodities. For example, in Thailand, Sriarunrungreauang (1989) found that farmers increasingly diversified into the production of other agricultural commodities, such as vegetables, seasonal crops, cattle, and poultry, as rice prices declined. However, many farmers in the rice-based agricultural systems of Southeast Asia often experienced unexpected problems. Their efforts to shift to other crops and to invest in other agricultural enterprises were frequently hampered by traditional irrigation facilities and limited improved technological options for other crops. In addition, government policies; services (research, extension, and credit); and incentives for food crops have discriminated against the growth of high-value crops (Timmer 1988; Barghouti et al. 1990; World Bank 1990a). In the Middle East and North Africa, similar conditions are confronting farmers, particularly regarding the changing role of the agricultural sector and its contribution to domestic and export markets. Diversification of agricultural systems in this region has been made possible because of declining profits from cereal production, increasing scarcity of fertile lands and water, and substantial growth in demand for high-value crops (fruits and vegetables) in both domestic and regional markets (Tsakok 1989; Barghouti 1990; Somel 1992; Jaffee, forthcoming). The agricultural sector in several Middle Eastern countries has also witnessed remarkable changes and modernization. Much of this modernization has been made in the irrigation subsector. Irrigated agriculture in this region has benefited from a wide variety of products derived from the recent scientific advances in water and crop management, biotechnology and crop improvement, and marketing and -6- processing facilities. These changes have been made possible because of effective and somewhat aggressive private sector companies that became involved in introducing (a) modern production technology, such as plastic houses, tunnels and mulches, soil fumigation chemicals, improved hybrid varieties, drip irrigation systems, and soluble fertilizers and herbicides; and (b) efficient marketing technology, such as suitable packaging and storage facilities, modern market information systems, and transport equipment with refrigeration for long hauls. These companies provided farmers with the necessary support to adopt modern technologies required for the cultivation of suitable, long-life varieties of fruits and vegetables. Dimensions of Diversification Diversification issues can be approached at four levels: farm, regional, sectoral, and intersectoral.' At the farm and regional levels, there may be valid reasons for specialization. The process of diversification at these levels, however, is highly responsive to modern crop production and water management technology. Technical issues at these levels cannot be analyzed without considering research priorities, extension strategies, incentives, marketing, and other macroeconomic issues determined at the sector or intersectoral level. Farm Level Diversification is not a new concept for farmers. There is ample evidence that farmers are skilled in adjusting the mix of products they produce to reflect changes in the relative profitability and risk of each activity. Recently, however, there has been a definite trend toward more diversified farming systems, largely in response to changes in technology and the economic environment. Mixing farm enterprises rather than specializing is advantageous because it: * enables resource requirements, e.g., labor, to be spread more evenly over the agricultural season; * provides some protection against production and price risks because low returns from an activity one year may be partly compensated by a higher return than usual from another activity; * enables farmers to retain some flexibility in their capital equipment and managerial skills for exploiting potential improved market opportunities. It also enables them to adjust more quickly if a market weakens or fails. These benefits, however, may be offset by the loss in average profits incurred by not specializing in the most profitable enterprises. The advantages of economies of scale from specialization are usually compromised when farmers diversify their enterprises. Where farmers are exposed to changes in world markets, many of them are presumably already making appropriate adjustments to their cropping patterns, at least to the extent permitted by technological and market considerations. Crop diversification at the farm level is also determined by on-farm constraints, the availability of extension services, the opportunities for contract farming and mechanisms for organizing farmers, and the responsiveness of farmers to the various signals they receive. The role of government in this process is twofold: to do whatever it can to facilitate appropriate changes at the farm level; and to deal with various externalities (e.g., unstable food prices, conflicts with food security objectives) that might result from changes in the production mix at the regional and aggregate levels. -7- Regional Level The scope for diversification is affected by the ability of each region (in the country) to specialize in specific enterprises based on comparative advantage; the formation of farm belts in the United States is a good example. Similar zoning of agricultural regions is also found in Australia and some of the EC countries. Regional diversification is influenced by several technical factors. Its scope is determined by agroclimatic conditions, including suitability of the natural environment (soil, weather, water) for expanding crop production or advancing specialized technology. Crops grown in marginal environments, such as wheat in the southern section of the Gezira Scheme in Sudan, or on marginal rainfed areas in Jordan and Syria, may not be attractive for further investment until maximum levels of production are achieved in the high-potential zones. In Southeast Asia, as reported earlier, large-scale irrigation systems were designed and constructed mainly for rice production: they are not geared toward crop diversification at the farm level. In irrigated areas, diversification from rice paddy to nonpaddy (upland) crops is limited by the structure of existing irrigation facilities, growing seasons, soil types, drainage capabilities, and land preparation requirements for rice. The prospects for diversification in the wet season are thus limited in most irrigated areas, particularly those with poor drainage, because the supply of water during the wet season significantly exceeds the water requirements of upland crops. It is possible, however, to increase the flexibility of irrigated farming in the dry season. There are other inherent difficulties in switching between irrigated rice and upland crops, but farmers manage to do so when adequate economic incentives exist. In areas where there are suitable soils and sufficient water control, farmers can diversify production during the dry season. Malaysia, Sri Lanka, Indonesia, and the Philippines have more dry-season irrigation than do other countries in Asia (Barghouti et al. 1990). Unlike farming in irrigated areas, however, farmers in rainfed areas must carefully adjust the choice of crops and timing of planting and harvesting to the weather. What crops can be grown depends critically on the amount and timing of rainfall. No cropping pattern is determined entirely by technology and ecological features. Part of the choice is determined by cost of production, market demand, and consumers' willingness to pay. If consumers are willing to pay enough, virtually any crop can be grown anywhere at any time of the year. Vegetables grown in plastic greenhouses in Jordan, Egypt, and Israel in the winter are profitable because high-income urban consumers are willing to pay high prices for fresh produce. The production environment can be made more flexible through sufficient investment in irrigation, water control, and relevant production technology and crop management under regional agroecological systems. The objective of diversification is to increase the flexibility of cropping systems so that a variety of activities can be undertaken in response to changing technology and market conditions with relatively low adjustment costs. The level at which such flexibility is created determines the costs involved: creating flexibility at the individual farm level is more costly than creating flexibility in cropping patterns for the country as a whole. Regional specialization is possible within an overall pattern of national diversification, thus drawing on differences in agronomic potential and economies of scale in marketing that are inherent in well-functioning commodity systems. Unfortunately, regional specialization cannot eliminate the vulnerability of farmers to price changes for the particular commodity they produce. The regional diversification problem is thus twofold: (1) What is the functional relationship between costs and flexibility? and (2) How should willingness to incur these costs relate to the degree of price stability for the various commodities that might be produced once the investments are made? Neither of these questions has general answers: the costs of creating flexibility depend on the regional environment. Irrigation systems in several parts of the world have created some of the most flexible farms: crop choice varies radically by season and by year. At the same time, the limited flexibility of rainfed agriculture forces farmers to grow only a limited number of crops regardless of the market prospect for these crops relative to that for other crops. - 8 - Concurrently, regional specialization may provide the opportunity for efficient development of entire commodity systems, from input production and marketing to downstream processing, thus fostering both vertical and horizontal diversification with extensive backward and forward linkages. The government has a positive role in stimulating diversification and increasing regional competitiveness by making broad-based investment in regional research and extension and rural infrastructure-such as irrigation, roads, and communications-to facilitate the movement of goods and information to small-scale and widely dispersed farmers. Intersectoral Level At the intersectoral level, diversification implies that rural populations seek better income-earning prospects off the farm or outside agriculture. Such changes are desirable as a long-run response but might be quite tumultuous in the short run. The implications at this level are complex. The countries involved have to increase their focus on human resources. Investing in human capital has a high social rate of return. Heavy investments should be directed toward extending nutrition programs to subsistence farmers and other rural poor; expanding lending for education at all levels; and increasing lending for rural services, industry, and technology. Increased investments in nutrition will significantly increase the labor productivity in these countries. Higher levels of educational attainment for the rural population are important to the efficient transfer of new technologies to the farm sector. They are also needed by the agricultural labor force to adjust to new labor market opportunities as development proceeds-the ultimate diversification problem but one which countries are already facing. For those without access to adequate land or on-farm employment, diversification out of agriculture altogether is needed. The rural nonfarm sector is increasingly becoming an important source of jobs for members of farm households. Although this change is consistent with desired patterns of structural transformation, such shifts in labor allocation may engender numerous problems. In economies burdened with inefficient, inward-looking industrial sectors, job opportunities outside agriculture are severely limited. If the rural economy is unable to absorb idle or displaced labor from agriculture, the flow of unskilled rural workers into low-wage urban informal sectors will increase. In countries that have to increase their capa^ity to absorb this labor, agricultural diversification may provide the appropriate mechanism. The diversification process has good potential to increase the role of small- and medium-scale rural industries, marketing, construction, and other labor-intensive services. Governments will need to give more attention to planning for expanded industrialization within rural areas, including policies required to create a favorable environment for private investors to move enterprises to rural communities. Diversification at the farm, regional, or intersectoral level is highly influenced by several factors, including: 1.. government policies. At all three levels of concern-the farm, the agricultural sector, and the economy-governments intervene to buffer or reinforce pressures and opportunities in world markets (Timmer 1988; World Bank 1990a). Sometimes, governments in developing countries adopt market-oriented macroeconomic (e.g., trade, exchange rate, monetary, and fiscal) and agricultural sector (e.g, price) policies that are designed to stimulate agricultural diversification. Often, however, these policies may conflict with the pursuit of other priorities, such as maintaining the stability of food prices and food security, and are thus not effectively pursued. In other cases, government policy interventions may actually impede the diversification process: successful cereal price stabilization programs seem to have that impact. Consequently, further studies are urgently needed to achieve greater consistency in policies to promote diversification as well as other policy objectives, such as food security. In particular, more studies are necessary in the design of effective agricultural diversification policies that are consistent with price stabilization policies for basic food staples; -9- 2. input supplies for all crops. A side effect of the heavy attention to cereal intensification programs in most developing countries has been the development of an input supply industry and infrastructure designed to provide small cereal farmers with fertilizer, pesticides, seeds, credit, and machinery. Creating a modern and efficient input, production, and marketing system has not been as easy as developing the high-yielding seeds themselves or building the irrigation systems that make them productive. Secondary crops have little of this supporting infrastructure to deliver crucial inputs. Diversification away from cereals to high-value crops, such as fruits and vegetables, will require more management for the distribution of inputs. Furthermore, quality and timely production is usually critical to the economic success of these high-value crops (World Bank 1990a); 3. institutional support. An agenda for diversification requires governments to take measures that would have an impact on various institutions and would encompass research and extension, irrigation and drainage, marketing, private sector involvement, and rural education (World Bank 1990a). Components of a New Agricultural Development Strategy The challenge to agriculture, therefore, is to sustain farming of traditional crops while expanding into a more flexible, diverse agriculture. The objective of this process should be to keep the adjustment costs relatively low and yet diversify the agricultural economy away from its heavy reliance on traditional crops and toward other items with more favorable income elasticities of demand: livestock products and fruits and vegetables, for example. Such demand-led diversification should be part of a broader process of developing an appropriate mix of agricultural policies and production technologies. Governments tend to allocate resources to develop commodity-specific programs and set production targets for the crops for which the country, judging from world market prices, might have a comparative advantage. Target-oriented diversification programs, however, risk diversifying farmers into crops whose prices may fall (or even worse, the added output causes prices to fall), with farmers no better off than before. To avoid this situation, it is necessary to regard diversification as a process of adjustment, rather than the establishment of targets for cropping patterns, and to shift resources at the margin from specific crop or irrigation investment to more general and less crop-specific rural and marketing infrastructure and institutions. Such efforts may also help close the gap that exists in most agricultural development strategies, i.e., the "intermediate" agricultural sector that involves neither a basic food staple nor a traditional export crop. The main lesson from the experiences in Southeast Asia and the Middle East is that investment in agriculture should preferably be for broadly based sectoral operations and less for commodity-specific, production-oriented projects. Furthermore, the success of government investment is contingent upon major changes in government policies, incentive programs, and agricultural services and upon the development of technological innovations promoting more efficient production. Policy Framework for Dlversification Agricultural diversification will be facilitated by the increased use of multidisciplinary approaches in designing investment strategies, the removal of distortionary policies, the elimination of land use restrictions and improvement of land tenure security, the expansion of private sector involvement, and the avoidance of a "pick-the-winner" project approach. A balanced mix of skills is needed to plan strategies for agricultural diversification. There should be a multidisciplinary approach to project design and appraisal that goes beyond economic policy analysis and includes technology assessment in - 10- production, processing, and marketing. Project design must incorporate advanced technological concepts that could increase farmers' options for production, as well as mechanisms for effective management of agricultural input supply, marketing, and processing. Expertise is required in irrigation and water systems design and management, natural resources management, rural industry and infrastructure, and integrated farming systems. Price stabilization policies should not be at the expense of secondary crops. Governments pursuing a policy of domestic price stabilization should minimize price distortions by maintaining appropriate relative prices between primary and secondary crops. Restrictions on land use and landholding size should be adjusted to encourage investments in land improvements and flexible farming systems. Land tenure arrangements should provide domestic and foreign investors land use security. Legal and institutional barriers that inhibit the economic activity of the private sector should also be removed. Attention should be given to the reduction of public sector monopolies on marketing, licensing requirements for agricultural production and processing, the strengthening of property rights, and the rationalization of foreign exchange. Private sector participation may also be fostered through the promotion of joint ventures and contract farming. Governments should provide the right set of incentives and legal and institutional arrangements to attract foreign and domestic investors to participate in joint ventures and contract farming for the production of high-value crops and livestock. To accommodate potential changes in farming patterns and to provide increased opportunities for off-farm employment, governments should devote more attention to promoting investment in small- and medium-scale industrialization in rural areas. Finally, it is necessary to perceive diversification as a process of adjustment, not the selection of specific cropping patterns. Farmers should be provided with options that encourage them to grow crops for which they have a comparative advantage based on agroecological and economic factors. Essential Support Services The focus of diversification is on improving the responsiveness of farmers to changing market conditions and new technologies. Therefore, emphasis should be placed on providing farmers with a wider range of technological options and various facilities to pursue diverse production patterns, as well as off-farm employment that will maximize household income. Thus, the scope of research, extension, marketing, and credit should be expanded from being crop-specific and production-oriented to cover both primary and secondary crops and livestock, supporting not only production but also processing and marketing. Extension workers, for example, should be able to deliver a range of agrotechnical messages for a variety of crops and to provide marketing and price information, as well as information about off- farm economic activities. Also, assistance should be provided to traders and processors. Governments should encourage investment in the modernization of marketing facilities: they can improve marketing by developing rural buying stations and commodity exchanges; by promoting the construction of wholesale markets; and more generally, by investing in improved roads, communications systems, and storage. Market information services and access to international markets should also be expanded. Governments should aim at improving the quality and availability of market information for export crops; radio broadcasts, for example, can provide market information on traditional and nontraditional commodities. Governments should finance market research and information dissemination, especially for small-scale farmers, and also encourage the establishment of growers' organizations. The identification of specific requirements for and market niches in both domestic and foreign markets is also needed. Governments should try to increase access to restricted international markets and devise a strategy that improves their bargaining position on agricultural trade issues. Standards for quality control will help facilitate private participation. Thus, governments should assist marketing agencies in establishing standards and grades to lower transaction costs and facilitate increased opportunities for domestic-foreign marketing and processing. - 11 - Greater emphasis on vertical diversification into the processing of both traditional and nontraditional agricultural commodities will help increase the value added of agricultural products and should be supported by governments. Investment in agriculture should not end at the farmgate, but should include postharvest handling and processing and continue to the final consumer. Last, diversification of the agricultural sector will require a wider variety of technical skills. Thus, investment in rural education is essential. Vocational education programs in rural areas should be expanded to provide improved training for on-farm jobs and also for off-farm jobs in the industrial and service sectors. Role of New Technology Agricultural diversification requires a comprehensive knowledge of agroecological resources, particularly information about agronomic potential and suitability for profitable production and marketing. Thus, lending for agriculture should include support for studies that improve the knowledge base regarding local agroecological resources (i.e., soil mappings, aerial surveys, hydrological surveys, climatological data, and past and present cropping practices). Effective diversification also requires the development of an appropriate analytic framework to evaluate its performance and progress. The economic analysis of diversification should include both the benefits and costs of increased flexibility of the agricultural system and the enhanced ability of the agricultural system to mitigate the effects of price fluctuations. To achieve success, diversification will often require new or improved technologies, such as modern water delivery and drainage systems. Investments will be necessary to modernize water control, upgrade existing irrigation systems, and develop new approaches to increase the flexibility of farming systems. The implementation of efficient water charge mechanisms is also required to improve the management of water resources for flexible cropping systems. Specific technologies and practices can enhance local comparative advantage in production and marketing, and research strategies should also emphasize the development of new innovations that reduce the unit cost of production and increase the quality of outputs to be more competitive in domestic and foreign markets. Concurrently, there is a need for more environmentally sound technologies to ensure sustainable agriculture. Attention should focus on the minimization of agrochemical requirements because of their negative environrnental impacts and on the design of farming systems that reduce soil erosion, such as appropriate intercropping and agroforestry. Emphasis should also be placed on integrating crop- livestock farming systems and improved feed grain production and feeding efficiency along with increased use of agricultural by-products and waste. The Next Step Without productive agricultural technology supported by responsive policies for crops and livestock, diversification programs cannot succeed in the long run. Policy reform and research in agricultural production, however, are not always sufficiently well articulated to effectively support the desirable process of diversification and modernization of agriculture. Even with more flexible and broader agricultural research programs underway, which crops to highlight is difficult to ascertain, and the potential to diversify into noncereals and to use price policy for the noncereal crops themselves to enhance their profitability and adoption could be somewhat paradoxical. Different countries have different approaches to analyzing and assessing the costs and benefits of increasing the flexibility of agricultural systems. Country-level research on diversification, currently being undertaken by the World Bank, will provide detailed information about how each country has approached the policy and technical issues of diversification and the analytic tools employed to articulate them. This subject has not received adequate attention from academicians and researchers. Demand for - 12 - drastic changes in agricultural trade and the increasing political pressure to remove agricultural subsidies may motivate researchers to pay more attention to the issues of diversification in both developed and developing countries. There have been dynamic changes in the agricultural sector, particularly during the last two decades. Increased assessment of the implications of these changes for investment in agriculture-and 'more generally for development policies-is needed, bearing in mind that the challenge is to present policymakers with options designed to sustain the profitability of the sector in a changing world economy. Notes 1. This section is drawn from Timmer 1988 and Barghouti 1990. 2. This discussion is drawn from World Bank 1990a and Barghouti 1990. References Baanante, C.A., B.L. Bumb, and T.D. Thompson. 1989. The Benefits of Fertilizer Use in Developing Countries. Muscle Shoals, Ala.: International Fertilizer Center. Barghouti, S. 1990. "Agricultural Diversification: Implications for Crop Production-The Case in the Middle East and North Africa Region." HortScience 25(12): 1499-1506. Barghouti, S., C. Timmer, and P. Siegel. 1990. Rural Diversification: Lessons from East Asia. World Bank Technical Paper 117. Washington, D.C. Flinn, J.C., and K. De Datta. 1984. "Trends in Irrigated-Rice Yields under Intensive Cropping at Philippine Research Stations." Field Crops Research 9: 1-15. Hayami, Y., and K. Otsuka. 1991. "Beyond the Green Revolution: Agricultural Development Strategy into the Next Century." Paper presented at the conference "Agricultural Technology: Current Policy Issues for the International Community and the Wcrld Bank," October 21-22, Airlie House, Virginia. International Rice Research Institute (IRRI). 1990. World Rice Statistics. Los Banos, the Philippines. Jaffee, S. Forthcoming. "Contract Farming in the Shadow of Competitive Markets: The Experience of Kenyan Horticulture." In M. Watts and P. Little, eds., Peasants under Contract: Contract Farming and Agrarian Transformation in Sub-Saharan Africa. Ithaca, N.Y.: Cornell University Press. Pingali, P., P.F. Moya, and L.E. Velasco. 1990. The Post-Green Revolution Blues in Asian Rice Production: The Diminished Gap between Experiment Station and Farmer Yields. International Rice Research Institute, Social Sciences Division Paper 90-01. Los Bainos, the Philippines. Somel, K. 1992. "Agricultural Diversification in the Middle East and North Africa." This volume, p. 39. Sriarunrungreauang, S. 1989. "The Impacts of the Declining Price of Rice on Production, Allocation of Farm Resources and Farm Incomes: A Case Study of Small Rice Farms in Amphoe Don Chedi and Amphoe U- Thong, Changwat Suphan Buri." Master's thesis, Department of Agricultural Economics, Kasetsart University, Bangkok. - 13 - Timmer, C.P. 1988. 'Crop Diversification in Rice-Based Agricultural Economies: Conceptual and Policy Issues." In Ray Goldberg, ed., Research in Domestic and International Agribusiness Management, vol. 8. Greenwich, Conn.: JAI Press. Tsakok, I. 1989. "Rationalization for an Agricultural Export Diversification Project: The Case for Tree and Selected Horticultural Crops." Ghana Agricultural Diversification Project Working Paper, West Africa Department. Washington, D.C.: World Bank. U.S. Department of Agriculture. 1990. World Agriculture: Trends and Indicators, 1970-89. Washington, D.C. World Bank. 1982. World Development Report 1982. New York: Oxford University Press. 1988. World Development Report 1988. New York: Oxford University Press. 1990a. Agricultural Diversification, Policies and Issues from East Asian Experience. Policy and Research Report 11. Washington, D.C. 1990b. World Development Report 1990. New York: Oxford University Press. TECHNOLOGICAL CONSIDERATIONS IN AGRICULTURAL DIVERSIFICATION H. G. Zandstra Agricultural diversification has several facets. Schuh and Barghouti (1988) recognize four levels of diversification: farm, sectoral (crop versus livestock), regional (through production specialization), and national (through urbanization and reduction of the agricultural labor force). Diversifying agriculture also has environmental aspects, relating to the stability of production systems, reduced pest pressures, and increased possibilities for soil nutrient cycling. Finally, diversification affects different members within a family or community in terms of their contribution to production as well as the distribution of benefits among them. This presentation deals primarily with the technological constraints to and opportunities for agricultural diversification. Because of the wide range of factors involved in agricultural diversification, it will be general, only touching on the most important environmental and technological aspects associated with diversification. Although agricultural diversification is not confined to rice lands or the introduction of alternative crops, because of the importance of rice, many of the technical aspects discussed here deal with diversification of rice-based cropping systems. Basis for Diversification A current hypothesis holds that reduced demand for staples, particularly rice, and reduced grower benefits justify special interventions that encourage agricultural diversification. This hypothesis assumes that there will be no problems with future rice supply and that demand will grow at lower rates than supply. Given the lead time for technological development and infrastructural change required for diversification, it is important to verify the assumptions made about future demand and the prospects for an increased supply of rice. The premise that consumer prices are low because of limited demand seems to be contradicted by the inability of an important part of the population to purchase the rice it needs. This unexpressed demand varies among countries in Asia, but it is undoubtedly increasing. In its strategy statement IRRI: Toward 2000 and Beyond, the International Rice Research Institute (IRRI) foresees that Southeast Asian rice-growing countries will have to satisfy a demand growth of 2.1 percent annually. If rice productivity continues to decline as a result of degradation of the paddy environment and the deterioration of irrigation systems, short supplies and higher rice prices are likely. Consequently, the wisdom of encouraging enterprise changes toward higher-value products, which also demand considerably greater infrastructural support and production inputs, may be questioned. Diversification or Specialization? Although the results of agricultural diversification are well understood, its precise definition and measurement present a challenge. Branch et al. (1988) employed Theil's diversity index, which is based H.G. Zandstra is director general, International Potato Center, Lima, Peru. - 16 - on the concept of entropy: minimum diversity is the practice of a single system over the universe, and maximum diversity is an equal distribution of all enterprises. When dealing with crop diversity, land utilization is generally the unit of measurement. This naturally becomes less attractive when the enterprise set includes aquaculture, forestry, on-farm utilization of agricultural produce, or crop-animal systems. In such cases, the use of other resource allocations or their aggregate may be more appropriate. The area of measurement, whether farm, province, region, or nation, also becomes relevant. One could as easily conceive of a diversity of regionally specialized, highly mixed farm systems as of a highly specialized, uniform, provincial land use pattern that differs greatly among provinces at the national level. In the first case, farms are highly diverse, but regional diversity is low because the farms are all similarly diversified. In the second case, farm and intraprovince diversification is low, but among provinces at the national level it is high. This may seem a spurious word game, but an understanding of specialization and diversification processes and their role in forming the structure of a country's agriculture is important. The forces that lead to diversification need to be defined. They may include low market demand for the staples produced, high production or market risks, high labor supply, and broad adaptation of crops and management techniques. The factors leading to specialization may be significant environmental constraints to production; high margins of economic return; narrow technological adaptation; and regulation of prices, inputs, or crop size. As economies become increasingly differentiated and developed, national agriculture diversifies, whereas regional specialization will generally increase. In considering the merits and technological potential of agricultural diversification, a better understanding of the factors controlling it may be needed. Opportunities for Diversification Farmer opportunity for diversifying production is limited by several land, climatic, and socioeconomic factors. It also depends greatly on the development of technologies that influence both the production process and the management of alternative crop and livestock enterprises. Input and Product Markets Modifications of land, crop varieties, animal housing and feeding techniques, and other practices can permit adaptation (at some cost) of many new enterprises to somewhat hostile environments. IRRI's experience, and that of several participants in the Asian Farming Systems Research Network, has been that the absence of product markets is the factor that has constrained diversification the most. This has been the case with sorghum in upland rice regions, and with cowpea and soybean in postmonsoon lowland conditions and upland rice regions in the Philippines. It is severely limiting the introduction of groundnut in several regions and has frustrated research on agronomically promising annual oilseed crops such as sesame. Consequently, the major cropping patterns associating upland crops with rice are rice/wheat, rice/soybean, rice/mungbean, and maize/rice. Crop diversification requires added inputs (figure 1, p. 18). Undoubtedly, investment in rural infrastructure can provide inputs to farmers, but they may only partly improve product markets. This particularly applies to commodities that are widely traded internationally and benefit from high production efficiencies or hidden subsidies in other countries. Crop diversification has therefore been predominantly in the form of horticultural crops in periurban areas, supplying vegetable markets in growing cities. Similar experiences abound in crop-livestock systems. These are often indigenous to rice-growing regions, and many technologies exist for their expansion, particularly in rainfed and upland rice ecologies (Calub and Roxas 1988). Farmer adoption is constrained, however, by high production costs and a demand suppressed by high product prices. - 17 - Water An absence or excess of water defines the cropping seasons of rice-based farming systems and has been widely studied for rainfed rice conditions (Zandstra 1982; Morris and Zandstra 1979). The accumulation of water in wetland rice fields of finer textures (silty clay loam, clay loam, and clay) and lower landscape positions limits planting of upland crops. During the monsoon period upland crops cannot be grown without major land modifications, such as the sorjan ditch and dike system or internal field drainage. The effects of rainfall distribution, soil texture, and landscape position on the duration of the pre- and postmonsoon period during which upland crops can be grown are well understood (figure 2). These measures show that in most wetland fields, even those with higher seepage and percolation rates, the duration of pre- and postmonsoon upland crop-growing periods is often too short to accommodate even short-duration grain legumes (table 1). Only in wetland fields that are high in the toposequence and approximate upland soil conditions can such crops be dependably grown, and then only when the onset of the monsoon and its termination are gradual (figure 3, p. 21). The availability of dry-season irrigation will extend the period for upland crop production. This period will be bound by the withdrawal of water at the end of the rice-growing season (e.g., the probability of the field being saturated should be less than 0.5 and the probability of flooding less than 0.3) and the onset of saturated conditions or flooding at the start of the rains. There is also the maintenance of the irrigation system to be considered. Day Length and Temperature Short day length and high temperatures can limit crop production in the severe dry seasons at latitudes above 15 degrees. Particularly for grain legumes, these conditions lead to excessively early maturity, early flowering, and short grain-filling periods, as well as sterility and other high-temperature crop damage. Woodhead (1989) indicated that on farmers' fields in Central Luzon, extreme air and soil temperatures reduce mungbean yield potential from 2.0 to 1.5 tlha. For wheat, both in the tropics and semitropics, planting after the optimal date (early December) typically reduces yields by 2.0 tlha at a rate of 0.1 t/ha per week of delay. Although these factors limit crop choice and place high demands on crop management, they can often be overcome by varietal improvement and special crop management techniques (Lantican 1982). Soil Conditions Soil conditions become especially important for upland crops after wetland rice (Syarifuddin 1982; Woodhead 1989). Because of puddling and flooding, the soil is saturated after the rice harvest and remains reduced for some time thereafter. Although the bulk density of rice soil after puddling at the end of the rice crop and certainly after field drainage decreases immediately, bulk densities nonetheless remain high. After drying and shrinking, accompanied by soil cracking, high soil densities impede root growth and water movement, particularly in fine-textured soils. These physical soil constraints are accompanied by a rapid shift in the soil's redox potential, which consequently changes the soil chemical conditions, particularly phosphate availability. These conditions greatly influence upland crop establishment after wetland rice and, even when irrigation is available, demand special management measures. They also bias farmer choices strongly toward sequential rice cropping where possible. During the 1970s, shorter-duration varieties and improved rice establishment techniques provided farmers in Iloilo, the Philippines, with new opportunities to intensify their cropping. They increased their double rice cropping more readily than the frequency of upland crops after rice (table 2). This is largely attributable to soil-related constraints, which bias land use toward sequential rice cropping. - 18- Figure 1. Costs and Returns of Farmers' and Experimental Cropping Patterns, Iloilo and Pangasinan, Philippines, 1975-78 12,000 - RRR(E) v _RRR(F) * uuR) 10,000 - RRU) Qv * RRU(E) RR(E U;~~~ RR(E) , 8,000 - RR /URU) v Ioilo RU(E) * Pnaia ; Omitted 6,000 - RRU RU(E) F = farmers' pattern RR(F)* R = rice - ~~~~~~~~~~~~~~~~U = upland crop 4,000 - URUR) E = experimental ptn _ RU(F) RU(F) Y p=2.19X 5(-F257 r2 =0.99 2,000 - R R(F) Ye =1.95X * -370 r 2 =0.97 0- I I I I 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000 variable costs (P/ha) Source. Price 1982. Figure 2. Growing-season Components of Wetland Rice-based Cropping Systems Field location Premonsoon Rkce Postmonsoon Soil High Lgh (well-drained) s l Medium Medium (silty loam) iRke (cLay loam) * 4 > ':g . * 0 (silty dayloam) Low Hg h humidity O. (enrichment) Heavy * ' ~~~~~~~~~~~(day) (Siltylday) Dry Mist U lablyvet Moist t D Onset Rain End of End of of rain causes period of rain (75 mm) saturadon santuon soil dry Source: Zandstra 1982. - 19 - TABLE 1. AVERAGE DURATION (DAYS) OF SELECTED GROWING-SEASON COMPONENTS FOR TYPICAL WETLAND FIELDS AT FIVE PHILIPPINE SITES CRITERION CABANATUAN ILOILO LAOAG IBA ZAMBOANGA Premonsoon dryland period (H = 5 mm) 51 53 45 27 136 Wetland rice period (H = 100mm) 181 202 146 192 106 Total rice period (H = 100 mm) 226 249 179 216 251 Postmonsoon dryland crop period (H =S mm) S1 78 76 71 102 Annual total 249 281 213 240 295 Note: SP index = 2. Representative of a typical clay loam wetland field. Source: Zandstra et al. 1982. Soil conditions most suitable for crop diversification occur in upland rice-growing regions. The constraints in these regions are predominantly access to markets, soil erosion, soil acidity, and infertility. Technologies for Agricultural Diversification There is a wide range of approaches that can be taken to increase farm-level diversity. These include (a) mixed-farming approaches (crop-livestock systems, rice-fish systems); (b) addition of nonrice crops (premonsoon, postmonsoon); (c) crops substituting for rice (upland systems, wetland systems); and (d) green manures. Crop-Livestock Systems Livestock's most important contribution may be that it provides a market for crops. Examples are the use of forage from field bunds and trees in wetland regions (Nitis 1982) and the recent work on using unsalable cowpea, which grows well in upland and postmonsoon wetland conditions, for making high-protein concentrates for pigs or poultry (Gerpacio 1988). There are numerous research needs and possibilities for increasing the livestock component of rice farms, primarily addressing the production of forage and the use of crop by-products for animal feed. These approaches almost always use land or short-growing-season components not used for crop production.' Rarely will smallholders substitute staple crops to produce animal feed, but they may use pre- and postmonsoon rice periods of 40 to 50 days for forage production. Introduction of these crop-livestock systems, however, places new demands on transport and processing facilities. Rice-Fish Systems The use of rice-fish systems has become less common since the intensification of rice production. In Asia, this system is most common in Indonesia, where about 100,000 ha, or nearly 1 percent of the wetland rice area, are dedicated to this system. As with crop-livestock systems, there are many - 20 - opportunities for rice-fish production, and the potential for TABLE 2. AREA CULTIVATED IN VARIOUS CROPPING diversification of the farm PATTERNS, ILOILO OUTREACH SITE, PHILIPPINES, 1974-79 enterprise can be greatly AREA CULTIVATED (%) enhanced. The International CROPPING PATTERN 1974-75 1975-76 1976-77 1977-78 1978-79 Centre for Living Aquatic 2 or more rice 5 20 38 49 45 Resources (ICLARM) and IRRI are collaborating with several I rice + I or national research institutions in more upland crops 11 28 30 17 31 Asia to increase knowledge of this system and to evaluate the 2 or more upland crops 2 5 12 6 8 system and to evaluate the potential for its improvement and 1 rice + fallow 82 47 20 27 1.4 wider application. 1 upland crop + fallow - - -- 1 2 Addition of Nonrice Crops Note: The 1974-75 data represent average results of a 205-farm baseline survey For the premonsoon conducted in January 1975. Data from 1975-79 came from a farm recordkeeping study of 45 farmers selected randomly from the baseline list. growing season, the main Source: Price 1982. challenges in rainfed systems are early land preparation in dry conditions, flooding during flowering and grain-filling stages, and high rainfall and humidity at harvest. These limit production to regions with a gradual onset of the rains, generally those within 10 degrees latitude of the equator. It also strongly favors crops tolerant of late flooding, with multiple product markets and low vulnerability to humidity. The most common premonsoon crop is maize. Other widely used crops are leafy vegetables and green beans. Techniques for wider adaptation of this system are ridging (as in Java) or surface drainage of the field. These can extend the effective flood-free growing season by several weeks. Ruminant production combines well with premonsoon maize because early flooded maize stands serve as a source of forage. It also provides a market for short-duration fodder crops. The premonsoon season can also be extended by installing wells (reservoirs) within the rice field, thus ensuring drainage and providing an opportunity for extension of the fish-rearing period prior to rice planting in fish-rice production systems. In wetland rice areas, the postmonsoon growing season provides the greatest potential for introducing nonrice crops. The length of this season has been extended by the adoption of short-duration rice varieties. As described above, for finer-textured soils in the lower part of the toposequence, flooding hazards are great. Rather than wait for the landscape to drain and lose the late-season rains needed for the nonrice crop, field drainage can effectively extend the growing season. Because of the high (2:1) clay content and puddling of the soil of the preceding rice crop, hydraulic conductivity will be low and drainage will be difficult. Modifying land preparation for rice (less or no puddling) to develop a better-structured soil for more rapid drainage has been considered. This leads, however, to serious increases in weed competition and a higher frequency of drought stress during the rice crop. Other technological aspects relate to the establishment of nonrice crops in postmonsoon wetland fields. Syarifuddin (1982) and Woodhead (1989) found that there is a severe yield penalty for sowing too soon after draining fields and for delaying establishment too long, with yield loss caused by poor stand establishment (figure 4). Farmers' management options, therefore, are to carefully time the last rice flood irrigation and the date of draining prior to rice harvest. In addition, the use of mechanical planters, such as the inverted T seeder and the prototype "slit seeder," can improve stand establishment. There are several additional technological aspects related to the production of postmonsoon crops after rice in wetland fields. These include deep tillage and placement of fertilizer to increase rooting depth; the use of mulches for conserving water and reducing soil temperatures; water conservation for - 21 - Figure 3. Rainfall Accumulated after April 1 (Onset) and Still to Occur before April 1 (Termination) for Northern (Laoag) and Southern (Zamboanga) Philippines 600 - 500- Onset Termination 300 - aa 12a 200- 100 Zamboanga 0 - I I I I I I I 15 15 15 15 15 15 15 15 15 15 15 15 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Note: p = .7. Source: Zandstra 1982. Figure 4. Relationship between Soil Moisture Content and Soybean Establishment on No-tillage Plots after Flooded Rice 100 - 80 - ~60- A Y 1073 + 159x 40 - R 2 = 0.92 Y = 16301 + 133 x R2 = 0.96 Permanent 20 - wilting Field point capacity 20 30 40 50 60 70 80 90 100 Soil moisture content (ODW) within the 0-5 cm layer at planting Note: ODW = oven dry weight. From IRRI, 1978 dry season. Source: Syarifuddin 1982. - 22 - dry-season cropping; and varietal selection for tolerance of early flooding, late drought stress, and diseases such as damping off for legumes and powdery mildew for mungbean (Lantican 1982; Mew and Elazegui 1982). Crop Substitution for Rice Although changing from rice to other crops is agronomically easiest in upland rice ecosystems, local consumption requirements and market constraints generally hamper the adoption of alternative crops. In addition, high humidity and rain limit the production potential of most annual crops, except root crops and maize. The major constraint in this ecology, however, is the severe land degradation and soil losses associated with annual cropping, particularly where slopes are steep. This leads to an instability of land use and the need for farmers to shift their cultivation to other fields (figure 5). The use of contour bunds suited to farmer labor and cash resources is a promising technology for regeneration of these soils (Fujisaka and Garrity 1989). The bunds reduce soil loss; lead to stable terraces; and can provide fodder, fuel, or other sources of income. Most important, the built-up soil productivity leads to reduced production risks and increased farmer incentive to invest in inputs. This again encourages the shift to higher-value crops and more frequent inclusion of perennial cash crops in the production system. As part of the Upland Rice Ecosystems Program, IRRI scientists continue to search for the most appropriate contour hedge systems for different land and farm resource characteristics. Forage crops, particularly perennials and forage trees, hold promise for diversifying upland rice farms. They can provide a high-value product that may be easier to market in the absence of strongly developed infrastructure. The technology development for mixed farms has, however, only recently received the attention it merits. In irrigated rice ecosystems, substitution of rice with other crops has greatest potential in lighter- textured soils with access to dry-season irrigation. The increasing cost of water and the system-wide benefits of reducing per hectare irrigation requirements make this an attractive alternative. The most common system of this sort is the rice/wheat system practiced on about 9 million ha in regions generally north of 20 degrees latitude. In most cases, the wheat-growing season is not suitable for rice because of low temperature constraints. The expansion of rice/wheat systems toward more tropical conditions is a distinct possibility as wheat varieties better adapted to high temperatures and humidity become available. IRRI and the International Center for Maize and Wheat Improvement (CIMMYT) collaborated with national agricultural research centers in surveys and analyses of this important system. They identified that low and unsustainable productivity was caused by crop-related constraints such as delayed wheat establishment, pest losses to both wheat and rice, inadequate soil and water management (mainly for wheat), and the low profitability of small-scale operations. System-level constraints included pest and disease carryover between crops, nutrient imbalances, and soil structural changes over time. Whereas the yield gap between farm and research conditions has rapidly declined in irrigated rice areas, the gap for rice/wheat systems is about 5 t/ha, evenly distributed between both crops. In the equatorial humid tropics, dry-season production of maize, soybean, mungbean, and sesame offers much promise. Land preparation, crop establishment, and fertilizer application techniques are required that compensate for the high bulk density, low hydraulic conductivity, and reduced soil conditions that remain after the rice crop (Woodhead 1989). Reduced puddling for the preceding rice crop or subsoil tillage prior to the nonrice crop may improve rooting depth and volume, thereby increasing nutrient and water availability to the crops following rice. The constraints imposed by high temperatures and high midday evaporative demands can be partly overcome by the use of straw or soil mulches and the selection of tolerant cultivars. Figure 5. 'I)pical Record of the Trend Maize Grain Production by a Small-scale Farmer in Claveria, Misamis Oriental, Philippines 250 - 200 - 150- T1000 B~~~~~~~~ *\ 50 - GY = 291-377YR (GY = tota production: YR :year). \ 1979 1980 1981 1982 1983 1984 1985 Year Note: I sack = 50 kg. Source: Fujisaka and Garfity 1989. - 24 - For rainfed lowland rice ecosystems, substitution of rice with nonrice crops requires sufficient drainage of the root zone, absence of flooding, and a gradual onset or termination of the rains that ensures that water supply reaches the nonrice crop. This predisposes lighter-textured soils with a shallow groundwater table and fields higher in the toposequence for such shifts. One technology that can increase the adaptation of nonrice crops is the early drainage of rice fields by using field ditches, reservoirs, or the sorjan system. The challenges for crop establishment and management are often similar to those for irrigated rice systems described above. Green Manures The increasing cost of nitrogen and the existence of growing-season components too short for a grain legume have led to research on the exploitation of the pre- and postmonsoon period for the production of green manure crops. These may also have promise in irrigated rice systems. This crop diversification strategy would result in lower costs of rice production, reduced dependence on imported goods, and the security of a production system based on renewable resources. The use of flood-tolerant, stem-nodulating green manures such as Sesbania rostrata and Ashaemenome afrespera has greatly increased possibilities for wetland green manures as a complement to rice production. The technological challenges of seed production, mechanized incorporation, and scheduling of the green manure crop are being researched at IRRI and collaborating research institutions. Summary Agricultural diversification can benefit farmers and consumers by providing a wider range of products, greater production stability, lower risks, and possibly more efficient land use systems. Given the difficulty researchers, change agents, and farmers will face, however, in meeting future rice requirements, the shifting of rice land to other enterprises may not be a fruitful policy direction. Rice land is lost every day, partly because of urbanization, partly because of salinization and erosion. Further reduction in seeded rice acreage will place greater pressure on remaining land and will force rice production into marginal areas where sustainable production is difficult to achieve. Crop diversification by further intensifying the use of rice land is more likely to succeed and will probably be more productive. In light of the potential benefits and technological constraints, this can be more effectively achieved by combining with rice the production of forages for livestock, fish, green manures, and at times nonrice vegetable or grain crops. Opportunities for such rice land diversification should be explored where the growing-season components are not suitable for rice, where water costs and the potential for increased water use efficiencies will lead to increased food production, and where soil and climatic constraints do not limit production. Crop diversification can be assisted by improved management and modification of land. In upland rice ecosystems the formation of contour terraces by hedgerow systems can regenerate depleted topsoil and has been found to increase farmer investment in the production of higher-value crops. Research and development of this type of land improvement should be given high priority. In rainfed lowland rice ecosystems, improved land and water management can lead to better utilization of the growing season. The use of field drainage, reservoirs within the field, and farm-level water storage reservoirs can lead to a lengthened nonrice crop period and increased potential for rice-fish production systems. On-farm research in these areas and in areas related to crop establishment after rice requires increased support. - 25 - The use of green manures for crop diversification should be more aggressively explored. The cost of nitrogen governs rice yields and will likely increase. For short growing-season periods, land can be dedicated to producing green manures adapted to wetland rice culture. This may be economically attractive and would avoid many of the market, input, and crop management constraints associated with the introduction of other nonrice crops. Note 1. For a wide range of approaches under study in Asia, see IRRI 1988. References Branch, J.W., D.S. Maher, and T.P. Zacharias. 1988. "Diversification of Louisiana's Field Crop Sector." Louisiana Rural Economist 50(2):8-11. Calub, A.D., and D.B. Roxas. 1988. "Philippine Crop-Animal Systems Research (Formerly Crop-Livestock Systems).' In Proceedings of the Crop-Animal Systems Research Workshop, Serdang, Malaysia. Los Banos, the Philippines: International Rice Research Institute (IRRI). Fujisaka, S., and D.P. Garrity. 1989. "Farmers and Scientists: A Joint Effort in Upland Soil Conservation Research and Technology Transfer." Paper presented at the International Workshop on Conservation Farming on Hillslopes, 20-29 March, Taichung, Taiwan. Gerpacio, A.L. 1988. "Utilization of Cowpea for Swine and Poultry Production." In Proceedings of the Crop-Animal Systems Research Workshop, Serdang, Malaysia. Los Banios, the Philippines: IRRI. International Rice Research Institute (IRRI). 1988. Proceedings of the Crop-Animal Systems Research Workshop, Serdang, Malaysia. Los Banios, the Philippines. Lantican, F. 1982. "Desirable Characteristics of Upland Crops for Planting before or after Wetland Rice." In Report of a Workshop on Cropping Systems Research in Asia. Los Banios, the Philippines: IRRI. Mew, T.W., and F.A. Elazegui. 1982. "Disease Problems in Upland Crops Grown before and after Wetland Rice." In Report of a Workshop on Cropping Systems Research in Asia. Los Banios, the Philippines: IRRI. Morris, R.A., and H.G. Zandstra. 1979. "Land and Climate in Relation to Cropping Patterns.' In Rainfed Lowland Rice: Selected Papers from the 1978 International Rice Research Conference. Los Banios, the Philippines: IRRI. Nitis, I.M. 1982. "Use of By-products to Feed Bali Cattle." In By-product Utilization for Animal Production. Proceedings of a workshop on applied research, 26-30 September, Nairobi. Price, E. C. 1982. "Adoption and Impact of New Cropping Systems in Iloilo and Pangasinan, Philippines." In Report of a Workshop on Cropping Systems Research in Asia. Los Bainos, the Philippines: IRRI. Schuh, G.E., and S. Barghouti. 1988. "Agricultural Diversification in Asia." Finance & Development 25(2): 41-44. Syarifuddin, A.K. 1982. "Tillage Practices and Methods of Seeding Upland Crops after Wetland Rice." In Report of a Workshop on Cropping Systems Research in Asia. Los Baiios, the Philippines: IRRI. - 26 - Woodhead, T. 1989. "Establishment and Production of Non-rice Crops Following Rice in Previously Puddled Soil." Paper presented at the 20th Asian Rice Farming Systems Working Group Meeting, 2 October, Bogor, Indonesia. Zandstra, H.G. 1982. "Effect of Soil Moisture and Texture on the Growth of Upland Crops after Wetland Rice." In Report of a Workshop on Cropping Systems Research in Asia. Los Banos, the Philippines: IRRI. Zandstra, H.G., D.E. Samarita, and A.N. Pontipedra. 1982. 'Growing Season Analyses for Wetland Rainfed Fields." IRRI Research Paper 73. Los Banos, the Philippines. AGRICULTURAL DIVERSIFICATION IN ASIA: LESSONS FROM THE 1980s AND ISSUES FOR THE 1990s C. Peter Timmer Policymakers in Asia are interested in agricultural diversification for three main reasons: First, a well-diversified and flexible agricultural economy provides more stable farm incomes when commodity prices, especially rice prices, are highly unstable. Budgetary costs of price stabilization programs can be reduced, with substantial savings. Second, the process of diversifying a rural economy can be a significant source of income growth for rural inhabitants, providing better living standards and reducing the flow of migrants to urban areas in search of jobs. Third, a diversified cropping pattern may prove more sustainable in the long run than intensive cultivation of a single crop. A balanced and shifting crop array presents a much more difficult target for pests and diseases, thus extending the effective usefulness of crop varieties with natural disease resistance. Of these three reasons, the first-the stability provided by a well-diversified and flexible agricultural economy-is the most important to the day-to-day concerns of policymakers. Instability is a short-run concern, but this is the typical time horizon of policymakers. Accordingly, policies designed to stimulate agricultural diversification are nearly always driven by short-run concerns over the impact of price instability on farmers-a concern that becomes acute when rice prices collapse. Not surprisingly, interest in agricultural diversification for rice-based economies surfaced in the early 1980s as world rice production began to outrun demand. Peaking in 1986, surpluses drove prices in the world markets to historic lows-just one-tenth the real level of rice prices at their peak in 1974. As drought in Asia and policy adjustments in the United States restored the world rice market to balance, the keen attention of policymakers to research on approaches to rural diversification evaporated. Only those with concerns about rural poverty and agricultural sustainability have continued to encourage such research. Many countries, however, are again concerned with attaining or maintaining self-sufficiency in rice, and different questions have arisen. Diversification is often seen as a threat to intensification programs designed to attain ambitious production targets. For agricultural diversification itself to be a concept relevant to changing conditions, it must be set in a comprehensive and longer-run policy framework. This paper aims to provide that framework. Agricultural diversification is a much broader process than simply finding new crops to grow instead of rice. It involves the entire rural economy and entails broadening and maintaining the income sources of rural households. The process extends from introducing new crops into traditional farming systems to developing off-farm jobs in small-scale rural industries and eventually to the exit of a significant proportion of the rural work force from agriculture (though not necessarily from rural areas) as part of a structural transformation of the economy. At this general level, rural diversification is a gradual and inevitable process engendering little controversy.' Background The emergence of crop surpluses can force output prices down sharply, pressuring farmers either to grow other crops or to leave agriculture. Diversification becomes a problem for both farmers and pol- C. Peter Timmer is Thomas D. Cabot Professor of Development Studies, At-Large, Harvard University. - 28 - icymakers when the process of adjusting to surpluses and depressed prices does not occur quickly enough to bring farm incomes back into balance with previous levels or with incomes being earned in other sectors of the economy. The sharp decline in rice prices in domestic and international markets in the mid-1980s, and the resulting low incomes of farmers who were not protected from price declines, caused Asian farmers to search for alternatives to rice cultivation. Countries that kept domestic prices above the low world market prices often faced large budgetary costs, and these governments sought nonprice programs to diversify their farmers out of the basic food staple. Donor agencies, especially the World Bank and the Asian Development Bank, found their agricultural portfolios heavily invested in rice-specific irrigation systems that generated low economic returns when evaluated at world prices. Rural diversification thus became a vehicle for alleviating the distress caused at three levels-farmer, government, and donor-by the collapse of world rice prices. Designing and implementing new policies and investment strategies to foster rural diversification turned out to be a complicated process. Two major tradeoffs surfaced quickly as governments attempted to respond to the "crisis of success." First, concern about income distribution-farm incomes were already lower than urban incomes-conflicted fairly directly with efficiency considerations, at least in the short run, and governments found it difficult to choose one or an appropriate balance of the two. Attempts to increase both created a second important tradeoff, between incurring large budgetary costs to stabilize rice prices and passing on the costs to consumers. The dilemma for these countries, especially the ASEAN-4 (Indonesia, Malaysia, the Philippines, and Thailand), is in reconciling their desire to minimize the adjustment costs to the rural sector of coping with low cereal prices, while keeping budgetary costs under control, with the need to ensure future patterns of resource allocation that are not severely distorted by policies and investments used to cope with the short-run problem. In principle, the long-run approach is to let the "pull" factors of higher incomes in other sectors attract resources out of agriculture, rather than let chronically low prices "push" farmers into urban jobs. Ultimately the process of rural diversification must be consistent with longer-run patterns of structural transformation. The most successful countries will find ways to use the diversification process to stimulate this transformation, thus laying the groundwork for more efficient resource allocation and better income distribution. In managing this process of balancing rural and urban incomes, especially through labor migration from rural to urban areas, government: policymakers in developing countries are often faced with another dilemma. It is shortsighted to allow farmers to be driven off their land by low prices for agricultural commodities, particularly in response to temporary price declines in world markets. Food security must be maintained and future supplies of rice guaranteed. Despite the changing patterns of food demand that can be expected in the process of economic development, the large Asian countries cannot rely on the world market to meet their requirements for their basic food staple: they must grow most of it themselves. Policymakers must find a balance among the appropriate level of incentives to farmers to grow rice, policies that encourage some farmers or agricultural regions to diversify their production and become less dependent on the production of a single commodity for their incomes, and policies that encourage resources to flow out of agriculture altogether. Diversification and Rural Incomes in Southeast Asia Since the mid-1960s, the performance of the agricultural sector in Southeast Asia has been impressive, particularly in response to the massive resources allocated to expanding rice production. By the mid-1980s, expansion of cereal production was successful, exceeding goals for self-sufficiency in several countries. Surpluses found few buyers in world markets, and governments were faced with high storage costs and the prospect of plummeting incomes for farmers if price support programs proved too costly to the government budget. - 29 - Growth in rice production had fueled growth in the agricultural sector and contributed to rising prosperity in rural areas. The impressive agricultural performance of the Asian developing countries since the early 1970s in the face of many difficulties must be attributed to government policies based on a recognition that the economic fate of their countries depended on the agricultural sector. The policies pursued by these countries have generally had similar objectives, but the main one was increased food production. Governments wanted increased food security through self-sufficiency in the staple food. Apart from its political and social justifications, the objective of food self-sufficiency has a sound economic rationale in countries that produce their staple foods at lower cost to the economy than the average cost of imports. A secondary objective of agricultural development was the promotion of export-oriented commodity crops, primarily agricultural raw materials. The ultimate objective of any agricultural sector, however, is to produce incomes for those employed in the sector that are comparable to opportunities in other sectors or activities. The link between agricultural diversification and longer-run structural change occurs mainly because diversification is a bridge between the declining income-earning opportunities from growing food staples and an exit from agriculture altogether. Slow employment growth in modern industry in most countries places a heavy burden on agriculture (and a labor-intensive residual service sector) to create jobs at a pace that matches the rate of new entrants into the labor force. These jobs are unlikely to be in cereal production. The countries of Southeast Asia vary widely in the success with which they have stimulated income growth in their agricultural sectors. In none of the countries are incomes in the agricultural sector as high as those outside agriculture. In Malaysia, the Philippines, Thailand, and Indonesia, the share of agriculture in GDP is always less, often sharply less, than the share of the labor force in agriculture, implying that labor productivity and incomes are lower on average than in nonagricultural sectors. But some countries have been more successful than others in maintaining parity of growth rates in per capita incomes among sectors. The degree of agricultural diversification is not a significant explanatory factor in the growth of agricultural incomes; more important is growth in the rest of the economy-the degree of structural change occurring-and the extent to which the agricultural sector is protected from the declining comparative advantage that arises from such structural change. Some interesting lessons can be learned by comparing these patterns across the region. The results of the diversification policies (or nonpolicies) pursued since the mid-1960s by Thailand illustrate the complexity of trying to achieve both goals of diversifying agricultural production and raising rural incomes. Protection of farmers has never been a significant element in Thai agricultural policy: domestic rice prices have been stabilized through the rice premium and other controls on exports, but the beneficiaries have been Thai consumers, not farmers. Other commodity prices have been transmitted fairly neutrally into the domestic economy from world markets, and Thai farmers have been among the most responsive in the world. Thai agricultural exports have been sharply diversified since 1965, and farmers have found crops to adopt as markets opened. Corn, cassava, and rubber compete with rice as the leading farm export, and broilers, shrimp, orchids, and processed fruits and vegetables are gaining. With surplus land, good roads, an aggressive private marketing system, and flexible farmers, crop diversification in Thailand has been successful-with little direct government intervention. Diversification has not, however, been a solution to the country's problem of low farm incomes. The low world price for rice in the mid-1980s caused major distress for Thai rice farmers. Although the aggregate pattern of Thai agricultural production has been broadly diversified, individual farmers tend to specialize in a single crop. This specialization has adverse consequences for income distribution when prices of different commodities rise and fall. If a major force inducing Thai farmers to diversify were sheer poverty, the success of the market-induced aggregate change in production patterns would hold little appeal for other countries. According to World Bank data, of the ASEAN-4 countries, Thailand was second to Malaysia in average per capita income in 1985, well ahead of the Philippines and Indonesia (table 1). In 1965, Thailand and the Philippines had identical per capita incomes, and both were considerably behind - 30 - TABLE 1. AGRICULTURE IN THE ASEAN4, 1965,1985 INDONESIA MALAYSIA PHILIPPINES THAILAND iTEM 1965 1985 1965 1985 1965 1985 1965 1985 Per capita income (1985 US$) 208.00 530.00 845.00 2,000.00 365.00 580.00 365.00 800.00 Average annual change, 1965-85 (%) - 4.80 - 4.40 - 2.30 - 4.00 Labor force in agriculture (%) 71.00 53.00 59.00 37.00 58.00 49.00 81.00 67.00 Agriculture share in GDP (%, current prices) 56.00 24.00 28.00 19.00 26.00 27.00 35.00 17.00 Labor force share minus agriculture share in GDP (%) 15.00 29.00 31.00 18.00 32.00 22.00 46.00 50.00 Share GDP/labor force share 0.79 0.45 0.47 0.51 0.45 0.55 0.43 0.25 Implied per capita income in agriculturea (1985 US$) 164.00 239.00 397.00 1,020.00 164.00 319.00 157.00 200.00 Avg. annual change in per capita agricultural income, 1965-85 (%) - 1.90 - 4.80 - 3.40 - 1.20 a. Assumes all GDP generated in agriculture accrues to the agricultural labor force. Source: World Bank data. Malaysia and ahead of Indonesia. The structure of production in these two years roughly corroborates these overall economic rankings: the poorer countries have higher shares of GDP from agriculture, and vice versa. The pattern is similar, but more severe, for the share of the labor force in agriculture. The disparities between the two shares can be used to reveal striking patterns of rural poverty and relative growth rates in agricultural incomes. In all four countries for both time periods, the share of the labor force in agriculture was substantially higher than GDP share generated by agriculture, indicating that rural labor productivity (and almost certainly per capita income) was lower than labor productivity in the rest of the economy. The differences are not uniform across countries, however, and the pattern has changed dramatically over time. In 1965, Malaysia, Thailand, and the Philippines had ratios that ranged from 0.43 to 0.47, implying that agricultural workers were approximately 45 percent as productive as nonagricultural workers. The ratio for Indonesia was 0.79, reflecting a much more equal distribution of income, but in an economy that at the time was far poorer and less-developed. The most interesting statistics are the implied per capita incomes for each country's rural population. Even in 1965 Malaysia was well out in front: its per capita rural income was more than double the level of any other country in the ASEAN4. Rural incomes in the other three countries were roughly equal, probably approximating a near-subsistence standard of living. Individuals substantially - 31 - below the indicated averages were likely to be highly vulnerable to disease or hunger. By 1985, all countries had made progress in improving rural productivity, but the changes were unequal. The implied growth rate in rural per capita incomes in Malaysia exceeded that of average national per capita incomes. This rapid rate of growth in rural incomes reflects a remarkably successful effort to channel resources to the agricultural economy. The Philippines also had a higher rate of growth in rural incomes than in nonrural ones, although this probably reflects an inefficient allocation of resources in a protected urban industrial sector rather than striking success in agriculture. Indonesia's growth in rural incomes was more modest, surprising in view of the rapid growth in agricultural output during this period though not surprising in view of the deteriorating terms of trade for agriculture that accompanied the "Dutch Disease" side effects of the boom in oil prices. Between 1965 and 1985, rural incomes in Thailand rose only 1.2 percent annually, from a level already slightly below those of Indonesia and the Philippines in 1965. Despite significant diversification into new export markets by 1985, Thailand's farmers lagged even further behind the rest of the ASEAN- 4, with Malaysian per capita farm income levels five times higher; Philippine levels, 50 percent higher; and even Indonesian levels, 20 percent higher. Thailand's success with crop diversification was not matched by rising real incomes in rural areas. Two factors may account for this: (1) exposure to low world commodity prices in the mid-1980s; and (2) the slow increase in the industrial work force, the outcome of a capital-intensive industrialization effort that left many workers in agriculture. Malaysia and the Philippines have been more successful in raising rural incomes, though somewhat less successful in diversifying agricultural output. Indonesia has been less successful in both, although its rice farmers were relatively well protected from low prices in world markets in the mid-1980s. A review of the experience in these countries suggests that income diversification is more a result of structural change, trade orientation, and export opportunities than of conscious govermnent policies. To stimulate diversification, agricultural policymakers must integrate policies that increase flexibility in production with those that maintain farm incomes. In resolving these problems of policy coordination, policymakers must juggle the complexities of finding optimal budgetary allocations for agricultural research and investment-as well as pricing policies and marketing structures that provide adequate agricultural incomes from production of cereals-and yet offer incentives to diversify. No country seems to have consistently managed all dimensions of this task effectively, and the tradeoffs in approach are quite apparent. What the market gives in efficiency, it can take in incomes; what the government can provide in incomes, it can easily lose in poor resource allocations. Policy Issues for the 1990s By the 1980s, most Asian countries knew how to stimulate their farmers to grow enough rice to meet domestic goals for food security-usually interpreted as self-sufficiency. With the real value of these commodities in world markets at low levels, the economic value of this achievement, especially at the margin, was questionable. Where, then, should governments put their resources? Since the mid-1960s, they have gone mainly into food production, especially rice. Production successes alone, however, did not solve the problem of food consumption, nor will they raise farm incomes in the future if prices return to their recent low levels. Other policies that address farm incomes-involving access to noncereal markets and off-farm employment-will be needed, and for these, broader diversification policies will play an important role. Production of rice can no longer fuel widespread growth in the agricultural sector. The need for sufficient grain for food security will remain, but policies that generate these grain supplies also must be consistent with the longer-run process of shifting resources out of grain production specifically, and out of agriculture generally, as part of the structural transformation. Many countries, especially in Latin America and Africa, have found it difficult to strike the right balance of support among (a) crops destined specifically for export and primary food grains such as wheat - 32 - and rice; (b) secondary foodstuffs such as starchy staples with low or negative income elasticities; and (c) higher-value pulses, vegetables, fruits, and livestock products. Allowing the market to be the sole determinant of this mix risks undervaluing foods for the poor and food security in general, whereas high-income consumers and export markets are well served. Overconcentrating on basic foods, however, can lead to discrimination against earning foreign exchange and to rigid production systems unable to adjust rapidly to surpluses. The ASEAN-4 have done reasonably well in striking this balance. They have stressed domestic food security by developing their rice sectors intensively and earning foreign exchange from tree crops and other commodities, which have relatively little interaction with the cereal economy. The problem with this strategy emerged only in the mid-1980s, and then only as a problem of success: there was too much rice, at least temporarily. The surpluses did, however, reveal a gap in most agricultural development strategies: the neglect of the "intermediate" agricultural sector that entails neither a basic food staple nor a traditional export crop. The important policy question is whether diversification of both the rice and more traditional export crop sectors into this intermediate sector is technically and economically feasible. Experience with this process varies, with respect to both individual commodities and countries. The nature of agricultural production in Southeast Asia dictates that most of the adoption of diversified cropping patterns will be by millions of small-scale farmers acting in accord with their private interests. State-owned plantations can be diversified by command, and public exhortations directed at farmers to switch in or out of rice might induce some response. At the level of adoption, however, diversification is essentially a private sector activity. At the other extreme, relatively few resources from the private sector will be invested in large-scale irrigation projects, roads, electrical and communications networks, or even agricultural research on alternative crops to rice that can be grown by smallholders. These activities fall almost entirely in the domain of the public sector, at least in the countries of concern. Between these two clearly defined domains lies the possibility of a significant hiatus with respect to implementation of crop diversification schemes. A basic premise of current diversification efforts is that further, large increases in rice production will not find a market at remunerative prices. Unless prices actually facing farmers signal this, they will try to continue expanding production. If the lower prices are passed on to farmers, they will actively seek alternative crops with better market prospects or better income-earning prospects off the farm or outside agriculture. These changes are desirable as a long-run response, but can be quite tumultuous in the short run. Poor farmers and rural landless laborers might crowd into urban labor markets seeking jobs, which can result in social unrest if they do not find them. Attempts to grow alternative crops may flounder because of untried technology or markets that are too thin to absorb profitably substantial increments in supply. Government extension workers tend to be blamed for pest and disease damage and middlemen blamed for falling prices. Confidence in the effectiveness of government development programs is shaken, and the willingness-of both farmers and policymakers-to trust the marketplace as an arena for easy and fair exchange of commodities is lost. To avoid these problems, governments are searching for alternatives to such a short-run market free-for-all. Stimulating Private Sector Growth The key to finding alternatives is a better understanding by public sector policymakers of the actual decisionmaking environment of private sector participants, that is, farmers, traders and processors, and consumers. With this understanding comes the capacity to design more sensitive and effective policies that draw on, rather than scare away, the initiative and investment resources of the private sector. Commensurate with this effort by public sector policymakers is the responsibility of private interests to cooperate in the development program by becoming informed about government objectives and plans. There are significant problems, however, with this approach to diversification. If the process is to be market-driven, knowledge of and access to those markets for nontraditional commodities are crucial. - 33 - When the markets are external, the easiest way to gain such knowledge and access is to tap the expertise, and possibly the capital, of a multinational enterprise. With many countries now bidding for this expertise, there is a risk that more incentives will be offered, either publicly or privately, than is justifiable on the basis of benefits to the country. A certain responsibility on both sides of the bargaining table is essential to keep the process within bounds. Most countries do not have well-developed mechanisms for conducting such a public-private interchange, and the analytical tools for understanding the process are blunt at best. Reducing the Burden of Adjustment An important lesson learned from postwar development experience is that economic growth depends on rising agricultural productivity. Economies that have invested financial, human, and policy resources in agriculture in the postwar era have grown faster than those that have neglected or discriminated against their rural economies. Wherever rapid agricultural growth stimulated rapid overall economic growth, however, societies have often had to pay a heavy price for their success in the form of severe structural lag. Human and financial resources have not been able to move out of the agricultural sector fast enough to prevent low economic returns to their utilization. Consequently, low incomes, even widespread poverty, exist in some rural areas. Those countries that have supported farm incomes through price supports have paid high budgetary costs, and these burdens were exacerbated in the 1980s by the collapse of commodity prices in world markets. Diversification of agricultural production patterns is a potential response to both problems-structural transformation and low commodity prices-but the relationship between the two problems is not "one-to-one," nor will potential solutions necessarily correspond. A successful structural transformation is painful for the agricultural sector in all societies. Agriculture declines in relative importance as societies become richer; eventually agricultural labor must find work in other sectors. The process of selling the family farm and uprooting households to look for jobs in the city is distressing, especially when family members have few skills applicable to modern urban life. Societies cushion this process whenever they become rich enough to afford it. The only puzzle is why all industrial countries seem to use the same mechanism-protection of domestic farmers from foreign competition for commodities that were historically imported. This protectionism may stem from the changing political "market," as food declines in importance in real wage rates and urban household budgets, and from the capacity to "hide" support costs in higher consumer prices rather than direct budgetary outlays. Whatever the reasons, they are quite powerful to account for the pervasive use of agricultural protection in industrial economies. Despite policy efforts to slow structural transformation and preserve the viability of family farms, the exit of agricultural labor to other sectors is inevitable, the only mechanism for coping with inexorable shifts in demand and supply conditions for agricultural output. In the long run, rural incomes will be depressed relative to urban incomes unless labor leaves agriculture. High commodity prices can temporarily slow this process, as in the mid-1970s, and low commodity prices can accelerate it, as in the mid-1980s. All experience argues, however, that transferring labor out of agriculture is inevitable if the economy is growing. The policy issue is whether efficient ways can be found to ease the pain of adjustment, perhaps through better educational programs in rural areas, more flexible land and credit markets so that farmers can avoid sales during hard times, and rural industrialization programs to create a better market for off-farm employment for members of farm families. Maintaining Rural Incomes To address the issues above, the analysis must return to the broader topic of structural change and adjustment. Greater off-farm employment opportunities are the surest way to cope with low incomes - 34 - from rice farming, even when most irrigated land continues to grow rice. Land consolidation may permit fewer farmers to continue to earn competitive incomes from rice farming even at lower prices, as in the United States and, to some extent, Western Europe. Alternatively, small farm size might coexist with lower cereal prices if most farm household members have full- or part-time jobs in rural labor markets. Taiwan and Japan have followed a variant of this path, with the South Korean pattern significantly different. Many farm household members in Taiwan and Japan work in the industrial sector, and average farm size remains minuscule by the standards of OECD countries, partly because of stringent controls on land accumulation enacted at the time of postwar land reforms. Rice prices in all three countries have been maintained far above those in the world market to provide added income to rice farmers, a politically powerful group, with Japan exhibiting the widest divergence. Poorer countries cannot afford the budget costs or the resource misallocations of the Japanese model, but they face similar pressure to adjust and support low farm incomes. In the developing countries of Asia, the process of rural diversification is likely to follow a path similar to that of other Asian countries that have already undergone agricultural transformation. Much has been learned from these countries, and the experiences of those based on a rice economy-Japan, South Korea, and Taiwan-are especially relevant to Southeast Asia. In particular, the role of small farm size in conditioning the path of diversification and structural change cannot be ignored. Widely dispersed small farms were a crucial factor in the adoption of high-yielding rice technology and the equitable distribution of its benefits. The pricing and marketing policies of countries that encouraged the small farmer to invest heavily in rice production, however, also gave them a strong vested interest in the continued use of pricing policy as an instrument for maintaining rural incomes relative to rapidly rising industrial wages. The large numbers of small farmers lent political importance to these concerns, and all successfully industrializing Asian countries have protected their rice farmers from foreign competition as they lost comparative advantage in rice production. Policies Consistent with Structural Transformation Diversification strategies have been used in Japan, South Korea, and Taiwan to cope with the budgetary costs and inefficient resource allocation that accompany protection of rice farmers. The unique role played by rice in the agricultural economies of all these countries, however, has been difficult to replace. Adjustment has been smoother where the rural economy had important nonagricultural activities for support, as in Taiwan. This role of a dynamic rural economy, as opposed to a dynamic agricultural economy, stresses the importance of nonagricultural policies in the diversification process. In particular, policies that help integrate urban and rural labor markets and facilitate the establishment of small-scale rural industries speed diversification and structural change by pulling resources out of agriculture. In most countries, however, pressures remain that exert a strong push, and part of any agenda to foster rural diversification must address the short-run problems generated by these pressures. The lessons from the already industrialized countries in Asia are not promising in this regard: heavy protection of rice farmers was an essential ingredient in coping with the short-run pressures on income distribution in Japan, South Korea, and Taiwan. The entire society has an important stake in diversification programs when they are designed and implemented within a policy framework that incorporates the changing structure of the economy during the process of economic growth. The impact can be positive or negative. Policies that attempt to freeze the allocation of resources and distribution of incomes to farmers growing particular commodities (especially rice) can gradually create serious distortions in the allocation of economic resources, and these distortions can endanger overall economic health unless the country is rich enough to afford them. Japan can afford to indulge rice farmers more than Malaysia or Indonesia can, although the spillover effects create serious problems for land prices and trade policy, even in Japan. There are also positive adjustment policies, however, that attempt to make the agricultural system more flexible, that train agricultural - 35 - workers for off-farm jobs, and that cushion the transition from a rice-based agricultural economy to an industrial and service-based economy. These policies include diversification programs at the research level, investment in infrastructure, improvement in marketing communications and information, and price interventions that create stable expectations and lessen the risk of adopting new technologies and products. To be successful, positive adjustment policies must be consistent across all three levels of policy concern: the farm, the agricultural sector, and the macroeconomy. Policies for the Agricultural Sector Farmers make the actual decisions that lead to a more diversified agricultural economy, but the agricultural policies and investments of government agencies determine the technical and economic environment in which farmers make choices. Diversification programs in particular are usually designed and implemented by agricultural ministries. Their primary dilemma is whether to diversify farm- or regional-level cropping patterns. The former approach requires flexible technologies and marketing capacity at the level of individual farmers; the latter allows farmers to specialize in a single crop while diversifying output and exports for the country as a whole. Specialization is the opposite of diversification: it is part of the problem rather than part of the solution to any overemphasis on a single crop. There are two reasons why regional specialization might be a crucial feature of any diversification policy. First, agronomic and climatic factors favor some crops over others. In the relatively large and diverse countries of Asia, soil types, temperatures, and rainfall patterns can vary radically from one part of a country to another and at different altitudes in hilly or mountainous regions. Volcanic soils support intensive crop cultivation: the acidic soils underlying much of Asia's rain forests are much thinner and more difficult to manage in terms of sustained fertility. Tree crops such as rubber or oil palm probably have substantial technical advantages relative to, for example, corn or soybean on such soils. Only if relative prices shifted sharply in favor of food crops would it make economic sense to invest in restructuring the acidic soils of the humid rain forest areas to support annual cultivation of rice, corn, root crops, or legumes. The second reason for regional specialization is more easily overlooked, but it may be as important as agronomic factors. Efficient development of entire commodity systems, from input production and marketing through to downstream processing and consumption of the final product, requires the formation of extensive backward and forward linkages from the producer level. These linkages can be technological-depending on, for example, engineering relationships and quality requirements-and financial-depending on investment patterns from profits generated by commodity production and consumption patterns from the incomes earned in the sector. Many of these linkages exhibit economies of scale and can be developed to efficient levels only if the commodity is produced in a relatively cohesive spatial pattern. This process of market deepening is a natural result of regional specialization and one of the major forces that gradually but persistently produces such specialization. Well-developed, low-cost marketing systems require sufficient supplies of the specific commodities being marketed to justify the full investments needed to capture any economies of scale to the system. Achieving this balance is a simultaneous process: as specialized production grows in a region, the marketing system expands to serve it but also creates the demand for further expansion by offering lower marketing costs than those in regions that would otherwise be competitive climatically and agronomically. The lower costs generated by specialization can confer significant competitive advantages on regions that are low-cost producers of a commodity and have an efficient marketing system with adequate volume to capture the economies of scale implicit in the forward and backward linkages. Thus, regional specialization in a range of agricultural products would seem to be the answer to the problem of diversification. Such specialization permits the cost economies of scale (and learning) to be captured, while diversifying the country's agricultural output. A problem, however, remains. Although the country may be well diversified, individual farmers and regions might not be. Significant price - 36 - instability, whether generated strictly in domestic markets or transmitted from international markets, still can lead to substantial income distribution consequences for farmers and regions, unless their output is sufficiently negatively correlated with prices that net revenue is stabilized by unstable prices. When large regions depend heavily on a single crop for their economic base, the resulting vulnerability is similar to that at the national level when rice cultivation is widespread. When rubber producers, coffee growers, or corn farmers specialize in production, each can encounter income stabilization problems in the face of unstable prices or yields. The consequences for income distribution of crop specialization at the farm or regional level are straightforward. With domestic price stability, small farmers can specialize in single crops, and regional diversification can keep surpluses from developing, but this strategy depends on price stabilization: otherwise, individual farmers must diversify to spread risks from price fluctuations. Such diversification is likely to incur high costs because of the foregone effects of "learning by doing" and scale economies inherent in marketing systems. Compared with national specialization in a single commodity, the macroeconomic consequences of regional vulnerability are not as great-unless all prices and yields move together. Policymakers should, however, pay attention to individual and regional problems, especially in countries with diverse regional interests, where appearing to ignore the economic plight of distressed regions can have devastating consequences for the political stability of the entire country. The Macroeconony and Agriculture Macroeconomic policies have a significant impact on agriculture, and their effects can often overshadow those of sector-specific policies. Trade, exchange rate, fiscal, and monetary policies are leading determinants of the movement of capital and labor between agriculture and the rest of the economy, the growth and composition of agricultural output, and the volume and composition of trade in agricultural products. Macroeconomic policies thus affect diversification both within the agricultural sector and outside it. Macroeconomic policy can similarly have a dominant impact on the process of rural diversification as it has on agricultural development. Evolving patterns of demand are the prime driving force behind the diversification process: the speed at which these patterns shift and thus support rapid diversification is largely a function of macroeconomic performance. Links to patterns of foreign demand are efficient only if the exchange rate correctly reflects the real opportunity costs of domestic resources in producing for export. A repressed financial system always discriminates against the rural sector, and diversification efforts are highly sensitive to credit availability for private traders if flexible marketing arrangements are to be created. Rural infrastructure is primarily a public good, not supplied in optimal amounts by the private sector. Unless a government's budgetary priorities and spending procedures support investments in and maintenance of this infrastructure, rural diversification can make little progress. The design and implementation of rural diversification programs and policies cannot rest solely in the ministry of agriculture. Cooperation among government agencies is essential to agricultural development efforts in general: for diversification, the planning agency, as well as other agencies, and the ministries of finance, trade, and public works must be actively involved. Facilitating the Process of Diversification Diversification through more flexible farm-level decisionmaking seems to be a highly effective solution to the many problems faced by farmers and policymakers. Why then are farmers so reluctant to diversify their cropping patterns? Why do so many rural areas remain heavily dependent on rice as the foundation of their local economies? Why do policymakers continue to design plans that require farmers - 37 - to comply with new, rigid guidelines on what to plant? General answers to these questions have been indicated above, but part of the answer must lie in country-specific institutions, resources, and attitudes.2 A summary of the two main general lessons learned from Asian experience with diversification away from rice will help set the stage for discussion. First, agricultural diversification is more than a production problem. It involves a process of change over time: it is not the setting of crop-specific production targets in annual or five-year plans. Raising and stabilizing rural incomes are the goal-not increasing crop production statistics. The process of agricultural diversification must be demand-driven. Because the process frequently involves vertical activities in processing and adding value through quality improvements, the marketing sector, not supply constraints on various crops, is usually the key bottleneck to success. Eventually, a developmental perspective on diversification shifts the focus from the rural economy altogether to a concern with reducing the human costs of migration, raising the level of human and financial capital that can be drawn out of agriculture, and speeding the structural transformation of the overall economy. Second, even when the production dimension alone is highlighted, the process of diversification involves more than technical constraints at the farm level. These technical constraints are important, imposed as they are by the existing level of research and technology, the knowledge of extension workers and farmers, the nature of soil types and climate, and the flexibility of local irrigation and drainage systems. Incentives and relative incomes are probably more important, however, in both the short and long run. Incentives depend on both costs and revenues: they are determined by these technical factors as well as by prices and marketing efficiency. In the short run, governments can and do alter incentives more quickly and effectively by manipulating prices than by changing technical constraints. When crop surpluses and deficits are the main forces causing policymakers to consider agricultural diversification as a potential solution, the short-run effectiveness of price interventions offers obvious political appeal. In addition to level of income, however, a farmer worries about the riskiness of alternative sources of income. Risk can come from yields and prices, and these two sources are seldom negatively correlated to a sufficient degree to produce income stability and neutrality of farmer decisionmaking. Especially in Indonesia, where price risks from growing rice are significantly smaller than for other crops, risk aversion is a significant impediment to what would seem to be rational diversification on the basis of average profitabilities of alternative crops. Behavior in the face of risk aversion is affected by farmer attitudes and the nature of technology, but the fundamental problem is the failure of local credit and risk institutions to provide any potential for farmers to transfer their risks to other parties. Avoiding risk thus becomes an internal household strategy, and many households concentrate on growing a familiar crop with known technology and guaranteed prices rather than risk their livelihood on new crops with sharply fluctuating prices. Finally, the ultimate influence on the degree and rate of crop diversification is the overall economic environment of the rural sector. As noted earlier, this environment depends on several key factors: macroeconomic growth as it drives the demand-led side of diversification; the efficiency of the marketing system and its capacity to transmit signals of consumer preferences back to farmers; and the foreign exchange rate as it translates signals from foreign consumers and competitors into domestic prices and hence into the terms of trade facing the rural economy. No "plan" for agricultural diversification can be successful if this overall economic environment is hostile to rural producers. It is debatable whether a plan is needed at all when the environment is positive. With or without a plan, all agricultural diversification strategies require increased flexibility on the part of farmers. Only part of the observed rigidity in farmer behavior with respect to crop choice is the result of technical and economic constraints at the farm level. At least in Indonesia, a substantial share of rigidity in farmer decisiomnaking can be traced to the government institutions (and donor agencies) ostensibly responsible for helping farmers diversify their cropping patterns. Despite this impediment, talented people are working in these organizations and in other institutional settings to build a more effective framework, both bureaucratically and with regard to general policy, for the long-run - 38 - diversification of the rural economy. A drastic and rapid switch from centralized planning and a commodity orientation to a new approach is not likely to be effective nor feasible: mind-sets are particularly slow to change. Yet not only must they change, however gradually, but they must change in the right direction: recognizing their starting point is the only way to begin. Notes 1. This perspective on agricultural diversification arises from work over the last several years on the process of structural change in agriculture and the role of diversification in this process. The most complete discussion of structural change in this context is in C. Peter Timmer, "The Agricultural Transformation," in Hollis B. Chenery and T. N. Srinivasan, eds., Handbook of Development Economics (Amsterdam: North-Holland, 1988), pp. 275-331. Work on diversification has been partly supported by the World Bank's Agriculture and Rural Development Department; early results are reported in C. Peter Timmer, "Crop Diversification in Rice-Based Agricultural Economies: Conceptual and Policy Issues," in Ray A. Goldberg, ed., Research in Domestic and International Agribusiness Management, vol. 8 (Greenwich, Conn.: JAI Press, 1988), pp. 95-163. See also Shawki Barghouti, Carol Timmer, and Paul Siegel, Rural Diversification: Lessons from East Asia, World Bank Technical Paper 117 (Washington, D.C.: 1990). 2. This section draws on the Indonesian experience with agricultural diversification to identify several of these more specific problems. Based on discussions and other papers in this volume, experience in other countries in the region reflects a similar range of problems, although the institutional details differ. AGRICULTURAL DIVERSIFICATION IN THE MIDDLE EAST AND NORTH AFRICA Kutlu Somel The Middle East and North Africa (MENA) region is a net importer of agricultural products, particularly basic food commodities.' On an individual country basis, there are few exceptions to the region-wide deficit in basic commodities such as wheat (the basic staple), barley (the basic feedstuft), cooking oil, milk, and sugar (Somel 1986; Alexandratos 1988). Historically, the governments in MENA have been preoccupied with meeting the growing food demand, arising from increasing incomes and high population growth rates, while contending with high climatic variability, which generates significant fluctuations in yields and production. These concerns have resulted in food security policies that have-again, with few exceptions-emphasized food self-sufficiency. Aside from the economic efficiency implications, food self-sufficiency in MENA usually constitutes political rhetoric rather than a technically feasible option. In addition, industrialization policies have sought to provide urban consumers with affordable food to maintain low and stable wage rates, sustain political balances, and supply commodities and raw materials for industry and exports. Consequently, the agricultural sector has been subjected to excessive controls in producing the needed basic staples and industrial raw materials, invariably at prices well below international levels. The results have been a stagnant agricultural sector, dependent on a few staples and industrial crops and with production and productivity far below potential. Increasing international and regional market integration in the 1980s induced changes in MENA governments' perspective on agriculture's role. The expanding opportunities for trade on the one hand and the increasing budgetary burdens of maintaining orthodox agricultural policies on the other exerted pressure on governments to increase the efficiency of the agricultural sector. Gradually, the focus is shifting from food security based on self-sufficiency to food security sustained through trade based on comparative advantage and more diversified agricultural production. This has heightened the need for more efficient utilization of agricultural resources, particularly in the irrigated and higher-rainfall areas. However, agricultural support services (e.g., rural credit, marketing, research, extension, and training) have not been able to meet the new demands of this diversification process. Furthermore, demand and price volatility in international markets is tempering governments' commitment to outward-oriented policies, resulting in risk management strategies whereby some resources continue to be allocated to supporting and controlling the production, marketing, and storage of traditional crops. Diversification in MENA, at both the micro- and macrolevels, is founded on the principle of risk management and should be analyzed from this perspective. Risk management emphasizes the assessment of a portfolio of diversified income-producing activities or options with respect to the magnitude and stability of the net benefits of the total portfolio. Focusing only on the income-increasing effects of agricultural diversification, therefore, will provide an incomplete picture: current as well as future options must be assessed in terms of their individual effects not only on income, productivity, and stability, but also on the income-generating potential and stability of the broader portfolio. Within this perspective, complementarities and conflicts between macro- and microeconomic objectives will become apparent. For Kutlu Somel is senior agricultural economist, Agriculture Division, Middle East and North Africa Region Technical Department, World Bank. The invaluable assistance of Aysen Tanyeri-Abur with literature and data review is acknowledged with gratitude. Many of the ideas presented here have been inspired by discussions and collaboration with colleagues in the Agriculture Division of MENA's Technical Department, though the sole responsibility for this paper rests with the author. - 40 - example, government income- and productivity-increasing efforts may not meet objectives when tempered by the stability concerns of farmers. This paper reviews the agricultural sector in MENA from both aggregate and farm-level perspectives. It then assesses the potential for and constraints to agricultural diversification in the region. Last, agricultural diversification is reviewed within the context of a general economic and agricultural strategy, and policy recommendations to induce effective diversification are proposed. The Aggregate Sphere General Overview Since the 1960s, agricultural growth in MENA typically has lagged behind overall growth. Moreover, agriculture's share in GDP has declined to 15 to 30 percent (table 1). Although agriculture's role is decreasing in MENA vis-a-vis its contribution to GDP, it remains an important sector in terms of employment. Except where irrigated agriculture is the norm (e.g., Egypt and Pakistan), agriculture in MENA depends on rainfall. Production in rainfed areas is dominated by cereals, mainly wheat for food (bread is a staple) and barley for feed (primarily for sheep and goats) (table 2). The relative importance of wheat and barley depends on rainfall levels; wheat is dominant in areas receiving over 350 mm average annual rainfall (with durum in the drier end of the spectrum and bread wheat in the wetter end), and barley dominates in the 250-to-350-mm rainfall areas. In rainfed areas, 25 to 50 percent of the land is left fallow each year for moisture and nutrient buildup. These fallow areas are used as grazing grounds during the fallow years except in areas such as the Anatolian Plateau, where clean fallow (as opposed to weedy fallow) practices are followed for moisture conservation. Fallow is decreasing due to pressure on the land. In certain areas, with appropriate technologies, this can result in sustainable production increases, but in others, particularly with cereal monoculture, there can be ecologically disastrous results: reduced soil fertility and soil degradation, which lead to yield decreases. Depending on irrigation availability, soil quality, and marketing infrastructure, other crops grown include pulses (lentil, chickpea, faba bean); industrial crops (cotton, tobacco); oil crops (sunflower); forage crops (berseem); winter and summer fruits and vegetables; and tree crops (fruits, olive, pistachio, almond, hazelnut). Major Issues At the macrolevel, there are 10 major areas of interest in the policy and physical-biological arenas: production controls, climatic variability, the natural resource base, irrigation, employment policies, domestic and international trade policies, agricultural research and extension, rural finance, and land tenure. Production controls. Controls over cultivated area, input allocation, price, marketing, and storage for basic commodities have been extensive, and many still persist. For example, in Syria, wheat and barley areas are controlled, and input allocations are centrally determined according to agroecological characteristics. The prices of these commodities are set by the government, and their movement is strictly regulated, with off-farm storage under government control. In Turkey, apparent support policies for individual commodities actually involve implicit taxation when exchange rate policies are considered (Somel 1987; Krueger, Schiff, and Vald6s 1988). Excessively dirigiste manifestations of these policies were (and some still are) also observed in area and production controls over even high-value crops in irrigated areas, e.g., in Jordan and Tunisia (World Bank data). The standard example of an irrigated crop - 41 - TABLE 1. AVERAGE ANNUAL PERCENTAGE GROWTH RATES OF GDP AND AGRICULTURE AND PERCENTAGE SHARE OF AGRICULTURE IN GDP AND RURAL POPULATION, SELECTED COUNTRIES AVERAGE ANNUAL GROWrH RATE SHARE SHARE IN RURAL GDP AGRICULTURE IN GDP POPULATION 1965-85 1980-89 1965-85 1980-89 1965 1989 1965 1990 Pakistan 5.2 6.4 3.3 4.4 40 27 76 68 Egypt 7.3 5.4 2.7 2.6 29 19 59 54 Syria 9.1 1.6 5.9 -0.9 29 22 60 50 Tunisia 6.5 3.4 5.5 1.6 22 14 60 46 Turkey 6.2 5.1 3.2 3.0 34 17 66 40 Jordan nla n/a n/a n/a n/a 6 54 33 Algeria n/a 3.5 n/a 5.3 n/a 16 62 49 Morocco 5.7 4.1 2.4 6.7 23 16 68 53 n/a Not available. Source: World Bank 1991. facing production and price controls is cotton in Egypt; farmers are obliged to cultivate cotton in TABLE 2. PERCENTAGE LAND USE irrigated areas and sell their production to IN SELECTED COUNTRIES, 1989 government agencies at centrally determined ARABLE AND prices, which are far below international levels. PERM. Not surprisingly, cotton yields have been CROPS 1RI- decreasing as farmers allocate resources away COUNTRY ('OOOHA) GATED WHEAT BARLEY from it-a trend the Egyptian government seeks to Algeria 7,605 4.42 18.67 13.15 reverse (World Bank data). Egypt 2,585 100.00 24.91 1.93 Morocco 9,241 13.69 28.46 25.96 Tunisia 4,700 5.85 11.85 8.77 Climatic variability. In MENA, a large Iran 14,830 38.77 39.03 16.93 proportion of agricultural land is rainfed.2 Iraq 5,450 46.79 10.77 16.75 Consequently, these areas experience significant Jordan 376 15.16 23.94 13.30 Syria 5,503 12.18 15.83 15.61 variability in the annual amount and distribution Turkey 27,885 7.96 33.09 11.95 of rainfall, leading to highly unstable agricultural Yemen 1,481 20.93 6.21 3.51 production and yields. The impact of this variability ranges from virtual crop failures to Source: FAO Production Yearbook 1990. gluts beyond existing storage capacities.3 Particularly in countries with large rainfed areas, production instability is striking: between 1961 and 1976, the coefficient of variation (cv) for staple food crop production was 27 percent in Morocco, 29 percent in Algeria, 38 percent in Syria, and 66 percent in Jordan. In Egypt, where irrigation dominates, the cv for the same period was only 4.5 percent (Oram 1988). The natural resource base. In MENA, exploitation of arable land is approaching or has exceeded ecologically sound limits. Cultivation is encroaching onto marginal lands, e.g., pastures, hilly or sloped lands, and steppes. Large tracts of land receive limited precipitation, with production usually limited to barley and livestock (small ruminants). Thus, the options for diversification in these rainfed areas are somewhat constrained, particularly because many are already overexploited. Fallow replacement is a diversification option, but it is proving to have adverse ecological consequences, especially in the lower- rainfall areas, because of continuous monoculture. There is greater potential for diversification in higher- rainfall areas, those receiving over 400 to 450 mm annually in a well-distributed and stable pattern. Here, the traditionally wheat-dominated systems can diversify into, for example, food and forage legumes, - 42 - oilseed, spices, and intensive livestock production. Generally, production increases can be achieved through increases in yield and cropping intensity and changes in the crop mix. Irrigation. Given the critical impact of water supply on agricultural production, another predominant concern for MENA governments has been harnessing surface and groundwater resources for irrigation.4 Irrigation investments continue to be the major component of public expenditures in agriculture. For example, in Tunisia, the value of agricultural production from irrigated areas is estimated to be about 30 to 40 percent of agricultural GDP. Although irrigated areas constitute only about 6 percent of arable land, and public irrigation perimeters account for slightly less than half of this, during the Seventh Five-Year Development Plan (1987-91), 55 percent of public expenditures for agriculture were in irrigation, drainage, conveyance systems, etc. Total public agricultural expenditures (excluding subsidies) per hectare of irrigated land were almost seven times those of rainfed areas (World Bank data). The exploitation of these valuable irrigated lands, however, is highly inefficient due to misguided policies and mismanagement. This leads to low cropping intensities, low system- and farm-level water use efficiencies, and misallocation of valuable land and water to low-value crops. Cost recovery is receiving increasing attention from MENA policymakers. This is because water charges often not only fail to reflect investment costs, but invariably do not even cover operation and maintenance (O&M) expenses. In Jordan, even allowing for the inefficiency of the Jordan Valley Authority, water charges reflect less than 50 percent of O&M costs and about 10 percent of total costs (World Bank data). This situation essentially benefits the few who have access to subsidized public irrigation water and creates rents that accrue to this limited group of beneficiaries.5 However, in private irrigation structures (e.g., tubewells), although equipment and power or fuel may be subsidized, investment and O&M costs must be, and invariably are, recovered. Moreover, through area, production, and marketing controls, MENA governments impose certain cropping patterns on irrigated areas that could otherwise be used more efficiently and profitably. In the former People's Democratic Republic of Yemen (PDRY), for example, about 5 percent of wheat requirements were domestically produced on part of the extremely scarce (only 60,000 ha) agricultural land irrigated from underground aquifers at an estimated investment cost of about US$6,000/ha (World Bank data.) In Jendouba, Tunisia, farmers who sign a contract to follow a government-determined rotation, including cereals and sugar beet, receive discounted water. In Jordan, however, the highly arbitrary cropping-pattern restrictions on certain horticultural crops in the Jordan Valley are being phased out. These policies reflect a governmental effort to ensure adequate and stable supplies of staples and industrial raw materials. The basis for these policies is the risk of shortages, particularly in urban areas, that can spark price increases and consequent political unrest and industrial disruptions. The costs of these policies, therefore, are the insurance premia for the risk management strategies of the MENA governments.' In Tunisia, for example, production subsidies during the seventh plan (1987-91) constituted 25 percent of public expenditures in agriculture, equal to an annual rate of US$1,235 per farmer, US$89 per hectare, and 29 percent of agricultural GDP. In Turkey, input subsidies between 1986 and 1988 ranged from 4.8 to 6.5 percent of the agricultural value added. In the early 1980s in Jordan, domestic wheat production was supported at 40 to 74 percent above import prices. Employment policies. MENA governments consider the agricultural sector the main depository of labor. Because the cost of maintaining labor is considered to be lower in rural areas, there have been concerted efforts to slow urbanization and maintain rural populations. A typical example is Tunisia's soil and water conservation activities, which are funded partly by donors and are designed primarily to slow migration to the developed countries. These are essentially public employment programs with little efficiency and productivity content. But if sectoral adjustment programs are to increase agricultural production efficiency, the sector must not be made to bear a disproportionate burden of absorbing excess labor. Furthermore, the perpetuation of existing employment policies also affects diversification options - 43 - by maintaining distorted factor proportions that promote labor-intensive technologies and prevent labor- productivity-increasing investments. The agricultural sector is also characterized by segmented labor markets, which has resulted in a bimodal farm size distribution. Small farms often rely on family labor for intensive exploitation of limited land. These farms supply surplus labor to the rural labor market, and this supply can be limited in peak seasons. The larger farms therefore face labor shortages and high wages, and thus opt for capital- intensive technologies. Domestic agricultural trade. Policymakers and bureaucrats in MENA have displayed a strong bias against commerce. This is reflected by frequently fixed producer prices; wholesale and retail price and margin controls (Jordan, Tunisia); pervasive state monopolies in the marketing of staples; controls over storage (Tunisia, though they are being relaxed); licensing controls; and restrictions on the movement of commodities (Syria, the former PDRY). Most of these controls cannot be enforced effectively and therefore are invariably a source of economic rents, first for those who can evade them and second for the underpaid bureaucrats who can facilitate such evasion. The controls also eliminate incentives for enhancing quality in production and marketing (packaging, grading, etc.). Moreover, fixed prices enforced throughout the year remove the incentive for production out of season and for on-farm storage, thereby increasing spoilage. Inferior, outdated packaging and storage techniques result in substantial waste. In addition, commodity grades and standards, particularly for horticultural products, either are nonexistent or unenforceable, resulting in ad hoc marketing practices at the retail level, i.e., preferential treatment of different customers. These policies stifle private trading and reduce the possibility of spatial and temporal arbitrage. Furthermore, they prevent the transmission of price information to producers and intermediaries that can induce diversification and quality enhancement. International trade. International trade policies involve a complex set of issues, which can only be discussed briefly here. In the MENA countries, export and exchange rate controls (invariably overvalued exchange rates), quantitative restrictions, and excessive tariffs have resulted in price distortions, which lead to the misallocation of resources in agriculture. These measures have also resulted in the implicit taxation of agricultural production, which in turn has reduced the incentive for export- oriented agricultural production and marketing by the private sector. The absence of quality standards has further handicapped promising export sectors. Nevertheless, recent economy-wide and sectoral structural adjustment efforts in, for example, Turkey, Jordan, and Tunisia, have incorporated attempts to gradually overcome these problems. Furthermore, the establishment of the Maghreb Arab Union (UMA) has served to facilitate the pursuit of a more common market approach. For efficiency, trade should be directed to those regions that optimize returns. In MENA, the share of intraregional agricultural exports is increasing; however, the share of intraregional agricultural imports is stable or decreasing, i.e., the region continues to depend largely on outside sources for its agricultural imports (table 3). This, as mentioned above, is partly the result of trade distortions within the region. For example, official trade within the Maghreb amounts to less than 3 percent; however, this is supplemented by "unofficial" trading that takes advantage of differential pricing and subsidy systems among the different countries. Distortions originating from developed countries, e.g., subsidized cereal exports from the EC or the United States, also compete with domestic production and threaten the potential of the region's cereal exports. Agricultural research and extension (R&E). R&E has consistently earned high rates of return. MENA governments had been allocating meager resources to agricultural R&E until the 1980s, when the situation improved somewhat with adequate expenditure growth, although the austerity measures toward the end of the decade may have slowed this growth (table 4). - 44 - TABLE 3. MENA AGRICULTURAL TRADE (THOUSAND US$) EXPORTS TO MENA EXPORTS TO WORLD % EXPORTS TO MENA COUNTRY 1975 1980 1985 1975 1980 1985 1975 1980 1985 Mororco 16,688 30,966 23,756 419,599 747,040 604,297 3.98 4.15 3.93 Algeria 1,086 4,384 703 174,832 121,139 57,541 0.62 3.62 1.22 Ti'aiaa 57,735 25,579 24,868 177,720 179,834 171,316 32.49 14.22 14.52 Libya 87 n/a n/a 398 n/a r)/ 21.86 n/a n/a E 70,964 60,963 43,023 78,4321 680,164 325,118 9.05 8.96 13.23 Tur}y 138,823 378,838 914,741 921,390 1,881,308 2,371,243 15.07 20.14 38.58 Syria 49,102 49,925 58,578 188,384 282,394 205,353 26.06 17.68 28.53 Jordan 36,576 10,573 114,916 36,999 102,179 119,026 98.86 98.43 96.55 Iraq 24,290 16,029 nia 58,307 63,086 Iva 41.66 25.41 n/a Lehan n/a 236,491 n/a n/a 27,706 MS nWS 84.55 n/a Saudi Arabia 22,017 89,690 n/a 26,545 116,237 n/a 82.94 77.16 Ar/ Yemen A.R. 1,185 9,523 n/1L 10,102 11,056 n/a 11.73 86.13 n/a Yemen P.D.R. /a 4,428 n/a n/a 31,670 n/a n/a 13.98 n/a Oman 91 25,672 42,775 3,098 31,054 62,689 2.94 82.67 68.23 Kuwait 31,853 108,058 n/a 40,079 122,562 n/a 79.48 88.17 n/a Babrain 13,134 1,579 n/la 18,018 22,449 n/a 72.89 7.03 n/a U.A.E. 33,179 n/a n/a 41,140 n/a n/a 80.65 n/a n/s Iran 28,750 26,752 n,a 306,277 229,751 n/s 9.39 11.64 n/a Toall 525,560 1,169,450 1,223,360 3,212,210 4,801,631 3,916,582 16.36 23.86 31.24 IMPORTS FROM MENA IMPORTS FROM WORLD % IMPORTS FROM MENA COUNTRY 1975 1980 1985 1975 1980 1985 1975 1980 1985 Morocco 7,377 8,698 5,488 868,789 1,081,522 865,400 0.85 0.80 0.63 Algria 55,768 39,152 95,689 1,412,596 2,525,579 2,837,338 3.95 1.55 3.37 Tunisia 7,855 42,362 18,659 309,851 602,728 464,504 2.54 7.03 4.02 Libya 80,809 62,217 n/a 675,818 1,396,493 n/s 11.96 4.46 n/a Ep 37,268 36,444 9,127 1,613,568 1,863,719 1,857,771 2.31 1.96 0.49 Turley 817 126 18,117 426,952 347,619 861,680 0.19 0.04 2.10 Syria 49,585 206,181 62,689 406,661 711,574 850,809 12.19 28.98 7.39 Jordan 57,712 104,646 86,587 177,956 471,073 537,161 32.43 22.21 16.12 Iraq 47,931 168,6D09 na 852,934 1,721,577 n/a 5.62 9.79 n/a Lebanon 117,796 82,498 n/a 520,289 545,668 nta 22.64 15.12 n/a Saudi Arabia 180,910 326,804 380,019 706,945 4,623,449 3,777,54 25.59 7.07 10.06 Yemen AR. 9,940 93,884 n/a 133,799 528,381 n/a 7.43 17.77 n/a Yemcen P.D.R. n/a 10,428 n/a nta 173,352 nla n/a 6.02 n1/a Oman 43,021 80,657 146,427 103,921 288,921 457,814 41.40 27.92 31.98 Kuwait 67,643 158,739 n/a 424,782 1,033,781 n/a 15.92 15.36 n/a Bahmin 11,037 15,861 n/a 80,596 260,034 n/a 13.69 6.10 n/a U.A.E. 57,595 62,724 n/a 319,137 1,010,951 n/la 18.05 6.20 n/a Iran 76,047 36,351 n/a 2,172,166 2,298,176 1/a 3.50 1.58 n/a Toal 909,110 1,536,378 S23,003 11,206,759 21,484,596 12,509,980 8.11 7.15 6.58 n/a Not avalable. Note: Export ame fo.b., and inpors are c.i.f. Conmoditly coerae inchale food, feed, and aicaltural raw materiab (SITC Rev. I categoriea 0+1+2+4-23-27-28). Figu4e mray not add exactly becauae of rounding. Scarce: UN Trade Data Systm fiks. The organization and priorities of agricultural R&E in MENA have to be reoriented from basic staples and export crops, such as cotton and tobacco, toward more diversified, higher-value crops. For example, in Jordan, between 1980 and 1984, field crops contributed 11 percent to agricultural GDP; vegetables and fruits, 64 percent; and livestock products, 24 percent. However, the three groups' respective shares in public research expenditures were 38, 54, and 8 percent (World Bank data). Current public R&E systems need to enhance their flexibility and expertise to service the needs of more diversified agricultural production patterns. This process will include not only public R&E institutions, but also the agricultural education systems, from technical/vocational to higher education. The private sector, particularly in horticultural products, can play an important role in R&E if an appropriate policy - 45 - environment can be developed (R&E has been successful when the private sector TABLE 4. AGRICULTURAL RESEARCH has been active, for example in Jordan, EXPENDITURES (THOUSAND 1980 US$) where it launched the diffusion of COUNTRY 1970 1975 1980 MOST CURRENT (YEAR) improved irrigation technologies). Areas that will require greater attention include Nonrh Afica breeders' rights, shortening of quarantine Algeria n/a n/a n/a 16,120 (1986) Egypt 20,854 21,248 30,214 32,793 (1983) and adaptive trial periods, acceptance of Libya n/a 1,702 2,793 2,988 (1986) international phytosanitary standards, Morocco 5,696 13,595 15,562 15,367 (1983) certification facilitation, marketing, and Tunisia 2,807 4,355 6,770 11,469 (1985) elimination of price controls over seeds Subtotal 29,357 40,900 55,339 78,737 and agricultural inputs. West Asia Iraq n/a 1,993 4,030 18,623 (1983) Rural finance. MENA's Jordan 383 750 850 850 (1980) Rural finance. MENA's Pakistan 4,696 30,620 31,757 17,903 (1986) agricultural credit systems are Syria 2,123 2,492 5,290 4,403 (1985) characterized by insolvency, dependence Turkey 18,530 22,900 30,000 29,315 (1981) on government funding and guarantees, Subtotal 25,732 58,755 71,927 71,094 and preferential rediscounting rates from the central banks (World Bank data). The n/a Not available. absence of viable agricultural insurance Source: Oram 1988. systems has resulted in the transfer of production risks to the agricultural credit system and ultimately to government budgets. The agricultural credit system is essentially outside the financial system and is distorted, as manifested by subsidized interest rates, low recovery rates, dependence on the government for funds and bailouts, and frequent government intervention (politically motivated) in granting amnesties or postponing interest and principal payments (World Bank data). The beneficiaries of these systems, however, have primarily been the large farmers. Because the transaction costs of servicing small farmers are high, lending to them is frequently commercially nonviable. Most grant schemes with social objectives that serve small farmers, however, operate as credit schemes, distorting the credit markets, reducing repayment discipline, and adversely affecting legitimate credit activities. The current focus is on rural finance: for example, deposit mobilization, commercial operations, integration into the financial sector, and lending for nonagricultural activities in rural areas. This perspective does have drawbacks: lending to the agricultural sector is risky and has high transactions costs. In a commercial lending portfolio, agricultural loans would inevitably have a small share, and agricultural interest rates would be high to reflect the risk premium and transactions costs. Land tenure. The rigid or virtually frozen land market, either for sales or rental, is an obstacle to development, particularly in the Maghreb. The lack of an effective land market prevents efficient use of land resources by those with the capacity and interest in bidding for land in such a market. The limited amount of land and the rigid land market frustrate not only farmers who want to expand their holdings, but also policymakers who see large tracts of absentee-owned, underutilized lands. This has repercussions beyond purely land use issues. For example, a farmer with a private well that provides more water than the farm needs but who is not able to expand farm area would have little incentive to incur the investment costs (e.g., sprinkler equipment) to economize on water or improve water use efficiency. Another issue is the absence of titles (or lack of legal value of existing ones), which makes land transactions and meeting collateral requirements in credit markets difficult. Fragmentation of farms into economically nonviable units, particularly through inheritance, is another serious problem. Other issues relate to costly and time-consuming cadastral surveys, complex documentation and legal regulations pertaining to private land transactions, and unplanned urban encroachment onto fertile agricultural land. - 46 - The Farm Level Farm lypologies Farmers in MENA can be classified according to agroecological zone, farm size, and dominant cropping pattern. The substantial variability in these typologies across the region will not be addressed here, but several characteristics do merit attention. Farm size distribution is skewed in most of MENA, with TABLES. LAND DISTRIBUTION IN the majority of farmers holding JORDAN (1983) AND TUNISIA (1986) (HA) small plots and controlling less AVG. than half of the cultivated land, FARM NO. OF FARM % OF % OF and a minority of large farmers SIZE HOLDINGS AREA SIZE HOLDINGS AREA controlling the remainder.7 For Jorn example, land distribution in 0-5 39,755 67,700 1.7 69.6 18.6 Tunisia (table 5) implies a Gini 6-10 8,981 59,200 6.6 15.7 16.3 coefficient of 0.65. This skewed 11-20 4,947 63,200 12.8 8.7 17.3 21-50 2,610 72,700 27.9 4.6 20.0 distribution, however, does not 51-100 569 35,600 62.5 1.0 9.8 imply "latifundia"- and 101+ 276 65,900 238.8 0.5 18.1 "minifundia"-type contrasts. Total 57,138 364,300 6.4 100.0 100.0 Larger farms typically constitute Tunsa hundreds of hectares, and small o-s 173,927 385,600 2.2 46.2 7.5 farms have less than 20. Farm size 6-10 78,128 573,000 7.3 20.8 11.1 also varies according to 11-20 67,704 984,600 14.5 18.0 19.1 21-50 43,094 1,401,800 32.5 11.4 27.3 agroecological zone. In some 51-100 8,720 618,800 71.0 2.3 12.0 countries, large farms are state-run 101+ 4,827 1,178,100 244.1 1.3 22.9 cooperatives or farms. Total 376,400 5,141,900 13.7 100.0 100.0 A myth about the MENA "subsistence" farmer must be Source: For Jordan, Department of Statistics, 1983 Agricultural Census, quoted in dispelled. A subsistence farmer is Qasem 1986; for Tunisia, Enquwes Agricoles, data files, 1986. typically considered to exploit the meager resources of the farm solely for food production for the farm family; to market very little, if any, of the farm production; and to be totally marginal to the commercial agricultural production system. Such farmers are rare in MENA, although there is poverty in the region's agricultural sector. The following discussion of farmer behavior pertains to both small and large farmers, but usually concerns small farms, i.e., "family farms." The family allocates its financial, labor, and managerial resources among the various activities in its portfolio, which invariably involves (a) agricultural production (crop and livestock), and (b) off-farm employment. The latter can be permanent, temporary, or seasonal, and in agriculture, other sectors, or abroad. Occasionally, it can also be a nonfarm enterprise managed by the family. The relatively meager agricultural resources Oand, livestock, etc.) of small farmers cannot usually support their daily needs and are complemented with off-farm activities. For large farmers, investments in nonagricultural activities such as trade and agroindustry are also common. The objective of the farm family is to optimize income with due consideration of risk. The composition of the farm family portfolio is a clear indication that risk management is a major concern. - 47 - Crop Production Large farmers usually practice monoculture, employ capital-intensive techniques and low cropping intensity,8 and are rarely involved in extensive livestock production.9 Small farmers use their land resources intensively, depending largely on family labor. For example, according to 1986 Tunisian figures, farms in the 0-to-5-ha and 6-to-10-ha categories had cropping intensities of 0.98 and 0.93, respectively, in contrast to the national average of 0.73. Average cropping intensities in the 0-to-S-ha category were as high as 1.20 in the northeast and 1.30 in the central-eastern regions of Tunisia. These average cropping intensities cover both irrigated and rainfed lands. High cropping intensities, however, may limit further intensification. As a corollary to these results, small farms allocate less land (in percentage terms) on average to fallow. Small farmers allocate a larger proportion of land than the national average to vegetable and fruit production and to tree crops, which also account for the high level of cropping intensity through alley cropping.'0 In terms of the national distribution of crop lands, small farms command small shares except for fruits and vegetables, tree crops, and pulses. Although the majority of livestock are raised in small farms, most forage production occurs in larger farms that can accommodate such product diversification. Livestock Production Livestock production in MENA is concentrated in the small farms; for example, in Tunisia 58 percent of livestock are in farms of less than 20 ha." This implies high stocking rates, but these livestock depend on natural resources outside the farm, such as common grazing areas (ranges, pastures, forests, and roadsides), as well as purchased feeds. The overexploitation of some of these resources, however, does have adverse agroecological effects. Small farmers generally raise small ruminants (sheep and goats) because they are relatively adapted to the harsher environments; are manageable by the less-active members of the family, such as the children and the elderly; and are able to utilize stubble and crop by- products efficiently. Livestock meat and dairy products provide valuable high-quality animal protein to the limited diets of rural families. Livestock production should not be viewed purely as a production activity; livestock are an important component of a farmer's risk management strategy because they also serve an income-stabilizing function as a store of wealth. Off-Farm Employment Off-farm employment is a major source of income in MENA's rural areas. For example, according to 1989 data, 41 percent of Tunisian farmers consider an occupation other than agriculture to be their principal activity (World Bank data).'2 The importance of off-farm income, particularly for small farmers, is reflected in other data. A 1984 survey of 240 farmers in Goubellat, Beja (Tunisia), found that 90 percent of the farmers had holdings of less than 30 ha and fell under the classification of small farmers based on the agroecological characteristics of the research area. Fifty-four percent of these small farmers had an off-farm activity (INRAT-INAT 1989). A 1982 study conducted under the auspices of the International Fund for Agricultural Development Rural Credit Project in Kef and Siliana (Tunisia) found that off-farm income was the major source of income for farms below 50 ha in the two provinces and constituted 39 percent of household income of farms in both rainfed and irrigated production systems and 49 percent of those under a rainfed production system only (FIDA 1982). Similar observations can be made in Syria, Turkey, Egypt, and other countries in the region (Adams 1991; Tully 1990a, 1990b). Off- farm employment utilizes family labor resources optimally. Disruption of this labor supply would adversely affect all economic sectors, as well as remittances. - 48 - Implications of Farmer Behavior The income from small farms in MENA is highhly variable. This uncertainty is addressed with a risk management strategy that integrates livestock with crop production and off-farm employment of family members.'3 The net balance of the flow of resources between the farm and nonfarm sectors is critical: is the small farm sector transferring net resources outside the farm or vice versa? The indications in MENA are that the net resource flows are toward the farm. Farmers may often indicate that they need off-farm income for food and clothing, but given the fungibility of funds, off-farm resources may be allocated to maintain the patrimony of the farm. The 1984 Goubellat study's conclusion supports this hypothesis: those with off-farm employment had better farm performance in terms of yields achieved. This could be attributed to the capacity to use purchased inputs, such as fertilizer, which are financed by off-farm income. The same situation has also been observed in Egypt (Adams 1991). In factor, input, and commodity markets, farmers invariably must operate with incomplete information and be subject to imperfections induced by policies and inadequate infrastructure. For example, in Tunisia, the organization of wholesale markets reduces their effectiveness in price discovery and hence in guiding farmers in their resource allocation decisions. In most of the region, input markets are controlled by the state or large state monopolies. Distortionary subsidy schemes implemented by these state enterprises are compounded by inefficiency in the timely delivery of adequate-quality inputs. Because the state plays a role in creating these market imperfections, reducing its economic activities, while increasing the private sector's role, is a prerequisite to improving market performance. The state is better equipped to focus on well-defined areas of infrastructure and regulation (discussed below). In many areas, the evolution of cooperatives, which would facilitate increased farmer bargaining power-particularly for small farmers-has been impeded by state intervention. For farmers, the arbitrary economic signals transmitted by imperfect markets under state control create uncertainty that hampers decisionmaking on investment and diversification. A serious implication relates to the effectiveness of policy interventions, used here to cover the entire spectrum of government actions directed toward the sector, from macroeconomic policies to research and extension to credit. For example, the aim of the government may be to influence crop production. But if calculations of the viability and effectiveness of the interventions are based solely on costs and benefits in the crop production sector, then the policy interventions may fail. In certain cases, the opportunity costs of adopting an innovation may be overlooked, particularly costs related to foregone off-farm income.'4 Because increasing production and enhancing food security motivate national actions, policy interventions rarely consider the stability of farm family income. In an uncertain world, stability of income is an overriding concern for most farmers, particularly the small ones. Newberry and Stiglitz emphasize that "producers are concerned not so much with price variability as income variability . . . [but] it is . . . variability . . . in consumption that is ultimately important" (1985). Attempts to stabilize one component of risk-averse farmers' portfolios may adversely impact their income stability. Newberry and Stiglitz demonstrate that because "the income from a given crop is not the only source of income for many individuals . . . even if one were successful in reducing the variability of income generated by a particular crop, the effect of this on the welfare of the producer may be minimal. . . . [Ilf the farmer is widely diversified and the income variability from one crop were uncorrelated with . . . other sources of income, then the variability in [the crop] income has a negligible effect on the welfare of the individual, and hence its elimination would have a negligible effect. For crops whose return is negatively correlated with income from all sources in the aggregate, the reduction in income variability would actually increase total income risk and hence have a deleterious effect on welfare" (1985). Agricultural policy interventions in MENA, therefore, must consider welfare effects on the total portfolio of activities from which the farm family income is derived. In the context of diversification, these implications are important. As indicated above, stabilization policies can actually have destabilizing effects on farmer incomes and thus be unpopular. Similarly, - 49 - efforts to diversify agricultural production in the aggregate must assess not only the potential production and farmer income increases, but also the stability of farmer income. Agroecological Constraints and Potential The basic premises underlying the potential for diversification are the reallocation of existing resources or the creation of additional resources through: * expansion of cultivable land; * reallocation of existing land, including changes in cropping mix and patterns; * increases in cropping intensity; * increases in productivity. Agricultural land in MENA has practically and ecologically been exhausted. There are limited possibilities for bringing marginal amounts of land under cultivation through (a) irrigation of previously uncultivated land, and (b) land reclamation. Because of the high investment costs required by such projects, their efficient use will be critical. Agriculture (crop and livestock production) in MENA is practiced under four broad agroecological conditions: (1) less than 200 to 250 mm annual precipitation; (2) 250 to 600 mm annual precipitation; (3) irrigated agriculture; and (4) microzones: oases, high-rainfall, and controlled (e.g., under glass or plastic) environments. In some areas, such as the Central Anatolian Plateau, the Atlas mountains of North Africa, and Baluchistan, winters are severe, thus requiring winter-hardy cultivars. Winters, however, are mild in the rest of the region. Low-Rainfall Zones The first zone (ess than 200 to 250 mm average annual precipitation) is not utilized for crop production except in years of high precipitation and is then primarily devoted to grazing barley (e.g., in Syria and Jordan). The dominant mode of utilization of these areas (steppes, deserts, marginal lands, etc.) is for grazing small ruminants. Because of specific policies such as subsidies for supplementary feeds (e.g., in Tunisia) that encourage herd sizes above the carrying capacity of the resource base, these lands are under tremendous pressure and the ecological balance is at risk. Efforts to promote diversification by bringing new lands (e.g., in Algeria) under cultivation in these areas, under rainfed or irrigated production systems, or by creating watering holes, need to be carefully assessed because such developments would entail adverse externalities for lands already under substantial stress. Variable-Rainfall Zones The second agroecological zone encompasses a wide spectrum of rainfall levels, with barley- livestock systems dominating the lower end and wheat-based systems dominating the upper end of the precipitation range. Large tracts of land in this area are left to fallow in two- or three-year rotations. The potential for diversification in this zone can be based on the reallocation of land, increased cropping intensity, reduction of fallow periods and areas, and increased yields. In MENA, there is movement in these directions. For many countries, because of land constraints, changes in the area allocated to individual crops are possible only through the reallocation of land. For farmers, this option is feasible, provided relative prices are right, the risks are acceptable, there is adequate R&E support, marketing is not problematic, and there are no binding resource constraints. Pakistan, for example, is implementing a project to expand - 50 - its oilseed area to increase domestic production of cooking oil, at the expense of other crops such as wheat and maize. Turkey's Thrace region also frequently shifts between sunflower and wheat areas in response to changes in the relative support prices of the two commodities. Diversified changes in the crop mix and the incorporation of new and exotic crops are technically feasible. For example, in Tunisia, the potential for castor bean, winged bean, amaranth, guar, buffalo gourd, jojoba, bambara bean, Acacia albida, Cassia sturtii, Atriplex, tamarugo, and cadelilla should be studied; eventual production, however, depends critically on the provision of technical marketing and support (World Bank data). In addition, in this agroecological zone some areas are subject to serious erosion problems resulting from the cultivation of annual crops on sloping or stony soils. These areas should be planted with trees (e.g., almond, pistachio, and olive); such developments are underway in Turkey, Syria, Jordan, and Tunisia. With reduced fallow periods and areas, the issues involved become more complex. One must distinguish between higher- and lower-rainfall areas and how fallow is reduced. Fallow has traditionally been part of cereal rotations in MENA, even under irrigated conditions. Fallow has several functions: to store moisture; to build-up nutrients; and in weedy fallow, to provide fresh grazing material for livestock after winter."5 The successful Central Anatolian Wheat Project in Turkey emphasized moisture storage techniques during the fallow period; however, they were less successful in other areas of the region, such as Jordan and Tunisia, because of soil and climatic patterns. Land limitations, however, inevitably strain the fallow system. In Turkey, a fallow reduction project in the Central Anatolian Plateau was able to increase the production of food legumes in the 1980s; food and forage legumes had positive effects on soils vis-a-vis nitrogen fixation, supplementing some of the nutrient buildup during the fallow. In other areas, however, fallow is being replaced by continuous cereal cultivation, as with wheat and barley in some areas of Syria. Continuous monoculture in lower-rainfall areas has adverse consequences, which include declining yields as a result of deteriorating soil structure and nutrient content. Efforts to introduce other crops such as food and forage legumes into the rotations to replace the fallow period or to break the continuous monocultural production of cereals have been considered. These efforts have been designed particularly to provide cheaper protein for human consumption and valuable feed for livestock. These efforts, however, face several problems: (a) under existing prices and production costs, continuous production of cereal appears to be the most profitable option, influenced partly by food self-sufficiency policies; (b) food legumes are labor-intensive, particularly for harvesting, and entail weed control problems; (c) forage legume seed availability is a major constraint; (d) some forage legumes, such as medics, have not succeeded, although they originated in the region, primarily because of the nontraditional management techniques required; and (e) unless wide contiguous areas are planted with forage legumes, there is substantial risk of severe bird damage. Evidence from Jordan indicates that the fallow system is also associated with farmers' risk-averse behavior, and the introduction of food legumes in rotations is undertaken by less-risk-averse farmers (Hassani 1988). This diversification alternative therefore needs to be carefully planned and assessed considering both economic and environmental effects. Another possibility for the reduction of the fallow period is using part of the fallow period for summer cropping, whereby a summer crop can follow winter fallow. This depends on the availability of residual moisture and the probability of summer rains. In certain areas (e.g., northwest Syria), however, groundwater is used to provide a minimum of irrigation. This water could be used for supplementary irrigation for the winter crops to increase and stabilize yields, but this practice is not yet widespread (see below). In other areas, the paucity of available water inevitably results in low-intensity summer cropping. The crops involved are cotton, tobacco, vegetables, and fruits, particularly low-density melons and watermelons, which depend on a viable marketing network or the proximity of large markets. Yield increases will promote diversification primarily by impacting the relative profitability of different crops and livestock. Evidence of development is reflected by the gradual substitution of barley for wheat in rainfed wheat areas as a result of increased feed requirements for livestock, which in itself - 51 - is the driving force affecting barley production (Mona 1986). The higher and better-distributed rainfall in the traditional wheat areas will primarily affect the stability of barley production and yields and in the longer run could enable the introduction of higher-yielding cultivars resistant to pests and disease. This shift would allow barley producers, for example in Syria, to change the minimum input (virtual mining) techniques and adopt fertilizer and other inputs that increase not only yields but also yield stability (Somel et al., forthcoming). Furthermore, because of the influence of livestock production, it would be more feasible to introduce forage legumes in these higher-rainfall areas. Wheat, however, is increasingly cultivated in even higher-rainfall and irrigated areas for two major reasons: (1) government risk management policies are encouraging wheat production to achieve self-sufficiency, and (2) the higher yield potential of the improved cultivars can be better exploited in these areas. In this agroecological zone, other crops can also be introduced or their current production increased, provided their yield potential is conducive to profitable and stable production. There are, however, constraints for crops such as food legumes (lentil, chickpea), oilseed, sesame, and groundnut, which are exacerbated by marketing problems. Another development (mentioned above) is the supplementary irrigation of crops, mainly cereals, to meet water requirements, particularly during critical stress periods or droughts. This is not irrigation per se, because the supplementary irrigation compensates for fluctuations in rainfall and ensures a minimum amount of moisture to stabilize yields. The supplementary irrigation of cereals invariably does not provide adequate economic justification for the investments required for wells and irrigation equipment; hence, such irrigation is possible only if the investments have been justified by other activities such as the summer production of fruits and vegetables. In some areas, however, such as the Deraa province in Syria, the government has invested in irrigation networks based on underground water for supplementary irrigation of cereals. Despite these efforts, supplementary irrigation remains a technology that has not yet gained wide acceptance because of (a) the unavailability of the necessary technical information to farmers in the cereal-growing areas; and (b) the high value of water, which may cause farmers to reserve it for higher-value crops. Irrigated Areas Technically, the irrigated areas have the highest potential for diversification. Although they constitute the majority of agricultural lands in a few countries, such as Pakistan and Egypt, they cover a minor portion of agricultural lands in the rest of the region. Nevertheless, a large part of agricultural production value originates in these irrigated areas. For example, in Tunisia, approximately 6 percent of agricultural lands are irrigated, yet they produce 30 to 40 percent of agricultural GDP; in Jordan, these percentages are 15 and 60 to 70 percent, respectively, and in Morocco, 14 and 40 to 50 percent, respectively. MENA's irrigation systems, unlike those in Asia, are not single-purpose systems; therefore, they are considerably more flexible in accommodating a wide variety of cropping systems, given adequate amounts of water. That irrigated areas receive high priority in public investment and expenditures can be justified to a certain extent by the considerably higher returns per unit area. Paradoxically, policies encourage or require the cultivation of lower-value crops, such as cereals, under irrigation. In Jordan, for example, citrus, other fruit trees, and vegetables under improved irrigation techniques yield a net income per hectare that is over five times that of irrigated wheat, which only yields a net income three times that of rainfed wheat under improved practices (Qasem 1986). Major problems in these irrigated areas are the low system- and farm-level water use efficiencies resulting from open canals and traditional irrigation technologies. For example, in the public irrigation perimeters of Tunisia, system efficiency has been estimated at about 75 percent and on-farm water use efficiency at about 45 percent. This implies that only one-third of the water harnessed by the public irrigation systems is actually used for agricultural production, challenging the calculations of economic -52 - returns to irrigation investments. As discussed above, water pricing does not provide adequate incentives for the efficient use of water. In the long run, when water becomes scarcer and has to be rationed, its use will be influenced by its shadow price rather than any administered price: this is already seen in areas with critical water shortages, such as the Jordan Valley. A water-pricing system that recovers full costs, however, will have far-reaching effects on irrigated agriculture, affecting comparative advantage and possibly producing radical changes in cropping patterns (e.g., eliminating low-value crops such as wheat and sugar beet) and prompting the production and marketing of high-value commodities. Currently, it does not make economic sense to allocate irrigated land, such as in the Disi aquifer area in south Jordan, to wheat, even as a risk management strategy. Using treated wastewater from urban systems for irrigation is an option. The potential for this is currently limited to less than 5 to 10 percent of the total water supply. Sanitary and quality controls must be fairly strict; in some export markets, such as the EC, there are already import restrictions on produce irrigated with treated water. There is some potential to increase irrigated areas, and many countries, such as Turkey, Tunisia, Algeria, Jordan, Yemen, and Pakistan, are making the necessary investments to exploit this potential, despite the riparian issues sometimes involved. Unfortunately, operation and maintenance of existing infrastructure and the more efficient use of harnessed water are being neglected, and for such valuable land, cropping intensities are dismally low. Therefore, the potential for diversification is high through increased cropping intensities in irrigated areas. There have been some successful experiences in Turkey with the introduction of maize and soybean in irrigated areas that were traditionally under cotton monoculture. Other possibilities are sugarcane in Egypt, and fruits and vegetables in many irrigated areas. The issues related to fruits and vegetables involve domestic market satiation, export market restrictions, quality, storage, and packaging standards. In some areas, such as Jendouba in northwest Tunisia, fruit and vegetable production is restricted in irrigated areas to prevent continuous cultivation, which may have adverse effects on soils and can lead to disease outbreaks, and rotations of sugar beet, forage crops, and cereals are promoted. These crops are produced primarily as a response to policymakers' risk management considerations, and incentives are provided to farmers through prices and lower water charges. The basic limits to yield increases under irrigation stem from the availability of profitable technologies and R&E support. Cotton yields in major production areas in Egypt, Syria, and Turkey are either stagnant or declining, due primarily to R&E deficiencies and adverse government controls, particularly in Egypt. The private sector, however, has been successful in diffusing advanced irrigation technologies in Jordan, and private sector involvement in the promotion of new varieties and crops, particularly in fruits and vegetables, would be promising in irrigated lands. With the relaxation of seed import regulations, foreign and domestic investment in the seed industry (maize, potato, fruits and vegetables, etc.) increased significantly in Turkey, and the R&E capabilities of large international seed firms filled the vacuum in the national public R&E systems for such commodities. Microzones Microzones with agroclimatic and agroecological advantages represent a small land area, but have significant potential. Areas such as the Jordan Valley, the eastern Black Sea coast of Turkey, and the oases of southern Tunisia (controlled-environment production under plastic or glass is also included) enjoy particular advantages that allow them to produce high-value crops, either early in the season or out of season, for major domestic and export markets. The returns from such activities are particularly attractive, even if they require relatively advanced cultural and management techniques. In Jordan, tomatoes and cucumbers under plastic yield a net income per hectare that is 2.5 to 4 times that of vegetables under improved irrigation. Satsuma, tomato, cucumber, asparagus, and strawberry can be produced profitably in these microzones, and dates are a typical product of the oases. In export markets, - 53 - however, there is increasing competition, particularly from Latin America, which can fill the same export niches. There are also quality and quantity restrictions in the major importing countries, particularly in the EC. The expansion of production under controlled environments must be monitored for excessive use of chemical pesticides in southern Turkey, Syria, Jordan, and Tunisia. The Policy Framework for Diversification Agricultural diversification must be viewed in the context of a general economic and agricultural strategy. For example, diversification will be influenced by macroeconomic parameters such as the exchange rate as well as other aspects of trade policies. Currently in MENA, the degree of diversification is a function of the tradeoffs between shifting the costs of agricultural production instability to farmers and relieving the budgetary pressures of state expenditures. It is neither politically nor economically feasible to shift all the risk to farmers without serious production disruption, nor is it feasible for governments to perpetuate the status quo of assuming a majority of the risks. In the agricultural sector, the outcome will gravitate toward less government direct intervention in the market, a greater governmental role as regulator, and a more active role for the private sector and the market in resource allocation. Comparative Advantage The primary concern regarding exploitation of comparative advantage is that economic policies affecting international trade should be reassessed in view of liberalizing trade; eliminating quantitative restrictions; installing a manageable tariff structure; simplifying foreign exchange management; aligning overvalued exchange rates to market rates; and eliminating price, margin, and marketing controls for export and import commodities. This liberalized environment will facilitate production diversification to exploit comparative advantage and explore diverse markets, including increased intraregional trade, which currently plays a small role in the region's total trade. For many commodities with fluctuating domestic and international prices, it is critical to facilitate an environment conducive to the flexible responses of producers and traders to market information, incentives, and opportunities. Economic policies that allow MENA countries to exploit their comparative advantage in the agricultural sector may produce specialization, but may also lead to diversification away from existing cropping patterns. In MENA, because of climatic advantages, horticulture is a prime area for diversification, although quality, standards, domestic market satiation, and export market constraints and restrictions are major obstacles. Furthermore, horticultural production for export markets has been sensitive to the public management of the foreign exchange rate: in Jordan, the large volume of exports to the Gulf countries drastically declined in the latter part of the 1980s as a result of the maintenance of an overvalued exchange rate, and market share was lost to competitors such as Turkey (Youngblood et al. 1989). The market for horticultural products is also more sensitive to price variations and does not have the benefit of the stabilizing effects of futures and options markets. Nevertheless, state interventions in these markets should be minimal to ensure that production can respond to changes in prices and market conditions. A related issue is the production of the major cereals, wheat and barley. Subsidized imports from the developed countries have affected cereal producers in the region. If, for the foreseeable future, MENA governments anticipate international prices that are lower than domestic production costs, they will continue to depend on imports. However, there have been changes in the policies affecting domestic cereal producers; they are now provided with price supports, subsidized inputs, etc., and even some protection against dumping (Tunisia). Cereal production will probably continue to be supported for two major reasons: first, governments will maintain the support for cereal production for buffer and strategic - 54 - stocks purposes; second, in some areas, cereal production is among the limited options available for crop production, if not the only option (particularly barley production in the drier rainfed areas). A challenge is to avoid the adverse ecological consequences of the predominance and continuous monocultural production of cereals. Policies that focus on the promotion of food and forage legume production, with their associated beneficial effects of nitrogen fixation, would be a desirable way of reducing cereal monoculture. Although food legumes would have a role in the rotations in rainfed areas, one country seems to dominate the market in the region. Turkey can effectively meet a major part of the region's demand and may reduce the comparative advantage of the other countries in the region through economies of scale in processing, storage, and marketing. Policies for forage legume production must also ensure that the increased availability of fodder substitutes for the dwindling capacity of common grazing areas and does not play a role in increasing livestock populations that put additional pressure on such areas. For the region to exploit its comparative advantage in small ruminants, more diverse feed sources need to be used, more resources must be allocated to livestock productivity rather than maintenance, and new management techniques must be explored. These new methods will entail managed grazing of fodder; enclosures; exclosures for rejuvenation of common resources; and "stratification," or new linkages among producers for raising lambs and fattening. It is, however, vital to reduce livestock production subsidies, e.g., subsidized feed and free veterinary services, that impose a sizable budgetary burden. In Tunisia, during the seventh plan (1987-91), livestock feed subsidies were expected to constitute 59 percent of agricultural production subsidies, 14 percent of current agricultural expenses, and 7 percent of public agricultural expenditures. In Turkey, between 1986 and 1988, feed subsidies ranged from 18 to 28 percent of input subsidies and between 0.9 and 1.8 percent of the value added in agriculture. Consequently, larger herds are maintained, particularly during drought, than can be supported by the natural resource base. Marketing Marketing is a critical constraint in the diversification process. More than two-thirds of the agricultural value added in MENA is created outside the farm. Price discovery is also critical in the diversification process because a market with price uncertainty will not be conducive to farmers' risk taking in changing product composition. Facilitating price discovery requires minimum intervention in markets so that they perform this function efficiently."6 This also implies infrastructural investments in communications facilities and regulation of central, wholesale, and retail markets as well as the establishment of norms and standards. Newberry and Stiglitz (1985) have emphasized the importance of the availability of information to the speed of adjustment. An important tool that would facilitate price discovery and also enhance price stability is somewhat inaccessible to MENA producers and traders. This mechanism is the futures and options markets (FOMs). Access to established FOMs in other countries is restricted specifically and generally by international trade policies and foreign exchange controls. The commodities for which contracts would be traded in FOMs are relatively limited compared with the currently and potentially more diversified list of commodities in MENA. Regardless, even if access to FOMs facilitates price discovery and stability only for some commodities, it would enhance overall decisionmaking in the allocation of resources among the more diversified product mix. Futures markets also have the advantage of achieving stabilization through the market, whereas the nonmarket price stabilization schemes currently preferred by most MENA governments, such as support prices and subsidies, intervene in the efficient operation of these markets. There are, nonetheless, indications in the region of interest in FOMs, such as the proposed cotton futures market at the Izmir Commodity Exchange in Turkey. Product differentiation through quality standards, processing, and packaging is an important diversification strategy. However, it first requires enforceable standards, and second, effective incentives and price information on which to base decisions. In MENA, standards for products and packaging either - 55 - do not exist or are not enforced. Furthermore, controls over prices and margins along the marketing chain discourage quality differentiation. Consequently, seasonality and shelf life are not major concerns for producers and traders. However, although domestic markets may not warrant investments for sophisticated standards, they are indispensable in export markets, particularly in high-income and developed countries. These standards span critical areas such as sanitary standards and agrochemical residues."' The lack of standards puts most of the burden of quality assurance on exporters, thereby increasing their costs. The last step in the transformation of many agricultural commodities involves agroindustry. In many MENA countries, this sector is a stronghold of public enterprises and is invariably characterized by inefficiencies, losses, waste, excess capacity, and overstaffing. Products are usually based on outdated technologies and are of substandard quality. Procurement processes do not reward quality. Furthermore, the agroindustrial sector is either excessively dependent on domestic production for the provision of raw materials (a major factor in excess capacity) or excessively dependent on imported raw materials, creating an enclave environment with little effect on domestic agricultural production. A more flexible agroindustry is desirable to promote diversification. First, the bureaucratic inertia of public enterprises must be replaced by a more active private sector. Where this is present, such as in the fruit and vegetable canning industry in Turkey, the situation is more dynamic. Second, trade controls must be relaxed so that joint ventures can be undertaken, more modern technologies adopted, and new products developed for domestic and export markets. Finally, the state must disengage from active participation in production, confining itself to a regulatory role. Infrastructure To promote diversification, the state must invest in infrastructure for marketing, rural roads, telecommunications, and irrigation systems. In particular, infrastructural investments that facilitate market integration and reduce the effects of limiting factors, such as water, will provide an environment conducive to diversification in agricultural production. More intensive marketing networks, improved communication of market and price information, and improved rural roads will reduce marketing costs and the risks associated with diversification. Concurrently, reducing the sources of variability resulting from climatic factors will permit a more stable production process on which to base diversification decisions. In financing infrastructural investments in agriculture, governments must not crowd out the private sector and must ensure that, where feasible, investments by the private sector are encouraged and that civil works are contracted to the private sector. Research and Extension As discussed above, agricultural research and extension systems are biased toward the basic cereals, mainly wheat, and some critical industrial crops and export commodities such as cotton. Institutional and structural changes to adapt the national public research and extension systems to a more diversified production pattern will entail substantial investment in time and training (including retraining). This may be more costly than generally envisaged and may not be timely enough to respond to imminent needs. The solution, which has successfully been implemented in some countries (e.g., Turkey), is to allow the private sector to conduct research and extension. This may sometimes involve close cooperation between the public and private sectors. Generally, however, it will involve licensing arrangements or joint ventures with international companies (particularly for horticultural products) and with livestock, agrochemical, or agromachinery firms, thereby allowing the vast research resources of these international companies to benefit the region. Another issue is to ensure that imported materials are well adapted to the local environment. For example, public livestock services in MENA have been invariably inclined to import breeder stocks of - 56 - sheep, goat, and cattle with prestigious pedigrees, though hardly adaptable to the local climatic conditions and management techniques. The results are inevitably a waste of the pedigree animals at the butcher block. It must be ensured that the private sector does not follow a similar track; rather, it should establish parallel or multiple-location and multiple-season adaptive trials and link certification to this process. There are several examples in the region of developments in private sector R&E activities. Turkey, by opening up seed imports and shortening the quarantine process, accelerated the flow of new materials. The private seed companies actually reaped the benefits of government investment in human capital by recruiting breeders, agronomists, etc., from the public sector. Many companies in Jordan and Saudi Arabia are benefiting from importing new technologies for horticultural and livestock operations. In some cases, however, some of these operations are viable due to substantial state subsidies, which many governments may find unsustainable in the long run. Despite short-term problems, many of these developments will gravitate toward a system emphasizing a higher degree of diversification in agricultural production. Rural Finance Theoretically, rural finance can facilitate the process of agricultural diversification by providing credit to farmers and traders. Generally, however, a current assessment of the demand for agricultural credit or rural finance in MENA has not been made; the extent to which credit can be used to finance diversification in MENA is therefore unclear. Generally, the move toward integrating rural finance into the financial system and eliminating preferential treatment (e.g., subsidized, even negative, real interest rates) and directed or supervised credit schemes will reduce the funds available for agriculture because of the higher risk and transaction costs associated with it. There are indications that farm families do generate funds to finance their agricultural operations and investments. Some of these resources can be allocated to financing diversification of crop and livestock production. Inevitably, however, some temporary programs to promote diversification can and will have to be financed with concessional credit. Social Services In the general provision of social services for education, nutrition, and health, state interventions have externalities for agriculture generally and diversification specifically. An enlightened rural population will make sounder decisions and be more efficient in agricultural production. Similarly, better health and nutritional conditions, particularly for women and children, enhance productivity. Although these may appear peripheral to diversification, they are important to an efficient and productive agricultural sector. Private Sector Role Diversification of agricultural production is ultimately a private sector matter. In most MENA countries, for example, public veterinary services operate with top-down approaches, emphasizing sophisticated methods and high pedigree breeds, whereas farmers' needs have to be assessed within a more adaptive strategy. Privatization of veterinary services will allow farmers to express their specific demands and allow market forces to determine the nature of veterinarians' work. Government services can then focus on crisis management for epidemics and similar public service endeavors that private veterinarians would not have the incentive to undertake. A similar observation can be made for rural cooperatives for production, marketing, etc. Typically, such organizations are state creations. For cooperatives to be effective institutions that promote farmer interests, they have to be the product of existing social forces, without state interference, and be able to operate as private sector entities. Individual farmers will determine the feasibility of new crops and changes in the cropping mix. Their decisions will be based on the effects of such diversification on the growth and stability of their - 57 - incomes, which in turn are also influenced by the availability of market information and technical support. Diversification will also depend on domestic and foreign trade regimes' effects on the marketing process. In marketing, the public sector's record in MENA is dismal, as seen in, for example, the control of produce marketing in the former PDRY, cotton marketing in Egypt, and fruit and vegetable marketing in Jordan. Marketing is the domain of the private sector, and it should be facilitated by government policies. Overall, the greater the private sector responsibility for resource allocation decisions, the more flexible the production environment, relatively free of rigidities typically introduced by state interventions and restrictions. Land tenure is a critical issue, and security of tenure must be ensured for farmers to justify investments in capital and inputs. Land markets (sales or rental) must become more flexible so that valuable land is not left idle, as with some irrigated land in Tunisia, and so that those capable of using the land most efficiently are able to bid for rights of access and ownership. Measures must be undertaken for voluntary land consolidation and prevention of fragmentation of holdings. If necessary, fiscal measures to achieve these ends should be assessed, provided the results are socially acceptable. Conclusions Diversification of agricultural production in MENA will not be costless, and its benefits are not ensured. Although agroecological constraints in different areas exist, technically there is substantial potential for agricultural diversification and many ecologically sound options to reverse overexploitation in large areas. Provision of technical services by R&E, from both the public and private sectors, as well as the stable and timely provision of inputs such as seeds and agricultural chemicals, will be critical to the diversification process. To counter increased costs and risks for farmers and traders, the state should emphasize policies and investments that reduce uncertainty and variability in agricultural production and marketing. For agriculture, these involve infrastructural investments and policies that improve marketing, communication facilities, price discovery, irrigation, and agricultural research and extension. However, the broader set of macroeconomic policies, such as those involving exchange rates, interest rates, and trade policies, also affects the potential of agricultural diversification. Because these macroeconomic policies have broad implications beyond the agricultural sector, agricultural diversification must be viewed in the general context of economic development policies. Current stabilization schemes, focusing on influencing the relative prices of specific commodities, may actually be counterproductive in that they could increase the instability of farmer incomes. Other mechanisms should be explored, such as the use of futures and options markets and improvement of the marketing process through the establishment and implementation of norms and standards. Whether MENA governments would be responsive to such directional changes remains an issue. MENA governments, with their historical concerns about food security, will continue to allocate resources to maintaining policies for the domestic production of basic staples as part of their risk management strategies. They must find a second-best solution in which such policies do minimal damage and resources are allocated optimally. Diversification under constraints of basic risk management policies, as well as the natural resource base, will be the inevitable outcome in MENA countries. Some will want to use established subsidization methods to provide incentives to promote diversification into certain commodities. These methods, however, will provide only temporary solutions that will be quickly reversed once the incentives are removed. For agricultural diversification to be sustainable and institutionalized, allowing market forces to operate under regulatory mechanisms will be more effective as well as efficient. - 58 - Notes 1. The regional classification is based on 1991 World Bank definitions: MENA comprises Afghanistan, Algeria, Babrain, Egypt, Iran, Iraq, Jordan, Kuwait, Lebanon, Libya, Morocco, Omsan, Pakistan, Saudi Arabia, Syria, Tunisia, Turkey, U.A.E., and Yemen. Pakistan and Afghanistan have since been organizationally transferred to the Bank's South Asia Region. 2. In MENA, average annual rainfall in rainfed areas ranges between approximately 200 and 1,000 mm; however, the dominant range in terms of area covered is restricted to 200 to 500 mm. 3. As discussed below, the sources of production and yield variability are not restricted to climatic factors; agricultural policies, particularly interventions in absolute and relative input and/or output prices, fiscal measures, and controls over international trade and exchange rates also affect production variability. 4. Some infrastructure (e.g., dams, barrages) is multipurpose; other functions include power generation and flood control. 5. The magnitude of these rents is seen in northern Jordan, where horticultural crops are produced profitably under plastic cover, using water from the urban drinking water system. 6. A typical example is storage costs for agricultural products (cereals in silos, horticultural and livestock products in cold storage, etc.) These costs can be burdensome. The recourse to intermational markets is always an option, and is eventually chosen, yet governments insist on investing in storage facilities for strategic stocks as part of their risk management strategies. 7. Farm size distribution refers to the distribution of farm holdings, not to the distribution of land ownership. 8. Cropping intensity is defined as the ratio of land cultivated with different crops (excluding fallow) to total land available, including fallow. 9. Some evidence may seem to, but actually does not, contradict this generalization. For example, in the lower-rainfall regions of Syria, landowners, who received land through the land reform, are generally livestock producers and remain so. Their lands are tilled and harvested by sharecroppers who are the decisionmakers vis-a-vis crop production. These sharecroppers use farm machinery extensively. 10. The calculation of cropping intensities uses the actual cultivated area for each commodity group; if the common practice of double-counting tree crops and land in fruits and vegetables were used, the intensities would be much higher in small farms. 11. In ovine equivalent units, 1 unit of bovine = 7 units of ovine. 12. The breakdown of this figure by farm size was not available. The questionnaires used in these surveys ask for the principal activity of the head of the farm household and thus miss the off-farm employment of family members. The percentages may therefore underrepresent the extent of off-farm employment. 13. An exception is the small group of absentee landowners who are composed primarily of city dwellers, functionaries, or professionals who maintain a small farm. Their primary source of income is their profession, with the farm serving more as a hobby or weekend retreat. The proportion of such holdings is unknown, but the owners would not be classified as small farmers by any definition of the term. 14. A hypothetical example could be the recommendation to top-dress fertilizer in the spring. Although this may appear profitable in itself, its feasibility will diminish if the farmer possibly has to relinquish more lucrative - 59 - off-farm employment to go through the time-consuming process of returning to the farm, seeking credit, obtaining and transporting the fertilizer, and applying it. 15. In the 1970s in Turkey's Central Anatolian Plateau, the value of weeds (measured on a dry-matter basis) on fallow was estimated to equal two-thirds of the value of wheat grain. This does not include an assessment of the value of the vitamins and minerals in fresh weeds. 16. Some policymakers consider guaranteed prices or support prices a substitute for market price discovery. These prices, however, reduce the production system's flexibility. 17. As mentioned above, a critical limiting input is that markets such as the EC restrict the importation of crops produced with reused water (e.g., treated urban wastewater). Instruments and policies to ensure quality, including quality certification, will be indispensable as water becomes more scarce and wastewater is increasingly used in agriculture. References Adams, Jr., Richard H. 1991. "The Effects of International Remittances on Poverty, Inequality, and Development in Rural Egypt." International Food Policy Research Institute Research Report 86. Washington, D.C. Alexandratos, Nikos, ed. 1988. World Agriculture Toward 2000. London: Belhaven Press. Aricanli, Tosun, and Kutlu Somel. 1979. "Observations on the Nature of the Changes in Agriculture and Land Distribution in Turkey." METU Studies in Development (Winter) 6(22-23): 91-110 (in Turkish). FIDA (International Fund for Agricultural Development). 1982. Direction de L'Assistance aux Petits et Moyens Exploitants. Rome. Hassani, Mohammed A. 1988. "Small Scale Farming and Agricultural Risks in the Semi-Arid Areas of Jordan." Doctoral dissertation, McGill University, Montreal. INRAT-INAT. 1989. "Projet de Recherche sur Les Syst6mes de Production Goubellat-Tunisie, Cooperation INRAT-INAT-ICARDA-CRDI, Rapport Final 1983-1989." Draft. Krueger A., M. Schiff, and A. Valdes. 1988. "Agricultural Incentives in Developing Countries: Measuring the Effect of Sectoral and Economywide Policies." World Bank Economic Review 2(3): 255-71. Mona, Noureddin H. 1986. "Structure and Responsiveness of Barley Production in Syria." Doctoral dissertation, Texas A&M University, College Station. Newberry, D., and J. Stiglitz. 1985. The Theory of Commodity Price Stabilization. Oxford: Clarendon Press. Oram, Peter. 1988. "Agricultural Research Objectives and Priorities: Constraints to the Development of Agricultural Research Institutions in Arab Countries." Paper prepared for the first meeting of the Council for Arab Agricultural Research (CAAR). Washington, D.C.: World Bank. Qasem, Subhi. 1986. "Agricultural Land Use Policies." In Allison Burrell, ed., Agricultural Policies in Jordan. Guilford and King's Lynn, U.K.: Biddles. - 60 - Somel, Kutlu. 1979. "Technological Change in Dryland Wheat Production in Turkey." Food Research Institute Studies 18(1): 51-65 (special issue, Pan A. Yotopoulos and Lawrence J. Lau, eds., Resource Use in Agriculture: Applications of the Profit Function to Selected Countries). . 1986. "Agricultural Support Policies in Turkey, 1950-1980: An Overview.' In A. Richards, ed., Food, States and Peasants: Analysis of the Agrarian Question in the Middle East. Boulder, Colo.: Westview Press. . 1987. 'Food and Agriculture in West Asia and North Africa-Projections to 2000." ICARDA Report 128. Aleppo, Syria. . 1988. "The Importance of Barley in Food Production and Demand in West Asia and North Africa." In E.F. Thomson and F.S. Thomson, eds., Increasing Small Ruminant Productivity in Semi-Arid Areas. The Netherlands: ICARDA. Somel, K., A. Mazid, and K. El-Haji. Forthcoming. "Fertilizer Use on Rainfed Barley and Risk." METU Studies in Development. Tully, Dennis, ed. 1990a. Labor and Rainfed Agriculture in West Asia and North Africa. Dordrecht, the Netherlands: Kluwer Academic Publishers. , ed. 1990b. Labor, Employment and Agricultural Development in West Asia and North Africa. Dordrecht, the Netherlands: Kluwer Academic Publishers. World Bank. 1991. World Development Report 1991. New York: Oxford University Press. Youngblood, C.E., G.M. Scobie, and M.H. El-Hurani. 1989. "Macroeconomic Policies in Jordan: Implications for Horticultural Exporting." Research Triangle, N.C.: Sigma One Corporation. Zijp, Willem. 1991. "From Extension to Agricultural Information Management: Issues and Recommendations from World Bank Experience in the Middle East and North Africa." EMENA Technical Department Paper, volume 1, no. 2. World Bank: Washington, D.C. ENHANCING AGRICULTURAL GROWTH THROUGH DIVERSIFICATION IN SUB-SAHARAN AFRICA Steven Jaffee Agricultural performance in most African countries has been poor over the past two decades. Yields have stagnated or declined, as have per capita food production and exports. For Sub-Saharan Africa as a whole, the average annual growth of agricultural production, agricultural value added, and agricultural export value (in nominal prices) was below 1.5 percent between 1975 and 1990. The volume of Africa's agricultural exports also declined over this period, resulting in a reduced world market share for Africa for most of the region's major exports. Population growth remained above 3 percent, however, and low growth in the dominant agricultural sector has been a major factor in stagnant or falling living standards. Agricultural performance in Sub-Saharan Africa has been strongly hampered by adverse world market conditions for traditional export crops; poor macroeconomic and sector policies; excessive market controls; and inefficient production support and marketing systems that have undermined traditional subsectors and inhibited farmer and trader response to new market opportunities, both locally and internationally. Frequent droughts (especially in the Sahel and southern Africa), together with civil strife in particular countries and subregions, have also contributed to poor agricultural performance. A recent study by the World Bank (1990) argues that Africa needs agricultural growth of 4 percent over the next 25 years to achieve food security, raise per capita incomes, and meet its import needs. It concludes that "transforming agriculture and expanding its productive capacity is the prerequisite for improving living standards in Sub-Saharan Africa." Although Africa's agricultural performance over the past two decades suggests that 4 percent growth will be difficult to achieve, other regions-notably Asia and several African countries-have shown that high growth can be achieved and sustained if economic incentives are strong and agricultural and other support services are strengthened.' Sub-Saharan Africa has several potential sources of agricultural growth. First, although land and water resources are unevenly distributed, there is significant scope for expanding areas under cultivation in southern and central Africa, and there is considerable unused irrigation potential in about a dozen countries (Harrison 1990). Second, the region has a large yield gap for most industrial and food crops when compared with Asia and Latin America; major gains in land and labor productivity are possible, though they will require more intensive input use, improved animal and crop husbandry practices, better integration of livestock and farming systems, and greater use of high-yielding varieties. Third, an increasing focus on market development and trade-rather than market control-could boost agricultural performance. With support from improved infrastructure, large income and food security benefits can be achieved from initiatives to integrate local and regional agricultural markets; reduce logistic and transaction costs in domestic markets; and enhance quality, reduce marketing costs, and improve sales targeting for traditional agricultural exports. Diversification-at the farm level, in processing, and in trade-can be another important avenue of agricultural growth in Africa. It can encompass introducing nontraditional, high-value crops and live- Steven Jaffee is a consultant, Agriculture and Rural Development Department, World Bank. During the preparation of this paper, Shawki Barghouti and Harry Walters provided helpful comments. The sole responsibility for the paper, however, rests with the author. - 62 - stock products; expanding production and trade in secondary food crops; generating new uses for and products from existing crops; adding value to primary commodities through processing; and developing new market outlets. Agricultural diversification, itself a process of structural change, is intimately tied to the other potential sources of agricultural growth: the basis for and efficiency of agricultural diversification lie in improved resource use, increased farm and marketing productivity, and improved responses to changing market demand. This paper approaches the issue of agricultural diversification in Sub-Saharan Africa through country, regional, and thematic case studies that examine diversification trends and barriers at the production level, at the interface between farmers and processor/traders, and in intra- and extraregional agricultural trade. A common theme is the central importance of technical and market risk and incentives in the decisions of farmers, processors, and marketing agents to diversify into (or specialize in) new patterns of production, products, and markets and in their ability to do so successfully. The case studies highlight the often different attitudes that governments, farmers, and processing/trading firms have toward agricultural diversification and the respective roles of the public and private sectors in supplying the incentives and support services necessary for production and/or trade diversification. After some background on agriculture in the region, the paper first examines the Gezira scheme in Sudan, the world's largest irrigation scheme. The analysis focuses on the incentives and institutional bottlenecks for crop diversification at the farm level, emphasizing the divergence between farmer and government objectives in the allocation of natural and human resources. Next, Africa's experience with agroindustrial and new crop development under the institutional framework of contract farming is reviewed, revealing a degree of dynamism within the agricultural sector that is often overlooked in analyses of wider agricultural stagnation or "crisis" within the region. The next section examines a case of successful export diversification-Kenyan horticulture-focusing on the roles of the public and private sectors in the subsector's development and response to favorable market conditions. The following section looks at the prospects for and barriers to intraregional agricultural trade, focusing on southern Africa and the policy, institutional, and infrastructural constraints to further market integration and trade diversification. Last, a summary and conclusions are presented. Agriculture in the Region Agriculture dominates the economies of most of Sub-Saharan Africa's 46 countries (table 1). For the region as a whole, agriculture accounts for 70 percent of total employment and 40 percent of total merchandise exports, with these shares considerably higher in some countries. Although agriculture accounts for only 34 percent of African GDP, one- to two-thirds of manufacturing value added in most countries is based on agricultural raw materials, and many services are linked to agriculture.2 Although there has been some industrial development, manufacturing sectors in most African countries remain small, protected, and highly import-dependent. Several countries have remained or recently become heavily dependent on mineral or oil exports. In most countries, however, manufacturing employs less than 10 percent of the labor force. Nevertheless, recurring drought, the vagaries of world commodity markets, and institutional failures have resulted in low agricultural productivity, widespread food insecurity, and stagnant or declining agricultural exports in Africa (table 2). At independence, most African countries inherited economies whose trade focused on supplying a few primary commodities to their colonial mother countries, mainly food, agricultural raw materials, and minerals for manufacturers and consumers in Europe. Local food production, manufacturing, and market development had received little attention. Most of the new African countries adopted development strategies emphasizing import-substitution industrialization, in which the state played major financial and managerial roles. The primary commodities that had been the export mainstays were expected to remain important only in the short - 63 - TABLE 1. AGRICULTURAL INDICATORS IN SELECTED SUB-SAHARAN AFRICAN COUNTRIES SHARE OF MANUFAC. VALUE ADDED IN INDUSTRIES AGRIC. EXPORTS AS % OF BASED ON AGRIC. AS % OF LABOR FORCE AS % OF GDP RAW MATERIALS' MERCHAN. EXPORTS (1987) COUNTRY (1989) (1988) (1990) FEMALE MALE Nigeria 31 62b, 2 65 67 C6te d'Ivoire 46 0 56 52 71 Ghana 49 46 43 54 48 Cameroon 27 658 27 58 74 Senegal 22 63 25 73 88 Kenya 31 50 71 75 83 Sudan 54 758 66- 59 82 Ethiopia 42 67 67 73 82 Tanzania 66 64" 63 76 89 Zimbabwe 13 50 48 64 80 Note: These countries account for about 60 percent of the region's population and GDP. a. Including food, beverage, tobacco, cotton. and textile industries. b. Data are for 1970. c. A large decline in this share has since occurred in Nigeria due to the development of its petroleum products industry. d. Dats are for 1980. e. Data are for 1986. Source: World Bank 1991a, 1992a. term-as sources of government revenue for the industrialization drive. Thus, little attention was given to adding value to agricultural commodity exports-or to diversifying them. Most agricultural markets were heavily controlled by the state or replaced by parastatal agencies, and agricultural exports and domestic agriculture were directly or indirectly taxed to provide resources for the industrialization drive. Except for specific projects, agricultural support services were weakened financially and administratively. African economies remain largely tied to a narrow range of export crops TABLE 2. AGRICULTURAL PERFORMANCE INDICATORS, and staple grains that mirrors the colonial SUB-SAHARANAFRICA,1975-90(AVERAGEGROWTH RATE[%]) period. In the mid-1980s, for instance, INDICATOR 1975-80 1980-85 1986-90 Sub-Saharan countries as a group derived 75 percent of their agricultural export Value added in agricultures 1.9 0.4 1.5 earnings from only six commodi- Agricultural production 0.5 1.0 1.6 Food production 0.6 0.9 1.6 ties-coffee, cocoa, cotton, sugar, Nonfood production -2.5 0.5 1.7 tobacco, and tea-with coffee and cocoa Per capita food production -2.3 -2.1 -1.4 alone accounting for more than one-half Volume of agricultural exports -0.8 -2.9 -2.5 of total earnings. Twenty-five years Value of agricultural exportsb 9.6 -2.9 -2.5 earlier, these six commodities accounted for only 53 percent of Africa's a. Based on constant 1987 prices. agricultural exports (Akiyama and Larson b. Based on nominal prices and exchange rates. 1989). There have been few significant Source: World Bank 1992a. cases of export diversification in Africa into higher-value fish, meat, and horticultural products. In domestic food markets, government food security strategies in most countries have focused on increasing maize production and trade, with other cereals, legumes, and roots and tubers relatively neglected in official support services and marketing. The world market trends for the major export crops on which the region depends have not been favorable since the early 1970s. Among Africa's six leading commodity exports (mentioned above), the growth rate in industrial-country imports over 1970-88 was negative for two commodities (sugar and cotton) and less than 0.5 percent for two others (tea and tobacco) (World Bank 1991b). Much of the - 64 - expansion in import demand for these commodities came from developing countries, to which African countries direct only a small proportion of their exports.3 During the 1980s, the real prices of five of the six crops fell by an annual average of 4.0 to 9.7 percent (World Bank 1991b), while their yearly prices exhibited coefficients of variation between 15.9 and 52.2 percent (Hazell et al. 1989). Market prospects for this narrow range of crops are unlikely to improve: World Bank projections to 2005 estimate that world imports of tea, cocoa, coffee, cotton, and tobacco will grow substantially more slowly than they did over the slow growth 1970-88 period. World prices for cotton and tobacco are expected to decline even further, and tea, cocoa, coffee, and sugar prices to stagnate during the 1990s (World Bank 199 lb). Most of these crops have yet another weakness: inflexibility. With the exceptions of sugar and cotton (which can be transformed into several other products using relatively small-scale equipment), these are low-value bulk goods that offer few processing possibilities, provide limited scope for new product development, and have severely limited local or regional markets. The beverage crops have another element of inflexibility: they feature a long gestation period for production, thereby increasing market risks and financing problems. With the exception of coffee, each of the other commodities requires "lumpy" investments in large processing facilities to produce derivative products meeting international quality standards. Agricultural diversification offers several potential benefits for African farmers, companies, and economies. First, production and/or trade diversification can increase the levels and reduce the variability of incomes and employment. Finding additional uses and market outlets for existing commodities reduces the income and employment effects of downturns in traditional markets. Having a broader portfolio of crops and tradable commodities enables farmers and firms to cope with and flexibly respond to changes in weather, market conditions, policies, etc. Diversification of production into high-value, labor-intensive commodities (e.g., horticultural crops) can increase both farm income and employment. New crop development can generate additional income and employment opportunities, both in input supply industries and in product distribution. Second, agricultural diversification can contribute to an improved balance of payments position by substituting local production of foods and industrial raw materials for imports, by providing additional sources of foreign exchange, by raising the value added from existing export commodities, and by adjusting the portfolio of export products in the direction of those with more favorable price and consumption trends. Diversification of a country's agricultural trade portfolio can also serve as a hedge against yearly fluctuations in export revenues stemming from international price movements and/or variations in production of specific commodities.4 Third, production and trade diversification can contribute to enhanced food security at the household, regional, and/or national levels, especially where diversification results in increased incomes and employment opportunities for the rural poor.5 Agricultural diversification is not a panacea for Africa's economic ills. Improving productivity and marketing of traditional export commodities and staple food crops will be a critical element of the region's economic performance in the 1990s. Still, with improvements in macroeconomic policies, the liberalization of markets, and measures to improve the physical and social infrastructure serving agriculture, agricultural diversification in Africa may be more viable and important in the 1990s than it has been for several decades. Inposed Monocropping and Controlled Diversification: Sudan's Gezira Scheme Sudan is Africa's largest country, with a land area of about 1 million square miles. Its most significant development investment has been in irrigated agriculture, and its 1.75 million irrigated ha represent one-third of the total irrigated area in Sub-Saharan Africa (Harrison 1990). The core of Sudan's irrigated subsector has been large-scale gravity flow schemes organized as parastatals and involving - 65 - "partnerships" among the government, scheme management companies, and individual tenant farmers (these schemes cover about 63 percent of the total irrigated area, with much of the remainder under pump irrigation). The most important irrigation investment in Sudan has been the Gezira Scheme-the world's largest irrigation scheme under single management, at nearly 900,000 ha. (The area of the Gezira scheme is equivalent to one-half or more the arable area in 25 individual African countries.) The Gezira scheme, begun in the 1920s, was the backbone of the commercial economy for colonial Sudan; its cotton was the primary source of foreign exchange earnings and tax revenues. In independent Sudan, the Gezira scheme has continued to be significant, accounting for 10 percent of the cultivated area, employment, and GDP and at least half of Sudan's exports in most years (Barnett 1977: 6; Euroconsult 1982: 8). The development of the Gezira scheme, and particularly the persistent attempt by the government to impose particular patterns of crop specialization on tenant farmers, offers considerable insight into the range of technical, institutional, and economic issues related to agricultural diversification at the production level. The Gezira experience illustrates the perils of a government's imposing a cropping pattern upon farmers whose own interests are better served by alternative production patterns. Developing the World's Largest Farm The Gezira is a vast, flat, triangular plain lying between the White and Blue Niles. Its climate is arid, and rainfall is low and erratic. The area historically supported a low-density population with an agricultural system based on seminomadic livestock herding and subsistence production of sorghum. Seasonal migration into and out of the area was common. Farming depended upon raising water from the Nile with buckets or ox-driven water wheels and retaining it in ridged fields (Barnett 1977: 3; Robertson 1987: 83). At the turn of the century, the British colonial government began to examine the potential for converting the Gezira area into a vast irrigation scheme, based on the damming of the Blue Nile at Semnar, to produce high-quality cotton to supply its Lancashire spinning mills. A feasibility study, contour survey, and land cadastre were conducted, and a group of British investors (the Sudan Plantations Syndicate [SPPI) was authorized to develop the area. The dam at Semnar was completed in 1925. The government obtained some 300,000 ha of land by directly claiming control over unoccupied areas or taking compulsory 40-year leases from local farmers and herders. As a result of labor recruitment difficulties faced by the SPP and the problems faced by the British in sponsoring cotton production by independent peasant producers in Egypt, the system adopted to develop cotton in the Gezira combined centralized decisionmaking and control with small, individual producing units. The Gezira scheme, ostensibly the largest farm in the world, was subdivided into tens of thousands of units, each 30 feddans (12.8 ha). The units were leased back to the former landowners, herders in the area, or migrants from within and outside the Sudan at a minimal cost. By 1930, some 20,000 tenants were installed over an irrigated area of 300,000 feddans (126,000 ha). Within a few years, both of these figures would more than double (Robertson 1987: 86). The Gezira scheme was established as a "partnership" among the government, the private SPP, and the tenants. The government owned or leased the land, was responsible for building and maintaining the irrigation system, and importing production inputs. The SPP's role was to manage the scheme, provide inputs and advice to farmers, undertake mechanized farm operations, distribute water in minor canals, and gin and market the cotton. Tenants were responsible for cultivating the cotton using family or hired labor. The "partnership" pertained only to cotton, and 35 percent of cotton revenues went to the government, 25 percent to the SPP, and 40 percent to the tenants. Although the technical and commercial risks from cotton were shared among the parties, the tenants, either individually or collectively, were given no role in scheme decisionmaking, other than how to recruit and utilize farm labor. The government-in consultation with the SPP-set the tenant cropping - 66 - pattern as a rotation of 10 feddans of cotton and 5 feddans each of sorghum and lubia (a fodder crop). Cotton was considered the "official" or "government" crop and was the only crop for which farmers received credit and inputs; the sorghum and lubia crops were considered "tenant" crops. The right to be a tenant and to grow these other predetermined crops was contingent upon tenant production of cotton, using household and/or hired labor. During the latter half of the 1920s, the Gezira scheme was highly successful, producing good cotton yields and significant profits for the government, the SPP, and tenants. This stimulated a large migration into the area in search of employment and tenancy contracts. By 1929, however, major outbreaks of plant diseases resulted in sharply reduced yields, leading management to change the size of individual tenanacies, introduce a longer fallow period, and restrict planting of lubia (as it had contributed to both disease and weed development). With the onset of the Depression, world cotton prices declined sharply. Tenant returns from cotton became minimal-and, in some years, negative-and indebtedness became widespread and serious. Many tenants neglected their unprofitable cotton crop, concentrating, as in the prescheme days, on subsistence sorghum production and herding their animals off the scheme. During the Depression, the government forgave tenant debt and continued production services for cotton, which helped maintain the area under cotton within the scheme and prevented a major decline in cotton yields as occurred in many of East Africa's cash crop industries during the Depression and World War II. Between 1929 and 1946, the appeal of a Gezira tenancy lay not in the potential income from cotton, of which there was very little, but in the ensured access to water to support subsistence sorghum production and maintain livestock holdings. Ironically, a government-sponsored cash crop program was attractive to producers primarily because it provided better food security for its participants. Only in the immediate postwar years, when world cotton prices recovered and then exceeded their pre-Depression levels, would tenants obtain significant incomes from cotton. Although its profitability fluctuated, the Gezira scheme was very important to the colony, accounting for over 50 percent of its exports between 1929 and 1946. In 1950, the SPP was replaced by the Sudan Gezira Board (SGB), a parastatal agency linked to Sudan's Department of Agriculture. The SGB inherited the SPP's management functions and was also assigned to promote social development programs within the scheme. After independence in 1956, the SGB progressively lost its financial and decisionmaking autonomy to bodies of the Sudanese government. Buoyant international cotton prices in the 1950s and the optimism surrounding Sudanese independence in 1956 led to a decision to nearly double the size of the Gezira scheme. Begun in 1957 and completed in 1964, the Managhil extension added 800,000 feddans to the scheme and enabled the settlement of an additional 50,000 tenants, reaching a total of approximately 100,000 tenants. The expansion was accompanied by a strategy of intensifying production and diversifying the cropping mix. Intensification was designed to expand export output and import-substitution commodities. The strategy reduced the area under fallow by introducing a winter crop-wheat-into the rotation, and between 1960/61 and 1970/71 the cropping intensity increased from 47 to 62 percent (Plusquellec 1991: table 5). Crop diversification included groundnuts as a new export crop to reduce Sudan's high dependence on foreign exchange earnings from cotton. The process of crop intensification and diversification was both dictated and controlled by the government through the SGB. The new rotation pattern put quarter shares of each holding under cotton and under wheat; another quarter combined groundnuts and sorghum, and a quarter was left fallow. The Managhilr rotation devoted one-third each to cotton and wheat, and one-third to groundnuts/sorghum, with no fallow areas. The rationale for the standardized and fixed area allocations was that water distribution, land preparation, and aerial chemical spraying could be better managed by the SGB if crop plots were centrally determined. The system, at least officially, provided tenants with no flexibility in adjusting their cropping patterns to match changing economic conditions, their resources, and their personal preferences. The system also did not permit livestock production. - 67 - Of the four crops in the fixed rotation, farmers received credit only for cotton and wheat, and only for cotton were production inputs supplied directly by the SGB. Tenants were required to sell all cotton and wheat to the SGB for subsequent supply to a government cotton trading company or to government-authorized grain mills. Tenants were not permitted to use SGB-supplied fertilizers and agrochemicals on their groundnut or sorghum crops. For these crops, farmers relied upon private traders for inputs and marketing. The SGB provided practically no technical assistance to tenants for wheat, sorghum, or groundnut production. Although the government did try to organize cooperatives for groundnut financing, decortication, and marketing, these arrangements collapsed due to mismanagement and nonpayment of loans (Magar 1986: 154-56). Despite the weak official support system, tenant plantings and production of groundnut, wheat, and sorghum expanded greatly during the 1960s and early 1970s, with groundnut production reaching 22,000 metric tons and wheat production nearly 34,000 metric tons by the mid-1970s. Although a large part of the increased output was the result of an expansion in planted area, the yields per feddan for each of these crops did increase during this period.6 Cotton plantings, production, and yields also expanded during the 1960s, but each began to fall by the early to mid-1970s. Although tenants successfully lobbied to obtain a larger share of the revenues from cotton sales, cotton continued to be less profitable than each of the other crops because of the growing costs of cotton production and the fact that the prices tenants received for their cotton were generally less than 20 percent of the f.o.b. value (Faki 1982: 49). Table 3 compares the net returns to farmers for different crops during the 1970s. As rational economic actors, tenants increasingly found ways to deviate from the TABLE 3. GEZIRA NET RETURNS scheme's controlled resource allocation system to TO FARMERS (LSD/FEDDAN) pursue their own objectives (Magar 1986: 153- GROUND- 54; Robertson 1987: 98-107). When possible, SEASON COITON WHEAT NUT SORGHUM tenants diverted resources from "the government's cotton" to crops that were more 1970/71 11 4 -4 -1 1971/72 4 7 -1 -3 remunerative and less labor-demanding. The 1972/73 14 8 53 18 expansion and growing complexity of the scheme 1973/74 30 23 55 18 prevented the SGB from effectively monitoring 1974/75 14 7 70 8 1975/76 -5 10 28 19 tenant activities, thus enabling tenants to divert 1976/77 9 16 30 20 water and fertilizer from cotton to other crops. 1977/78 28 20 63 17 The majority of tenants also developed 1978/79 29 -1 26 14 1979/80 -4 14 52 33 subtenancy or sharecropping arrangements with 1980/81 -4 -3 92 27 hired laborers to work their groundnut and Average 11 10 42 20 sorghum crops and to manage small vegetable plots in and around villages. These sharecropping LSd = Sudanese pounds. arrangements eroded SGB control over Note: Highest figures for the season are in bold. production and enabled tenants to cope with both Source: Sudan Gezira Board, as reported in Magar 1986. labor and cash shortages. With effective demand exceeding supply, a vibrant unofficial market in wheat developed, with tenants bypassing the SGB-and repayment of their wheat credits. From the tenants' perspective, the most appropriate strategy in allocating resources and effort increasingly became "ABC"-anything but cotton. Not only were the returns higher for other crops, but payments were received on a timely basis, rather than in multiple installments six months or more after cotton delivery. Payment delays for cotton became common during the 1970s as the SGB's financial position deteriorated. Unlike cotton, there were no water charges or taxes on the other crops. Even when water charges were introduced in the 1980s, a large proportion of tenants succeeded in evading such payments for crops other than cotton. Also, unlike for cotton, tenants had several marketing options for - 68 - the other crops and could-subject to short-term cash requirements-store the harvested crop in anticipation of higher prices. From the government's perspective, cotton has remained the lead crop. The government has continued to invest in Gezira with the aim of maintaining or expanding Sudan's cotton exports. Despite a large investment in the 1980s to rehabilitate the Gezira irrigation infrastructure and improve productivity, cotton plantings, production, and yields have all declined. With a seriously overvalued exchange rate and with continued efforts to stimulate tenant interest in cotton, for much of the 1980s the price paid to tenants for cotton was higher than the Sudanese pound equivalent of export revenues (World Bank 1990: 18). Gezira's Unrealized Potential A half century of investments in dams and canals should have made the Gezira an extraordinarily valuable asset in Sudan's economic development. The area's potential is evident in the dramatic increase in wheat and groundnut production during the 1960s and in a small vegetable production boom during the latter part of the 1980s. It is, however, a potential that has seldom been realized; indeed, there have been years in which Gezira cotton, wheat, and other crops have had negative returns both for tenants and the country as a whole. For many years, there has been an inherent conflict between government and farmer objectives and strategies. The extraction of a cotton crop, although favorable for the country's balance of payments, has yielded profits for farmers in relatively few years. For food security and income generation, farmers have preferred sorghum, groundnut, wheat, vegetables, and livestock raising. The lack of effective support services for these other crops, however, has kept farmer yields in the Gezira to only 25 to 33 percent of their potential, and weak services have also limited the potential for livestock development in the area (Euroconsult 1982; Blench 1987). Even with relatively low yields, these other crops have generally had a comparative advantage-demonstrated by favorable domestic resource cost ratios- especially compared with cotton (D'Silva 1986). Although bringing Sudanese agriculture up to its potential will be a slow process that critically depends on factors outside the agricultural sector-from macroeconomic policy to political stability-it is unlikely to advance at all until government policies are much more in harmony with the needs of farmers. And, ironically, the government would probably be better served if it deemphasized cotton and devoted more of its research, extension, input supply, and other support efforts to raising the productivity of "tenant" crops-particularly sorghum, which is a major medium for paying agricultural laborers, and groundnuts and vegetables, which have potential domestic and export markets and considerable scope for value-adding processing. Higher yields for these crops would mean that farmers could raise their incomes while devoting less land and labor resources to them. If the appropriate incentives were put in place, this would free up both land and labor for voluntary and dedicated-as opposed to forced-cotton cultivation on the part of the tenants. Agricultural Diversification through Contract Farming Contract farming has proven to be an effective institutional mechanism for increasing agricultural productivity and stimulating agroindustrial expansion and diversification in Africa (Watts et al. 1988). It has served as a mode for reducing the technical and market risks faced by farmers, processors, and traders and for coordinating their activities in operating environments generally characterized by poorly functioning input and commodity markets and limited agricultural support services. Farmer risk is reduced because the intermediary-the contractor-links the growers and markets. Having an ensured market outlet as well as a dedicated source of inputs and technical advice, farmers are provided with both - 69 - an incentive and an improved capacity to adopt a new crop or to diversify livestock holdings. Contract farming also reduces costs and risks faced by processors and traders, enabling them to expand and diversify their activities. Using the land, labor, and skills of existing farmers enables the firms to forego the costs of buying and developing their own land and recruiting and managing an agricultural labor force. Rather than centralizing raw-material production on a single farm, production risks related to weather, disease, and pests can be spread among many contracted farmers over a wider geographical area. Having an ensured supply at a contracted price reduces uncertainty and allows more confident forward planning. Contract farming is not, however, risk-free; success depends on correctly judging both technological and market factors and on carefully balancing production with processing or market requirements. Contract farming is widespread in Africa (Watts et al. 1988; Jaffee 1987). Its incidence and influence have been greatly underestimated in discussions of African agriculture and, in many cases, have suggested a dynamism at the level of production and in the interface between agriculture and agroindustry that contradict the overwhelming sense of stagnation and decline that dominates the literature on Africa's economic "crisis. " A mid-1980s review of the literature on African agriculture and agribusiness identified 88 contract farming schemes in 21 countries; this probably underestimates their incidence as there is a paucity of information on small-scale, privately organized contracting schemes. Both in Kenya and Cameroon, major contract farming schemes have been developed for a half-dozen or more commodities that occupy a central role in these countries' agroindustrial sectors. In at least five African countries (Kenya, Cameroon, Burkina Faso, Senegal, and Mali), contract farming schemes involve the participation of at least 75,000 smallholder farmers. In Kenya alone, more than 250,000 smallholders-over 15 percent of all smallholders-are contracted to grow one or more export or industrial crops.7 Table 4 shows the incidence and selected characteristics of documented contract farming schemes in Africa as of the mid-1980s. The table indicates that the largest number of individual schemes has been for horticultural commodities, either for fresh produce marketing or for processing. Multiple schemes in several countries have also been developed for major export and industrial crops, such as sugar, palm oil, tea, tobacco, and cotton. These commodities share certain technical and economic characteristics that ordinarily contribute to a high level of risk for producers and processor/traders. These commodities are generally highly (a) perishable in their raw form, (b) valuable per unit weight or volume, and (c) heterogenous in quality. They also require specialized skills and inputs to produce; have extended production cycles or gestation periods; and feature large economies of scale in processing, though not in farm-level production. Under these circumstances, farmers are likely to require specialized services and value market assurances, while processors/traders will seek commodity/raw material supply assurances and can potentially appropriate significant benefits from providing production support services to farmers. In contrast, there have been relatively few contract farming schemes in Africa for livestock and dairy products and for staple food grains. This is surprising, given the widespread contracting of such products in North America, Western Europe, Asia, and Latin America. With livestock products, the reason may stem from the still limited effective demand for high-quality meats and the limited availability of refrigeration and other technology to maintain the quality of high-priced meats.' For dairy products, the reason may derive from the widespread application of price controls in the subsector and the extensive reliance upon highly subsidized powdered milk imports for urban supply rather than the development of internal supply systems. Food grains in Africa lack most of the technical and market characteristics that would lead to a demand for and supply of market and production contracts. For example, food grains have low value per unit weight or volume, have widely controlled markets and prices, do not require specialized inputs, have relatively short production cycles, can be processed in several alternative forms, and have readily identifiable quality variations. - 70 - TABLE 4. DOCUMENTED CONTRACT FARMING SCHEMES IN AFRICA, MID-1980s AVG. OWNERSHIP PAITERNS NO. OF NO. OF NO. OF MARKET (% OF ALL SCHEMES) COMMODITY SCHEMES COUNTRIES FARMERS ORIENTATION PRIV. PUB. JVa Fresh fruit/veg. 25 + 6 < 100 Mostly export 90 5 5 Sugar 12 6 5,000 Domestic/export 33 33 33 Processed fruit/veg. 9 7 3,0QQb Export/domestic 33 22 45 Palm oil 8 5 2,000 Domestic/export 0 75 25 Cotton 7 7 64,000- Domestic/export 0 57 43 Tea 7 5 26,000 Export/domestic 28 72 0 Tobacco 6 6 5,500 Domestic 17 33 50 Dairy 3 3 4,000 Domestic 0 100 0 Rubber 2 2 1,250 Export 0 50 50 Poultry 2 2 n/a Domestic 50 0 50 Rice 2 2 2,000 Domestic 0 0 100 Spices 2 2 n/a Export 100 0 0 Oilseed 2 1 n/a Domestic 50 0 50 Gari (cassava) 1 1 141 Domestic 100 0 0 Total 88 21 41 33 26 (Without fresh fruit/veg.) 24 43 33 n/a Not available. a. Joint venture between the public and private sectors. b. Based on infornation for seven schemes only. c. Based on information for four schemes only. Source: Watts et al. 1988; Jaffee 1987. Contract farming has been used as an institutional framework for developing new export crops, for promoting import-substitution of foods and agricultural raw materials, and for expanding and adjusting existing subsectors to make them more competitive. There have been several striking success stories, such as the schemes for tea in Kenya, for palm oil in Cote d'Ivoire, for cotton in several Francophone countries, and for sugar and tobacco is several countries in different subregions. Many schemes, however, have also failed or been nonsustainable, undermined by poor management, commodity price declines, and/or competition from alternative local market outlets. Two examples will be used to illustrate successful crop diversification and agroindustrial development within the framework of contract farming. The first example relates to cotton industry development, with its stark contrast between (a) the rapidly developing industries in Francophone West Africa under state-supported contract farming systems managed by the Compagnie Francaise pour le Developpement des Fibres Textiles (CFDT), and (b) the stagnant or declining industries of most other African countries, which have relied on nonintegrated systems of support and marketing services. Lele et al. (1989) have highlighted this difference in performance, attributing it primarily to institutional factors. Although the official producer prices for cotton have generally been lower in the Francophone West African countries than in other African countries, the former countries have experienced better long- term trends in production yields that have amply compensated for the lower prices. Lele et al. (1989) attributed the better Francophone country performance to the contract-based system, which has facilitated more effective input supply, technical assistance, and payment systems and better coordination among research, extension, and production operations. Table 5 updates and extends the comparison to cover several additional countries and to also include trends in export volumes. The data indicate generally better yield performance for the countries - 71 - whose cotton industries are based on contract farming. The greatest difference is in long-term export performance; the industries based on contract farming experienced average increases of 7 to 15 percent in export volumes over 1975-89, while the other countries (except Zimbabwe) experienced export declines or ceased to export cotton altogether. The second illustration of favorable agricultural performance TABLES. COTTON INDUSTRY PERFORMANCE under contract farming comes from IN SUB-SAHARAN AFRICA, 1975-89 Kenya. Table 6 compares long-term SYSTEM/ AVG. GROWTH RATE (% yield trends in Kenya among COUNTRY IN YIELDS IN EXPORT VOLUME subsectors based on contract farming Contractfanning and those based either on marketing Burkina Faso 3.7 7.3 board/general support services or Cameroon 4.5 14.8 vertically integrated agribusiness C6te d'Ivoire 5.3 15.7 operations. Over a 15- to 20-year Mali 0.8 8.9 Senegal -0.2 8.0 period, smallholder production of Togo 8.5 14.0 some commodities under contract has achieved greater yield improvements Other systems Central African Republic 5.2 -6.8 than smallholder production under Malawi 3.1 marketing board/general support Nigeria -3.9 service systems. In addition, with the Tanzania -2.2 -0.9 exception of tea, contracted Uganda n/a -5.4 Zimbabwe -1.7 5.1 smallholders have achieved greater productivity gains than have certain vertically integrated agribusiness -- No longer exporting. systems. Even with tea, the quality of Source: Varangis et al. 1990; World Bank 1992a. smallholder production has exceeded that of estate production. The expansion and improved productivity in smallholder tea, tobacco, sugar, and french bean production has been directly linked to the dedicated support and supervisory system for such crops (Lamb and Muller 1982; Allen 1983; Buch-Hansen and Kieler 1983; Jaffee 1987), while weak support systems and payment delays have been major factors in the weaker performance of smallholder coffee, cotton, maize, and cashew nut production (Schluter 1984; Ommeh 1984; Lele et al. 1989). There have been enormous differences in the size, ownership, and management of contract farming schemes in Africa (Watts et al. 1988). Although the vast majority of schemes have involved less than 1,000 contract farmers, several schemes, including the KTDA tea scheme and several of the West African cotton schemes, have included 50,000 or more farmers. Most contractor enterprises have established criteria for restricting entry into their schemes. Such criteria normally relate to location, landholdings, and other assets. Although there have been exceptions, such provisions have restricted the poorest farmers and the least favorable areas in most countries from participating in contract farming schemes, though landless persons and poorer farmers have migrated in significant numbers to work in many contracting schemes. Successful contracting schemes have enabled better resource-endowed commercial smallholders to either diversify their income sources and production operations or to move into more specialized production of one or more contracted crops. Ownership and management patterns in contract farming schemes in Africa differ from those commonly found in Asia and Latin America. While schemes privately initiated and run dominate in these regions, a majority of the documented schemes in Africa have featured direct investment by government. Direct government participation is found in nearly all the larger schemes, i.e., involving 5,000 or more farmers. Government-initiated contract farming schemes in Africa date to the colonial period, but it was only in the 1960s and 1970s when a large number of such schemes were developed as part of export - 72 - development or import-substitution programs or in conjunction with land/area TABLE 6. COMPARATIVE LONG-TERM YIELD settlement programs. PATTERNS IN KENYAN AGRICULTURE A large proportion of AVG. government-initiated schemes have ANNUAL involved financial, managerial, and SYSTEM CATEGORY/ RATE OF technical assistance from international COMMODITY TIME PERIOD CHANGE (%) financial institutions, especially the Contractfanning' CommonwealthDevelopmentCorporation Sugar" 1966-85 6.7 and the World Bank. A large proportion Tea' 1971-86 2.7 Tobaccob 1966-85 13.8 of these schemes, including most of those French beans' 1982-86 37.9 involving sugar or tree crops and the construction of large processing facilities, General suppont have also featured the development of marketing board' Cashew nutsb 1961-84 -1.6 nucleus estates to guarantee minimum Coffee 1971-86 2.1 throughput into factories, undertake Cotton, 1965-86 -1.3 technical research, and develop high- Maize, 1966-87 1.3 quality planting materials. Another Vertically integrated common feature of government-initiated agribusiness contract farming schemes has been the Pinespple' 1973-86 2.5 participation of major multinational Estate teal 1971-82 5.6 corporations in a technical and managerial capacity and, frequently, with a minority sharholdng.Such hybrid institutions, a. Smnallholders only. shareholding. Schyrdistuon, b. Data from FAO production series. combining government ownership, donor c. Data from KTDA annual reports. financing, and private management, are d. Data from Njoro Canners Ltd., production farmer records. common in Africa's major sugar, cotton, e. Data from Coffee Board of Kenya, annual reports. common ~~~~~~~~~~~~~f. Data from Lele et al. 1989: table 3. oil palm, and tobacco industries. Their g. Data from Kenya Canners Ltd. presence and indeed their significance h. Data from Kenya Tea Board, as reported in Lamb and Muller 1982. belie the stark analytical dichotomy of the public and private sectors found in most of the literature on African agriculture. The organization of contract farming schemes by local African companies has generally been limited to the fresh fruit and vegetable trade, although there are cases of local companies teaming up with foreign partners to develop outgrower schemes and processing operations for food, beverage, and industrial products (Jaffee 1987; Horton 1987). For the most part, local companies have lacked the technical capacity and financial resources to organize larger contract farming schemes, which may necessitate extended credit to producers. Also, private local companies are possibly less successful than either state- or foreign-owned companies in lobbying government to gain protection against competitive imports or competing firms and to acquire preferential access to financing and physical infrastructure (e.g., roads, utilities). As noted earlier, contract farming may contribute to specialization rather than diversification at the farm level. This will depend upon resources available, returns, local food markets, and contractual stipulations. Although there have been instances where contract farming has resulted in a virtual monocropping pattern for farmers and for entire areas, in most cases the contracted crop occupies only a small proportion of a farmer's total landholding. In addition, there is evidence of spillover effects from contracted production onto other crops as a result of the availability of inputs and the acquired technical skills. For example, technological spillovers have been observed in selected cases of contract farming of tobacco, tea, cotton, and vegetables (Buch-Hansen and Kieler 1983; Jaffee 1987; Lele et al. 1989). - 73 - Nontraditional Export Development: Kenyan Horticulture Over the past two decades, horticultural products, such as fresh and processed fruit and vegetables and cut flowers, have become one of the most dynamic components of international trade, with a 10 percent average annual increase in value (GATT 1989; Islam 1990). With this growti in horticultural trade and stagnation or decline in world trade for many other agricultural commodities during the 1980s, the value of fruit and vegetable products imported by OECD and developing countries now exceeds those for any other category of agricultural products, including cereals, oilseeds, fish products, meat products, and tropical beverages (GATT: table IV.2).9 Many African countries have a potential comparative advantage in the production of fresh and processed horticultural products as a result of their agroecological conditions, location, and relatively low labor costs. Very few African countries and entrepreneurs, however, have effectively translated these resource advantages into competitive and profitable horticultural trade with Europe, the Middle East, or other regions. The reason is that African countries have remained poorly endowed with the physical infrastructure and the technical, managerial, and marketing skills necessary for competitive horticultural export development. For the vast majority of Sub-Saharan African countries, nascent horticultural export industries have floundered due to continually weak production support systems, international transport and other infrastructural bottlenecks, weak marketing institutions, and poor trade linkages with overseas distribution systems. Kenya is one of a very few Sub-Saharan African countries that have emerged as major participants in international horticultural markets (the others are Cote d'Ivoire and South Africa). Kenya's horticultural exports are an anomaly in another context, that of the wider patterns of agricultural and industrial development in Kenya since its independence in 1963. While achieving import-substitution for many consumer goods and food products, Kenya has been largely unsuccessful in diversifying its commodity export base, either in industry or agriculture. With many of its traditional agricultural export industries experiencing stagnation or decline (e.g., sisal, pyrethrum, meat products, and cotton), Kenya has become increasingly dependent on highly variable foreign exchange earnings from coffee and tea (table 7). For many of its traditional agricultural exports and for a wide range of "minor" commodities, the development or maintenance of efficient commodity systems has been undermined by official marketing monopolies of parastatal enterprises and the failure of government to provide these organizations with either the mandate or the necessary technical and financial resources to support production and trade of these commodities. The major exception to this lack of export diversification has been the development of a large and highly diverse trade in horticultural products. Although minimally important during the colonial period, Kenya's horticultural exports have expanded enormously since independence. By 1988, aggregate horticultural exports reached over 110,000 metric tons, valued at $107.5 million. Horticultural exports have expanded virtually every year since the early 1970s, with an average annual increase in trade volume of 11.6 percent. If current trends continue, horticultural products will surpass both coffee and tea during the 1990s to become Kenya's leading commodity export subsector. The expansion of Kenya's horticultural exports has been based upon a wide range of individual commodities, in recent years numbering over 75. This trade has consisted of (1) canned fruit and vegetable products (including canned pineapple and french beans); (2) fruit and vegetable juices (including pineapple, passion fruit, orange, and tomato); (3) fresh temperate, subtropical, and tropical fruits and vegetables (including french beans, chilies, okra, mango, avocado, strawberry, pineapple, and passion fruit); and (4) cut flowers (including carnations, roses, alstromeria, chrysanthemums, statise, and orchids). (Kenya has also developed a small export trade in vegetable and flower seeds.) The development of Kenya's horticultural trade has been uneven across commodities, with the different patterns providing certain insights into the constraints to and opportunities for agricultural trade diversification in Africa. In the pineapple canning and cut flower industries, production and trade have - 74 - expanded significantly since the early 1970s, placing TABLE 7. COMMODITY SHARES IN KENYA'S AGRICULTURAL Kenya among the world's EXPORTS (% OF AGRICULTURAL EXPORT EARNINGS) five leading exporters. COMMODITY 1914 1935 1955 1964-66 1974-76 1987-88 Recently, exports of canned pineapple, pineapple juice, Coffee 6 37 38 39 39 38 and cut flowers have Tea 0 9 12 17 18 30 Horticultural accounted for over 75 percent products' 0 0 3 3 8 16 of Kenya's horticultural Hides and skins 48 7 5 5 4 4 export earnings. As discussed Maize 3 7 7 0 3 4 Sisal 13 17 8 I11 7 2 below, Kenya's trade in these Pyrethrum 0 0 5 6 4 2 products is primarily a result Meat, dairy, wool 5 6 4 6 5 0 of large-scale foreign Oilseeds 18 2 0 1 1 0 investment. Kenya has also Other producta' 7 14 18 12 11 4 achieved significant expansions in trade and a. Including fresh and processed fruits and vegetables and cut flowers. market shares for several b. Including cashew nuts, potatoes, wheat, tobacco, sugar, legumes, cotton, fish products, and wattle bark/extract. other products, including Source: Calculated from data in Department of Agriculture annual reports and Kenyan fresh and processed french statistical abstracts. beans, strawberries, and a wide range of "Asian" vegetables (e.g., chilies, okra, karela) consumed in large quantities by the South Asian and other immigrant communities in the U.K. and elsewhere in Europe. In contrast, for most other fresh and processed horticultural products, Kenya has either experienced reduced trade levels over the past decade or not taken advantage of major expansions in West European and Middle Eastern import demand. Although for a few temperate commodities (e.g., courgettes and sweet peppers) Kenya's earlier market position was eroded by expanded lower-price or higher-quality supplies within Europe, for most other products, supply side constraints have been more important. They have included Kenya's limited horticultural research capacity, its lack of high-quality planting materials of preferred varieties, poor coordination between production and processing operations, risk aversion on the part of most producers and traders, and a failure to develop infrastructure to export fresh produce by sea. In examining patterns of diversified trade development in Kenyan horticulture, separate treatment will be given to the fruit and vegetable processing industry, the fresh produce trade, and the cut flower industry. Fruit and Vegetable Processing Industry The first significant investments in fruit and vegetable processing facilities in Kenya occurred during and immediately following World War II. These investments were geared toward producing import-substituting jams, juices, and canned products behind steep protective import walls. Still, the colonial government regarded this industry as having export potential and thus undertook a program of horticultural research to identify crop varieties suitable for processing and subsequent export. The government also encouraged several British firms to invest in export-oriented operations, particularly for canning pineapples and vegetables. Despite government subsidies and market regulations, these ventures were not commercially successful because the firms (a) were not effective in generating reliable supplies of high-quality raw materials from settler or smallholder farmers; (b) incurred high costs for intermediate inputs (e.g., tin cans, sugar) and international transport; and (c) encountered strong competition from other international suppliers with better quality and well-established brand names. Government subsidies and other support measures simply kept the export-oriented processors from going bankrupt. - 75 - The government of newly independent Kenya followed a similar pattern after 1963, seeking foreign investors in processing ventures and underwriting a significant proportion of the risks and investments for these ventures. One joint venture project in vegetable dehydration was a commercial failure because the foreign partner had little technical expertise or trading experience in this field. The government bore most of the financial loss to maintain factory operations and thus provide for a continued sales outlet for recently resettled smallholder farmers. This negative experience, together with continued technical and marketing problems in the remainder of the processing industry, led the government to condition its further lending to the private sector or its own direct investments on the participation of major multinational corporations in future ventures. In the pineapple canning industry, this multinational corporation participation initially took the form of a management and marketing contract with the California Packing Company (Del Monte), but by the early 1970s the government had entered into joint ventures with major multinationals for pineapple canning, vegetable dehydration, and passion fruit juice production. In each case, the Kenyan government took a minority equity share or provided loan financing, provided low-cost access to land and services, and helped organize and finance raw material production schemes involving smallholder farmers. The foreign partner acquired exclusive or predominant control over factory management and marketing activities. With capital investments and equity shares being phased or altered over several years, the foreign partners generally bore little initial financial risk. From a production and trade perspective, each of these joint ventures experienced a significant expansion during the 1970s, with Kenya becoming one of the leading world suppliers of canned pineapples, passion fruit juice, and dehydrated vegetables."0 In both the vegetable dehydration and passion fruit juice industries, several thousand smallholder farmers became raw material suppliers. Most undertook production of new crops with a guaranteed market outlet and a package of production inputs and technical advice provided by the processing company. By the early 1980s, however, both of these industries neared collapse due to increased competition for raw materials from the fresh produce trade and/or the unwillingness of the Kenyan government to continue to underwrite company losses for the sake of smallholder outgrowers. In the most important case-pineapple canning-the growth and subsequent stability of Kenyan production and trade were based on a shift by Del Monte from an outgrower-based raw material procurement system to one based on the company's own large plantation. With the Kenyan government providing low-cost access to 20,000 acres of land, Del Monte developed an extensively vertically integrated system, combining input supply, farm production, processing, transport, and marketing. By altering its specific mix of products, changing its market destinations, and heavily promoting its products, Del Monte has been able to retain the viability and scale of the Kenyan industry despite an overall stagnation in the European canned pineapple market. During the 1980s, Kenya also emerged as a leading supplier of premium-quality canned green beans to the French market. This trade has been the result of a joint venture between a private Kenyan firm and a French company, which assumed the venture's management and marketing responsibilities. In sharp contrast with the Del Monte nucleus estate system, this processing operation has relied on raw material supplies from some 20,000 resource-poor smallholders in western Kenya with whom the company has developed an intensive contract farming scheme. The scheme has featured the supply of a package of production inputs and continuous technical support and crop policing by extension agents hired and trained by the company. Although the overall market for canned green beans in France experienced little growth during the 1980s, the Kenyan operation was provided an ensured sales outlet by the French partner that positioned the Kenyan product in the high-quality, high-price segment of the market. - 76 - Fresh Produce Trade The patterns of production and export trade development for fresh fruit and vegetables sharply contrast with those discussed above for the processing industry. Although a small export trade in fresh produce from the Kenyan coast dates at least as far back as the mid-19th century, a steady long-distance trade in these commodities was built up only in the late 1950s, when air- and sea-freighted supplies of a few tropical and temperate commodities were sent to London and Aden. This trade, which essentially pioneered the European import market in off-season fresh produce, was initiated by a few European and Asian wholesale/retail firms or cooperatives with prior experience serving the small local market catering to the European settler community and regional markets for less perishable commodities (e.g., potatoes and onions). The export trade involved considerable experimentation by European settler farmers and by trading companies in new crops and in new packaging and storage methods. The government provided minimal support for this trade, regarding it as too technically complex and with little growth potential for Kenya. A similar government position was held for much of the first decade after independence. Due to limited supplies of high-quality produce and severe shortages of air-freight facilities, this trade remained limited through the 1960s. During the 1970s, Kenya's fresh produce trade experienced considerable expansion and commodity diversification, stimulated by strong West European import demand for tropical commodities and "off-season" produce. Kenya's supply capability was enhanced by large farmer investments in irrigated vegetable production, the expanded production of pineapples on Del Monte's plantation, informal production support services provided by several leading exporters, and the introduction of wide-bodied carriers with greater air-freight capacities on Kenyan-European routes. Although many commodities were exported, the core of the trade was fresh pineapples and selected temperate and "Asian" vegetables. Basing their trade on these commodities, individual firms and farmers could experiment with additional commodities. Still, the basis of trade remained narrow. While most pineapples were obtained from the Del Monte estate and a few neighboring large farms, supplies of temperate and "Asian" vegetables were obtained from a few dozen medium- to large-scale European-, Asian-, or African-owned farms. Despite indications of favorable market conditions, few farmers or trading companies were prepared to invest in most fruit crops because of their extended gestation periods. The lack of high-quality planting materials, the absence of infrastructure for sea shipments of fresh produce, and uncertainty about Kenya's future political stability inhibited such investments. Many firms and entrepreneurs had entered the export trade; failure rates, however, were high, and only a few medium-size and product-diversified Asian-owned companies accounted for the bulk of trade. Each of these firms developed highly personalized, long-term trading links with one or several European market importers, enabling them to reduce the risks and transaction costs associated with fresh produce trade. During the 1980s, the fresh produce trade continued its rapid expansion and developed a much broader production and trading base. Increased competition among exporters, together with increased foreign demand, resulted in attempts by exporters to encourage and provide basic technical support to large numbers of new producers, particularly smallholder farmers. Various informal and formal institutional arrangements were developed to coordinate this smallholder production with marketing requirements. The expansion of production and trade has had a beneficial effect on incomes and employment opportunities in several regions, both in the center of the country and in more remote locations. Concurrently, several of the largest exporters have integrated backward into horticultural production on their own or leased farms. Production of individual crops has tended to cluster in a few locations, as successful pioneers have been copied by neighboring farmers. Most of this investment and most of the expansion in trade occurred in temperate and subtropical fruit and vegetables with short production cycles. Only a few large and diversified local or foreign companies made considerable investments in fruit orchards, with Kenya subsequently not in a position - 77 - during the 1980s to take advantage of the greatly expanded European and Middle Eastern import demand for commodities such as mango and avocado. Attempts to diversify production and trade into additional high-value commodities (e.g., strawberries, mangetout peas, cherry tomatoes) took several years of trial and error by private firms and farmers in the absence of a well-functioning horticultural research system. Nevertheless, with the pull of strong market demand and with the government at least not interfering in these ventures, many of these experiments have succeeded and enabled farmers and firms to diversify their mix of crops and traded products. Cut Flowers Between World War II and the late 1960s, commercial cut flower production and trade developed on a small scale in Kenya, with most production directed to the local market. A few small European- owned nurseries accounted for the bulk of production and sales in international markets. By the 1960s, a growing number of African smallholders located near Nairobi began to experiment with flower production for local sale. No official research and extension in floriculture was undertaken. In 1969, Kenya's nascent cut flower trade was completely transformed by a large-scale investment by the Danish firm Dansk Chrysanthemum Kultur (DCK), which was then the world's largest producer of chrysanthemum cuttings. As with foreign investors in fruit and vegetable processing, DCK was offered highly favorable investment terms by the Kenyan government, including a low-cost lease on 15,000 acres, exclusive growing and trading rights for several flowers for eight years, and a 25-year guarantee of exemption from any changes in the law regarding foreign-investor taxation and transfer of profits. The Danish government also supported the venture, providing a cash grant equivalent to one-third of the investment cost. Within a few years, DCK had over 100 ha under several types of flowers at its original estate and had acquired two additional farms. By the mid-1970s, cut flower exports reached nearly 2,500 metric tons. In contrast with the canned pineapple industry, the cut flower industry did not maintain its near monolithic structure. In 1976, the Danish investor suddenly pulled out of Kenya and all its holdings were spun off, first to firms headed by high-ranking Kenyan government officials, and later to Brooke Bond (a tea company) and to a government parastatal. Within a few years, several additional firms had developed, most of which involved either foreign investments or management arrangements or individual expatriates from the original DCK project branching off to develop their own production/trading operation. By the mid-1980s, there were over a dozen flower producing/trading companies, although only three of them controlled 90 percent of exports. Although flower production and exports have been dominated by a few European-owned and managed firms, several African-owned firms have also developed profitable operations (drawing upon technical, managerial, and/or marketing advice from experienced flower firms), and several thousand smallholder farmers have been incorporated into the trade through informal production support and crop procurement systems of the smaller firms. Much of the technology for flower production, including seeds, drip irrigation equipment, and specialized fertilizers/agrochemicals, continues to be imported. As with specialized fruit and vegetable production, cut flower production has tended to be clustered in a few locations (especially around Lake Naivasha and on the outskirts of Nairobi), facilitating learning through example as well as informal exchanges of technical information. Lessons The Kenyan experience indicates that there are potentially high payoffs in demand-driven, export- oriented horticultural development. Expanded horticultural exports have been important for Kenya's balance of payments position in light of significant recent declines in world coffee and tea prices. Export- - 78 - oriented horticulture has provided employment and at least supplemental incomes for tens of thousands of people. Both smallholders and larger farmers and companies now widely participate in the subsector, although only a few companies account for a large proportion of total trade. The Kenyan experience also indicates that production and trade diversification in this direction is risky given the high perishability and heterogeneity of the products, the heavy logistical and service demands, and the highly competitive and changing markets. Product and company failure rates have been high in Kenya. In the fresh produce trade, those firms that have overcome these risks and developed sustained and profitable trade have had extensive prior experience in horticultural marketing and/or production, have dealt with a diversified mix of products and suppliers, and have developed long-term international trading links. In the processing industry, government infant-industry support was critical during the 1950s and 1960s, while linkages to multinational trading companies eventually proved vital for sustained trade development. In the cut flower industry, the initial transfer of technology and build-up of local skills were orchestrated by one multinational company. Within a decade, this investment spawned a much broader industry comprising small, medium, and large companies, both expatriate- and Kenyan- run. The product mix of the industry has evolved steadily with changing market opportunities. Kenya's horticultural sector has developed despite limited and belated investments in official research, advisory services, and input supply. Whereas the governmnent has largely allowed the private sector to respond to changes in relative prices and overall market conditions, it has provided farmers and firms with relatively weak support in making such responses. In the absence of effective support services, the private sector has had to fill the vacuum. It has often done so, although the process has sometimes been very costly, limited entry of smallholder farmers into the trade for many years, and contributed to a bias in the direction of crop diversification away from crops with extended gestation periods. Part of the vacuum has been filled by foreign technical support and imported technologies, without which Kenya could not have developed its cut flower industry nor its trade in several fresh and processed fruit and vegetable products. Still, support services for most small-scale horticultural producers and traders remain weak, constraining their productivity and limiting their capacity to respond competitively to new market opportunities. Diversification of Regional Trade: The Case of SADCC Over the past two decades, official trade among Sub-Saharan African (SSA) countries has stagnated. Recorded SSA imports from other SSA countries increased in nominal terms from $2.15 billion in 1981 to $2.71 billion in 1990, representing a decline in real terms. In 1990, only 7.4 percent of SSA imports originated from other SSA countries (World Bank 1992a). The vast majority of SSA countries conduct less than 10 percent of their external trade within the region (World Bank data). Regional trade within Africa has been inhibited by several factors, including: 1. natural barriers (e.g., distance and low population densities); 2. infrastructural bottlenecks; 3. market imperfections (especially in finance and information); 4. macroeconomic instability and policy failures; 5. administrative barriers (including trade controls, licensing, currency controls) (World Bank data). Pending significant policy and institutional changes, up to an additional estimated $5 billion of Africa's current imports from the rest of the world could be supplied by African countries already exporting similar products outside the region. Although commodity aid programs and special bilateral arrangements - 79 - could prevent the regional trade potential from being realized, there remains considerable scope for efficient intra-SSA trade expansion. These general patterns also pertain to Africa's agricultural trade, particularly agricultural inputs and high-value food commodities. Over 1981-84, only 9 percent of Africa's recorded agricultural exports and 8 percent of its agricultural imports were destined for or originated from other African countries. Only 2 percent of Africa's beverage crop exports (e.g., coffee, tea, and cocoa) were traded within the region. The only major exports for which 10 percent or more was traded within the region were live animals (59 percent), tobacco (21 percent), sugar (17 percent), and fruits and vegetables (11 percent) (Valdes 1987, in Koester et al. 1990). In several subregions, however, there is anecdotal evidence that considerable unofficial, cross-border trade in various agricultural commodities occurs (Jaffee 1985; Barad 1990; Kingsbury 1989). Within Africa, at least eight regional groups have been established since the 1960s to pursue a combination of political and economic objectives. One such group, the Southern Africa Development Coordination Conference (SADCC), was established primarily to reduce its members' economic and political dependence on South Africa. SADCC has 10 member countries: Angola, Botswana, Lesotho, Malawi, Mozambique, Namibia, Swaziland, Tanzania, Zambia, and Zimbabwe. Over the past decade, however, there has been growing interest in expanding trade among SADCC members to spur more rapid agricultural and industrial development, promote agricultural diversification, and increase regional and national food security. To highlight both the potential and barriers to expanded intra-SSA agricultural trade, SADCC's recent experience in this area is reviewed below. T7he Role, Structure, and Performance of SADCC Country Agriculture The agricultural sector plays an important, if not predominant, role in the economies of all SADCC countries (table 8). At least two-thirds of the active labor force in each country is employed in agriculture. Agriculture's share of GDP is considerably lower, exceeding 50 percent for only two countries (Mozambique and Tanzania) and being less than 20 percent in Botswana, Zambia, Namibia, and Zimbabwe. With the exception of Angola, Botswana, Namibia and Zambia-all major mineral exporters-agricultural products are the leading source of foreign exchange. Most SADCC countries depend on a single commodity for 50 percent or more of their agricultural exports. For SADCC as a whole, only five commodities (coffee, tea, tobacco, sugar, and cotton), each with fair-to-poor, medium-term international market prospects, account for about 80 percent of total agricultural exports." Food crop production within the region also has a narrow base, with maize alone accounting for 56 percent of the planted area for food crops and only three crops (maize, sorghum, and cassava) accounting for 80 percent of this area (Jaffee 1991). Agricultural services, including research, extension, and marketing, have been heavily focused on maize and the core export crops. Poor agricultural performance since the early 1970s, both in terms of exports and food production, has been an important factor in the low levels of economic growth and the development of chronic food insecurity in many SADCC countries.'2 Caught in historical trading patterns with the former colonial powers and with North America, the market potential of SADCC's 80 million inhabitants has been largely ignored. Despite low levels of per capita income in much of the region, this market potential is significant. Even under severe foreign exchange constraints, SADCC imports of agricultural products (including inputs) exceeded $1 billion in 1987 and were at a level only slightly lower than SADCC's total agricultural exports for that year. In 1987, official SADCC imports exceeded $200 million for both agricultural inputs and cereals and approached $100 million for meats/animals, dairy products, and animal/vegetable oils (Jaffee 1991). Official trade statistics indicate that trade among SADCC countries constitutes less than 5 percent of the total external trade for SADCC as a whole."3 This proportion, which remained relatively constant through the 1980s, is far lower than the intraregional trade shares recorded for other regional groupings - 80 - TABLE 8. SADCC AGRICULTURAL INDICATORS % of % of merchan. % of food % of GDP exports % of agric. cropped area Averaze arowth rate (%) labor in from from exports from under maiz Per capita Agric. export agric. agric. agric. 'big five"' + sorghum food produc. volume Country (1987) (1987) (1989) (1987) (1986) (1975-90) (1975-88) Angola 71 20 1 93 62 -2.9 -15.0 Botswana 65 3 7 1 81 -2.3 4.5 Lesotho 82 21 53 0 86 -1.1 -13.4 Malawi 78 37 93 92 87 -1.5 +4.8 Mozambique 83 62 38 20 55 -2.6 -10.0 Namibia n/a 11 21 n/a n/a -2.0 +1.3b Swaziland 69 nla 40 70 n/a -1.5 +2.5 Tanzania 82 66 79 72 63 -0.6 -5.0 Zambia 70 14 2 82 84 -3.8 -0.2 Zimbabwe 70 11 47 80 74 -2.9 -5.3 n/a Not available. a. 'Big five' = coffee, cotton, tea, tobacco, and sugar. b. Figure is for 1980-88. Source: World Bank 1991a, 1992a; Akiyama and Larson 1989. in Africa (World Bank data). Approximately one-fourth of official intra-SADCC trade is in food products. For all agricultural products, including foods, agricultural raw materials, and inputs, the share of intra- SADCC trade in total external trade is on the order of 10 percent. Characteristics of Existing Intra-SADCC Agricultural Trade Intra-SADCC agricultural trade consists of a wide range of commodities. In most years, the largest amount of trade is conducted in maize, a commodity for which several countries have had large, persistent or periodic deficits and for which one country (Zimbabwe) has frequently produced an exportable surplus. Malawi has also been an irregular exporter of maize to other SADCC countries. While the volume of intra-SADCC maize trade has fluctuated greatly by year (primarily the result of variable production in Zimbabwe), in most years this trade accounts for over 10 percent of SADCC maize imports. During the 1980s, an increasing proportion of intra-SADCC maize was financed by bilateral or multilateral development agencies (World Bank/World Food Programme 1991). Intra-SADCC trade in food grains other than maize is insignificant, at least according to official statistics. There may be considerable unrecorded trade among SADCC countries and between SADCC countries and their neighbors, especially Zaire (Kingsbury 1989). Of the member countries, only Malawi has consistently exported rice, although in very small quantities. Intraregional trade in wheat is virtually nonexistent. A large number of additional agricultural commodities and inputs are traded among SADCC countries, although this trade is generally minor and irregular. Table 9 includes those trade flows that exceeded $1 million for any year over 1980-84. The table shows that most SADCC countries have exported only one or a limited few agricultural commodities on any significant scale to other member countries. The primary exception is Zimbabwe, whose agricultural trade within the region is broadly based. In addition to being relatively small overall (perhaps no more than $100 million), official intra- SADCC trade in agricultural products is characterized by (1) wide variations in volumes by year; (2) ad - 81 - hoc, one-time trades, rather than regular trade flows; TABLE9. SIGNIFICANTINTRA-SADCCAGRICULTURAL TRADEFLOWS,198084 and (3) third-party (e.g., donor) financing. These PRODUCT ECPORTER(S) lMPORTER(S) patterns reflect the limited Live animals Zimbabwe Tanzania degree of market integration Processed meat Botswana, Zimbabwe Angola, Mozambique within the region. There is Fish/fish meal Angola, Malawi, Tanzania Mozambique, Zambia, Zimbabwe also evidence that at least Vegetables Malawi Zimbabwe some and perhaps much of Animal oils Botwana Zimbabwe intra-SADCC agricultural Vegetable oils Zimbabwe Botuwana Sugar Zimbabwe Botswana, Mozambique t r ad e i s current l y Tobacco Zimbabwe, Malawi, Zambia Angola, Mozambique unprofitable in strict Cotton Zimbabwe, Zambia Malawi financial terms for Wood/pulp Swaziland Zambia, Zirnbabwe Crude veg. mat. Zimnbabwe Mozambique exporters, but iS undertaken Agric. implements Zimbabwe Tanzania to reduce storage costs and losses in surplus production Note: Trade flows exceeding $1 million for any year between 1980 and 1984. The table does years or to gain scarce not include cereals. foreign exchange made Source: Data from Chr. Michelsen Institute 1986. available through export promotion schemes (Kingsbury 1989; Rusike 1989; Food Studies Group 1990). The Potential For Expanded and Diversified Trade The weak agricultural performance of the SADCC countries and their high vulnerability to short- and medium-term climatic and market changes highlight the need for a growth-oriented agricultural development strategy that particularly emphasizes higher value commodities (e.g., vegetable oils, meats, dairy, fish, horticulture). There is scope for directing some of the diversified production to local markets, given the size and current structure of demand within SADCC countries. The success of diversification efforts, however, will be closely linked to international trading opportunities and performance, both within southern Africa and between the SADCC countries and the rest of the world. The expansion and diversification of intra-SADCC agricultural trade can potentially: * increase regional and national food security; * enable regional firms to achieve economies of scale in production, processing, and distribution; * enable regional firms to gain production and international marketing experience, thereby enhancing their competitiveness in other markets; * improve the balance of payments position of several countries. To date, most analysis of the potential for expanding intra-SADCC trade has focused on basic food grains to enhance national and regional food security and to enable SADCC members to profitably dispose of their (periodic) grain surpluses (Koester 1986; Stackhouse 1987; Lipton 1988; Kingsbury 1989). Although there is no complete consensus, it is agreed that there is potential for intraregional and cross-border trade based on (1) transport cost advantages, (2) agroecological factors resulting in comparative advantage for different food grains, and (3) regional stability of food grain production that - 82 - exceeds individual country stability. There are, however, several important political, economic, and technical barriers to increased intra-SADCC trade in food grains. Perhaps most important is the overall reluctance of governments to rely more on trade to meet basic national demands given the importance of reliable food supplies and stable prices to political stability and social welfare. Important economic and technical barriers to such trade are discussed below. There may be equal or greater scope for increased specialization and intraregional trade in a wide range of agricultural inputs, including seeds, locally produced farm machinery and equipment, tools, compound fertilizers and fertilizer raw materials, livestock remedies, crop drying and dehulling equipment, and packaging materials. Not only are agricultural requisites currently the most important agricultural imports (in value terms) for several SADCC countries, but several of the countries in the region have already developed specialized industries for many such commodities (Rusike 1989; Conroy 1990; Friis-Hansen 1990). For some agricultural inputs (e.g., equipment, tools, and seeds), international producers have relatively little incentive to develop or adapt their products specifically to suit the conditions and requirements within the SADCC region. Although individual country markets may be small, the regional market may be sufficient to enable SADCC producers to achieve economies of scale. The similarities in agricultural systems, combined with the differences in the scientific and manufacturing capabilities among the SADCC countries, point toward trade potential for agricultural inputs and technologies. Table 10 indicates potential agricultural input product lines and suppliers for intraregional trade. With increased commercialization of agriculture, urbanization, and per capita TABLE 10. POTENTIAL AGRICULTURAL income come changes in food demand INPUT SUPPLIERS WITHIN SADCC patterns, away from basic grains and PRODUCT POTENTIAL SUPPLIER(S) toward higher value products such as milk, eggs, meats, fish, edible oils, fruits and Seeds Zimbabwe (maize, sorghum, cowpeas) vegetables, and processed foods. In Zambia, Malawi (maize) Fruit seedlings Zimbabwe Angola, Botswana, Mozambique, Catle genetic material Zimbabwe Swaziland, and Tanzania, cereals now Animal health products Botuwana, Zimbabwe account for less than 55 percent of total Livestock scales/ calorie consumption (FAO 1986). The hand, equip. Zimbabwe Swaziland above high-value foods account for more Fertilizer/raw materials Zimbabwe,Zambia,Angola,Botswana,Malawi than one-third of total SADCC agricultural Agric. tools/implements Zimbabwe, Zambia, Swaziland imports (includinginps. STractor spares Zimbabwe, Botawana imports (including inputs). Small and Packaging mnaterials Zinmbabwe, Swaziland fluctuating levels of intra-SADCC trade have already been developed for many such products, but there is considerable Source: Based on Rusike 1989; Coroy 1990; and Friis-Eansen 1990. scope for expanded production and intraregional trade even taking into account the subsidized or free supplies offered by donor agencies for some products (e.g., milk powder, edible oils). In contrast with the prognosis for cereals or agricultural inputs trade, which assumes that Zimbabwe would be the dominant regional supplier (together with a regionally integrated South Africa), different countries have the potential to serve the regional market for different products. This is especially true for fish, animal products, and fresh and processed fruits and vegetables. Obstacles to Divers.ifed and Expanded Intra-SADCC Agricultural Trade Although there is considerable potential for SADCC countries to diversify their agricultural systems and expand their trade, there are major obstacles that must be removed for trade to reach its fullest potential. Some of the most serious obstacles to diversified intraregional trade are the intended and - 83 - unintended effects of selected macroeconomic and agricultural sector policies. The single most important constraint to trade faced by all SADCC countries (except Botswana) is probably the severe shortage of foreign exchange. Although the result of various structural, policy, and performance factors, one important common element has been prolonged exchange rate misalignments. Despite recent currency devaluations in several countries, SADCC country currencies have remained overvalued, some to a considerable degree. The general shortage of foreign exchange has directly or indirectly hindered trade. In the majority of SADCC countries, foreign exchange shortages have led to the imposition of strict import licensing/foreign exchange allocation systems, administered by government. These systems have directly restricted the levels and timing of imports of particular products and to particular firms. The imposed policy of import contraction during the 1980s contributed to shortages of equipment, spare parts, and raw materials serving the agricultural sector, reducing productivity and raising costs and thus diminishing the competitiveness of regional production and trading ventures. The implicit or explicit prioritization of import needs by governments has tended to be biased against imports from within the region, especially for higher value food commodities that might be construed as inessential (this may explain the wide yearly fluctuations in intra-SADCC trade in many [non-food grain] agricultural commodities). On the import side, uncertainty regarding the availability of foreign exchange has led both public and private sector firms to prefer foreign suppliers who can provide extended payment or credit terms. These suppliers are normally from South Africa or from outside the region. On the export side, foreign exchange shortages have led official trade promotion activities to focus on hard currency countries. Foreign exchange shortages have also resulted in severe limitations on capital outflows for investment in neighboring countries. Such investments might be expected to result in added intraregional trade. The misalignment of exchange rates has inhibited intra-SADCC trade for several reasons. First, currency overvaluation serves as an implicit tax on exports and other tradable commodities, inducing shifts in resources into the production of nontradables. Second, significant currency misalignment engenders uncertainty among economic agents, with future production and trading plans affected by expectations of sudden and potentially large devaluations. Third, large misalignments in the exchange rate among member countries may negate trading opportunities associated with differences in production and transport costs. Countries with relatively more overvalued currencies price themselves out of regional markets. Certain government agricultural policies have also served as a barrier to diversified production and intraregional trade. In several countries, official pricing systems exhibited a strong antitrade bias through part or all of the 1970s and 1980s, discriminating against exportables in favor of potential importables, especially maize. For example, in both Tanzania and Zimbabwe, the production of cereals has been subsidized, whereas that of exports has been heavily taxed; agriculture on the whole has been taxed in Malawi and Zambia, although the levels of taxation have been especially high for certain export crops. In several countries, panterritorial pricing has been applied for maize. Govermnent policies designed to promote maize self-sufficiency have not brought about the most efficient mix of agricultural production and trade in SADCC countries and has served as a disincentive for diversification into higher value (and higher risk) commodities. Panterritorial pricing for maize has inhibited crop diversification by subsidizing the transport costs for maize produced in remote regions where otherwise market-oriented producers would direct resources to higher value commodities that can bear higher transport costs. Panterritorial pricing may also inhibit intraregional trade by (a) placing a severe burden on already strained transport systems, tying up rail and road transport services that might be used for intraregional trade, and (b) discouraging cross-border trade by stimulating the flow of food grains to more distant local markets rather than to nearby deficit regions in neighboring countries. Throughout the region, agricultural marketing has been subject to a high degree of official regulation and administration. Strong official preference has been given to public over private marketing institutions with the presumption that major policy concerns-national food security, exports of key - 84 - commodities, income stabilization, promotion of new crops-are better served by an expanding public marketing system. Until recently, parastatal marketing agencies have had major trading functions in most important food grain and export-oriented subsectors, and parastatal agencies were assigned the primary task of promoting new crop development. Support services for secondary food crops and "minor" export commodities have remained weak. Private sector marketing has, until recently, been regarded as a residual, permitted when complementary with public trading activities (e.g., as "approved" buyers), when public marketing is infeasible (e.g., for local fruit and vegetable trading), or when no major public policy is at stake (e.g., for minor domestic or export crops). Private firms have been constrained by a web of administrative and trade regulations. The lack of an adequate transportation system in southern Africa is another major constraint to intraregional trade as six of the SADCC countries are landlocked and depend upon either technically deficient or costly rail, road, and port systems for their external trade. Transport barriers are associated with inadequate access to facilities, poor quality of services, and the high costs of such services.'4 The rail system, the major means of goods transport within the region, is biased against intraregional trade because most rail lines lead to either South Africa or the sea; for example, there are no direct rail links between Harare and either Lusaka or Blantyre. Most of the investments earmarked for improving regional rail systems have gone to corridors used primarily for nonregional trade.'" Transporting goods within the region is frequently more costly than transporting goods between regional and European ports." High transaction costs are another major barrier to trade among SADCC countries. Transaction costs are the direct and hidden costs a trader incurs when buying or selling a commodity. Such costs are associated with identifying market outlets and sources, transferring the goods and the property rights to them, and enforcing agreed stipulations. In the SADCC context, high transaction costs are a product of: 1. extensive trade and financial restrictions, which require major expenditures and personnel time lobbying officials and cutting through red tape; 2. complicated customs-clearing hurdles involving nonuniform procedures and discretionary decisionmaking; 3. the absence or inadequacy of export credit insurance; 4. the absence or inadequacy of information on regional markets, traders, and products (the result of weak communications linkages, restrictions on regional business travel, etc.). For the most part, traders incur far lower transaction costs when dealing with suppliers in South Africa or outside the region. With these countries, there are better communications linkages; more transparent customs and financial procedures; and more developed sources of product, market, and company information. In conclusion, the countries of southern Africa face slow growth, limited employment opportunities, and stagnant or declining exports. Agricultural trade diversification, especially within the region, can be an important vehicle for stimulating growth and reducing the vulnerabilities inherent in dependence on a narrow range of exports and food crops. The diversification strategy for the SADCC countries should not entail the selection of specific crops or cropping patterns for promotion or the setting of production or trade targets for specific commodities. Rather, government (and donors) should provide broadly based technical and institutional support that will enable farmers, cooperatives, and private firms to develop domestic and regional market demand and adjust to changes in that demand. This support should include both policy change and investments. Among the necessary policy changes are changes in industrial policies and trade regimes that make available supplies of intermediate inputs of higher quality and lower cost, the liberalization of commodity and input markets with relaxation - 85 - of government monopolies on all commodities, the adoption of crop-neutral pricing policies, and the removal of legal and institutional barriers on production and trade. Potentially important investments include national or regional investments in agricultural research; expansions in agricultural, industrial, and trade credit; improvements in information systems; and improvements in road, rail, and other transport facilities. Summary and Conclusions During the latter half of the 1980s, many African countries experienced an improvement (over the prior decade) in their food production and agricultural export performance. Over half of the countries in the region are undertaking structural adjustment programs, many of which involve reforms in agricultural pricing, input supply, marketing, and trade policy (Jaeger 1991). These reforms, if combined with measures to improve physical infrastructure and social and production services, could stimulate increased investment in agricultural production, technology development, and agroindustry, and could provide the basis for an expansion and diversification of intra- and extraregional agricultural trade. Agricultural diversification could provide an important basis for economic growth for many African countries. The country and thematic case studies provided in this paper offer several lessons regarding the diversification process. The case of the Gezira Scheme in Sudan indicates that although investment in irrigation can provide the potential for more diverse and flexible farming systems, it will be undermined where the government attempts to impose a particular crop or cropping pattern on farmers. Crop diversification is more likely to be efficient when it is demand-driven rather than imposed for convenient irrigation management or to meet predetermined production targets. The Gezira experience also indicates that the provision of irrigation facilities must be combined with effective support services-either by the private or public sectors-to enable farmers to attain high levels of productivity in new food or commercial crops. Finally, the Gezira experience illustrates the ingenuity of farmers in bypassing a command economy when resource allocation patterns other than those dictated by government result in both higher incomes and reduced risk. The case study on contract farming indicates that single source technical and input supply and product marketing can be a second best solution to poorly functioning markets and support services in Africa. Not surprisingly, many of Africa's successful cases of new agroindustrial development and trade diversification have featured institutional arrangements such as contract farming and vertical integration, which have internalized, bypassed, or supplemented existing product and input markets and general support services. These institutional frameworks have frequently provided sufficient incentives and risk- reducing or -sharing mechanisms to stimulate investments by farmers, processors, and traders in new crops, new facilities, and new trading relationships. Whereas many contract farming schemes in Africa have been initiated and managed by private firms, most of the larger schemes, especially those involving tree crops and investments in large processing facilities, have featured public sector participation, frequently in joint ventures with international companies. The hybrid institutional features of many contract farming schemes in Africa belie the stark dichotomy of the public and private sectors often portrayed across the African economic landscape. Although contract farming (and vertical integration) operations have been an effective vehicle for new agroindustrial development, this institutional arrangement is not viable for some crops, especially staple food crops. Wherever possible, efforts should be made to incorporate smallholder farmers into contracting schemes and to supplement crop- (or livestock product)-specific support systems with technical assistance for other crops and farming activities. Kenya's horticultural export experience illustrates the large potential for demand-driven, nontraditional exports in Africa, along with the high risks and long gestation period in developing such exports. The case also' shows the important role of traders/intermediaries in linking African producers with specialized international markets. In the fresh produce industry, the inadequate provision of public - 86 - goods (e.g., research, extension, port facilities) by the government lengthened the development process, added to production and trading risks, delayed the entry of small farmers and African exporters, and biased the direction of crop and trade diversification away from more technically demanding crops or crops with extended gestation periods. With cut flowers, it took a major foreign investment to fill the vacuum of a nonexistent floricultural support system. This foreign investment eventually fostered the development of an entire industry, featuring large- as well as small-scale producers and traders. In the fruit and vegetable processing industry, "lumpy" investments in large-scale processing facilities deterred private firms, and thus the government played an important risk-bearing and financial role in the early development years. Although the government proved to be a poor manager of commercial ventures, it recognized this and encouraged foreign investment, management, and marketing linkages. The SADCC story indicates that while there is considerable scope for expanded production and intraregional trade in many agricultural commodities and implements, such diversified production and trade are blocked by a combination of misguided agricultural, trade, and macroeconomic policies; infrastructural bottlenecks; and poor informational flows within the region. In this context, traders are prevented or severely hampered from encouraging new forms of production and developing new markets. Economies of scale in the production or processing of many products cannot be realized when servicing only the local market. Capturing the benefits from (diversified) intraregional agricultural trade will require not only significant policy changes at national levels, but investments in production and trade infrastructure and support at the national and regional levels. Efficient agricultural diversification in Africa could take many forms, including: 1. expanding the production, processing, and marketing of secondary food crops; 2. developing new uses and products from existing raw materials; 3 . expanding subregional trade in food products and agricultural inputs; 4. adding value to traditional exports through processing and/or the targeting of special market niches; 5. expanding production and trade of high-value commodities with favorable domestic and international market prospects. Although the diversification of production, processing, and trade should form part of a growth- oriented strategy for African agricultural development, attention must also be given to improving productivity in traditional crop production and marketing and to regaining international competitiveness in traditional exports. An efficient process of agricultural diversification is unlikely to build upon a stagnant or declining agriculture. There is scope for increased foreign investment and international technology transfer for African agriculture and agroindustry, but a large proportion of the financial and technical resources, managerial and entrepreneurial skills, and social institutions necessary for product, agroindustrial, and market diversification will have to be supplied locally or regionally within Africa, from existing firms and subsectors. Both the private and public sectors have important roles to play in the diversification process. The private sector-cooperatives; trade associations; and small-, medium-, and large-scale firms and farmers- should play the central, if not exclusive, role in commercial production, processing, and (input and commodity) marketing activities. The public sector needs to provide the wider institutional and infrastructural basis for efficient agricultural market development. At a minimum, this will entail policy reforms that improve the macroeconomic environment; measures to create a more secure legal and institutional setting for trade and investment; investments to expand and maintain transport, - 87 - communications, and other infrastructure necessary for market development; and an improved supply of (or financing for) agricultural research, extension, and other services with public good properties. Notes 1. Over 1960-85, eight African countries-C6te d'Ivoire, Kenya, Liberia, Malawi, Rwanda, Swaziland, Tanzania, and Uganda-each achieved average annual growth rates of 3 percent or more in agricultural production (Harrison 1990: table 1). 2. Food processing, beverage, tobacco, and textile industries are among those dependent on agricultural raw materials. Agricultural input supply and food marketing services are among the leading service activities in many African countries. 3. Over 1981-84, 74 percent of Africa's agricultural exports were directed to OECD countries. For some major commodities, this share was even higher (coffee, 85 percent; cocoa, 87 percent) (Valdes 1987, in Koester et al. 1990). 4. Whether trade diversification actually stabilizes export earnings depends upon the correlation between prices and the annual production of the different commodities in the portfolio. Diversification can stabilize export eamings where the earnings from the new products are either uncorrelated or inversely correlated with those from traditional products (Siegel 1990). 5. Cash crop and part-time employment earnings may enable smallholder farmers to improve their productivity in food grain production (via increased fertilizer use) or to invest in improved food storage facilities. Where vertical diversification involves the processing of secondary food crops, the increased palatability and storability of such items may serve to diversify local diets and reduce household reliance on the primary food grain. These processing activities may stimulate increased commercially oriented production of drought-resistant crops. At the national level, the potential expansion and/or increased stability of foreign exchange earnings associated with trade diversification increases the capacity to import food as well as inputs used in food production and marketing. 6. Between 1961 and 1970, groundnut yields increased by an average of 9.2 percent annually, whereas wheat yields increased by 0.7 percent a year (Mohammed 1983: 156). 7. The origins, scale, and performance of contract farming schemes in Kenya are examined in Jaffee 1990. 8. See Billings 1988 for a review of producer-buyer links in Senegal's poultry subsector. 9. This case study is based primarily on Jaffee 1990, 1992, and forthcoming. 10. Although foreign exchange earnings increased, each of these ventures showed steady financial losses during the decade, resulting in no corporate taxes or investment dividends for the government. The foreign partners were satisfied with earning their profits overseas, through the marketing and distribution of the products. 11. Among SADCC countries, only Mozambique (prawns, cashew nuts), Botswana (meat), and Lesotho (mohair) rely on commodities other than the 'big five' for the bulk of their agricultural exports. 12. Over 1973/77 to 1984/88, all SADCC countries (except Tanzania) recorded negative average rates of change in per capita cereals production. Between 1980 and 1988 the real value of SADCC country agricultural exports declined by some 39 percent. See Jaffee 1991 for these and other indicators of SADCC agricultural performance. - 88 - 13. Four countries-Zimbabwe, Mozambique, Zambia, and Botswana-accounted for three-fourths of the total value of recorded intra-SADCC trade in the early to mid-1980s. Only a minuscule proportion of the external trade of Angola, Lesotho, Swaziland, and Tanzania has been conducted with other SADCC countries. 14. See Rusike 1989 and Gaviria 1991 for discussions of transport constraints and costs. Rusike found that 60 percent of the firms he surveyed reported a shortage of reliable and reasonably priced transport services as a significant trading constraint. 15. Compared with the rail system, road transport accounts for a relatively small proportion of international goods transport. In addition to requiring more fuel, the region's road transport systems are impaired by poor maintenance, truck shortages, and the lack of spare parts and repair facilities. 16. Rusike's (1989) Zambian respondents reported that one could bring containerized cargo from Britain to Dar es Salaam for $80/mt, whereas the final leg of the journey to Lusaka cost $103/mt. Agricultural machinery exporting firms in Zimbabwe also reported their costs for transporting equipment to Tanzania to be roughly twice those of firms in Brazil or Norway (p. 32). References Akiyama, T., and D. Larson. 1989. "Recent Trends and Prospects for Agricultural Commodity Exports in Sub- Saharan Africa." World Bank PPR Working Paper 348. Washington, D.C. Allen, G. 1983. 'Development of the Mumias Company, Kenya." Oxford Agrarian Studies 12: 63-93. Barad, R. 1990. "Unrecorded Transborder Trade in Sub-Saharan Africa and Its Implications for Regional Economic Integration." In Long-Term Perspective Study of Sub-Saharan Africa. Vol. 4. Washington, D.C.: World Bank. Barnett, T. 1977. The Gezira Scheme: An Illusion of Development. London: Frank Cass. Barnett, T., and A. Abdeilcarim. 1991. Sudan: The Gezira Scheme and Agricultural Transition. London: Frank Cass. Billings, M. 1988. "Contract Poultry Farming in Senegal." Binghamton, N.Y.: Institute for Development Anthropology. Blench, R. 1987. Livestock in the Gezira Scheme-1986. Overseas Development Institute, Agricultural Administration Unit Paper 23c. London. Buch-Hansen, M., and J. Kieler. 1983. "The Development of Capitalism and the Transformation of the Peasantry of Kenya." Rural Africana 15-16: 13-39. Chr. Michelsen Institute. 1986. SADCC Intra-regional Trade Study. Prepared by the Department of Social Science and Development for SADCC. Bergen, Norway. Conroy, A. 1990. "Fertilizer Use and Distribution in Zimbabwe." Background paper prepared for the World Bank Zimbabwe Agricultural Sector Memorandum. Washington, D.C. D'Silva, B. 1986. Sudan's Irrigated Subsector. U.S. Department of Agriculture, Economic Research Service Report AGE860811. Washington, D.C. - 89 - Davis, J. 1979. Capitalist Agricultural Development and the Exploitation of the Propertied Laborer." In F. Buttel and H. Newby, eds., The Rural Sociology of Advanced Societies. London: Croom Helm. Euroconsult. 1982. Gezira Rehabilitation and Modernization Project 1. Finial report to the Government of the Democratic Republic of the Sudan. Khartoum. Faki, H. 1982. "Disparities in the Management of Resources Between Farm and National Levels in Irrigation Projects: Example of the Sudan Gezira Scheme." Agricultural Administration 9: 47-59. FAO. 1986. Food Supply Situation and Crop Prospects in Sub-Saharan Africa. Rome. Food Studies Group. 1990. "Agricultural Marketing and Pricing in Zimbabwe." Background paper prepared for the World Bank Zimbabwe Agricultural Sector Memorandum. Washington, D.C. Friis-Hansen, E. 1990. "The Zimbabwe Seed Industry." Background paper prepared for the World Bank Zimbabwe Agricultural Sector Memorandum. Washington, D.C. Gaitskell, A. 1959. Gezira: A Story of Development in the Sudan. London: Faber and Faber. Gaviria, J. 1991. Rural Transport and Agricultural Performance in SSA: 6 Country Case Studies. Washington, D.C.: World Bank and the Economic Commission for Africa. GATT (General Agreement on Tariffs and Trade). 1989. International Trade, 1988-89. Geneva. Harrison, P. 1990. "Sustainable Growth in African Agriculture." In Long-term Perspective Study of Sub-Saharan Africa. Vol. 2, Economic and Sectoral Policy Issues. Washington, D.C.: World Bank. Hazell, P., M. Jaramillo, and A. Williamson. 1989. "How Has Instability in World Markets Affected Agricultural Export Producers in Developing Countries?" World Bank PRE Working Paper 263. Washington D.C. Horton, J. 1987. "Characteristics of the Horticultural Export Enterprises Utilizing Contract Farming Schemes in Senegal." Binghamton, N.Y.: Institute for Development Anthropology. Islam, N. 1990. Horticultural Exports of Developing Countries: Past Performances, Future Prospects, and Policy Issues. Research Report 80. Washington, D.C.: International Food Policy Research Institute. Jaeger, W. 1991. "The Impact of Policy in African Agriculture: An Empirical Investigation". World Bank PRE Working Paper 640. Washington, D.C. Jaffee, S. 1985. "The Marketing of Livestock and Livestock Products in Somalia." Working Paper for the World Bank/Government of Somalia Agricultural Sector Survey. Washington, D.C. 1987. "Case Studies of Contract Farming in the Horticultural Sector in Kenya." Binghamton, N.Y.: Institute for Development Anthropology.. 1990. "Alternative Marketing Institutions for Agricultural Exports in Sub-Saharan Africa with Special Reference to Kenyan Horticulture." Unpublished doctoral dissertation, Oxford University. . 1991. "Agricultural Diversification and Trade in SADCC Countries." Paper prepared for the World Bank, Agriculture Operations Division, Southern Africa Department. Washington, D.C. . 1992. "How Private Enterprise Organized Agricultural Markets in Kenya." World Bank PRE Working Paper 823. Washington, D.C. 90 - Forthcoming. 'Marketing Africa's Horticultural Exports: A Transaction Cost Approach." World Bank, Africa Technical Papers series. Kingsbury, D. 1989. "An Analysis of Price and Non-price Barriers to Agricultural Marketing and Trade in Southern Africa." Unpublished doctoral dissertation, Michigan State University. Koester, U. 1986. "Regional Cooperation to Improve Food Security in Southern and Eastern African Countries." International Food Policy Research Institute, Research Report 53. Washington, D.C. Koester, U., H. Schafer, and A. Valdes. 1990. Demand-Side Constraints and Structural Adjustment in Sub-Saharan African Countries. Washington, D.C.: International Food Policy Research Institute. Lamb, G., and L. Muller. 1982. "Control, Accountability, and Incentives in a Successful Development Institution: The Kenya Tea Development Authority." World Bank Staff Working Paper 550. Washington, D.C. Lele, U., N. van de Walle, and M. Gbetibouo. 1989. "Cotton in Africa: An Analysis of Differences in Performance." Working paper in World Bank study on Managing Agricultural Development in Africa. Washington, D.C. Lipton, M. 1988. "Regional Trade and Food Security in Southern Africa." In C. Bryant, ed., Poverty, Policy and Food Security in Southern Africa. Boulder, Colo.: Lynne Rienner. Magar, W. 1986. "Farms in the Gezira.' In A. Zahlan, ed., The Agricultural Sector of Sudan: Policy and Systems Studies. London: Ithaca Press. Mohammed, H. 1983. "The Impact of the Transfer of Agricultural Technology on the Sudan." Unpublished doctoral dissertation, Boston University. Ommeh, M. 1984. "An Investigation into the Kenyan Cashew Nut Industry." Unpublished master's thesis, University of Nairobi. Plusquellec, H. 1991. The Gezira Irrigation Scheme in Sudan. World Bank Technical Paper 120. Washington, D.C. Robertson, A. 1987. The Dynamics of Productive Relationships. New York: Cambridge University Press. Rusike, J. 1989. "Trader Perceptions of Constraints on Expanding Agricultural Input Trade Among Selected SADCC Countries." Department of Agricultural Economics and Extension, University of Zimbabwe, Working Paper AEE 5/89. Schluter, M. 1984. "Policies to Increase Production of Commodities with Export Potential to Oil Exporter Markets." Institute for Development Studies, Working Paper 406. Nairobi. Siegel, P. 1990. "Diversification of Agricultural Commodity Exports in SADCC Countries." Paper prepared for the World Bank, Agriculture Operations Division, Southern Africa Department. Washington, D.C. Stackhouse, L. 1987. "Analysis of Trade Flows in Staple Agricultural Commodities in the SADCC Region." Unpublished master's thesis, Michigan State University. Tait, J. 1983. "The Modernization of the Colonial Mode of Production in the Gezira Scheme." In P. Oesterdiekhoff and K. Wohlmuth, eds., The Development Perspectives of the Democratic Republic of Sudan. London: Weltforum Verlag. -91 - Varangis, P., T. Akiyama, and E. Thigpen. 1990. "Recent Developments in Marketing and Pricing Systems for Agricultural Export Commodities in Sub-Saharan Africa." World Bank PRE Working Paper 431. Washington, D.C.. Watts, M., P. Little, C. Mock, M. Billings, and S. Jaffee. 1988. Contract Farming in Africa. Vol. 1, Comparative Analysis. Binghamton, N.Y.: Institute for Development Anthropology. World Bank. 1990. Sub-Saharan Africa: From Crisis to Sustainable Growth. Washington D.C. 1991a. World Development Report 1991. New York: Oxford University Press. 199 lb. Price Prospects for Major Primary Commodities, 1990-2005. Vol. 2. Washington, D.C. 1992a. African Development Indicators. Washington, D.C. 1992b. World Development Report 1992. New York: Oxford University Press. World Bank and World Food Programme. 1991. Food Aid in Africa: An Agenda for the 1990s. Washington, D.C. CENTRAL AMERICA AND SOUTH AMERICA'S PACIFIC RIM COUNTRIES: EXPERIENCE WITH EXPORT DIVERSIFICATION Panfilo C. Tabora, Jr. Central America constitutes a small land mass, but its location has provided it with richly diverse physical and biological resources for agriculture (Leonard 1987). A volcanic range divides Central America into dry Pacific and wet Atlantic sides, rendering cool subtropical highlands and hot tropical lowlands. Social, political, and economic factors also make the region perpetually dynamic. Central America and several South American Pacific Rim countries are dominated by agriculture, particularly a few export crops. This paper examines export diversification in Central America, focusing on the five Central American Common Market countries: Costa Rica, El Salvador, Guatemala, Honduras, and Nicaragua. Particular attention is given to public and private roles in the diversification process, with the efforts of nongovernmental organizations (NGOs), transnational corporations, and governments highlighted. The diversification experience of several South American Pacific Rim countries-Chile, Colombia, and Ecuador-is also reviewed and compared with that of Central America. Agriculture in the Region In 1989, the share of agriculture in total GDP amounted to 26 percent in Guatemala, 23 percent in El Salvador and Nicaragua, 24 percent in Honduras, and 20 percent in Costa Rica. In Central America as a whole (excluding Panama), it averaged 23 percent (CEPAL 1989). In Colombia and Ecuador, the figure was 22 percent (1991 data from the Colombian National Coffee Federation [FEDERECAFE] and the Ecuadorian Agricultural Foundation [FUNDAGRO]), and in Chile and Peru, 10 and 14 percent, respectively (BID 1987). In 1985, the share of agriculture in employment in both Ecuador and Colombia was 34 percent; in Chile, 20 percent; in Costa Rica, 28 percent; in Nicaragua, 37 percent; in Honduras, 54 percent; in El Salvador, 40 percent; and in Guatemala, 58 percent (lICA-FLACSO 1991). In all five Central America Common Market countries, exports of primary agricultural products constitute over 50 percent of total exports (IMF 1990) (table 1). Bananas and coffee have experienced adverse price fluctuations, and the profitability of other commodities such as sugar, beef, tobacco, and cotton has also been decreasing due to increasing costs of inputs (e.g., imported fertilizers and chemicals) and noncompetitiveness in the market (Baumeister 1990). Since the late 1980s, however, export patterns have been changing in most Central American countries, and the contribution of nontraditional agricultural products has been increasing. In Costa Rica and Guatemala, the combined exports of nontraditionals have begun to surpass the single largest export, coffee (table 2). These two countries have vigorous agricultural diversification thrusts that have proven effective in promoting nontraditionals. Agricultural diversification involves spreading investments and benefits among various commodities to increase the variety of commercial products and benefits from agriculture. This definition implies not only an economic perspective, but also cultural, social, political, and technological aspects. Panfilo C. Tabora, Jr., is leader, Diversification Program, Honduran Foundation for Agricultural Research, La Lima, Honduras. - 94 - TABLE 1. SHARE OF FIVE MAJOR AGRICULTURAL EXPORTS TO TOTAL EXPORTS OF THE CENTRAL AMERICAN COMMON MARKET COUNTRIES, 1970-89 EXPORT 1970 1975 1980 1984 1985 1986 1987 1988 1989 Bananas 29.7 22.7 34.7 34.6 40.9 47.1 45.0 35.5 30.5 Coffee 15.9 11.4 10.5 12.9 13.9 12.4 14.2 14.0 13.1 Cotton 7.8 10.6 5.4 8.2 8.4 4.1 3.6 5.2 02.6 Meat (beef) 5.6 3.7 5.9 2.6 2.6 2.2 3.0 3.3 04.7 Sugar 3.4 12.6 3.6 4.4 2.8 3.5 3.5 3.1 04.1 Total 62.4 61.0 60.1 62.7 68.6 69.3 69.3 61.1 55.0 Note: Central American Common Market countries are Costa Rica, El Salvador, Guatemala, Honduras, and Nicaragua. Source: UICA-FLACSO 1991: 62-64. Economically, diversification entails TABLE 2. PERCENTAGE PARTICIPATION OF spreading risks and benefits and NONTRADITIONAL AND COFFEE EXPORTS FOR counterbalancing fluctuations GUATEMALA AND COSTA RICA, 198488 encountered in exchange economies. EXPORTS 1984 1985 1986 1987 1988 Bananas, coffee, sugar, and meat account for about 70 percent of all Nomradirdonals exports for most Central American Guatemala 37.5 35.6 29.4 40.7 42.1 countries, and for a large percentage of Costa Rica 34.1 35.9 33.8 39.5 43.3 exports for Colombia and Ecuador. Coffee Price fluctuations, therefore, greatly Guatemala 31.8 42.6 48.1 36.3 36.0 concer policmakersand ecnomic Costa Rica 26.8 33.0 36.1 30.2 25.4 concern policymakers and economic administrators because export revenues are an important source of financing Source: nCA-FLACSO 1991:62. the national budget and foreign debt. When commodity prices drop as much as 50 percent within a year (as with coffee and bananas), gross receipts fall drastically and dollar reserves are often depleted. That several commodities could develop to a level sufficient to counteract the income drops in bananas and coffee is therefore very appealing. New commodities, however, also entail new technologies. Improved varieties, cultural management, irrigation, harvesting and postharvest technologies, and marketing are all necessary for the success of new commodities. Familiarization witi and utilization of such technologies also require time and resources. Because Central American countries have scarce resources (e.g., skilled labor), they can pursue only a few commodities at a time. Thus far, each country has pursued a different strategy. Honduras and Colombia have leaned toward technology development in a few commodities, whereas Chile and Guatemala have a more direct technology transfer strategy. In between is Costa Rica, with a more integrated approach to technology and agribusiness development (see box 1). These strategies correspond to the differences in the human capital resources among these countries (table 3). Costa Rica has the highest component of administrators and managers in its labor force (13.7 percent) and also the highest percentage of technical professionals (8.4 percent). Honduras has a higher technical professional component than Guatemala (4.1 and 3.6 percent, respectively). The reverse is true for administrators and managers, with Guatemala having 1.1 percent to Honduras's 0.9 percent (IICA-FLACSO 1991). El Salvador and Nicaragua have lower percentages of managerial profes- - 95 - sionals, 0.6 and 0.9 percent, respectively; they do have BOX 1. AN R&D APPROACH IN COSTA RICA higher percentages of technical professionals (4.2 Costa Rica has diversified into commodities such as palmito, chayote, and 5.2 percent, respectively) guanabana (soursop), black pepper, strawberries, coconut, annatto (achiote), , macadamia root crops and mango. This process has included research trials than~ ~ ~ ~ ~~~~acdma Guatmal crops and mago Thi prcs ha includedresearchtria.s' than Guatemala and validation of results, preparation of enterprise studies, and agribusiness planning Honduras. conducted by several notable institutions, including (a) CATIE (the Center for Costa Rica's high Instruction and Research in Agriculture), a regional center based in Costa Rica; managerial and sales and (b) the association of independent banana growers (ASBANA, now professional components CORBANA), who supply the major brands. The University of Costa Rica also reflect its policies to promote participates in diversification efforts, and INCAE (the regional business management school with links to the Harvard School of Business) has moved from an outwardly linked system Nicaragua to Costa Rica. Private corporations such as Del Monte and Dole have that enables it to exploit research facilities in Costa Rica, while Chiquita's NUMAR division has technical opportunities in world groups in the country. The less visible, though more numerous, small markets as well as acquire entrepreneurs, both local and foreign, have also added a vibrant dimension that technology from global has tumed many commodities with small volumes into export commodities. technoloy .The heart of palm from pejibaye exemplifies the research process. The sources. Costa Rica's drive University of Costa Rica, CATIE, and ASBANA conducted independent research for new investments and its that concluded that pejibaye was a feasible crop for palm heart (palmito) aggressive plan for new production. Subsequently, the research results were adopted by the Del Campo investor residents have group, a processing firm based in Costa Rica (with Mexican investments). Palmito attracted a wealth of tech- has come a long way: 17 years ago it was simply a research undertaldng; now nologies and new products Costa Rica exports some $6 million worth of palmito from some 3,000 ha. The process has been repeated for black pepper, mango, guanabana, coconut, lemons, for the market. Guatemala plantains, chayote, root crops, ornamentals, and strawberries. and Honduras are following , __ _, Costa Rica's example, and the new settlers (Europeans, North Americans, and a few Asians) have invigorated the economy with many new activities. These new settlers have access to world markets and therefore have confidence in establishing new export product lines. The peaceful and stable political climate has doubtlessly been a factor. In Chile the military junta provided some form of stability for about 20 years, which nurtured entrepreneurial activity favorable to the acquisition of technology. Since 1965, Chile has been developing its managerial and technical capacity with strong linkages to the universities of Chile, California, and Chicago. Agricultural Diversification Programs Bananas and coffee were fortuitous diversification commodities in the early 1900s (as colorants and cocoa were the principal commodities from the colonial days in the early 1800s) (Martinez et al. 1986). Sugarcane, beef, and cotton were aggressively promoted in the 1950s and 1960s, but never achieved similar importance. These three commodities were supported by major state promotion programs, including state enterprises and related state-supported agrarian cooperatives. During the economic expansion of the 1940s to 1960s, the private sector was also active in diversification. Both the United Fruit Co. and the Standard Fruit Co., American transnationals, had established major facilities and research stations (see box 2) (Rosengarten 1991). The United Fruit Co. sought to diversify with African oil palm, Manila hemp, and rubber, with major plantations in Honduras, Costa Rica, and Guatemala. The Standard Fruit Co. diversified into pineapple in Honduras. Del Monte followed suit, investing in tropical fruits and vegetables in Costa Rica (Attenburg et al. 1990). Cooperatives and other entrepreneurial outfits invested in Manila hemp in Ecuador and in ornamental crops in Colombia. All these endeavors occtirred within the context of agribusiness activities, with no - 96 - special government incentive. TABLE 3. ACTIVELY EMPLOYED POPULATION BY A major impetus PRINCIPAL OCCUPATION, 1980 (%) toward diversification EL COSTA occurred during the OCCUPATION GUATEMALA SALVADOR HONDURAS NICARAGUA RICA 1980s, when rising Technical/ petroleum prices led to professional 3.6 4.2 4.1 5.2 8.4 an economic slowdown Managerial 1.1 0.6 0.9 0.9 13.7 in Central and South Salesperson 6.2 14.1 5.8 7.1 13.9 America (except in Agricultural worker 56.7 39.5 59.4 46.7 47.7 Laborer Ecuador, which does not Production 18.2 26.4 17.2 21.9 O.00 import oil). Export Other 14.2 15.2 12.6 18.2 16.3 earnings shrank while oil Total 100.0 100.0 100.0 100.0 100.0 import costs increased. At the same time, these a. Included in previous category. countries become heavily Source: WVilkie 1982, as cited in ICA-FLACSO 1991. indebted (including Ecuador), and declining export revenues exacerbated difficulties in debt repayment. Debt servicing surged as a share of total exported goods and services (table 4). This situation stimulated the drive to increase exports of other commodities, i.e., nontraditionals. In the late 1970s and - early 1980s, the U.S. Agency BOX 2. THE LANCETILLA GARDENS IN HONDURAS for International Development (USAID) was involved in The United Fruit Co. (now Chiquita Brands International) had an p r o j e c t s p r o m o t i n g important research facility, its Tropical Research Division, based in Honduras. This division maintained a full botanical garden and research center called the agroexport development and Lancetilla Gardens, the most extensive collection of tropical fruits, spice and oil socioeconomic welfare, crops, and ornamentals in the Western Hemisphere. Lancetilla, established in particularly in Honduras and 1925, maintained a full complement of scientists, in addition to the support it El Salvador. USAID helped received from 30 to 45 scientists from the Tropical Research Division, and create new NGOs such as the produced many notable publications. Honduran Foundation for Plants from around the world were collected by the staff, and, of course, bananas were the most extensively collected. To prevent transfer of Agricultural Research disease, a USDA-assisted quarantine program was established on an island off the (FHIA); the Federation of Honduran coast. The first studies of African oil palm and the first major research Producers and Exporters of undertaking on rubber in Latin America were performed at Lancetilla. Manila Agricultural and Livestock hemp was also introduced there. Today, there are 25,000 ha of oil palm in Costa Products (FEPROEXAAH); Rica and Honduras, some 1,000 ha of black pepper in Honduras and Costa Rica, the F dsome 5,000 ha of rubber in Guatemala, over 50,000 ha of citrus in Central the Foundation for Industrial America as a whole, and 10,000 ha of Manila hemp in Ecuador. Mango, Development (FIDE); the rambutan, litchi, breadfruit, and specialty bananas, initially colected and grown Livestock Fund (Fondo at Lancetilla, are gradually spreading. Ganadero); the publication El The founder of LancetiUa, Wilson Popenoe, was an American Agricultor; and AVANCE, a horticulturist who also established the Panamerican School of Agriculture in socialy oriented project. Honduras. Numerous graduates of the school have duplicated LancetiUa socially oine prjc. throughout South America. There were also USAID- In 1974, Lancetilla was tumed over to the Honduran govemment and funded projects with the became a public garden. Many of the plants collected earlier continue to spread Honduran Coffee Institute to other Central and South American countries. The garden remains a potent (IHCAFE) and Ministry of technical instrument for observing plant products that may have diversification Agriculture's PRODIVERSA potential in Latin America. (Project for Agricultural - 97 - Diversification). The Panamerican School of TABLE 4. PERCENTAGE SHARE OF EXTERNAL DEBT Agriculture also had SERVICING TO TOTAL EXPORTS OF GOODS AND SERVICES USAID funds for some COUNTRY 1979 1980 1984 1985 1986 1987 1988 1989 projects in melons. As a result of USAID's Guatemala 7.7 3.4 33.9 43.8 42.3 34.8 42.4 nla development assistance El Salvador 7.8 4.4 53.8 48.1 51.8 47.1 44.9 32.7 dee n aHonduras 22.6 20.2 39.4 41.6 47.6 53.2 50.2 nla by the mid-1980s the Nicaragua 8.9 16.2 18.3 19.4 10.9 7.7 4.0 3.5 banana industry was no Costa Rica 33.0 39.8 56.9 58.9 52.1 53.5 54.4 58.0 longer the largest Average 16.0 16.8 40.5 42.4 40.9 39.3 39.2 n/a foreign exchange contributor to the Source: BCA-FLACSO 1991. Honduran economy. In El Salvador, FUSADES (the Foundation for Salvadoran Development) organized its diversification program, DIVAGRO, in 1987. DIVAGRO now has cross-linked projects with the country's water management project and agrarian reform financing sector. The Regional Organization for Central America and Panama (ROCAP), the USAID's regional arm, had myriad projects devoted to fostering nontraditional products for exports. The private sector was also active, particularly those working as groups, for example banana, coffee, and rubber producers.' Toward the mid-1980s, every country in the region had formed new special programs for exporting numerous nontraditional agricultural products.2 Current Issues in Diversification Although there is agreement in principle on the importance of diversification, the instruments for implementation, such as the approach to technology, financial and investment policy, infrastructure, human resource development, and the role of government, have become major issues. Technology Development versus Technology Transfer New commodities and the attendant technologies face the problem of diversity in the physical and cultural environment. New varieties have to be tested for adoption; cultural management has to be suited to the climate, soils, irrigation, and pest management systems; and postharvest handling, packaging, and storage require site-specific techniques. On the one hand, technology development is often the first essential step, as with melons in Honduras and pejibaye hearts of palm and chayote in Costa Rica. On the other hand, the successful direct transfer of technology in snowpeas and asparagus in Guatemala, temperate fruits in Chile, and Manila hemp in Ecuador supports the technology transfer proponents. There are numerous flaws in both approaches, however, which render the issue unresolved. Governments in Central and South America have built research stations and laboratories; staff, however, are poorly paid, and ever-changing government administrations do not allow a continuity and stability in the civil service that could sustain long-term research and technology development. In response, NGOs have been organized with generous grants from international development agencies and the private sector. The Chile Foundation and the Honduran Foundation for Agricultural Research are two examples; yet these institutions must also face the issue of sustainability. Private sector funds are often tied to financial gains generated quickly, and international grants also have limited time frames within which to produce an impact. Because a research institution's ability to generate substantial funds is limited, these time constraints can easily affect its viability. - 98 - Public versus Private Sector Investments During the 1960s and 1970s, many governments directly set out to establish public and parastatal enterprises to produce new products (table 5). Honduras's CONADI and Chile's CORFO, both national development corporations, were designed to invest in new commodities. Numerous cooperatives were also established, many of which were a response to the agrarian reform policy of the 1950s and 1970s. Because these programs were primarily socially and politically oriented, an overriding agribusiness orientation was absent, thus stunting their development. In the 1980s, the focus shifted toward private sector initiatives, partly the result of the Caribbean Basin Initiative that promoted the participation of the private sector. U.S. embassies, through USAID, directed the thrust with an annual allocation amounting to as much as 50 percent of Central America's annual dollar receipts from exports of goods and services (table 6). USAID's participation highlights the critical role of international development institutions as catalysts for investments. New foreign-owned companies have also been encouraged TABLE 5. DEVELOPMENT OF COOPERATIVES INVOLVED to increase production. These IN NEW NONTRADITIONAL COMMODITY PROGRAMS companies have had a tremendous NO. OF impact on expanding the production COUNTRY COOPERATIVES COMMODrrIES and export of cucumbers, melons, mangos, snowpeas, and winter Costa Rica 16 Chayote, orange, macadamia pineapple, black pepper, passion squash. Smaller foreign enterprises fruit, oramentals, strawberries, have also flourished in Central tubers, plantain, cacao America and are beginning to move into South America's Pacific Rim. El Salvador 16 Melons, cardamon, soybeans, other The established transnationals, however, found that-with few Guatemala 32 Cucumbers, snowpeas, chayote, exceptions-crop diversification did broccoli, melons, cauliflower not meet their profit expectations. Honduras 9 Cotton, plantain, melons, cucumber, Consequently, they became oil palm, squash increasingly irtvolved in expanding Panama 4 Honeydew melons, pineapple, black their banana production capacities. pepper, passion fruit, strawberries, Bananas are still the top grosser per omamentals unit area, and there are few crops with extensive markets that are more profitable (Tabora 1989). Infrastructural Development Little of Central and South America's lowland fertile valleys are irrigated, although streams and rivers are abundant and located at ideal elevations for irrigation. (The major agricultural activities such as coffee and livestock production and the larger farming communities are in the highlands.) The lowland plains were settled relatively recently (early in the century), and the few irrigated areas encompass extensive plantations of bananas and sugarcane. There are few major infrastructural projects, though some are being planned. They are, however, mired in social, political, and environmental controversies. Other infrastructure that could stimulate diversification, such as roads, ports, and depots, has not been fully exploited. For example, the modern seaport of Puerto Castilla in northern Honduras, the irrigation project in Guanacaste in Costa Rica, and the seaport of San Jose in Guatemala have all been underused for over nine years. Roads have also been underused, even those that lead to newly opened fertile valleys (Bajo Aguan in Honduras, San Carlos in Costa Rica, and southeastern Guatemala). - 99 - Nontraditional crops have been slow to develop in these areas because the population influx to such areas TABLE 6. USAID ALLOCATION AS mainly consists of subsistence-level farmers who raise PERCENT OF EXPORT INCOME, 1985, 1986 traditional food crops and cattle. COUNTRY 1985 1986 Substantial crop diversification has occurred, however, near urban areas and in areas close to major Honduras 28 14 export-farming activities (as with coffee farmers). ElGSalvador 64 18 Guatemala's and Costa Rica's export production areas Costa Rica 23 14 for major vegetables, ornamentals, and root crops are within one hour of their capital cities. Guatemala's Source: HCA-FLACS0 1991: 86, 62. cardamom production is integrated with existing I coffee areas. This illustrates some form of complementarity or integration with preestablished services and facilities, which can be readily tapped without the additional burden of new investment. To accelerate development, new commodities require a full complement of services and facilities that require enormous investments. Although small investors may view new commodities as risky, large transnationals can be more daring since they are able to establish support structures such as roads, housing and villages, ports, communications, transport (shipping systems), and some human resource development in isolated locations, as with Del Monte's pineapple plantation in Costa Rica and Chiquita's oil palm plantations in Costa Rica and Honduras (Attenburg et al. 1990). Insufficient Expertise Even if technology were available and appropriate and all the facilities and services in place, the medium for new commodity development is still the entrepreneur and his or her staff. Without adequate human capital, little can be accomplished. Human resource development is vital, but who will provide the training? Local technology centers (mostly academia) may have some expertise, but their mandates do not give them enough flexibility, time, and resources for training entrepreneurs and their staff-a completely new clientele. This weakness is being remedied by developing centers for continuing education and technology transfer to channel in-house and borrowed expertise. Only major development programs with extensive government and private sector support have sufficient expertise and clientele for cost- effective training. (Most international research and development funds are for basic food crops: potato, rice, corn, wheat, cassava, etc.) Difficulties with the Export Market Bad experiences with export marketing have discouraged many new entrepreneurs from pursuing nontraditionals. Language, from a technical, business, and cultural perspective, has led to disagreements on quality consistency, adequate volumes, timely deliveries, payment schedules, and pricing arrangements (Dominguez and Sequeira 1990). Exporters and importers may find it impossible to guarantee deliveries and payments, respectively, due to climatic variability on the export side and to a changing consumer market on the import side. Misunderstandings arise, resulting in mistrust and even disillusionment. This has led to several approaches that have encouraged entrepreneurial families to establish their own brokerage operations in the market. Conversely, many importers have pursued backward integration to ensure their input supply lines. This is especially crucial in perishable commodities trading, which requires precision in determining planting and harvesting schedules, quality control, volume, and transport. - 100 - With their fully integrated systems, the multinationals are not hampered by this problem. Others must rely on good faith coupled with regular communication and constant supervision. Some institutions have representatives in the market and sell the service to smaller firms. This solution is still in its infancy, and its cost-effectiveness has yet to be determined. Inadequate Financial Incentives and Excessive Foreign Exchange Controls The presence of a financing program is an important feature in the promotion of nontraditionals and often draws more borrowers than there are funds. The constraint, however, is not availability, but inadequate financing packages and excessive foreign exchange controls. Thus, many credit funds are not utilized. Nontraditionals, including annual crops, often require longer development periods, new facilities, and a trial period before becoming viable. This is often overlooked in the financial packages offered; grace periods are therefore insufficient and interest rates stiff. The financial package estimates are based on the first crop; a slight miscalculation or deviation in output mars the image of the recently introduced nontraditional crop. The second or third crops, which can profitably gain and help recover the initial losses, are not given the chance to correct start-up problems. The tight control over foreign exchange is also a disincentive to TABLE 7. OFFICIAL AND BLACK MARKET RATES exporters who need dollars as an input to OF LOCAL CURRENCIES TO THE US$,1989 their enterprises. Dollars are particularly EL COSTA necessary because many new crops need RATE GUATEMALA SALVADOR HONDURAS RICA inputs that are not yet locally available. During the late 1980s, the central banks Black market 3.40 6.48 3.90 89.00 of Central American countries imposed Difference in controls that led to high black market value(%) 240.00 29.60 95.00 342.00 rates (table 7). Liberalization followed in 1989, but the tight control had driven Source: IICA-FLACSO 1991:104, 105. away investors in new enterprises that required imported equipment and materials such as tractors, irrigation motors, hormones, and planting materials to get underway. The Benefits of Diversification Diversification has several positive effects, some proven and some subjective perceptions. These include: * broadening and expanding the base of economic activity; * counterbalancing fluctuations in free market economies; * enhancing the elements of integration in resource utilization; * developing a healthy competitive atmosphere among industries to enhance efficiency; * expanding the job market alternatives of the community; * enriching entrepreneurial opportunities for business generation. How has diversification, particularly in export crops, benefited the ordinary worker, farmer, or family? The Guatemalan experience with the export of snowpeas shows that farmers received slightly higher incomes raising snowpeas rather than corn. The associated increase in labor requirements from snowpea production also meant that there was more employment for family members and the community. In Chile, - 101 - the effect was largely the same. For Chilean and Guatemalan producers, the gross margins per hectare increased more than tenfold (Von Braun et al. 1989; Lowell Jarvis 1991 [personal communication]), although these figures have been diminishing recently. The results of the Guatemalan study by Von Braun et al. (1989) found that the higher family income led to an increase in the family's net caloric intake and improved children's health and nutritional status. A side effect has been the improved yields of the traditional corn crop through rotation with the leguminous snowpeas, which resulted in residual soil fertility. The largest effect, however, was on farmers' confidence, as they became more entrepreneurial and daring in integrating export operations (see box 3). The study also found significant impacts on the economy, particularly the changing composition of employment and public services. Greater economic stability and a sense of equity have emerged because of lower unemployment and higher incomes. This healthy picture also averts the depredation of the physical environment, which is normally overexploited during hard times. The Guatemalan and Chilean experiences have engendered many other benefits: 1. increased economic productivity and efficiency. In the process of study and evaluation, many new products were tried and proved feasible. Thus, Guatemala added as many as 10 new lead crops, and Chile added as many as 12 new commodities for expansion; 2. surplus production and expansion of the economy. Surplus is a natural incentive for creative energies. Sufficient surplus can invite additional work and value added for processing low- cost products. Chile has increased its export income by 8 to 10 percent through the processing of surplus products; 3. an enriched quality of life. This concept is embodied in the capability to enjoy the variety of commodities on the market. Additionally, as one diversifies food consumption, there is less pressure on staple food crops, which in many countries are often in short supply. The Inherent Problems of Diversification and the Role of Government A poor evaluation of relative opportunities and advantages can lead to disastrous consequences for new investments. An example is annatto (Bixa orellana), or achiote, which was appealing to investors in Guatemala and Honduras. In 1989, annatto was viewed as an attractive substitute for the artificial red food coloring (Appropriate Technology International 1989). Unfortunately, the natural red color in annatto is normally degraded when heated and becomes yellowish. Thus, it is classified as a yellow colorant in commerce, where it joins many competitive colorants such as saffron, turmeric, and marigold. Conversely, an extensive study often means that many producers are already established. The younger cashew nut programs in Guatemala, Honduras, and Costa Rica, for instance, are not as competitive as the highly developed programs in India, Brazil, and Tanzania. Diversification is really a leap of faith for many investors in nontraditionals. This leap can be cushioned with a supportive government or a large private sector program, which can provide expertise, physical resources, and other facilities to nurture infant projects. The transnationals have provided this cushion in such projects as melons in Guatemala and Honduras (Chiquita International Brands), cucumbers in Honduras (Seaboard Corp.), and mangos and chayote in Costa Rica (Del Monte Corp.). Government can provide this cushion for medium- and small-scale agribusinesses by facilitating the acquisition of expertise, equipment, materials, and foreign investments. Diversification is also influenced by market imperfections, which are always looming as power players in both government and the private sector tinker with either the supply of or demand for major agricultural inputs (fertilizers, chemicals, petroleum, etc.). Artificial subsidies for favored and - 102 - noncompetitive sectors have led to disincentives for BOX 3. AN ENTREPRENEURIAL COOPERATIVE: sectors shouldering the GUATEMALA'S CUATRO PINOS subsidies. These imper- Successful small-farmer cooperatives are rare. Guatemala's Cuatro Pinos fections should be minimized, (Four Pines) Cooperative is not only successful, but its success has been in the with government promoting a risky business of nontraditional exports. Cuatro Pinos was organized after the shift to more viable 1976 earthquake that shocked the Guatemalan highlands. A nongovernmental nontraditional products. Swiss group began an assistance program that included food aid, literacy courses, Overall, agricultural reconstruction, institutional strengthening, and development of employment diversification has achieved opportunities. Meanwhile, ALCOSA, a firm exporting frozen and fresh some measure of success in vegetables, had been bought in 1975 by Hanover Brands, Inc., of the United States. It established buying stations in the highlands, a strategy that encouraged Central America and South small growers to cultivate exportable vegetables. Armed with a firm market outlet America's Pacific Rim and with funds from the Latin American Agribusiness Corporation (USAID countries. Although loans), ALCOSA joined with the Swiss group in assisting the highlanders. Cuatro diversification programs have Pinos was established from these forces and had a contract with ALCOSA. In declined at one time or 1980, three years into the contract, ALCOSA was swamped with production beyond its capacity. In 1981, Cuatro Pinos began exporting directly to various another in Ecuador, U.S. markets and discovered, not painlessly though, that it could grow into a Honduras, and Guatemala, major enterprise with its own processing and freezing plant, as well as marketing they resurge during system. Cuatro Pinos then went on to the European market. With the Caribbean downturns in the coffee and Basin Initiative, opportunities became more favorable. banana industries and when The cooperative has been operating for 14 years, and for the past 12 years has been a direct exporter of snowpeas, pod beans, broccoli, cauliflower, Imported petroleum prices and brussels spouts. This experience has encouraged the expansion of production rise. Other countries have in the Guatemalan highlands, including new products such as roses, baby squash, had sustained efforts, such as onions, garlic, and asparagus. The success of Cuatro Pinos has also resulted in Chile (see box 4), Costa the establishment of numerous enterprises, both large and small, and other Rica, and Colombia, where cooperatives involving raspberries, french beans, sweet bell peppers, celery, baby diversification programs were corn, baby carrots, escarole-endive, iceberg lettuce, peas, and strawberries. Additionally, an annual exposition, AGRITRADE, was established in 1987 and insulated from local political showcases Guatemalan and Central American products. changes and where there was a provision for maintaining the stability of the lead diversification institution. Colombia's FEDERECAFE, the Chile Foundation, and Costa Rica's National Banana Association (CORBANA) have been continuously operational for at least 16 years, encompassing research, extension, and enterprise development promotion. Moreover, in these countries, the government also set up parallel and linked diversification programs that shadowed (though did not duplicate) the NGOs and their programs; these include the National Institute for Agriculture and Livestock Research (NIA) in Chile and diversification programs in the ministries of agriculture in Costa Rica and Colombia. This pattern is being emulated in other Central America countries and in Ecuador. Governments should encourage the private sector to invest in future technologies. This could take the form of organized funding (for technology development centers) or in-house technical projects within private firms that could exert major entrepreneurial and economic impacts. A model of organized funding for a successful diversification institution is the cofounding of the Chile Foundation by the International Telephone and Telegraph Co. and the Chilean government, which each donated $25 million. The private sector offers sustained effort over a longer period than the terms of most governments in the region (four to six years). Support to technical institutions should be provided as early as possible (as in Chile); instability, resulting from changing political influences or rapid personnel turnover, impedes technology and enterprise development and should be minimized. Thus, while the private sector possesses the resources to pursue agribusinesses for diversification, the government can be a major supporter in developing favorable policies. New commodities may need special incentives and may require some special privileges normal to infant industries. Additionally, - 103 - because new commodities involve foreign-as well as BOX 4. AN AGRIBUSINESS BOOM IN CHILE local-investors who could bring technology, financing, Chile's current agribusiness boom can be traced back to 1962, when a and markets, governments national plan was drawn up by the government and supportive laws enacted to survey production potential, analyze markets, introduce and propagate seed will have to provide an materials, construct facilities, provide a credit line for orchard development, and attractive climate. establish a research facility (INIA). In 1965, the University of Chile and the University of California at Davis separately agreed on a technical cooperation program that would train Chilean technologists with Ford Foundation grants. Summary and Chilean technologists realized the large potential of transferring technology from California to Chile so that Chile could produce during California's off-season. On Recommendations their return to Chile, these technologists assumed positions in the dominant public sector institutions (e.g., INIA, CORFO [the national development corporation], Central America and and the University of Chile) that had been pursuing this scheme. South America's Pacific Rim In 1973, a military junta assumed power, and the new economic countries have been involved managers favored agribusiness entrepreneurs. Even with high interest rates, the costs of labor and capital were low, and the profitability of fruit cultivation rose in export diversification over dramatically. There was sufficient incentive to expand production and ample the past 40 to 50 years with technically skilled labor to support such production. The new economic mixed success. The management (the "Chicago boys") was able to borrow capital for Chile, and importance of a diversified investments rose rapidly. This expansion went smoothly, even as there were shifts economy, however, is well in technologies and also administrative problems. The expansion was fueled by accepted. Bananas and coffee some aggressive producer-exporters such as David del Curto, who convinced many Chileans to go into the fruit business. The INIA and CORFO technologists dominate exports, with beef, were encouraged to move to the private sector as the fruit business needed them. sugar, cotton, tobacco, and What was initially a fruit boom spilled over into vegetables and other several other commodities agribusinesses. In 1976, the Intemational Telephone and Telegraph Co. and the also exported. The climate, Chilean government cofounded the Chile Foundation, a private arm devoted to soils, and proximity to large developing high-value ventures by transferring to Chile the technologies to utilize its resources and productive capacity. The foundation has gone on to develop markets, however, favor the more agribusiness enterprises, including salmon industries, that are likely to production of many other surpass $1 billion in exports during the 1990s. commodities. The fruit boom almost ended in the 1980s, but a devaluation in 1983 Costa Rica, Chile, fueled another expansion. More fruits, vegetables, and aquaculture and forest and Colombia have been products have been added. Recently, multinationals have also entered the picture. successful due to sustained European, American, Middle Eastern, and Japanese companies have bought into many Chilean agribusiness corporations. Doubtlessly, this will fuel another export efforts in high-value crops boom. that are competitive in the market. These sustained efforts have been directed by the private sector, though always fully supported in many ways by governments, even as political administrations changed. Some countries have pursued a rapid transfer of technology to enterprises, whereas others have developed technology first and later developed the enterprises. Some countries are experimenting with a combination of these two approaches; it is too early to indicate results. International development institutions, such as USAID, the Canadian International Development Agency (CIDA), Japan International Cooperation Agency (JICA), and German Agency for Technical Cooperation (GTZ), have supported these diversification efforts. Since 1981, the Caribbean Basin Initiative has provided a philosophical base for this effort, which supports the private sector and NGOs. NGOs have proliferated, but their general stability and sustainability over the long term are still doubtful due to lack of sufficient financing that could be used to develop and transfer technology, especially for new commodities being introduced into a new environment. A successful diversification program is dependent on several factors, but early integration of efforts and long-term commitment are key to advancing technology and expertise, human resource and - 104 - enterprise development, infrastructure, and foreign investments. Diversification into high-value crops can improve family incomes, health, and quality of life (as in Guatemala), but real-wage increases are still elusive. This is where government has a role as a distributor of benefits and risks. Governments also have a special role in diversification through project investment and technology transfer. They can also provide incentives for infant industries through specially designed financing programs. Above all, governments should help all those seeking a shift to high-value commodities, rather than protect inefficient, noncompetitive, low-value industries. As the successful experiences illustrate, diversification requires a sustained effort, even in the midst of political change. This is best achieved by identifying organizations with long-term stability to pursue long-term objectives. Notes 1. Organizations involved in the promotion of nontraditionals in the late 1970s include, in Costa Rica, the National Council for Science and Technology, National Banana Association, Costa Rican Coffee Institute, Institute for Agrarian Development, Ministry of Agriculture, and University of Costa Rica; in El Salvador, the Salvadoran Institute for Coffee Research and the Cooperative League of the USA; in Guatemala, the Bureau of Agricultural Services, Guild of Rubber Growers of Guatemala, and Institute for Agricultural Science and Technology; in Honduras, the Honduran Coffee Institute and the Ministry of Agriculture's Diversification Project; in Nicaragua, the Ministry of Agriculture; in Panama, the Ministry of Agriculture; in Colombia, the Colombian National Coffee Federation; in Ecuador, the Ecuadorian Agricultural Foundation and the National Institute for Agriculture and Livestock Research; in Chile, the Chile Foundation and the National Institute for Agriculture and Livestock Research. 2. New organizations working on nontraditionals include the Costa Rica Coalition for Development Initiatives, Salvadoran Foundation for Agricultural Development, Honduran Federation of Producers and Exporters Associations, Honduran Foundation for Agricultural Research, Guild of Guatemalan Exporters, Association of Producers and Exporters of Nontraditionals of Nicaragua, Colombian National Coffee Federation, Ecuadorian Agricultural Foundation, and the Chile Foundation. These organizations are independent, but have some degree of coordination with other national and international development institutions. References Appropriate Technology International. 1990. 'Annatto Processing." Bulletin 21. Washington, D.C. Ardila, Jorge, Belen Arcila, and Hector Lopez. 1986. 'Cambio Tecnico en el Sector de Pequenos Productores Campesinos." In Martin Pineiro and Ignacio Llovet, eds., Antioquia: Colombia en Transicion Tecnologica y Diversificacion Social. San Jose, Costa Rica: IICA. Arnade, C., and David Lee. 1990. Risk Aversion through Non-traditional Export Promotion Programs in Central America. Washington, D.C.: U.S. Department of Agriculture, Economic Research Service. Attenburg, T., Wolfgang Hein, and Jurgen Weller. 1990. El Desaflo Economico de Costa Rica: Desarrollo Agroindustrial Autocentrado como Alternativa. Berlin: Institute of Latin American Studies, University of Berlin. Baumeister, Eduardo. 1991. "Desarrollo Agropecuario, Participacion Campesina y Diversificacion Agricola." In Raul Ruben and Govert van Oord, eds., Mas alla del Ajuste. San Jose, Costa Rica: DEI. BID (Inter-American Development Bank). 1987. Progreso Economico y Social en America Latina: Fuerza de Trabajo y Empleo. Washington, D.C. - 105 - Camacho, Daniel, and Manuel B. Rojas. 1984. La Crisis Centroamericana. San Jose, Costa Rica: FLACSO. CEPAL (Economic Commission for Latin America and the Caribbean). 1989. Anuario Estadistico de America Latina. Santiago, Chile. CMC (Central American Monetary Council). 1989. Boletin Estadistico. San Jose, Costa Rica. Dickson, John D. 1977. Check List and Uses of Plants in the Wilson Popenoe Botanical Garden. La Lima, Honduras: SIATSA. Dominguez, Luis, and Carlos G. Sequeira. 1990. "La Empresa Exportadora Centramericana: Evolucion, Estrategia y Desempeno." Revista INCAE 4(1): 11-22. Fundacion Chile. 1991. Manual del Exportador Hortofruticola. Santiago, Chile. IICA-FLACSO. 1991. Centro America en Cifras. San Jose, Costa Rica: IICA. IMF. 1989. World Economic Outlook. Washington, D.C. . 1991. World Economic Outlook. Washington, D.C. INIA (National Institute for Agriculture and Livestock Research). 1989. Informe 1989. Santiago, Chile. Kaimowits, David. 1991. Cambio Tecnologico y la Promocion de Exportaciones Agricolas no Tradicionales en America Central. San Jose, Costa Rica: IICA. Leonard, Jeffrey H. 1987. Natural Resources and Economic Development in Central America. New Brunswick, N.J.: Transaction Books. Martinez, Manlio, Luis Corrales, Emily de Alvarado, Tito Ordonez, and Salomon Ordonez. 1987. "Cambio Teenologico en la Agricultura de Centro America." Editorial Guaymuras, pp. 19-29 (Honduras). Mora-Urpi, Jorge, Anabelle Bonilla, Charles Clement, and Dennis Johnson. 1991. 'Mercado Internacional de Palmito y Futuro de la Exportacion Salvaje vs. Cultivado." Boletin Informativo (Serie Pejibaye) 3(1-2) (San Jose, Costa Rica: University of Costa Rica). OICD-USDA. 1985. Agricultural Market Handbook for Caribbean Basin Products. Washington, D.C.: U.S. Department of Agriculture. Rosengarten, F. 1991. "Wilson Popenoe." Honolulu, Hawaii: National Tropical Botanical Garden. Secretariat of Agriculture, Colombia. 1989. Antioquia Agropecuaria. Antioquia. Tabora, Panfilo C. 1989. Evaluation of 25 Crops with Potentialfor Honduras. La Lima, Honduras: FHIA. Von Braun, J., David Hotchkiss, and Maarten Immink. 1989. Non-traditional Export Crops in Guatemala: Effects on Production Income and Nutrition. Washington, D.C.: International Food Policy Research Institute. Weeks, John. 1985. The Economies of Central America. New York: Holmes and Meier. I DIVERSIFYING ASIAN RICE-FARMING SYSTEMS: A DETERMINISTIC PARADIGM Prabhu L. Pingali Declines in rice production profitability and farmer incomes have increased interest in crop and income diversification. Policy discussions on diversification have often been preceded by farmer initiatives to sustain their incomes by moving at least partially out of monocultural rice production into other crops or enterprises. Thailand, a predominant rice exporter, has exhibited the ASEAN region's highest levels of diversification out of rice production, and there has been smaller, though significant, movement out of rice monoculture in the Philippines and Indonesia. This paper views diversification out of rice monoculture into a multicrop enterprise system as an essential component of agricultural development, induced by the changing relative profitability of rice and nonrice enterprises. Diversification out of rice to nonrice crops will not always be profitable and will face both physical and economic constraints. This paper attempts to identify and evaluate these constraints for each of the major rice-growing environments. It ends by assessing research priorities for rice and rice- based farming systems, considering the relative profitability of rice production in each environment.' The Changing Profitability of Asian Rice Production The long-term profitability of rice production depends on three factors: the long-term price of rice, current and potential yields, and input costs. This section argues that the prospects for sustaining income primarily through rice monocropping are bleak, given low rice prices, stagnant yields, and high input costs. The Long-Term Decline in the Real Price of Rice Despite the recent increase in world rice prices, several analysts predict a downward trend in real price over the longer run. Figure 1 (adapted from Mitchell 1987a) shows trends in real world rice prices between 1900 and 1987 using 1964-66 as the base period. This graph shows an almost 50-year declining trend in real rice prices despite frequent and prolonged price fluctuations.2 Although many Asian governments protect domestic producers from international rice market fluctuations, these long-term trends are passed on to them in direction, if not in magnitude, with the relative profitability of rice production declining. Where alternatives to rice production are not easily available, this decline leads to a sharp decline in the welfare of rice producers. This trend can be arrested by reducing the unit cost of rice production or reallocating resources from rice to crop or noncrop nonrice enterprises. The unit costs of rice production can be reduced by increasing either farm yields or input use efficiency. Prospects for these options will be discussed below. Also discussed later in the paper are the prospects for crop and income diversification in predominantly rice monocultural systems. Prabhu L. Pingali is agricultural economist and leader of the Irrigated Rice Program, International Rice Research Institute (IRRI), Los Bafios, the Philippines. Figure 1. Trends in World Rice Prices, 1900-87 300 - 250 200- 150- 00 100 -~ ~ ~ ~ ~ ~~ ~~~~~~~~~~~~~~~~~~~~~~~~~0 50 - 0- l l l l 1900 1910 1920 1930 1940 1950 1960 1970 1980 1987 Year Source: Adapted from Mitchel 1987a. - 109- The Diminishing Yield Gap Recent evidence indicates that the technological yield frontier has stagnated and shows signs of long-term decline (Pingali et al. 1990; Flinn and De Datta 1984). Farm-level evidence indicates that farmer yields are catching up to the yield frontier and that further exploitation of the yield gap is not economical, as the incremental costs of further yield gains exceed incremental returns (Pingali et al. 1990). Figures 2 and 3 show the highest-yielding entries in the maximum yield trials for the wet season at the IRRI farm and at the Maligaya Rice Research and Training Center, respectively. These figures show that wet-season rice yields per hectare on the experimental stations have declined from a high of 6.2 tons in 1966-70 to a 1986-1987 level of 5.3 tons at Maligaya and 4.9 tons at IRRI. Similar declining rice yield trends have been observed in other experimental stations in India, Thailand, and Indonesia (Nambiar and Ghosh 1984; Gypmantasiri et al. 1980; International Network for Sustainable Rice Fertility [INSURF] 1987). The long-term decline in the irrigated yield frontier under intensive rice monoculture can be attributed to increased pest pressure, rapid soil micronutrient depletion, changes in soil chemistry brought about by intensive cropping, and increased reliance on low-quality irrigation water. Although the rice research system has been generating varieties with increasingly higher genetic yield potential, the rate of degradation of the paddy environment has been even greater (Pingali et al. 1990). Concurrently, farmer yields have caught up with those of experimental stations. IRRI's Economics Department has been following randomly selected farmers in Laguna (monitored over 1966-88) and Central Luzon (1970-88) (see Herdt 1987 for details). In Central Luzon the gap between the average sample farmer and the experimental station yield in 1970 was almost 4 tons per hectare in the wet season. Comparing the top third of the sample and the experimental station is even more striking: the gap was approximately 2 tons in 1970, diminished to less than half a ton within a decade-and in 1986 these farmers outyielded the experimental station by almost half a ton. The highest-yielding Laguna farmers show a similar pattern: the yield gap between the average sample farmer and the IRRI farm in 1984 was less than half a ton per hectare. The top third in Laguna started off with a 2.5 ton difference in 1965 and have consistently outyielded the IRRI farm since 1975. There is one important pattern in both the Laguna and Central Luzon samples that merits attention: although average farm yields have been rising, the top third yields in both cases have peaked and are declining. This is similar to the trend in the experimental station yields. An extrapolation of average farm yields shows that a similar peak and decline can be expected on those farms.3 Given current technology, the exploitable yield gap between the experimental station and farmer yields from irrigated lands in the Philippines is small, and long-term prospects are for stagnation or decline in average irrigated farm yields. Three other implications arise from this analysis: 1. The trend toward stagnant or declining yields, if widespread, questions the long-run sustainability of intensive irrigated monocultural rice production, as currently practiced in the Asian tropics. In this context, crop diversification needs to be examined in much greater depth as a mechanism to reverse rice yield declines in intensive systems; 2. If yield per crop stagnates, increased productivity and profitability are limited. Major attention must be focused on increasing annual crop production and income through intensification or diversification. This may involve accommodating additional rice crops every year or partially replacing rice with other crops or other enterprises, or both; - 110- Figure 2. Experimental and Farmers' Yelds (Wet Season), Laguna, 1966-87 7.0 6.0 5.0 4 .0 Hlghyielders 3.0 - , 2.0- 0 I , I I I 1966 1970 1975 1978 1981 1984 1987 Year Note: IPRR farm data are from the maximum yield trial, with N = 60 kg/ha. High yielders arc the highest- yielding 10 fanmers out of a sample of 35. "AU farms" is the sample average. Source: Ill Agronomy and Economics departments. Figure 3. Experimental and Farmers' Yields, Nueva Ecija, 1970-86 6 6-\ / \ NIRR~~~~A TC 5 4 - \/ 3 2- °- ' ~ ~ ~ ~~I I I I 1970 1974 1979 y 1982 1984 1986 Note: MRRTC maximum yield trial data with N = 60 kgba. High yielders are the highest-yielding 5 farmers out of 20. "Al farms" is the sample average. Source: IRRI Agronomy and Economics departments. - 111 - 3. If declining rice productivity becomes a long-term trend in Asia's rice bowls, rice production and supplies will be affected, and projections of a long-term decline in rice prices will no longer be valid. Degradation of Irrigation Infrastructure The degradation of irrigation infrastructure in Asia is contributing to the expansion of areas under nonrice crops. Since the mid-1960s the global growth rate of irrigated area has declined by about 60 percent; in Asia, it has declined by 72 percent (Rosegrant and Pingali 1990). This reflects a sharp reduction in irrigation investments, caused partly by the relatively favorable food security situation in Asia and by the collapse of world rice prices. The problem is exacerbated by poor maintenance of irrigation infrastructure, despite a relative shift in overall irrigation investment in the 1980s from new construction to rehabilitation and maintenance. An analysis of 92 irrigation systems in the Philippines shows that almost a third of them have declining trends in wet- and dry-season irrigated areas and yields (Masicat et al. 1990). Between 1979 and 1989, the wet- and dry-season irrigated areas in Luzon declined by 20,466 ha and 36,175 ha, respectively. Even in areas that continue to be irrigated, the quality of irrigation in terms of the amount and reliability of water supplied has deteriorated. Where irrigation water reliability is low, there is a strong case for dry-season crop diversification to increase the efficiency of water use and sustain farm incomes. Increasing Input Costs Per hectare costs of inputs-land, labor, purchased inputs, and supervision time-rise as a result of unit cost or quantity increases. Agricultural intensification, measured in the Asian rice context in terms of cropping intensity, leads to an increase in input use per hectare, per crop (Pingali and Binswanger 1987; Herdt 1987). Sustaining yields and soil fertility over time in rice monocultural systems with double- and triple-cropping requires greater levels of labor, fertilizers, other chemicals, and mechanical power than for single-crop systems. For the Philippines, a panel of 132 irrigated rice farmers in Nueva Ecija monitored in 1980 and 1988 showed that a 13 percent increase in yield per hectare required a 21 percent increase in nitrogen fertilizer, a 34 percent increase in seeds, and a 24 percent increase in hired labor. For Suphan Buri, Thailand, the average irrigated rice yields of 146 farmers increased by 6.5 percent between 1982 and 1988-at a cost of a 24 percent increase in nitrogen fertilizers, a 53 percent increase in pesticides, and a 35 percent increase in seeds. Yield increases of 23 percent for 71 West Java, Indonesia, farmers required increases of 65 percent in fertilizer and 69 percent in pesticides. Real returns to rice production were stagnant for each of the three countries during these periods (Pingali et al. 1990). Land values are positively associated with agricultural intensification (Binswanger and Rosenzweig 1986), and the population growth rates in these countries imply relatively higher opportunity costs of land and higher land rental values than other countries. Costs for hired and family labor are also expected to be higher as a result of increases in the opportunity cost of labor due to increased off-farm employment opportunities. Family labor, in addition to facing higher opportunity costs due to off-farm employment opportunities, has to provide greater levels of management and supervision. If rapid efficiency gains in the use of chemical inputs are not achieved, there could also be a significant increase in per hectare use of chemical fertilizers and pesticides. The Declining Long-Tern Profitability of Rice Production Low rice prices, declining or stagnant yields, and increasing input costs have steadily reduced the profitability of rice production (figure 4 [p. 1211 shows the Philippine situation) and farmers' incomes and welfare. The prospects for improvement are not encouraging. Improvement in the profitability of rice - 112 - production requires either a substantial increase in experimental station yields that would reverse a 20- year trend, or substantial increases in input use efficiencies. Sustaining and increasing the incomes of rice farmers will therefore depend largely on diversifying crops and income sources. This progression to crop and income diversification has occurred smoothly in countries where product markets operate relatively freely. In Suphan Buri, Thailand, for instance, the adoption of nonrice enterprises has been closely associated with rice price trends. Rice prices in Thailand were on a declining trend between 1980 and 1986, reaching their lowest level during 1985- 86. Between 1985 and 1988, 79 percent of a sample of 146 households adopted nonrice enterprises (table 1). The nonrice enterprises adopted included vegetables, fruit orchards, noncrop commodities such as shrimp and livestock, and off-farm activities such as rural industries or urban employment. By 1987, 91 of the 146 households had adopted diversified farming systems. A third of these households switched back to exclusive rice production in 1988 when rice prices rose after the 1987 drought. The process of diversification has been slower in the Philippines and Indonesia where rice profits were buffered to a greater extent by government intervention. Opportunities for Diversification out of Rice The potential for diversification out of rice production depends on both physical and economic factors. The feasibility and cost of substituting other crops vary across the region's five basic agricultural ecosystems: irrigated and rainfed lowlands, tidal and deep-water wetlands, and uplands. Each of these systems also presents different rainy- and dry-season profiles and requires different levels of physical and human capital investment to switch from rice to nonrice crops and back. Nonrice crops are grown year-round in Indonesia in a sorjan system that requires high levels of investment in drainage control. (The sorjan system is a ditch with a flat bottom and notched edges. Rice is usually planted on the ditch, and upland crops may be grown on the flatbeds between ditches.) Flexibility is low because moving out of monocultural rice to upland crop production on elevated dikes or back into monocultural rice production involves high physical investment. Upland areas, however, can switch between rice and nonrice crops with minimum additional investment. Table 2 presents the flexibility of crop choice by ecosystem and season. Wet-season crop flexibility is low in all but the upland environments because investment requirements for drainage are high and not easily reversible in the lowlands. Switching between rice, maize, and other crops is possible in the uplands where fields are not bunded and do not require puddling before crop establishment. During the dry season, crop choice is constrained by water availability and drainage. The irrigated lowlands have the most reliable water supply. These areas, depending on the severity of the drainage constraint, have the highest flexibility in dry-season crop choice. Switching from dry-season rice to nonrice crop production involves investing in temporary drainage structures and learning about nonrice technology, cultivation practices, and irrigation water management. Onion farmers in the Philippine Upper Talavera River Irrigation System (UTRIS), for example, build multipurpose ditches and levees in their rice paddies to drain excess water (Pingali et al. 1989). Heavy soils require elaborate drainage structures, whereas light sandy soils may not require drainage. The returns to these investments are highest for the irrigated lowlands with moderate-to-well-drained soils, and these areas tend to diversify more than the other ecosystems as the relative profitability of nonrice crops improves. Irrigated lowland soils can be classified as well, moderately, or poorly drained. Table 3 shows crop choice flexibility for each of these soils by season. Only well-drained soils have potential for nonrice crop production during the wet season, as other soils require investments in bed and furrow or sorjan systems. Dry-season crop choice in irrigated rice lands is high for all but heavy waterlogging-prone irrigated rice soils that require prohibitive drainage investments. Irrigated areas in South and Southeast - 113 - TABLE 1. HOUSEHOLD-ADOPTED NONRICE ENTERPRISES BY ENTERPRISE AND YEAR ADOPTED, SUPHAN BURI, THAILAND YEAR FIRST ADOPTED (NO.) TOTAL TYPE OF PRE- HOUSE- ENTERPRISE 1980 1980 1981 1982 1983 1984 1985 1986 1987 1988 HOLD Cattle 2 0 1 0 2 3 2 6 5 13 34 Poultry 4 0 0 3 2 1 2 4 1 2 19 Prawn & fish 0 0 0 0 0 0 0 2 3 0 5 Vegetables 0 0 0 2 1 0 5 10 13 5 36 Fruit trees 0 0 0 2 1 0 1 1 0 1 6 Seasonal crops (short periods) 0 0 0 0 1 1 3 6 8 2 21 Sugarcane E 0 0 0 0 0 0 1 2 7 3 13 Sugarcane C 0 0 0 0 0 0 0 0 5 2 7 Off-farm work (avg.) 0 0 0 1 0 1 0 0 0 0 2 Total 6 0 1 8 7 6 14 31 42 28 143 Price of paddy (baht/ton) - _ n/a 2,470 2,415 2,273 2,230 2,398 3,122 3,726 Note: Sample constitutes 146 households. Source: Sriarunrungreauang 1989. Asia with dry-season diversification have all TABLE 2. FLEXIBILITY OF CROP limited their nonrice crop production to well- CHOICE BY ECOSYSTEM AND SEASON drained soils, while intensive rice production has SEASON continued concurrently on poorly drained soils. ECOSYSTEM SET DRY The period of irrigation water availability is also an important determinant of dry-season Irrigated lowlands Low Moderate to high' Rainfed lowlands Low Low to moderateb diversification. Partially irrigated areas that cannot Deep-water and tidal support a dry-season rice crop have a natural wetlands Low Low to moderate advantage in diversifying into upland crops during Uplands High Moderateb the dry season. But crop choice may again be limited by heavy, poorly drained soils, in which a. Includes the postrice period late wet season) or the prerice water control to avoid waterlogging or drought is period (dry-wet transition). difficult. The relative speed of diversification of b. Conditional on rainfall level and distribution. rice lands in Suphan Buri took two forms, dry-season diversification and year-round diversification. Dry-season diversification involved vegetables and other seasonal crops, such as maize and sweet potato, and 39 percent of households adopted a dry- season nonrice crop. Land investment requirements for establishing these crops are minimal, and when the rice price improved in 1988 these lands quickly returned to rice production. Year-round diversification involved sugarcane, shrimp and fish farming, and fruit orchards. Investment requirements for year-round diversification out of rice are high and are made only if prospects for relative long-term profitability are favorable. For fish and shrimp production, for instance, initial investment costs are about 110,000 baht ($4,400) per hectare, with input and labor requirements high. Table 4 provides data on the relative input requirements and the profitability of rice and nonrice enterprises. Sriarunrungreauang (1989), using the above panel data for Thailand, found that if the rice price drops by 20 percent, dry- season nonrice crops will be relatively more profitable than rice, but year-round diversification will not - 114 - be a profitable alternative to rice in the irrigated lowlands. TABLE 3. FULLY IRRIGATED LOWLANDS: Opportunities for dry-season FLEXIBILITY OF CROP CHOICE diversification in rainfed lowlands and WELL- POORLY deep-water areas are limited by water DRAINED MEDIUM DRAINED availability for postrice crop production. SEASON SOILS DRAINAGE SOILS In the humid and subhumid zones rainfall Wet Moderate LOW LOW level and distribution are such that a post- Dry High High Low to moderateb or prerice crop in the rainfed lowlands is possible. Postrice cropping of legumes, a. Conditional on rainfall levels and effective water control. cereals, or vegetable crops may be b. Depending on the level of drainage investments. possible with late-season rains and residual moisture. This practice has become much more feasible on rainfed rice lands that produce earlier-maturing rice cultivars harvested before the onset of the dry season. In the Cagayan Valley of the northern Philippines, the replacement of traditional rainfed rice varieties with a six-month duration with early maturing modem varieties has led to rice double-cropping in lower elevations and the introduction of a prerice crop of mungbean on the upper elevations (Garrity et al. 1988). Prerice crops in the lower elevations are only possible on ridges that prevent waterlogging (Pernito and Garrity 1988). The strategy of increasing cropping intensities in the rainfed lowlands will only be successful if modern rice varieties adapted to problem hydrologies, including drought and flood, are available. In rainfed environments where there is a sharp and prolonged dry season (especially the semiarid zones) postrice crops are not possible without supplementary irrigation. In the rainfed lowlands of South Asia, northeast Thailand, Cambodia, and Lao People's Democratic Republic, dry-season crops cannot be grown on residual moisture. There is potential, markets permitting, for short prerice crops followed by short-duration rice crops. Where supplementary irrigation is available, as with pumps, opportunities exist TABLE 4. RELATIVE ANNUAL INPUT REQUIREMENTS for a dry-season rice or nonrice crop. In AND PROFITABILITY OF RICE AND NONRICE Nueva Ecija, Philippines, where there is a ENTRPRISES, THAILAND, 1988 (BAHTIHA) six-month dry season, the introduction of ENTERPRISE deep tubewells has led to a rotation of maize RICE- RICE- SUGAR- followed by mungbean in the dry season after RICE VEG. CANE PRAWN a rainfed, wet-season rice crop (Gines et al. irnput 1988). This diversification occurred only on Labor (man-day/ha) the upper paddies with light-textured and Farnily 42 595 17 90 Hired 41 445 60 2 easily drained soils. The lower paddies with Total 83 1,040 77 92 heavy-textured soils prone to waterlogging Labor 5,739 71,916 5,325 6,362 were kept in dry-season rice. While two rice Fertilizer 2,915 27,174 995 627 Pesticide 964 19,224 280 - crops are also possible on upper paddies, with Other costs 294 17,037 8,389 18,589 the dry-season crop irrigated by pumps, Feed - - - 42,049 private and social returns to a diversified Total costs 9,912 35,351 14,989 67,627 cropping system outweigh those of the Gross returns 28,427 160,517 32,399 104,485 rice/rice cropping system, primarily because Net returns 18,515 25,166 17,410 36,858 the costs of irrigating rice are higher (Gines et al. 1988). In the semiarid tropics of India, Engelhardt (1984) reported the emergence of Sourc: Sranrungrauang 1989. a diversified cropping system with the introduction of deep well pumps. Rainfed rice in the wet season is followed by either groundnut, sorghum, or vegetables. - 115 - Dry-season diversification in upland areas similarly depends on the level and distribution of rainfall. In areas with a sufficient growing period, a postrice crop can be grown. Maize, sweet potato, and vegetables are common sequential crops. In northern Mindanao, Philippines, for instance, where the average annual rainfall of 2,350 mm is evenly distributed over eight months, double-cropping of maize is practiced on a quarter of the upland area (Mandac et al. 1987). For the lower-rainfall upland areas in much of the subhumid and semiarid zones, rice production is generally not profitable because of the risk of drought stress. Where irrigation is not available, wet-season sorghum, millet, and pulses such as pigeonpea and chickpea are commonly grown. Diversification out of rice in response to changes in the relative profitability of rice and nonrice crops is most feasible in the dry season. The rice ecosystems in which this diversification is most profitable and feasible are the irrigated lowlands, because of greater water supply reliability and higher returns to diversification investments. Diversification Constraints The profitability of diversification is constrained by both markets and physical infrastructure. If market access is good, output demand is relatively elastic, and returns to investments in land, education, and technology are relatively higher. Table 5 shows the physical and market constraints to diversification in the irrigated lowlands. In the dry season, in areas with good market access, the profitability of diversification is high on well-drained soils and moderate to low on poorly drained soils. In areas with poor market access, the profitability of diversification on well-drained soils is moderate to low, depending on the nature of output demand. If demand is highly inelastic-due perhaps to the high cost of transporting output to markets-the profitability of diversification is low. For poorly drained soils with poor market access, the profitability of diversification is very low. Table 6 shows similar data for upland soils according to their susceptibility to TABLE S. MARKET INFRASTRUCTURE VERSUS erosion. If market access is good, the PHYSICAL CONSTRAINTS AS A DETERMINANT profitability of diversified field crop OF DIRSIFICATION'S PROFITABILI production on soils not highly susceptible to SOIL erosion is high. For soils susceptible to ACCESS WELL-DRAINED POORLY DRAINED erosion, profitability of field crop production Good market High Moderate to low' is determined by the erosion control Poor market Moderate to lowb Very low investments required. Where high levels of erosion control investments are required, tree Note: For irrigated lowlands, dry season. crops may be a more viable option than field a. Conditional on the level of investment requirements for drainage crops, particularly after land degradation has control. been allowed to occur through field crop b. Conditional on input supply conditions. production. In upland areas with poor market access the returns to diversification out of subsistence rice production are limited in areas of either type of soil. The relationship between the flexibility of crop choice and erosion control investments becomes pronounced on sloping uplands, which are extremely susceptible to erosion. There are various options for erosion control to maintain permanent cropping, ranging from grassy strips to stone wall terraces. Farmers' choice of an erosion control strategy depends on population pressure on the land, market access, and the erosion control techniques available. Pingali (1990) and Fujisaka and Garrity (1988) argue that farmer interest in erosion control measures is directly related to land values and market access and is conditional on the availability of suitable technologies. - 116 - Options for Sustaining Incomes TABLE 6. MARKET INFRASTRUCTURE VERSUS During the wet season, rice PHYSICAL CONSTRAINTS AS A DETERMINANT OF will continue to be the dominant DIVERSIFICATION'S PROFITABILITY source of income in all but upland SIUOUS SOIL environments. Table 7 shows the CHEMICAL' dominant income-generating activities CONS/RAINT WMAHOUT for each season and environment. ACCESS EROSION HAZARD CONSTRAINT Empirical evidence on the sources of income by rice environments is Good market High-input, Diversified farming provided in tables 8 and 9 for the cropping systems Philippines and Thailand, oragroforestry respectively. This does not imply that systems rice is not an important source of Poor market Shifting Subsistence income for the uplands, but rather cultivation cropping stresses that the uplands have always systems been very diversified. In the irrigated lowlands, dry-season rice will Note: For uplands, wet season. continue to be the major source of a. Includes highly acidic soils with potential aluminum toxicity/P deficiency. income. Areas with good market access and those near urban centers increasingly diversify to nonrice crops and vegetable production. The dominant dry-season activities in the rainfed lowlands and deep-water areas are off-farm employment, livestock production, and cottage industries. There is scope for postrice crops on residual moisture or prerice crops during the early wet season. The share of total income from these activities, however, will be relatively low. Dry-season cropping activities in rainfed areas are limited by technical problems related to timely and effective crop establishment, limited (or sometimes excess) moisture, and generally modest yields and high yield instability. Off-farm activities are often more dependable income sources, suggesting that dry-season cropping intensities will remain low even if technical problems in crop production are solved. The above discussion leads to the conclusion that irrigated environments, although they have an absolute advantage (relative to the other environments) in a rice/rice cropping pattern, may also have a comparative advantage in a rice/nonrice cropping pattern. The extent of comparative advantage for the irrigated lowlands in dry-season diversification depends on the physical constraints and the market opportunities for nonrice crop production. During the wet season, upland environments have both an absolute and a comparative advantage in nonrice crop production. Dynamics of Crop Diversification Farrner Land Preferences Within an irrigated microenvironment, lands with the greatest potential for rice production are heavy clay soils and those that have the best access to irrigation water in the head section and paddies close to irrigation canals.4 As the relative returns to dry-season nonrice crops rise, there is an increased preference for lands that are marginal vis-a-vis rice production and on which investment requirements for drainage are lower. Wardana et al. (1990) documented differences in yields and net returns for rice and nonrice crops for the Cikeusik Irrigation System in West Java, Indonesia, finding the relative profitability of nonrice crops higher on lands further away from the head of the system, to a point where water scarcity is a problem. Pingali et al. (1989) documented similar cropping pattern choices for the UTRIS - 117 - project in the Philippines. Two crops of rice are grown on the upper TABLE 7. DOMINANT CROP AND NONCROP portions of the system, with onions, OPTIONS FOR SUSTAINING INCOMES, BY ENVIRONMENT chilies, and vegetables common in the ENVIRONMENT WET SEASON DRY SEASON middle section. Dry-season crop choices at the tail of the system are Irrigated lowlands Rice Rice/nonrice crops conditioned by reliability of water Rainfed lowlands Rice Off-farm employment Deep-water and tidal supply: where farmers have access to wetlands Rice Off-farm employment pumps, nonrice crops are grown. Uplands Diversified production systems In the irrigated lowlands, when dry-season returns to nonrice crops dominate the returns to TABLE 8. SOURCES OF HOUSEOLD NCOM E 1 riceprodctio, deand WITH RICE FARMS BY ENVIRONMENT, PHILIPPINES, 1988 rice production, demand and price are highest for RAINFED FARMS UPLAND FARMS IRRIGATED FARMS lands with the fewest % OF NO. OF % OF NO. OF % OF NO. OF INCOME TOTAL HOUSE- TOTAL HOUSE- TOTAL HOUSE- constraints to diver- SOURCE INCOME HOLDS INCOME HOLDS INCOME HOLDS sification out of rice. Over the past five years Rice 48.1 48.0 3.0 39.0 56.7 129.0 the Phlippin UTRIS Nornice 6.3 41.0 31.0 39.0 0.4 1.0 the Philippine UTRIS Noncrop 9.0 39.0 16.0 50.0 6.5 60.0 project has witnessed Off-farm 1.0 20.0 10.0 28.0 10.2 35.0 dramatic changes in Nonfarm 35.6 53.0 40.0 15.0 26.2 95.0 preferences for dry- Total value preferences for dry- of incomea 22,748.0 49 0b 15,777.0 54.0" 33,975.0 132.0b season cultivation land and land values (Pingali et al. 1989). The system a. In Philippine pesos. b. Total sample number for each category; some households have two or more sources of income. consists of areas of Source: Social Sciences Division, IRRI. heavy clay and sandy loam soils that are more suitable for dry-season onion and vegetable production. In the last five years land preferences have switched from the heavy clay soils to the sandy loam soils, and prices have followed suit. Credit, Labor, and Risk Constraints to Diversification Switching from rice monoculture to diversified farming requires substantial start-up investments and operating expenses that require long-term and seasonal credit arrangements. Where diversification has occurred successfully, farmers have managed to acquire credit through private or public sources. In the Philippine UTRIS project, the main alternative to dry-season rice production is onions. The credit constraint to onion production has been alleviated by arrangements with onion traders who provide credit for inputs in exchange for exclusive rights to purchase all output at the market price. No interest is charged for this credit, but traders benefit considerably from substantial price increases between the harvest and postharvest months that more than offset foregone interest charges and storage costs. Similar credit arrangements have been observed for vegetable and sugarcane production in Suphan Buri, Thailand, where longer-term credit is also provided by the government and agricultural cooperatives. Diversified cropping increases labor requirements relative to rice. Labor for temporary drainage structures is an essential activity immediately following rice harvest. Planting, weeding, harvesting, and postharvest operations are also extremely labor-intensive for these crops. Recent research by the International Irrigation Management Institute (IIMI) in the Philippines estimated the mean labor demand for rice, mungbean, onion, and garlic as 85.7, 68.7, 468.5, and 241.0 man-days per hectare, - 118 - TABLE 9. NET HOUSEHOLD INCOME BY PROVINCE, THAILAND, 1980-81 SOURCES OF NET HOUSEHOLD INCOME (BAHT) REGION NO. OF NO. OF NONFARM AND PROVINCE VILLAGES HOUSEHOLDS FARM OTHER WAGE OTHER TOTAL Northeast Khon Kaen (rainfed) 8 141 13,275 3,385 6,627 4,713 28,000 (47.4) (12.1) (23.7) (16.8) (100.0) Roi Et (rainfed) 5 75 4,889 6,047 5,514 5,404 21,854 (22.4) (27.7) (25.2) (24.7) (100.0) North Chiang Mai (upland) 9 163 6,046 10,629 11,417 4,095 32,187 (18.8) (33.0) (35.5) (12.7) (100.0) Center Suphan Buri (irrigated) 3 42 29,232 -409 9,027 3,461 41,311 (70.8) (-1.0) (21.8) (8.4) (100.0) All provinces 25 421 10,643 6,284 8,544 4,481 29,952 (35.5) (21.0) (28.5) (15.0) (100.0) Note: Percent of total in parentheses. Source: Onchan and Chalamwong, forthcoming. respectively. Labor requirements for nonrice crops are higher at the head of the system where drainage investments are greater (Wardana et al. 1990). Given the higher crop and drainage labor requirements, nonrice crops on irrigated lands are grown on extremely small plots, in general about one-fourth the paddy area. Does diversified cropping aggravate labor peaks between the harvest of the rice crop and the planting of the nonrice crop? As discussed above, additional labor is required to build temporary drainage structures, and additional labor or mechanical power is also required for land preparation. Land preparation for nonrice crops following rice requires breaking the compact soil surface caused by puddling to allow root penetration by a nonrice crop (Zandstra 1992). Power requirements for breaking up this "hardpan" are higher on heavy clay soils than on the lighter soils. Although mechanization can alleviate this labor peak, the machine power required for upland crops is substantially greater than for puddling rice paddies. Contract hire operations can be used to fill this gap, but are only profitable when large areas are grown to nonrice crops. Expansion of nonrice area is also constrained by the nature of the output market, the supply of labor, the prevalence of credit contracts, and farmer aversion to production and price risks. In addition to crop labor requirements, the supervision time required of the farmer is significantly higher: this may be the dominant labor constraint to high-value, nonrice crop production given the highly inelastic nature of management labor available in the farm household compared with hired labor augmented by seasonal migrants. In the Philippine UTRIS project, onion farmers overcame the supervision constraint by dividing their farms in half, cultivating one part and renting the other to seasonal tenant farmers. Unlike the case of rice, price risks outweigh production risks in nonrice crop production, and the system's seasonal tenancy arrangements are a method of diffusing the price risks associated with nonrice crop production. This way landowners receive a certain income from part of their land and gamble on the remainder. - 119- Collective Water and Land Management Irrigated environments with diversified cropping patterns require collective action to ensure adequate water supplies, regulate their timing, and prevent flooding of nonrice crops. Water user associations have similar objectives and face similar operational constraints. The main problem with organizing a viable association is that farmers at the head of the system with better water access do not have as much incentive to join as farmers at the lower parts. But farmers at the tail who join find their access to water improves only marginally, since inefficient water use and water stealing by the head farmers continue. Thus, only farmers in the midsection of irrigation systems benefit significantly from the formation of water associations. In the Philippine UTRIS project, rice-growing farmers in the upper portions of the system have adequate water supply during the dry season, little need for in-season regulation of timing, and no excess water problems. Several attempts to organize into an irrigators' association have failed. Farmers in another area grow exclusively nonrice crops (onions) during the dry season. In-season regulation of the timing of water supply differs for onions and is important. Water flow has to be regulated to prevent excess water in the onion fields. This need for collective action has led to success in organizing into an irrigators' association. Collective action, although desirable, may not always be feasible. In the Philippine UTRIS project, the onion-growing farmers at the tail found that forming an association did not result in increased water allocations to their farms because there was not enough dry-season water to service them. After two years these farmers stopped paying membership fees to the association and began depending exclusively on pumps for meeting their water needs. Collective land management action is equally important in upland areas. In the uplands, group action for making watershed-level investments for erosion control is essential to developing long-term sustainable cropping systems. Sloping land management systems in the Philippines and terraces in West Java are examples of such collective efforts. In the irrigated lowlands, crop choice decisionmaking requires consensus on which crops to grow at the system or lateral level. Without such consensus, the ability of farmers to influence system management to change water allocation rules for nonrice crops will be limited. Finally, land tenure security is crucial for making the long-term land investments required for diversification from rice to nonrice crops and enterprises. Formal land ownership also helps farmers acquire credit. Evidence on land ownership and investment is provided for Thailand by Feder and Onchan (1987) and Chalamwong and Feder (1986). Research Priorities Given current technology, farmer crop management practices, and the long-term decline in real rice prices, the profitability of rice production is expected to continue to decline. Rice farmers will continue to face pressure to seek alternative income opportunities. Sustaining the profitability of rice production in the face of competing opportunities for resources will require farmer access to technologies that increase yields, input efficiency, or crop intensity to reduce the unit cost of rice production. There are assorted prospects for supplying technologies in each of these categories for the various rice-growing environments. Irrigated Lowlands Yield sustainability. In the short to medium term, understanding the causes of the decline in experimental station rice yields is essential to arresting and reversing the trend. If the trend toward - 120 - declining yields is not reversed, the implications for future national production trends and the economic viability of rice cultivation are serious. Perhaps this issue, which has not received significant research attention to date, must rank as high as the traditional focus on increasing yield ceilings. Long-term research will, of course, concentrate on breaking the current yield ceiling. The relevance of a higher yield ceiling, however, is conditional on crop husbandry techniques that can sustain these yield gains. Sustaining current yields requires identifying optimal crop management techniques and understanding the net effects of the interactions of the component technologies. It is unlikely that there will be one general prescription for achieving incremental yield gains. Rather, the answer might differ by location. This highlights the importance of close collaborative research between national programs and IRRI in sustaining the yield gains already achieved. Research into appropriate crop management techniques should investigate the comparative long-term productivity of continuous cropping of rice versus alternative rice-based cropping patterns. Providing break crops in a rice cropping system helps maintain or regenerate soil fertility, reduces weeds and pest buildup, and provides more diversified options to sustain farm incomes (Westcott and Mikkelson 1988). Grain legume crops such as mungbean or cowpea, leguminous green manures, or vegetable crops may be particularly suitable rotation crops with rice. Wheat and maize are popular rotation crops, but are also nutrient- demanding cereals. There is concern that yields in rice/wheat rotations are also declining in some areas. Input use efficiency. Input-saving technical change can take several forms, including integrated TABLE 10. RICE FARMS SWITCHING FROM pest management, integrated nutrient management, TRANSPLANTING TO DIRECT SEEDING, DRY direct seeding techniques in place of transplanting, SEASON 1980-88, PHILIPPINES AND THAILAND and more efficient water use. (Tables 10 and 11 1980 1988 % CHANGE provide examples for direct seeding.) For instance, pesticide use in Laguna declined significantly over Philippinaes 153 71 -54 1984-87 without reducing yields. The average Direct seeding 14 112 700 number of applications per season dropped from three to two, and the average dosage per application 1982 1988 % CHANGE also declined. Integrated management research that critically addresses the contribution of every Thailand production factor to the overall cost and productivity Transplanting 75 19 -75 can often substantially reduce input costs while Direct seeding 71 127 79 maintaining yields. Figure 5 shows how production costs were reduced in Colombia by the equivalent of 1.2 t/ha without affecting yields, with dramatic effects on the profitability of rice production. Such work may be essential in countries where there is a real concern that rice production levels cannot be maintained if the profitability of rice remains low. Given the decline in real rice prices and stagnant yields, an incentive exists for farmers to adopt efficient input use technologies. The growing importance of off-farm income and other nonrice activities, resulting from increasing income diversification, also makes labor-saving technologies attractive. Technologies that provide efficiency gains, however, are knowledge-intensive and require substantially greater levels of farmer judgment and supervision. The generation and adoption of these technologies would consequently require high levels of national program involvement. Diversified cropping patterns in irrigated environments that maximize the carryover effects of inputs from one crop to the next can increase the efficiency of input use. The common example is rice/legume systems that allow for lower levels of nutrient application for the subsequent rice crop. In wheat/rice systems, the phosphorous applied to wheat is efficiently available to rice (since phosphorous availability increases in the flooded soil). Also, rice/break crop systems are available for reducing pesticide demand for a subsequent rice crop. An issue that has not received sufficient research interest is the optimization of input use over the entire cropping pattern rather than on a per crop basis (Kundu - 121 - Figure 4. Changes in Net Revenue, Central Luzon and Laguna Farms 6% ;laguna 58% ,costsX is7\ 1970 1987 7\ Central A8 Luzon 1970 1986 Figure 5. Savings in Production Costs in Irrigated Rice by Changing from Conventional to Improved Crop Management, 1986 Conventional Management Improved Management 9.0%644 Pests 6.8% 1.7% 1 457edsol27weed 10.7% Seeding 5.4% 10.7% ~~~~~~~~6.4% Total costs = 5,162 P/ha Tlbtal costs = 3,994 P/ha; average yield = 5,234 kg4ha. average yield = 5,234 kg/ha. Source: CIAT 1987. - 122 - and De Datta 1988). In diversified agricultural TABLE 11. COMPARATIVE INPUT USE, TRANSPLANTED AND systems, this ought to be DIRECT-SEEDED RICE, DRY SEASON 1988, NUEVA ECUA farmers' strategy. DIRECT- % INPUT TRANSPLANTED SEEDED CHANGE Rainfed Lowlands No. of sample (parcel) 71.00 112.00 58 N (kg/ha) 95.21 103.20 8 Increasing yields. The P (kg/ha) 27.01 20.06 -26 rainfed lowlands are diverse, K (kg/ha) 22.03 13.22 40 but generally rice yields and Seed (kg/ha) 196.04 161.62 -18 Pesticide (P/ha) 551.00 708.00 28 further intensification in these Yield (kg/ha) 4,966.57 4,963.83 - environments are constrained Labor use (man-day/ha) by production instability from Farnily 24.74 14.09 -43 Hired 70.83 36.02 -49 highly variable field water Total 95.57 50.11 -48 regimes. Drought, submer- gence, and prolonged water- logging seriously affect rainfed - Negligible. lowland rice. To raise yields it is essential to introduce technical innovations to alleviate these constraints. The development of more stress-resistant cultivars can significantly improve yield stability, and the increased submergence tolerance of late-generation cultivars is encouraging. Increasing input efficiency. A more holistic diagnostic approach is required in research on rainfed rice to accelerate yield improvement and reduce yield variability. Improved crop and water management practices are essential to achieving and sustaining higher yields. Input use efficiency in the rainfed lowlands, however, is closely related to the reliability of the water regime. Where the reliability of water supply is low, the efficiencies of input use will necessarily be low. Increasing cropping intensity. Less than 20 percent of the rainfed rice area is cropped to something other than a single rice crop per year, and intensification of land use is being induced by population densities and market demand. The real potential for increasing cropping intensities lies with nonrice crop production. Two supply side constraints to crop intensification in the rainfed lowlands that deserve special attention are poor crop establishment practices for the pre- and postrice crops and insufficient qualitative understanding of the competition for labor between crop and noncrop activities during the non-rice-growing season. Pingali (1987) provides an example of the latter for northeast Thailand, where attempts to encourage a prerice green manure crop are hampered by the high cost of foregone wages from off-season work in Bangkok. Uplands In the Asian uplands rice is grown primarily as a subsistence crop. Little upland rice is marketed because upland farmers use a wide range of other crops for cash income, and the relative profitability of rice is low. These factors provide a backdrop to the unique research imperatives for the upland farming systems in which rice is grown. Upland rice yields are highly unstable due to drought, blast disease, weeds, and other stresses. Rice yields are often unsustainable due to production on highly acidic, erosive soils that drastically lose their production potential after a few years. Technology development for upland rice must be directed primarily to stabilizing and sustaining yields as it attempts to improve them. The development of ecologically sound yet economically attractive crop rotations within which upland rice - 123 - is produced will be a major vehicle for meeting these objectives. Upland rice research in most environments ought to be conducted within a framework of an overall land management strategy that includes investments in erosion control at the farm level. Diversified cropping in the uplands will be an essential component of a viable and sustainable upland farming system. Notes 1. Panel data sets for the Philippines, Thailand, and Indonesia collected by IRRI's Social Sciences Division for 1980 and 1988 were used to examine changes in farmer crop and noncrop enterprise choices over time and to examine the changing profitability of rice versus nonrice enterprises. These data sets were complemented with data from other published sources to provide an Asia-wide perspective for the conclusions reached. 2. The major causes of the long-term decline in rice prices are discussed by Mitchell (1987a, 1987b); Mitchell and Duncan (1987); Schuh (1987); and David (1987). 3. See Pingali et al. 1990 and Pingali and Moya 1989 for further details on yield comparisons between experimental stations and farmers' fields. 4. Yields almost always decline from the head to the tail of the irrigation system. (See Chambers 1988 for a review of South Asian evidence. Also, Pingali et. al. 1990 provide empirical evidence for Southeast Asia.) Incomes and net returns to labor decline more sharply than yields (Chambers 1988). References Bayacag, P.G. 1989. "Comparative Economic Analysis of Diversified Crops under Irrigated Condition and Their Performance versus Irrigated Rice." In Proceedings of a National Workshop on Crop Diversification in Irrigated Agriculture in the Philippines. Colombo, Sri Lanka: International Irrigation Management Institute. Binswanger, H., and M. Rosenzweig. 1986. "Behavioral and Material Determinants of Production Relations in Agriculture." Journal of Development Studies 22: 503-539. Binswanger, H., and P.L. Pingali. 1988. "Technology Priorities for Farming Sub-Saharan Africa." World Bank Research Observer 3(1): 81-98. Chalamwong, Y., and G. Feder. 1986. "Land Ownership Security and Land Values in Rural Thailand." World Bank Staff Working Paper 790. Washington, D.C. Chambers, R. 1988. Managing Canal Irrigation: Practical Analysis from South Asia. Cambridge, U.K.: Press Syndicate of the University of Cambridge. CIAT (International Center for Tropical Agriculture). 1987. ClAT Annual Report, 1987. Cali, Colombia. David, C.C. 1987. "The Global Rice Situation." Paper presented at the International Rice Research Conference, September 21-25, Hangzhou, China. Engelhardt, Thomas. 1984. "Economics of Traditional Smallholder Irrigation Systems in the Semi-Arid Tropics of South India.' Doctoral dissertation, University of Hohenheim. Feder, G., and T. Onchan. 1987. "Land Ownership Security and Farm Investment." American Journal of Agricultural Economics 69(1): 311-20. - 124 - Flinn, J.C., and S.K. De Datta. 1984. 'Trends in Irrigated-rice Yields Under Intensive Cropping at Philippine Research Stations." Field Crops Research 9(1): 1-15. Fujisaka, J. S. 1990. Targeting Research to Improve Sustainability and Productivity of Shifting Cultivation: Northern Laos. Los Bafios, the Philippines: International Rice Research Institute, Social Sciences Division. Fujisaka, J.S., and D.P. Garrity. 1988. "Developing Sustainable Food Crop Farming Systems for the Sloping Acid Uplands." Paper presented at the 4th Southeast Asian Universities Agroecosystems Network Research Symposium, Khon Kaen, Thailand. Garrity, D.P., A. Garcia, Patricio C. Agustin, and R. Dacumos. 1988. "Methods for the Extrapolation of Cropping Systems Technologies." Paper presented at the 19th Asian Farming Systems Network Working Group Meeting, November 11-15, Baguio City, Philippines. Gines, H.C., T.B. Moya, R.K. Pandey, and V.R. Carangal. 1989. "Crop Diversification: Problems and Prospects in Partially Irrigated Rice-Based Farming Systems." In Proceedings of a National Workshop on Crop Diversification in Irrigated Agriculture in the Philippines. Colombo, Sri Lanka: International Irrigation Management Institute. Gyprnantasiri, P., A. Wiboonpongse, B. Rerkasem, I. Craig, K. Rerkasem, L. Ganjanapan, M. Titayawan, M. Seetisarn, P. Thani, R. Jaisaard, S. Ongprasert, and T. Radanachaless. 1980. "An Interdisciplinary Perspective of Cropping Systems in the Chiang Mai Valley: Key Questions for Research." Paper prepared by the Faculty of Agriculture, University of Chiang Mai, Thailand. Hakim, M.A., et al. 1990a. "Irrigation Service Fee and Collection Efficiency." Paper prepared for the Workshop on Applied Research for Increasing Irrigation Effectiveness and Crop Production (organized by BRRI, BWDB, IIMI, and IRRI), Dhaka, Bangladesh. . 1990b. "Role of Farmer Organizations in Irrigation Management." Paper prepared for the Workshop on Applied Research for Increasing Irrigation Effectiveness and Crop Production (organized by BRRI, BWDB, IIMI and IRRI), Dhaka, Bangladesh. Herdt, R. W. 1987. "A Retrospective View of Technology and Other Changes in Philippine Rice Farming, 1965-1982." Economic Development and Cultural Change 35(2): 329-49. . 1988. "Increasing Crop Yields in Developing Countries." Paper presented at the 1988 meeting of the American Agricultural Economics Association, July 30 to August 3, Knoxville, Tenn. INSURF (International Network for Sustainable Rice Fertility). 1987. Annual Report of the International Network for Sustainable Rice Fertility. Los Bainos, the Philippines: International Rice Research Institute. Kundu, D.K., and S.K. De Datta. 1988. "Integrated Nutrient Management in Irrigated Rice." Paper presented at IRRI, November 7-11, Los Bafios, the Philippines. Mandac, A.M., R.D. Magbanua, and M.P. Genesila. 1987. "Multiple Cropping System in Northern Mindanao, Philippines." Philippine Journal of Crop Science 12(2): 71-85. Masicat, P.B., M.V.M. de Vera, and P.L. Pingali. 1990. "Philippine Irrigation Infrastructure: Degradation Trends for Luzon, 1966-1989." Paper presented at the Sixth Annual Scientific Conference of the Federation of Crop Science Societies of the Philippines, May 16-18, Naga City. Mitchell, D.O. 1987a. Factors Affecting Grain Prices. Washington D.C.: World Bank. - 125 - 1987b. "Rice Market Prospects to the Year 2000.n Paper presented at the Seminar on Recent and Future Movements in World Rice Prices, January 13-14, Jakarta, Indonesia. Mitchell, D.O., and R.C. Duncan. 1987. "Market Behavior of Grain Exporters." World Bank Research Observer 2(1): 3-21. Mondall, M.K., et al. 1990. "Water Regimes and Crop Diversification." Paper prepared for the Workshop on Applied Research for Increasing Irrigation Effectiveness and Crop Production (organized by BRRI, BWDB, IIMI, and IRRI), Dhaka, Bangladesh. Moya, T.B., and S.M. Miranda. 1989. "Socio-Technical Issues in Diversifying Rice-Based Irrigation Systems." In Proceedings of a National Workshop on Crop Diversification in Irrigated Agriculture in the Philippines. Colombo, Sri Lanka: Intemational Irrigation Management Institute. Nambiar, K.K.M., and A.B. Ghosh. 1984. "Highlights of Research of a Long Term Fertilizer Experiment in India (1971-82)." LTFE Research Bulletin, no. 1 (Indian Agricultural Research Institute, New Delhi). Onchan, T., and Y. Chalamwong. Forthcoming. Rural Off-farm Income and Employment in Thailand: Current Evidence, Future Trends and Implications. Pemito, R., and D.P. Garrity. 1988. "Mungbean Response to Surface Drainage When Grown as a Pre-rice Crop on Waterlogging-Prone Ricelands." Paper presented at the 4th Annual Scientific Meeting of the Federation of Crop Science Societies of the Philippines, April 27-29, Davao City. Pingali, P.L. 1987. "Trip Report to Northeast Thailand." Los Banos, the Philippines: International Rice Research Institute, Economnics Department. Pingali, P.L., Yves Bigot, and Hans Binswanger. 1987. Agricultural Mechanization and the Evolution of Farming Systems in Sub-Saharan Africa. Baltimore: The Johns Hopkins University Press. Pingali, P.L., and H.P. Binswanger. 1987. "Population Density and Agricultural Intensification: A Study of the Evolution of Technologies in Tropical Agriculture." In D. Gale Johnson and Ron Lee, eds., Population Growth and Economics Development. Madison, Wis.: University of Wisconsin Press. Pingali, P.L., P. Masicat, P. Moya, and A. Papag. 1989. "The Micro-economics of Crop Diversification in a Diversion Irrigation System: A Progress Report from the UTRIS." In Alfredo Valera, ed., Crop Diversification in Irrigated Agriculture in the Philippines. Colombo, Sri Lanka: International Irrigation Management Institute. Pingali, P.L., and P.F. Moya. 1989. "Has the 'Green Revolution' Played Itself Out on Asia's 'Best' Farms?: Micro-Level Evidence from the Philippines." Paper presented at the AAEA meetings at Louisiana State University, July 30 to August 2, Baton Rouge. Pingali, P.L., P.F. Moya, and L.E. Velasco. 1990. The Post-Green Revolution Blues in Asian Rice Production: The Diminished Gap between Experiment Station and Farmer Yields. IRRI Social Sciences Division Paper 90-01. Los Banios, the Philippines. Rosegrant, M.W., and P.L. Pingali. 1990. Sustainability of Rice Productivity Growth in Asia: A Policy Perspective. IRRI Social Sciences Division Paper 91-01. Los Bantos, the Philippines. Schuh, G.E. 1987. "Agricultural Research: Still a Good Investment?" In CGL4R Annual Report, 1986/87. Washington, D.C.: World Bank. - 126 - Soemarwoto, O., and I. Soemarwoto. 1984. "The Javanese Rural Ecosystem." In T. A. Rambo and P.E. Sajise, eds., An Introduction to Human Ecology Research on Agricultural Systems in Southeast Asia. Los Bafios: University of the Philippines. Sriarunrungreauang, Sernsri. 1989. "The Impacts of Declining Price of Rice on Production, Allocation of Farm Resources and Farm Incomes: A Case Study of Small Rice Farms in Amphoe Don Chedi and Amphoe U- Thong, Changwat Suphan Buri." Master's thesis, Department of Agricultural Economics, Kasetsart University, Bangkok. Walker, T.S., and J.G. Ryan. 1990. Against the Odds: Village and Household Economics in India's Semi-Arid Tropics. Baltimore: The Johns Hopkins University Press. Wardana, P., A.M. Fagi, M.A. Lantican, and S.I. Bhuiyan. 1990. "Water Relations to Dry Season Crop Choice and Profitability in the Cikeusik Irrigation System, West Java." Paper presented at IIMI's Intercountry Workshop on Irrigation Management for Rice-based Farming Systems, November 12-14, Colombo, Sri Lanka. Wescott, M.P., and D.S. Mikkelson 1988. "Effect of Green Manure on Rice Soil Fertility in the United States." In Green Manuring and Rice Farming. Los Banos, the Philippines: International Rice Research Institute. Zandstra, H.G. "Technological Considerations in Agricultural Diversification." This volume, p. 15. IRRIGATION MANAGEMENT FOR CROP DIVERSIFICATION Senen M. Miranda Since the mid-1960s, the primary aim of irrigation projects in Asia's humid tropics has been to provide more reliable, supplementary wet-season water supplies for wetland rice cultivation. Dry-season irrigation has been a secondary consideration, and the reduced water supplies of that period are usually used to grow additional rice crops in more limited areas. Although nonrice crop production has sometimes been considered at the design stage, most irrigation projects have been planned on the basis of their suitability for rice production. When the International Irrigation Management Institute (IIMI) was established in 1984, several traditional rice-importing countries in Southeast Asia were on the verge of becoming self-sufficient or had already attained self-sufficiency in rice. Traditional exporting countries, such as Thailand, were beginning to feel the pinch of a shrinking international rice market and declining prices (table 1). This glut in rice supply has been ascribed to the new rice/seed/fertilizer technology introduced by the International Rice Research Institute (IRRI) and the rapid expansion of irrigated rice areas. The heavy investment in irrigation produced the controlled flooded water environment considered necessary to exploit the potential of the new high-yielding and early-maturing rice varieties. The success in increasing rice production has, however, spawned a second-generation problem of declining farmer incomes. It is in this context that IIMI, shortly after its establishment, TABLE 1. WORLD MARKET RICE PRICES (1985 = 100) responded to an Asian Development PRICE (US/Mn Bank (ADB) request to explore the CROP 1970 1980 1985 1990 1995 2000 potential for, and constraints to, accommodating nonrice crops in Rice 395 414 216 233 214 206 irrigation systems in the Philippines and Indonesia that had been designed, Source: Schuh and Barghouti 1987. built, and operated to grow rice. UMI began the study in both countries in 1985, and similar research began in Sri Lanka at the request of the Sri Lanka Consultative Committee. In November 1986, with funding from the Bureau of Sciences and Technology of the U.S. Agency for International Development, a workshop on irrigation management for crop diversification was held at IIMI headquarters in Sri Lanka. Participants from several countries reviewed research findings, compared differences in irrigation management, and proposed future activities. In December 1988, with support from the ADB, IIMI sponsored a workshop to establish a research network on irrigation management for diversified cropping in rice-based systems. Participants from eight Asian countries-Indonesia, Malaysia, the Philippines, Thailand, Bangladesh, India, Nepal, and Sri Lanka-agreed to organize the network to facilitate the development and dissemination of needed irrigation management innovations in their countries to bring about agricultural diversification. The network was to be overseen by a steering committee, which subsequently met in Kuala Lumpur in late 1989 to review and plan activities, including an annual workshop. This paper summarizes the highlights of IIMI's research results in irrigation management for crop diversification in Indonesia, the Philippines, and Sri Lanka. The research was conducted in collaboration Senen M. Miranda is senior irrigation specialist, International Irrigation Managernent Institute (IIMI), Lahore, Pakistan. - 128 - with staff of various national agencies and IRRI, with secondary materials used to reinforce findings. The focus of HMI research is on identifying the physical, managerial, and institutional changes in irrigation management needed to more intensively cultivate nonrice crops in irrigation systems developed primarily for rice production. Research also examines the technical and socioeconomic factors that constrain or encourage more intensive rice production during the drier part of the year, when water is insufficient to grow rice over entire systems. In its ongoing collaboration with IRRI on irrigation management in rice- based farming systems, IIMI has further concentrated on the main system management concerns in which it has comparative advantage, while leaving on-farm irrigation questions to IRRI. Major Findings It became apparent early in IIMI's research that rainfall distribution, rather than total annual rainfall, was the major factor influencing cropping patterns. These patterns are determined by the prevailing monsoons and the presence or absence of mountain barriers. During the first wet season, the primary crop grown is rice, essentially an aquatic plant. During the second or third cropping seasons, when water supply becomes relatively scarce, the irrigation and cultivation practices observed vary across countries and conditions. In countries with only two cropping seasons, service areas are reduced to a fraction of those served during the wet season. This reduced irrigation area is rotated each year in the Philippines, whereas institutional sharing of the reduced area is practiced in Sri Lanka. In Indonesia, where cropping intensity is highest, the government has decreed that the third cropping period be devoted exclusively to nonrice crops. Rice is still being grown in poorly drained, low-lying areas with a high groundwater table. The Philippines is currently encouraging cultivation of nonrice crops in diversified and dual-land class soils, while Sri Lanka is promoting nonrice crops in well-drained soils. In irrigation systems management, the intensity of operational planning and implementation, monitoring, and plan evaluation was generally found to vary across countries and across sites within each country. Regardless, the basic principle in planning is to match water supply as closely as possible with water demand or soil and crop requirements. The planning process can be simple or complex, depending on the scope for manipulating supplies according to demand. Irrigation officials and related government personnel meet with farmers to decide formally on plans before they are finalized and implemented. Plan implementation is dictated by the availability of water at the start of the main season. The type of system-run-of-river or storage-influences the availability of water and, consequently, allocation, delivery, and distribution schedules. Monitoring of delivery implementation is inadequate, tending to break down toward the lower end of systems. Except where there is active farmer participation, monitoring is geared more toward office reporting than day-to-day operations. Consequently, it is generally found that the tail ends of systems are deficient, whereas the heads and upper portions have excess water. On-farm operations are relatively flexible in terms of farmer capability to cope with different conditions of water availability. With reliable-though not necessarily adequate-water supply at the turnout level, water sharing among farmers is better organized, with greater equity and fewer conflicts. The reverse situation, however, triggers an undesirable chain reaction. Some potentials and incentives for more diversified crop production in rice irrigation systems are describe below: * There are nonrice crops grown in each country that show higher and more consistent profitability than rice, as seen in table 2; * There are well-drained and coarse-textured soils in parts of the commands of the irrigation systems that are well-suited to diversified crops; - 129 - TABLE 2. AVERAGE DRY-SEASON RICE PRODUCTION COSTS AND RETURNS COMPARED WITH OTHER CROPS, 1986 INDONESIA (Rp/ha) RATIO OF NONRICEIICE CROP ITEM RICEi MAIZE' PEANUT' SHAIITO SOYBEANd CUCUMBER Cash cost Labor 237,792 0.86 0.97 2.60 0.41 2.53 Seeds 16,500 1.13 1.00 13.33 2.29 1.90 Ferdlizer 51,770 1.20 0.12 2.61 0.39 1.30 Chezecala 13,000 0.50 1.33 0.70 4.43 2.02 Equipment 35,000 0.36 -- -- 0.21 1.43 Rent, taxes, etc. 50,638 0.76 0.83 0.46 1.02 1.57 Total 404,700 0.87 0.79 2.50 0.71 2.17 Gross retuns 688,578 0.78 0.89 3.60 1.29 2.90 Net retun 283,878 0.64 1.04 0.52 2.12 3.90 THE PHlI.PPINES (pesofis) RATIO OF NONRICE/RICE CROP ITEM RICE TOBACCO COTTON TOMATO MUNGBEAN ONION GARIIC Cash costs Hired labor 2,842 0.60 0.78 0.53 0.22 2.56 1.39 Seeds 435 0.00 0.00 1.15 1.79 9.84 29.17 Fertilizer 1,243 1.82 1.51 1.95 0.00 0.00 1.25 Chemicalb 290 5.60 8.57 9.15 2.00 3.72 1.78 Totals 7,507 1.22 1.39 1.26 0.39 3.22 1.94 Grow retur 11,035 1.86 1.89 0.62 0.25 3.70 2.19 Net retun 3,528 3.48 2.59 -- - 4.77 2.69 SRI LANKA (SL Rs5/ha) RATIO OF NONRICE/RICE CROP DEWAHUWA KALANKUTIIYA ITEM RICE CHIU MUNGBEAN SOYBEAN RICE CHIU Cash costs Hired labor 1,345 4.70 1.80 1.80 1,432 3.40 Seeds 406 1.40 2.00 1.80 671 0.90 Ferilizer 788 2.60 0.10 0.20 804 2.50 Chemicals 181 14.30 5.80 2.10 420 5.20 Equipment 1,323 1.10 1.00 1.30 1,930 0.80 Irrigationwater 7 11.10 -- -- 56 1.00 Total 4,050 3.20 1.40 1.30 5,313 2.10 Gross return 7,814 3.36 1.64 2.16 10,436 2.43 Net return 3,764 3.60 1.94 3.00 5,123 2.76 a. Rp 1,514 = US$1. b. Wet season. c. East Java. d. Central Java. e. West Java. f. Peso 20 = US$1. g. Inclhding other cash costs. h. SL Rs 30 = US$1. i. Fee not inchlded. Source: ILUI 1986, various reports. - 130 - * Limited water supplies that are inadequate to meet the requirements of rice during the dry season are observed in many schemes with favorable soils. Related to this are distinct unimodal rainfall patterns that make well-aerated soil conditions during the dry season possible; * No major land movement or reshaping is needed to irrigate nonrice crops, although farmers have to introduce rudimentary systems of on-farm supply and drainage ditches in their plots to facilitate the timely application and removal of water from their fields. In the Upper Talavera River Irrigation System (UTRIS) in the Philippines, for instance, additional farm ditches are dug for growing onions and erased during land preparation for the subsequent wet-season rice; * Rice basin bunds are retained where the appropriate seedbeds are prepared according to the water application requirements of specific nonrice crops, as indicated in table 3 and shown in figure 1; * Irrigation systems properly designed and constructed for supplementary irrigation of wet- season rice that can meet the land-soaking and land-preparation requirements have enough canal capacity for the intermittent flow of water needed to irrigate nonrice crops, although the need for greater canal water regulation is apparent. This is partly the result of the smaller water requirement of nonrice crops, which permits a reduction in the total volume of water supplied per delivery interval. Moreover, increasing the "free-board" depth by 10 percent for typical canal cross-sections provides an increase in flow area. Thus, the extra canal capacity can be used for occasional excess flows (Levine 1986; Ko 1986); - Greater interest among all concerned, from farmers to policymakers and the donor community, is being generated. There is increased attention to the issues inherent in evolving a viable strategy for rural diversification, of which irrigated crop diversification is a key element. There are, however, the following constraints: * Water control for nonrice crops is more demanding in terms of supply and removal because of stricter soil moisture requirements. The intermittent delivery of limited and uncertain water supplies during the dry season requires greater joint management efforts and, in turn, effective communication between irrigation staff and farmers; * The presence of regulation and measuring facilities is essential to provide functional water control and enable effective monitoring of the water supply; * Farmers who have grown only irrigated rice are unfamiliar with the agronomic and irrigation practices for nonrice crops; * There are greater economic risks associated with nonrice crops than with rice: cash and labor inputs can be three or four times higher for nonrice crops. Institutional credit is scarce, and noninstitutional credit carries usurious interest rates as high as 20 percent a month; * Unstable prices and the absence of organized marketing for nonrice crops increase the risks for farmers involved in their production. - 131 - Figure 1. Irrigation Application Practices for Nonrice-Crop Fields A. Basin flooding B. Flusfi basin flooding with rudimentary shallow field ditches C. Broad bed and furrow D. Raised beds E. Furrowed basin A Source: Adapted from ILMI 1987. - 132 - Recommendations TABLE 3. IRRIGATION FIELD APPLICATION Following are some PRACTICES FOR RICE AND NONRICE CROPS, BY COUNTRY recommendations for APPLICATION managing irrigation systems PRACTICE INDONESIA THE PHILIPPINES SRI LANKA to provide the flexibility Bs flooding Basin flooding Rice Rice ~~~~~Rice required to accommodate the more stringent water Flush basin flooding Soybean Garlic Soybean requirements of nonrice with rudimentary field Peanut Mungbean/white bean Mungbean crops. Irrigation systems ditches Mungbean Onion (with mulch) with dry-season water supplies that are not Broadbed and furrow Cucumber Garlic Chili sufficient to grow rice Onion across the entire command Raised beds Red onion Onion Bombay onion area, and with substantial areas of well-drained and Furrow Corn Corn Chili Cassava White bean coarse-textured or diversified soils, should be selected first. The irrigation Source: UMI reports, various years. system should be in a physical state to enable satisfactory water delivery and control at various levels, especially at the turnout level. Wherever additional checks are placed by farmers, these indicate a need for greater canal regulation, which should be considered in establishing check structures. Repeated overflow points suggest a lack of capacity brought about by siltation or embankment settlement. There is an urgent need to improve the interaction between irrigation staff and farmers in managing irrigation systems, from seasonal planning and implementation to water delivery monitoring. Some form of joint management that enhances farmer group participation in the irrigation management process is needed to meet the more demanding requirements of nonrice crops, given limited and uncertain water supply. When water rotation is needed, a clear plan must be drawn up for each season. It may be possible to establish a progressive rotational plan by having rotation between tertiaries when water supplies are moderately deficient and rotation between secondaries or lateral canals when water supplies are severely deficient. Reducing excess moisture and improving drainage have been shown to encourage more crop diversification. Not only is stricter control over water deliveries implied, which may require additional management inputs, but complementary on-farm drainage facilities for fast removal of excess water and lowering of the water table may have to be installed and integrated with the main drainage systems. Water table observations in 1987 in a catenary transect of a turnout service area in the Hahaweli H system in Sri Lanka illustrate dry-season water table behavior. Some form of intercepting field drain is required to cut off the seepage flow from the higher, adjoining allotment. Farmers try to provide these facilities with rudimentary systems of dual-purpose field ditches and beds. Irrigation and drainage systems that enable the timely application and removal of water should be envisioned. This could be achieved by evaluating a spectrum of available technology, ranging from the ingenious rudimentary technologies of farmers to the capital-intensive practices in Japan and Taiwan in which each field plot is served by an irrigation ditch, a drainage ditch, and a farm road. Assessments of water supply and demand need to be improved to better match diversified cropping conditions. For nonrice crops, supply includes not only canal flow and rainfall but also subsurface supply from the moisture stored in the soil profile and from the groundwater table. There is also the augmenting supply from shallow tubewells installed by farmers. The contribution of the - 133 - fluctuating water table is more difficult to determine. Demand information on the type, extent, and growth stage of nonrice cropping needs to be collected periodically in the field. This could be facilitated by more organized farmer participation in the management of irrigation systems. More vigorous extension programs that disseminate information on irrigation as well as agronomic practices for nonrice crops with good potential profitability can help farmers decide which nonrice crops to grow. Finally, an assured and stable market, competitive prices, and readily available institutional credit are essential to promoting and sustaining crop diversification. Conclusion Irrigated crop diversification has the potential to increase farmer incomes, achieve stable national food supplies, and earn or save foreign exchange. It may be a starting point for more general and all- encompassing agricultural diversification programs. Through its research network on irrigation management for crop diversification in rice-based systems, IIMI is collaborating with irrigation, agricultural, and policymaking organizations to bring about the desired agricultural diversification. Acknowledgments The author acknowledges the contributions of IIMI colleagues who conducted the irrigated diversification studies, particularly C.R. Panabokke, H. Murray-Rust, A. Valera, D. Vermillion, E. Martin, and M. Kikuchi. The author is also indebted to IIMI's field research staff and the staff of national collaborating irrigation management agencies and research organizations in the three countries where the research is being conducted. Bibliography International Irrigation Management Institute (IIMI). 1987. 'Final Report on the Technical Assistance Study of Irrigation Management. Indonesia." Report TA No. 673-INO. Digana Village, Sri Lanka. 1989a. 'Final Report of ADB RETA, Vol. 2." Digana Village, Sri Lanka. 1989b. Research Network on Irrigation Management for Diversified Cropping in Rice-based Systems. Proceedings of the Organizational and Planning Workshop for a Research Network on Irrigation Management for Diversified Cropping in Rice-based Systems, Asian Institute of Technology Center, November 30 to December 3, Bangkok. Colombo, Sri Lanka. Ko, H.S. 1987. "Adoption of Diversified Cropping in Rice Irrigation Projects." In IIMI, Irrigation Management for Diversified Cropping. Digana Village, Sri Lanka. Levine, G. 1986. "The Challenge of Rehabilitation and Betterment." In Fowler, ed., International Conference on Irrigation System Rehabilitation and Betterment, vol. 2. Water Management Synthesis II Project. Fort Collins, Colo.: Colorado State University. Miranda, S.M. 1989. Irrigation Management for Crop Diversification in Indonesia, the Philippines, and Sri Lanka: A Synthesis of lIMI's Research. Colombo, Sri Lanka: IIMI. Miranda, S.M., and C.R. Panabokke. 1987. 'Irrigation Management for Diversified Cropping: A Concept Paper." In IIMI, Irrigation Management for Diversif ed Cropping. Digana Village, Sri Lanka. - 134 - Panabokke, C.R. 1988. 'On-farm Water Management for Crop Diversification in Sri Lanka." Draft final report of IMI Crop Diversification Group. Digana Village, Sri Ianka. Schuh, G.E, and S. Barghouti. 1987. 'Meeting the Challenges of Diversification out of Rice Production in Asia: Towards a Research Agenda." Paper presented at a meeting of the Consultative Group for International Agricultural Research, Montpelier, France. Tabbal, D.F, G.S. Bumanlag, M.K. Mondal, and S.I. Bhuiyan. 1988. 'Farm Level Water Control and Management Practices for Rice and Non-rice Crop Sequence." Progress report presented at the Review and Planning Workshop on the IIMI-IRRI Collaborative Project on Irrigation Management for Rice-Based Systems, Los Baios, the Philippines. Wen, L. 1987. "Irrigation Management for Diversified Cropping in Taiwan." In IIMI, Irrigation Managementfor Diversified Cropping. Digana Village, Sri Lanka. World Bank. 1988. "Policies and Issues for Diversification in Asian Agriculture." Agriculture and Rural Development Department. Washington, D.C. INDONESIA: DIVERSIFICATION AS AN AGRICULTURAL POLICY INSTRUMENT Faisal Kasryno Indonesia's five-year development plan Repelita V (1989-93) calls for agricultural development to encompass four major processes: diversification, intensification, extensification, and rehabilitation. Diversification is the most difficult of these terms to interpret. Usually, it is understood in the micro "horizontal" sense as combining crop and livestock operations to minimize risk and optimize the utilization of family resources to maximize income. Another common definition adds a "vertical" dimension of product diversification through agroindustrial development. The World Bank expands diversification to include off-farm employment by defining it as "a process of broadening and maintaining the sources of incomes of rural households" (World Bank 1988). This paper identifies and clarifies diversification as a potential policy instrument in Indonesia's agricultural development. It then presents operational implications for the country's agricultural institutions. Patterns of Diversification in Indonesia Attempts to measure diversification in Indonesian agriculture show that diversification-measured in terms of sources of rural income-varies across Indonesia's districts and provinces, although no changing geographic trend in diversification can be observed over 1982-87. Generally, Indonesian income sources are less diversified in lowland than in upland ecosystems. Thus, low diversification indices in West Java, South Sulawesi, and South Kalimantan reflect the domination of lowland rice, for which these regions are ecologically well suited. Lampung, West Sumatra, and Central Java show more diversification because of their diverse agroecologies and dryland ecosystems. The analysis shows no systematic relation between the degree of diversification and employment. An examination of regional shifts in food crop production shares for all crops other than rice provides some indication of the increasing diversification of Indonesia's agricultural economy (area and production figures for rice are presented in table 1). The harvested area in corn increased only slightly between 1969 and 1988, although production doubled as a result of high yield increases. Concurrently, there was a significant shift in corn area and production away from Java, especially to Sumatra and South Sulawesi, though Java's share remained high, particularly in East Java (table 2). Soybean production shifted significantly away from Java to Sumatra, Sulawesi, and east Indonesia (table 3). Demand increased at an even higher rate, particularly when the demand for feed is included. Java's share of soybean production (particularly East Java) dropped dramatically, whereas Sumatra, Sulawesi, and east Indonesia experienced significant increases. Between 1969 and 1988, the harvested area of cassava declined as a result of the sharp decline in Java, although the harvested area outside Java increased slightly. Yields increased, so that cassava production generally increased for all regions, whereas Java's share of production declined. The highest Faisal Kasryno is director, Bureau of Planning, and senior agricultural economist, Center for Agroeconomic Research, Agency for Agricultural Research and Development, Ministry of Agriculture, Indonesia. - 136 - TABLE 1. AVERAGE HARVESTED AREA AND PRODUCTION OF RICE BY REGION, 1969-88 HARVESTED AREA ('000 HA) PRODUCTION ('000 TON) 1969-74 1981-85 1986-88 1969-74 1981-85 1986-88 REGION AREA % AREA % AREA % PROD. % PROD. % PROD. % East Java 1,239 (15) 1,523 (16) 1,558 (16) 3,689 (18) 7,259 (20) 7,663 (19) Central Java 1,410 (17) 1,548 (16) 1,620 (16) 3,903 (20) 6,891 (19) 7,698 (19) West Java 1,778 (22) 1,934 (21) 2,054 (21) 4,697 (23) 8,022 (23) 9,297 (23) North Sumatra 533 (7) 570 (6) 637 (6) 1,490 (7) 1,894 (5) 2,290 (6) Other Sumatra 1,394 (17) 1,665 (18) 1,807 (18) 2,860 (14) 5,119 (14) 6,001 (15) South Sulawesi 477 (6) 595 (6) 680 (7) 1,109 (6) 2,237 (6) 2,715 (6) Other Sulawesi 192 (2) 220 (2) 249 (2) 347 (2) 593 (2) 766 (2) Other Indonesia 1,191 (14) 1,371 (15) 1,412 (14) 2,047 (10) 3,749 (11) 4,025 (10) Total 8,214(100) 9,426(100) 10,017(100) 20,142(100) 35,764(100) 40,455(100) Source: Central Bureau of Statistics (CBS), various publications. TABLE 2. AVERAGE HARVESTED AREA AND PRODUCTION OF CORN BY REGION, 1969-88 HARVESTED AREA ('000 HA) PRODUCTION ('000 TON) 1969-74 1981-85 1986-88 1969-74 1981-85 1986-88 REGION AREA % AREA %9 AREA % PROD. % PROD. % PROD. % East Java 1,163 (43) 1,096 (40) 1,134 (37) 1,127 (40) 1,943 (43) 2,349 (40) Central Java 624 (23) 605 (22) 663 (22) 752 (27) 1,118 (25) 1,415 (24) West Java 107 (4) 83 (3) 114 (4) 129 (5) 138 (3) 219 (4) North Sumatra 24 (1) 41 (2) 68 (2) 39 (1) 74 (2) 139 (2) Other Sumatra 99 (4) 127 (4) 256 (8) 110 (4) 196 (4) 502 (8) South Sulawesi 253 (9) 281 (10) 306 (10) 201 (7) 397 (9) 267 (11) Other Sulawesi 154 (6) 162 (6) 155 (5) 159 (6) 240 (5) 476 (5) Other Indonesia 286 (10) 244 (13) 262 (12) 263 (10) 429 (9) 542 (6) Total 2,710(100) 2,639(100) 2,958(100) 2,780(100) 4,535(100) 5,909(100) Source: CBS, various publications. increase was observed in Sumatra (table 4). The harvested area of sugarcane increased significantly, with production more than doubling (table 5). Between 1970 and 1988, estate crop production shifted less, with Sumatra retaining its dominant role, especially for oil palm, rubber, coffee, and pepper. The share of the rest of Indonesia, primarily the east Indonesian islands of Irian Jaya, Maluku, and Nusa Tenggara, and East Timor, remained small for oil palm, rubber, and pepper. Cocoa production between 1970 and 1988 increased rapidly, from just 12,000 ha in 1970 to 186,000 ha in 1988. Despite this increase, Java's share declined from 47 to 16 percent during this period, whereas the share of the rest of Indonesia (mainly Sulawesi) increased from 43 to 56 percent (table 6). Cashew grows well in relatively marginal land and has been used for land rehabilitation to increase farmers' incomes. Between 1980 and 1988, cashew nut production more than doubled. A large part of the area expansion was in Sulawesi and Nusa Tenggara. Oil palm was once concentrated in North Sumatra (90 percent in 1970), but has spread to many provinces in Sumatra, maintaining the high Sumatra share of 90 percent in 1988. The growth of oil palm in the remaining islands was also significant between 1970 and 1988 (table 7). - 137 - TABLE 3. AVERAGE HARVESTED AREA AND PRODUCTION OF SOYBEAN BY REGION, 1969-88 HARVESTED AREA ('000 HA) PRODUCTION ('000 TON) 1969-74 1981-85 1986-88 1969-74 1981-85 1986-88 REGION AREA % AREA % AREA % PROD. % PROD. % PROD. % East Java 378 (55) 337 (44) 394 (34) 291 (57) 322 (47) 419 (34) Central Java 167 (24) 180 (24) 206 (18) 116 (23) 150 (22) 219 (18) West Java 29 (4) 40 (5) 68 (6) 21 (4) 31 (5) 67 (6) North Sumatra 7 (1) 7 (1) 26 (2) 6 (1) 6 (1) 27 (2) Other Sumatra 34 (5) 91 (12) 271 (23) 27 (5) 80 (12) 270 (22) South Sulawesi 6 (1) 13 (2) 37 (3) 4 (1) 12 (2) 38 (3) Other Sulawesi 2 (0) 21 (3) 40 (3) 1 (0) 17 (2) 41 (4) Other Indonesia 67 (10) 71 (9) 135 (11) 43 (9) 61 (9) 138 (11) Total 690(100) 760(100) 1,177(100) 509(100) 679(100) 1,219(100) Source: CBS, various publications. TABLE 4. AVERAGE HARVESTED AREA AND PRODUCTION OF CASSAVA BY REGION, 1969-88 HARVESTED AREA ('000 HA) PRODUCTION ('000 TON) 1969-74 1981-85 1986-88 1969-74 1981-85 1986-88 REGION AREA % AREA % AREA % PROD. % PROD. % PROD. % East Java 464 (32) 374 (28) 301 (25) 3,377 (30) 3,710 (28) 3,500 (24) Central Java 415 (29) 356 (27) 324 (26) 2,939 (26) 3,621 (27) 3,895 (27) West Java 245 (15) 178 (13) 147 (12) 1,989 (18) 1,955 (15) 1,846 (13) North Sumatra 20 (1) 23 (2) 23 (2) 190 (2) 259 (2) 275 (2) Other Sumatra 84 (6) 136 (10) 175 (14) 837 (8) 1,504 (11) 2,161 (15) South Sulawesi 39 (3) 30 (2) 37 (3) 301 (3) 300 (2) 416 (3) Other Sulawesi 42 (3) 47 (4) 38 (3) 330 (3) 421 (3) 407 (3) Other Indonesia 158 (11) 179 (14) 187 (15) 1,151 (10) 1,610 (12) 1,880 (13) Total 1,467(100) 1,323(100) 1,232(100) 11,114(100) 13,380(100) 14,380(100) Source: CBS, various publications. TABLE 5. AVERAGE HARVESTED AREA AND PRODUCTION OF SUGARCANE BY REGION, 1969-88 HARVESTED AREA ('000 HA) PRODUCTION ('000 TON) 1969-74 1981-85 1986-88 1969-74 1981-85 1986-88 REGION AREA % AREA % AREA % PROD. % PROD. % PROD. % East Java 125 (86) 171 (51) 167 (48) 943 (91) 1,078 (62) 1,166 (55) Central Java 8 (5) 83 (25) 64 (18) 59 (6) 381 (22) 455 (22) West Java 0 (0) 37 (11) 24 (7) 2 (0) 144 (8) 113 (5) Total Java 133 (91) 291 (87) 255 (73) 1,004 (97) 1,603 (92) 1,734 (82) Outside Java 13 (9) 43 (13) 96 (27) 27 (3) 145 (8) 380 (18) Total Indonesia 146(100) 334(100) 351(100) 1,031(100) 1,748(100) 2,114(100) Source: CBS, various publications. - 138 - Sumatra dominates coffee production, and its share in coffee TABLE 6. CASHEW AND COCOA AREAS BY area between 1970 and 1988 slightly REGION, SELECTED YEARS ('000 HA) increased from 55 to 62 percent. CASHEW COCOA Java's share in coffee area dropped REGION 1980 1988 1970 1980 1988 sharply, from 28 to 14 percent during the same period. For the North Sumatra 4 (3) 3 (1) 1 (10) 6 (16) 26 (14) remaining islands, the share in Java 73 (63) 79 (31) 6 (47) 15 (41) 30 (16) coffee area increased (table 8). West Kalimantan 1 (1) 4 (2) - (0) - (0) 3 (2) Sumatra, and more OtherKdlinantan I (1) 4 (1) - (0) 1 (3) 13 (7) Other Indonesia 35 (30) 163 (64) 5 (43) 14 (38) 105 (56) particularly South Sumatra, Total 117(100) 256(100) 12(100) 37(100) 186(100) dominates rubber and pepper cultivation. Its share in pepper declined frm 95 percen in 1970 to Note: Percent of total in parentheses. declined from 95 percent in 1970 to Source: Estate Crops Statistics, Directorate General of Estates, Ministry of 82 percent in 1988, a result of the Agriculture, 1990. growth of pepper cultivation in the remaining islands (table 8). Sumatra's share in rubber production increased slightly, from 70 percent in 1970 to 72 percent in 1988. In Java, rubber declined, whereas some increase was observed in Kalimantan, except in West Kalimantan (table 9). Java's share in coconut production declined TABLE 7. OIL PALM AND COCONUT slightly from 33 percent in AREAS BY REGION, SELECTED YEARS ('000 HA) 1970 to 28 percent in 1988 OIL PALM COCONUT (table 7). Coconut is grown REGION 1970 1980 1988 1970 1980 1988 throughout Indonesia and is essentiall ndnevenl distri- North Sumatra 119 (90) 240 (82) 434 (49) 60 (3) 119 (4) 150 (4) e t Y enY distri- Other Sumatra 13 (10) 51 (17) 362 (41) 375 (21) 599 (22) 848 (26) buted in proportion to Java 1 (0) 4 (1) 16 (2) 599 (33) 826 (31) 889 (28) population. Coconut WestKalimantan - (0) - (0) 38 (4) 56 (3) 99 (4) 72 (2) production in Java mainly Other Kalimantan - (0) - (0) 19 (2) 36 (2) 98 (4) 132 (4) Other Indonesia - (0) - (0) 20 (2) 680 (38) 939 (35) 1,150 (36) entails selling fresh coconut Total 133(100) 295(100) 889(100) 1,806(100) 2,680(100) 3,241(100) milk or drinks in the market. Tobacco is locarketion-sp c Tobaccuse Note: Percent of total in parentheses. location-specific because Source: Estate Crops Statistics, Directorate General of Estates, Minstr of Agriculture, 1990. quality is affected by climatic conditions, and the shifts shown in table 9 may be temporary. Java's share in tobacco has remained high, nevertheless, with 77 percent in 1988. Between 1970 and 1988, the regional distribution of livestock production changed. Cattle population is highest in East Java. Despite rising figures, East Java's percentage share relative to the total cattle population declined from 43 percent in 1970 to 30 percent in 1988 (table 10). The carabao population declined significantly in Java during 1970-80. In 1988, the highest share of carabao population was in Sumatra, rising from 21 percent in 1970 to 36 percent in 1988. The goat population is highest in Java, particularly Central Java and the Yogjakarta area. Java's share in the goat population declined from 73 percent in 1970 to 59 percent in 1988 (table 11). The sheep population is highest in West Java, though the province's share declined from 54 percent in 1970 to 45 percent in 1988. Java's total share of sheep fell from 93 percent in 1970 to 88 percent in 1988 (table 11). The hog population, shown in table 12, varies with the local inhabitants' religious beliefs. - 139 - Diversification of Household Income TABLE 8. COFFEE AND PEPPER AREAS BY REGION, SELECTED YEARS ('000 HA) The growth rate of COFFEE PEPPER farm household per capita REGION 1970 1980 1988 1970 1980 1988 income has reached SouthSumatra 67 (17) 182 (26) 235 (23) 7 (15) 19 (27) 48 (43) approximately 3.8 percent Other Sumatra 149 (38) 296 (42) 391 (39) 36 (80) 43 (63) 43 (39) annually. Rural household Java 112 (28) 110 (15) 140 (14) 1 (1) 1 (1) 1 (1) incomes are diversifying, West Kalimantan 2 (1) 6 (1) 8 (1) - (0) 2 (2) 4 (4) with aruOther Kalimantan 3 (1) 9 (1) 22 (2) 1 (2) 3 (5) 10 (9) with agriculture's share Other Indonesia 62 (15) 105 (15) 212 (21) 1 (2) 1 (2) 4 (4) declining and the shares of Total 395(100) 708(100) 1,008(100) 45(100) 69(100) 110(100) the service sector and off- farm employment rising. The Note: Percent of total in parentheses. role of rural processing Source: Estate Crops Statistics, Directorate General of Estates, Ministry of Agriculture, industries remains weak, 1990. contributing only 6 percent to rural incomes. Table 13 presents TABLE 9. RUBBER AND TOBACCO AREAS BY REGION, rural household incomes SELECTED YEARS ('000 HA) for several regions in RUBBER TOBACCO Indonesia based on 1987 REGION 1970 1980 1988 1970 1980 1988 data from the Household South Sumatra 493 (21) 474 (20) 549 (18) 1' (1) 4' (3) 3a (1) Expenditure Survey Other Sumatra 1,133 (49) 1,246 (52) 1,624 (54) 9 (5) 7 (5) 12 (5) (SUSENAS), conducted Java 244 (10) 169 (7) 167 (6) 124 (73) 87 (62) 182 (77) by the Central Bureau of W. Kalimantan 289 (13) 317 (14) 373 (12) - (0) - (0) - (0) Other Statistics. The share of Kalimantan 152 (7) 173 (7) 283 (10) - (0) - (0) - (0) farm income relative to Other Indonesia 4 (0) 8 (0) 13 (0) 36 (21) 43 (30) 39 (17) total income was smallest Total 2,317(100) 2,387(100) 3,009(100) 170 (100) 141(100) 236(100) for West and East Java. In West Java the role of Note: Percent of total in parentheses. wage/salary income was a. Figures are for North Sumatra. relatively large because Source: Estate Crops Statistics, Directorate General of Estates, Ministry of Agriculture, 1990. of increased employment opportunities outside agriculture and outside villages in industries around urban areas such as Jakarta, Bandung, Karawang, Bekasi, Tanggerang, and Bogor. Only in East Java did industrial employment provide a significant share of income (13.4 percent). The salient industries in East Java are cracker (kerupuk) making and machine shops producing farm equipment such as power threshers and tillers. In coastal areas the most significant industry is fish processing. The top five provinces by level of total farm household income are Riau, Maluku, North Sulawesi, South Sumatra, and Central Kalimantan. Common characteristics of these provinces are a lower contribution of food crops, combined with a higher contribution of other crops and of salary and wage income. The bottom five provinces are West Nusa Tenggara, South Kalimantan, Lampung, East Java, and Central Java. These provinces show a high concentration of food crops (mainly rice, corn, soybean, and cassava) and the lowest contribution of other crops. During the last 20 years, the domestic real price of rice has been declining. The average price of rice in 1987 was about 80 percent of the 1971 price level, largely the result of increased supplies, which can be attributed to a large degree to the significant growth in yields. Paddy yields have increased - 140 - TABLE 10. CATTLE AND CARABAO POPULATIONS BY REGION, SELECTED YEARS ('000 HEAD) CATrLE CARABAO REGION 1970 1980 1988 1970 1980 1988 East Java 2,590 (43) 2,576 (35) 2,913 (30) 256 (8) 232 (8) 222 (7) Other Java 1,273 (21) 1,187 (16) 1,472 (15) 1,233 (39) 773 (27) 806 (24) Sumatra 781 (13) 1,195 (16) 1,694 (17) 664 (21) 916 (32) 1,194 (36) Bali 277 (5) 342 (5) 433 (4) 9 (0) 7 (0) 9 (0) Southeast Nusa 499 (8) 636 (9) 927 (10) 342 (11) 309 (11) 400 (12) South Sulawesi 284 (5) 817 (11) 1,198 (12) 340 (10) 471 (17) 520 (15) Other Sulawesi 248 (4) 431 (6) 752 (8) 24 (1) 43 (1) 54 (2) Other Indonesia 129 (2) 226 (3) 417 (4) 316 (10) 102 (4) 134 (4) Total 6,081(100) 7,410(100) 9,806(100) 3,184(100) 2,853(100) 3,339(100) Note: Percent of total in parentheses. Source: Livestock Statistics, Directorate General of Livestock Services, Ministry of Agriculture, 1990. TABLE 11. GOAT AND SHEEP POPULATIONS BY REGION, SELECTED YEARS ('000 HEAD) GOAT SHEEP REGION 1970 1980 1988 1970 1980 1988 Sumatra 1,074 (15) 1,741 (21) 1,776 (17) 113 (3) 376 (10) 450 (8) West Java 1,190 (17) 934 (11) 1,613 (15) 1,818 (54) 1,730 (43) 2,473 (45) Central Java 2,166 (31) 2,174 (27) 2,612 (24) 949 (28) 1,122 (28) 1,342 (25) East Java 1,714 (25) 1,769 (22) 2,061 (20) 361 (11) 613 (15) 973 (18) Sulawesi 380 (5) 699 (9) 999 (10) 7 (0) 41 (1) 31 (1) Southeast Nusa and Bali 340 (5) 486 (6) 750 (7) 114 (3) 84 (2) 128 (2) Other Indonesia 129 (2) 339 (4) 743 (7) 27 (1) 32 (1) 49 (1) Total 6,993(100) 8,142(100) 10,554(100) 3,389(100) 3,998(100) 5,446(100) Note: Percent of total in parentheses. Source: Livestock Statistics, Directorate General of Livestock Services, Ministry of Agriculture, 1990. from 2.56 mt/ha in 1971 to 4.04 mt/ha in 1988. The real prices of other agricultural commodities, however, have risen in real terms. Table 14 shows rural household income and the rate of growth in employment opportunities in agriculture for 1978-85. In all provinces of Java food crops were dominant. Rice yields in Java have increased rapidly-especially irrigated rice-yet in absolute terms employment in agriculture has declined. This is mainly the result of the concentration of urban and industrial development in Java, prompting migration and urbanization of the rural labor force. This has contributed to increasing real wage rates in agriculture, which, together with the high profitability of rice farming, have induced the use of agricultural machinery as a substitute for labor. Advances in mechanized farm technology, such as power tillers, threshers, and rice mills, are primarily occurring in the rice-producing centers. Rice farming's profitability is mainly the result of improvements in irrigation systems; favorable input-output price ratios (until recently, inputs were subsidized); and the adoption of modem varieties. Thus, in Java and other - 141 - provinces where rice is important, such as Aceh, North Sumatra, West Sumatra, South Sulawesi, TABLE 12. HOG POPULATION BY and South Kalimantan, the growth of agricultural REGION, SELECTED YEARS ('000 HEAD) employment was small, below 1.0 percent REGION 1970 1980 1988 annually between 1978 and 1985 (table 14). In the irrigated rice-producing centers, Sumatra 472 (18) 1,042 (24) 1,645 (26) farm income reflects a tendency toward rice West Java 35 (1) 41 (1) 30 (1) Central Java 65 (3) 154 (3) 142 (2) monoculture, probably strengthened by the East Java 58 (2) 73 (2) 86 (1) government's rice-centered policy. In upland Sulawesi 377 (15) 586 (13) 777 (12) ecosystems, the sources of income tend to Southeast Nusa and Bali 626 (25) 1,248 (28) 1,889 (30) diversify. Thus, diversification depends on Other agroecosystems, resource potential, and Indonesia 932 (36) 1,255 (29) 1,797 (28) govermnent policy. Between 1976 and 1983, Total 2,565(100) 4,399(100) 6,366(100) income diversification was higher than in 1987, probably the result of better economic conditions Note: Percent of total in parentheses. in the early 1980s that offered higher Source: Livestock Statistics, Directorate General of Livestock employment opportunities for the unskilled labor Services, Ministry of Agriculture, 1990. TABLE 13. RURAL HOUSEHOLD INCOME BY REGION, 1987 ('000 RP) AGRI- SERVICE/ REGION CULTURE INDUSTRY TRADE SALARY OTHER TOTAL No. Sumatra 474.7 (49.5) 24.2 (2.5) 30.1 (3.1) 419.2 (43.7) 11.4 (1.2) 959.6 (100) Other Sumatra 485.4 (52.7) 31.1 (3.4) 80.1 (8.7) 310.5 (33.6) 14.5 (1.6) 921.6 (100) W. Java 175.5 (26.9) 16.5 (2.5) 84.4 (12.9) 364.0 (55.7) 12.7 (2.0) 653.1 (100) Cent. Java 254.4 (48.2) 17.9 (3.4) 56.3 (10.6) 191.5 (36.4) 7.3 (1.4) 527.4 (100) E. Java 251.7 (36.7) 89.6 (13.4) 47.8 (7.1) 264.8 (39.7) 14.8 (2.2) 668.7 (100) Kalimantan 415.8 (50.6) 31.0 (3.8) 79.1 (9.6) 283.8 (34.5) 12.1 (1.5) 821.8 (100) So. Sulawesi 439.4 (59.3) 52.6 (7.1) 52.5 (7.1) 189.9 (25.5) 6.0 (1.0) 740.4 (100) Other Sulawesi 486.3 (60.4) 49.7 (6.2) 71.0 (8.8) 194.7 (24.2) 3.7 (0.4) 805.4 (100) Other regions 481.4 (42.3) 96.4 (8.5) 21.8 (1.9) 412.4 (36.4) 123.7 (10.9) 1137.7 (100) Note: Percent of total in parentheses. Source: Processed 1987 SUSENAS (Household Expenditure Survey) data, Central Bureau of Statistics. force. Economic conditions worsened after 1984 when oil prices fell, and the resulting decline in off-farm employment was reflected in relatively lower income diversification. The growth rates of industrial crops, livestock, brackish-water fisheries, and horticultural crops have been encouraging, particularly in the provinces of Sumatra, such as Riau, Jambi, South Sumatra, Bengkulu, and Lampung; in the provinces of Kalimantan, except South Kalimantan; and in the provinces of Sulawesi, except South Sulawesi. Export Diversification Indonesia has been successful in diversifying exports to reduce its overdependence on oil and gas products. Table 15 shows a shift from agriculture to other sectors in nonoil exports over 1974-88, despite the more than threefold increase in agricultural exports over this period. Table 16 shows the changes by - 142 - agricultural subsector. Fisheries exports experienced particularly TABLE 14. RURAL HOUSEHOLD INCOME AND ASSOCIATED high growth. GROWTH OF EMPLOYMENT IN AGRICULTURE BY PROVINCE, 1984 Cattle and buffalo ceased % OF FOOD CROP INCOME to be exported in 1979; thereafter, RELATIVE TO ANNUAL EMPLOY. Indonesia became a net importer. TOTAL INCOME GROWTH Hogs began to beexportedAGRIC. TOTAL LEVEL RATE Hogs began to be exported only in PROVINCE INCOME INCOME ('000ORP) (1978-85) the last two years, after the establishment of a hog farm in D.I. Aceh 44.6 26.9 791 0.1 Batam exclusively for export. wes Sumatra 5.58 29.6 735 0.2 West Sn-iatr 58.830.5 735 0.2 Fisheries exports are dominated by Riau 17.4 9.9 1,086 3.8 shrimp, particularly from Jambi 40.5 28.0 724 8.5 aquaculture. The share of fresh South Sumatra 33.8 22.5 878 5.9 Bengkulu 37.7 26.1 821 7.0 fish is also increasing rapidly, Lanpung 39.7 25.1 590 6.5 although the absolute amount remains low compared with west Java 68.0 28.3 640 -0.8 Indonesia's extensive marine Central Java 61.6 31.0 609 -1.0 D.I. Yogjaksrta 57.1 24.3 750 -2.0 resources. East Java 58.4 33.9 593 0.1 Table 17 shows export values for food crops. The Bali 39.1 23.0 847 2.1 West Nusa Tenggara 63.4 39.8 523 -1.1 dominance of low-value cassava in East Nusa Tenggara 45.9 33.3 621 1.6 food crop exports implies that opportunities to significantly West Kalimantan 42.7 27.2 655 3.8 Central Kalimannan 36.9 22.3 853 1.5 increase value added-and rural South Kalimantan 58.9 29.4 574 0.4 incomes-are seriously EastKalimantan 35.3 21.6 702 8.1 underexploited. Exports of estate commodities are presented in table Central Sulawesi 44.1 33.4 836 6.6 18. The share of traditional export South Sulawesi 55.8 36.4 634 0.6 products (rubber, coffee, tobacco, Southeast Sulawesi 42.9 26.1 659 9.2 and tea) has been slowly Maluku 29.5 19.7 909 5.5 decreasing, with palm oil's export Ian Jaya 30.8 22.7 770 8.6 share remaining relatively small, primarily because of the high Source: ILO/UNDP 1989; SamIle Census, 'Income of Farm Households," 1984 domestic market share. Agricultural Census, CBS Series. The data also show that Indonesia's vast marine resources, particularly in east Indonesia, are seriously underutilized and heavily dependent on shrimp exports to markets that are becoming saturated. Offshore marine resources are virtually untapped, and a stronger fishing fleet is needed. A bilateral agreement with Japan may be needed to strengthen Indonesia's fishing fleet and ensure a reliable market. Without a modem fishing fleet and significant marine resources utilization, it will be difficult for east Indonesia to develop because marine resources dominate the region. As a result of the regional diversification of estate crops through a heavy government investment program, Indonesia has been broadening the spectrum and deepening the development of this subsector. The expansion of estate crops production, however, has mainly relied on area expansion, whereas productivity in smallholder rubber, coconut, coffee, and tea has remained stagnant. Furthermore, the Indonesian rehabilitation program for smallholders has not been successful. Consequently, unless measures are taken to improve smallholder productivity, the development of Indonesian agriculture will remain a critical issue in the near future. - 143 - Breaking Out of a Java- and Rice- TABLE 15. INDONESIAN EXPORT VALUE BY SECTOR, 1974-88 (MILLION USS) centered SHARE OF Agriculture NONOIL AGRICULTURE IN NON- TOTAL NONOIL YEAR AGRIC. AGRIC. OIL TOTAL EXPORTS EXPORTS Java consti- 1974 1,072.4 1,132.5 5,221.4 7,426.3 14.4 48.6 tutes only 7 percent 1975 909.1 882.7 5,310.8 7,102.6 12.8 50.7 of Indonesia's land 1976 1,283.5 1,258.9 6,004.2 8,546.6 15.0 50.5 area, and over- 1977 1,868.3 1,606.2 7,378.1 10,852.6 17.2 53.8 rea, l1978 2,054.8 1,603.0 7,985.4 11,643.2 17.6 56.2 reliance on this small 1979 2,399.6 3,026.9 10,163.7 15,590.2 15.4 44.2 island for major 1980 2,743.0 3,425.8 17,781.6 23,950.4 11.5 44.5 commodities repre- 1981 1,843.4 2,657.9 20,663.2 25,164.5 7.3 41.0 1982 1,605.8 2,322.4 18,399.1 22,327.3 7.2 40.9 sents a serious bias in 1983 2,087.7 2,917.5 16,140.7 21,145.9 9.9 41.7 resource utilization, 1984 2,517.7 3,352.0 16,018.1 21,887.8 11.5 42.9 signaling a need to 1985 2,538.0 3,330.8 12,717.9 18,586.7 13.7 43.2 1986 2,749.9 3,778.4 8,276.6 14,804.9 18.6 42.1 establish a regional 1987 3,096.0 5,483.5 8,556.1 17,135.6 18.1 36.1 diversification pol- 1988 3,830.7 7,706.2 7,681.6 19,218.5 19.9 33.2 icy. The only way Indonesia will be able Source: Monthly Statistical Bulletin of Indonesia, July 1980, July 1986, July 1991, Central Bureau of to sustain food crop Statistics. self-sufficiency over the long term is by broadening the land resource base for major food crops. This requires broadening diversification's definition beyond mere crop diversification to include such concepts as intensification, rehabilitation, and extensification, which in Indonesia include resettlement programs such as transmigration. This new definition broadens diversification to include not only changes in cropping and crop-livestock mixes, but also diversification in processing; final products; employment; the use of technology, resources, and energy; and the institutions that support these processes. Given Java's fast industrial growth and urbanization, there is increasing pressure to produce more food and industrial products in an ever-decreasing land area. There is strong competition for land, and agriculture has had to yield to the stronger sectors of the economy. A rough estimate suggests that about 30,000 ha (a significant part of which comes from fertile agricultural land) are converted annually to industrial, real estate, and other urban uses. Urbanization and industrialization also compete with agriculture for water. Over the next century, half of Java's present irrigation water is likely to be appropriated by urban and industrial uses. Thus, in the long run, food crop production in Java is highly vulnerable. Indonesia's transmigration program was initially designed to shift food production outside Java, and the support for transmigrants has focused on food production. About 2 million ha have been developed, yet the expected shift in food crop production has not been obvious. Although small-scale irrigation schemes were built in many regions, overall rice production has remained concentrated in large- scale irrigation areas in Java. The tidal swamp and swamp ecosystem irrigation schemes used in the transmigration program have not proven effective in sustaining food crop production. Most transmigration areas in upland ecosystems also do not provide a strong base for sustainable food production given the current level of technology available to farmers. Transmigration's opening of new land areas, mainly dryland and tidal swamp ecosystems, can be regarded as an example of both land resource and regional diversification, but lacking well-adapted technology, there has been little progress in productivity. Thus, land resource diversification needs the backing of a strong technology diversification capability. - 144- The Institutional Context for Diversification TABLE 16. EXPORT VALUE OF AGRICULTURAL PRODUCTS BY SUBSECTOR, 1974-88 (MLLION US$) Defining and implementing an FOOD LIVE- FISH- ESTATE Indonesian agricultural diversification YEAR CROPS STOCK ERIES CROPS TOTAL strategy have always been a "no man's 1974 51.2 16.8 92.3 912.1 1,072.4 land," for which no ministry or agency 1975 35.4 15.8 88.2 769.7 909.1 has been given clear responsibility. The 1976 14.5 22.7 131.4 1,114.9 1,283.5 Ministry of Transmigration is now seen 1977 18.9 20.1 163.0 1,666.3 1,868.3 as the lead diversification agency, but 1978 125.0 22.7 193.4 1,713.7 2,054.8 1979 74.8 41.3 236.8 2,046.7 2,399.6 diversification cuts across regions and 1980 55.0 25.1 226.4 2,436.5 2,743.0 sectors: a coherent strategy would include 1981 18.8 26.5 225.4 1,572.7 1,843.4 industry and infrastructure as well as 1982 26.7 24.2 249.4 1,305.5 1,605.8 1983 54.5 24.0 257.0 1,752.2 2,087.7 agriculture and would encompass the 1984 94.1 37.9 248.1 2,137.6 2,517.7 overcrowded inner islands as well as the 1985 133.5 35.8 259.4 2,109.3 2,538.0 "sleeping lands" in the outer islands that 1986 104.2 40.2 374.1 2,231.4 2,749.9 1987 135.8 53.4 475.5 2,431.3 3,096.0 the ministry has been assigned to open 1988 185.6 74.1 712.2 2,858.8 3,830.7 and settle. This task-basically narrow e T ic rather than barroad Source: Monthly Statistical Bulletin of Indonesia, July 1980, July 1986, extensification rather than broad July 1991, Central Bureau of Statistics. diversification-was earlier vested in the Ministry of Public Works, and, before that, in the Ministry of TABLE 17. EXPORT VALUE OF SELECTED Agriculture. The arbitrary division of FOOD CROPS, 1974-88 (LLION US$) tasks across ministries has confined the agriculture ministry's YEAR SAVA SEED TABLE FRUIT OTHER TOTAL diversification efforts to attempts to broaden crop and crop-livestock 1974 23.8 19.8 1.7 - 5.9 51.2 mixes, with no responsibility for 1975 23.5 6.4 2.1 - 3.4 35.4 1976 11.3 0.4 1.7 - 1.1 14.5 vertical or employment-related 1977 13.2 1.2 2.4 - 2.1 18.9 diversification. This does not 1978 119.4 2.4 2.7 - 0.5 125.0 necessarily mean that transmigration, 1979 69.1 0.9 4.2 - 0.6 74.8 irrigation, agricultural mechanization, 1981 427 3814 637 0.12 3.3 55.0 and the other vertical and horizontal 1982 16.0 4.8 5.1 0.1 0.7 26.7 dimensions of diversification should 1983 28.1 21.4 3.5 0.1 1.4 54.5 be assigned to the Ministry of 1984 33.0 48.7 8.6 1.5 2.3 94.1 1985 45.6 75.1 6.4 5.5 0.9 133.5 Agriculture. Better coordination 1986 49.8 30.1 15.2 8.9 0.2 104.2 among ministries is needed if each is 1987 89.8 14.0 16.1 15.1 0.8 135.8 to take advantage of the others' 1988 126.2 7.2 26.0 15.9 10.3 185.6 expertise to mount a full-fledged diversification strategy. Source: Monthly Statistical Bulletin of Indonesia, July 1980, July 1986, July Agriculture has always 1991, Central Bureau of Statistics. been-and will remain for the foreseeable future-Indonesia's leading sector for reaching economic and equity goals. Integrated regional development and diversification will be key tools in the country's agriculture-based growth strategy. This will require increased coordination that integrates planning across sectors and regions and strengthens both top-down and bottom-up capabilities. The most important aspect of top-down planning is regional development, and the most relevant sector in regional development is agriculture. Unless regional planning is coordinated by - 145 - sector, there is likely to be concentration rather than TABLE 18. EXPORT VALUE OF SELECTED ESTATE CROPS, 1974-88 diversification, because the MLLION USS) easiest way to develop is RUBBER, often to concentrate new COFFEE, PALM OIL investments in existing areas TOBACCO, &PALM COPRA or programs. YEAR TEA PEPPER KERNEL CAKE SPICES OTHER TOTAL As a first "take-off" 1974 656.2 24.3 166.9 25.9 7.6 31.2 912.1 step, it is necessary to 1975 545.5 22.9 156.0 24.7 4.8 15.8 769.7 establish a long-run regional 1976 869.0 47.0 139.3 33.4 14.1 12.1 1,114.9 1977 1,348.9 65.0 189.5 37.4 16.0 9.5 1,666.3 master plan, which should be 1978 1,358.7 68.7 210.4 34.1 18.0 23.8 1,713.7 carried out stepwise, first in 1979 1,690.1 46.6 211.6 41.3 18.3 38.8 2,046.7 the agricultural sector, to be 1980 1,992.8 50.1 265.9 47.4 23.7 56.6 2,436.5 1981 1,325.2 47.1 111.4 35.8 18.7 34.5 1,572.7 expanded later to other 1982 1,070.9 44.8 97.5 36.1 26.4 29.8 1,305.5 sectors. Two of the plan's 1983 1,429.4 52.0 114.6 31.8 30.8 93.6 1,752.2 major objectives should be 1984 1,773.4 64.2 76.0 14.2 38.5 171.3 2,137.6 shifting food production from 1985 1,456.9 78.4 243.3 30.5 36.6 263.6 2,109.3 shifting ~~~~~~~1986 1,691.6 113.7 113.1 35.7 56.6 220.7 2,231.4 Java to the other islands and 1987 1,733.4 148.2 176.8 39.4 70.3 263.2 2,431.3 identifying the agricultural 1988 1,961.6 144.5 333.9 47.4 70.6 300.8 2,858.8 research needed to bring this about, particularly the main Source: Monthly Statistical Bulletin of Indonesia, July 1980, July 1986, July 1991, Central issue of technology Bureau of Statistics. diversification. Institutional Problems in the Ministry of Agriculture One of the agricultural sector's major technical weaknesses has been a lack of civil and mechanical engineers in the Ministry of Agriculture. Technical problems are often handled by "agricultural engineers," who are usually graduates in such fields as economics, rural sociology, agronomy, livestock, or fisheries. This lack of engineering expertise is one of the major reasons the ministry focuses strongly on intensification and has had to reduce its irrigation arm to a mere subdirectorate. The ministry's approach to farm mechanization is equally weak, because there are no mechanical engineers working in agriculture, rendering Indonesia's farm mechanization the weakest among Asian nations. Farm machinery is often perceived as displacing labor, which may be temporarily true in the crowded areas of Java. Many Indonesian agricultural economists opposed the introduction of rice mills and sickle harvesting equipment in the 1960s, believing that these innovations would eliminate employment for rice pounders and ani-ani harvesters. In fact, this attitude toward farm mechanization hampered the development of agriculture in upland and dryland ecosystems in sparsely populated areas outside Java. Intensification in Java requires hand and minitractors, underscoring the real need to introduce mechanical power in agricultural extensification. Credit Institutions and Diversification One of the major constraints to the movement of farmers and rural firms toward diversification is the shortage of capital, which can be overcome by credit. Imperfections in Indonesia's capital markets call for government intervention through monetary and fiscal policy and the establishment of rural financial institutions. BIMAS, INMAS, and INSUS are the main institutions that provide capital, technology, and other services to farmers, usually as an integral component of commodity programs. The - 146- TABLE 19. BANK LENDING BY SECTOR, SELECTED YEARS (MILLION RP) TREND SECTOR 1977 1980 1985 1988 (%) Agriculture 270 (6.7) 526 (6.7) 1,656 (7.5) 2,891 (8.2) 18.3 M;ning 1,062 (27.0) 1,866 (23.7) 394 (1.2) 294 (0.8) -14.5 Industry 1,154 (29.4) 2,563 (32.5) 7,592 (34.3) 11,928 (34.0) 17.6 Trade 911 (23.1) 1,977 (26.0) 7,255 (32.7) 10,997 (31.4) 18.7 Services 319 (8.1) 482 (6.1) 4,183 (18.9) 5,978 (17.0) 23.1 Other 218 (5.5) 466 (5.9) 1,213 (5.5) 2,993 (8.5) 23.0 Total 3,937 (100.0) 7,880 (100.0) 22,157 (100.0) 35,081 (100.0) 17.1 Note: Percent of total in parentheses. Source: Bank Negara Indonesia. Small Investment Credit (KIK) and Small Working Capital Credit (KMKP) programs are other sources of capital for production of both agricultural (nonrice commodities such as pepper, tobacco, and clove) and nonagricultural (such as warehouses, rice mills, and water) commodities. There are various financial institutions and policies that provide capital for various purposes. The total amount of bank lending, by economic sector, is presented in table 19. Between 1977 and 1988, the amount of agricultural credit extended in nominal terms grew by about 18 percent annually, slightly higher than the growth of total bank lending. The proportion of bank credit that accrued to the agricultural sector, however, remained around 7 to 8 percent. The volume of agricultural credit in 1988 was Rp 2.891 billion, about 10 times the 1977 level. The shares of credit that accrued to industry, trade, and services in 1988 were 34.0, 31.4, and 17.0 percent, respectively, higher than in 1977. Thus, the share of credit for the agricultural sector was smaller relative to other sectors, except for mining, which implies that investment opportunities were broader in nonagricultural sectors. This situation, holding other factors constant, will increase a region's economic diversity. Before 1965, there were no modern institutional credit schemes available for small farmers in Indonesia. Their only source of credit was local moneylenders, who charged high interest rates and who remain an important source of credit in rural areas. In 1964/65 the BIMAS (mass-guidance) program was instituted after a successful experiment carried out by a team from Bogor Agricultural University in villages in the Karawang district. Under this program, farmers received extension services in conjunction with loans from Bank Rakyat Indonesia (BRI) for operational expenses and farm inputs such as fertilizer, seeds, and agrochemicals. The augmentation of rural capital as a result of the BIMAS agricultural credit program has played a major role in increasing rice output. Nonetheless, although almost all BIMAS credit was originally directed toward rice, the share of other crops, shrimp, and poultry has also increased since 1973. In 1986, around 11 percent of the Rp 14.2 billion lent by BIMAS accrued to palawija (table 20). BIMAS remains, however, a limited diversification tool, because it is still confined to specific commodities, such as rice, soybean, corn, sugarcane, shrimp, and poultry. In 1973 broader medium- and long-term financing was made available to farmers when Bank Negara Indonesia introduced the KIK and KMKP programs, which were not restricted to specific commodities. At that time there was no government measure enforced against noninstitutional farm credit. Instead, BRI extension activities were expanded to village units, and the development of farmer cooperatives was launched as a means of strengthening the rural financial market. The KIK and KMKP programs also aimed to increase the capacity of small business firms and farmers, with rural firms having the freedom to allocate capital as they wish. This setup is designed to facilitate economic diversification. The development of small-enterprise credit in Indonesia is presented in table 21. The volume of KIK, - 147 - KMKP, and KUPEDES (the general credit system to rural areas extended by BRI) credit has been TABLE 20. BIMAS LENDING FOR RICE increasing. In 1978, for example, KIK and KMKP AND PALAWIJA, 197348 (MILLION RP) credit constituted Rp 79 million and Rp 125 YEAR RICE PALAWIJA TOTAL million, respectively, out of a total rural small- enterprise credit volume of Rp 204 million. 1973 17,514 (99.2) 150 (0.8) 17,664 In 1965 all government-owned banks were 1974 41,320 (93.7) 3,400 (6.3) 44,720 1975 49,871 (86.3) 7,900 (13.7) 57,771 merged into one institution, the Bank Negara 1976 68,616 (88.5) 8,900 (11.5) 77,516 Indonesia (State Bank). This conglomerate banking 1977 58,707 (88.4) 7,700 (11.6) 66,407 institution had existed for only three years when in 1978 55,494 (89.1) 6,800 (10.9) 62,294 1979 50,019 (90.6) 5,200 (9.4) 55,219 December 1968, it was divided into the Central 1980 44,934 (88.9) 5,600 (11.1) 50,534 Bank and seven state banks, consisting of five 1981 45,061 (87.6) 6,400 (12.4) 51,461 commercial banks, one savings bank, and one 1982 49,931 (86.5) 7,804 (13.5) 57,735 bank. Bank RakyatIndonesia is one 1983 49,429 (85.6) 8,351 (14.4) 57,780 development bank. Bank Rakyat Indonesia is one 1984 13,364 (85.2) 2,330 (14.8) 15,694 of the five commercial state banks specializing in 1985 4,140 (86.3) 660 (13.7) 4,800 loans for agriculture, cooperatives, and rural 1986 12,585 (88.9) 1,573 (11.1) 14,158 development. BRI is Indonesia's main agricultural 1987 - - 137,000 credit institution, with the largest network of branches throughout the country. It is also the first bank to open branch offices in rural areas. BRI's Note: Percent of total in parentheses. main objective is to provide general commercial Source: BIMAS. banking services, primarily for the benefit of farmers, those involved in fisheries and small- TABLE 21. SMALL-ENTERPRISE scale industries, traders, and other CREDIT DEVELOPMENT IN INDONESIA, entrepreneurs. It also assists the government SELECTED YEARS (MILLION RP) in implementing its agricultural policies and rural achievement programs by administering 1978 1981 1985 1988 TREND (%) government credit and subsidy programs for KIK 79 336 946 1,301 19.7 rural development and supervises secondary KMKP 125 656 2,448 3,895 23.3 rural banks in accordance with Central Bank KUPEDES directives. BRI's subsidized credit is funded K - 6 11 16 10.6 by Central Bank liquidity credit, for which KMKP - 44 153 446 30.2 BRI pays 3 percent annual interest and Total 204 1,042 3558 5,658 228 charges its clients 10.5 to 15 percent. The 1 , 5 main credit issues are related to equity problems: the proportion of credit between (K = Small Investment Credit Program; KM"P = Small Working Capital Credit Program; KUPEDES = General credit system to rural rich and poor, formal and informal, rural and areas extended by Bank Rakyat Indonesia (1R1. urban, agriculture and nonagriculture. Source: Bank Negara Indonesia. Although Indonesia's economy has been growing rapidly in the past few years, the growth of small-enterprise credit for agriculture has not kept pace. A huge increase in loan portfolios, mobilized by banking institutions and the newly developing stock market, reflects the burgeoning economy that has resulted from government deregulation and export promotion policies.' Concurrently, according to Hazad and Basalim (1990), the proportion of small-enterprise credit, which was 20 percent before the June 1983 banking reform, dropped to only 3 percent as of October 1988. Despite government policy calling for commercial banks to allocate 20 percent of their loan funds to small-enterprise credit programs, it is difficult for banking institutions to carry out this task, even with a rather loose definition of small-enterprise credit-that going to enterprises with less than Rp 600 million in assets (land and buildings excluded). Government credit policy now focuses on allocating subsidized - 148 - credit primarily to sustain food self-sufficiency, rather than the commercial crops and agroindustry that can foster diversification and increase rural incomes. One important step that could be taken to increase rural credit availability would be to transform the village unit cooperative (KUD) into a rural bank with links to the BRI while retaining its cooperative nature at the village level. Toward a Strategy for Diversification The regional shift in production of agricultural commodities-except rice-over the last 18 years shows that Indonesia has broadened its agricultural resource base. The country's agricultural sector is diversifying, but the process is slow, greatly hampered by Java's entrenched rice monoculture and the lack of policy and technical support needed to allow diversification to reach its full potential quickly and effectively. Action is needed at several levels if diversification is to become an operational strategy for increasing production, productivity, and rural incomes. The most basic change may seem somewhat paradoxical: the Indonesian government needs to do both less and more. Indonesia's economy-wide movement toward a macroeconomic framework of deregulation and market incentives over the past few years is perhaps the most important step it can take to boost all sectors of its economy. Doing less encompasses continuing to remove bureaucratic hurdles and subsidies-from fertilizer to credit-that will free rural entrepreneurs, from farmers to nascent agroindustrialists, to make investment and production decisions based on real-world risks and rewards. This will require a steadfast political commitment; government agencies and those who benefit from the present system will have a strong interest in maintaining distortions that benefit them. The government must also do more. Its agricultural strategy is now fragmented among uncoordinated ministries and financial institutions whose functions and competencies show both great gaps and areas of overlap and duplication. This leads to redundancy and "turf protection" that block success, while allowing program initiatives to fail because their components cannot be coordinated. Thus, transmigration has had limited success in extending food crops outside Java, largely because the full support packages transmigrants need-from investment in infrastructure to capital resources (credit and agricultural machinery)-and the technological packages suitable for dryland and swamp ecosystems too often could not be assembled and delivered by the fragmented ministries and agencies who share responsibility for the program. An absence of mechanical and civil engineering skills in the Ministry of Agriculture has, similarly, crippled that agency's ability to plan infrastructure and contribute to farm mechanization. The ministry's lack of mechanical engineers is evident in the low level of "energy diversification" in agricultural development, making agriculture dependent on high population density to develop. It is also visible in the uncertain development of transmigration areas and the weaknesses in small-scale irrigation project implementation. Indonesia cannot focus on production and employment in agriculture alone, because there are much greater opportunities outside the narrow range of agricultural production. To launch an integrated effort in agricultural and rural development, it is necessary to develop a strong planning capability at the regional and national levels to develop agriculturally based regional urban centers throughout the country. Upstream and downstream industries should be developed along this line to ensure economies of scale and to impart positive external effects within and among regions. The financial system supporting rural development is equally uncoordinated. The plethora of-often subsidized-credit programs offered farmers and rural entrepreneurs has a poor record of making financing available when and where it is needed. Its collection record for the loans it does make has been equally poor. Revamping rural credit to make the banking system an effective change agent in Indonesia's agricultural evolution will be a major task. The agricultural sector should, therefore, be approached broadly, including support services such as trade, packaging, transportation, banking, extension, input supply, upstream industries that produce - 149 - farming machinery, production itself, and downstream processing. This would give the sector great potential for creating employment and improving income distribution. Improvements in technology and infrastructure will push farming system development and create employment opportunities in rural areas. The government's role in this diversification strategy should be to provide integrated technology packages and a macroeconomic atmosphere that uses market mechanisms to provide incentives while minimizing distortions. Arbitrary price supports for some commodities-and the neglect of others-are likely to prevent agricultural diversification from proceeding optimally. Income sources are less diversified in irrigated lowland ecosystems than in the uplands. Because irrigated lands cannot be greatly expanded, and with slow technological change and relatively constant real prices, farmers' incomes can increase slowly at best. To improve the situation, efficiency in factor inputs should be increased to reduce production costs; modern technology should be introduced as quickly as possible; and structural changes should be introduced to move rural employment out of agriculture to agroindustry and services. Production of secondary crops such as corn, soybean, and cassava should be shifted from Java to Sumatra, Sulawesi, and Nusa Tenggara, because these commodities cannot compete with rice and sugarcane at the level of technology now used by farmers. Sugarcane should be shifted to dryland areas in Java and outside Java since it cannot compete with rice on irrigated land. To speed the regional shift in food crop production, including rice, away from Java, technology must be diversified and investment in farm mechanization outside Java rapidly increased. Ultimately, this should also occur in Java as that island's agriculture is encouraged to shift from food crops to horticultural crops and from cattle to dairy cows. To improve and speed the diversification process-and especially to increase rice and sugarcane production outside Java-Indonesia needs to increase investment in infrastructure and land development to support production of these crops and speed technology generation and adoption for dryland or upland and swamp ecosystems outside Java. To further broaden the agricultural resource base and the spectrum of agricultural development, integrated commodity development should become a part of regional and area development programs. This approach will add a vertical dimension of agribusiness and agroindustry that would increase labor productivity and income equitability. Most of the potential for raising rural household incomes lies outside direct agricultural production, with nonfarm wage employment the most important potential source. Because the decline in demand for unskilled labor in the nonagricultural sectors will seriously impact rural incomes, it is important to strengthen income and employment generation potentials in the nonfarm and nonagricultural sectors by supporting investment in agro-based industries and, at the same time, encouraging agribusiness to create employment and expand nonfarm income sources. Finally, the agricultural diversification strategy should devote resources to adapting and developing farm machinery for Indonesia's various farming and ecosystems to increase productivity. This will involve phasing out the current heavy reliance on animal and human power. Greater efforts also need to be put into preparing off-shore fisheries, which will be a key to east Indonesia's economic development. In a parallel effort, a major thrust is needed to overcome stagnation in smallholder productivity in such estate crops as rubber, coconut, coffee, and tea. Rapid improvement in smallholder productivity will became a critical challenge in future agricultural development. Rehabilitation programs for smallholders should be given priority, and agroprocessing complexes for estate crops need to be further developed to diversify markets for these products. Last, large agricultural estates should be given greater flexibility to develop industrial complexes, because there is particular potential for the development of a paper industry by sugarcane estates and the development of frying oil, soap, and related industries by oil palm estates. - 150 - Note 1. Between the end of October 1988 and the end of December 1989, banking institutions increased the mobilization of money by 40.4 percent, reaching a volume of Rp 51.8 trillion. During the same period, stock market transactions grew 216 percent (see Hazad and Basalim 1990). References ILO/UNDP. 1989. Prospect for Labour Absorption in Agriculture in Repelita V. Report of ILO Study INSl84/006. Hazad, Hamzah, and Basalim, Umar. 1990. "Pengaruh Deregulasi Sektor Finansial Tahun 1983 dan 1988 Terhadap Mobilisasi Sana Masyarakat" (Impacts of Financial Sector Deregulations of 1983 and 1988 on Mobilization of Societal Financial Assets). Kompas (newspaper), February 12. Indonesian Central Bureau of Statistics. Various years, various publications. Jakarta. Indonesian Ministry of Agriculture, Directorate General of Estates, Subdirectorate of Data and Statistics. 1990. 'Estate Crops Statistics (1970-1990)." Jakarta. -, Directorate General of Livestock Services, Subdirectorate of Data and Statistics. 1990. 'Livestock Statistics, 1990." Jakarta. World Bank. 1988. "Policies and Issues for Diversification in Asian Agriculture." Agriculture and Rural Development Department. Washington, D.C. AGRICULTURAL DIVERSIFICATION IN INDONESIA: THE POTENTIAL ROLES OF RESEARCH, EXTENSION, AND IRRIGATION Ibrahim Manwan, Ahmad Dimyati, A.M. Fagi, and M. Oka Adnyana Over the past two decades Indonesia has made great progress in agriculture, with agricultural output having increased significantly for almost all crops. Indonesia has become self-sufficient in rice and no longer needs to spend a large amount of foreign exchange earnings to import rice. The agricultural sector continues to be a significant force in the country's economic growth, and its role has been greatly enhanced through the application of improved technology and management practices. Increased agricultural productivity and cropping intensity were achieved as a result of investments in research, construction and rehabilitation of irrigation systems, proper use of agricultural chemicals, improved crop varieties, extension activities, and other support schemes. Despite the success in attaining rice self-sufficiency, however, there is growing public concern over environmental degradation, widening income and prosperity gaps, underemployment, and sustainabiity of agricultural production. The question, therefore, is how to exploit the benefits of science and technology in agricultural production to achieve better farming, better business, and better living. The Indonesian government is paying increased attention to agricultural diversification-an approach that has long been practiced by Indonesian farmers aiming to increase land productivity, broaden their sources of food and income, intensify the use of land resources, and mitigate the unpredictable risks of farming activities. Diversification first became an agricultural development policy instrument during Indonesia's mid-1970s Second Five-Year Development Plan (PELITA II). Relatively little progress was made during PELITAs II and III, however, as efforts were concentrated on increasing rice production. Beginning with PELITA IV (1984-88), significant attention has been devoted to agricultural diversification. Diversified agriculture at farm, sectoral, and regional levels is commonly found in irrigated and rainfed wetlands, rainfed drylands, and swampy lands. Because of population pressure and the need to optimize the use of land resources, most small farmers have practiced diversified agriculture, growing multiple crops on their lands to maximize earnings and minimize risks. Various technologies and farming models that integrate several activities and/or crops in a farming system have already been made available, and research scientists continue to develop new technology packages to support agricultural diversification. Several nontechnological factors also determine the success of agricultural diversification. This paper discusses these determinants within the context of diversification's technological aspects. It assesses the status and challenges of opportunities for agricultural diversification in Indonesia and examines the contribution of research, extension, and irrigation activities to the overall agricultural diversification process. It reviews the findings of studies pertaining to soil survey and classification, land ownership and use, and farming systems in different agroecological zones. It also draws upon commodity studies, interviews, and field observations. Statistical data on agricultural commodity production, published by Ibrahim Manwan is director; Ahmad Dimyati is head, Research Programming Division; and M. Oka Adnyana is head, Development Research Group, all of the Central Research Institute for Food Crops, Agency for Agricultural Research and Development, Ministry of Agriculture, Indonesia. A.M. Fagi is director, Sukamandi Research Institute for Food Crops. - 152 - Indonesia's Central Bureau of Statistics (CBS), were studied and analyzed. Extension workers, agricultural service officers, and farmers were interviewed. The interviews and field observations were conducted in five districts of West Java and at the Upland Agriculture and Conservation Project in Salatiga, Central Java. Challenges and Opportunities Agriculture plays a major role in the Indonesian economy. More than two-thirds of the population lives in rural areas, over half of whom depend on agriculture. The sector accounts for 24 percent of GDP and more than 60 percent of the value of nonoil exports. Over the past two decades, agricultural output has grown by about 4 percent annually. Rapidly expanding rice production has provided much of this growth, for rice accounts for more than 40 percent of agricultural output, land use, and employment. Rice production rose from 12 million mt in 1969 to approximately 29 million mt in 1989. Food crops are the most important agricultural subsector, followed by industrial and estate crops, livestock, and fisheries. Additionally, the adoption of improved technology and management has been largely responsible for the significant increases in the production of major agricultural commodities during the past 10 years. Vast areas of underutilized land in Java and other islands provide scope for future investments in agriculture. Millions of hectares in the upland and swampy areas can still be made more productive, and the productivity and cropping intensity of irrigated lands can still be increased. Several problems, however, have to be confronted, including destruction of upper watersheds and forest lands and irrational application of inorganic fertilizers and pesticides, which cause nutrient imbalances and environmental pollution. The return to each unit of input such as fertilizer and pesticide is declining. Thus, despite agriculture's great potential, serious efforts have to be undertaken to sustain productivity, especially under steady population pressure. All these technical and socioeconomic concerns have to be considered in pursuing agricultural diversification, which is expected to become the main vehicle for achieving increased productivity and cropping intensity in the future. Scope for Diversification The Indonesian government's strategy envisions three dimensions to diversification: vertical, horizontal, and regional (Hedley 1988). Vertical diversification includes processing and the multiple use of agricultural products and by-products to improve their quality and value, generate higher prices, and increase comparative advantage. Upstream and downstream agroindustries play an important role in this aspect. The horizontal dimension-for which vertical diversification is often a prerequisite-involves crop diversification. Many nonrice food crops, vegetables, fruits, fish, and dairy products deteriorate easily under storage and transportation. Immediate processing after harvesting and before transportation or storage is often necessary. Because of the limited absorption capacity of local markets, increasing the production of agricultural commodities above local demand may result in locally depressed prices. Diversified processing and utilization increase the market potential of agricultural products. Horizontal diversification can involve a farm unit, irrigation command, watershed area, or agroecological zone. Diversification at one level may be combined with specialization at a lower level. Thus, national diversification may be implemented through regional specialization or regional diversification. The model used in pursuing diversification at any level is expected to be the one with the highest comparative advantage technically and socially, and it may differ by region. In Java, farm-level diversification is most suitable, as small farm units require intensive cropping. In other islands where man-land and land-capital ratios are lower, regional specialization in mechanized production of food crops, estate crops, and livestock involving the private sector is considered the best option. - 153 - Agricultural diversification requires a thorough inventory and analysis of Indonesia's biophysical, social, and economic resources. In addition, global determinants that affect the supply, demand, and price of commodities should be predicted and anticipated. Land Utilization TABLE 1. AGRICULTURAL AND In 1986, 144 million ha of the country's 2 FORESTRY LAND AREA, 1986 million km2 of land constituted forest lands-including AREA reserved, conservation, and production forests and LAND TYPE ('000 HA) parks-and 66 million ha were devoted to agricultural activities (table 1). Thirty-eight percent of the Agnculture 4,836 agricultural land was devoted to woody plants, 17.1 Rainfed bareland, garden 12,235 percent to estate crops, 23.5 percent to seasonal crops Steppe pasture-grass land 2,922 and bareland under rainfed dryland conditions, 15 Temporarily unutiized land 9,508 Snailholder tree plants 19,688 percent to wetland paddy, and 5.6 percent to steppe Estate tree plants 8,910 pasture, leaving only 0.4 and 0.3 percent, Paddy field 7,774 respectively, for dike and freshwater ponds (Central Coastal dike 211 Freshwater pond 121 Bureau of Statistics data). These figures indicate great Total 66,205 potential for the production of food and estate crops, particularly on barelands and "sleeping lands" largely Forests devoted to seasonal animal grazing with potential for Park an reserved 310837256 crop agriculture. Limited production 30,525 Fish culture in paddy and freshwater ponds is Permanent production 33,867 less developed and limited to certain areas. The Conversion 30,537 potential for fish production in natural lakes and man-made reservoirs is vast. Marine fish production can be increased through better fishing technologies Source: Central Bureau of Statistics 1989. and storage and transport facilities. More work is required to improve fish production, utilization, and trade. Regarding seasonal crops, the degree of water control and availability affects yield potential and farmers' ability to adopt and use improved technology. Inadequate water control is a physical impediment that aggravates the effects of the entire range of technical and economic constraints. Based on water control and availability, Indonesia's agricultural lands are divided into four major agroecological zones: (1) irrigated wetland, (2) rainfed wetland, (3) rainfed dryland, and (4) tidal and nontidal swampland. Substantial variation in the onset and withdrawal of the west monsoon rains and in rainfall intensity is the major constraint to successful rice-based farming under rainfed conditions. Land is the most important asset in the rural areas, but many farmers do not own land. Land scarcity, largely the result of the increasing urbanization of Java, has fostered inequitable land and income distribution, while creating landless farmers and nonfarmer landowners. Control over agricultural land has increasingly accrued to those who have capital and access to credit and are often not supportive of diversification. For example, wealthy landowners in the Jatiluhur irrigation area are not interested in cultivating soybean, although sharecroppers working their holdings perceive its profitability. The owners may allow growers to plant soybean, but if irrigation water arrives, they are often required to sacrifice the unharvested crop and begin preparing the land for growing the rice favored by the landowners. - 154 - Soil and Water Conservation Indonesia's rapidly growing population generates continued pressure to raise agricultural production, which can be dealt with in four ways: intensifying crop production in cultivated areas, opening new lands, rehabilitating degraded land, and diversifying into new commodities. Intensification and the opening of new-often marginal-lands can, however, be costly, both economically and ecologically. New and better technology derived from well-planned and well-organized research is needed, particularly to meet the threat of erosion. In Java, for instance, about 75 percent of the land is mountainous or hilly. These mountains and hills straddle the central parts of the island from east to west and are the major source of water flowing to the rice fields through rivers and reservoirs. Population and economic pressure has driven many farmers to these steep lands to plant annual food crops such as rice, cassava, maize, and vegetables-without following proper soil conservation techniques-instead of perennial tree crops for which the terrain is more suitable. This has been occurring for years and has been a major factor in the degradation to critical levels of 22 of Indonesia's 33 watersheds-I1 of which are in Java. The Indonesian government pays serious attention to the rehabilitation of degraded land. In the late 1970s, Indonesian agricultural research and development strategists began looking beyond the primary goals of productivity to consider factors such as stability, sustainability, and equity. They considered that (a) the intensive production systems of Java and the extensive low productivity systems of the outer islands are vulnerable to irreversible environmental degradation; (b) intensively cultivated irrigated areas are susceptible to damage from improper soil and input management and cultural practices; and (c) extensively used rainfed fragile soils and swampland are subject to degradation from erosion and the loss of organic matter. The watershed areas in Java were considered highly susceptible to degradation as a result of deforestation and poor cultivation practices. The major initiative to prevent further degradation has been the introduction of high-sustainability/high-productivity systems, so that increased agricultural production is not in conflict with environmentally sound social goals. About 70 percent of Indonesia's agricultural area is rainfed, and large areas are in transition from shifting cultivation to sedentary agriculture. Erosion, soil acidity, aluminum toxicity, and removal of organic matter and nutrients have been the key concerns and constraints regarding sustainable farming in these areas. Swamp areas pose another set of challenges of the same magnitude. To achieve and sustain self-sufficiency in food grains, production strategies were initially directed to fertile areas with ensured infrastructure such as irrigation. In the absence of sound management practices, these sawah areas are vulnerable to nutrient deficiency and toxicity; pesticide residue buildup; and the outbreak of disease, insect biotypes, and new insect pests. Intensive use and continuous cropping without considering the nutrients removed from the soil by the crops will result in the depletion of natural fertility and capacity to produce. To some extent, holistic approaches are being implemented to optimize the use of available resources. Unless watersheds are properly managed, however, the irrigated rice lands in Java and other islands face various problems. Thus, interdisciplinary and interinstitutional approaches that stress hydrological concepts are needed. Soil Survey and Mapping Efficient agricultural investment and the development and use of new technology require a systematic evaluation of soil resources. Indonesia's Center for Soil Research, in cooperation with FAO, has inventoried, mapped, and evaluated land resources at the reconnaissance level (1:250,000). Their findings can be used in agricultural planning at the provincial level and in the selection of potential and priority areas for agricultural diversification. Smaller-scale soil maps (1:10,000 or 1:5,000) are needed for hilly land with heterogeneous landforms for designing soil and water conservation measures and - 155 - farming systems. A detailed soil map, along with a detailed inventory of climatic factors, can be used to determine the amount of land suitable for different classes of crops and to conduct agroecosystem analysis. This is important in identifying constraints and opportunities in target areas. To meet this need, the Regional Physical Planning Program for Transmigration has developed land system and land suitability maps, land use and forest status maps, and recommended development area and land status maps for most areas outside Java. Land system and land suitability maps describe land systems within each province and their suitability for the production of various agricultural commodities. Land use and forest status maps describe land utilization in various land systems and their land status. Diversified Agriculture in Perspective Agricultural diversification is a highly complex process. Government policies and programs play an important role in fostering successful diversification. Price policies, external and internal support, taxes and subsidies, foreign exchange rates, industrial protection, inflation, export promotion, and budgetary priorities directly affect the development of agricultural enterprises. Policy interventions can have positive or negative effects on agricultural diversification. Government policy also acts to buffer or reinforce pressures and opportunities in world markets. Agricultural diversification requires flexibility in the choice of commodities and in the technologies to support those choices. Research institutes, therefore, work to make available a wide range of technology options that are accessible and affordable to farmers. An effective and efficient institution- building program is needed for Indonesia's agricultural research system to generate improved technologies to support agricultural diversification. Rice research has long received strong support from the government, but more attention and larger research resource allocations are now being given to other crops as well. A great variety of technologies is available to increase productivity, cropping intensity, and the quality of the nonrice crops that will play an important role in expanding agricultural diversification. Because marketing leads farmers and rural entrepreneurs toward activities with better market and profitability potentials, it has been and will continue to be important in promoting new agricultural commodities for diversification. Several other factors determine the success of this market-oriented diversification process. Better infrastructure reduces marketing costs and improves profit margins for farmers and household entrepreneurs in rural areas. Financial assistance for traders, entrepreneurs, and farmers plays a central role. Lack of this assistance affects the rate of adoption of improved practices and technologies introduced to farmers and small entrepreneurs. Experience indicates that sustainable agricultural diversification needs support from six different standpoints: (1) government policy, (2) quality and suitability of new technologies, (3) farmers' response to change, (4) effective extension activities, (5) participation and motivation of farmers, and (6) other factors such as market structures for agricultural commodities. Studies on the interactions of these factors are important. Wide-ranging analysis of these factors and how technology can improve farmer welfare is essential. Moreover, without the necessary political will, the change from traditional to modem agricultural diversification will not proceed as quickly as desired. Comparative Advantage Analysis Analyzing the comparative advantage of commodities is essential to planning and policymaking aimed at raising the efficiency of agricultural production. Comparative advantage analysis is needed to help policymakers, planners, and researchers set agricultural diversification priorities for crops in different agroecological zones and develop the technologies needed to promote them. - 156 - Agricultural diversification includes introducing new crops and production systems. If it is to succeed in the long run, it is essential that these crops be profitably produced without government subsidies. From an efficiency perspective, these crops and systems must be both privately and socially profitable. Table 2 shows the private and social returns to soybean, corn, and cassava production. Cassava production in East Java and Lampung provinces is privately and socially profitable-the social profitability of cassava production in Lampung is almost double that of East Java, whereas private profitability is about equal. Under an export regime, corn is not privately and socially profitable in Aceh and socially unprofitable in Lampung. Under an import substitution regime, corn is socially profitable in all regions, but privately unprofitable in Aceh. The derived domestic resource cost (DRC) of corn production in East Java is 0.99 (table 3), making even the social profitability of corn production in East Java questionable. The derived DRC of corn production under the export regime in Lampung and Aceh is 1.26 and 1.70, respectively. These values indicate that the DRC of corn production in the two regions exceeds its value added, making corn production in these regions inefficient. Research is needed to develop technology that can increase productivity without neglecting concerns about environmental sustainability or increasing production costs. Soybean production is economically and socially TABLE 2. PRIVATE AND SOCIAL RETURNS profitable in East Java and Aceh, TO SOYBEAN, CORN, AND CASSAVA but socially unprofitable in PRODUCTION UNDER SELECTED REGIMES Lampung, mainly because of the SOYBEAN CORN CASSAVA high opportunity cost of land and (IMPORT) (IMPORT) (EXPORT) (EXPORT) low yields. Soybean production, Ear Java however, is privately profitable in Profit (Rp/ha) all regions, mainly due to the Private 224,664 42,546 42,546 259,603 floor price system protecting its Social 197,634 160,039 1,380 214,800 domestic price. One reason for Policy effect 27,030 -117,493 41,166 44,803 the high soybean profitability in Lampwg East Java is the low production Profit (Rp/ha) Private 46,990 2,227 2,227 254,437 costs that reflect the low wages Social -58,221 79,985 -47,353 434,571 paid in East Java compared with Policy effect 105,211 -77,758 49,580 -180,134 Aceh and Sumatra in general. High yield per hectare is another Aceh factor that makes soybean Private 64,768 -33,573 -33,573 n/a production in East Java more Social 69,166 12,574 -107,920 n/a efficient. Soybeans in Lampung Policy effect 4,398 -46,147 74,347 n/a require more input, particularly n/a Not available. labor and fertilizer, than in the Source: Adapted from Altemeier and Gijsbers 1988. two other regions. Low yields also render soybean production in Lampung socially unprofitable. The DRC coefficient of soybean production in Lampung is 1.28, indicating that soybean is inefficient in terms of resource allocation unless low-cost technology can increase yields and reduce the DRC coefficient. Another way of evaluating the comparative advantage of a crop or production system is by comparing the minimum yield of one crop that produces a net profit equal to that of others. The Central Research Institute for Food Crops (Manwan et al. 1989) conducted a simple analysis of comparative advantage in soybean production with respect to other palawia crops such as corn, peanut, and mungbean. Rice is also included in the analysis to provide an idea of how soybean competes with rice, but it is not aimed at replacing rice with soybean in yearly production systems. Instead, soybean is viewed - 157 - as one component of diversified agriculture or a farming system, and TABLE 3. DOMESTIC RESOURCE COST (DRC) COEFICIENTS the development of soybean FOR SOYBEAN, CORN, AND CASSAVA PRODUCTION production must consider the SOYBEAN CORN CASSAVA sustainability of rice production. (IMPORT) (IMPORT) (EXPORT) (EXPORT) Results of the analysis under different scenarios are presented in tables 4 Laffipun 1.28 0.74 1.26 0.34 and 5. Aceh 0.87 0.95 1.70 nla On irrigated wetlands, soybean must yield more than 2 n/a Not available. mt/ha to provide a net income Source: Altemeier and Gijabers 1988. comparable to that of rice, which can be attained at the present level of technology. A soybean yield of 1.1 mt/ha, the national average, produces profits comparable to those of peanut. However, when compared with mungbean, a crop with good endurance, soybean yields must be 1.8 mt/ha to be equally profitable. On rainfed lowland and upland areas, the expected yields of soybean are lower than the standard required to achieve an income comparable to rice, corn, and mungbean in irrigated wetland. Peanut has comparative advantage on these two types of land over irrigated wetland, meaning that soybean should produce higher yields to provide an equal income. Comparative advantage analysis facilitates the process of making choices among different commodities or different farming models. Thus, it can play an important role in setting agricultural diversification strategies. Present and Future Direction The first four of Indonesia's five-year development plans gave priority to agriculture, particularly rice production. This priority continued in PELITA V (1989-93). Expansion and improvement of irrigation facilities have ensured water availability and supported the expansion of new rice areas and the extensive use of modern rice technologies at the farm level. This was accompanied by a disproportionate allocation of resources to research, extension, credit, subsidies, and other supporting services in favor of rice production. Considering the limited resources and the need for adequate rice supplies, emphasis on rice was a rational strategy. This emphasis, however, has led to a substantial disparity between the technology available for rice and for other crops. In addition, almost all irrigation facilities were designed mainly for rice, although many farmers grow two or more commodities on the same piece of land at the same time to reduce risk, diversify income sources, and maximize land use. The introduction of other crops in the irrigated areas is possible given proper technology. Proper use of irrigation water may be able to increase crop stability, cropping intensity, and farmer incomes. The distribution of major commodities in Indonesia is not well patterned. Crop sequences representing the distribution of major food crops in irrigated areas are summarized in table 6. The cropping patterns commonly found in irrigated areas are rice-rice-palawia, rice-rice-vegetable, and rice-rice-rice, sometimes interrupted by sugarcane and tobacco. Different cropping patterns are frequent within irrigation areas, causing difficulties in water allocation during dry seasons. The traditional cropping pattern on rainfed wetlands was rice-palawUja. Introduction of wet-season gogorancah rice and walikjerami rice in the dry season has changed the pattern to gogorancah rice-walik jerami rice-palawija in several areas (table 6). Crop distribution on rainfed drylands is the most complicated because of heterogeneous land capability and land suitability classes, aggravated by high variability in rainfall patterns. In the - 158 - TABLE 4. SOYBEAN YIELD REQUIRED TO ACHIEVE COMPARATIVE ADVANTAGE OVER OTHER FOOD CROPS IN IRRIGATED AND RAINFED AREAS PRODUC. NET FARM YIELD PRICE COST INCOME (KGIHA) (RP/KG) (RP/HA) (RP/IHA) (RP/HA) In imgated lowlands Rice 6,000 250 1,500,000 600,000 900,000 (El) Corn 3,500 200 700,000 475,000 445,000 (E2) Peanut 1,200 1,000 1,200,000 225,000 750,000 (E3) Mungbean 900 700 630,000 300,000 330,000 (E4) Soybean 600 (B5) 350,000 (D5) with respect to Rice 2,083 (Fl) 1,249,800 900,000 Corn 958 (F2) 1,099,800 445,000 Peanut 1,133 (F3) 679,800 750,000 Mungbean 1,325 (F4) 795,000 330,000 In rainfed lands Rice 4,000 250 1,000,000 625,000 375,000 (El) Corn 3,000 200 600,000 500,000 100,000 (E2) Peanut 1,200 900 1,080,000 600,000 480,000 (E3) Mungbean 800 700 560,000 325,000 235,000 (E4) Soybean 600 (B5) 325,000 (D5) with respect to Rice 1,166 (Fl) 700,000 375,000 Corn 708 (F2) 625,000 300,000 Peanut 1,342 (F3) 1,075,000 750,000 Mungbean 933 (F4) 560,000 235,000 Note: Fl = (E1+D5)/B5; F2 = (E2+D5)/B5; F3 = (E3+D5)/B5; F4 = (E4+D5)/B5. Source: Manwan et al. 1989. Jratunseluna watershed area in Central Java, the common food crops planted are upland rice, corn, cassava, groundnut, and vegetables. Tree crops are rarely grown on the uphill parts of this watershed area and are usually planted downhill close to villages in homeyards or gardens. Cropping patterns in tidal swamp areas differ according to farmers' origins. Native inhabitants and spontaneous transmigrants usually use coconut-based farming systems with extensive cultivation, whereas government-sponsored transmigrants use more food-crop-based systems with intensive cultivation. The policy shift toward supporting crops other than rice under PELITA V has led to some gains in increasing the production of other crops. The intensification programs for crops such as soybean and corn are based on the lessons learned from rice. Farmers in several areas are growing crops such as shallot and garlic, or even perennial crops such as banana and citrus in their paddy fields to increase their incomes. In some cases they have to convert irrigated wetlands into sorian ditch and dike systems, or even to completely rainfed agriculture. Technology Development and Transfer The development and transfer of modem rice technologies have dominated Indonesian agriculture for the last three decades. The success of this effort has brought self-sufficiency and served as a catalyst - 159 - TABLE 5. SOYBEAN YIELD REQUIRED TO ACHIEVE COMPARATIVE ADVANTAGE OVER OTHER FOOD CROPS IN UPLAND AREAS PRODUC. NET FARM YIELD PRICE COST INCOME (KGIHA) (RP/KG) (RPMHA) (RP/HA) (RPIHA) Nonpodsolfic soils Upland rice 2,500 250 625,000 425,000 200,000 (El) Corn 3,500 200 700,000 400,000 300,000 (E2) Peanut 1,300 900 1,170,000 500,000 670,000 (E3) Mungbean 900 700 630,000 350,000 280,000 (E4) Soybean 600 (B5) 350,000 (D5) with respect to Rice 917 (F1) 675,000 325,000 Corn 1,083 (F2) 650,000 300,000 Peanut 1,700 (F3) 1,020,000 850,000 Mungbean 1,050 (F4) 630,000 280,000 Podsolic soils Upland rice 2,000 250 500,000 450,000 50,000 (El) Corn 2,500 200 500,000 425,000 75,000 (2) Peanut 1,000 900 900,000 525,000 375,000 (E3) Mungbean 700 700 490,000 375,000 115,000 (E4) Soybean 600 (B5) 375,000 (D5) with respect to Rice 708 (F1) 550,000 175,000 Com 750 (F2) 600,000 225,000 Peanut 1,250 (F3) 750,000 375,000 Mungbean 817 (F4) 490,000 115,000 Note: Fl = (El+D5)/B5;F2 = (E2+D5)/B5;F3 = (E3+DS)1B5;F4 = (E4+D5)1B5. Source: Manwan et al. 1989. for changes in the attitudes of farmers and their readiness to adapt improved varieties and other related technologies. This achievement should serve as a foundation for achieving similar results in the diversification into other commodities. Indonesia's Agency for Agricultural Research and Development has generated a wide range of technologies related to various commodities and problem areas to meet the needs of farmers ready to diversify. Rapid progress in technology generation and transfer has been a catalyst for increasing productivity and cultivated areas. Indonesia's need to spur diversification means that these efforts must be extended to a wider range of commodities. This does not signify reducing allocations to rice research and technology development-Indonesia must sustain, or increase, its ability to produce rice to keep pace with the growing demand from a growing population-but the amount devoted to other commodities must be increased. Although soybean has recently been accorded priority second only to rice, an agricultural diversification program requires that more weight be given to research, extension, financial support, marketing, and other internal and external mechanisms to support other crops. Research resources should be flexible enough to respond to the dynamic and changing needs for agricultural technologies, and there should be much greater flexibility in the assignment of professional staff to commodity-oriented and problem area research activities. This will be difficult, as most extension staff are specialized and tend to become even more specialized as their careers advance, particularly personnel in highly staffed research programs such as those focusing on food crops. - 160- TABLE 6. MAJOR FOOD CROP SEQUENCES IN INDONESIA IRRGATED AREA PROVINCE CROP SEOUENCE Mojokerto East Java Rice-palawija Ngawi East Java Rice-rice-tobacco-soybean Rice-rice-rice Rice-rice-corn Lumajang East Java Rice-rice-rice Rice-rice-comn Rice-palawija Tegal and Brebes Central Java Sugarcane-palawija-rice-rice Sugarcane-palawija-rice-shallot Sugarcane-vegetable-rice-shallot Rice-rice-shallot Palawija-rice-shallot Palawija-rice-palawija Majalengka, Garut, Kadipaten, and Sumedang West Java Rice-nice-soybean Soybean-rice-rice Soybean-rice-rice Soybean-rice-soybean Kuningan West Java Rice-rice-sweet potato Rice-sweet potato Sweet potato-rice-palawija CROPPING PATTERNS IN SOUTH SUMATRA AND JAMBI AVERAGE FARM FARMER TYPE SIZE (HA)& CROP Local residents Muara Telang, S. Sumatra 5.86 Rice, coconut, banana, palawija Rantau Rasau, Jambi 2.70 Rice, banana, coconut, palawija Penuguan, S. Sumatra 2.79 Rice, banana, coconut, palawija Migranus (Buginese) Sungai Senang, S. Sumatra 3.98 Rice, coconut, banana Sungai Cawang, S. Sumatra 3.50 Rice, coconut, banana Muara Telang, S. Sumatra 4.82 Rice, coconut, pineapple, banana, palawija Penuguan, S. Sumatra 5.14 Rice, coconut, banana, palawija Teluk Bayur, S. Sumatra 4.72 Rice, banana, coconut, citrus, coffee, palawija Mendahara, Jambi 4.86 Rice, coconut, banana, coffee, palawija Governent-sponsored ransmigranss (Javanese) Purwosari, Delta Upang 2.98 Rice, coconut, banana Damarwulan and Bintaran, Saleh Delta 1.25 Rice, banana, corn Bangun Karya, Rantau Rasau 2.49 Rice, banana, coconut, palawija Telang 1, Unit 1 2.08 Rice, cassava, banana, palawija, coconut Telang 1, Unit V 2.22 Rice, cassava, banana, palawija a. Includes wetland rice and homeyards/homegardens. - 161 - Technology generation for farm-level diversification has occurred through farming systems research activities. Rice-based cropping systems research was initiated in 1973, and models of cropping systems for various agroecological conditions have been developed. Early on, little attention was given to perennial crops, fish, and livestock, although farmers in many places had been integrating these components into their farming activities. More holistic farming systems research was initiated in 1983. Research topics cover all farm-related activities, from crop production to marketing, animal and mechanical power, and soil and water management. Farming models for several agroecological conditions have been developed. For example, to prevent further conversion of irrigated wetlands without sacrificing farmer incomes, farming models have been developed to achieve high rice production in combination with other crops/commodities with higher value added potential. Continued research on several commodities will be important to spur diversification. Water management research is needed to increase the efficiency of water use. Improved technologies for major agricultural commodities are also essential to provide farmers with options not available for previous crops or varieties. Furthermore, improved cultural practices and new processing methods change the comparative advantage of commodities. One important channel for transferring the results of this research is the extension service. Within Indonesia's Ministry of Agriculture, however, extension agents are still housed in different directorates general, based on commodity categories such as food crops, industrial and estate crops, livestock, and fisheries. Extension activities for farming systems and agricultural diversification are still weak, and the responsibility for extension activities promoting diversification has not been clearly delineated. Extension activities have also been heavily focused on rice production, and the activities related to rice have been better equipped and better staffed than those for other commodities. A slight change in current policy-recognizing the role of other crops in sustaining rice self-sufficiency-has been made in which other crops such as soybean are receiving more attention. This policy shift, however, has been slow in reaching implementing personnel at the village level where actual farming activities occur. Surveys conducted for this paper show that district- and local-level extension personnel have not been prepared for supporting agricultural diversification and that most extension materials are still related to rice production. Out of 40 extension agents interviewed, 37 reported that rice production technologies remain the primary topic in their field discussions. Few agents stated that they dealt with other commodities such as palawMa, vegetables, and fish, and even fewer with livestock and estate crops. Rice also dominated group meetings and field demonstrations. Much more time was allocated to other commodities during home visits during which the farmer is requested to set the agenda for discussions. Field extension workers are also not well informed about technologies for commodities other than rice and a few other major palawija crops, such as soybean and maize. District and subdistrict extension offices, the main source of information for field agents, offer little advice on other crops. Farmers interviewed complained about the need for more information on nonrice commodities and on other aspects of agricultural production such as seed production, postharvest processing, and market information. Many field agents do not have practical training in farming systems technologies, and courses on farming systems usually emphasize theoretical aspects. Field agents' primary source of knowledge and skill vis-a-vis farming systems is their own unguided experience. Most field extension workers reported low farmer participation in extension activities, especially in field discussions and group meetings where extra effort is required for farmers to attend. When this matter was brought to farmers' attention, they responded that extension activities were routine and that the same things were discussed repeatedly, especially regarding rice. Many other problems hamper effective extension activities at the rural level: lack of transportation for agents to visit farms, lack of facilities for field demonstrations, and unfavorable field conditions have prevented field extension workers from aggressively pursuing activities. Although research and extension - 162 - activities have tried to shift their focus from rice self-sufficiency to more diversified agriculture, the vehicles currently available are still inadequate to facilitate rapid progress. Role of Irrigation Indonesia has some of the most sophisticated irrigation systems in Southeast Asia. Technical irrigation systems cover more than 1.6 million ha, or about 39 percent of total irrigation systems. Nearly 60 percent of irrigation systems are located in Java. Cropping system experiments were conducted in Indramayu, West Java, between 1975 and 1978 to determine the potential for diversified irrigation use. Areas that receive seven to nine months' irrigation annually and used the walikjerami method for a second rice crop were able to shorten turnaround times so that soybean could be planted after rice using zero tillage. In areas receiving five months' irrigation, the use of the gogorancah method for the first rice crop made it possible to grow walikjerami rice in the dry season and an additional crop such as cowpea as a third crop. The overall data indicate that modification of cultural practices could be used to increase the efficiency of available irrigation water to improve productivity of land and increase net returns by increasing cropping intensity. Similar experiments were conducted at the Sukamandi Experimental Farm and in farmers' fields in Bekasi between 1979 and 1983. Adoption of recommended cropping patterns has been slow because the timing of irrigation water delivery is often unreliable. Giving farmers the confidence to try such new patterns requires much better irrigation management and interagency coordination to schedule and monitor reliable water deliveries. In the Maneungteung irrigation system in Cirebon, West Java (21,000 ha), farmers have developed their own technology for using residual soil moisture. They increase the level of the water table by creating an artificial confining layer that keeps the water in the field. Farmers essentially create a small pond with plants growing out of it, and irrigation practices during the wet season are planned to maintain the pond at a certain level-when it rains they drain the pond, and during dryer periods they irrigate. During the first dry season, farmers decide whether to maintain the pond for another season so that they can grow rice. In the absence of rainfall, irrigation water is used to recharge the pond. Other common practices include irrigating only one to four times during the dry season. Thus, farmers use subirrigation techniques, through which they survive the uncertain irrigation allocations of water masters and ditch tenders. Using this method, cropping intensities in the Maneungteung irrigation system range between 220 percent when water is scarce to 350 percent when water is abundant. This type of liberal irrigation water management, however, causes irregular cropping sequences. Irrigation commands sometimes cover more than one district; the Jatiluhur system, for example, covers the districts of Bekasi, Karawang, Subang, and Indramayu. This can lead to even greater unreliability, as irrigation committees work only during the planning of irrigation distribution and assume no responsibility after the plan is agreed upon and signed. Thus, there is no authority that can change water delivery to a given district, even if its actual planted areas are less than projected, or that can reallocate excess water to other districts where planting exceeds projections. The management of small-scale irrigation systems begins at the village level, where a farmer designated as a village water master collects information on the amount and mix of crops grown in the village portion of the irrigation block. He or she then determines the amount of water that should be allocated to the block. As management begins at the village level, so do the problems. Village water masters usually have little education, and their reports are very rough estimates, which, when transferred up the system, often lead to excessive water being diverted to the blocks. Agricultural diversification in irrigated areas has great potential, but is hampered by poor management systems. Diversifying from rice to palawija crops in rice-based irrigation systems during - 163 - dry seasons requires vast increases in managerial input. Recently, there has been considerable discussion about whether it is necessary to build additional irrigation canals to facilitate diversification to palawija crops, which use much less water than the 1.5 1/ha/s needed for rice. This paper concludes that management shortfalls are a much greater constraint than those related to technology or hardware, around which farmers have shown they can work. Shifting much of the responsibility for irrigation management to farmers themselves-in addition to being the least costly method for irrigation authorities and the government-may be the best short-run way to improve management. Farming Systems Research Farming systems research traditionally carried out within Indonesia at various agricultural research centers for different commodities and disciplines was largely directed by policy and circumstances (McIntosh 1986). However, the farming systems research approach has been gradually improved to meet the emerging need of farmers to improve income and optimize land resource use as well as increase agricultural production. To meet these objectives, research has gradually become comprehensive, covering various commodities and farm activities. The main thrust of farming systems research has been accomplished, and the results are widely acclaimed. The methodology has been developed and widely used, and there are trained research staff at various levels of operation. The farming systems approach to research has evolved in response to the notion that traditional research did not always generate appropriate farm technology as quickly as possible. In the current approach, it is assumed that appropriate technology can be generated and diffused through correct problem identification and implementation of research by competent researchers in cooperation with farmers and extension agents. Thereafter, technology design and testing in cooperation with national and regional extension services will be strengthened. Farming systems research in Indonesia has evolved gradually, from the cropping pattern testing stage at research stations in 1971 to the current complex, multisystem holistic projects with an agroecological orientation. Based on massive amounts of data and experiences gained over the past 15 years, future work will systematically deal with whole farms as the components of simple or complex matrices based on edaphic divisions, commodity production systems, and specific disciplinary concerns. Research on component technology, cropping systems, and specific issues as subsystems of farming systems research will continue when warranted. In other cases a single commodity or concern may be the primary factor. Secondary data in project planning and ex ante analysis will assume greater importance. One of the important achievements of farming systems research in Indonesia has been to intensify and diversify cropping systems. This has generally consisted of adding additional farm enterprises such as livestock, fruit trees, or estate crops to food crops. This approach can play an important role in promoting agricultural diversification, particularly if it is broadened to include diversification in commodities, technologies, resources, regional agricultural products, institutions, and job opportunities. Conclusion Continued agricultural diversification requires further government policy reforms and actions, external support, participation of farmers, and improved technology. Some rearrangement of infrastructure and institutional mechanisms within the agricultural sector, as well as in related sectors such as industry, trade, banking, and public works, may be needed. Improvements-and the resources to support them-will be needed in research, extension, regulation, irrigation management, and other supporting systems. -164- The differences between Java and the outer islands should be considered in formulating an agricultural investment policy. For example, land-man and land-capital ratios in the outer islands are higher than in Java. Success in previous approaches that relied heavily on rice production in Java will not be appropriate for the more complex issues of agricultural diversification. Instead, the government will have to focus on supporting increased land and labor productivity across Indonesia by providing technical services and rural infrastructure while allowing farmers greater latitude in choosing appropriate cropping patterns in response to market signals. There must also be technical support to improve the management and flexibility of irrigation systems and water collection mechanisms for dry-season farming in rainfed areas. Investment in farmers' education to improve their skills and knowledge in managing irrigation systems professionally may also be warranted. Farm mechanization is one means of overcoming the labor shortage in the outer islands. Some investment in farm mechanization can be managed and organized by farmers themselves. The government should provide them with technical support services to improve their skills and knowledge so that farm machinery can be operated efficiently. Private sector sales, rental, and service of farm machinery should be promoted, as this can greatly speed mechanization while easing financial and organizational burdens on the government. Technologies generated by research are important inputs to agricultural diversification. Research, however, is a long-term investment and will require continuous government support. In addition, research findings do not benefit farmers unless they are put to use. The transfer of technology to farmers is not automatic regardless of how good the technology is. Further steps need to be taken to transmit findings to farmers. Many researchers used to think that they had completed their responsibility when they published their findings in professional journals and that dissemination of new technology was the responsibility of others. This situation is changing, and researchers are now more concerned about how improved technology can be optimally adopted by farmers. Considering this issue, Indonesia's Ministry of Agriculture has begun to carry out research in technology development on farmers fields in areas of 50 to 100 ha. Researchers, extension agents, farmers, and related agencies work together in this research activity. Several approaches have been taken to transfer technology from the research system to farmers quickly enough to meet their continuing need for new technology. Farmers who intend to improve their farming activities will be able to accomplish their goals and objectives with sufficient knowledge and skill. There are still problems, although feedback from farmers to researchers now flows more easily and researchers are better informed of farmers' needs so that they can make research activities relevant to their problems. But to achieve a more effective and efficient extension program capable of transferring technology available from the research system to farmers and channeling farmer feedback to researchers, there is a need for a stronger linkage between Indonesia's research and extension systems. The linkage between research and extension is already being strengthened by broadening the surface of contact between the two. A ministerial decree has built a foundation for effective research-extension links, and overlapping though coordinated activities are being promoted. Areas of overlap include on-farm trials, pilot tests of technology packages, and diagnosis of farmer needs and circumstances. Additional channels for strengthening the linkage between the two have been intensified and include technical meetings, seminars, workshops, and field days. Commodities with the highest potential for increasing farm incomes should be given priority. Although rice farming will continue to be the main activity in irrigated areas, other commodities should be given more attention because of their potential for increasing farmer incomes. Future rice production is projected to meet ever-increasing demand, but the grain's concentration in the most suitable areas and the small scope for further productivity increases mean that targets may not be easy to accomplish. Some resources may have to be taken out of rice and other food crops and put into commodities with higher value. - 165 - Flexible and efficient use of water is important to enhancing diversification in irrigated areas. Studies on water management show that water use efficiency is still low, and improvements in the management of irrigation water are important. In addition, farming technologies for irrigated areas such as rice-fish culture, rice-fish-duck culture, and rice-fish-palawia crops culture could be expanded to achieve optimum use of resources and increase the feasibility of growing seasonal crops in addition to rice in the wet season. Better control of water supply and drainage is the key element to the improvement of the irrigation and water management system. Comparative advantage analysis must be put to much greater use to determine the most profitable areas for commodities and farming systems. This should be based on a comprehensive understanding of the distribution of existing resources, including a land inventory and classification by crop suitability and an inventory of infrastructure in relation to industry and markets. This will provide the information base needed to determine the potential for the opening of new lands for certain crops, agricultural diversification, and agroindustry and to help develop more realistic production and area targets. Production and trade of agricultural commodities should be monitored and disseminated to producers. In certain cases the information should be given to farmers as is. In other cases it should be interpreted in the form of production and area targets. Finally, most of the Indonesian officials implementing agricultural development programs at provincial, district, and rural levels have extensive experience with BIMAS (the government rice production program). Their knowledge of and skill in more diversified agriculture and market-oriented activities, however, need to be strengthened further. Intensive retraining of staff and farmers is needed to enhance their competence in and professional commitment and responsibility to agricultural diversification. References Affandi, Ahmad. 1986. Agricultural Development in Indonesia. Jakarta: Indonesian Ministry of Agriculture. Altemeier, K., and G. Gijsbers. 1988. Soybean in Indonesia: A Case Study on Comparative Advantage. Bogor, Indonesia: CGPRT Centre. Central Bureau of Statistics (CBS), Indonesia. 1970-89. Statistical Yearbook of Indonesia (for 1969-88). Jakarta. Hedley, D.D. 1988. Diversification: Concepts and Direction in Indonesian Agricultural Policy. Jakarta. Manwan, Ibrahim, et al. 1989. Technology for Soybean Production Improvement in Indonesia. Special Report Pus/02/89. Bogor, Indonesia: Central Research Institute for Food Crops. McIntosh, J.L. 1986. "Directions and Strategy for Future Farming Systems Research in Indonesia." In Proceedings of the International Farming Systems Workshop (held December 10-13, 1985). Sukarami, Indonesia: Sukarami Research Institute for Food Crops. RATIONALIZING CROP DIVERSIFICATION IN MALAYSIA'S NONGRANARY IRRIGATED AREAS Jaafar Mat and Ng Chau Chen Irrigation development efforts in Malaysia have long been geared toward the provision of facilities for the cultivation of paddy, primarily to meet government self-sufficiency objectives and to raise the income of paddy farmers, who account for a large proportion of the country's poor. Recently, however, the general trend in rice production has been a decrease in paddy planting, coupled with an increasing incidence of idle land. The government is taking steps to evaluate the performance of irrigated schemes, particularly inefficient small-scale schemes, as a prelude to examining and preparing crop diversification programs as a means of achieving sustainable farm incomes and reducing vegetable food imports. Concurrently, a national agricultural rice policy holds that rice production should come mainly from existing large-scale irrigation schemes (so-called granary areas), whereas nongranary irrigated areas will be phased out and the land converted to the cultivation of diversified crops. To avoid fragmented development programs for crop diversification, there is an urgent need to study the prevailing conditions of irrigated schemes to formulate strategies and plan and implement projects effectively. The Malaysian government has therefore embarked on a crop diversification study via a technical cooperation program with the government of Japan. This paper describes the study, outlines its approaches, and presents some interim findings for three representative schemes.' Background Of Malaysia's 650,000 families with annual incomes below the M$4,800 poverty line, one-quarter are paddy farmers, with holdings on average of less than 2.4 ha. This size makes irrigated rice cultivation financially inviable and has led to a significant idling and underutilization of paddy lands outside the country's eight major granary areas (figures 1 and 2). By 1987, cropping intensity had fallen to 49 percent in these nongranary areas. Young people are migrating to the cities, and the average age of farmers has risen to 46 years. The Fifth Malaysia Plan (1986-90) set a major goal of revitalizing the paddy subsector to reduce the number of paddy farm households with incomes below the national poverty line and to increase the productivity of idle and underutilized paddy fields. Malaysia's strategic objective in the paddy subsector is diversification through value added production alternatives and vertical movement into processing and other forms of agroindustry. To reach this target, strengthening the farm management capabilities of full-time paddy farmers, particularly poor paddy smallholders, by enhancing farm income sources must be stressed. The fifth plan encourages crop diversification through private sector involvement and promotion of profit-oriented farm management. It limits support for large-scale paddy cultivation to eight granary areas in the Malaysian Peninsula to meet a rice self-sufficiency target of 60 to 65 percent of domestic demand. Paddy production in the country's 924 other irrigated areas will gradually be phased out and more remunerative crops pursued. Jaafar Mat is with the Malaysian Ministry of Agriculture, and Ng Chau Chen is senior engineer, Irrigation Branch, Department of Irrigation and Drainage, Malaysia. - 168 - Figure 1. Land Use in Nongranary Irrigated Areas 100 80 60 ~40 20 0-- 1983 1984 1985 1986 1987 Year Paddy Upland crops Idle land Source: Adapted from JICA 1990 (interim report). Figure 2. Proportion of Upland Crops and Paddy in Nongranary Irrigated Areas, by State, 1987 F ~ ~ ~ ~ ~ ~ ~ ~ ~~I IIII Peris Kedah P. Pinang Perak Selangor N. Sembilan Melaka Johor Pahang Terengpanu Kelantan Sabah Sarawak I iI I I I 0 20 40 60 80 100 Proportion (%) - Annual crops/fruit crops [1 Tree crops Paddy Source: Adapted from JICA 1990 (interim report). - 169 - Malaysia's agricultural diversification strategy has four major objectives: 1. widening the demand and production bases of agriculture and the economy to minimize fluctuations in export earnings and income at national and farm levels; 2. generating new sources of growth and employment through horizontal diversification; 3. ensuring adequate food security through domestic production; 4. attaining wider intersectoral linkages for balanced growth and greater employment elasticity between agriculture and industry through vertical diversification. The Japan International Cooperation Agency's Inventory/Survey In the late 1980s, the Malaysian government recognized that its strategy of promoting crop diversification required a diagnostic base that could only be established by identifying the agricultural and economic constraints and potentials of its 924 nongranary irrigated areas. The government sought assistance from the Japanese government in creating a series of studies to define the paddy subsector's problems, as well as in selecting countermeasures and formulating a crop diversification program. A 20- month program of farm-level studies, funded by the Japan International Cooperation Agency (JICA), was undertaken in February 1989. Its objectives were to: * carry out a nationwide inventory/survey of land use, constraints, and potentials in the 924 nongranary irrigation schemes managed by the Ministry of Agriculture's Department of Irrigation and Drainage (DID); * develop a data base founded on the inventory/survey's results; * undertake a feasibility study on improving agricultural production infrastructure and farming practices to promote crop diversification in 12 areas selected through the inventory/survey. The inventory/survey suggested classifying the 924 nongranary irrigation schemes into four broad categories: paddy cultivation and crop diversification areas, idle land, and nonagricultural land. Several areas were designated for maintenance as paddy lands, some because of their high productivity, others because the aged farmers working them are set in their ways and unlikely to participate in a crop diversification program. Deciding whether lands should be continued in paddy was based on criteria such as the availability of water for irrigated double-cropping; suitability for less labor-intensive mechanisms, such as direct seeding, mechanized plowing, and mechanized harvesting; soil suitability for nonpaddy crops; and the profitability and marketability of such crops. There are several diversification alternatives for these paddy lands, including (1) a double- cropping system, planting paddy in the main season and short-term annual crops in the off-season; (2) irrigated annual cash cropping; (3) perennial crop cultivation; (4) freshwater fish pond culture; and (5) livestock production. Eight potential categories of diversification schemes were established: 1. conversion to high-value crop cultivation under irrigated conditions; 2. conversion to tree crop cultivation; - 170 - 3. introduction of a double-cropping system: planting paddy during the main season and short-term annual crops during the off-season; 4. conversion to livestock crop cultivation or livestock fields; 5. conversion to freshwater fish culture ponds; 6. maintenance as minigranary areas; 7. maintenance as paddy cultivation areas for a defined period for social welfare objectives and thereafter to be recategorized; 8. conversion to housing/industrial and other uses. State Profiles The survey produced the following profiles and recommendations for nongranary paddy land use in Malaysia's 13 states (tables 1 and 2): Perlis-22 nongranary irrigation schemes, covering 4,215 ha. All schemes are planted to paddy during the main cropping season, but are left uncultivated during the off-season because of climatic limitations. Most farmers rely heavily on paddy for farm income and have a strong desire to continue growing rice. Sixteen schemes were categorized as minigranary areas, and diversification into irrigated high-value crops, tree crops, and alternating paddy and short-term annual crops also has good potential. Off-season irrigation should be developed and water-saving irrigation practices promoted to increase farm incomes by encouraging the off-season cropping of either paddy or high-value upland crops. Kedah-75 nongranary irrigation schemes, covering 17,133 ha. Forty-four of these areas have year-round irrigation water. Twenty-eight schemes were designated as best suited for minigranary development, and 44 for paddy-annual crop rotations and fruit tree cultivation, with some converting to irrigated high-value crops as demand increases. A rapid increase in nonirrigation water use in Kedah's lower Muda River basin has led to increasing conflicts, requiring urgent examination of future irrigation water demand and formulation of a water distribution and utilization plan for the basin. Pulau Pinang-14 nongranary irrigated areas, totaling 3,541 ha. The state shares the Muda River watershed with upstream Kedah and faces the same problems of conflicting water demands. Perak-63 nongranary irrigation schemes, totaling 12,722 ha. Irrigated double-cropping of paddy is possible in 43 schemes and is being undertaken in 5,343 ha, or 42 percent of the total irrigable area. Fifteen schemes are completely idle, and main-season paddy is grown in less than half of the irrigable areas in another 15, so that 39 percent of the total irrigable area is fallow. Fifteen schemes should be converted to high-value irrigated crops and 33 schemes to tree crops. Eight schemes should be put into paddy-annual rotation. Utilization plans for nongranary irrigation schemes should stress integrated approaches to revitalizing idle schemes. Selangor- 17 nongranary irrigation schemes, covering 939 ha. Six of these schemes are fully idle, and for all but one, which is adjacent to the Barat Laut Selangor granary area, it is difficult to justify continued support for paddy cultivation. Tree crop miniestates should be examined as the best alternative for the remaining areas. - 171 - Negeri Sembilan-156 nongranary irrigation schemes, TABLE 1. AREA DISTRIBUTION OF covering 10,934 ha. Forty-three of PADDY FIELDS BY STATE, 1990 (HA) these schemes are completely idle, NON- NO. OF with 83 schemes less than half- GRA- GRA- NONGRANARY utilized for paddy cultivation. STATE NARY NARY TOTAL SCHEMES Diversification should focus on tree Perlis 19,500 4,215 23,715 22 crops in 136 schemes, and a study Kedah 75,500 17,133 92,633 75 should be undertaken on the human Pulau Pinang 13,000 3,541 16,541 14 and financial resources currently Perak 39,568 12,722 52,290 63 Selangor 19,022 939 19,961 17 available to irrigation authorities in Negeri the area to facilitate the shift of Sembilan 0 10,934 10,934 156 operation and management systems Melaka 0 7,149 7,149 54 from irrigation supply to drainage Johor 0 4,010 4,010 23 Pahang 0 17,430 17,430 290 control. Terengganu 5,100 9,083 14,183 39 Kelantan 38,807 10,667 49,474 77 Melaka-54 nongranary Sabah 0 17,163 17,163 56 Sarawak 0 15,136 15,136 38 irrigation schemes, totaling 7,149 ha, Total 210,497 130,122 340,619 924 all of which should be devoted to cultivation of high-value irrigated Sourc: JICA 1990 crops. Crop diversification in Melaka is progressing well, but technical studies should be conducted on overcoming poor drainage conditions. Johor-23 nongranary irrigation schemes, covering 4,010 ha. Eight of these are idle, and another eight are less than half-utilized. High-value crop cultivation is the preferred use for 16 of these areas. Constant water shortages in the Muar River basin, where 18 of the schemes are located, have led to the promotion of conversion to crops other than paddy, but progress has been slow because of insufficient agricultural support services. Finding ways to improve project implementation is indispensable to accelerating the crop diversification program. Pahang-290 nongranary irrigation schemes, covering 17,430 ha. For the past three years, there has been no cropping in 195 of these areas. Despite this high incidence of idle schemes and the low cropping intensity of 12 percent during the main season, supervisory staff and annual recurrent budgets for operation and management are continually allocated to the entire area. High-value irrigated crops are the preferred category for 32 schemes. Fifty-eight schemes should be put into tree crops, and 16 into a paddy-annual crops rotation. Another 135 schemes in areas that have not been farmed in the past three years should be converted to nonagricultural uses. A basic strategic plan needs to be formulated for the entire area, emphasizing effective reallocation of human and budgetary resources. Under the framework of this strategic plan, the present operation and management system should be rationalized and agricultural support services in specific areas should be strengthened to promote crop diversification. Terengganu-39 nongranary irrigation schemes, covering 9,083 ha. Continued and intensified paddy cultivation is attractive to farmers in the northern part of the state, where minigranary areas should be maintained. In the southern and central parts, however, land utilization is low, and these areas should be converted to tree crops, although a detailed physical and socioeconomic study will be needed to implement this scheme. - 172 - Kelantan-77 nongranary irrigation TABLE 2. DISTRIBUTION OF NONGRANARY schemes, covering 10,667 IRRIGATED SCHEMES BY STATE, 1990 ha, of which 23 should be HIGH- MAIN- converted to tree crops and VALUE PEREN- DUAL MINI- TAINED 29 to a paddy-annual crops UPLAND NIAL CROP- GRA- AS NON- rotation. The remaining STATE CROPS CROPS PING NARY PADDY AGRIC. TOTAL areas on the east coast Perlis 3 2 5 11 1 0 22 require further analysis to Kedah 4 21 9 11 30 0 75 identify a proper balance Pulau Pinang 6 0 5 0 3 0 14 between improving paddy Perak 6 37 6 10 4 0 63 production and investing in Selangor 0 10 0 0 7 0 17 other crops. Negeni Senibilan 14 140 0 1 1 0 156 Melaka 9 39 1 0 5 0 54 S a b a h-5 6 Johor 9 5 3 4 2 0 23 nongranary irrigation areas, Pahang 77 45 0 2 18 148 290 covering 17,163 ha. Terengganu 0 13 5 8 12 1 39 17,163 ~~~Kelantan 16 18 12 11 20 0 77 Although off-season Sabah 0 3 0 16 32 5 56 irrigation water is available Sarawak 0 1 0 0 37 0 38 in 45 schemes, only 15 Total 144 334 46 74 172 154 924 percent of the area of off- season paddy is cultivated. Source: JICA 1990. Thirty-six schemes should be maintained in paddy for social welfare objectives while further studies are undertaken. To improve rice self-sufficiency within the state, however, it will be necessary to raise cropping intensity and increase yields in these schemes. Work should focus on improving farmers' potential to double-crop irrigated paddy through improved production fields and the upgrading of agricultural support services. Sarawak-38 nongranary irrigation schemes, covering 15,136 ha. Twenty-six schemes, however, are in coastal swamp areas and would be difficult to convert to crops other than paddy. Thirty-five of the schemes should be maintained in paddy. A comprehensive study of the state's agricultural sector, along with natural resource and management evaluations, will be conducted shortly, forming the basis for future paddy land utilization plans. Crop Diversification Potential in Representative Schemes More comprehensive follow-up studies were undertaken in Pulau Pinang, Negeri Sembilan, and Kelantan states. The Pulau Pinang study-in the Sungai Kulim I and II irrigation schemes-examined the potential for converting paddy lands to irrigated high-value crops and fruit trees, using a paddy-annual crops rotation as a transitional step. The Negeri Sembilan study focused on crop diversification using perennial trees in seven schemes along the Mampong River. The Kelantan study examined the potential for a crop diversification program through the introduction of a double-cropping system, planting paddy during the main season and short-term annual crops during the off-season. - 173 - Sungal Kulim, Pulau Pinang State Pulau Pinang State's Kulim area is in the Perai River basin and has an irrigable area of 3,223 ha. Recent water shortages have resulted in the cancellation of a planned water diversion to the Kulim area and the formulation of a revised development plan for the Pinang Integrated Area Development Project (IADP). Although the revised plan provides water from the Kulim and Jarak tributaries of the Perai River, it is not sufficient to meet irrigation water demand for growing paddy biannually. There is, therefore, a pressing need to rationalize the use of limited irrigation water and formulate an agricultural development plan that includes crop diversification. The Kulim area is divided into six irrigation blocks, Kl through K6. Based upon physical characteristics such as topography, direction of slope, and block shape, the K2, K3, and K4 blocks-with the highest paddy yields-are best suited to on-farm irrigation and drainage service facilities. Under the Pinang IADP, therefore, these three blocks are included in the Balik Pulau/Seberang Perai granary area. The Kl block is maintained as a nongranary irrigation area of Sungai Kulim I, and the K5 and K6 blocks, in Sungai Kulim II, are target areas for promoting crop diversification. Over one-third of Kulim's irrigated area is no longer actively used for paddy: 861 ha are idle, and 410 ha have been converted to other crops. Only about 2,000 ha remain in double-cropped paddy production. Examination of the water resources in the Sungai Kulim area shows insufficient off-season supplies to extend paddy double-cropping much beyond its present scope. A detailed survey was made of 209 of the 800 farm households in the Kulim schemes to determine their attitudes toward crop diversification. A profile of respondent farmers found 61 percent to be over age 45, and for 88 percent of farm households, paddy cultivation was the most important income-generating activity. Fifty-six percent of responding farmers rated the TABLE 3. FARMER RATING OF RIGATION A" schemes' irrigation and drainage DRAINAGE FACILITIES IN SUNGAI KULIM I AND 11 SCHEMES, 1990 facilities as less than good (table 3). s. KULIM I S. KULIM II TOTAL Most respondents viewed RATING NO. (%) NO. (%) NO. (%) rehabilitation and upgrading of irrigation (63 percent) or drainage (74 Good 325 518 48 33 809 349 percent) systems as the basic Fairly good 17 27 53 36 70 33 requirements for implementing crop Not good 8 13 40 27 48 23 diversification programs, along with No response 1 1 1 1 2 1 the improvement of farm road Total 63 100 146 100 209 100 networks (table 4). Almost all farmers responded that it would be Source: JICA 1990. difficult to participate in a crop diversification program unless certain problems were overcome, such as frequent pest and disease outbreaks, lack of machinery, insufficient irrigation water supplies, lack of postharvest facilities, labor shortages, and poor drainage (table 5). Fifty-five percent of the farm households favored a crop diversification program. Forty-one percent, however, viewed promotion of crop diversification as "difficult" or "impossible," mainly because they thought that the land was suitable only for paddy. In terms of higher-value crops, respondents showed interest in sweet corn, okra, cucumber, and chili, with 17 percent considering their marketing prospects "very promising." In contrast, 47 percent of respondents felt "a little bit worried" about the marketing prospects of these crops, and 27 percent felt "uneasy" because of anticipated marketing difficulties. Regarding possible markets, 53 percent of respondents would target a local market, and 30 percent considered neighboring markets promising. Given the existing drainage facilities in the Kulim area, which are designed both to control soil moisture and the groundwater table and reduce the time required for draining surface water from paddy - 174 - fields, the prevailing soil types in the Kulim schemes are suitable TABLE 4. FARMER PREREQUISITES FOR CROP for cropping vegetables and DIVERSIFICATION IN SUNGAI KULIM I AND II SCHEMES, 1990 fruits. Possible vegetable crops include bittergourd, cucumber, S. KULIM I S. KULIM II TOTAL cauliflower, cabbage, okra, NO. (%) NO. (%) NO. (%) longbean, and chili; fruits that Irrigadonfacilides would grow well include banana, Rehabilitation/upgrading starfruit, pineapple, guava, and of existing irrigation systems 36 57 95 65 131 63 mango. Asparagus is also an Introduction of sprinkler alternate crop if drainage is irrigation facilities 35 56 60 41 95 45 improved. Crop profitability, as Construction of irrigation shown in table 6, is estimated facilities according to Department of Dri,nagefacifities Agriculture data, taking into Rehabilitation/upgrading account farmers' familiarity with of existing drainage facilities 47 75 107 73 154 74 growing these new crops. Construction of new Expected crop yields are 75 drainage facilities 10 16 32 22 42 20 percent of the Department of Construction or improvement Agriculture standard. Under these of flood protection structures 5 8 12 8 17 8 conditions, 13 of the alternate crops should bring higher profits Fann roads Rehabilitation/upgmading than single-crop cultivation of of roads 29 46 59 40 88 42 paddy. Table 6 also estimates market demand for these crops Construction of new farm roads 12 19 24 16 36 17 and the cropping areas required. From a marketing perspective, to Note: Some multiple answers were given. reduce the effects of fluctuating Source: JICA 1990. market prices on selling prices, diversified cropping, rather than monoculture of specific crops, should be introduced. Cropping systems combining two or more crops should also be introduced to avoid pests and diseases associated with continuous monocropping and to reduce the peak labor requirements at planting and harvesting by harmonizing farm operation periods for each crop. Given these parameters, five cropping systems have been designed. The systems combine eight vegetables-asparagus, bittergourd, cabbage, cauliflower, chili, cucumber, longbean, and okra-with short- and long-term (vis-a-vis growing period) fruits. Irrigation structures will need upgrading for the vegetable-based cropping schemes; a furrow irrigation system is recommended to keep initial investments low and to facilitate maintenance. Total required expenditures for irrigation works, including on-farm development in the Kulim areas, are estimated between M$11.5 million and M$17.9 million, depending on cropping choices. Benefits are estimated between M$14.8 million and M$20.5 million, given internal rates of return ranging from 28.6 to 32.9 percent. Three field-level farm management systems are proposed for the Kulim schemes: * management by individualfarmers, under which all decisionmaking is undertaken by owner- operators, is appropriate when farms are isolated and support services have difficulty reaching them; - 175 - * management by group fanning, with a group of farmers responsible for all aspects of planning and decisionmaking. Small lots are grouped together to form a contiguous or noncontiguous block of land of economic size, managed by a selected body. This allows agricultural agencies to provide services more effectively to the group and to solve technical and logistical problems at the farm level in an integrated manner; * management by miniestates, formed by grouping land lots of farmers in an area to be developed using plantation management practices. The resulting miniestates would then be run by a central management body, with landowners holding shares in the miniestate. Contiguous lots of idle land are expected to be converted into miniestates. Existing management arrangements, especially the Pinang IADP, should be used to ensure smooth implementation of crop diversification programs while establishing a special task force to demonstrate crop diversification to farmers. Marpong, Negeri Sembilan State Forty-three of the 156 nongranary irrigation schemes in Negeri Sembilan State are idle, and in 83 schemes paddy is grown in less than half of each command area. An idle paddy land revitalization project covering rainfed paddy fields and six nongranary irrigation schemes is being implemented. There are 11 nongranary irrigation schemes in the small Mampong River basin. Crop diversification centering on conversion to oil palm has been promoted to overcome labor shortages stemming from rural-to-urban migration. Seven schemes in the upstream section of the basin have been selected for possible accelerated diversification into tree crops. The study area contains 517 ha of irrigable land, of which 280 ha are now idle. Paddy accounts for 173 ha, tree crops for 58 ha, and upland crops for 6 ha. The seven schemes have gravity irrigation systems, with most main canals concrete-lined. No secondary or tertiary irrigation canals are lined. Only one 64-ha area has irrigation water sufficient for paddy double-cropping throughout the year. Women in the Mampong area have usually TABLE 5. FACTORS INFLUENCING FARMER CROP assumed a greater role in DIVERSIFICATION IN SUNGAI KULIM I AND 11 SCHEMES, 1990 paddy cultivation, from s. KULIM I S. KULIM 11 TOTAL plowing to harvesting, than NO. (%) NO. (%) NO. (%) men. Because the uneconomic size of paddy farms provides Finance 51 81 81 55 132 63 Technique 28 44 57 39 85 41 insufficient returns from Marketing 34 54 61 42 95 45 farming alone, Mampong Labor 21 33 40 27 61 29 farm families earn Other 11 17 19 13 30 14 None 2 3 30 21 32 15 supplemental income from No. of respondents 63 146 209 temporary rubber-tapping Total no. of responses 147 288 435 work and receive cash remittances from their a. Some multiple answers were given. children working in urban Source: JICA 1990. areas. In this situation most paddy farmers intend to convert their farming pattern from labor-intensive paddy cultivation to labor-saving crops such as oil palm and cocoa, for which the coarse local soils are well suited. Both crops are principally grown for export and have well-established marketing channels. - 176 - TABLE 6. PROFITABILITY OF ALTERNATE CROPS IN KULIM I AND II SCHEMES, 1990 REQUIRED GROSS PROD. NEr TOTAL CROPPING YIELD INCOME COST INCOME DEMAND AREA CROP (NfT/HA) (MS/HA) (MS/HA) (MS/HA) (MT) (HA) Bittergourd 10.0 7,400 4,600 2,800 1,900 230 Cucumber 19.0 7,600 6,400 1,200 4,000 250 Cauliflower 5.5 15,000 5,300 9,700 4,800 1,040 Cabbage 14.0 14,200 5,200 9,000 11,100 960 Okra 20.0 15,300 5,400 9,900 2,000 120 Longbean 11.0 7,200 5,700 1,500 6,000 650 Chili 11.0 16,900 6,800 10,100 7,300 800 Asparagus 3.0 16,500 9,000 7,500 1,100 440 Banana 5.0 3,800 2,500 1,300 10,800 1,350 Starfruit 28.0 28,600 3,300 25,300 700 30 Pineapple 24.5 2,800 1,500 1,300 7,700 370 Guava 20.0 5,000 1,800 3,200 2,000 120 Mango 5.0 10,100 1,600 8,000 1,100 260 Paddy 3.5 2,300 1,200 1,100 - - M$ = Malaysian ringgit. Note: Required cropping area is estimated assuming postharvest losses of 20 percent of yield. Source: JICA 1990. A relatively long period is required for oil palm and cocoa to bear their first harvests, and farmers growing these crops need to follow instructions on farming practices given by responsible agencies. Crop diversification plans, especially those for converting to oil palm, should be incorporated into the Federal Land Consolidation and Rehabilitation Authority schemes. For cocoa, individual farm operation is possible with extension services from the Department of Agriculture. In the initial stage of conversion, mixed planting of cocoa with other tree crops, such as banana, is preferable to protect young cocoa trees from the sun. No irrigation water is required for planting oil palm and cocoa. In the Mampong scheme, year-round water supply can be expected, and there is some scope for converting paddy fields to freshwater fish ponds. In Negeri Sembilan, fish ponds cover more than 400 ha, and there is also a tin- mining pool used for fish culture. A water balance study has shown that freshwater fish pond culture technically can be introduced to the Mampong area. Meeting the drainage requirements for growing oil palm and maintaining freshwater fish ponds would require construction costs estimated at M$1,000/ha for conversion to oil palm and cocoa and M$21,000/ha for conversion to freshwater fish ponds. Total investment requirements for oil palm and cocoa (each area = 517 ha) are M$1.04 million, and for freshwater fish pond culture (area = 64 ha) in the Mampong scheme, total investments are M$1.34 million. Annual incremental benefits are estimated at M$1.74 million for oil palm, M$0.98 million for cocoa, and M$0.22 million for freshwater fish ponds, providing a financial internal rate of return of 26.3, 19.6, and 3. 1, respectively. The following steps need to be taken to consolidate Mampong-area farming groups and organizations and improve their productivity: * introduce new technologies for double-cropping and planting of high-yield varieties; • establish regular crop maintenance operations and systems; - 177 - * introduce nonpaddy crops during the off-season; * strengthen extension services through training in demonstration plots; * consolidate farmer organizations by increasing membership to ensure more efficient and rational management; * promote group farming to overcome labor shortages; * improve transportation for marketing of products and for finding new markets. Nongranary irrigation schemes should be incorporated into Federal Land Development Authority (FELCRA) schemes when oil palm is the main vehicle for crop diversification, and the Ministry of Agriculture's idle paddy land revitalization program should be used for cocoa planting on a group farming basis. Freshwater fish pond culture needs to be operated on a commercial, private basis. Kelantan In Kelantan, 77 nongranary irrigation schemes are scattered throughout the middle-to-upstream sections of the Kelantan River basin, and paddy farmers have adjusted their planting schedules to the annual flooding patterns of the river and its tributaries. Consequently, single-cropping of paddy is common even in irrigated areas, and there is scope for introducing a two-season rotation of paddy and short-term annual crops to increase off-season cropping intensity and farm incomes. A study of water balances in the three study areas, Hilir Sat I, Repek, and Rawa Bechah Laut, showed that they have sufficient supplies to serve as paddy-annual or minigranary areas. A detailed survey was conducted in the three representative scheme areas to determine paddy farmer attitudes toward crop diversification. Of the approximately 1,300 farm households in these areas, 105 were interviewed. Of respondent farmers, 28 percent were over age 55. Although all farm households cultivate paddy, 45 percent also had "part-time jobs," and 27 percent cultivated tree crops. Two-thirds of those interviewed rated irrigation and drainage facilities "fairly good" or "less than good" (table 7). Farmers pointed out seven major limitations to diversifying crops: lack of postharvest facilities, insufficient irrigation water supply, lack of funds and machinery, labor shortages, frequent pest and disease outbreaks, the occurrence of natural disasters, and poor drainage. Moreover, 66 percent of respondents stated that they did not intend to convert present paddy area to other crops. The respondents' basic requirements for implementation TABLE 7. FARMER RATING OF IRRIGATION of crop diversification programs AND DRAINAGE FACILITIES IN KELANTAN, 1990 were rehabilitation and upgrading RAWA of the existing irrigation and HILIR BECHAH drainage systems, introduction of SAT I REPEK LAUT TOTAL sprinkler irrigation systems, and NO. (%) NO. (%) NO. (%) NO. (%) improvement of the farm road Ver good 0 0 0 0 0 0 0 0 network. Regarding possible Good 28 43 3 11 3 27 34 32 markets, 77 percent of Fairlygood 20 30 6 21 5 46 31 30 Notgood 18 27 19 68 3 27 40 38 respondents would target a local Total no. of market, and 11 percent respondents 66 100 28 100 11 100 105 100 considered neighboring markets promising. Source: JICA 1990. - 178 - In the three representative schemes, 93 percent of the respondent farmers replied that they intended to continue paddy cultivation. Their preferred cropping schemes were a double-cropping system, planting paddy during the main season and short-term annual crops during the off-season (28 percent), and conversion to livestock crop cultivation or livestock fields (19 percent). With some irrigation and drainage improvements, the prevailing soil structure in these areas is suitable for upland crops, TABLES. PROnTABILITY OF including tobacco, maize (as either a cash or ALTERNATE CROPS IN KELANTAN, 190 fodder crop), chili, longbean, and groundnut. GROSS PROD. NET The profitability of these crops is estimated in INCOME COST INCOME table 8, with cropping levels set at 75 percent CROP (MS/HA) (MS/HA) (MS/HA) of standard Department of Agriculture levels Tobacco 6,300 1,800 4,500 to compensate for farmers' initial lack of Maize familiarity with the new crops. Even with this Feed 12,300 12,250 50 limitation, all alternate crops except maize Consumption 1,300 1,250 50 Chili 16,900 6,800 10,100 produce higher profits than a single paddy Longbean 7,200 5,700 1,500 crop. Federal Agricultural Marketing Groundnut 2,900 1,400 1,500 Authority (FAMA) projections show a market demand for the alternate crops that would MS = Maaysin riiggit. require (assuming 20 percent postharvest Source: JICA 1990. losses) the cultivation of 420 ha in chili, 440 ha in longbean, and 870 ha in groundnut. (Tobacco was omitted from this computation because the annual allocation of planting areas is controlled by a quota system.) In introducing a double-cropping system, the cultivation period of off-season upland crops is primarily fixed, considering the optimum timing of planting and harvesting. Main-season paddy cultivation, therefore, is restricted to a particular duration, and the planting of main-season paddy consequently needs to begin as early as possible, even in areas affected by flooding. Furthermore, completing the harvest of off-season upland crops before the main monsoon begins should be stressed, and it is necessary to finish harvesting main-season paddy prior to planting off-season upland crops. In and around the three representative schemes in Kelantan, the Department of Agriculture has encouraged farmers to establish farming groups, and 299 groups have been set up, representing 17 percent of the state's farmers. Total construction costs for all three areas, representing 930 ha, are estimated at M$10.4 million. The annual incremental benefits from adopting dual-crop rotations in all three areas are estimated at M$4.4 million, providing a financial internal rate of return ranging from 20.1 to 24.3 percent in the three areas. Crop diversification should be implemented by introducing a double-cropping system through existing committees, such as the Ministry of Agriculture's Planning and Implementation Committee, state agricultural action and planning coordination committees, and district agricultural planning coordination committees. Also, the linkages among related departments and agencies should be consolidated, particularly the Irrigation and Drainage and Agriculture departments, the Farmers' Organization Authority, the Federal Agricultural Marketing Authority, and the Bank Pertanian Malaysia. Conclusion Agricultural diversification, with its aim of growing additional export crops and avoiding overdependence on a single export such as rubber, has long been the policy of the government of Malaysia. For crop diversification within irrigated areas, the study presented in this paper represents an - 179 - important step for the governnent before formulating strategies and programs for implementation. The interim findings of the study indicate an encouraging trend in the diversification process, especially for those areas with idle land. The result of the categorization of 924 nongranary irrigation schemes reveals that 144 schemes have high potential for promoting crop diversification through high-value upland crop cultivation under irrigated conditions. If markets are available, 410 schemes have high potential through fruit-tree planting and short-term upland-crop growing, and 67 schemes are most suitable for growing paddy in the main season and high-value nonpaddy crops in the off-season. The 410 schemes also have potential for conversion to industrial crops such as oil palm and cocoa rather than fruit-tree planting, though these industrial crops are likely to be less profitable than fruit crops. The well-operated nongranary irrigation schemes where paddy productivity and profitability are higher than those of alternative crops should be maintained as minigranary areas to increase rice production. The schemes on which aged farmers depend for their main income sources need to be maintained as temporary paddy schemes if successors within the farm household or the capability of learning new farming practices to grow diversified crops is lacking and there is no scope for encouraging group farming in these communities. Seventy-four schemes have been identified as minigranary areas and 172 as temporary paddy cultivation areas. The study also provides a comprehensive data base necessary for the formulation of crop diversification programs within nongranary irrigated areas. The area studies indicate that with provision of integrated support services, good economic performance can be expected from implementing crop diversification programs in Pulau Pinang's Kulim schemes and the study areas in Negeri Sembilan and Kelantan. Note 1. For a discussion of the macroeconomic aspects of diversification in Malaysia, see S. Barghouti, C. Timmer, and P. Siegel, Rural Diversification: Lessons from East Asia, World Bank Technical Paper 117 (Washington, D.C., 1990). Reference Japan International Cooperation Agency. 1990. Feasibility Study on the Rationalization and Crop Diversification in Nongranary Irrigated Areas in Malaysia. Interim and final reports. AGRICULTURAL DIVERSIFICATION IN THE PHILIPPINES Marietta S. Adriano and Jocelyn L. Cedillo For countries with a dominant agricultural sector, such as the Philippines, agricultural diversification is an important strategy in the reallocation of resources to stabilize and raise rural incomes and should be viewed within the entire economic development process. This paper discusses the status of the Philippine agricultural sector, focusing on the policies that shape and affect it. It then reviews government programs and projects supportive of agricultural diversification, with recommendations for optimal implementation of this strategy. Diversification, Resource Movements, and the Philippine Market Diversification involves the movement of resources across activities, industries, and sectors. Farming households in the Philippines have attained a significant degree of diversification. The labor and capital resources of most Philippine farm families are allocated among crop and noncrop production as well as farm and nonfarm and sometimes even off-farm activities. A rice-based farm, for instance, may have a small livestock/poultry production unit. During the dry season, farmers and members of farm households in rainfed farm areas take off-farm jobs in irrigated areas with ensured supplies of irrigation water. Nonfarm activities such as basket weaving can sometimes be observed in these rainfed farm areas, especially during the dry season. Diversification is more pronounced at the district level, as rural households engage in specialized activities. Trading, however, is limited because the market at this level is too small to allow significant gains from exchange. Moreover, production within districts is usually fairly homogeneous and offers fewer opportunities for trade. In Central Luzon, for instance, the crop produced by most farms is rice: rice trading within the region is thus limited. The benefits of diversification are more profound at the national level. Regional specialization engenders more significant resource movements in both labor and capital. Specialization based on comparative advantage induces economic efficiency that results in significant increases in production, thus creating opportunities for mutually advantageous interregional trade. The Philippines can be divided into broad areas of production based on comparative advantage. Central Luzon has a comparative advantage in rice production, as irrigation allows a second rice crop. Negros has a comparative advantage in sugar production, and Bicol (in Southern Luzon) in abaca production. Table 1 shows regional efficiency of production for several food crops. The comparative advantage of Southern and Central Mindanao is in corn production. These regions, however, do not have enough support from forward linkages such as feedmills to allow them to maximize this advantage through their own production of pork and chicken. Most of the corn produced in Mindanao is transported to Cebu and the metropolitan Manila area, where the majority of the feedmillers are located. Although these areas produce little corn, the proximity of feedmillers to urban markets allows them to produce pork and chicken at a comparative advantage. Weak transport links, how- Marietta S. Adriano is director, and Jocelyn L. Cedillo is senior economic development specialist, Agriculture Staff, National Economic and Development Authority (NEDA), the Philippines. The views expressed in this paper are those of the authors and do not necessarily reflect those of NEDA. - 182 - ever, often make it more economical for TABLE 1. ECONOMIC EFFICIENCY OF these producers to import corn rather SELECTED FOOD CROPS BY REGION, 1985 than purchase that grown in Mindanao. This highlights how the presence of REGION' support facilities and other forward CROP A B C D E linkages affects an area's overall Rice 1.1 0.8 0.8 0.8 0.7 production performance. Corn 1.1 0.7 0.9 0.8 0.7 The provision of support Peanut 2.9 5.6 - 3.4 - facilities and small agro-based industries Coaon 0.5 - - - 0.4 within regions should follow primary Sorghum - 0.6 - 1.3 0.6 production activities. Agricultural Wheat 1.4 - - - - Tobacco 0.2 - - - - zonification in accordance with Garlic 0.4 0.3 - - - comparative advantage in production Onion 0.4 0.3 - - - may further improve economic efficiency aish potato 1.0 - - - Cassava 3.1 1.9 2.2 - 0.8 within the country. Government support Sweet potato 2.3 2.1 2.2 - 6.9 to establish postharvest facilities would Abaca - - 0.5 - 1.2 complement the gains from specialization, which may lead to Note: Efficiency coefficients calculated as domestic resource cost/shadow increased rural incomes. exchange rate. Because of the nature of a. For this analysis, the 12 political regions have been regrouped into five agricultural production, continuing regions based on physical proximity to markets, resource endowment, and state of development: A = Ilocos and Cagayan Valley; B = Central Luzon increases m productivity levels cannot and Southern Tagalog (subregion = metro Manila); C = Bicol and Central sustain increases in incomes due to an and Eastern Visayas; D = Western Visayas and Western Mindanao; and "immiserizing growth" process. E = Northern, Southern, and Central Mindanao. For a map of this regrouping, see p. 2.55, Rosegrant et al. 1987. Increases in production beyond self- Source: Adapted from Rosegrant et al. 1987. sufficiency result in falling agricultural prices and, consequently, falling incomes. Balisacan (1989) reports that in the years immediately following World War II, area expansion was the major source of growth in agriculture. Partly as a result of the increasing pressure on land, high- yielding varieties of rice that contributed to significant increases in rice production were developed. In the late 1960s, these high-yielding varieties were introduced and led to substantial increases in production. These increases in productivity resulted in decreases in the real price of rice, which lowered farm incomes, and failed to bring substantial growth rates to Philippine agriculture. Table 2 shows that GDP growth in agriculture has been relatively constant, though consistently lower than in industry. Although the agricultural sector has a significantly large share of the labor force, the growth rate in agricultural employment has also been consistently lower than that in industry. This means that with the expansion of industry, resources generally moved toward industry. However, this movement did not prevent agricultural incomes from falling, and disparity between urban and rural incomes continually increased. To prevent a drop in farm incomes, the process of agricultural diversification must always be linked with activities in other economic sectors. The linkage of agriculture and industry should be highlighted for successful diversification. The industrial sector should be viewed as a market for agricultural products, through both the primary demand for food and the derived demand for the inputs used by agro-based industries. Continuing increases in agricultural productivity have to be matched by an expansion of agro-based industries to sustain agricultural growth. This means that significant labor resources must be reallocated from agriculture to industry and the service sectors. Capital, however, must continually flow to the agricultural sector to maintain increases in labor productivity. The Philippines must continue to strive for greater productivity in rice because of its status as a wage good and because of the country's population growth rate. - 183 - TABLE 2. PERCENTAGE PRODUCTION AND EMPLOYMENT GROWTH IN AGRICULTURE AND INDUSTRY, 1965-85 1965-70 1970-75 1975-80 1980-85 PROD. EMPLOY. PROD. EMPLOY. PROD. EMPLOY. PROD. EMPLOY. Agriculture 3.54 1.3 3.85 4.9 4.97 1.7 2.12 2.8 Industry 4.94 1.2 7.52 7.1 6.19 1.9 2.62 5.2 Source: NEDA Stadistical Yearbook, vatious years; Paderanga 1989. Policies and programs must be formulated to promote agriculture and agro-based industries, ensuring that the production and processing technologies adopted are comparable to those of foreign competitors. This introduces another dimension of agricultural diversification: the international market. The size of the Philippine market may cap the growth of the agricultural sector. It then becomes imperative that a diversification program be anchored in economic efficiency similar to that of other countries. Although agriculture was the major contributor to export earnings throughout the early 1970s, its share declined from almost 90 percent of exports in 1965 to just under 50 percent in 1980, falling to 29 percent by 1988. The Philippines has great potential to benefit from exporting agricultural products, which can be achieved by establishing a strong agro-based industrial sector that would make agricultural products competitive in international markets. The conditions described above are the result of government programs and general economic polices both inside and outside agriculture. The following sections review these programs and policies. Economic Policies Past Policies One of the Philippine government's major goals has been to achieve self-sufficiency in basic food commodities, especially rice and, to some extent, corn. Importation has been perceived as government failure. Thus, the agricultural sector's priority is to ensure, at a minimum, the availability of rice, with diversification as a secondary strategy for increasing farm incomes. The country was a net importer of rice from 1970 to 1976, when new rice technology and various productivity-increasing programs helped achieve rice self-sufficiency. The Philippines became a net exporter of rice beginning in 1977. Since the foremost objective had been achieved, planners began to consider programs to diversify agriculture. Net exports of rice, meanwhile, kept increasing until 1979, after which a slight decrease was experienced in 1980. This continued until 1983; then, the Philippines again became a net importer of rice. Consequently, interest in developing a concrete framework for diversification diminished, so that agricultural diversification as a visible national program never materialized. Although the government did not abandon the idea of diversifying agriculture, none of the several small projects involving nonrice crops and crop-livestock-fisheries farming schemes graduated to a national scale. Current Policies Recognizing the urgent need to develop agriculture, the government has refocused its programs and polices in rural development to support the pro-poor, Countryside Agro-Industrial Development - 184 - Strategy (CAIDS). This strategy recognizes the complementarity and interdependence of agriculture and industry. Some of the measures identified to support the promotion of agroindustrial development are agricultural modernization, improved productivity, and aggressive diversification. These measures are needed if the objective is to raise and stabilize farm incomes. One of the key features of CAIDS is the promotion of site-specific small- and medium-scale agro-based industries that will allow farms to produce commodities with higher value added. Support via the required physical infrastructure, as well as macroeconomic policy environment, will be extended by the government to ensure full implementation. Price. The level of agricultural prices directly affects farm incomes. High real prices for agricultural commodities increase profitability and generally induce a movement of resources toward agriculture, thereby increasing production. A rice-based agriculture, however, has slightly different characteristics. Table 3 shows that support and farmgate prices for rice have been increasing. Real prices, however, fell in the late 1980s, an indication of the government's failure to maintain stable real prices. This may be a crude indication of decreased real incomes of rice farmers. Lower real income should push the average farmer to concentrate on rice production, the family's staple food. This is perhaps why rice production in the Philippines is, at minimum, near self-sufficiency. Any attempt at crop diversification in rice-based farms would therefore depend on the economics of rice production. If rice production becomes profitable, the risk attached to diversification decreases, and farmers may be induced to include nontraditional crops in their cropping patterns and, perhaps, to include noncrop production in their farming activities. This is confirmed by the findings of several studies on diversification. The government also maintains a support price for corn, although since early 1989 this has TABLE 3. YEARLY AVERAGE PRICES been lower than what traders offer. Corn OF PADDY RICE, 1981-88 (PESOSNMT production is relatively stable, but poor transport NOMINAL PRICES REAL PRICES facilities have led to supply shortages in areas YEAR FARM SUPPORT FARM SUPPORT where feedmills are located and surpluses in production areas. This has increased the price 1982 1,370 1,703 791 983 differences between production and distribution 1983 1,530 1,800 803 944 areas, with the former having lower prices and 1984 2,470 2,508 862 876 thereby contributing to a reduction in corn farmer 1985 3,240 3,438 918 975 1986 2,820 3,500 794 985 incomes. This has also affected the profitability of 1987 3,070 3,500 832 949 the livestock and poultry subsectors, which use 1988 3,440 3,500 858 873 corn-based feeds. The increase in real prices for corn has increased the cost of producing pork and Source: Bureau of Agricultural Statistics, National Food poultry, thus constraining growth in these Authority, and NEDA Siisrical Yearbook, various years. subsectors. Sugar is a major export commodity, and the government's sugar price policy is affected by world market conditions. Domestic prices are kept stable and higher than world prices. The domestic market acts as a cushion for the industry. When the world market for sugar slumps, domestic consumers pay a premium over world prices. When it booms, most sugar is exported (some smuggling has been reported as well). Importation is sometimes resorted to during boom periods. This may be considered a paradox in an export-oriented industry. In the mid- 1980s most sugar planters in Negros shifted to prawns when prices for sugar were low and those for prawns high. This situation has been reversed, with the farmers having gone back to sugar production. Unlike sugar, there is no distinct domestic price policy for coconut. Production is affected, however, by international prices because it is also a major export commodity. As a perennial crop, coconut production is not very responsive to changes in world markets, or even domestic market prices in the short run. In the longer run, the price effect on production may be significant as it affects decisions - 185 - to expand or replant coconut areas. The lack of a price policy supporting increases in production may have contributed to coconut's low productivity in the Philippines. Fertilizerprices. In June 1989 the tariff for fertilizer products not locally produced was adjusted downward to 5 percent ad valorem. Quantity restrictions had been lifted earlier. This is an indication of the government's commitment to reducing agricultural production costs to make agriculture and agro- based industries competitive in the world market. Trade. In 1982 the government began a trade liberalization program aimed at increasing economic efficiency by encouraging competitive domestic production. The program has two components: the Tariff Reform Program and the continuing removal of nontariff restrictions on imports. Tariff reform over 1981-82 cut average tariffs for about 730 commodities, mostly producer goods, from 43 to 28 percent. The Philippine Department of Agriculture reports that the average tariff for primary agriculture fell from 6.23 to 4.95 percent, while processed agricultural goods fell from 2.1 to 1.17 percent (table 4). This is a 20 percent reduction in tariffs for primary agriculture and a 44 percent reduction for processed agriculture. Industrial goods tariffs, however, fell 25 percent from 22.61 percent, so that the overall reform was biased against agriculture. Given the low protection rates accorded producers of processed agricultural products, one TABLE 4. AVERAGE TARIFF RATES BY SECTOR would expect a movement of resources toward the more protected industrial sector. Production of REFORM REFORM nonagricultural industrial goods has become SECTOR (%) (%) relatively more profitable. The domestic supply of industrial goods increased at the expense of agro- aricre based firms. This policy works against the agriculture 6.23 4.95 efficient utilization of resources in an agricultural Processed country. agriculture 2.10 1.17 T'his bias arose from a piecemeal Industry 22.61 16.99 approach to the tariff adjustment process because there was no overall tariff policy framework. Reonzn ths th governmen adpeda. Source: Philippine Department of Agriculture 1988. Recognizing this, the government adopted as a strategy, under CAIDS, the formulation of a framework that would result in a uniform effective tariff structure. Concurrently, key industries, including agro-based industries, are to be accorded appropriate protection to encourage their growth in line with the policy of supporting agricultural diversification. The adoption of a competitive exchange rate policy has also been identified as a strategy for promoting agricultural growth. This aims to correct the penalty imposed on export commodities by an overvalued currency. With this new policy, agricultural products are expected to be more competitive in the world market, translating into higher export earnings. Value added tax (VAT). The Philippine government has recently adopted the value added tax system, replacing the gross receipts tax. This new scheme aims to simplify the country's tax system and at the same time allow a uniform rate of taxation across sectors and industries. In addition, it serves as an incentive to industries to increase production because, for several commodities, the new effective tax is actually lower. Exports are also promoted because export taxes are effectively removed. The tax rate adopted is either 0 or 10 percent on the value added. Producers of primary agricultural products benefit through a 0 percent VAT. The VAT also benefits agricultural processors because primary agricultural products are used as inputs to their production activities. The reduction in the price of primary products via the reduction in - 186 - taxes decreases the cost of processed agricultural commodities and allows agro-based industries to be more competitive. Cedit. Government policy is to make credit available where it is needed. In the past, the government has attempted to alleviate poverty in the agricultural sector by providing subsidized credit. This strategy, however, had negative effects on the viability and efficiency of the rural financial system, as well as on income distribution because most low-interest loans benefited not small farmers, but rather owners of large farms. In 1986, the government slowly began phasing in a policy of interest rate liberalization, avoiding direct lending to farmers. Its primary role is to ensure a macroeconomic and policy environment that supports private sector activities in the rural financial market. Accordingly, and in recognition of the high risk faced by banks in lending for agriculture-related loans, the government has developed a scheme whereby agricultural loans are guaranteed up to 85 percent of default risks. The program, the Consolidated Agricultural Loan Fund (CALF), is an incentive for rural and commercial banks to increase their lending portfolios for agricultural projects. The program increases available credit funds and results in a uniform effective interest structure across sectors. This scheme is also beneficial to agricultural processors because primary agricultural products are used for diversification and/or increased production. Similarly, another major program, the Livelihood Enhancement for Agricultural Development (LEAD), has been implemented to assist small farmers in establishing site-specific, agro-based projects to be run as business propositions, with guidance from the Management Association of the Philippines (MAP). Generally, this is a credit program that allows farmers to shift from subsistence to market- oriented production activities, in line with the requirements for diversification. The Integrated Rural Financing Program (IRFP), managed by the Land Bank of the Philippines, extends financial services to small farm households in support of profitable projects, including production, postharvest activities, marketing, and small rural nonagricultural enterprises. Because the loans are based on the financing requirements of farm households and not on a per commodity or activity basis, the financing usually covers crop-livestock production, small-scale agricultural processing businesses, or even nonfarm enterprises-all funded by a single loan to the farm household. The IRFP therefore encourages diversification at the farm level. Programs and Projects Involving Agricultural Diversification The 1987-92 Medium-Term Philippine Development Plan identifies agricultural diversification as an agricultural development strategy, and there are several programs and projects being implemented by the Philippine government that are supportive of agricultural diversification. These projects, however, are primarily limited to crop diversification in rice-based systems. Irrigation Management for Crop Diversification This project, implemented by the International Irrigation Management Institute (IIMI) in collaboration with the National Irrigation Authority (NIA) and researchers in several state colleges and universities, identified the constraints to and opportunities for crop diversification in irrigated areas. Among others things, the study involved determining irrigation practices that would enhance the cultivation of selected diversified crops in areas receiving limited irrigation water during the dry season. Technically, the project identified strict control over the allocation and distributionof water as a necessary condition for successful crop diversification at the farm level. Without it, farmers will be more inclined to grow rice despite the possibility of earning higher profits with diversified crops. Short-maturing - 187 - varieties of nonrice crops were found to be essential in providing farmers with greater flexibility and productivity in the choice of cropping patterns that would suit water supply conditions. Adriano (1988) reports that production costs of nonrice crops are much higher than those of rice. Credit support at market rates is therefore needed to induce farmers to diversify their production. Farmers with good experience in producing diversified crops were found to be willing to shift production to these crops. The provision of postharvest facilities to process and store nonrice crops is essential, considering their specific technical requirements. Generally, the availability of these facilities contributes to price stabilization. Bicol Farming Systems Research and Development Project This project was designed to develop institutional mechanisms and a policy framework for community-based management of settled upland forest, rainfed agricultural areas, hilly lands, and coastal zones. Generally, farmer exposure to the production of nontraditional crops improved their rate of technology adoption, thus enhancing their options for coping with weather variabilities. KABSAKA Rainfed Fanming Systems Project This was a community-based project aimed at increasing cropping intensity and crop-livestock production in monocrop rice areas in the Philippines. This was the first major project supportive of multiple cropping to raise agricultural productivity in rainfed areas and was considered successful in terms of introducing technologies that increase cropping intensity. Fisheries Subsector Program This is a subsectoral reform program begun in 1989 to support a realignment of the Philippine government's policies and institutional support to the subsector for (a) sustainable fisheries resource management; (b) regeneration/rehabilitation of coastal zones; (c) poverty alleviation and occupational diversification among marginal fishermen; (d) intensification and diversification of aquaculture production for the domestic and export markets; and (e) optimum exploitation of offshore, deep-sea resources. Livestock Subsector Program This program aims to alleviate poverty by focusing on the development of the livestock and poultry industries. Technical assistance is currently being undertaken in preparation of this program to identify investment opportunities, support services, and incentives designed to encourage increased production in livestock products. Emphasis will be given to the needs and opportunities of smallholder farmers to increase income. Horticulture Development Program The Philippine Department of Agriculture is in the process of packaging this program, aimed at developing the country's horticulture industry. The agricultural products/areas identified under this program include sericulture, abaca product diversification, other fibers, vegetables, fruits, spices, and cut flowers. This program will contribute to poverty alleviation in rural areas via the market-oriented horticultural production activities of small farmers. A technical assistance grant is being proposed to study how such a program may be effectively implemented. - 188 - Small Coconut Farms Development Program This is a World Bank-funded project aimed at rehabilitating small coconut farms in the Philippines. As mentioned earlier, coconut farms in the Philippines have low productivity levels, mainly because of the inability of farmers to replant and rehabilitate the existing plantations. This project has just been approved for implementation by the Philippine Coconut Authority, in collaboration with the Department of Agriculture. Conclusion The approach to agricultural diversification in the Philippines has been limited, since the emphasis is on crop diversification after the rice requirement has been fulfilled. One way to solve this problem is to expand irrigated areas. The increase in wet-season rice production would release dry-season irrigated land to diversified cropping. Considering that nonrice crops require only 60 percent of the water requirement for rice, the benefits of irrigation can be spread in more areas, leading to an expansion of the area for diversified cropping. This approach to the process of agricultural diversification, however, is narrow. A comprehensive approach should consider the livestock, poultry, and fisheries subsectors, all integrated into the system. This is important for a poverty alleviation program because the demand for these commodities increases with income. These commodities-for example, prawns-also have the potential to bring in more export earmings. In addition, they provide possibilities for the development of a stronger linkage between agriculture and industry, through its feed manufacturing requirements, which in turn need inputs from agriculture, specifically from the crops subsector. Future action in agricultural diversification should consider these linkages. References Abiad, V., and G. Llanto. 1989. Credit Policies Affecting the Small Farmer: Lessons from the Past and Present Strategy. Agricultural Credit Policy Council Staff Paper Series 89-01. Manila: Philippine Department of Agriculture. Adriano, M.S. 1989. lImplications for Policy on the Studies on Profitability of Irrigated Non-rice Crop Production: A Synthesis." In A. Valera, ed., Crop Diversification in Irrigated Agriculture in the Philippines. Colombo, Sri Lanka: IIMI. Adriano, M.S., and V.E. Cabezon. 1987. "Economic Policies Affecting Crop Diversification in the Philippines.' In IIMI, Irrigation Management for Diversified Cropping. Digana Village, Sri Lanka. Balisacan, R.C. 1989. 'Philippine Agricultural Development in Historical Perspective." In M. Montes, ed., Philippine Macroeconomic Perspective: Developments and Policies. Tokyo: Institute of Developing Economies. Cedillo, J.L. 1989. 'Policy Options for the Philippine Corn Industry." National Economic and Development Authority (NEDA) Staff Paper Series, vol. 1, no. 4. The Philippines. Clarete, R. 1989. Correcting the Anti-Agriculture Bias of Government Policies. Research and Training for Agricultural Policy Project Technical Paper. Manila: Philippine Department of Agriculture. - 189 - Corpuz, Jovita. 1989. "Agricultural Diversification in the Philippines." Paper presented at the Study Meeting on Agricultural Diversification, October 17-26, Tokyo, Asian Productivity Organization. David, C., et al. 1986. Agenda forActionfor the Philippine Rural Sector. Los Bahos, the Philippines: Agricultural Policy and Strategy Team, University of the Philippines at Los Bafios-Agricultural Policy Research Program. Miranda, S.M. 1989. Irrigation Management for Crop Diversification in Indonesia, the Philippines, and Sri Lanka. A Synthesis of IIMI's Research. International Irrigation Management Institute Technical Paper 1. Colombo, Sri Lanka. National Food Authority, the Philippines. Various unpublished data. National Statistical Coordination Board, the Philippines. Various years. NEDA Statistical Yearbook. Nolledo, J.N., and M.S. Nolledo. 1988. National Internal Revenue Code. 12th ed. Manila: National Bookstore. Paderanga, C.W., Jr. 1989. "Employment in Philippine Development." In M. Montes, ed., Philippine Macroeconomic Perspective: Developments and Policies. Tokyo: Institute of Developing Economies. Philippine Bureau of Agricultural Statistics. Various unpublished data. Philippine Department of Agriculture. 1988. Trade and Philippine Agriculture. Manila. Rosegrant, M.W., L.A. Gonzales, H.E. Bouis, and I.F. Sison. 1987. Price and Investment Policies for Food Crop Sector Growth in the Philippines. Washington, D.C.: International Food Policy Research Institute. Schuh, G.E., and S. Barghouti. 1988. 'Agricultural Diversification in Asia." Finance and Development 25: 41-44. I THE DIVERSIFIED CROPS IRRIGATION ENGINEERING PROJECT IN THE PHILIPPINES Alberto S. Adrias Agricultural diversification is an essential element in the process of upgrading the living conditions of Philippine agricultural workers, estimated at 50 percent of the national work force. Recognizing the need to develop irrigation technology for crop diversification, the Philippine government has embarked on a technical cooperation endeavor, the Diversified Crops Irrigation Engineering Project (DCIEP), with the Japan International Cooperation Agency (JICA). The project's main objective is to formulate comprehensive engineering technology for diversified crop irrigation through applied research that considers water management requirements, soil texture and agronomic characteristics, and the required physical facilities for irrigation and drainage. This paper provides a brief overview of the Philippine agricultural sector and then describes the components of the DCIEP. Background The main objective of rural development is to improve rural living standards. Promoting agricultural productivity strengthens community infrastructure and creates more jobs in agriculture and other related industries, contributing to national efforts to attain economic stability and prosperity. The majority of the Philippine population depends on agriculture, either directly or indirectly: agricultural production accounts for 30 percent of GDP, and agricultural products and processed foodstuffs account for 60 percent of total exports. Agricultural progress is thus a major concern of the state, and self-sufficiency in staple crops remains a top priority of the government. Efforts to effectively advance agricultural development include efficient management of irrigation water, in conjunction with improved agricultural practices. Consequently, the development of irrigation in the Philippines has concentrated mainly on ensuring an adequate supply of water during the growing season. In the 1970s, the Philippines attained self-sufficiency in rice production and has since promoted the cultivation of nonrice crops. Based on 1988 statistics from the Bureau of Agricultural Statistics, the agricultural harvested area is 13.4 million ha, or about 45 percent of the Philippines's estimated 30 million ha of land area. Major crops planted are rice and corn for domestic consumption, and coconut and sugarcane for export. Table 1 lists the area planted in major crops. Rice area in the Philippines is classified as irrigated or rainfed. In 1988, the total irrigated and rainfed areas amounted to 1.96 million ha and 1.44 million ha, respectively. For the first half of the year, the dry season, the area supplied with irrigation totaled 1.37 million ha, or 40 percent of paddy area. Between July and December, the wet season, only 2.02 million ha, or about 60 percent of paddy area, were irrigated. The less-than-100 percent provision of irrigation water throughout the service area, even during the wet season, occurs because of low water levels in the rivers being tapped for irrigation. Water resources in the Philippines are difficult to develop at low cost because rivers do not have wide catchment areas and irrigation projects face obstacles in benefiting large areas. Furthermore, exten- Alberto S. Adrias is an engineer, National Irrigation Administration, the Philippines. - 192 - sive deforestation has reduced the soil's absorptive capacity, leading TABLE 1. AREA PLANTED to erosion of fertile topsoil and severe flooding during heavy rainfall. IN MAJOR CROPS (HA) Other important factors include the untimely and uneven occurrence of effective rainfall and the inadequate management of usable water CROP AREA supplies. Rice 3,392,700 Statistics for 1988 from the National Irrigation Administration Corn 3,745,100 (NIA), the government agency charged with developing water Coconut 3,523,300 resources, showed that the agency's irrigation facilities covered 1.45 Sugarcane 272,600 Coffee 149,000 million ha, or 46.4 percent of the total potential irrigable area Cacao 17,200 regionwide (tables 2 and 3). Despite these efforts, only 73 percent of Fruits/nuts 751,200 the NIA total service area in the wet season and 55 percent in the dry Vegetables 1,217,600 Nonfood 319,000 season were actually irrigated, and less than 1 percent of that area was Total 13,387,700 planted with nonrice crops (table 4). In this context, the suggested solution to upgrading the living Source: Bureau of Agricultural conditions of agricultural workers is the cultivation of inherently less- Statistics. water-demanding crops with higher market value on areas with I insufficient irrigation water supply or on completely nonirrigated rice lands. This would provide additional benefits to farmers in terms of increased profitability, particularly during dry-season croppings, and extra time to tend to other matters. Recognizing the need to develop irrigation technology for crop diversification, the Philippine government, through the NIA, asked the Japanese government for technical assistance in the agency's first attempt at crop diversification activities. In response, Japan, through the Japan International Cooperation Agency (JICA), provided technical assistance through a cooperation program to launch the Diversified Crops Irrigation Engineering Project (DCIEP). The Diversified Crops Irrigation Engineering Project DCIEP activities were officially launched with the signing of the Records of Discussions in May 1987 between the Japanese governrnent (through JICA) and the Philippine government (through the NIA), TABLE 2. TOTAL SERVICE AREA BY IRRIGATION SYSTEM, 1979-88 IRRIGATION SYSTEMS SERVICE AREA (HA) % IRRIGATION YEAR NATIONAL COMMUNAL PUMPS TOTAL DEVELOPMENT 1979 475,174 552,092 152,128 1,179,394 37.7 1980 472,182 579,751 152,128 1,204,061 38.5 1981 491,729 602,081 152,128 1,245,938 39.9 1982 514,334 634,102 152,128 1,300,564 41.6 1983 549,930 648,837 152,128 1,350,895 43.2 1984 548,345 658,807 152,128 1,359,280 43.5 1985 567,678 665,099 152,128 1,384,905 44.3 1986 595,902 668,828 152,128 1,416,858 45.3 1987 596,953 673,119 152,128 1,422,200 45.5 1988 614,164 684,639 152,128 1,450,931 46.4 Source: National Irrigation Administration. - 193 - TABLE 3. TOTAL SERVICE AREA BY REGION, 1988 POTENTLAL IRRIGABLE IRRIGATION SYSTEMS SERVICE AREA (HA) % IRRIGATION REGION AREA (HA) NATIONAL COMMUNAL PUMPS TOTAL DEVELOPMENT 1 309,810 45,382 132,139 5,520 183,041 59.1 2 539,710 153,287 81,105 36,593 270,985 50.2 3 263,220 173,394 84,260 22,946 280,600 58.2 4 263,590 54,982 69,019 27,948 151,949 57.7 5 239,650 16,506 50,296 16,943 83,745 34.9 6 197,250 53,500 31,733 21,677 106,910 54.2 7 50,740 - 18,426 2,481 20,907 41.2 8 84,380 15,633 34,421 2,176 52,230 61.9 9 76,500 12,991 21,327 2,804 37,122 48.5 10 230,150 20,213 43,957 2,045 66,215 28.8 11 290,250 34,186 61,495 6,872 102,553 35.3 12 362,080 34,090 56,461 4,123 94,674 26.2 Total 3,126,330 614,164 684,639 152,128 1,450,931 46.4 Source: National Irrigation Administration. TABLE 4. SERVICE AND IRRIGATED AREAS, CALENDAR YEAR 1988 (HA) IRIUGATED AREA SERVICE WET DRY THIRD ANNUAL REGION AREA SEASON SEASON CROP CROP TOTAL 1 45,382 28,180 13,311 0 14 41,505 2 153,287 98,864 95,821 52 0 194,737 3 173,394 136,092 83,298 0 0 219,390 4 54,982 36,204 24,817 0 0 61,021 5 16,506 13,181 13,173 0 0 26,354 6 53,500 42,532 29,269 0 0 71,801 7&8 15,633 11,399 10,255 0 0 21,654 9 12,991 11,349 10,135 658 0 22,142 10 20,213 11,988 13,790 - - 25,778 11 34,186 28,321 24,348 0 2,915 55,584 12 34,090 28,104 18,041 0 0 46,145 Total 614,164 446,214 336,258 710 2,929 786,111 Benefited area (%) - 73 55 0.11 0.48 - Source: National Irrigation Administration. establishing a five-year technical cooperation program that ends in May 1992. Project implementation is based on the findings and recommendations of a joint study conducted by the International Food Policy Research Institute and the International Rice Research Institute, commissioned by the Asian Development Bank. The shift toward diversified irrigated agriculture will focus on planting low-water-requirement crops, especially in areas that have insufficient water supplies during the dry months. The NIA's existing irrigation systems nationwide will be considered for this planting. - 194 - Project Conponents Under the terms of the Records of Discussion, the main responsibilities and undertakings assigned to both govermnents are as follows: Philippine government: 1. creation of a joint committee comprising Japanese experts, JICA representatives, and NIA's management officials to assess project performance and formulate job plans; 2. appointment of a full-time project manager and staff; 3. assignment of counterparts to expatriate JICA experts; 4. provision of office space, facilities, and agricultural land for research; 5. provision of transportation services. Japanese government: 1. deployment of Japanese experts in different fields of irrigation engineering, headed by a team leader; 2. provision of equipment and materials; 3. acceptance of Philippine personnel for training in Japan; 4. construction of a trial farm. Objectives As noted earlier, the project's main objective is to develop a comprehensive engineering technology package for diversified crop irrigation. The specific project objectives are to: 1. study the most economical method of providing irrigation to diversified crops; 2. formulate guidelines for planning and designing irrigation and drainage facilities for nonrice crops; 3. measure the effects of crop diversification on land areas with different soil characteristics; 4. develop appropriate farm management techniques for nonrice crops; 5. conduct technical training for NIA personnel and an intensive information campaign on the DCIEP and its technology. - 195 - Organization The project is assigned to the Philippine Office of the Assistant Administrator for Systems Operations and Equipment Management, who chairs the joint committee. The committee reviews project progress and approves the proposed annual job plans. The project is headed by a manager, who oversees nine sections and the irrigation engineering experts' team leader. Sectional Concerns The different project sections perform the following functions: Project Management * provides technical and administrative services to the project; * coordinates work in the head and field offices; evaluates and monitors activities of the other sections. Planning Criteria * studies adaptability of diversified crops under various conditions; prepares agricultural development studies and compiles statistics; * evaluates physical facilities suited for different crops. Design Criteria * establishes irrigation systems and components; oversees installation/observation of hydrometeorological stations; 3 designs irrigation and drainage facilities and procedures applicable to nonrice crops. Water Management * determines water requirements for each crop through research and field tests; * recommends appropriate methods of irrigating and maintaining low-water-requirement crops. Pedology * describes and delineates soil characteristics nationwide; * identifies suitable soil types for each diversified crop through field tests and laboratory experiments. Agronomy * prepares diversified crop production reports and compiles relevant statistics; * determines farm management techniques and cropping calendars and patterns; * specifies proper management and cultural requirements of diversified crops. - 196 - Tial Farm Management * supervises farming activities; * provides physical facilities for sectional field tests, experiments, and research studies; * coordinates other services such as service/repair of laboratory apparatus, storage of harvested crops, and operation and maintenance of the farm and machinery. Economics * conducts studies on the suitability of diversified crops, considering profitability and markets; * carries out cost-benefit analyses and socioeconomic studies on the effects of diversified cropping in a particular area. Training * prepares training designs, programs, presentation papers, reports, and facilities; * conducts training on the different criteria each section will formulate; also handles information campaign and training evaluation. Research Programs Among the research/field-test activities being conducted at the trial farm, topics include: * consumptive use rates for 17 common nonrice field crops; irrigation scheduling for furrow and border irrigation methods, including soil infiltration; * on-farm irrigation distribution design for efficient water application; * significant soil layer and its relationship to plant-rooting depth and consumptive use rates; * soil water-holding capacity after 24- and 48-hour standard draining periods; * crop growth and yield response to cropping calendars with November and January as planting months; * percolation rates under flooded conditions and water table behavior under on/off flooding conditions; * amount of irrigation water utilized in relation to actual cropping works. The sample crops being planted as part of the research studies include bittergourd, cabbage, carrot, corn, cucumber, eggplant, garlic, mungbean, okra, onion, peanut, pepper, radish, squash, sweet potato, tomato, and watermelon. Expected Benefits A second JICA grant-in-aid project, the Diversified Crops Irrigation Engineering Center, will be used to disseminate the technologies that evolve from the studies and field tests of the DCIEP. After - 197 - project completion and eventual field implementation, increased cropping intensity in irrigation service areas will usually mean improved opportunities for the NIA to attain and maintain agency and system viability. Additional nonrice crops will indicate increased benefits in the form of higher income for farmer beneficiaries, thus improving economic conditions and minimizing under- and unemployment in rural areas. The nationwide promotion of the technologies derived from the project will enhance agricultural development in the country, leading toward its ultimate objective-food self-sufficiency. PUBLIC POLICIES TOWARD AGRICULTURAL DIVERSIFICATION IN THAILAND Amnar Siamwalla, Direk Patamasiriwat, and Suthad Setboonsarng In the immediate postwar period, rice was central to the Thai economy, contributing about half of Thailand's foreign exchange earnings and employing more than two-thirds of the labor force. Revenues from the rice export tax constituted almost one-third of government revenues in the late 1940s and about one-fifth in 1958 (Ingram 1971: 226). For most of the postwar period, rice was the largest export item, and it remains the largest agricultural export. Despite its continuing importance, however, its share in the economy and in crop agriculture has been steadily declining and, within this context, Thai farmers have been steadily diversifying away from rice. (Throughout this paper, unless specified, agriculture refers to crops agriculture only.) The factors contributing to this diversification are complex and still unclear. In the 1960s Thailand was a land-surplus country, with population growth that exceeded 3 percent annually; thus, there was natural pressure to bring more land under cultivation. Much of this was upland with clear comparative advantage in crops other than rice. With these new lands coming under cultivation, new crops were grown, each of which had its golden age of rapid expansion, followed by a leveling off of the growth rate. Maize was the first such crop in the late 1950s and early 1960s; kenaf in the mid-1960s; sugarcane in the late 1960s and early 1970s; and cassava in the 1970s. Table 1 shows the growth rates of cultivated areas under different crops. Rapid land expansion and automatic diversification TABLE 1. AVERAGE ANNUAL ended in the 1980s as a result of two factors. First, the land PERCENTAGE GROWTH RATES surplus that had fueled the rapid expansion of agricultural FOR SELECTED CROPS, 1%917 growth and diversification ended in the late 1970s-the GRowrH upward trend in land per agricultural worker reversed around cRoP RATE (%M 1978. Second, the high prices that characterized the markets for most agricultural products in the 1970s, particularly for Ruice 1395 carbohydrate products, reversed sharply after 1982. Vegetables 1.49 The decline of agricultural prices during the 1980s, Trees 2.81 and of rice prices in particular, led many to advocate diversification of Thai agriculture away from rice. Before Note: Trends are estimated using examining public policies toward diversification, this paper wemilogarithmic regresions. discusses the rationale for public sector intervention to Source: Cultivated area statistics collected by promote diversification. Individual sections then focus on the Ministry of Agriculture and Agricultral promote ~~~~~~~~~~~~~~~~~Cooperatives and compiled by the Thailand public policies in price and marketing, credit, irrigation, Development Research Institute. research, and extension. In addressing these issues, however, the paper takes no position on whether public policies to force diversification per se are desirable. Ammar Siamwalla is president, Thailand Development Research Institute, Bangkok; Direk Patamasiriwat is vice rector, Saresuan University, Pitsanuloke, Thailand; and Suthad Setboonsarng is visiting faculty member, Asian Institute of Technology, Pathum Thani, Thailand. - 200 - Rationale for Public Intervention The term diversification is normally used in the investment literature. Investment risks can be reduced by commitment to a diversified portfolio of assets. As agriculture is inherently a risky activity, and involves an investment of resources, one can naturally apply the idea of diversification to its analysis. As risk markets are often absent, there is a prima facie case for government intervention. What sort of government intervention is required, however, remains murky. A first step in the analysis is the distinction between the problem of diversification as faced by the economy as a whole and as faced by the individual farmer. For an agriculture-based economy, overconcentration on a few crops may be socially undesirable because fluctuations in the production and prices of these crops will have severe spillover effects on tax revenues and on other sectors of the economy. If agriculture plays a smaller role, however, it is not clear that the risk-minimizing benefits of diversification are worth a major government effort. In the Thai context, the spillover effects of fluctuations in Thai agricultural production on the rest of the economy are slight, as agriculture contributes only one-sixth of GDP. The sector's direct contribution to tax revenues is probably less, as most export taxes have been removed, and land and income taxes levied on agriculture are minimal. At the farm level, without risk markets, one would expect that farmers themselves would have to adopt risk-reducing strategies, among which diversification would feature prominently. Diversification enables farmers to cope with uninsurable price and quantity risks on their own. Relative to a scenario in which these risks can be insured, diversification will be socially inefficient, as risks are not tradable across farms, and the value of the aggregate output would be less than if such markets did exist. Thus, left to their own devices, farmers will overdiversify rather than the reverse. In Thailand, slightly more than half the farmers grow more than one crop (table 2). Many farmers do not diversify within their own farms, although the average farm size is about 18.8 to 25 rai (3 to 4 ha) for paddy, and 31.2 to 50 rai (5 to 8 ha) for upland crops. Why do these farmers not diversify? For paddy, the nature of the land itself and a lack of water control limit the scope for diversification. Even in upland areas, however, where many crops can potentially be grown, farmers usually do not diversify. Most diversification has occurred where farmers grow multiple crops, usually rice and another upland crop. As land use substitution between rice and upland crops is limited, this pattern of landholdings reveals that farmer choices have been dictated by the need to reduce weather risks by owning both rice and upland areas, and that the reduction of price risk is merely a by-product. Should the government nonetheless be concerned about the extent of diversification beyond what farmers themselves wish? There are several reasons why there should be such concern. Classically, it has been argued that, for an open agricultural economy with a degree of monopoly power in world markets-such as Thailand's-there is a limit to the amount of any product it can sell in the world market without having an adverse impact on its terms of trade. There is, in other words, an optimal level of exports above which production costs will be greater than the marginal revenue extra exports generate. Under these circumstances, to allow private competitive markets to perform their function would lead to suboptimal results. For Thailand, rice clearly belongs in this category, as does cassava since Thailand's voluntary export restraint agreement with the European Community, which requires government intervention to control export volumes. Another reason involves current government interventions in agricultural production and markets and the nonneutrality of these interventions in their impact on various crops. Several questions are germane: How much influence have government policies had on crop diversification? Because shifting international prices for many crops will affect the returns from government expenditures to promote their production, how flexible has the government been in adapting its activities? Specifically, as the world rice market weakened in the 1980s and is expected to resume its downward trend, what role should the government take in facilitating a move by farmers away from rice? -201 - TABLE 2. CROPS GROWN BY FARM SIZE, 1982 FARM SIZE (RAD CROPPING ALL SYSTEM SIZES < 6 6.0-9.9 10.0-39.9 > 40 Single crops 2,032,474 455,680 280,897 1,076,345 219,552 Rice 1,458,104 276,795 206,587 816,783 157,939 Upland 272,560 66,767 31,955 138,595 35,243 Vegetables 25,329 18,296 3,340 3,412 281 Trees 276,481 93,822 39,015 117,555 26,089 Multiple crops 2,397,869 167,890 270,485 1,502,711 456,783 Rice-other 2,196,461 132,987 248,077 1,397,186 418,211 Other combinations 201,408 34,903 22,408 105,525 38,572 Uncultivated holdings 33,546 31,468 381 1,576 121 Total 4,463,889 655,038 551,763 2,580,632 676,456 Note: 1 rai = 0.16 ha. Source: National Statistical Office, 1983 Intercensal Survey of Agriculture. Price and Marketing Policies Throughout the developing world, internal marketing of agricultural crops is handled to some extent by government agencies, with a significant impact on market prices and on the availability of marketing services in rural areas. Thailand has been almost unique in that the role of state enterprises in marketing has been negligible. Although the government's Marketing Organization for Farmers was established in 1975 to engage in various marketing activities, and was assigned occasionally to help implement various government price support programs, the effectiveness of these programs on actual farm prices is questionable (Siamwalla and Setboonsarng 1989). Consequently, internal marketing in Thailand is almost entirely the domain of private traders, who have over the years created a sophisticated marketing system from the farm level to export markets. Except for some commodities (notably sugar), the marketing systems are highly competitive; where they are not, govermnent policies helped allow the trade to "cartelize" itself (Usher 1967). The government, however, has not refrained from pursuing effective price policies. Since Thailand is an exporter of most of its agricultural commodities, the government has had powerful and smooth means to influence the domestic prices of these commodities through border taxes-powerful, because the government can induce substantial changes in prices by simply varying export taxes; smooth, because there is no need for public capital to be expended on warehouses and transport facilities, and the intensity of the intervention can thus be varied at will. Because the government has had more interest in lowering than in increasing domestic prices relative to world prices for most of the period under consideration, these taxes usually yield revenue for the government as a by-product. This revenue has given the government financial flexibility that has allowed it to pursue several interests without the discipline of a budget constraint. Without this discipline, many government agricultural policies are based on political rather than economic or financial considerations. Only recently, when its objectives changed and the budget constraint became binding, did the government begin to have trouble with this mode of intervention. - 202 - The history of border intervention in agricultural products has been much studied and documented (Siamwalla 1975 for rice; Jessadachatr 1977 for sugar; Siamwalla and Setboonsarng [1989] updated both these studies and did additional work on maize and rubber). The following summarizes this earlier work from the perspective of diversification. The Thai government's budgetary system is highly inflexible and permits few expenditures on any subsidies, particularly if subsidies will fluctuate a great deal and move upward. Fluctuations in tax revenues are more tolerable, but even if a tax's contribution to the budget were large, there would be great reluctance to allow the total intake to vary. Thus, in the 1950s and the early 1960s when the rice premium constituted more than 10 percent of the government budget, there was considerable reluctance to vary the rice premium, which, significantly, was set as a specific tax. Only when government dependence on these taxes decreased was the government willing to adjust the rates to achieve other objectives. After 1966, when the premium began to contribute a smaller share of the total revenue, the government began to vary the premium rates frequently and introduced a progressive export tax on rubber, signaling a shift to a greater emphasis on price stabilization as a policy objective. The impact of these rigid government policies can be seen in table 3, which delineates the deviation of domestic prices from world prices for eight commodities. With the exception of sugar, the pattern in the past is clear and consistent: exportables tend to have their prices pressed below world prices, and producers of importables enjoy prices above world levels. Both these deviations generate revenues for the government. With sugar, the concentrated nature of the industry allows the government to pursue a two-price policy, taxing domestic consumers to finance the subsidization of exports (with zero impact on the treasury), thus lifting the domestic price above the world level, except in years when world prices are at exceptionally high levels. Subject, then, to the constraint that pricing should have either positive or zero impact on the budget, what objectives has the government pursued, and what impact have they had on diversification? For many of the commodities whose history of price intervention is shown in table 3, an important objective has been stabilization of domestic prices. Specifically for rice, domestic real price instability TABLE 3. EFFECT OF DIRECT INTERVENTION ON DOMESTIC RELATIVE PRICES OF SELECTED AGRICULTURAL COMMODITIES, 197046 PROPORTIONATE DEVIATION FROM REAL WORLD PRICE SUGAR PALM YEAR RICE MAIZE GROWER MILLER CONSUMER RUBBER CASSAVA COTTON SOYBEAN OIL 1970 4.129 -0.026S 0.23S2 0.4085 0.4939 .0.1303 0.0X00 1.1780 n/a nia 1971 4.253 4.0358 0.0216 0.2799 0.4287 40.1058 0.0000 1.1658 nal nl 1972 4.2682 .0.0622 40.2201 0.0252 0.1451 .0.1101 0.0000 1.0385 nta nta 1973 4.4044 4.0971 4.2932 4.1114 4.1547 4.1712 0.0000 1.1054 n/a nts 1974 4.4755 4.0258 4.6224 4.4092 4.56S5 4.1B88 0.0X0 0.3104 iA i14 1975 4.3176 4.0650. 4.578 4.4082 4.6353 4.1661 0.0000 0.4198 n/A n/a 1976 4.1576 4.0327 4.2078 4.1127 4.2495 .0.2083 0.0X0 0.1774 4.0753 nta 1977 4.2199 0.0033 4.0256 4.0280 4.1067 4.2223 0.0X0 0.0258 0.1824 4.0198 1978 4.3106 4.0229 03491 0.0423 0.1462 4.2275 0.0X0 0.1001 0.1550 4.0390 1979 4.2392 4.0461 0.3921 0.1047 0.2117 4.2449 0.0000 0.0740 0.16S8 4.0766 1980 4.2611 4.0545 4.0282 0.2105 0.3565 4.2611 0.0000 0.0870 0.1398 4.0820 1981 4.2573 4.0073 4.1017 4.0093 4.0057 4.1890 0.0X0 0.0301 0.1183 0.0001 1932 4.1155 0.000 03609 0.0841 0.6M05 4.1312 0.0000 0.0254 0.2886 0.0715 19s3 4.0805 0.000 0.5158 03512 1.3564 4.1795 0.0000 0.3791 0.2455 0.0934 1964 4.0611 0.0000 0.5310 0.4857 1.2781 4.1509 4.0034 0.2234 0.2606 0.1395 1965 4.0442 0.0x0 0.9672 0.2380 1.8329 4.1056 4.1734 0.4538 0.1553 4.1186 1966 4.0030 o0. 030 .5 0.0632 1.6537 4.1107 4.02D 0.6423 0.3413 0.3176 Note: For rice and cassava, of which Thailand is a major producer, the deviation is based on Thailand unilaterally elimninating intervention. Source: Siamwalla and Setboonaarng 1989 for 1970-85 figures for rice, maize, sugar, and rubber; additional calculations made for the renmaining commodities and for 1986. - 203 - between 1975 and 1984 was about one-quarter of the world price instability, achieved through variations in border intervention. In interpreting these numbers, however, one has to keep in mind that Thailand is a major participant in the world rice market. Stabilization of the domestic rice price through border intervention, therefore, implies an automatic destabilization of the world price. To the extent that Thai rice farmers have thus been insulated from world price instability, they have been spared the need to diversify out of rice into other crops to minimize price risks. As the prices of most other crops (except sugar) have been less insulated from international prices, farmers have less incentive to diversify away from rice toward these other crops. The policy of insulation could be claimed a successful effort to compensate for the absence of risk markets for farmers. Often heard, however, is that the level of past rice export taxes had the effect of pushing farmers away from rice toward a more diversified cropping pattern. Normatively, why the government should want to do this is difficult to understand. From a positive point of view, the impact is probably small, as the cross-elasticity in production of upland crops (as a group) with respect to rice price is only 0.03. Regardless, although this may have been the impact in the past, one clear point that emerges from table 3 is the broadly declining trend in the impact of negative price intervention, particularly for rice. Among exportables, which are the bulk of Thai agriculture, the persistent bias of price policies against rice ended in 1986, when all taxes on rice exports were removed after more than half a decade in which they had been continually decreased. This more recent government stance with respect to agricultural (particularly rice) prices results from its policy objective since the mid-1970s of supporting producer interests. At the same time, one could argue that the decline and final elimination of border taxes for rice and other commodities have been a response to the declining world prices that occurred in the early and mid-1980s. But this would be belied by the government's action when the world market for most agricultural commodities tightened after 1987, caused first by the 1987 Thai drought that affected rice prices, and then by the 1988 North American drought, which affected the prices of other agricultural commodities. The government permitted a rise in domestic rice prices of about 50 percent over a period of 18 months without intervening. Surprisingly, there was little political pressure against the surge in prices, whereas a few years earlier, when the rice price was low, the government was under considerable pressure to lift it. One possible explanation for this approach is that rice now constitutes only 4 percent of the average Bangkokian's budget. Also, perhaps public opinion may have been anesthetized by the rapid growth rates of the last three years. Meanwhile, pressure from producers has gathered steam. Recently, there has been a series of farmer protests demanding more support for various crops, to which the government response has tended to be more cosmetic than effective. The Thai government must now tackle how to shift from taxing agriculture to protecting it. For importables, such protective policies are already practiced, as is partly indicated by figures for palm oil, soybean, and cotton (in table 3), and dairy products. For exportables, particularly rice, the problem is more serious. Export taxes have been eliminated, and hence the government can no longer support farmers by taking less away from them: its only alternative is export subsidies, and here the budget constraint has finally caught up with the government. Funds now have to be earmarked from the budget for agricultural subsidization and procedures need to be established to implement an ongoing and permanent subsidization program, instead of a panoply of ad hoc programs more typical of past measures for agricultural prices. In designing such a permanent program, several issues must be addressed. Use of border measures has to be carefully considered. Although using them as a lever to lift domestic prices would minimize budgetary cost, international complications (e.g., the GATr, the American Export Enhancement Program) will have to be taken into account. The alternative path of intervention in the domestic market for tradable goods would both involve a bigger drain on the budget and require improvements in the efficiency of the government's marketing organization and its ability to deal in domestic as well as international markets. - 204 - Such protectionist policies, although politically expedient, would freeze Thai agriculture within its present structure and would fly in the face of any reasonable definition of diversification. Indeed, if Thailand has monopoly power in rice, it is difficult to justify an export subsidy rather than an export tax. Yet if diversifying means facilitating a transition to subsectors with a potential for rapid growth, it is then difficult to avoid choosing the "sunrise" subsectors to which resources should be devoted, including support from government marketing and price policies. Such speculation has in recent years settled on high-value horticultural crops, such as fruits and vegetables. Although high recent growth in the Thai economy should boost domestic demand for these income-elastic food crops, most govermnent effort has gone into promoting their export-although this has not involved explicit subsidization. Regardless, it is difficult to predict an aggressive role for the government in marketing and pricing these crops (additionally, the number of holdings cultivating these crops is small [see table 2]). Because these crops are highly capital-intensive, though, there may be some role for credit policies. Credit Providing working capital through credit has been one of the government's more successful policy thrusts. The Bank for Agriculture and Agricultural Cooperatives (BAAC), established in 1969, has branches in every province in the country and has on its rolls more than 2 million farmers, or almost half the country's total. BAAC operations, and a 1975 Bank of Thailand directive to commercial banks to lend a percentage (presently 14 percent) of their deposits to the agricultural sector, including livestock, fisheries, and forestry, have pushed the share of the formal sector in total agricultural credit up from less than 10 percent in the mid-1960s to close to half (Siamwalla et al. 1990). BAAC credit is subsidized, with three modes of subsidization. Forcing banks to lend to the agricultural sector is an implicit tax on the nonagricultural sector used to subsidize agricultural loans (assuming the banking sector is competitive). More important, BAAC has been following an average-cost pricing rule, rather than using the more correct marginal-cost pricing, to set the interest rate it charges. Finally, since 1987, BAAC has embarked on a paddy mortgage scheme, in response to the government's policy. All these subsidies add up to a small annual amount (1 billion baht, or $40 million) (Siamwalla and Nettayarak 1990). Has the government been using credit to steer farmers in the direction it desires? To answer this question, one can examine special lending programs in BAAC's portfolio, usually funded from abroad (e.g., the European Community), designed to direct farmers away from crops that the government and foreign donors may want to discourage (e.g., cassava). These special programs are, however, only a small part of BAAC's loan portfolio: the bulk of loans (75 percent) is for short-term working capital loans, where BAAC has had the most impact on Thailand's agricultural sector. Table 4 reports the percentage share of various crops in value added to the crop subsector and in BAAC's short-term loan portfolio. Obviously, BAAC has been following a conservative lending policy. BAAC personnel repeatedly indicated in interviews that in their choice of clients, they tend to discriminate against monocultural farmers growing traditional crops, particularly in rainfed areas, indicating that they direct their funds away from most of the traditional crops. Yet the figures in table 4 reveal that, at a minimum, rice and cassava growers still receive more lenient treatment from BAAC loan officers than they would if the bank or the government had been following a policy of diversification. Tree crops, however, have not had a share commensurate with their importance. The figures in table 4, however, underestimate the credit received by rubber growers, whom the government supports through the separate Rubber Replanting Fund. An examination of BAAC's lending operations, particularly its short-term loans, indicates why it has not succeeded in changing the structure of its loan portfolio. At the beginning of each year, it provides working capital loans to farmers; at the end of the year, farmers are supposed to repay the loans - 205 - TABLE 4. PERCENTAGE VALUE ADDED (VA) AND BAAC AGRICULTURAL LENDING (L) FOR SELECTED CROPS, 197848 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1986' CROP VA L VA L VA L VA L VA L VA L VA L VA L VA L VA L VA L Rim 44.7 52.6 41.5 51.1 40.7 52.0 45.1 54.4 40.7 53.3 41.2 51.2 38.2 53.6 38.0 54.2 34.7 59.5 36.0 59.9 42.2 59.0 Cu.ava 7.7 13.9 9.8 15.5 10.1 16.0 5.6 14.9 6.9 16.1 9.3 17.0 6.9 13.1 6.2 12.6 12.6 11.7 8.8 12.4 5.6 12.5 Maim 4.3 10.4 5.1 12.0 5.5 12.5 5.2 11.8 3.7 12.0 5.4 12.9 6.7 12.6 6.1 13.3 4.0 12.8 3.2 11.8 6.8 12.3 SUga 3.8 8.3 3.8 7.7 2.2 6.2 8.0 7.9 10.3 7.0 4.9 5.9 6.7 6.8 4.1 5.7 5.4 3.2 4.3 2.3 4.4 2.0 Coudt 0.6 0.7 1.1 0.9 1.3 0.8 1.1 0.8 0.7 0.4 0.7 0.4 0.5 0.4 0.8 0.4 0.3 0.2 0.5 0.2 0.6 0.2 cr 16.8 2.5 19.9 1.4 19.5 1.4 17.9 2.7 17.6 3.6 17.9 3.4 21.9 4.0 24.0 4.5 24.7 4.5 27.3 5.1 24.1 6.1 Tobacco 1.0 1.5 1.1 1.5 1.0 1.1 0.9 1.2 1.8 1.6 1.1 1.4 1.0 1.4 1.1 1.6 0.9 1.5 0.8 1.5 0.5 1.4 Pinzapple 2.0 1.4 1.4 1.1 1.5 1.0 0.5 0.4 0.8 OA 1.1 0.5 1.7 0.6 1.4 0.9 1.1 0.2 1.7 0.3 0.8 0.3 Kenafrjue 1.3 1.4 0.8 1.0 0.9 0.8 0.7 0.9 0.8 0.9 0.7 0.8 0.8 1.5 1.4 1.8 0.6 1.0 0.4 0.8 0.4 0.6 Lagmcs 2.5 3.0 1.6 2.0 2.0 2.3 2.5 1.7 2.2 1.5 2.5 1.7 2.9 1.8 3.6 1.9 3.2 2.1 2.7 1.8 3.9 1.8 Otbhrs 15.5 4.4 13.9 5.8 15.3 5.9 12.5 3.4 14.6 3.1 15.2 4.8 13.0 4.1 13.4 3.2 12.7 3.3 14.2 4.0 10.8 3.9 Note: Lending portfolio comprises loans to farmers and cooperatives. a. Estimated. Source: National Economic and Social Development Board, National Accounts Division; annual report of the Bank for Agriculture and Agricultural Cooperatives, various years. in full-even if they will return in a matter of weeks seeking loans for the following year. Nondefaulting farmers expect almost automatic rollovers. In this case, BAAC cannot redirect its loans at will toward sectors that it favors at any given moment, unless its clients themselves wish to shift toward those sectors. In the past, the bank's rapid growth could have facilitated such redirections, at least with new clients: apparently, it has not chosen this option. In addition to BAAC's loans, twice as great a sum is lent to the agricultural sector by commercial banks. This figure, however, is somewhat inflated, as commercial banks appear to evade the Bank of Thailand's directive considerably. irrigation The development of modem irrigation in Thailand dates from a series of ambitious proposals in 1903 by van der Heide, a Dutch engineer hired by the government to design an irrigation system for the entire lower Chao Phraya with a series of storage and diversion dams (Ingram 1971; Small 1973; Feeny 1982; Brown 1988).1 Van der Heide's proposal was rejected, and a far more modest proposal made by Sir Thomas Ward was accepted in 1916. Van der Heide's grand conception of the Chao Phraya scheme, however, remained and was adopted with some major changes as the Royal Irrigation Department's (RID) central program of work in the early postwar period. Milestones in the evolution of this system include completion in 1969 of the ditch and dike system enabling dry-season rice cultivation and a series of dams: Bhumipol (1957), Chao Phraya (1958), Sirikit (1972), and Naresuan (1979). This complex and the adjoining Mae Klong system are the core of the Thai irrigation system, with a command area covering 40 percent of the country's irrigated area. - 206 - Irrigation is the most important government activity that has influenced the relative quantities of various agricultural goods produced. Since the end of World War II, the Thai government has regularly spent large sums on it. Every year, RID has had the largest budgets among the various departments of the Ministry of Agriculture and Cooperatives. The cumulative capital expenditures of the department total close to $5 billion (1984 prices). Certain peculiarities separate irrigation from other agricultural inputs. At the level of the economy, water can be regarded as a private good, which could in principle be bought and sold, particularly if it is as scarce as it has been in the dry season. Because of system design, however, it is extremely difficult to collect a water charge to enhance the efficiency of water use. This problem arises because at the system and subsystem levels, irrigation may be considered a type of club good: it is normally better utilized collectively by a group than by individual farmers. This is the case only when water is not scarce. If it is, the dependence of a downstream farmer on the behavior of those upstream is well known. Conflicts among various water users without clearly established property rights are then common and require state intervention. With the government the main actor planning and investing in irrigation with public money, there is a distinction between users of irrigation services and the taxpayers who bear the costs. Balancing the desire to have an efficient irrigation system and the means to pay for it is full of political pitfalls and is complicated because irrigation structures are durable capital goods that are costly to convert to other uses. In Thailand, this problem is further compounded by distribution structures' specificity for rice production. Thai irrigation expenditure patterns reveal TABLE 5. INVESTMENT COST OF IRRIGATION that projects that cost less in PROJECTS BY REGION, 195645 real terms to develop have INVESTMENT been favored; newer COST COMMAND COST projects, however, have NO. OF (MILLION AREA PER RAI proven much more costly REGION PERIOD PROJECTS BAIHT ('000 RAI) (BAHT) (table 5). Costs have been Nonh 1956-65 9 4,481 1,927 2,326 increasingprimarilybecause 1966-75 34 7,064 1,327 5,326 the recent need to increase 1976-85 94 9,546 1,440 6,628 farmer incomes in more Nortieast 1956-65 31 395 511 772 remote areas has led to 1966-75 21 1,291 176 7,323 investments in smaller-scale 1976:85 47 12,105 1,367 8,852 projects. Table 6 provides Central 1956-65 76 3,453 5,580 619 summary statistics as of 1966-75 36 2,516 1,795 1,401 1988. The command area, 1976-85 67 5,904 1,959 3,015 constituting about 25 South 1956-65 11 27 131 205 million rai (4 million ha), 1966-75 12 681 282 2,418 accounts for about 20 1976-85 44 6,248 1,448 4,316 percent of arable land, and, Total 1956-65 127 8,356 8,150 1,025 assuming that the over- 1966-75 103 11,552 3,580 3,227 whelming proportion is 1976-85 252 33,803 6,214 5,440 used to grow rice, for about 40 percent of the area under , , rice. Table 6 provides four classifications of irrigation Source: Siamwalla and Na Ranong 1991. projects. For large-scale projects, the construction budget exceeds 200 million baht, and projects may take many years to complete; medium-scale projects require a budget of 2 million to 200 million baht and two to five years' construction. Small-scale projects - 207 - TABLE 6. SUMMARY IRRIGATION STATISTICS, 1988 PROJECT NO. OF CAPACllY COMMAND AREA TYPE PROJECTS (MILLION MI) (RAD Large- and medl um-scale Northerm 127 630.07 (22,463.36) 2,981,162 (676,450) Northeast 239 3,563.22 (0.57) 2,412,593 (15,300) Central Plains 162 1,913.10 (41.68) 11,985,835 (61,250) Southern 76 63.88 (0.42) 1,714,615 (22,400) Total 604 6,170.27 (22,506.03) 19,094,205 (775,400) Small-scale Northern 1174 159.54 2,279,221 Northeast 2582 575.93 1,468,988 Central Plains 744 107.81 1,335,527 Southern 488 7.05 799,190 Total 4988 850.33 5,882,926 Royal development projects (Office of the Coordination Commiuee for Royal Development Projects) Northern 33 49.62 94,570 Northeast 17 25.47 28,320 Central Plains 40 19.72 26,434 Southern 18 0.86 56,663 Total 108 95.67 205,987 Self-defens wUage projects Northern 20 5;57 20,900 Northeast 43 12.05 17,700 Central Plains 10 15.83 8,228 Southern 6 0.00 200 Total 79 33.45 47,028 Note: Figures in parentheses represent projects entrusted to other organizations for operation and mnaintenance. Source: Royal Irrigation Department, 1989 irrigation project statistics. cost less than 2 million baht and take only one year or less to construct. Royal development projects, are initiated as a result of the king's visit and suggestions. Self-defense village projects are minor. Returns on Irrigation Investments There have been different estimates of the efficiency of investments in the vast irrigation system in the Central Plains region. Small (1973) estimated the real rate of return in the Chao Phraya system to be about 5 to 7 percent a year, when dry-season rice cultivation was still marginal. Trung (1980) later reestimated the rate of return and found it to be 10 percent. The efforts expended on tank irrigation systems yielded a return of 10 to 16 percent (Tubpun et al. 1982), but the return was high largely because of the benefits from drinking water and fisheries: the return on agricultural benefits alone was -8 to -7 percent. Because returns on investments in groundwater range from about 12 percent to over 200 percent (World Bank data), there are serious questions about the efficiency of surface water irrigation. An exercise carried out at the Thailand Development Research Institute estimated the impact of irrigation on agricultural production between 1961 and 1987 by using a nested supply function. The first nested model examined the supply of agricultural output as a whole, given various fixed and quasi-fixed factors, of which irrigation was one. Irrigation was measured by the cumulated expenditures on irrigation in each province. The impact of irrigation was relatively small and its significance somewhat ephemeral. - 208 - Only by introducing a multiplicative term with research could a reasonable estimate be obtained. In a growth-accounting exercise based on this econometric supply function, the impact of both variables combined added 0.59 percent to the growth rate of agricultural productivity per worker in Thailand (Patamasiriwat and Suewattana 1990; Siamwalla et al. 1990). Only in the second nested model, which explained crop composition, could the impact of irrigation be seen clearly: its key role was in shifting the crop composition toward rice (table 7). For the important Central Plains region, the introduction of irrigation led to an increase in the share of rice by almost 4 percentage points, much of which was at the expense of upland crops. The rice share of another major rice-supplying region, the Lower North, increased by 4.3 percentage points. Thus, past irrigation expenditures have tended to reinforce specialization in rice. TABLE 7. PERCENTAGE CHANGE IN SHARE Irrigation investments, therefore, are a poor OF RICE IN TOTAL CROP VALUE CAUSED BY A 1 PERCENT INCREASE IN CUMULATED vehicle for inducing diversification, particularly IRRIGATION EXPENDITURES, 1961-87 toward upland crops. Some planners, however, have indicated that irrigated areas have more ABSOLUTE CHANGEDUE flexibility to move to high-value horticultural and REGION SHARE IN TO aquacultural activities. The data used to estimate the results in table 7 do not permit a verdict on Upper North 44.5 0.0 this issue, as they vastly underestimate the Lower North 58.2 4.3 Central Plains 73.6 3.9 contribution of horticulture and ignore aquaculture East 55.8 -5.9 altogether. West 39.8 -3.6 Northeast 65.3 0.7 Government Responsiveness South 25.8 1.0 Figure 1 shows the movement of real Source: Estimates from Thailand Development Research gover-nment investments in irrigation over time. Institute, Agriculture and Rural Development Program. The two peaks in the figure may reflect the heavy investments in the Chao Phraya system and the Mae Klong and Pitsanulok (Naresuan) systems, or they may reflect the price peaks of the mid-1960s and the early 1970s, indicating government responsiveness to price movements. An unpublished econometric analysis by Rosegrant and Mongkolsmai of government irrigation investments ("Impacts of Government Investment on Irrigation in Thailand") indicates that the government does respond significantly to movements in world prices in its budgeting of irrigation investments and is inhibited by the increasing costs of irrigation. These costs, however, have been partially compensated by the growth in GNP, which presumably means that the government can afford an expenditure that is often regarded as a means to help the poor. Current Policy Issues Because of macroeconomic factors and low rice prices that led to belt tightening in the early 1980s, no major irrigation projects have been initiated since the completion of the Naresuan Dam. Another factor underlying this hiatus is that the Naresuan project (financed by the World Bank) came as close to a zero-return investment as a project can. The Chao Phraya system could not be expanded without additional storage capacity. By 1977, the Bhumipol and Sirikit main storage dams were having trouble meeting water demand for the dry-season rice crop. During that year, the government decided to invest in the Naresuan Dam. The dry-season water for this project area could only be obtained by diverting water from other parts of the Chao Phraya system. Because much of the return from the irrigation system is from the dry-season crop, what the Naresuan command area has gained, the other parts of the system lost, with a near-zero net result for the entire system. Figure 1. Irrigation Investment, 1962-85 1.6 - 1.4- 1.2- 0 1.0 0.8X 0.6- 0.4 - II I I I I I I I I I I I I I I I I I I 1962 1965 1968 1971 1974 1977 1980 1983 Year - 210 - Despite the clear excess demand (admittedly at zero price) for water during the dry season, the government has been reluctant to invest in new storage dams: reasons include increasing public opinion against such dams because of potential adverse environmental impacts and opposition from inhabitants who may be displaced by dam sites. (The government has used water rationing as a tool to limit rice production when the rice price is low.) As a substitute for major new investments, the government has increasingly concentrated on small-scale projects. A "hard-headed" calculation by the Bureau of Budget indicates considerable waste in these small projects. An alternative view holds that government agencies in charge of small-scale projects have tended to ignore the need to involve local people. Because these projects are ultimately dependent on local management skills, their failure rate is not surprising (Lawson 1984). Given the difficulties in engaging in new construction, there should have been an interest in reforming the management of the existing system. A legislative attempt was made to collect water charges (on a fixed per-acre basis), but the bill was withdrawn (by the Ministry of Agriculture and Cooperatives) after a brief show at the National Assembly. It is doubtful, however, whether a water charge would improve poor management. The biggest drawback of the system is that it does not reduce farmer uncertainty as to water supplies, surely the raison d'etre of an irrigation system. Dry-season water is rotated each year in much of the command area; the rotation system, however, is not yet well defined, and the government often reneges on its promises. Also, the system does not allow even partial control over water levels at individual farms. Thus, in areas with second-cropping, in addition to the inability of farmers to choose what crops they can grow, they have little choice regarding what variety of rice they can grow. This situation has created severe labor constraints. High rice prices during the past two years have led farmers to invest substantially in tubewells-a delayed response to major public investments in rural electrification in the late 1970s and early 1980s. This development is promising because it will allow farmers greater independence in their cropping decisions. Agricultural Research The Thai government began engaging in agricultural research after World War 11 to enhance returns to rice farming, although prewar governments selected and propagated rice seeds with good eating quality. Kasetsart University, the main agricultural university, was established in 1942. Later, primarily in the 1960s and early 1970s, regional universities with sizable agriculture faculties were established in the north (Chiang Mai), northeast (Khon Kaen), and south (Songkhla and Pattani). Until the 1970s, public research was fairly low-key. There was relatively little pressure to increase crop yields because of Thailand's relatively abundant supply of land. Crop research is the domain of the Department of Agriculture (DOA). Research on livestock, fisheries, and forestry is conducted by the departments responsible for each of these subsectors. Other institutions engaged in crop and other research, most notably universities, together annually spend one- quarter of the amount spent by the DOA. The DOA and the universities together represent more than 95 percent of the total research expenditure for crops. Crop research under the DOA constitutes the major portion of research expenditures because crop production is the largest component of agriculture, although when compared with its value added, crop research receives proportionately less attention than research in livestock, fisheries, and forestry. The Thai government's allocation of research expenditures is biased toward importables and against exportables and nontradable goods. (Ihis and subsequent paragraphs of this section draw on Setboonsarng et al. 1990.) Although there is some rigidity in the crop-wise allocation over time, for the reasons described previously, there is also some movement: an econometric analysis of the determinants - 211 - of research expenditure allocations revealed that the main explanatory factor for the share of research going to each major crop is the amount of land devoted to its cultivation. Although the Thai government has never used its research funds to foster diversification, its bias toward import-competing crops tends to support diversity in cropping patterns, in the sense that such crops, because they have to be supplemented by imports, tend to be underproduced. However, the tendency of the research funding (at least among the major export crops) to follow cropping areas implies that it has followed rather than led farmer choices that generally reflect market signals. This conclusion also finds support from an examination of the procedures followed by the DOA in allocating its research funds. In evaluating research proposals for funding, the department follows guidelines heavily weighted by expected economic benefits, although it has at times somewhat idiosyncratic criteria for what constitutes relative "economic benefits" among crops, e.g., share of export values or relative import values (Setboonsarng and Khaobarisuthi 1990). Agricultural Extension The Department of Agricultural Extension (DOAE), established in 1967, underwent a major change in 1977, when the government decided to adopt the training and visit (T&V) approach recommended by the World Bank. This led to a rapid rise in the number of the department's employees, reaching 13,292 in 1988. The Thai government now expends greater resources in its agricultural extension network than in its research by approximately 1.7 to 1. The DOAE is the primary mechanism by which the government introduces new technology to farmers. In addition to conventional extension work using T&V, the DOAE also produces significant amounts of open-pollinated seeds for most major crops. In response to low agriculture prices in the early and mid-1980s, the government strongly pushed the DOAE to promote agricultural diversification. DOAE's agricultural plan under the 1982 national development plan proposed five strategies: 1. reducing rice planting in low-yielding areas, particularly the northeast; 2. promoting substitutes for rice, sugarcane, tobacco, and cassava; 3. limiting production areas for crops for which demand is likely to be saturated, while raising productivity within existing areas; 4. replacing rubber trees with modern high-yielding varieties; 5. promoting production of nontraditional crops. To achieve these objectives, the so-called Four Cooperation Program was initiated by the Ministry of Agriculture and Cooperatives in 1986 to foster crop diversification and stimulate technology transfer. The scheme brings together four sets of actors: the government, private firms, financial institutions, and farmers. Government agencies and private firms technologically assist participating farmers and assure them of minimum prices or guaranteed incomes. Farmers are then trained by the DOAE to plant the promoted crops, adopt the technology, and use the inputs suggested by the firm. The government will ensure that input prices are reasonable, i.e., not higher than market prices. To induce farmers to participate, the scheme subsidizes interest costs by 3.5 percent, with participating farmers paying 8.5 percent instead of the 12 percent normally charged by BAAC. The scheme was allocated 250 million baht - 212 - (about $10 million) from the Farmers Aid Fund (which obtains its revenue by levying export taxes on agricultural goods) and used the earned interest to finance the interest subsidies. The program has attractive features, including guarantees from private firms to ensure that participating farmers are not left in the lurch by market failures. Because the new technology may require cash inputs and be risky, the scheme provides credit or income guarantees. The program, however, is not likely to have much impact on cropping choices. Theoretically, the most effective subsidy should work only for marginal farmers who would be induced by the subsidy to switch to a crop preferred by the government. The program depends on firms to select farmers, however, and they are not likely to choose marginal farmers. On the contrary, they would tend to pick areas and farmers who already have comparative advantage in growing the promoted crop. The subsidy merely adds to the rent of farmers who already have this comparative advantage. Therefore, this scheme cannot be expected to have a major impact on the cropping pattern. Furthermore, the conditions set by the government may be too restrictive and unrealistic. For example, a firm must set up buying points all over the promoted areas. This requirement ignores the possible (and indeed likely) advantage of existing local middlemen. In sum, the only tool that the Thai government has deployed specifically to induce diversification is its extension system. There are some resources via the credit system to change the relative profitability of various crops beyond what the market dictates, but the amount of subsidy involved is minor. Conclusion The Thai government influences crop diversification through price and marketing policies (primarily exercised through border measures), credit, irrigation, research, and extension. The Sixth (1986-91) Five-Year National Economic and Social Development Plan specified diversification as a guiding concept. Concurrently, world market conditions for most Thai agricultural products have declined, creating additional pressure to diversify Thai agriculture away from not only rice, but all the major food crops, particularly cassava, maize, and sugarcane. The key factor driving the Thai nonrice agricultural sector and the government's role in it is the world market. Particularly since the early 1980s, but even before that for all crops other than rice, the agricultural sector has been fully exposed to world markets. When these become adverse, not only must traders and farmers respond, but the government itself is subject to strong pressures to adjust. At this point conflicting objectives come into play. On one hand, the government is under pressure to provide support as a safety net to farmer incomes. This, given the strong price responsiveness of Thai farmers, implies an antidiversification move. On the other hand, in its supply of support services-for example, credit, irrigation, research, and extension-the government has the tools to shift the production structure away from those goods adversely affected. The question becomes: Is the government structure itself free enough to induce shifts? This paper's answer is essentially no. In some cases (research, extension, and credit) procedural rigidity impedes the needed changes. Irrigation is a capital expenditure and thus can be and actually is much more flexible than research and extension, which, from an accounting perspective, are regarded as recurrent expenditures. But with capital expenditures, the problem is physical: once the investment has been made, the structures are fixed and cannot be easily changed to produce other commodities. Throughout the postwar period, irrigation has been almost exclusively concerned with rice, and the scheduling of water supplies is still entirely dictated by the rhythm of rice production. The severe battering Thai farmers took as a result of the fall in world prices in the early and mid- 1980s has raised concerns that a diversified cropping pattern should be the standard practice for each farmer. Within the government this concern is felt only in the Department of Agriculture (the agency in charge of research), and then only within the Farming Systems Institute, which receives only about one- tenth of the total research allocation. BAAC claims that its lending policies avoid monocropped farms, - 213 - but on closer examination, the only group that comes under this definition seems to be rainfed-rice farmers. They are discriminated against because relative to their incomes, they market less of their products and therefore are not as 'bankable." Could the government, although itself rigid, encourage a diversified cropping pattern? For most annual crops, farmers have tended to diversify only to cope with environmental risks, not with market fluctuations: most farmers produce annual crops and therefore already have the needed flexibility, at least over the medium term (more than one crop cycle). Tree crop growers also do not diversify, and among fruit growers there is a strong trend toward specialization at the farm level, as fruit production has become somewhat more "professionalized' and less of a sideline activity. Thai farmers still look to the markets primarily for their signals to produce or shift away from a crop. The choice for the government, given the unfavorable prospects for agriculture in the long term, is whether to provide a safety net for farmers growing traditional crops. The alternative, to push for intensification by investing heavily in research and to promote modem varietal technology, may be appropriate in the near term (5 to 10 years). However, given the rapid growth of the nonagricultural economy during the late 1980s-growth expected to continue into the 1990s-the absolute size of the labor force in agriculture will probably begin to shrink by the end of the decade. The relevancy of the varietal technology that Thailand lacks will then be a major question. Note 1. Prior to the arrival of van der Heide, a canal system was built in the late 1880s by a private company (owned by members of the Thai elite) in what is now the Rangsit area in exchange for a land grant from the king. This cannot be regarded as an irrigation project, as there was no attempt to control the water supply that would be deliverable to the farms. The main benefit of this project was to open a large new area for cultivation by having access to water transport provided by the canals (Asawai 1987). The only part of van der Heide's project to have been implemented immediately improved the water conditions of the Rangsit canal system. References Asawai, Soontaree. 1987. The History of Rangsit Canal: Land Development and Its Social Impact. Bangkok: Thammasat University Press (in Thai). Brown, Ian. 1988. The Elite and the Economy in Siam, c. 1980-1920. Singapore: Oxford University Press. Feeny, David. 1982. The PoliticalEconomy ofProductivity: lhaiAgricultural Development, 1880-1975. Vancouver: University of British Columbia Press. Ingram, James C. 1971. Economic Change in Thailand, 1850-1970. Stanford, Calif.: Stanford University Press. Jessadachatr, Phitsanes. 1977. 'A History of Sugar Policies in Thailand 1937-75." Master's thesis, Faculty of Economics, Thammasat University, Bangkok. Lawson, Evan. 1984. "Assisting Villages to Build Their Own Irrigation Scheme-An Example from Thailand." Washington, D.C.: Irrigation Management Network, Overseas Development Council. Patamasiriwat, Direk, and Sakeddao Suewattana. 1990. 'Sources of Growth of Agriculture Production, 1965-1985: Analysis based on a TDRI Model." Warasarn Settasat Thammasat 8(1): 43-69 (in Thai). - 214 - Setboonsarng, Suthad, et al. 1990. 'Objectives of Public Agricultural Research in Thailand.' Research Priorities for Agriculture in Thailand Project Discussion Paper. Thailand Development Research Institute, Thai Department of Agriculture, and Australian Centre for International Agricultural Research. . 1991. Research Prioritiesfor Agriculture in Thailand. Research report submitted to the Australian Centre for International Agricultural Research (ACIAR). Setboonsarng, Suthad, and Chareon Khaobarisuthi. 1990. 'Public Research Resource Allocation: Procedure and Outcome." Research Priorities for Agriculture in Thailand Project Discussion Paper. Thailand Development Research Institute, Thai Department of Agriculture, and Australian Centre for International Agricultural Research. Siamwalla, Ammar. 1975. "A History of Rice Policies in Thailand." Food Research Institute Studies 14(3). Siamwalla, Ammar, et al. 1990. 'The Thai Rural Credit System: Public Subsidies, Private Information, and Segmented Markets." World Bank Economic Review 4(3). Siamwalla, Ammar, and Viroj Na Ranong. 1991. A Fact Book on Rice. Bangkok: Thailand Development Research Institute. Siamwalla, Ammar, and Prayong Nettayarak. 1990. "Estimating the Cost of Subsidies for Agricultural Credit." Mimeo. Bangkok: Development Research Institute. Siamwalla, Ammar, and Suthad Setboonsarng, 1989. Trade, Exchange Rate, and Agricultural Pricing Policies in Thailand. World Bank Comparative Studies on the Political Economy of Agricultural Pricing Policy. Washington, D.C. Siamwalla, Ammar, Suthad Setboonsarng, and Direk Patamasiriwat. 1990. "Thai Agriculture: Resources, Institutions and Policies." Mimeo. Bangkok: Thailand Development Research Institute. Small, Leslie E. 1973. "Historical Development of the Greater Chao Phraya Water Control Project: An Economic Perspective." Journal of Siam Society volume 61, part 1: 1-24. Trung, Ngo Quoc. 1980. "An Economic Analysis of Irrigation Infrastructure Development in Deltaic Regions of Asia: The Case of Central Thailand." Discussion Paper 79-06. Quezon City, the Philippines: Council for Asian Manpower Studies. Tubpun, Yuavares, et al. 1982. "Tank Irrigation in Northeastern Thailand: The Returns and Their Distribution.' Economics Report 82-6. Minneapolis, Minn.: University of Minnesota, Department of Agricultural and Applied Economics. Usher, Dan. 1967. "The Thai Rice Trade." In T.H. Silcock, Thailand: Social and Economic Studies in Development. Canberra, Australia: Australian National University Press. Distributors of World Bank Publications ARGENnNA EL SALVADOR JAPAN SOUTH AFRICA, BOTSWANA Caro Hirsch, SRL Fusada Easrn Book Sevice Feor singl Utic Galira Cuomn Alam Dr. Manuel Enrique Aimujo #3430 Hongo 3-Ciome, Bunkyo.m 113 Oxford UniversttyPI Florida 165,4th FloorOfk W/465 Edificio SISA, lr. Pilo Tokyo Southen Africa 1333 Buenos AJius San Salvador 01l PO Box 1141 KENYA Cape Town 800 AUSTRALA, PAPUA NEW GUINEA. FINLAND Afrlca BookService (LA.) A.dT Fiji, SOLOMON ISLANDS, Aateeuinn IKierlkuFpp Quaran House, Mfargano Stivot ForsuMcr**omonderm: VANUATU, AND WESTERN SAMOA PO. Box 128 P.O. Box 45245 Inteational Subesripon Service D.A. Books I Joruals S40101 Hesnid 10 NaIrobi P.O. Box 41095 648 Wiutehoe Roud Cigh Mitrham 3132 FRANCE KOREA, REPUBUC OF joharousburg2024 Victona World Bank Publlmlion Pan Kore Book Coepon 66, avenue dl6s P.O. Box 101, Kwangwhamun SPAIN AUSTRIA 75116 Paris Seoul Mundi-Preln.Ibroa,S.A. Cerold and Co Cstello37 Graben 31 GERMANY MALAYSIA 2801 Madrid A-1011 Wien UNO-VerHag Univerasity of Malya CoPetve Poppesdorer All. 55 Bookihop, Limited Librert Internadonal AEDO5 BANGLADESH D5300Bonn P.O. Box 1127,J Jlan Panta iBu Conel de Cent, 391 Mcro Indu Developmnt 59700 Kuala Lurnpur 009 Barelona Assistance Society (MIDAS) GUATEMALA House 5, Road 16 Librar Piedra Santa MEXICO SRI LANKA AND THE MALDIVES Dhanmozndi R/Ae Se. Call 7-55 INFOTEC Lake House Boobahop Dhaka 1209 Zone I Aparlado Pasl 21-80 P.O. Box 244 Guatemal City 1406 TMapan. M DOF. 1005, Sir Cithmpalam A. ammA ame9s: GardinerMaWath1k 16, Nur Ahbmed Sark HONG KONG, MACAO NETHERLANDS Colontbo2 Chitaong 4000 Asa 200 Ltd. De Undeboom/InOr-Publikade 448 Wyndham Sthe P.O. Box 20 SWEDEN 76, KXDJ Avenue Winning Centr 7480 AE Haksbegn For single tiflI Kuln 9100 2nd Foor Fritzes Fackboksforetsget Central Hong Kong NEWZEALAND Rgeprisgtan 12, Bx 16356 BELGIUM Hills LIbrary end Inomation Sevie S-103 27 Stockholm lean De Lannoy INDiA Private BSag Av. du Rod 2M Allied Publishei Private Ltd. New market ForsnsWhoft on dm. ID60 Bruses 751 Mount Road Auckland Wennerl e-Williant AS Madrs - 600 002 Box 30DO4 CANADA NIGERIA S l04 25 Stockholm Le Diffuseur EancA rin Univerity Prm Limited CP. 85, ISIB rue Ampbr 15 J1N. Herdia Mrg Thee Crowns Building Jricho SWITZERLAND Boucherville Qulbe Balard Este Private MAll ag 5095 Fer5 1titIaW J4B SE6 Bombay - 400 ibadan Libraine Payot 1, rue de Bourg CHINA 13/14 Asf All Road NORWAY CH 1002 LAusan China Financial & EconomIc New Delhi -110 00 Narwvesen irmtion Center PublishingHouse Book Deparntmt Persufcr*e onkio: 8. Da Fo Si Dong Ji 17 Chiit jte n Avenue P.O. Box 6125 Etuterd Librie Payot Beiing Calcutt - 700 072 N 0602 Oslo 6 Service des Abonnements CaG poettle 3312 COLOMBIA Joyadev Hol Buiding PAKISTAN CH1002 LAusanne Infoenlace Lida. 5th Main Road Gandlinagpr Minx Book Agency Aparlado Aeres 70 B ngalore-S60 009 65, Sh&hrah-e