ENVIRONMENTALLY AND SOCIALLY SUSTAINABLE f1j\ .^ DEVELOPMENT STUDIES AND MONOGRAPHS SERIES 14 Work in progreas for public discussion S ' 1lqihh Rural Finance Lsies, De.', (n1 Best PriYÍ(/s - .7 Ii _Dn1( ?.1 *en/(/Ñ9 Ld Pie ESSD Proceedings Series 1 Culture and Development in Africa: Proceedings of an International Conference (Also in French) 2 Valuing the Environment: Proceedings of the First Annual International Conference on Environmentally Sustainable Development 3 Overcoming Global Hunger: Proceedings of a Conference on Actions to Reduce Hunger Worldwide 4 Traditional Knowledge and Sustainable Development: Proceedings of a Conference 5 The Human Face of the Urban Environment: A Report to the Development Community 6 The Human Face of the Urban Environment: Proceedings of the Second Annual World Bank Conference on Environmentally Sustainable Development 7 The Business of Sustainable Cities: Public-Private Partnershipsfor Creative Technical and Institutional Solutions 8 Enabling Sustainable Community Development 9 Sustainable Financing Mechanismsfor Coral Reef Conservation: Proceedings of a Workshop 10 Effective Financing of Environmentally Sustainable Development: Proceedings of the Third Annual World Bank Conference on Environmentally Sustainable Development 11 Servicing Innovative Financing of Environmentally Sustainable Development 12 Ethics and Spiritual Values: Promotíng Environmentally Sustainable Development 13 The Self and the Other: Sustainability and Self-Empowerment 14 Meeting the Challenges of Population, Environment, and Resources: The Costs of Inaction 15 Rural Well-being: From Vision to Action ESSD Studies and Monographs (formerly Occasional Paper) Series 1 The Contribution of People's Participation: Evidencefrom 121 Rural Water Supply Projects 2 Making Development Sustainable: From Concepts to Action 3 Sociology, Anthropology, and Development: An Annotated Bibliography of World Bank Publications 1975-1993 4 The World Bank's Strategyfor Reducing Poverty and Hunger: A Report to the Development Community 5 Sustainability and the Wealth of Nations: First Steps in an Ongoing Journey 6 Social Organization and Development Anthropology: The 1995 Malinowski Award Lecture 7 Confronting Crisis: A Summary of Household Responses to Poverty and Vulnerability in Four Poor Urban Communities (Also in French and Spanish) 8 Confronting Crisis: A Comparative Study of Household Responses to Poverty and Vulnerability in Four Poor Urban Communities 9 Guidelinesfor Integrated Coastal Zone Management (continued on the inside back cover) Copyright © 1997 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 September 1997 This report has been prepared by the staff of the World Bank. The judgments expressed do not necessarily reflect the views of the Board of Executive Directors or of the governments they represent. Cover photograph: Thai fariner. From Bank for Agriculture and Agricultural Cooperatives, "Annual Report 1989," Bangkok. Jacob Yaron is rural finance advisor in the Rural Development Department of the World Bank. McDonald P. Benjamin, Jr., is a financial economist in the East Asia and Pacific Region, Finance and Private Sector Development Unit, of the World Bank. Gerda L. Piprek is a consultant to the World Bank. Library of Congress Cataloging-in-Publication Data Yaron, Jacob. Rural finance: issues, design, and best practices / Jacob Yaron, McDonald P. Benjamin, Jr., Gerda L. Piprek. p. cm. - (Environimentally and socially sustainable development studies and monographs series ; no. 14) Includes bibliographical references. ISBN 0-8213-3737-8 1. Rural credit. 2. Agricultural credit. I. Benjamin, McDonald P., Jr. II. Piprek, Gerda L., 1962- . III. Title. IV. Series. HG2041.Y37 1996 332.7'1-dc2O 96-34492 CIP ®3 The text and the cover are printed on recycled paper, with a flood aqueous coating on the cover. ENVIRONMENTALLY AND SOCIALLY SUSTAINABLE DEVELOPMENT STUDIES AND MONOGRAPHS SERIES 14 Rural Development Rural Finance Issues, Design, and Best Practices Jacob Yaron McDonald]B Benjamin, Jr Gerda L. Piprek The World Bank Washington, D. C. Contents Foreword ix Acknowledgments x Abstract xi Abbreviations xii Executive Summary 1 Introduction 11 PART ONE AN OVERVIEW OF THE TRADITIONAL APPROACH 15 Chapter 1 The Challenge of Rural Financial Intermediation 17 Obstacles to Financial Intermediation 17 Urban-Biased Policies 17 Systematic Weaknesses of Rural Financial Markets 18 Poorly Designed Interventions 18 Conclusion 19 Chapter 2 The Traditional Approach to Rural Finance 20 The Traditional Perspective on Rural Finance 20 Assessing the Traditional Approach 22 Conclusion 28 PART Two POLICIES FOR IMPROVING RURAL FINANCIAL INTERMEDIATION 29 Chapter 3 New Perspectives on Rural Finance 31 Policy Goal I: Expand Rural Income 31 Policy Goal HI: Reduce Rural Poverty 41 Conclusion 45 liii iv Rural Finance Chapter 4 Creating a Favorable Environment 47 - Establishing a Favorable Macroeconomic Environment 47 Removing Policy Biases against Agricultural and Rural Sectors 48 Establishing Integrated and Efficient Financial Markets 50 Conclusion 53 Chapter 5 The Legal and Regulatory Framework for Rural Financial Markets 54 Problems That Limit the Use of Collateral 54 Problems That Limit Unsecured Lending 59 Problems That Limit All Lending 60 Solution Options 61 Benefits and Costs of Reforms 62 Conclusion 63 Chapter 6 Direct Interventions 65 Targeted Interventions 65 Guiding Principles for Public Interventions 66 Credit Guarantee Schemes 78 Crop Insurance 80 Conclusion 82 PART THREE AN OVERVIEW OF RURAL FINANCIAL INSTITUTIONS 85 Chapter 7 Performance Criteria for Rural Financial Intermediation 87 Methodological Problems in Assessing the Impact of Rural Credit Projects 88 Assessing the Performance of Rural Financial Institutions 90 Financial Rate of Return and the Rate of Return on Investors' Equity 97 Conclusion 98 Chapter 8 Building Institutional Capacity 99 Definition of Institutional Capacity 100 Problems Experienced by RFIs 100 Key Elements of Capacity Building 101 Information and Incentives 116 Conclusion 116 Chapter 9 Best Practices: Three Success Stories 117 Summary of Main Features 117 Overview of the Institutions 118 Performance Assessment 120 Modes of Operation 123 Potential Improvements 128 The Environment 129 Conclusion 130 Annex. Summary of Performance and Operating Methods of BRI-UD, BAAC, and GB 132 Appendix The Subsidy Dependence Index of Rural Financial Institutions 139 Notes 143 References 149 Contents v Boxes 2.1 Merits of long-term finance 21 2.2 Assessing the performance of agricultural credit projects 23 2.3 Credit and income redistribution in Costa Rica 23 2.4 Impact of formal finance on the rural economy of India 24 2.5 Impact of wheat subsidies in Saudi Arabia 25 2.6 Fiscal burden of rural finance in Mexico 26 2.7 Focus of the intervention: rural versus agricultural development 27 3.1 Consultative Group to Assist the Poorest 36 3.2 Sustainable banking with the poor 37 3.3 How information externalities can generate market failures 38 3.4 Two obstacles to financial transactions: moral hazard and adverse selection 38 3.5 The problem of social barriers to trade 39 3.6 Two perspectives on market failure: the argument for intervention 40 3.7 Credit beneficiaries below the poverty line 43 3.8 ls credit the solution to poverty? 44 3.9 Targeting the poor: infrastructure 45 4.1 Eight pillars of urban-biased policies 49 4.2 Mitigating the risk of banking crises 51 4.3 Two good reasons for market-determined interest rates: equity and efficiency 52 5.1 Titling versus registration 56 5.2 High-tech land titling 56 5.3 Creation of a security interest: gaps in coverage 57 5.4 Enforcement: canaries, jail, and death 58 5.5 Warehouse receipts: an instrument that facilitates trade and inventory financing 58 5.6 Nonbank lenders 59 5.7 Fragmentary reform 60 5.8 Land titling and registration: a private initiative in Peru 64 6.1 The importance of empirical evidence in justifying and designing intervention 67 6.2 Rural finance in Central and Eastern Europe 70 6.3 Village credit funds in Albania 72 6.4 Financial systems that work for the majority 72 6.5 The World Bank and nongovernmental organizations 73 6.6 Rehabilitation of credit unions in Guatemala 74 6.7 Piloting savings and loan associations in Madagascar 75 6.8 Proposed pilot rural financial project in Croatia 76 6.9 Guaranteeing noncommercial risks: a viable approach for rural finance in Russia? 78 6.10 Moldova pre-export guarantee 79 6.11 A primer on credit guarantees 79 6.12 Liquidity options for diversifying risk in rural financial markets 81 6.13 Effective public crop insurance: the example of sugar in Mauritius 83 6.14 A drought insurance proposal 84 7.1 Merits of computing the subsidy dependence index for a rural financial institution in Jamaica 93 7.2 Calculating the subsidy dependence index 94 7.3 Cost comparisons in public spending 95 7.4 Performance of credit line operations: collection experience of on-lending institutions 96 vi Rural Finance 7.5 Definitions 97 7.6 Illustration of the financial rate of return and the investor's return on equity 98 8.1 The need for institutional reform 99 8.2 Potential impact of interest rate subsidies 100 8.3 Subsidized interest rates: good intentions, disappointing results 101 8.4 Rural credit institutions as a political tool: debt forgiveness in India 102 8.5 When does informal become formal? South Africa 103 8.6 Rural savings and loan associations in Benin 104 8.7 Mysore Resettlement and Development Agency 105 8.8 The formal meeting the informal in South Africa 106 8.9 Village bank methodology 107 8.10 Groups in rural financial intermediation 108 8.11 Technology in servicing the poor: introducing the SmartCard in Swaziland 109 8.12 Effective management information systems 110 8.13 Impact of failure to write off bad debt 111 8.14 Quality of assets: a healthy loan portfolio 112 8.15 Cooperative Development Foundation: strengthening women's thrift groups in India 113 8.16 Risk and its consequences 115 Figures 3.1 Decision tree for expanding rural income 34 3.2 Financial depth in 1960 and GDP growth per capita from 1960 to 1989 35 3.3 Decision tree for reducing poverty 41 4.1 Foreign exchange premiums and economic rates of return on bank-supported agricultural products 48 5.1 Impact on credit markets of well-designed legal and regulatory reform 62 7.1 Criteria for assessing the performance of rural financial institutions 90 7.2 Optimizing performance 91 7.3 Loan recovery profile for short-term loans of the Bank for Agriculture and Agricultural Cooperatives, 1981-86 97 9.1 BAAC average annual total loans and total deposits, 1987-93 118 9.2 BRI-UD average annual total loans and total deposits, 1987-94 119 9.3 GB average annual total loans and total deposits, 1987-94 120 9.4 Return on equity for BAAC, BRI-UD, and GB, 1987-94 121 9.5 Return on assets for BAAC, BRI-UD, and GB, 1987-94 121 9.6 Sectoral distribution of BRI-UD loans, 1993 123 9.7 Sectoral distribution of GB loans since inception 123 9.8 Personnel costs as a percentage of average annual total loans and deposits for BAAC, BRI-UD, and GB, 1987-94 123 9.9 Average annual value of total deposits as a percentage of total loans for BAAC, BRI-UD, and GB, 1987-94 127 9.10 Financial costs as a percentage of average annual total loans for BAAC, BRI-UD, and GB, 1987-94 128 Contents vii Tables 3.1 Characteristics of old and new approaches to rural finance 32 5.1 Explaining high interest rates for loans in Argentina 63 6.1 Variables and policy options for direct interventions 68 7.1 Past-due age analysis of short-term working capital of Bank for Agriculture and Agricultural Cooperatives, 1981-86 96 9.1 Operating costs 118 9.2 Distribution of BAAC loans, selected years 125 9.3 Loan terms and performance 126 9.4 Repayment performance of BAAC clients, selected years 126 9.5 Nominal and real average arnual GB interest rates on loans, 1989-94 129 9.6 Economic and demographic indicators for regions served by the BAAC, the BRI-UD, and the GB 130 Foreword T bhis volume is intended to serve as a sectors, rural development, and the legal and sourcebook to aid decisionmakers as regulatory framework. they seek to create an environment con- Moving from the macroeconomic and ducive to the development of rural financial sectoral to the institutional level, the report of- markets. As a sourcebook, it explores both the fers two primary criteria-outreach and self- traditional and new approaches that govern- sustainability-as the bases for assessing the ments have taken to rural finance. performance of rural financial institutions. It First, the report outlines the traditional reviews and analyzes the modes of operation approach to rural finance, which relied heavily and performance of three successful rural on supply-led, state-owned agricultural credit financial institutions: the Grameen Bank of institutions. Second, it highlights how wide- Bangladesh, the Bank of Agriculture and Ag- spread, urban-biased policies impeded rural ricultural Cooperatives (BAAC) in Thailand, development and the promotion of rural finan- and the Bank Rakyat of Indonesia, Unit Desa cial markets. These policies included overval- (BRI-UD). The analysis illustrates the phe- ued exchange rates, inflexible price controls on nomenal growth and development that these food produce, underinvestment in rural infra- institutions underwent over the past decade. structure, and protection of domestic indus- An understanding of their modes of operation tries against import competition. could be useful in designing interventions in The volume then illustrates the emerging rural financial markets elsewhere. new approach, which focuses on creating an environment to promote viable rural financial markets. Specifically, it focuses on creating a Ismail Serageldin favorable policy environment in terms of Vice President the macroeconomy, agricultural and financial Environmentally and Soc:lly Sustainable Development ix Acknowledgments T his study would not have been possible Anandarup Ray, Richard Roberts, Michel without collegial effort. The authors Simeon, Khalid Siraj, David Steeds, Laura thank Heywood Fleisig and Nuria de la Tuck, Panos Varangis, and J. D. Von Pischke. Pena for writing chapter 5. We also extend We are happy to acknowledge the special our thanks to the many people who reviewed attention and assistance of Muhammad and commented on the paper, particularly Yunus, managing director, Grameen Bank; those who wrote summaries of the case studies, M. Masud Isa, deputy general manager, which were included as boxes in the text: Grameen Bank; M. Shahjahan, deputy general Lynn Bennett, Hans Binswanger, Colin Bruce, manager, Grameen Bank; Don Johnston, Har- Silvia Castro, Rodrigo Chaves, Arie Chupak, vard Institute for International Development, Stephanie Charitonenko Church, William Rural Banking Project in Indonesia; and Pit- Cuddihy, Carlos Cuevas, Sunil Datt, Ramesh tayapol Nattaradol, executive vice president, Deshpande, Lynn Engstrand, Gershon Feder, Bank for Agriculture and Agricultural Coop- Delbert Fitchett, Elena Folkerts-Landau, Vinod eratives. Goel, Michael Goldberg, Syed M. Hashemi, The authors also thank Leo Demesmaker for Peter Hazell, Shahidur Khandker, Daniela editing the paper, and Lisa Barczak, Gunnar Klingebiel, Richard Lacroix, Fred Levy, Mohini Larson, Tetsutaro Muraki, Wendi Maloney, Malhotra, Alexander McCalla, Luciano Mosele, Alison Strong, Donna Allen, Alicia Hetzner, Dan Mozes, Jupiter Ndjeunga, Maria Nowak, and Virginia Hitchcock for their editorial Nwanze Okidegbe, Jo Ann Paulson, Glenn contributions. Gaudencio Dizon desktopped Pederson, Sita Ramaswami, Martin Ravaillon, the volume. x Abstract T he provision of affordable financial serv- ronment. It first outlines the traditional ap- ices to the rural population has been a proach to rural finance, which relied heavily prime componenrt of development strat- on supply-led, state-owned agricultural credit egy for several decades. Governments, devel- institutions. Next, it presents the emerging opment agencies, and other donors have new approach and proposes a framework for supported various rural financial institutions, determining the most appropriate role for with the aim of accelerating growth and reduc- governments, identifying the most effective se- ing poverty. quencing of policy reforms, and selecting the In recent years it has become clear that the most suitable forms of intervention. On the in- promotion of rural financial markets should stitutional level the report presents a frame- be done in the overall context of rural devel- work for assessing the performance of rural opment and integrated financial markets. This finance institutions and suggests guidelines for approach implies an important, market- the development and effective management of friendly role for governments in creating an these institutions. Finally, the volume reviews environment conducive to the development of and analyzes the performance of three success- rural financial markets. ful rural finance institutions. These three case This volume is offered as a sourcebook for studies provide valuable insights for the de- decisionmakers seeking to create such an envi- sign of future interventions. xi Abbreviations ACB Agricultural Credit Bank of IFAD International Fund for Agricul- Jamaica tural Development BAAC Bank for Agriculture and Agricul- IMF International Monetary Fund tural Cooperatives (Thailand) IREQ internal rate of return on BIS Bank for Intemational Settlements investor's equity BKK Badan Kredit Kecamatan MIS management information systems BRI Bank Rakyat Indonesia NGO nongovernmental organization BRI-UD Bank Rakyat Indonesia- OECD Organisation for Economic Unit Desa Co-operation and Development CGAP Consultative Group to Assist OED Operations Evaluation Depart- the Poorest ment, World Bank DFI development finance institution PSD Private Sector Development EBRD European Bank for Reconstruction Department, World Bank and Development RFI rural financial institution ERR economic rate of return RFM rural financial market ESW economic and sector work ROA return on assets FI financial intermediary ROE return on equity FIL financial intermediary loan ROSCAs rotating savings and credit FRR financial rate of return associations FSD Financial Sector Development SACI specialized agricultural credit Department, World Bank institution GAAP generally accepted accounting SAL structural adjustment loan principles SBP Sustainable Banking with the Poor GATT General Agreement on Tariffs SDI subsidy dependence index and Trade SECAL sectoral adjustment loan GB Grameen Bank (Bangladesh) SLA savings and loan associations GDP gross domestic product TAL technical assistance loan IDB Inter-American Development TSE transitional socialist economy Bank WWB Women's World Banking xii Executive Summary R ural financial markets, in particular similar to those that have surfaced over the those involving state-owned agricul- past decade, by carefully adapting them to the tural credit institutions, have been at socioeconomic and cultural setting, and by re- the center of government interventions in most ducing or fully eliminating the heavy fiscal developing countries over the past five dec- costs associated with inefficient and inequita- ades. With the support of various donor agen- ble agricultural credit programs, it is possible cies governments have channeled sizable to further promote efficient rural financial mar- resources into these programs to ensure a con- kets in many more countries. tinued flow of cheap credit to agricultural en- In this report the emphasis is on identifying trepreneurs through an abundance of financial the challenges facing rural financial interme- intermediation mechanisms. The outcomes of diation, evaluating the traditional approach to these interventions have been generally disap- rural finance, looking at how rural finance can pointing. In contrast, developments during the contribute to both income expansion (that is, past decade in the provision of rural financial growth) and poverty reduction, asking what services (both savings and credit) have demon- constitutes a favorable policy environment and strated that proper institutional design and a supportive legal and regulatory framework adherence to appropriate policies pay off for improved rural financial intermediation, handsomely and have the potential to generate and reviewing the role of government. The per- substantial achievements in terms of greater formance of rural financial institutions (RFIs) institutional outreach and self-sustainability. is assessed and guidelines are proposed for Several challenges continue to face rural fi- building institutional capacity in RFIs. Three nancial intermediation-challenges of select- successful REIs are examined and the operat- ing the set of institutional design and policy ing methods and common features that con- alternatives that is most effective in different tribute to their success are identified. socioeconomic environments. These challenges More than 1.3 billion people around the include the selection of proper modes of op- world live in poverty-the vast majority in ru- eration that differ significantly from the old ral areas. Although most of the rural poor de- pattern of (a) using concessional interest rates pend on agriculture and agriculture-related (often negative in real terms), (b) favoring ag- activities for their meager incomes, under- ricultural rather than rural operations, (c) ig- nourishment is severe in many rural areas. noring or oppressing the creation of savings General economic growth is often insufficient deposits, and (d) implementing costly and in- to eradicate poverty and undernourishment. efficient service delivery mechanisms. By im- Growth must be shared by poor rural commu- plementing best practices in rural finance nities, particularly by smallholders and 1 2 Rural Finance women. Efficient rural financial services con- The Challenge for Rural Financial tribute to poverty reduction, particularly when Intermediation supported by hospitable rural development policies. Rural Financial Intermediation Faces Several Like their urban counterparts, rural commu- Challenges nities have a bankable demand for credit, sav- ings, and insurance services. Insurance and * Obstacles to financial intermediation. Finan- access to income-smoothing financial services cial markets are often highly integrated can contribute significantly to the welfare of with (and affected by) other markets. The rural communities by mitigating the impact of main problems preventing financial mar- seasonality and natural disasters on their in- kets from operating efficiently are un- comes. Savings and credit facilities can help to sound macroeconomic policies, distorted make lumpy investments affordable and allo- financial policies and market rigidities, cate resources to potential investments with and legal and regulatory constraints. the highest returns. Indeed, the central role that * Urban-biased policies. The past decades financial systems play in most economic activi- were characterized by policies that were ties makes them an essential component of implemented in pursuit of accelerated in- economic development, and the depth of the dustrial development. These policies often financial sector is an excellent predictor of adversely affected rural areas and ham- long-term economic growth. pered rural development. The following Yet the rural poor have been largely ne- "eight pillars" of urban-biased policies glected by formal financial intermediaries. This have often hampered the development of neglect stems from distorted macroeconomic rural communities and the promotion of and sectoral policies, the lack of economic and rural financial markets: political power of the rural poor, and the per- 1. Overvalued exchange rates ception among formal for-profit financial in- 2. Low, controlled, and seasonally invar- termediaries that rural markets are not iant prices for agricultural products potentially profitable. Successful rural finan- 3. High effective rates of protection for cial intermediation requires a favorable policy domestic industry, the outputs of environment, a supportive legal and regula- which are used as agricultural inputs tory environment, and strong, autonomous ru- 4. Disproportionately high budgetary ral financial institutions. allocations for urban over rural infra- This volume identifies the challenges facing structure (roads, electricity, and rural financial intermediation, evaluates the water supply) traditional approach, looks at how rural fi- 5. Disproportionately high investment nance can contribute to income expansion and in human resources in urban over ru- poverty reduction, asks what constitutes a fa- ral areas (health and education) vorable policy environment and a supportive 6. Usury laws that rule out the loans legal and regulatory framework for improved typical in rural areas: small, risky, rural financial intermediation, and reviews the and high-cost loans role of government in rural finance. The report 7. Underdeveloped legal and regula- reviews and analyzes how the performance of tory provisions regarding land ti- rural financial projects is assessed and pro- tling and collateral for typical rural poses guidelines for building institutional ca- assets (land, crops, and farm imple- pacity in RFIs. In closing, the report examines ments) relative to urban assets (cars, the operating methods of three successful RE;s durables, and homes) and identifies common features that contrib- 8. Excessive taxes on agricultural ex- uted to their success. ports. Executive Summary 3 • Distinguishingfeatures of ruralfinancial mar- pursued short-term objectives aimed at achiev- kets; Poverty, low population density, iso- ing agricultural production gains rather than lated markets, highly covariant risk, and long-term objectives aimed at sustained rural seasonality often result in high transaction income expansion. The sole, disproportionate costs, lack of traditional collateral, high emphasis in RFIs on disbursing agricultural credit income fluctuations, and limited opportu- led to general neglect of portfolio quality, non- nities for risk diversification. These fea- farm rural development, savings mobilization, tures differentiate rural financial markets and efficient rural financial intermediation. from urban ones and often scare off tradi- * Rural communities have been perceived as tional for-profit financial intermediaries. too poor to save, so efforts have been con- However, these challenges also under- centrated almost exclusively on the provi- score the benefits the rural poor would sion of credit, ignoring the perhaps more gain from access to efficient consumption- crucial benefit of rural monetary savings smoothing mechanisms and financial serv- facilities. ices that could help them out of poverty. * Subsidized interest rates have led to RFIs To this end, special products and operating often being perceived as governmental methods are required. disbursement windows rather than solid * Poorly designed interventions. Well- financial institutions. Such perceptions intended rural financial interventions, have led to a poor loan repayment culture. such as subsidized credit targeted exclu- Subsidies in the form of concessional lend- sively to agricultural development, have ing rates and a high tolerance for defaults generally hampered the development of have often been captured by well-to-do rural financial markets. As a consequence, and influential farmers, thereby crowding the poor still have inadequate access to out poor farmers and further reducing the credit, savings, and insurance services. poor's access to credit. o The traditional approach has focused on The Traditional Approach to Rural Finance lending for agricultural purposes while ig- noring small nonagricultural rural enter- The traditional approach usually implied a high prises and neglecting opportunities for level of governinent intervention in the form of risk diversification and growth. targeted credit, with government-owned and a Subsidized agricultural credit has at managed RFIs receiving concessional loans and times resulted in production inefficiencies on-lending to customers at below-market inter- by targeting the wrong products. For ex- est rates. This form of intervention was often ample, by encouraging the adoption of based on serious misconceptions of the real excessively capital-intensive farming tech- challenges facing rural communities and was nologies, low-interest loans displaced ag- directed toward the symptoms rather than the ricultural laborers and increased rural causes of inadequate rural financial intermedia- unemployment. tion. Por example, low participation in formal * The poor design and performance of state- financial activities was erroneously seen as re- owned RFIs and their access to con- sulting from the inability of the poor to save or cessional funds discouraged private, to pay market rates of interest. for-profit financial intermediaries from en- While these interventions had some positive gaging in rural financial intermediation. effects, the objectives of income expansion and The disappointing outcome of the traditional poverty reduction were not met. In fact, some of approach led critics to question whether scarce these interventions have compounded the dis- public resources could be applied more effi- tortions, thereby worsening the condition of ru- ciently in other ways in pursuit of rural income ral communities. Traditional strategies usually expansion and poverty reduction. 4 Rtural Finance The New Approach Removing the causes of market failure may require direct government intervention. Com- A radical change had long been advocated by a mon causes of market failure in rural financial few lone voices. The new approach gradually markets are imperfect information and poor became mainstream thinking, however, as evi- communication of information. Cost-effective dence of "success stories" became knowrn during methods of dealing with market failures the 1980s. The new approach focuses on the pri- should be ídentified and evaluated. If a market mary goals of rural development: income expan- failure cannot be removed cost-effectively, di- sion and poverty reduction. It recognizes that rect interventions aimed at income expansion providing rural finance may not always be the are not justified. The issue of direct interven- most cost-effective way of reaching these goals tion remains controversial because the socially and that effective rural financial intermedia- optimal level of financial services for a particu- tion should often be complemented by other lar community is hard to determine. government actions, such as increased invest- ment in rural infrastructure and in human Poverty Reduction development. The new approach proposes an active role Rapid growth in rural economies is often the for government in establishing a favorable pol- most pronising way to reduce rural poverty. icy environment to facilitate the smooth func- Creating a favorable policy environment and tioning of rural financial markets, but a more pursuing reforms in the legal and regulatory limnited role in direct interventions in rural fi- framework should therefore receive the same nancial markets. priority as they do for income expansion. However, raising average rural incomes may Income Expansion not be sufficient to reduce poverty if economic growth is not appropriately shared. A program Approaching the objective of income expan- of targeted interventions for poverty reduction sion calls for knowledge of the underlying interventions may be justified, irrespective of causes of depressed growth. Obtaining such whether market failures (or cost-effective inter- knowledge requires an assessment of the effi- ventions to remove the market failures) have ciency of markets, íncluding rural financial been identified. The severity of poverty may markets, and a determination of the causes of compel immediate (perhaps short-term) ac- market inefficiencies. These inefficiencies can tions, such as food relief or financial assistance be the result of some combination of poor following natural disasters. macroeconomic and sectoral policies, a weak legal and regulatory framework, and market Creating a Favorable Policy Environment failure. Since markets are highly interrelated, distortions in one market often have spillover The main measures in maintaining a favorable effects in other markets. policy environment for rural financial interme- Once the causes of market inefficiencies have diation are maintaining macroeconomic stabil- been identified, the government can review the ity, removing urban-biased policies, and options available to promote more efficient ru- promoting integrated and resilient financial ral financial markets and use cost-benefit markets. analysis to select the most appropriate actions. Policy reforms should take first priority to en- Establishing a Favorable Macroeconomic sure the effective functioning of markets. Legal Environment and regulatory reforms can also be undertaken immediately, even if policy reforms are yet to Macroeconomic instability affects rural finan- be implemented. cial markets directly through monetary vari- Executive Summary 5 ables such as real interest rates, or indirectly ing public investment priorities toward rural through effects on the RFIs' clients. Inappro- areas, and increasing participation in commu- priate macroeconomic policies have often un- nity development. These reforms will improve dermined efforts to strengthen financial the efficiency of the rural sector and enable sectors. Persistent distortions in key macroe- rural communities to earn a higher return on conomic variables (such as overvalued ex- their investments. change rates) distort price signals, causing financial markets to channel excessive re- Promoting Efficient Financial Markets sources to inefficient sectors instead of to sec- tors with comparative advantage. Rural financial intermediaries are exposed to Both external and internal factors may gen- risks that can lead to crises, as are other finan- erate macroeconomic instability. External risks, cial intermediaries. These risks include inade- such as terms of trade shocks, can be diversi- quate prudential regulation and supervision, fied internationally by opening domestic mar- inappropriate direct interventions in financial kets to foreign investors, or they can be hedged markets, and financial repression. Govern- in international options and futures markets. ments can promote financial markets by Internal factors include volatile fiscal and strengthening the supervision and prudential monetary policies, as well as political risk. The regulation of financial intermediaries, and by goals of prudent fiscal and monetary policies introducing measures supporting financial lib- should be price stability and sound, well- eralization, including deregulating interest aligned exchange rates. rates, reducing high reserve requirements, and relaxing credit controls. Removing Policy Biases against Agriculture In rural markets governments can establish and Rural Development special regulatory frameworks for semiformal institutions. Inefficient policies, such as requir- Rural development in most developing coun- ing that banks lend a large proportion of their tries has been slowed by policies that favor funds to unremunerative sectors of the econ- industry over agriculture and urban over rural omy (such as agricultural parastatals), should sectors. Such policies reduce the profitability be abolished. of agriculture and nonfarm rural enterprises and devastate rural financial markets. Coun- The Legal and Regulatory Framework tries with the highest degree of discrimination for Rural Financial Markets against agriculture have had the lowest rates of economic growth. In many countries deficiencies in laws, regula- In addition to the adverse effects of urban- tions, and institutions prevent the formal sec- biased policies on rural welfare and promotion tor from delivering credit to farmers and rural of rural financial markets, policies often fa- businesses. These impediments may make it vored large wealthy farmers over smallholders difficult for the formal sector to lend to the and poor subsistence farmers, thereby worsen- informal sector and for banks and other finan- ing income distribution in agricultural and ru- cial institutions to lend to nonbank creditors ral sectors. (typically traders) who have many advantages Working principles for creating an improved in efficiently reaching poor rural borrowers. environment for agricultural and rural devel- Lenders need a system through which claims opment include establishing a more nearly against property can be created, publicly estab- neutral trade regime between agricultural ex- lished, or perfected and enforced. The more portables and importables, removing nontariff uncertain and expensive this process, the less barriers, realigning overvalued exchange rates, willing are lenders to lend. Problems can arise reducing excessive industrial protection, shift- in creating a mortgage or a claim on movable 6 Rural Finance property because of untitled land, high regis- Government interventions should aim at re- tration costs, and the absence of legal provi- moving the causes of market failure or poverty. sions for future interests. When perfecting a For example, market failure caused by imper- claim, there may not be clearly designated, eas- fect information can be addressed, where ap- ily accessible registries, and search costs may propriate, by providing seed capital for the be high. Enforcement of claims on mortgaged establishment of RFIs in remote areas. The high land or property can be extremely costly, transaction costs associated with the small loan lengthy, and uncertain. Other laws and regu- and deposit amounts of the poor can often be lations that constrain rural financial interme- lowered through subsidies sharing a portion of diation include exempting property provisions the transaction costs on a declining basis, or by that prevent rural smallholders from using assuming the costs of training staff in modes their smallholdings as collateral and usury of operation known to reduce transaction costs. laws that rule out small, high-interest loans Interventions should always aim to comple- from formal financial intermediaries. ment, facilitate, or improve the market over the The legal, regulatory, and institutional long term. The management of a state-owned changes needed to expand access to credit in RFI requires full autonomy and should be ac- rural areas include titling and registering land; countable: no special privileges should be ac- reforming the law on secured transactions; re- corded to such an institution, and competition forming legal registries and expanding the from private RFIs should be encouraged. Sub- scope for private operation; lowering the costs sidies or grants could be restricted to seed capi- of registration and foreclosure; drafting spe- tal or have a sunset clause. The cost of an cific, clear, and limited homestead provisions; intervention should be monitored continu- and removing interest rate ceilings. ously in order to assess the cost-effectiveness Well-designed programs that have reformed of the intervention while maintaining a level the laws of secured transactions have resulted playing field among participating financial in- in increased supplies of credit and lower lend- termediaries, irrespective of their ownership. ing interest rates. These benefits have pro- Analysis of the informal financial market duced gains that have been estimated at will help to determine the demand for formal several percentage points of gross domestic financial services and methods of operation product (GDP); the costs of such legal reform that may be acceptable to the specific clientele. programs are usually remarkably low. The presence of rotating savings and credit as- sociations (ROSCAs) may indicate an accep- Direct Interventions tance of group-based schemes and demand for savings facilities. Interventions could be de- There is no single optimal level or form of in- signed to complement these schemes, for ex- tervention. The most appropriate intervention ample, through savings accounts with positive will depend on the objective of the intervention returns for the deposits of ROSCAs. and on such variables as the demographics Two widely used products that have gener- of the target clientele and the socioeconomnic ally performed poorly are credit guarantee environment. In designing interventions, schemes and crop insurance. Guarantees have governments can mix and match instruments- often failed because they did not adequately such as by supporting pilot programs or pro- cover credit risk and administrative costs and, viding seed capital; institutions-such as in some instances, discouraged financial disci- specialized agricultural credit institutions, pline and collection efforts. Crop insurance nongovernmental organizations (NGOs), or schemes generally failed because they pro- state-owned RFIs; and products-such as credit, vided multiperil coverage for uninsurable savings, guarantees, or insurance to targeted risks; some schemes have also been under- clientele. mined by political interventions. Given the Executive Summary 7 alternative uses to which funding for guarantee with cheap credit to developing institutional schemes and crop insurance could be put, it is capacity and improving institutional perform- often difficult to justify either with confidence. ance for a broader range of RFIs. Institutional capacity refers to the degree to which operating Perforrnance Criteria for Rural Financial procedures and regulations are embedded in Intermediation the daily activities of an organization. Building this capacity is a continuous process aimed at Evaluations of the impact of agricultural credit improving the performance of an RFI and re- projects on the farm level are fraught with quires adherence to sound business principles. methodological problems relating to the fungi- There is no single formula for a successful bility of money and the assessment of addition- RFI. The most appropriate modes of operation ality. Rigorous econometric studies are costly will be determined by the needs and socioeco- and often highly specific, making generic ap- nomic characteristics of the target clientele, as plication of their findings impossible. well as by the physical, economic, and regula- A framework was introduced in 1992 for as- tory environment. sessing the performance of RFIs. It has gained A growing number of successful RFIs wide acceptance among practitioners and aca- provide potentially replicable or adaptable demics. The framework proposes two primary methodologies. These successes indicate ways criteria, outreach and self-sustainability. It is of developing institutional capacity that will based on the assumption that RFIs that provide ensure responsive, sustainable, and efficient a broad range of services to the target clientele services to rural clients. An understanding of in an efficient manner are likely to have the informal financial markets in a particular desired impact of expanding incomes and re- community can provide guidance on the range ducing poverty. Evaluating the performance of of workable practices within the community RFIs based on these criteria could serve as an and identify market opportunities to better easily quantifiable proxy for the impact of rural meet the demand for financial services in that financial intermediation. community. Outreach is measured by several indicators, Perhaps the most important factor that such as number of clients, average loan size (as determines the success of RFIs is ensuring ap- proxy for income level), and percentage of fe- propriate governance. The powers and respon- male clients. Self-sustainability is assessed by sibilities of all decisionmaking entities and calculating an RFI's subsidy dependence index individuals must be clearly defined, and the (SDI), which indicates the percentage by which boundaries of roles and responsibilities must an RFl's prevailing average on-lending interest be fully understood and consistently enforced. rate would have to increase to make it self- Management must have autonomy and be held sustainable (that is, subsidy independent). The accountable for operational decisions, and cli- SDI also indicates the cost to society of subsidiz- ents' interests must be fully represented. The ing an RFI measured against the interest earned appropriate form of supervision and prudential by the RFI in the marketplace. The main factors regulation in an RFI will depend on the size, that contribute to self-sustainability are adequate type, and ownership structure of the institu- on-lending rates and interest rate spreads, high tion. External supervision becomes particu- rates of loan collection, high levels of savings larly important when the institution mobilizes mobilization, and low administrative costs. voluntary deposits fromn the general public. Other key requirements for successful RFIs Building Institutional Capacity involve personnel, lending policies designed specifically for local clienteles, savings mobi- The focus has shifted dramatically from nur- lization, and monitoring and information turing subsidized government-run institutions systems. 8 Rural Finance * Clearly defined strategies and objectives. While their operating methods differ, reflect- * Motivated and skilled staff. Staff training ing their dient needs and the particular socioeco- and incentive systems are essential for im- nomic and legal environment of the countries proved performance. they serve, certain comrnmon features can be iden- * Innovative low-cost systems and proce- tified. These include a high degree of manage- dures that meet the special needs of the ment autonomy in the formulation of operational target clientele. Examples are mobile bank- policies, intensive staff training and advanced ing for improved delivery of financial serv- staff incentive programs, innovative low-cost ices and joint liability lending. These distribution networks, positive real on-lending systems reduce transaction costs and pro- interest rates, dose monitoring of loan perform- vide security for loans in the absence of ance and high loan recovery rates, increasing traditional forms of collateral. savings mobilization and declining reliance on * Positive interest rates on loans and high donor funding, relatively low and declining ad- loan repayment rates. Flexible loan terms ministrative costs, effective MIS, and macro- and conditions, careful monitoring of loan economic stability and growth. repayments, and incentives to clients for early repayment could improve loan re- Conclusion payment performance. • Savings mobilization through flexible and The traditional approach to rural finance accessible savings facilities and positive in- proved ineffective in achieving income expan- terest rates. sion and poverty reduction. It was charac- • Risk diversification through lending for terized by an active direct role for government both agricultural and nonagricultural ac- through interventions involving the provision tivities, geographic diversification, and in- of cheap agricultural credit through heavily tegration with the broader financial subsidized, state-owned RFIs. The new ap- system. proach suggests a minimalist, market-oriented • Advanced management information sys- role for government and focuses on creating a tems (MIS) that enable the monitoring of favorable policy environment and a supportive loans, individual client records, and staff legal and regulatory framework for rural finan- performance. cial intermediation. A consensus is forming on this new direc- Best Practices: Three Success Stories tion, but the appropriate level and form of di- rect government intervention remains a Three RFIs widely perceived as successful with contentious issue. In pursuit of income expan- respect to the primary performance criteria of sion direct interventions are justified only if an outreach and self-sustainability are the Bank identified market failure can be removed cost- for Agriculture and Agricultural Cooperatives effectively. In practice it may be difficult to (BAAC) in Thailand; the village banks (unit identify and measure the extent of market fail- desas) of Bank Rakyat Indonesia (BRI-UD); and ure. In pursuit of poverty reduction direct gov- Grameen Bank (GB) in Bangladesh. All three ernment interventions may at times be justified institutions have attained high levels of market by social objectives. While interventions aimed penetration in their target markets, and their at poverty reduction do not invariably produce clientele now numbers in the millions. The BRI- net (unweighted) economic benefits, they UD became financially self-sustainable in 1988, should at a minimum reflect the most cost- and both the BAAC and the GB have made effective way of reducing poverty for the target substantial progress towards self-sustainabil- group. ity, as measured by the remarkable decline in Much agreement has been reached on im- their SDI in recent years. proving financial intermediation at an institu- Executive Summary 9 tional level. The building of institutional capac- to expand the product lines of the ex- ity is essential for RFIs striving toward greater isting institutions? outreach and self-sustainability. The most im- * What should be the role of apex portant requirement for developing institu- on-lending institutions, commercial tional capacity is a high level of management banks, NGOs, and rural savings autonomy and accountability for the perform- schemes? ance of an RFI. No special privileges should be * Role of government. Develop guiding prin- extended to state-owned RFIs-rather, all RFIs ciples for the types of intervention re- should be guided by commercial imperatives, quired, if any, in view of a country's and competition should be strongly encouraged. socioeconomic conditions, the type of cli- If subsidies are warranted, these should be pro- entele targeted, and the products to be de- vided irrespective of the RFI's ownership. livered. Research implementation and sequencing issues for the reform of rural Proposed Areas for Future Research financial and legal sectors and for institu- tional development. This summary has provided a static view-a * Informal market. Determine what lessons landscape of an important subsector. From this can be learned from the performance of tour d'horizon one can, however, identify key informal markets. How can formal finance issues and opportunities that will have to be best complement informal financial activi- explored: ties? Institutional development. Explore strategies * Cost-benefit analysis and performance indica- for selecting the appropriate institutional tors. Develop a framework for conducting design and modes of operation to serve a cost-benefit analyses of financial interme- specific target clientele given the political diation and nonfinancial intermediations, and economic setting, the degree of matur- interventions, and alternative rural finan- ity of the rural and financial sectors, and cial intermediation aimed at income ex- the social and cultural conditions. This re- pansion and poverty reduction. Define key search requires a focus on the following performance indicators to measure the issues: cost, effectiveness, and impact of projects * Is there an appropriate institution in over time. operation that could be approached to * Impact studies of selected schemes. Conduct expand its operations with respect to econometric studies to better understand the target clientele? the impact of selected rural financial * What policy measures could be taken schemes on income, production, and con- to facilitate improved performance or sumption smoothing. Introduction R ural financial markets in developing The often disappointing performance of countries are different from other finan- many rural financial projects can be ascribed cial markets. Four main differences largely to (a) insufficient knowledge and mis- prevail: the underdevelopment of complemen- conceptions about the real problems of rural tary markets and related institutions, the mea- financial markets, such as information asym- ger availability or absence of reliable collateral metry, segmented markets, and high covariant security, the covariant risks and seasonal fluc- risk; (b) the effects of historically distorted, tuations in demand for and supply of short- urban-biased macroeconomic and financial term financial resources, and the high cost policies on the rural sector; and (c) poor man- associated with serving rural clients in low agement of rural financial institutions. density areas. These problems are far more se- However, a few recent successful develop- rious in rural financial markets in developing ments in rural financial markets that empha- countries than they are in the sophisticated fi- size the efficient delivery of savings and credit nancial markets of industrial countries. services to the rural population have demon- In recognizing the limitations of rural finan- strated that adherence to appropriate policies, cial markets in developing countries, states proper institutional design, autonomy of man- pursuing both growth and more equitable in- agement, creative incentives, and efficient come distribution have often considered inter- modes of operation can generate substantial vention to be inevitable. These interventions social returns. have taken many forms, but have focused in Improved rural financial intermediation has particular on augmenting the availability of ag- been receiving increased attention from gov- ricultural credit and on reducing the burden of ernments and donors as a policy tool with farmers' interest payments. which to address the plight of rural popula- Agricultural credit programs, including the tions. The importance of encouraging healthy widespread establishment of and ongoing state rural financial markets is increasingly acknow- support to specialized agricultural credit insti- ledged in the quest to foster rural develop- tutions, have been at the center of government ment, accelerate growth, and reduce poverty. interventions in rural financial markets. Do- nors have supported these interventions by as- Purpose of This Volume sisting in their design and by channeling substantial financial resources for on-lending The purpose of this volume is to develop broad in pursuit of accelerated growth in agricultural guidelines for improved rural financial inter- production and poverty reduction. mediation and to highlight areas that require 11 12 Rural Finance further research. The findings presented here the form of a "sourcebook," with chapters have been derived from an analysis of past that are largely self-contained to enable read- rural financial intermediation efforts and re- ers to draw directly on sections of particular cent developments in economic theory. The re- interest. port sets out to answer the following questions: Layout of the Volume At the policy level * What are the specific challenges to rural Part One. An Overview of the Traditional financial intermediation that differentiate Approach it from urban financial intermediation? * Under what circumstances should support Chapter 1 describes the problems associated with for rural financial intermediation be con- financial markets in general and rural finan- sidered as a policy option for the expan- cial markets in particular. Chapter 2 gives an sion of rural incomes and the reduction of overview and assessment of the traditional ap- rural poverty? proach to rural financial interrnediation, which • What is the impact of macroeconomic, fi- dominated the world until the mid-1980s. nancial, and agricultural sector policies on rural financial intermediation? Consider- Part Two. Policiesfor Improving Rural ing these policies and the legal and regula- Financial Intermediation tory framework of a country, what would constitute a favorable policy environment Chapter 3 discusses the new approach to rural for rural financial intermediation? financial intermediation and provides a sys- • What type of direct and indirect measures tematic approach to determining the role of are available to governments to facilitate government in pursuing rural income expan- improved rural financial intermediation? sion and poverty reduction. The approach is illustrated by two decision trees: "Income Ex- At the institutional level pansion" and "Poverty Reduction." Chapters * How can the economic impact of rural fi- 4 and 5 develop recommendations on the role nancial intermediation be evaluated? of government in creating a favorable policy * What constitutes sound business princi- environment. Chapter 4 focuses on the macro- ples for conducting rural financial inter- economic, rural, and financial sector, and mediation at an institutional level? chapter 5 on creating a supportive legal and * How can the performance of rural financial regulatory framework. Chapter 6 discusses the institutions be assessed? types of direct rural financial interventions that can be implemented by governments to im- Audience prove rural financial intermediation in pursuit of income expansion and poverty reduction. The report is intended for government policy- makers, task and project managers at the Part Three. An Overview of Rural Financial World Bank, other donor agencies and non- Institutions governmental organizations, and managers of rural financial institutions (RFIs) who wish to Chapters 7 to 9 review successful rural finan- expand or improve their operations in rural cial intermediation at the institutional level. areas. Policy-related issues may be of particular Chapter 7 discusses the problems inherent in interest to policymakers; issues at the institu- assessing the impact of rural finance projects tional level may be of greater interest to practi- and proposes a framework for assessing the tioners. The report is therefore arranged in performance of RFIs as a proxy for evaluating Introduction 13 the success of rural finance projects. Chapter 8 desas of Bank Rakyat Indonesia, the Bank for proposes guiding principles for managing RFIs Agriculture and Agricultural Cooperatives of and developing institutional capacity to im- Thailand, and the Grameen Bank of Bangla- prove the performance of RFIs. Finally, chapter desh. The common features identified in their 9 analyzes the performance and operating operations can prove useful in the design of methods of three successful RFIs: the unit future interventions. PART ONE An Overview of the Traditional Approach Tf ihe formulation of policies for improved policies and methodologies that contributed to rural financial intermediation calls for a achievements and failures. thorough analysis of the problems un- Chapter 1 develops the problem statement, derlying rural financial markets and a clear un- which can be used as a checklist in the design derstanding of the dynamics of financial of future projects. Chapter 2 provides an over- markets in general and rural financial markets view of the policies and outcomes of the tradi- in particular. In addition, past performance tional approach, which largely dominated should be analyzed to identify the underlying thinking on rural finance until the late 1980s. 15 y CHAPTER 1 The Challenge of Rural Financial Intermediation he challenges of rural financial markets the performance of financial markets, and are broadly related to (a) obstacles to negative external shocks could aggravate and shortcomings of financial markets the resulting situation. in general, (b) urban-biased policies, (c) sys- * The sectoral policy context. Government temic weaknesses of rural financial markets, price controls, trade policy, and public in- and (d) poorly designed interventions. vestment priorities frequently distort the allocation of resources by financial inter- Obstacles to Financial Intermediation mediaries. Financial market constraints. Financial mar- Various obstacles prevent financial markets in ket rigidities, imperfect information, and general, and rural markets in particular, from in some instances social barriers to finan- operating efficiently.1 These obstacles all have cial transactions preclude an optimal allo- a direct impact on the effectiveness of rural cation of resources. financial markets. Most of the obstacles are * Legal and regulatory constraints. Problems related to the promissory feature of financial with enforcing claims increase uncertainty contracts. For example, a bank will lend to a and reduce the expected returns to credi- client in exchange for a promise of repayment tors from financial transactions, thereby with interest over a given period. Because increasing transaction costs and reducing of the promissory feature and the time di- the supply of credit and deposits. mension of financial contracts, financial in- These constraints are generally more binding termediaries and their clients require good on financial transactions in the rural sector be- information to determine the riskiness of trans- cause of the effects of widespread, distorted actions, need a reasonably stable political and sectoral policies and the unique characteristics economic environment to extend contract ma- of rural markets. turities over time, must be free to price per- ceived risks appropriately, and must be able to Urban-Biased Policies exercise remedies when contractual terms are not honored. Widespread urban bias exists in sectoral poli- These conditions are rarely met. The main cies and in the orientation of the legal and regu- problems are: latory framework. (This bias is outlined in * The macroeconomic environment. Unsound chapter 4, box 4.1: "Eight Pillars of Urban- macroeconomic policies adversely affect Biased Policies.") Examples of these policies 17 18 Rural Finance include restrictive price controls on agricul- risk and low profitability), commercial banks tural produce; excessive agricultural export have largely avoided servicing rural areas. taxes; overprotection of domestic industry; low Often, the only financial services available are investment in rural infrastructure and human those provided by the informal sector, which resources; and collateral laws that are better offers a limited range of services along with defined for urban assets, such as consumer limited availability of funds. Cases in which durables, than for farm implements. In many households have paid to have their savings countries with a comparative advantage in ag- placed in safe keeping are well documented. riculture macroeconomic policies that result in overvalued exchange rates are also urban- Poorly Designed Interventions biased because the policies encourage exces- sive food imports and thus depress agricul- Governments responded to the perceived tural producer prices and rural incomes. shortage of financial services in the rural sector by creating a range of different institutions, Systemic Weaknesses of Rural such as specialized agricultural credit institu- Financial Markets tions, intended to channel government and do- nor funds to rural clients, especially farmers. Rural populations are generally poorer than Though well intended, some of these initiatives their urban counterparts. They generally work were misdirected and did not address the real in agriculture or agriculture-related activities, problems. Indeed, some exacerbated the prob- and they live in areas in which overall popu- lems. A few examples: lation density is low. These factors, combined * Subsidized below-market ceiling rates of with poor infrastructure and lack of integra- interest often led to the crowding out of tion with urban markets, cause rural inhabi- poor farmers because the subsidies were tants in many countries to live in relative captured by wealthier, better-connected isolation. These characteristics are related to farmers, increasing the income disparity the following problems in rural financial inter- between rich commercial farmers and poor mediation: subsistence farmers and reducing the • Low population density, small average poor's access to credit. In many instances loans, and low household savings increase subsidized credit also became highly poli- transaction costs. ticized, and it was consequently difficult to * Rural clientele often lack the traditional eliminate.3 forms of collateral required by commercial * Inefficient business practices often re- banks. sulted in substantial losses and further re- * Poor communication and lack of integra- duction in access to financial services by tion with other markets result in highly the poor. Rural financial institutions were fragmented markets, which create infor- often treated as disbursement windows mation barriers and limit risk diversifica- rather than as financial institutions, and tion. recipients sometimes viewed "soft" loans * Seasonality of the rural agricultural busi- as grants that did not have to be repaíd. ness cycle and the high probability of co- * The focus on lending exclusively for agri- variant production price and income culture increased the risk carried by RFIs shocks add to the risks of rural financial and reinforced the perception that lending intermediation. to rural areas is a specal activity, not to Because of inadequate collateral (and the per- be integrated with the broader financial ception that rural financial markets have a high market. The Challenge of Rural Financial Intermediation 19 Conclusion integrated solution, taking into consideration macroeconomic, microeconomic, sectoral, and These problems have resulted in rural commu- institutional issues. nities, particularly their poor members, being Chapter 2 discusses in more detail the suc- underserved by the formal financial sector. cesses and shortcomings of the traditional ap- Consequently, rural households often have less proach to rural finance. Part 2, chapters 3 access to adequate credit, savings, and insurance through 6, proposes a framework for develop- services than urban ones. The underlying causes ing an integrated approach to improved rural of the problems are interrelated and call for an financial intermediation. CHAPTER 2 The Traditional Approach to Rural Finance I T hroughout the world, financial markets Government-intervention instruments in- are subject to government intervention. clude financial accounting standards, pruden- During the 1980s government loans and tial regulations, deposit insurance, loan guarantees accounted for about 25 percent of guarantees, interest floors or ceilings on loans all lending in the United States (Schwarz 1992). and deposits, interest subsidies, fiscally moti- The largest Indian commercial banks are gov- vated preemption of funding (for example, ernment owned, and smaller private commer- through unremunerative reserve asset require- cial banks in India have had up to 60 percent ments or liquidity ratios), and supervisory of their resources preempted by the govern- mechanisms (contracted or implemented di- ment or locked into directed lending. Until rectly by governments). Governments have 1988 the Soviet Union's entire banking system also intervened in financial markets by acting consisted of just two state-owned banks: a sav- as lenders of last resort to maintain the liquid- ings bank (Sberbank) and a lending monobank ity of the system; prohibiting loans to particu- (Gosbank). The degree of intervention has var- lar groups (for example, shareholders or ied by country, but in no case has any govern- insiders); restricting entry and exit of banks ment refrained from intervening in the rural from particular sectors; directing concessional financial sector. loans to given sectors; directing loans to par- Governments have intervened in financial ticular clients; establishing new government markets to protect depositors, ensure the sol- banks; and nationalizing, expropriating, or vency of financial intermediaries, promote prohibiting privately owned banks.5 competition, ensure macroeconomic stability It is widely accepted that without indirect and growth, and pursue social and political interventions there would be widespread fail- aims. To achieve these ends, governments have ures in financial markets. However, the case used various instruments to regulate, direct, for direct interventions is more controversial increase, complement, and supplant the provi- and is at the core of the difference between the sion of financial services by privately owned traditional approach that dominated economic financial intermediaries.4 These interventions thought on rural finance until the 1980s and have sometimes been indirect, aimed at im- the new perspectives on rural finance that have proving the policy environment (by promoting emerged since 1980 (see chapter 3). macroeconomic and financial sector stability, for example), in other cases they have been The Traditional Perspective on Rural Finance direct and have included actions such as estab- lishing state-owned financial intermediaries The traditional perspective on rural finance en- and subsidizing interest rates. visages a govermnent active in rural financial 20 The Tradítional Approach to Rural Finance 21 markets and leans heavily toward direct inter- lished RFIs, notably specialized agricultural ventions. It dates from the 1950s, when Keynes- credit institutions, to provide credit to farmers. ian economic thinking was inspiring successful Some countries (India in 1969 and 1980, and interventions at the macroeconomic level. Over Mexico in 1982) nationalized commercial the next thirty years policymakers identified a banks to increase lending to agriculture. Other host of problems in rural financial markets, countries (Morocco añd Thailand) mandated which they tried to resolve by targeting, pro- that commercial banks allocate a specified por- viding subsidies, and applying government tion of their portfolio to agriculture or to gov- controls. They hoped that these measures ernment-owned specialized agricultural credit would stimulate growth and reduce rural pov- institutions for on-lending to agriculture. erty. The (perceived) problems included: Lack of Modern Technology in Agriculture Lack of Credit in Rural Areas Rapid growth in agricultural productivity was Private banks avoided lending in rural areas essential to keep pace with rapid population because they viewed the sector as risky and growth. Policymakers believed that the best unprofitable. They particularly avoided long- way to increase agricultural productivity was term loans with grace periods, which are im- to encourage the use of modern pesticides, portant for certain agricultural investments fertilizers, and farm equipment. Because (see box 2.1). As a result, goverrnents estab- farmers were viewed as risk-averse and cash- Box 2.1 Merits of long-term finance Does lack of long-term finance prevent firms in de- whereas a similar firm in a more competitive econ- veloping countries from investing in potentially prof- omy may require external financmg to grow at the itable growth opportunities? The answer appears to same rate. be yes. An empirical study investigated the question To control for this endogeneity, the authors of of using firm-level data on thirty industrial and de- the study estimated for each firm in their sample a veloping countries for the 1980-91 period. The evi- predicted growth rate that could be attained by rely- dence shows that use of long-term credit by firms is ing on self-financing and short-term credit only. They important in facilitating their growth. Interestingly, show that the proportion of firms that grew at rates the study finds no evidence that government subsi- exceeding this predicted rate is related to the devel- dies to firms are associated with the ability of firms opment of financial markets-both long-term debt to grow faster. To the contrary, the evidence indicates and equity-and legal institutions in the economy. that although the ability of firms to enter into long- Thus the underdevelopment of financial markets and term debt contracts is associated with greater num- institutions prevents firms in developing countries bers of firms growing at higher than predicted rates, from investing in potentially profitable growth op- the result is reversed when govenmments subsidize portunities. However, goverrment subsidies to in- credit. dustry do not increase the proportion of firms A firm's external financing needs depend on the growing at higher than predicted rates, and the magnitude of its intemal cash flows relative to its positive effect of long-term debt is reversed to the investment opportunities. Both the firm's cash flows. extent that the debt is subsidized. Thus, although and its optimal investment level are endogenous. The the study found evidence to support the premise ratio of cash flows to optimal investment may differ that "availability of term finance affects firm systematically across countries, even for similar firms growth" (an argument used to justify many directed- employing the same technology. Thus, for example, credit programs), it also found that government inter- a firm with capital-intensive technology may need to vention in providing such finance has generally not finance large investment expenditures in order to been successful, possibly due to weaknesses in design grow. If the firm has sufficient market power or and underlying institutional infrastructure. faces high demand, it may be able to generate suffi- cient cash flow internally to finance investment, Source: Dernrgii-Kunt and Maksimovic 1996. 22 Rural Finance constrained, interest rate subsidies were used populations. Interest subsidies and debt forgive- to accelerate adoption of these technologies. ness offered visible means of meeting this objec- tive and placating rural voters. The benefits of Pursuit of Industrial Growth at the Expense the subsidies and loan write offs accrued primar- of Agriculture ily to wealthy and influential rural borrowers, but this was an acceptable outcome to many gov- Influenced by dualist theories of economic de- ernments (Ladman and Tinnermeier 1984). velopment and by the example of the wealthy industrial economies, policymakers in devel- Assessing the Traditional Approach oping countries looked to industry rather than to agriculture as the vehicle for rapid growth.7 Experience has varied across countries, but the Agriculture was often heavily taxed through results of the traditional approach have gener- price and budgetary mechanisms to subsidize ally been poor or modest at best. This con- industrial expansion. However, for equity rea- clusion is based on an assessment of the sons (and to ensure adequate food supplies for traditional approach against three criteria: urban consumers) policymakers sought to (a) the success of traditional strategies, such as compensate farmers for this heavy taxation by targeting and subsidization, in addressing the providing credit subsidies. perceived problems; (b) the cost-efficiency of traditional methodologies; and (c) the success Prevalence of Usurious Moneylenders of the traditional strategies and objectives in achieving the broader goals of income expan- Interest rates in the informal sector were con- sion and poverty reduction. sidered to be unconscionably high (often more than 100 percent a year). It was argued that Success of Traditional Strategies in Addressing poor agricultural borrowers could not afford the Perceived Problems these rates of interest and should not have to pay them. Subsidized, government-directed It is difficult to assess the effects of targeted línes of credit were introduced to drive usuri- credit because of several methodological prob- ous moneylenders out of business and reach lems (see box 2.2). First, it is not clear how farmers with loans at "reasonable" rates of much liquidity a borrower might have had in interest. the absence of a particular "directed" loan. That is, in the absence of the loan the borrower Low Savings Capacity in Rural Areas might have obtained the funds from other sources.. This lack of clarity is a problem in Rural communities were traditionally viewed assessing "additionality." Other methodologi- as too poor and subject to too many shocks of cal problems are examined in chapter 7. nature to be able to save. Policymakers relied on budgetary funds and on external donors to Impact on access to credit. The additionality of provide loanable funds for government rural directed credit programs for agriculture cannot credit programs. be quantified, but in the short run the programs often result in increased investment and sea- Political Pressure sonal credit that benefit agriculture. However, the subsidies are often captured by the rich (see Although immediate political pressure on gov- box 2.3). In addition, because of government ernments tended to come primarily from urban budget constraints the potential number of bor- populations, governments had to do some- rowers decreases as the subsidy per borrower thing, or at least be seen to attempt to do increases, limiting the potential impact on ac- something, to address the concerns of rural cess to credit. The Traditional Approach to Rural Finance 23 Box 2.2 Assessing the performance of agricultural credit projects States, donors, and managers of rural financial insti- Attributing "production gains" to access to formal tutions have used different methods to assess the per- credit might be misleading unless inherent measure- formance of agricultural credit projects. The World ment difficulties are overcome. Frequently, produc- Bank's Operations Evaluation Department (OED) tion gains are attributed to formal credit in an provided several guidelines on such matters: exaggerated and unjustifiable manner. The financial When there are strong and well-substantiated performance of the participating intermediaries, doubts about the comnmtitments or capability of however, has often been unambiguous: the financial intermediary or government to im- Traditional credit projects sought to provide plement a project, the World Bank should defer credit to farmers in need of support. Criticisms lending until doubts are dispelled. Going ahead that these projects failed to address systentic with a project prematurely can lead to far problems are substantiated, as are assertions greater problems than those caused by delaying about the poor performance and lack of finan- it. (World Bank/OED 1989) cial viability of many credit institutions repre- The OED has highlighted the need for a thorough sented in the portfolio. Failure to raise collection assessment of institutional strengthening at the com- rates was nearly universal. Some institutions, pletion of projects, particularly of the financial impact dependent on portfolio and government trans- of Bank operations on intermediary institutions and fers that are no longer reliable, are at risk. Criti- on farming. In a 1990 report the OED asserted that cism of the banking side of traditional "as a working proposition, the OED approach is to agricultural credit operations is mostly valid. require favorable outcomes on both of these counts (World Bank/OED 1996) in order for a credit project to be evaluated as satis- factory" (World Bank/OED 1990). Source: Authors' findings. Impact on agricultural production. The effect of finance does not invalidate attempts to relate traditional strategies on agricultural production credit to incremental physical output and that has been hotly debated. The "Ohio School" difficulties in measuring the impact of credit do (Adams, Graham, and Von Pischke 1984) believes not mean that credit has no impact (World that the methodological problems of evaluating Bank/OED 1993). The study notes that directed the impact of credit render such an assessment credit programs were associated with the adop- futile because of the fungibility of money (see tion of modem technologies, such as green- chapter 7). However, the World Bank's Opera- houses in Morocco and tube wells in northwest tions Evaluation Department (OED) argues in a Bangladesh, and that these irtnovations were study on rural finance that the fungibility of farm associated with production gains. Box 2.3 Credit and income redistribution in Costa Rica Subsidized agricultural credit has been extended in The distribution of the agricultural credit (and hence Costa Rica by four commercial banks, all government subsidies) disbursed by Banco Nacional was actually ownéd and all well positioned to reach the small more skewed than the distribution of income and farmer. By the mid-1970s the three smaller banks had land. Low-interest rates on the loans implied subsi- more than 30 regional offices, and the largest bank dies equivalent to 4 percent of gross domestic prod- (Banco Nacional) had more than 100. Two-thirds of uct and almost 20 percent of value added in these branches were rural offices that specialized in agriculture. This suggests an average credit subsidy credit for small farmers. of $10,000 on each of the big loans that áccounted for Data for Banco Nacional, which accounted for 60 80 percent of Banco Nacional's credit. In 1974 a family percent of agricultural credit disbursed in Costa Rico with an income of $10,500 was in the top 10 percent in 1974, showed that only 10 percent of the bank's of income distribution. agricultural loans that year accounted for 80 percent of the total agricultural credit extended by the bank. Source: Adapted from World Bank 1989. 24 Rural Finance In India state-owned RHIs were associated velopment impact. Box 2.5 presents an extreme with significantly deeper rural financial mar- example of how production gain could result in kets and with measurable production gains. net economic loss. These gains were, however, achieved at tre- mendous cost in efficiency (see box 2.4). The Impact on production technologies. Traditional cost of forgone intermediation by private rural credit was often provided as part of a package financial intermediaries that result from distor- of services, including agricultural extension, in- tions in the state-owned rural finance system puts (fertilizers, pesticides, and high-yielding cannot readily be estimated, even in the most plant varieties), and farm equipment (irriga- solid empirical work. tion pumps, tractors, and combines). Formal credit often accelerated the adoption of modern Impact of misdirected credit. There is evidence technologies when large investments were that governments have often directed credit to involved. However, adopting modern tech- the wrong products. In some formerly planned nologies, such as improved seeds, fertilizers, economies credit was distributed as a matter and pesticides, requires capital in amounts that of policy to the weakest performers to sustain are often available and affordable to farmers their operations. In other cases distorted policy even without access to concessional, directed environments made some crops profitable for credit schemes. farmers, even when countries lacked a com- These interventions might have achieved parative advantage in those products. Under considerably more had attention been paid to these circumstances increasing the supply of designing mechanisms to encourage sound in- credit is harmful to the economy, and produc- vestments. The below-market interest rates tion gains may be a misleading measure of de- and leniency on loan collections that still char- Box 2.4 Impact of formal finance on the rural economy of India A recent study of rural finance in India analyzes data declined. This implies that the policy of forced lending for the agricultural years 1972-73 to 1980-81 for to agriculture had more impact on the substitution of eighty-five rural districts (Binswanger and Khandker capital for labor than on agricultural output, a clearly 1995). The study found that the Indian government counterproductive outcome when the abundance of la- pursued a policy of rapid, forced expansion of com- bor in India is considered. mercial banks into rural areas. This expansion re- The modest increase in agricultural output is com- sulted in a rapid increase of nonagricultural rural pared in the study with the costs to the government employment and a modest increase in the rural wage of the directed and modestly subsidized credit to ag- rate. By inference the study concludes that the policy riculture. The govemment costs include rough esti- must have significantly increased the growth of mates of the transaction costs, interest subsidy, and the nonagricultural rural sector. These findings sug- loan losses in the system. Assuming a default rate of gest that the expansion of commercial banks into ru- 10 percent, the findings suggest that the value of the ral areas eliminated severe constraints in rural extra agricultural output triggered by the targeted financial markets and led to significant rural financial credit almost covered the government's cost of pro- deepening. viding it. Of course, farmers have additional costs in The Indian government also pursued a policy of making their investments that are not included in the directing commercial banks and the cooperative credit government's costs. institutions to lend specifically for agriculture. The These findings indicate that deepening the system study found that the expansion of rural and agricul- of rural financial intermediation in India had high tural credit volumes had a small positive impact on payoffs in rural growth, employment, and welfare, aggregate crop output. This small increase is ac- but that specifically targeting credit to agriculture counted for by a large increase in fertilizer use and by was of doubtful benefit. increased investments in animals and irrigation pumpsets. At the same time agricultural employment Source: Binswanger and Khandker 1995. The Traditional Approach to Rural Finance 25 Box 2.5 Impact of wheat subsidies of all the rural producers affected by dis- in Saudi Arabia criminatory urban-biased policies. In Mexico, notwithstanding a network of more than Until the early 1980s Saudi Arabia imported most of 500 agricultural bank branches and billions of the wheat that it consumed. In 1981 it produced U.S. dollars in directed credit programs 187,000 tons. The Saudi governnent then began to through both a state agricultural development encourage wheat production by guaranteeing farn- bank and nationalized commercial banks, a re- ers artificially high prices, free land, and low-interest loans for machinery, fertilizer, and seed. Less than cent Bank study found that formal credit a decade later, production had increased fifteen foid reached only 8 percent of rural enterprises, to 3 million metric tons per year, two-thirds of which and direct government loans reached less than was exported. Domestically produced wheat cost 1 percent of rural enterprises (Chaves and about US$1,000 per ton when wheat could be im- Sanchez 1995). ported for about US$80 per ton. Agriculture con- sumed more than 80 percent of the kingdom's water, mostly from aquifers that are not naturally replen- Impact on activities of usurious moneylenders. In ished and may soon be exhausted (in twenty years many countries directed credit programs have or sooner, according to some experts). largely failed to displace moneylenders. The Source: Excerpted from Yarbrough and Yarbrough 1991. share of informal finance remains large or even dominant, and interest rates remain high despite below-market rates on formal loans (Ramola and Majahan 1996).81n India, however, acterize most traditional credit programs in- evidence from credit surveys suggests that the crease incentives for substitution of formal share of formal finance grew from about 33 loans for other resources, or even outright di- percent in the 1950s to 55 percent in the 1980s as version of credit to nontargeted, nonproduc- a result of state efforts to curtail the operation tive uses. Some critics of the traditional of moneylenders. Nonetheless, despite the de- approach note that subsidized agricultural cline in the share of informal rural lending, the credit has often resulted in the adoption of absolute volume of such lending rose over the excessively capital-intensive technologies. The same period. The prevailing high interest rates classic example is concessional lending for charged by moneylenders, despite the expan- farm mechanization in Pakistan, where tractors sion of formal (often concessional) credit, can be displaced agricultural laborers and deepened explained at least partly by the adverse selec- poverty in rural areas (Khan 1977). It is now tion of clients: the more creditworthy clients widely accepted that if the goal is to reduce opted to borrow from the formal sector. poverty, and the principal asset of the poor is labor, then a subsidy on capital through below- Impact of political objectives. Stiff political resis- market interest rates is counterproductive. tance in many countries (Bangladesh, Tunisia, the United States) to the removal of agricultural Impact of urban-biased industrial growth prac- credit subsidies suggests that credit subsidies tices. Credit subsidies clearly provided some have achieved their explicit and implicit politi- compensation to rural producers who were cal objectives. However, the subsidies are part powerful or lucky enough to gain access to of an overall urban-biased strategy that has, them. However, total compensation from input even if unintentionally, depressed rural in- subsidies amounted to only a fraction of the comes and accelerated rural-urban migration. value of losses in agriculture that occurred be- cause of policy distortions (Schiff and Valdés Cost-Efficiency of Traditional Methodologies 1992). Compensation in the form of subsidies through the credit system reached only a small The financial performance of virtually all gov- proportion (usually much less than 30 percent) ernment-owned RFIs has usually been ex- 26 Rural Finance tremely poor. Most RFIs have remained highly Lack of managerial autonomy. Lack of manage- subsidy-dependent. In India arrears as a pro- rial autonomy enables politicians to respond to portion of amounts due and overdue hover at populist political objectives rather than sound around 50 percent in most states. The recovery management principles. For example, favored rate of Mexico's BANRURAL was around 25 clients obtain special treatment, and promises percent in the late 1980s (ignoring recoveries of debt forgiveness are made during elections. from the loss-making national agricultural in- Such practices have led to the collapse of RFIs surance company). Recoveries for the Small- during election years in Jamaica (1980), Malawi holder Agricultural Credit Agency in Malawi (1994), and elsewhere. plummeted from almost 90 percent to less than 20 percent during the most recent elec- Poor operating procedures. The consequences of tions; the agency was subsequently declared in- poor operating procedures, such as lack of in- solvent. Inflation eroded the real value of the centives for the clients and staff of RFIs, below- equity of government-owned RFIs throughout market interest rates, and poor information Latin America during the 1980s because of poor systems, include: loan collection and agricultural on-lending * Insufficient interest rate spreads to cover rates that failed to keep pace with inflation. RFIs' overall costs, including provisions The economic cost of this dismal perform- for doubtful loans. ance has been enormous and has often put * Negative real on-lending rates (or below- macroeconomic stability at risk. For example, market rates) that oblige RFIs to pay agricultural credit subsidies totaled 2.2 percent below-market returns on savings, resulting of Brazil's GDP in 1980, and 1.7 percent of in the decapitalization of RFIs. Subsidies Mexico's GDP in 1986 (see box 2.6). In several inherent in below-market rates encourage cases, the subsidies could not even be meas- rent seeking by borrowers and RFI staff. ured because of poor accounting practices. * Obligatory lower on-lending rates for Many governments do not know the full fiscal poorer clients-who actually demand ac- and quasi-fiscal cost of maintaining directed, cess to credit, not subsidized loans-that cheap rural credit prograrns, let alone the as- have reduced lending to the poor and sociated opportunity costs.9 made it less rewarding for lenders. The reason for the poor performance is evi- * Support for state-owned cooperatives with dent: the interventions invariably have been, compulsory membership, that ignores the and generally still are, characterized by a lack merits of voluntary association based on of managerial autonomy for the RFIs and by self-selection principles and discourages poor operating procedures. financial discipline. Box 2.6 Fiscal burden of rural finance in Mexico The Mexican government transferred almost US$23 bil- ment policies led to an almost 30 percent real decline lion (rneasured in constant 1992 U.S. dollars) to its agri- in the federal budget between 1986 and 1992. In the cultural finance institutions over the perod 1983-92 (World process federal funds were reallocated away from Bank 1994c). Most of these transfers consisted of budge- RFIs, causing fiscal and quasi-fiscal transfers for rural tary support from the Ministry of Finance and below- finance to fall from the equivalent of 4 percent of the market rate rediscounts from the Central Bank to the gov- budget to less than 1 percent. Relative to agricultural ernment's two agricultural development banks (FIRA and total GDP, these transfers declined from a peak and BANRURAL). About US$3. billion (17pe ~ent) wentto of 18 percent of agricultural GDP (1.7 percent of total the loss-maldng agricultural insurance companies owned GDP) in 1986 to 2.7 percent of agricultural GDP by the govemment (ANAGSA and AGROASEMEvI. (0.2 percent of total GDP) by 1992. During the late 1980s the transfers became unsus- tainable. Macroeconomic stabilization and adjust- Source: World Bank 1994c. The Traditional Approach to Rural Finance 27 Tolerance for financial indiscipline in man- The true dichotomy is between a narrower, aging RFIs and for high loan losses over short-term approach that focuses on immedi- many years. ate agricultural production gains and a broader, To sum up, the poor design of traditional long-term approach that focuses on large and in- terventions has seriously compromised the sustained expansion in rural incomes.'0 Directed efficiency of the interventions themselves and credit may achieve the former, but efficient ru- hindered the promotion of rural financial ral financial markets are essential for the latter. markets. A key criticism of the traditional approach is that by neglecting financial sector imperatives, Success of the Traditional Approach it placed insufficient emphasis on longer-term in Expanding Income and Reducing Poverty rural development objectives. Traditional focus on disbursing agricultural Critics of the traditional approach have often credit had three misplaced emphases: suggested that policymakers focused on pro- 1. Disbursement. In focusing on disburse- duction objectives rather than on financial ment, the traditional approach has empha- sector objectives and that this imbalance sized quantity of lending over quality. should be redressed (Adams, Graham, and However, treating banking institutions Von Pischke 1984; González-Vega 1984; more as disbursement windows rather World Bank 1993). This dichotomy between than demand-driven, full-service institu- financial and real sector objectives is more tions has resulted in loan portfolios apparent than actual since improving the fi- plagued by defaults and high arrears. nancial sector ensures that resources are al- This approach has reduced financial located optimally to achieve the greatest deepening and constrained long-term possible gains in real income and poverty economic growth (Levine 1994).1' Only by reduction. emphasizing the quality of lending can Box 2.7 Focus of the intervention: rural versus agricultural development Best practice in rural financial markets indicates that recent World Bank study in Mexico showed that in efforts should be aimed at increasing access to finan- the rural area surveyed only 37 percent of self-em- cial services in rural areas at large. This objective ployed people were devoted exclusively to agricul- diverges from widely observed policies directed at ture and directly related activities (Chaves and providing credit to agriculture itself. These policies Sanchez 1995). The majority of rural entrepreneurs may have promoted an unbalanced overall develop- (55.6 percent) were dedicated to other nonagricul- ment. The idea behind rural-as opposed to agricul- tural activities, and the remaining 7.5 percent of tural-development is that development should the self-employed combined agriculture activities occur on a regional or location basis and not on a with other nonagricultural businesses. The survey sector or activity basis. There is nothing magical also found that agriculture was not the main about agriculture itself. The assumption seems to source of rural cash income, generating only 32 per- have been that the majority of people who live in cent of the aggregate nonwage income of rural rural areas work in agriculture and that this activity entrepreneurs. thus generates most of rural income. Even if these Targeting credit to agriculture may not even be assumptions were true, the implied strategies may be possible because of the fungibility of credit. Attempts counterproductive because they could cause an un- to prevent borrowers from diverting credit resources balanced development of the rural sector while pro- to other activities will most likely be futile and will moting underdiversification of rural households' impose unnecessary costs on borrowers (avoidance sources of income. costs, for example) and to lenders (supervision costs). In any event, because of recent interest in under- In brief, the policy is unsound and may not be im- standing the entire rural economy, evidence has been plemented at a reasonable cost. found that agriculture may not be as important in rural areas as traditionally thought. For example, a Source: Chaves and Sanchez 1995. 28 Rural Finance the quantity of lending be expanded in a allocation of loans (Cuevas and Graham rapid, efficient, and sustainable manner. 1984). 2. Agriculture. In focusing on agriculture, the traditional approach has neglected sig- Conclusion nificant opportunities for growth and risk diversification (both for RFIs and for rural Even though the traditional approach was not clients) in nonfarm rural enterprises (see entirely without benefits, particularly at the box 2.7 and chapter 4). farm level, it had serious shortcomings. Essen- 3. Credit. In focusing on credit, the tradi- tially, the goals of rural income expansion and tional approach has ignored savings mobi- poverty reduction were not met, and in some lization, the "forgotten half of rural cases interventions may have exacerbated the finance" (Vogel 1984). A1 three successful problems of rural areas and worsened the RFIs analyzed in chapter 9 mobilized plight of rural communities. large amounts of savings, and there is lit- New perspectives on rural finance have tle doubt now that the rural poor can emerged from recent developments in eco- save. RFIs that have a high share of sav- nomic theory and from the analysis of current ings relative to government borrow- best practices in rural financial intermedia- ings in their total liabilities are likely tion. These new perspectives are examined in to be more careful and efficient in their part two. PART Two Policies for Improving Rural Financial Intermediation The new perspective on rural finance that optimizing the efficiency of markets. A sum- emerged during the 1980s has enabled mary of the differences and commonalities of the formulation of broad guidelines on the old and the new approaches appears in rural financial interventions. These guidelines table 3.1. are by no means conclusive; many aspects re- Chapter 3 provides a systematic approach to quire further testing and research. assessing the role of government in pursuing The overriding goals of the new approach- the goals of income expansion and poverty re- income expansion and poverty reduction-do duction. The approach is illustrated by two de- not deviate from those of the traditional ap- cision trees (one for each goal). Chapters 4 and proach. But important differences can be found 5 develop the government's role in creating a in the objectives, strategies, and methodologies favorable policy environment through indirect applied to attain these goals. The new ap- interventions, and chapter 6 proposes forms of proach proposes a more limited, market- direct interventions aimed at promoting im- friendly role for government, with the focus on proved rural financial intermediation. 29 ¡ CHAPTER 3 New Perspectives on Rural Finance T bhis chapter identifies some of the basic Importance of Efficient and Complete Markets empirical analyses and policy decisions needed to develop rural financial mar- Besley (1994) defines a market as efficient kets with the primary goals of (a) increasing "when it is not possible to make someone bet- growth by expanding rural incomes and (b) re- ter off without making someone else worse ducing rural poverty. The policies recom- off." Thus, in efficient markets participants mended to attain each of these goals are undertake all possible exchanges that yield similar. The key differences are in the approach a positive net economic benefit at the mar- to direct government intervention. This chap- gin until there is no scope for further im- ter first considers the approach required to provement.12 Perfectly efficient markets achieve income expansion; it then recommends require participants to be fully informed policies for reducing poverty. and to fully account for all the costs and benefits of their actions in maximizing their Policy Goal I: Expand Rural Income returns. The growth rate of an economy is highly To expand rural incomes, it is essential to ensure dependent on the competitiveness and effi- that markets are efficient and complete (figure ciency of its markets. For example, average 3.1). This report focuses on ways of promoting growth rates are typically 1 to 2 percentage deep and efficient rural financial markets. Gov- points higher for economies with relatively ernments can of course play an important role undistorted foreign exchange markets in promoting nonfinancial markets (for inputs, than for economies with high exchange rate goods, and services), but an analysis of policy premiums (World Bank 1991). Economies options in those markets is beyond the scope of with deep financial markets are likely to this volume. achieve significantly higher long-term If markets are found to be inefficient, meas- growth rates than economies with limited fi- ures may be required to create a more favor- nancial intermediation (figure 3.2). This as- able policy environment, to improve the legal sociation suggests that thefirst objective of any and regulatory framework that supports the program that aims to expand rural incomes must markets, or to address identified market fail- be to enhance the efficiency and completeness of ures in a cost-effective manner. This chapter markets. This objective can be achieved considers the importance of efficient and com- through improved risk diversification and re- plete markets, identifies typical sources of in- source allocation, thus facilitating the inte- efficiency, and proposes measures that can be gration of the rural economy with the rest of taken to improve the operation of markets. the economy. 31 32 Rural Finance Table 3.1 Characteristics of old and new approaches to rural finance Old New Primary goais * Growth and income expansion (frequently pursued by intro- * Growth and income expansion. ducing modem technologies with concessional credit). * Poverty reduction. * Poverty reduction. Worcing assumptions * Accelerated economic development requires controlled com- * Accelerated economic development requires enhanced modity and financial markets (such as control of food prices competition in goods and finandal markets (through flexible and interest rates). prices, for example). * Smail farmers and rural entrepreneurs cannot pay commercial * Small farmers and rural entrepreneurs can pay commercial, interest rates. market rates of interest. * Small farmers and rural entrepreneurs cannot save. o Small farmers and rural entrepreneurs can and want to save. * Access to concessíonal credit is essential to growth and * Access to nonsubsidized financial services is essential to poverty reduction. growth and poverty reduction. Role of government * To directly intervene in and control the agricultural sector and * To create a favorable policy environment, while minimizing agrcultural credit. direct intervention in and control of the agricultural sector and agricultural credit. Mechanisms of govemment intervention Policy environment Policy environment Rely on the eight pillars of urban-biased policies: Introduce and maintain a polícy environment to promote rural financial markets: * Maintain an overvalued exchange rate to ensure a cheap * Support macroeconomie stability through market-determined supply of agricultural produce. exchange rates. * Control prices of agricultural products. * Maintain a level playing field among economic subsectors (inciuding the agricultural and rural sectors) and enhance competition. * Levy excessive taxes on agricultural exports. * Deregulate the financial sector and support a competitive environment. * Protect domestic industres whose outputs are used as * Introduce legal, regulatory, and enforcement mechanisms that agricultural inputs. address the specific requirements of the rural population. • Enforce usury laws that (despite their good intentions) harnper * Eliminate urban-biased policies that discourage rural the provision of financial services to the rural poor. development and impede the development of rural financial markets. * Enforce laws and regulations that do not take into account the unique features and requirements of the rural economy. Direct rural financial interventions Dírect rural financial interventions * Introduce a legal ceiling on deposit and on-lending interest * Remove ceilings on deposit and on-lending interest rates; rates. encourage market-determined rates. * Establish state-owned rural financial institutions, primarily * Provide financial services through various RFIs (and not specialized agricultural credit institutions (SACis): formal exclusively through SACIs). financial services to rural communities are mostly provided by these institutions. * Provide financial services primarily to the agricultural sector, * Provide financial services to all rural entrepreneurs (not only to thereby discriminating against rural nonagricultural entrepre- agriculture-related activities). neurs. * Focus on providing (agricultural) credit; savings in monetary * Encourage domestic savings mobilization by providing savings instruments is discouraged in rural areas. facilities with poshive interest rates. New Perspectives on Rural Fínance 33 Old New * Provide special benefits and concessional funds to state- o Revitalize and restructure SACIs and other RFIs to encourage owned RFls; subsidize on-lending interest rates to RFI clien- sound management principles; transform SACIs to RFIs to tele to compensate for urban-biased policies. - serve a broader rural clientele. * Cover loan losses of RFIs and frequently bailout loss-making * Privatize RFIs (or segments of their operations) where institutions. appropriate and shut down inefficient and unsalvageable RFis. * Support poorly administered crop insurance and credit * Support eariy innovators; provide seed capital and auction guarantee schemes. subsidies to new (pilot) or existing credit unions, NGOs, and other RFIs that meet strict eligibility criteria and are capable of providing efficíent rural financial services; partially cover start- up costs (the 'infant industry" approach). * Support institution building; assist in staff training, development of management information systems, research, and dissemination of information about successful institutions or practices in specific socioeconomic and cultural settings. * Extend limited subsidies to RFIs to compensate for distorted policies (now being removed); cap and gradually phase out subsidies. * Introduce adequately priced and well administered crop insurance and review effectiveness of credit-guarantee schemes. Policy variables and outcomes • Underinvestment in rural public infrastructure (for example, * Improved rural infrastructure and educational and health roads and water supply) and in rural human resources (for facilities are necessary. e example, education and health) is acceptable. • Subsidized interest rates are used primarily as a compensa- * Positive real interest rates serve as an allocative mechanism. tory mechanism rather than an allocative one. * Subsidies mostly benefit the agricultural sector, and mainly * Interest rates are sufficiently high to ensure sound spreads well-to-do, influential farmers. between lending and deposit rates. * Nonagricultural rural entrepreneurs have limited access to * Ail rural entrepreneurs have access to financial services. financial services, causing rural development to slow down. * Insufficient provision of savings facilities and artificially low * RFIs' dependence on borrowed funds from donors and deposit interest rates (because of interest rate ceilings) result governments is reduced as domestic savings mobilization in limited savings mobilization. becomes the main source of finance, improving financial self- sustainability. * RFIs depend on rediscounting facilities and donor and budget o RFIs enjoy autonomy in introducing efficient operating methods. funds to back their (subsidized) loan portfolios. * SACIs and other RFIs do not enjoy autonomy; most opera- * No special privileges are extended to state-owned RFIs; a level tional decisions (such as on-lending interest rates, cost of playing field is maintained and competition among RFIs is borrowed funds and staff policies) are dictated. encouraged; access to subsidies (when warranted) is not contingent on an RFI's ownership. * Special privileges are often extended to SACIs, resulting in * Institution building and financial discipline is encouraged dependence on concessional funds, lack of competition, and through management's accountability for RFI performance; no incentives to improve performance; SACIs are often per- poor loan collection is not tolerated. ceived as disbursement windows. * No commercial imperatives exist for (state-owned) RFls; * Improved performance and productivity by RFI staff is management is not accountable for RFI performance; finan- encouraged; staff is motivated by advanced performance cial indiscipline and poor loan collection prevail. incentive systems. * Lack of staff incentives result in poor performance by unmoti- vated staff. * RFI performance is evaluated in terms of traditional financial * RFI performance is evaluated in terms of outreach to its target profitability ratios (return on equity and assets), ignoring the clientele and financial self-sustainability (as measured by the cost of subsidies; the real cost to society of maintaining an subsidy dependence index): these crtera make it possible to RFI is not known. assess the real impact and cost of an RFI. Source: Authors' findings. 34 Rural Finance Figure 3.1 Decision tree tor expanding rural income Goal 1: Expond rural income Enhance the efficiency and completeness of markets má | Evaluate poliy options Promote deep and efficient rural Promote nonfinancial markets financial markets (beyond the scope of this paper) l ls the market efficient? Yes No Maintain exsting set of policies ldetif probable Is there a poor policy environment? Is there a weak legal and regulatory Is there an identifiable market ffamework? failure? (cihapter 3) Yes NoY No v Ves Yes No 4 Can the market failure be removed Ma intain existing set cost-effectively with public of policies ~ ~ NointervenUon? I Yes Create a favorable policy Improve the legal and regulatory Based on cost-benefit ana4lsis selct environment framework direct inteiventions to remove market failure * Ensure macroeconomic stability * Improve land titiing and registrabon * Remove urban-biased policies * Reform the law of secured * Support social intemmediation * Promote broad financial-sector transactions * Provide incentives for innovation reforms * Deregulate lending by non- * Disseminate information (chapter 4) deposit-taking institutions (hape 6) (chapter 5) Source: Authors' findings. New Perspectives on Rural Finance 35 Efficiency of Rural Financial Markets Figure 3.2 Financial depth in 1960 and GDP growth por capita 1rom 1960 to 1989 Rural financial markets are efficient if they enable participants to take advantage of all opportunities GDP per capita (perceno for exchange that yield positive net economic bene- 3.5 - fits. Thus credit markets are efficient when loans cannot be reallocated to make anyone 3.0 better off without making someone else worse off, and a borrower would therefore have no 2.5 - incentive to resell a loan to another borrower and in so doing become a lender (Besley 1994). Efficiency requires that markets be competi- 1.5 - tive, that participants be fully informed, and that there be a full set of markets (for example, 1.0 no externalities). When these requirements are violated, as is generally the case, rural financial 0.5 markets are likely to be inefficient. Because markets are highly integrated, inefficiencies in 0.0 rural markets may also stem from (or be exac- Very srnall SmaII Large Very large erbated by) inefficiencies in other markets. Financial depth (1960) Assessing the efficiency of rural markets. The Source: Levine 1994. efficiency and level of development of rural financial markets can be assessed on the basis to the most productive potential uses, the focus of criteria such as: of these questions is on the efficiency with Are rural clients áble to insure themselves which funds flow through rural financial mar- against variations in their individual in- kets and not on the application of funds. Eco- comes? Can they easily defer consumption nomic studies, such as Udry (1990) and Rashid (save) or advance it (borrow) and so and Townsend (1994) can be used to check for smooth consumption as their individual the symptoms of inefficiency suggested by incomes vary?14 these questions. * Are financial markets highly segmented? Market research and the evaluation of pre- Are rural financial institutions (RFIs) read- vious or current interventions in rural financial ily able to address problems of liquidity markets can also be useful in identifying the surpluses or shortfalls through financial factors that contribute to the efficiency or inef- transactions with other RFIs? ficiency of rural financial markets. Assessing * What are the differences in spread and the circumstances under which interventions on-lending rates for loans to similar clients have been welfare-enhancing could highlight from similar types of RFIs in a given area? the inefficiency addressed by the intervention. * Do spreads reflect the real economic costs Some of this work is currently being spear- (including credit risk) of engaging in finan- headed by the World Bank through its efforts cial intermediation, or merely rent seeking in microfinance and banking with the poor. or market power? The two main initiatives are a comprehensive * Can financial products not currently pro- World Bank study on sustainable banking vided but demanded by rural dients be with the poor, and the establishment of the profitably provided by RFIs? Consultative Group to Assist the Poor by the Because money is fungible and efficient rural World Bank and other donors (boxes 3.1 and financial markets will channel loanable funds 3.2). 36 Rural Finance Box 3.1 Consultative Group to Assist thesignificantweight loss of Gambian farmers the Poorest during the "hungry season" (World Bank 1990). Hettige and Steel (1996) found evidence of The Consultative Group to Assist the Poorest-A significant financial market fragmentation in Micro-Finance Program (CGAP) is the product of Africa. an agreement reached by donor agencies at the 1993 Binswanger and Rosenzweig (1993) found Conference on Hunger. CGAP was launched in that covariant shocks to income are reflected June 1995 by nine donor agencies; its membership has since increased to nineteen. Its objectives are to in reductions in food consumption in poor strengthen donor coordination in mierofinance, dis- households. Even when households use the semninate microfinance "best practices" to policy- available risk-diffusion mechanisms (extended makers and practitioners, improve and increase family networks, remittances from migrants, World Bank activity in this area, and support ino- diversified production, and reliance on equity vative microfinance institutions through a grant fa- cility. The ultimate goal is to expand access to financial services for the poor. unable to smooth their consumption of food. The CGAP secretariat administers a core fund, Poor households alter their investment be- supported by a World Bank contribution of US$30 havior to reduce their income risks by choos- million, from which it disburses grants to micro- ing investments that have relatively low risks finance instítutions that meet eligibility criteria A and low returns. Large farmers are suffi- policy advisory group of practitioners from leadinr microfinance institutions around the world advises ciently well insured through their assets, so- the donor group and the secretariat on strategies cial networks, and access to financial markets for providing support to the poor. The current to avoid such sacrifices. Their agricultural in- chairperson is Muhammad Yunus, managing direc- vestment portfolios are designed to maxi- tor of the Grameen Bank. mize expected profits rather than to mitigate CGAP issues a note series called Focus, and a risk. biannual newsletter on key policy and technical is- sues in microfinance. CGAP plans to conduct pol- Money-lending is ubiquitous in rural finan- icy-level workshops on a regional and national level cial markets and is often associated with ex- on best practices in microfinance. The notes, news- tremely high interest rates. Although it has letter, and other announcements are posted on the been argued that moneylenders may just be CGAP home page on the World Wide Web: covering the costs of their operations (Bottom- www.worldbank.org/htm-tl/cgap/rgap.htm/ ley 1963), it is likely that the information ad- Source: Mohini Malhotra, CGAP. vantages they enjoy allow them to extract monopoly rents (Vírmani 1982). The market imperfection results in loans that are priced Assessing inefficiencies in rural markets. Effi- above the marginal cost to the lender and in cient rural financial markets are rare. Rashid the provision of insufficient credit from an and Townsend (1994) found evidence in Indian overall econonic (as opposed to privately op- villages of índividual consumption variations timal) perspective. associated with idiosyncratic shocks to in- comes. These consumption variations were Causes of Ineificiencies in Rural Markets exacerbated by the absence of credit and insur- ance markets. Rural areas are also affected by When evidence is found of inefficiencies in ru- highly covariant risks of crop yields, prices, ral financial markets, the cause of the ineffi- and incomes. The covariance makes it more ciencies should be identified to enable the difficult to insure against crop risks locally, goverrmment to select the most appropriate pol- and the problem of intertemporal risk diffu- icy to improve market efficiency. The main sion is likely to be severe for farm households. causes of rural market inefficiencies are a poor policy Missing markets for risk sharing at the na- environment, a restrictive legal and regulatory tional level in the Gambia are evidenced by framework, or market failure. New Perspectives on Rural Finance 37 Box 3.2 Sustainable banking with the poor The objective of the World Bank's study of sustain- identified about 750 programs or institutions able banking with the poor is to improve the ability with at least three years of operations and 1,000 of donors, govemments, and practitioners to design clients. Thus far, about 10 percent have re- and implement policies and programs to strengthen sponded to survey questionnaires. Further up- the financial sector and build sustainable financial dates of the study are planned when more institutions that reach the poor. A systematic gather- questionnaires are received. ing of evidence is required to identify and under- . A Sourcebook on sustainable banking with stand the factors that deternine successful financial the poor, as the final product of the study. intermediation with the poor. Although the main The sourcebook, which will be completed in audience is Bank operational staff, the study will 1997, will integrate the findings that emerge reach policymakers, managers of financial institu- from the case studies and distill lessons for tions, govemmental and nongovemmental organiza- program design, implementation, and policy tions involved in providing financial services to the formulation. A "tool kit' section of the source- poor, women, and other uriderserved groups. book will provide a decision tree presenting The study program includes: a step-by-step approach to the design of fi- • More than two dozen field-based case studies nancial services projects. of innovative rural financial institutions in o A monthly seminar series. Asia, Africa, and Latin America that have re- o Working papers and technical papers based on duced the costs and risks of providing financial the case studies, which will precede the com- services to a large number of low-income clients. pletion of the sourcebook. • The preparation of a worldwide inventory of microfinance institutions. A draft report has Source: Bennett, Cuevas, and Yaron 1996. Assessing the policy environmentfor ruralfinan- of financial intermediaries can provide (see cial intermediation. The first potential cause of chapter 5). The main measures required to inefficiencies in rural financial markets is an improve the legal and regulatory framework unfavorable policy environment. The main are: problems associated with inefficient rural finan- * Improvement of land titling and registra- cial markets, as well as recommendations and tion sequencing for implementing reforms, are dis- * Reform laws governing secured trans- cussed in chapter 4. The main measures required actions to establish a positive policy environment are: o Deregulation of lending by non-deposit- * The introduction of macroeconomic stabil- taking institutions. ity and the reduction of macroeconomic In terms of sequencing, legal and regulatory distortions reform is an area in which corrective measures * The removal of urban bias in pricing, trade can be undertaken immediately, even if policy policies, and budgetary allocations measures have yet to be implemented to re- • The promotion of competitive, integrated, move major distortions in the macroeconomy, transrarent, and resilient financial mar- the real goods sector, and the financial sector. kets. Assessing evidence of identifiable market failure. Assessing the legal and regulatoryframeworkfor Although the policy environment and the legal ruralfinance. The second potential cause of inef- and regulatory framework may be favorable, ficiencies in rural financial markets is an unsup- there may still be failures in rural financial mar- portive legal and regulatory framework. An kets. About financial markets Besley (1994) unsupportive framework can lead to problems notes, "A market failure occurs when a com- with enforcing contracts; it can also limit the petitive market fails to bring about an efficient scope of financial services that various types allocation of credit." 38 Rtural Finance Market failures arise in rural financial mar- Box 3.4 Two obstacles to financial trans- kets when RFIs are unable to internalize the actions: moral hazard and adverse selection full costs and benefits of establishing rural banking services. In the presence of externali- Moral hazard arises when one party to a contract ties and public goods, there are likely to be (usually referred to as the principal) cannot observe potential gains from trade that are not captured the actions of the other party to the contract (usually by private exchanges. Even if the government referred to as the agent), and the agent's actions (or lack of action) affect the expected returns to both faces the same information constraints that the parties. For example, a borrower's effort may be private sector faces, it can intervene in ways largely unobservable to a lender; the lender cannot that increase welfare because it is able to inter- readily determine to what extent unfayorable out- nalize social costs and benefits (Greenwald and comes of projects can be attributed to the borrower's Stiglitz 1986). Because financial markets cannot lack of effort by the client, and to what extent to bad operate efficiently without good informatluck. A lender who could observe the borrower p e eo , could relate the terms of the loan contract to the and because there are significant extemality effort put in by the borrower. Because the borrower's and incentive problems associated with infor- effort is not observable at reasonable cost, the best mation (free riding, moral hazard, adverse se- the lender can do is to set the terms of the contract lection), financial markets are by their nature so that they are acceptable to the borrower and en- likely to be characterized by market failures courage behavior that the lender prefers. Doing so likeey box 3.3)usually means putting more risk on the borrower (see box 3.3). (by being less willing to forgive loans when things Direct measurement of the existence and go wrong) than the lender would do if provided extent of market failure is difficult and uncom- with better information. mon.16 Some studies have shown that con- Adverse selection arises when clients' charac- sumption is volatile for given individuals or teristics (rather than their actions) are unobservable. groups. Buí sore volatility may be unavoid- For example, an insurance company may not be able to tell risky clients from safe ones without in- able, given the need to provide effective incen- curring enormous costs. Risky clients have an in- tives to avoid moral hazard (see box 3.4). For centive to pretend to be safe in order to pay lower example, measures aimed at smoothing con- premiums. Because the premiums that the company sumption and production patterns might dis- charges affect the mix of clients in ways that the company cannot readily observe (higher premiums drive out safe clients more rapidly than risky ones), the company cannot price risks optimally by tailor- Box 3.3 How information externalities can ing contracts (premiums and deductibles) to the generate market failures characteristics of its clients. It is widely and wrongly believed that if a given Source: Rees 1985a, 1985b. activity were profitable, someone in the private sec- tor would have done it already. As Besley (1994) notes, "An inefficiency might develop if individuals hang back waiting for others to try things out. The courage farmers from maximizing their pro- slow diffusion of certain agricultural technologies duction. Often, uncovering the degree of mar- has often been attributed to a reluctance to be the ket failure, and thus the scope for improving first user. An obvious role for government interven- welfare, requires considerable research and pi- tion is to subsidize early innovators. Thus experi- lotíng under a wide variety of circumstances. ments in institutional design, such as the Grameen more ageabletai o documsthe Bank iri Bangladesh, might serve as prime candi- A more manageable task is to document the dates for subsidization. Such arguments appear existence of factors that economic theory sug- only to justify subsidizing new ventures, however, gests lead to market failures. These include and subsidies should be phased-out along the barriers to trade in existing markets, including way." imperfect information, imperfect enforcement, Source: Besley 1994. and social barriers; externalities; and market power (see box 3.5). Where these factors pre- New Perspectives on Rural Finance 39 Box 3.5 The problem of social barriers ing information, the benefits accrue to that in- to trade stitution and to others. Because a single institu- tion is not likely to internalize all of the benefits, There are abundant exanples of the effects of social investment in obtaining information is likely to barriers on trade. Akerlof (1984) developed a model be undersupplied. Acquiring information usu- to explain how caste barriers can persist as a result ally involves incurring fixed costs (for an RFI the of sanctions against trade across castes, even cost of establishing a bank branch in a relatively though such trade would lead to gains. The caste sparsely populated rural area), which can lead barriers could be broken if a sufficient number of to imperfectly competitive market structures agents were to generate additional exchanges caus- (Stiglitz 1993) ing the net loss from violating the sanctions to be outweighed by the benefits of exchanges across The heterogeneity of production conditions castes. Violating the sanctions is however unprofit- and the spatial dispersion of producers result able for the initial few agents who do so. in severe information problems. It is often Social barriers take many forms and are often costly and difficult to determine whether farm- reinforced by sanctions such as social ostracism and ers' crops failed because of factors beyond the economic boycott. The most important barriers are farmers' control or because of negligence, lack based on differences in socioeconomic class, caste, ethnicity, and gender. Gender barriers are particu- of skills, or other inadequacies. Asymmetry of larly difficult to overcome because they operate at information between RFIs and clients creates the societal level and within the household. Other severe moral hazard and adverse selection important impediments to trade include language, problems in crop insurance and credit transac- literacy, and numeracy barriers, which aggravate tions (see box 3.4). information barriers. The lack of interaction across social groups adds to the information barriers that hinder effective in- Optionsfor Addressing Market Failure termediation, particularly for poor and female cli- ents whose access to financial services is already If an identifiable market failure exists, the govern- severely constrained by limited acceptable collat- ment should consider interventions that would di- eral. Restricted access is sometimes the result of metsodcnserievntoshawodd- social barriers or legal and regulatory deficiencies. rectly address thefailure and conduct a cost-benefit Chapter 5 surveys these problems and makes rec- analysis to determine (a) whether the marketfailure ommendations for improving restrictive laws and could be dealt with cost-effectively and if so (b) what regulations. the most cost-effective intervention would be. Inter- ventions may include subsidizing the start-up costs of RFIs, or providing innovative financial vail, policy and intervention options derived services in rural areas, rediscounting loans pro- from research and feasibility studies can be vided by private RFIs to target clients, or es- weighed in addressing them. tablishing and operating state-owned RFIs to address the problem of missing markets. Assessing the problem of imperfect information. The interventions should directly address Imperfect information results in less than fully the cause of the market failure. For example, competitive markets because information has the problem of imperfect information can be public good features: nonexcludability (the addressed by implementing mobile banking, very act of a bank lending to a particular dient initiating group lending, or providing financial gives other banks signals about the quality of incentives or technical assistance to RFIs open- that client) and nonrival consumption (if one ing branches in remote areas. These and other depositor learns that a given bank is weak, an- interventions will succeed only if governments other depositor can acquire the same informa- are fully committed to developing rural finan- tion). Because of these features, the acquisition cial markets. Options for direct interventions of information leads to externalities. When a by governments are discussed in detail in particular institution bears the costs of acquir- chapter 6, and initiatives that can be taken by 40 Rural Finance RFIs to overcome market failures are discussed account the probable incentive effects on private in chapters 8 and 9. participants in the market to avoid crowding them out or creating spillover effects for them Assessing the effects of imperfect communication (such as weakened credit discipline resulting of information. Even when there is good local from poorly executed state interventions). Fi- information about the actions or characteristics nally, potential interventions should be evalu- of certain people, that information may not be ated against the benefits and costs of general widely available outside the community. Lim- financial and nonfinancial interventions, such ited access to information can lead to the seg- as public investment in rural infrastructure, mentation of financial markets into "islands of education, and public works. information." Information about rural areas Government interventions are frequently with few infrastructure links to the wider econ- subject to major, costly failures arising from omy is especially likely to be limited. Under incentive problems, poor governance, and such circumstances local creditors with infor- other factors (see chapter 2). A poorly designed mation better than that available to lenders out- intervention usually does more harm than side the community can extract monopoly rents good and needs to be addressed urgently, but from their clients (Virmani 1982). discontinuing the intervention cannot elimi- nate the market failure that triggered the inter- Assessing the viability of intervention. Market vention in the first place. The problem of how failure (and the resultant inefficiency) is not best to address the market failure remains. always amenable to correction. So it is useful to While there is broad consensus that market introduce the concept of constrained efficiency. failures are likely to arise in financial markets, That is, sometimes the source of the market what governments should do in addressing failure is not changeable in an economically these failures remains an area of contention viable way, and consequently intervention can- (see box 3.6). not be justified economically. A cost-benefit When the primary goal of rural financial in- analysis should therefore always be conducted termediation is income expansion, and there is when considering the correction of a market no identifiable market failure (or no known failure. Cost-benefit analyses of rural finan- cost-effective intervention to address an iden- cial interventions should take into account their tified market failure), the govermnent should opportunity costs. Analyses should also take into maintain existing policies. The government Box 3.6 Two perspectives on market failure: the argument for intervention Stiglitz (1993) and Besley (1994) provide two per- cannot be derived ... from economic theory or spectives on market failure and government inter- be justified by an examination of a broad base vention: of experience (Stiglitz 1993). There is a role for the state in financial mar- In summary, there may be good arguments kets; it is a role motivated by pervasive market for intervention, and some may be based on failures. In developing countries, market fail- market failure. But as one unpacks each argu- ures are almost undoubtedly greater than in the ment, the realization grows that, given the cur- more developed countries. While limitations on rent status of empirical evidence on many markets are greater in less developed countries relevant questions, it is impossible to be cate- than in developed countries, so too, many gorical that an intervention in the credit markets would argue, are limitations on government. is justified. Empirical work that can speak to We have argued that government policies can these issues is the next challenge if the theoreti- be designed which are attentive to those lirnita- cal progress on the operation of rural credit mar- tions. What is clear is that a simple ideological kets is to be matched by progress in the policy comrnitment to financial market liberalization sphere (Besley 1994). New Perspectives on Rural Finance 41 should not, however, necessarily do so when This trickle-down approach failed to reduce the overriding goal is poverty reduction. poverty substantially in rural areas. More re- cent approaches have placed greater emphasis Policy Goal II. Reduce Rural Poverty on addressing poverty at its roots in rural ar- eas. In general rapid growth in the rural econ- The second overriding policy goal of most gov- omy is the most promising avenue for reducing erriments is to reduce poverty in rural areas poverty. Expanding rural incomes by promot- (see figure 3.3). During the 1950s and 1960s the ing efficient and complete markets is essential theory of economic dualism suggested that the to addressing the goal of poverty reduction. way to eliminate poverty was to assimilate la- Increasing average incomes in rural areas bor from the backward, rural subsistence sec- may not be sufficient to decrease poverty be- tor into the modern, urban industrial sector. cause the growth in incomes may not be shared Figure 3.3 Decision tree for reducing poverty Goal ll: Reduce rural poverty Conduct poverty assessment to characterze the rural poor and the binding constraints they face. Evaluate policy options 4 Are marketoriented measures sufficient to reduce poverty? Yes 0/ \X No Enhance the efficiency and Based on a cost-effectiveness completeness of markets analysis, select a program of direct (see fnure a2). interventions to reduce the poverfy of the target group. Pursue a comprehensie strategy for Rural finance intervenfions: Alternative interventions: developing markets: * Provide matching grants for * Provide targeted food support. * Promote deep and efficient rural financial community-based funds * Generate employment through pubrlc works markets * Support social internediaUion programs * Provide start-up subsidies for hnovative * Invest in rural infrastructure (chaptu's 3, 4, 5) savings and credit programs * Develop human resources in rural areas * Make available lines of credit at market * Promote nonfinancial markets (beyond rates (chapter 6) the scope of this papar) (chapter 5) 42 Rural Finance by all. Consequently, poverty reduction may to increase the earnings capacity of poor house- call for a special intervention that is a program holds is often more cost-effective than provid- of targeted interventions may be justified, irrespec- ing financial services, particularly where tive of whether (a) the policy environment is sound, infrastructure is underdeveloped. Similarly, (b) market failures can be identified, or (c) cost- providing or upgrading rural primary educa- effective ínterventions can be identified to address tion and health facilities can achieve longer- market failures. term benefits for the rural poor, especially for The potential benefits and costs of targeted the poorest of the poor, by increasing the pro- interventions to reduce rural poverty should be ductivity of labor, the main asset of the poor. evaluated in relation to an intervention's effect Empirical research by Mosley (1995) suggests on the target group. Sometimes an intervention that credit interventions to reduce poverty may produce a benefit for the target group but have failed to raise the incomes of the poorest not result in a net economic benefit for the coun- beneficiaries-those below the poverty line. try as a whole because no cost- effective way of The same interventions have, however, been improving the efficiency of the market could be successful for clients above the poverty line identified. However, measures that address in- (see box 3.7). efficiency in rural financial markets will be par- When the initial objective is to ensure that a ticularly beneficial to those who have the least drought or other calamity will not cause a sharp access to financial services, such as the rural downswing in rural incomes, public works pro- poor. Income expansion and poverty reduction grams are often cost-effective vehicles for re- measures will frequently coincide. ducing poverty. When the programs are properly designed (when they create efficient Policy Optionsfor Poverty Reduction incentives that attract only the target clientele), they also offer a more efficient long-term safety Governments, donors, and nongovernmental net for the ultra-poor than do interventions in organizations can draw on several instruments financial markets, such as targeted microfi- to address poverty effectively in target groups. nance or rural credit programs (see box 3.8).' These instruments include rural financial inter- Just as too high a wage on public work pro- ventions and nonfinancial public interven- grams attracts less needybeneficiaries who often tions, such as labor-intensive public works crowd out poorer ones, subsidized interest rates projects; means- tested food supplements; food on loans designed to reduce poverty are likely subsidies; family planning; rural primary edu- to attract less poor clients and crowd out the cation and health care projects; rural roads, target cientele. The incentive problems, and the electricity, and water projects; and support for leakages associated with targeted benefits, apply low-income housing. to both financial and nonfinancial interventions The rural poor aré not homogeneous, and and should be considered when developing a given instruments will vary in efficiency under program of interventions to reduce poverty. different social and economic conditions. For An analysis to determine the cost-effective- example, credit and savings facilities may help ness of a program must assess the probable poorer clients smooth production and con- reduction in poverty that will occur as a result sumption patterns, but there may be limited of the program, measured by poverty gap or opportunities for clients to use credit and sav- headcount indices per dollar invested in the ings facilities productively unless a broader set program, compared with the reduction in pov- of nonfinancial services and rural infrastruc- erty that other interventions would achieve. ture is also provided. When the benefits are similar, cost-minimiza- Improving rural roads and electricity, or pro- tion must be analyzed. If the nature of the bene- viding matching grants for village-determined fits differs for different interventions, then a investments, such as watershed management, full cost-benefit analysis is required. Such an New Perspectives on Rural Finance 43 Box 3.7 Credit beneficiaries below the poverty line Since the 1980s credit schemes have been advocated shifted by upgrading the financial performance as a way to reduce poverty. Recent findings by of the microfinance institution. The frontier re- Mosley (1996) call the effectiveness of such schemes lates potential impact on clients' incomes to into question. Using data from microfinance institu- their preloan incomes. In this study it was up- tions in seven countries, Mosley compared increases ward sloping and concave. in borrowers' incomes with increases in the incomes It appears that the poorest clients use loans for of a control group. To minimize problems with com- risk-reducing purposes (consumption and low-risk paring similar clients, the control group consisted of production technologies). The study does not explore successful loan applicants who had not yet received whether the risk-reducing benefits of the loans out- their loans. Mosley found that: weighed the costs to the poorest clients of relatively * All loans in the sample increased average bor- lower incomes (access to credit might have come at rower incomes. a cost in terms of income but might have yielded * More profitable RFIs generally had a larger im- more stable incomes than for the control group). Most pact on their clients' incomes. of the less poor clients (those above the poverty line) * Richer clients generally had larger increases in used loans for more risky and productive invest- income than poorer clients. ments, including technological transformation. * Clients below the poverty line were worse off While credit ray be an effective vehicle for boosting than the control group. the incones of the poor, it may be less effective, or even * Clients above the poverty line did better than counterproductive, in helping the poorest of the poor to the control group. raise their living standards. Altemative poverty reduc- * The impact possibility frontier for each micro- tion mechanisms are probably advisable for this finance institution in the study could have been group. Box figure 3.7 Loan impact in relation to borrower income: within-scheme data Poverty line ^ 30 Banco Sol ch 30 _ -s F 20 *1 (3 co 8 Malawi SACA O _l . @ r 2 L]anca Solborrowers O K-REP bonrowers [1 Malawi Mudzi o. -1A BKKUJRKi»nrowers O Malawi SACA bonrowers Fnbnwr -20 .- j 022 100 200 300 400 Boaower incorne as percentage f poverly lin b incorne Note: Only a few specimen data points, together with the regression line for each organization, are indicated. Full data are available from the authors on request. Source: Hulme and Mosley 1996. 44 Rural Finance Box 3.8 Is credit the solution to poverty? Credit schemes almed at reducing poverty have been a Even the better microfinance schemes have had popular response to the perceived lack of credit for poor limited impact on the poorest groups, who are often people and to the perceived inadequacies of indigenous risk-averse and incapable (for reasons of education village-level credit and insurance arrangements. Many or health) of taking advantage of credit for self- different credit schemes have been tried, induding (first employment activities. The poorest may have no ac- generation) agricultural credit banks and (second gen- cess to complementary inputs, mcluding infrastruc- eration) microenterprise finance schemes. ture. Gains tend to be higher (as a proportion of The evidence on effectiveness of these schemes in income as well as absolutely) for the less poor. It is reducing poverty is fragmentary. Ex post evaluations also unclear whether targeting credit is the best way have often been biased. For example, when a means to promote investment by poor people; even the poor test is applied to determine eligibility, some loan re- can save out of labor income to finance self-employ- cipients confirm ex post that their income was indeed ment activities, and they can consume out of credit. below the means test cut-off at the time of the loan, Other (noncredit) direct interventions to relieve pov- but that their income is now far higher. It is conse- erty can be designed to encourage the poor to escape quently difficult to assess what incomes would have poverty by their own means in the longer term. been without the loans. It should not be assumed that Credit schemes are clearly not the only way to foster the investment would have otherwise been impossi- investment by poor people. ble; an approximation to credit and insurance markets From what we now know about the causes of pov- is often possible without any intervention, as a trip to erty, it is hard to maintain that poverty can be solved almost any village will confirm. Assessments of tar- by targeted credit schemes alone. Chronic poverty geting have often ignored the heterogeneity of the does not appear to be due mainly to market failure poor, by assuming, for example that all the rural poor in credit or other markets, but rather to low factor are landless. Estimated rates of return can be seriously productivity and low endowments per person of biased unless the evaluation is carefully designed. nonlabor factors. If such conditions prevail, even per- The evidence suggests that efforts at targeting fect markets may have substantial chronic poverty. credit to the rural poor have had mixed results. Even To the extent that well-designed credit interventions when a means test is applied, the outcome can fall can efficiently improve access to credit by those far short of the targeted goal. For example, a com- among the poor who are liquidity-constrained, the parison of participation by consumption group in the interventions should be welcomed. Sound program credit-based and income means-tested Integrated Ru- design can often enhance an intervention's effect on ral Development Program, in Maharashtra, India, poverty. But this type of intervention seems unlikely shows that the scheme is far less well targeted than to be the main route out of poverty. that state's Employment Guarantee Scheme, which involves neither credit nor a means test. Source: Ravaillon 1996. analysis may be difficult to conduct because active markets. They are also more likely to information on the benefits of an intervention succeed when RFIs can hire staff with a good may be lacking. Solid empirical work (such as knowledge of the community and solid finance an assessment of poverty) is required to deter- and accounting skills. Finally, credit interven- mine what blend of interventions will be most tions are most likely to be cost-effective when cost-effective under various scenarios.15 When directed at entrepreneurial clients, although rural finance is found not to be the most cost- properly designed credit schemes based on effectiveapproach,alternativepovertyreduction savings and mutualist principles can be programs should be adopted and their effective- used effectively for poorer clients (as shown by ness should be observed and evaluated. Grameen Bank). Evidence suggests that, in general, credit in- The selected intervention should be closely terventions are likely to be more cost-effective monitored during implementation so that poli- in areas with relatively high population den- cymakers can judge whether it makes eco- sities, profitable agriculture (irrigated, double- nomic sense to maintain a particular level of cropped lands, for example), and access to intervention. For example, for food stamps New Perspectives on Rural Finance 45 the relevant indicator is how much it costs to rural incomes and significant and irreversible deliver one dollar of benefits to a target bene- reductions in rural poverty. This chapter has ficiary. For financial services the subsidy advocated a market friendly approach that un- dependence index offers a way of assessing the derscores the importance of creating a positive social costs of supporting a rural financial in- environment for promoting rural financial termediary that is not self-sustaining (Yaron markets. It has also called for a judicious choice 1992a). Box 7.3 in chapter 7 illustrates how the of well-targeted interventions. merits of credit and other interventions might When the objective is to expand incomes, the be compared using this index. focus should be on enhancing the efficiency and completeness of markets. In most cases pro- Financial Interventionsfor Poverty Reduction grams for developing rural financial markets should focus on creating the best envirorunent Targeted financial interventions include rural for rural financial intermediation. Analyses of credit programs directed at small-scale farmers rural financial markets (particularly of barriers and fishers and microfinance programs directed to exchange or of degree of volatility in con- at nonfarm rural households, particularly rural sumption patterms owing to income shocks) can women. When rural financial programs are seen point to the existence of market failures and as a cost-effective means of reducing poverty, in- inform policy decisions on the best approach to terventions should be designed to minimize resolving them. Direct interventions through problems with information and incentives. targeted loans or grants or through state-owned Guidelines for the design of these interventions RFIs should be considered only when market are discussed in chapter 6. failures are clearly identified and when the net economic outcome of government interven- Conclusion tions is firmly expected to be positive. When the objective is to reduce poverty, the New perspectives on rural finance emphasize first step should be to conduct poverty assess- the need to build efficient rural financial mar- ments of rural areas with vulnerable groups kets to achieve rapid and lasting increases in and identify some of the key constraints to Box 3.9 Targeting the poor. infrastructure "Although the relationship between infrastructure and "Price subsidies to infrastructure almost always poverty is pivotal, infrastructure is nevertheless a blunt benefit the nonpoor disproportionately. In Bangla- instrument for intervening directly on behalf of the desh subsidies on infrastructure services are poor. Adequate budgetary allocations to particular sec- roughly six times larger for the nonpoor than for tors or to poor regions, removal of price distortions the poor. A study of five Latin American countries which support biases against the poor, and the selection found that water and sewerage subsidies are di- of appropriate standards and design are generally the rected more to richer than to poorer households. most effective ways to ensure that infrastructure real- Even in formerly centrally planned Algeria and izes its potential for fostering labor-intensive growth Hungary, the rich have received more than the and helping the poor to participate in the growth proc- poor in the way of infrastructure service subsidies. ess. Subsidized provision of infrastructure is often pro- When the poor lack access, as is frequently the case, posed as a means of redistributing resources from they do not receive the lifeline rate and typically higher-income households to the poor. Yet its effective- end up paying much higher prices for infrastruc- ness depends on whether subsidies actually reach the ture services or their substitutes. Subsidizing ac- poor, on the administrative costs associated with such cess to public infrastructure services is often more targeting, and on the scope for allocating budgetary useful for the poor than price subsidies." resources to this purpose without sacrificing other so- cially beneficial public expenditures." Source: World Bank 1994a. 46 Rural Finance increasing the well-being of these groups. In programs to reduce poverty should be based addition to improving the efficiency of markets on cost-benefit analysis. to expand incomes in rural areas, a variety of The selected program should be reviewed programs can be used to directly target the regularly. Carefully designed indirect and direct poor. Many programs (food stamps, public goverinent interventions have been successful works, and so on) do not involve rural financial under a wide range of circumstances. The key intermediation. The decision about whether to features of successful indirect and direct inter- use rural finance interventions or alternative ventions are described in chapters 4, 5, and 6. CHAPTER 4 Creating a Favorable Environment R ural financial markets require a favor- side, since it may not be easy to curtail or recall able environment in which to develop. loans to meet depositors' withdrawals without This includes establishing the right jeopardizing borrowers' cash flows and the macroeconomic conditions, removing policy quality of the portfolio as a whole. biases against agriculture and the rural sector, Even macroeconomic changes that appear and establishing an integrated and efficient fi- positive can have adverse effects on RFIs in the nancial market. short run. A sharp decline in the rate of infla- tion can lead to defaults on loans if farmers are Establishing a Favorable Macroeconomic locked into long-term debt at nominal interest Environment rates adjusted to a previously expected rate of inflation (Corden 1989). Indeed, since macroe- There are important, two-way links between conomic stabilization almost invariably leads the macroeconomy and rural financial markets. to high real interest rates, macroeconomic mis- management can severely reduce farm profits Macroeconomic Volatility Hurts and RFIs' collection performance. This chain of Rural Financial Markets events was illustrated in the United States dur- ing the early 1980s when poor macroeconomic Instability in the macroeconomy affects rural policies led to unusually high real interest financial intermediaries directly through rates, which triggered massive farm failures monetary variables, such as real interest rates, (especially for more highly leveraged farms) and indirectly through effects on clients.19 and a crisis in the farm credit system (Draben- Volatility affects both the assets and the liabili- stott and Barkema 1993). ties side of the balance sheet. On the assets side domestic financing of fiscal shocks in thin fi- Persistent Macroeconomic Distortions Hurt nancial markets can lead to credit crunches, Rural Financial Markets especially for rural clients, because profit mar- gins on rural loans are often thin. These credit Exchange rate distortions are particularly dam- crunches will in turn reduce the loan collection aging to rural financial markets. Exchange rate rates of financial intermediaries if anticipated pegs that do not reflect macroeconomic funda- rollovers of farm credit are not forthcoming. mentals distort foreign price signals, causing On the liabilities side uncertainty in the domes- financial markets to channel excessive re- tic economy may lead to capital flight, liquidity sources to inefficient sectors and insufficient crises, and increases in the cost of funds for resources to sectors with a comparative advan- RFIs. These shocks can spill over to the assets tage. The liquidity and solvency of RFIs may be 47 48 Rural Finance undermined if they base credit decisions on clearly in Mexico in the late 1980s, when relative prices that are later realigned signifi- macroeconomic stabilization made transfers cantly. The same is true if a significant share of to the state agricultural banks unsustainable RFIs' liabilities are denominated in foreign cur- (see box 2.6). Containing government out- rency (often true for many state-owned RFIs lays on interest subsidies and other such that benefited from external donor lending). items is warranted for macroeconomic reasons, The risks of supporting agricultural sector pro- particularly because these policy instruments jects in the presence of persistent foreign ex- are usually dysfunctional (they tend to ob- change distortions are illustrated in figure 4.1. struct rather than assist in achieving their Other things being equal, the higher the black objectives). market premium for foreign exchange, the lower the retum on agricultural investments.21 Prudent Fiscal and Monetary Policies Reduce Volatility Poorly Performing Rural Financial Markets Hurt the Macroeconomy Policymakers can pursue measures to reduce macroeconomic volatility and persistent dis- The link between macroeconomic and rural fi- tortions in macroeconomic variables (Khan nancial sector performance goes both ways. and Knight 1982; Corden 1989; and Mills and Crises in rural financial markets can strain gov- Nallari 1992). The key is to pursue prudent errment budgets and thus macroeconomic sta- fiscal and monetary policies to achieve price bility. The principal sources of fiscal burdens stability and to maintain a sound, well-aligned and deterioration are interest subsidies, loan exchange rate policy. Pursuing such policies delinquencies and defaults, and actuarially are warranted even without regard to their unsound crop insurance and loan guarantee positive effect on rural financial intermedia- schemes. These problems were illustrated tion. Until reasonable stability is achieved (until, for example, inflation and foreign Figure 4.1 Foreign exchange premiums and economic rates exchange premiums fall below 30 percent of retum on bank-supported agricultural products and real interest rates fall below 10 percent but remain positive), interventions in rural finan- Ecoiiomic rate of retum (perceno cial markets should generally focus on piloting and providing support for institutional devel- 16 opment-rather than on introducing large- 14 - scale credit programs. Because the links between rural financial markets and macro- 12 - economic stability go in both directions, the 10 _ fiscal impact of rural financial policies, especially with regard to subsidies, should be 8 - minimized. 6 Removing Policy Biases against 4 - Agricultural and Rural Sectors 2- o _ Rural development has been held back in al- Low Medium High most all developing countries by policies that <20% 20-200% >200% favor industry over agriculture and urban ar- Exchange rate overvaluation {premiums) eas over rural ones. The effect of these policies on formal rural financial markets has been Source: Isham and Kaufmann 1995. calamitous. Creating a Favorable Environment 49 Urban-Biased Policies Hurt Rural blend of policies they pursued.3 Indeed, poli- Financial Markets cymakers often regard government controls as the solution to depressed agricultural produc- The performances of financial and real goods tion, not the source of the problem. (Knudsen markets are closely interrelated. Because com- and Nash 1990). Nonetheless, the use of gov- petitive financial markets are guided by price ernment controls has clearly contributed to the signals, distortions in prices for real goods lead misgivings of formal financial intermediaries to the misallocation of resources by financial about extending rural credit and has severely markets. For years most developing countries reduced the flow of resources (in both direc- subjected their rural sectors to heavy taxation tions) between formal financial intermediaries through a blend of direct and indirect policy and rural clients. measures. The eight pillars of urban-biased policies have had a devastating effect on the Subsidized, Directed Credit Does Not Correct profitability of agricultural and nonfarm rural the Adverse Effects of Urban-Biased Policies enterprises (box 4.1). For example, Schiff and Valdés (1992) estimated that in a sample of To mitigate the adverse impact of urban-biased eighteen countries during 1960-84, direct and policies, governments have established sub- indirect interventions depressed domestic ag- sidized credit programs for the agricultural ricultural terms of trade by 30 percent and ef- sector. These programs have often been admin- fected an income transfer out of agriculture istered by state-owned specialized agricultural equivalent to 46 percent of agricultural gross credit institutions. Governments have also sub- domestic product. They found this "plunder- sidized agricultural inputs, such as fertilizers, ing of agriculture" shortsighted, because the pesticides, seeds, water, and electricity. This countries with the highest degree of discrimi- "compensation" for a poor policy environment nation against agriculture had the lowest rates has not worked. of economic growth, and vice versa.22 Schiff and Valdés (1992) found that from Because indirect, rather than direct, interven- 1960 through 1984 income transfers tions accounted for most (74 percent) of the through all input subsidies averaged unfavorable terms of trade effects, policymak- only 2 percent of agricultural GDP for the ers may not have been aware of the full extent eighteen countries in their study. Where- of the damage to agriculture caused by the as indirect interventions resulted in Box 4.1 Eight pillars of urban-biased policies Urban-biased government policies and public invest- 5. Disproportionately high investment in human ment priorities pervade development efforts. This ap- resources in urban over rural areas (health and proach to development is usually linked to the goal education) of rapid industrialization as well as to political pres- 6. Usury laws (that rule out the loans typical in sure for low food prices from a vocal urban popula- rural areas: small, risky, and high-cost loans) tion. Here are the eight pillars: 7. Underdeveloped legal and regulatory provisions 1. Overvalued exchange rates regarding land titling and collateral for typical ru- 2. Low, controlled, and seasonally invariant ral assets (land, crops, and farm implements) rela- prices for agricultural products tive to urban assets (cars, durables, and homes) 3. High effective rates of protection for domestic 8. Excessive taxes on agricultural exports. industry, the outputs of which are used as ag- This approach has stifled agricultural and rural ricultural inputs development in most developing countries for sev- 4. Disproportionately high budgetary allocations eral decades. Eliminating these urban-biased policies for urban over rural infrastructure (roads, elec- and investment priorities is essential to increasing the tricity, and water supply) well-being of most of the world's population. 50 Rural Finance negative transfers of 15 percent of agricul- policy reforms are in place. Institution building tural GDP. for RFIs, which takes longer to implement, can . Policy distortions affect entire agricultural proceed simultaneously with these broader sectors, but compensation is usually cap- sectoral reforms.25 tured by only a few farmers. For example, formal agricultural credit on concessional Establishing Integrated and Efficient terms rarely reaches more than 30 percent Financial Markets of the farming sector. . Wealthy, well-connected farmers have The links in an economy between rural finan- captured most of the subsídies (see cial markets, which are often segmented, and box 2.2); very few subsidies have reached broader financial markets are far from seam- the most needy farmers. less. Nonetheless, the performance of rural fi- * Inefficient state-owned enterprises inter- nancial markets cannot be divorced from that vene in areas that the private sector can of broader financial markets. Governments can handle better (such as input supply and take several measures to strengthen the per- marketing), while investment in essential formance of financial intermediaries at the ru- public goods (such as rural roads, agricul- ral and national levels. tural research and public registries) have been neglected. Problems in Broader Financial Markets Hurt Rural Financial Intermediaries Well-Established Principles Help to Create a Preferred Policy Environment Banks and other intermediaries in urban and for Development rural areas face incentive problems associated with using other people's money. Major moral There are well-established principles for creat- hazard problems for financial intermediaries ing the preferred policy environment for agri- are associated with high leverage ratios, lim- cultural and rural development. These ited downside risk for shareholders, and the principles include establishing a more nearly potential for insider loans and fraudulent prac- neutral trade regime between agricultural ex- tices. Externality problems are associated with portables and importables, realigning overval- depositors free-riding on the monitoring of fi- ued exchange rates, reducing excessive nancial intermediaries, runs on weak banks industrial protection, shifting public invest- spilling over to sound ones, illiquid financial ment priorities toward public goods and rural intermediaries participating in the payments areas, and increasing participation in commu- system, and high-risk financial intermediaries nity development. Taking such measures will drawing depositors away from prudent ones. increase efficiency and greatly enhance the These moral hazard and externality problems value of the principal or only asset of most exist for a wide range of rural and nonrural poor people-their labor.24 financial intermediaries. Governments and Policy reforms will benefit RFIs by improv- their monetary authorities should address such ing the welfare of their clients. However, be- incentive problems. cause policy distortions have not been uniform, Government policies that increase risks or some protected subsectors may become much financial repression hurt RFIs. Governments less profitable when the distortions are re- can reduce the risk of financial crises, but they moved (as has happened with maize in Mex- are often the source of those crises, owing to ico). In making lending decisions, RFIs should poor macroeconomic policies or harmful direct consider the effects of policy reforms on loan interventions in the financial sector.26 The three quality; substantial increases in credit flows to most damaging interventions are excessive re- rural areas should be deferred until major price serve requirements, a large volume of directed Creating a Favorable Environment 51 credit programs, and subsidized interest rates enterprises and sectors to the financial system, or interest rate ceilings. but in doing so, they have taxed the financial Governiments have often established exces- sector more heavily than other sectors at tre- sively high reserve requirements for fiscal rea- mendous cost to future growth (Giovannini sons. In the early 1990s Argentina's reserve and de Melo 1993 and Levine 1994). requirements were higher than 40 percent. Do- Governments have administratively fixed ing so drives up interest rates for borrowers interest rates and established interest ceilings ar-d discourgges deposit mobilization. for vast amounts of bank lending in general Governments have directed banks to lend and for rural credit in particular. This form large shares of their portfolios to unremunera- of intervention has perhaps been the most tive sectors of the economy. In Tunisia and Mo- damaging, because interest rates are critical rocco mandated RFI loans have been used to in the mobilization and allocation of re- subsidize inefficient state-owned agricultural sources. The market determination of interest enterprises. The mandate itself constitutes a rates is an essential ingredient of a competi- subsidy to the targeted beneficiaries because tive financial sector. Government restrictions administratively directing more funds to the on interest rates restrict the levels and types beneficiaries than the market would have allo- of participation by financial intermediaries in cated to them enables them to pay less than the rural financial markets because interest rates opportunity cost of capital for those funds. on directed agricultural credit are usually fixed Governments have used directed credit to below market rates. Restricting interest rates shift the fiscal burden of propping up preferred discourages monetary savings and impedes Box 4.2 Mitigating the risk of banking crises Govemments can address the moral hazard and ex- country, not restricting foreign investors ternality problems that characterize the financial sec- from owning shares in domestic banks, and al- tor by increasing the transparency, accountability, lowing domestic banks to open branches and risk-bearing capacity of financial intermediaries. abroad. Possible measures include the following: . Developing advanced central clearíng systems • Supervising financial intermediaries through an that allow for early detection of payment prob- autonomous, efficient, and motivated banking lerns. Doing so is important because daylight superintendency. borrowing (interbank loans through the clear- * Requiring audited financial statements that ad- ing system) is the single largest segment of so- here to generally accepted accounting principles, phisticated credit markets. especially with regard to internationally ac- . Establishing clear barikruptcy laws and rapidly cepted asset classification and provisioning executable provisions for dealing with failing norms. institutions. In general these provisions should • Establishing clear, simple, rule-based criteria for require replacing management, eliminating entry that reduce the potential for political inter- shareholders' interests, and carving out bad ference and promote competition in the industry. debts. The institutions may be financially re- • Establishing asset diversification requirements structured, recapitalized, or dissolved. that reflect credit and market risk, credit limits . Enforcing prudential regulations through puni- per borrower or per group of related borrowers tive fines and prosecution of banks and manag- (expressed as a proportion of the bank's capi- ers that fail to adhere to appropriate prudential tal), and restrictions on loans to insiders. standards. • Setting capital adequacy ratios that go well be- Taken together, these measures can greatly in- yond the 8 percent Bank for Intemational Set- crease the information and incentives available to tlements standards for more volatile economies. participants in financial markets. * Defining permissible or prohibited activities. • Allowing international diversification of risks Source: Baer and Caprio 1995; Dale 1982; and Polizatto by permitting foreign banks to operate in the 1991. 52 Rural Finance access to formal credit by small-scale entre- could be taken to improve the performance of preneurs. the banking sector and reduce the probability of a banking crisis are listed in box 4.2. Strengthening the Supervísion and Regulation Programs for restructuring or recapitalizing of Financial Intermediaries Promotes fragile institutions should be undertaken only Market Development after adequate prudential and supervisory mechanisms have been put in place. These fi- To promote the development of deep and effi- nancial reforms should be applied at the same cient financial markets, governments should time to all the financial institutions to be in- strengthen the supervision and prudential cluded in the program, including RFIs super- regulation of financial intermediaries, and im- vised by the monetary authority. Until prove the legal and regulatory framework for financial reforms are put in place and applied contracts (see chapter 5). Governments should to RFIs, there is a significant risk of loan losses also support programs of financial liberaliza- associated with formal credit programs in rural tion and improved risk diversification. Such areas. In the interim the focus should be on programs could deregulate interest rates, re- developing institutions and piloting schemes duce reserve requirements to sound levels, and to increase outreach and self-sustainability, relax credit controls. Further measures that rather than on expanded lending. Box 4.3 Two good reasons for market-determined interest rates: equity and efficiency Equity. Directed credit programs face a characteristic resources and higher rates of econonic growth (King dilemma. The dilemma is whether to lend to more and Levine 1993; Jaramillo, Schiantarelli, and Weiss clients with no subsidy or to lend to fewer clients 1994; and McKinnon and Shaw 1975). Other studies with a high per dollar subsidy. lf the goal is to point to a positive relationship between savings and real resolve the inadequate access to formal credit of the interest rates in developing countries (Fry 1988 and rural masses, then (on equity grounds) the program McDonald 1983). The importance of financial inter- should pursue inereased outreach-a choice that re- mediaries in offering and charging positive real in- quires eliminating or minimizing the subsidy per dol- terest rates is clearly indicated in box figure 4.3, which lar lent. compares deposit interest rates with GDP growth Efficiency. Several studies show that liberalized fi- for four groups of countries sampled over the period nancial markets generate a more efficient allocation of 1974-89. Box figure 4.3 Real deposit interest rates and GDP growU per capita, 197449 Real growth rates per capata (percent 2 o- _ 15 0.5 -0.5 -5.00 -2,49 0.16 2.88 Real deposil interest rales (perceno Creating a Favorable Environment 53 Conclusion made. Chile and Mexico have instituted policy measures that have dramatically changed the Building efficient rural financial systems is es- terms of trade for agriculture and created a sential for improving the livelihood of most of more level playing field within the sector (ag- the rural population. Progress toward this goal ricultural employment has been increasing as a has been held back for decades by policies bi- share of total employment in Chile). Several ased against agriculture and rural develop- countries have embarked on wide-ranging fi- ment. There is little hope of developing strong nancial sector reforms and economies have rural financial markets unless the urban-biased achieved greater macroeconomic stability with policy context is changed. Similarly, prudent the support of the International Monetary fiscal and monetary policies are essential to re- Fund and the multilateral development banks duce macroeconomic volatility and increase (Caprio, Atiyas, and Hansen 1994). But extensive the profitability of rural investments. Finally, reforms are still warranted in many economies the depth and viability of rural financial mar- if the overriding goals of expanding incomes kets will depend on conditions in the broader and reducing poverty are to be achieved. financial markets. Governments must avoid fi- Measures to create a favorable policy envi- nancial repression and address the incentive ronment for rural finance are necessary, but and externality problems that can undermine they may not be sufficient in themselves. For- the performance of financial intermediaries. mal rural financial markets require a suppor- The importance of creating the right environ- tive legal and regulatory framnework, and direct ment is increasingly appreciated by policy- interventions may be needed to accelerate the makers, and considerable progress has been building of robust rural financial markets. CHAPTER 5 The Legal and Regulatory Framework for Rural Financial Markets In many countries shortcomings in laws, to impose penalties on the borrower. Secured regulations, and institutions prevent for- lending rests on the collateral that a borrower mal sector institutions from delivering puts forward for a loan. This collateral gives credit to farmers and rural businesses.27 These credence to the borrower's promise to pay and problems also make it difficult for banks to lend gives comfort to the lender should the bor- directly to farmers, for the formal sector to lend rower not pay. Because evaluating collateral is to the informal sector, and for banks and other often cheaper than acquiring detailed knowl- financial institutions to lend to nonbank credi- edge about a lender, requiring borrowers to tors (such as dealers and nongovernmental or- provide collateral permits lenders to offer ganization group lenders), who might be better larger loans at lower interest rates. placed to reach rural borrowers, especially the Problems in laws and regulations, however, poor. Many of these legal and regulatory prob- can impede both systems of lending, resulting lems cause or aggravate the endemic problems in private lenders denying loans or offering of rural credit: information barriers, social bar- only small amounts of credit at very high in- riers, adverse selection, spatial dispersion of cli- terest rates. Placing restrictions on collateral ents, and covariance of returns. can have serious effects on investment. In in- There has been little analysis of these problems dustrial countries such as the United States, despite their major implications for rural finance. about 45 .percent of the private capital stock is This chapter identifies key legal and regulatory residential real estate, 25 percent is nonresi- impediments and ways of resolving them. The dential real estate, and 30 percent is movable proposed solutions can have an enormous im- capital-including livestock, farm equipment, pact on rural financial markets at a modest cost. fertilizer, standing crops, and invento- ries. Even the economically important class of Problems That Limit the Use of Collateral movable capital is only a subset of what U.S. law views more broadly as personal property. Every private lender worries about the bor- Intangible personal property-accounts receiv- rower's prornise to repay. Over the years pri- able and chattel paper, for example-is not vate markets have developed two broad means counted in the capital stock. Nonetheless, this of giving comfort to private lenders: unsecured property is critical in credit sales and non- lending and secured lending. Unsecured lend- bank financing. When real estate and mov- ing rests on the lender's detailed knowledge able property cannot serve as collateral for about the borrower or on the lender's ability loans, property owners cannot use their prop- 54 The Legal and Regulatory Frameworkfor Rural Financial Markets 55 erty to raise funds for other purposes. Those countries have provisions for afuture interest, who own no property will have difficulty buy- consequently, few farmers can get loans ing on credit. against newly planted crops because the out- To offer loans secured by real or personal put does not yet exist. Few countries have property, lenders need laws and institutions provisions for continuation in proceeds that through which claims against property can be would permit the security interest to be created, publicly established or "perfected," maintained as the collateral is transformed. and enforced. The more uncertain and expen- A lender with a security interest in wool in sive this process, the less willing lenders are to a warehouse will lose that security interest lend. Lenders can cover some of these costs when the wool is sold. Sometimes high costs with higher interest rates. However, as risk and prohibit certain transactions. In Uruguay it cost rise, lenders will reduce the amount lent costs 6 percent of the amount of the instru- relative to the value of the collateral. Under the ment to register a pledge; in Russia it costs 3 least favorable circumstances, they will lend no percent. Such registration fees alone, calcu- more when collateral is offered than they lated at an annual rate, will exceed the inter- would without it. The result? Small loans at est rate on short-term loans for storing farm high interest rates. inventory. Creation of Security Interests in Collateral Perfection of Security Interests in Collateral Creating a security interest requires a frame- Perfection, or publicity, is the process by which work that permits both parties to enter into a lenders establish the priority of their security legally enforceable agreement that applies to interest in the collateral. Sometimes a borrower the parties, the collateral, and related circum- lets the lender keep the collateral until the loan stances that have economic importance. For is paid-as with a watch in a pawnshop, grain real estate many problems can arise in creating in an elevator, or securities deposited with a a mortgage-a security interest in real estate. brokerage house. There is little danger that the The owner may not have clear title. Or the borrower will pledge the same merchandise to owner may lack a title recognized in the main another lender-the first lender's possession of body of law in the country. Or the owner the merchandise serves as effective notice to all may have established title to the land by other parties that the borrower has encum- adverse possession-by living in a place for the bered his or her goods. Other lenders will pre- period of time that the law specifies is neces- sume that the circumstances of storage imply sary to establish title. But that owner may not that the first lender has a security interest in have actually established the title and filed it the goods. in a real estate registry. Since the law typically The situation is not so clear when the collat- states that only the owner of the land may eral remains in the hands of the borrower. How mortgage it and that registered title is the can a lender know whether prior encum- only acceptable proof of ownership, the lender brances exist? Prior encumbrances crucially will not be able to create an enforceable mort- affect the value of an asset as collateral. A gage. For personal property creation of security $100,000 house that seems sufficient collateral interests is usually even more difficult and for a $50,000 loan would be far less attractive risky-as well as expensive and cumber- to a lender if it had a first trust of $60,000. One some. solution to this problem is public registration The law can fail to envision instruments to of security interests. However, many practical cover transactions of key economic impor- problems liniit the registration and perfection tance. In some countries, some commodities process in most developing countries (see can serve as collateral and others cannot. Few boxes 5.1 and 5.2). 56 Ruiral Finance Box 5.1 Titling versus registration The govemments of Bolivia and Peru conveyed title gage could be issued. The deeds registered in the through an agrarian reform law that delivered special agrarian reform registries were sufficient to give deeds to land. These deeds were registered in agrar- the possessors some security of tenure because le- ian reform registries. However, they were not regis- gal processes would probably fail to dislodge the tered in the real estate registries specified for that occupiers. But the system would not permit lenders purpose in the civil code. Consequently, when a to use the land as collateral because there was no lender went to register a mortgage against a property, feasible way to enter the mortgage. there would be no record of ownership and no prop- erty against which to register the mortgage. No mort- Source: Fleisig and de la Peña 1996. For real estate lack of a registered title means Once a title is registered, perfecting a mort- that a lender will not have a legally certain place gage still requires a registry that functions and to file the mortgage on the property that the is open to the public. But some public registries borrower offers as collateral, and so cannot be are expensive to use. High filing costs, long sure that the land is not already mortgaged to delays in registration, and fees and con- another lender. Lacking that certainty, the prop- straints on obtaining information all add to erty will have no value as collateral. This prob- the expense. Public operators sometimes re- lem can arise even when the ownership of the strict public access to registries, thus prevent- property is not disputed-as with clear denion- ing private firms from remedying these stration of ownership by means of adverse pos- problems. Since many of these costs are fixed, session or titles granted through governrnient they can represent a high percentage of the actions, such as agrarian reform programs. value of a small plot. Perfecting a mortgage Box 5.2 High-tech land titling When strict requirements for collateral are main- extremely accurate map. The satellite information is tained in developing countries despite the low avail- fed into a computer, after which CARE advisers work ability of collateral among the rural poor, there is an with the community to decide how to divide the land inherent conflict between the lender's need for secu- into individual plots. Areas that are unsuitable for rity and the government's policy of improving rural grazing or growing crops are then measured and di- households' access to credit. Being unable to provide vided. Because the GPS system is computerized, tech- sufficient collateral, rural borrowers may qualify for nicians are able to divide the land quickly, using laser little or no credit. The absence of sufficient collateral beams to measure distances. The computer program is, in turn, a result of their inability to obtain credit can then print a detailed wrtten description of each to enable them to increase their incomes and reduce plot, a tedious job when done by hand. their poverty-a vicious cycle. One way out of this "What we are doing will lead to a real transforma- collateral-based poverty trap is the new Global Posi- tion in the nation's land tenure system," said CARE's tioning System (GPS) for high-tech land titling. Javier Molina, director of the project, which is being The GPS is a network of twenty-four U.S. Defense financed by the U.S. Agency for International De- Department satellites that orbit the earth, enabling users velopment. "We are helping give people title to to plot their exact position anywhere on the globe. New, land in a way that has not been done before." "Now detailed maps are being drawn up using GPS technol- that everyone knows where his land is, everyone is ogy. Over the past six months CARE, a private aid already beginning to work," said Severino Chico, a organization, has used the GPS system for a pilot project Salvadoran community leader. The new GPS sys- aimed at dividing 11,900 acres into individual plots in tem for titling land has already led to investment Salvadoran communities. In its initial phase, the project in the land and will make credit for development shaved years off the time it takes to get land titles. efforts easier to come by. Carrying a radio transmitter that beams signals to the satellites, workers walk around a property, creating an Source: Farah 1996. The Legal and Regulatory Frameworkfor Rural Financial Markets 57 will raise the total borrowing costs faced by in order for the collateral to comfort the lender smalloperators. and demonstrate the good faith of the bor- For personal property procedures for perfection rower. In most World Bank borrowing member may not be well defined. When they are not, countries many problems encumber reposses- those taking security interests cannot be certain sion and sale. of their priority, cannot easily discover In many countries repossession and sale to whether collateral has already been pledged, enforce a mortgage requires court actions that and cannot be certain of the legal standing of can take six months to two years. Homestead their security interests (see box 5.3). In many provisions often make it illegal to repossess countries it is unclear where a pledge should and sell small parcels of land, making them be filed. Some countries limit security interests useless as collateral. in movable property to items for which regis- Slow enforcement procedures can further tries already exist for other purposes-claims diminish the value of movable property as can be registered against automobiles, ships, collateral. Movable property often has a and planes for example, but not against rice lower unit value than real estate. Unlike real mills, grain harvesters, or hay balers. In other estate, movable property usually depreciates countries laws establish priority by the date on in value over time. Enforcement that takes six the instrument as shown in the notarial regis- months to two years can exceed the economic try-but there may be hundreds of scattered life of much inventory-fruits, vegetables, notarial registries, making it impossible to livestock, meat, poultry, standing crops- search them to determine whether the property and most of the economic life of industrial has already been encumbered by pledge, judg- and farm equipment. Under these circum- ment lien, or other security interest. stances, the collateral will provide little com- fort to a potential lender. Enforcement of Security Interests in Collateral All countries have laws that limit the amount that a creditor may claim from a debtor in de- The value of collateral in securing a loan de- fault. Destitution of the borrower rarely serves pends heavily on the certainty and speed of society's interest. However, when carelessly enforcement and the ease of repossession and conceived, homestead and exempt property provi- sale of the collateral. The importance of en- sions can prevent rural smallholders, usually forcement arises not because repossession and the poor, from using their small holdings as sale are common or desired by lenders but be- collateral or purchasing small parcels of land cause contracting parties must believe that re- or small quantities of equipment on credit. possession and sale will be certain and rapid Similar problems arise for rural businesses Box 5.3 Creation of a security interest: gaps in coverage Often property cannot serve as collateral or be sold sizes, can. For similar reasons, wheat in an Argentine on credit because of gaps in the ability to create a silo cannot secure a loan but sugar in a warehouse security interest. For example, in Uruguay the can. pledge is a lawful contract for cattle but not for These problems are fatal to lending because the chicken or financial institutions. In Argentina and lender knows that in the event of default the bor- Bolivia something that does not yet exist cannot be rower can clairn that the underlying contract has no the object of a loan. Consequently, farmers cannot get legal foundation. These legal problems have no valid credit against the eggs from their poultry, the milk basis in public policy. Some may be accidents of draft- from their cattle, or the wine from their grapes. In ing, others may reflect the old notion of collateral Peru a rotating inventory requires redefinition of the property being retained physically by the creditors. loan, so fruit extract in warehouses cannot serve as collateral but fish meal, stored in containers of fixed Source: Fleisig and de la Peña 1996. 58 Rural Finance Box 5.4 Enforcement: canaries, jail, and death As long as there has been lending, there has been and when it bounces, the lender has the borrower concern about enforcing debt contracts. Rome exiled imprisoned for check fraud. In 1993 one-quarter of nonpaying debtors-transported them across the the jail inmates in La Paz were imprisoned for writing Tiber River-and in some cases dismembered them such checks as security loans. and distributed the pieces to their creditors on a pro In the United States and Canada the creditor is rata basis. The Canarios (Canaries) of Costa Rica allowed to seize the collateral of the defaulting bor- dressed in yellow bird suits and followed nonpaying rower without a court order on condition that in do- debtors to publicly hurniliate them. In Bolivia, Chile, ing so the creditor does not "breach the peace." and Thailand creditors get postdated checks to secure loans. lf the borrower cannot pay, the check is cashed Source: Fleisig and de la Peña 1996. Box 5.5 Warehouse receipts: an instrument that facilitates trade and inventory financing Warehouse receipts are widely used in industrial strategic reserves and support prices without countries as collateral for loans. However, they are physically holding stocks. rarely used in developing countries and econornies * Warehouse receipts can be combined with price in transition, mainly because of the lack of appropri- risk management, so that uncertainty is reduced ate legal and institutional environments, low levels for the collateral on agricultural loans and for of awareness, and govemment policies that reduce the value of that collateral. The combination en- private sector incentives to store commodities-for ables banks to discount the warehouse receipt example, fixed crop prices that discourage storage less than they otherwise would, and may re- and later sales. The overall efficiency of markets can duce the cost of lending, since loans would be be greatly enhanced by enabling producers and com- more secure. mercial entities to convert agricultural production To establish a warehouse receipt system, there must into a negotiable instrument that can be traded, sold, be an appropriate legal and institutional environrnent: swapped, used as collateral for export or inventory warehouse receipts must be functionally equivalent to financing, or used for delivery against a derivative stored conmmodities; the rights, liabilities, and duties instrument, such as a futures contract. of each party to a warehouse receipt (producer, bank, There are some ways in which the introduction of warehouse manager) must be clearly defined; ware- a warehouse receipt system can benefit developing house receipts must specify essential terms; ware- countries and economies in transition: house receipts must be freely transferable by delivery * By providing secure collateral to lending institu- and endorsement; and the holder of a warehouse re- tions, warehouse receipts increase the availability ceipt must have priority to receive the stored goods of local and foreign funds for agricultural credit or their fungible equivalent on liquidation or default and reduce the cost of lending. of the warehouse. Other preconditions for creating a • For countries in the process of market liberali- warehouse receipt system include verification and zation, warehouse receipts can provide the ba- physical controls, warehouse certification, consistent sis for the development of a domestic spot grading, property and casualty insurance, and per- market and possibly (at a later stage) for the formance guarantees by the warehouse manager. creation of forward and futures markets. When these conditions are met, warehouse receipt * During liberalization warehouse receipts pro- systems can contribute to increased and more efficient vide an alternative to direct government rural financial intermediation. involvement in physical markets. The govem- ment can use warehouse receipts to manage Source: Varangis, IECCP, World Bank 1996. when exempt property regulations that osten- on credit because small properties are. legally sibly protect the poor have the opposite effect, exempt from seizure. when for example, small businesses are pre- Problems in the framework for secured trans- vented from purchasing tools and equipment actions make it difficult for formal sector lend- The Legal and Regulatory Frameworkfor Rural Financial Markets 59 ers to make loans secured by small properties. tion is important in such transactions. Com- The problems also make it difficult for formal pared to the bank, the store owner knows more sector lenders to provide financing through the about the customers and can more safely select natural conduits of rural credit, such as dealers, good risks. However, in most developing suppliers of farm inputs, and lenders speciahiz- countries problems in the framework for se- ing in unsecured loans, such as solidarity groups. cured transactions prevent banks from creat- ing, perfecting, and enforcing security interests Problems That Limit Unsecured Lending in accounts receivable. These problems limit the access to credit of store owners and choke Unsecured lending can be an important means off a potentially promising source of rural of providing credit to rural borrowers. Exten- credit. The same story, with small variations, sions of credit by dealers in input and personal can be told for all rural credit sellers and non- and solidarity group loans (with more than one bank lenders (see boxes 5.6 and 5.7). guarantor) are used in many areas. However, several features himit the expansion of these Information unsecured lending systems. Information about borrowers is crucial in reduc- Frameworkfor Secured Transactions ing risk in unsecured lending. Because rural borrowers are more dispersed than urban bor- Problems in the framework for secured trans- rowers, only local lenders and credit sellers may actions can limit the access to credit of the know the repayment records of such borrowers. lender or credit seller. For example, once a farm This special knowledge leads to domination of input store has extended credit out of the rural lending by a series of small monopolies store's own capital, the store operator will have that make small loans at high interest rates. a portfolio of "accounts receivable" repre- Competition can break this link, but competi- senting the credit extended. This portfolio can tors will require information about borrowers. stand for a substantial amount of money even though the individual loans may be small. In Credit Reporting Systems an industrial country with a good framework for secured transactions, such a store operator In many developing countries credit reporting could use that portfolio of accounts receivable systems are only now being developed. Banks as security for a loan at the bank. With that have no systems for making credit decisions loan the dealer could then extend more credit based on lending records. Nonbank lenders to customers on an unsecured basis. Informa- and credit sellers cannot have access to credit Box 5.6 Nonbank lenders Nonbank creditors, such as dealers in feed, fertilizer, and fertilizer dealers in Bangladesh have all reported insecticides, and machinery, can channel credit to the that they have been unable to do so. The absence of rural sector. Often they are willing to extend credit a good framework for secured transactions can break without collateral. However, because they lack the any possible link between rural supplier credits and deposit base of banks, they must be able to borrow urban formal sector lending. themselves in order to offer credit. In a well-function- In this way defects in the secured transactions sys- ing system of secured transactions, such credit sellers tem reduce the availability of credit to those who could use their inventories and accounts receivable borrow in small amounts or who cannot provide land to secure loans from the formal sector to extend more as collateral. credit. But herbicide dealers in Bulgaria, equipment dealers in Uruguay and Argentina, and insecticide Source: Fleisig and de la Peña 1996. 60 Rural Finance Box 5.7 Fragmentary reform Reforms of systems of secured transactions are rela- automobile registry functions well and is the place tively new, and there is no example of successful where security interests in cars can be recorded. But change in developing countries. But different efforts in those same countries the same individual cannot work in different countries, showing that, taken to- get a loan for a tractor because the registries do not gether, such reforms should have an impact in devel- allow the filing of security interests in tractors. In Bul- oping countries similar to the effect they have had in garia, where the automobile registry does not file se- industrial countries. curity interests, car loans are unknown except when Creation. Argentine grain dealers cannot get loans guaranteed by the state or when the cars remain ware- against grain inventories, but Argentine sugar deal- housed and in the possession of the lender. Again, ers can get loans against sugar inventories; Peruvian removal of legal limitations can increase the scope for fruit exporters cannot get credit on inventory, but valuable exchanges. Peruvian meal exporters can. Why? Because the laws Enforcement. Banco Agricola Boliviano, now closed of those countries do not easily permit the writing of after a spotty record in collecting its loans, managed contracts secured by rotating inventory. If the inven- collection rates in the nud-90 percentile on tiny loans to tory does not rotate, the goods can serve as collateral. poor farmers on the altiplano. Why? Because under its Lenders in those countries are not reacting to the charter, the bank could repossess and sell, without judi- underlying economic attractiveness of the collateral; cial intervention, the collateral for loans in default. The the failure of the law to provide a useful contract for bank could readily resell repossessed equipment on rotating inventory limits their lending. When these credit in other altiplano villages. Knowing this, farmers limitations are removed, lending can flourish. serviced their loans. Credible enforcement mechanisms Perfection. In Argentina, Uruguay, and Bangladesh, are essential for recovening loans and can help to im- salaried workers who do not own land can get car prove the performance of otherwise weak institutions. loans-they make a down payment and the car itself serves as collateral. The system works because the Source: Fleisig and de la Peña 1996. information. Consequently, borrowers (espe- derground, preventing poor rural borrowers cially the poor) who are not well known out- from getting the benefits of competition and side their areas have no way to publicize a proper regulation of lending practices. strong payment performance and gain access to credit. The expansion of credit bureaus is National Identity Cards limited by lack of public access to registries and lack of demand for information about unse- These are sometimes required by formal sector cured borrowers because credit sellers cannot lenders. However, these cards can be difficult refinance their portfolios and take advantage and expensive to get in many countries, par- of profitable credit sale opportunities. ticularly for people who have not served in the army or who do not have birth certificates. In- Problems That Limit All Lending ability to gain an identity card can critically limit action. Single women heads of house- A variety of legal and regulatory restrictions holds and those born in rural areas are espe- affect the cost and volume of both secured and cially affected by this problem. Governments unsecured lending. These include: must ease access to these cards, or bank regu- lations must be changed to permit banks to Usury Laws accept other reasonable forms of identification. Limits on interest rates are among the oldest Age of Majority forms of economic regulation. Usually these measures do not achieve a public policy pur- It is common in countries with rapidly growing pose. Instead, informal lenders are driven un- populations to find households in which all The Legal and Regulatory Frameworkfor Rural Financial Markets 61 members are younger than twenty-one. Were taking institutions are not regulated by the a member of such a household to sign a con- monetary authorities and where the laws rec- tract, including a loan agreement, it would ognize a broader array of acceptable collateral, not be enforceable in court. Such households only about 50 percent of total credit is provided often deal largely with informal sector lend- by banks. ers. Nonetheless, the age of majority problem There is a clear justification for regulating inhibits the development of linkages between banks and other deposit-taking institutions, informal lenders and the formal sector through particularly if the government explicitly or im- refinancing accounts receivable and chattel plicitly insures their deposits, but the rationale paper. for regulation and supervision is much weaker regarding lenders and deferred payment sell- Written Contracts ers who do not take deposits and who are not insured. Such regulation often serves only to Most laws require that contracts have in- increase the market power and dominance of formed consent-the contracting party must the banks by restricting their competitors. read and affirm agreement by signing. This Since these competitors often have peculiar ad- implicit requirement that loan contracts be in vantages in lending to agriculture, such super- writing makes a lawful contract impossible for visory policies can burden agriculture heavily. the illiterate. Within the banking system, personal (unse- cured) lending may not be well developed. In Supervision and Regulation the absence of norms for such lending, super- of Financial Institutions visors may not have an efficient system for distinguishing sound from unsound personal Supervision and regulation should be firmly loans. The ability to do so can bear heavily on aligned with the legal system. A bank regulator promising group lending programs but permit cannot remedy underlying defects in the legal leeway for unsound lending to favored groups. system by issuing regulations. If the legal proc- ess requires six months to two years to repos- Slow and Uncertain Legal Systems sess and sell collateral, then loans secured by perishables should be treated no differently Slow and uncertain legal systems-to which from unsecured loans. creditors must turn to enforce debt contracts- Supervision and regulation sometimes mili- raise the cost and risk of making loans. Small tate against rural credit in subtle ways. Non- loans and loans secured by movable property deposit-taking financial institutions and are particularly affected. In severe cases, nonfinancial credit-granting institutions are lenders may refuse domestic real estate col- often regulated, or affected by regulations ap- lateral arrangements and accept only offshore plied to banks, in ways that discourage rural guarantees. lending. On occasion the leasing and lending activities of nonfinancial institutions such as Solution Options equipment dealers are also so affected. Some- times special rapid collection procedures are The legal, regulatory, and institutional changes reserved for banks. needed to expand access to credit in rural areas Such regulation obviously limits the devel- are well defined. Many developing countries opment of nonbank formal sector lenders who have experienced more flows of credit to both could channel credit to many rural borrowers urban and rural clients after remedial measures who might not be natural clients of banks. have been undertaken, even on a partial or frag These effects can be substantial. In Bolivia Reform usually requires greater flexibility in about 95 percent of total credit is granted by the private enforcement of contracts. The judi- 62 Rural Finance ciary often acts as the final arbiter but need not lower interest rates and more credit. These ef- be involved in each step of the enforcement fects can be achieved by reducing the risk and procedure. The pattern is found in Canada, the cost involved in lending without resorting to United Kingdom, and the United States. In subsidies. other countries, particularly Germany and the The reforms recommended in this chapter Netherlands, reforms have focused on the permit lenders to lend profitably at lower rates powers and ability of the judiciary and other en- of interest. Figure 5.1 shows a downward shift forcement offices to act quickly. Developing in the credit supply caused by reform, which countries face a broad range of practical options produces lower interest rates and more credit and can implement them when appropriate. in total. Society gains from the reduced cost of supplying the currently demanded credit (ci) Changes in Laws and Regulations and from the gain in total output that arises from using the expanded credit (c1 - c2). The The following list gives examples of changes reductions in interest rates can be substantial required in laws and regulations: (see table 5.1). * Title land and register it in a registry; lower The combination of lower interest rates and the costs of registration and of foreclosure. more credit that results from well-designed • Reform the law of secured transactions; programs to reform the laws of secured trans- perrnit repossession and sale without ex- actions produces gains estimated at several tensive judicial intervention. percentage points of gross domestic product. • Adjust the age of majority to conform to The costs of such legal reform programs would prevalent family-rearing practices. run under US$1 million. Even if the registries • Draft specific, clear, and limited home- were operated publicly and not privately, the stead provisions. cost of the reform would typically run less than • Permit witnesses to give legal standing to US$10 million. The World Bank is supporting contracts signed by illiterates. such reforms with different combinations of • Deregulate the lending activities of non- economic and sector work and lending in deposit-taking institutions. Figure 5.1 lmpact on credit markets of well-designed legal Changes in Institutions and regulatory reform Institutional changes that would aid in ex- Interest rate panding credit in rural areas include: * Reform legal registries and expand the scope for private operation. * Remove barriers to the operation of Reduction incosts credit bureaus and use the ratings of attabutabletoreforms credit bureaus in bank supervision and S2 regulation. r2 -------- ---------- * Where national identity systems are used, make these universal; where they are not used, set the standards for privately sup- plied identification. Benefits and Costs of Refonns Reforming the legal, regulatory, and institu- Credi cl c2 tional systems that govern credit can lead to Source: Contrbuted by Heywood FIeisig. The Legal and Regulatory Frameworkfor Rural Financial Markets 63 Table 5.1 Explaining high interest rates for loans in Argentina (percent) Factor Unifted States Argentina Difference Effect of greater macroeconomic risk in Argentinma Government borrowing rate on 3-month dollar- denominated notes 5.5 11.4 5.9 Effect of higher bank intermediation spreads and greater difticulty collecting against real estate collateral Mortgage interest rate 7.5 18.0 10.5 Estimated impact of higher local bank spreads and collection costs 4.6 (10.5-5.9) Effect of the difference in the framework for secured transactionsb New car loan 9.7 21.0 11.3 Used car loan 14.1 26.0 11.9 Equipment loan 14.1 60.0 45.9 Personal unsecured loan 16.2 60.0 43.8 Estimated impact of the framework for secured transactions New cars 0.8 (11.3-10.5) Equipment and personal loans 33.3 to 35.4 (43.8 or 45.9-10.5) a. There is no risk that the U.S. govemment will be unable to pay its bonds in dollars because it has a legal monopoly on printng dollars. The Govemment of Argentna must get its dollars by raising taxes or cuttng spending. Such actons are of course politicall difficult The difference between the U.S. interest rate on dollar bonds and the Argentinean interest rate on dollar bonds measures the markets' percepton of this macroeconomic risk. b. Differences in macroeconomic risk and intermediation cost apply equally to loans secured by real estate and loans secured by movable property. In the Unted States loans secured by movable propert have interest rates close to he interest rates on mortgages; in Argentina banks do not make loans secured only by movable property, and nonbank lenders charge rates of about 60 percent. Source: Fleisig and de la Peña 1996. Argentina, Bolivia, Bangladesh, Bulgaria, El legal and regulatory framework is particularly Salvador, Honduras, Peru (see box 5.8), and important to reduce risks and build confi- Uruguay. The Inter-American Development dence. Bank has projects under way in Guyana and This chapter has drawn attention to the im- Haiti. The European Bank for Reconstruction portance of collateral as an instrument for se- and Development has projects under way in curing financial transactions, and to the many several countries with economies in transition ways in which governments can facilitate the from controlled to market-driven systems. creation, perfection, and enforcement of secu- rity interests in collateral by improving laws, Conclusion regulations, and institutions. The chapter pointed out the potential for increased private A fundamental obligation of the state is to pro- participation in these areas, and it highlighted vide a secure legal framework that enables its other regulatory and institutional reforms that citizens to engage in exchanges that increase can increase the scope for both secured and their welfare. This framework requires efficient unsecured lending. Such measures include re- and impartial institutions to settle claims moving constraints arising from usury laws among disputing parties. Because of the prom- and from regulations governing nonbank fi- issory nature of financial transactions, a sound nancial institutions. 64 Rural Finance Box 5.8 Land titling and registration: a private initiative in Peru Where does this property end and that one begin? recent reforms, getting a deed involved 207 bureau- And who owns what? Land titling is not the world's cratic steps, divided among 48 government offices; most fascinating topic. Yet bringing secure property on average, it took 43 months and cost 10 weeks' rights to the poor is a central issue in many develop- worth of the official minimum wage. There, however, ing countries. with support from President Alberto Fujimori, Mr. Traditionally, the answers to such questions have de Soto's institute began to simplify and decentralize come from bureaucrats. Or rather, have not come. the process by creating its own computerized register More than four-fifths of the countryside in develop- of land and commercial property. ing countries has no legally recognized owner. Now, This went beyond titling: African experience has a Peruvian, Hernando de Soto, and his Institute of shown that titling alone will not expand credit if links Liberty and Democracy are trying to bring market to the financial sector are missing. So the institute's forces to the task of titling. Peruvian database is designed to be used by lenders Why does titling matter? Governments gain be- and utilities. Easier credit appraisal and the assurance cause they can collect property taxes. Utilities can of more certainty in foreclosure or cutting off serv- more easily identify, and then charge, those who use ices, if need be, mean less risk. The institute's scheme their services at a given address. For ordinary citi- is said to cost a fraction of officialdom's, and still be zens, recognized ownership brings pride, security, faster. The Peruvian government has helped by legal- and credit: in the United States, over half the loans izing 270,000 squatter businesses. to new businesses depend on formal titles as collat- The institute claims that property given deeds eral. The result is more investment and better upkeep under its scheme doubled in value in the plan's for houses, workshops, and farms. Better value too: first three years. Mr. de Soto sees potential partners a recent study in the Philippines found that providing in financial firms: the more titles, the more mort- deeds for property raised its value by a third. gages, and in time, a secondary market in mort- Aid donors have given (or lent) developing coun- gage-backed securities. try govemments several hundred million dollars in Not everyone is convinced. Time and the attitude the past three decades for this cause. This funding of governments will decide. But at least Peru is bring- has produced satellite-derived maps and elaborate ing fresh thinking to the old problem of helping the databases but not many useful registries. The World poor to join the formal economy. Bank admits that nearly all its rural titling schemes have produced poor results. In Indonesia onerous titling regulations add 10-30 Source: "A Matter of Title," The Economist, December 9, percent to the cost of buying land. In Peru, before 1995. Although the costs associated with the im- the reform aspect of building rural financial plementation of legal and regulatory reforms markets and improving the performance of ru- are moderate, the long-terrn benefits are enor- ral financial institutions. mous. Much more emphasis is warranted on CHAPTER 6 Direct Interventions T he new approach to rural financial intér- Targeted Interventions mediation suggests a more limited, market friendly role for government in- Targeting specific market segments is not un- terventions. It specifies strict criteria regarding common in the private sector. But in the public the conditions under which direct public inter- provision of goods, the situation differs in two ventions are warranted. important respects. Budgetary resources used • If the overriding goal is rural income expan- to target specific activities are (ultimately) ob- sion, direct public interventions arejustified tained by imposing taxes on other economic onlyiftheyaddressidentifiablemarketfail- activities. The allocation of public funds, in- ures, and then only if the expected benefits cluding subsidies, is usually not guided by of the interventions outweigh the associ- market forces but by social objectives. Because ated costs of the interventions. public funds can be put to a variety of uses, • If the overriding goal is poverty reduction, the opportunity cost of each intervention the benefits of economic growth may fail should be strictly analyzed. to reach many of the rural poor and inter- Subsidies to rural financial institutions ventions based on prevailing social norms should be measured, and the social costs of may still be justified, even in the absence given interventions should be compared with of market failure (or of an economically the cost of alternative interventions (see chap- justifiable intervention to address such ter 3, and Squire 1995). The impact and per- failure). formance of targeted or subsidized programs Differentiating between indirect and direct should be evaluated regularly against their interventions is often difficult. In this report stated objectives. Chapter 7 presents the poten- the term indirect refers to the policy environ- tial problems associated with rural financial in- ment (macroeconomic, sectoral, and legal and termediation and proposes a methodology for regulatory), and direct refers to interventions assessing its effectiveness. that normally involve the direct application of Rural financial interventions should always public funds through targeted credit, the be guided by the fundamental objective of financing of technical assistance to rural finan- complementing or facilitating the workings of cial institutions, and so forth. Although con- the market. All interventions should aim to sensus has developed among policymakers reduce government involvement over time, regarding optimal indirect interventions (see while increasing the private provision of finan- chapters 4 and 5), the appropriate role for gov- cial services and competition. The unit desas ernments in direct interventions is still much (village units) of the Bank of Indonesia debated. (BRI-UD) were created with government assis- 65 66 Rural Finance tance through the provision of seed capital. Be- easily spread risks geographically, provide ade- cause BRI attained economies of scale and in- quate payments, transfer services across large troduced efficient modes of operation, it has areas, or provide sufficient savings services at become highly profitable and financially self- positive real interest rates (because lenders sustainable. Its success has encouraged other must maintain high equity-to-assets ratios to financial intermediaries to explore profitable manage covariance risk). lines of activity in rural areas, following the Examples of interventions to rectify market BRI-UD methodology. failure resulting from imperfect information are government or donor assistance in the form of Guiding Principles for Public seed capital for establishing rural financial in- Interventions stitutions in remote areas or for introducing mobile banking; support for piloting innova- No single form of intervention is optimal in all tive schemes; and incentives for existing finan- cases. The selection of the most appropriate cial institutions to expand their range of intervention will be determined by the initial services or to extend services to a certain clien- objective of the intervention and should take into tele or specific geographic area. These interven- consideration several related variables. These tions would enable institutions to test methods include the characteristics of the target clien- that could systematically reduce the cost of ob- tele, the physical and socioeconomic environ- taining information about a specific clientele, ment, the country's or community's culture, making it feasible to extend financial services and the nature and efficiency of the prevailing to them. In Indonesia the government pro- informal financial market (see box 6.1). vided seed capital for the formation of BRl's The intervention could draw on several village units on the clear understanding that kinds of institutions, forms of assistance (such any unit failing to become independent of sub- as seed capital or training), products or serv- sidies would be closed. ices (such as savings or insurance), and modes Social barriers to successful financial inter- of operation (such as group lending or mobile mediation could be overcome through simi- banking). Table 6.1 summarizes some of the lar interventions. Examples include financial main variables and policy options and pro- incentives to institutions that service disen- vides a framework for the discussion in this franchised segments of the market (women chapter. or the poor, for example) or cost-sharing to cover the initial expense of researching and Objectives of the Intervention piloting effective methods for servicing the target clientele. Although social intermedia- Direct interventions should be implemented tion could help to further overcome these either to address specific market failures or to barriers, its cost can seldom be fully covered reduce poverty (see chapter 3). The effective- by an institution through its interest income. ness of the intervention should always be Social intermediation can take several forms, measured against the objective. including advocacy, training, and group- based approaches. Group-based lending or Marketfailure. Informal providers cater to sev- saving can also help to reduce the transaction eral aspects of the demand for financial services costs of financial institutions (Bennett and in rural communities. However, these provid- Goldberg 1993). ers all rely on community-based information to screen poor risks, and the costs of obtaining this Poverty reduction. In designing interventions information rise rapidly with geographical dis- to reduce rural poverty, the causes of poverty tance. Partly as a result, rural financial markets and obstacles to income expansion should be are almost always segmented. They cannot carefully analyzed, and the intervention should Direct Interventions 67 Box 6.1 The importance of empirical evidence in justifying and designing intervention A recent World Bank study demonstrated that rural government's new objective is to increase the avail- financial markets are very shallow in three important ability in rural areas of viable, competitively priced, regions of Mexico, where rural entrepreneurs have only and untargeted deposit and credit services from for- limited access to financial services. The markets are also mal financial intermediaries. This effort, if successful, highly segmented because the types of borrowers and will increase the outreach and sustainability of formal lenders are so closely matched that funds cannot flow financial intermediaries in rural areas. It will also in- across regions or groups of individuals. These markets crease overall access to rural financial markets and are either noncompetitive or highly inefficient. High change the composition of that market niche by serv- interest rates imply that there are abnormal returns or ing entrepreneurs whom the formal sector did not that the ex ante risk premiums charged by some lend- previously serve. ers are socially inefficient because of the lenders' inabil- The government is attempting to create an environ- ity to diversify geographically. ment in which property rights, contracts, and financial Because of this poor performance the rural econ- services can prosper. Such an environment requires omy in Mexico will continue to have difficulty ad- an improved legal framework for enforcing contracts justing to the major policy reforms of recent years and making use of collateral, as well as networks for and to the aftermath of the exchange rate crisis of sharing information between lenders (for example, early 1995. Rural entrepreneurs will have to adjust credit checks). In particular, it is necessary to (a) have factor proportions, modify output mixes, change a more efficient legal framework for lending secured their scale of operations, and invest in new technolo- by real estate, equipment, inventory receivables, and gies. Their success will depend on the performance consumer goods; (b) make the system of public regis- of all factor markets, particularly financial markets. tries efficient; and (c) introduce mechanisms for expe- Segmentation, however, will force rural investors to ditious recovery of collateral. rely mainly on limited local resources. Therefore, any The government will also fund-with World Bank local negative income shocks or low initial endow- assistance-experiments in developing sustainable ments of resources will have long-term negative ef- technological packages for delivering financial serv- fects on wealth accumulation and hence on poverty. ices to small rural entrepreneurs. The objective is to Extreme interest rates-up to an average of 36 per- establish general guidelines for providing financial cent per month in real terms for credit in kind-and services to medium-size and small entrepreneurs in peculiar collateral requirements have negative dis- rural areas based on examples of successful informal tributional consequences. The evidence also suggests lenders in the regions surveyed and in other coun- that traditionally disadvantaged groups in the rural tries. The guidelines will recommend sensible busi- population may be trapped in low risk-low return ness practices that can be applied to the provision of investment strategies that in the long run will prob- financial services in rural areas. That is to say, the ably widen the income distribution gap. Lack of ac- services offered and the technologies used must be cess to rural financial markets is contributing to this appropriate, and the people involved must act on a state of affairs. proper set of incentives within a conducive policy The newly available evidence seems to have framework. helped the govemment arrive at a better under- standing of the relevant issues and to formulate Source: Contributed by Rodrigo Chaves and Susana strategies for dealing with them. For example, the Sanchez. be directed toward overcoming these problems. overcome the problem of insufficient collateral. For example, financial institutions often refrain Credit guarantee funds have also helped to en- from lending to the poor because small loan and courage institutions to lend to high-risk clients deposit amourits increase relative transaction with no credit record and insufficient collateral. costs. These costs could be lowered through Sound financial reporting and transparency in subsidies aimed at sharing a portion of the disclosing the costs of such interventions are transaction costs or through dissemination of essential for choosing the most efficient inter- information on institutional practices known to vention, whether through financial intermedia- reduce transaction costs. One successful prac- tion or through nonfinancial interventions tice is group-based lending, which helps to aimed at poverty reduction. Table 6.1 Variables and policy options for direct Intementions Viariales Forms of public financiaiíntermnediation Target cientele (demographics, cultural, Physlca/ and Nature of existing Modes of eperation andsockctkl socioeccouic (informal) prcrducts, Dellvery vehicle lnstruments for (InstiutkoHal or ObNectives consideraffc7s) environment schemes, and instltut,ons and institutions targeted interventions Products and services product ievel) Address Income level Population density Moneylenders Nongovernmental Seed capital Credit (working capital Group lending: market organizations, prvate and personal) - Collective failure rural financial institutions - Peer pressure Reduce Occupations in rivestment in Rotating savings and Commercial banks Matching grants Savings Mobile or village banking poverty agriculture or infrastructure credit associations microenteiprises Rural, nonagricuituraj Agricultural growth trends Nongovemmental Speciallzed agricultural Subsidized transaction Credit guarantees Incentives to staff and enterpises organizations credit hstitutions costs clients Agrcuitural conditions, Commercial banks Rural financial institultons Incentives br innovations Crop hnsurance such as double"cropping or threat of droughts or floods Access to formal Postal servies Insurance companies Piloting Other insurance financiai services Access to informal Cooperatives Information dtssemination financiai seMoes Social intennediation Direct Interventions 69 The poor often have limited skills and mar- savings deposits as a precondition to extending kets, and the profitability of their investments credit. could be improved through skills-based train- ing, including training in effective agricultural Culture. Cultural factors are important to the methods. It remains debatable whether such success or failure of given methodologies. For services should be provided by institutions spe- example, joint liability groups may be more cializing in financial services. The govermnent successful in some cultures than in others (see or donors could, however, work in association box 8.9). Practices prevalent in the informal with financial institutions or nongoverrumental market, such as the use of rotating savings and organizations to provide these services. credit associations (ROSCAs) can provide a good indication of the types of products and Target Clientele delivery mechanisms that may work in a given community. There may be several ways to reach a given target clientele. Some factors to consider in Physical and Socioeconomic Environment designing rural financial intermediation pro- jects are the income-generating activities of the The most appropriate form of direct interven- target clientele, their income levels, and their tion can be determined by the physical and culture. socioeconomic environment of the target clien- tele. Factors to be considered include level of Income-generating activities. Clients involved education (of the population and potential in agricultural activities may require longer- staff), population density, the quality of the term loans than those involved in microenter- rural infrastructure, growth trends in agricul- prise activities. The potential impact of ture, and climatic conditions. seasonality and natural disaster on incomes should also be taken into account, for example, Well-educated, low-cost staff. The availability of by introducing crop or livestock insurance well-educated, low-cost staff with a good against specific risks or allowing for reschedul- knowledge of the community would improve ing of loan repayments (at full financial cost) the efficiency of an RFI and perhaps lower under unwillful default. transaction costs. Most successful rural finan- Loans are often extended to provide working cial institutions, such as the BRI, the GB, and the capital. Poor clients or those with wide fluctua- Bank for Agriculture and Agricultural Coop- tions in income may at times require loans to eratives (BAAC) in Thailand, invest heavily in meet personal expenses. Short-term revolving ongoing training. The GB selects and trains staff credit or easily accessible savings could meet who have a unique knowledge of local condi- these needs, particularly for clients above the tions. Government assistance to support the in- poverty line. itial training of employees or early innovators in rural finance could facilitate the development Income levels. Credit alone is unlikely to draw of institutional capacity. the very poor out of poverty because they have limited access to markets, networks, and skills. Population density. Credit interventions are Access to nonfinancial services, such as skills- likely to be more successful in areas with a rela- based training, could help to improve the poor's tively high population density because average return on their investments, just as access to administrative costs are likely to be lower in savings facilities with a positive return would such areas (see chapter 9 for a discussion of suc- start to smooth their consumption. The cessful interventions in Bangladesh, Indonesia, Grameen Bank (GB) in Bangladesh requires and Thailand). 70 Rural Fínance Box 6.2 Rural finance in Central and Eastern Europe Rural financial interventions in countries in Central sources of credit to turn to when the formal banking and Eastern Europe (CEE) have been characterized system collapsed under the weight of bad loans. Ru- by potentially high social return on the one hand and ral finance in CEE faces the following challenges: high risk on the other. The high social return refers to * Low profitability of agriculture. Rural finance can- the strong supply response of private enterprises that not expand rural incomes if borrowers do not might have been realized following implementation have profitable investment opportunities. Sec- of reforms in macroeconomic, trade, agricultural, and toral reforms to remove distorted policies that financial policies-but which were constrained by tax the agriculture sector should be given pri- insufficient agricultural credit and lack of reform in ority. Such reforms include elinrinating food rural finance. The high risk stems from poor policies price controls and aligning input-output prices which lead to distorted input-output prices and the to international prices. intensive use of concessional agricultural credit (to * Lack of information. The rapid and substantial compensate for the inadequate profitability of agri- changes in the CEE econonies have made old cultural activities), which have resulted in misalloca- information about creditworthiness irrelevant. tion of resources. New information is difficult to compile. Emerg- ing financial institutions are generally too inex- Challenges perienced to assess the risks associated with rural finance, and there are no credit rating sys- Under central planning state and collective farns de- tems through which to share and obtain credit pended on credit from state banks, which acted as information. Potential measures to improve in- disbursement windows rather than as profit maxi- formation include training farmers to keep re- mizing financial institutions. Frequent loan forgive- cords and develop business plans, providing ness took place, and the agricultural banks were not technical assistance to banks to help them price expected to screen clients for their credit worthiness risks and evaluate agricultural projects, and but rather to channel funds to facilitate the realization promoting credit rating agencies. of central plans. Because the state had filled all credit * Overcoming the high risk. The rapidly changing "needs" in the past there were virtually no informal macroeconomic and policy environments in Agricultural trends and infrastructure. Credit Karnataka, India, is a good example of a sav- would also be more effective in areas with prof- ings-led financial intervention (see box 8.6). Cli- itable agriculture (such as irrigated, double- ents' risk in areas with major climatic changes cropped lands), well-developed infrastructure, could also be reduced through well-designed and access to active markets, because these fac- crop or drought insurance schemes. tors would facilitate a higher (and often more The prevailing economic system and the na- stable) return on investments. ture of existing financial intermediaries are ex- tremely important. The form and sequence of Climaticconditions. Inpoor, semi-arid areas with interventions would differ markedly in a highly volatile climatic conditions and limited sparsely populated and underserved rural infrastructure, a credit program is unlikely to be African community than in a former command the preferred option because it may impose sig- economy in which a central government had nificant financial risks on its clients, who may be previously assumed much of the risk related unable to generate a regular cash flow to service to agricultural activities and in which a well- debts. But financial interventions that support established network of state-owned supply- savings mobilization and offer clients liquid and driven financial institutions was in place. remunerative savings instruments with which to smooth consumption may be warranted. The Informal Finance Market Mysore Resettlement and Development Agency (MYRADA), which operates group-based sav- Rural populations often rely mainly on savings ings and credit schemes in semi-arid parts of and credit services from informal sources: trad- Direct Interventions 71 CEE countries foster high levels of uncertainty that lend to the agricultural sector at market that adversely affect financial intermediation by rates without state budget subsidies. To promote making potential creditors and producers reluc- rural financial intermediation through a competi- tant to engage in financial intermediation. To tive banking system, measures are required to reduce risk, macroeconomic stabilization must strengthen the banking system as a whole and to be achieved and mechanisms must be devel- address the multiple policy, and legal and regu- oped to allow reduced risk through hedging latory constraints to rural financial intermnediation instruments, such as adequately priced crop in- in the former command economies. surance and guarantees. * Credit union model. This approach relies on local, Poorframeworkfor secured lending. Secured lend- member-owned institutions that are locally di- ing is facilitated when clients are allowed to versified or nationally federated and thus able own property and hold titles to their assets. To to overcome seasonality and high covariance improve secured lending, cadastre and land problems by pooling and sharing risk. registration efforts should begin as soon as pos- * Stop-gap models. Interim solutions to prevent sible, collateral laws for real and personal prop- unwarranted decline in agricultural produc- erty should be updated, and registries and tion include the use of government-sponsored procedures for out-of-court settlements should nonbanking institutions to channel funds on be established. In the absence of adequate col- a temporary basis to the agricultural sector at lateral, social capital in the form of joint and market rates and according to commercial im- several liability may be created by supporting peratives while the financial system is being the voluntary formation of small, homogeneous rebuilt. groups of borrowers. * Sovereign risk guarantee model. Government- sponsored sovereign risk facilities can offset the Strategies sovereign risk associated with policy setbacks that negatively affect credit contracts for the sup- The World Bank financial intermediary loans to CEE ply of agricultural inputs from intemational countries are of four types: suppliers. - Competitive banking model. This approach calls for a network of financially sound commercial banks Source: Tuck and Yaron 1996. ers, landlords, moneylenders, money keepers, of operation and may help to identify innova- savings and loan associations, and rotating tive modes of operation applied by the infor- savings and credit associations (ROSCAs). Sig- mal market to overcome problems inherent to nificant use of informal mutual insurance has rural financial markets (see chapter 8). Encour- also been observed in rural communities (Plat- aging private or state-owned financial institu- teau and Abraham 1987). tions to provide group-based savings facilities These informal schemes provide many nec- at positive interest rates for ROSCAs or for essary financial services, but they are often un- members of funeral savings and insurance derdeveloped. For example, savings usually funds is another way of complementing the earn no interest-in some places depositors informal market. pay to have their savings safeguarded. Simi- larly, markets are often segmented and Kinds of Institutions isolated, which may allow usurious money- lenders to extract monopoly rents. The infor- Financial sector policies have generally mal market has limited ability to diversify been tailored to urban banks, brokerages, covariant risk or to deal with seasonality and and other financial institutions that serve pri- natural catastrophe. marily larger corporate clients and urban de- Governments can facilitate the functioning positors.28 In rural areas governments have of these markets (see box 6.4). Researching the focused on specialized agricultural credit insti- informal market can provide important data tutions and government-run cooperatives. Nei- about the needs of clients and accepted norms ther set of institutions has been attuned to the 72 Rural Finance Box 6.3 Village credit funds in Albania Background. The dismantling of Albania's agricultural (25 percent), trade and services (6 percent), and crafts cooperatíves in 1992 resulted in the creation of (4 percent). The interest rate, which was initially 380,000 farms with an average size of 1.4 hectares. pegged to the U.S. dollar because of an inflation rate Albania regressed to a subsistence economy. The of 200 percent in 1992, has been increased from 6 banking sector was in disarray and unable to provide percent to 10 percent in local currency. Inflation was the new, risk-averse, small-scale farmers with credit. estimated at 8 percent in 1995. A credit system was started with the participation of Follow-up. The objectives of the follow-up Rural De- the farmers; it was developed and tested using a velopment Project which began in 1995, are to create leam-by-doing approach. The farmers refused to use 720 VCFs in the next five years, integrate savings with solidarity group methodologies, which reminded credit services once a cooperative law is passed that them of the former top-down cooperatives, but they provides a legal structure for VCFs, transform the wanted to restore village identity, which had been ADF's credit department into an apex institution sub- destroyed during the*communist period. ject to the banking law, and move the new rural fi- Start-up. Village credit funds (VCF), managed by nancial institution toward financial self-sustainability. elected village credit comnmittees (VCC), were set up Lessons. Four main lessons can be drawn from the under a Rural Poverty Alleviation Pilot Project fi- Albardan experience, which is especially relevant for nanced by the International Development Associa- post-socialist countries. First, the strength and dura- tion to provide financial intermediation services at bility of local tradition must be considered. In Alba- the village level. The Albanian Development Fund nia local traditions survived about 50 years of a (ADF), an autonomous government foundation, is centrally planned command economy. Second, the responsible for implementing the project, which has active participation of villagers and regular dialogue a rural works component. The ADF has recruited and with the village credit committees are essential for trained credit officers, who participate in VCC deci- success. Third, by decentralizing decisionmaking and sions and extend the VCF network. promoting entrepreneurship, credit becomes a pow- According to the basic rules of the VCFs, if a bor- erful tool for promoting a market economy and fos- rower fails to repay a loan, the line of credit to the tering democracy. Fourth, using a multipurpose borrower's entire village may be suspended. Only a project unit with a soft budget constraint makes it part of a village's population can receive loans at a more difficult to achieve the goal of financial sustain- given time; those who expect to obtain credit in the ability. At the beginning of the pilot project it was not future exert strong social pressure on borrowers to known whether the VCFs had the potential to be- repay on time. come sound financial institutions. Now that their Performance. Ninety-three VCFs had been created potential is known, the autonomy of the credit pro- by December 1995, and more than 4,300 loans had gram-and the imposition of a hard budget con- been disbursed, with an average loan size of US$350 straint-is an urgent necessity. and a repayment rate of nearly 100 percent. The loans have been used for livestock (65 percent), agriculture Source: Nowak 1996. Box 6.4 Financial systems that work for the majority In April 1995 Women's World Banking (WWB) organ- * Allowing microfinance institutions to operate ized the Global Policy Forum on Financial Systems That as recognized financial institutions, possibly Work for the Majority (WWB 1995). Fifty leaders of under separate supervisory and regulatory ar- finance ministries, central banks, financial institutions, rangements and intemational development agencies agreed on key * Encouraging lending institutions that meet pru- principles and practices for promoting financial institu- dential standards to mobilize savings and other tions that serve the poor. These practices include: domestic resources * Encouraging a range of institutions with differ- * Permitting financial intermediaries that serve ent legal structures to provide sound financial the poor to rely on competition rather than gov- services to the poor emment fiat to contain their on-lending interest * Keeping entry thresholds low and developing rates to clients. simple supervisory and reporting requirements for financial intermediaries that serve the poor Source: WWB 1995. Direct Interventions 73 financial needs of the vast majority of rural agricultural credit institutions. Finding the inhabitants. most appropriate ways of encouraging the ef- fective provision of rural financial services by NGOs and private formal and informal institu- commercial banks requires further research. tions. Public support for rural financial interme- diation does not necessarily imply public State-owned rural financial intermediaries. The provision of credit. NGOs or locally sponsored state may decide to revitalize specialized agri- credit unions are frequently more adept at sup- cultural credit institutions, convert them to porting grassroots community development RFIs, or privatize them. Governments may also than central or even local governments. NGOs have to establish new RFIs. However, no pref- often establish ties with the community that erences should be given to state-owned RFIs reduce information barriers and discourage ir- that would not be granted to other (new) appli- regularities. Governments can support the pro- cants interested in servicing the same target vision of credit through these institutions by clientele-competition with state-owned rural providing grants to cover transaction costs re- financial institutions should not only be wel- lated to these agencies' start-up activities or by comed but encouraged. assisting in training and disseminating infor- The special treatment often accorded to mation (see box 6.5). state-owned rural financial institutions, which Governments can provide incentives to com- were run as disbursement windows rather mercial banks to expand their services to rural than as autonomous financial institutions, usu- communities. The banks often have stronger ally led to extremely poor financial perform- financial skills and can reach significantly more ance. The political and fiscal costs of closing a clients than can most NGOs. The involvement state-owned RFI can be significant (large staffs of commercial urban-based banks would help must be laid off, unfunded pension liabilities to mainstream rural finance. However, the po- accrue). In addition, such rural financial insti- tential place of commercial banks in rural tutions often have a sunk cost represented by finance has been largely neglected, often be- a large branch network in otherwise unders- cause of the preference accorded to specialized erved areas that could be put to more efficient Box 6.5 The World Bank and nongovemmental organizations As the World Bank focuses on investing in human technical capacity building. There is particular con- development, it is increasingly recognizing the im- cern that NGOs asked to scale up their operations portance of local and international nongovernmental may lose their innovative quality and grassroots par- organizations (NGOs) in achieving poverty reduction ticipatory base. The challenge for organizations such and environmentally sustainable growth and in effi- as the World Bank and its partners in civil society is ciently and equitably delivering services. NGO in- to develop a strategy for choosing from among the volvement has significantly improved beneficiary wide range of NGOs working in a particular sector participation, and thus helped to ensure that goals and country, those best suited for collaboration. and objectives of projects truly reflect beneficiary The World Bank is committed to achieving a close needs. NGOs reach poor communities and remote and productive working relationship with NGOs. areas where there are few basic resources and little Opportunities for collaboration exist on the opera- infrastructure. They operate at low cost by building tional level (identifying, designing, implementing, on existing resources and transfer technologies devel- and evaluating projects), the analytical level (par- oped elsewhere. ticipating in the Bank's research and economic But there are factors that make it difficult for some and sector work), and the policy level (discussing NGOs to collaborate on Bank-financed operations. and formulating Bank policies on development Many NGO-sponsored activities are too limited and issues). localized to have important regional or national im- pact, and many could benefit from managerial and Source: World Bank 1996a. 74 Rural Finance use. It may be worthwhile to re-engineer rather rural financial institutions must have both than close down the operations of unsuccessful autonomy in operational decisions and ac- RFIs by restructuring incentives and sources of countability for performance. For privately finance to enhance compliance with commer- owned institutions, subsidies should be limited cial imperatives and by introducing a strict to providing seed capital, helping to lower in- program for reaching financial targets and itial transaction costs, and assisting with train- phasing out targeted credit. ing to facilitate building institutional capacity When establishing a new RFI or rehabilitat- and achieving economies of scale and scope. ing a poorly performing one, it is essential to Once RFIs become profitable, they can be pri- use market-oriented interventions that aim to vatized in part or in total. complement and not replace the informal rural The initial rehabilitation and conversion of a financial market. State-owned rural financial specialized agricultural credit institution into institutions should be run according to the a diversified RFI may require substantial in- same sound business principles that guide pri- vestment in institution building (see box 6.6), vate and commercial financial institutions (see including the retraining of staff and upgrading chapter 8). The management of state-owned of management information systems. Other Box 6.6 Rehabilitation of credit unions in Guatemala Before 1988 Guatemalan credit unions had a strong clients using (interest-earning) savings and credit social orientation; their main purpose was to provide facilities. cheap rural credit. The credit unions were financed Several factors contributed to this turnaround: by subsidized external credit and by compulsory, * A change in modes of operation was imple- zero-interest share deposits from members. Because mented through the use of a business plan that loans were issued at below-market interest rates, incorporated institutional development, finan- members were in effect penalized for saving and re- cial stabilization, savings mobilization, and warded for borrowing. The credit unions had serious credit administration. Firm financial targets operational problems (management information sys- were set and an effective management informa- tems were undeveloped, the credit unions carried a tion system was put in place. Interest rates on large volume of nonearning assets, the loan delin- deposits and loans were increased. quency rate was about 20 percent, and loan loss re- * A change of mind was effected to achieve a serves were underestimated by more than 50 change in operations and management. To en- percent). Liquidity reserves were so low (about 3 per- sure long-term sustainability, the changes had cent) that the credit unions could not always honor to be embedded in organizational practices and cash withdrawals from members. culture. To make sure that they would be, the The World Council for Credit Unions (WOCCU), WOCCU emphasized institutional develop- funded by the U.S. Agency for International Devel- ment. An agreement of participation was opment, implemented an institutional develop- signed by all parties to demonstrate their com- ment program for the credit unions between 1987 mitment, and credit unions were required to and 1994. The WOCCU Cooperative Strengthening contribute staff and financial resources during Project worked with the National Credit Union the stabilization process. Federation of Guatemala and twenty of its thirty- * Compliance from the credit unions was ob- nine affiliated members, of which nineteen were tained through positive incentives. Under the in rural areas. project, financial assistance for the stabilization By 1994 the situation had completely reversed: process was provided to the credit unions in the deposits had grown from 24 percent of assets form of non-interest-bearing one-year loans. (1988) to 55 percent (1994), the loan delinquency The loan principal was placed in high-yielding rate had decreased to 8 percent of the loan portfo- Guatemalan investments, and the ínterest lio, loan loss and liquidity reserves were increased earned went to the credit unions to offset non- to adequate levels, and nonearning assets had performing assets. been halved. The result was phenomenal growth in assets, profitability, and service with many more Source: Branch and Richardson 1995. Direct Interventions 75 reforms that are usually required include mer,t. Care should, however, be taken to mini- increasing reliance on voluntary savings, with- mize the distortions inherent in these methods. drawing operating subsidies, initiating aggres- Some guidelines on the design and application sive loan collection campaigns, and developing of targeted credit follow. staff incentives. Ensure that targetedfunding remains the excep- Instrumentsfor Targeted Rural tion. When directed credit targets are binding Financial Interventions on financial intermediaries (when they allo- cate more to particular activities or clientele Governments and donors can help to dissemi- than they would in the absence of the lending nate best practices by supporting training and requirements), an economic subsidy inevita- funding study trips to successful rural financial bly accrues to the target group because it pays institutions in other parts of the world. They can less than the economic opportunity cost of fund initial training costs or support piloting funds. schemes that promote innovation and establish ways of reducing transaction costs through a Phase out government involvement over time. In- learning-by-doing approach. For example, the terventions should aim to reduce government World Bank is now involved in two promising involvement over time, while increasing the pri- pilot rural financial projects in Madagascar and vate provision of financial services and compe- Croatia (see boxes 6.7 and 6.8). tition. When financial assistance is provided to Targeting and explicit subsidies should be private or state-owned rural financial institu- used to address barriers to financial interme- tions, clear goals for outreach and self-sustain- diation and to accelerate institutional develop- ability should be set, as should a sunset clause Box 6.7 Piloting savings and loan associations in Madagascar In 1993 the World Bank approved an International ingly accepted, even in the poorest rural areas. The Development Association Credit of SDR2.7 rillion for pilot project is drawing considerable interest from a pilot operation to help rural groups in Madagascar crop and livestock producers, who have requested develop savings and loans associations that could help to set up savings and loan associations in their eventually be grouped into unions and linked with villages. New schemes will be established in the dairy the formal banking sector. The project has piloted new production areas of the country under the IDA-fi- schemes under the guidance of two experienced in- nanced Livestock Sector Program: more are likely to ternational organizations-the World Council of follow. Credit Unions and Développement International Des- The overall process is being coordinated by the jardins-and is financing the expansion of existing Association for the Development of the savings and schernes supported by foreign nongovemmental or- loans associations movement (ADMMEC), a small ganizations-the International Development and Re- private apex organization that is also managing the search Company and the Foundation for Land IDA credit. ADMMEC is now assessing experience, Development. organizing workshops to produce recommendations About forty savings and loans associations have for improving the regulatory framework for SLAs, been established in less than two years, exceeding ap- and steering a participatory process that should lead praisal estimates. Each has from thirty to fifty mem- to the formulation of a national rural financial strat- bers with average savings of US$20-$30 per person. egy and a five-year action plan. The outcome of this Savings and loans associations committees manage work will be discussed with the government and do- loans and establish interest rates that are invariably nors during the midterm review of the project, pav- positive in real terms. Recovery rates have been high, ing the way for the development of a full-scale and project implementation is highly satisfactory. follow-up operation. Although the concept of savings as a prerequisite for credit activities is new to the country, it is increas- Source: Contributed by Michael Simeon in 1996. 76 Rural Finance Box 6.8 Proposed pilot rural financial project in Croatia Objectives. The proposed project aims to identify the National Bank of Croatia. Funds will be provided at reasons for the lack of rural credit in Croatia, to the central bank's interbank rate. strengthen viable rural financial intermediaries, and 3. Borrower support. Rural business advisory serv- to provide long-term credit resources suitable for ru- ices for farmers and other small-scale entrepreneurs ral investments. If approved, the project will (a) sup- will be established to assist them with evaluating in- port participating commercial banks committed to vestment projects, developing business plans, and serving rural borrowers, (b) strengthen the network presenting loan applications to banks and savings of savings and credit associations as rural financial and credit associations. A farmer association service institutions, (c) assist rural entrepreneurs in seeking will advise farmers and other rural entrepreneurs of access to bank credit, and (d) develop solutions for how to join associations and apply for group loans collateral- and savings-related constraints facing ru- from commercial banks. ral financial intermediation. 4. Studies and evaluation. Studies on rural collateral Components. The project would consist of the fol- will be carried out. These studies will deal with rural lowing four components: real estate, movable property, and mutual credit 1. Commercial banks. Long-term credit funds guarantees. A study for a possible savings and credit (US$12 million) will be given to three Croatian association deposit guarantee facility and a formal commercial banks. These banks will engage in evaluation study of the pilot project will also be rural small-scale lending at market rates to the completed. private sector and, through a learning-by-doing Project implementation. A project coordination unit in approach, find ways of overcoming initial infor- the Ministry of Agriculture and Forestry will be respon- mation constraints and reducing high unit costs sible for all components (except the project coordination in the market segment. In addition, a training unit components, which will be administered by the program will be developed for credit officers of three banks). A unit will be established in the central participating commercial banks dealing with ru- bank to supervise savings and credit associations. Par- ral small-scale loans. The program will aim to ticipating commercial banks and savings and credit as- develop substitutes for the screening and en- sociations must comply with eligibility criteria, which forcement mechanisms used by informal lenders will include managerial and financial soundness and in Croatia and to learn from the experiences of the quality of portfolio. Only creditworthy, private cli- banks in other countries. Project funds to partici- ents will be eligible for subloans, which will be evalu- pating commercial banks would be allocated atedaccordingtothesubprojects'commercialviability, through an auction at which the bank offering to available collateral, and environmental impact. on-lend at the lowest rates (linked to a long-term Expected outcome. The project will provide evidence interest rate index in Croatia) will receive the of the ability of commercial banks and member-based most funds. savings and credit associations to attract rural savings 2. Savings and credit associations. Long-term credit and deliver recoverable rural credit. After the project's funds (US$2 million) will be given to five to seven completion, commercial banks and savings and credit selected savings and credit associations for on-lend- associations will be expected to undertake rural finan- ing to small and rural borrowers at market rates. cial intermediation on commercial principles at their Technical assistance and training will be offered to own risk. savings and credit associations and to the new sav- ings and credit associations supervision unit of the Source: Contributed by Vinod Goel in 1996. for phasing out support completely, when do- parent and be reflected in the financial state- ing so is clearly warranted. ments of institutions. Assess performnance. The performance of tar- Focus on access. If the aim is to expand the geted credit programs should be evaluated provision of financial services to underserved regularly against their stated objectives and clients, then the problem of access must be ad- against other potential interventions. To facili- dressed. Because budget constraints are nearly tate measurement, subsidies should be trans- always an issue, underpricing interest rates in- Direct Interventions 77 variably compromises the access objective and other financial intermediaries. The appropriate should be avoided. degree of targeting requires judicious balancing of cost and incentive factors on a case-by-case Be alert to undesirable incentive effects of target- basis.29 ing. The potential for incentive distortions for target and nontarget clients increases with the Products and Services size of the unit subsidy provided through the targeted program. Both financial institutions The provision of financial services to rural and their clients have an incentive to misrepre- communities, especially poor communities, sent clients' status if doing so gains them access has traditionally focused almost exclusively on to subsidies (Besley and Kanbur 1990). Ad- extending credit. Savings, insurance, and other dressing this problem can involve significant financial products and services were largely administrative costs and invasive bureaucratic ignored. This emphasis on credit was due procedures. Therefore, avoiding subsidized in- partly to the perception that the poor cannot terest rates is essential. Mechanisms such as save. This perception has subsequently been small loans should be designed to encourage proved wrong by many institutions, including self-selection of targeted clients into programs. the GB, the BRI, the BAAC and the MYRADA project in India (see chapters 8 and 9). Create a level playingfield. A level playing field Often, the need for safe, liquid, and remu- can be achieved at the sectoral level by ensuring nerative savings facilities takes precedence competitive access to lines of credit among over the need for credit, because saving im- qualifying institutions that provide targeted ac- proves clients' ability to smooth consumption tivities (Guasch and Glaessner 1992), by apply- with their own resources, and allows them to ing hard budget constraints to state-owned avoid having to carry the burden of debt re- rural financial institutions, and by providing no payments during income downswings. Simi- special privileges, such as coverage of loan larly, the presence of insurance schemes in losses. informal markets shows that there is demand for a broad array of financial services among Focus on providing services to the target clientele, the rural poor. not to the owners of the RFI. If the objective is to Chapter 8 offers some guidelines for provid- provide savings and credit services to a target ing traditional financial products, such as sav- clientele, any reputable RFI capable of doing so ings facilities, on an institutional level. Two could be supported with direct interventions special and much-debated products that de- for institution building. Particular institutions serve further discussion are credit guarantee should not be favored. An impartial approach schemes and crop insurance; these products is needed to enhance competition. are discussed later in this chapter. Evaluate the appropriate degree of targeting on a Modes of Operation case-by-case basis. Determining how narrowly public funds should be targeted is difficult. Tra- The mode of operation most suited to the target ditional approaches have often been too nar- clientele must be considered in selecting the rowly targeted to specific subsectors within products and services provided, the instru- agriculture, exposing specialized agricultural ments used, and the delivery vehicles sup- credit institutions to unnecessarily large covari- ported. For example, it may be cost-effective ant risks. Similarly, donors who have focused for NGOs or state-owned rural financial insti- assistance on one institution have denied op- tutions to provide credit or savings services in portunities for institutional development to remote areas through mobile banking agents 78 Rural Finance rather than through full-scale branch offices. credit guarantees. They cover commercial The National Agricultural Bank of Morocco risks associated with loans to clients who doubled its banking network by opening sea- have insufficient enforceable collateral. Other sonal banking windows in existing local offices major forms of guarantee cover noncommer- of the Ministry of Agriculture. Several rural cial risks such as war, civil disturbance, or financial intermediaries have employed agents breach of contract by host governments (boxes who regularly visit villages by motorcycle or 6.9 and 6.10). on foot to provide financial services. Guarantees are widely used by the private Another issue is whether group-based joint sector. For example, it is common for banks to liability schemes are preferable to loans to in- require cosigners on loans to clients, and loans dividuals. lf clients have better information to cooperatives or other group-based loans about each other's investments than the lender often involve mutual guarantees (see chapter has about them, and the clients can engage in 8). Governments have also sponsored public cooperative behavior, interlinked contracts guarantee schemes to encourage commercial based on mutual guarantees can result in better banks and other financial intermediaries to loan terms for the clients with no reduction in lend to farmers who have viable projects expected income for lenders. This is so because but inadequate collateral or credit histories the co-guarantees can lead to higher effort lev- (see box 6.11). Guarantees allow for greater els and lower default rates on loans. The merits leverage and thus greater potential outreach of group-based lending and other modes of than rediscounts of subloans. It is often ex- operation are discussed in chapter 8. pected that lenders, once induced to lend, ill discover that the target clientele is really not Credit Guarantee Schemes unacceptably risky and that future loans can be made without guarantees. It is assumed that The most important form of guarantees for without the guarantees the supply of credit transactions in rural financial markets are will be suboptimal from a social perspective. Box 6.9 Guaranteeing noncommercial risks: a viable approach for rural finance in Russia? Guarantee schemes that cover noncommercial risks input guarantee facility was proposed because of an are usually designed to reduce political risk and in- acute shortage of short-term working capital loans crease the flow of foreign private investment into an and medium-term equipment loans to achieve these economy. Such schemes can provide incentives for objectives. Although government credit programs foreign companies to engage in increased trade with have been cut back sharply, private banks and input rural farm and nonfarm enterprises. suppliers have yet to fill the gap with credit or de- That is precisely the aim of a proposed World Bank ferred payment transactions. input guarantee facility for the Russian Federation. The facility will provide incentives for clients to The facility is designed to place production inputs into use increased inputs and repay their input supply the hands of Russian farmers, producers, processors, credits in a timely way to retain access to scarce in- and manufacturers when no cash, credit, or collateral puts. The facility will also offer incentives for multi- is available to them to purchase these inputs. The fa- national companies to establish trading relationships cility will be used to induce international and domes- with rural enterprises and incentives for the govern- tic suppliers to provide inputs to clients on deferred ment to avoid triggering payment of guarantees by payment terms by providing the suppliers with a gov- avoiding actions proscribed under the guarantee con- ernment guarantee against certain fornis of interfer- tracts. The guarantee facility will be marketed ag- ence in deferred payment transactions and against gressively among multinational input suppliers, nilitary action, war, and civil disturbance. and a public information campaign will precede the There is an urgent need to restore production, im- launching of the facility in Russia. prove quality, and develop new markets for agricul- tural products within the former Soviet Union. The Source: Contributed by Lynn Engstrand in 1996. Direct Interventions 79 Box 6.10 Moldova pre-export guarantee Outcomes of Guarantee Schemes The World Bank has so far provided loans to fund The results of credit guarantees have not been guarantees and other forms of coverage for only a well documented, but the empirical evidence few projects. The 1995 Moldova Pre-export Guar- on the performance of guarantees that does ex- antee Facility Project has both a govermment guar- ist has been mixed. There is plenty of skepti- antee facility and a third-party guarantor. Moldova established a guarantee facility to cover foreign s supplier credit provided to its exporters against guarantee performance. Most crop insurance specified political risks. In addition, an offshore programs that cover specific insurable risks are bank (the agent bank), acting as a third-party guar- subsidized, and a comprehensive credit guar- antor on behalf of Moldova, issued standby letters of antee with its severe adverse selection and credit to foreign creditors, backstopping the govern- moral hazard problems would probably also ment facility's guarantees. The agent bank cou.d draw down the World Bank loan to fund payments to e need to be subsidized. Skeptics conclude that made under its letters of credit to the foreign creditors. guarantees are simply a form of subsidized credit dressed in new clothes. But many gov- Source: World Bank 1996b. emments, donors, and bankers favor guaran- tees as a way to induce lending to targeted clientele. Advocates in the NGO ACCION In- Mexico and India have devoted consider- ternational claim that guarantee schemes are a able resources to publicly supported credit major reason for the expansion in microenter- guarantee schemes. Mexico's guarantee and prise lending in Latin America. technical assistance fund covers some commer- Guarantee schemes in Mexican and India cial bank loans to agriculture and subsidizes have suffered significant losses; the schemes' transaction costs for commercial bank loans to benefits in terms of additionality cannot read- low-income producers. India's Deposit Insur- ily be discemed.30 The reasons for the poor per- ance and Credit Guarantee Corporation offers formance of most guarantee schemes include credit guarantees for loans by commercial and failure to take into account incentives for par- cooperative banks to targeted borrowers, in- ticipating lenders and clients, inability to cope cluding farmers and small-scale industrial effectively with problems of moral hazard and firms. adverse selection, failure to price guarantees at Box 6.11 A primer on credit guarantees A typical credit guarantee scheme links three agents ally banks, to undertake profitable lending to an un- or participants-a guarantor, a lender, and a bor- derserved clientele (Stearns 1993). The challenge of rower-who strive to maximize their objective func- meeting these objectives can be analyzed by examin- tions through the guarantee contract (Meyer and ing five major issues for each of the three participants Nagarajan 1996). An ideal credit guarantee scheme in credit guarantee schemes: additionality, collateral shares risks among all the agents in an incentive-com- requirements, viability, costs and fees, and learning. patible way so as to increase lending to a targeted If credit guarantee schemes are prudently managed, (rationed) clientele at low cost and in a sustainable lenders will eventually realize that the targeted clien- fashion. A contract is incentive-compatible when no tele is less risky than was believed. Properly devel- other contract gives the participants a higher ex- oped screening and monitoring would reduce the pected utility. It satisfies the objectives of all partici- cost of supporting the scheme. Optirnally, transaction pants and is self-enforcing (Philips, 1988). unit costs would become low enough for the guaran- Credit guarantee schemes generally have two ma- tee scheme to be gradually withdrawn. jor objectives: (a) to improve access to financial senr- ices for a targeted sector by reducing risks and Source: Meyer and Nagarajan 1996; Philips 1988; and transaction costs and (b) to encourage lenders, usu- Stearns 1993. 80 Rural Finance levels that encourage participation while cov- ture, and within agriculture on crops much ering indemnity payments and administrative more so than on livestock.31 The results of these costs, inefficient bureaucracy that reduces con- schemes have often been disappointing. This fidence in the guarantees, and provision of section discusses the lessons from these guarantees to public lenders that do not adhere schemes and identifies principles for crop in- to commercial imperatives so that loan losses surance that can be applied more broadly to have to be covered by the public coffers. rural insurance services. Lessonsfor Reform Insurable and Uninsurable Risk In deciding whether to use government guar- Rural producers face a variety of risks associ- antee schemes, other uses of public funds must ated with input supplies, production, market- be considered. When guarantees are warranted, ing, health, and asset management. These risks the following principles should be followed: can be categorized as insurable or uninsurable. The most insurable risks have to do with events Client selection. Establish how far to extend the having three characteristics: their likelihood of risk frontier and thus the target loss rate-the events is readily quantifiable, the damage they ratio of daims paid for defaults over loans guar- cause is easy to attribute and value; and the anteed (Levitsky and Prasad 1989). The focus insuree can affect neither the probability of should be on clients with sound projects who fail their occurrence nor the damage they cause to meet lenders' collateral requirements rather (there is no moral hazard). Typhoon damage than on risky clients or projects. Assessment of is usually an insurable risk; other forms of crop risk should largely determnine the guarantee fee damage may be uninsurable. Some production and the leverage ratio of the guarantee fund. and asset-management risks and most health risks are insurable, but most market and re- Process issues. Appoint competent and well- source risks and a wide array of production trained staff to investigate requests for guaran- risks are not. tees and to build credibility with financial intermediaries. Such efforts should be balanced Limitations of Private Risk-Coping against the need to contain administrative costs and Risk-Reducing Strategies to keep guarantee fees low (3 percent or less). Farmers reduce risk in advance by diversifying Incentive issues. Settle claims quickly to in- crops, intercropping, planting on dispersed crease lenders' incentives to participate, while plots of land, entering into crop-sharing ar- reserving the right to reopen and severely penal- rangements, and seeking nonfarm sources of ize fraudulent claims. Responsibility for collec- income. Once losses occur, farmers cope by seil- tion efforts should be clearly demarcated, and ing assets, consuming a larger share of output, processes for allocating recoveries between the borrowing, or seeking temporary off-farm em- guarantee scheme and the financial intermedi- ployment. In many rural societies mutual aid ary should be clearly established. Banks should and kin-support systems provide an important bear a portion of the risk on principal and all the safety net for member households (Posner 1981; risk on interest due. This risk will provide an Udry 1990). However, these mechanisms may incentive to pursue recoveries vigorously. significantly reduce average incomes because of forgone returns from specialization without Crop Insurance effectively coping with covariate shocks that af- fect most members of the community. Insurance schemes for production in rural ar- Formal rural financial institutions (lenders eas have focused almost entirely on agricul- and insurers) risk incurring losses as a result Direct Interventions 81 of the risks faced by given clients and because reduce lenders' administrative costs to encour- of the mix of clients that they serve (see box age greater provision of credit to clients who 6.12). To reduce these risks, rural financial in- would otherwise be classed as too risky to stitutions usually diversify across different qualify for loans.33 types of crops, farms, regions, and sectors of To be financially viable in the long term (Z), the economy. They adjust interest rates and insurers' income from premiums (P) must premiums to reflect risk. These practices can cover indemnities (I) plus administrative costs be effective in managing risks, but because (A), plus reserves for the catastrophic losses (C) they increase administrative costs, most banks that occasionally strike agricultural producers focus on large commercial farms rather than (Hazell 1992; Roberts and Dick 1991). The in- on smaller farms or more risky clients. Where surer must satisfy: crop insurance has been offered voluntarily, private insurers such as the Chilean Consorcio z A + I + C 1 Nacional de Seguros, have focused on larger p farmers.32 Public ctop insurers rarely meet this condi- Public Agricultural Insurance tion, and Z ratios are usually well above 100 percent even without reserves for catastrophic The rationale for public agricultural insurance losses. Indeed, loss ratios alone (the ratio of schemes is that they stabilize incomes for farm- indemnity payments to income from premi- ers by pooling risks at a regional or national ums) have often exceeded 100 percent. They level. The schemes also aim to improve recov- were often as high as 519 percent in Bangla- eries for rural financial institutions and to desh and 687 percent in India on a cumulative Box 6.12 Liquidity options for diversifying risk in rural financial markets Governments and formal financial intermediaries ing a crisis, nonagricultural banks could be encour- have applied many methods to complete markets in aged to make an "option" available to rural financial rural areas and to diversify the agricultural covariant institutions. The RFI would be able to opt for seasonal risk of rural financial institutions. Most of these at- liquidity in the financial market. Such an option tempts have had poor or indeterminate cost-benefit would have a cost similar to an insurance fee and results. Among the most notorious failures have been would be based on the probability of the associated crop insurance and credit guarantee schemes. risk. The benefits of such an option-which have yet What altemative financial options are available to to be tried-would be threefold. First, the option rural financial institutions to ensure their long-term would ensure access to funds for individual rural fi- viability and to minimize the impact of covariant risk nancial institutions during a bad year or slow season; on their liquidity? An RFI would be better able to second, it would increase the flow of loanable funds; cope with a liquidity shortage in a bad agricultural and third, it would improve both the economies of year and the impact on long-term liquidity if it met scale and scope of the financial sector by diversifying the following conditions: covariant risk throughout the economy and by reduc- * Maintained a high equity ratio ing the risk premium. * Applied on-lending interest rates high enough To qualify for such an option, an RFI would have to to reflect the related credit risk premiuwr asso- meet all three of the conditions listed above-an ex- ciated with the probability of bad agricultural tremely rare accomplishment in the rural financial years (liquidity shocks) arena. If the RFI appeared insolvent and in search of a * Collected principal and interest payments in an bailout, the cost of the option (even if offered to the REI) efficient manner. might be prohibitive. Unfortunately, not many rural An agricultural crisis may nonetheless prevent an financial mstitutions currently meet the three condi- RFI from obtaining funds when liquidity is needed tions; these RFIs would have to improve their perform- most. To ensure that an RFI has access to funds dur- ance before being considered for such an option. 82 Rural Finance basis during the late 1980s (FAO 1992). Admin- When public insurance schemes are war- istrative expenses are rarely below 30 percent ranted, the following operating principles of premium incomes in public insurance and incentive issues should be taken into con- schemes. In the Philippines administrative ex- sideration: penses amounted to 180 percent of premium incomes during the 1980s (Hazell 1992). Operating principles. Write coverage only on insurable risks (on named perils over which Reasonsfor Disappointing Performance clients have little control and on cash crops). Use available weather records and information The main reasons for the disappointing per- about clients to calculate actuarially based pre- formance of most public insurance schemes are miums, and diversify to the extent possible that (a) they have attempted to provide multi- across regions and insured activities. peril coverage for essentially uninsurable risks; (b) they have caused moral hazard problems Incentive issues. Address incentive problems by among insurers, who have relied on govern- adjusting indemnities for visibly negligent ment bailouts and engaged in poor premium- husbandry practices; by tailoring premiums, setting and loan-adjustment practices; (c) they indemnities, and deductibles to the risk levels have led to moral hazard problems among cli- of individual clients (orto fairly narrowly defined ents, who failed to follow sound husbandry client groups); and by ensuring financial and po- practices because of reduced downside risks; litical autonomy for the insurer (by applying a and (d) they have been undermined by politi- hard budget constraint and avoiding exploita- cal interventions. tion of insurance funds for political patronage). When sound business practices are pursued Lessonsfor Reform by public insurance agencies the outcomes can be highly beneficial (see box 6.13). Given the other uses to which funding for ag- A creative approach to crop insurance that ricultural insurance could be put-uses that seems to overcome the major shortfalls of could increase agricultural productivity and many crop insurance schemes is detailed in reduce risks (through irrigation, watershed box 6.14. management, and so on-there may not always be a rationale for publicly funded in- Conclusion surance schemes, especially for larger commer- cial farmers who could be served by private New perspectives on rural financial intermedia- agencies. For small-scale farmers who produce tion suggest a limited, market-friendly role for cash crops, private, industry-based insurance government. Direct public intervention in rural schemes may meet the most important in- financial markets is warranted only if it is the surance requirements. An example is the pro- most cost-effective way of addressing identifi- vision of insurance by banana producer able market failures or reducing poverty. The associations against wind damage to their expected benefits of the interventions must out- members' crops. Governments can facilitate weigh the associated costs of intervening. the private provision of insurance services by No single forin of intervention is universally removing entry barriers to the industry for do- optimal. The choice of an appropriate interven- mestic and foreign firms, by ensuring transpar- tion through a targeted rural finance program ency and accountability in the firms' or an alternative nonfinancial market pro- operations, and by facilitating reinsurance do- gram will depend on the initial objectives of mestically or abroad. the intervention (addressing market failure or Direct lnterventions 83 Box 6.13 Effective public crop insurance: the example of sugar in Mauritius Sugar is the dominant agricultural product and main Only cyclones, fire, and excess rainfall are covered- earner of foreign exchange in Mauritius. The Mau- the MSIF has resisted covering damage from pests and ritius Sugar Insurance Fund (MSIF), which was set diseases, which are viewed as management problems. up in 1945, covers all growers (about 34,000, from As a public agency the MSIF has been unable to backyard ventures to commercial estates) and all the accumulate the reserves that a private comnmercial sugar mills (19) in the country. Sugar has so domi- venture might have accumulated. However, its goal nated the economy that there have been few alterna- is the accumulation of sufficient reserves to handle tive sources of funds from which to subsidize the back-to-back declared events (or 40 percent of aver- industry. The MSIF has always been run on commer- age annual crop value), and it has earned more than cial principles. sufficient investment income from reserves to cover The MSIF has amassed considerable, accurate in- administrative costs (which are very low at 6.5 per- formation over the years on which to base premiums cent of premium income). As a public agency it has and indemnities. It has a highly representative board enjoyed the right to limit its financial liabiity in the and a small headquarters. Two-thirds of its staff are event of massive claims to reserves outstanding plus field agents who send information back to cartogra- reinsurance indemnities, and it has negotiated for phers and to the MSIF computer. Since 1964 preni- levies on sugar tonnages when major shortfalls were ums, first loss percentage, and indemnity percentages expected. Thus, even though the historic loss ratio have been tied to clients' individual claim histories has been slightly negative (at 120 percent), the MSIF to increase fairness and reduce moral hazard. For has had access to emergency funding from the sugar millers indemnities are based on outcomes for the industry itself without seeking recourse to public growers that supply the mills and adjusted for the subsidies. extractive efficiency of the sugar mill. Indemnities for The key to the MSIF's success has been a commer- growers are related to insurable sugar per hectare, cial orientation to insurance, actuarially fair pricing calculated according to recorded individual yields based on carefully gathered information, and atten- and known area harvested. Each plot is inspected at tion to potential incentive problems in designing all least four times a year, and adjustments are made for elements of the scheme. negligence by growers (for example, a 10 percent re- duction in insured sugar for inadequate pest control). Source: Excerpted from Roberts and Dick 1991. reducing poverty). The selection of an inter- ments (such as matching grants and credit vention will also depend on the characteristics lines), products and services (such as savings, of the target clientele, the physical and socio- credit, and insurance services), and modes of economic environment, the cultural features of operation (such as group-based schemes and the target group, and the existing situation in village banking). rural financial markets. Public support for insurance schemes may be Credit interventions are likely to be more warranted under certain conditions. The insur- effective in areas with relatively high popula- ance could be for loans to clients (guarantees) tion densities, open access to markets, produc- or for the primary risk (crop insurance, for ex- tive agricultural sectors, and entrepreneurial ample). Publicly supported guarantee and in- target groups who are slightly above the pov- surance schemes have often proved ineffective erty line. In less favorable conditions alterna- in the past because of design deficiencies and tive projects, such as public works schemes or weak implementation. The merits of direct sup- savings-driven; group-based financial services, port for insurance and guarantees need to be are likely to be more effective. evaluated through rigorous cost-benefit analy- Interventions can be conducted through a sis in the same way that the merits of credit- wide variety of institutions (such as NGOs and and savings-oriented interventions must be state-owned rural financial institutions), instru- carefully assessed before implementation. 84 Rural Finance Box 6.14 A drought insurance proposal Drought threatens the welfare of many poor people of a drought (the expected premium collected over in the developing world. It ought to be insurable. But ten years minus the 10 percent administration cost). to design effective drought insurance, several objec- The calculation is approximate, because no allowance tives must be met. has been made for expected interest earnngs on ac- First, the insurance must be readily accessible to cumulated premiums held by the insurer, for reinsur- all kinds of households: small and large farmers, lan- ance costs, or for the need to build a financial reserve. dless laborers, shopkeepers, agricultural merchants, Drought insurance tickets could be marketed like processors, and artisans. That means that insurance lottery tickets, and low-income people could sell the contracts cannot be tied solely to crop or livestock tickets on comnission. Unlike standard insurance, production. however, drought insurance would provide all ticket Second, the insurance must be affordable, particu- holders for a given weather station with an indem- larly by poor people, suggesting that administration nity in a drought year but not in nondrought years. costs must be kept very low and that only rare If the scheme were managed by a commercial bank, drought events (say in one or two years out of ten) the indemnities could be issued through local branch can reasonably be insured. offices after being announced in the newspapers, and Third, because drought damage within a region on radio and television. tends to be highly covariate, drought insurance will Because all participants would pay the same pre- be financially viable only if a rnechanism exists to mium and receive the same indemnity, drought in- spread the risk beyond the insured region. In a large surance would avoid moral hazard and adverse country it might be possible to spread risk by insur- selection problems. And because the insurance does ing many regions, particularly if these regions have not require the writing of individual contracts, field low or even negatively correlated rainfall patterns. inspections, or loss assessments, administration costs But it is necessary to establish arrangements for re- could be kept low (perhaps to 2 to 3 percent of the insurance or contingent loans with the govenmment ticket value). These features could make drought in- or with private banking annd insurance. surance an attractive proposition for reinsurance in The drought insurance proposed here would be the intemational market and thus overcome the co- weather-station specific, with rainfall monitored variate risk problem of regional droughts. from satellites. All persons insuring against drought To increase its benefits to rural households, at a specific station would pay the same premium drought insurance should be sold freely to all kinds and receive the same indemnity per dollar of insur- of people, who should be able to purchase insurance ance. Indemnities should be paid whenever the sta- from any insured weather station. Allowing them to tion's cumulative rainfall for some specified period do so would enable them to exploit less than perfectly of the year (an agricultural season, for example) fell correlated drought risks and to tailor insurance port- below an agreed-upon level (70 percent of the aver- folios to better match their own individual risk. Pri- age, for example). Premiums would be calculated ac- vate firms might even offer such portfolio insurance. cording to the probability of a drought occurring, the The emergence of secondary markets should be tol- size of the indemnity to be paid, and administration erated to allow cash-short individuals to obtain their costs. For example, for a station faced with the prob- indemnities earlier in a drought year, perhaps at a ability of one drought in ten years and an insurance market discount. administration cost of 10 percent, a $1 insurance ticket would pay out approximately $9 in the event Source: Hazell 1996. PART THREE An Overview of Rural Financial Institutions P Dart three provides an overview of rural Chapter 9 discusses in detail the perform- financial intermediation on the institu- ance and operating methods of three rural fi- tional level. Chapter 7 considers the nancial institutions that are widely viewed as problems inherent in measuring the success of highly successful: the Bank for Agriculture and rural credit projects and proposes that the per- Agricultural Cooperatives in Thailand, the formance of rural financial institutions be used Bank Rakyat Indonesia-Unit Desa System in as a proxy for assessing the success of rural Indonesia, and the Grameen Bank in Bangla- finance intermediation. A framework is intro- desh. The institutions' target clientele and the duced to measure the performance of rural fi- environments within which they operate differ nancial institutions in terms of outreach and markedly. Nonetheless, certain common fea- financial self-sustainability. tures can be identified in the management Chapter 8. proposes guiding principles for principles and operating methods of all three helping rural financial institutions to improve institutions. These shared characteristics can outreach and self-sustainability. The experi- contribute to the development of useful guide- ences of rural development donors and govern- lines for designing future rural financial inter- ments in rural financial intermediation projects mediation projects. throughout the world are shared in boxes throughout the chapter. 85 CHAPTER 7 Performance Criteria for Rural Financial Intermediation T he decisions of private financial inter- Financial institutions that provide a wide mediaries regarding their services and range of financial services to a broad range target markets are generally guided by of clients in an efficient manner are likely to profit motives. However, for government- contribute to income expansion and poverty supported rural financial institutions (RFIs), reduction. Efficient RFIs should therefore development impact is usually a more impor- achieve the desired development impact. tant consideration than profit. Rural financial Measuring the performance of RFIs can serve programs or institutions are considered suc- as a proxy for assessing development im- cessful if they expand rural incomes and re- pact. Measuring RFI performance has the added duce rural poverty. benefit of havingfewer methodological problems In evaluating the performance of rural finan- than measuring the impact of rural credit cial initiatives, it is essential to determine first schemes. whether they have met their goals of expanding The performance of rural financial institutions income and reducing poverty, and then to can be measured by the extent of their outreach evaluate their opportunity cost. A concerted ef- and self-sustainability. Measuring performance by fort by governments or other agencies to pro- these criteria has two important benefits. First, it mote rural finance usually includes directed enables policymakers to better evaluate the cost credit, which, to the extent that it changes the to society of supporting rural financial institu- allocation of resources, involves a subsidy and tions; second, it establishes key performance entails an economic opportunity cost (Ray benchmarks for rural financial intermediation. 1995). Given these characteristics, it is reason- These benchmarks can assist in the formulation able to require an estimate of the total cost to of policies and modes of operation for future society to facilitate the optimal allocation of initiatives (World Bank 1994e). The shortfalls in scarce public resources both within and be- the capacity of rural financial institutions are tween rural financial markets and other sectors often cited as major constraints for successful ru- of the economy. ral financial intermediation (Women's World Measuring the impact of rural credit projects, Banking 1995). programs, or institutions on rural incomes and The first part of this chapter reviews the poverty levels is fraught with methodological methodological problems associated with problems. However, a logical framework has measuring the impact of rural credit projects, been developed to assess the performance of and the second part develops a framework to rural financial institutions (World Bank 1992). assess rural financial institutions. 87 88 Rural Finance Methodological Problems in Assessing Fungibility of money. Attributing benefits to the Impact of Rural Credit Projects credit projects is complicated by the question of fungibility. Subloan resources can be pooled This section reviews assessments of the impact with other resources at a client's disposal, and of agricultural credit projects, in line with the these combined resources can be used for mul- traditional methodology of evaluating and at- tiple purposes. It is hard to discern whether tributing benefits to agricultural projects.3 The loans result in additional resources for the pur- complex problems associated with this meth- poses specified under projects, and it is equally odology can be categorized as microeconomic hard to establish the purposes for which loans and macroeconomic problems. are applied. The terms substitution and diversion are often Microeconomic Problems used to clarify the problems that fungibility poses for evaluating the impact of credit pro- Microeconomic problems consist of (a) the as- jects (Von Pischke and Adams 1980). Since sumption of a representative farm model, (b) farmers who obtain project loans may have the limited comparability between project par- been able to mobilize other resources even ticipants and nonparticipants, and (c) the without the project, loans may have substi- fungibility of money (loans provided under the tuted for these alternative sources of financing. credit scheme may substitute for other sources Audits of many agricultural credit projects of finance or may be applied to "unintended` have shown that alternative sources of funds investments). were available for financing project-related in- vestments (World Bank/OED 1981). Diversion The representativefarm model. In establishing a of funds is a form of substitution in which project's anticipated financial and economic credit funds are used for purposes not author- rates of return, an imaginary representative ized by the loan covenants. Close supervision farm is used as a model. Usually, the economic of numerous rural borrowers can be costly, and rate of return (ERR) on farm model invest- diversion can occur even in projects that are ments, which is used to quantify the project well managed. benefits, is estimated for investments in the The problem of substitution can be concep- model form and assumed to apply to the uni- tualized along a continuum. At one end, no verse of subborrowers. This assumption may be substitution is possible (as in the case of a erroneous, given the large variations in farmers' farmer in a remote area with no access to other characteristics. sources of funds). At the other end, a farmer's investment could have been implemented Limited comparability of project particípants and without the credit project or with less than the nonparticipants. Comparing a group that received amount lent. Except in rare instances when no project funds with a control group that did not substitution is possible, calculations of eco- may be misleading. The farmers that participate nomic and financial rates of return must be in most agricultural credit schemes are self- adjusted to reflect the possibility of limited ad- selected. Usually they are more entrepreneurial, ditionality under the project. It is difficult to less risk-averse, and more receptive to new tech- make these adjustments accurately (because nologies than their nonparticipating counter- substitution is not readily quantifiable), and parts, and they have stronger links with financing the problem of attributing benefits to credit agencies. In light of this reality, efforts to create a projects becomes increasingly complex as one "without project" situation established ex post moves along the continuum. through random sampling of project partici- A 1983 report by the World Bank's Opera- pants and a control group of nonparticipants tions and Evaluation Department (OED) notes are biased in favor of the project. that "obtaining adequate farm survey data to Performance Criteria for Rural Financial Intermediation 89 measure actual production impact of invest- cash flows, and debt servicing capacity of ments financed under the project is difficult; borrowers-which would allow for ex post problems related to fungibility may make such comparisons, which are preferable to ex ante attempts even superfluous as they will not nec- ones. RFIs should generally abstain from costly essarily measure project impact. What might impact assessments that could more appropri- be needed is in-depth farm survey/evaluation ately be conducted by research institutes with research to measure the impact of credit pro- the requisite skills and resources. jects, but this should be limited to a few pro- jects only; in other cases, economic analysis Macroeconomic Problems should keep the attractiveness of the develop- ment packages under review mainly during Attributing benefits to credit projects is even implementation, not ex post" (World Bank more tenuous when considering issues of 1983, paragraphs 5.150, 5.155). fungibility at the macroeconomic level. If, for Another OED study found that because of example, a donor were to provide a loan de- fungibility problems, "high economic returns nominated in foreign exchange for an agricul- as quoted in the completion and audit reports tural credit project to a country that maintains overestimated the real economic return" on ten free convertability of the currency but that is agricultural projects in India (World Bank/ already flooded with foreign exchange re- OED 1981), and in a review of twenty credit serves, it is plausible that the volume of agri- operations in five countries, rates of substitu- cultural credit loans in the country would tion were found to range between 25 and 75 increase by only a fraction of the amount of the percent (World Bank/OED 1976). In fact, sub- donor funding. The overall volume of agricul- stitution may be present to varying degrees in tural credit lending in the country is likely to most agricultural credit projects. be determined by monetary, financial, institu- Quantifying and attributing the benefits of tional, and rural-sector considerations, rather credit projects call for economic evaluation than by an externál loan from a donor. and econometric studies significantly more An opposite situation might arise in an econ- complex than the one-dimensional methods omy in transition. An external agricultural frequently used in assessing project impact credit loan may relax constraints on farmers' (World Bank/OED 1983). Researchers associ- and input suppliers' access (licenses) to foreign ated with Ohio State University question exchange for imports of agricultural inputs and whether even the use of rigorous economet- equipment. In such economies, the shortage of ric methods can provide conclusive results, foreign exchange may constitute a constraint given the fungibility of money and the diffi- on farm-level investments. As large traditional culty of establishing an appropriate control borrowers (like India and Mexico) move to- group (Von Pischke and Adams 1980; David ward free convertibility of their currencies, and Meyer 1984). However, more recent the direct link between the volume of exter- econometric studies have attempted to deal nal agricultural donor credit and domestic with some of these methodological concerns lending becomes much less discernible. In fact, (Binswanger and Khandker 1995). donor lending intended to increase domestic Since such econometric studies are costly agricultural credit may result only in balance and since findings cannot readily be general- of payments support. Under these circum- ized beyond the contexts in which the studies stances, the appropriate action is to focus on are conducted, it is not feasible for practitioners changing policies that hamper agricultural and to engage in regular, rigorous econometric rural growth and to invest in institution build- impact assessments. To engage in prudent ing in rural financial intermediaries. lending, rural financial institutions should To sum up, in assessing the benefits of gather information on the assets, liabilities, donor-financed agricultural credit projects, the 90 Rural Finance convertibility of currencies and availability of RFI operations or the special objectives often foreign exchange should be considered. assigned to rural financial institutions. The performance assessment framework in- Assessing the Performance of Rural troduced by Yaron in 1992 has since been Financial Institutions widely accepted by academia and practitioners (Christen and others 1995; Chaves and For many years no agreed-upon criteria existed González-Vega 1994; Ramola and Mahajan to evaluate the performance of rural financial 1996). The framework has two primary criteria: institutions. RFI performance was usually as- the level of outreach achieved among target cli- sessed using standard financial ratio analyses. entele and the self-sustainability of the RFI (figure These standard measuring tools do not take 7.1). These criteria do not provide a full assess- into account the various subsidies involved in ment of the economic impact of the operations Figure 7.1 Criteria for assessing the pertormance of rural financial institutions Primary assessment criteria SeIf- Outreach sustainability to target clients T_ Composite lndex: Subsidy Dependence Index HybEtd Index: - Measures subsidies Evaluates outreach to received against interest clients and quaW of eamed by RFI servEces offered Examples of subsidies Examples of indicators * Interest rate subsidy on concessionally borrowed o Market penetration funds * Number and annual growth rate of savings and loan * Opportunity cost of equity accounts * Other, including - e Value and annual growth rate of the loan portfolio * Reserve requirement exemptions and deposits * Free equipment by govemmenVdonors * Number of branches and staff e Government's assumption of loan losses o Relative income level * Free training for staff provided by government * Value of average loan and range of loan amounts donor * Percentage of rural clients * Government assumption of foreign exchange * Percentage of women clients loans o Quality of services * Transaction costs to clients * Flexibility and suitability of services * Distrbution network Perfornance Criteriafor Rural Financial Intermediation 91 of an RFI, but they serve as quantifiable proxies ample, if gender does not limit access to credit, of the extent to which an RFI has reached its the weight attached to the number of women objectives and make transparent the social costs clients may be low or even zero. The assess- associated with supporting the institution. ment of an RFI can only be as good as the information available on (and to) the RFI. Outreach Measuring the performance of an RFI requires that the RFI adhere to generally accepted ac- Outreach is a hybrid measure that assesses the counting principles and be subject to regular extent to which an RFI has succeeded in reach- extemal auditing. ing its target clientele and the degree to which The primary assessment criteria can be rep- the RFI has met that clientele's demand for fi- resented on two axes, as in figure 7.2. Some nancial services. The indicators of outreach are theorists argue that there may be a trade-off both qualitative and quantitative and can be between outreach and self-sustainability, but used to measure both depth (type of client many institutional policies that improve out- reached and level of poverty) and breadth of reach also improve self-sustainability (see outreach (number of clients served with differ- chapter 8 for details about building institu- ent kinds of instruments). tional capacity). Improvement in the perfor- Figure 7.1 gives examples of the indicators mance of an RFI requires improvement in at that can be used to measure outreach. These least one criterion, while performance in the and other indicators are presented for the Bank other is at least monitored. for Agriculture and Agricultural Cooperatives in Thailand, the Bank Rakyat Indonesia-Unit Self-Sustainability Desa System, and the Grameen Bank in Bang- ladesh in the annex to chapter 9. The indicators Previous interventions in rural financial mar- can be weighed and quantified according to kets have placed insufficient emphasis on the their relevance to a particular society. For ex- self-sustainability of state-owned rural finan- Figure 7.2 Optimizing performance High seif-austainability Sustainable financial Sustainable sevices with low financial services outreach to the that reach the target dients target clients Low High outreach outreach Highly subsidized Highly subsidized financial services financial services with low outreach that reach the to the target dients target clients Low seftsustainability Source: Mahajan 1994. 92 Rural Finance cial institutions. Attempts to track their per- Subsidy Dependence Index formance have been improved by the use of standard accounting measures to assess the To determine the real cost to society of main- special circumstances of rural financial insti- taining an RFI, measuring its performance tutions. should take into account the subsidies received The financial performance of profit-maxi- by the RFI. The subsidy dependence index mizing organizations can be measured using (SDI) provides such a measure. conventional financial profitability ratios, such as return on equity and return on assets. How- Definition and rationale. The SDI is a composite ever, these measures were not designed to measure of an RFI's financial performance (see measure the financial performance of state- box 7.2). Because it takes into account the sub- owned development finance institutions, sidies received by an RFI and shifts the focus which are not profit-maximizers and which al- away from traditional profitability ratios, the most always benefit from subsidies that carry SDI provides a more appropriate measure for an opportunity cost to society. Nor do conven- the performance assessment of an RFI than tional financial ratios take into consideration standard financial indicators. the effect of subsidies on the profitability of The SDI is expressed as a ratio that indicates rural financíal institutions. The financial prof- the percentage increase required in the on- itability ratios of an RFI can increase because lending rate to completely eliminate all subsi- of an infusion of subsidized funds or other dies received in a given year.35 It allows for the forms of subsidy, even as the true performance calculation of: of the RFI is deteriorating (see box 7.1). The . The volume of subsidies required to reverse is also true: an RFI's financial profit- keep an RFI afloat (the total amount of ability can appear to decline, not because it has subsidies received by an RFI in a given become operationally less efficient, but because period) subsidies granted to the institution have been * The ratio of subsidies received to interest reduced. earnedby an RFI on its loan portfolio in the The major forms of subsidy to state-owned marketplace. RFIs are usually the opportunity cost of the RFIs' equities (considered costless by the insti- Interpretation. The SDI has a lower bound of tutions) and concessional borrowing. In most -100 percent but no upper bound (Benjamin cases the equity of an RFI determines neither 1994). An SDI of zero means that an RFI has its ability to borrow nor the cost of its borrow- achieved full self-sustainability. An SDI of 100 ing. The concessional borrowing interest rate percent indicates that a doubling of the prevail- that benefits the typical state-owned RFI is usu- ing average on-lending interest rate would be ally determined by considerations of political required to eliminate subsidies. A negative SDI economy through a process involving a central indicates that an RFI has achieved full self-sus- bank and related ministries and agencies-not tainability and that its annual profits exceeded by market forces. As a result, distinguishing the total annual value of any subsidies received between a state-owned RFI's equity and its ex- by the RFI. Such an RFI could lower its average ternal financial liabilities often does not make on-lending interest rate, eliminate all subsidies, sense in the way that it does for profit-maxi- and remain self-sustainable. The SDI should, in mizing financial intermediaries. Although an some cases, be seen as a lower bound for the RFI's concessional borrowing rate has a major required increase in the on-lending rate, be- influence on its return on equity and return on cause full financing of RFI activities is likely to assets, its management has little, if any, influ- be difficult at the prevailing market reference ence over this exogenous determinant of the deposit interest rate if the RFI's financial per- institution's financial profitability. formance is dismal. Perforannce Criteria for Rural Financial Intermediation 93 Box 7.1 Merits of computing the subsidy dependence index for a rural financial institution in Jamaica The Agricultural Credit Bank of Jamaica was estab- dependence index (SDI) to the bank reveals that a total lished in 1981 as an apex institution supporting ag- of US$84 million was provided to the bank in implicit ricultural credit operations through commercial and direct transfers between 1983 and 1991 and that banks and rural cooperatives. The government in- transfers far exceeded profits every year during that tended that the bank should become self-sustaining period. The bank made solid progress toward self- within three to four years of its inception. A review sustainability between 1984 and 1987 (see figure below). of the bank's financial statements using standard fi- However, its SDI doubled between 1987 and 1991, ris- nancial indicators suggests that this goal was at- mg from 166 to 328 percent, and thus outpacing the tained. The retum on equity had become positive by growth m return on equities. Other things being equal, 1985 and by 1987 cumulative retained earnings had the net return on equity exduding transfers would have entered the black. Indeed, between 1987 and 1991 the been increasingly negative. bank's retum on equity rose by 75 percent-from 13.8 It is clear that (a) notwithstanding the attractive to 24.4 percent-and the retum on assets hovered standard financial ratios, the goal of achieving self- between 3.6 and 5.3 percent. sustainability within three to four years of the bank's However, more detailed analysis suggests that the inception was not attained, and (b) the SDI can offer bank has been dependent on subsidies for its opera- a clearer picture of a development financial institu- tions and that its reliance on these subsidies has tion's true financial position than is revealed by increased over time. Application of the subsidy standard financial analysis. Box figure 7.1 Comparing the SDI with the return on equity for the Agricutural Bank of Jamaica, 1984-91 50 9% 9% 14%/9 13% 21% 23% 24% 0 d - ROE 100 * - - - - - - ~~~~~~~~~~~~~~~~~~~~~~~~~SDI 150 _ 200 _ 250 - 197% 187% 250 - _ -____.__ 300 291% 350 - 328% 1984 1985 1986 1987 1988 1989 1990 1991 Source: World Bank 1993. Application. By applying the imputed cost of 1. By making explicit the slubsidies received subsidized resources extended to an RFI, the by an RFI, the SDI provides an estimate of SDI enables the assessment of the real financial the total cost involved in supporting an RFI. cost of intervention in financial markets and This estimate does not indicate whether the focuses on the real viability and longevity of the RFI has accomplished its socioeconomic ob- institution. The SDI enables governments, do- jectives, nor is it sufficient for a full economic nors, and RFI managers to better allocate and cost-benefit assessment. However, it pro- apply resources in three ways: 36 vides a starting point for comparing the costs 94 Rural Finance Box 7.2 Calculating the subsidy dependence index To calculate the subsidy dependence index (SDI) for a rural financial institution (RFI), all the subsidies received by the RFI must be aggregated. Total subsidies are then compared with the RFI's average on-lending interest rate multiplied by its average annual loan portfolio (the interest earned as presented in the RFI's income statements). The ratio of an RFP's annual subsidies to its annual interest income indicates the percentage by which that RFIs interest income would have to increase to eliminate the need for any subsidies. The SDI can be expressed as follows: SDI = Total annual subsidies received (S) Average annual interest income (LP*i) = A (m-c) + [(E*m) -P] +K n (LP * i) where: A = Annual average outstanding concessionally borrowed funds m = Interest rate the RFI would probably pay for borrowed funds if access to concessionally borrowed funds were to be eliminated. This is generally the market reference deposit interest rate, adjusted for reserve requirements, and the administrative cost associated with mobilizing and servicing additional deposits c = Weighted average annual concessional rate of interest actually paid by the RFI on its average annual outstanding concessionally borrowed funds E = Average annual equity P = Reported annual profit before tax (adjusted for appropriate loan loss provisions, inflation, and so on) K = The sum of all other annual subsidies received by the RFI (such as partial or complete coverage of the RFI's operational costs by the state) LP = Average annual outstanding loan portfolio of the RFI i = Weighted average on-lending interest rate of the RFI's loan portfolio = Annual interest ramed Average annual loan portfolio. Source: Yaron 1992b. of altemative public interventions (see box 7.3 rates of loan collection, savings mobilization, and Squire 1995). and control of administrative costs.37 This sec- 2. The SDI enables comparison of the finan- tion provides an overview of these four factors. cial performance and degree of subsidy depend- Chapter 8 provides details about improving ence of rural financial institutions that provide RFIs' operational efficiency more generally. comparable services to a similar clientele. 3. The SDI serves as a long-term planning Adequate on-lending rates and spreads. To com- and monitoring tool for governments, donors, pletely eliminate subsidy dependence, an RFI's and managers to track a specific RFI's progress on-lending rate should be positive in real terms toward self-sustainability over time. An analy- and set at a level sufficient to cover all operating sis of the sources and application of an RFI's costs and financial expenses. This is a main con- subsidies could further assist in determining tributing factor to financial self-sustainability. the merits of subsidizing an RFI (Yaron 1992a). High rates of loan collection. Loan losses have Factors Contributing to a Reduction often been the largest single cost borne by rural in Subsidy Dependence financial institutions and the principal manifes- tation of financial distress (see box 7.4). Appro- Four factors critical to eliminating subsidy de- priate monitoring of loan performance should pendence are adequate on-lending rates, high be an integral part of sound financial manage- Performance Criteriafor Rural Financial Intermediation 95 Box 7.3 Cost comparisons in public spending Box figure 7.3 shows the allocation of US$340 million on preventive health in a country with a high infant in public subsidies to a state-owned RFI in an African mortality rate. country over a four-year period. The amount of the The information provided by the SDI calculation subsidy was determined by using the subsidy de- enriches the public debate on the allocation of scarce pendence index (SDI) methodology. The annual av- resources at a time when state funds are shrinking erage amount of subsidy for the period was US$85 and the role of the state is being reexamined. This nillion. This amount was equivalent to 20 percent of cost comparison enables the following question to be public expenditure on basic education during the put: "Does this support for the RFI represent an op- same period and 165 percent of public expenditure timal allocation of scarce public resources?" Box figuro 7.3 Comparng subsidibs for rural finance wti funding for basic educadion and preventive heam care (mons of U.S. dotais) 450- 400- 350- 300- 250- 200- 150- 107 100 67 76 9 85 _ ~ - - 52 50- 0 , , , , _ 1989 1990 1991 1992 Average Preventive Basic Credit Credit Credit Credit credit health educaion subsidies subsidies subsidies subsidies subsidies Source: Adapted from Squire 1995. ment. The main indicators of loan performance standing loan portfolio may expand rapidly as are: the average loan size increases in line with in- * The ratio of arrears to amounts due, meas- flation. The percentage of arrears to outstand- ured for loans payable tracked by their ing loans may appear less alarming than it original maturity dates (see table 7.1). reallyis. This deceptiveeffectwillbedeepened * Loan collection as measured against total whenlong-termloans(withmaturitiesbeyond collectibles (the sum of old arrears and the one year) are issued with multiyear install- amount due in the reporting period). ments and grace periods or when arrears are For these ratios to be meaningful, an RFI measured against the overall loan portfolio of must clearly define arrears and adhere to which only a small part has fallen due. Under generally accepted accounting principles such circumstances a more useful ratio would with regard to writing off bad debt (see box be the contaminated loan portfolio as a per- 8.13 on the impact of failure to write off bad centage of the total loan portfolio. (The con- loans). Caution is needed when interpreting taminated portfolio refers to all loans loan collection ratios. In a highly inflationary outstanding with at least one installment over- environment the volume of the current out- due.) 96 Rural Finance Box 7.4 Performance of credit line calculate separate recovery curves for each pe- operations: collection experience riod to facilitate a realistic estimate of future of on-lending institutions annual provision for loan losses (see figure 7.3). Since this analysis takes account of historical "An important indicator in the design and monitor- trends in collecting both current dues and over- ing of credit line operations is the loan collection dues, it allows for: experience of on-lending institutions. There is no o Sound assessment of the adequacy of the strict one-to-one relationship between the profit- provision for loan losses ability and the repayment as businesses might orprvsnfrlonoss might not repay loans independent of the profit or o Appropriate pricing of the cost of loan losses they make on the loan. As an empirical mat- losses when setting lending interest rates ter, however, it appears that loan profitability and * Proper income recognition repayment are linked. Borrowers who expect to o Improved cash flow management. have to repay their loans seem more careful in their Despite these advantages, such an age analy- choice of projects than those who do not expect to sespita theseavntagely chndage a y repay. Not having to repay loans may lead to capi- sis of loan collection is rarely conducted by tal misallocation, since the borrower will make rural financial institutions. money even from socially unprofitable projects. This broad empirical association between loan col- Savings mobilization. Active savings mobiliza- lection and efficient capital allocation supports the tion helps to ensure a continuous source of concem of the Bank in improving bank supervision funds for an RFI. An increase over time in the and regulation in its borrowing member countries." ratio fotal RFI. of vover savin to ratio of the total value of voluntary savings to Source: World Bank 1995b, 5. the loan portfolio will indicate the extent to which an RFI has been successful in replacing concessional funds from the state (or from lTo calculate loan collection ratios and make donors) with savings from clients. adequate provisions for bad debts, an RFI should track loan performance continuously, Containment of administrative costs. To ensure as shown in the past-due aging analysis of the subsidy independence, administrative costs must Bank for Agriculture and Agricultural Coop- be reflected and covered in the final on-lending eratives in Thailand (table 7.1). The analysis rate of rural financial institutions. High transac- uses information routinely presented in the tion costs can reduce outreach and render an RFI bank's audited financial statements. It demon- uncompetitive, or they can increase dependence strates how arrears, measured in relation to on subsidies if the on-lending rate is not adjusted their original maturity dates, become loan appropriately. When analyzing administrative losses after several years of late payments of costs, the nature of the services provided by an overdue amounts. This analysis can be used to RFI must be considered; these services may in- Table 7.1 Past-due age analysis of short-term working capital of Bank for Agrculture and Agricultural Cooperatives, 1981-86 (percent) Amount due Overdue Overdue Overdue Overdue Overdue Overdue during the at year- end of end of end of end of end of Fiscalyear year end year + 1 year + 2 year + 3 year + 4 year + 5 1982 100 24 15 7 5 3 3 1983 100 22 9 4 3 2 1984 100 22 11 6 4 1985 100 24 11 5 1986 100 20 7 Source: Authors' findings. Performance Criteria for Rural Financial Intermediation 97 Figure 7.3 Loan recovery profile for short-tbrm sional loan to the ultimate borrower is viewed loans of the Bank for Agriculture and Agricultural as a necessary condition for undertaking the Cooperatives, 198146 investment. (percentage of original loan amount)ivstet The financial rate of return (FRR) alone does __________ 1984_______1983______ not allow the project analyst to judge the at- 1985 ... .................. tractiveness of the proposed investment to po- 95 1982 tential entrepreneurs (for definitions, see box so- / 7.5). The entrepreneur's decision to invest is 85 / motivated by the level of the (risk-adjusted) internal rate of retum on the investor's equity 80- / (IREQ). The IREQ should therefore be calcu- 75 lated whenever the FRR is computed. The po- tential entrepreneur is motivated to invest in 70 j "priority investments" by a high IREQ, not by a Due l 2 3 4 high FRR, although the two are closely inter- date 1 2 3 4 5 related. The IREQ is derived from calculating Years overdue the FRR with two modifications. Source: Authiors' findings. Often a large gap exists between the FRR and the IREQ. The gap will widen if there is a large clude (mobile) banking services, savings mobi- difference between the FRR and the on-lending lization, social intermediation services, techni- interest rate, the loan is granted for a long pe- cal assistance, and other nonfinancial services. riod (including the grace period), and the in- Administrative costs should be monitored vestment is financed by a large share of against total operating costs, annual average borrowed funds and a small share of owner assets, and the annual average loan portfolio equity. Therefore, the IREQ may provide some (see annex 9.1 for figures on the Bank for Ag- guidance in setting the adequate on-lending riculture and Agricultural Cooperatives, the interest rate to charge on the loan. Computing Bank Rakyat Indonesia-Unit Desa System, and the IREQ may be particularly helpful when the Grameen Bank). Analysis of the individual concessional directed credit schemes have to components of administrative costs could help to identify the major cost factors and changes Box 7.5 Definitions in the cost structure of an RFI. IRR The internal rate of return is a discounted Financial Rate of Return and Rate measure of a project's worth. It is the dis- of Return on Investors' Equity count rate that just makes the net-present value of the incremental net benefit On-lending rates are often set too low, result- stream, or incremental cash flow, equal zero. ing in a high subsidy per dollar outstanding ERR The economic rate of return is the IRR cal- loan portfolio. In the end low rates benefit culated using economic values. fewer borrowers than adequate access to for- FRR The financial rate of return is the IRR cal- mal credit would. culated using financial values. By definition, directed credit is always IREQ The internal rate of return on equíty is de- budget-onstraned. Cncessinal tems arerived by calculating the FRR with two budget-constrained. Concessional terms are modifications: (a) only the investor's eq- used to induce potential investors to borrow uity is considered (borrowed funds are and invest. The donor or the state acts as subtracted from total investment funds) financier of last resort when the borrower- and (b) the interest payment and principal investor is unable to obtain a loan elsewhere repayments are subtracted from the an- with the same or better terms. The conces- _ nual_cash_flow_ 98 Rural Finance Box 7.6 Illustration of the financial rate of return and the investor's retum on equity The example described here refers to a loan from a borrower as a result of the decrease in the value of large, state-owned agricultural bank in Asia to fi- the loan repayment and interest payments caused by nance a tractor. The figures demonstrate the merits inflation of 10 percent a year, and (b) the profitability of calculating the IREQ of the investor whenever the of the investment as judged by the IREQ was far FRR ís computed. higher than the on-lending interest rate, which was A farmer borrows $7,500 to buy a tractor that costs about 5 percent inflation adjusted (15 percent nomi- $10,000 (an investment of about thirty times the gross nal). What could be the justification for lending at domestic product per capita). The balance constitutes such a low interest rate when the investment yields equity. The financial rate of return (FRR) is estimated an expected IREQ equal to about twenty times the by the bank to be at least 28 percent. The loan has a borrowing rate? ten-year maturity with an interest rate of 15 percent This illustration points to the need to compute the a year. The country has had slightly more than 10 IREQ whenever the FRR is computed, as well as de- percent annual inflation in four out of the last five termining on-lending interest rates that are sound years. Reviewing the investment cash flow reveals and do not waste scarce resources and encourage that when the estimated FRR in constant prices was corruption because of the huge grant element embed- 28 percent and inflation was 10 percent a year, the ded in the concessional lending terms. Lending on IREQ was about 97 percent. the terms described in this illustration should be se- There are two lessons to be leamed from this illus- riously questioned, particularly in a country that tration: (a) the FRR calculated by the agricultural aims officially at increasing employment while actu- bank did not reflect the benefits that accrued to the ally enormously subsidizing capital. resist widespread pressure to lend at relatively funds, it is important to ensure that these goals low interest rates. are achieved and that the opportunity cost in- Calculating the difference between the esti- volved has been assessed. mated FRR and the IREQ of the investor may Because of the methodological problems as- help more clients benefit from scarce public sociated with impact assessments, the perform- funds and obtain access to formal credit than ance of rural financial institutions and rural is possible through heavy subsidies that benefit credit programs should be assessed with the a smaller number of borrowers. use of proxies for development impact. The pri- mary criteria proposed are outreach and self- Conclusion sustainability. The SDI, which is used to assess the self-sustainability of a program (or insti- Government-supported interventions in rural tution), also measures the social cost of financial markets usually aim to achieve devel- supporting a program. Using the SDI thus opment impact by expanding incomes and re- enables policymakers to wéigh support for ducing poverty. Since these interventions the program against alternative uses of public frequently involve the use of scarce public resources. CHAPTER 8 Building Institutional Capacity F ew rural financial institutions (RFIs) have been successful in both outreach Box 8.1 The need for institutional reforr and self-sustainability. Most state- -"Not even an extremely far-reaching and successful owned RFIs are the product of traditional in- financial sector reforrn would eliminate the need for terventionist approaches that emphasized specific efforts to initiate and support target-group- disbursements rather than the quality of lend- oriented financial institutions: There does not seem to ing, low interest rates rather than broad access be a 'market mechanism' which would induce estab- to credit, and agricultural production rather lished financial institutions to start providing financial services to the lower-income target groups as soon as than diversified, market-oriented rural econo- a financial sector reform has taken place. . . In the mies. The resulting poor financial performance medium- to long-term (banks) will have other strate- of RFIs has made them dependent on subsidies gic options, such as financing extemal trade, and ex- that have drained government budgets and de- perience indicates that they prefer these other options prived other sectors of funds. The soft budget and are reluctant to serve the poorer segmehts of the constraints applied to state-owned rural finan- population if they are not given specific incentives and techinical assistance to motivate and enable them to cial institutions have also undermined the fi- address the 'mass market' of the poor.' nancial discipline of the institutions and their clientele and have reduced incentives to de- Source: Krahnen and Sdcmidt 1995,s8. velop sound institutional practices and institu- tional capacity. The new approach to rural finance-limited, self-sustainability (see chapter 7). To attain a market-friendly government interventions that high level of outreach and self-sustainability, address clearly defined market failures-calls RFIs require appropriate governance, which en- for policies aimed at removing the causes of tails clearly defined and limited roles and pow- market failures rather than their effects. How- ers for government, donors, the central bank, ever, addressing macroeconomic, sectoral, and other agencies. Capable management with a legal, and regulatory constraints to rural finan- high degree of autonomy is also required, as are cial intermediation does not automatically re- innovative and efficient operating procedures sult in the emergence of efficient and effective guided by a well-developed management infor- RFIs. Reforms will facilitate the process, but a mation system. concerted effort is required to build strong in- This chapter provides guidelines on institu- stitutions (see box 8.1).39 tional policies and practices necessary for Building institutional capacity is a continu- building institutional capacity in rural finan- ous process aimed at improving the perfor- cial institutions. The chapter begins with a defi- mance of RFIs in terms of outreach and nition of institutional capacity, which is 99 100 Rural Finance followed by a description of the main problems * Motivated and skilled staffwith the ability to experienced by RFIs. It then discusses in detail effectively execute and continuously refine the key elements of a comprehensive strategy and improve the operational methodology for building institutional capacity. to better meet the organization's objectives o The ability to secure appropriate resources: Definition of Institutional Capacity human,financial, and technical. Institutional capacity refers to the degree to Problems Experienced by RFIs which an organization's objectives and proce- dures are embedded in its day-to-day activi- Few state and donor-sponsored RFIs have been ties. Edgecomb and Cawley (1994) define able to provide financial services to their target capacity as an organization's ability to move clientele in a sustainable manner. Access to a from thinking to action and to structure itself. continuous flow of inexpensive funds, lack of Capacity includes the following elements: commercial imperatives, and insufficient atten- * A strong management team capable of com- tion to information and incentive issues in de- municating the missionof the organization signing institutional mechanisms and and translating the organization's objec- operating procedures have resulted in inade- tives into actions quate credit evaluation, poor enforcement, * An organizational structure that supports poor loan collection, and costly administrative the organization's objectives, is responsive procedures. The perception that donor-funded to client needs, and engenders appropriate loans were grants that did not have to be re- participation paid has led to a lack of financial discipline by • Systems and procedures that are customer- both clientele and institutions (see box 8.2). friendly, facilitate the smooth and efficient Such factors explain the failure of many rural flow of information, and ensure transpar- financial institutions. ency in the institution's operations and Because of ineffective policies, only a small policies portion of the target market has been reached. Box 8.2 Potential impact of interest rate subsidies Aside from the effects that subsidies can have on the o Below market on-lending interest rates attract overall economy, they can have the following (unin- rent-seekers. They also result in rationing, tended) effects on rural financial intermediation at which typically favors the wealthy and politi- the institutional level: cally connected. • Subsidized on-lending interest rates could re- o Below market on-lending interest rates can en- sult in a lack of financial discipline by the rural courage institutional corruption and fraud be- financial institution (RFI) or its clientele because cause staff may try to capture the difference in cheap loans are often regarded as grants that the interest rate by eliciting side payments from do not have to be repaid. clients. • Low-cost loans to clients discourage savings . Subsidies prevent an RFI from attaining finan- mobilization because the loans preclude remu- cial self-sustainability, which requires that the nerative interest rates on deposits, and the low- interest rate margin should cover all the RFI's cost rediscounts associated with cheap loans administrative and financial costs over the long usually entail less administrative effort than is term. A higher than market interest rate is usu- required to mobilize and manage deposits. ally required to cover the higher transaction • Government budget constraints mean that if costs and greater risk associated with serving loans are extended to clients at subsidized in- an RFI's target clientele. terest rates, there will be fewer borrowers * Withdrawal or cancellation of subsidies by the (fewer people are served as the subsidy per dol- government or donors could cause financial lar lent increases). distress for an RFI and its clientele. Building Institutional Capacity 101 Box 8.3 Subsidized interest rates: good intentions, disappointing results There have been many attempts to reach target clien- interest rate (official interest plus baksheesh) paid by teles by heavily subsidizing on-lending interest rates. the ultimate borrowers. But the rural poor and small-scale farmers value ac- The accommodation of officials capturing the bak- cess to timely, reliable formal credit and savings fa- sheesh impoverishes the financial intermediary, ad- cilities much more than they value subsidized rates versely affects financial discipline, and undermines of interest. A recent World Bank staff appraisal report the repayment culture. A borrower forced to pay bak- describes a proposed project to provide credit in one sheesh will not be motivated to repay the loan, hav- country in transition: ing never received the full value of the loan. Failure Potential microenterprises do not have access to to repay will generate further deterioration in the formal bank credit, which has a current rate of inter- financial position of the financial intermediary. So an est of 24 to 32 percent (plus, reportedly, a 10 percent increase in the on-lending interest rate seems desir- baksheesh, or "irregular" payment to a bank official able on all counts. for authorizing a loan). Informal credit at an interest rate of 5 to 10 percent monthly is available mainly One year Six months for trade activities greater than US$10,000. The proposed project intends to introduce an on- Annual lending rate 24% 24% lending rate of 24 percent (about 3 percent in real Baksheesh (paid in advance) 10% 10% terms). The box table illustrates the effective lending Effective annual lending rates that the borrower will face once the baksheesh interest rate 38% 55% is taken into account. Under these prevailing circum- stances, and assuming that the scarcity of capital in Clearly there is no point in attempting to reduce the country is well reflected in the effective rates paid the on-lending interest rate through interest rate sub- by the borrower (inclusive of baksheesh), any inter- sidies when any reduction in the rate below the rate vention in the form of a lower interest rate will enrich prevailing in the market is likely to result in increased only the rent seeker-the bank official who can con- rent-seeking in the form of baksheesh. trol the availability of loans and approve applications with almost no trickle-down effect on the effective Source: Authors' calculation based on World Bank data. The wealthy and politically connected have Key Elements of Capacity Building been the main beneficiaries of most pro- grams. Subsidized interest rates have encour- There is no single formula for developing a aged rent-seeking behavior, and the benefits successful RFI: factors such as an RFI's target of cheap funds have seldom reached poorer clientele and the prevailing socioeconomic clients (see box 8.3). Weak governance ex- conditions will determine the most suitable plains the failure of most rural financial in- operational approach. This section discusses stitutions to detect and prevent rent-seeking the main components of institutional capac- behavior. ity and provides some principles for institu- RFI management teams have been held back tional development in formal and semiformal by limited autonomy in determining operational RFIs. policies. Because of their limited accountability, they have also often lacked the motivation to Ensure Appropriate Governance pursue viable policies. RFls have frequently been used as vehicles for patronage, with politicians The powers and responsibilities of the entities taking advantage of them to gain electoral sup- involved in managing and supervising an RFI port rather than helping them to serve the finan- should be clearly defined. These entities usu- cial needs of disadvantaged rural communities. ally indude an external auditing or supervi- Box 8.4 provides a typical example of political sory agency, the owners (who may be private interference. citizens, nongovernmental organizations, local, 102 Rutral Finance Box 8.4 Rural credit institutions as a political tool: debt forgiveness in India Rural financial institutions (RFIs) that are associated 'The 'willful' defaulters are, in general, socially with governments often become the target of politi- and politically important people whose example oth- cians. The influential clientele of these institutions ers are likely to follow and, in the present democratic makes them a particularly attractive political target. set-up, the credit agency-bureaucracy is reluctant to In India the government-appointed Agricultural touch the influential rural elite who wield much for- Credit Review Committee reported in 1989: mal and informal influence and considerable power. "During the election years, and even at other Farmers' agitation in many parts of the country is times, there is considerable propaganda from po- taking a virulent form, and banners are put up in litical platforms for postponement of loan recovery many villages declaring that no bank officer should or pressure on the credit institutions to grant ex- enter the village for recovery purposes. This dampens tensions to avoid or delay the enforcement process the enthusiasm of even the conscientious members of of recovery. In the course of our field visits, it was -the bank staff working in rural areas in recovery ef- often reported that political factors were responsi- forts. The general climate, therefore, is becoming in- ble for widespread defaults on the ostensible plea creasingly hostile to recoveries." (Part III, para. 39) of crop failures in various regions." (Part III, para. 36) Source: World Bank 1994b. or national governments, or donors), and the its and grows in scale. The supervision need RFI's management team. not be by the monetary authorities. External auditing or supervision by the central offices Supervision and prudential regulation. The de- of a federated credit union may suffice. Pru- gree and appropriate form of regulation re- dential regulations may be limited to separat- quired will depend on the size, ownership ing the roles and powers of the management structure, and nature of an institution. At one from those of a board or supervisory commit- end of the spectrum are large formal banking tee, although many countries require adher- institutions, whose owners and clients are sepa- ence to prudential laws if the RFI takes rate entities and whose many depositors are rela- deposits (see box 8.5). tively uninformed. These institutions should be When determining the appiopriate level of subject to prudential laws and to fairly intensive supervision, the potential costs involved in for- prudential supervision by the monetary mal supervision must be assessed. Regulatory authorities to protect the interests of creditors requirements (such as minimum-capital re- and ensure financial sector stability. However, quirements) may thwart the operations of a the supervisory body should not be involved small semiformal RFI, forcing it to restrict the in managing the day-to-day operations of the kinds of products offered or to quit the market institution. altogether. At the other end of the continuum are small credit unions or village banks. Because the cli- Management. A strong management team, ca- ents and depositors of these institutions are pable of formulating and communicating the often also the owners, self-regulation and vision and objectives of the institution, is critical monitoring may be acceptable. Even so, the po- to the success of an RFI. Management must be tential for rent seeking or conflict between the accountable for the performance of the institu- interests of borrowers and depositors may stil tion. It should therefore have a high degree of exist (Chaves and González-Vega 1994). Since autonomy in setting the operational policies of this potential is related to the size of the insti- the institution. Autonomy is of special impor- tutions and to the amount of information avail- tance to RFIs because funds are often provided able to depositors, some external supervision by government or other donors who may want is warranted if the institution mobilizes depos- to influence operational policies to benefit a Building Institutional Capacity 103 Box 8.5 When does informal become formal? of operation, and the costing structure of the South Africa RFI. Clearly defined goals and objectives are nec- "By the time the South African banking regulators essary to enable the RFI to measure its pro- became aware of Sun Multi Serve's operations, it gress. Once the objectives of the RFI have been had already collected approximately R50 million defined, a detailed plan and schedule should (US$13 million) from 53,000 black investors-well be developed that includes intermediate objec- in excess of the R9.9 million (US$2.7 million) statu- tives oped tat and q titate orm~ tory limit above which a bank must comply with tives and qualitative and quantitative perform- banking laws. South African authorities stepped in ance indicators. If the goal is to improve and froze the assets of Sun Multi Serve, pending an savings mobilization, related actions and ob- investigation. The central bank's chief of banking jectives should be specified, together with in- supervision insists that Sun Multi Serve is a pyra- dicators such as the target number of clients, mid scheme and that the banking regulators have volume of annual average deposits, and ratio the responsibility to protect depositors. However, Sun Multi Serve claims to be a stokvel (an informal of savings to loans to be achieved by a specific savings scheme or credit union) in mine with African date. When appropriate, explicit agreement tradition, and insists that the scheme is the victim should be reached on interim objectives with of 'cultural intolerance.' Some of its depositors external agencies. If an RFI does not meet its marched on the central bank to protest the central objectives, management will know that it needs bank's actions."Eolcle,mngmn ilko hti ed to improve the RFI's operational performance. Source: `South African Finance: Not a Formality," ie In defining the RFI's goals and objectives, Economist, January 13, 1996. clear roles must be established for the various stakeholders. Box 8.7 provides an example of the importance of clearly defined roles and specific group or attain political or other (dis- objectives. bursement) objectives (see box 8.6). Management should focus on reaching the Learnfrom the Informal Sector institution's target clientele. Outreach can be encouraged by linking management incentives Knowledge about a country's informal finan- to RFI performance in serving the defined cial markets can provide guidance on the kinds target clientele or by limited client participa- of practices that are acceptable and workable tion in the management structure, which the within a community. it also helps in identify- Grameen Bank in Bangladesh has done ing opportunities available in the market to successfully. better meet the financial needs of that commu- nity. An RFI that offers products or uses meth- Define the Institution's Strategies ods familiar to the target market is likely to and Objectives gain wider product acceptance and achieve greater market penetration (see box 6.3). An RFI should explicitly define its target clien- However, the objective of a formal RFI tele and objectives. Unless this is done up front, should not be to supplant the informal finan- the conflicting objectives of different stake- cial market but to compete with it in certain holders may prevent the institution from for- areas and complement it in others by providing malizing the most suitable policies and a broader range of products to a wider clien- procedures for fulfilling its mission. For exam- tele. For example, the existence of rotating sav- ple, if social intermediation or the provision of ings and credit associations (ROSCAs) or social services are institutional objectives, the voluntary group-based savings schemes in a institution should say so explicitly, because the community indicates that group-based finan- objectives will at the very least partially deter- cial schemes are feasible in that community. mine the types of services offered, the modes ROSCA members often, however, have no 104 Rural Finance Box 8.6 Rural savings and loan associations in Benin Benin's rural savings and loan associations were cre- and procedures, accounting, auditing, manage- ated in 1975 as local member-managed banks. They ment, and financial control. were soon placed under the jurisdiction of a public agricultural bank that functioned as an apex institu- Results. The network was completely revitalized tion for the associations. This move undermined their between 1990 and 1995. Membership increased almost mutualist principles and shifted decisionmaking fivefold to 127,000, and share capital rose by 160 per- from elected association members to govemment- cent to CFAF 390 million. Deposits increased 240 per- appointed managers. The apex institution used the cent to CFAF 9.4 billion, and credits rose from CFAF network to transfer savings from the rural sector to 200 million to CFAF 6 billion. The recovery rate was the urban (mostly public) sector. Because of lax credit consistently above 98 percent, and more than 90 per- policy under public management, the financial situ- cent of local banks reported operating profits with ation of the network deteriorated in the 1980s until -support from extemal donors. Measures have been it reached the brink of collapse. taken to ensure the medium-term sustainability of the regional associations and the national federation. A strategyfor reform. Several donors, including the World Bank, recognized the importance of savings New chal lenges. Does anything overshadow this pic- and loan associations in rural financial intermedia- ture? With success and popularity the same dangers tion and helped to devise a rehabilitation plan that that caused failures in the 1980s are back. Government included: interest in the network, which had waned during the * Reestablishing a positive macroeconomic and crisis period, is increasing steadily, and govemment sector policy environment and donors alike are lining up to propose new lines * Earning the confidence of rural dwellers by re- of credit and new activities in which the network has verting to mutualistic principles and empower- no experience or comparative advantage, such as ing members-government agreed to restrict medium-term loans. "Market failures" are once again itself from intervening in network affairs used as the rationale for these interventions. The mas- * Making external financial resources comple- sive (2,900 percent) expansion in credit has been det- ments to savings rather than the main source of rimental to security. The loans to savings ratio has loan funds risen from 50 percent to over 60 percent, average loan * Instituting more rigorous credit policies focus- amounts and reliance on external lines of credit have ing on security and sustainability increased, and group lending is decreasing in relation * Establishing a "full recovery" principle to main- to individual lending. Is history repeating itself? tain repayment discipline * Using group lending, familiarity between bor- Lessons. It is clear that appropriate rehabilitation rowers and elected association officials, and so- measures work, but success is not necessarily perma- cial pressure to ensure debt repayment nent. Any mistakes quickly forgotten are likely to be * Applying positive real interest rates and ade- repeated. It is therefore imperative in times of success quate spreads to increase sustainability to advise circumspection and restraint. * Building capacity and training staff and elected officials, especially in institutional mechanisms Source: Contributed by Luciano Mosele. access to interest-earning facilities where funds Determine Which Services Clients Need can be deposited, and recipients of savings therefore often hoard funds or spend them im- Successful outreach means meeting the tar- mediately upon receipt. The provision of inter- get clientele's demand for financial services est-bearing group savings accounts for ROSCA in a way that is as beneficial as possible for and individual accounts for ROSCA members the clientele. Krahnen and Schmidt (1995) could therefore encourage savings and im- note that "target-group orientation is more prove members' return on savings. Box 8.8 a matter of an institution's general policy or shows how innovations implemented by com- 'corporate culture' than of the clever appli- mercial banks in South Africa helped to com- cation of techniques or 'tricks of the trade."' plement informal saving schemes. In commercial terms that means that devel- Building Institutional Capacity 105 Box 8.7 Mysore Resettlement and Development Agency The Mysore Resettlement and Development Agency This poor performance can be ascribed to several (MYRADA) was established in 1968 to assist with causes. First, overlap of responsibilities and different Tibetan refugee resettlement. In 1984 MYRADA in- development objectives have led to confusion among itiated a group-based community development and group leaders and members. Second, there is enor- financial services project with credit management mous variation in the financial services provided and groups. Today, the program has a staff of sixty-seven in the way that groups are organized. Each group servicing 388 villages and more than 10,000 credit wrestles independently with issues such as member- management group members. ship requirements, interest rates, and loan size and use. Third, the Indian Bank offers, through its Inte- Kamasamudran District. The MYRADA savings-led grated Rural Development Program, subsidized model in Kamasamudran emphasizes the institu- loans to members selected by a credit management tional development of the credit management group, with the average loan size twelve times larger groups. The results have been encouraging. The over- than loans from the group's own funds. The link to all repayment rate in early 1993 was 85 percent, with Integrated Rural Development Program funds dra- women's groups outperforming men's. Accounts matically changes the incentive structure for mem- were updated and accurate, meetings were well or- bers, notably with regard to repayment of loans. The ganized, and individual members were well in- links are also a disincentive to increased mobilization formed about their rights and responsibilities. of savings from members. Finally, the use of paid village animators who tend to control the groups Dharmapuri District. In this area of Tamil Nadu undermines the credit group's principle of self- MYRADA has collaborated with many government reliance-an essential characteristic of other more institutions and nongovemmental organizations (in- successful MYRADA projects. cluding Plan Intemational, the International Fund for Agricultural Development, the Indian Bank, and the Conclusion. The effectiveness of MYRADA's highly Tarnil Nadu Women's Development Program) to participatory approach in Kamasamudran results provide financial and other services to groups of poor from the clear roles, responsibilities, incentives, and villagers. Activities include assisting in the formation performance standards set for MYRADA staff and and development of community savings and credit the leaders and members of the credit management groups; providing subsidized credit from the govern- groups. It also results from the effective sequencmg ment's Integrated Rural Development Program; pro- of organizational and financial activities. In Dhar- viding youth skills training; providing technical mapuri the MYRADA approach has become un- assistance in agriculture, forestry, and animal hus- wieldy because of the uncoordinated involvement of bandry; and assisting with resettlement, watershed various institutions. Financial performance and management, and land reclamation. group cohesion has suffered from conflicting objec- The results in the Dharmapuri District have been tives and incentives and from the availability of ex- disappointing. The credit management groups have ternal subsidized loan funds linked to a govemment displayed a general lack of financial discipline: the program. In sum, the complexity and multiple objec- repayment rate is estimated at only 54 percent, sav- tives of the Dharmapuri approach have adversely ings have stabilized at less than 50 percent of the loan affected financial performance and the strength of the portfolio, financial service delivery costs have in- groups. creased to at least 31 percent with the involvement of new institutions, and members are confused about Source: Contributed by Lynn Bennett and Mike their repayment schedules and obligations. Goldberg. opment institutions need to be customer- refining products, policies, and modes of op- oriented. eration to better meet clients' evolving needs. Development planners should not assume Client needs can be determined by learning that they know what clients want more than from the informal market; consulting commu- the clients themselves do. The financial re- nity leaders; interviewing current, former, and quirements of the target clientele should be re- potential clients; and analyzing defaults and searched as part of a continuous process of product performance. 106 Rural Finance Box 8.8 The formal meeting the informal in South Africa A stokvel is an informal savings scheme or credit union Some of the large commercial banks in South to which group members usually contribute monthly. Africa have recently developed savings instru- Stokvels have been part of the financial landscape in ments to accommodate these schemes. Stokvels are South Africa for decades. The National Stokvels As- able to open a single bank account for the group, sociation of South Africa represents about 11,000 with certain designated members having signature stokvels. There are different kinds of stokvels. Some rights to the account. Loans can also be issued are run as "funeral insurance schemes," with proceeds against the accumulated savings. These savings used to pay the expenses of members' funerals (costly and loan instruments enable stokvel members to affairs among black South Africans). Other stokvels safeguard their money, earn a return on their sav- pay out proceeds in full on a rotating monthly basis ings, and gain access to additional funds. The to mnembers to purchase otherwise unaffordable com- banks obtain access to funds for on-lending in a modities. Still others simply keep proceeds (with no regulated environment. interest eamed) in a safe place. Membership is often confined to a small group of people, but some stokvels Source: Adapted from "South African Finance: Not a For- have grown to a few hundred rnembers. mality," The Economist, January 13, 1996. The traditional view that credit was the only banking, mobile banking, or a combination of financial need of rural populations has been the two. Village banking relies on community- disproved. Rural populations, including the based savings and credit associations that poor, have financial demands similar to those may be linked with formal financial institu- of urban populations (access to credit, savings, tions (see box 8.9). Mobile banks may consist insurance, and payment facilities). of small, low-cost branch offices at the village However, the needs of rural clients may dif- level, which may be open only on selected fer with respect to the structure or nature of days. The offices are operated by a small staff products required and in delivery mecha- that spends much of its time visiting clients. nisms. For example, RFIs often focus on Meetings with clients can take place at the medium-term, production-oriented loans with market, at organized outdoor meetings, or at fixed-term repayment requirements. Liquid clients' homes. savings instruments, overdraft facilities, or other forms of quickly accessible funds may be Time and mobility constraints. These con- more appropriate for a creditworthy client dur- straints should be considered carefully, espe- ing personal financial difficulties. cially the lack of mobility among the poor, particularly women. RFIs should determine the Establish Appropriate Modes of Delivery most suitable time of the day or week to meet with potential clients, and they should arrange Efficiently and effectively delivering financial access and schedule meetings accordingly. services is particularly challenging in rural ar- eas. RFIs must overcome many obstacles, in- Literacy and numeracy constraints. To deal with cluding the geographic dispersion of clients, illiteracy and lack of numeracy among clients, underdeveloped infrastructure and lack of witnesses should be allowed to give legal stand- transport, the time constraints faced by the ru- ing to contracts signed by illiterate or innumer- ral poor, high illiteracy rates and other social ate clients. For smaller transactions, written barriers, limited information, and inadequate contracts could be replaced by oral contracts, formal enforcement mechanisms. joint liability, and regular meetings with clients. All clients should be told the effective rate of Geographic constraints. Some rural financial interest, not just the nominal rate, and the institutions have successfully used village amount and timing of installments should Building Institutional Capacity 107 Box 8.9 Village bank methodology Village banks are "community-managed credit and leaders with bank management training in leadership savings associations that are established to improve styles and responsibilities, membership eligibility members' access to financial services, build a com- rules, administrative procedures, and basic savings munity self-help group, and help members accumu- and loan recordkeeping. Some NGOs provide addi- late savings" (Holt 1991). The size of individual loans, tional training and support services, such as technical backed by the solidarity guarantee of the village training to improve productivity and nutrition and bank, is increased when the borrower repays in a child health care. timely way and saves money. The loan amount is CARE/Guatemala. CARE/Guatemala initiated a increased according to the amount of savings gener- pilot village bank project in April 1988 and launched ated by the member during the four- to six-month a full-scale program in early 1991. The program loan cycle. The participant is usually a rural low- is designed to respond to rural women's lack of ac- income woman who has no access to affordable fi- cess to affordable working capital and technical sup- nancial services and is dependent on informal port for income-generating productive and moneylenders and market intermediaries. commercial activities. The program provides a short- Different international nongovernmental organi- term loan to each participating village bank, requires zations (NGOs) pursue different institutional ap- mandatory savings, and provides organizational proaches to village banking. Katalysis works with training and limited production-oriented techni- local NGOs to create village banks; Freedom from cal assistance. The approach has gained widespread Hunger works with established credit unions. Catho- acceptance because of peer pressure, shared eco- lic Relief Services has apex institutions in Peru, Thai- nomic risk, and the use of delivery mechanisms that land, and Honduras to work with local NGOs and address the constraints and productive opportunities credit unions. World Relief, Save the Children, and of women in rural Guatemala. In late 1994 the pro- Project HOPE have local offices, and the Foundation gram had more than 10,000 women members and for Intemational Community Assistance has local af- CARE reported a 99 percent on-time loan recovery filiates with branch banks. rate. Training in village banking projects is usually car- ried out by extension specialists who provide elected Source: Contributed by Lynn Bennett and Mike Goldberg. always be clearly explained. Training in finan- stage links these groups to formal RFIs for two- cial management, budgeting, interest rates, way financial transactions. The role of groups collateral, and foreclosure may help to avoid in rural financial intermediation is described in misunderstanding, improve repayment per- box 8.10. formance, and ensure that a positive long-term relationship is formed between the client and Information and enforcement barriers. The prob- the institution. lem of asymmetric information is significantly more pronounced across rather than within rural Other social constraints. Some barriers to trade communities. Villagers often have information are based on social factors, such as gender and that is not available outside the community. RFIs ethnicity (see box 3.5). Various forms of social should aim to obtain information by consis- intermediation may be warranted to make for- tently deploying appropriate contractual de- mal financial services available to economically signs and clear operating procedures, including or socially marginalized members of society. reputation mechanisms, interlinked credit and One of the most important kinds of intermedia- savings transactions, group-based mechanisms tion overcoming social barriers is the provision for screening and selecting clients, the use of of group-based savings and credit services. local staff, and intensive direct monitoring (see First, capacity must be built among target cli- Stiglitz and Weiss 1983, Bell 1988, Varian 1990). ents to establish and manage groups. Building The approaches used to acquire information this capacity may require largely unremuner- are also often extremely effective in enforcing ated technical assistance from RFIs. The second contracts. 108 Rural Finance Box 8.10 Groups in rural financial intermediation When should groups be used? * Joint liability makes up for absence of individ- * When communities have strong group cohe- ual collateral siveness * Improved loan collection-better screening and * When initial start-up costs and transaction costs selection through peer pressure and joint liabili- are high ty, especially if group penalties and incentives * When individuals (potential group members) are incorporated in the loan terms can obtain information about each other at a o Improved savings mobilization, especially if in- lower cost than the bank can centives are incorporated in the group scheme * When individuals do not have collateral. * Lower administrative costs (selection, screen- ing, and loan collection). Guidelinesfor more effective use of groups * Focus on small, homogeneous groups that will Risks associated with using groups assume ~ .. son.epn.iit o uevsn * Risk of poor records and lack of contract en- assume some responsiblity for supervising f rcmn their funds * Impose penalties (such as prohibiting access to * Risk of corruption and control by a powerful further loans while an individual is in default) nucleus or leader within the group and províde incentives (such as promising * Risk of covariance because of similar produc- higher future loan amounts to those who repay tion activities on time) * Risk of generalized repayment problems (con- * Institute sequential lending (to allow groups to tagion) screen out bad risks). o Risk of limited participation by women because of gender mix o Risk of spending excessive time and money to Advantages of using groups form viable groups * Economies of scale-a larger clientele served by * Risk that departure of a group leader may jeop- a fixed investment in operating assets ardize group viability * Economies of scope-increased capacity to de- * Risk of potential of free-riders. (This risk can be liver multiple services through the same group reduced if the group can impose penalties on mechanism individuals.) * Less information asymmetry through group's knowledge of potential borrower-savers Source: Adapted from Pederson 1995; Bennett, Goldberg, * Reduced moral hazard risks through group and Hunte 1996; Besley and Coate 1995; Huppi and Feder monitoring and peer pressure 1990; Rashid and Townsend 1994. Contain Transaction Costs time and cost. The Swazi Business Growth Trust in Swaziland has no physical branch net- The transaction costs to both clients and RFIs work; it uses SmartCard technology, which al- should be minimized to improve the outreach lows clients to conduct transactions at regular and financial performance of RFIs. commercial banks (see box 8.11). The delivery system is a key determinant of A simple, decentralized loan-processing sys- transaction costs. A solidarity group can reduce tem can speed loan approval and reduce pro- the screening, selecting, and enforcement costs cessing cost by minimizing red tape and thus of an institution if the group carries most of increase dient satisfaction. the cost of these functions. Innovative tech- Efficient, appropriately trained, and moti- niques like mobile banking may further reduce vated staff will further improve operational ef- transaction costs for rural financial institutions ficiency and contain costs. (Staff policies are and their clients. discussed in greater detail toward the end of the chapter.) Appropriate technology can substitute for An effective management information sys- physical representation and reduce processing tem is vital for reducing transaction costs be- Building Institutional Capacity 109 Box 8.11 Technology in servicing the poor. introducing the SmartCard in Swaziland Swaziland is a small country with a fairly well-devel- make repayments at various commercial bank oped banking industry. However, commercial banks branches in Swaziland. The SBGT holds a line of in Swaziland have not shown much interest in pro- credit with the participating banks, and collects trans- viding financial services to small businesses and the action information from the banks and settles net poorer segments of the market. This lack of interest transfers to and from its line of credit daily. Commer- is caused by the usual constraints, such as lack of cial banks provide this service free of charge becaise collateral and the high transaction costs associated the SBGT is a nonprofit organization that does not with servicing these clients. compete for the banks' clients. SmartCard credit The Swazi Business Growth Trust (SBGT), a pri- purchases can be made at authorized dealers and vate nonprofit Swazi-directed nongovernmental or- suppliers, who are provided with card-reading ganization, was established in 1992 with support equipment by the SBGT. from Development Alternatives to facilitate the de- A SmartCard transaction takes a fraction of the velopment and growth of small Swazi businesses by time of a regular teller transaction. The SmartCard helping them to obtain access to key inputs. To mini- provides SBGT clients with access to funds through rnize the transaction costs associated with extending a convenient countrywide distribution network at and recovering loans, the SBGT introduced the SBGT minimal cost to the SBGT. To further reduce costs, SmartCard. Each customer receives a SmartCard, the SBGT plans to put the SmartCard network on-line which has a memory chip on its surface that contains in commercial bank branches as soon as the Swazi all information pertinent to the client's loan. Com- telephone system allows this. mercial bank branches are supplied the equipment to read the cards and SBGT clients can draw funds and Source: Adapted from McLean and Gamser 1995. cause it speeds and improves the quality of deposits can be extended as credit. Appropri- decisionmaking, enables staff to detect and im- ate, positive real on-lending rates will also mediately deal with potentially costly prob- discourage rentseekers, who might crowd lems, and affords management greater out the target clientele if interest rates were oversight over decentralized activities. subsidized. Cover Costs with Appropriate, Positive Customize Loan Terms and Conditions On-lending Interest Rates for the Target Clientele Interest rates charged by the informal sector Loan terms and conditions should be simple, are usually much higher than those charged flexible, and comprehensible to clients. The by commercial banks, showing that the rural special circumstances of the clientele should be poor are willing and able to pay high interest considered in structuring the terms of a loan. rates. What they need is improved access to funds, not subsidized loans that hardly reach Loan amount. Loan amounts should be deter- them. mined according to clients' projected cash flow Both savings outreach and the quality and and repayment capacity. Debt capacity calcula- volume of lending can benefit from a positive tions should take into account the likelihood of real on-lending interest rate that covers the true adverse shocks to clients' incomes-a major is- risk and full administrative and financial costs sue for many supply-led specialized agricul- associated with lending to the target group. A tural credit institutions (Von Pischke 1991). positive interest rate will enable an RFI to pay positive competitive interest rates on deposits. Loan type. Often loans are offered only for Paying competitive interest rates can simulta- investment or working capital purposes, but neously stimulate both savings mobilization clients may have a more immediate demand for and the volume of lending, since additional consumption loans to address personal finan- 110 Rural Finance cial crises or to finance lumpy consumption collateral requirements of mainstream banking. expenditures. More flexibility in loan condi- Another is to create implicit or explicit multipe- tions and types of lending instruments is riod contracts that make the size of new loans needed because of the fungibility of money and contingent on prevíous repayment performance. the frequent lack of clear distinction between target dients' personal and business finances, Monitor and Maintain the Quality of Assets especially for smaller enterprises and farmers with low surplus production. Poor loan collection is a major cause of finan- cial distress in RFIs. Special attention must be Installment amount and timing. The repayment paid, first, to monitoring, and second, to main- ability of clients may be affected by seasonal- taining the quality of assets. ity or be highly vulnerable to natural events. Fluctuations in repayment ability could be Tracking and reporting loan performance. The provided for without sacrificing financial dis- absence of a process to define and systemati- cipline by allowing a greater choice of repay- cally account for arrears and loan losses often ment terms at the associated cost of funds. results in the misrepresentation of assets, prof- Indeed, allowing for frequent repayment of its, and net worth. RFIs often understate loan small amounts should improve repayment per- losses and so overstate profitability. The ap- formance and the ability of RFIs to monitor pearance of profitability can entice politicians to repayments and detect problems early. make greater use of the institutions for patron- Krahnen and Schmidt (1995) found that shorter age and lead governments to tax rural financial terms to maturity were associated with im- institutions more heavily. Operationally, the proved collection performance on loans to non- appearance of profitablity can hamper careful farm enterprises. screening and'selection of clients and prevent RFIs from pricing loans appropriately. To sys- Collateral. When customers are not capable tematically monitor loan performance, RFIs of providing formal collateral, RFIs should be should: innovative. The use of joint liability groups * Develop an effective management infor- and compulsory savings are two examples of mation system (see box 8.12) to track pay- ways to overcome clients' inability to meet the ments, dues, and overdues daily. Box 8.12 Effective management information systems The importance of a sound management information branch level to provide incentives to clients or system in a rural financial institution (RFI) cannot be investigate problems and either reschedule exaggerated. It enables informed decisionmaking loans or implement punitive measures. Client by management, sends warning signals to RFI staff and 'profiles' could also be developed to improve external evaluators, and contributes to the transparency future loan selection and screening. of operations. A good management information system RFIs should strive to continuously improve the will at a ninimum provide information about: technology and methodology they employ, but * The financial performance of the overall insti- management information systems do not have to tution be high-tech to be efficient. The focus should be on * The performance of products, especially in loan simplicity and on the integration of the system into collection and arrears the daily activities of the RFI. As McNaughton * Individual branch and staff performance, re- (1992, 120) notes: "[Information technology] wards for performance may be required as in- should support the management process, not pre- puts for the staff incentive system - cede it." * Repayment performance of individual clients or groups. This information will enable staff at the Source: Authors' findings. Building Institutional Capacity 111 Box 8.13 Impact of failure to write off bad debt Rural financial institutions often fail to write off bad Instead, it is often measured as loan collection over loans because of legal restrictions or the desire to demand (the amount falling due plus the cumulative show exaggerated profits. When bad debts are not total of previous years' bad debt). Several countries, written off, reported loan recovery performance may including India, have followed this practice. be highly distorted. Assume that the RFI lends 100 units of currency The calculation below provides an example of the every year of which 10 units are never collected and type of distortion that can arise when a bad debt is not never written off. The denominator (demand) conse- written off but is carried over as part of the outstanding quently increases by 10 units annually, resulting in a loan portfolio. Loan collection performance should be misleading deterioration of the recovery ratio while measured as follows: loan collection over the amount actual performance has remained constant-at a 90 falling due during the reported year net of bad debt. percent annual collection ratio. Year 1 2 3 ... 10 Collection x 0% 90 =9% 90 = 82% 90 75 .. 90 47 Demand x 100% = 90% 10+100 - (10x2) + 100 (10x9) + 100 Source: Yaron 1994b. Adopt a transparent and informative meth- Manage and Diversify Risks odology by which to monitor loan perfor- mance. A past-due-aging analysis like the The longevity of an RFI is largely a function of one used by the Bank for Agriculture and its ability to manage risk successfully. A de- Agricultural Cooperatives in Thailand is tailed discussion on risk management in finan- highly recommended (see figure 7.3, table cial institutions is beyond the scope of this 7.1). paper. However, the three categories of risk • Pursue transparent accounting procedures that characterize RFIs do warrant mention. that follow generally accepted accounting principles for income recognition related Subsidy risk. Dependence on subsidies threat- to nonperforming loans, provisions for ens the longevity of RFIs and subjects them and bad debts, and write-offs of unrecoverable their clients to risk in the form of possible vari- loans. Past-due-aging analysis should ations, and especially reductions, in future sub- form the basis for realistic estimates of an- sidized funds (Krahnen and Schmidt 1995). ticipated loan defaults, although current This risk can be mitigated by measures that and expected future trends should be con- reduce dependence on subsidies. sidered. Box 8.13 provides an example of the distortion that can arise in reporting on Covariant risk. RFIs are subject to high covari- loan collections when bad debts are never ant risks in lending to their target clientele, since written off. the rural economy is highly influenced by!agri- culture, which is seasonal and subject to covari- Maintaining the quality ofassets. Several factors ant shocks. The solution is to diversify to the can cause high arrears on loans: poor screening extent possible and to forge links with financial and selection of clients, insufficient monitoring intermediaries in other areas (including urban of loans, and inadequate incentives for clients areas). to repay. Box 8.14 provides guidelines for re- ducing arrears and improving loan collections Default risk. Challenging covariant risks and in RFIs. information and enforcement barriers in rural 112 Rural Finance Box 8.14 Quality of assets: a healthy loan portfolio Guidelines for reducing loan defaults include the * The RFI should mobilize savrings where permit- following: ted-this may give the client a sense of owner- * Lending must be initiated to meet the bankable ship of the loanable funds and of the institution demand for credit (as opposed to being supply-led). (Bennett, Goldberg, and Hunte 1996). * Clients must be carefully selected and screened, * Clients should be rewarded (through interest asking for character references and forming rate rebates or access to larger future loans) for groups can offset the lack of information about timely repayment and penalized (through fore- potential borrowers. closure and denial of future access to loans) for * Altemative forms of collateral, such as joint Ii- defaulting. ability, should be accepted. * Staff should be given incentives (through profit * Foreclosure should be exercised (where legally sharing and promotion for performance) for possible) in instances of default. Although fore- high loan collection performance. closing may not always be cost-effective, par- * Loan perfornance should be monitored daily ticularly for small loans, it can deter willful through a well-developed management infor- default by other clients. mation system-arrears should he investigated * Loan size should be carefully evaluated and immediately by the loan officers responsible for gradually increased according to the client's managing the account. repayment performance, earnings, and debt- * Rent-seeking by staff (fraud and bribery) should carrying capacity. be detected and prevented by requiring manag- * Loan terms should meet the needs of the client, ers and internal auditors to monitor staff, by be flexible, and be convenient in relation to the providing staff incentives and competitive re- repayment schedule, installment amounts, and muneration, and by eliminating opportunities location of payments. for arbitrage afforded by below-market on- * The loan contract must be clear, simple, and lending interest rates. explicit. * Covariance risk associated with loan portfo- * The effective interest rate should be quoted to the lios should be reduced through diversification client. of loans in relation to clients and client * A positive real on-lending interest rate must be groups, lines of activity (by lending for agri- charged so that the loan is not perceived as a grant. culture and nonfarm activities), loan maturi- * Arrangements with the government (especially for ties, and location of clients. Prudent maximum interest payments) should be transparent ratios should be established relative to net (Krahnen and Schmidt 1995). worth. areas expose RFIs to higher default risks. These Mobílize Savings Resources in the Market risks can be reduced by carefully screening and selecting clients, using joint liability mecha- Savings have been referred to as "the forgotten nisms, providing incentives for timely repay- half of rural finance" (Vogel 1984). Indeed, un- ment (an interest rate rebate, for example), and til recently, little effort had been made by rural carefully monitoring loan performance. Mak- financial institutions to mobilize savings. Ac- ing adequate provisions for doubtful loans and cording to Robinson (1994), this lack of effort including appropriate risk premiums in on- can be attributed to misconceptions about rural lending rates will help RFIs to cope with this clients (the notion that the poor cannot and do risk and reduce the likelihood of financial dis- not save), ineffective policies (cheap lines of tress. Credit guarantor schemes have been in- credit to rural financial institutions), and op- troduced by some governments to reduce RFIs' erational inadequacies (limited, low-cost ac- risks and encourage them to lend to customers cess to liquid and secure savings instruments). with insufficient collateral. Credit guarantees In recent years different kinds of rural fi- are discussed in chapter 6. nancial institutions around the world have Building Institutional Capacity 113 proliferated, and informal savings schemes erative Development Foundation in India (notably ROSCAs), have had tremendous (box 8.15) and the pilot savings and loans pro- success with savings mobilization. These de- gram in Madagascar (box 6.7) are young velopments prove unambiguously that rural programs with savings as the major source communities can and want to save. The Bank of funds. Rakyat Indonesia and the Bank for Agricul- ture and Agricultural Cooperatives in Thai- The benefits of savings mobilization. The provi- land are examples of established rural sion of savings facilities by rural financial insti- financial institutions with fast-growing sav- tutions can contribute to improved financial ings mobilization (see chapter 9). The Coop- intermediation by: Box 8.15 Cooperative Development Foundation: strengthening women's thrift groups in India The Cooperative Development Foundation (CDF) credit, 4 percent were used to cover interest on sav- has been an active force in forming and strengthening ings, and the remainder was used to cover expenses. women's thrift and credit societies in Andhra By September 1995 the system was serving 13,297 Pradesh, India, for more than ten years. In 1991 the women members in 238 groups. Loan portfolio CDF launched a project to organize women's groups growth in the CDF system has been rapid, with a 425 and to develop a cost-effective, sustainable, and rep- percent increase in 1992 and a 235 percent increase licable model for women's thrift cooperatives. CDF in 1993. The share of portfolio backed by savings was believes that when women have access to credit, they 98 percent in 1993. The balance of loanable funds was can take part in local economic growth by estab- Rs 16.4 million (US$530,000) by Decernber 1993. By lishing their own enterprises. There is no conscious 1995 the portfolio had reached Rs 7.4 million effort to target low-income women-any woman (US$239,000), savings of Rs 100,000 were being mo- with a business opportunity and no affordable source bilized each month (about $US3,400). Financial per- of financing is considered to be a good candidate formance indicators showed a highly successful for membership. However, the strategy is to pro- cooperative program, with a repayment rate of above mote women's thrifts only in areas of active paddy 95 percent and a real lending interest rate of 12 per- cooperatives. cent. The deposit interest rate was slightly positive, The CDF has used a "saturation" approach in the unlike other similar programs in India. Administra- Warangal and Karimnagar Districts to achieve its ob- tive costs are low for a new program, amounting to jective. There are four reasons for choosing these dis- an estimated Rs 0.24 per rupee of loans outstanding. tricts: (a) successful experience with men's paddy Four factors have contributed to the success of the cooperatives, (b) growing markets in the area, (c) the CDF system. First, other prograras have promoted ac- existence of farm-to-market roads and other basic in- cess to subsidized credit from external sources, but the frastructure, and (d) the receptiveness of women in CDF has emphasized savings mobilization as the exclu- the villages. Even though the CDF has a staff of only sive source of funds for loans to members. Second, the five and has received modest support from donors, security of deposits has been demonstrated through the it has been a catalyst, emphasizing cooperative group use of transparent record keeping, efficient financial organization as the first step in developing a sustain- management systems, and periodic reporting. Third, able financial services system. groups control credit decisions, contributing to owner- Individual cooperatives depend largely on inter- ship and comnitment, with the CDF providing assis- nally generated savings to provide a pool of loanable tance in devising installment repayment plans. Finally, funds, since the CDF does not offer a credit facility. the CDF promotes the development of improved, com- Direct CDF assistance to the thrift groups has de- puterized financial management systems and a repre- clined from a peak of Rs 21,500 in 1991 to Rs 12,900 sentative apex body. It also has a clear "exit strategy," in 1992. The emphasis on self-reliance has led to enabling it to maintain a role as consultant to leaders efficient cooperatives. Mulkanoor cluster's while avoiding management decisions that might lead Kasiviswanadha Savings Society illustrates the effi- to dependency. ciency of the individual thrift and credit cooperatives. Of funds allocated, 92 percent were extended as Source. Bennett and Goldberg 1993. 114 Rural Finance * Helping clients to smooth consumption * A high level of client confidence in the patterns while earning higher real returns institution. than they might by hoarding or saving real goods Institutional capacity and regulatory require- * Enabling clients to build reputations and mentsfor accepting deposits. Savings mobilization collateral with rural financial institutions is not always feasible or desirable for rural fi- • Enhancing clients'perception of "owning" nancial institutions. Schmidt and Zeitinger an RFI and thus potentially increasing (1996) found that savings mobilization may not their commitment to repaying loans40 always be an optimal source of funds for insti- * Encouraging rural financial institutions to tutions: the administrative complexities and intensify efforts to collect loans because of cost associated with mobilizing savings, espe- pressure from depositors (which becomes cially of small amounts of it, may be prohibitive especially strong when depositors lose for some rural financial institutions. Institutions confidence in the RFI) may find it difficult to comply with prudential * Providing a source of funds for rural finan- regulations that apply to deposit-taking finan- cial institutions that can contribute to cial intermediaries. When mobilizing savings is improved loan outreach, increased auton- not an option, an RFI may have to rely on lines omy from governments and donors, and of credit from domestic financial institutions or reduced dependence on subsidies. on foreign sources of funds. An RFI may at times be prevented from ac- Improved savings mobilization. Different meth- cepting deposits by limitations on the issuance ods can be used to improve savings mobiliza- of government licenses for savings mobiliza- tion, but the most suitable approach for an RFI tion. There may not always be a particular need will depend on the cultural and socioeconomic for more depository institutions, because the circumstances of its target clientele. The Gra- supply of savings facilities in the market may meen Bank in Bangladesh, the Promotion of be adequate. However, where there are no sav- Rural Initiatives and Development Enterprises ings facilities close to the target clientele or no in Kenya, and numerous cooperatives require institutions interested in serving that clientele, mandatory savings by client-members who governments may consider adjusting the regu- want to gain access to loans. Access to these lations in rural areas to facilitate increased savings is often restricted. However, the Bank savings mobilization. Donors can draw gov- Rakyat Indonesia and the Bank for Agriculture ernment attention to regulatory constraints. and Agricultural Cooperatives in Thailand Any RFI that accepts deposits should have have mobilized voluntary deposits that are the financial strength and institutional capacity separate from credit transactions. to do so. When neither the government nor Some broadly recognized requirements for donors are willing or able to be lenders of last effective voluntary savings mobilization are: resort, a poorly run RFI may experience finan- * Positive real deposit interest rates (which cial distress that can lead to substantial losses require positive real on-lending interest for depositors and loss of confidence in finan- rates) cial intermediaries in general. Financiers of ru- * Flexibility and diversity of savings instru- ral financial institutions should ensure that ments they meet minimum requirements before sup- * Easy access to deposits for clients porting their aspirations to mobilize savings. * Easy access to the financial institution (through its branch network or through Motivate and Invest in Staff mobile banking staff) • RFI staff incentives linked to savings Because rural financial institutions are in a mobilization service industry, staff is their main asset: the Building Institutional Capacity 115 services delivered can only be as good as the Performance indicators for staff should be staff that provides them. Many successful rural selected carefully. For example, the volume or financial institutions appoint well-educated number of loans as a singular indicator could staff and invest substantially in staff training.41 induce nondiscretionary loan approvals and Rural financial institutions should ensure even coercion of dients. However, if staff in- that staff policies conducive to high productiv- centives are tied to loan repayment perform- ity are put in place. Appropriate financial and ance, the volume of loans supervised and nonfinancial incentives related to institutional deposits mobilized, and the number of bor- objectives will motivate staff and reduce the rowing and saving clients, staff behavior will incentive for rent seeking. Transparent proce- better meet the long-term objectives of the dures and strong controls (internal auditors RFI. The discussion of three successful rural and an effective management information sys- financial institutions in chapter 9 provides tem) will help to ensure staff behavior that con- some examples of staff incentives and training tributes to meeting the objectives of RFIs. undertaken by successful institutions. Box 8.16 Risk and its consequences Project failure is easily explained by risk. The most Unexpected events reduce this capacity. Awareness obvious manifestation of risk in credit projects is poor that this is so may lead to measures to manage the portfolio quality that leads to bad debt losses that real risk faced by farms, industrial concerns, trading erode the capital of the lending institution. firms, and microenterprises. On the farm, for exam- At one level failure occurs because of unexpected ple, changes in cropping pattems or input use may events that were not forecast in a project's design. In make the farmer more secure. credit projects these events are often bad agricultural Efforts are also required to manage the lender risk. yeáts; unforeseen increases in the cost of industrial Lenders can reschedule loans, provide additional production; and lack of markets for the skills, prod- credit, develop savings facilities to provide altema- ucts, and services supported by credit. Opportunistic tives for borrowers who are also depositors, alter loan behavior by borrowers who refuse to repay and the sizes to diminish future credit risk, raise interest rates unwillingness or inability of lenders to exert mean- to offset credit risk, and use any leverage that may ingful sanctions against such borrowers can also lead be available to foreclose on security. to failure. Risk management by lenders requires attention to At a more profound level the manifestations of risk liquidity. When adversity strikes, borrowers are less suggest shortcomings in credit project design. It is often likely to repay, and depositors may want to draw helpful to ask why these shortcomings from expecta- down their savings balances. Funding of new loans tions were not foreseen and why the project design becomes especially important during crises because failed to include plans to deal with the unexpected. credit is useful when the borrower's resources have At the outset of a project it is not necessary to been depleted. identify potential catastrophic situations, such as po- Risk exists because things do not always work out litical upheaval or the collapse of civil order. Instead, as planned. Planning how to manage the risk that project designers should ponder the more mundane affects credit projects has positive benefits for all con- things that are likely to go wrong. They should ask cerned. Planning helps the lender and others to ac- the question, "What is most likely to go wrong?" This cumulate knowledge about the problems faced by effort to identify the major credible threat is highly borrowers and how these problems can be managed. useful, in part because it encourages an expansive It encourages the lender to build robust systems for view. Once the thing most likely to go wrong is iden- managing relationships with borrowers and the port- tified and plans are made to manage the risk, the folios the relationships generate. Efforts at risk man- questions become what else is most likely to go agement by donors through improvements in credit wrong, what damage can it do, and what response project design enable them to contribute to develop- would be the most effective. ment in a more sensible manner by making project The effects of potential adverse events should be design more responsive. estimated and the estimate used to test borrowers' cash flows, which determine their capacity to repay. Source: Von Pischke 1991. 116 Rural Finance Information and Incentives Conclusion Two recurrent themes emerge in almost all the Targeted credit without institution building in recommendations proposed for building insti- rural financial institutions is almost always a rec- tutional capacity. The first is the need to in- ipe for prolonged dependence on donor or state crease the availability of relevant information funds and bailouts (Yaron 1992c). It is not access to enable RFI managers, staff, and dients to to cheap credit that makes financial institutions make informed decisions about financial successful, but competent people and sound transactions. The second is the need to provide management policies and practices. Institutional appropriate incentives for parties to financial development should be an integrated and evolv- contracts. The most successful rural financial ing process. It requires that the RFI's leadership institutions have developed effective manage- develop a clear mission, which is communicated ment information systems to meet their infor- throughout the organization. The processes and mation requirements and have built incentive procedures must be embedded in the daily ac- features into many aspects of their institu- tivities of the istitution, and the staff must have tional design and operating procedures (see the necessary skills and incentives to ensure sta- box 8.16). bility and continuous improvement of the modes of operation of the RFI. CHAPTER 9 Best Practices: Three Success Stories T his chapter analyzes the management tion's policies, performance indicators, and practices and modes of operation under- modes of operation. lying the success of three rural finance institutions (RFIs):42 Summary of Main Features * Bank for Agriculture and Agricultural Co- operatives (BAAC), Thailand43 The BAAC, the BRI-UD, and the GB each have * Village Banks (Unit Desas, BRI-UD) of millions of clients. These institutions are Bank Rakyat Indonesia (BRI)44 important in both rural and national finan- * Grameen Bank (GB) in Bangladesh.45 cial sectors. Here are some of the features All three institutions are widely perceived as they share that have contributed to their successful in achieving the primary perfor- success: mance objectives of outreach and self-sustain- o Favorable macroeconomic, agricultural, ability, and for all three excellent information and rural policies is available about their operations. These are o Adequate investment in the physical and not, however, the only successful RFIs. Several social rural infrastructure (Thailand and other promising rural financial programs have Indonesia) emerged in recent years, some of which are o High degree of management autonomy in described briefly in previous chapters. It is not the formulation of operational policies yet known, however, whether they will achieve * Staff policies that stress training, account- long-term success in terms of outreach and ability, and reward through monetary in- self-sustainability. centives and promotion The relative performance of RFIs should be * Innovative, low-cost distribution systems analyzed in the context of their individual ob- and mobile banking jectives and target clientele. The three institu- * Innovative and flexible loan terms and tions selected for this chapter differ from each conditions other in many respects, but all three have con- a Close monitoring of loan performance, sistently followed basic principles. The chapter high collection rates, and low arrears provides an overview of each institution, fol- * Strong records of domestic savings mobi- lowed by an assessment of the institution's lization as a growing source of funds, re- performance in terms of outreach and self- sulting in diminishing or eliminating the sustainability. It also analyses the underlying need for donor funds factors (both internal and external) that con- o Positive and often relatively high on-lend- tributed to the institution's success. Annex 9.1 ing rates presents additional details about each institu- * Control over administrative expenses 117 118 Rural Finance Table 9.1 Operating costs (percent) Rural financial institution 1987 1990 1993 1994 Percentage of average annual value of total assets BAAC 3.6 3.7 4.0 3.3 BRI-UD 12.3 8.0 5.9 5.4 GB 8.6 8.8 9.7 8.4 Percentage of average annual value of outstanding loan portfolio BAAC 5.3 5.6 5.2 4.3 BRI-UD 16.1 12.9 15.1 13.8 GB 19.2 16.5 14.1 12.2 Percentage of average annual value of outstanding loan portfolio plus deposits BAAC 4.0 3.7 3.3 2.6 BRI-UD 9.9 5.8 4.7 4.3 GB 13.9 11.9 10.0 8.9 Note: BAAC's operating costs include annual provisions for doubtful amounts. Source: Authors' findings. * Advanced management information sys- growth rate in agriculture was 5.2 percent dur- tems that facilitate effective planning, ing 1965-73 and 3.7 percent during 1973-88. control, and timely monitoring of loan re- Real agricultural GDP showed phenomenal payment growth during the 1960s and 1970s, thanks to * Operating in areas of high population the availability of land, extensive road construc- density. tion, and an active private sector increasingly involved in commodity exports. Agricultural Overview of the Institutions Figure 9.1 BAAC average annual total loans and total This section briefly reviews the evolution and deposits, 1987-93 basic institutional features of the three institu- (millions of baht) tions. Table 9.1 summarizes data on their op- erating costs from 1987 through 1994. 80,000 _ Bankfor Agriculture and Agricultural 60,0.0_ Cooperatives, Thaíland Thailand has enjoyed relative political stabil- 40,0M ity and unprecedented economic growth over the past decade. Real annual growth in gross 20,000 domestic product (GDP) averaged 8 percent * per year during 1990-94, while inflation av- ° eraged 5 percent during 1984-94. Almost 80 1987 1988 1989 1990 1991 1992 1993 percent of Thailand's population of 60.3 mil- lion lives in rural areas. The average annual Source: Authors' flindings. Best Practices: Three Success Stories 119 growth slowed in the 1980s as expansion Figure 9.2 BRI-UD average annual total loans and total neared the limits of the cultivable land frontier. deposits, 1987-94 While agriculture has become increasingly (millions of rupiah) intensive and commercial, nonagricultural 5,000 activities in rural areas have also been 4,OW expanding rapidly, and income from these 4,000 activities is increasing faster than that from 3,000 agriculture. Tctal deposits The Thai government started implementing 2,000 financial policy reforms in 1989. By 1992 all i,ooo interest rates were liberalized and the Bank of "" Tca k0rE5 Thailand's rediscounting rates were also ad- ° u U justed to reflect market-oriented rates. How- 1987 1988 1989 1990 1991 1992 1993 1994 ever, some priority sectors, including Source: Authors' findings. agriculture, still enjoy preferential rates. The BAAC was founded in 1966 to stimu- late agriculture by extending financial serv- BAAC lent mostly through large agricultural ices to the agricultural sector. It replaced the cooperatives, but repayment problems led the Bank for Cooperatives, whose funding was bank to increase its direct lending to individual limited and whose lending activities were re- farmers. stricted to agricultural cooperatives. The BAAC operates as a state-owned bank under Unit Desas of Bank Rakyat Indonesia the supervision of the Ministry of Finance; it can provide loans only for activities related Real GDP growth in Indonesia averaged 6.8 to agriculture.46 percent during 1990-94. This growth resulted The BAAC enjoys privileges because it lends in an increase in the demand for credit, which to agriculture. For example, it is exempt from contributed to the rapid growth and success of certain taxes (including income tax) and from the unit desa (village bank) program of the BRI. reserve requirements on deposits. The Bank The BRI is a state-owned bank. It ran a pro- of Thailand requires commercial banks to in- gram of directed subsidized credit for rice vest at least 20 percent of their deposits in ag- farmers until 1983. The unit desa system (BRI- riculture, either directly or through the BAAC. UD) was established in 1984 as a separate profit Banks have opted mostly for the latter, provid- center within the BRI. A general manager who ing BAAC with access to a large and consistent reports directly to the BRI board of directors source of funds. The administrative costs asso- supervises the BRI-UD. ciated with mobilizing and servicing these de- The BRI-UD is a nationwide network of posits have been borne by the commercial small village banks. The founding objectives of banks and not by the BAAC. the BRI-UD were to replace directed agricul- The BAAC enjoys substantial autonomy in tural credit with broad-based credit for any setting its operational and financial policies. It type of rural economic activity; to replace sub- has focused mainly on lending to borrowers in sidized credit with positive on-lending rates the low- to medium-income range. This strat- with spreads sufficient to cover all financial egy has been supported by a progressive cross- and operational intermediation costs; and to subsidizing interest rate policy, with higher provide a full range of financial services (sav- interest rates charged on larger loans, ceilings ings as well as credit) to the rural population. placed on loan amounts, and loans offered to All these objectives were achieved only a few small farmers without traditional collateral years after the program's inception. The BRI- through joint liability groups. At first the UD's phenomenal success in savings mobiliza- 120 Rural Finance tion is its distinguishing achievement. Other Figure 9.3 GB average annual total loans and total deposfts, financial institutions, including the Badan 1987-94 Kredit Kecamatan, which targets the extremely (millions of taka) poor among Indonesia's rural population, have 1,0 guidelines similar to those of the BRI-UD. 10,000 8,000-i Grameen Bank, Bangladesh ,0 Bangladesh is one of the poorest countries 4,000 in the world. In 1994 annual per capita in- come was US$223, and more than 70 percent 2,000 of the population lived in poverty. Over the 0 past few years, however, Bangladesh has shown steady economic growth. From 1990 to 1987 1988 1989 1990 1991 1992 1993 1994 1994 average annual real GDP grew 4.2 per- Source: Authoars findings. cent, and inflation remained below 10 percent per year. Most of the population live in rural areas: lack of access to traditional collateral, lending agriculture accounts for more than 30 percent is done exclusively through joint liability of the GDP and more than 60 percent of em- groups, tied to compulsory savings. The GB ployment. However, the population of almost has achieved phenomenal success with this ap- 130 million is growing at a rate of more than 2 proach, inspiring many other countries to copy percent per year, resulting in increasing land its efforts. pressure and decreasing farm size. Professor Muhammad Yunus started the GB Performance Assessment as an experimental project in 1976 in the village of Jobra. The project was financed by a com- This section assesses the performance of the mercial bank and was personally guaran- three institutions in achieving the primary teed by Professor Yunus. In 1983 Grameen was goals of outreach and sustainability. established as a specialized financial institu- tion under the Grameen Bank Ordinance. The Outreach Grameen Bank is not subject to the Banking Companies Ordinance or to any other law All three institutions have achieved significant related to financial institutions in Bangla- outreach in the depth and breadth of their cov- desh, nor is it subject to interest rate ceilings. erage. Their success can be measured using It has also been partially insulated from several indicators. other government policies. Today 92 percent of the Grameen Bank is owned by members; Average size of loans and deposits. The average the remaining 8 percent is owned by the size of (outstanding) GB loans was $141 in 1995. government. In the same year BRI-UD loans averaged $541, From the start the bank's main goal has been and BAAC loans averaged $1,285. The value to improve the conditions of the rural poor by of the average deposits for each bank fol- providing them with access to credit, savings lowed a similar pattern. The small size of GB facilities, and some nonfinancial social pro- loans and deposits indicates that the bank has grams. Its focus is on the lowest social strata, succeeded in its objective of reaching a poorer and the income level of its clientele is lower clientele. The BRI-UD and the BAAC target a than that of the BAAC and the BRI-UD. Be- different market (low- to middle-income). The cause of the low incomes of GB clients and their average size of loans and deposits in all three Best Practices: Three Success Stories 121 institutions has been increasing gradually. An Figure 9.4 Return on equity for BAAC, BRI-UD, and GB, impact study conducted by the BRI found that 1987-94 both the loan size and income levels of BRI-UD (percent) borrowers have increased, showing that repeat borrowers (the majority of the clients) have beeneconomicallysuccessful(Yaron1992b). 150 Women's participation. Because women usu- loo ally have less access to resources than men, a lL higher percentage of female clients usually cor- responds with a lower average loan size. About o BRI_UD 94 percent of Grameen's clients are women, sij compared to 25 percent of the BRI-UD's clients. 1 (Numbers are not available for the BAAC.) 19891990 B 1901991 t Market penetration. All three institutions have I92 1993 1994 attained high levels of penetration in their tar- get markets. The BAAC serves about 76 percent Source: Authors findings. (borrowers) of Thai farming households, the BRI-UD provides credit to approximately 5 per- By 1993 the volume of deposits was more than cent (1.9 million borrowers) of all Indonesian double that of the outstanding loan portfolio. households and extends savings facilities to as This ratio has since declined somewhat to about many-as 14.5 million households.49The GB has 1:1.8. 2.06 million clients and provides an estimated 36 percent of the total credit extended to poor, Distribution network. All three institutions em- landless people living in rural Bangladesh. ploy forms of mobile banking, which enables them to better serve clients and to keep over- Growth. The three institutions have greatly head and transaction costs down. The institu- expanded their assets and deposits over the tions have vastly increased their branch past decade. The BRI-UD deposits overtook the networks and increased the size of their staff volume of outstanding loans in the late 1980s. and the number and value of loans and deposits handled per branch and per staff member. Figure 9.5 Return on assets for BAAC, BRI-UD, and GB, These improvements have led to easier access 1987-94 for clients and gains from economies of scale (percent) because of lower per unit transaction costs. Loan processing is efficient; all three institutions dis- burse funds within two weeks of receiving an I t I I I I *¡ application. 4 3 Loan and deposit terms and conditions. All three 2 institutions apply innovative methods to over- i I _: come information and collateral problems. The 0 | j _ f f i BRI-UD BRI-UD requires full collateral, but it often uses -l BAAC local character references and has highly flex- 1987 1988 18 1 ible loan terms and various savings instru- 13019911992 1 994 ments. The BAAC and the GB both use joint liability to overcome screening, collateral, and Source: Authors' findings. enforcement problems, but they apply more 122 Rural Finance rigid loan terms. The BAAC offers various sav- ministrative costs are very low; its subsidy- ings instruments; savings in the GB are compul- dependence results mainly from its on-lending sory and tied to loans and membership. rate, which is deliberately set below that of commercial banks. The GB has much higher Transaction costs. The transaction costs of all admninistrative costs than the BRI-UD or the three banks compare favorably to those of simi- BAAC because it provides nonfinancial serv- lar institutions. The banks' administration costs ices that are not fully covered by income correspond to the type of clientele serviced, the from lending. average size of loans and deposits, the kinds of Another factor that could affect the financial services rendered, and the ages of the institu- viability of an institution is the degree of risk tions. For example, the small average size of GB associated with its loan portfolio. The BRI-UD loans and deposits and the bank's provision of and the GB both lend to clients engaged in a nonfinancial services make the GB's adminis- broad range of activities (not only agriculture- trative costs per outstanding loan portfolio and related activities) and have accordingly par- deposited higher than those of the BAAC or the tially diversified the high covariant risk BRI-UD. The larger average loan and deposit associated with lending strictly for agricul- size of the BAAC has enabled it to have lower tural purposes (see figures 9.6 and 9.7). The average operating costs than the BRI-UD and BAAC lends exclusively to agriculture-related the GB. activities, but has diversified its risk somewhat by extending loans to farmers throughout Self-Sustainability Thailand. Although the financial performance of all three Modes of Operation institutions has improved over recent years, each institution has performed differently (see This section analyses the organizational prin- figures 9.1-9.3). After incurring initial start-up ciples and modes of operation that have con- losses, the BRI-UD steadily increased its prof- tributed to the success of the three institutions. its. As indicated by the subsidy dependence It includes data on management; staff and in- index (SDI), the BRI-UD has reached its objec- centive policies; delivery mechanisms; lending tive of becoming financially self-sustainable; it policies, terms, and conditions; quality of as- has not required a subsidy since 1988. Its suc- sets-loan performance; savings mobilization cess is attributable mainly to its (a) high posi- policies; interest rate policies; administrative tive on-lending interest rate, (b) high-quality and transaction costs; and management infor- loan portfolio with minimal loan losses, mation systems. (c) high financial spreads, (d) relatively low administrative costs, and (e) extremely suc- Management cessful savings mobilization. The BAAC is still marginally subsidy- All three institutions have significant (but dependent, and the GB, while still subsidy- varying) degrees of autonomy in determining dependent, has markedly reduced its subsidy- operating policies and systems. A high level of dependence over recent years.50 The bank's transparency and accountability contributes to dependence suggests that the on-lending inter- the effective management of the institutions. est rates do not fully cover the costs of financial Although the BRI-UD forms an integral part intermediation. To reduce subsidy-depend- of the state-owned BRI, it operates as a separate ence, an RFI could increase its income by in- profit center and its management has a free creasing its on-lending rate or reduce its hand in determining its interest rates and other expenses by improving loan recovery and operating policies. The Indonesian government trimming operating expenses. The BAAC's ad- implemented financial sector reforms in the Best Practices: Three Success Stories 123 Figure 9.6 Sectoral distribution of BRI-UD loans, 1993 Figure 9.7 Sectoral distribution of GB loans since inception Other Agrícufture Ot8% 20%.... Agrculture and forestry Trading 21% L .. ~~~~~~~~~~~~~16% Al Other .6% 4% Fixed income. 28%ed income Trading Processing and 48% manufactunng e -n: 20% \ Lvestock and 20% --ñfisheries 35% Source: Authors' findings. Source: Authors' findings. 1980s that deregulated certain interest rates staff incentives and the prospects of promotion and abolished interest rate ceilings. As the reinforce individual accountability. BRI-UD struggled to become financially self- The BAAC is also state owned. The Thai Min- sustainable, market forces drove its decision- istry of Finance holds 99 percent of its paid-up making and it became increasingly subject to capital. The bank may sell up to 10 percent of competition from other financial institutions. its paid-up capital to private investors, but it Management has a high degree of autonomy has not yet attracted much interest and has only and full accountability for the performance of paid dividends since 1992. It receives some the BRI-UD. This accountability is pushed privileges (tax and reserve requirement exemp- down the line; every unit is responsible for its tions) designed to facilitate its primary goal of own lending decisions and profits. Monetary stimulating agricultural activity, but it is also subject to directives (such as defining lending Figure 9.8 Personnel costs as a percentage of average activities). Within this framework the BAAC annual total loans and deposits for BAAC, BRI-UD, enjoys relative autonomy. and GB, 1987-94 The roles and powers of different entities in the bank are clearly defined. The board of di- rectors is responsible for setting the overall pol- icy direction; an internal executive committee lo is responsible for setting operational directives. Although it is state-owned, the BAAC is run 8 largely according to the managerial principles 6 of a typical commercial enterprise. Strict con- 4 trols, a well- developed management informa- 42 o | G B tion system, branch-level accountability, and 2 ll lil l __ staff incentive policies linked to the overall ob- O BRI-UD jectives of the institution all contribute toward 1987 1971988 9919 BAC the excellent performance of the BAAC. Unlike the BAAC and the BRI-UD, which are 19 1994 both state owned, the GB is owned mostly by Source: Authors' findings. borrowers. The government owns only 8 per- 124 Rural Finance cent of the bank, borrowers own 92 percent. mine the bonuses received. During 1994-95 However, the government appoints four of the each BAAC staff member received almost five nine members of the board of directors. Even months' salary as incentive bonuses. so, the GB has much autonomy in setting op- With the rapid expansion of its business, the erational policies. In doing so it is guided pri- GB hired new employees. In 1985 it had 2,777 marily by its original financial and social employees; by December 1995 it had 12,268 principles. Decentralized loan approval and (Khandker, Khalily, and Khan 1995). The GB staff incentives tied to performance facilitate gives annual awards to the best branch man- staff accountability. agers and bank assistants according to their overall performance. Promotions depend on Staffand Incentive Policies efficiency and seniority. Because of the GB's broader social agenda and the close relation- The three institutions all employ educated staff ship it wishes to maintain with its clients, the and invest in the further development of staff GB requires branch staff to live in the villages through intensive training. The BRI-UD spent they serve. The GB provides training to its bor- 1.9 percent (1995) of its total administrative rowers and center chiefs. costs on training, the BAAC spent 1.1 percent, and the GB spent 1.5 percent (see Annex 9.1).51 Delivery Mechanisms To further improve motivation and overall productivity, all three institutions have staff All three institutions have a wide and growing incentive programs that tie employee bonuses branch network and use mobile banking to pro- and promotions to quantifiable performance vide clients with easier access to services and indicators. reduce transaction costs. The BAAC and the BRI-UD staff are among the higher-paid em- GB rely on self-help groups to achieve financial ployees in rural Indonesia (World Bank 1995a). discipline and high loan collection. The bank's staff incentive system is linked to The processing of loans is efficient in all three transparent and well-defined performance cri- institutions, although different processes are teria, such as unit profits, loan portfolio quality, applied by each. The BRI-UD allows managers and savings mobilization. Rewards are tied of village units to approve loans of up to about both to individual and group performance. US$2,800, depending on their experience. Staff can receive 10 percent of a unit's profits and Loans over this amount must be approved by a maximum of 1.5 months' salary in bonuses a branch manager. Simplicity of terms and con- per year. A unit is given a car when it reaches ditions also helps to facilitate fast (and consis- the level of Rp 2 billion in savings deposits. tent) loan approval. BAAC branch managers Competitions are run between village units. review all loan applications, although joint li- About two-thirds of the BAAC's credit of- ability groups are involved in the initial ap- ficers are college graduates. They have mul- proval stage. The GB has a participatory loan tiple roles, but each manages up to 500 clients approval process that involves group mem- (Sacay, Rhandawa, and Agabin 1996). They bers, village branch staff, and area managers. are able to do so because the bank relies on Personal banking technologies such as auto- joint liability groups and provides the offi- mated teller machines have not yet been used cers with excellent training and transporta- by any of the three institutions, although tion (motorcycles). The employee incentive electronic banking is under development. system assesses performance by the number of clients officers serve and the number of Lending Policies, Terms, and Conditions loans they extend. The officers' loan recovery and savings mobilization records, as well as All three institutions are restricted by lack of several qualitative indicators, largely deter- information about potential borrowers, prob- Best Practices: Three Success Stories 125 lems with collateral, and limitations on their The GB does not require any collateral. It is ability to exercise foreclosure. Innovative not cost effective for the GB to exercise foreclo- modes of operation have enabled the institu- sure. The bank uses joint liability and group tions to largely overcome these limitations. pressure by extending loans only to individu- The BRI-UD has no restrictions on the types als within a binding group context. No new of operations it finances, but it requires full loans are issued to the individuals in a group collateral for most loans. It is generally not while an existing loan to one group member is cost-effective for the GB or the BRI-UD to ex- in arrears. Loans are also tied to compulsory ercise foreclosure. The BRI-UD has highly flex- savings (individual and group-based savings) ible loan terms and conditions: minimum and which serve as a form of collateral. Goods pur- maximum loan amounts are almost the only chased with funds borrowed from the GB es- restrictions the bank imposes. The problem of sentially remain the property of the bank until not having adequate information about poten- full repayment of the loans. tial borrowers is sometimes overcome by solic- iting recommendations from village leaders or Quality of Assets and Loan Performance from other borrowers. The BAAC has three categories of lending: The collection performance of all three institu- normal lending, special-projects lending, and tions is exceptionally good. Each institution policy lending. Normal loans and special-pro- defines arrears differently, which makes com- ject loans are initiated by the BAAC and ad- parison complex. Table 9.3 summarizes the ministered according to the BAAC's internal policies and procedures the banks apply. (Fig- procedures; policy loans are implemented in ure 7.3 and table 7.1 in chapter 7 show the response to government directives. The BAAC BAAC's system for monitoring loan collection lends through large cooperatives and directly through an advanced age analysis system.) to individuals. The performance of cooperative The loan collection rate in 1995 (as a percent- loans has been poorer than that of direct loans, age of current maturities falling due and old and the BAAC has been phasing-out coopera- overdues) was estimated at 99 percent for the tive loans, as illustrated in table 9.2. Full col- BRI-UD, 90 percent in 1995 for the BAAC lateral is required from individual borrowers, (much of the BAAC arrears are repaid belat- but the BAAC makes allowance for individual edly), and 99 percent for the GB. The BRI-UD loans with joint group liability when borrowers and the BAAC consider a loan to be in arrears are unable to provide traditional collateral. if the latest installment is not paid on time, but Loan terms and conditions are rigid to instill the GB allows a one-year grace period. The financial discipline. The BAAC has no mini- details of the three institutions' operation poli- mum loan amount requirement, but it does cies differ, but they share some important simi- have a maximum amount that varies depend- larities with regard to loan collection: all three ing on the type of client and investment. institutions hold clients accountable for loan Table 9.2 Distribution of BAAC loans, selected years (percentage share of total) Client 1975 1983 1992 1994 Farmers and other individuais 55.1 77.7 90.8 92.0 Agricultural cooperatives 35.8 22.2 9.0 7.8 Farmers' associations 9.1 0.2 90.8 0.2 Total 100.0 100.0 100.0 100.0 Source: Sacay 1996. 126 Rural Finance Table 9.3 Loan tebms and pertorrnance BRI-UD BAAC GB Definition of arrears Latest installment not paid on Latest instaliment not paid on One-year grace perod due date due date Estimated loan collection rate 98.9% 90% 98.6% on-time collection for 1995 Management information Well developed WelI developed Well developed system Kind of character references Local residents, other Group lending Group lending required borrowers, local government officiais Collateral requirements 100% (but no cost-effective 100% on individual loans; none None, but compulsory savings option of foreciosure) for groups tied to loan Group lending with joint liability No Yes Yes Individual lending (without joint Yes Yes No liability) Future loan eligibility None if in default None (for entire group) if None (for entire group) if individual defaults individual defaults Client incentives/penalties Rebate of 0.5% per month on Penalty of 3% a year on arrears None orginal loan amount for prompt payment Staff incentives Yes Yes Yes Source: Auttors' findings. repayment through collateral or joint liability, participants in cooperative loans appear to all three have established various monetary benefit from the free-rider syndrome, with lim- and other incentives and penalties for both ited accountability and little incentive to repay staff and clients to encourage timely repay- loans. ment, and all three have excellent management information systems, which enable them to Savings Mobilization Policies track loan performance and implement and ef- fectively manage their incentive systems. The BRI-UD's and the BAAC's success in sav- The difference in performance among the ings mobilization can be partly attributed to the various types of loans extended by the BAAC various savings instruments they offer. In ad- is worth noting. The performance of BAAC dition to individual savings, the BRI-UD at- loans extended directly to clients has been tracts savings from community groups (such as very good, but the performance of loans to co- schools and municipalities). In 1985 the BAAC operatives and farmers associations, while im- launched a successful special savings deposit proving, has been poor (table 9.4). Some facility targeted at middle- and higher-income Table 9.4 Repayment performance of BAAC clients, selected years (percentage of amount scheduled to be repaid by year) Client 1986 1989 1992 1994 Farmers and other individuais 77.1 86.8 89.0 86.0 Agricultural cooperatives 45.3 68.9 68.8 68.4 Farmers' associations 41.2 72.8 71.3 60.0 Source: Sacay, Rhandawa, and Agabin 1996. Best Practices: Three Success Stories 127 Figure 9.9 Average annual value of total deposits UD to mobilize substantial savings with as a percentage of total loans for BAAC, BRI-UD, attractive deposit interest rates. As a result, and GB, 1987-94 the BRI-UD has become completely finan- cially self-reliant and subsidy- independent. The BAAC has enjoyed the privilege of not being subject to interest rate ceilings, which 250 / / g f were applicable in Thailand until 1992 (Sacay, Rhandawa, and Agabin 1996). However, the 200 / .BAAC has always tried to keep interest rates 150 one or two percentage points below that of 100 commercial banks. Tax and reserve-require- ment exemptions and special discount rates re- so ceived from the Bank of Thailand enabled the 0 AA - BMC BAAC to maintain a relatively low on-lending 1 9 ~~~~~~~~rate. However, the BAAC remaíns partially 1988 1989 1991 GB subsidy-dependent, although remarkable cuts 19923 1994 in administrative costs, improved loan collec- Source: Authiors' findings. tion, and increases in voluntary savings have led to a reduction in its dependence on subsi- dies in recent years. groups. The BAAC's staff incentives are linked The GB is not subject to the Banking Com- to savings mobilization performance. panies Ordinance that is applicable to com- The GB has a different policy for deposits, mercial banks and other financial institutions (which are primarily compulsory). Members in Bangladesh, so interest rate regulations en- are required to add one taka to their savings forced by the Central Bank of Bangladesh are account every month, and 5 percent of any not applicable to the GB. The GB serves a amount borrowed is placed in an interest-bear- much poorer clientele than the BRI-UD or the ing savings account owned by the group they BAAC and provides various nonfinancial belong to. These savings act as a kind of insur- services. It has a much higher costing struc- ance fund from which group members can lend ture, and its interest rate policy does not aim to each other; they also serve as a form of com- to recover all of its financial and operational pensating balance for the GB. Positive interest costs. The GB has remained subsidy-depend- rates are paid on the savings, although the rates ent, despite charging a relatively high interest are significantly lower than those charged on rate on the bulk of its loan portfolio (20 per- loans. This policy should be seen in the context cent, an increase from 16 percent in 1991). of the GB's target clientele, which is the lowest However, this subsidy-dependence has re- economic strata of the population. duced markedly in recent years. It is difficult to determine exactly what the GB's effective Interest Rate Policies interest rate is, because borrowers' obligatory compensating balances effectively reduce the The BRI-UD charges a flat interest rate of 1.5 loan amount and (ignoring the savings' poten- percent per month, equivalent to an effective tial benefit as insurance) increase the effective rate of 32.7 percent per year. An additional 0.5 lending interest rate (see figure 9.10). percent per month is charged, which is returned to the client (through the savings account) for Administrative and Transaction Costs timely repayment of installments. The on-lend- ing rate has been sufficient to cover the full costs The administrative and transaction costs of of mobilizing funds and has allowed the BRI- RFIs are often high because of small loan and 128 Rutral Finance Figure 9.10 Financial costs as a percentage of average Management Information Systems annual total loans for BAAC, BRI-UD, and GB, 1987-94 All three institutions have well developed management information systems, but the sys- tems,differ markedly. Each institution has de- veloped a system that suits its particular 30 ~~~~~~~~~~~methods of operation, facílitates transparency, and supports decisionmaking. The effective- 20 - / i li l | l - l ness of financial policies and operating proce- dures can be determined and adjusted as 10 needed. The tracking of clients' repayment per- BRI-UD formance enables the institutions to improve O BAAC loan collection, penalize defaulters, and speed 1988 8 9 1the loan approval process. The institutions' 1988 1989 199 GBsystems facilitate staff incentive programs by 993 1994 enabling management to track the perform- Source: Authors' f¡ndings. ance of individual branches and staff members and to tie that performance to the organiza- tion's overall performance. deposit amounts, the geographic dispersion of clients, the efforts needed to obtain information Potential Improvements about clients, and the use of nontraditional col- lateral. Nonetheless, controlling costs is crucial The three RFIs reviewed in this chapter have to the financial performance of RFIs be- made exceptional progress in improving their cause their income-generating capacity is outreach to their target clienteles and in reduc- often restricted by interest rate ceilings or other ing or eliminating their dependence on subsi- limitations. dies. Could altering some of their operations The BAAC has succeeded in keeping its costs result in even stronger performance? exceptionally low by offering relatively large Long-standing populist political pressures in loans and group-lending operations, which re- Thailand that continuously test the level of the sult in lower transaction costs. (It has also re- BAAC's managerial autonomy are (a) should the ceived funds that incurred almost no BAAC become a rural financial institution administrative costs in the form of obligatory or remain a specialized agricultural bank, and deposits from other banks). Because of its ex- (b) should interest rate policies be designed to ceptionally high market penetration, the BAAC achieve optimal resource allocation or should the is well known. It does not engage in costly BAAC continue to cross-subsidize small loans promotional activities, and its large clientele using distorted interest rates as a "compen- enables it to gain from economies of scale. In sating" mechanism? Eliminating cross-subsidi- contrast, the BRI-UD is growing fast and has zation, changing the BAAC from a specialized placed much emphasis on voluntary savings, agricultural credit institution to an RFI and ap- which are costly to mobilize. The GB has a plying interest rates that reflect the nonsubsi- much smaller average loan and deposit size dized cost of capital, loan-specific credit risk and than either the BRI-UD or the BAAC, and it administrative costs, would benefit the rural provides nonfinancial services and extensive economy and enhance the efficiency of rural fi- training to staff and clients. These factors lead nancial intermediation in Thailand. to substantially higher administrative costs for The case of the GB is unique (and more com- the GB. plex) because it provides nonfinancial services Best Practices: Three Success Stories 129 to a poorer clientele. However, the adequacy self-sufficiency without damaging its outreach of the GB's on-lending interest rate in light of performance. its dependence on subsidies is a concern. As- The BRI-UD's performance and modes of sessing the adequacy of the bank's on-lending operation seem to reflect the best-case sce- interest rates is of widespread and crucial im- nario in rural financial intermediation with portance to all credit schemes that target pov- respect to subsidy-independence. Moreover, erty alleviation and look toward the GB as a the BRI has had a negative SDI in recent years successful role model.52 (that is, the BRI-UD could have markedly re- Increasing on-lending interest rates is usu- duced its on-lending interest rate while still ally the most effective way to decrease the sub- maintaining subsidy-independence). This situ- sidy-dependence of an RFI. Such an increase ation prompts the question: what would have (when not curtailing loan demand or having been the impact on the rural economy, and an adverse impact on loan collection) usually particularly on BRI-UD clients, if the value of results in better outreach because scarce re- negative subsidies had been used to decrease sources reach a larger number of clients. The the spread between the BRI-UD's on-lending GB applied an annual 16 percent interest rate and deposit interest rates, instead of subsidiz- on its regular loans until the middle of 1991, ing other BRI non-unit desa activities? This when it raised the rate to 20 percent a year. question is of the utmost importance because There is no indication that the increased inter- the cross-subsidization within the BRI (as est rate reduced loan demand or increased reflected in the negative SDI) results in adverse arrears. income distribution, with small-scale rural The issue becomes more complex when the entrepreneurs subsidizing the more affluent GB's lending rates are adjusted for inflation. The clientele. inflation adjustment presents a much more volatile picture of real lending interest rates (see The Environment table 9.5). The data suggest that real lending rates Understanding how the three institutions ranged from 5.5 percent to 20.0 percent- have overcome limitations in their social, po- a wide range. If the high real lending rate of litical, and economic environments may fur- 1993 had no discernible adverse impact on ther guide the design of future rural finance repayment rates or on clients' demand for efforts. credit, then a plausible inference is that the real The three RFIs operate under relatively sta- lending rate could have been higher in other ble political and economic conditions, with in- years. That is, the GB could have increased flation rates consistently below 10 percent. The its nominal lending rates to improve its economic bases of Bangladesh, Indonesia, and Table 9.5 Nominal and real average annual GB interest rates on loans, 1989-94 (percent) Interest rate 1989 1990 1991 1992 1993 1994 Inflation in Bangladesha 9.9 8.1 7.2 4.3 0.0 3.6 GB average annual interest rate charged on regular loans Nominal lending rate 16.0 16.0 18.0 20.0 20.0 20.0 Real lending rate 5.5 7.3 10.1 15.1 20.0 15.9 a. As measured by the consumer price index. Source: Authors' findings. 130 Rural Finance Thaíland differ substantially. Per capita GDP Although the legal and regulatory frame- in 1994 amounted to US$223 in Bangladesh, works that govern the three RFIs' operations US$892 in Indonesia, and US$2,417 in Thai- differ substantially, each has overcome the land. The countries' annual average GDP grew shortcomings of its environment or devised at different rates over the past five years, with ways to insulate itself from certain aspects of Bangladesh growing slowest at 4.2 percent, the surrounding environment. The institu- Indonesia growing at 6.8 percent, and Thailand tions have developed joint liability groups, leading at 8.3 percent. obligatory savings, or other methods to over- Agriculture is important to the economies come collateral and enforcement problems. of all three countries. Agriculture accounts for In 1994 and 1995 the GB overcame the ex- 65 percent of the labor force and 30 percent of ceptionally poor repayment culture in Bang- the GDP in Bangladesh, 55 percent of the labor ladesh-most other Bangladeshi financial force and 17 percent of the GDP in Indonesia, intermediaries had very poor recovery rates and 64 percent of the labor force and 10 percent on agricultural loans during 1994 and 1995- of the GDP in Thailand. As would be expected, and both the BAAC and the GB managed to the percentage of the population living in rural prevent interest rate ceilings and other finan- areas is high for all three countries. More than cial sector restrictions from being imposed on 80 percent of the population lives in rural areas them (see table 9.6). in Bangladesh and Thailand; close to 70 percent of Indonesia's population is rural. To im- Conclusion prove rural access to the financial markets, the three RFIs instituted various forms of mo- The success of the three institutions can be bile banking. They continue to bring financial ascribed partly to the relatively favorable markets closer to rural clients by developing macroeconomic conditions under which they and planning the implementation of electronic have been operating. Nonetheless, their abil- banking. ity to overcome limitations in their environ- Table 9.6 Economic and demographic indicators for regions served by the BAAC, the BRI-UD, and the GB (percent unless otherwise specífied) BRI-UD BAAC GB Indicator (Indonesia) (Thalland) (Bangladesh) Economic Average annual inflation, 1984-94 8.90 5.00 6.60 Inflation, 1994 8.53 4.37 3.58 Average annual GDP per capita, 1994 (U.S. dollars) 892 2,417 223 Average annual GDP real growth, 1990-94 6.81 8.27 4.18 Agriculture as a percent of GDP, 1994 17 10 30 Average annual agrculture growth, 1980-93 3.20 3.80 2.60 Agriculture as a percent of labor force, 1990 55 64 65 Demographic Population in millions, JuIy 1995 203.6 60.3 128.1 Population growth, 1995 1.56 1.24 2.32 Population in rural areas, 1994 66 80 82 People per square kilometer, JuIy 1995 33 398 957 Literacy rate for population, 1995 84 94 38 Literacy rate for men, 1995 90 96 49 Literacy rate for women, 1995 78 92 26 Source: Authors' findings. Best Practices: Three Success Stories 131 ments by implementing innovative policies Guiding principles can be deduced from the and operating methods provides invaluable institutions' operations, but care must be taken lessons for the design of future rural fi- in replicating their operations: a solution ade- nance projects. The elimination or remark- quate in one socioeconomic environment will able reduction of subsidy-dependence by all not necessarily be suitable in another. It should three institutions can be attributed to effi- not be forgotten that the income levels and ciency gains rather than to changes in the income-generating activities of the target clien- business environment in which the institu- tele largely determine the effectiveness of spe- tions operate. cific modes of operation. 132 Rural Finance Annex. Summary of performance and operating methods of BRI-UD, BAAC, and GB Institution Year of establishment 1980 1966 1976b Fiscal year ends Fiscal year ends Fiscal year ends December 31 March 31 December 31 Objectives To provide credit to small, To provide financial To improve the economic income-generating activities assistance to farmers for conditions of the rural poor (pñmarly rural) and savings agricultural and agriculture- facilities for households related activities Type of institution Autonomous rural savings Financial institution Financial institution and credit program within a financial institution Ownership Bank Rakyat Indonesia State-owned Independent: (state-owned) 92% borrowers; 8% state-owned Financial services Loans and savings deposits Loans and savings deposits Loans and (compulsory) savings deposits Other services Payment of school fees and Marketing services; Social intermediation and electrical bilis technical assistance and training on various issues advice (on fertilizers, for such as dowry, sanitation, example) education, and nutrition Target clientele Rural low- and middie- Low- and middie-income Rural poor income; broad-based farmers and farmers' associations Country Indonesia Thailand Bangladesh GDP per capita, 1994 Rp 1,963,136 (US$892) B 60,258 (US$2,436) Tk 8,988 (US$223) GDP average annual real growth, 1989-94 (percent) 6.8 8.5 4.2 Annual inflation rate, 1994 (percent) 8.5 5.0 3.6 Average inflation rate, 1991-94 8.9 4.2 4.6 Clients and staff Number of clients or members 14.5 million0 4.3 million clientsd March 1996: 2.1 million members (cumulative since 1983) Agriculture or related loans (including livestock and fisheries) (percent) 18 100 1994: 61e Total target clientele serviced Percentage of total Registered members: 76% About half of allí villages in (percent) households: of farming population; share Bangladesh Loans: 5 of total credit to farmers: Deposits: 20 36%1 Women borrowers (percent) 25 Not available March 1996: 94 Number of staff 16,916 11,379 March 1996:12,268 Best Practices: Three Success Stories 133 Town units 3,5209 365 112 (area offices) Village posts or units 437 840 1,056 Ratio of village posts to town units 0.12:1 2.3:1 9.43:1 Mobile banking in use?h Yes Yes Yes Loan outreach Number of loans outstanding (millions) 2.3 million' 3.1 million 2.1 millioni Volume of loans outstanding (annual Rp 2,825 billion B 94,453 million Tk 11,798 million average) (US$1,224 million) (US$3,817.8 million>k (US$289 million) Real annual average growth rate of total 41.6% (1993-95) 22% (1993-95) 35% (1992-94) assets durng last three years Minimum loan size Rp 25,000 Not applicable Not applicable Maximum loan size Rp ?5 million B 60,000 to B 5 millioni Tk 15,000 (US$372) (US$11,364) (US$2,425 to $202,000) Average outstanding loan size Rp 1,247,673 (US$567) Individual farmers: Tk 5,708 (US$142) B 31,800 (US$1,285) Average outstanding loan as percentage 54 52 64 of GDP per capita Value of loans per staff member Rp 167 million B 9.3 million Tk 958,428 (US$75,909) (US$377,527) (US$23,812) Number of loans (al types) per staff 134 270 245m member Savings outreach Average annual volume of savings Rp 5,624 billion B 68.8 billion Tk 5,366 million (US$2,556 million) (US$2,780.9 million) (US$133.3 million) Average annual savings as a percentage of 1995:199 1994: 66.5 1995: 45.6 average annual outstanding loan portfolio 1989:131 1988:42 1989: 55 Number of savers 14.5 million 3.1 million 2.1 million" Value of average savings account Rp 388,383 B 22,389 Tk 2,605 (US$177) (US$905) (US$65) Value of savings deposits per staff member Rp 332.5 million B 6.0 million Tk 437,374 (US$151,136) (US$244,391) (US$10,866) Number of savers per staff member 856 270 167 Nominal annual deposit interest rate Not established 5.0 to 10.75 8.50 (percent) Profitability Return on assets = 3.6 4.8 0.52 0.55 0.1 0.14 (Net income before tax) (Average annual assets) (percent) 134 Rural Finance Annex (continued) Return on equity = 45 127 7.6 7.1 3.4 1994:10.4 (Net income before tax) (Annual average equity) (percent) Interest rate spread (a) Income from lending as percentage of average outstanding loan portfolio 28.0 30.9 14.0 10.4 12.0 16.0 (b) Financial expenses as percentage of average outstanding loan portfolio 11.2 9.3 9.1 6.3 4.7 8.1 (a) - (b) spread 16.7 21.7 4.9 4.1 7.3 8.0 Expenses Total operating costs as percentage of: Annual average assets 10.9 5.7 3.0 2,9 9.3 7.0 Annual average savings 12.4 6.1 - 5.8 - 23.1 Annual average outstanding loan portfolio 15.2 13.5 4.7 3.5 16.7 10.6 Annual average outstanding loans plus savings 6.8 4.2 2.9 2.2 - 7.6 Personnel expenses as a percentage of: Average annual assets 7.0 2.9 1.9 1.7 5.0 4.7 Average annual outstanding loan portfolio 9.7 6.9 3.0 2.0 9.0 7.2 Other administrative expenses as percentage of: 2.6 2.2 1.1 1.2 4.3 2.2 Annual average assets Annual average outstanding loan portfolio 3.6 5.2 1.7 1.5 7.7 3.4 Costs of training staff as percentage of:° Total administrative costs 7.3 1.9 0.58 1.08 28.1 1.5P Annual average assets 0.8 0.1 - 0.03 2.6 0.1 Annual average outstanding loan portfolio 1.1 0.3 - 0.04 0.6 0.2 Financial expenses as percentage of: Annual average assets 9.9 8.8 9.1 5.5 3.0 5.3 Annual averagesavings 11.2 9.5 - 10.9 - 17.7 Annual average outstanding loan portfolio 13.7 20.9 9.6 6.5 5.3 8.1 Terms and conditions of loans Eligibility requirement and types of activities No restrictions on Farmers and farmers' Households owning less financed operations financed associations for agriculture than 0.5 acres of land, or and related activities assets worth less than 1 acre of land Best Practices: Three Success Stories 135 Collateral requirement 100%. Usually land (difficult Joint liability and personal Indirectly through forced to repossess); others guarantee for loans not savings, joint liability, and include 80% of deposit exceeding B 100,000 per peer pressure amounts, or lease contracts person; over B 100,000 titie to tangible assets Loans provided to cooperatives or farmers' associations? No Yesq No Lending to individuais? Yes Yes Yes-to individuais within groups, with joint liability Lending to groups? No Yes Yes' How are groups formed - Not applicable By members By members Size of groups Not applicable 5-O 5 Loan application processing (from submission to disbursement) For new borrowers 2 weeks (maximum) 1 week 1-2 weeks For repeat borrowers 2 days 1 week 1-2 weeks Loan approval by Manager of unit desa, Credit officer less than Area manager on usually to Rp 5 million; B 60,000; branch manager recommendation of group branch manager over Rp 5 less than B 300,000; members and bank workers millions president less than B 5 million; board less than B 5 million Local leadership or local officiais involved in approval of applications? Yes Yes No How? Character reference for Character reference for Not applicable borrower borrower Repayment frequency Flexible-mostly monthly or Short-term loans: upon Weekly quartery matuñty (11 months) Gradual increase in borrower's eligibility to Yes Yes Yes subsequent lending? How are loans monitored? Special department at Visits by credit officers; Weekly loan supervision headquarters in charge of client meetings meetings held openly in monitoring plus normal village centers branch regional office supervision Loan repayment incentives orpenaities Social or peer pressure on borrowers to repay loans? Yes-partialu For group loans Yes-all loans Disciplinary measures for non-repayment of loans? Yes Yes No Penalty interest rate? Yes, loss of 0.5% per Yes, 3% per year on arrears None month interest rate (on original loan amount) for timely repayment Access to future loans? No Nov Now 136 Rural Finance Annex (confinued) Foreciosure or repossessions? Yesx Not established Not applicable Staff incentives for loan collection? Yes-annual profit bonus Yes-annual bonus and Yes-based on overall up to 1.5 month's salary promotionsY performance, profit sharing and special awards to staff (up to 10% of branch profits) of top performing units On-lending interest rates Nominal quoted lending interest rate (percent) 1.5 per monthZ 8.0 to 14.5 per year 2.0.0 per yearaa Effective (nominal) annual interest rate (percent) 32.7 8.3 to 15.5 20bb Real interest rate (percent per year) 23 2.6 to 6.9 16.0 Typical loan maturity 1-3 years 11 months (for short-term 1 year loans) Formal annual nominal on-lending interest rate prevailing in the country (percent) 1994: 21 1 Oto 14.75 14 Informal annual interest rate prevailing in the country (percent) Not established 25-60 106 Maximum legal on-lending interest rate? No No Yescc Loan performance Definition of arrears used by the Latest instaliment not paid Latest installment not paid One year overdue institution on time on time (Annual loan collection) Estimated 95 Approximately 90dd 98.2ee (Old overdues + current maturities that fall due in the year) (percent) (Arrears) 6.5 8.3 3.6 (Total outstanding loan portfolio) (percent) (Annual provisions for loan losses) 4.9 1.0 1.2 (Total outstanding loan portfolio) (percent) (Interest paid) 77 61 46 (Interest eamed) (percent) Note: Uniess otherwise indicated, the figures and financial data in this annex were obtained from the insttut¡ons concemed or are the authors' calculations. The corresponding end-of-perod exchange rates are as follows (uniess stated differentiy): a. BRI-UD developed the General Rural Credh, or the 'KUPEDES," Program in 1983, but it did not provide financiai services until 1984. b. The Grameen Bank originated as a project of the Rural Economics Programme of the University of Chittagong in 1976, financed by a nationalized commercial bank and personally guaranteed by Professor Muhammad Yunus. In 1983 a govemment ordinance transfommed it to a specialized financial institution for the rural poor. c. Based on number of savings accounts. d. BMC financial statements, Fiscal 1994: farmer clients: 3,071,545; cooperatives and associations: 1,321 with membership of 1,238,712; total members: 3,072,866 or 4,310,257 clients. e. Reflects disbursement of loans during 1983-1994. f. BAAC financial statements, Fiscal 1994, and Sacay, Rhandawa, and Agabin 1996. g. Numbers of 'subdistrict" units or 'unit desas." h. This measure refers to the ability of staff to be in villages and poor neighborhoods daily or weekly, vishting borrowers, explaining requirements to potential clients, disbursing loans, and collecting repayments. i. Based on number of borrowers. j. Based on number of members. Best Practices: Three Success Stories 137 k. Individual farmers: 1994: B 97,680 million; 1995: B 75,068 million Associations: B 8,474 million; B 7,684 million Total: B 106,154 million; B 82,752 million. 1. The loan size depends on the type of client: individual, group, or agricultural association. m. Reflects number of borrowers per employee. n. Aithough no data on the number of savers are available, this should be the minimum number since saving is obligatory. o. Whereas the BRI-UD provides training only to staff, the GB and (to a lesser extent) the 8AAC provide training to borrowers. p. The number of new recruits, and therefore training expenditure on new recruits, dropped substantially after 1992. q. Cooperatives have had dismal loan collection records, and the BAAC has taken measures to reduce the share of cooperatives and fammers' associafions in its total lending portfolio. r. Loans provided to indMduals within group. s. Unit desa ¡s the village unit. The average unit desa serves 17 to 18 villages. t. Each area manager supervises about 10 to 15 Grameen Bank branches. The GB's whole loan approval process ¡s complex and participatory, involving group members, a bank worker, and an area manager. u. The eflecfiveness of social pressure varies widely among different geographic regions. v. Entire group is disqualified if individual member defaulfts. w. Entre group is disqualifled if individual member defauits. x. Foredosure is legally possible, but the fime-consuming, costiy process ls rarely used. y. In 1994195 BAAC staff received up to five months' salary based on the overail performance of the BAAC. z. Flat rate of 1.5 percent per month, excluding the upfront penalty fee of 0.5 percent per month, returned if all payments are made on time. aa. Benjamin 1994. Calculation excludes impact of compulsory saving, which will effectiveiy increase the rate. bb. For irregular borrowers the GB calculates interest at 20 percent on the basis of a declining balance. For the regular borrowers GB collects Tk 2 every two weeks as interest against Tk 1,000 of loan disbursement with a grace period of two weeks. cc. The temms of the Grameen Bank Ordinance of 1983 exempt the GB from interest rate ceilings. dd. BMC's direct loans to individual farmers have a much better loan performance (close to 90 percent) than the loans to fatmerse cooperatives and asso- ciabons (60 to 70 percent). ee. Khandker, Khalily, and Khan 1995. This figure ¡s for general loans (excluding collective loans) in 1992. APPENDIX The Subsidy Dependence Index of Rural Financial Institutions T wo main problems face the analyst who subsidized RFI paid only 2 percent a year on must rely on conventional accounting central bank rediscounting facilities instead of data to measure the financial perform- the prevailing market deposit rate of 12 percent ance of rural financial institutions (RFIs). First, a year, the accounting profit and the financial on top of the expenses and income (including ratios measuring the RFI's profitability would reimbursements of specific expenses by states not convey that such ratios were obtained only or donors) that appear in the income state- because of the significant subsidy embodied in ments of RFIs, there are other expenses and the cheap central bank rediscounting facilities. incomes that remain unrecorded. Second, there Providing an RFI with concessional funds is is no design in conventional accounting prac- the most common method of subsidization, yet tices to reflect and appropriately report on all calculating the value of the subsidy implicit in types of subsidies received by RFIs. 54 the RFI's concessionally borrowed funds re- Conventional accounting practice measures quires information not included in the RFI's the cost of funds at actual cost. The opportunity financial statements. The same is true for the cost of an RFI's borrowed funds-that is, the RFI's equity. cost the RFI would have to pay for its funds if The profit maximizer does not differentiate access to concessional funds were eliminated- between profit that is partially subsidy- is not taken into account. The subsidy depend- dependent and profit that is fully subsidy- ence index (SDI) calculation assumes that the independent so long as continued subsidi- volume of an RFI's outstanding loan portfolio zation is ensured. It is, however, crucial in remains unchanged. Hence, the change in the assessing an RFI's performance. Determining cost of funds is caused by substituting conces- the social cost of RFI operations, of which sub- sionally borrowed funds with voluntary sav- sidies constitute a significant share, is essential ings, obtained at a market deposit interest rate. in determining the social justification for RFIs' Thus, if the central bank loans to an RFI at 2 continued existence and operations, because percent, conventional accounting practices list rural RFIs are often public or quasi-public in- the cost of the loan at 2 percent a year. How- stitutions. To understand the futility of the cur- ever, if the cost of alternative nonconcessional rent financial reporting system, one might funds is 12 percent a year, the SDI considers consider the meaning of a 20 percent return on the 10 percent difference in interest rates on equity when 50 percent of the RFI's financial those funds and identifies this as a subsidy obligations constitute concessional borrowed received by the RFI. The rationale is that if the funds from the central bank (rediscounting 139 140 Rural Finance facilities); the funds carry an interest rate sig- ity level, represented by interest earned on nificantly below market deposit interest rates. its loan portfolio (similar to calculations of Moreover, one-third of the RFI's payroll cost, effective protection domestic resource cost 80 percent of its loan losses, and all training or job creation cost) expenses are assumed by the government. * Tracking an RFI's subsidy dependence Breaking away from applying the financial over time prices of inputs and outputs and instead using o Comparing the subsidy dependence of shadow prices reflecting the social cost of in- RFIs providing similar services to a similar vesting in the real goods sectors has become clientele. common in assessing the social desirability of Lenders' dialogue with borrowing countries investments. Applying economic shadow can be enriched by using the SDI routinely to prices permits calculation of the economic rate measure an RFI's performance during ap- of return, which often differs from the financial praisal, supervision, and completion of pro- rate of return. Applying the SDI calculation jects. As with any other financial measurement achieves a similar goal. The SDI measures more tool, however, the SDI is only as accurate as accurately the social cost involved in an RFI's the data used to compute it. continued operation. There is, however, a dif- The SDI requires the application of certain ference between calculating the economic rate procedures as well as judgment. Consistency of return and applying the SDI. The SDI does from period to period is more important than not fully assess and measure the social benefits the absolute accuracy of the figures used to of resource allocation through an RFI to the compute the SDI. The SDI is a ratio that meas- real goods sectors. The SDI, however, better ures the percentage increase in the average on- estimates the social cost of subsidies by apply- lending interest rate required in a given year ing approximate market interest rates to the to allow an RFI to eliminate subsidies and to financial resources used by the RFI. achieve a return on equity equal to the approxi- The objective of the SDI methodology is to mate nonconcessional borrowing cost. The in- provide a comprehensive method of measur- dex assumes, for simplicity, that an increase in ing the overall financial costs of operating an the on-lending interest rate is the only change RFI and quantifying its subsidy-dependence. needed to compensate for loss of subsidy, but The SDI methodology avoids overreliance on it is clear that any saving in operational costs the financial profitability ratios of conventional or reduction in loan losses would result in a accounting procedures in the financial analysis decreased SDI. Although removing the subsi- of RFIs. The SDI aims at providing a public dies received by an RFI is not always politically interest analysis of RFI financial performance feasible or desirable, measuring them is always and subsidy dependence. Using the SDI allows warranted, economically and politically. for a full accounting of the overall social costs Calculating the SDI requires aggregating all entailed in operating an RFI, including the full the subsidies received by an RFI. The total value of all subsidies received by the institu- amount of the subsidy is then measured tion. The SDI makes explicit the subsidy against the RFI's on-lending interest rate mul- needed to keep an institution afloat; this infor- tiplied by its average annual loan portfolio, be- mation is not gained through conventional ac- cause lending is the prime activity of a counting reporting. The SDI is a user-friendly supply-led RFI. Measuring an RFI's annual device that is simple to calculate because it subsidies as a percentage of its interest income does not require collecting detailed informa- yields the percentage by which interest income tion about an RFI's operational costs. The SDI would have to increase to replace the subsidies; helps in: the measurement also provides data on the * Placing the total amount of subsidies re- percentage points by which the RFI's on-lend- ceived by an RFI in the context of its activ- ing interest rate would have to increase to Appendix 1Te Subsidy Dependence Index 141 eliminate subsidies, and on the total value of I = The weighted average on-lending in- the subsidy. terest rate earned on the RFI's loan In the computation of the SDI the amount of portfolio. the annual subsidy received by an RFI is de- LP * I = Annual average interest rate earned fined as: on annual average loan portfolio = in- terest earned on the loan portfolio as S = A (m - c) + [(E * m) - pl + K presented in the income statement, (adjusted if need be). where The SDI cannot clarify how a subsidy was S = The annual subsidy received by the used or whether it benefited most clients or RFI went mainly to an inefficient bureaucracy. A = The RFI's outstanding concessional Determining how much of a subsidy is con- borrowed funds (annual average) sumed by bureaucracy requires far more de- m = The interest rate the RFI would be as- tailed data and even then is often subject to sumed to pay for borrowed funds if interpretation. The advantage of the SDI is its access to borrowed concessional simplicity. It focuses exclusively on the in- funds were eliminated take subsidy, that is, the value of subsidy c = The weighted average annual conces- received by the RFI. The SDI should be seen sional rate of interest actually paid by in some instances as a lower bound, because the RFI on its average annual conces- full financing of the activities of an RFI whose sional borrowed funds outstanding financial performance is dismal is likely to be E = The average annual equity difficult at current market borrowing rates. P = The reported annual (before tax) profit However, calculating this lower bound is vi- (adjusted, when necessary, for loan loss tal for ascertaining the RFI's progress toward provisions, inflation, and so on) self-sustainability and the social desirability K = The sum of all other annual subsidies of its continued subsidy-dependence. received by the RFI (such as partial or An SDI of zero means that an RFI has complete coverage of the RFI's opera- achieved full self-sustainability. An SDI of tional costs by the state). 100 percent indicates that a doubling of the average on-lending interest rate would be re- The financial ratio that is suggested as an quired to eliminate subsidies. An SDI of 200 SDI is: percent means that a threefold increase in the on-lending interest rate would be required to SDI S eliminate subsidies. A negative SDI indicates LP * I not only that an RFI has achieved self-sus- tainability, but that its annual profits, minus where its capital (equity) charged at the approxi- mate market interest rate, exceeded the total SDI = The index of the RFI's subsidy de- annual value of subsidies, if subsidies were pendence received by the RFI. A negative SDI also im- S = The annual subsidy received by the plies that the RFI could have lowered its RFI average on-lending interest rate while simul- LP = The RFI's average annual outstanding taneously eliminating any subsidies received loan portfolio in the same year. Notes 1. The imperfect operation of financial markets is stemmed more from uncertainty regarding the distribu- indicated by cases of incomplete diversification of idi- tion of potential returns (given the vagaries of nature osyncratic risks (Rashid and Townsend 1994). and of commodity markets) than from lack of knowl- 2. Savings are deposited in exchange for a promise to edge about the technologies. return the deposits ata given date, usually with interest, 7. Other influences were W. W. Rostow's (1963) and premium payments are made in exchange for a characterization of the take-off into self-sustained promise to indemnify insurees under contractually growth by industrialized economies and Prebisch's specified circumstances. Under an Islamic banking re- (1950) identification of a long-term decline in the terms gime, fees or a share of returns may be promised in lieu of trade for producers of agricultural commodities rela- of interest. However, the promissory feature of the tive to producers of industrial consumer products. transaction still holds. 8. Recent theoretical workby Hoff and Stiglitz (1993) 3. The adverse impact of imposed interest rate ceil- offers one reason for the prevalence of high interest ings is analyzed by González-Vega in "Credit Rationing rates. If there are fixed start-up costs to moneylending Behavior of Agricultural Lenders: The Iron Law of In- and government loans cannot reach the rural poor di- terest Rate Restrictions," in Adams, Graham, and Von rectly, then subsidized credit to traders, landlords, and Pischke 1984. other moneylenders reduces their cost of capital. How- 4. The term "formal financial services" refers to serv- ever, the resulting profits encourage new entry into the ices provided by banks and other institutions recog- moneylendingbusiness, which decreases the number of nized as formal financial intermediaries. Such clients per moneylender and increases moneylenders' intermediaries are defined as legally chartered financial average fixed costs. The result may be higher interest institutions supervised by the monetary authorities. rates to the rural poor than before the govemment Semiformal financial intermediaries are legally char- subsidies. Similarly, govemment subsidies on formal tered institutions that engage in a narrower range of credit may increase informal interest rates if entry into financial transactions (deposit mobilization is often the moneylending business and greater competition proscribed) not supervised by the monetary authorities. among moneylenders weakens borrowers' bonds with Informal financial intermediaries, such as moneylend- given lenders and so reduces the effectiveness of repu- ers, are financial intermediaries that are neither legally tation mechanisms in maintaining credit discipline. chartered nor supervised by the monetary authorities. 9. The figures for Brazil and Mexico reflect the fiscal 5. Varying degrees of similar interventions have and quasi-fiscal transfers. However, they do not reflect been made in capital markets, bond markets, insurance the additional costs to the economy of foregone invest- markets, and pension markets. ment activity in nonpriority sectors that is crowded out 6. Thisissuewasnotsomuchoneofpromotingnew, by targeted lending, or of lower savings rates in the untested technologies, but of capital deepening and economy. These costs arise from the need to charge expanding the use of traction equipment, terracing, higher interest rates to other sectors (thus crowding out storage facilities, and other technologies previously "nonpriority" borrowers) and to offer lower deposit available only to a small number of farmers. Conse- rates to savers to cross-subsidize targeted lending quently, the risks associated with these technologies programs. 143 144 Rural Finance 10. Proponents of the traditional approach have cross-sectional variations in consumption. This sug- often taken simple increases in physical production as gests that United States financial markets do not ensure indicators of success, without evaluating the value of constrained Pareto optimal resource allocations. Rashid those production gains in economic terms at border and Townsend (1994) point out that Phelan's findings prices. are not conclusive, because the findings depend on the 11. Financial depth is generally defined as the ratio way the models are specified (more sophisticated mod- of M2 or M3 to GDP, where M2 comprises notes and els might do a better job of explaining the consumption coins in circulation, plus current and (relatively liquid) variations). deposit accounts; M3 broadens the definition of M2 to 17. For an overview of the incentive problems asso- include less liquid forms of deposits. ciated with targeting, see Besley and Kanbur (1990) or 12. This concept is referred to as "Pareto optimum." Sen (1992). Net economic benefit refers to the private costs and bene- 18. For research along these lines, see Subbarao, Ah- fits of participants and to the costs and benefits of the med, and Teklu (1996). whole society. 19. Although this section focuses on the links be- 13. Deep financial systems invite a large share of tween the macroeconomy and rural financial markets, potential participants and provide them with a wide a growing empirical literature links reduced growth in range of opportunities to increase returns and reduce overall gross domestic product and lower investment risks. There is a substantial literature on the relationship rates to macroeconomic instability (Ramey and Ramey between finance and growth. See Bencivenga and Smith 1994, EBRD 1995, IDB 1995). This suggests that, even (1991), Giovannini and de Melo (1993), and King and withoutreferencetotheimpact ofmacroeconomicvola- Levine (1993). tility on rural financial markets, establishing macroe- 14. Economic agents seek to equate the marginal conomic stability should be an overriding priority for utility of consumption across time periods. If incomes policymakers. are subject to idiosyncratic shocks, then state-contin- 20. Chile has shown that passing on foreign ex- gent contracts with other agents (credit, savings, or change risk to clients does not resolve the problem, insurance) can be used to insulate consumption from doing so merely converts exchange rate risk into credit idiosyncratic variations in personal incomes. In credit risk. transactions even risk-averse borrowers must bear 21. While Isham and Kaufmann (1995) findings sug- some risk on loans if lenders have limited information gest that some projects may be viable even in the pres- about those borrowers. This risk of lower returns when ence of modest exchange rate distortions, the major a project is less successful entails a departure from full risks to financial intermediation in distorted policy en- insurance of consumption against idiosyncratic vari- vironments should be addressed before significant ex- ations in income, but it gives borrowers an incentive to pansions in credit are initiated. increase efforts to succeed and to repay the loans. The 22. These distortions have been widely documented question is whether fluctuations in borrowers' con- (World Bank 1986, Bautista and Valdes 1993, and sumption levels can be explained by this incentive Goldin and Knudsen 1990) Other subsectoral distor- mechanism. If the fluctuations are greater than those tions in resource allocations are less well documented. required to achieve the incentive effects, then financial For example, in the rural sector, insufficient attention markets are inefficient. has been paid to the potential for generating incomes 15. Although market failures can occur in urban ar- and diversifying risk through nonfarm rural enterprises eas too, they are particularly severe in rural areas be- (World Bank 1990, and Chaves and Sanchez 1995). In cause of a legacy of urban-biased policies that reduced agriculture the disproportionate attention paid to larger the profitability of rural investments and because of farmers may not have been warranted given the rela- misguided government interventions in rural financial tively high productivity of smaller farmers (Berry and markets. Cline 1979; Binswanger, Deininger, and Feder 1993; van 16. An exception is Phelan and Townsend (1991) Zyl, Binswanger, and Thirtle 1994). Finally, resource who apply three models of markets with constrained allocations (for example, agricultural extension) have information to United States data on consumption and systematically favored men over women, although find that these models explain consumption variations women account for a major share of production in most better than a model that assumes full information can. rural areas. Nonetheless, according to Phelan and Townsend, even 23. Direct interventions, such as price controls on these models do not fully explain intertemporal or agricultural products, immediately affect agricultural Notes 145 terms of trade. Indirect measures, such as tariffs on 30. In Mexico the imputed cost of FEGA's equity and industrial imports or disproportionate public invest- net reported profits, budgetary transfers to the guaran- ment in urban areas, have an indirect effect on relative tee and technical assistance fund amounted to US$63 prices of agricultural products. million in 1992. Part of this cost went to cover support 24. For example, reducing industrial protection is for technical assistance to commercial banks. In India likely to shift the terms of trade toward products for the Deposit Insurance and Credit Guarantee Corpora- which countries enjoy a comparative advantage, and to tion (DICGC) had an accumulated deficit of Rs 1,200 increase returns on labor-intensive products, thus crore (US$340 million) on the credit guarantee fund in boosting rural employment and wage incomes. 1995, which was largely covered by cross-subsidies 25. For a discussion of best practices in trade policy from the deposit insurance fund. Morever, the current reform, including the sequencing of macroeconomic process of rehabilitating regional rural banks is likely to and trade reforms, see Thomas and Nash (1991). bring further substantial losses to the DICGC in the 26. Poor fiscal and monetary policies induce macroe- immediate future. conomic shocks and amplify shocks to the system from 31. This section draws heavily on Hazell (1992 and extemal factors. These shocks can have damaging spil- 1995). lover effects in financial markets, particularly when the 32. In Pakistan a pilot cotton insurance scheme shocks are large in proportion to the size of the market. launched in 1986 by a private insurance company was For example, the ratio of fiscal volatility to GDP in Latin obliged to accept all clients in the pilot area who had America averages 3.4 percent, compared to 1.2 percent received loans for inputs from the Agricultural Devel- in industrialized countries. The effect on the monetary opment Bank of Pakistan. In practice the insurees' farms system in Latin America is however ten times greater were among the largest farms in the country, in the top than that in industrialized countries, because the ratio 8 percent in terms of size (Roberts and Dick 1991). of financial depth to GDP is much lower (20 percent 33. Borrowers are usually required to purchase in- compared to 70 percent). (IDB 1995). surance to qualify for loans, and the bank often acts as 27. This chapter was contributed by Heywood Flei- an intermediary, collecting the premiums and receiving sig, based on the background paper prepared for this indemnities on behalf of borrowers. volume by Heywood Fleisig and Nuria de la Peña in 34. This section draws heavily on the findings of the 1996. The background paper is also the source for all World Bank's Operations Evaluation Department boxes in this chapter unless otherwise indicated. (World Bank/OED 1981). 28. The fact that financial sector policies have been 35. There are several ways in which an RFI might urban-biased has major implications for nontraditional respond to decreased subsidization. For example, the clients.IntheDominicanRepublic,commercialbanklend- RFI might eliminate its loss-generating activities and ing to microenterprises is effectively ruled out by the apply stricter loan approval procedures or more aggres- combination of a 20 percent reserve requirement and high sive loan collection. A significant increase in the on- provisioning requirements (10 to 60 percent) for loans to lending interest rate may also influence loan demand clients with limited fornal documentation. Unlike corpo- and loan losses (Yaron 1992a). rate clients, most microenterprises do not have acceptable 36. In cases in which monopolies or otherwise un- documentation. Since the same norms apply to all finan- competitive markets exist, a low, zero, or negative SDI cial intermediaries that mobilize savings, any financial does not necessarily imply efficient financial interme- intermediary that lends to thepoor musteffectivelychoose diation. Complementary ratios (such as the number and between offering savings services and offering credit serv- volume of loans and deposits per staff member and ices to its target clientele-it cannot profitably do both at relative interest rate spreads) should be considered to interest rates that the market will bear. adequately assess a rural financial intermediary's per- 29. As Sen (1992) notes, "There is not going tobe any formance under conditions of imperfect competition. general formula here, and much would depend on par- 37. This section draws heavily on Yaron (1994a). For ticular circumstances. I should say in passing that I do further details about financial ratio analysis of RFIs see not doubt that some expert in modem economics would Mould (1991) and Barltrop and McNaughton (1992). find it helpful to say that targeting should be pushed 38. The IREQ and the FRR are closely interrelated exactly to the point at which the marginal benefit from when other variables are held constant, namely the it equals its marginal cost. Anyone who gets enlighten- investment cash flow stream before debt service, the ment from that wonderful formula fully deserves that percentage in investment financed by debt, loan maturi- enlightenment." ties, and the lending interest rate. 146 Rural Finance 39. Strong institutions can sometimes go far in suc- 49. The BAAC's extremely high penetration rate cessful rural financial intermediation even in unfavor- could be partially attributed to the Thai government's able environments. Successful RFIs have been found in legislated requirement that commercial banks lend a set a wide range of environments. However, it is not known minimum of their portfolio to the agricultural sector (5 how these institutions would have performed in a more percent in 1975, increased to 20 percent in 1987). Many (or less) favorable environment. A more enabling policy commercial banks prefer to use BAAC as an intermedi- environment usually facilitates improved institutional ary in extending these loans, their use of the BAAC has performance, although a sound environment obviously contributed to the bank's phenornenal growth and high does not guarantee institutional success (see Pederson level of market penetration. 1995 and Christen and others 1994). 50. The GB's SDI computations are based primarily on 40. With group savings and credit programs involv- calculations made by Professor Syed M. Hashemi of the ing joint liability, the prospect of losing one's savings Grameen Trust. The authors adjusted the market reference because of nonrepayment by another group member is deposit interest rate to better reflect the opportunity cost likely to lead to significant peer pressure for loan repay- of the concessional funds used by the GB. ment (see Bennett, Goldberg, and Hunte 1996). 51. We counted BAAC bonuses to employees as ex- 41. Extermal agencies sometimes attempt to manage penditures rather than as profit appropriations, which local RFIs, but the systems and procedures that they is how they are represented in BAAC financial state- develop will not become embedded in the organization ments. and will not be transferable unless local RFI staff sup- 52. A 1995 World Bank study, Grameen Bank. Per- port the systems and procedures. formance and Sustainability (Khandker, Khalily, and 42. This chapter draws heavily on Yaron (1992b and Khan 1995), proposes a measure of the GB's subsidy 1994a). dependence, the "subsidy dependence ratio" (SDR), 43. Information on BAAC has been obtained primar- which does not differentiate between interest eamed on ily from BAAC (1994 and 1995), and Sacay, Rhandawar, the loan portfolio and interest and dividends earned on and Agabin (1996). capital market investments. This results in an under- 44. The unit desa, or village bank, is a sepa ate profit statement of Grameen's dependence on subsidies, par- center of the Bank Rakyat Indonesia (BRI) and is re- ticularly during its initial years of operation, when a ferred to throughout this report as BRI-UD. Unless larger share of its financial resources was invested in the otherwise indicated, the figures presented in this chap- capital market. The measure therefore also underesti- ter have been obtained from the World Bank (1995a). mates the subsequent progress the GB made in reducing 45. Grameen means village in the Bangla language (Gib- its dependence on subsidies, as the share of funds in- bons 1992). Reference to the Grameen Bank in this publi- vested in the capital market declined relative to the cation draws heavily from Gibbons (1992); Grameen Bank share of funds loaned to clients. Following this logic, a (1993); and Khandker, Khalily, and Khan (1995). microfinance institution could appear increasingly in- 46. A proposal to allow the BAAC to lend up to 20 dependent of subsidies simply by reducing its loans percent of its loan portfolio for nonagricultural activi- outstanding. If a subsidy measurement device indicates ties is currently awaiting govemment approval (Sacay, a lower dependence on subsidies simply when an inter- Rhandawar, and Agabin 1996). mediary shifts resources away from loans to its target 47. Income tax in Thailand is currently 30 percent. clientele to investments in the market, then something The reserve requirement on deposits for commercial is basically wrong with the device. banks is 7 percent, of which at least 2 percent must be Suppose a microlender receives additional soft loans deposited with the Bank of Thailand (Sacay, Rhan- that are fully invested in the capital market at market dawar, and Agabin 1996). rates. Its profits increase, but so do the subsidies it 48. The figures and financial data quoted and the receives. Therefore the numerator of the SDR, which is corresponding end-of-period exchange rates are as fol- identical to that of the SDI (that is, subsidies net of lows (unless otherwise indicated): profits), remains basically unchanged. On the other BRI-UD: December 1994: US$1 = Rp 2,200 hand, the denominator-which unlike the SDI includes December 1995: US$1 = Rp 2,308 interest and dividend earned on capital market invest- BAAC: December 1994: US$1 = B 25.09 ments-increases by the gross amount eamed in the March 1995: US$1 = B 24.74 capital market financed by the new concessional GB: December 1994: US$1 = Tk 40.25 funds.The SDR therefore falls, thereby providing a mis- December 1995: US$1 = Tk 40.75 leading indication that subsidy independence has in- Notes 147 creased, even though there has been absolutely no regular loans, thereby substantially cross-subsidizing change in the microlender's activities with its clients. housing loans. 53. The GB charges interest on housing loans at a rate 54. This appendix is based on Gurand, Pederson, of 10 percent a year-about half the rate charged on its and Yaron (1994). References Adams, Dale W., Douglas Graham, and J. D. Von Bell, Clive. 1988. "Credit Markets and Interlinked Pischke. 1984. Undermining Rural Development Transactions." In Hollis Chenery and T. N. with Cheap Credit. Boulder, Colo.: Westview Srinivasan, eds., Handbook of Development Eco- Press. nomics. Amsterdam: North Holland. Akerlof, George. 1984. An Economic Theorist's Book Bencivenga, Valerie, and Bruce D. Smith. 1991. of Tales: Essays That Entertain the Consequences of "Financial Intermediation and Endogenous New Assumptions in Economic Theory. New York: Growth." Review of Economic Studies 58 (April): Cambridge University Press. 195-209. Aryeetey, Emest, Hemamala Hettige, Machiko Benjamin, McDonald P., Jr. 1994. "Credit Schemes Nissanke, and William F. Steel. 1997. Financial for Microenterprises: Motivation, Design and Market Fragmentation and Reforms in Sub-Saharan Viability." Ph.D. dissertation. Georgetown Uni- Africa. World Bank Technical Paper 356. Wash- versity, Department of Econonúcs, Washing- ington, D.C. ton, D.C. BAAC (Bank for Agriculture and Agricultural Co- Bennett, Lynn, and Mike Goldberg. 1993. Provid- operatives). 1994. "Annual Report 1993." Bang- ing Enterprise Development and Financial Services kok, Thailand. to Women: A Decade of Bank Experience in Asia. . 1995. "Annual Report 1994." Bangkok, World Bank Technical Paper 236. Washington, Thailand. D.C. Baer, Herbert, and Gerard Caprio. 1995. "Bank Bennett, Lynn, Mike Goldberg, and Pamela Recapitalization: If and When." DEC Policy Hunte. 1996. "Ownership and Sustainability: Review Note 2. World Bank, Policy Research Lessons on Group-Based Financial Services Department, Finance and Private Sector Devel- from Five NGO's in South Asia." World Bank, opment Division, Washington, D.C. Asia Technical Department, Human Resources Barltrop, Chris, and Diana McNaughton. 1992. and Social Development Division, Washington, Banking Institutions in Developing Markets. Vol. 2, D.C. Interpreting Financial Statements. Washington, Benoit, Phillipe. 1996. "Mitigating Project Risks- D.C.: World Bank. World Bank Support for Government Guaran- Barro, Robert J., and David B. Gordon. 1985. tees." Viewpoint 79, May. "Rules, Discretion and Reputation in a Model of Berry, R. A., and W. R. Cline. 1979. Agrarian Struc- Monetary Policy." NBER Working Paper 1079. ture and Productivity in Developing Countries. National Bureau of Economic Research, Cam- Geneva: International Labour Office. bridge, Mass. Besley, Timothy. 1994. "How Do Market Failures Bautista, Romeo, and Alberto Valdes. 1993. The Justify Interventions in Rural Credit Markets?" Bias against Agriculture: Trade and Macroeconomic World Bank Research Observer 9:27-48. Policies in Developing Countries. San Francisco, Besley, Timothy, and Stephen Coate. 1995. Cal.: International Center for Economic Growth. "Group Lending, Repayment Incentives and 149 150 Rural Finance Social Collateral." Journal of Development Eco- Analysis of Successful Microfinance Pro- nomics 46 (February): 1-18. grams." USAID Program and Operations Besley, Timothy, and Ravi Kanbur. 1990. "The Prin- Assessment Report 10. United States Agency ciples of Targeting." Policy Research Working for International Development, Washington, Paper No. 385. World Bank, Office of the Re- D.C. search Administrator, Washington, D.C. Corden, Max. 1989. "Macroeconomic Adjustment Binswanger, Hans, and Mark Rosenzweig. 1989. in Developing Countries." World Bank Research "Wealth, Weather, Risk, and the Composition Observer 4 (1): 51-64. of Agricultural Credit Investments." World Cuevas, Carlos E., and Douglas H. Graham. 1984. Bank, Latin America and Caribbean Country "Agricultural Lending Costs in Honduras." In Department, Agriculture Operations Division, Adams, Graham, and Von Pischke, eds., Under- Washington, D.C. mining Rural Development with Cheap Credit. Binswanger, Hans, Klaus Deininger, and Gershon Boulder, Colo.: Westview Press. Feder. 1993. "Power, Distortions, Revolt, and Dale, Richard. 1982. "Issues in Bank Supervision Reform in Agricultural Land Relations." Policy around the World." The World of Banking (Sep- Research Working Paper 1164. World Bank, tember-October): 14-24. Latin America and the Caribbean Technical De- David, Cristina, and Richard Meyer. 1984. "Meas- partment and Agriculture and Rural Develop- uring the Farn-Level Impact of Agricultural ment Department. Washington, D.C. Loans." In J.D. Von Pischke, Dale Adams, and Binswanger, Hans, and Shahidur Khandker. 1995. Gordon Donald, eds., Rural Financial Markets in "The Impact of Formal Finance on the Rural Developing Countries. Baltimore, Md.: Johns Economy of India." Journal of Development Stud- Hopkins University Press. ies 32 (2): 234-62. Demirguç-Kunt, Asli, and Vojislav Maksimovic. Branch, Brian, and David Richardson. 1995. "Re- 1996. "Financial Constraints, Uses of Funds, tooling Credit Unions in Guatemala." Paper on and Firm Growth: An International Compari- a World Council of Credit Unions project pre- son." Policy Research Working Paper 1671. sented at the Sustainable Banking with the Poor World Bank, Policy Research Department, Fi- Seminar Series, October 19, World Bank, Wash- nance and Private Sector Development Divi- ington, D.C. sion, Washington, D.C. Caprio, Gerard, Izak Atiyas, and James Hansen. Edgcomb, Elaine L., and James Cawley. 1994. "The 1994. Financial Reform: Theory and Experience. Process of Institutional Development: Assisting New York: Cambridge University Press. Small Enterprise Institutions to Become More Chaves, Rodrigo, and Claudio González-Vega. Effective." In Maria Otero and Elisabeth Rhyne, 1994. "Principles of Regulation and Prudential eds., The New World of Microenterprise Finance. Supervision and Their Relevance for Microen- West Hartford, Conn.: Kumarian Press. terprise Finance Organizations." In Maria Otero The Economist. 1995. `A Matter of Ti¡te." December 9. and Elisabeth Rhyne, eds., The New World of . 1996. "South African Finance: Not a For- Microenterprise Finance. West Hartford, Conn.: mality." January 13. Kumarian Press. EBRD (European Bank for Reconstruction and De- . 1996. "The Design of Successful Rural Fi- velopment). 1995. Transition Report. London. nancial Intermediaries: Evidence from Indone- Farah, Douglas. 1996. "Satellites Solve Salvadoran sia." World Development 24: 65-78. Farm Disputes." Washington Post. July 14. Chaves, Rodrigo, and Susanna Sanchez. 1995. Fleisig, Heywood, and Nuria de la Peña. 1996. "Mexico: Rural Financial Markets." Report "Creating a Legal and Regulatory Framework 14599-ME. World Bank, Latin America and the to Promote Access to Rural Credit." Center for Caribbean, Country Department II, Natural Re- the Economic Analysis of Law, Washington, sources and Rural Poverty Division, Washing- D.C. ton, D.C. FAO (Food and Agriculture Organization of the Christen, Robert, Elisabeth Rhyne, Robert Vogel, United Nations). 1992. Crop Insurance Compen- and Cressida McKean. 1995. "Maximizing the dium 1991. Rome: Agricultural Services Divi- Outreach of Microenterprise Finance: An sion, Marketing and Credit Service. References 151 Giovannini, Alberto, and Martha de Melo. 1993. IMF (International Monetary Fund). 1993. Interna- "Government Revenue from Financial Repres- tional Capital Markets: Systemic Issues in Interna- sion." American Economic Review 83: 953-63. tional Finance. Washington, D.C. Gibbons, David S. 1992. The Grameen Reader. Isham, Jonathan, and Daniel Kaufmann. 1995. Dhaka, Bangladesh. "The Forgotten Rationale for Policy Reform: The Goldin, Ian, and Odin Knudsen. 1990. Agricultural Productivity of Investment Projects." Policy Trade Liberalization: Implicatíons for Developing Research Working Paper 1549. World Bank, Euro- Countries. Paris: Organisation for Economic Co- pean and Central Asian Country Department IV, operation and Development. Regional Mission in Kiev, Washington, D.C. González-Vega, Claudio. 1984. "Credit-Rationing Jaramillo, Fidel, Fabio Schiantarelli, and An- Behavior of Agricultural Lenders: The Iron Law drew Weiss. 1993. "The Effect of Financial of Interest Rate Restrictions." In Adams, Gra- Liberalization on the Allocation of Credit: ham, and Von Pischke, eds., Undermining Rural Panel Data Evidence for Ecuador." Policy Re- Development with Cheap Credit. search Working Paper 1092. World Bank, Grameen Bank. 1993. Annual Report for 1993. Country Economics Department, Washing- Dhaka, Bangladesh. ton, D.C. Greenwald, Bruce, and Joseph Stiglitz. 1986. Ex- Khan, Azizur Rahmah. 1977. Poverty and Landless- ternalities in Economies with Imperfect Infor- ness in Rural Asia. Geneva: The International mation and Incomplete Markets." Quarterly Labour Office. Journal of Economics Cl (2): 229-64. Khan, Mohsin, and Malcolm Knight. 1982. "Some Guasch, J. Luis, and Thomas Glaessner. 1992. Theoretical and Empirical Issues Relating to "Auctioning Credit." Vol. I, Conceptual Issues. Economic Stabilization in Developing Coun- Regional Studies Program Report 15. World tries." World Development 10: 709-30. Bank, Latin America and the Caribbean Techni- Khandker, Shahidur, Baqui Khalily, and Zahed cal Department, Washington, D.C. Khan. 1995. Grameen Bank: Performance and Sus- Gurand, Marc, Glenn Pederson, and Jacob Yaron. tainability. World Bank Discussion Paper 306. 1994. Outreach and Sustainability of Six Rural Fi- Washington, D.C. nance Institutions in Sub-Saharan Africa. World King, Robert, and Ross Levine. 1993. "Finance and Bank Discussion Paper 248. Washington, D.C. Growth: Schumpeter Might Be Right." Quar- Hazell, Peter. 1992. "The Appropriate Role of Ag- terly Journal of Economics 108: 717-37. ricultural Insurance in Developing Countries." Knudsen, Odin, and John Nash. 1990. Redefining Journal of International Development 4(6): 567-81. the Role of Government in Agriculturefor the 1990s. =. 1995. "Managing Default Risks in Rural World Bank Discussion Paper 105. Washington, Finance: The Experience with Credit Insurance D.C. Programs." Paper presented at the Workshop Krahnen, J. P., and Reinhard H. Schmidt. 1995. on Finance against Poverty, March 27-28, Read- Development and Finance as Institution Building: ing University, United Kingdom. A New Approach to Poverty-Oriented Banking. Hoff, Karla, and Joseph Stiglitz. 1993. "Theory of Boulder, Colo.: Westview Press. Imperfect Competition in Rural Credit Markets in Ladman, Jerry, and Ronald Tinnermeier. 1984. Developing Countries." Institute for Policy Re- "The Political Economy of Agricultural Credit: form, Working Paper Series 49. Washington, D.C. The Case of Bolivia." In Adams, Graham, and Hulme, David, and Paul Mosley. 1996. Finance Von Pischke, eds., Undermining Rural Develop- against Poverty. London: Routledge. ment with Cheap Credit. Huppi, Monica, and Gershon Feder. 1990. "The Levine, Ross. 1994. "Does the Financial System Role of Groups and Credit Cooperatives in Ru- Matter?" World Bank, Policy Research Depart- ral Lending." World Bank Research Observer 5 ment, Finance and Private Sector Development (July): 187-204. Division, Washington, D.C. IDB (Inter-American Development Bank). 1995. Levitsky, Jacob, and Ranga Prasad. 1989. Credit Economic and Social Progress in Latin America: Guarantee Schemes for Small and Medium Enter- Overcoming Volatility. Baltimore, Md.: Johns prises. World Bank Technical Paper 58. Wash- Hopkins University Press. ington, D.C. 152 Rural Finance Mahajan, Vijay, and Bharti Gupta Ramola. 1996. United Nations Department of Economic Af- `Financial Services for the Rural Poor and fairs, Econonúc Comnmission on Latin America. Women in India: Access and Sustainability." Ramey, Garey, and Valerie Ramey. 1994. "Cross- Journal of International Development 8(2): 211-24. Country Evidence on the Link between Volatil- McLean, Douglas, and Matthew Gamser. 1995. ity and Growth." NBER Working Paper 3755. "Hi-Tech at the Frontier: Reducing the Transac- National Bureau of Economic Research, Cam- tion Cost of Lending in Swaziland with the bridge, Mass. SmartCard." Presentation delivered at the Sus- Ramola, Bharti, and Vijay Mahajan. 1996. "Finan- tainable Banking with the Poor Seminar Series, cial Services for the Rural Poor and Women in October 11, Washington, D.C., World Bank. India: Policy Issues on Access and Sustainabil- McNaughton, Diana. 1992. "Banking Institutions ity." Sustainable Banking with the Poor Occa- in Developing Markets." Vol. 1, Building Strong sional Paper 6. World Bank, Washington, D.C. Management and Responding to Change. World Rashid, Mansoora, and Robert Townsend. 1994. Bank, Washington, D.C. "Targeting, Credit and Insurance: Efficiency, Meyer, Richard, and Geetha Nagarajan. 1996. Mechanism Design and Program Evaluation." "Credit Guarantee Schemes for Developing ESP Discussion Paper 47. World Bank. Wash- Countries: Theory, Design, and Evaluation." ington, D.C. United States Agency for International Develop- Ray, Anandarup. 1995. "Directed Credit." DEC ment, Office of Economic and Institutional Re- Policy Review Note 1. World Bank, Develop- forn, Center for Economic Growth, Washington, ment Economics Vice Presidency, Washington, D.C. D.C. Mills, Cadman Atta, and Raj Nallari. 1992. Analytical Rees, Ray. 1985a. "The Theory of Principal and Approaches to Stabilization and Adjustment Pro- Agent, Part 1." Bulletin of Economi* Research 37 grams. Economic Development Institute Semi- (1): 3-26. nar Paper 44. Washington, D.C.: World Bank. _ _ . 1985b. "The Theory of Principal and Mould, M. C. 1991. Financial Information for Man- Agent, Part 2." Bulletin of Economic Research agement of a Development Finance Institution. 37(2): 175-95. World Bank Technical Paper 63, Industry and Roberts, Richard, and W. J. A. Dick, eds. 1991. Finance Series. Washington, D.C. Strategies for Crop Insurance Planning. Rome: Pederson, Glenn. 1997. The State of the Art in Mi- Food and Agriculture Organization of the crofinance: A Summary of Theory and Practice. United Nations. Washington, D.C.: World Bank. Robinson, Marguerite S. 1994. "Savings Mobiliza- Philips, Louis. 1988. The Economics of Imperfect In- tion and Microenterprise Fínance: The Indone- formation. New York: Cambridge University sian Experience." In Maria Otero and Elisabeth Press. Rhyne, eds., The New World of Microenterprise Platteau, Jean-Philippe, and Anita Abraham. 1987. Finance. West Hartford, Conn.: Kumarian Press. "An Inquiry into Quasi-Credit Contracts: The Rostow, W. W. 1963. The Economics of Take-Offinto Role of Reciprocal Credit and Interlinked Deals Sustained Growth; Proceedings of a Conference Held in Small-Scale Fishing Communities."Journal of by the International Economic Association. Lon- Development Studies 23: 461-90. don: Macmillan. Polizatto, Vincent. 1991. "Prudential Regulation Sacay, Orlando, Bikki Rhandawa, and Meliza and Banking Supervision: Building an Insti- Agabin. 1996. "The BAAC Success Story: A Spe- tutional Framework for Banks." In Philippe cialized Agricultural Bank under Government Callier, ed., Financial Systems and Develop- Ownership." World Bank, Financial Sector De- ment in Africa: Collected Papers from an EDI velopment Department, Washington, D.C. Policy Seminar held in Nairobi, Kenya,from Janu- Schiff, Maurice, and Alberto Valdés. 1992. The ary 29 to February 1, 1990. Washington, D.C.: Plundering ofAgriculture in Developing Countries. World Bank. Washington, D.C.: World Bank. Posner, Richard. 1981. The Economics of Justice. Schmidt, Reinhard H., and Claus-Peter Zeitinger. Cambridge, Mass.: Harvard University Press. 1996. "Prospects, Problems, and Potential of Prebisch, Raúl. 1950. The Economic Development of Credit-Granting NGOs." Journal of International Latín America and Its Principal Problems. New York: Development 8(2): 241-58. References 153 Schwarz, Anna. 1992. "Effectiveness of Directed ture." World Bank, Agriculture and Natural Credit Policies in the U.S.: A Survey of the Lit- Resources Department, Washington, D.C. erature." World Bank, Policy Research Depart- Varian, Hal. 1990. "Monitoring Agents with Other ment, Finance and Private Sector Development Agents." Journal of International and Theoretical Division, Washington, D.C. Economics 146 (1): 153-74. Sen, Amartya. 1984. Resources, Values, and Develop- Virmani, Arvind. 1982. The Nature ofCredit Markets ment. Oxford: Blackwell. in Developing Countries: A Framework for Policy .1992. "The Political Economy of Targeting." Analysis. World Bank Staff Working Paper 524. Paper presented at the World Bank Conference on Washington, D.C. Public Expenditures and the Poor: Incidence and Vogel, Robert. 1984. "Savings Mobilization: The Targeting, June 17-19, Washington, D.C. Forgotten Half of Rural Finance." In Adams, Squire, Lyn. 1995. "What We Expect from a Bank Graham, and Von Pischke, eds., Undermining Project: The Rural Finance Context." Seminar Rural Development with Cheap Credit. presentation at the World Bank, December 8, Von Pischke, J. D. 1991. Finance at the Frontier: Debt Washington, D.C. Capacity and the Role of Credit in the Private Econ- Steams, Katherine. 1993. Leverage or Loss: Guaran- omy. Washington, D.C.: World Bank. tee Funds and Microenterprise. Washington, D.C.: Von Pischke, J. D., and Dale W. Adams. 1980. ACCIÓN International. "Fungibility and the Design and Evaluation of Stiglitz, Joseph. 1993. "The Role of the State in Agricultural Credit Projects." American Journal Financial Markets." Proceedings of World Bank of Agricultural Economics 62(4): 719-26. Annual Conference on Development Economics. WWB (Women's World Banking). 1995. "The Washington, D.C. Missing Links: Financial Systems That Work Stiglitz, Joseph, and Andrea Weiss. 1981. "Credit for the Majority." Report on Women's World Rationing in Markets with Inperfect Informa- Banking Global Policy Forum, April, New tion." American Economic Review 71(3): 393- York. 410. World Bank. 1989. World Development Report 1989: . 1983. "Incentive Effects of Terminations: Financial Systems and Development. New York: Applications to the Credit and Labour Mar- Oxford University Press. kets." American Economic Review 73(5): 912-27. . 1990. World Development Report 1990: Pov- Subbarao, Kalanidhi, Akhter Ahmed, and Tesfaye erty. New York: Oxford University Press. Teklu. 1996. Selected Social SafetyNet Programs in - . 1991. World Development Report 1991: The the Philippines: Targeting, Cost-Effectiveness and Challenge of Development. New York: Oxford Options for Reform. World Bank Discussion Pa- University Press. per 317. Washington, D.C. . 1992. "Effective Implementation: Key to Thomas, Vinod, and John Nash, with Sebastian Development Impact." Report of the Portfolio Edwards, Thomas Hutcheson, Donald Keesing, Management Task Force. Washington, D.C. Kazi Matin, Garry Pursell, and Alexander Yeats. . 1994a. World Development Report 1994: In- Best Practices in Trade Policy Reform. 1991. New frastructure and Development. New York: Oxford York: Oxford University Press. University Press. Tuck, Laura, and Jacob Yaron. 1996. "Rural Fi- _ . 1994b. "Management Response to OED: nance in the Transition Economies of Central A Review of Bank Lending for Agricultural and Eastem Europe." Paper presented at the Credit and Rural Finance." The Secretary, Joint Annual Meeting of the American Agricultural Audit Committee, January 5, Washington, Economics Association, San Antonio, Tex. D.C. Udry, Christopher. 1990. "Credit Markets in . 1994c. "Mexico: Agricultural Sector Memo- Northern Nigeria: Credit as Insurance in a Rural randum." Latin America and the Carib- Economy." The World Bank Economic Review 4(3): bean, Country Department II, Natural Re- 251-69. sources and Rural Poverty Division, Washington, van Zyl, Johan, Hans Binswanger, and Colin D.C. Thirtle. 1994. "The Relationship between Farm . 1994d. Global Economic Prospects and the Size and Efficiency in South African Agricul- Developing Countries 1994. Washington, D.C. 154 Rural Finance . 1995a. Assessing Development Effectiveness: . 1995. "Performance Audit Report- Evaluation in the World Bank and the International Indonesia: BRI/Kupedes Small Credit Project Finance Corporation. Washington, D.C. (Loan 2800-Ind)." Report 14511. Washington, D.C. _ .___ 1995b. "The Use of Sectoral and Project Yarbrough, Beth, and Robert Yarbrough. 1991. The Performance Indicators in Bank-Financial Sec- World Economy: Trade and Finance. Chicago: Dry- tor Operations." Financial Sector Development den Press. Department, Washington, D.C. Yaron, Jacob. 1992a. Assessing Development Finance . 1996a. "World Bank Collaboration with Institutions: A Public Interest Analysis. World NGOs." Gendernet 2(1). Bank Discussion Paper 174. Washington, D.C. World Bank, Operations Evaluation Department. . 1992b. Successful Rural Finance Institutions. 1976. "Agricultural Credit Programs.` Report World Bank Discussion Paper 150. Washington, 1357. Washington, D.C. D.C. . 1981. "Agricultural Credit Projects: A Re- . 1992c. "Rural Finance in Developing view of Recent Experience in India." Report Countries." Policy Research Working Paper 3415. Washington, D.C. 875. World Bank, Agriculture and Natural Re- .1983. "Annual Review of Project Perform- sources Department, Washington, D.C. ance Audit Results," Vol. 2, para. 5.150, Report - . 1994a. "What Makes Rural Finance Insti- 4720, September 16. Washington, D.C. tutions Successful?" World Bank Research Ob- . 1993. "A Review of Bank Lending for server 9(1): 49-70. Agricultural Credit and Rural Finance (1948- . 1994b. "The Assessment and Measure- 1992)." Report 12143. Washington, D.C. ment of Loan Collections and Loan Recovery." _ ._-- 1994. "An Overview of Monitoring and World Bank, Agriculture and Natural Re- Evaluation in the Bank." Report 13247. sources Department, Washington, D.C. Distributors of COLOMBiA GERMANY ISRAEL NEPAL PORTUGAL SWEDEN lntoenlace Ltda. UNO-Vedag Yozmot LBeroture Ud. Everea Media Intemational Services (R) Ltd. Livrara Portugal Wennergren-Williams AB W orld Bank Carrera 6 No. 51-21 PoppeladorferAllee55 RO. Box 56055 GPO Box 5443 Aparado 2681, RuaDoCarmoo70-74 P 0.Box 1305 Apartado Aereo 34270 53115 Bonn 3 Yohanan Hasandiar Street Kathmandu 1200 Lisbon S-171 25 Soina Publications Santalé de Bogotá, D.C. Tel: (49 228) 949020 Tel Aviv 61560 Tel: (9771) 472152 Tel: (1) 347-4982 Tel (46 8) 705-87-50 Prices and credit terms varyfrom Tel: (571) 285-2798 Fax: (49 228) 217492 Tel: (972 3) 5285-397 Fax: (977 1) 224 431 Fax: (1) 3470264 Fax: (46 8) 27-00-71 cotntr- to country Consult your Fax:(571)285-2798 URL: hp://www.uno-vedag.de Fax: (9723)5285-397 E-mail: mail@wwi.se (oca) distributor p~~~~~~~~acirog an ~~~E-mail: unovedag@aol.com NETHERLANDS ROMANIA ioralde stributrbefarei COTE DIWVOIRE R..Y lntemational De LindebooroinOr-Publikaties Compani De Librafn Bucuresti S.A. SWITZERLAND order. Center d'Eddion et de Ditausion Atricaines GREECE PO Box 13056 PO. Box 202,7480 AE Haaksbergen Str. Lipscani no. 26, sector 3 Librairie Payot Service lnsthutionnel (CEDA) Papasotidou S.A. Tel Aviv 61130 Tel: (31 53) 574-004 Bucharest Cotes-de-Montbenon 30 ARGENTINA de libro Intemacional04 BP 541 35, Stoumara Str. Tel: (9723)5461423 Fax: (31 53)572-9296 Tel:(401)6139045 1002Leusanne Olicna del Libro lotemucional Abidjan 04 10682 Athens Fax: (9723)5461442 E-mal: lindeboo@workdnline.nt Fax: (401)312 4000 Tel: (41 21) 341-3229 Av. Cordobe 1877 Tel: (225)246510:246511 Tel: (301) 364-1826 E-mail: royil@netvision.net.il URL: hotp://avww.woddonline.nV-lindeboo Fax: (41 21) 341-3235 T120 8uenos15-3 Fao: (225)259567 Fax: (301) 364-8254 RUSSIAN FEDERATION Fax: (54 1) 81543156 Palestinian Authodty/Middle East NEW ZEALAND lsdateistvo ADECO Van Diermen EdlionsTechniques Fao: (54 1)815-0156 CYPRUS HAITI índex Intormation Services EBSCO NZ Ltd. 9a, Lolpachniy Pereulok Ch. de Lacuez 41 AUSTRALIA, FtJt, PAPUA NEW GUINEA CenternorAppliedResearch CubureDiftusion PO.B. 19502 Jerusalem Private Mail Bag 99914 Moscow 101831 CH1807Blonay SOLOMONISLANDS, VANUATU, AND CyprusCollege 5, Rue Capois Tel: (972 2)6271219 New Markel Tel: (7095) 9178749 Tel: (4121) 9432673 WESTERN SAMOA 6, Diogenes Street, Engomi C.E 257 Fax: (972 2) 6271634 Auckland Fax: (7 095) 917 92 59 Fax: (41 21) 943 3605 WESTElnfoRmaiN ServMOA s PO. Box 2006 Port-au-Prince Tel: (64 9) 524-8119 D4A. lnhmatinhServceo Nicosia Tel (509)239260 ITALY Fax: (649)524-8067 SINGAPORE, TAIWAN, THAILAND 648 Whitehorse Road Tel (357 2) 44-1730 Fax: (509) 23 4858 Licosa Conmissbinaria Sansoni SPA MYANMAR, BRUNEI Central Books Distribution Victoria Fax: (357 2) 46-2051 Vía Duca Di Calabra, 1/1 NIGERIA Aaahgate Publishing Asia Pactic Pte. Ltd. 306 Silom Road Tel: (61)392107777 HONG KONG, MACAO Caoella Potale 552 University Press Limhted 41 Kallang Pudding Road #04-03 Bangkok 10500 Fáx: (61)392107788 CZECH REPUBLIC Auia 2000 Ltd. 50125 Firenze Three Crowns Building Jericho Golden Wheel Building Tel: (662) 235-5400 E-mail: )ervice@dadirect.comnau National Intormation Center Sales & Circulation Department Tel: (55) 645-415 Private Mail Bag 5095 Singapore 349316 Fax: (66 2) 237-8321 URU http:/e .dadirect.com.au prodeina, Konviktska 5 Seabird House, unHt 1101-02 Fax: (55) 641-257 lbadan Tel: (65)741-5166 CS -11357 Prague 1 22-28 Wyndham Street, Central E-mail: licosa@tfbcc.R Tel: (234 22)41-1356 Fax: (65) 742-9356 TRINIDAD & TOBAGO, AUSTRIA Tel: (422) 2422-9433 Hong Kong URL: htpJ/wa.Htbcc.iti¶cosa Fax: (23422) 41-2056 E-mail: ashgate@asianconnect.com AND THE CARRIBBEAN Gemod and Co. Fax: (42 2) 2422-1484 Tel: (852) 2530-14089 Systematics Studies UnH Weihburggasae 26 URL: http:Uwww.nis.czf Fax: (852) 2526-1107 JAMAICA NORWAY SLOVENIA 9 Wa8s Street Ae1011ggaWien 26 E-mail: salesOasia2000.com.hk lan Randie Publishers Ltd. NIC Ino ANS Gospodarski Vestnlk Pubilshing Group Curepe A-10 Wie 7 DENMARK URL: htip1/www.asia2000.com.hk 206 Oid Hope Road, Kingston 6 Book Department, Postboks 6512 EKerstad Dunaiska cesta 5 Trinidad, WestIndies TFL: (431) 512-47-31-0 SamiundsLtteratur Tel: 876-927-2085 N-0606 Gsul 1000 Lub)iana Tel:(809)662-5654 URLa: tip1/54.geroid.col-t.onlin- RosenoemsAiée1 HIUNGARY Fax: 876-977-0243 Tel: (47 22) 97-450 TeL (38661) 13383 47; 1321230 Fax: (809)682-5654 UPU hfl://víe .gemt.co/atonline DK-1970 Fredeñksberg C Euro mífo Service E-mail: iml@~culiu.com Fáx: (47 22) 97-4545 Fao: (386 61) 133 80 30 E-mail: tobe@trinidadnet BANGLADESH Tel: (45 31) 351942 Marghszgeti Europa Haz E-oail: repansekjogvestnik.si Micro industries Develapment Fao: (45 31) 357822 H-1 1,38 Budapeut JAPAN PAKISTAN UGANDA Ascoitance Society (MIDAS) URL: hhp.i/www.sl.cbs.ck Tel: (36 1B1dt m Earotem Book Semvice Miza B AN SOUTH AFRiCA, BOTSWANA Gustro Ltd. hsunse 5, Poad 16 Fax: (36 1) 302 5035 3-13 Hongo 3-chome, Bunkyo-ku 65, Shahrah-euaad-e-Azam For single t9/s: PO Box 9997, Madhvani Building DhanmondAió AreaECUADOR E-mail: euroinuo@mail.matav.hu Tokyo 113 Lahore 54000 Oxford Universfty Press Southem Africa PlIt 1614 Jn)a Rd. Dhanmond Area Libri Mundi Tel: (81 3)3818-0861 Tel: (92 42) 735 3601 Vasco Boulevard, Goodwood Kampala Tel: (900 2) 326427 Libreria Intemacional INDIA Fax: (81 3) 3818-0864 Fax: (92 42) 576 3714 PO. Box 12119, Nl City 7463 Tel: (256 41) 251467 Fax. (880 2) R11188 PO. Box 17-01-3029 Allied Publishers LId. E-mail: orders@svt-ebs.cojp Cape Town Fax: (256 41) 251 468 Fax: (98002)8011188 Juan Leon Mer 851 751 Mount Road URL:httpliww.bekkoame.or)p/-svl-ebs Oxford UniversityPreos Tel: (27 21) 595 4400 E-mal: gus@swituganda.com Quito Madras - 600 002 5 Bangalore Town Fax: (27 21) 595 4430 BELGIUM Tel: (5932)521-606 (5932)544-185 Tel: (91 44)852-3938 KENYA Sharae Faisal E-mail: oxtord@oup.co.za UNITED KINGDOM Jean De Lanuoy Fax: (593 2) 504-209 Fax: (91 44) 8524649 Atrica Book Service (E.A.) Ltd. PO Box 13033 Microinmo Ltd. Av. do Roi 320 E-mail: librimuo @librimundi.com.ec Quaran House, Mlangano Street Karachi-75350 For subscr(prion orcers: RO. Box 3, Ahon, Hampshire GU34 2PG Tel: (322)538-5169 E-mail: libnmu2@Iibrimundi,com,ec INDONESIA PO. Box 45245 Tel: (9221)446307 intemationat Subscrptlon Service Engand Fax: (32 2) 538-0841 ~~~~~~~Pt. Indira Limoited Nairobi Fax: (92 21) 4547640 RO. Box 41095 Tel: (44 1420) 86848 Fax: (322)5i,0-0841 EGYPT, ARAB REPUBLIC OF Jalan Borobudur 20 Tel: (254 2) 223 641 E-mail: ouppak@TheOKlice.net Craighail Fáx: (44 1420) 89889 BRAZIL Al Aharm Distnbulfin Agency PO. Box 181 Fax: (254 2) 330 272 Johannesburg 2024 E-mail: wbank@ukminho.demon.co.uk Al Galaa Street Jakarta 10320 Pak Book Corporation Tel: (2711)880-1448 URL: httpJ/www.microinf.co.uk Publicacóes Tecnicas lntemacionaosLtda. Cairo Tel: (6221)390-4290 KOREA, REPUBLIC OF Aziz Chambers 21, Queens Road Fax: (27 11)880-8248 Rua Peixoto Gomide, 209 Tel: (20 2) 578-8083 Fax: (62 21) 390-4289 Dae)on Trading Co. Ltd. Lahore E-mail: iss@is.co.za VENEZUELA Te: 5 51a259 6644 Fax: (20 2) 5784O833 . Box 34, Youida, 706 Seoun Bldg Tel: (92 42) 636 3222; 636 0885 Tecni-Ciencia Ubros, S.A. Fal: (55 11) 258-90 IRAN 44-6 Youido-Dong, Yeongchengpo-Ku Fax: (92 42) 636 2326 SPAIN Centro Cuidad Comercial Tamanco Eai: pi tm tr 2 tp olbr The Middle East Observer Ketab Sara Co. Publishers Seoul E-mail: pbc@bran.net.pk Mundl-Prensa Libros, SA. Nivel C2, Caracas Rmal: posmasterp .ol.br 41, Sherf Streel Khaled Eslamboli Ave., 6@h Street Tel: (82 2) 785-1631/4 Castelo 37 Tel: (58 2) 959 5547; 5035; 0016 URL: htipJDwww .uol.br Cairo Delafrooz Alley No. 8 Fax: (82 2) 784-0315 PERU 28001 Madrid Fax: (58 2) 959 5636 CANADA Tel: (20 2) 393-9732 FO. Box 15745-733 Edftorial Desarrola SA Tel: (34 1) 431-3399 CAeNuol PAbiiuhiug Ca. Ltd. Fax: (202)393-9732 Tehran 15117 MALAYSIA Apartada 3824, Lima 1 Fax: (341)575-3908 ZAMBIA Renou5 Publishing Co. Ltd. Tel: (98 21) 8717819,8716104 University of Malaya Cooperative Tel: (51 14) 285380 E-mail: librerb&mundiprensa.es Univerity Bookshop, University of Zamba 5369 Canotek Road FINLAND Fax: (98 21) 8712479 Bookshop, Limited Fax: (51 14) 286628 URL: htpJiwww .mundipronsa.esi Great East Road Campus OTawa Onta( o KlJ) 93 Akateeminen Kirjakauppa E-mail: ketab-sara@neda.net.ir RO. Box 1127 PO. Box 32379 Fax: (613) 745-7680 PO. Bou 128 Jalan Pantai 6amu PHILIPPINES Muodi-Frensa Borcelona Lusaka Fax: (613) 7e@e76c FIN-00101 He7sinkE i Kowkab Publishers 59700 Koala Lumpor intematinai Booksource Center Inc. Conoell de Cent, 391 Tel: (260 1) 252 576 E-mail: order.deptrenoulbooks.com Tel: (3580)121 4418 PO. Box 19575-511 Tel: (603) 756-5000 1127-A Antipolo St, Barangay, Venezuela 08009 Borcelona Fax: (260 1)253952 URL: http:D www .renoulbooks.com Fax: (358 0)121-4435 Tehran Fax: (60 3) 755-4424 Makati City Tel: (34 3) 488-3492 CHINAE-mail: akatilaus stockmann.li Tel: (98 21 ) 258-3723 Tel: (63 2) 896 6501: 605; 6507 Fax: (34 3) 487-7659 ZIMBABWE China Finaucial & Economic /URL: http://www .akateeminen.oornl Fax: (98 21) 258-3723 MEXICO Fax: (63 2) 896 1741 E-mail: barcelona@mundiprensa.es Lorigmn Zimbabwe (Pv)Ld. Publishing House INFOTEC Toado od rbni 0,Fublshin HousgJie FRANCE IRELAND Av. San Femando No. 37 POLAND SRI LANKA, THE MALDIVES RO. Box ST125 8, Da F Si Dong Jie Wodd Bank Publications Goveomment Supplies Agency Col. Torello Guerra Intemational Publishing Service Lake House Bookshop Southenon Tul: (8610)6303-8257 66, avenue diéna Odig an tSoláthair 14050 Mexico, D.F 1/1. Piekna31/37 100,SirChRtampalamGardinerMawatha Harare Fax: (86 10) 640137365 75116 Paris 4-5 Harcour Road Tel: (525)624-2800 004677 Warzawa Colombo 2 Tel: (2634)6216617 Tel: (33 1) 40-69-30-56557 Dublin 2 Fax: (52 5) 624-2822 Tel: (48 2) 628-8089 Tel: (94 1) 32105 Fax: (263 4) 621670 Fax: (331) 40-69-30-68 Tel: (353 1) 661-3111 E-mail: infotec@rtn.net.mx Fax: (48 2) 621-7255 Fax: (94 1)432104 Fox: (353 1)475-2670 URL: htpJ/rtn.net.mx E-mail: books%ips@ikp.atm.com.pi E-mail: LHL@sd.lanka.net URL: hotp:l/www .ip8cg.waw.pVips/exponrt w7 10 Enabling the Safe Use of Biotechnology: Principles and Practice 11 Biodiversity and Agricultural Intensification: Partnersfor Development and Conservation 12 Rural Development: From Vision to Action. A Sector Strategy 13 Integrated Pest Management: Strategies and Policiesfor Effective Implementation 14 Rural Finance: Issues, Design, and Best Practices 15 The Economics of Involuntary Resettlement (forthcoming) 16 Social Assessmentsfor Better Development: Case Studies in Russia and Central Asia 17 Expanding the Measure of Wealth: Indicators of Environmentally Sustainable Development 18 Five Years after Rio: Innovations in Environmental Policy 19 Advancing Sustainable Development: The World Bank and Agenda 21 Related ESSD Publications Monitoring Environmental Progress: A Report on Work in Progress Nurturing Development: Aid and Cooperation in Today's Changing World Toward Sustainable Management of Water Resources Water Supply, Sanitation, and Environmental Sustainability: The Financing Challenge The World Bank Participation Sourcebook THE WORLD BANK 18 1811 Strect, N.X. Iashington, I).C. 20433 USA Iclephiolle: 202-477-1234 Faesimile: 20>2-477-6391 Tele\1: N1CI 64145 VORI )AMANK X1ICI 248423 W)RIA)ISANK Cable Address: INTBAF.\RAI) NVA.'SI UNGTONMC)l)( Wnorid nide \\Nh: http:!/¡w\.\ v )rldban k. rg¡ lI-niail: hooks rlhank.org SB0-8213- 737-8