41459 D I R E C T I O N S I N D E V E L O P M E N T Finance Increasing Access to Rural Finance in Bangladesh The Forgotten "Missing Middle" A study led by Aurora Ferrari Increasing Access to Rural Finance in Bangladesh Increasing Access to Rural Finance in Bangladesh The Forgotten "Missing Middle" © 2008 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved 1 2 3 4 5 10 09 08 07 This volume is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The bound- aries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with com- plete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. ISBN: 978-0-8213-7333-0 eISBN: 978-0-8213-7334-7 DOI: 10.1596/978-0-8213-7333-0 Cover photos: S. M. A. Hye Shapan, Daily Menabzamin and Andrew Biraj. Cover design: Quantum Think. Library of Congress Cataloging-in-Publication Data Ferrari, Aurora. Increasing access to rural finance in Bangladesh : the forgotten "missing middle"/by Aurora Ferrari. p. cm. Includes bibliographical references and index. ISBN 978-0-8213-7333-0--ISBN 978-0-8213-7334-7 (electronic) 1. Rural credit--Bangladesh. 2. Agricultural credit--Bangladesh. 3. Banks and banking-- Bangladesh. 4. Financial institutions--Bangladesh. 5. Rural industries--Bangladesh. I. Title. HG187.B3F37 2007 332.7'1095492--dc22 2007039297 Contents Acknowledgments xi Abbrevations and Acronyms xiii Executive Summary 1 Access to Finance: A Major Constraint on Rural Growth 2 Lending to the Missing Middle: Supply-Side Evidence 5 Lending to the Missing Middle: Demand-Side Evidence 5 Why Have Government Efforts to Increase Access to Finance Failed for the Missing Middle? 7 How Can the Government Facilitate Increased Access to Finance for the Missing Middle? 7 Chapter 1 The Missing Middle: Supply-Side Evidence 11 The Financial Sector 11 The Banking Sector 13 v vi Contents Government Efforts to Increase Rural Credit 16 Lending to the Missing Middle 25 Chapter 2 The Missing Middle: Demand-Side Evidence 29 Access to Credit for MSMEs 29 Access to Credit for MSMFs 32 Chapter 3 Constraints on Increasing Access to Finance for the Missing Middle 39 BKB's and RAKUB's Deep Insolvency 40 Challenges for Private Banks 43 Challenges for Microfinance Institutions 51 Challenges for Insurance Providers Serving MSMFs 56 Chapter 4 Increasing Access to Finance for the Missing Middle 61 Transforming BKB and RAKUB into the Main Providers of Financial Services in Rural Areas 62 Helping Banks Serve Rural MSMEs 68 Helping Microfinance Institutions Scale Up Lending to the Missing Middle 70 Promoting Weather Risk Management for MSMFs 70 Appendix 1 Definitions Used in This Report 77 Rural Areas 77 Micro, Small, and Medium-Size Enterprises 77 Marginal, Small, and Medium-Size Farmers 79 Appendix 2 Comparing Rural Lending in Bangladesh and India 81 Appendix 3 The Rural Micro, Small, and Medium-Size Enterprise Finance Survey 85 Survey Questionnaire 85 Sampling Frame 86 Contents vii Sample Selection 86 Focus Group Discussions 88 Basic Sample Data 88 Appendix 4 Recapitalization Scenarios and Costs of Closing BKB and RAKUB 91 Costs of Recapitalizing BKB 91 Costs of Rehabilitating BKB 94 Costs of Recapitalizing RAKUB 95 Costs of Rehabilitating RAKUB 98 Costs of Closing BKB and RAKUB 98 Appendix 5 Movable Collateral System 101 Creation of Security Interests 101 Priority of Security Interests 102 Publicity of Security Interests 105 Enforcement of Security Interests 110 Appendix 6 Case Studies of Six Microfinance Institutions 115 Methodology 115 Selection of Institutions 115 The Six Selected Institutions 116 Lending by the Six Institutions 117 Financial Performance of the Six Institutions 117 Growth Prospects of the Six Institutions 125 Bibliography 127 Index 133 Boxes 1 How Do Natural Disasters Affect Rice Production? 3 1.1 Terms and Conditions of Loans Refinanced under the Agriculture Refinancing Facility 18 2.1 How Big Is the Potential Market for Providing Credit to MSMEs? 33 2.2 How Big Is the Potential Market for Providing Credit to MSMFs? 36 3.1 BRAC Bank--Serving MSMEs 50 viii Contents 3.2 ASA's Microenterprise Programs 52 3.3 BEES's Product for Marginal and Small Farmers 54 4.1 Khan Bank of Mongolia 63 4.2 Benefits of Having Development-Oriented Investors as Minority Shareholders 66 4.3 Financial Costs of Alternatives to Rehabilitating BKB and RAKUB 67 4.4 Banco Procredit's Expansion into Rural Areas 71 4.5 Using Technology to Improve Financial Access in Uganda 73 4.6 BASIX and ICICI's Drought Index Insurance Product 74 4.7 Mexico's Fund for Natural Disasters 75 Figures 1 Most Rural Enterprises Rely on Internal Sources for Financing 4 1.1 Bank Assets, 1995­2006 13 1.2 Change in Bank Loans and Deposits, 1995­2005 14 1.3 Estimated Rural Loans by Source, 2002 and 2005 20 1.4 Urban and Rural Shares of Bank Loans, 1996­2005 22 1.5 Proportion of Loans to Deposits by Bank Branch Location, 1996­2005 23 1.6 Distribution of Bank Loans by Purpose, 1996­2005 23 2.1 Collateral Requirements of Banks and Assets of MSMEs 32 3.1 Classification Criteria for Term Loans of Less than Five Years 45 3.2 Transaction Costs on a 12-Month, 300,000 Taka Bank Loan 49 A2.1 Proportion of Loans to Deposits by Bank Branch Location, 1996­2005 82 A2.2 Urban and Rural Shares of Bank Loans 82 A2.3 Distribution of Bank Loans, by Purpose 83 A2.4 Estimated Rural Lending by Source, 2005 83 A3.1 Bangladesh, with Survey Areas Highlighted 87 Tables 1 Rural Farming Households and Landholdings 2 1.1 Indicators for Bangladesh's Financial Sector, 2000­05 12 1.2 Capital Adequacy Ratios of Banks, 2000­05 15 1.3 Returns on Assets and Equity for Banks, 2000­05 16 Contents ix 1.4 Nonperforming Loans of Banks, 2000­05 17 1.5 Bank Lending and Deposits in Rural Areas, 2000 and 2005 21 1.6 Average Annual Growth in Urban and Rural Bank Loans, 1996­2005 22 1.7 Summary of Audited Financial Statements for BKB and RAKUB, 2005 24 1.8 Access Indicators for Private Clients of BKB and RAKUB, 1996­2005 24 1.9 Microfinance Lending to Micro and Small Enterprises, 2006 25 1.10 Microfinance Lending to Marginal and Small Farmers, 2006 25 2.1 Sources of Credit for MSMEs, 2003­06 30 2.2 Returns on Capital for Small and Medium-Size Enterprises, 2005 30 2.3 Profitability Measures for Small and Medium-Size Enterprises, 2005 31 2.4 Average Loan Size for MSMEs, 2003­06 31 2.5 Average Lending Terms from Banks, Microfinance Institutions, and Family and Friends 31 2.6 Top Reasons MSMEs Do Not Apply for Credit 32 2.7 Borrowing Patterns of Farmers, 2000­02 34 2.8 Borrowing Patterns of Farmers with Access to Formal Credit, 2000­02 35 2.9 Average Size of Most Recent Loan Taken by Farmers, 2000­02 35 2.10 Transaction Costs of Most Recent Loan Taken by Farmers, 2000­02 35 3.1 Selected Data from BKB's Audited Balance Sheet, 2004­05 42 3.2 Selected Data from RAKUB's Audited Balance Sheet, 2004­05 43 3.3 Differing Loan Product Preferences for MSMEs 44 3.4 Application Process for Bank Loans to MSMEs 48 3.5 Features of Traditional Crop Insurance and Index Insurance 57 A1.1 Definition of Small and Medium-Size Enterprises Used by Bangladesh Bank, by Sector 78 A1.2 Definition of Small and Medium-Size Enterprises Used by the Ministry of Industry, by Sector 78 x Contents A1.3 Definition of MSMEs Used in This Report 79 A3.1 Sectoral Distribution of Sampled Enterprises 88 A3.2 Average Number of Employees of Sampled Enterprises, by Enterprise Size 88 A3.3 Legal Status of Sampled Enterprises 89 A3.4 Location of Business of Sampled Enterprises 89 A3.5 Average Years of Schooling for Owners of Sampled Enterprises 89 A4.1 BKB Recapitalization: Realistic Scenario with 50 Percent of Standard Loans Carved Out 92 A4.2 BKB Recapitalization: Optimistic Scenario with 25 Percent of Standard Loans Carved Out 94 A4.3 RAKUB Recapitalization: Realistic Scenario with 50 Percent of Standard Loans Carved Out 95 A4.4 RAKUB Recapitalization: Optimistic Scenario with 25 Percent of Standard Loans Carved Out 96 A4.5 BKB and RAKUB Consolidated Estimated Cost of Closure 99 A5.1 Attributes of Creation in Secured Finance Systems in Bangladesh and Good Practice Environments 103 A5.2 Attributes of Priority in Secured Finance Systems in Bangladesh and Good Practice Environments 106 A5.3 Attributes of Publicity in Secured Finance Systems in Bangladesh and Good Practice Environments 109 A5.4 Attributes of Enforcement in Secured Finance Systems in Bangladesh and Good Practice Environments 111 A6.1 Activities of the Six Selected Microfinance Institutions 118 A6.2 Loan Products for MSMFs 121 A6.3 Loan Products for MSMEs 123 A6.4 Financial Viability Indicators for SOJAG, CARB, and BEES, 2005­06 125 A6.5 Financial Viability Indicators for ASA, BRAC, and PMUK, Various Years 126 Acknowledgments This report was prepared by a World Bank team led by Aurora Ferrari under the overall guidance of Simon Bell, Sadiq Ahmed, Christine Wallich, and Xian Zhu. Major contributors to the report were Shamsuddin Ahmad, Mukta Joshi, Sadruddin Muhammad Salman, and Andrew Lovegrove. In addition the team was composed of Richard Meyer, Olivier Mahul, Mehanz Safavian, Sheela Rahman, Willian Dick, Henry Bagazonzya, Gabi George Afram, Guillemette Sidonie Jaffrin, Shah Nur Quayyum, Sukhinder Arora, and Shahnila Azher. Administrative support was provided by Maria Marjorie Espiritu and Bridget Rosario. The report was funded by the U.K. Department for International Development and the World Bank. The report was edited by Paul Holtz. The report draws on a Rural Micro, Small, and Medium-size Enterprises (MSMEs) Finance Survey undertaken jointly by the World Bank and Business and Finance Consulting (BFC) in association with HB Consultants. It also draws on the following background papers: a) "Delivering Finance to MSMEs in Bangladesh," August 2006, by Michael Kortenbusch from BFC; b) "Microfinance Products for Small and Medium Farmers and Micro and Small Enterprises: Challenges and Opportunities for MFIs in Bangladesh," April 2007, by Dewan Alamgir xi xii Acknowledgments and Richard Meyer, which is based on six MFI case studies conducted by the authors; and c) "Bangladesh Crop Index-Based Insurance: A Feasibility Study," July 2006, by William Dick, Ornsaran Pomme Manuamorn, and SARMAP. The peer reviewers were Renate Kloeppinger-Todd, Nathan M. Belete, and Akbar Ali Khan. The report incorporates extensive suggestions and comments from representatives of the government of Bangladesh, Palli Karma Sahayyak Foundation (PKSF) staff and its participating organizations, and Bangladeshi financial institutions who attended workshops in Dhaka in September 2006, April 2007, and July 2007, at which the findings of the report were presented and discussed. Abberviations and Acronyms ASA a prominent Bangladeshi microfinance institution; the word means "hope" BASIC Bank of Small Industries and Commerce BEES Bangladesh Extension Education Services BFC Business and Finance Consulting BKB Bangladesh Krishi Bank BRAC Bangladesh Rural Advancement Committee BRDB Bangladesh Rural Development Board BSB Bangladesh Shilpa Bank BSBL Bangladesh Samabaya Bank Limited BSRS Bangladesh Shilpa Rin Sangtha CARB Center for Action Research­Barind GDP gross domestic product ICICI India's largest private commercial bank (formerly Industrial Credit and Investment Corporation of India) MELA Micro Enterprise Lending and Assistance MFI microfinance institution MSME micro, small, and medium-size enterprise MSMF marginal, small, and medium-size farmer PKSF Palli Karma Sahayyak Foundation xiii xiv Abbrevations and Acronyms PMUK Padakhep Manobik Unnayan Kendra (Padakhep Center for Human Development) RAKUB Rajshahi Krishi Unnayan Bank SBC Sadharan Bima Corporation SBL Small Business Loan SEL Small Enterprise Loan SOJAG Samaj O Jathi Gathan SREESTI Sustainable Refinancing for Enterprises and Technology Improvement (product of PMUK) Currency Equivalents US$1 = 69 taka (as of June 17, 2007) Executive Summary Three-quarters of Bangladesh's people live in rural areas, and more than 40 percent of this population lives below the poverty line--making the rural sector central to the country's development. Micro, small, and medium-size enterprises (MSMEs) and marginal, small, and medium-size farmers (MSMFs)--together referred to in this report as the "missing middle"--are the engines of growth in rural Bangladesh, in terms of both current employment and contributions to gross domestic product (GDP) and possible prospects for future development. In 2004 nonagricultural activities employed 40 percent of rural work- ers and accounted for more than 50 percent of rural households' income (World Bank 2004). Most of these activities are conducted by MSMEs. Such enterprises help diversify rural incomes by reducing households' vulnerability to weather and seasonality in agriculture. With limited opportunities for creating employment in agriculture, and if an appropri- ate enabling environment is developed, these enterprises could continue to grow and will lead Bangladesh's rural economy. Though becoming less important, farming still plays a major role in the rural economy.The sector employs 54 percent of rural workers (Bangladesh Bureau of Statistics 2000a) and accounts for nearly 21 percent of national GDP.1 Crops account for the lion's share (57 percent) of agricultural GDP. 1 2 Increasing Access to Rural Finance in Bangladesh Table 1. Rural Farming Households and Landholdings Land owned Rural households Share of operated Classification (acres) Number (millions) Share (percent) land area (percent) Landless <0.05 6.40 33.8 0.0 Landless 0.05­0.49 3.56 18.8 4.5 Marginal 0.5­0.99 2.58 13.7 8.5 Marginal 1.0­1.49 1.86 9.9 10.3 Small 1.5­2.49 1.99 10.5 17.9 Medium 2.5­7.49 2.20 11.6 41.5 Large 7.5+ 0.32 1.7 17.3 Source: Agricultural Census of Bangladesh 1996/97. About 45 percent of Bangladeshi farmers own between 0.5 and 7.5 acres of land--making them marginal, small, or medium-size (table 1). More than half of farmers own less than 0.5 acre of land and are considered landless. With their extremely limited landholdings and highly volatile incomes, such farmers can make only a limited contri- bution to rural growth. Indeed, future agricultural growth will depend on intensifying current practices, expanding irrigated areas, and diversi- fying into higher-value crops--all of which require landholdings larger than 0.5 acre. Access to Finance: A Major Constraint on Rural Growth Considerable investment is needed for Bangladesh's agriculture sector to become more commercial and for its nonagricultural activities to grow. MSMEs and MSMFs not only need access to finance for investment, they also need it in a timely fashion to take advantage of market and investment opportunities. And given the high weather risk inherent in agricultural investments, financial institutions need to be able to transfer part of that risk to profitably lend to the segment. Yet 53 percent of rural enterprises (both agricultural and nonagri- cultural) consider access to finance a major or very severe obstacle to their operations (World Bank 2003). Making matters worse, no agricul- tural insurance scheme exists in Bangladesh. About two-thirds of the financing (working capital and new investment) needs of both small and medium-size enterprises are met through internal funds or retained earnings (figure 1). Thus most enterprises rely on household and business savings to expand--even for working capital. Although this might be the best solution in the current environment, there are limits to equity-financed growth. Executive Summary 3 Box 1 How Do Natural Disasters Affect Rice Production? Flooding is a recurrent event in Bangladesh. Most of its territory consists of floodplains, and up to 30 percent of the country experiences annual flooding during the monsoon season--while periodic extreme floods affect 60 percent of the national territory. Although annual flooding is beneficial, severe flooding hurts the population and causes major losses in rice production. Rice crops have, however, become less susceptible to flood damage as the area planted with deepwater aman rice (grown on flood-prone land during the monsoon season) has fallen and boro rice (which can be produced five to six months after a flood-damaged rice harvest) has been planted more extensively. Production losses caused by floods mainly affect aus and aman (rainfed) rice, while losses for boro rice are limited to unusual events. For example, the 2000 flood caused an estimated loss of 305,482 metric tons of aman and only 7,920 met- ric tons of boro. Losses caused by cyclones follow the opposite pattern, with boro rice most affected. For example, it is estimated that cyclones in 2000 caused a loss of 318,460 metric tons of boro and no damage to aus and aman crops. Bangladesh is also vulnerable to recurrent droughts. Some 2.3 million hectares are prone to drought, and between 1960 and 1991 droughts occurred 19 times. Two critical dry periods are distinguished: · Rabi and pre-Karif drought (between January and May), due to the cumulative effects of dry days, higher temperatures during pre-Karif, and low availability of soil moisture. · Karif droughts (between June/July and October), created by sub-humid and dry conditions in the country's highlands and medium highlands. Rainfall shortages affect the critical reproductive stages of the transplanted aman crop in December, reducing its yield--particularly in areas with where soil has low moisture-holding capacity. Western regions are especially vulnerable to droughts. During the Rabi sea- son 1.2 million hectares of cropland face droughts of various magnitudes, and a severe drought can damage more than 40 percent of broadcast aus. During the Kharif season drought causes significant damage to the transplanted aman crop on about 2.3 million hectares. In addition to causing agricultural losses, droughts significantly increase land degradation. Source: Bureau of Statistics data. 4 Increasing Access to Rural Finance in Bangladesh Figure 1. Most Rural Enterprises Rely on Internal Sources for Financing 70 small 60 medium-size 50 40 percent 30 20 10 0 internal family informal equity NGOs private remittances trade state fund and friends sources bank credit institution Source: World Bank 2003. Increasing access to rural finance is often the last frontier for financial sector development in developing countries. Financial institutions aim- ing to operate in rural areas in these countries usually have to deal with high transaction costs, low population densities, remote areas, and a heavy focus on agriculture, with related weather and commodity risks (box 1). Although Bangladesh is highly vulnerable to natural disasters-- including floods, droughts, and cyclones--its high population density lowers transaction costs considerably (though not entirely). This report seeks to: · Measure the quantity and quality of access to finance by rural MSMEs and MSMFs. In Bangladesh these are also referred to as the "missing mid- dle"--a segment not served by banks or microfinance institutions (MFIs). · Identify constraints that financial institutions face in serving MSMEs and MSMFs. The institutions analyzed in detail include, Bangladesh Krishi Bank (BKB), Rajshahi Krishi Unnayan Bank (RAKUB), MFIs, private banks, and insurance companies. These institutions have been selected because they are among the largest providers of finance to the missing middle. Moreover, given their size, they have a considerable impact on the current dysfunction of rural financial markets. · Develop a realistic strategy and options for sustainably increasing access to finance for the missing middle. Executive Summary 5 Lending to the Missing Middle: Supply-Side Evidence Bangladesh's banking sector has grown quickly since 1995, with bank assets growing by nearly 23 percent a year between 1995 and 2005. Moreover, the banking sector grew faster than GDP during this period, with loans rising from 21 to 30 percent of GDP and deposits jumping from 26 to 39 percent of GDP.Yet this expansion has not really benefited the missing middle. Over the years the government has introduced several initiatives aimed at facilitating rural credit in general, and lending to the missing middle in particular. These include: · Creating BKB in 1973 and RAKUB in 1987 to serve rural areas and the Bank of Small Industries and Commerce (BASIC) in 1988 to pro- mote small-scale industries in rural and urban areas. · Promoting two systems of cooperatives since 1972. · Providing refinance facilities to institutions interested in financing agriculture. · Introducing a policy encouraging banks--especially private ones--to expand their branches and activities in rural areas. · Creating a multiperil crop insurance scheme and waiving principal and interest on small agricultural loans from state banks and coopera- tives in case of natural disasters. But government efforts to increase access to finance have had mixed results. The importance of banks and cooperatives in rural lending has declined, while the importance of MFIs has increased. Bank intermediation in rural areas remains limited, with banks transferring 0.5 taka to urban areas for each 1 taka in deposits collected in rural areas. The largest providers of agricultural loans, BKB and RAKUB, are deeply insolvent, and BASIC and the cooperative system are irrelevant in terms of rural loans. Finally, since the closure of the government-sponsored crop insurance scheme, insurance companies have been almost completely absent from rural markets. No data are available on bank lending to MSMEs and MSMFs, but evi- dence indicates that such lending is limited. Moreover, the microfinance sector--traditionally the largest provider of small loans in rural areas--has remained focused on the landless poor with group loans involving weekly repayments. Such products are not suitable for MSMEs and MSMFs. Lending to the Missing Middle: Demand-Side Evidence A 2006 survey of rural MSMEs found that despite substantial use of bank services among these enterprises (43 percent had an account, and 6 Increasing Access to Rural Finance in Bangladesh 75 percent made transfers), during 2003­06 only 32 percent borrowed from banks. During the same period 16 percent of these enterprises bor- rowed from MFIs and 8 percent from informal sources (mainly family and friends)--while 44 percent did not borrow at all. When borrowing from banks, the average loan for such enterprises is 400,000 taka, while their average loan from MFIs is 38,000 taka. These firms financed 88 percent of their working capital needs using retained earnings and internal funds, and only 33 percent used debt to finance new investments. Thus, instead of borrowing, most MSMEs expand their activities only after building up internal resources. Micro, small, and medium-size enterprises do not seem to be excluded from financial markets because of poor financial perform- ance. Indeed, such enterprises have strong returns on capital and robust long-term profitability. Instead, they seem to be excluded due to a gap in the financial market. When asked why they did not apply for credit, 40 percent of these enterprises cited the high costs, direct and indirect. Direct costs include interest rates and other transaction costs (such as for documentation, including financial statements, titles, and the like). Indirect costs include long processing times (which translate into missed business opportunities) and intensive application processes requiring many meetings between borrowers and banks (again translating into missed opportunities). Once all the required documents have been submitted, it takes an average of 40 days to get a loan from a bank and 28 days from an MFI. In addition, the survey found that banks mainly require immovable assets as primary col- lateral in 86 percent of cases, while movable assets account for 73 percent of the average asset value of MSMEs. Similarly, a 2002 agricultural credit survey found that 34­40 percent of medium-size farmers (depending on their acreage) have no access to credit. The survey showed that larger farmers tend to borrow from formal and informal sources more than small farmers do. About 22 percent of landless farmers had formal or informal loans, compared with 38 percent of large farmers. Not surprisingly, access to bank credit increases with farm size, while access to microcredit decreases. But the average loan per acre declines with increasing farm acreage. In line with this, farmers that use both bank and informal loans borrow less from banks than from informal sources.As with MSMEs, the total cost of borrowing for MSMFs is much higher, on average, than is advertised by banks. Advertised annual interest rates range from 8 to 12 percent, while the total cost of borrowing (including transport, documentation, and bribes) is closer to 25 percent. Executive Summary 7 Why Have Government Efforts to Increase Access to Finance Failed for the Missing Middle? BKB and RAKUB, which are supposed to be the main providers of financial services in rural areas, are deeply insolvent. A government refinance facility that should have facilitated agricultural lending has instead become the main source of funding for BKB and RAKUB's growing capital deficits, and agricultural lending is falling--compromising their capacity to serve the missing middle. Private banks have high enough earnings from corporate lending that they have limited appetite to expand into rural areas. Although required to open one rural branch for every four in urban areas, private banks use rural branches to collect deposits that are then used for loans in urban areas. Most lending opportunities in rural areas involve MSMEs and MSMFs, which banks cannot serve profitably under the current legal and regulatory framework and with current lending methodologies. Although lending profitably to these enterprises is not easy, doing so for such farm- ers is even more challenging because of the difficulties that agricultural lending presents--including weather risks. But such markets are harder to serve profitably with traditional approaches to agricultural insurance due to the lack of financial capacity among rural households, lack of distribu- tion channels, low unit size of transactions, and high operating costs. Finally, while MFIs operate in rural areas, they focus on the landless and do not have appropriate lending methodologies to serve larger, more complex clients. How Can the Government Facilitate Increased Access to Finance for the Missing Middle? A number of policies could increase access to finance for the missing middle. This report offers suggestions that are considered the most prag- matic in the current environment, that have been proven in similar circumstances in other countries, and that could lead to tangible improvements for the missing middle. Given the historical significance of BKB and RAKUB, any solution to increasing services in rural areas--especially for MSMFs--will necessarily involve these institutions. Their current market-distorting behavior only reinforces the importance of tackling their underlying problems. In addition to reforming the publicly sponsored financial sector, other useful steps could be taken to facilitate the participation of pri- vate financial institutions and MFIs in increasing access to financial 8 Increasing Access to Rural Finance in Bangladesh services for the missing middle. If the right enabling environment were created with appropriate lending methodologies, private banks could become the main lenders to MSMEs--especially for the top end of the market. But given the current stage of development of the rural banking market, it seems unlikely that private banks will lend to MSMFs in the near future. With suitable products and technology, MFIs could scale up their lending to MSMEs and MSMFs--especially micro and small enterprises and marginal and small farmers. Indeed, such institutions are likely to focus on the bottom end of these market segments, at least in the medium term. Finally, with the right enabling environment and development of the right products, the insurance sector could absorb some commercial agri- cultural risk. In addition, uninsurable financial risk could be covered by the government in a disciplined fashion. To help banks increase lending to MSMEs, the government could: · Reform the enabling environment for MSME lending by reviewing rules on provisioning requirements, reforming and implementing a new legal framework for secured transactions, and strengthening the credit bureau's operations. · Create a technical assistance fund to help selected banks with potential to develop appropriate products and procedures for MSMEs. To help MFIs increase lending to MSMEs and MSMFs, the government could: · Set up a technical assistance fund to increase lending to the missing middle, by introducing new products and lending technologies and appropriate management information systems. To help BKB and RAKUB become the main providers of rural financial services, the government could: · Reform and recapitalize both banks. To facilitate risk transfer for institutions lending to MSMFs, the govern- ment could: · Remove the legal and regulatory obstacles to the development of index-based weather insurance. Executive Summary 9 · Promote the creation of a technical assistance unit that helps interested institutions develop new index-based insurance products. · Create a Fund for Natural Calamities to cover risk that is not commer- cially viable. Note 1. The industrial sector accounts for 27 percent and the service sector, 52 percent (World Bank, World Development Indicators database 2006). C H A P T E R 1 The Missing Middle: Supply-Side Evidence Over the past 10 years, Bangladesh's banking sector has grown considerably, while overall access to financial services has increased moderately. Despite the booming banking sector and government efforts to increase access in rural areas, rural financial markets have shrunk in relative terms. In addition, access to finance by MSMEs and MSMFs--the "missing middle"-- remains limited. The Financial Sector Over the past decade Bangladesh's financial sector has grown steadily, with the ratio of M2 (broad money) to GDP--which indicates the degree of monetization with respect to the real economy--reaching 38 percent in 2003 (table 1.1).1 This level is comparable to that in Pakistan (43 percent) but lower than in India (60 percent). Although Bangladesh's ratio is low in absolute terms, the improving trend suggests that there is growing capacity in the financial sector to channel surplus savings into productive investments. In 2005 banks in Bangladesh provided domestic credit equal to 44 percent of GDP--again, comparable to Pakistan (also 44 percent) but considerably behind India (61 percent). Bangladesh compares favorably with its peers in terms of domestic credit to the private sector, which in 2005 was 11 12 Increasing Access to Rural Finance in Bangladesh Table 1.1. Indicators for Bangladesh's Financial Sector, 2000­05 (percent) Indicator 2000 2001 2002 2003 2004 2005 Broad money (M2/GDP) 32 35 37 38 -- -- Liquid liabilities (M3/GDP) 36 37 39 40 42 45 Private credit/GDP 23 24 26 27 -- -- Domestic credit provided by banks/GDP 37 39 40 38 41 44 Domestic credit to private sector/GDP 26 27 29 29 30 32 Financial sector assets/GDP -- -- -- 46.5 51.9 55.4 Bank assets/financial sector assets -- -- -- 87.6 90.3 90.4 Source: World Bank, Financial Sector Development Indicators and World Development Indicators database. Note: -- = not available. 32 percent of GDP, versus 28 percent in Pakistan and 41 percent in India. (Similarly, in 2003 private credit as a share of GDP was 27 percent in Bangladesh, 22 percent in Pakistan, and 31 percent in India.) Pakistan's lower share shows that there is more "crowding out" in credit markets by state institutions. By the end of 2005 financial sector assets were equal to 55 percent of Bangladesh's GDP. Banks account for 90 percent of financial sector assets. Other parts of the financial sector remain largely undeveloped. For exam- ple, the market capitalization of equity markets averaged just 4 percent of GDP during 1999­2004. And in the insurance sector, premiums equaled only 0.61 percent of GDP in 2003--the lowest level in Asia. (In India this share was 3.14 percent; in Pakistan it was 0.67 percent.) For products other than life insurance, the insurance sector is dominated by Sadharan Bima Corporation (SBC), a state-owned insurance corpora- tion with 21 percent market share--and 51 percent market share when reinsurance is included. But the non-life insurance sector is very small (total premiums equaled 0.2 percent of GDP in 2005), highly segmented (with 43 companies, 10 of which account for two-thirds of market share), and has mixed profitability (with the 10 largest companies generating net profits of 12­14 percent of gross premiums in 2005, while the average return for the entire industry was about 6 percent [AXCO 2006]). Access to insurance is quite limited. In 2005 estimated premium income per capita was $0.8 for non-life and $1.7 for life insurance. By comparison, estimated premium income per capita for non-life and for life insurance were $4.4 and $18.3 in India and $1.8 and $2.9 in Pakistan (Swiss Re 2006). The Missing Middle: Supply-Side Evidence 13 The Banking Sector In 2005 Bangladesh had 48 banks, 9 of which were owned by the state.2 The nine state-owned institutions include four nationalized commercial banks (Agrani Bank, Janata Bank, Rupali Bank, and Sonali Bank) and five development financial institutions, two of which (BKB and RAKUB) were created to meet credit needs in rural areas, two (Bangladesh Shilpa Bank [BSB] and Bangladesh Shilpa Rin Sangtha [BSRS]) to target the industri- al sector, and the fifth (BASIC) to meet the needs of small and medium- size enterprises in both urban and rural areas. In 2005 the nationalized commercial banks and development financial institutions accounted for 47 percent of total bank assets, or 812 billion taka, and 48 percent of deposits, or 683 billion taka.3 In addition, 30 private commercial banks and 9 for- eign banks had 915 billion taka in assets and 733 billion taka in deposits. The banking sector has grown quickly since 1995, with bank assets growing by nearly 23 percent a year between 1995 and 2005--though this slowed to average annual growth of 11 percent between 2000 and 2005 (figure 1.1). The banking sector grew faster than GDP between 1995 and 2005, with loans rising from 21 to 30 percent of GDP and deposits jumping from 26 to 39 percent of GDP (figure 1.2). Bank credit grew by an annual average of 18 percent between 2000 and 2005, outpacing the growth in bank assets during this period by nearly Figure 1.1. Bank Assets, 1995­2006 2,000 30 1,800 25 1,600 assets (left axis) 1,400 20 1,200 taka of 1,000 15 percent 800 billions 10 600 year-over-year change 400 (right axis) 5 200 0 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source: Bangladesh Bank 2006c. 14 Increasing Access to Rural Finance in Bangladesh Figure 1.2. Change in Bank Loans and Deposits, 1995­2005 45 40 deposits 35 GDP 30 of loans 25 20 15 percentage 10 5 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 Source: Bangladesh Bank 2006c. 57 percent and indicating that banks have been steadily redeploying funds from non-credit assets into loans. Despite the rapid growth of banks, indicators of access--such as deposit accounts and loan accounts per 1,000 people--grew only moderately between 2001 and 2005. Deposit accounts increased from 232 to 237 per 1,000 people, while loan accounts rose from 57 to 60 per 1,000 people. In India during this period, deposit accounts per 1,000 people grew from 417 to 432 and loan accounts per 1,000 people from 51 to 71. Moreover, between 1996 and 2005 total deposits as a share of broad money (M2) remained flat, at 86 percent. Ideally, this ratio should increase as more people are exposed to the formal financial sector and decide to open deposit accounts rather than hold cash. Bangladesh's banking sector is in effect two banking systems: a private one that is reasonably healthy, profitable, and focused on urban markets and a state-owned one that is largely insolvent, unprofitable, and the main provider of banking services in rural areas. Indeed, the four nation- alized commercial banks account for 56 percent of rural branches, and the two rural banks (RAKUB and BKB) for 28 percent. Further analysis highlights the contrast between state and private banks. Since 2000 the capital adequacy ratios of nationalized commer- cial banks have steadily deteriorated, while private commercial banks and development financial institutions have generally maintained theirs slightly above the minimum of 9 percent (table 1.2). And foreign banks have had high and steadily rising ratios, reaching 26 percent in 2005. In addition, since 2000 returns on assets and equity have steadily declined for state banks, reaching ­0.1 percent and ­6.9 percent for The Missing Middle: Supply-Side Evidence 15 Table 1.2. Capital Adequacy Ratios of Banks, 2000­05 (percent) 2000 2001 2002 2003 2004 2005 Nationalized commercial banks 4.4 4.2 4.1 4.3 4.1 ­4 Development finance 3.2 3.9 6.9 7.7 9.1 9.2 institutions Private commercial banks 10.9 9.9 9.7 10.5 10.3 9.1 Foreign banks 18.4 16.8 21.4 22.9 24.2 26 Average 6.7 6.7 7.5 8.4 8.7 7.3 Source: Bangladesh Bank 2006a. nationalized commercial banks and ­0.1 percent and ­2.0 percent for development financial institutions in 2005 (table 1.3). During the same period private banks improved their performance, with private commercial banks increasing their returns on assets and equity to 1.1 percent and 18.1 percent and foreign banks raising theirs to 3.1 percent and 18.4 percent. The poor and deteriorating condition of state banks is due to: · Weak governance--for example, boards of directors are political appointees,4 and periodic announcements of interest and sometimes principal repayment waivers encourage a culture of nonpayment among borrowers (in the expectation of future waivers).5 · Weak management, often also politically appointed. · A wide variety of operating weaknesses, including inadequate or non- existent information technology, ineffective control systems, and anti- quated credit and risk management. Finally, despite the seemingly better performance of private banks, reported capital adequacy ratios and earnings for both state and pri- vate banks do not reflect their real status. The discrepancy between banks' reported data and real performance is caused by three factors. First, classified loans and assets are widely underreported due to weak controls and poor information technology. Second, Bangladesh's accounting standards are much less stringent than international stan- dards. Third, some banks fail to make even the loan loss provisions required by the country's weak regulations.6 For the banking sector as a whole, Bangladesh Bank estimates that in 2005 the shortfall in provisions (that is, provisions required by regulation but not made) totaled 45.7 billion taka--just over half of the sector's aggregate capital and reserves. For example, nationalized commercial banks and development financial institutions have reported declining 16 Increasing Access to Rural Finance in Bangladesh Table 1.3. Returns on Assets and Equity for Banks, 2000­05 2000 2001 2002 2003 2004 2005 Return on asssets (percent) Nationalized commercial 0.1 0.1 0.1 0.1 ­0.1 ­0.1 banks Development finance ­3.7 0.7 0.3 0 ­0.2 ­0.1 institutions Private commercial banks 0.8 1.1 0.8 0.7 1.2 1.1 Foreign banks 2.7 2.8 2.4 2.6 3.2 3.1 Average 0 0.7 0.5 0.5 0.7 0.6 Return on equity (percent) Nationalized commercial 1.7 2.4 4.2 3.0 ­5.3 ­6.9 banks Development finance ­68.0 12.3 5.8 ­0.6 ­2.1 ­2.0 institutions Private commercial banks 17.0 20.9 13.6 11.4 19.5 18.1 Foreign banks 27.3 32.4 21.5 20.4 22.5 18.4 Average 0.3 15.9 11.6 9.8 13.0 12.4 Source: Bangladesh Bank 2006a. nonperforming loans since 1998 (table 1.4), but these data are incon- sistent with their reported returns on assets and equity (see table 1.3). The latter returns strongly suggest that the cash flows of state banks have become increasingly negative, indicating that nonperforming loans are actually increasing.7 Government Efforts to Increase Rural Credit To ensure that the growth of the banking sector is balanced between urban and rural areas and between smaller and larger clients, Bangladesh's government has introduced quite a few initiatives aimed at facilitating rural credit in general, and lending to the missing middle in particular. Creating banks to serve rural areas The government created BKB in 1973 and RAKUB in 1987 to serve rural areas, and BASIC in 1988 to promote small industries in urban and rural areas.8 BKB and RAKUB play a key role in delivering financial serv- ices to rural households. As noted, in 2005 they accounted for 28 percent of bank branches in rural areas, as well as 23 percent of outstanding bank rural loans and 8 percent of deposits (by volume). In addition to their focus on supporting agricultural activities, both banks engage in com- mercial and retail business. The Missing Middle: Supply-Side Evidence 17 Table 1.4. Nonperforming Loans of Banks, 2000­05 (Percentage of total loans) 2000 2001 2002 2003 2004 2005 Nationalized commercial banks 38.6 37.0 33.7 29.0 25.3 21.4 Development finance 62.6 61.8 56.1 47.4 42.9 34.9 institutions Private commercial banks 22.0 17.0 16.4 12.4 8.5 5.6 Foreign banks 3.4 3.3 2.6 2.7 1.5 1.3 Average 34.9 31.5 28.0 22.1 17.6 13.6 Source: Bangladesh Bank Annual Report 2005­06 Note: Data are reported net of loan loss provisions. With 947 branches, BKB covers all of Bangladesh except the Rajshahi division--which is covered by RAKUB, with 357 branches. Both banks are entirely owned by the Ministry of Finance, with state-appointed boards of directors and management.9 Both banks are also deeply insolvent (see below).With 27 branches across the country, BASIC was created to be the main provider of credit to small industry in urban and rural areas. The bank's bylaws require that half of its loanable funds go to small-scale industries.10 The bank is the only profitable state-owned bank. Promoting cooperative networks The government has promoted two cooperative networks, covering both traditional cooperatives under the Registrar of Cooperatives and financed by Bangladesh Sambaya Bank Limited (BSBL) and two-tier cooperative systems under the Bangladesh Rural Development Board (BRDB).11 Cooperatives in the two networks are financed through Bangladesh's Bank agricultural refinance facility (for BSBL) and budget allocations (for BRDB). Cooperatives under BSBL are organized in three tiers: BSBL is the apex cooperative bank, under which there are central cooperative societies and primary societies. BSBL on-lends to the central societies, which in turn on- lend to the primary cooperatives. BRDB, by contrast, organizes cooperative societies into a two-tier structure, with primary cooperatives at the local level and others at the thana level. Although allowed, in practice neither cooperative network takes deposits. Rather, both act as credit delivery mechanisms. The Department of Rural Cooperatives and Development oversees the activities of cooperatives. Providing refinance facilities for agriculture Through Bangladesh Bank, the government provides refinance facili- ties to financial institutions interested in financing agriculture. The Agricultural Extension Department sets the terms and conditions for 18 Increasing Access to Rural Finance in Bangladesh agricultural refinance loans between Bangladesh Bank and participating financial institutions, and between these institutions and farmers (box 1.1). BKB and RAKUB are the only active borrowers from the refi- nance facility. Although most of the outstanding amount is overdue, Bangladesh Bank continues to extend the facility because BKB and RAKUB's exposure is explicitly guaranteed by the Ministry of Finance. BSBL and BRDB have outstanding (and overdue) loans from the facility but have not been borrowing recently, as they have been unable to repay the amounts outstanding and have no government guarantee for these loans. Box 1.1 Terms and Conditions of Loans Refinanced under the Agriculture Refinancing Facility Bangladesh Bank provides a refinance facility for agriculture. At the beginning of each fiscal year, interested banks determine their goals for agricultural lend- ing and submit applications to Bangladesh Bank.The limits for each participating institution are approved by Bangladesh Bank's board of directors, and repay- ments to the bank are guaranteed by the government. Between 1996 and 2005 annual disbursements from the facility ranged from 6 to 8 billion taka, while cumulative overdue loans increased from 7 billion taka in 2001 to 34 billion taka in 2005. Refinance facilities are offered for three types of loan products: · Short-term loans (one-year maturities) to cultivate crops (aman, boro, sugar- cane, wheat, oil seed, vegetables), jute, maize, cotton, tobacco, fisheries, and the like. · Medium-term loans (five-year maturities) for livestock and poultry produc- tion (dairy, cattle, goat, buffalo, sheep, ducks), fisheries, nurseries, betel leaf cultivation, beef fattening, fruit gardening, and the like. · Long-term loans (more than five-year maturities) for agricultural equipment, power tillers, deep and shallow tube wells, low lift pumps, hand tube wells, rubber cultivation, and the like. The final annual interest rate is 8 percent for crop loans, 9 percent for other agricultural loans, and 11.5 percent for medium-size and large agricultural activities. Bangladesh Bank charges the participating financial institutions an annual interest rate of 5 percent. The Missing Middle: Supply-Side Evidence 19 Encouraging banks to expand activities in rural areas A recently introduced government policy encourages banks, especially private ones, to expand their activities in rural areas. Bangladesh Bank requires that banks have one rural branch for every four urban ones. Providing crop insurance and waiving payments on small loans In the late 1970s the government designed the traditional multiperil crop insurance scheme that was administered by SBC, a state-owned insurance corporation. The program had to be ended in the early 1990s because the scheme was financially unsustainable, with claims 20 times higher than pre- miums (see chapter 3). Since the program was discontinued, after natural disasters and before elections the government has regularly waived interest, and sometimes principal, on agricultural loans below 5,000 taka disbursed by BKB, RAKUB, BSBL, BRDB, and nationalized commercial banks. In theory the Ministry of Finance should compensate the financial institutions for half of the amount waived; in practice this has almost never happened. Outcomes of government efforts Government efforts to increase access to finance have had mixed results. Lending in rural areas has increased, though at a much slower rate than in urban areas. The importance of banks and cooperatives in rural lending has been declining, while the importance of MFIs has increased. Bank intermediation in rural areas remains limited, with banks transferring 0.5 taka to urban areas for every 1.0 taka in deposits collected in rural areas. The largest providers of agricultural loans, BKB and RAKUB, are deeply insolvent; while BASIC, BSBL, and BRDB are irrelevant in terms of rural loan volumes. Finally, since the closure of the government-sponsored crop insurance scheme, insurance companies have been almost totally absent from rural markets. (Appendix 2 pres- ents a comparison of rural lending markets in Bangladesh and India.) Outstanding loans in rural areas rose 60 percent (by volume) between 2002 and 2005 (the more recent year for which data are available), and rural lending is estimated to account for a third of lending volumes nationwide. During this period MFIs were the fastest-growing financial institutions, with government microfinance programs increasing their loan portfolio by 240 percent and MFIs by 99 percent. Total outstanding bank loans also increased, though by much less--57 percent. Loans from BRDB and BSBL increased by just 5 percent. In 2005 banks accounted for 74 percent of rural lending volumes, MFIs for 20 percent, BRDB and BSBL for 4 percent, and government 20 Increasing Access to Rural Finance in Bangladesh microfinance program for 2 percent (figure 1.3). Between 2002 and 2005 the shares in rural lending of banks and of BRDB and BSBL fell, while those of MFIs and government microfinance programs increased. Within the banking sector, between 2000 and 2005 private banks became more important at the expense of state-owned ones. Still, state- owned banks account for 70 percent of bank loans and 66 percent of bank deposits in rural areas. Sonali Bank is the largest provider of loans and deposit services (by volume) in rural areas. BSB, BSRS, and BASIC account for less than 1 and 0 percent respectively--and their shares have been declining (table 1.5). Since 1996 growth in rural bank credit has been well below that in urban areas (table 1.6). As a result, rural lending as a share of total bank lending fell from 37 percent in 1996 to 28 percent in 2005 (figure 1.4). The government's policy of encouraging banks to expand in rural areas has had mixed results. Bank lending has increased in rural areas, yet between 1996 and 2005 the proportion of loans to deposits in rural branches deteriorated sharply as banks increasingly drained deposits from rural areas to finance urban loans (figure 1.5). In 2005, for every taka collected as a deposit by a rural bank branch, only a little over half a taka (0.57) was lent.This proportion was similar for both state (0.6) and private banks (0.5). Moreover, between 1996 and 2005 agricultural loans fell from 17 percent to 10 percent of total bank lending (figure 1.6). BKB and RAKUB, which were created to be the main providers of financial services in rural areas, are deeply insolvent, as evidenced by growing capital deficits (table 1.7). Given the importance of both banks for rural areas, their insolvency severely undermines the stability of rural financial markets. Both banks' growing capital deficits have been funded by Bangladesh Bank's agriculture refinance facility and 90-day demand Figure 1.3. Estimated Rural Loans by Source, 2002 and 2005 2002 2005 1% 2% 16% 20% 7% 4% 76% 74% banks BRDB+BSBL MFIs government microfinance programs Source: Authors'estimates based on Bangladesh Bank and PKSF (Palli Karma Sahayyak Foundation) data. Table 1.5. Bank Lending and Deposits in Rural Areas, 2000 and 2005 Rural Lending, 2000 Rural Lending, 2005 Millions of Share of Millions of Share of Type of bank taka total (percent) Type of bank taka total (percent) Nationalized commercial banks 109,561 55 Nationalized commercial banks 141,979 46 BKB, RAKUB 53,353 27 BKB, RAKUB 72,923 23 BSB, BSRS, BASIC 1,038 1 BSB, BSRS, BASIC 2,737 1 Private commercial banks 32,861 17 Private commercial banks 92,079 30 Foreign banks 668 0 Foreign banks 1,000 0 Total 197,481 100 Total 310,719 100 Rural Deposits, 2000 Rural Deposits, 2005 Millions of Share of Millions of Share of Type of bank taka total (percent) Type of bank taka total (percent) Nationalized commercial banks 207,662 68 Nationalized commercial banks 317,380 58 BKB, RAKUB 26,125 9 BKB, RAKUB 46,432 8 BSB, BSRS, BASIC 1,050 0 BSB, BSRS, BASIC 2,693 0 Private commercial banks 69,132 23 Private commercial banks 180,341 33 Foreign banks 1,207 0 Foreign banks 2,650 0 Total 305,176 100 Total 549,496 100 Source: Bangladesh Bank data. 21 22 Increasing Access to Rural Finance in Bangladesh Table 1.6. Average Annual Growth in Urban and Rural Bank Loans, 1996­2005 (percent) 1996­2000 2000­05 2004­05 Urban 17.4 20.7 18.4 Rural 10.8 11.5 15.0 Source: Bangladesh Bank, scheduled Statistics April­June 2005. Figure 1.4. Urban and Rural Shares of Bank Loans, 1996­2005 100 80 60 percent 40 20 0 1996 2000 2004 2005 urban rural Source: Bangladesh Bank 2006c. facility.12 In 2006 BKB and RAKUB's outstanding loans from Bangladesh Bank were equal to its equity. One of the main causes of BKB and RAKUB's insolvency is an implicit government policy of using the banks as insurers of last resort and as vote banks.13 As noted, after natural disasters and before elec- tions, interest and sometimes principal on agricultural loans under 5,000 taka are often waived. Possibly due to waivers of principal and interest and rollovers of over- due loans, the average loan size at BKB and RAKUB fell sharply in real terms between 1996 and 2005 (table 1.8).14 This fact is also evident from the number of borrower accounts, which is growing more than The Missing Middle: Supply-Side Evidence 23 Figure 1.5. Proportion of Loans to Deposits by Bank Branch Location, 1996­2005 100 90 urban 80 national average 70 percent 60 rural 50 40 1996 2000 2004 2005 Source: Bangladesh Bank data. Figure 1.6. Distribution of Bank Loans by Purpose, 1996­2005 100 80 60 percent 40 20 0 1996 2000 2004 2005 agriculture, forestry, and fishing industry construction transport and communication trade working capital financing others Source: Bangladesh Bank data. three times faster than the rural population and is partly responsible for the drop in average loan size. In the 20 years since its creation, BASIC has made no dent in the MSME market segment. During that time the bank has acquired just 2,000 clients, of an estimated 6.8 million MSMEs. Between 2000 and 2005 it added an average of 200 new clients a year, or just over 1 client a month per branch. 24 Increasing Access to Rural Finance in Bangladesh Table 1.7. Summary of Audited Financial Statements for BKB and RAKUB, 2005 (millions of taka) BKB RAKUB Total assets 95,284 30,385 Total loans (private and state sectors) 65,675 21,428 Total deposits 56,344 11,578 Reported capital ­15,519 ­98 Total capital shortfall per auditors ­25,989 ­2,275 Adjusted capital/total assets (percent) ­16 ­7 Borrowings from Bangladesh Bank/total assets (percent) 35 51 Sources: BKB 2005; RAKUB, 2005 audited financial statements. Table 1.8. Access Indicators for Private Clients of BKB and RAKUB, 1996­2005 Change, 1996­2005 1996 2000 2004 2005 (percent) Number of borrower accounts BKB 1,724,811 2,647,206 2,872,232 3,050,224 76.8 RAKUB 603,421 725,726 809,862 867,002 43.7 Total 2,328,232 3,372,932 3,682,094 3,917,226 68.2 Outstanding loans in real terms (millions of taka) BKB 32,259 40,519 42,038 44,044 36.5 RAKUB 14,322 10,418 13,115 14,592 1.9 Total 46,581 50,937 55,153 58,637 25.9 Average loan size in real terms (taka) BKB 18,703 15,307 14,636 14,440 ­22.8 RAKUB 23,735 14,355 16,195 16,831 ­29.1 Total 20,007 15,101 14,979 14,969 ­25.2 Source: BKB and RAKUB data. The cooperative movement, which was created to increase lending to rural areas, has never really taken off; it accounts for a mere 4 percent of rural lending volumes. In 2006 BSBL was estimated to have just 32 million taka in capital. Moreover, the cooperative movement is almost entirely dependent on external sources--namely, Bangladesh Bank's agricultural refinance facility and, more recently, budget allocations. Finally, since the end of the government-sponsored crop insurance scheme,insurance companies have almost totally abandoned rural markets.15 To fill the gap, 60 or so MFIs (including Bangladesh Rural Advancement The Missing Middle: Supply-Side Evidence 25 Committee [BRAC] and ASA) have developed micro-insurance products, mainly in rural areas, to accompany their lending activities.An estimated 81 micro-insurance plans cover 20 million rural households. But most of these plans are for life and health insurance; only four MFIs offer livestock insur- ance, and none offers crop insurance. Micro-insurance schemes are self- managed by the institutions and funded by premiums from participating members. The premiums are usually invested in revolving loan funds, with a limited portion invested in banks' fixed term deposits. Premiums are based on specific assumptions, not actuarial calculations. Lending to the Missing Middle No data are available on bank lending to the missing middle. But evidence indicates that such lending is limited. First, agricultural lend- ing has been declining for many years. Second, BKB and RAKUB--the largest providers of agricultural loans--have progressively reduced their average loan sizes in real terms to 15,000 taka. Finally, it is estimated Table 1.9. Microfinance Lending to Micro and Small Enterprises, 2006 Outstanding loans Number of loans (millions of taka) ASA 215,126 3,126 BRAC 87,272 4,064 Grameen Bank -- 5,872 PKSF and partner MFIs 73,869 757 Total 376,267 13,819 Source: ASA, BRAC, Grameen Bank, and PKSF data. Note: Data for ASA are as of June 2006 and are for both small business and small entrepreneur loan products, BRAC data are as of April 2006, Grameen data are as of March 2007, and PKSF and partner data are of December 2006. -- = not available. Table 1.10. Microfinance Lending to Marginal and Small Farmers, 2006 Outstanding loans Outstanding loans of PKSF to MFIs of MFIs to clients (millions of taka) (millions of taka) Number of loans Microfinance for marginal and small farmers 338 329 56,182 Seasonal loans 241 243 22,955 Total 579 572 79,137 Source: PKSF data. Note: Data for the Microfinance for Marginal and Small Farmers Program are as of December 2006, while data for the Seasonal Loans Program are as of October 2006. 26 Increasing Access to Rural Finance in Bangladesh that bank lending to both urban and rural MSMEs accounted for just 2 percent of total lending in 2005.16 The microfinance sector--traditionally the largest provider of small loans in rural areas--has stayed focused on the landless poor using group loans and weekly repayments. Such products are not suitable for MSMEs and MSMFs. Only in recent years have some MFIs started targeting the missing middle, in response to demand for larger loans from existing clients. In 2006, 11 percent of loans from MFIs (by volume) went to enter- prises. Enterprise loans ranged from 20,000­500,000 taka, though most were for 20,000­50,000 taka and had a maturity of six months to two years. The pioneer in this segment is BRAC, which in 1996 launched the Micro Enterprise Lending and Assistance (MELA) program. Since then other large MFIs (such as ASA) and up to 100 smaller ones with loans from the PKSF have launched this type of credit product (table 1.9). Although microfinance lending to enterprises (especially small ones) is limited, microfinance lending to MSMFs is almost irrelevant. In 2006 only 1 percent of microfinance lending (by volume) was disbursed to such farmers, for a total of 79,000 loans (table 1.10).17 Although some tradition- al microfinance lending might be extended to or contribute to households borrowing for crops, most is not targeted at MSMFs. This is mainly because most MFIs limit their lending to households that own less than 0.5 acre of land, which excludes MSMFs. Notes 1. Most of the analysis in this chapter is based on data from Bangladesh Bank.The Bangladeshi financial year ends on 30 June for state banks and on 31 December for private banks. Most of the data in this chapter refer to the 30 June financial year. When data for private and state banks are presented together, they are combined to produce figures for the years concerned. 2. For statistical purposes, Bangladesh Bank treats Grameen Bank as a microfi- nance institution. This report uses the same designation throughout and so excludes Grameen from bank statistics. 3. The percentages in this sentence are based on data as of December 2005, while the absolute numbers are as of June 2005. 4. Board members are senior civil servants, most of whom lack the skills and time required for effective governance. 5. The most recent waiver was announced in early 2007. 6. According to Bangladesh Bank, this is true for 5 of 30 private commercial banks. The Missing Middle: Supply-Side Evidence 27 7. As discussed in chapter 3, for example, BKB and RAKUB cover up the extent of their loan portfolio problems by booking credit losses in "other assets" and "investments" accounts. 8. Bangladesh Bank also manages the Small Enterprise Fund, a refinance facility for small enterprises.This facility is not discussed here because it provides loans only to small urban businesses. 9. BKB's most recent board of directors included representatives of the Ministry of Finance, Telecom and Telephone Board, Ministry of Fisheries and Livestock, Bangladesh Rural Development Board (the government's microfinance pro- gram), Bangladesh Army, Ministry of Water Resource, Ministry of Establishment, Ministry of Agriculture, and BKB management. Members of RAKUB's most recent board of directors included representatives of the Ministry of Finance, Ministry of Agriculture, Rajshahi district, and RAKUB management. 10. Loanable funds equal 65 percent of the bank's customer deposits, 90 percent of its total deposits, and 100 percent of its equity. Because there are different def- initions of small businesses in Bangladesh (see appendix 1), BASIC bank has adopted the Small and Medium Size Enterprise Definition of the Ministry of Industry for manufacturing businesses and the definition of Small and Medium Size Enterprise of Bangladesh Bank for nonmanufacturing enterprises. 11. The second system is referred to also as the "Comilla Model," from the name of the town where this approach was piloted by the Bangladesh Academy for Rural Development. 12. In June 2005 an estimated 71 percent of outstanding loans from RAKUB and 61 percent from BKB were financed by the agriculture refinance facility and 90-day demand facility. 13. Interest waivers apply to BKB and RAKUB as well as nationalized commercial banks. But this report focuses only on BKB and RAKUB, because the others are being privatized and corporatized. 14. In an effort to get some cash, BKB and RAKUB recently implemented a recovery program. The program is technically a rollover of the existing loan portfolio, because borrowers are required to pay only the interest on their loans, but the loans are recorded and provisioned as new performing loans. 15. The only exception is Delta Life Insurance Company, which offers a micro life insurance policy that is very popular in rural areas. 16. Although there is no international benchmark for such lending, 2 percent seems low given that Bangladesh Bank estimates that enterprises with fewer than 10 employees (most of which would be considered micro, small, and medium-size enterprises as defined here) account for an estimated 87 of GDP. 28 Increasing Access to Rural Finance in Bangladesh 17. This is mainly the result of a recent program sponsored by the International Fund for Agricultural Development and implemented by PKSF partner organizations. None of the three largest microfinance institutions (Grameen Bank, BRAC, and ASA) offers these specialized products to marginal, small and medium-size farmers. C H A P T E R 2 The Missing Middle: Demand-Side Evidence This chapter discusses sources of finance for Bangladesh's "missing middle"--that is, MSMEs and MSMFs. It also estimates the potential size of the rural market for both, in terms of number of borrowers and volume of credit. To complement the broad picture of rural access provided by the supply indicators in chapter 1, this chapter measures access from the demand side. The data presented here are from the 2006 Rural MSMEs Finance Survey and from the 2002 Agricultural Credit Survey by Khalily and others. Because the two surveys were among the first attempts to compre- hensively measure rural access to financial services for MSMEs and MSMFs, it is not possible to provide comparisons with earlier periods. The 2006 survey covered 226 MSMEs in the Rajshahi and Moulvibazar districts (see appendix 3 for details on the survey's building blocks, methodology, and sampling),1 while the 2002 survey covered 806 farmers in all districts.2 Access to Credit for MSMEs In 2006, 43 percent of sampled MSMEs had an account in a bank; among these, more that 90 percent saved in their accounts. Savings averaged 67,000 taka for microenterprises and 90,000 taka for small 29 30 Increasing Access to Rural Finance in Bangladesh enterprises.3 In addition, three-quarters of all the enterprises used banks regularly to make domestic transfers. Despite the widespread use of bank services, during 2003­06 only 32 percent of MSMEs borrowed from banks (table 2.1). During the same period 16 percent borrowed from MFIs and 8 percent from informal sources (mainly family and friends)--while 44 percent did not borrow at all. Small enterprises borrowed significantly more from banks than did microenterprises. Micro, small, and medium-size enterprises financed 88 percent of their working capital needs using retained earnings and internal funds, and only 33 percent of new investments were financed using any type of debt.Thus, instead of borrowing, most MSMEs do not try to expand their activities until they have built up internal resources. This pattern is confirmed by their low ratios of debt to assets; this ratio is 2 percent for microenterpris- es and 20 percent for small businesses, for an average of about 8 percent.4 These enterprises do not seem excluded from financial markets due to poor financial performance. Indeed, such enterprises have strong returns on capital (table 2.2) and robust long-term profitability (table 2.3). Instead, many MSMEs seem excluded from financial markets because there is a gap in the market and because lending requirements are not adapted to the business needs of this segment. When borrowing from banks, the average loan for such enterprises is 400,000 taka--while their average loan from MFIs is 38,000 taka (table 2.4). Table 2.1. Sources of Credit for MSMEs, 2003­06 (percent of total credit) Type of credit Micro Small Average Formal 37 77 48 Bank 25 71 32 Microfinance institution 12 6 16 Informal 6 9 8 None 57 14 44 Source: 2006 Rural MSME Finance Survey. Table 2.2. Returns on Capital for Small and Medium-Size Enterprices, 2005 (percent) Return Small 33 Medium 13 Average 15 Source: SEDF 2006. The Missing Middle: Demand-Side Evidence 31 Table 2.3. Profitability Measures for Small and Medium-Size Enterprises, 2005 (percent) 2004 Average, 2002­04 Gross margin 24 23 Operating margin 13 12 Return on invested capital 15 n.a. Source: SEDF 2006. Table 2.4. Average Loan Size for MSMEs, 2003­06 (taka) Loan source Micro Small Average Bank 169,459 590,588 408,081 Microfinance institution 26,000 75,000 38,250 Informal 43,500 43,750 38,929 Source: 2006 Rural MSME Finance Survey. Table 2.5. Average Lending Terms from Banks, Microfinance Institutions, and Family and Friends Microfinance Family and Indicator Banks institutions friends Processing time (days) 40.4 28.0 5.0 Annual interest rate (percent) 13.6 16.4 14.8 Collateral requirements (value of collateral 311.4 56.3 0 as percentage of loan) Source: 2006 Rural MSME Finance Survey. Note: It is better to consider average (mean) values than median values because they give a clearer picture of the impact from an institution's perspective. Getting a loan from formal sources, especially banks, is a long process. Once all the required documents have been submitted, it takes an average of 40 days to get a loan from a bank and 28 days from a MFI (table 2.5). Banks also require considerable collateral. This requirement partly reflects the difficulties that banks face in repossessing collateral and the fact that borrowers rarely consider discounted asset values. In addition, banks almost exclusively require immovable assets as collateral, while MSMEs have mainly movable assets (figure 2.1). When asked their top reason for not applying for credit, 40 percent of these enterprises cited the high costs, direct and indirect (table 2.6). Direct costs include both interest rates and other transaction costs (such as for required documents, including financial statements, titles, and the like). Indirect costs include the long wait to get the money (which trans- lates into missed business opportunities) and an intensive application 32 Increasing Access to Rural Finance in Bangladesh Figure 2.1. Collateral Requirements of Banks and Assets of MSMEs Assets possessed by micro, small, Bank collateral requirements and medium-size enterprises movable assets accepted 14% immovable assets 27% immovable assets required movable assets 86% 73% Source: World Bank 2002. Table 2.6. Top Reasons MSMEs Do Not Apply for Credit Reason Percentage of firms citing High direct and indirect costs 40 Don't need credit 28 Excessive collateral requirements 11 High risks--uncertain of ability to repay debt 6 Other 15 Source: 2006 Rural MSME Finance Survey. process requiring many meetings between borrowers and banks (again translating into missed opportunities, because of the time spent away from the business). Focus group discussions with MSMEs have found that direct costs equal 6­8 percent of loan values. The market size for loans to MSMEs is estimated to be nearly 400 billion taka, with 1 million potential clients (box 2.1). Access to Credit for MSMFs When considering all types of farmers, almost equal proportions have no access to credit, access only to formal credit (from banks or MFIs), access only to informal sources, and access to both formal and informal sources (table 2.7). Larger farmers are less exclusively dependent on informal sources (15 percent of large farmers compared with 26 percent of the landless) and have better access to credit (just 17 percent of large farmers have no access to credit, compared with 33 percent of the landless). Except for the landless, access to formal credit is almost evenly distributed. Finally, access to both formal and informal credit sources, though common for most types of farmers, is more common among large farmers. The Missing Middle: Demand-Side Evidence 33 Box 2.1 How Big Is the Potential Market for Providing Credit to MSMEs? The market size for loans to MSMEs is estimated to be nearly 400 billion taka, with 1 million potential clients (see the table). By comparison, in 2005 total out- standing loans for banks in Bangladesh were 1,117 billion taka. The estimate of market size is based on the following calculations: · An estimated 5.9 million micro, small, and medium-size enterprises in 2003-- which, based on population growth, suggests 6.8 million such businesses in 2006. · Only 2.5 million of the 6.8 million enterprises are estimated to have annual profits above 50,000 taka. · Survey findings indicate that 84 percent of enterprises want to borrow. Of these, only about half will be both willing and able to borrow at a given time (based on the assumption that 70 percent of firms that want to borrow are creditworthy and that 70 percent of these are borrowing at the same time). This means that 1 million micro, small, and medium-size enterprises are potential borrowers. · About78percentoftheseenterprisesaremicro,17percentsmall,and5percent medium-size. · The potential lending volume for each segment is calculated by multiply- ing the number of enterprises in that segment by the average loan for that segment. Average loans are determined by the mean for each segment as defined by this report--for microenterprises, 50,000­300,000 taka; for small, 300,000­1,000,000 taka; and for medium-size, 1­5 million taka. This result assumes that the midpoint of each segment corresponds to the average expected loan for each segment. Estimated Credit Market for MSMEs Share of Lending volume Market total Number of Average loan (billions of share Segment (percent) enterprises (taka) taka) (percent) Micro 78 805,477 175,000 140 35 Small 17 179,366 650,000 117 30 Medium 5 46,420 3,000,000 138 35 Total 100 1,031,263 395 100 Source: Authors'calculations based on 2006 Rural MSME Finance Survey, Bangladesh Bank data, and Daniels 2003. 34 Increasing Access to Rural Finance in Bangladesh Table 2.7. Borrowing Patterns of Farmers, 2000­02 (percent) Access to only Access to Access to formal only formal and Type of farmer Size of farm credit (banks informal informal (acres) No credit and MFIs) credit credit Landless Landless 33.33 18.52 25.93 22.21 Up to 0.5 24.64 26.57 28.5 20.29 Marginal and small 0.5­2.5 25.62 26.48 19.56 28.38 Medium 2.5­5.0 34.31 30.66 13.87 21.17 5.0­7.5 40 12 24 24 Large 7.5+ 17.2 29.79 14.89 38.3 Total 27.05 26.67 20.97 25.31 Source: Khalily and others 2002. The biggest anomaly in these data is among medium-size farmers-- especially those with 5.0­7.5 acres of land, who have the worst access to credit. Only 12 percent of these farmers have access to formal credit (the lowest share among all the farmer groups), and 40 percent have no access to any credit (the largest share among the groups). Like MSMEs, medium-size farmers seem to build up their equity before trying to expand. Among farmers with access to formal credit, MFIs focus on the landless while banks focus on medium-size and large farmers (table 2.8). Probably using bank loans to top up their loans from MFIs, 20 percent of marginal and small farmers borrow from both banks and MFIs. Not surprisingly, the average size of bank loans to farmers increases with farm size (table 2.9). But the average loan per acre declines with increasing farm acreage, and is much lower per acre than loans offered by PKSF's partner organizations participating in its Seasonal Loan Program (9,000 taka an acre). In line with this, farmers that use both bank and informal loans borrow less from banks (though in the range of what those borrowing only from banks do) than from informal sources. In other words, farmers seem to be topping up what they bor- row from banks with larger amounts from informal sources. Similarly, informal borrowing grows with farm size. All this seems to indicate that the loan amounts offered by banks are not appropriate, especially for the largest farms.5 Although the annual interest rates advertised by banks range from 8­12 percent, the real cost of borrowing from banks is much higher for The Missing Middle: Demand-Side Evidence 35 Table 2.8. Borrowing Patterns of Farmers with Access to Formal Credit, 2000­02 (percent) Size of farm Access to Access to Access to Type of farmer (acres) only banks only MFIs banks and MFIs Landless Landless 18.18 72.73 9.09 Up to 0.5 29.90 55.67 14.43 Small and marginal 0.5­2.5 60.30 20.10 19.60 Medium 2.5­5.0 81.69 8.45 9.86 5.0­7.5 88.89 11.11 0.00 Large 7.5+ 81.25 6.25 12.50 Total 58.00 26.49 15.51 Source: Khalily and others 2002. Table 2.9. Average Size of Most Recent Loan Taken by Farmers, 2000­02 (taka) Access to Access to only only Access to both informal Size of farm informal bank and bank sources Type of farmer (acres) sources sources Informal Banks Landless Landless 25,471 n.a. 8,000 2,500 Up to 0.5 9,127 4,938 6,585 3,385 Marginal and small 0.5­2.5 18,681 6,327 11,461 5,020 Medium 2.5­5.0 17,752 10,157 15,550 6,525 5.0­7.5 13,833 13,333 5,040 10,300 Large 7.5+ 79,428 16,461 28,715 14,423 Source: Khalily and others 2002. Note: n.a. = not applicable. Table 2.10. Transaction Costs of Most Recent Loan Taken by Farmers, 2000­02 Cost Type of farmer Size of farm (acres) (percentage of loan principal) Landless Landless 42.98 Up to 0.5 14.48 Marginal and small 0.5­2.5 10.14 Medium 2.5­5.0 9.52 5.0­7.5 3.42 Large 7.5+ 7.7 Total 13.36 Source: Khalily and others 2002. 36 Increasing Access to Rural Finance in Bangladesh Box 2.2 How Big Is the Potential Market for Providing Credit to MSMFs? The market size for loans to MSMFs is estimated to be nearly 88 billion taka, with 4.3 million potential clients (see the box table). By comparison, in 2005 banks in Bangladesh disbursed 42 billion taka for crop production loans. (The disbursed amount is reported rather than the outstanding amount because the latter includes a large share of overdue loans that should be written off, and because most crop loans are annual anyway.) The estimate of market size is based on the following calculations: · The average farm size for each segment was calculated as the midpoint acreage in each segment. · Farm credit requirements were estimated at 9,000 taka per acre, which is the same amount used by PKSF in its Seasonal Loan Program. · The number of farmers willing and able to borrow was calculated by taking the total number of farmers from the 1996/97 Agricultural Census (though the population has grown since then, the proportion of people working in agriculture is decreasing and agricultural land has not increased), then discounting the number by 50 percent for each farmer segment (based on the assumption that 70 percent of those who want to borrow are creditwor- thy and that 70 percent of these are borrowing at the same time). · The potential demand from each segment was calculated by multiplying the average farm size in that segment by estimated loan size per acre and the es- timated number of potential borrowers. Estimated Credit Market for MSMFs Average Loan Number of Lending Market farm size size per farmers volume (billions share Segment (acres) acre (millions) of taka) (percent) Marginal (0.5­1.49 acres) 1 9,000 2.2 20 23 Small (1.5­2.5 acres) 2 9,000 1.0 18 20 Medium (2.5­7.5 acres) 5 9,000 1.1 50 57 Total 4.3 88 100 Source: Authors'calculations based on the 1996/97 Agricultural Census and PKSF data. The Missing Middle: Demand-Side Evidence 37 farmers--on average, 13 percentage points higher (table 2.10). These additional transaction costs mainly involve bribes, costs of required documents, and transportation costs to reach banks.Transaction costs fall as farm size increases, because large farmers usually possess what banks need to assess their creditworthiness (such as formal documents).6 The market size for loans to MSMFs is estimated to be nearly 88 billion taka, with 4.3 million potential clients (box 2.2). Notes 1. Although the sample included 226 enterprises, 37 had an annual profit below 50,000 taka.Thus they are not included in the analysis here because they do not fit the definition of MSME used in this report. 2. See chapter 2 of Khalily and others (2002) for details on the survey's method- ology and sampling. 3. Only seven medium-size businesses were included in the survey, so they are not listed in separate categories in the discussion in this section. But averages for the entire sample in the text and tables based on the survey include MSMEs. 4. By comparison, in many developing countries this share is 30-40 percent. 5. That banks offer loans that are too small may explain why large farmers with access only to informal sources borrow extremely large amounts. That land- less farmers in the same position also borrow large amounts from the infor- mal sector could not be explained and probably is not representative of the category, as there are only seven observations in the sample. 6. Although it is not surprising that the transaction costs of borrowing from the formal sector are highest for landless farmers, it should be highlighted that the sample contains very few observations of landless with access to formal credit. Hence the number may not be representative. C H A P T E R 3 Constraints on Increasing Access to Finance for the Missing Middle This chapter examines the legal, regulatory, and institutional constraints that a critical group of financial institutions--BKB, RAKUB, private banks, and MFIs--face in providing financial services to rural MSMEs and MSMFs. It also analyzes the constraints that traditional, multiperil crop insurance schemes face when serving such farmers. The analysis is based on data from chapters 1 and 2 and case studies of two banks (and detailed discussions with others) and six MFIs. Despite extensive efforts by Bangladesh's government, access to finance in rural areas remains limited--especially for MSMEs and MSMFs. BKB and RAKUB, which are supposed to be the primary providers of financial services in rural areas, are deeply insolvent. Bangladesh Bank's refinance facility, designed to facilitate agricultural lending, has instead become the main source of funding for BKB and RAKUB's growing oper- ating losses, and agricultural lending is falling as a share of total lending. This situation has severely compromised the two banks' capacity to serve the missing middle. Other potential sources of rural finance, such as private banks, earn enough from corporate lending that they have limited incentives to 39 40 Increasing Access to Rural Finance in Bangladesh expand in rural areas. Although required to open one rural branch for every four opened in urban areas, rural branches opened by private banks are mainly used to collect deposits that are then used for loans in urban rather than rural areas. Most lending opportunities in rural areas involve MSMEs and MSMFs--which banks cannot serve profitably under the current legal and regulatory framework and the constraints imposed by current lending technologies. Although lending profitably to such enter- prises is not easy, doing so for such farmers is even more challenging because of the difficulties that agricultural lending presents, including weather risks, limited financial capacity among rural households, lack of distribution channels, low unit size of transactions, and high operating costs. Finally, while MFIs operate in rural areas, they focus on the landless and do not have appropriate lending methodologies to serve larger, more complex clients. Although a variety of financial institutions (such as BRDB and BSBL) serve rural areas, this chapter focuses on the most critical sub- set of institutions for three reasons. First, the institutions analyzed here are the largest rural lenders in terms of loan volumes--BKB accounts for 13 percent of rural lending, RAKUB for 5 percent, private banks for 22 percent, and MFIs for 20 percent. Only nationalized commercial banks have a larger lending volume, and they are not analyzed here because they are being privatized (or in the case of Sonali Bank, cor- poratized). Second, BKB, RAKUB, and private banks have large rural deposit portfolios, so their financial health can affect the stability of the entire rural system. Third, BKB and RAKUB are the main cause of distortions in rural financial markets due to their importance and the government policies to which they are subjected (such as interest rate caps and waivers of principal and interest after natural disasters and before elections). BKB's and RAKUB's Deep Insolvency BKB and RAKUB are the largest providers of financial services in rural areas. In 2005 they accounted for 28 percent of bank branches, 23 percent of outstanding bank loans (by volume), and 8 percent of bank deposits (by volume) in rural areas. But as noted, both banks suffer from increasingly severe liquidity constraints. The banks' financial distress is driven by politically influenced gover- nance and weak internal operating, accounting, and control systems. Politically influenced governance has led to repeated waivers of interest, Constraints on Increasing Access to Finance for the Missing Middle 41 and sometimes principal, on small agricultural loans (less than 5,000 taka) after natural disasters and before elections. These waivers--and a lack of government compensation for them--generate a vicious circle that makes lending unprofitable for the two banks.1 This happens because the banks' already low interest income on such loans falls by a quarter or half (depending on the waiver),2 and even the partial compensation due to the banks for waivers is never actually paid by the government. Moreover, the waivers create a culture of nonrepayment among other clients, who also feel entitled not to make loan payments. Finally, the large amount of overdue loans accumulated over time has made both banks entirely dependent on the agricultural refinance facili- ty for liquidity and cash with which to pay their operating expenses. (Their deposits have been used to fund old nonperforming loans, and their operating cash flows are deeply negative.) This makes the banks' loan portfolios more exposed to waivers, which are provided only for agricultural loans, for which the refinance facility is used. Despite efforts by BKB's management, the bank's condition contin- ued to deteriorate between July 2004 and June 2005, and the scale of its insolvency is such that the bank cannot solve its problems using its own resources. Continued growth in the bank's loan portfolio appears simply to be driving up losses, with the regulatory capital shortfall growing at about the same rate as the portfolio. Moreover, the total cap- ital shortfall grew at about twice that rate, reaching 26 billion taka in 2005 (table 3.1). Despite a capital injection of 420 million taka in 2004, RAKUB reported negative capital in 2005 (table 3.2). The total regulatory capi- tal shortfall identified by its auditors rose to 2.3 billion taka due to extremely negative net income, driven by excessive operating costs and low net interest income, and despite the fact that the bank again reversed provisions against its classified loans in 2005. Although the performance of BKB and RAKUB as indicated by their audited statements is very poor, their actual performance is likely to be much worse. First, in addition to reported losses, both banks' balance sheets include large amounts of questionable items booked as good assets. These include accumulated credit losses (some more than 11 years old), booked as "other assets" in anticipation of compensation from the government, that have not been provisioned for or written off and that are unlikely to be repaid. Second, the fact that interest income on nonperforming loans is being recorded as accrued interest (rather than being suspended, as it 42 Increasing Access to Rural Finance in Bangladesh Table 3.1. Selected Data from BKB's Audited Balance Sheet, 2004­05 Millions of taka Item 2004 2005 Change (percent) Total assets 87,417 95,284 9.0 Total loans 59,793 65,675 9.8 Total deposits 49,982 56,344 12.7 Borrowings from Bangladesh Bank 30,951 33,389 7.9 Reported capital -14,101 -15,519 10.1 Regulatory capital shortfall -20,275 -22,278 9.9 Additional provisions identified by auditors -1,745 -3,711 112.6 Total capital shortfall identified by auditors -22,020 -25,989 18.0 Adjusted capital/total assets (percent) -16 -16 0 Borrowings from Bangladesh Bank/total assets (percent) 35 35 0 Gross loans and advances 59,793 65,674 9.8 Gross unclassified loans and advances 35,418 40,885 15.4 Gross classified loans and advances 24,375 24,789 1.7 Net loans/total assets (percent) 68 69 0 Net loans/deposits (percent) 120 117 0 Gross classified loans/gross loans (percent) 41 38 0 Source: BKB audited financial statements 2004­05. Note: The regulatory capital shortfall is calculated using the bank's audited balance sheet (that is, before adjusting for underprovisioning of loan losses and overreporting of income), while the total capital shortfall identified by au- ditors takes into account identified (but not taken) additional provisions, but not overreporting of income. should be) reflects the banks' lack of effective governance, accounting, management, control, loan administration, internal audit, risk control, and risk assessment systems.3 Finally, the quality of auditing is dubious at both banks. BKB auditors sign off on financial statements that they acknowledge present a com- pletely inaccurate picture of the bank's condition (and, under interna- tional auditing guidelines, would likely require the auditors to express a negative opinion on BKB's viability as a going concern)--reflecting low auditing standards. Similarly, despite improvements in 2005, RAKUB's audit report was also weak and failed to appropriately address the signif- icant problems of capital, earnings, and systems that it identifies. In par- ticular, the bank's auditors identified its weak systems as a major cause of poor financial reporting, yet accepted the bank's books more or less as is. Most troublesome was that the auditors allowed to pass without com- ment the bank's low level of provisions (amounting to about 26 percent of classified loans and 8 percent of all loans) and practice of reversing them to generate "earnings." Constraints on Increasing Access to Finance for the Missing Middle 43 Table 3.2. Selected Data from RAKUB'S Audited Balance Sheet, 2004­05 Millions of taka Item 2004 2005 Change (percent) Total assets 27,989 30,386 8.6 Total loans 18,327 21,428 16.9 Total deposits 10,807 11,579 7.1 Borrowings from Bangladesh Bank 13,402 15,607 16.5 Reported capital 268 ­98 Regulatory capital shortfall ­1,904 ­2,271 19.2 Additional provisions identified by auditors 0 -4 0.0 Total capital shortfall identified by auditors ­1,904 ­2,275 19.4 Adjusted capital/total assets (percent) ­7 ­7 0 Borrowings from Bangladesh Bank/total assets (percent) 48 51 0 Gross loans and advances 18,327 21,428 16.9 Gross unclassified loans and advances 11,025 15,020 36.2 Gross classified loans and advances 7,302 6,407 ­12.2 Net loans/total assets (percent) 65 71 0 Net loans/deposits (percent) 170 185 0 Gross classified loans/gross loans (percent) 40 30 0 Source: RAKUB audited financial statements 2004­05. Note: The regulatory capital shortfall is calculated using the bank's audited balance sheet (that is, before adjusting for underprovisioning of loan losses and overreporting of income), while the total capital shortfall identified by au- ditors takes into account identified (but not taken) additional provisions, but not overreporting of income. Once their loan portfolios have been fully provisioned for losses, and worthless assets written off, the real losses of the banks are estimated at 51­61 billion taka for BKB and 12­16 billion taka for RAKUB. (Appendix 4 provides detailed estimates for both banks, including the assumptions underlying these estimates.) As deeply insolvent institutions, neither BKB nor RAKUB is in a position to scale up lending to the missing middle. Challenges for Private Banks Although private banks have been increasing lending in rural areas, it remains limited. This is mainly because in rural areas most potential bank clients are MSMEs and MSMFs, which have very different fea- tures from large corporations--the traditional clients of private banks--which are easier to serve profitably. Despite the challenges that MSMEs present (including low-value transactions, limited formality, and sensitivity to loan delivery time), lending to farmers presents even more obstacles (including exposure to weather risks). This section focuses on the obstacles that banks face in serving smaller enterprises 44 Increasing Access to Rural Finance in Bangladesh as well as the additional challenges of targeting smaller farmers. In addition, the section on insurance below discusses how risk transfer mechanisms could be developed to facilitate agricultural lending. Micro, small, and medium-size enterprises are usually family businesses, run and managed by one or two people who take full responsibility for all aspects of the business. Although these individuals typically know their business, they are often weak at producing written business and financial plans and are discouraged by the documentation requirements of banks. Moreover, management rarely plans operations far in advance--so when such enterprises require financing, they usually need it immediately. These enterprises usually have limited financial state- ments and few immovable assets. Business and household finances are often strongly intertwined, and such businesses work mainly with cash-- that is, they deposit only part of their revenues in banks. Finally, the banking needs of MSMEs are unlike those of large corpo- rations, and even among these enterprises product preferences differ considerably. Focus group discussions with MSMEs found that they placed varying importance on interest rates, loan processing times, collat- eral requirements, and transaction costs. The smaller the enterprise, the more sensitive the borrower becomes to loan processing times and collateral requirements, and the less sensitive to interest rates (table 3.3). To profitably serve MSMEs, banks need to minimize transaction costs and generate a large number of high-quality loans. As in similar activities with small profit margins, banks need to increase revenue by making many loans while lowering expenses--for example, by making loan offi- cers more productive and avoiding bad loans. For several reasons, banks in Bangladesh find it difficult to serve MSMEs profitably. First, the legal and regulatory framework (especially rules on provisioning, credit bureau reporting, and movable collateral) makes it expensive to serve this market segment. Second, when lending, banks make little or no distinction between large corporations and small- er enterprises in terms of the products they offer or the demands and Table 3.3. Differing Loan Product Preferences for MSMEs Relative importance Product criteria Micro loans Small loans Interest rates Low High Loan processing time High Low Collateral requirements High Low Transaction costs High Low Source: Focus group discussions with MSMEs. Constraints on Increasing Access to Finance for the Missing Middle 45 procedures they apply. Third, banks do not segment MSMEs. Finally, banks do not have organizational structures and monitoring tools con- ducive to achieving high efficiency. Legal and regulatory framework Bangladesh's provisioning rules make lending to MSMEs more expensive than other types of lending. First, general loan loss provisioning for these enterprises is 2 percent, instead of the 1 percent applied to consumer loans and SBLs refinanced under Bangladesh Bank's facility for small and medium-size enterprises (Bangladesh Bank 2006b). In addition to raising the cost of lending to MSMEs--for no reason4--this requirement dis- criminates against banks that make such loans using their own funds, by giving preferential treatment to banks that use the refinance facility. Second, Bangladesh Bank's provisioning rules require that the amount provisioned be calculated as the overdue loan amount minus the discounted value of selected collateral. Selected collateral includes immovable assets and only movable assets that are fully controlled by the bank, making it impossible for MSMEs to pledge assets that they need for daily operations. Hence if such an enterprise has neither immovable property nor movable assets that can be given up for bank possession, lending to it is more expensive because the base for calcu- lating provisioning is higher. Although provisioning requirements for MSMEs are too stringent, overall provisioning rules are too lax and do not create the right incen- tives for careful monitoring. Specifically, term loans for less than five years are classified based on the criteria listed in figure 3.1.5 Given that banks need to test the repayment capacity of previously unbanked clients and that to make small business lending profitable the portfolio Figure 3.1. Classification Criteria for Term Loans of Less than Five Years Substandard Doubtful Loss · Installments in · Installments in · Installments in default > sum of default > sum of default > sum of installments due installments due installments due within next within next within next 6 months 12 months 18 months 46 Increasing Access to Rural Finance in Bangladesh at risk above 30 days has to be low, current provisioning requirements could lead banks to undertake lax supervision of loans to MSMEs. Bangladesh's credit bureau takes a long time to generate reports-- usually 7­10 working days--which slows growth in loans to MSMEs. Moreover, the information provided on such loans is outdated because bank reporting on loans of 50,000­10 million taka is done only quarter- ly. The information provided is also limited. For example, other out- standing loans and guarantees provided by borrowers are not recorded, and only information for the current month is reported. Although the fact that the credit bureau provides outdated and partial information is a problem for lending in general, the issue is more serious for lending to smaller enterprises. The lack of timely, accurate information on these enterprises is a bigger concern due to their limited credit history, less accurate statements, and lower formality. Finally, though not compulsory, Bangladesh Bank's regulations on small enterprise financing (which also cover medium-size enterprises) do not facilitate lending to MSMEs. First, the recommended organiza- tional structure requires that banks split loan sales from monitoring--an approach that does not facilitate in-depth client knowledge, which is crucial in this market.6 Second, the policy requires that loans be reviewed by a special unit, slowing the growth of lending to MSMEs.7 Such enterprises have limited immovable collateral but a wide array of movable assets (see chapter 2). Yet Bangladesh's legal framework on movable property precludes these assets from being used productively in loan contracts, making these assets dead capital. A well-functioning regime for secured transactions is based on the fol- lowing related principles: · Creation of security interests over movable assets is easy and allowed on most assets and by every entity (both physical and juridical persons). · Priority rankings are clearly defined among those who might have claims on property offered as collateral. · There is a place (such as a registry) for making priority interests pub- licly known. · Enforcement of security interests for all assets is fast and cheap. Though this is important for all types of property, it is particularly so for movable assets subject to rapid depreciation. The current framework for movable collateral has major deficiencies that limit businesses' ability to use movable assets as collateral.8 The fact Constraints on Increasing Access to Finance for the Missing Middle 47 that most MSMEs are sole proprietorships makes this an even bigger problem. Appendix 5 presents a detailed analysis of the secured transac- tions regime and its shortcomings. The system's main deficiencies are: · Legal provisions that determine priority among types of creditors, continu- ations of proceeds, and modifications to charges are unclear.9 Different security interest laws have different priority structures, creating confu- sion about the order of priority and so increasing lenders' risk of loss and reducing the leverage value of property used as collateral.The sys- tem permits a security interest to continue in proceeds only in a lim- ited way. For example, if the security interest is an intermediary input, such as cotton, the security interest no longer applies once the cotton is transformed into a garment. In addition, Bangladesh has no clear pri- ority rules for future advances or modifications to a registered charge. Such advances or modifications could include increasing the amount secured or adding new parties to the charge.As a result taking assets as security is risky, because lenders never know where they stand in the priority structure. · Only corporations creating certain types of security interests can register them in the Registrar of Joint Stock Companies and Firms.10 This approach heavily discriminates against MSMEs (as well as MSMFs), most of which are not set up as corporate entities and so are not registered at the registry, and which own mainly movable assets. Moreover, the fact that some interests are not registered makes the registration system in- effective. In addition, for movable property attached in some way to immovable property, there is only case law that the intention of the landowner and occupier determines whether the property will remain the property of the occupier--and not form part of the "land"--and so be considered movable property. This affects lending against fixtures such as generators, machinery installed on property, and certain tem- porary structures. · The transaction costs of filing security agreements are high. First, given the registry's low capacity, firms often must use facilitators to ensure that registration occurs--both because the process is long and because facil- itators know how to expedite it. Second, because the registry is paper- based, verifying collateral is a time-consuming process that does not allow potential lenders to conduct real-time searches of existing liens.11 This substantially increases the risk of secured lending for lenders, who cannot be sure that the collateral offered by borrowers is unencum- bered.Third, laws on evidence require corroboration of any documents 48 Increasing Access to Rural Finance in Bangladesh that are not originals and that do not bear an original signature. This means that only "original" evidence is admissible, increasing the cost of filing and, given the state of the registry, making loss of or damage to original documents probable. · Enforcement of debt recovery is slow. Despite the creation of the money loan court to speed up debt recovery for banks and financial interme- diaries regulated by Bangladesh Bank, enforcement takes one to three years--by which time most movable property will have lost all or most of its value. Although the court must issue a judgment within three months of filing, enforcing a judgment requires filing an execu- tion case.This process makes it difficult for financial institutions to sell collateral because by then the property is usually absent or because such actions are stayed by judicial order due to challenges from debtors. Debtors have wide scope to seek injunctions requiring judi- cial review by the Supreme Court of Bangladesh, and hearings of debt recovery cases are stayed until injunctions have been completed. Moreover, enforcement of orders for repossession is uncertain. This is partly because there is no legislation or procedure for access to, seizure, and sale of secured movable property (except pledges) and partly because enforcement relies on understaffed and undermotivated police--court police officers or ordinary police officers. Complex, expensive lending procedures and inappropriate products Bank procedures for lending to MSMEs are too complex for them, mak- ing such lending unnecessarily time consuming and costly for both the banks and the businesses. On average it takes a bank 29 steps, 9 staff, and 50 documents to issue a loan to such an enterprise--leading to a total approval time of 30­80 days once all the documents have been submitted (table 3.4). Such long procedures mean that banks can approve at most Table 3.4. Application Process for Bank Loans to MSMEs Indicator Average number Steps 29 Meetings with borrower 9 Borrower payments 10 Documents 50 Document length (pages) 200 Staff involved 9 Borrower time (hours) 20 Source: Authors' calculations. Constraints on Increasing Access to Finance for the Missing Middle 49 3­4 loans a month per loan officer. For banks to make a profit serving such enterprises, under current cost structures, each loan officer would have to make at least 8­10 loans a month. Enterprises seeking bank loans must attend an average of nine meet- ings, taking 20 hours. This translates into missed business opportunities for MSMEs both because it takes a long time to get loans and because the owner or manager (and often one of the few employees) must spend time away from the business. In addition, borrowing from banks involves high transaction costs beyond interest payments (see also chapter 2). For a 300,000 taka loan the borrower must pay, on average, an additional 6.6 percent for legal fees, collateral registration, and documentation (such as preparation of financial statements, business plan, title, and the like; figure 3.2). Because of the weak framework for movable collateral, banks require immovable assets as collateral (except BRAC Bank; see box 3.1). Yet MSMEs have mostly movable assets. Finally, banks mainly provide continuous credit products to these enterprises, which restricts the banks' pool of such clients. Continuous credit products are more risky for clients that are often unbanked and that rarely deposit their revenues in banks; hence banks often require that MSMEs have deposit accounts for a few months before they will extend loans to them. Figure 3.2. Transaction Costs on a 12-Month, 300,000 Taka Bank Loan 7 20,000 6 15,000 5 taka 4 10,000 percent3 2 5,000 1 0 0 amount share of loan documentation collateral registration legal fees Source: Authors'calculations. 50 Increasing Access to Rural Finance in Bangladesh Box 3.1 BRAC Bank--Serving MSMEs BRAC Bank is possibly the only bank in Bangladesh that focuses on MSMEs. BRAC Bank was founded in 2001 by BRAC (a nongovernmental organization), the International Finance Corporation, and Shorecap. By 2006 it had 200,000 deposit accounts and 45,000 loan accounts. The average loan was 28,000 taka, while portfolio at risk above 30 days was 6 percent. The bank has 22 branches, 350 unit offices, and 19 automated teller machines (ATMs) across the country. The unit offices only market loans, which are approved at the bank's branches or headquarters. BRAC Bank has developed six products for micro, small, and medium-size enterprises. Most loans are secured with movable assets and, for sole proprietors, with postdated checks from borrowers. Since defaulting on a check is a criminal offense, borrowers have a strong incentive to repay loans secured that way. Although possibly the best approach in the current environment, for sev- eral reasons this is an imperfect way to secure loans. First, defaulting borrow- ers may face criminal offenses--and sending debtors to prison is an archaic, objectionable way of enforcing loan contracts. Second, the amount of the checks used to secure loans cannot be large if they are to be credible for small borrowers. So, while this practice may work for small loans, it will not be enough to ensure that loan sizes are scaled up as enterprises grow. Finally, using postdated checks means that loans are still provisioned as unsecured loans, resulting in higher provisioning requirements. Lack of market segmentation Most commercial banks engage in little segmentation and simply lump all MSMEs into one group, ignoring differences in the market (again, except for BRAC Bank). But as noted, and as highlighted in focus group discussions, there are significant differences among these enterprises.12 Limitedspecializationandnomeasuresofsmallbusinessperformance As in other business activities involving low-value transactions, effi- ciency is crucial when lending to MSMEs. To achieve high efficiency, specialization of staff, training, and procedures is essential. Yet most Constraints on Increasing Access to Finance for the Missing Middle 51 banks in Bangladesh do not have specialized loan officers, training, and procedures for such lending. To increase efficiency, this type of lending needs to be measured and staff incentives aligned accordingly. Although banks in Bangladesh have sophisticated management information systems, they generally do not use them to monitor the efficiency of individual staff members and lend- ing departments, and staff rewards are usually unrelated to efficiency or the quality of the portfolios they manage. Additional challenges in serving MSMFs Banks face additional challenges when lending to MSMFs, including: · Training loan officers in appraising agricultural activities. · Developing products with flexible repayment schedules that fit agri- cultural cycles. · Coping with weather risks. · Developing delivery and monitoring mechanisms that keep costs low while serving sparsely populated areas (though due to its high population density, this is less of an issue in Bangladesh than in many other countries). Challenges for Microfinance Institutions Although MFIs have started expanding access to the missing middle, their outreach remains limited. To scale up lending to MSMEs and MSMFs, these institutions must address a number of challenges. These market segments require new and bigger loan products, mainly targeted at male borrowers, who want to borrow individually. In addition, such clients often have more complex finances--requiring rigorous financial analysis. All this increases lending costs, making the segment more expensive to serve than traditional microfinance borrowers. Given the larger loans and increased risks involved, MFIs must introduce basic management information systems. Finally, MSMFs are exposed to weath- er risks and require financing simultaneously, making loans from MFIs more seasonal. (See appendix 6 for a more detailed analysis of lending by selected MFIs to MSMEs and MSMFs.) Limited numbers of target borrowers Maintaining low costs requires that lenders locate a sufficient number of creditworthy clients in the market area being served, so that loan officers 52 Increasing Access to Rural Finance in Bangladesh specialized in specific market segments can achieve large loan portfolios. Although Bangladesh has a huge number of MSMEs, the number in any given market area may be limited. ASA, for example, had a hard time locating enough small, production-oriented enterprises to meet its goal of 150 borrowers for each of its specialized branches supporting the Small Entrepreneurs Loan program (box 3.2). This situation required a change in strategy, leading to the opening of special windows for this product in ASA's regular branches rather than operating specialized branches. In some areas MFIs may encounter too few potential clients to warrant offering specialized products. Box 3.2 ASA's Microenterprise Programs ASA, one of the three largest microfinance institutions in Bangladesh, offers two specialized microenterprise programs: the Small Business Loan program and Small Entrepreneurs Loan program. The Small Business Loan program was launched in 1993 to serve traders, shopkeepers, and graduates of ASA's small loan program. The first cycle of loans ranges from 15,000­50,000 taka, and suc- cessive loans are increased by a maximum of 5,000 taka a year.The program uses a group lending methodology, with weekly repayments. The Small Entrepre- neurs Loan program, started in 2003, targets production-oriented microenter- prises. Loans range from 30,000­200,000 taka, for a maximum of 24 months, with monthly rather than weekly payments. No collateral is required for either product, but a guarantor is required for the SBLs program. By June 2006 this program had nearly 200,000 loans, with 2.4 billion taka outstanding. At the same date, the Small Entrepreneurs Loan program had issued 15,805 loans worth 723 million taka. ASA face two challenges in trying to scale up the Small Entrepreneurs Loan program. First, the requirement that such loans be made only to production-ori- ented businesses limits the universe of potential clients. Second, the fact that ASA does not take physical collateral limits the possibility of increasing the max- imum loan size under the program. Moreover, it would be difficult for smaller microfinance institutions to adopt this program model, because they may not have enough capital to offer such large loans or highly skilled loan officers able to estimate cash-flows of potential borrowers. Constraints on Increasing Access to Finance for the Missing Middle 53 Serving male borrowers Most MFIs lend mainly to women, because they are more amenable than men to joining loan groups and attending meetings, accepting responsibility for repaying loans of defaulting group members, and meeting other responsibilities of group membership. The opportunity to leave home and interact with others in weekly or monthly meetings offers a socially accepted reason to leave the confines of the home and helps women build confidence to engage in business activities. In addi- tion, women may be more susceptible than men to pressure from loan officers and other group members to make loan payments on time. By contrast, crop farming and large-scale trading and production activities are dominated by men.As a result, when lending for such activ- ities, MFIs must decide whether to continue to target women (while knowing that men will be the ones who actually use the loan funds) or, alternatively, adjust their lending approaches to lend directly to the men who operate many farms and MSMEs (box 3.3). Adjusting lending methodologies Microfinance institutions in Bangladesh traditionally extend loans that are "secured" using the Grameen Bank's group lending approach. But many MSMEs and MSMFs prefer individual loans. Given the larger loans involved, and from a purely financial perspective, MFIs need to shift to individual loans for larger clients. In fact, only by shifting to individual lending will MFIs be able to analyze individual borrowers and set loan terms and conditions based on the risks they present. Group members in most MFIs have little liability for loan payments by other group members, though loan officers may urge group leaders to pressure delinquent borrowers to repay. Few MFIs have started shifting to individual loans and different forms of collateral. Making individual loans even at the micro level requires considerable planning and building of institutional capacity, followed by testing, evaluation, and mainstreaming of new products (Dellient and others 2005). These challenges will be even greater if MFIs try to expand lending to farmers. Small loan sizes and lending for household activities Microfinance institutions currently focus on seasonal and farm loans ori- ented to the enterprise, activity, and person being funded. Because farms are small, the demand for loans for a specific crop--especially rice--is small even when improved production techniques are used. Moreover, average farm sizes rarely grow over time, so there is little opportunity for 54 Increasing Access to Rural Finance in Bangladesh Box 3.3 BEES's Product for Marginal and Small Farmers Through 10 of its 83 branches, Bangladesh Extension Education Services (BEES), a microfinance institution, targets marginal and small female farmers. As of 2006 BEES had 11,400 farmer borrowers (or 12 percent of its total borrowers). During the same year it had disbursed 72 million taka in farm loans (equivalent to 13 percent of its disbursements for the year). These loans are essentially an adaptation of the Grameen Bank's group lend- ing approach applied to agriculture, with women engaged in a variety of agri- cultural activities as clients. Loans are to be repaid within a year in 45 weekly installments. BEES forms groups of 15­40 women that meet for one hour a week to engage in financial transactions and occasionally receive other services, such as monthly health checkups. Mandatory weekly savings earn an annual interest rate of 5 percent. One percent of the loan is deducted as a premium for micro life insurance and to cover other service expenses. Although this product has succeeded in increasing lending to small farmers, there are limits on BEES's ability to scale up its lending--in terms of both loan sizes and volumes. To be able to target larger farmers, BEES will have to start lending to men, who are likely to prefer individual loans. Making individual loans requires estimating households' monthly cash flows and evaluating borrowers' repayment capacity. Moreover, BEES's group-based weekly repayment lending approach is not particularly well suited to the seasonality of crop farming. MFIs to increase loan sizes by lending for individual crops. Relaxing this constraint would require these institutions to develop greater capacity to appraise and lend against the entire portfolio of household economic activities, including all crops and products produced for consumption and sale as well as nonagricultural activities undertaken by all household members who will contribute to loan repayment. Training loan officers in financial analysis Microfinance institutions are accustomed to using a Grameen-style methodology to serve clients capable of generating the regular cash flows needed for weekly payments. As a result, microloan portfolios are dom- inated by borrowers engaged in small-scale trading, poultry, and dairy Constraints on Increasing Access to Finance for the Missing Middle 55 production. This methodology does not require loan officers to have the advanced skills needed to evaluate borrowers' businesses. But offering loans for larger crop or livestock operations, or for produc- tion loans for nonagricultural activities, requires that loan officers have more business and technical knowledge about the enterprise being financed and the economic subsector involved, as well as greater capacity to evaluate clients, their businesses, and their competition. It also requires that loan officers have a social and economic status approximately equal to their borrowers. Introducing management information systems and new technology Loans to MSMEs, whether to graduates of standard microfinance pro- grams or entirely new clients, will by definition be larger on average than traditional microfinance loans. Making larger loans raises the risk of larger losses for MFIs in cases of default. To manage such a portfolio effectively, introducing even a basic management information system, including care- ful client monitoring, is essential. The strong profitability of Bangladesh's microfinance is based on their low costs--reflecting, among other things, their low cost of funds, good portfolio quality, and group lending methodology. In particular, group lending has allowed MFIs to generate a high number of borrowers per staff and so low costs per borrower. But serving MSMEs using individual lending requirements, more complex financial analysis, and better- trained staff will be more expensive, and these costs may be only partly offset by larger loan sizes. To keep staff productivity high and costs low, increasing the use of technology may be essential. Managing liquidity When serving MSMFs, MFIs must address the liquidity management challenge created by the seasonality of credit demand from farmers. Financing crops implies making a lot of loans at the beginning of the planting season--followed by several months of low or no repayments while crops grow or cattle fatten (unless borrowers have other sources of income). Then at harvest time, borrowers earn income and prefer to repay their loans in lump sums. The resulting large flow of funds into MFIs may accumulate and then largely sit idle until the start of the next crop or fattening season. Crop lending also introduces seasonality into the workloads of loan officers, so MFIs may need to adjust employee compensation levels, 56 Increasing Access to Rural Finance in Bangladesh require compulsory furloughs during slack periods, and use improved technology to accelerate data processing. Addressing weather risks Weather risk is the main source of income instability for farmers borrow- ing from MFIs, because their agricultural activities depend on rainfall and access to water for irrigation is limited except in areas where tube wells have been installed. Accordingly, MFIs become financially vulnerable to weather risk as well. Weather-induced crop failures, usually due to low rainfall, often cause overdue repayments on crop term loans--threatening the portfolio quality of MFIs. To address this problem, these institutions should diversify their risks by developing loan portfolios that cover large areas. But most MFIs do not have the opportunity to diversify geograph- ically; hence the importance of agricultural insurance. Challenges for Insurance Providers Serving MSMFs In the late 1970s Bangladesh's government designed a multiperil crop insurance scheme that was administered by SBC. The scheme failed because of the standard drawbacks of farm-level, multiperil crop insurance: · Adverse selection.The insurance product was voluntary, but due to lim- ited resources it was offered only in selected, higher-risk thanas. · Poor due diligence and supervision. Many farmers insured unproductive land or filed claims when crops had been destroyed by causes other than natural disasters. · Weak control over loss assessments. It proved difficult to assess losses using an objective methodology. A variety of methods was used, including crop cutting and eye estimation, and there were frequent disputes. · High administrative costs. · High crop yield guarantee. The scheme guaranteed 80 percent of expected crop yields, which is high by international standards.13 · Individual farmer insurance. Insured yields were based on thana-level average yields, while losses were adjusted and indemnified based on actual yields on individual farms. This approach exposed the scheme to claims arising from poor management by individual farmers. This insurance product was--like all traditional crop insurance--more appropriate for large farmers. Such programs rely on examinations of Table 3.5. Features of Traditional Crop Insurance and Index Insurance Function Traditional Index Establishing yield of insured Insurers must establish farm- or district-level yield. Payout based solely on the measured index (such as rainfall). Cost: High Cost: Low Underwriting Requires assessment of individual or local risk. Not required, though insurers need to screen clients for insurable interest (for example, to ensure that the farmer is growing a crop). Cost: High Cost: Low Policy sales Sales process requires high skills because it involves Sales process requires good product knowledge, but underwriting decisions. no major underwriting decisions occur in the sale process. Cost: High Cost: Medium Paperwork and technology Generally complex. Need to record yield history, Simple certificates or coupons. crop establishment during the season, and farm and field details Cost: High Cost: Medium Field inspection Need to check for emergence of crop. Not required. Cost: High Cost: Low Loss adjustment Need to inspect crop damage and adjust claims. Not required: Payments based on measured index. Cost: High Cost: Low Claims payment Settlement of claims. Settlement of claims. Cost: Low Cost: Low 57 58 Increasing Access to Rural Finance in Bangladesh individual farms and evaluations of yield, driving up costs considerably. This approach is not suited to serving MSMFs, as this is a low-value transaction business. Costs aside, this type of weather risk management suffers from moral hazard problems--especially for MSMFs, who are less likely to use fertilizers and pesticides. Indeed, when they know they are insured, farmers often take less care of their crops. Instead, commercially viable approaches to agricultural insurance for low-value transactions are based on the use of objective factors. So-called index insurance relies on an index (such as the amount of rainfall), based on historical events and yields, to set objective triggers that determine if and when there are payouts to farmers (table 3.5). Index-based crop insurance is actuarial-based and is priced on the probability that a certain type event, and hence amount of damage, will occur within a cer- tain period. This type of insurance reduces moral hazard, does not require indi- vidual evaluations, and can more easily be reinsured into international markets. Moreover, to further reduce transaction costs, insurance com- panies often do not distribute their products directly, but sell them through existing networks (such as those of MFIs). In addition to the technical difficulties of developing an index, insurance companies in Bangladesh face two key legal and regulatory obstacles. First, the existing insurance law allows only licensed insurers to issue policies, and forbids institutions such as MFIs from acting as agents or brokers.14 Second, under current legislation insurance companies must define insurable inter- ests as under conventional property insurance, where losses are indemnified based on actual losses or damages to specified insured property. But with index insurance, there are no measures of actual losses--only the proxy payment triggered and measured by the index. Notes 1. The government is supposed to repay half of the waived amounts, but that rarely happens. 2. For loans refinanced under the agriculture refinance facility, and before waivers, BKB and RAKUB have a spread of 3­4 percentage points. 3. Even under the more permissive provisioning rules for loans financed under the agriculture refinance facility, agricultural loans must be provisioned at 100 percent 60 months after their due date (Bangladesh Bank 2006b). 4. International experience shows that, if done properly, the risk of lending to micro, small, and medium-size enterprises is low, with a portfolio at risk above 30 days below 3 percent. Constraints on Increasing Access to Finance for the Missing Middle 59 5. This section focuses on term loans because many MSMEs have no experience with bank borrowing. Thus term loans requiring frequent repayments are usually less risky and allow banks to test new clients. But provisioning is also lax for other types of loans (such as continuous loans). 6. While every bank chooses the organizational structure that suits it best, all successful lending programs for MSMEs have full-cycle loan officers. Personal relationships and knowledge are key for clients with a high degree of informality. 7. Although checks and balances must be in place and loan officers should not approve loans on their own, review by a central unit slows the growth of the loan portfolio. Most successful programs for lending to MSMEs use branch committees for loans under a certain amount--allowing for checks and balances while also ensuring speedy approval. 8. This report does not analyze the regime for immovable collateral because this complex topic requires separate analysis. As an indication of the complexity of the problem, an estimated 80 percent of cases pending in Bangladesh are related to land disputes. 9. Examples of competing creditors include tax authorities for tax arrears, lenders of overdue debts, suppliers of unpaid-for goods, receivers in bankruptcy, and judgment creditors to enforce court awards in their favor. 10. For example, financial leases on assets cannot be registered in this registry, and any liens on motor vehicles are registered in the motor vehicle registry. 11. Even with an upcoming computerization project supported by the Small Enterprise Development Facility, computer systems will be desktop- or network- based. Such systems require a lot of maintenance and training and are unsuit- able for use by the general public. 12. For example, Arab Bangladesh Bank has one segment for all loans under 100 million taka; BASIC Bank has one segment for loans of 300,000­1,500,000 million taka; Dhaka Bank has one segment for loans of 300,000­5 million taka; Islamic Bank has two segments, one for loans of 200,000­5 million taka and one for loans of 5 million­10 million taka; and United Leasing Company has three segments, one for loans of less than 10 million taka, one for loans of 10 million­100 million taka, and one for loans of 100 million­300 million taka. 13. Although profitable guarantee levels for individual, multiperil crop insurance are related to a number of factors--including crop type, agro-ecological conditions, and the risk profile of a location--they usually do not exceed 65­75 percent. 14. The regulator seems to draw a distinction between potential index insurance products, which ideally should be reinsured, and microinsurance products issued by microfinance institutions (see chapter 1). The latter are allowed because the risks are retained by the microfinance institution, without the involvement of an insurance company.With index insurance the risks are trans- ferred externally, because insurance companies transfer them to reinsurers. C H A P T E R 4 Increasing Access to Finance for the Missing Middle This chapter offers suggestions on increasing access to finance for the rural missing middle. Government efforts should include improving the enabling environment and assisting institutions with the potential to expand access. A number of policy options have the potential to increase access to finance for the missing middle. This chapter provides suggestions that are consid- ered the most pragmatic in the current environment, that have proven successful in similar circumstances in other countries, and that could lead to relatively rapid, tangible improvements for the missing middle. Given the historical significance of BKB and RAKUB, any effort to expand financial services in rural areas--especially for marginal, small, and medium size farmers--will necessarily involve these institutions.Their cur- rent market-distorting behavior only reinforces the importance of tackling their underlying problems. This chapter offers concrete suggestions for rehabilitating BKB and RAKUB and, for means of comparison, also esti- mates the costs of leaving them unreformed or closing them. This is not to say that increasing access to financial services for the missing middle could not also involve actions involving other institutions, such as cooperatives and BASIC. Exploring such options would be useful but would not detract from the need to tackle the challenges facing BKB and RAKUB. 61 62 Increasing Access to Rural Finance in Bangladesh In addition, steps could be taken to facilitate active participation by private banks and MFIs in increasing access to financial services for the missing middle. With the right enabling environment and appropriate lending methodologies, private banks could become major providers of loans to MSMEs--especially at the top end of the market. Unfortunately, given the current stage of development of the rural banking market, it seems unlikely that private banks will lend to MSMFs in the near future-- especially if there continues to be no crop insurance system. With suitable products and technology, MFIs could scale up their lend- ing to MSMEs and MSMFs, especially to micro and small enterprises and marginal and small farmers. Given their experience and orientation, MFIs are likely to focus on the bottom end of these markets, at least in the medium term. Finally, with the right enabling environment and development of the right products, insurance companies could begin to purchase at least part of the commercial agricultural risk pool, while commercially unin- surable risk could be covered by the government in a financially disci- plined fashion. The following sections describe these recommendations in greater detail. Transforming BKB and RAKUB into the Main Providers of Financial Services in Rural Areas If rehabilitation of BKB and RAKUB is undertaken with the goal of restoring their long-term sustainability, it should focus on: · Transforming the governance of the two banks, to protect them from political interference. · Transferring the function of insurer of last resort for the agriculture sector to more appropriate institutions. · Recapitalizing the banks in a way that frees them from the legacy of decades of losses--allowing them to make a fresh start. · Restructuring operations so that the banks can provide cost-effective, well-designed financial services to rural markets. Rehabilitating BKB and RAKUB would cost about 96 billion taka (see appendix 4).This undertaking would be a major challenge for the govern- ment and Bangladesh Bank--but if it succeeded, the two banks could sig- nificantly increase rural access to financial services. A few other countries, including Mongolia, have managed to reform agricultural banks (box 4.1). Increasing Access to Finance for the Missing Middle 63 Box 4.1 Khan Bank of Mongolia Founded in 1991, the Agricultural Bank of Mongolia was designed to be the main provider of financial services in rural areas. But by 1999 the bank was in receivership due to many years of operating and credit losses. Management was poor, political interference was felt at every level of the bank, and it was failing to deliver financial services and products needed by rural communities. In 2000 the bank was recapitalized and a new, donor-financed management team was installed. The bank's governance was reformed by creating an inde- pendent board of directors, with two members appointed by the donor, two members appointed by the government, and an independent chairperson. The new management team focused on turning the Agricultural Bank into a full-service financial institution that could attract deposits and relend them in rural areas. The bank's staff was retrained, and in many cases underperforming managers and staff were replaced. Still, the number of staff members more than doubled as the bank opened new branches and offices in the countryside, with the number growing from 269 in 2000 to 379 in 2004. A key component of the bank's new strategy was to encourage growth in deposits by renewing lending (people tend to deposit where they can borrow, if needed). Many new lending products tailored to the needs of rural communities were introduced, as were other services (such as payments). In late 2003 the Agricultural Bank of Mongolia was sold to a Japanese invest- ment group and renamed Khan Bank. Between July 2000 and February 2004 the bank was transformed. During this period: · Half of Mongolian households began using (and continue to use) Khan Bank. · 900,000 new loans were made, averaging $382, with a default rate consistently below 2 percent. · 90 percent of the bank 's lending shifted to rural areas. · Deposits grew 740 percent, while the share of government deposits in Khan Bank's total deposits fell from 50 percent to 7 percent. Source: Dyer, Morrow, and Young 2004. A comprehensive rehabilitation process for BKB and RAKUB would include: · Immediately placing the banks under tight supervisory controls to prevent them from incurring new losses, and installing independent temporary 64 Increasing Access to Rural Finance in Bangladesh boards and competent temporary management.1 Although both banks are prime candidates for being placed under administration, doing so would likely be counterproductive as such announcements could trigger a run on deposits and, potentially, the collapse of both banks. It might be more practical to place both banks under tight supervisory control using consensual arrangements--possibly in the form of supervisory agree- ments--and, given the banks' state ownership, the government could compel them to cooperate. A team of senior supervisors would be placed in each bank to monitor its activities and all management deci- sions would be subject to confirmation by the supervisors. The banks would not be allowed to take further deposits until after recapitalization and a new management structure were in place.Tight liquidity controls would be placed on branches and loans, and expenditures would also become subject to approval by the supervisory teams. If the government wants to continue lending to farmers during this period, and in order to explore new lending methodologies, special offices could be created to pilot new agricultural products. These pilots would ensure that farmers continue to receive credit and that new lending methodologies are trialed, so that once the banks are in a position to start lending again, the pilots could be rolled out to other branches. · Updating financial data and conducting diagnostics for both banks. Audits--using international financial reporting standards--should be conducted to determine the up-to-date costs of rehabilitating the two banks. Diagnosing each bank's systems and governance would generate detailed recommendations for new systems and governance as well as opportunities for reducing operating costs, primarily (though not exclusively) from automation. · Deciding whether to merge the two banks. Given the potential efficien- cies of a merger, and the fact that the combined bank could use a single information technology package to support new systems and controls, there seems to be a strong argument in favor of merging BKB and RAKUB during the rehabilitation process. Moreover, rehabilitating a merged bank would simplify the technical assistance required and reduce the management challenges involved. · Preparing a detailed recapitalization plan. Using information from the financial diagnostic, this plan would model the restructuring of each bank (or the merged bank) and determine the financial instruments and amounts required to restore them to full compliance with regula- tory capital requirements, prudent liquidity levels (cash and cash Increasing Access to Finance for the Missing Middle 65 equivalents equal to about 20 percent of deposits), and at least zero pretax net income and positive operating cash flows. The design of recapitalization instruments would also have to reduce and restructure the terms of the banks' borrowing from Bangladesh Bank to sustainable levels. The recapitalization would mostly be in the form of bonds to provide the banks with liquidity only to the extent that they become able to redeploy it prudently, by complying with detailed financial and operational indicators, and to reduce the immediate fiscal impact. As part of the recapitalization plan, a decision would need to be made on whether to invite development-oriented investors (such as the Inter- national Finance Corporation, Asian Development Bank, Agha Khan Foundation, or Japan Bank for International Cooperation) to make minority investments in BKB and RAKUB. The goals of such invest- ment would be to reduce the short-term cost of recapitalization and improve the banks' long-term governance and operations (box 4.2). In many countries development-oriented investors interested in holding minority shares have made meaningful contributions by lowering recapitalization costs, strengthening governance, reducing political interference and ensuring sustainability in financial institutions with a developmental mandate. · Developing a new business strategy and business plan to allow the banks to operate prudently and profitably and to serve rural markets. Experi- ences in Mongolia, Tanzania, and Thailand indicate that rural-focused banks with suitable governance and investors with the standing to re- sist political pressure can function prudently while serving a broader social mission. To ensure that the banks maintain their focus on rural areas and on agriculture, their activities in large cities could be limited and the banks could be required to lend a certain percentage of the loan portfolio to the agricultural sector. As part of their new business models, BKB and RAKUB's current quasi-insurance functions should be excluded from their mandates, as the banks lack the technical, operational, and financial capacity required to provide insurance services. While it is legitimate for the government to protect farmers against the financial consequences of natural disasters, this function should be provided by specialized entities, such as insurance companies (see below, in the section on weather risk management). · Creating a bank rehabilitation unit. Such a unit, in the Ministry of Finance or at Bangladesh Bank would oversee and control the banks' activities during the rehabilitation process. 66 Increasing Access to Rural Finance in Bangladesh Box 4.2 Benefits of Having Development-Oriented Investors as Minority Shareholders Inviting development-oriented investors to be minority shareholders can bring two major benefits: lowering recapitalization costs and improving governance and operations. Lowering recapitalization costs. Development-oriented investors can make a significant contribution to the capital of the bank concerned. Although they would not help cover historical losses (which should be covered by the gov- ernment during the restructuring process), they would invest in the new capital structure of the bank--giving it additional financial strength and better enabling it to fulfill its development function. In addition, development investors may be willing to provide additional financial resources (such as by extending credit lines to the bank or financing technical assistance for opera- tional restructuring) alongside their capital investment to provide liquidity so that the restructured bank can build its new business. Improving governance and operations. To invite development-oriented minor- ity shareholders, the government would develop and commit to: · A detailed business plan for the restructured bank. Once the plan has been agreed to, it becomes a legally binding document. · Appointingaboardofdirectorscomposedmainlyofindependentprofessionals. This arrangement is designed to increase the board's competence and inde- pendence, and in turn improve the quality of its management supervision and ability to exercise independent governance. · Having large credit decisions and changes in credit policies unanimously approved by the board. This approach provides further insulation from political interference in the bank's decision making. Rehabilitating BKB and RAKUB would likely take up to three years.The rehabilitation period is long for two reasons. First, it would take time to implement comprehensive operational and system improvements. The banks require a complete overhaul, and modernization of their information technology and staff--from tellers in branches to senior management-- would require extensive training in how to use the new systems. Second, the banks require an overhaul of their business strategies, involving the Increasing Access to Finance for the Missing Middle 67 development and implementation of new policies and procedures for all aspects of bank operations and the development and introduction of new credit, deposit, and service products, all requiring extensive training of staff and management and the introduction of new technology. There are two alternatives to rehabilitating BKB and RAKUB: doing nothing or closing the two banks (box 4.3). Box 4.3 Financial Costs of Alternatives to Rehabilitating BKB and RAKUB The government has three policy options for addressing BKB and RAKUB's insol- vencies.The first is a"do nothing"strategy that would leave the banks unreformed. The second is to close and liquidate the banks. The third is to rehabilitate them. The costs of rehabilitation are estimated at 96 billion taka; the costs of the first two options are estimated below. Doing nothing The government could simply do nothing to address the problems posed by BKB and RAKUB. Under this option the two banks would remain open but unre- formed, and Bangladesh Bank or the Ministry of Finance would need to keep ad- vancing additional funds to pay the banks' operating expenses and keep some credit flowing to their customers. This approach might be feasible for up to three years--but at that point accumulated operating losses would be so high that the banks would have to be liquidated or recapitalized. Assuming that no less than 10 billion taka a year in new funds were provided to BKB and RAKUB by Bangladesh Bank or the Ministry of Finance, the banks would incur additional credit losses of 7­8 billion taka a year (see the table). This estimate, based on the banks'track records, assumes that for every taka of addi- tional financing provided to the banks, 70­80 percent would turn into nonper- forming loans. Closing and liquidating BKB and RAKUB In addition to the social and economic costs, closing and liquidating RKB and RAKUB would have a high one-time fiscal cost.These costs are driven by the banks' state ownership and their extensive entanglement with the state on both the asset and liability sides of their balance sheets. State ownership means that there is an implicit guarantee of the banks' deposits, so these would have to be paid off (Continued) 68 Increasing Access to Rural Finance in Bangladesh in full and in cash. The banks' second largest source of liabilities, Bangladesh Bank, has a government guarantee of its lending to the two banks and would be bankrupted if the government did not pay these guarantees. Finally, the banks' employees are civil servants, and so are entitled to compensation and pension benefits for which the government is responsible. It is difficult to quantify this liability, but it would be substantial given the size of the banks' workforces and payrolls. Liquidating the banks would cost more than 100 billion taka, ex- cluding the staff compensation and pension benefits. Government Costs of Options for Resolving BKB and RAKUB (millions of taka, based on 30 June, 2005, balance sheets) Cost Overall Option Year 1 Year 2 Year 3 Total Total Doing nothing but provide refinancing 13.5 13.5 13.5 40.5 141.5 Operating losses 6.0 6.0 6.0 18.0 18.0 Losses from new lending 7.5 7.5 7.5 22.5 22.5 Liquidation costs 101.0 Closing and liquidating 129.0 ­14.0 ­14.0 101.0 101.0 Paying off liabilities 143.0 143.0 143.0 Recoveries from assets ­14.0 ­14.0 ­14.0 ­42.0 ­42.0 Source: Authors'calculations based on BKB and RAKUB audited financial statements 2005. Helping Banks Serve Rural MSMEs To facilitate bank lending to MSMEs, the government could promote an enabling environment that makes such lending less costly. The gov- ernment could also support grant-funded technical assistance to help some banks adapt their lending procedures to significantly increase such lending.2 Improving the enabling environment To reduce the costs of serving MSMEs, Bangladesh Bank could reduce provisioning requirements on such loans to 1 percent--the same share it requires for consumer loans and loans issued under the refinance facility for small and medium-size enterprises. Bangladesh Bank could also allow movable collateral not in the possession of banks among assets whose value can be discounted when calculating provisions. In parallel, the government could reform the legal framework for Increasing Access to Finance for the Missing Middle 69 secured transactions and upgrade the registry to enable all businesses to safely register pledges on movable assets. Bangladesh's credit bureau should be strengthened to reduce informa- tion asymmetries between lenders and borrowers. Bangladesh Bank could upgrade the services provided by the bureau by: · Ensuring that the bureau's data are electronically accessible by financial institutions, as this will shorten delivery times for credit reports and ensure that the bureau's information is up to date. · Requiring that banks provide reports in real time on all loans above 50,000 taka. · Increasing the amount of information in the credit bureau, including guarantees provided by borrowers, other information on outstanding loans, and a list of other businesses of borrowers. Strengthening lending technology Technical assistance to improve lending technology should focus on the key obstacles to increasing lending to the missing middle (see chapter 3). Because not every bank will be a suitable candidate for lending to MSMEs, technical assistance should be given only to those fully committed to it. Serving such enterprises profitably requires more than just introducing new products and procedures: it also requires changing the entire corpo- rate culture and the way that banks operate. Hence technical assistance should go only to banks whose investors' mandates mesh with these changes. Banks with large branch networks and a focus on retail lending usually have a comparative advantage in entering this market segment, but smaller private banks could also be targeted to serve as catalysts. Although many countries have developed lending programs for MSMEs in state-owned banks with large networks, success usually takes a few years because the banks are often slow in implementing the required changes.This is why technical assistance should also be given to a few other banks--including smaller, faster-moving private banks. But while such banks are faster in implementing changes at the beginning, acting as a catalyst for the state-owned banks, they reach a plateau after a few years due to their smaller networks. Like Khan Bank in Mongolia, reformed BKB and RAKUB could be suitable candidates for technical assistance to state banks. The assistance program should be comprehensive and cover the entire loan cycle. It should last at least two years and involve resident advisors. 70 Increasing Access to Rural Finance in Bangladesh And it should include a redesign of bank products to meet client needs, a robust management information system, and a scheme to ensure that staff incentives are aligned with banks' efficiency targets. Finally, the technical assistance program should include performance agreements setting minimum targets, signed by the partner bank and the technical assistance provider (and its funder, if different). These agree- ments should specify the number and volume of loans disbursed and outstanding to be achieved by a given date, as well as an indicator of portfolio quality (such as keeping the portfolio at risk over 30 days to less than 3 percent). These targets should be monitored monthly, and remedial actions should be taken if they are not met. Helping Microfinance Institutions Scale Up Lending to the Missing Middle Meeting the needs of MSMEs and MSMFs requires that MFIs transform their operations and themselves--from social mobilizers to commercial providers of financial services. Changes are needed in their staff training, lending methodologies, and use of management information systems. The challenge will obviously be greater for smaller MFIs, but even larg- er ones will have to make major adjustments. Like banks, it is easier for MFIs to serve MSMEs than MSMFs. Private banks are unlikely to serve these farmers in the near future. So, whether BKB and RAKUB are left to languish or are restructured, MFIs will need to step up lending to this segment.A few MFIs, such as Banco Procredit, have managed to lend profitably to MSMFs (box 4.4). To help the microfinance industry make this shift, the government could promote a technical assis- tance fund to help selected institutions upgrade their lending methods and management information systems, and even use technology to cut the cost of serving remote areas (box 4.5). Promoting Weather Risk Management for MSMFs To help banks and MFIs lend to MSMFs, the government could pro- mote--and in some cases, sponsor--risk transfer mechanisms. To ensure that private insurance companies are not crowded out, the risk should be segmented, and insurance products that are commercially viable should be sold by the private sector. Insurance products that are not viable, either because they are too expensive or because they cover events that, though very destructive, occur very rarely, can instead be covered through direct government intervention. Increasing Access to Finance for the Missing Middle 71 Box 4.4 Banco Procredit's Expansion into Rural Areas Like many Bangladeshi microfinance institutions, El Salvador's Calpia began as an urban nongovernmental organization in 1988 to provide emergency relief after a natural disaster caused by an earthquake. It transformed into a formal regulated finance company in 1995 and into a bank in 2004, when it was renamed Banco Procredit El Salvador. A pilot agricultural product was tested successfully in 1993­95, and a refined version of the product was mainstreamed in Calpia's activities. The product's design includes the following elements: · Interview, application, and verification. Loan officers interview potential clients, asking about farming or nonagricultural economic activities, sources of income for repayment, purpose of the loan, sources of collateral, and credit history. Officers also visit potential clients' home or place of business (sometimes both) to view household relationships, potential collateral, and documentation (such as receipts for purchases and sales and titles for property). · Credit report. Loan officers evaluate the information obtained and use computers to prepare appraisal reports. A cash-flow projection is pre- pared based on the expected quantity and timing of household revenues and expenses. A balance sheet is also prepared, listing household, farm, and business assets, their values, and outstanding liabilities. Collateral is identified and valued, and the capacity to pay of any proposed co-signers is evaluated. · Credit committee. The credit committee approves, modifies, or rejects loan applications presented by loan officers. The process is rapid for loans under about $15,000, with the decision often made the day the report is submitted. Larger loans take longer because they must be reviewed in the head office. · Loan terms and conditions. Agricultural loans are made for an average of 10 months and livestock loans for 15­18 months. Some interest and partial prin- cipal payments are requested from clients projected to have the necessary cash flow. Otherwise a single payment is required at loan maturity, for both in- terest and principal. Annual nominal interest rates are 27 percent, charged on unpaid principal. A 4 percent origination fee is charged for single payment loans, and 2 percent for loans with multiple payments. (Continued) 72 Increasing Access to Rural Finance in Bangladesh · Disbursement, monitoring, and contract enforcement: To minimize fraud, loans are usually disbursed as lump sums in a branch office, and all pay- ments must be made there. Penalty interest of up to 5 percent of the loan balance may be assessed for late payments. A late payment agreement can be arranged if the delay is expected to be temporary, but collateral may be seized if late payment persists for 30 days. With almost 140,000 depositors, more than 70,000 borrowers, and profitable operations since the early 1990s, Banco Procredit is one of Latin America's largest microfinance institutions. Moreover, its rural portfolio grew from 18 percent of its total portfolio in 1996 to 32 percent in 2006. Source: Gonzalez-Vega, Rodriguez-Meza, and Pleitez-Chaves 2002; Navajas and Gonzalez-Vega 2003; Buchenau and Meyer 2007. The government could launch two initiatives to this end. First, it could promote the development of a pilot commercial index insurance product by: · Removing legal and regulatory restrictions on agents distributing insurance policies and allowing the proxy payment triggered to be considered the actual loss (see chapter 3). · Promoting the creation of a technical assistance unit that helps interested institutions develop the technical side of the product and encourages interested participants (insurance companies and MFIs) to participate in the pilot phase. Because insurance companies might be reluctant to enter this market segment individually, an organization- al arrangement could be developed between insurers and reinsurers.3 Initially a drought insurance product, similar to that developed in India by BASIX-ICICI (box 4.6), could be piloted. If successful, it could be followed by a flood index product requiring more detailed feasibility analysis. Second, given the weakness of the insurance sector, the government could create a Fund for Natural Calamities to cover adverse weather events such as floods. To avoid excessive fiscal exposure to adverse natural events, a contingent debt facility could be created to cover any contingent pay- ments required should floods occur.4 The fund would, in a sense, substitute Increasing Access to Finance for the Missing Middle 73 Box 4.5 Using Technology to Improve Financial Access in Uganda Three Ugandan microfinance institutions--Uganda Microfinance Union, FINCA UGANDA, and FOCCAS Uganda--have introduced a remote transaction system to provide more convenient services to their clients, reduce costs for their clients and themselves, and strengthen their management information systems. The remote transaction system hardware consists of a battery-powered point of sale device with an interface, smart card reader, numeric keypad, and printer. Using cell phone infrastructure, the system can convey data to a transac- tion server connected to each microfinance institution's back office. Staff mem- bers then use a Web-based server to view transaction histories and manage client accounts. The three institutions take different approaches in using the remote transaction system to reach clients. FOCCAS's loan officers take the point of sale terminal to group meetings, where they conduct loan payments, savings, and other financial transactions related to the group fund. FINCA UGANDA piloted the system in one of its sub-branches. FINCA tellers travel to collect money from group leaders at the sub-branch twice a week and use the terminal to capture group information elec- tronically. The cash is then deposited in the nearest commercial bank. Finally, the Uganda Microfinance Union has third-party agents (local merchants) using the system to provide financial services to its clients. As cash transactions are carried out between clients and agents, electronic cash transactions are executed at the institution. Client and agent accounts are then debited or credited as applicable. At the end of each day, all the transactions on the point of sale terminal are uploaded on the remote transaction system. Source: Firpo 2005; Hewlett-Packard 2005. for BKB and RAKUB's role of insurers of last resort. It would also provide needed technical, operational, and financial discipline in the financing of natural disasters. The fund should comply with basic insurance principles, including disbursement rules that allow farmers affected by natural disas- ters to get immediate payouts. The payments would be triggered by an objective and transparent index (for example, using remote sensing tech- nology), and farmers could use them to repay loans. As a risk financing mechanism, the Fund for Natural Calamities would not be effective at covering recurrent losses generated, for example, by 74 Increasing Access to Rural Finance in Bangladesh Box 4.6 BASIX and ICICI's Drought Index Insurance Product BASIX is an Indian MFI, based in Andhra Pradesh, with operations in more than 10,000 villages and about 200,000 customers. In 2003, in collaboration with the ICICI Lombard General Insurance Company, BASIX designed a rainfall-indexed insurance product. The pilot pursued a business model in which the insurance company, along with a reinsurance arrangement, takes on all the risk (see figure). BASIX simply acts an intermediary that receives commissions by selling the product to its customers. BASIX and the insurance company both incur administrative expenses from the operation. Structure of the Product offer distribute products insurance rural insurer BASIX customer inputs for feedback on products needs During the 2003 pilot the insurance contracts were designed to protect farmers from drought during the groundnut and castor growing season. ICICI Lombard underwrote the products, which BASIX extension officers marketed and sold in four villages. In total, 230 farmers bought the insurance for the Kharif monsoon season (June­September)--and most were small farmers, with less than 2.5 acres of land. ICICI Lombard reinsured the risk using a major reinsurance company. In 2004 the program was significantly modified in terms of geography, prod- uct design, and scope. To minimize basis risk (the mismatch between insurance payouts and crop losses), the product included a three-phase payout structure that reflects the weighted importance of rainfall during crop sowing, growth, and harvest. In 2005 BASIX and ICICI Lombard further improved the product by adding features (such as a varying starting date) recommended by farmers. In addition, instead of crop-specific policies, BASIX sells area-specific generic weather insurance products that suit all major rainfed crops in the same agro- climatic region. (Continued) Increasing Access to Finance for the Missing Middle 75 Between June 2003 and March 2006 BASIX sold 7,653 rainfall insurance poli- cies in six Indian states. For the first two years it did not make profits, but since the program was scaled up in 2005, BASIX has earned about 350,000 rupees from its commission fee (equal to 15 percent of collected premiums). Total administrative expenses account for 22 percent of premiums, which is compa- rable to that for other insurance products. The product developed by BASIX and ICICI has been scaled up by other insurance companies--mainly AICI, which in 2006 sold more than 250,000 such policies. Source: Manuamorn 2005. Box 4.7 Mexico's Fund for Natural Disasters In 1996 Mexico created a Fund for Natural Disasters (FONDEN) as a last resort source of financing to rebuild public infrastructure and compensate low-income producers for crop and livestock losses after natural disasters. The fund is a disas- ter relief program rather than an insurance program, but the agricultural compo- nent has characteristics similar to crop insurance programs--for example, loss claims trigger field inspections. Farmers' claims are settled at a predetermined level per hectare, and assistance is capped at 5 hectares per farmer, which is both efficient and weights the assistance to smaller holders. Irrigated or insured land is ineligible. The fund was initially financed through annual budget allocations, exposing it to fiscal uncertainties. This fund's reliance on budget contribution was reduced in 2007 with the introduction of a commercial risk transfer strategy, including the purchase of reinsurance and the issuance of a catastrophe bond. This market- based strategy ensures that FONDEN can access up to $450 million if a major earthquake hits Mexico City. nonviable farming practices. Hence mitigation projects and projects should be developed to promote viable farming practices, and the implementing agencies should work closely with the fund. The Fund for Natural Calamities would not exclude the commercial insurance industry. On the contrary, it would develop new risk transfer products and pass them to the insurance industry once they became 76 Increasing Access to Rural Finance in Bangladesh commercially viable. Thus the fund would fill the gap between farmers' needs for protection against natural disasters and the lack of agriculture insurance. Accordingly, the fund should focus on experimental business activities, while viable agricultural insurance activities would be left to commercial insurance companies. Such a public-private partnership in the financing of natural calamities has been implemented in several countries, including Mexico (box 4.7). Notes 1. Both banks have signed memorandums of understanding with Bangladesh Bank that include, among other things, capping loan portfolio growth at 5 percent a year. But neither bank is in compliance with these agreements. 2. A number of countries, including China and several in Eastern Europe and the Commonwealth of Independent States, have adopted similar schemes for both state and private banks­­generally with very successful results. 3. A successful pool of insurance companies was developed in Mongolia, where index livestock insurance has been developed. 4. Mongolia recently established a similar facility to support its index-based livestock insurance scheme. A P P E N D I X 1 Definitions Used in this Report Rural Areas Because the ultimate goal of this report is to formulate a realistic strategy for sustainably increasing rural access to financial services and because, as of 2005, 72 percent of lending (by volume) occurred in the cities of Dhaka and Chittagong, the report uses a broader definition of rural than is normally used in Bangladesh. Here rural lending refers to all interme- diation activity that happens outside those two cities. Like everywhere, in Bangladesh financial institutions operating profitably gradually enter more difficult market segments, as competition intensifies in segments where they are already active. Hence it is not realistic to expect financial institutions to jump from operating mainly in Dhaka and Chittagong to operating in rural villages. Micro, Small, and Medium-Size Enterprises There are two official definitions of small and medium-size enterprises, one used by Bangladesh Bank and the other by the Ministry of Industry. Both use economic sector as the main criterion for grouping these enter- prises: Bangladesh Bank's definition differentiates between three sectors (services, trading, manufacturing), while the Ministry of Industry uses 77 78 Appendix 1 only two (manufacturing and nonmanufacturing). Both definitions use asset and employment criteria to further classify these enterprises (tables A1.1 and A1.2). When considered from a perspective of increasing access to financial services, both of these definitions present problems. First, both were cre- ated to classify small and medium-size enterprises as one group within the national economy, divided primarily by sector rather than size. Thus the definitions are of limited value for service providers--such as banks-- seeking to understand and cater to the needs of specific segments of small and medium-size enterprises. Second, the two official definitions are very broad. For example, almost all small and medium-size enterprises employ far fewer employees than the maximum thresholds set by the definitions.1 Similarly, smaller enterprises in these sectors average much lower asset values than those stipulated by the definitions. A definition that is useful for financial institutions should describe a unique target market and its specific demands. In Bangladesh, as else- where, loan size is the most important factor considered by financial institutions when designing new products. Hence this report defines small and medium-size enterprises based on their capacity to absorb credit. Because a business's ability to absorb credit is closely related to its Table A1.1. Definition of Small and Medium-Size Enterprises Used by Bangladesh Bank, by Sector Total asset value excluding Economic sector land and buildings (taka) Maximum number of employees Service 50,000­3,000,000 30 Trading 3,000,000­5,000,000 20 Manufacturing 5,000,000­10,000,000 60 Source: Bangladesh Bank data. Table A1.2. Definition of Small and Medium-Size Enterprises Used by the Ministry of Industry, by Sector Total asset value excluding Maximum number Economic sector Category land and buildings (taka) of employees Manufacturing Small <15,000,000 n.a. Medium-size 15,000,000­100,000,000 n.a. Nonmanufacturing Small n.a. <25 Medium-size n.a. 25­100 Source: Ministry of Industry data. Note: n.a. = not applicable. Appendix 1 79 Table A1.3. Definition of MSMEs Used in This Report (taka) Segment Credit absorption capacity Annual profit Micro 50,000­300,000 50,000­170,000 Small 300,001­1,000,000 170,001­540,000 Medium-size 1,000,001­3,000,000 540,001­5,000,000 Source: Authors. Note: The credit limit for medium-size enterprises can potentially increase to 5,000,000 taka as institutions gain experience with lending. profitability, here the latter is used as a proxy for credit absorption capacity. Moreover, to allow for homogeneous market segments, and given the different nature of the bottom and top ends of this sector, this definition is further divided into micro, small, and medium-size enter- prises (table A1.3).2 The floor for the microenterprise definition is set at a loan size of 50,000 taka, as supply-side data show that microfinance institutions provide good coverage below that level. Marginal, Small, and Medium-Size Farmers Marginal, small, and medium-size farmers are defined using the defini- tion from the Bangladesh Agricultural Census: marginal farmers are those with landholdings of 0.5­1.49 acres, small farmers those with 1.5­2.49 acres, and medium-size those with 2.5­7.5 acres. In terms of loan size, although the amount to be borrowed by acre depends on the level of mechanization, type of crop, and other factors, 9,000 taka an acre (which is what PKSF partner organizations offer under the recently created farmers lending program) is a realistic estimate. Based on this estimate, average loans to this market segment would range from 4,500­67,500 taka. Notes 1. The average MSME has 4.3 employees according to Daniels (2003), or 4.4 employees according to the 2006 Rural MSME Finance Survey. 2. For example, microenterprises are highly sensitive to loan processing time and collateral requirements, and less sensitive to interest rate costs. Microenterprises also often have low levels of formality--for example, they rarely have written financial statements. A P P E N D I X 2 Comparing Rural Lending in Bangladesh and India This appendix compares rural lending patterns in Bangladesh and India. Data for Bangladesh refer to financial activities outside the cities of Dhaka and Chittagong, which include cities with up to 700,000 people, while data for India refer to rural and semiurban areas based on the Reserve Bank of India's definition--that is, towns with fewer than 100,000 inhabitants. Thus the data for India paint a conservative picture of the country's financial intermediation in rural areas. Given this, the rural-urban divide in bank lending is likely to be bigger in Bangladesh than in India. In 2005 the difference between loan-deposit ratios in rural and urban India was 26 percentage points, while in Bangladesh it was 37 percentage points (figure A2.1). In India rural lending accounted for 21 percent of bank loans, compared with 27 percent in Bangladesh (figure A2.2).And between 1996 and 2005 agricultural bank lending fell by 7 percentage points in Bangladesh, while in India it remained stable--and in 2005 increased by 1 percentage point over 2004 (figure A2.3). Banks are the largest providers of rural loans in both countries, accounting for 72 percent in Bangladesh and 61 percent in India (figure A2.4). In India cooperatives are the second largest provider of rural loans (38 percent), while in Bangladesh MFIs are (20 percent)--and coopera- tives account for just 4 percent of the total. 81 82 Appendix 2 Figure A2.1. Proportion of Loans to Deposits by Bank Branch Location, 1996­2005 (percent) Bangladesh India 100 100 90 urban 90 80 80 urban national cent 70 nt national average 70 average per 60 60 rural perce 50 50 rural 40 40 30 30 1996 2000 2004 2005 1996­97 2000­01 2004­05 2005­06 Source: Bangladesh Bank data. Source: Reserve Bank of India data. Figure A2.2. Urban and Rural Shares of Bank Loans Bangladesh India 100 100 80 80 60 60 percent 40 percent40 20 20 0 0 1996 2000 2004 2005 1996­97 2000­01 2004­05 2005­06 urban rural urban rural Source: Bangladesh Bank data. Source: Reserve Bank of India data and World Bank 2006a. Appendix 2 83 Figure A2.3. Distribution of Bank Loans, by Purpose Bangladesh India 100 100 80 80 60 60 percent 40 percent 40 20 20 0 0 1996 2000 2004 2005 2001 2002 2003 2004 2005 agriculture, forestry, others agriculture commerce and industry and fishing industry real estate services construction transport and communication others trade working capital financing Source: Bangladesh Bank data. Source: Reserve Bank of India data and World Bank 2006a. Figure A2.4. Estimated Rural Lending by Source, 2005 Bangladesh India government microfinance programs MFIs 2% 1% rural coop MFIs banks and 20% credit ocities 38% BRDB + BSBL commercial 4% banks 52% banks 74% regional rural banks 9% Source: Authors' calculations based on Bangladesh Source: Authors' calculations based on Bank and PKSF data. World Bank 2006a and Sa-Dhan 2006. A P P E N D I X 3 The Rural Micro, Small, and Medium-Size Enterprise Finance Survey In April and May 2006 the World Bank, Business and Finance Consulting (BFC), in cooperation with HB Consultants, surveyed 226 MSMEs in the Rajshahi and Moulvibazar districts on their demand for financial services. Respondents were evenly divided between the two districts. The survey used a multistage sampling technique. Survey Questionnaire The survey had three sections. The first gathered basic information, such as district and location of business, and gender of the respondent. The second section collected demographic information such as age and education of the respondent, as well as data on the enterprise's primary economic sector of activity, number of employees, and ownership struc- ture and legal status. Information was also collected on the revenues and cost structures of the enterprises surveyed. The third section compiled information on use of and access to financial services such as savings, loans, and remittances. Questions were asked on whether the respondent saves; if so, whether he or she saves in a financial institution; and, where applicable, the barriers to saving. On the use of payment services, questions were asked about the frequency of use, average amount exchanged, travel time required to visit the 85 86 Appendix 3 institution if financial institutions were used, and so on. Respondents were then asked whether they had sought a loan in the three years prior to the survey and, where applicable, reasons for not applying to formal sources. Those who had received loans (whether from formal or informal sources) were asked to provide details on the terms and conditions of their loan contracts. Finally, respondents were asked to estimate the amount of external financing they will need in 2006, which external financing sources they would seek, and their preferred terms and conditions for loan sizes, terms, and interest rates. Sampling Frame The sampling frame was based on a list compiled using data from cham- bers of commerce and local authorities. The list included 199,552 enterprises in the Rajshahi district and 148,970 in the Moulvibazar district. These districts were chosen for two reasons. First, agribusiness is impor- tant in both (agribusiness being one of the sectors with the biggest growth potential in rural Bangladesh). Second, the team conducted case studies on bank branches in the districts, allowing for cross-checking of information (such as the estimated time required to get a loan from a bank). Sample Selection The survey used a multistage sampling technique. · The first stratum coincided with district; 150 enterprises were surveyed in each of the two districts selected for the study (Rajshahi and Moulvibazar). · The second stratum was defined by three locations of enterprises: urban, semi-urban, and rural.To ensure a sufficient number of responses from each group, 25 percent, 50 percent, and 25 percent of respondents came from each respective category. The urban, semi-urban, and rural locations were chosen using a cluster analysis weighted by population within each district. · The third stratum was the economic sector. In rural areas all selected enterprises were agricultural enterprises. Elsewhere, enterprises were divided into three sectors, with the number of respondents from each sector proportional to the sector's share of production for micro, small, and medium-size enterprises as a whole--50 percent trade, 24 percent Appendix 3 87 services and 26 percent manufacturing. Within each district and economic sector surveyed, enterprises were selected randomly using lists developed by HB Consultants, based on information provided by chambers of commerce and local authorities. Map A3.1. Bangladesh, with Survey Areas Highlighted 88 Appendix 3 Focus Group Discussions In addition to the survey, BFC and the World Bank organized four focus group discussions, each with 7­10 participants. Two discussions were conducted prior to the survey, and two afterward. Discussion group par- ticipants were selected using information provided by chambers of com- merce, banks, and local authorities. Basic Sample Data The tables below report the sample composition by sector, number of employees by enterprise size, legal status and location of enterprises, and years of schooling of enterprise owners. The average enterprise had been in operation for 11.6 years. Table A3.1. Sectoral Distribution of Sampled Enterprises Sector/subsector Number Share of total (percent) Trade 114 50.4 Wholesale 36 15.9 Retail 78 34.5 Services 54 23.9 Hotels, restaurants, catering 19 8.4 Transportation 19 8.4 Other 16 7.1 Manufacturing 58 25.7 Textiles and apparel 5 2.2 Wood and furniture 18 8.0 Food and beverages 11 4.9 Light engineering 17 7.5 Leather goods 4 1.8 Other 3 1.3 Total 226 100.0 Source: Rural MSME Finance Survey 2006. Table A3.2. Average Number of Employees of Sampled Enterprises, by Enterprise Size Size Number Micro 4 Small 8 Medium-size 10 Source: Rural MSME Finance Survey 2006. Appendix 3 89 Table A3.3. Legal Status of Sampled Enterprises Status Number of enterprises Share of total (per cent) Not registered with authorities 3 1.3 Privately held limited company 1 0.4 Sole proprietorship 218 96.5 Partnership 4 1.8 Total 226 100.0 Table A3.4. Location of Business of Sampled Enterprises Location Number of enterprises Share of total (per cent) Dedicated premises 60 26.5 Home 7 3.1 Marketplace 112 49.5 In the street 47 20.8 Total 226 100.0 Table A3.5. Average Years of Schooling for Owners of Sampled Enterprises Number of years Number of enterprises Share of total (per cent) None 4 1.8 1­5 32 14.2 6­10 100 44.2 11­12 44 19.5 13 or more 46 20.3 Total 226 100.0 A P P E N D I X 4 Recapitalization Scenarios and Costs of Closing BKB and RAKUB Costs of Recapitalizing BKB Efforts to recapitalize BKB should aim to restore the bank to complete financial health. Tables A4.1 and A4.2 provide two scenarios for the recapitalization of BKB. The first is a realistic scenario where, based on a diagnostic audit,1 50 percent of loans graded standard in June 2005 are found to be classifiable. The second is an optimistic scenario where only 25 percent of standard loans are found to be classifiable after a diagnostic audit. In both scenarios the entire portfolio of the bank's substandard or worse loans (after classification by a diagnostic audit) are carved out of the bank, leaving it with a fully performing (payments of principal and interest made on time) loan portfolio. Other items are also carved out: · BKB's claims on the government that have not been met (1,716 million taka) are booked as investments. · A wide variety of assets representing claims for remission of interest, principal waivers, capitalized losses, and miscellaneous items that appear to be of no value (15,590 million taka) are booked as other assets. · BKB borrowing from Bangladesh Bank (33,389 million taka). · BKB borrowing directly from the government (148 million taka). 91 92 Appendix 4 The difference between the book value of the assets removed and the value of liabilities transferred results in a reduction of the bank's capital of 29,399 million taka in the realistic scenario and 19,178 taka in the optimistic scenario (with the difference between the two outcomes being caused by the reduced quantity of standard loans removed in the optimistic scenario). After carving out bad assets and offsetting some of the cost to the bank of doing so by removing liabilities, BKB's capital would be extremely negative. To restore it to financial health, the bank would then be recapitalized to about a 12 percent capital adequacy ratio with an injection of 47,983 million taka in government bonds in the realistic scenario and 38,988 million taka in the optimistic scenario. These calcu- lations use a 12 percent capital adequacy ratio rather than the regulatory minimum of 9 percent to allow for the absorption of continued operating losses while BKB is restructured and to provide some scope for renewed lending, which would increase BKB's risk assets. The exact amount of capital needed would be determined only after diagnostic work has been completed and a business plan prepared, but would not be less than the amounts shown in tables A4.1 and A4.2. Table A4.1. BKB Recapitalization: Realistic Scenario with 50 Percent of Standard Loans Carved Out (based on 30 June 2005 balance sheet; millions of taka) Recapitalize Carve out to 12% capital Item Audited bad assets Adjusted adequacy ratio Recapitalized Total liquid assets 6,802 0 6,802 0 6,802 Cash on hand 4,024 4,024 4,024 Deposits with other banks 2,778 2,778 2,778 Investments (1) 1,719 1,716 3 47,983 47,986 Net loans and advances 65,673 45,231 20,443 0 20,443 Standard 40,885 20,443 20,443 20,443 Substandard 4,763 4,763 0 0 Doubtful 4,738 4,738 0 0 Loss 15,287 15,287 0 0 Fixed assets 935 935 935 Other assets (2) 20,154 15,990 4,164 4,164 Total assets (3) 95,283 62,936 32,347 47,983 80,330 Appendix 4 93 Table A4.1. BKB Recapitalization: Realistic Scenario with 50 Percent of Standard Loans Carved Out (continued) (based on 30 June 2005 balance sheet; millions of taka) Recapitalize Carve out to 12% capital Item Audited bad assets Adjusted adequacy ratio Recapitalized Total borrowings 34,243 33,537 706 0 706 Borrowing from Bangladesh Bank 33,389 33,389 0 0 Borrowing from government 148 148 0 0 Borrowing from other banks 706 706 706 Deposits and bills payable 56,345 56,345 56,345 Other liabilities 20,216 20,216 20,216 Total liabilities 110,804 33,537 77,267 0 77,267 Paid-up capital 2,400 2,400 2,400 Statutory reserves 0 0 0 Other reserves 1,305 1,305 1,305 Retained earnings 19,224 29,399 48,623 47,983 640 Total capital 15,519 29,399 44,918 47,983 3,065 Total liabilities and capital (3) 95,285 62,936 32,349 47,983 80,332 Notes: (1) Investments Government bond 1,715,879,478 (2) Other assets suspense adjustment for remission of interest Receivable from government 5,785,705,527 Interest on flood affected loans 20,426,576 Interest on frozen tea loan 5,322,078 Pre-liberation loans interest 39,962,414 Weavers credit 60,621,926 Transferable remission of interest 320,731,946 Loss on remission of interest 861,650,555 Double principle interest 3,325,079,640 Receivable from RAKUB 1,564,990,206 Remission against 1st claim loan 50,684,233 Swaninver credits 5,246,766 Interest on loans <5,000 taka 647,298,437 Remissionable interest account 1,734,383,356 Waiver credits 79,241,162 Loan board credit 1,416,604 20% remission of classified >4 years 70,110,628 Fructional power pump loans 15,210,732 Legal expenses of written off loans 13,244,765 50% of loans <5,000 taka 7,999,943 Accrued interest receivable 1,380,255,074 Total 15,989,582,568 (3) Imbalance of 2 million taka between assets and liabilities and capital due to rounding 94 Appendix 4 Table A4.2. BKB Recapitalization: Optimistic Scenario with 25 Percent of Standard Loans Carved Out (based on 30 June 2005 balance sheet; millions of taka) Recapitalize Carve out to 12% capital Item Audited bad assets Adjusted adequacy ratio Recapitalized Total liquid assets 6,802 0 6,802 0 6,802 Cash on hand 4,024 4,024 4,024 Deposits with other banks 2,778 2,778 2,778 Investments (1) 1,719 1,716 3 38,988 38,991 Net loans and advance 65,673 35,009 30,664 0 30,664 Standard 40,885 10,221 30,664 30,664 Substandard 4,763 4,763 0 0 Doubtful 4,738 4,738 0 0 Loss 15,287 15,287 0 0 Fixed assets 935 935 935 Other assets (2) 20,154 15,990 4,164 4,164 Total assets (3) 95,283 52,715 42,568 38,988 81,557 Total borrowings 34,243 33,537 706 0 706 Borrowing from Bangladesh Bank 33,389 33,389 0 0 Borrowing from government 148 148 0 0 Borrowing from other banks 706 706 706 Deposits and bills payable 56,345 56,345 56,345 Other liabilities 20,216 20,216 20,216 Total liabilities 110,804 33,537 77,267 0 77,267 Paid-up capital 2,400 2,400 2,400 Statutory reserves 0 0 Other reserves 1,305 1,305 1,305 Retained earnings 19,224 19,178 38,402 38,988 587 Total capital 15,519 19,178 34,697 38,988 4,292 Total liabilities and capital (3) 95,285 52,715 42,570 38,988 81,559 Notes: See table A4.1. for notes. Costs of Rehabilitating BKB The cost to the government of rehabilitating BKB would have four components: the costs of repaying BKB's borrowing from Bangladesh Bank, of writing off BKB's borrowing from the government, of recapitalizing BKB, Appendix 4 95 and of technical assistance to support the rehabilitation process ($5 million over three years) and new information technology systems ($15 million). If responsibility for collecting the loans carved out of BKB is transferred to a competent and appropriately motivated work-out company, there is some prospect of recovering amounts due from borrowers. In table A4.2 these are shown as reducing the net cost to the government of rehabilitating BKB, but it should be recognized that the costs of rehabilitation are certain and immediate, whereas the costs of recoveries are uncertain at best. Costs of Recapitalizing RAKUB Efforts to recapitalize RAKUB should aim to restore the bank to com- plete financial health. Tables A4.3 and A4.4 provide two scenarios for Table A4.3. RAKUB Recapitalization: Realistic Scenario with 50 Percent of Standard Loans Carved Out (based on 30 June 2005 balance sheet; millions of taka) Recapitalize Carve out to 12% capital Item Audited bad assets Adjusted adequacy ratio Recapitalized Total liquid assets 2,891 0 2,891 0 2,891 Cash on hand 946 946 946 Deposits with other banks 1,945 1,945 1,945 Investments 3,450 3,378 72 4,341 4,413 Net loans and advances 21,428 13,918 7,510 0 7,510 Standard 15,020 7,510 7,510 7,510 Substandard 869 869 0 0 Doubtful 1,112 1,112 0 0 Loss 4,427 4,427 0 0 Fixed assets 65 65 65 Other assets (1) 2,552 1,894 658 658 Total assets 30,386 19,190 11,196 4,341 15,537 Total borrowings 15,949 15,935 14 0 14 Borrowing from Bangladesh Bank 15,607 15,607 0 0 Borrowing from BKB 328 328 0 0 Borrowing from other banks 14 14 14 Deposits and bills payable 11,578 11,578 11,578 (continued) 96 Appendix 4 Table A4.3. RAKUB Recapitalization: Realistic Scenario with 50 Percent of Standard Loans Carved Out (continued) (based on 30 June 2005 balance sheet; millions of taka) Recapitalize Carve out to 12% capital Item Audited bad assets Adjusted adequacy ratio Recapitalized Other liabilities 2,956 2,956 2,956 Total liabilities 30,483 15,935 14,548 0 14,548 Paid-up capital 1,500 1,500 1,500 Statutory reserves 0 0 0 Other reserves 209 209 209 Retained earnings 1,807 3,255 5,062 4,341 721 Total capital 98 3,255 3,353 4,341 988 Total liabilities and capital 30,385 19,190 11,195 4,341 15,536 Notes: (1) Other assets suspense adjustment for remission of interest Crop loans 14,759 Loans