Download - Micro Finance - Emerging Trends and Challenges

Microfinance

MicrofinanceEmerging Trends and Challenges

Edited by

Suresh Sundaresan

Professor of Economics and Finance, Columbia BusinessSchool, USA

Edward ElgarCheltenham, UK • Northampton, MA, USA

© Suresh Sundaresan 2008

All rights reserved. No part of this publication may be reproduced, stored ina retrieval system or transmitted in any form or by any means, electronic,mechanical or photocopying, recording, or otherwise without the priorpermission of the publisher.

Published byEdward Elgar Publishing LimitedThe Lypiatts15 Lansdown RoadCheltenhamGlos GL50 2JAUK

Edward Elgar Publishing, Inc.William Pratt House9 Dewey Court NorthamptonMassachusetts 01060USA

A catalogue record for this bookis available from the British Library

Library of Congress Control Number: 2008935924

ISBN 978 1 84720 920 7

Printed and bound in Great Britain by MPG Books Ltd, Bodmin, Cornwall

Contents

List of contributors viPreface xi

1 The changing landscape of microfinance 1Suresh Sundaresan

2 The role of international capital markets in microfinance 25Brad Swanson

3 Securitization and micro-credit backed securities (MCBS) 46Ray Rahman and Saif Shah Mohammed

4 Cell phones for delivering micro-loans 71Anand Shrivastav

5 How should governments regulate microfinance? 85Richard Rosenberg

6 Gender empowerment in microfinance 108Beatriz Armendáriz and Nigel Roome

Index 123

v

Contributors

Beatriz Armendáriz is a Lecturer in Economics at Harvard University, anda Senior Lecturer at University College London. She is also a researchaffiliate at the David Rockefeller Center for Latin American Studies atHarvard University. She has taught at the London School of Economicsand has worked as a Visiting Associate Professor at MIT, and as a VisitingProfessor at the Toulouse School of Economics. Her research focuses oneconomic development and finance. Having published numerous articles onmicrofinance, notably with Christian Gollier and Jonathan Morduch, sherecently co-authored The Economics of Microfinance, a book published byMIT Press in 2005.

Her current research includes fieldwork on microfinance and genderempowerment, with researchers from the Innovations for Poverty Action(Yale and Harvard), and the Financial Access Initiative (Harvard, Yale,NYU). She is also co-editing a handbook on microfinance, and a book, TheContemporary Latin American Economy, also for MIT Press. LecturerArmendáriz grew up in southern Mexico where she founded AlSol andGrameen Trust, Chiapas, the first Grameen-style microfinance organiza-tions in the region.

Saif Shah Mohammed is co-founder of MR Analytics. In 2005–06, heworked extensively on the BRAC micro-credit securitization, having movedto Bangladesh to complete and implement the transaction. Prior to joiningMF Analytics, he worked as an analyst at Cornerstone Research, assistingindustry and faculty experts in developing economic and financial analysesin litigation contexts. Saif graduated from Harvard College in 2002 with aB.A. magna cum laude in economics. He currently attends the ColumbiaUniversity School of Law.

Ray Rahman is the founder and CEO of MF Analytics and one of the chiefarchitects of the BRAC micro-credit securitization. Prior to starting MRAnalytics, Ray was at Lehman Brothers, working on the cutting edge ofasset securitization in the fledgling commericial mortgage-backed secu-rity (CMBS) industry in the 1990s. He has also working in privateequity, restructuring multinational companies with sales of over $1 billion.Rahman is a cofounder of Potenco, a green energy company that is creatingthe world’s most efficient hand-held power generator, targeted at the needs

vi

of developing countries. The generator was featured in Wired magazine inthe summer of 2007. Rahman is also founder of Fast Forward India, a non-profit organization that links graduate students from MIT, Stanford andIIT with local mentors and NGOs, to identify and solve key problems thatdirectly affect the life of disadvantaged populations in India. FFI helpsteams pilot test proposed solutions on the field with local NGOs, obtainfunding to grow and widen the scale and scope of impact. Rahman gradu-ated with a BS degree in mechanical engineering (high distinction) and aBA in economics, (cum laude) from the University of Rochester. He has anMBA from MIT Sloan School of Management.

Nigel Roome holds the Daniel Janssen Chair in Corporate SocialResponsibility at Solvay Business School, Free University of Brussels, andthe University Chair in Corporate Global Responsibility and Governanceat TiasNimbas Business School, Tilburg University, Netherlands. He previ-ously held chairs in The Netherlands and Canada, and academic positionsin Britain. He is widely published on topics that relate business strategies,innovation, and technology to issues of corporate responsibility, sustainabledevelopment, and global change. His books include Management Educationfor Sustainability and Corporate Environmental Management (1994);Sustainability Strategies for Industry (1998); and The Ecology of Informationand Communications Technologies (2002).

Professor Roome’s career has involved innovations in both educationand research. His achievements as a European Faculty Pioneer over 20years were acknowledged by The Aspen Institute in 2006. Roome is cur-rently Chair of the Academic Board of the European Academy of Businessin Society and is a member of the academy’s management board. In addit -ion to these responsibilities Roome has served as an environment coun-cilor, an advisor to the president of Ontario Hydro and the Office of theCanadian Commissioner for Environment and Sustainable Development,a member of ABB’s global stakeholder advisory group, an expert to theEuropean Commission’s “Futures” project, and Chair of the EuropeanCommission’s expert group on competitive and sustainable productionsystems to the year 2020. He has been invited to contribute his views oncorporate responsibility and sustainability to meetings of the EuropeanUnion, under successive presidencies of Sweden, The Netherlands,Finland, and Germany.

Richard Rosenberg holds a law degree from Harvard University. He hasworked in international development since 1984 with USAID and theWorld Bank, concentrating on development finance, especially micro -finance. He has authored or co-authored two dozen publications onmicrofinance, including several on issues associated with regulation and

Contributors vii

supervision of microfinance, and teaches regularly on that topic at theBoulder Institute of Microfinance Training.

Anand Shrivastav is the chief software architect and founder of Suvidha. Hehas over 25 years of experience in marketing and distribution of consumer,nutrition, and health care products besides telecom and banking services.He has set up projects, implemented marketing strategies, and managedlarge distribution channels. Shrivastav currently serves on the boards of thefollowing organizations: Suvidha Starnet (mobile transaction services), asChairman; and Genesis Biogen (stem cell technologies), as Chairman.

In the past Shrivastav has served Intercorp Biotech (biotechnology,nutrition) as managing director, Hiperworld (information technology/software) as managing director, The Home Store India (retail chain)as director, Intercorp Industries (hybrid seeds) as director, IEPCL(telecom, chemicals) as executive director, as marketing consultant forpublic sector units Kerala State Drugs & Pharmaceuticals and KeralaSoaps & Oils (consumer products), and as resident manager with RGSoft Drinks. Shrivastav attended Harvard University Graduate School ofEngineering Design.

Suresh Sundaresan is the Chase Manhattan Bank Professor of Economicsand Finance at Columbia University. He is currently the Chair of theFinance subdivision. He has published in the areas of Treasury auctions,bidding, default risk, habit formation, term structure of interest rates, assetpricing, pension asset allocation, swaps, options, forwards, futures, fixed-income securities markets, and risk management. His research papers haveappeared in major journals such as the Journal of Finance, Review ofFinancial Studies, Journal of Business, Journal of Financial and QuantitativeAnalysis, European Economic Review, Journal of Banking and Finance, andJournal of Political Economy, among others. He has also contributed arti-cles in the Financial Times and to World Bank conferences. He is an associate editor of Journal of Finance and Review of Derivatives Research.His current research focus is on default risk and how it affects asset pricingand sovereign debt securities.

Sundaresan has consulted for Morgan Stanley Asset Management andErnst and Young. His consulting work focuses on term structure models,swap pricing models, credit risk models, valuation, and risk management.He is the author of the book Fixed-income Markets and their Derivatives(1996). He has served on the Treasury Bond Markets Advisory Committee.He has served as a resident scholar at the Federal Reserve Bank of NewYork. He has worked with the Center for Microfinance Research (CMFR)on issues relating to interest rates on micro-loans, the effects of repeatedborrowings, and contract design.

viii Contributors

Brad Swanson is a partner in Developing World Markets, a socially respon-sible investment bank and fund manager, which he joined in 2003. As abanker and a diplomat, he has worked in emerging markets for 25 years andhas done business in more than 50 countries. He began his career in the USState Department and was posted to West Africa. He then worked forDonaldson, Lufkin & Jenrette Securities Corp. in New York, Bankers Trustand Banque Nationale de Paris in London, and Global Environment Fundin Washington, DC. During spring and summer of 2004, he was on assign-ment with the US Department of Defense in Iraq running private-sectorfinance programs for the occupation government. He is a member of theInvestment Committee of Partners for the Common Good, a non-profitcommunity development finance organization, and has served on theboards of investment banks in Poland and the Ivory Coast. He has a BAfrom Princeton University and an MBA from Columbia University.

Contributors ix

xi

Preface

This book—Microfinance: Emerging Trends and Challenges—bringstogether recent practical innovations and policy questions in the field ofmicrofinance, and is largely based on the contributions made by differentscholars and practitioners to the conference hosted by the Social EnterpriseProgram of Columbia Business School on 20 April 2007 at ColumbiaUniversity. This conference, entitled Credit Markets for the Poor: Focus onMicro-Finance, examined recent developments in the field, researchfindings, and the challenges that lie ahead. The contributions in this book,after the introductory chapter, focus on (1) integration of capital marketswith microfinance, (2) securitization and micro-credit backed securities, (3)technological innovations such as the delivery of banking services to the“unbanked sector” using mobile phone technology, (4) the regulatory chal-lenges and opportunities as the landscape of microfinance undergoes a seachange, and (5) the consequences of gender empowerment, where women,especially among the poorest, predominate micro-loan borrowings.

1. The changing landscape ofmicrofinanceSuresh Sundaresan

INTRODUCTION

The ability of households to save, access capital, and manage risk exposuresof various kinds, such as life, property, and health through insurance is aprerequisite for their economic and social development. Access to basicfinancial services (such as credit, savings, and insurance) is most likely todevelop the entrepreneurial skills and opportunities among those poor whoare currently outside the perimeter of such financial markets and services.Furthermore, over time, such access will promote better risk managementcapabilities and promote the economic aspirations of the poor.

The World Bank uses two reference benchmark levels of consumption/income to measure poverty: a consumption level of (US) $1.08 per day anda consumption level of $2.15 per day.1 These levels are measured in 1993purchasing parity terms. As of 2001, the World Bank estimated 1.1 billionpeople had consumption levels below $1 a day, and 2.7 billion lived on lessthan $2 a day. While these figures are very stark, it is also true that the pro-portion of people living under $1 a day has fallen from 28 percent in 1990to 21 percent in 2001. This progress not withstanding, it remains clear thatpoverty alleviation should be a major priority, especially for countries wherea great proportion of people live under $1–2 a day.

It is difficult to visualize how countries such as Brazil, China, and Indiacould truly emerge as developing economies until their numerous poor citi -zens find easy access to essential financial services, which is critical to themin climbing out of poverty. Microfinance, which has emerged and evolvedover the past 35 years, is one such mechanism that has attempted to deliverthe core financial services to the poor. This mechanism along with somerecent developments in this field is the focus of this book.

It should be made clear that while the provision of low-cost access tofinancial services is an important ingredient in alleviating poverty, thereare a number of other basic services that are unavailable to the poor. Theseconstraints include (1) absence of primary education, (2) absence of

1

primary health care, and (3) relatively primitive technologies used by thehouseholds. These constraints need to be tackled in parallel as the momen-tum to deliver financial services gathers speed. There are significant com-plementarities between access to financial services and the ability of thepoor to access education, health care, and better technologies.

FORMAL AND INFORMAL MARKETS

Capital markets and financial institutions help provide mechanisms forthrifty households to save and provide access to households and institu-tions seeking capital for funding their consumption and investment plans.In addition, they offer financial services such as life insurance, propertyinsurance, and health insurance. The successful development of economiesin Europe and the United States is in no small measure due to the servicesprovided by such institutions and markets, which allocate savings from onepart of the economy to finance the capital requirements in other parts.

Organized stock exchanges, bond markets, dealer markets, commercialbanks, insurance companies and other institutions, constitute a vital partof financial architecture of most economies, whereby households andinstitutions are able to allocate their savings and have access to capital.These markets and institutions are known as formal markets for financialservices. Formal credit markets consist of institutions such as commercialbanks, credit unions, rural banks, and other financial institutions, whichare subject to public and private oversight and regulation. They are gov-erned by bankruptcy codes, investor protection laws, and disclosurerequirements. In many developing economies, however, a significant percentage of the population is simply unable to access these organ -ized financial institutions and markets due to the fact that they areextremely poor, ill-educated, and have no sustained opportunities forgainful employment. From the perspective of the lenders in the formalcredit markets, delivery of small loans or provision of insurance to coververy small individual exposures, or accepting tiny savings deposits issimply not an economically viable proposition. This is due in no smallmeasure to the transactions costs associated with extending such tinyscale of services to the poor, who are numerous.

Yet another important reason for lack of access of markets to the pooris the fact that lenders and service providers in the formal markets have verylittle information about the potential consumers of their services. As aresult, such poor households are unable to tap meaningfully into the formal(organized) markets for financial services or capital markets for financingtheir consumption and investment needs.

2 Microfinance

The properties of financial markets where there are significant differ -ences between the information possessed by the lender and the informationpossessed by the borrower have been well documented in a number ofinfluential papers.2 In particular, the following insights emerge from thisstrand of research: first, with asymmetric information, there will be creditrationing, leading to households being shut out of credit markets. Second,contracts will have to be designed so that the lender (who has limited infor-mation about the borrower’s ability or willingness to pay), can assure himor herself a fair rate of return on loans. In fact, microfinance is an areawhere the concept of “group lending” or “self-help groups” (SHGs) iswidely practiced: loans are extended to a group of borrowers who arejointly liable for each of the loans extended to the members of the group.Even the formation of the group itself is a matter of considerable import -ance: since the borrowers have better information about each other, it is recognized that the borrowers should be allowed to form the groups them-selves. The presence of joint liability will then imply that “high-risk” bor-rowers will not be accepted in the group, and the pool of “low-risk” groupswill then also be conservative in the choice of projects.

The loan contract itself may be structured to reflect ground realities: loancontracts may tend to have short maturity in order to set strong incentivesfor the group to service the debt payments, where the income stream issteady. Payments of initial rounds of loans may qualify borrowers to obtaina loan of higher size in the subsequent rounds: this provides an additionalincentive to borrowers to acquire good credit history. Interest payments maybe collected on a weekly basis, thereby assuring the presence of loan officersat site levels at that time. In seasonal projects, loan payments may be indexedto reflect the revenue patterns. Loans may be bundled with insurance pro-grams to mitigate aggregate risk exposure such as monsoon or droughts.

In the past and to a significant extent even now, poor households havetended to rely extensively on informal markets for their capital needs andother financial services such as insurance or savings. Indeed, these otherfinancial services have only recently become available to poor householdsthrough microfinance. Informal credit markets typically operate outside theperimeter of regulators and are often not subject to monitoring and super-vision by governments or agencies of governments. The rights and respon-sibilities of lenders and borrowers in such markets do not come under aformal bankruptcy code.

Players living in close proximity to such markets often have detailedknowledge of each other. Examples of such informal markets and participants might include the following: (1) family members, (2) friends,(3) trade credit from local shops, (4) professional moneylenders in theregion, (5) pawnbrokers, (6) local landowners, and so on. These markets

The changing landscape of microfinance 3

have existed for a long time, and indeed have preceded the development offormal banking systems and capital markets. Because they provide a veryuseful function in poorer sectors of world economies in keeping the savingsflow from lenders to borrowers who are unable to access formal creditmarkets such markets will continue to exist. In this context, it is worthpointing out that even in a well-developed economy such as the UnitedStates, pawnbrokers and payday lending institutions exist and serve a clien-tele of borrowers.3

Without necessarily detracting from the useful functions provided bysuch informal credit markets, it is also important to examine more carefullythe interest rates that prevail in them. Roughly, interest rates in such infor-mal markets provide an upper bound on the interest rates that the borrow-ers would be willing to pay in microfinance markets. While detailed andreliable micro-level data on such markets are usually not available, severalstudies have documented that the effective borrowing costs in such infor-mal credit markets are rather high. The effective annualized interest rates inpayday lending runs well into three digits, often in excess of 200 percent onan annualized basis! For example, interest rates estimated in pawnbrokingand local moneylenders run into triple digits, on an annualized basis.In the United States, interest rates charged by pawnbrokers ranged from36 percent (in New Jersey and Pennsylvania) to 240 percent (in Oklahoma)during 1987–88.4 In addition, the supply of capital from informal creditmarkets tends to be rather limited. These observations suggest that a preva-lence of high interest rates in informal credit markets where the privatesector individuals possessing asymmetric information play the role oflender of last resort to liquidity-constrained households.

EVOLUTION OF MICROFINANCE

Informal credit markets described in the previous section have precededmicrofinance. Institutions such as credit unions and specialized lending pro-grams targeted to agricultural sectors have also been in existence since theearly 1900s.5 The seeds for microfinance in its current form were plantedduring the period 1950–80, when small loans were extended to poor bor-rowers who could not post meaningful collateral. Major organizations, whichpioneered this initiative, were ACCION International in Latin America,SEWA Bank in India and the Grameen Bank founded by MuhammadYunus in Bangladesh. These initiatives demonstrated for the first time thatpoor borrowers, especially women, were not only willing to take on small-scale projects funded by loans, but were also capable of chalking up excel-lent payment records.

4 Microfinance

Over the period 1980–90s many microfinance institutions began todevelop and found sustaining models of lending to the poor: non- governmental organizations (NGOs), non-bank financial institutions(NBFI), rural banks of nationalized banks, and village banks began todevelop.6 In the early stages of evolution, microfinance was largelyrestricted to loans and was funded by either governments or aid agenciesand was thus based on “soft capital.”

Since the 1990s, microfinance has branched out both in terms of therange of financial and economic services extended, as well as in terms ofhow capital is raised. Banks began to access this market in a moresignificant way than ever before. Financial services ranging from savingsdeposits, loans, insurance to cover life, health, crop, and properties are cur-rently offered. Many microfinance institutions access capital markets eitherby issuing equity or debt capital in order to raise capital. There are otherswho have been able to securitize their loans and thus attract capital byissuing micro-credit backed securities.

Technological innovations have also paced the evolution of microfinance:widespread availability of mobile phones, access to community-level kiosksof computer terminals with access to the Internet, biometric technology toobtain loan approval and credit history, and correspondent banking havedramatically changed the landscape of microfinance.

The year 2005 was declared as the “Year of Microfinance,” and a numberof private sector enterprises and foundations have now dedicated pools ofcapital for exclusive investments in the area of microfinance: in November2005, Pierre Omidyar (founder of eBay) announced a $100 millionmicrofinance fund in partnership with Tufts University for exclusive invest-ments in the microfinance initiatives. TIAA-CREF created in September2006 a $100 million Global Microfinance Investment Program (GMIP) toinvest in selected microfinance institutions worldwide. TIAA-CREF madea $43 million private equity investment in ProCredit Holding AG, amicrofinance company.7 These are high profile, large-scale investment ini-tiatives. Given the current size and status of microfinance institutions it willbe interesting to see how these initiatives pan out.

As Swanson and Rahman and Mohammed point out respectively inChapters 2 and 3, concepts such as securitization are rapidly integratingmicrofinance with the capital markets of the developed world, therebyaltering the landscape of microfinance. As we will demonstrate later in thischapter, traditional institutions such as NGOs still remain an importantforce, especially for the poorest of micro-borrowers. In the upper tier of themicrofinance, capital markets are actively helping microfinance institutionsto tap debt and equity capital. New market institutions have developed,which are likely to improve the transparency and potentially reduce the cost


of accessing financial services: for example, a recent study has reported thatthe microfinance institutions (MFIs) in countries with credit bureaus tendto have a 5 percent lower operating expense ratio than the ones in countrieswithout credit bureaus.8 Credit rating agencies have developed in countriesthat rate institutions in microfinance. Two rating agencies, MicroRate andM-CRIL are well established in the field.

QUESTIONS ADDRESSED IN THIS BOOK AND AWORD ABOUT DATA

What are the costs of accessing credit in microfinance? Does it depend onthe nature of contracting? Does the nature of organization of themicrofinance institution affect the interest rates, and the characteristics ofthe borrowers? How have the capital markets and financial intermediariesinfluenced the character of microfinance? What are the risks and returns toinstitutions in microfinance? What are the implications of technologicalinnovations such as biometrics, the Internet, mobile phones, and so on, onmicrofinance practice? Is there a strong gender bias? What are the positiveand adverse consequences of such a bias?

These questions form the intellectual basis for this book. To motivatethese issues, I would like to examine some of the studies that have used MIXdata, in order to shed some light on the issues studied here. Before delvinginto these questions, it is useful to note the following properties of MIXdata. MIX data is based on voluntary reporting by 704 institutions to adetailed survey conducted by MIX. As of 2006, the survey covered morethan 52 million borrowers with $23 billion in loans. The survey also covered56 million depositors with over $32 million in deposits. Clearly, thesefigures are a downward-biased estimate of the true size of the microfinancemarket. It is also conceivable that the participating institutions are qualita-tively different from the non-participating institutions, which are likely tobe smaller and more donor-financed. With these caveats, let us examine thequestions in turn in the following sub-sections.

Cost of Access to Credit

Lacking micro-level data, it is difficult to estimate the costs of obtainingfinancial services to the borrower in the microfinance area with great pre -cision. Still, we have some anecdotal evidence about the levels of interestrates that micro-loan borrowers encounter. In a recent paper, CGAP(Consultative Group to Assist the Poor) reports that the interest yield forCompartamos (a lender in this market) stood at 86.3 percent!9 CGAP

6 Microfinance

estimates that the cost to the micro-loan borrower is about 100 percentwhen taxes are taken into consideration. This data corresponded to the2005 period when the median interest rates charged by village bankingMFIs stood at 47.2 percent, and the interest rates charged by low-end MFIsstood at 35.4 percent.

Gonzalez (2007) provides an analysis of the MIX database, whichenables us to shed some light on this issue at a more aggregate level,10 andusing this data, we can estimate the interest costs faced by the borrower.Our first estimate is the yield on gross loan portfolio. This measure takesthe ratio of adjusted financial revenue from the loan portfolio to theadjusted average gross loan portfolio. Financial revenue includes therevenue from the loan portfolio and other assets plus revenue from otherfinancial services. Such services may include insurance, passbooks, smartcards, and so on. For our purposes these expenses are relevant to the mea-surement of the costs of financial services. Gross loan portfolio excludeswrite-offs. In Table 1.1, I provide an estimate of the costs for differentlending institutions for the period 2003–05.

Note that the costs range from a low of about 20 percent to a high of over42 percent. The estimated costs are the highest for NGOs, which service thepoorest of the borrowers, and lowest for credit unions. Transactions costsassociated with delivery of loans, monitoring, collection, and the risk ofdefault are the primary reasons for such high interest rates. Gonzalez (2007)in his analysis also provides another way to look at this measure: acrossdifferent target group of borrowers,11 as shown in Table 1.2.

Table 1.2 confirms our finding that the low-end borrowers with a loanbalance of less than $150 are the ones who face the highest cost of obtain-ing financial services. Their costs range from 35 percent to 38 percent. Thisraises the question of whether at these rates it is reasonable to think thatmicrofinance can play an effective role in alleviating poverty? There areseveral reasons to harbor such a hope. First, it is safe to assume that the


Table 1.1 Costs of obtaining financial services—variations acrossinstitutions (in %)

Institution 2003 2004 2005

Banks 33.7 36.9 28.3Credit unions 22.0 20.2 20.4NBFI 32.6 32.1 29.3NGOs 42.5 40.2 38.6Rural banks 27.3 26.8 23.2

Source: Gonzalez (2007).

alternative sources of credit and other financial services are even moreexpensive.12 Since the participation in microfinance is purely voluntary, itis reasonable to conclude that for the participants this avenue must becost-effective.

Second, the costs appear to decline for the broad and higher-end borrowers. Presumably these are seasoned borrowers, who by repeated borrowing have established a good credit reputation and honed their entrepreneurial skills. This has in turn dramatically declined the costs ofaccess.

Finally, there are reasons to think that the average costs of participatingin this market is expected to go down significantly due to the advent of tech-nology such as biometric screening, smart cards, and the delivery of loansby mobile phones. This said, it is clear that the borrowing costs must comedown significantly in order for microfinance to be a credible and sustainingavenue for poverty alleviation.

Contracting and cost of accessAs noted earlier, lenders utilize different contracting methods to extendloans. Some borrowers get “individual loans” and others are part of a “sol-idarity group” where the group is jointly liable for the loans taken by themembers. Table 1.3 reports the cost of accessing loans during the period2003–05 across different contracting methods.

Table 1.3 shows that the interest rates are highest for village bankinggroups which consist of poorest borrowers formed into solidarity groups.We note that “individual borrowers” face borrowing costs ranging from 28percent to 30 percent, among the lowest.

Capital structure and default riskThe cost of extending financial services such as credit and insurance reflectsthe costs of accessing capital for different lending institutions and the risks

8 Microfinance

Table 1.2 Costs of obtaining financial services—variations across targetgroups (in %)

Target Group 2003 2004 2005

Low end 37.8 37.8 35.4Broad 34.5 33.7 31.1High end 24.1 24.1 22.7Small business 24.8 22.7 22.6


of the consumers of financial services. We now turn to these issues. In theanalysis of trend lines, Gonzalez (2007—see note 10) reports that duringthe period 2003–05, the debt/equity ratios of banks varied from 3.6 to 5.6.The corresponding range for credit unions was 3.4 to 4.4. For the ruralbanks, the debt/equity ratios ranged from 4.6 to 5.2. In contrast, non-bankfinancial institutions and NGOs had much smaller debt/equity ratios. Thelow debt/equity ratios of NGOs (ranging from 0.9 to 1.6) may reflect thepervasive use of donor capital by NGOs. High debt/equity ratios of banks,credit unions, and rural banks may reflect their inability to get sufficientequity capital and the relative ease of access to debt capital. In the next sub-section, we provide some evidence concerning the ability of microfinanceinstitutions to tap into debt and equity capital.

Role of Capital Markets in Financing and Investment in Microfinance

Chapters 2 and 3 explore the role of capital markets in the field ofmicrofinance. Microfinance began largely as a philanthropic effort or aquasi-philanthropic effort. Government-mandated programs such as ruralbanks, and branch expansion by nationalized banks into rural areas, areexamples of such effort. Non-governmental organizations (NGOs), whosesupply of capital is largely donor-based, dominated the scene. In fact, ananalysis of the MIX database shows that there are over 400 NGOs and over100 rural banks that extend financial services as of 2006. Indeed, this evi-dence demonstrates the very key roles that these groups continue to play inmicrofinance. But we have seen a very active interplay between capitalmarkets and microfinance: increasingly, capital markets are being used tosource capital for providing microfinance services. Indeed, we tend to see atiered market: some of the top-tier microfinance institutions are able toaccess capital markets for fairly large chunks of capital.


Table 1.3 Costs of obtaining financial services—variations acrosscontracting (in %)

Contracting Method 2003 2004 2005

Individual 30.1 30.7 28.5Individual/Solidaritya 33.8 33.4 30.7Solidarity 35.3 40.4 35.9Village banking 52.0 49.6 47.2

Note: a. Group lending.


There are other microfinance institutions that appear to rely on domes-tic and local markets for their funding needs. In Chapter 2, Swansonexplores this development and the challenges and opportunities that lieahead. In Chapter 3, Rahman and Mohammed explain in detail the firstmicro-credit backed securitization, which tapped capital from differentfinancial institutions to the microfinance sector in Bangladesh.13 Thesedevelopments show that microfinance is starting to establish itself as amajor sector in the eyes of big financial institutions. In due course, this willmean that the assets backed by micro-loans may potentially establish them-selves as a separate asset class.

In addition to the securitization deals that are described by Swanson andRahman and Mohammed, we have also seen other major developments inthe interaction between capital markets and microfinance institutions. In2006, shares of Equity Bank in Kenya were listed on the Nairobi StockExchange. This is reputed to be the first microfinance IPO in Africa, sig-naling the ability of some institutions in microfinance to access equity (risk)capital.14 On 20 April 2007, Banco Compartamos, a microfinance institu-tion launched in 1990, made the first initial public offering (IPO).15 Thisimplies the flow of equity capital into microfinance from the capitalmarkets. Major international financial institutions (IFIs) such as EBRD,IFC, and KfW raise significant amounts of capital and invest in the field ofmicrofinance. These IFIs are private-sector arms of public financial insti-tutions. A recent report suggests that IFIs have a portfolio of $2.4 billionin microfinance, with the average investment size at $4 million.

Microfinance investment vehicles (MIVs) have been created to channelprivate sector funds into microfinance. The MIVs have a portfolio of $2billion, with the average investment size of $1 million. This is a topic ana-lyzed by Swanson in Chapter 2.

Securitization techniques are used to create debt and equity financingpossibilities for microfinance institutions. For example, Citigroup recentlycreated $165 million of collateralized debt obligations (CDOs), which willbe initially backed by 30 MFI loan portfolios in 13 countries. Privateinvestors are expected to invest $69 million of senior tranches rated AA byFitch. IFIs and Citigroup will invest in the subordinated tranches andequity.16 Swanson analyzes the CDO structures and the benefits that accruefrom such innovative financing methods. Rahman and Mohammed inChapter 3 analyze the BRAC securitization deal in sufficient detail so thatthe reader can appreciate the benefits of such innovative ways to tap intocapital markets.

A migration has also started to occur, whereby NGOs are starting to shedtheir “not-for-profit” status and become regulated providers of financialservices.17 Stephens (2007—Note 17) reports the growth of deposit-taking

10 Microfinance

services and increased commercialization, with the median commercialfunding of loan portfolios increasing from a level of 40 percent in 2003 to60 percent in 2005. This also presents interesting regulatory challenges onissues such as governance, investor protection, disclosure requirements, andso on.18

Technology and Microfinance—Opportunities and Challenges

In an earlier sub-section, we noted that the borrowing rates in microfinancemarkets can be rather high. In addition, we noted that the delivery offinancial services such as tiny loans and insurance payments, and acceptingtiny savings deposits require that the service providers overcome (1)significant transactions costs, (2) adverse selection problems, and (3) moralhazard issues. In fact, a skeptic could reasonably ask how microfinance couldbe the mechanism for alleviation of poverty, if the interest rates are so high,and if scalability is simply unattainable due to the factors outlined above.

Technological developments over the last decade hold much hope inovercoming these obstacles. I will review some of these developments inorder to set the stage for Chapter 4, in which Shrivastav explains howmobile phone technology is used in India to potentially deliver loans thatare as low as $20.

Electronic matching of borrowers and lenders: search and delivery costsOne of the factors that contributes to the relatively high interest rates onloans is the cost associated with search and delivery of loans, and the sub-sequent efforts that are needed to manage the default risk of the loan port-folio. The Internet has opened up new vistas for matching borrowers andlenders electronically, and the developments that we have seen in thelast decade have the potential to lower the costs of very small loans. InTable 1.4 we highlight some of the recent innovations that have occurred inthe electronic matching of borrowers and lenders.19 There are firms such asProsper.com in the United States, and Zopa.com in the United Kingdom,which permit peer-to-peer micro-lending and essentially avoiding any inter-mediaries.20 These options are alternatives for the small borrower to pilingup credit-card-type debt. On these websites potential lenders are able toview and evaluate loan requests, ranging from about $1000 upwards, andbid on them. Most of the loans are of short duration, not exceeding threeyears. Legally binding contracts are entered into. Information about creditscores is used in contracting. Monthly payment schedules are enforced.Unsurprisingly, interest rates on these websites are well above what onesees in the “formal credit markets,” ranging anywhere around 15 percent to25 percent or more.


Table 1.4 shows a group of organizations that have exploited the Internetto reduce dramatically the time it takes to match borrowers and lenders,and to arrange the flow of loans between them as well as to promote sociallending at a profit.21 This trend may well expand more in to the field ofmicrofinance, although there are barriers such as illiteracy, and inability toaccess community networks of computers to access the Internet. These bar-riers may well yield due to other technological advances, which we sketchnext.

Correspondent banking and biometric authorization of credit22

A development that has accelerated the access of financial services to thepoor is the concept of “correspondent” banking or branchless banking.This has been especially successful in Brazil, and the concept is bound tohave a major effect in other regions.23

Under this concept, local post offices and shops are equipped withbarcode-reading point-of-sale (POS) terminals. These local entities then act

12 Microfinance

Table 1.4 Matching borrowers and lenders—some innovations

Organization Services Provided Some Attributes

KIVA Kiva enters into a Lenders are able to selectpartnership with existing their borrowers and theirmicrofinance institutions loans are electronicallyworldwide to link lenders transmitted to the localand borrowers MFIs

Zopa Zopa partners with credit Zopa is able to offer theunions thus, it offers attractive features of creditdeposits as investments, and unions through the Internet lending opportunities to its investors, lenders, and

borrowers

Prosper market Provides an open market Has several specializedplace place to match lenders and lending programs targeted

borrowers. Diversification at different communitiesof loans is encouraged

Lending Club Borrowers complete a Accessible to members inpersonal loan request and the networkare screened based on aminimum FICO score, and so on. Interest rate is fixed forthree years

Source: Author’s compilation.

as agents for banks. It is estimated that currently there are nearly 100 000such correspondent entities in Brazil alone. CGAP estimates that nearly 13million customers have been brought into the fold of the banking systemthrough these correspondent networks.

It is well recognized that the literacy levels of microfinance borrowers arerather low. This presents unique challenges in reducing the costs of lendingto them. One technological development that has entered the field ofmicrofinance is the application of fingerprints for the purpose of iden t -ifying and validating financial transactions, through ATM networks.Biometrics and smart cards are already in use in microfinance in India andIndonesia. Biometric teller machines (BTMs) reduce the administrativecosts of extending small loans in communities where the literacy levelsare low. The smart card contains the credit history of the borrower. Bankssuch as ICICI in India and Danamon in Indonesia are employing such technologies.

Village Internet centers and eChoupalsOne of the developments that has occurred in the delivery of services topoor citizens in remote corners of the world is the proliferation of com-munity Internet centers, which deliver both the infrastructure and infor-mation that may be utilized by small agricultural producers to get the bestpossible price and thus eliminate intermediaries.24 Some see this develop-ment as using the technology to empower as well as significantly increasethe range of services to the poorer sections of society. One such innovationis the concept of eChoupals, pioneered by India Tobacco Company (ITC),which deliver price information and an infrastructure for buying the agri-cultural production from small farmers. eChoupals are village Internetkiosks run by local entrepreneurs who provide pricing information fordifferent delivery dates to local farmers and enable the farmers to sell theirproduce directly to ITC, bypassing the intermediaries. Farmers benefitfrom a known price schedule, and ITC benefits from the elimination ofcommissions and transactions costs that intermediaries would havecharged.

Mobile phone delivery of financial services25

It is estimated that there are over 2 billion mobile phone users who not havebank accounts.26 In recent times, many companies have been able to utilizemobile phone technology to provide transfer of cash, make loans, andextend basic financial services. This promises to bridge the “unbanked”sector with the market for financial services at a speed that could not havebeen imagined a decade ago. In South Africa, it is reported that banks havebeen able to link a debit card and a bank account to a cell phone. This


enables the owners of cell phones to make a deposit at a bank or any postoffice, and the deposits are then credited to an account and confirmed viatext message.

USAID has pioneered mobile-phone-based access to financial servicesfor the poor in the Philippines. Microfinance customers make loan pay-ments by a text messaging system, dramatically lowering the transactionscosts and eliminating intermediaries in the process. A concept known as G-cash has been implemented in the Philippines, which effectively allows theusers of cell phones to send and receive cash via text messages. As of March2006, over 1.3 million customers are using the G-cash system, whichhandles $100 million a day. This avenue promises to cut the transactionscosts, time, and effort for both borrowers and lenders. Since the growth ofcell phone customer base has been exponential in the last decade, the poten-tial for a steep drop in the cost of delivery of financial services throughmobile phones is promising.

This particular technological breakthrough is the topic explored byAnand Shrivastav in Chapter 4. Shrivastav examines the growth ofmobile phone technology in India and its potential for delivery for financial services. He also examines the different players in the market. He thendescribes a technology that he has developed for the delivery of micro-loans.

Regulating Microfinance

We have traced some of the major changes that have occurred in the land-scape of microfinance. These changes present regulatory challenges, whichform the focus of Chapter 5 in which Rosenberg explores a framework forregulating microfinance. One such challenge is the question of how to inte-grate mobile-phone-based delivery of credit into the overall bankingsystem with appropriate safeguards. In addition, other challenges arisefrom lender–borrower relationships. We have noted that the costs of access-ing credit in informal credit markets in general and in microfinance in par-ticular can be in the range of 20 percent to in excess of 50 percent. Ouranalysis also showed that the poorest of the borrowers are often subject tothe highest of the interest rates, and they are often in group lending (self-help groups—SHGs), obliging them to expend greater resources in peermonitoring.

A question that naturally arises is whether there must be some oversighton the interest rates charged by microfinance lending organizations. Sincethe borrowers are often not well educated, another question is whether theyfully understand the effective interest rates that are being charged by thelending institutions. Such effective interest rates will have to reflect a

14 Microfinance

number of factors such as: (1) frequency of compounding and payments,(2) transactions costs charged, (3) costs of any mandated insurance poli-cies, and (4) efforts associated with compliance—peer-monitoring efforts,time spent with loan officers, and so on. Full disclosure of these factors maynot always take place.

A related issue is the social pressure that may be placed upon the bor-rowers to enforce timely payments: a borrower in a group-lending contractfaces considerable pressure from the community to service the loan pay-ments, as there is (1) joint liability, and (2) he or she lives in the community,and default can qualitatively and adversely affect his or her social life.

Absence of credit bureaus and other institutions may also contribute toa situation whereby the same borrower may end up obtaining multipleloans well beyond his or her ability to service the loan payments. This mayresult in defaults, which may adversely affect otherwise financially healthymicrofinance lenders.

Defaults may arise due to many reasons. A borrower may default becausehe or she is unable to service the loan contract due to insufficient income.Such poor realizations of income may be the result of factors outside thecontrol of the borrower (such as drought, monsoon, and so on) or owingto poor incentives or insufficient effort. Defaults may also be strategic andmay be coordinated by the entire borrowing group. The risk of default ismanaged by the lenders through various methods: (1) forming groups thatare less risky, (2) contracting with built-in incentives for peer monitoring,(3) frequent (often weekly) payments where feasible, (4) extensive lendermonitoring, and so on. Some recent developments, related in the next sub-section, point to some underlying problems that call for prudent oversightof microfinance practice.

Defaults in 2006 on micro-loans in Andhra Pradesh27

The state government of Andhra Pradesh (in India) shut down 50branches of two major MFIs in the Krishna District in March 2006. Thisextreme action was precipitated when many micro-loan borrowers com-plained to the state government about the “usurious interest rates” and“forced loan recovery” practices. There was an allegation that ten borrow-ers of MFIs in the Krishna District committed suicide because of theirinability to repay the loans taken from the MFIs. MFIs operating inAndhra Pradesh reached an agreement with the state government on MFIinterest rates, product portfolio, inter-MFI competition, credit disburse-ment, and loan recovery methodologies. As per the terms of the agree-ment, MFIs have agreed to an interest rate ceiling of 15 percent. They haveagreed to desist from providing multiple credit to an existing borrower andrecovery of loans at a pace compatible with the borrower’s income level.


MFIs are also to remain strictly within the micro-credit domain, avoidingmicro-insurance products.28

This episode raises several important regulatory questions: should thegovernment set a ceiling on interest rates charged by microfinance lenders?How often should such ceilings be reviewed and reset to reflect creditmarket conditions? Should the government set standards and enforce suchstandards on acceptable loan recovery practices? This episode is a water-shed in focusing attention on the need to have appropriate institutions andlending and loan recovery standards in place in order to promote privatecapital flow into underserved communities.29

Savings and intermediationMany microfinance institutions also accept savings deposits and provideother services such as insurance. The direct and indirect costs associatedwith the provision of these services should be transparent to the consumersof these services.30 Perhaps more importantly, this raises the question ofdeposit insurance. Should there be a deposit insurance program? What willhappen to the deposits if a microfinance institution were to default? To geta sense of the magnitude of this problem, let us take a look at the questionof how pervasive deposits in microfinance are. Gonzalez (2007—seenote 8) reports that the ratio of deposits to loans ranges from 30 percent to46 percent for individual loans. For solidarity groups and village banking,deposits are not reported, and may be presumed to be negligible. Depositsare a big part of banks, credit unions, and rural banks. Obviously, they arenot an issue for non-bank financial institutions and NGOs. How should thedeposits of the poor be protected?

Technological innovationsMobile phone delivery of financial services, community Internet portalsthat enable farmers to sell their crops and so avoid intermediaries, and othersuch innovations also raise important regulatory questions. Should mobilephone deliverers of loans be given a banking license? We noted that in cor-respondent banking, third parties interact with customers to deliverbanking services. Some of the issues that arise from a regulatory standpointinclude the following:

1. Should third parties (correspondents, for example) be permitted toexecute banking transactions and interact directly with customers? Inthis context, how should electronic money transfers and mobile phonetransfer of cash be treated?

2. After 9/11, the regulatory practices in organized credit/bankingmarkets have brought into play risk-based anti-money laundering

16 Microfinance

(AML) rules, as well as rules for combating the financing of terrorism(CFT) adapted to the realities of remote transactions conductedthrough agents.31 How should they be adapted in microfinance?

Finally, issues of consumer protection also have to be addressed.Regulation is costly from the perspective of both formulation and enforce-ment from the regulator’s point of view. If poorly framed, it can also becostly to comply with from the standpoint of the lending institutions inmicrofinance. Rosenberg addresses this very important issue in Chapter 5and articulates how we must evaluate the challenges in this regard.

Gender Empowerment and its Characterization

One of the striking facts about the field of microfinance is that an over-whelming number of the borrowers are women. In Chapter 6, Armendárizand Roome examine this strategy of targeting women or the so-called issueof gender empowerment in microfinance and note the salient fact thatindeed, in the aggregate, seven out of ten microfinance clients are women.This predominant empowerment in lending has been examined in the liter-ature, and many beneficial effects that arise from such an empowerment havebeen documented. More recent evidence, though anecdotal, points to somepotential dysfunctional consequences of such an empowerment. In theirchapter, Armendáriz and Roome observe that we have no reliable empiricalevidence to examine the results of this gender strategy on the extent andquality of economic and social development, and argue for future researchto examine this issue in greater detail. They also explore the potentialbenefits of women bringing their male partners into the fold of microfinanceon a voluntary basis. Such efforts may reduce domestic friction, reducedefault, and potentially increase overall welfare. These and other importantissues raised by Armendáriz and Roome warrant additional research.

Rather than summarizing the potentially important issues thatArmendáriz and Roome raise in Chapter 6, it is useful to characterize thenature of gender empowerment in greater detail:

● Is this empowerment related to the nature of the lending institution?● If so, what are the potential causal factors?● Do we tend to see greater gender empowerment when the average

pool of borrowers is poorer?● Is it the case that poorer women borrowers tend to gravitate to certain

types of lending institutions?● Are there geographical variations in gender empowerment? What are

the cultural factors that might explain such geographical variations?


Evidence on gender empowermentEstablishing such stylized facts may enable us to further expand on thequestions that are analyzed by Armendáriz and Roome. To this end, weanalyze the MIX data of 2006 to shed some light on these questions.32

Table 1.5 documents the distribution of the fraction of women bor-rowers in the sample, split across geographical regions and the types oflending institutions.33 While this is admittedly a very aggregated picture,which may potentially mask country-specific sociocultural dimensions, itstill provides an interesting breakdown. The data clearly show that thegender empowerment is extremely strong in South Asia, and to a lesserextent in East Asia. It is much less so in Africa and Latin America. Indeed,for micro-loans extended by banks, cooperatives, and credit unions,and non-bank financial institutions, we find that the borrowers are verynearly evenly split across the gender in Latin America, and roughly so inAfrica.

Sorting the results by the type of lending institution leads to a very inter-esting stylized fact: NGO borrowers are predominantly women across allregions. We will see later, unsurprisingly, is that NGO borrowers are typi-cally the poorest of micro-borrowers and this is where we find the greatestof gender empowerment. Gonzalez (2007—see note 8) in his trend lineanalysis has also documented a very similar picture for the period 2003–05,which is summarized on a more aggregated level in Table 1.6.

Microfinance borrowers are targeted for credit either on an individual(stand-alone) basis or as a part of a solidarity group (group-lending) basis.The benefits of a solidarity group and the associated peer monitoringeffects have been addressed in the literature.34 In addition, there is theconcept of village banking, which is another self-help group (SHG) consisting of very poor borrowers, with low per capita income. SHG-based

18 Microfinance

Table 1.5 Gender empowerment—percentage of women borrowers acrosslending institutions and regions

Region Banks Cooperatives and Non-bank NGOsCredit Unions Financial

Institutions

Africa 52 39 62 81East Asia 61 N/A 86 98Latin America 52 52 52 71South Asia 97 99 65 100

Source: MIX data of 2006 and author’s calculations.

lending has the beneficial impact of peer monitoring and assortativematching, which leads to lower default risk and hence should result in lowerborrowing rates. What we see in the data as documented in this chapter,however, is that the SHG borrowers are among the poorest, and this causesthe interest rates to actually go up.

Table 1.7 clearly shows the concentration of women borrowers in theSHG category and much less so in individual loan programs.

Median loan sizeTable 1.8 shows that the median loan size of the NGO borrowers is, by anorder of magnitude, smaller in each region when compared with other bor-rowers. South Asian borrowers are all uniformly poor, with a median loanbalance ranging from $100–130 across different lending institutions. Inlight of this it is interesting that the gender empowerment is so different fornon-bank financial institutions in South Asia.

The median borrower in Latin America has a loan balance, which is fourto ten times the median loan balance of borrowers in South Asia: thisappears to suggest that the gender empowerment declines with increases inaverage loan balance, which may be an instrumental variable for seasoned(proven) borrowers, lower default risk, and greater entrepreneurial skills.These assertions remain to be tested.


Table 1.6 Percentage of women borrowers across different institutions2003–05

Year Banks Credit Unions NBFI NGO

2003 52.8 73.0 53.8 79.02004 50.0 56.5 60.8 82.12005 52.5 60.0 56.1 79.7


Table 1.7 Percentage of women borrowers across different contractingarrangements 2003–05

Year Individual Individual/Solidarity Solidarity Village Banking

2003 47.9 67.9 82.0 90.32004 53.9 66.2 92.0 94.52005 51.8 62.0 100.0 90.2


Median total assetsTable 1.9 is provided to give the reader a sense of the relative dominanceof different lending institutions in the microfinance space. A caveat is inorder in this context: only four banks reported from South Asia, andGrameen Bank, which operates out of Bangladesh, reported assets worth$819 830 340 and is by far the biggest player. But this is an exception and notthe rule, as the median total assets of banks in South Asia is only$33 709 260. In general, there are fewer reporting banks in the market, butthey dominate the assets (loans, for the most part) in the market. In SouthAsia, the median assets of banks are nearly five times the median assets ofNGOs. In Latin America the corresponding multiple is 46, and in East Asia,the corresponding multiple is 550! These numbers point to the increasinglydominant role played by banks in extending financial services to poor bor-rowers. As we noted in the section on “Technology and Microfinance,” withthe advent of Internet and mobile phones, we will begin to see an increasingpresence of banks and new players in this market. This has far-reaching regu -latory implications: the median loan sizes are far higher for banks, and thegender empowerment is far lower for banks, with the exception of SouthAsia.

The issue of gender empowerment in microfinance requires a carefulstudy at two levels: first, at a suitable level of aggregation (perhaps at acountry or a regional level) in order to better understand how it comesabout and what the evidence is from a development perspective. Severalinteresting questions arise at such a broad level of aggregation: perhapscertain types of lending institutions prefer women borrowers. Or, it couldbe the case that women borrowers gravitate towards certain lending insti-tutions. Are there differences at the level of countries? If so, what are thedeterminants of such empowerment? For example, with the exception ofBangladesh, in countries with a majority Islamic population, we tend to see

20 Microfinance

Table 1.8 Median loan size in US$ across lending institutions and regions


Institutions

Africa 529 373 215 129East Asia 562 N/A 235 90Latin America 1445 1510 1039 437South Asia 127 124 126 107

Source: MIX data (2006) and author’s analysis.

fewer women borrowers in the MIX database, but NGOs seem to have suc-ceeded in attracting women borrowers in these countries.35

At a second level, we need to address this issue at a very micro level: dowe see group-based lending and greater gender empowerment going handin hand? Is there a significant positive association between first-time bor-rowers (who are likely to be among the poorest) and gender empowerment?

I find the issues studied in the chapter by Armendáriz and Roome to bevery important in many respects. First, we want to document clearly thedevelopmental and social benefits arising from targeting women inmicrofinance to guide future efforts. Second, we need to better understandwhy this empowerment does not appear to be present with some lendinginstitutions in some countries, at least in the context of the evidence pre-sented from the MIX database. The potential benefits of voluntary intro-duction of male partners by women into a microfinance program as a wayto alleviate the number of important issues highlighted by Armendáriz andRoome (in Chapter 6) seems to be well worthy of more detailed investiga-tion in different regions of the world.

CONCLUSION

The landscape of microfinance has changed dramatically, especially at theupper end, where capital markets are rapidly becoming integrated with thefinancing and investment needs of the microfinance markets. This rapidgrowth and inflows of comparatively large amounts of capital presents itsown set of challenges. On the other hand, at the lower end, NGOs remainthe mainstay for the poorest of borrowers, where the loan sizes are verysmall and the interest rates remain high. Technological innovations may


Table 1.9 Median total assets in US$ across lending institutions andregions


Institutions

Africa 19 980 000 1 915 690 3 006 357 1 809 247East Asia 742 857 915 N/A 6 237 324 1 350 415Latin America 166 225 000 11 394 423 29 777 549 3 560 909South Asia 33 709 260 1 143 689 8 516 418 6 098 412

Source: MIX data (2006) and author’s analysis.

well hold the key to reducing the costs of delivering small loans and accept-ing very small savings deposits. They may in turn pave the way to makingthe microfinance approach more scalable. The regulatory challenges asso-ciated with the rapid change in the landscape deserve greater attention frompolicy-makers and researchers.

An issue that the book does not consider but is perhaps extremely import ant is the fact that the poor lack very basic services that the rest ofsociety takes for granted. Such services include: (1) access to primary edu-cation, (2) access to primary health care, and (3) access to and training inoperating more modern tools and technology in day-to-day activities.Access to these is as important if not more so than access to financial ser-vices. On the other hand, it is clear that the access to financial services willenhance the ability of poor households to access these important basic ser-vices and skills as were. A number of practitioners have already recognizedthe need to deliver in parallel both these basic services and the financial services.

NOTES

1. World Bank, PovertyNet website: http://go.worldbank.org/33CTPSVDC0.2. Akerlof, G.A. (1970), “The market for ‘lemons’: quality uncertainty and the market

mechanism”, The Quarterly Journal of Economics, 84 (3), 488–500; Stiglitz, J. and A.Weiss (1981), “Credit rationing in markets with imperfect information”, AmericanEconomic Review, 71 (3), 351–66.

3. Flannery, M. and K. Samolyk (2005), “Payday lending: do the costs justify the price?”,FDIC Center for Financial Research working paper, Chicago University, Chicago.

4. Caskey, John P. (1991), “Pawnbroking in America: the economics of a forgotten creditmarket”, Journal of Money, Credit and Banking, 23 (1), 85–99.

5. See Global Envision (14 April 2006), “The history of microfinance”, www.globalenvision.org/library/4/1051/.

6. Many nationalized banks in India extended credit to “priority sectors” such as agriculture.In Indonesia, Bank Rakyat Indonesia (BRI) extended micro-savings and credit products.

7. The motivation for this initiative underscores the view of many financial institutions:“We subscribe to the view that microfinance investing can contribute to a double bottomline,” said Ed Grzybowski, TIAA-CREF’s Chief Investment Officer, “GMIP, and thisinvestment in ProCredit, gives us an opportunity to seek competitive returns throughsocially responsible investments that we believe have a low correlation to traditionalequity and fixed income markets,” quoted in “TIAA-CREF launches $100mmicrofinance program”, in SustainableBusiness.com (2006), accessed 18 October atwww.sustainablebusiness.com./index/cfm/go/news.feature/id/1384.

8. Gonzalez, A. (2007), “Efficiency drivers of microfinance institutions: the case of oper-ating costs”, MicroBanking Bulletin, 15 (Autumn).

9. CGAP (2007), “CGAP reflections on the Compartamos IPO: a case study onmicrofinance interest rates and profits”, Focus Note, 42 (June).

10. Gonzalez, A. (2007), “Resilience of microfinance to national macroeconomic events: alook at MFI asset quality”, MicroBanking Bulletin, 14 (Spring).

11. The criteria used for the target group classification and the participating institutions aredetailed in Gonzalez (2007) (see Note 10). Median values are reported in the tables.

22 Microfinance

12. Local moneylenders, who operate with limited capital, provide the outside borrowingoptions for the poor, besides friends, relatives, and trade credit from local shops.Anecdotal and scattered evidence suggests that the interest rates charged by the money-lenders are significantly higher than the rates under the microfinance alternative.

13. See also Zaman, S. and S.N. Kairy (2007), “Building domestic capital markets: BRAC’sAAA securitization”, MicroBanking Bulletin, 14 (Spring).

14. Reddy, R. (2007), “Microfinance cracking the capital markets II”, Insight, 22 (May).15. See Note 9.16. CGAP (2008) “Microfinance capital markets update”, 23 (January), www. cgap. org/

mcm/archives/V23_0108.html.17. Stephens, B. (2007), “Commercialization continues apace”, MicroBanking Bulletin, 14

(Spring).18. Hishigsuren, G. (2006), “Transformation of microfinance operations from NGO to

regulated MFI”, IDEAS, www.microfinancegateway.org/content/article/detail/36733.19. Pearlstine, J. Boon (2006), “Lenders, borrowers hook up over the web: Prosper.com and

other sites provide forum for individual bidders willing to offer small loans”, Wall StreetJournal, 20 May; also see Credit Union Magazine, January 2008.

20. Zopa now has operations in the United States.21. Current websites of these organizations may be found at www.kiva.org/; www.prosper.com;

www.lendingclub.com/home.action; https://us.zopa.com/.22. Milbrandt, J. (2008), “Biometrics and smart cards serve as successful microfinance inno-

vations in Asia”, Microfinance Report, 9 January.23. www.cgap.org/policy/branchlessbanking.24. Kumar, R. (2004), “eChoupals: a study on the financial sustainability of village Internet

centers in rural Madhya Pradesh”, Information Technology and InternationalDevelopment, 2 (Fall), 45–73.

25. Milbrandt, J. (2008), “The rise of mobile phone banking”, Microfinance Report, 17January; Chemonics (2006), “Mobile-phone banking expands into rural Philippines”, 24May, www.chemonics.com.

26. Global Envision (2007), “Microfinance goes mobile: cell phone banking revolutionizesfinancial services for the poor”, 3 August, www.globalenvision.org/library/4/1708.

27. Shylendra, H.S. (2006), “Microfinance institutions in Andhra Pradesh: crisis and diag-nosis”, Economic and Political Weekly, 20 May; Microcapital (2006), “Microfinanceinstitutions reach crucial agreement with government in Andhra Pradesh, India”,11 October, www.apmas.org/pdf%5Cn.pdf; Microcapital (2006), “Indian Bank – ICICIreaches arrangement with provincial government on micro-loan interest rates”,www.microcapital.org/?p=580.

28. Shylendra, H.S. (2006) as Note 27.29. This episode also called attention to the potential for increased default risk, and the polit-

ical-economic nature of microfinance. The Andhra Pradesh state government had takena tough stance following allegations that usurious interest rates and heavy-handed loanrecovery procedures contributed to farmer suicides. Four lending institutions—Spandana, Asmita, Umduma Poddu Pedatha, and SHARE Microfin came in for exten-sive scrutiny. In addition, ICICI Bank, which writes most of these loans using these MFIsas disbursement and collection agents, faced potential write-offs worth $100 million.

30. Some microfinance institutions require savings as a condition for lending, and imposecosts for providing such services. Products such as savings, insurance, and loans arebundled in a way that the overall costs to the consumer can become prohibitive. SeeShylendra (2006), note 27.

31. “Branchless banking: rapid growth poses regulatory challenges,” CGAP Focus Note,accessed 31 January, 2008 at www.cgap.org/p/site/c/template.rc/1.26.2154.

32. The analysis is based on MIX database, which is in the public domain. There is no entryin Table 1.5 for credit unions and cooperatives for East Asia as there was only one suchinstitution there during the sample period considered. MIX data only included institu-tions that voluntarily report their financial and outreach data, and hence is necessarily abiased and under-represented sample of the population of the lending institutions and


borrowers in reality. We nevertheless believe that the MIX data can provide very usefulstylized facts about the issues studied in this book.

33. The sample size of reporting institutions is as follows: the data have 37 reporting banks,124 cooperatives and credit unions, 158 non-bank financial institutions, and 295 NGOs.The data are not necessarily contemporaneous and may have a time lag of up to one year.

34. Stiglitz, J.E. (1990), “Peer monitoring in credit markets”, World Bank Economic Review,4 (3), 351–66.

35. In Pakistan, of the six non-bank financial institutions reporting, the percentage ofwomen borrowers ranged from 4.1 percent to 34 percent. On the other hand, two NGOsfrom Pakistan report a near 100 percent empowerment in favor of women! InAfghanistan, of the three non-bank financial institutions reporting, the percentageof women borrowers ranged from 45 percent to 65 percent. One NGO reports 100percent women borrowers and the other two report 25 percent and 40 percent participa-tion by women.

24 Microfinance

2. The role of international capitalmarkets in microfinanceBrad Swanson

INTRODUCTION

In 2004, international capital markets awoke to the attractiveness of invest-ing in microfinance. Since then, debt and equity security issues formicrofinance have raised an estimated US$1 billion from private sectorfinancial institutions seeking commercial returns.1 The deals have takenforms that are familiar in developed markets such as initial public offerings,bond issues, collateralized debt obligations (CDOs), and securitizations ofthe underlying micro-loans. In addition, private sector debt and equitymicrofinance funds have sprung up—for investors who prefer to give dis-cretion to professional managers—and are now thought to control morethan $2 billion, of which more than $300 million is “mainstream” com-mercial investment.2

Overall, cross-border investment in microfinance surged to $1.4 billion in2006, triple the rate only two years previously.3 While traditional suppliersof microfinance capital—non-profit organizations, governmental develop-ment agencies, and individuals—are contributing to this surge,4 the noveltysince 2004 is the participation by private sector institutional investorsseeking full market returns. These mainstream commercial investors, mostlocated in Western Europe and the USA,5 are driving the opening of capitalmarkets to microfinance.

How and why commercial mainstream investors have come intomicrofinance and the likely evolution of capital markets funding formicrofinance is the topic of this chapter.6

THE NEED FOR CAPITAL MARKETS FUNDING INMICROFINANCE

When properly conducted, microfinance is a profitable, low-risk, andexpanding financial activity. For example, from January to June 2007, the

25

26 widely dispersed microfinance institutions (MFIs) in MicrofinanceSecurities XXEB, a $60 million CDO sponsored by Developing WorldMarkets (DWM), had an aggregate annualized return on equity of morethan 25 percent and were growing their loan portfolios by more than 50percent on an annual basis, while their “PAR-30” (total amount of “port-folio at risk,” or loans with payment delays, beyond 30 days) was only 2.9percent.7 This is a performance that any commercial bank would be proudto announce.

Already, the number of borrowers served by MFIs is globally estimatedat 100 million.8 With an average loan size of $170, the total market size isestimated at $17 billion. Yet the potential demand is 15 times the currentmarket—estimated at 1.5 billion, or half the 3 billion global working poor.Thus, microfinance represents a total commercial market of more than$250 billion.

Currently more than three-quarters of the $17 billion funding total israised from domestic markets. However, this number is skewed by theamount—almost $8 billion—coming from deposits in the few countrieswhere MFIs are allowed to take deposits. Most of the estimated 10 000existing MFIs are not deposit-taking institutions—and are unlikely tobecome so, given the cost and complexity of complying with regulationstypically applied to institutions taking deposits from the public. Futurefunding for MFIs is thus unlikely to be sourced mainly from deposits.

Domestic emerging country commercial banks, which should be majorfunding sources for MFIs, are typically averse to lending to them (see thesection on “Local Currency” below). Moreover, capital markets in mostdeveloping countries are thin and the major institutional players are averseto or legally constrained from significant investment in microfinance. Forthese reasons, it is unlikely that domestic sources in emerging countries willgenerate more than a fraction of the more than $200 billion that will needto be raised to satisfy potential demand.

Moreover, while non-commercial investors account for 80 percent of the$4 billion in funding now sourced internationally, this is a legacy of theorigin of microfinance in charitable and officially sponsored developmentactivity. As MFIs’ appetite for capital grows exponentially, it is unlikelythat government agencies and non-profit organizations will increase theirflow of funding proportionately: first, they will be faced with competingdemands for assistance; and, second, they will begin to question whethertheir mission is best served by funding financial enterprises that areprofitable and are increasingly transforming into privately owned com -panies able to attract commercial investment. (However, this realizationmay not have begun to sink in yet—see discussion of role reversal below inthe section on “The Contribution of Non-commercial Investors.”)

26 Microfinance

The only available source of funding for commercial lending of this mag-nitude is the international capital markets. Already, microfinance invest-ment vehicles, which typically include private sector institutional investors,are growing their investment portfolios at 233 percent per year, while officialdevelopment agencies are lagging at 150 percent.9 For the internationalcapital markets, funding a $200 billion industry is routine.

FROM FUND TO CDO

The first10 microfinance fund to reach beyond socially responsible investorswas established in 1998. The Dexia Microcredit Fund, sponsored by Dexia,a Franco-Belgian bank, and advised by BlueOrchard Finance SA, based inGeneva, offers investors a return above their cost of funds (typically 1–2percent over a benchmark rate11) and an ability to redeem their investments.In November 2007, funds under management were $233 million.12

As a fund (a Luxembourg-based SICAV13) offering redemption rights toinvestors, Dexia needs to keep its maturities to MFIs relatively short and alarge portion of its assets in cash (typically 20 percent or more). This limitsreturns to investors and the attractiveness of Dexia’s funding to MFIs,many of which need longer-term maturities on a portion of their liabilitiesto better manage risk.

In 2004, after six years of operations and with $45 million under man-agement, BlueOrchard wanted to provide longer-term funding to MFIsand more attractive rates to investors. It partnered with DWM, an emerg-ing markets fund manager and advisor based in Connecticut, to create thefirst CDO in the microfinance industry. In this transaction, loans weremade to MFIs for seven years from the proceeds of issuing fixed rate bonds.As the bond investors were not entitled to their principal until the bonds’maturity, there was no need to keep large quantities of cash on hand to dealwith redemptions. Furthermore, MFIs had use of the funds for the fullperiod with no interest rate uncertainty.

The CDO was named BlueOrchard Microfinance Securities I (BOMSI).The first closing of $40 million occurred in July 2004 and a subsequentclosing of $47 million was held in April 2005.

This transaction looked very different from any existing microfinanceinvestment vehicle and it marked the beginning of mainstream capitalmarkets investment in microfinance. The major innovations in micro -finance funding pioneered by BOMSI include the following:

1. BOMSI is not a fund—investment decisions are not handed off toa professional manager. There is no asset substitution or active

The role of international capital markets 27

management. Investors in BOMSI have a single source of repayment,a static pool of 14 loans to MFIs taken on at closing. When investorscame into BOMSI, they did so on the basis of their own assessment ofthe credit risk of the underlying MFIs—and they have to live with thisdecision for seven years. Legally, BOMSI is a special-purpose (legal)vehicle (SPV)—a limited liability corporation—registered in the busi-ness-friendly state of Delaware. The vehicle is limited by its constitu-tional documents solely to servicing its loans to MFIs and repaying itscreditors. Cashflows from debtors to creditors pass transparentlythrough the vehicle. When the loans pay off and the liabilities mature,BOMSI will make its final payments to investors and be liquidated (seeFigure 2.1).

2. BOMSI’s funding is stratified in five levels of risk—senior, three classesof subordinated, and, at the bottom, equity. (Both BlueOrchard andDWM are equity investors in BOMSI.) The cashflow from BOMSI’sloans to MFIs is applied according to a strict order of precedence,known in structured finance as the “cash waterfall.” Senior investorsare paid completely first, then the other classes in order of precedence.Equity investors do not get a current return on their investment but if,after all MFI loans have reached maturity and all other investors havebeen repaid, there is residual cash left in the BOMSI SPV it will be allo-cated to the equity investors.

3. BOMSI’s investors do not hold units in a fund and have not made loansto BOMSI. Rather, they have purchased securities—bonds and equityinterests. As we will see later, this distinction was important in attract-ing institutional investment.

28 Microfinance

BOMSISpecial-purpose

legal vehicle

1st priority

2nd priority

3rd priority

Senior

Subordinated(three classes

A,B,C)

EquityMFIs

InvestorsDebt service

Service providers

Trustee AdvisorServicer

Figure 2.1 Cashflows from loan repayments

These elements are common to CDOs and other forms of securitization inmore developed asset classes such as mortgages, corporate loans, autoloans, or student loans. But these are asset classes with substantial datagoing back a number of years describing default performance under anumber of economic scenarios. In the microfinance industry, by contrast,MFI write-off policies vary widely and data on micro-loan defaults typi-cally are not recorded consistently by different MFIs. Moreover, these datatypically are neither independently audited nor rigorously modeled todetermine likely performance under varying circumstances. (Althoughrecently, a non-profit research firm, Center for the Development of SocialFinance, did a static pool analysis of more than 600 000 micro-loans fromtwo MFIs—SKS in India and IMON in Tajikistan—using developed worldmethodology, in order to demonstrate that at least some MFIs are rigorousenough in their record-keeping to permit this style of analysis.)14

Moreover, BOMSI securitized loans to only 14 institutions in nine countries—much less diversification than typical CDOs or other securit -ization transactions in developed markets, where the asset pool may com-prise many hundreds or thousands of loans.

Given these factors, implementing a CDO for the microfinance industryrequired changing the way investors viewed both microfinance and theCDO product.

INTRODUCING COMMERCIAL INVESTORS TO THEMICROFINANCE CDO

Despite the relative paucity of data and diversification, DWM, which tookprimary responsibility for structuring the transaction, encouraged investorsto compare BOMSI to mainstream commercial investments. DWM held theview that to attract sufficient investor interest, BOMSI had to reach beyondthe circle of funders primarily motivated by social, not financial, returns.

To distinguish BOMSI as a commercial investment—different frominvestment funds, donations to NGOs, or other means then available tosupport microfinance—DWM highlighted the following:

● Low default rate in MFI loan portfolios. All participating MFIsreported default rates below 1 percent. Although reporting systemswere not consistent or their results independently verified, the pro-fessionalism and the track record of the MFIs themselves addedcredibility to their findings.

● Favorable risk/return ratios. The tiered capital structure enabledBOMSI to offer high returns to the higher-risk tranche investors,


while providing the lower-risk investors with a substantial degree ofcollateralization, enabling them to feel satisfied with a low creditspread over the benchmark Treasury bond because their notes hadthe highest priority of repayment. Investors were not asked to dis-count their return expectations in view of the presumed social valueof microfinance. With a variety of securities offering different riskand return parameters, DWM was able to segment the internationalinvestor base and thus appeal to a wide spectrum of potentialinvestors.

● Familiar investment instruments. BOMSI debt investors purchasedbonds drafted in their language, and carrying features common tocommercial bonds. They benefited from the appointment of a trusteeto safeguard their interests, as is the case in most bond issues. Thebonds are transferable and each series is endowed with a uniqueCUSIP15 number that facilitates record-keeping, valuation, and per-mitted transfers. (However, the bonds were privately placed, are notlisted, and are not intended to be actively traded.) These featureshelped to ensure that investors had a high comfort level with the formof the investment and could focus clearly on the underlying risk andreturn.

In one important respect BOMSI was differently structured from othercommercial transactions: Overseas Private Investors Corporation (OPIC),a US government development agency, purchased the most senior trancheof securities. Note that OPIC’s ownership of the senior tranche conveyedno protection to more junior investors—by virtue of the cash waterfall,they were exposed to risk in the MFI loan portfolio ahead of OPIC.However, the participation by a large and well-respected developmentagency—often referred to as “the halo effect”—encouraged investors whootherwise might have been unwilling to consider the transaction.

GROWING PARTICIPATION BY COMMERCIALINVESTORS

In the event, the first closing of BOMSI attracted only $1.5 million, or 4percent of the capital raised, from private sector investors seeking a fullmarket return (see Table 2.1 below). However, by the time of the secondclosing, nine months later in April 2006, interest in the transaction hadspread and commercially motivated institutional investors accounted for 41percent of the amount invested. Moreover, the commercial investmentcame from a wider spread of investor types.

30 Microfinance

A little over a year later, in June 2006, DWM closed its third CDO trans-action, Microfinance Securities XXEB (MFS), for which it was solesponsor. This $60 million securitization of loans to 26 MFIs had moreinvestment primarily commercially motivated than primarily socially moti-vated. Moreover, for the first time commercial investors (besides thesponsor) purchased equity. By this time, not only had market familiaritywith microfinance grown, but DWM had also obtained an investmentgrade rating—A—on the MFS senior notes from MicroRate, a specializedmicrofinance rating agency. This heightened commercial investors’ comfortwith the senior tranche. In addition, DWM had sponsored a study indicat-ing that microfinance is less correlated to economic downturn than otheremerging markets assets, making portfolios including microfinance, intheory, less volatile (see the section below, “On the Path to an Asset Class”).This development was of interest to commercially motivated investors.

Table 2.1 shows the amount of investment in three CDO transactionscontributed by institutional investors seeking full market returns, withsocially positive impact a desirable additional benefit. The remainder of the


Table 2.1 Commercially motivated CDO investors by type and riskcategory

Investor type BOMS 1 BOMS 2 MFS TotalUSD USD USD USD

Bank 500 000 9 139 640 9 639 640Money manager 1 000 000 500 000 3 036 000 4 536 000Insurance company 500 000 1 000 000 1 500 000Pension fund 18 000 000 20 500 000 38 500 000University endowment 100 000 100 000

Total 1 500 000 19 100 000 33 675 640 54 275 640% of total investment 4% 41% 56% 37%

Risk categoryEquity 1 500 000 1 500 000Juniora 500 000 600 000 125 320 1 225 320Mezzanineb 1 000 000 18 500 000 1 125 320 20 625 320Senior 30 925 000 30 925 000

Total 1 500 000 19 100 000 33 675 640 54 275 640

Notes:a. For BOMS 1 and 2, subordinated notes C and B.b. For BOMS 1 and 2, subordinated notes A.

Source: DWM.

investment came from investors whose primary motivation was social—thus, for these, financial return was of secondary importance.

High net worth individuals (HNWIs) constituted 10 percent of theinvestment amount in the first BOMSI close. (They are not shown in Table2.1 as we do not characterize them as commercial investors.) This percent-age fell in the second close and by the closing of MFS, HNWIs as a groupwere down to under 5 percent of the total capital invested.

While there is doubtless a significant potential market among HNWIs,and among retail investors generally, for microfinance risk, financial insti-tutions are the bellwethers as they have greater sophistication, moreresources, and stronger tolerance for volatility and illiquidity.

The market for microfinance CDOs has continued to grow. Notably,BlueOrchard has sponsored two more CDOs, with Morgan Stanley asplacement agent, in April 2006 and May 2007, raising $99 million and $110million respectively. The entrance of Morgan Stanley, a “bulge bracket”investment bank, is another signal that microfinance funding is gainingcredibility as a capital markets activity.

However, in the latter half of 2007, and continuing into 2008, the failureof a number of CDOs based on sub-prime mortgages in the USA made thevery term “CDO” suspect in the eyes of many institutional investors andslowed the pace of growth in microfinance CDOs, even though the twotypes of assets are unrelated. As MFIs in microfinance CDOs have contin-ued to perform well, capital markets intermediaries believe that receptivityamong investors to this asset class will improve with increasing recognitionof the inherent robustness of microfinance credit risk.

The relatively small volumes outstanding to date and the legal restric-tions on trading privately placed securities mean that secondary marketshave not developed. Secondary markets should not be anticipated until thenumber of participants and amounts outstanding increase significantly,including several listed issues to act as price indicators for the markets.

CDOs VS. FUNDS

CDOs were the first non-fund capital markets products in microfinance forseveral reasons:

● MFIs typically have balance sheets that are too small to justify trans-actions of the scale required to access international capital markets—aggregating MFI loans is necessary.

● On the other hand, MFIs are used to borrowing internationally—cre-ating loans to international standards and packaging them into the

32 Microfinance

asset side of a special-purpose financing vehicle does not presentinsuperable challenges.

● Capital markets investors predominantly demand instrumentsdenominated in USD or euros (although the decline of the dollar andrelative stability of many emerging market currencies in recent yearsare persuading investors to become more open to local currencyrisk). On the other hand, MFIs typically (but not always) lend in localcurrency; therefore, they are used to borrowing in hard currency andpassing the risk on to their clients, whose demand for loans is rela-tively interest rate inelastic.

● Top-quality MFIs are found throughout the emerging markets sogeographic diversification can be achieved.

● MFIs typically have very few loan products, so their risk on the assetside is relatively easy to analyze and can serve as a proxy for theunderlying risk of the micro-borrower—highly diversified, highlygranular—that strongly draws the typical capital markets investor.

Despite these factors, the relative scarcity of top-quality MFIs may actto brake the growth of this asset category. Of an estimated 10 000 MFIsworldwide, fewer than 100 have qualified for inclusion in a CDO to date. Asmarket demand for CDOs grows, CDO arrangers will need to push farther“down the pyramid” to tap MFIs of lesser size and credit quality to gener-ate assets. But, given the absence of data in the microfinance industry, asnoted above, the analysis of risk in CDOs is not a function of statistics butrather of individual assessment of MFIs.

Investors find it difficult to make the time necessary to take individual creditdecisions on numerous MFIs, especially given that the investment representsonly a very small part of the investor’s portfolio responsibility. Up to now, thepresence in CDOs of MFIs that are mostly top-ranked—demonstrated eitherthrough ratings or performance over time—has served to ease these creditdecisions. But with the top tier of MFIs growing “overbanked” (see thesection below, “Is Microfinance Riding for a Fall?”), CDO arrangers will needto persuade investors to take risks on MFIs that are less known or appearfinancially weaker. Part of this persuasion may come through education—some smaller MFIs may be as creditworthy as their larger peers—but struc-tural features such as credit guarantees or higher collateralization levels maybecome necessary in some deals to assuage investor concern.

While the CDO has broken new ground as an investment instrument inmicrofinance, investment funds have also been growing, and, as previouslynoted, are thought today to control more than $2 billion of capital. Ofcourse, investment funds in microfinance are not new. Traditionally, low-return or no-return funds sponsored by non-profit organizations have been


a major source of funding for microfinance. What is new is an emphasis onfunds that actually offer a return to investors.

Even as of January 2008, of the 89 microfinance funds listed byMicroCapital, a microfinance news and research service, only 26 are char-acterized as actually seeking a financial return.16

Would-be institutional microfinance fund managers have several hurdlesto overcome in persuading clients to invest:

● Investment fee “cascades”. Institutions that manage funds make it apractice for their funds not to invest in other funds in order to avoida build-up of fees that erodes the ultimate returns to their investors.Also, ceding investment discretion to others may appear to weakentheir own standing as managers.

● Lack of transparency. Investors may find it difficult to understand thepricing, volatility, and performance of assets that exist primarily infunds, as the portfolio effects and the manager’s screening activitiescould mask the underlying data.

● Liquidity. Funds typically trade off liquidity (that is, redemption) forreturn. If they provide an easy exit for investors, funds that invest inilliquid assets like microfinance will find it necessary to keep a rela-tively large percentage of their portfolio in low-yielding cash. Investorswith long time horizons, as many institutions have, may prefer to investdirectly in the underlying assets and run the liquidity risk.

However, funds do play an important role in the growth of capital marketsaccess for microfinance. For example, many institutions will choose a fundas their first investment in a new asset category, relying on the manager’sexperience and knowledge of the market to enhance the investor’s comfortlevel, as well as to gain familiarity with a multiplicity of MFIs through asingle investment.

MICRO-LOAN SECURITIZATIONS

CDOs and funds that specialize in senior loans to MFIs are not the onlycapital markets instruments in microfinance. Direct securitization of micro-loans has attracted a great deal of interest, as micro-loans are relativelyhomogeneous and vastly diversified. As the spectacular growth in recentyears of asset-backed securities in international markets makes clear,investors welcome a “pure play” risk on granular financial assets. However,several important constraints are slowing the emergence of a true asset-backed notes product in microfinance:

34 Microfinance

● Short maturity of micro-loans. As opposed to 30-year mortgages,most micro-loans mature in less than a year and feature frequentamortization, so that all but the shortest-term micro-loan securitiza-tions will need to incorporate a mechanism to roll over or substitutethe underlying assets, which greatly increases the structuring com-plexity and administrative cost.

● Origination risk. Because the portfolio of underlying micro-loansneeds constant replenishment, the ability of the MFIs to originatecontinually a sufficient volume of micro-loans is a significant ad -ditional risk.

● Important role of servicer. Successful MFIs cultivate intimate rela-tionships with borrowers. Thus, the MFI role in servicing securitizedmicro-loans is a critical element in the performance of the securitizedportfolio. This makes it difficult to portray micro-loan securitizationsas pure borrower risk. In effect, the performance risk of the MFI ser-vicer is a key component in the overall risk profile—and a difficultone to quantify, much less hedge against.

● Government regulation. Many emerging market jurisdictions havenon-existent, rudimentary, or inflexible regulatory structures thatpose daunting obstacles to the legal structuring necessary to set upsecuritization vehicles, execute true sales of micro-loans, and trans-fer payments transparently to offshore investors.

Given these constraints, there have been only two cases of micro-loan secu-ritization in international capital markets (as opposed to CDOs that secu-ritize loans to MFIs), and both of them have featured substantial creditenhancement by non-commercial investors. First in May 2006, ProCreditBank Bulgaria, a subsidiary of ProCredit Holding AG, sold €48 million ofits loan portfolio to institutional investors in a deal rated BBB by FitchRatings. The European Investment Bank and KfW, the German develop-ment agency, provided partial guarantees.17 Four months later, BRAC, alarge Bangladesh MFI, held the first close of a program, backed by micro-loans, which will issue $15 million (local currency equivalent) of six-monthmaturity notes twice a year for six years. The issue was rated AAA by a localrating agency. The partial guarantors were KfW and the Dutch develop-ment agency FMO.18

EQUITY

As MFIs mature and transform from non-profit organizations into com-panies, including in some cases regulated institutions, their need for equity


grows. With the high return on equity and fast growth of the industry, theinternal rate of return of MFI equity investment looks compelling onpaper. Consequently, at least 15 private equity funds mobilizing $620million (much of it from non-commercial sources, however), have been setup to address this need.19

The major uncertainty in commercial equity investment in MFIs is thesmall number of “exits,” that is, portfolio investment liquidations, to date.Most private equity investors look more to capital gains upon sale of theirstakes and less to dividends as the principal component of their return. Thisis appropriate in microfinance as MFIs need to retain earnings in the busi-ness to finance further growth if they are to escape an endless cycle ofsourcing fresh equity. But without a deep track record of successful exits,the private equity investor is entitled to puzzlement if not skepticismregarding the prospective return on MFI equity investment.

The only private equity fund that has gone through a complete cycle ofinvestment and liquidation is ProFund Internacional SA, which from1995–2005 invested approximately $20 million total in ten Latin AmericanMFIs for an annual average return of 6 percent. ProFund is of interest herenot for its financial returns—it was sponsored by socially motivatedinvestors and did not set out to maximize profits—but for its success in re -alizing all ten exits within its allotted ten-year life.20

All but one of ProFund’s exits came from sales to shareholders or spon-sors of portfolio MFIs, several of them pursuant to a put (that is, a con-tract requiring one counterparty to purchase an asset at a specified pricefrom another party at the seller’s option) or pursuant to an agreementamong existing shareholders. While effective in the case of ProFund, exitsto insiders (management, major shareholders, and sponsors) are worrisometo private equity investors if they are the only feasible means of liquidatinginvestments. Investors prefer a mix of mechanisms including those thatbring in third-party buyers, such as initial public offerings (IPOs), mergers,and acquisitions, in order to set arm’s-length pricing and foster competi-tion. Moreover, puts to insiders expose the put-holders (that is, investors)to the credit risk of put-writers (that is, insiders), and expose the put-writersto substantial future liabilities they may not be willing to take on, or mayaccept only at very conservative valuations. If a put can be agreed, and thecredit risk of the counterparty is acceptable, the risk-adjusted return is notlikely to excite the private equity investor.

Acquisitions by financial or strategic investors are more welcome path-ways to exit, but there have been very few examples of this in microfinance.Microfinance networks might seem to be likely acquirers but most, whetherfor-profit or non-profit, prefer to build their own operations in new coun-tries from the ground up or to partner with smaller, non-corporate MFIs.

36 Microfinance

No substantial organization has attempted a “roll-up,” or a growth strat-egy through acquisition to date.

IPOs have been used to provide exits to investors in two significantcases—Equity Bank in Kenya and Compartamos in Mexico. The lattertransaction, in April 2007, garnered much publicity, some of it unfavorable,for putting $450 million into the hands of existing investors who had paidapproximately $2 million for these shares originally.21 Equity Bank alsorewarded its early investors, but on a smaller scale. These examples havegiven MFI owners and private equity investors hope that IPOs will providelucrative exit opportunities. However, few emerging country stock marketshave sufficient liquidity to provide assurance of full valuation. In addition,both Compartamos and Equity Bank are relative giants in their jurisdic-tions. As market leaders and first-movers, they represented unique invest-ment opportunities that by definition later entrants to these domesticpublic markets will not provide.

It is more likely that MFI acquisitions will provide consistent exit pathsfor private equity investors. Strategic investors such as commercial banks,leasing companies, and insurance companies will see the value in MFIs notjust as lenders but as delivery vehicles for other financial services to a pro-prietary and loyal customer base. These potential investors in many caseswill come cross-border, recognizing that the fundamentals of micro-lendingare roughly similar in most countries, as shown by the success of networksthat apply a common methodology across the developing world. Alreadysome Western European banks have purchased Eastern European banksthat specialize in small and micro-enterprise lending in order to extend theirfootprint into the European Union hinterland.

Another likely source of acquisition is by a competitor. Already somecountries, including Bolivia, Ecuador, India, Nicaragua, and Peru, areseeing competition among MFIs that previously relied for growth on anunder-penetrated market.

THE CONTRIBUTION OF NON-COMMERCIALINVESTORS

As we saw in the BOMSI case, an official development agency can providecredibility and ease market acceptance of a product even without directenhancement of risk. But as international capital markets grow more famil-iar with microfinance, the value of the “halo effect” is diminishing. Yet non-commercial investors are not superfluous in microfinance. They can play avaluable role in taking on risks that commercial investors don’t understandor are uncomfortable with, and in so doing, leverage this investment. For


example, the Global Commercial Microfinance Facility (GCMF), spon-sored and managed by Deutsche Bank, is a $75 million fund whoseinvestors include socially responsible HNWIs, official development agen-cies (from the USA, the UK, and France), foundations, and also a numberof commercially motivated investors such as banks, insurance companies,and pension funds. The facility is designed to make it easier for MFIs toobtain local currency loans from local banks.22 While we deal with localcurrency issues later in this chapter, the importance of the facility for thissection is to recognize that most institutional investors are uncomfortabletaking local currency risk, especially inasmuch as many currencies inemerging markets either cannot be hedged or can only be hedged at unac-ceptable cost.

In essence, the GCMF takes advantage of the ability of non-commercialinvestors to shield commercial investors from risks they are unwilling totake on, thus leveraging the risk capital of the non-commercial investors tothe benefit of both.

Another, less obvious, example of the catalytic role of non-commercialinvestors is the $11.4 million bond issued by Microfinance Bank ofAzerbaijan (MFBA) in August 2007, managed by DWM. This was the firstcase of a bond issued in international capital markets by an MFI withoutcredit enhancement. While MFBA had a growing and profitable business,the issue’s attractiveness was bolstered by investors’ perception that theAAA-rated development agencies owning a majority of MFBA shares—European Bank for Reconstruction and Development, InternationalFinance Corporation, and European Investment Bank—would step in ifthe issuer faced financial difficulty rather than face the embarrassment ofa default in a portfolio investment.

Whether providing a halo to comfort commercial investors or actuallytaking on risk that commercial investors feel uncomfortable with, non-commercial investors can significantly speed up access to capital marketsinvestment for MFIs. But it appears that bilateral and multilateral devel-opment agencies are going beyond this role and are actually crowding outprivate sector investors in commercially credible deals.

MicroRate, a Washington, DC-based MFI rating agency, has publisheda study23 claiming that “development agencies are today heavily concen-trating their funding on the largest and most successful MFIs, exactly thetarget investment market of private investors.” The study posits that devel-opment agencies tend to make easy choices and are squeezing privateinvestors out of the market with their subsidized finance rates.

In 2005 (last full year of data), the study found that the developmentagencies increased their direct funding to top-rated MFIs by 88 percent. Atthe bottom of the pyramid, where MFIs are most in need of the “patient

38 Microfinance

capital” and technical assistance that these agencies provide at taxpayerexpense, the development agencies actually cut their funding to the lowest-rated MFIs by 25 percent.

Shortly after the appearance of the publication, a number of privatesector microfinance funders joined together to appeal to the developmentagencies to change this practice, but the results have been inconclusive.Clearly, if development agencies see their roles as competing with privatesector investors, they will slow the access of microfinance to capitalmarkets.

LOCAL CURRENCY

One of the largest constraints to growth of microfinance funding is the illi -quidity and volatility of many local currencies in the developing world. Ofcourse, if MFIs were able to rely on local funding sources, this would notbe a problem. But, as we noted earlier, the bond markets of most develop-ing countries are thin and poorly regulated. Moreover, institutionalinvestors, the largest capital sources in these countries, are often highlyrestricted in their permitted range of investments.

Paradoxically, local commercial banks, which should be a major sourceof funding for MFIs, in many countries are less likely to accept MFI riskthan foreign banks. This is symptomatic of the larger problem of risk-aversion among these banks. In many countries, capital-hungry governmentscrowd out private lender borrowers. In some countries, banks are content tolend to large corporations, state-owned entities, and foreign businesses andare under no pressure to expand their presence into smaller indigenous busi-nesses. In some countries, banks have simply not made the effort to under-stand and analyze MFI risk, assuming that “banking the unbankable,”whether directly or indirectly through MFIs, cannot be prudent.

Foreign investors typically are uncomfortable with local currency riskthat cannot be hedged. This means that many MFIs must borrow in dollarsor euros and push the risk onto their borrowers. Fortunately for the MFIsthe short maturities of their loans gives them flexibility to effectively re-price their assets to account for currency fluctuations. Even more fortu-nately for the MFIs, most borrowers are unable to access capital from othersources and so accept interest rate hikes that a more affluent and competi-tive market would challenge. Nevertheless, adjusting constantly to unfore-seeable shifts in exchange rates is a strain on MFI operations and imposesadditional risk on borrowers.

On occasion, MFIs and offshore lenders hedge by depositing the hardcurrency loan in a local commercial bank, which then lends to the MFI in


local currency, secured by the deposit. (In a variant of this technique, thedeposit-taking bank is different from the local bank but issues the localbank a standby letter of credit to secure the risk of the MFI local currencyloan.) Although the local bank’s loan to the MFI is effectively risk-free, thelocal bank frequently will not reduce the interest rate to the MFI by a largeenough quantum so that the combination of the local currency interest rateplus the guarantee fee paid to the offshore lender for taking the risk worksout as a feasible financing cost for the MFI.

A number of initiatives are underway to provide unorthodox hedgingfacilities for capital markets investors in thinly traded currencies. TheDutch development agency FMO, for example, is putting together a swapvehicle capitalized with $350 million in equity that would support $1.5billion outstanding in currency swaps that are beyond the maturity avail-able commercially. By acting as swap counterparty for a basket of emerg-ing market currencies, the facility aims to achieve risk mitigation throughdiversification while providing a substantial return to equity investors.24

Ultimately, local currency markets will mature and provide efficient andflexible hedging tools. In addition, by that time, local capital markets mayhave sufficiently matured to lessen the strain put on foreign investment tomeet MFIs’ growing capital needs.

ON THE PATH TO AN ASSET CLASS

The term “asset class” has a number of definitions. From an institutionalinvestor’s standpoint, an asset class is a kind of asset that is suitable forinclusion in an investment portfolio. In order to be suitable, the asset classmust fulfill certain requirements. Fundamentally, it must be recognizable asa distinct kind of asset, such that different investments in the same assetclass can be analyzed together, can substitute for each other, and can berelied upon to perform similarly in similar circumstances.

Crucially, the asset must be liquid, so that portfolio managers can tradeinto and out of the asset easily according to their changing viewpoint andtheir portfolio’s cashflow. Liquidity is a function of several factors includ-ing volume, exchange listings, ratings, research, and so on.

Additionally, it is important that the asset has a track record, data thatcan be analyzed to make predictions about price changes in response tomarket conditions. If the asset is relatively less correlated to other assets inthe portfolio, that is, of course, a positive as the overall volatility of theportfolio will be reduced by including the new asset in the mix.

Overall, microfinance funding is a long way from meeting these require-ments. It approaches the definition most closely in its distinctiveness and

40 Microfinance

relative homogeneity. But it is extremely illiquid and likely to remain so foran extended period of time while volumes build up. Secondary markets arenot likely to develop until there is a critical mass of exposure among a largenumber of investors so that willing buyers can be matched with willingsellers.

Interestingly, a case can be made that microfinance is largely uncorre-lated to other emerging market assets and so would reduce portfolio volatil-ity, or beta. In a study sponsored by DWM and carried out by New YorkUniversity25 the operating performance under different economic scenariosof 283 MFIs in 65 developing countries was compared with that of 112commercial banks from 33 developing countries. The findings were thatMFI financial results are less sensitive to economic downturn than that ofemerging market commercial banks. While the authors concede that thestudy is based on somewhat inconsistent and incomplete data, it neverthe-less serves as a useful indicator and will likely lead to further useful inves-tigation of the characteristics of microfinance as a prospective asset class.

IS MICROFINANCE RIDING FOR A FALL?

Looking ahead, some microfinance investors see events on the horizon thatworry them:

How will microfinance perform in turbulent economic conditions? Theconcern is that the risk of default in the event of global or even localizedrecession is unknowable, and may be substantial, for MFIs that have onlyoperated during periods of prosperity.

This fear overlooks the fact that microfinance as a financial servicesegment is not nascent, even though capital markets only recently “discov-ered” the asset. Many MFIs have been in business for 10–20 years and haveweathered significant economic and political instability in countries such asIndonesia and Bolivia. Experience and research, such as the correlationstudy noted earlier, indicate that MFIs are inherently less vulnerable to eco-nomic shocks than other finance providers. (Of course, a sovereign eventsuch as rescheduling or capital controls, or a breakdown of law and order,could force default on even the strongest and most liquid MFI, as well asany other debtor to external markets.)

The top tier of MFIs shortly may be “overbanked” The fear is that toomuch investment is chasing too little opportunity and that returns arefalling to the extent that investors will lend imprudently to lower-qualityMFIs in order to meet return expectations. The current compression of


emerging market spreads relative to higher-rated paper, while cyclical, high-lights this concern.

However, while many of the best-known and largest MFIs are attractivecandidates for investment, many smaller and more obscure MFIs also havehigh-quality credit risk. This stems from the underlying robustness of themicrofinance business model.

Most micro-enterprises operate “under the radar” of the formaleconomy. The level of economic activity they engage in is so basic as to beimmune from the normal ebb and flow of the economic and politicalsystems they operate in. Their operating margins are commonly quite high(although, of course, small in absolute terms). Their employees are familymembers or close associates whose terms of employment are informal andflexible. Their owners’ liability for business debts is not limited by a legalform—micro-borrowers take personal responsibility for the loans made tothem, and they know their ability to continue to make a living, and oftento maintain the respect of their community, is intrinsically tied to theirpunctual payment of all amounts due.

For the MFI, administering the loan book is time-consuming and labor-intensive, but once the procedures are carefully designed, inculcated, andtested in practice, operations are usually stable, and extending the customerbase of the MFI by opening new branches becomes almost routine.Financial controls need to be strict and minutely observed, however.

In fact, it is difficult to find instances of default by MFIs that seek self-sufficiency (that is, do not view themselves as charitable operations) andhave been in business several years. Certainly some MFIs may have soughtsupport from the international networks they belong to in order to shoreup a weakened balance sheet or improve faulty operations. In addition,some MFIs are believed to understate their portfolio at risk numbers byroutinely extending the maturity of overdue loans. But MFIs that practicethis usually do end up collecting close to 100 percent of the principal andinterest from the overdue borrowers.

Are MFIs abandoning their core constituency? A third concern is the moveof some MFIs “up market” along with their more successful clients. Whilethe vast bulk of MFI activity currently consists of small loans to individ-ual micro-entrepreneurs some MFIs have begun to offer more sophisticatedservices to larger clients involving more substantial risks—small businesslending, mortgages, factoring, leasing, insurance, and so on—and alsoenhanced revenue. Generally speaking, a larger loan is more profitable to afinancial institution than a smaller one, as administration costs do notincrease proportionately with loan size. This is a controversial develop-ment. Some observers denounce MFI “mission drift” and worry that MFIs

42 Microfinance

will abandon their low-income clients as they progress upstream. Othersbelieve MFIs can continue to remain committed to poverty alleviation andstill retain their more successful clients as they accumulate wealth.

As these products take on more importance on MFIs’ balance sheets, theanalysis of the MFIs’financial strength will grow more complicated, and theirperformance vis-à-vis other emerging markets assets may grow more highlycorrelated, reducing their value in lowering portfolio beta. On the other hand,as these MFIs grow to more resemble mainstream financial institutions, bothin terms of size and structure, they may attract the attention of some main-stream analysts, traders, and investors, further enhancing investment sourcesand liquidity. Ultimately, while some MFIs may turn their backs on theirorigins, most will keep their focus on micro-loans even while providinghigher-level services, both because microfinance is good business in itself andbecause it will provide the breeding ground for the higher-value customers.

CONCLUSION

The rush of capital markets investment in microfinance is unprecedentedand it is wise to question its sustainability. Certainly, risks to continuedgrowth abound, and we have noted a number of them, including:

● eventual exhaustion of investment opportunities at the well-knownand accessible tip of the MFI pyramid;

● structural obstacles to providing investors with direct exposure tomicro-loans via securitization;

● scarce track record of equity exits;● lack of clarity regarding the role of non-commercial investors;● underdeveloped local capital markets, coupled with insufficient

hedging tools for foreign currency investment;● illiquidity, sparse data, and small volumes slowing the journey

toward achievement of “asset class” status;● “mission drift” eroding MFIs’ distinctive risks and returns, and less-

ening their value in reducing portfolio volatility.

Many of these risks reflect the fact that microfinance has only recently beenintroduced to capital markets. They should ease over time as investors accu-mulate exposure to this asset. Extrapolating current trends points tofinancial products that are more numerous, more standardized, and morefitted to capital markets norms. At the same time, secondary markets willmake their appearance, and ratings agencies and researchers (both com-mercial and academic) will focus more attention on the sector. Specialized


hedging tools will ease the distortions of too much lending in foreign cur-rency. These developments should abet liquidity and help to give investorscomfort that microfinance is suitable for regular allocations of portfolioinvestment. In effect, investor demand for assets itself will become animportant and self-fulfilling driver of progress in microfinance.

Moreover, as MFI owners and managers grow accustomed to an envi-ronment in which a deep pool of commercial funding is available for thewell-run, expanding MFI, we can expect strategic transactions—mergers,acquisitions, buy-outs, roll-outs, listings, and so on—to become integralelements in the life cycle of successful MFIs. This will result overall instronger, more efficient, and more skilled institutions better serving clients’needs.

Of course, too rapid growth could also lead to speculation, overheating,and a crash, as we have seen many times before in financial markets, fromjunk bonds to high tech to sub-prime. And certainly some MFIs willexpand too quickly and lose control of their costs and their loan books, orcut rates too aggressively for competitive reasons, or push their clients intoover-indebtedness. Microfinance is no more immune to excess than anyother business activity. But the inherent robustness of the microfinancebusiness model lays down a strong foundation for solid growth, and thesizable potential market ensures absorption capacity for substantial freshfinancing.

Overall, the distinctive focus of microfinance on “banking the unbank-able”—bringing financial services to customers outside the formal financialsystem—gives it a unique and attractive profile of risk and reward that candraw institutional investors seeking diversification and absolute return—even those who are unmoved by the prospect of promoting social values.

NOTES

1. Estimate by the author.2. Littlefield, E. (2007), “Building financial systems for the poor”, presented at Cracking

the Capital Markets Conference, New York, 19 March. The author adjusted the data toremove commercial investment in CDOs (which are treated in this chapter as transac-tions, not funds) and to add more recent funds aimed at the commercial market, such asthe €160 million SNS Institutional Microfinance Fund, managed by DWM, which hadits first closing in May 2007.

3. CGAP (2007), “Microfinance investment vehicles”, CGAP Brief, April, www.cgap.org/portal/site/CGAP/menuitem.95cb370f4995240167808010591010a0/.

4. For example, the CGAP brief cited in Note 3 states that portfolio investment by inter-national financial institutions doubled from 2004 to 2006, while microfinance investmentchannels targeted to individuals report increasing assets, such as Kiva (www.kiva.org).See Walker, R. (2008), “Extra helping”, The New York Times Magazine, 27 January.

5. DWM research.

44 Microfinance

6. In this chapter, “capital markets” means transactions or funds in which all or a majorportion of the investment is raised from private sector institutional investors seekingfully risk-adjusted returns.

7. DWM research.8. Data in this paragraph and the next three come from “Optimizing capital supply in

support of microfinance industry growth”, a presentation by McKinsey & Company tothe Microfinance Investor Roundtable in Washington, DC on 24 October 2006. Theexception is the reference to 10 000 MFIs, an estimate widely quoted in the literature; seefor example, Odell, Anne Moore (2008), “Microfinance: catch the swelling SRI wave”,Sustainability Investment News, 11 January.

9. Littlefield (2007), p. 3—see note 2.10. According to the Dexia website, 27 January 2008, www.dexia.com/e/discover/

sustainable_funds 2.php.11. BlueOrchard uses a standard short-term rate as its benchmark, six-month LIBOR

(London Inter-bank Offered Rate), the rate that prime banks charge each other for liquidity.

12. Dexia Micro-Credit Fund Monthly Newsletter, BlueOrchard Finance, November2007, www.blueorchard.org/jahia/Jahia/Site/blueorchard/Products/pid/42/dexiannewsletter.

13. A SICAV (société d’investissement à capital variable) is an open-ended fund common inWestern Europe especially Luxembourg, Switzerland, Italy, and France, comparable toa mutual fund in the USA.

14. CDSF (2007), “Capital markets-style risk assessment: testing static pool analysison microfinance”, Center for the Development of Social Finance, March,www.cdsofi.org/downloads/MFIStudy-CDSF-Mar 07.pdf.

15. CUSIP is an acronym for the Committee on Uniform Securities and IdentificationProcedures, a standards body. A CUSIP number uniquely identifies a specific security tofacilitate custody and trading of securities.

16. MicroCapital, 27 January 2008, www.microcapital.org/?page_id=7.17. “Press Release”, ProCredit Holding AG and Deutsche Bank, 15 May 2006.18. Rahman, R. and S. Shah Mohammed (2007), “BRAC micro credit securitization series

I: lessons from the world’s first Micro-credit backed security (MCBS)”, MF Analytics,Ltd, Boston, 20 March, www.microfinancegateway.com/files/45785_file_11.pdf.

19. DiLeo, P. and D. FitzHerbert (2007), “The investment opportunity in microfinance”,Grassroots Capital Management LLC, June, www.grayghostfind.com/industry_insights/viewpoints/the_investment_opportunity_in_microfinance.

20. “ProFund Internacional, SA (2008)” www.calmeadow.com/profund.htm.21. Rosenberg, R. (2007), “CGAP reflections on the Compartamos initial public offering: a

case study in microfinance interest rates and profits”, Focus Note (42), June.22. USAID (2007), “The Deutsche Bank global commercial microfinance consortium and

USAID’s DCA guarantee”, United States Agency for International Development,January, www.microlinks.org/ev_en.php?ID=17450_201&ID2=DO_TOPIC.

23. Abrams, J. and D. von Stauffenberg (2007), “Role reversal: are public development insti-tutions crowding out private investment in microfinance?”, MicroRate, February,www.microrate.com/pdf/rolereversal.pdf.

24. “TCX–the currency exchange” (2008), www.fmo.nl/smartsite.dws?id=88.25. Krauss, N. and I. Walter (2006), “Can microfinance reduce portfolio volatility?”, Stern

School of Business, New York University, November.


3. Securitization and micro-creditbacked securities (MCBS)Ray Rahman and Saif Shah Mohammed

INTRODUCTION

BRAC Micro Credit Securitization Series I, closed in August 2006, wasthe world’s first securitization of micro-credit receivables and the first ofa new type of investment called micro-credit backed security (MCBS).This securitization was also the first AAA-rated transaction withinBangladesh.

A number of innovative transactions have taken place in the micro-creditindustry in the last few years. In 2004, ICICI bank purchased 25 percent ofSHARE Microfin Ltd’s micro-loans in a US$4.9 million transaction. Thepurchased microfinance receivables were valued at their net present value atan agreed-upon interest rate. A portion of the transaction was guaranteedby Grameen Foundation USA. Also in 2004, BlueOrchard issued its firstcollateralized debt obligation (CDO). In these, and subsequent transac-tions, BlueOrchard pooled collections of loans made out to a number ofmicrofinance institutions around the world.

Unlike the ICICI transaction, the BRAC securitization is scalable andallows for tranching. It is not a one-time sale of receivables at a discount.Unlike the BlueOrchard CDOs, the BRAC securitization involves the directpooling of micro-credit receivables instead of the pooling of loans disbursedto microfinance institutions. The investments are thus directly linked to theperformance of the underlying portfolio of micro-credit receivables insteadof the risk of the originating institution. The transaction is entirely in localcurrency—thus removing any currency risk from the originating institu-tion. As such, the BRAC securitization represents a step in the evolution ofthe linking of microfinance to capital markets.

In general terms, the transaction size is US$180 million in localBangladesh currency. It consists of 12 equal tranches with the asset poolbacking each tranche, mirroring the overall risk profile of BRAC’s micro-credit portfolio. The tenor of the transaction is 6.5 years, with each tranchematuring in 12 months. It is a fixed rate private placement sold pre -

46

dominantly to local Bangladesh investors. BRAC is the servicer, andEastern Bank Limited (EBL) is the trustee.

Having arranged and structured this transaction, we discuss the storybehind it, and lessons from it for the future. The next section dealswith the transaction rationale that convinced BRAC to agree to goingforward with a securitization. The third section examines various politi-cal economy considerations that had to go into arranging the transaction.The fourth looks at the actual structure of the transaction. It highlightsvarious data and logistical constraints that were reflected in the final struc-ture. Working with BRAC’s micro-credit portfolio data allowed us to spotcertain risks that were not anticipated at the beginning of the structuringprocess, and these risks were mitigated in the final structure. Entering theeighth month of the transaction, we are able to draw some lessonsabout how well the structure has held, and what can be taken from it infuture transactions. The fifth section thus examines the performance ofthe bond and the underlying pool of securitized micro-credit receivablessince the beginning of the transaction. We scrutinize the various creditenhancements that were put in place, and explore which of them could bereduced or removed in future securitizations. The final section, as the con-clusion, relates some final lessons from the transaction and ideas for thefuture.

TRANSACTION RATIONALE

Established in 1972, BRAC is Bangladesh’s largest NGO, with nearly100 000 employees. BRAC takes a holistic approach to development, com-bining micro-credit activities with health care, education, and social advo-cacy. With its 1381 area offices, 35 000 primary schools and 35 healthcenters, BRAC has a presence in every district in Bangladesh.

As of March 2007, BRAC’s micro-credit program had more than 5.3million borrowers with outstanding loans. BRAC’s micro-credit activitiesare carried out through three programs targeted at different borrowergroups:

1. Dabi—with loans ranging from around US$50 to US$500 in size, tar-geting individuals with less than 1 acre of land;

2. Unnati—with loans ranging from around US$150 to US$850 in size,targeting individuals with more than 1 acre of land, or individuals whohave graduated from the Dabi group;

3. Progati—with loans ranging from US$350 to US$5000 in size, target-ing small enterprises.

Securitization and micro-credit backed securities 47

Loans generally have a one-year maturity. The interest rate is a flat 15percent, with collections taken on a weekly basis for Dabi and Unnati, andon a monthly basis for Progati.

The BRAC portfolio has been growing rapidly in the last few years.Between January 2006 and January 2007, for example, the principal out-standing on loans grew by 35 percent from US$280 million to more thanUS$380 million, and the number of borrowers grew by more than 10percent to more than 5 million borrowers.

BRAC has been funding its micro-credit activities from a number ofsources, including the savings of its group members/borrowers, donorgrants, and funding from the government and some donor agenciesthrough loans from Palli Karma Sahayak Foundation or PKSF, the apexmicrofinance institution. In early 2004, BRAC borrowed the equivalent ofnearly US$30 million from the local capital markets through its first syndi-cation. Table 3.1 gives the breakdown of BRAC’s microfinance fundingsources prior to the transaction.

These financing options had severe drawbacks. Donor funding wasvolatile, and could not be relied upon in long-term planning. BRAC alsofelt itself under increasing pressure from the government to reduce theinterest rates charged to borrowers as a precondition for further loansthrough PKSF. However, BRAC considered its current interest rates nec-essary for the sustainability and viability of its micro-credit program andfelt that they could not be reduced further. Financing through syndicationswas problematic in the long run as interest rates on future syndicationswould be driven by the volatile credit market and would not adequately

48 Microfinance

Table 3.1 BRAC microfinance sources of funding 2002–05

Source (%) 2002 2003 2004 2005 CAGRa (three-year)

Members’ savings 36.70 45.70 48.33 46.30 8.05Loan from PKSF(quasi-government) 28.50 23.85 15.71 8.14 –34.14

Grant from donors 16.30 12.97 10.35 8.53 –19.42Retained earnings 11.90 14.05 15.02 16.34 11.15Loan from banks(includingsyndication) 6.10 3.24 9.50 20.69 50.25

Others 0.00 0.00 1.09 0.00

Note: a. Compound annual growth rate.

Source: BRAC.

reflect the historical performance of BRAC’s micro-credit program.Besides, syndications could not be carried out indefinitely without fallingfoul of regulatory limits on borrowing.

Securitization presented itself as an attractive alternative, or at least com-plement, to the existing sources of funding. Securitization would allowBRAC to have a more efficient balance sheet. It would improve BRAC’sasset/liability management, and reduce leverage. Properly structured, thetransaction could obtain a higher rating than BRAC, and allow BRAC toraise lower-cost funds. A longer-term transaction would give BRAC theability to plan out its explosive growth. And significantly, BRAC wouldhave a broader investor base. Through the transaction, BRAC would beable to remove its dependence upon PKSF funding, thereby alleviatingsome of the pressure to reduce interest rates.

POLITICAL ECONOMY CONSIDERATIONS

The Approval Process

Bangladesh BankThe central bank initially welcomed the concept of securitization as a wayto deepen the capital markets. In December 2004, around the time theBRAC transaction was initially discussed, Bangladesh Bank arranged awell-attended workshop on securitizations as a source for funding for infra-structure and development projects. And yet, over the next two yearsBangladesh Bank did little to actually implement many of the proposalsthat came out of the December 2004 workshop. For whatever reason, thebank simply refused follow-up on many of the changes one might deem nec-essary for having more efficient capital markets. For example, in the work-shop (and in other public and private discussions with the bank), the needfor a yield-curve, was noted by the local financial community. As we writethis report, the yield-curve is still non-existent in the Bangladeshi bankingsector, and there are no official efforts in place to create such a benchmark.

Our own discussions with Bangladesh Bank about securitizing micro-credit receivables were well received in the months after December 2004.And based on these interactions with Bangladesh Bank, the structure forthe transaction was created and submitted to the bank for approval aroundJune/July 2005. The original signals from the bank indicated that theapproval process would be a short one. After all, the structure addressed theconcerns that it had raised, namely currency risk concerns and the involve-ment of local financial institutions. The entire currency risk of the foreigninvestment was borne by FMO of the Netherlands, and two-thirds of the


transaction would be subscribed by local investors. Yet, Bangladesh Bankwithheld approval until the end of December 2005/January 2006.

Bangladesh Bank raised three objections. First, we would have tofurther reduce the involvement of foreign investors by reducing the size ofthe tranche covered by a KfW/FMO guarantee. Second, the guaranteeswould have to be removed completely after the first year. Third,Bangladesh Bank was worried about income from the bonds and feesbeing remitted outside of Bangladesh by FMO and KfW. This concern wasdirectly related to the overheating of the local US dollar market in thesecond half of 2005, which saw the exchange rate move up in a matter ofmonths from around taka 60/USD to more than taka 70/USD. Due to theincrease in global oil prices, the government-owned Bangladesh PetroleumCorporation borrowed US dollars heavily from nationalized commercialbanks.

The structure was changed to meet the first two concerns. For the thirdconcern, we were able to show that the impact on the local dollar marketfrom remittances would not be as strong as anticipated by BangladeshBank. The structure required FMO to buy a new tranche every six months,and this would actually help in the local currency crisis. Further, the regu-lators were simply an order of magnitude off in their understanding of thesize of fees and other remittances.

Securities and Exchange CommissionThe approval process at the Securities and Exchange Commission (SEC)was similarly fraught with delays. Since the bond issued in each tranchewould have a one-year maturity under our structure, the existing rules didnot require an SEC approval. The SEC confirmed this in our conver -sations with it while the transaction was being structured. However, theBangladesh Bank approval letter required us to approach the SEC againon this question. In mid-January 2006, we approached the SEC for a finalconfirmation that the approval process was not required under the rules.The SEC now decided that it would require us to get approval after allbecause of the six-year term of the transaction, even though the securi-ties issued themselves would be of one-year maturity. Then in earlyFebruary 2006, we were again surprised with a long request for docu-ments and analyses. We submitted these documents and analyses almostimmediately. In March 2006, we received an AAA rating for the transac-tion from the Moody’s-affiliated Credit Rating Agency of Bangladesh(CRAB). After this, the SEC did not logically have a reason to withholdapproval. However, it was only in June 2006 that we received theapproval.

50 Microfinance

What Explains the Delays in the Approval Process?

We can only speculate about the reasons behind the delays in the approvalprocess. Some of the concerns expressed by the regulators—such as pres-sures on the local dollar market—were genuinely felt. And regulators mayvery well have been uncomfortable with the transaction due to their unfam -iliarity with both microfinance and with securitizations. After all, this wasa first-of-a-kind transaction with a number of moving parts (as describedin the next section). Further, the bureaucratic regime is not one known forits transparency or nimbleness. Based on informal discussions, we cansuggest that some of the delays may have been related to deeper concernsof the government.

PKSFEstablished in 1990, PKSF is the apex micro-credit organization inBangladesh, a private–public partnership between the microfinance NGOsand the government. The organization receives funding from the govern-ment, the International Development Association (IDA)/World Bank,USAID, the Asian Development Bank (ADB), and the International Fundfor Agricultural Development (IFAD). It lends these funds to partnermicrofinance organizations at a below-market rate.

PKSF has successfully utilized its funding clout to induce microfinanceorganizations to strengthen their reporting and auditing processes, as wellas implement computerized management information systems (MIS) at thehead office (and in some cases, area office) level, thereby strengthening insti-tutional capacity. The organization has played some role in coordinatingthe activities of microfinance organizations to better target groups that maynot have been receiving microfinance services. It has also facilitatedresearch on microfinance activities.

BRAC itself had received US$30–40 million in loans from PKSF. But asdiscussed previously, BRAC felt itself under increased pressure to reduce theinterest rates it charged as a precondition for being able to borrow againonce the PKSF loans were paid down. One of the rationales for the securi-tization was thus to move away from dependence on PKSF funding.Further, PKSF may have felt its role as the apex microfinance body inBangladesh threatened by the prospect of a microfinance organization beingable to raise funds through securitization without PKSF involvement.

The government2005 was the UN International Year of Microcredit. In what is arguably thehome of micro-credit, the government’s reception of the year was schizo-phrenic. On the one hand, the government sponsored a number of


workshops and events to celebrate the year and highlight the achievementsof microfinance institutions. On the other, the headlines in Dhaka were attimes dominated by statements from the Minister of Finance publiclydoubting whether the activities of microfinance institutions had any posi-tive role to play in development. In the last quarter of 2005, there was apublic disagreement between the Chairperson of BRAC and the Ministerof Finance over the issue of interest rate caps.

But the government’s apathy to microfinance activity in Bangladesh wasnot animated solely by policy concerns about high interest rates. Since2001, the centre-right BNP had ruled Bangladesh in alliance with the right-wing Jamaat-i-Islami. Jamaat represented many of the most conservativeelements of Bangladeshi society, and historically the relationship betweenthese constituents of Jamaat and the activities of the development NGOshas been a tense one. Further, the NGOs have been vocal about the rightsof minorities under the BNP–Jamaat coalition. The influence of Jamaat ongovernment policy since 2001 has been widely reported.

A less well-publicized aspect of the relationship between the governmentand the microfinance sector in Bangladesh has been concerns about theinvolvement of microfinance institutions in political activities. With mil-lions of borrowers and beneficiaries of health care and education pro-grams, microfinance institutions have an unprecedented ability to mobilizevoters. In past years, prominent organizations such as Proshika have beenaccused of openly partisan activities.

The Proshika case is worth looking at closely to get a sense of govern-ment concerns. Like BRAC, Proshika was involved in education and socialadvocacy along with its microfinance activities. Like BRAC, Proshika hadalso been able to scale up its microfinance activities to reach millions ofborrowers in nearly every part of Bangladesh. In the mid-1990s, Proshikaplayed a prominent role in the opposition Awami League-led movement forelections under a neutral caretaker government. In the 1996 and 2001 elec-tions, BNP accused Proshika of slanting its voter education material infavor of the Awami League. And in April 2004, Proshika was accused oftrying to mobilize thousands of its borrowers to assemble in the capital tolaunch an opposition platform. In May 2004, the BNP/Jamaat governmentcracked down upon Proshika, arresting its chairperson and stopping theflow of donor funding to its programs.

Whatever the merits of the Proshika case, government misgivings aboutthe influence of the large microfinance institutions and their potential forquickly mobilizing thousands for political causes led in 2004 to the draft-ing of laws deepening government control of NGOs. It was in this climatethat the BRAC securitization was proposed to the regulators. Concernsabout the perception of Proshika’s activities had led BRAC to form the

52 Microfinance

Federation of NGOs of Bangladesh (FNB) with other NGOs, breakingaway from the Proshika-led Association of Development Agencies inBangladesh (ADAB), the apex development group. But the governmentlikely felt that BRAC’s dependence upon the whims of government fundingthrough PKSF was a vital lever for controlling its activities.

STRUCTURE OVERVIEW

The Structure

The partiesBRAC is the originator in this transaction. BRAC also plays the role of ser-vicer, depositing the collections from the securitized receivables and collat-eral to the special-purpose vehicle’s (SPV’s) accounts on a monthly basis,updating the pool of securitized receivables and collateral, and reportingthe performance of the pool to the investors. The trustee for the SPV isEastern Bank Limited—a leading local bank.

The investors in this transaction are FMO, Citibank, and two leadinglocal banks, Pubali Bank and The City Bank. Citibank’s investment shall(for the first year) be guaranteed by FMO and counter-guaranteed byKfW of Germany. As in most securitizations, BRAC is also the residualbeneficiary of all cash flows after fees, principal payments, and interest pay-ments are paid out each month by the SPV.

MF Analytics shall provide continuing support to BRAC to maintainthe securitized pool of receivables and collateral and assist in reporting performance to the investors. RSA Capital was the lead arranger of thetransaction. FMO, KfW, and Citibank were co-lead arrangers.

The transaction was rated by the Credit Rating Agency of Bangladesh,the local Moody’s affiliate. BRAC’s MIS system and MF Analytics’ poolingand reporting algorithms were audited by PriceWaterhouseCoopers.BRAC itself is audited by Ernst & Young Malaysia.

Tranche structureThis US$180 million transaction is divided into 12 equal tranches over sixyears. Every six months, the originator shall sell US$15 million worth ofmicro-credit receivables to the trust/SPV created for this transaction in returnfor cash. It shall also earmark another US$7.5 million worth of micro-creditreceivables as additional collateral. The trust shall issue US$15 million worthof certificates or bonds of one-year maturity backed by the pool of securi-tized receivables and collateral to the investors. In return, the investors shallinvest US$15 million in the trust on or before the date of issue for the tranche.


Each tranche is divided into three sub-tranches. Sub-tranche A is theFMO investment, amounting to US$5 million. Sub-tranche B is theCitibank investment, also US$5 million. As noted already, Sub-tranche Bis guaranteed by FMO for the first two tranches, and counter-guaranteedby KfW. Sub-tranche C, the remaining US$5 million, is issued to the localinvestors. The certificates issued to the investors in each sub-tranche arepari passu (equitable). It should be noted that the cost of funds for BRACfrom this transaction is between 150 to 200 basis points below what wouldhave been available to it had it gone for a straight loan or syndication.

The securitized asset poolThe underlying asset pool for the securitization consists entirely of BRAC’smicro-credit receivables. At the outset of each tranche, the pool or the col-lateral underlying the tranche is selected to reflect the distribution of loansin BRAC’s current portfolio (excluding all loans that have liens from PKSFand syndications) along three variables:

1. program (Dabi/Progati/Unnati);2. geography (identified by area offices);3. activity or purpose of the loan (identified by a “schematic code”).

The securitized loans are selected within these categories randomly. As a result,the pool is not biased towards including loans of a particular size or age.

The pool is over-collateralized by 50 percent, that is, at the beginning ofeach tranche, US$22.5 million (in equivalent Bangladesh taka) of receiv-ables are pooled as per the criteria noted above. Furthermore, the asset poolis replenished with additional collateral from month to month if the fore-casted cash flow from the pool is less than 140 percent of the SPV liabilityin the following month.

Bond pay-down structure for each trancheThe certificates or bonds issued for each tranche are of one-year maturity.The bonds are amortized monthly based on a predefined pay-down sched-ule. Interest is also paid on the outstanding principal outstanding of thebonds on a monthly basis. The pay-down schedules of the certificates reflectsthe actual pay-down of the underlying securitized receivables. The principalpay-down schedules for the first two tranches is presented in Table 3.2.

With new tranches issued every six months over a course of six years,BRAC is provided a committed, long-term source of funding. The tranchesdisburse funds at a rate that BRAC can absorb without trouble for distri -bution to its borrowers. The one-year maturity of the bonds reflects theshort-term nature of the underlying micro-credit receivables, allows the

54 Microfinance

securitized pool to track BRAC’s portfolio, and provides the investors anadditional level of comfort for investing in this first-of-a-kind transaction.

Credit enhancementsA number of credit enhancements were included in this transaction. The 50percent over-collateralization of the securitized pool and the replenishment ofthe pool with additional receivables in the event of projected cash flows fallingbelow140percentof thefollowingmonth’sSPVliability,havealreadybeendis-cussed. A few additional credit enhancements were added for good measure:

Substitution Loans identified as delinquent and loans with missing or cor-rupted data are removed on a monthly basis from the pool. (The delinquentloans are under our definition of delinquency. Loans identified as defaultedloans under BRAC’s definition are not removed from the securitized pool.)In their place, loans from the same program, geographical region (areaoffice), and type of activity as far as possible are purchased from BRAC asreplacements by the SPV. These replacement loans are selected so that theywould mature later than delinquent and missing data loans.

Removing prepayment risks Instead of being used to pay the investorsor released to the residual beneficiary, prepayments are captured in a provisional account. Once loans with prepayments within the securitizedpool mature, these prepayments are released at the rate at which these


Table 3.2 Principal pay-down schedules for tranche 1 and tranche 2

Montha % Principal Paid Down

Tranche 1 Tranche 2

1 12.25 12.0002 12.25 12.0003 12.25 12.0004 12.25 12.0005 12.00 11.7506 9.85 11.5007 9.00 9.7508 7.50 7.5009 5.90 4.500

10 4.00 4.25011 1.90 2.00012 0.85 0.750

Note: a. September 2007–February 2008.

amounts would have been collected had no prepayment taken place.Prepayment risk is thus totally hedged out from the transaction.

Debt service reserve account A debt service reserve account (DSRA) ofUS$2.5 million was provided by BRAC as an additional credit enhance-ment at the beginning of the first two tranches. The DSRA amount isexpected to be negotiated down after tranche 2.

MF Analytics pool maintenance and reporting modulesThe structure required the creation of a software package by MF Analyticstailored to this transaction. The package included modules that create thepool of securitized receivables, and forecast the cash flow from the securitizedpools and bond pay-down structure at the beginning of each tranche. It alsoincluded modules that on a monthly basis substitute delinquent and missingloans, trap and release prepayments to completely hedge out prepaymentrisk,and replenish the pool if necessary with additional collateral.Additionalmodules fulfill all of the reporting needs of BRAC for this transaction.

Issues

The need for disaggregating dataThe investors desired the securitized pool at the time of pooling to reflectthe risk characteristics of BRAC’s entire micro-credit portfolio. However,once the pooling took place, the portfolio would not track the evolution ofBRAC’s portfolio. Rather, because of substitution, prepayment trapping,and replenishment, the pool would have its own risk characteristics. Therisk characteristics of the pool would thus diverge from those of BRAC’sportfolios over the course of the transaction. Further, as already noted, thesubstitution of loans on a monthly basis and the tracking and controlledrelease of prepayments above would require the processing of individualloans. Thus, it simply would not do to analyze BRAC’s historical portfoliodata at an aggregated level for structuring.

The structuring of the transaction required two levels of data analysis:

1. We had to identify the drivers of risk in the available data, and design theselection process of the securitized pool so the distribution of identifiedcharacteristics in BRAC’s portfolio would be reflected in the selectedpool. This involved identifying and analyzing the qualitative and quan-titative data available on the loans and borrowers in BRAC’s portfolio.

2. Once the pool was selected, we had to run simulations of the dynamicpool on the available historical data as well as on simulated data toreflect various stress conditions to test and refine the structure.

56 Microfinance

Disaggregating the loan data required us to process and analyze enormousamounts of data. At the time of the structuring, BRAC had nearly 5 millionborrowers, and the transactions of these loans were tracked monthly. Eachsecuritized pool contained between 250 000 and 300 000 loans.

Our analyses of BRAC’s information system and portfolio also revealeda number of issues that needed to be mitigated in the final structure:

Information and logistical issues

Time lag BRAC updates the collection and disbursement data at its 1381area office computers daily. However, these data are transferred to the Dhakahead office only once a month. The infrastructure simply does not currentlyexist for more frequent transfers of information to the head office. At thehead office, it takes around a week to complete the process of updating andchecking the database. (At the time of the structuring, this process tooknearly 10–12 days.) As a result, the longest gap between a transaction in thefield and the information reaching the head office is nearly 42 days. A delin-quency or prepayment will in many cases be reported to the head office nearly42 days after they took place. The structure would have to “solve” this lag.

Changing collection dates BRAC’s loans to its borrowers are collected ona weekly or monthly schedule. However, the exact collection dates for a par-ticular loan cannot be known until after the end of the month for a numberof reasons.

No collections take place on local holidays, and these holidays are oftendependent upon the sighting of the moon. BRAC’s system will, during themonth, allow the collection schedules to be updated to reflect these localholidays. Similarly, at a country-wide level, national holidays—many ofthem based on the lunar calendar—require shifts in collection dates.

Furthermore, changes in BRAC’s personnel itself can lead to changes incollection dates. At BRAC, each collection officer is assigned two villagegroups a day to meet borrowers, examine their activities, collect payments,and make new disbursements. The transfer or promotion of a collectionofficer can result in a change in the day of a week that a particular villagegroup will be visited.

Forecasts of collections from the pool of securitized receivables wouldhave to reflect such uncertainty about collection schedules.

Missing, inconsistent, and unusable data The monthly data transferred to thehead office arrives in the form of CDs or zip drives carried by couriers fromthe 1381 area offices. These data are uploaded to the head office servers beforeany reporting can take place. To say that the transfer process is not fail-safe is


an understatement. Each month, BRAC’s head office finds missing informa-tion and corrupted data, and has to ask individual area offices to transfer theinformation again. But such checks do not spot all missing information.

At each of the area offices, the information for thousands of collectionsand disbursements are entered each day by a BRAC-trained computeroperator hired solely for this purpose. While the software used by BRACdoes incorporate some checks on whether information was entered cor-rectly, some mistakes inevitably creep in. As a result, BRAC’s data includedsome inconsistent information.

Further, the software used for data entry at BRAC itself generated someinconsistencies. We were able to spot some of these inconsistencies in ourdue diligence process, and BRAC corrected them.

Finally, at the field offices, there is no way to fix malfunctioning com-puters without sending the machines to Dhaka. Sometimes malfunctioningmachines are not spotted early, and some of the information transferred tothe head office is corrupt and unusable. Because of fires, theft, or naturaldisasters, there is always some underlying risk of spoilage of data.

Notwithstanding these constraints, we found that errors and inconsist -encies in BRAC’s dataset were relatively rare, less than 1 percent of thedata every month. We also noticed that over time, BRAC’s monthly datasetgot better, for a number of reasons. BRAC removed the software-generatedinconsistencies that we spotted. Its internal checks on data qualityimproved. And the process of transferring data became more streamlined.However, the structure would have to take into account the risk of somemissing, corrupt, or inconsistent data.

Data issues in the risk analysis process

Unavailability of historical data One of the limitations that we had to dealwith in the structuring process was the incompleteness of the historicalinformation available to us. BRAC area offices had been computerized atdifferent times, and it was only near the end of 2005 that the computeriz -ation process was completed. Around the same time, BRAC was also in themidst of upgrading the data-entry and database software used at the areaoffices. While accessible computerized data for some area offices went backa few years, for most area offices accessible data went back only a fewmonths. And because the upgrades to the area office systems had not beencompleted, at the time of structuring the deal we were only able to analyzeindividual loans on a monthly basis, instead of being able to look at theactual dates of transactions. Thus, we would only be able to know if a par-ticular loan failed to pay in a particular month, but not the exact date whenit failed to pay.

58 Microfinance

Absence of data on borrower characteristics The BRAC data available tous at the time of the structuring process included information on where theloan was being used and what the loan was being used for. But importantdemographic characteristics were missing, such as the estimated age of theborrower, the length of time they had been borrowing from BRAC, infor-mation on defaults on previous loans, and even the number of installmentsmissed in the current loan. These constraints severely limited our ability tocreate credit scores for individual borrowers.

While the upgrades to BRAC’s system allow some of this information tobe captured, BRAC’s database still does not include useful information thatcould be easily incorporated. BRAC’s health care and education programs,for example, are a rich source of information on borrower households. Yetthese databases still do not speak to each other.

Deciphering the meaning of defaults At the beginning of the transaction,there was some inconsistency in the way that BRAC and the investorsunderstood what constituted a defaulting loan. BRAC defines a defaultingloan as one that failed to pay its total obligation of principal and interestby the end of the one-year period of the loan. Thus, a loan that missed allweekly payments for months but was still to reach the end of its one-yearterm is considered a “current loan” by BRAC. For BRAC, it is thisdefinition that generates the astonishing, well-publicized nearly 100 percentrepayment rates.

Such a definition was simply not palatable for the investors. But aproblem in creating a definition of default or delinquency more consistentwith the conventional understanding of the concept was the fact that onlymonthly aggregates of the transactions of individual loans were availableat the time of structuring the transaction. Taking into account data limi-tations, we defined delinquent loans as those loans that failed to pay theiraggregate monthly installments in the immediately preceding month.

The definition was stricter than BRAC’s own in that many BRAC“current” loans were identified as delinquent. But on the other hand, it wasless strict than BRAC’s in that a loan that was beyond its one-year matu-rity period but was making its payments on time in the immediately pre-ceding month would not be identified as delinquent.

Table 3.3 shows the rate of delinquencies (defined as the principal out-standing of delinquent loans over the principal outstanding of BRAC’stotal current portfolio) for the period January 2005 to April 2006. It alsoshows the cash flow impact of delinquencies in this period. It must be notedthat Bangladesh did not suffer from any natural disasters in this period. Buteven by our arguably stricter definition of delinquencies, the rate of delin-quencies has never been greater than 8.25 percent.


Analysis of risk

Risk variables identified In the available data, our analysis identified thelocation of the borrower and the type of activity of the loan as relevantrisk variables. Borrowers in peri-urban and urban areas, for example, wereless likely to be delinquent than borrowers in remote parts of the country.We also identified the age of the loan as a relevant risk variable. Loansthat were six to eight months old were more likely to be delinquent thannewer or older loans. We also found that loan size did have an impactupon delinquencies. The larger loans given in the Progati program to smallenterprises almost never missed payments. Delinquencies were much morefrequent for the smallest loans in the Dabi program.

Prepayments Our analysis also revealed the existence of risks to the struc-ture that had not been anticipated in the beginning of the structuring process.We noticed that a significant number of loans prepay before their maturitydate. Table 3.4 shows the cash impact of these prepayments between January2005 and April 2006. It is worth noting that in some months, the positive cashimpact of prepayments is comparable to the negative impact of delinquen-

60 Microfinance

Table 3.3 Delinquency rate and cash impact of delinquencies, January2005 to April 2006

Month Year Delinquency Rate (%) Cash Impact (%)

1 2005 8.14 7.692 2005 7.48 7.333 2005 8.25 7.494 2005 7.81 7.355 2005 6.62 6.496 2005 6.14 5.537 2005 6.42 6.098 2005 5.93 5.989 2005 5.47 5.88

10 2005 6.51 6.6211 2005 4.77 5.7212 2005 4.24 4.571 2006 4.83 5.182 2006 4.45 4.833 2006 4.51 4.774 2006 4.89 4.88

Average 6.03 6.02

Source: BRAC data.

cies. Our structure would need to take into account the risk of prepaymentsnot leaving enough cash flow for bond payments in future months.

PERFORMANCE

At the time of the writing of this chapter, 11 (of 12) payment dates on Tranche1 of the transaction had passed, and five for Tranche 2. Over 99 percent ofTranche 1 has already been paid down, and 60 percent of Tranche 2. We arenow in a position to look closely at the efficacy of the structure.

Delinquency Rates

Bangladesh has not witnessed any natural calamities in the last few months.However, there has been some political turmoil over the parliamentary elec-tions, which were cancelled on 11 January 2007 with the declaration of aState of Emergency caretaker government.

Table 3.5 looks at the delinquency rate of loans in the securitized pool.(Delinquency rate for a particular month is defined as the principal


Table 3.4 Prepayment rate and cash impact of prepayments, January 2005to April 2006

Month Year Prepayment Rate (%) Cash Impact (%)

1 2005 7.73 4.222 2005 11.14 5.043 2005 5.84 2.814 2005 5.63 2.955 2005 8.39 4.366 2005 3.26 2.747 2005 2.69 8.838 2005 2.83 8.819 2005 5.17 10.82

10 2005 4.82 19.0211 2005 10.49 21.1712 2005 7.65 11.991 2006 4.58 10.112 2006 4.99 11.863 2006 5.07 10.014 2006 2.70 9.12

Average 5.81 8.99

Source: BRAC data.

outstanding of loans that missed a single payment in the previous monthdivided by the principal outstanding of the securitized pool.) Delinquencies,in the politically volatile period between November 2006 and January 2007never rose above 6.10 percent. The relatively high delinquency rate in October2006 (just after the flood season) is comparable to the rate from previous years.

Delinquencies for Tranche 1 hovered above 15 percent in March 2007. Thiswas because of the aging of the pool. A jump in the delinquency rate wasexpected, as older loans have a much higher probability of being delinquent.

Bond Performance

Tables 3.6a and 3.6b show the excess cash flow in the first two tranchesreturned to BRAC as residual beneficiary, and the future SPV liabilitycovered by the amounts collectible from the securitized pool and cashtrapped in the prepayment accounts.

It is clear that the SPV is awash in liquidity. This was expected, and wasone of the reasons that the credit rating report cited the AAA rating for thebond. For both tranches, at least 150 percent of the SPV’s liability in futuremonths is covered by the underlying securitized receivables and the cashtrapped in the prepayment account without taking into account the US$2.5million deposited in the debt service reserve account.

But what would have happened had the credit enhancements not been putin place? Tables 3.7a, 3.7b, 3.8a and 3.8b examine the impact of removing

62 Microfinance

Table 3.5 Delinquency rate for Tranche 1 and Tranche 2 securitized pool,August 2006–June 2007

Month Tranche 1 (%) Tranche 2 (%)

August 2.60September 2.80October 8.20November 6.10December 4.00January 5.80February 6.90 2.90March 14.00 5.30April 18.00 7.30May 15.00 6.40June 16.00 7.90Average 9.04 5.96

Source: BRAC Data.

the different credit enhancements—over-collateralization, the trapping andcontrolled release of prepayments, and the substitution and replacement ofdelinquent loans.

Removing over-collateralization would have resulted in the SPV beingunable to cover the bond obligations beyond the first two months for bothtranches. The securitized pool barely covers future SPV liabilities in the first


Table 3.6a Tranche 1 excess cash flow and SPV liability covered by pooland trapped prepayments, September 2006–July 2007

Payment Year % Excess % Future SPVMonth Cash Flow Liability Covered

9 2006 99 15710 2006 87 15711 2006 62 16612 2006 79 1721 2007 56 1842 2007 72 1993 2007 78 2224 2007 96 2865 2007 125 4136 2007 174 7627 2007 339 1823

Note: DSRA (debt service reserve account) amount not included in any calculation.

Source: BRAC data.

Table 3.6b Tranche 2 excess cash flow and SPV liability covered by pooland trapped prepayments, March 2007–July 2007

Payment Year % Excess % Future SPVMonth Cash Flow Liability Covered

3 2007 70 1614 2007 84 1715 2007 87 1786 2007 84 1907 2007 76 209

Note: DSRA amount not included in any calculation.

Source: BRAC data.

64

Tab

le 3

.7a

Per

form

ance

of

diff

eren

t cr

edit

enh

ance

men

ts fo

r T

ranc

he 1

: %

exc

ess

cash

flow

of

SP

V li

abili

ty r

etur

ning

to r

esid

ual b

enefi

ciar

y un

der

diff

eren

t sc

enar

ios

Pay

men

tY

ear

All

Cre

dit

Ove

r-co

llat.

,O

ver-

colla

t.,

Ove

r-co

llat.

,N

o O

ver-

colla

t.,

Mon

thE

nhan

cem

ents

No

Cr.

w/P

repa

y.w

/Sub

st.

w/N

oE

nhan

ce.

Tra

p.C

r. E

nhan

ce.

920

0699

102

8789

3510

2006

8783

6470

2111

2006

6240

2444

�7

1220

0679

4930

54�

11

2007

5649

1630

�1

220

0772

4118

37�

63

2007

7837

1237

�9

420

0796

24�

244

�17

520

0712

513

�18

54�

256

2007

174

11�

2371

�26

720

0733

926

�14

154

�16

Not

e:C

alcu

lati

ons

do n

ot in

clud

e D

SRA

am

ount

of

US$

2.5

mill

ion.

Sou

rce:

BR

AC

dat

a.

65

Tab

le 3

.7b

Per

form

ance

of

diff

eren

t cr

edit

enh

ance

men

ts fo

r T

ranc

he 2

: %

exc

ess

cash

flow

of

SP

V li

abili

ty r

etur

ning

to r

esid

ual b

enefi

ciar

y un

der

diff

eren

t sc

enar

ios

Pay

men

tY

ear

All

Cre

dit

Ove

r-co

llat.

,O

ver-

colla

t.,

Ove

r-co

llat.

,N

o O

ver-

colla

t.,

Mon

thE

nhan

cem

ents

No

Cr.

w/P

repa

y.w

/Sub

st.

w/N

oE

nhan

ce.

Tra

p.C

r. E

nhan

ce.

320

0770

7561

6517

420

0784

6448

739

520

0787

4729

70�

26

2007

8432

1563

�12

720

0776

16�

152

�23

Not

e:C

alcu

lati

ons

do n

ot in

clud

e D

SRA

am

ount

of

US$

2.5

mill

ion.

Sou

rce:

BR

AC

dat

a.

66

Tab

le 3

.8a

Per

form

ance

of

diff

eren

t cr

edit

enh

ance

men

ts fo

r T

ranc

he 1

: %

of

futu

re S

PV

liab

ility

for

Tra

nche

1 c

over

edby

am

ount

col

lect

ible

fro

m s

ecur

itiz

ed p

ool a

nd c

ash

trap

ped

in p

repa

ymen

t ac

coun

ts

Pay

men

tY

ear

All

Cre

dit

Ove

r-co

llat.

,O

ver-

colla

t.,

Ove

r-co

llat.

,N

o O

ver-

colla

t.,

Mon

thE

nhan

cem

ents

No

Cr.

w/P

repa

y.w

/Sub

st.

w/N

oE

nhan

ce.

Tra

p.C

r. E

nhan

ce.

920

0615

715

215

515

610

110

2006

157

147

153

154

9811

2006

166

148

158

161

9912

2006

172

148

165

165

991

2007

184

148

181

173

982

2007

199

150

202

182

100

320

0722

215

624

319

910

44

2007

286

152

306

251

101

520

0741

318

550

135

312

36

2007

762

266

1092

633

177

720

0718

2356

633

2015

1737

8

Not

e:C

alcu

lati

ons

do n

ot in

clud

e D

SRA

am

ount

of

US$

2.5

mill

ion.

Sou

rce:

BR

AC

dat

a.

67

Tab

le 3

.8b

Per

form

ance

of

diff

eren

t cr

edit

enh

ance

men

ts fo

r T

ranc

he 2

: %

of

futu

re S

PV

liab

ility

for

Tra

nche

2 c

over

edby

am

ount

col

lect

ible

fro

m s

ecur

itiz

ed p

ool a

nd c

ash

trap

ped

in p

repa

ymen

t ac

coun

ts

Pay

men

tY

ear

All

Cre

dit

Ove

r-co

llat.

,O

ver-

colla

t.,

Ove

r-co

llat.

,N

o O

ver-

colla

t.,

Mon

thE

nhan

cem

ents

No

Cr.

w/P

repa

y.w

/Sub

st.

w/N

oE

nhan

ce.

Tra

p.C

r. E

nhan

ce.

320

0716

115

515

715

910

34

2007

171

137

142

167

915

2007

178

134

144

172

896

2007

190

126

142

182

847

2007

209

127

152

198

84

Not

e:C

alcu

lati

ons

do n

ot in

clud

e D

SRA

am

ount

of

US$

2.5

mill

ion.

Sou

rce:

BR

AC

dat

a.

few months. The trapping and controlled release of prepayments wouldhave allowed the securitized pool and trapped cash to cover future SPV liab-ilities more comfortably, particularly after the fifth month of the transac-tion. Substitution and replacement of delinquent loans have a similarimpact. However, the SPV still would not have been able to cover its liabili -ties between the second and sixth month.

With over-collateralization but no other credit enhancements, Tranche 1would have had some excess cash flow in all months, and the SPV would alsohave been able to cover at least around 150 percent of its future liabilities inany month. While the excess cash flow would have fallen to 15 percent in thelast few months of Tranche 1, the structure could easily have been adjusted toaccelerate payments in the earlier months where excess cash flow was higher.

With over-collateralization and substitution (but no trapping of prepay-ments), there would have been a slight decrease in excess liquidity, whichwould still hover around 30 percent in the fifth to seventh months of thetransaction.

Over-collateralization without substitution but with the trapping andcontrolled release of prepayments would have meant less SPV liquiditythroughout the life of Tranche 1. In fact, the structure with the currentamortization schedule would have been unable to pay its obligations in latermonths. However, the future liability covered by the pool and the trappedprepayments would increase significantly in later months. The trapped pre-payments could be used to extend the maturity of the bond to beyond 12months. A viable structure could have been created with accelerated pay-ments in the early months, lower months in the later months (where cashflows would have been negative) and the release of prepayments to paybond obligations beyond 12 months.

The Tranche 2 experience is consistent with that of Tranche 1.The trans-action till date has not required the replenishment of the pool with new col-lateral, as excess cash flow has never dipped below 40 percent. We have thusnot analyzed the impact of replenishment upon the robustness of the structure.

Clearly, much of the robustness of the structure can be attributed toover-collateralization. The process of trapping prepayments and releasingthem upon maturity adds to the comfort that the SPV shall be able to coverfuture obligations. However, even without the trapping of prepayments, atleast 150 percent of future obligations are covered by the receivables fromover-collateralized pool at any point in time for Tranche 1. Substitution hasa larger impact on SPV liquidity than the trapping of prepayments, but isnot crucial for a viable structure. These results suggest under normal cir-cumstances, over-collateralization may be enough as a credit enhancement.

68 Microfinance

LOOKING FORWARD

Servicer Risk

We have discussed many of the informational and logistical risks that wereidentified and addressed by the structure. However, a major risk noted by theinvestors was servicer risk. In the event of BRAC ceasing to exist, it would takea few months for another institution to step in and restart collections fromBRAC’s borrowers.

A few factors mitigate some of the servicer risk. PriceWaterhouseCoopers(PWC) ran a systems analysis of BRAC’s MIS. PWC noted that BRAC’ssystems were quite robust. In particular, they noted the presence of databack-ups and system redundancies at both the head office and area officelevels. They also independently checked the accuracy of the data. Further,the BRAC database includes the name and location of the individual bor-rower. Thus, in case the servicer needs to be replaced, the new servicer will beable to identify and locate the borrowers whose loan has been securitized.

Yet, it is unavoidable that a new servicer will likely take some time to sendits own collection officers. Additionally, while all funds collected from thesecuritized pool can be identified in the current BRAC system, the 42-daytime lag for information reaching the head office means that there is alwayssome commingling of funds in BRAC’s accounts. In the event of BRACgoing bankrupt, there is some risk of not being able to retrieve all the col-lected funds for the immediately preceding month. However, the presenceof a DSRA fund may mitigate some of these risks.

Moving Beyond the BRAC Micro Credit Securitization Series I

We have already discussed how information on borrower characteristics andhouseholds was limited. This constrained the risk analyses that could be done.Further, historical data were unavailable. In future transactions, more will beknown about the risk profile of BRAC’s borrowers from its existing databases.

But, as noted, BRAC is not tapping into rich sources of information thatare readily available to it. BRAC’s health care and education programscollect detailed information on BRAC’s group members and their house-holds. However, currently these databases cannot be linked to the micro-credit database. Further, BRAC also collects information on the savings ofthe borrowers. These data were unavailable to us, and can yield valuableinformation about the risk profiles of BRAC’s borrowers.

As more is learnt about micro-credit borrowers, investors also have to bemore flexible in their approach to conditions they place on future securiti-zations. The 50 percent over-collateralization, for example, was dictated by


the investors at the start of the transaction without much analysis. The 40percent excess cash flow requirement and DSRA amount were also a func-tion of some investors’ (and regulators’) risk aversion and the need to guarantee an AAA rating in this first-of-a-kind transaction. In futuretransactions, investors should allow analyses of the risk profile of the bor-rowers to guide optimal structures. Such analyses may also create the pos-sibility of tiering risks to meet the appetites of different investors.

The sharing of information about borrowers among different micro -finance institutions may also help these institutions better understand therisk profiles of their own borrowers. It may also allow smaller institutionsto pool their portfolios to achieve the necessary scale for accessing fundsthrough securitizations. Sharing of information—particularly in a countrylike Bangladesh with no national ID cards, let alone credit rating reports—may help mitigate residual servicer risk.

We have noted how BRAC’s own MIS in the area and head offices is quiterobust, particularly given the difficult operational conditions. But the 42-daylag between actual collections and information reaching the head office isproblematic. BRAC has made tremendous strides in getting the informationto the head office quicker. But more frequent updates would reduce the riskof in vestors in future transactions, while also allowing BRAC to improve itsown cash flow management. BRAC is considering the possibility of linkinga few regional offices to the head office through the Internet, and transfer-ring information to these regional offices from the area office weekly. It mayalso be possible to leverage mobile technology for live updates on collectionsand disbursements.

During the course of the transaction, the very fact that BRAC had to exposeand explain its processes and systems to investors, auditors, and a credit ratingagency helped advance transparency and accountability at BRAC. Over thecourse of the transaction, issues spotted with BRAC’s systems or data wererectified as far as possible. BRAC has also made tremendous strides in the lastyear or so in strengthening its reporting systems. We predicted at the start ofthis transaction in 2004 that BRAC would mature as a result of this securiti-zation. We were pleasantly surprised when this prediction was confirmed.

A securitization such as this is really a learning process for the origin -ator, investors, and regulators. It contains lessons for future transactionsat many different levels, from political economy considerations to risksidentified. The structure that was created for BRAC Micro CreditSecuritization Series I incorporated many of these lessons. The structurehas proved to be robust in the first six months of Tranche 1. As the trans-action moves forward into its new tranches, we hope to learn more lessonsabout what works in the current structure for replications of this transac-tion and the creation of new MCBS structures elsewhere.

70 Microfinance

4. Cell phones for deliveringmicro-loansAnand Shrivastav

I will give you a talisman. Whenever you are in doubt, or when the self becomestoo much with you, apply the following test. Recall the face of the poorest andthe weakest man whom you may have seen, and ask yourself, if the step you con-template is going to be of any use to him. Will he gain anything by it? Will itrestore him to a control over his own life and destiny? In other words, will it leadto Swaraj [freedom] for the hungry and spiritually starving millions? Then youwill find your doubts and your self melt away. (Mahatma Gandhi)1

INTRODUCTION AND BACKGROUND

Cutting-edge technology can provide a transformational role in deliveringfinancial, health, and educational services to the poor. This chapter focuseson India and the potential for providing financial services to the poorthrough mobile phone technology. Such a delivery has the potential to cutsignificantly the transactions costs of accessing small loans and this canlead to an improvement in the overall welfare of the poorer sections ofsociety who are currently unable to access financial services from the organ -ized banking sector and credit markets. To provide the reader with a back-drop, it is first useful to review some background data on India.

Demography and Economy2

India is a large country with an area of 3.3 million square kilometers. Thepopulation of India has exceeded 1.1 billion, and is showing a year-on-year growth rate of about 2 percent. Of immediate interest to the subjectmatter addressed in this chapter is the fact that India has a fairly highsavings rate (2004–05) of 29 percent, and a growing literacy rate. Moreover,there is a large pool of professionals, and the youth population is significantas attested by the fact that more than 50 percent of the population is under25 years old. An important fact that has been recognized by the policy-makers is that there is significant unemployment, especially in the age group20–24 in rural and urban India.3 The unemployment rate of educated

71

women in rural and urban India is very high as well. Increased availabilityof credit and other financial services is likely to be an important factor inalleviating unemployment, and in empowering women.

Mobile Telecom4

Mobile telecom has already reached where other sectors are yet to reach.Telecom statistics and growth projections are spectacular: mobile sub-scribers have a cumulative annual growth rate of 86 percent and wereexpected to reach 143 million by November 2006. The Department ofTelecom—DOT—puts forward a target figure of 500 million mobile con-nections and mobile access to every village over 1000 population by 2010.Despite this remarkable growth, the mobile penetration is only 10 percentand is one of the lowest in the world although, even with the 400minutes/subscriber/month usage, which is close to the United States figure,Indian telecom is well placed in the world markets.

Looking at the future, potential projections vary. However, even a con-servative estimate sees 300 million subscribers by 2010. Considering thatthe mobile penetration is lower than many other comparable economies,the growth rate of mobile phones in India is likely to remain very high forthe foreseeable future.

Microfinance5

The RBI (Reserve Bank of India) Internal Group’s “Report on MicroFinance” has noted that the outreach of the Indian banking system hasseen rapid growth in rural areas. Forty-eight percent of the total branchesof the scheduled commercial banks (SCBs) and regional rural banks(RRBs) cater to the rural areas (32 303 branches translates to a populationof about 23 000 people per branch). Of these, 31 percent (136.7 million) ofdeposit accounts and 43 percent (25.50 million) of borrower accounts arein the rural areas. This expansion of the organized financial infrastructurehas reduced the dependence of the rural population on the unorganizedmoneylending sector from 68.3 percent in 1971 to 36 percent in 1991.

In spite of this growth, there continues to be wide gaps in the availabil-ity of banking services in the rural areas as the SCBs cover only 18.4 percentof the rural population through savings/deposit accounts and even a lowerpercentage of 17.2 percent of the rural households by way of loan accounts.Though the primary agriculture credit societies (PACS), with about 100 000outlets, have a deep and wide presence in rural India, their impact in termsof extension of deposit and credit products has not only been minimal butconcentrated in a few states only.

72 Microfinance

The decline in productivity of the rural branches of the commercialbanks, fragility of the cooperative credit structure, and weakness of RRBswitnessed since the early 1990s has further accentuated the problem ofinaccessibility of banking services for a large part of the rural population.Furthermore, as the banking sector has shown, a propensity towards thelarger-size accounts has meant that the number of loan accounts of smallborrowers with a credit limit range of less than Rs 25 000 decreased from58.8 million in 1991 to 36.9 million in 2003.

A vast majority of the rural population remains unintegrated with theorganized banking sector. On analyzing the supply side, we observe someof the reasons for this, such as: (1) people are unbankable in the evalu -ation/perception of bankers, (2) the loan amount is too small to invite theattention of the bankers, (3) the person is bankable on a credit appraisalapproach but distances are too far for servicing and supporting theaccounts, and expanding branch network is not feasible and viable,(4) there are high transactions costs, particularly in dealing with a largenumber of small accounts, (5) there is a lack of collateral security, (6) aninability to evaluate and monitor cash flow cycles and repayment capaci -ties due to information asymmetry, lack of data base, and absence ofcredit history of people with small means, (7) there are human resources-related constraints both in terms of inadequacy of personnel and lack ofproper orientation/expertise, (8) there is an adverse security situation pre-vailing in some parts of rural India, (9) a lack of banking habits andcredit culture, (10) information-shadow geographical areas (geographicalareas where information on identity and residential proof of persons isyet to be completed, which makes know your customer (KYC) compli-ance difficult for banks; hence the population in these areas remainsunbanked), and (11) an inadequacy of extension services, which arecrucial to improving the production efficiency of the farmers, leading tobetter loan repayments.

Similarly on the demand side, there are several reasons for the rural poorremaining excluded from the formal banking sector, such as: (1) high trans-actions costs at the client level due to expenses such as travel costs, wagelosses, incidental expenses, (2) lack of awareness, (3) lack of social capital,(4) non-availability of ideal products, (5) very small volumes/size of trans-actions, which are not encouraged by formal banking institutions,(6) hassles related to understanding documentation and procedures in theformal system, (7) easy availability of timely and doorstep services frommoney lenders/informal sources and (8) prior experience of rejection byindifference of the formal banking system.

Under the microfinance program, loans are extended to the self-helpgroups (SHGs) who pool a part of their income into a common fund from

Cell phones for delivering micro-loans 73

which they can borrow. The members of the group decide on the minimumamount of deposit, which ranges from Rs 20–100 per month dependingupon the size of the group. The group funds are deposited with amicrofinance institution (MFI) against which they usually lend and thedeposits are usually placed with a bank by the MFI. The group funds arethe way “micro-savings” are enforced, though it may seem like collateral.The loan ticket sizes are usually Rs 2000–15 000. Although loan repaymentis a joint liability of the group, in reality individual liability is emphasized.Maintaining group reputation leads to the application of tremendous peerpressure.6

In India and other Asian countries the majority of SHGs typicallyconsist of women because, in these countries, self-employment throughmicrofinance was perceived as a powerful tool for the emancipation ofwomen. A World Bank report (2001—see note 6) observes that genderequality is a necessary condition for economic development. It reports thatsocieties that discriminate on the basis of gender are in greater poverty, haveslower economic growth, weaker governance, and lower living standards.And the results are encouraging. Loans obtained from MFIs are utilized inagriculture and small businesses. Independent incomes and modest savingshave made women self-confident and helped them to fight poverty andexploitation.

This can be seen from a statement by a woman beneficiary: “Previouslywe had to cringe before our husbands to ask for one rupee. We do not haveto wear tattered sarees anymore and, today, we have the confidence to comeand talk to you without seeking permission from our husbands” (as told tothe author of the UNPAN’s Field Survey [note 6]).

Suvidha7 proposes to play a significant role in microfinance using theSWIFT mobile transaction platform and providing Beam (see “TheProduct” below) services, leveraging its distribution network and customerprofiles. Banks have the opportunity to partner with Suvidha to supplementtheir extension banking services too. Likewise, MFIs, RRBs, SCBs, NBFCs(non-banking finance companies), and banks can also partner Suvidha toextend their product deliveries and manage both inbound as well as out-bound payments through the network of Beam Mobile Entrepreneurs.

Payment System Inefficiencies

The GSM Association (GSMA) launched a pilot program in January 2006,aimed at tapping the ubiquity and ease-of-use of mobile technology toenable the world’s 200 million international migrant workers to send remit-tances easily and securely to their dependants, many of whom did not havebank accounts.8 By exploiting the extensive reach of the mobile networks,

74 Microfinance

the program complemented existing local remittances channels and madetransferring money internationally significantly more affordable.

In India, the situation of payment realization for different modes is asshown in Table 4.1.

It can be seen from Table 4.1 that the present payment systems in Indiaare still quite inefficient and can contribute to lower than optimal rates ofeconomic growth.

GLOBAL TRENDS

As per a report by IFC Washington-GSMA, the advantage of developing amarket for micro-payments (also referred to as m-commerce), is that it continues to drive the economic system toward a cashless transaction envi-ronment.9 Elimination or minimization of physical cash has many advan-tages, including less opportunity for fraudulent or criminal activity,reduction of cash handling costs, and, for the user, less reliance on having theright amount of cash when needed. It also allows the value of money to bebetter utilized. Cash held outside the banking system is not available forshort-term investment so that the time-value of the cash asset is lost.


Table 4.1 Payment realization for various payment modes in India

Nature of Financial Location of Time to CompleteTransaction Transaction Transaction

Cheques Local (same bank) Three daysOutstation 10–30 days plus seven

days by post.Demand draft Local (same bank) One to two days, plus seven

days by postOutstation Three days, plus seven days

(other bank) by postElectronicCredit card Seven days (merchant

payments)Debit card Seven days (merchant

payments)Electronic clearing system Four days (limited

coverage)Postal money order 10–45 days

Source: Data collected by author.

In the more affluent economies, there is already a good infrastructure fora cashless environment, with most people having bank accounts and anarray of both debit and credit cards. Nevertheless there is an underlyingneed for cash for minor purchases and there is little incentive to eliminatecash entirely. These economies can manage quite well and there is nospecific interest group that feels sufficiently under pressure to developsystems aimed at eliminating cash from the environment. Systems that havebeen developed in such markets are often expensive and hence not particu-larly attractive to the customer.

In the developing economies however, there is a very large “underclass”that is totally reliant on cash for all their day-to-day expenses. Moreover,this underclass makes no use of the banking sector and so is “invisible” interms of its cash value. At the same time, the need for cash forces theproviders of goods and services in these markets to have adequate cash-handling facilities and this comes at some cost. In these cases, the com-mercial organizations have much more to gain by addressing the problemof cash transactions. Not only is the risk associated with cash holdingsmuch greater, but the time-value of the cash being held outside the bankingsector is entirely lost. Furthermore, the population in this category is lost,that is, unseen by the banking sector. For these reasons, there is likely to bemore incentive in developing economies to move the population at largeaway from cash, than exists in developed economies. That being so, a solu-tion that meets the needs of developing economies will also have extensiveapplication in the developed economies. This arises because the solutionmust be accompanied by very low costs as, if it were otherwise, the solutionwould have no appeal in those developing economies. The resulting low-cost solutions can then be applied in the developed economies, resulting infurther efficiency gains.

Further, according to the IFC report, the most successful micro-paymentapplications are to be found in the Philippines, with over 3.5 millionm-commerce users on two mobile networks. The key success factors for thatmarket included the ability to load prepaid airtime credits as well as theability to transfer both cash and airtime credits between customers.Coupled with these were the low values set by the operator for such prepaidtop-ups or credit transfers. Typical top-ups of 47–57 (US) cents wereallowed by the networks (equivalent to around four to five minutes of calls)while transfers between customers of both cash and airtime credits werepermitted as low as 4 cents.

The target market surveyed by IFC is attuned to “sachet purchasing” orthe practice of purchasing goods in very small quantities packed in sachets.This phenomenon is known to be common in other developing marketswhere the populace rely on cash for all trading and can afford to buy

76 Microfinance

provisions for just a few days’ consumption. This market does not exhibitbulk purchase tendencies and m-commerce that involves a significant cashdeposit or payment will be unlikely to find any significant uptake from thetarget market.

While the application of m-commerce to developing markets was notconstrained to the Philippines, African (South Africa and Kenya) marketdevelopments seemed to reflect the Filipino views, indicating that the targetmarkets in these geographically diverse areas were very similar in their useof cash and their expectations.

The range of features available in each market showed significant uni-formity as to be expected if the target markets were similar. With minorvariations, the features of all systems included:

● provision for cash deposits and withdrawals;● the ability for third parties to make deposits into a user account

(employer, family member, or MFI);● the ability to make retail purchases at selected outlets;● over-the-air prepaid top-ups using the cash already in the account;● the ability to transfer cash between users’ accounts;● the ability to transfer airtime credits between users;● provision for bill payments.

These features could be used for microfinance applications involving bothloan repayments as well as loan advances, and this area in particular isbeing exploited in the Kenya trials and in Philippines services in conjunc-tion with the Rural Bankers Association.

Apart from the use of the services by MFIs all services studied by IFCoperated on a debit account basis, that is, the account could only be oper-ated in credit. As a result, bad debt is not an issue other than loss causedby fraudulent activity. No operator indicated any serious concerns in thisarea and provided the overall system security was ensured the possibilityfor fraud could be managed. In that regard, most of the systems studiedinvolved a bank with normal banking systems in place as this arrangementresults in the fraud issue being restricted to the bank’s area of involvementfor which it will be well-equipped.

The possibility of money laundering was considered by all serviceproviders and it was noted that in all jurisdictions the banking regulatoryauthority had established appropriate policies governing the activitiesof the banks. These policies included monitoring transaction levels and frequency, looking for transaction patterns and stipulating bothmaximum account balances and daily transaction levels. While it is poss -ible for a network operator to take almost full responsibility for the entire


micro-payment service, only one service in the Philippines was operating inthat manner. Even then, the actual cash float generated was held in one ofthe country’s regular banks. All other cases studied involved cooperativearrangements between banks and networks. In view of the regulatory issuessurrounding the banking industry, this method of operation is more likelyto appeal to intending service providers, given that the banks can bringadditional advantages including the availability of debit cards throughissuers such as MasterCard.

While there were no quantifiable figures available on system costs,various estimates placed a likely cost in the range of US$5 million to US$10million with an expectation that an m-commerce system could be profitablewith as few as 25 000 users connected but that would depend on the overallinvestment and service operating costs. Various estimates placed the trans-action level at around two per customer per day and average transactionvalues at between US$15 and US$30 per customer and airtime top-ups ofaround US$4 per time. These figures can only be regarded as indicative ofthe type of activity that may be encountered.

According to a study by Portio Research, SMS (short message service)traffic in the Asia Pacific (AP) region is expected to increase from 434billion messages in 2004 to over 1.2 trillion by 2010.10 (See Table 4.2.)

One of the key growth drivers is attributed to higher future mobilephone penetration in India and China. The report projected that the com-bined markets of India and China would account for over 800 billionSMS messages per annum by 2010. The study predicts that India, with alow mobile penetration rate has “massive potential.” SMS in India isexpected to grow from 12.3 billion messages in 2004, to 180 billion in2010.

In summary, the situation provides an excellent and growing opportunityto Suvidha and its SWIFT technology platform, which uses SMS toperform both inbound as well as outbound transactions. Suvidha’s networkof Beam Mobile Entrepreneur women become self-employed by providingthe front end services at the localities where they live, and thus gain bettercontrol over their lives.

78 Microfinance

Table 4.2 Increase in SMS traffic in Asia Pacific, 2004–10

SMS 2004 2005 2006 2007 2008 2009 2010

Traffic (billions) 434.1 540.1 672.8 802.4 935.9 1072.1 1212.7

Source: Portio Research, accessed atwww.adnetasia.com/news/ communications/ 0,39044 192,39252956,00.

THE INDUSTRY

India currently has around eight domestic companies operating in thetransaction space as shown in Table 4.3.

Paymate is SMS-based but exclusively for use by customers havingaccounts with banks, for example, Citibank. JiGrahak requires downloadof software to a mobile handset and is thus limited to certain types of hand-sets. MChek works on GSM technology, as it is USSD (unstructured sup-plementary source data) technology-based solution and provides servicesto the subscribers of Airtel and some banks. ITZ Cash and Done Card areprepaid card-based solutions that work on the Internet. Wallet 365 is alsoan Internet-based service for banked customers. Fino is a closed-user pro-prietary technology service provider for MFIs. Last but not least, the tech-nology focus of A Little World is on biometrics-based ID, RFID (radiofrequency identification) smart cards (Java, PKI) and NFC (near field com-munication) mobile phones as acceptance and enabling devices (with mer-chants, field forces of MFIs, and at cashless ATMs).

While ITZ Cash, Done Card, Paymate, and JiGrahak market themselvesas m-commerce, their services can only be used if the subscriber has Internetaccess—microfinance is not their focus. Further, none are interoperable orneutral as they are exclusively tied either to a specific bank (Paymate, MChek,A little World) or telecom (MChek, JiGrahak) or can be used by special typesof phones (A Little World, JiGrahak, MChek via GSM), thus excluding asignificant number of CDMA (code division multiple access) subscribers.Except for A Little World, none are focused towards microfinance.

Fino, on the other hand, acts only as an application service provider(ASP) for microfinance institutions requiring core banking application and


Table 4.3 Domestic transaction companies operating in India

Company Website

Paymate www.paymate.co.in/web/Default.aspxJiGrahak www.jigrahak.com/siteMChek www.mchek.com/ITZ Cash www.itzcash.com/Done Card https://www.donecard.com/index-1.aspxWallet 365 www.timesofmoney.com/tomsvc/jsp/home.jspFino www.fino.co.in/A Little World www.alittleworld.com/

Source: Gathered from market intelligence.

point-of-sale devices. The parentage of Fino is ICICI Bank, hence possi-bilities of interoperability with other banks will be a challenge.

Suvidha-Beam is focused on the micro-payments for the unbanked andenabling microfinance offered by MFI/banks. Beam does not require theInternet, or software to download, change of SIM, or require special equip-ment. It is interoperable with subscribers of any telecom.

THE PRODUCT

Beam is an innovative and simple way of transacting money using mobilephones. It takes advantage and plugs the inefficiencies in the paymentsystems of the economy. It enables subscribers to register and use a host ofother services, anytime, anywhere, using short message service (SMS).

The product has two parts. A robust, future-ready technology platformcalled SWIFT is at the backend; and a stored value prepaid card that con-sumers purchase for using services, called Beam.

Backend Technology

The backend SWIFT (subscriber wireless interaccount financial transac-tion system) is a mobile commerce platform that took several years todevelop. It leverages the cumulative knowledge and experience gained bySuvidha from product distribution and providing transaction managementservices to banks as well as telecom services. SWIFT is a sophisticated,robust, secure, and scalable application having disaster management andbusiness continuity system too. It lets subscribers use the Beam services viaSMS, IVRS (interactive voice response system) or the Internet.

ServicesBeam as a service allows mobile phone subscribers to send money, givegifts, pay each other, make purchases from merchants besides a host ofother services—take credit, make deposits, obtain insurance, make invest-ments, all using their mobile phone.

Services can be accessed as soon as a mobile phone customer registerswith Beam. This can be done by sending a simple SMS message to Beamand assigning him or herself a secure personal identification number(SPIN). The subscriber’s Beam account is established automatically bySWIFT and the subscriber receives an SMS to this effect within a fewseconds.

Beam prepaid cards can be purchased from any retailer, Beam Merchant,Beam Express (shops) franchisee or Beam Mobile Entrepreneurs (individ -

80 Microfinance

uals). Additionally the Beam prepaid cards can also be purchased fromSuvidha’s alternative channels comprising SCBs, cooperative banks, RRBs,NBFCs, MFIs, and India Post.

Subscribers not having a bank account can purchase Beam prepaid cardsto top up their account and perform a variety of transactions. Money canbe gifted via Beam to another subscriber. A Beam Merchant can be paid bya subscriber using Beam. Similarly, refund of the residual amount in thesubscriber’s Beam account can also be taken from any Beam MobileEntrepreneur or Express franchisee.

Besides micro-payment services, Beam Mobile Entrepreneurs can alsoextend microfinance, micro-insurance, micro-investment as well as internat -ional money transfer services of Suvidha partners to the Beam subscribers.

Additionally, the Beam Mobile Entrepreneurs can act as service deliveryvehicles and extend microfinance, micro-insurance, micro-investment, aswell as international money transfer services in his or her locality. These willbe to the customers located anywhere who may not have a mobile phoneand/or may not be registered with Beam but are clients of banks, coopera-tive banks, RRBs, NBFCs, India Post, MFIs, SHGs, and Ladies’ KittyClubs (LKCs).

Transaction Ecology

Figure 4.1 shows the human ecology of the various types of transactions.Subscribers can be seen sending money, giving gifts, paying each other,making purchases from member Beam Merchants and also taking refundsfrom Beam Mobile Entrepreneurs using their mobile phones.

Similarly, Beam Mobile Entrepreneurs can be seen providing refund ser-vices and also extending microfinance, micro-insurance, micro-investment,and international money transfer services to Beam subscribers, as well asto the customers of banks, SCBs, RRBs, NBFCs, MFIs, SHGs, India Postand LKCs, who may have mobile phones and/or are not registered withBeam.

THE FUTURE AND CHALLENGES

Suvidha is starting off with micro-payment services. As it moves forward,it will use the flexibility and scalability of SWIFT technology platform,leverage the distribution network and the profiles of Beam subscribers tooffer microfinance products, micro-insurance (life and general insu -rance products like crop insurance and so on) of its partners. As the feet-on-street, that is, Beam Mobile Entrepreneurs mature, it will offer the


micro-investment products of partners. Suvidha will also offer internat -ional money transfer services of partner money transfer organizations(MTOs) through the Beam Mobile Entrepreneurs and Beam Expressfranchisees. In addition to offering partner products, Suvidha will providemicro- payments transaction management services to customers of banks,SCBs, RRBs, NBFCs, microfinance, micro-insurance, micro-investment,and MTO companies. It will move to other countries at an appropriatestage.

Challenges

1. Regulation. The regulatory environment for payments, microfinance,micro-insurance, micro-investment, and international money transferis still evolving and there are no clear guidelines.

2. Taxation. While service tax is well understood, VAT is administered byindividual states who are not clear on what the treatment should bewith regard to charging or not.

3. Anti-money-laundering (AML). Customers may not yet have beenissued the required documents with regard to anti-money-laundering.

4. Combating the financing of terrorism (CFT). Here again no clear dis-semination of information has occurred at the enforcement level.

82 Microfinance

Subscriber

Self-help groupsLadies-kitty Clubs

CREDIT - F2S

DEPOSIT - S2F

INSURANCE - S2F

INVESTMENT - S2F

INTL MONEY XFR - F2S

GIFT - S2S Subscriber

Blank

Insurance

REFUND-S2F

Membermerchant

Mobile entrepreneurfranchisees

Mobile entrepreneurfranchisees

PAY - S2M

Mutual fund

Intl money transfer

Note: OS2M = Subscriber to Merchant; S2F = Subscriber to Franchisee; S2S = Subscriberto Subscriber; F2S = Franchisee to Subscriber.

Source: © Anand Shrivastav.

Figure 4.1 The ecology of Beam transactions

Nevertheless, Suvidha believes its services will not only improve the lives ofits customers and the Beam Mobile Entrepreneurs, but will enable themhave a greater control over their life and destiny, and in this manner makeits humble contribution to the economic prosperity of India.

NOTES

1. Mahatma Gandhi (1958), Last Phase, vol. II, p. 65. From a note left behind in 1948.2. Census of India.3. Government of India-MOSPI (2004), “Socio-economic dimension of unemployment in

India”; Government of India-MOSPI (2005) “Employment and unemployment situ -ation in Cities and Towns in India 2004–05 Part 1”.

4. Sources include Telecom Regulatory Authority of India—TRAI June 2006;Government of India Department of Telecommunication (2006), “Telecom Vision2010” and Merrill Lynch (2006), “Global Mobile Matrix 2006”.

5. Sources include RBI (2005), “Internal Group Report of the RBI on issues relating torural credit and micro-finance 2005”; RBI Banking Statistics 2003; National SampleSurvey Organization (NSSO) (1999), “All India debt and investment survey 1991”;National Federation of State Cooperative Banks.

6. Sources include UNPAN/Ghosh, R. (???) “Field Survey”; Sinha, F. and team (2003),“Impact assessment of microfinance in India”, EDA Rural Systems Pvt Ltd, World Bank(2001), “Engendering development through gender equality in rights-resources and voice”;Sampark (2003), “Mid-term impact assessment study of CASHE project in Orissa”.

7. Suvidha (Sue – vee – dhaa) is a Sanskrit word that means convenience. Suvidha is amobile transaction service provider and is ISO 9001 certified. It was incorporated on 11


Microfinance

Micro-insurance

Micro-investment

Micro-payments

International money transfer

Source: © Anand Shrivastav.

Figure 4.2 Services to Beam subscribers

December 2002 and is headquartered in New Delhi. Its global transactions managementsystem (GTMS) is used by corporate banking for cash, tax management, and coopera-tive payment management services via the Internet. Some of the banks using GTMS ser-vices are HSBC, Deutsche Bank, ICICI Bank, HDFC Bank, and UTI Bank. Suvidha isabout to launch the Beam services in FY 2007–08, with SWIFT as its backend.

8. GSMA (2007), “Global money transfer uses pilot mobile to benefit migrant workers andthe unbanked”, press release, www.gsmworld.com/news/ press_2007/ press 07_ 14.shtml.

9. IFC Washington-GSMA-infoDev (2006), “Micro-payment systems”, infoDev report,www.infodev.org.

10. ZDNet-Portio Research (2005), “SMS traffic to double in AP by 2010”, September,www.zdnetasia.com.

84 Microfinance

5. How should governments regulatemicrofinance?Richard Rosenberg1

INTRODUCTION

Powerful new microfinance techniques are being developed that allowformal financial services to be delivered to low-income clients who havelong had no access to such services. But the microfinance industry will notreach its full potential unless many of its service providers can eventuallyenter the arena of licensed, prudentially supervised financial intermedi-aries. Regulations must eventually be crafted that allow this to happen.Dozens of developing and transition country governments are now atearlier or later stages of addressing this challenge.

Many different actors are pushing for regulatory adjustments, frommicrofinance institutions themselves (MFIs), to international developmentagencies, to government officials who want to democratize finance orprotect against perceived risks for the financial system (or perhaps clampdown on annoying non-governmental organizations—NGOs). The inter-ests and objectives of these actors diverge considerably. Thorny technicaland practical issues are involved. We do not have decades of experiencewith regulated microfinance to guide us—most of the countries withmicrofinance regulations have only a few years of experience with imple-menting them. And country-specific circumstances loom large, so there canbe no standard model for microfinance regulation.

Nevertheless, among people working on these topics there are sur-prisingly wide areas of consensus on some general principles that shouldbear on regulatory design for microfinance. The author is confidentthat most of the material in this chapter is consistent with the views ofmost of the technical advisors who have multi-country experience andwho do not represent the interests of any particular combatant in thefray.

The discussion will begin with an important definitional distinctionbetween “prudential” and “non-prudential” regulation. Non-prudentialregulation will be treated in the third section, prudential regulation in the

85

fourth, and the challenges of prudential supervision in the fifth. The sixthsection concludes.

PREAMBLE: PRUDENTIAL AND NON-PRUDENTIALREGULATION

Governments regulate the behavior of all businesses. Such regulation maybe aimed at protecting consumers, or employee safety, or the environment,but it usually does not try to protect the financial health of the business—that concern is generally left to the owners, at least where the owners areprivate. But in almost every country in the world, banks are treateddifferently. Governments impose an elaborate regime of “prudential” regu -lation whose aim is to protect the solvency of banks. Various reasons areadvanced for this. The main one is that financial systems depend criticallyon confidence, so that the failure of one bank can hurt many other banksand provoke a systemic crisis, with dire effects for the economy at large.Another reason is that banks are financed predominantly by money ofpeople other than the shareholders, which creates incentives for bank man-agers to take imprudent risks: when the gamble succeeds the bank and itsshareholders capture the gain, but when the gamble fails, others— especially depositors—may bear a large part of the loss. Finally, govern-ments believe that small, unsophisticated depositors need protectionbecause they are in no position to appraise the riskiness of a bank on theirown. Examples of prudential requirements include capital adequacy rules(how much of other people’s money a bank can use), restrictions on riskyuncollateralized lending, limits on insider lending, or requiring main -tenance of reserves for loans that are likely to go bad.

“Non-prudential” regulation is an inelegant name for all the otherbanking rules—the ones that don’t involve the government in assessing andprotecting the financial health of banks. Such rules are sometimes referredto as “conduct of business” regulation. Examples include limits on interestrates charged to borrowers, other consumer protection like truth-in-lendinglaws, or anti-money-laundering rules that require screening and reportingof customers. Securities regulations are another non-prudential example:these rules usually require banks, just like other firms, to disclose all mate-rial information about their business to potential investors. The rules havebeen complied with when all the information about the bank’s business isdisclosed. Investors are left to fend for themselves when it comes to weigh-ing their risks.

Implementing prudential regulation, where the government in somesense is vouching for the financial soundness of each licensed bank taking

86 Microfinance

deposits, tends to be much more complex, difficult, expensive, and intrusivethan implementing non-prudential regulation. Enforcing prudential regu-lation always requires a specialized banking authority, whereas many non-prudential regulations apply to non-deposit-taking firms as well, and mightnot necessarily require a specialized banking supervisor to enforce them.

The reason for emphasizing this distinction is that in a number of coun-tries, governments are applying burdensome prudential rules to non-deposit-taking MFIs whose failure would cause neither loss of depositors’funds nor material disruption of the national financial system. Prudentialregulation has high costs not only for the supervisory authority but also forthe supervised institution, which will eventually pass these costs along to itscustomers. Prudential regulation of credit-only MFIs uses a cannon—avery expensive cannon—where a rifle would be more than adequate in viewof the risks involved.

It is especially important to focus on the implications of regulation for theadministrative costs of MFIs. In the centuries-old effort to improve financialaccess for poor and low-income people, the critical factor is cost, more thanthe motivations of financial service providers. Major, long-lasting improve-ments in access are usually associated with new ways to lower costs. Concernfor the poor has played an important part in the microfinance revolution ofthe last three decades. But concern for the poor has been around for a longtime. The revolution became possible when Grameen Bank and other pion -eers in Indonesia and Latin America discovered less costly ways to deliverand collect tiny uncollateralized loans, and mobilize and manage smallsavings. Some of the regulatory requirements discussed in this chapter havesignificant cost implications for microfinance providers. Decisions aboutsuch practices need to be handled carefully.

NON-PRUDENTIAL ISSUES

Usury Limits

Lending a million (US) dollars in 10 000 loans of $100 each entails admin-istrative costs that are hugely greater than the cost of lending out the sameamount in one or two big loans. As a result, it is usually impossible to domicro-lending on a financially sustainable basis without charging interestrates that are very substantially higher than what banks charge to largerborrowers. In 2005, the median annual interest rate collected by the hun-dreds of MFIs reporting to the MIX Market (www.themix.org) was about31 percent. Rates above 50 percent are not uncommon, and a few MFIscharge more than 80 or even 100 percent. In most cases, these rates reflect

How should governments regulate microfinance? 87

not high profits but high costs of micro-lending: the smaller the loan size,the higher the administrative costs are for lending a given amount. But thisanalysis of lending costs is fine print that is usually too small to show up onthe screens of politicians or the general public. Charging poor borrowers30 percent when fat cats pay only 10 or 15 percent shocks most consciences.

Not all microfinance interest rates can be explained by the costs of lending.In a recent well-publicized instance, the Mexican MFI Compartamos wascharging interest of about 100 percent a year and producing annual profitsthat gave its shareholders more than a 50 percent return on equity. This causecélèbre, despite being a highly exceptional case, has fanned the flames of agrowing backlash against high micro-credit interest rates, a backlash that hasalready been underway in Latin America and the rest of the world for severalyears now.2

Limits on interest rates can hurt rather than help low-income borrowersif the interest cap is set too low for certain types of lending to be profitable:providers will withdraw from the business and potential borrowers will loseaccess to services. In theory, an interest rate cap would avoid this result if itwere set at just the right level. As a practical matter, however, finding theright level is hard, not least of all because loan products, clienteles, andcosts vary. Moreover, it is politically difficult for governments to set inter-est caps at appropriate levels: a reasonable interest rate for tiny, high-costmicro-loans will inevitably seem exploitative to most people, because theydo not understand the reasons for the high rates.

Consumer Protection and Borrowers’ Rights

When governments are concerned about microfinance interest rates thatsound abusive, they are sometimes advised to avoid interest rate caps andfocus instead on other borrower protection issues such as truth-in-lending(disclosure of the full cost of loans in a format that makes it easy tocompare rates offered by various lenders) or prohibition of certain unac-ceptable lending and collection practices.

Most MFIs today do not quote their loan charges in the form of anannualized effective interest rate that includes all costs. In some cases theremay be a legitimate concern that explicit quotation of rates this way wouldlead to a political backlash, resulting in interest rate caps that would makeit impossible to continue serving their clients. In most cases, though, micro-lenders’ opposition to truth-in-lending policies and requirements probablystems mainly from normal, less noble motives. Loan-cost disclosure maynot be a panacea, however: there are some indications that low-incomeclients have trouble understanding and using the information. This concernhas led to scattered efforts to provide financial education for consumers, but

88 Microfinance

most of these programs are still too young to be assessed, at least in devel-oping and transition countries.

Consumer protection regimes may also include restrictions on the wayloans can be made and collected. Obvious examples would include pro-hibiting the use of violence or other heavy intimidation to collect loans, butother less dramatic practices may also be deemed abusive. In South Africa,for instance, so-called “micro-credit” consists mainly of firms makinghigh-interest consumer loans to a clientele consisting mostly of salariedformal-sector employees.3 Taking possession of a borrower’s ATM card orrequiring delivery of a post-dated check for the loan amount were commonpractices, which were prohibited under a non-prudential regulatory regimecreated by the Micro Finance Regulatory Council (MFRC), a governmentagency lodged outside of the banking authority. MFRC rules were giventeeth by stipulating that any loans issued in violation of those rules wouldbe legally unenforceable.

The Bolivian microfinance sector suffered huge losses when profligateChilean-backed consumer lenders started marketing to unsalaried micro-entrepreneurs, passing out loans that bore no relation to the borrowers’repayment capacity. Many borrowers got in over their heads, and since alarge percentage of these were also borrowing from more responsibleMFIs, those sound MFIs were badly hurt by the ensuing wave of defaults,not to mention borrowers who lost their credit rating. The BolivianSuperintendency of Banks responded by requiring all uncollaterizedlending to include an assessment of repayment capacity.

A the time of this Bolivian crisis, the government’s credit referencebureau was not working well, so it was hard for lenders to know whether apotential borrower had loans outstanding from another source, or had ahistory of repayment problems. After the crisis, all of the actors foundthemselves considerably more enthusiastic about the credit bureau, andunlicensed lending-only MFIs were allowed to participate for the first time.As a general matter, credit reference bureaus are a powerful tool for extend-ing credit access to previously excluded groups, because the bureaussignificantly lower the costs of appraising borrower creditworthiness, andstrengthen borrowers’ motivation to repay. Credit bureaus make it possibleto lend to customers who would have been unprofitable otherwise.

Other consumer protection measures include privacy protection andaccessible dispute-resolution systems.

AML/CFT Regulation4

The Financial Action Task Force (FATF) is an international body that rec-ommends standards for national legislation on anti-money-laundering and


countering the financing of terrorism (AML/CFT). The FATF standardsdo allow room for adjustment to fit individual country circumstances, butdeveloping and transition countries are often nervous of straying far fromthe standards, because winding up on the list of non-complying nations canhave severe consequences. Among the FATF standards are know-your-client rules (ascertaining and documenting the customers’ true identitiesand addresses), heightened surveillance of transactions, preserving trans-action records, and reporting suspicious transactions to national authori-ties. Many banks complain loudly about the additional costs generated bythese requirements when dealing with their normal customers. In the worldof microfinance, where transactions and balances are much tinier, fullenforcement of regular AML/CFT standards would make it uneconomicto serve large groups of customers. The additional administrative costs areparticularly problematic when transactions are so small, and compliancecan be impractical for some kinds of clients. For instance, it is challengingto document identity and address for people who have no national identitycard, are illiterate, and have never seen any document that specifies wherethey live.

In its early years FATF was dominated by people coming from a law-enforcement perspective, who were not always instinctively sympatheticwith concerns about how FATF rules might exclude low-income clientsfrom services. More recently, this problem is getting more attention at inter-national and local levels. Taking a risk-based approach to AML/CFT, itwould not seem that transactions of, say, $30 or loan or savings accountsof $300 create substantial security risks. Governments should considersofter requirements, or waiving them altogether, for accounts and balancesbelow defined limits. After this was done in South Africa, for instance,banks were able to offer basic, no-frills transaction accounts that in a fewshort years reached 1.6 million customers, most of whom would haveremained unbanked if the AML/CFT rules had not been relaxed.

PRUDENTIAL ISSUES

Whether, When, and How to Open Prudential Licensing Regimes forMicrofinance

Regulation as promotionIn more than a few countries, the microfinance sector consists mainly ofweak non-governmental organizations that provide lending only, as well ascredit unions or similar savings and loan cooperatives, few of which arelarge and stable. When a government is confronted with this situation, and

90 Microfinance

wants to catalyze a large expansion of quality financial services for itslower-income population, one plausible-sounding response is to develop anew licensing window that allows institutions to become prudentially regu -lated, and take deposits, without facing minimum capital requirements ashigh as those required for a full banking license. (Deposit-taking is doublyimportant: not only does it provide a large funding source for expansion oflending, but it also gives the institution’s low-income clients access to asavings service that is often even more valuable to them than credit.)

In such countries, this “build it and they will come” approach is based onthe hope that the special licensing window will attract new private sectorentrants to the business, or encourage weak existing MFIs to tighten uptheir operations so as to meet the prudential standards for licensing. Thereis considerable controversy over this approach. It is premised on the beliefthat the binding constraint is absence of appropriate regulation, rather thanscarcity of competent retail operators. International experience to date hasbeen too limited to produce a general answer to the question; if anything,it suggests that the answer will vary from country to country.

For instance, Tanzania spent a great deal of time, effort, and money onthe development of a well-conceived licensing regime, but years afterwardthe results were disappointing. South Africa does not offer a micro-bankinglicense, but the government took steps a decade ago to make it possible formicrofinance institutions to offer small loans at relatively high interest rates.Despite this change, South Africa still does not have many solid institutionsoffering uncollateralized loans to unsalaried micro-entrepreneurs.

Pakistan has a huge unserved microfinance market, and it’s hard to findmany countries with as good a licensing regime for microfinance as the onePakistan enacted in 2001. But until very recently, institutions licensedunder this law contributed hardly at all to the growth of microfinance in thecountry, and the overall financial condition of licensed and unlicensedproviders was worse than it had been when the law was passed. Within thelast year, however, things are looking brighter: some of the newly licensed,privately owned microfinance banks are expanding aggressively, and theoverall financial performance of the sector is improving.

Regulation that follows, rather than leads, the marketIn most of the countries with effective prudential regimes in place today formicrofinance, the regulation came after, not before, the development of acritical mass of strong MFIs that were delivering loans on a sustainable basis.Bolivia has the longest and most solid record of successful microfinance regulation, and this experience is often held up as an example where a newlicensing window for microfinance was a powerful contributor to the successof the industry. But the Bolivian licensing regime was put in place only after


the country already had a number of strong NGO MFIs who had shown theycould manage their lending business stably and profitably. BancoSol, theleading MFI, did not use the microfinance licensing window—it got a regularcommercial banking license well before the specialized microfinance law waspassed. Most of the other MFIs who got licenses under the new law couldprobably have raised the money for a banking license if the easier andcheaper MFI license had not been available. In BancoSol’s first year ortwo, there was a certain amount of “supervision by winking,” as theSuperintendent waived application of some prudential rules that didn’t fitmicrofinance very well. But the time the microfinance law was passed andnew prudential norms formalized, the Superintendency already had experi-ence from supervising BancoSol. It’s possible to argue that Bolivianmicrofinance did not need a new specialized microfinance window to reachits present vital state, and that a few adjustments to the country’s bankinglaw and regulations would have created the necessary regulatory space.

When countries design a new licensing window for microfinance on theexpectation that licenses will go mainly to existing NGO MFIs during theearly years, the regulators sometimes don’t pay enough attention to the con-dition of those MFIs and their loan assets. In Zambia, for instance, theforeign aid agencies of the United States and Sweden financed developmentof a prudential regime in 1999 that would allow MFIs to take deposits. Butat that time, sources say, the country had few if any MFIs whose cost recov-ery and loan collection would make them safe custodians of customers’deposits. There may have been some expectation that donor-funded tech-nical assistance would turn the MFIs into strong institutions, but it is hardto find many examples of weakly managed MFIs that have been turned intovibrant, stable MFIs by outside technical assistance. This is not to suggestthat such assistance is useless: MFIs that already have strong managersmake good use of such support, but technical assistance can seldom turn aweak manager into a strong one. Political considerations prevented enact-ment of the Zambian law at the time. A set of microfinance regulations wasfinally put into force in 2006, but a review of the MIX Market database asof the beginning of 2007 shows only a single sustainable Zambian MFI,and that one had only 12 500 clients.5

Adjusting Prudential Norms to Fit Microfinance Products and Institutions

Some regulations common in traditional banking need to be altered toaccommodate microfinance. Whether microfinance is being developedthrough specialized stand-alone deposit-taking MFIs, or as a product linewithin retail banks or finance companies, the following norms will usuallyneed re-examination:

92 Microfinance

Minimum capitalThe kind of investors who are willing and able to finance MFIs may not beable to come up with the minimum capital required for a full bankinglicense, especially as minimum capital requirements trend upward aroundthe world. Setting a low minimum capital bar is often the central objectiveof those pushing for new licensing regimes for microfinance.

When banking authorities set minimum capital, bank safety and sound-ness may not be their primary concern. Rather, minimum capital is oftenused as a rationing device to manage the number of separate institutionsthat have to be supervised. The arguments for and against low minimumcapital for MFIs will be treated in the next section, which deals with super-visory challenges.

Capital adequacyUnder the Basel Accords, the relationship between shareholders’ equityand bank risk assets is the foundation of prudential regulation. Equity istreated as a cushion that protects depositors and other creditors of thebank: the more of its assets are funded by shareholders’ money, the higherthe losses the bank can sustain and still be able to repay its depositors.

There has been controversy about whether solvency (capital adequacy)requirements should be tighter for specialized MFIs than for banks. Ifwe want a level playing field in the financial sector, should microfinancebe penalized with tougher solvency requirements that lower shareholderprofitability?

Several theoretical arguments point in the direction of higher equity-to-risk-assets ratios for MFIs. In the first place, deposit-taking microfinance isa new business in most countries, which supervisors—and some MFI man-agers as well—do not have decades of experience with. Second, most MFIloan assets are not collateralized. Normally, MFI portfolio quality is verygood, but if an MFI starts to have problems with loan delinquency, theycan balloon out of control much faster than would be normal with collat-eralized loans. Third, administrative costs for MFIs are much higher thanfor commercial banks. When a significant part of the MFI’s loans are notbeing paid, the uncompensated administrative cost on those loans decapi-talizes the MFI faster than would be the case with a normal bank. All ofthese considerations suggest tighter solvency requirements, at least in theearly years.

Balanced against those theoretical arguments is the clear empirical factthat licensed MFIs suffer fewer loan losses than commercial banks do.There is also emerging evidence that licensed MFIs are more resilient thancommercial banks in times of financial or economic emergencies. In arecent banking crisis in Bolivia, all the commercial banks went insolvent


and MFIs came through in good shape. During the financial meltdown inIndonesia, repayment plummeted on the commercial loans of Bank RakyatIndonesia (BRI), while there was hardly a blip in the repayment of itsmicro-loans. When times are uncertain, low-income people are especiallyanxious to maintain their access to a credit facility, which can be a life-saverif an unexpected shock hits. BRI’s micro-borrowers understand that theonly way to keep access to a future loan if and when they need it is to faith-fully repay today’s loan.

Some microfinance is delivered through credit unions and other financialcooperatives. Application of capital adequacy norms to these institutionspresents a particular issue with respect to the definition of capital. All creditunion members have to invest a minimum amount of “share capital” intothe institution. But unlike an equity investment in a bank, share capital canusually be withdrawn whenever a member decides to leave the credit union.From the vantage of institutional safety, such capital is not very satisfac-tory: it is impermanent, and is most likely to be withdrawn at precisely thepoint where it would be most needed—when the credit union gets introuble. Capital built up from retained earnings, sometimes called “institu-tional capital,” is not subject to this problem. One approach to this issue isto limit members’ rights to withdraw share capital if the credit union’scapital adequacy falls to a dangerous level. A more conservative approach,now recommended by the World Council of Credit Unions, is to requirecredit unions to build up a certain level of institutional capital over a fewyears, after which time capital adequacy is based solely on those retainedearnings.

Unsecured lending limits and loan-loss provisionsThe experience in normal banking is that loans are more likely to defaultwhen they are not backed by collateral or guarantees. Thus, banking regu-lations often put tight limits on unsecured lending—for instance, cappingit at no more than 100 percent of the bank’s equity base. Such a limitationwould make a portfolio of uncollateralized micro-credit impossible, at leastin a specialized MFI. Some regulators have avoided the problem by treat-ing group guarantees as collateral. But not all micro-lenders use a groupmethodology, and group guarantees are less effective than is commonlysupposed. Many MFIs do not really enforce such guarantees, and loanlosses in MFIs that use such guarantees are not markedly lower than loanlosses in MFIs that do not.

A more straightforward approach is to recognize the empirical evidence.Worldwide, with relatively few exceptions, uncollateralized loans in acountry’s licensed MFIs suffer less delinquency and default than collater-alized loans in normal bank portfolios. The reasonable response to this

94 Microfinance

evidence is to put no collateral requirements on micro-loans, but instead toconcentrate on close supervision of the MFI’s lending systems and repay-ment history.

Banks in some places have been required to book a loan-loss provisionexpense to cover the full value of uncollateralized loans they make, evenbefore they become delinquent. Again, this is unworkable for micro-credit.Even if the provision expense is later reversed when the loan is collected,the accumulated charges for loans that are showing no problems wouldproduce a massive under-representation of the MFI’s real net worth. Andsuch a requirement has no empirical justification in the case of micro-credit. Thus, general provisions (provisions booked when the loan is made,and so not based on the presence of any repayment delays) should be nomore stringent for micro-credit than for normal portfolios.

The picture changes, however, once a micro-loan has actually fallen late.When one narrows the focus down to the micro-borrowers who do run intorepayment problems, experience shows that ultimate collection of theirloans is less likely than collection of collateralized loans that fall late by thesame amount of time. As a result, provisioning of already-delinquent loansneeds to be more aggressive for micro-credit than for normal collateralizedloans.

Loan file requirementsGiven the nature of microfinance borrowers, their informal businesses,and their loan sizes, it would be unreasonable to make micro-lenders gen-erate the same loan documentation that is required for normal bank loans.This is particularly true in the case of financial statements for the bor-rower’s business, evidence that the business is formally registered, or col-lateral documentation. On the other hand, micro-loan files should alwayscontain at least the loan contract, a record of the borrower’s repaymenthistory on prior or concurrent obligations, and a simple estimate of theborrower’s income, expenses, and repayment capacity, at least wherethe MFI’s methodology relies on loan officers rather than fellow groupmembers to determine repayment capacity. However, MFIs that makerepeated short-term loans, for instance every three months, should not berequired to do a fresh analysis of borrower cash flow before every singleloan.

Restrictions on co-signers as borrowersRegulations sometimes prohibit a bank from lending to someone whohas co-signed or otherwise guaranteed a loan from that same bank. Suchrules would need to be waived for MFIs that do group lending with cross- guarantees among the group.


Insider lendingLoans made to owners, directors, or managers of a bank are not likely toreceive the same objective scrutiny as loans to unrelated parties. In recog-nition of this fact, most banking authorities now restrict insider lending tosome limited percentage of the bank’s assets or equity. This author’s viewis that insider lending should be completely prohibited in licensed MFIs,with the exception of small welfare loan programs for employees. Whenspecialized MFIs are receiving favorable regulatory treatment for the solereason that they are extending financial access to low-income customers, itis hard to see any reason or need for insider lending.

Frequency and content of reportingBanks may be required to report their financial position frequently, evendaily. In many settings, the undeveloped state of transportation and com-munication infrastructure may make this difficult or impossible for ruralbanks or branches. In addition, frequent or voluminous reporting to thebanking supervisor can add substantially to the administrative costs of anintermediary, especially one that specializes in very small transactions. Thechief financial officer of BancoSol once estimated that compliance with thebanking supervisor’s reporting requirements cost the bank 5 percent of itsassets the first year, and 1 percent or more a year thereafter. On the other,effective supervision is impossible without adequate reporting. Specializedmicrofinance banks or branches usually present a less complex set of risksthan normal banking, so it should be possible to supervise them well basedon reporting that is somewhat less burdensome and expensive than whatconventional banks have to provide.

Physical security and branching requirementsBanks’ hours of business, location of branches, and security requirementsare often strictly regulated in ways that could impede service to amicrofinance clientele. For instance, the convenience of clients who arerunning their micro-businesses all day might require operations outsidenormal business hours, or cost considerations might require that staffrotate among branches that are open only one or two days a week. Securityrequirements such as guards or expensive vaults, or other normal infra-structure rules, could make it too costly to open small-volume branchesin poor areas. Branching and physical security requirements merit re- examination—but not necessarily elimination—in the microfinance con -text. Clients’ needs for financial services have to be balanced against thesecurity risks inherent in managing cash.

96 Microfinance

Ownership requirementsSome countries have ownership-diversification rules that prohibit anysingle party from controlling more than 20 percent (for instance) of abank’s shares. Also, NGOs may not be eligible to own bank shares. Both ofthese rules serve legitimate prudential objectives, but they can cause seriousdifficulty in the common case where the assets of a newly licensed MFIcome almost exclusively from an NGO that has built up the business overa number of years. In recent years, commercial and quasi-commercialinvestors are showing greater interest in buying shares of newly licensedMFIs, but there are many transforming MFIs for whom attracting suchinvestment is not a practical option, at least not at the time that they convertto licensed form. If the original NGO has to find new owners who will pur-chase 80 or 100 percent of the business, transformation into licensed formcould be delayed a long time. Some banking supervisors are allowing theoriginal NGO to own most or all of the shares of a newly licensed MFI,with a requirement that the ownership structure has to be brought into linewith normal banking rules over a reasonable period of years.

Deposit insuranceIn order to protect smaller depositors and reduce the likelihood of runs onbanks, many countries provide explicit insurance of bank deposits up tosome size limit. Some other countries provide de facto reimbursement ofbank depositors’ losses even in the absence of an explicit legal commitmentto do so. There is considerable debate about whether public deposit insu -rance is effective in improving bank stability, whether it encourages inap-propriate risk-taking on the part of bank managers, and whether suchinsurance would be better provided through private markets. In any event,if deposits in commercial banks are insured, the presumption probablyought to be that deposits in other institutions prudentially licensed by thefinancial authorities should also be insured, absent strong reasons to thecontrary.

Branchless bankingIn a growing number of developing and transition countries, financial ser-vices are being provided outside of conventional bank branches. The use ofautomated teller machines (ATMs) has been spreading for years. Morerecently, payment, transfer, and savings services are being offered throughpost offices or retail outlets like groceries, pharmacies, or gas stations. Suchservices may be used mainly by the middle class, but they hold promise forpoor people as well, especially the rural poor. Using such “retail agents,”banks can reach places where building and staffing a branch wouldbe unprofitable because of remoteness, low client density, or low client


transaction sizes. In addition, mobile phone operators in countries like thePhilippines, South Africa, and Kenya are exploiting their networks toprovide fast and convenient payment and transfer services to their sub-scribers, who include increasing numbers of low-income people.

Some branchless banking is bank-led: all of the clients’ transactions arewith a licensed commercial bank, and the retail agents or mobile phoneoperators are acting as third-party agents to handle cash on the bank’sbehalf. The bank remains responsible for any cash received. These arrange-ments raise some regulatory risk, including security of cash-handling andproper training of agents, but in general they do not add materially to therisks that are present in normal branch-based banking.

More of a regulatory challenge is presented by non-bank-led models,where the client’s cash is taken and held by a company like a mobile phoneoperator that is not licensed and prudentially supervised by the bankingauthorities. When such companies are holding significant amounts of cus-tomers’ cash, should they be required to meet the same prudential stan-dards as banks? South Africa’s answer to that question is a conservativeone: any mobile phone operator that wants to provide “e-money” servicesis required to partner with (that is, operate under the license of) a commer-cial bank. This raises the mobile operator’s costs considerably, and thesecosts must eventually be passed along to the client.

The Philippines is starting with a more liberal approach, allowing mobileoperators to operate independent of a banking license. However, theamount of such transactions and the size of outstanding balances owed anindividual customer are capped at low levels. So far this arrangement hasnot created significant problems, but the central bank is moving to tightenrestrictions further.

It is not yet possible to identify best practices in dealing with this issue:so far, the European Community has not been able to agree on a commonapproach.

SUPERVISORY ISSUES

The Burden of Supervising Small Intermediaries

For bank supervisors in many developing countries (though certainly notall), the central fact of life is responsibility for supervising a commercialbanking system with severe structural problems, often including somesizable banks teetering dangerously close to the edge of safety. The collapseof one—or a half-dozen—of these banks could threaten the country’sfinancial system with implosion. In trying to manage bank risk, supervisors

98 Microfinance

may have to work in a political minefield, because the owners of banks areseldom under-represented in the political process. The supervisors’ legalauthority to enforce compliance or manage orderly clean-ups is often inad-equate. They may not have enough control over the tenure, qualification,and pay of their staff. Monitoring healthy banks is challenging enough, butthe real problems come when it is time to deal with institutions in trouble.When a sick bank finally crumbles, its president can start sleeping again(though perhaps in a different country), while it is the supervisor who hasto stay awake at night worrying.

If bank supervisors sometimes display resistance to adding MFIs—mostly small, mostly new, mostly weak on profitability—to their basket ofresponsibilities, we should recognize that their reasons may be nobler thannarrow-mindedness or lack of concern for the poor.

The Philippines licenses hundreds of small intermediaries as “ruralbanks.” Originally, the minimum capital requirement for a rural bank wasvery low. These banks are not microfinance institutions, but their opera-tions do include credit and deposit services for lower-income clients. Theyhave access to the national payments system and are supervised by thecentral bank. As of September 1997, 824 rural banks were serving half amillion clients. These banks had only about 2 percent of the bankingsystem’s assets and deposits, but they made up 83 percent of the institutionsthe central bank had to supervise.

Supervising the rural banks severely stretched the resources of thecentral bank’s supervision department, tying up as much as one-half of itstotal staff and budgetary resources at times. In the early 1990s one in everyfive rural banks had to be shut down, and many others had to be mergedor otherwise restructured. An unpublished 1996 analysis reported thatabout 200 inspectors were assigned to the rural banks, but even this level ofresources was viewed as inadequate. Each on-site examination consumedup to three person-weeks or more. At one point the supervisory departmentfound that this burden, combined with its budget limitations, was severelyendangering its ability to function.

One of the responses to the crisis was to raise the minimum capital sub-stantially. But a knowledgeable observer has guesstimated that as of late2007, perhaps half of the rural banks are still technically insolvent, thoughthese tend to be the smaller ones. The larger rural banks have most of theassets and customers are said to be strong and expanding aggressively.

The occurrence of a supervisory meltdown doesn’t necessarily mean thatlicensing rural banks has been a failure in the Philippines. Hundreds ofthousands of people are still getting services that would otherwise havebeen unavailable to them. But the experience, and similar ones in Indonesiaand Ghana, teaches us to be realistic about the difficulty of supervising


large numbers of small new financial institutions. Some would argue thatineffective supervision is worse than no supervision at all, because it mis-leads depositors and tarnishes the credibility of the banking authorities.

Minimum Capital as a Rationing Device

When regulators set minimum capital requirements for licensing MFIs, amajor consideration should be limiting new licenses to a number that isconsistent with available supervisory resources. Obviously, this has to bebalanced against the objective of opening access to financial markets, anobjective that tends in the direction of keeping minimum capital as low aspossible. There is a strong argument to be made that regulators should startwith relatively high minimum capital for a new licensing window, and grad-ually relax the requirement after there has been more supervisory experi-ence, and supervisors are better able to judge what they can take on.

A country does not necessarily need large numbers of licensed MFIs inorder to serve its market well. In most countries, a few large MFIs accountfor the vast majority of the outreach. In 2000, Bangladesh probably had over1000 MFIs, but the largest ten served all but about 15 percent of the clients.

Small Community-based Intermediaries

Smaller institutions may not require as much supervision as big ones, butthere are lower limits to how far supervision can be watered down. At somepoint, “supervision lite” is no longer effective, if effectiveness means thatthe supervisor can expect to flag most problems before they have gotten tooserious to fix.

Some member-owned intermediaries take deposits but are so small, andsometimes so geographically remote, that they cannot be supervised on anycost-effective basis. This poses a practical problem for the regulator. Shouldthese institutions be allowed to operate without prudential supervision, orshould minimum capital or other requirements be enforced against them sothat they have to cease taking deposits? Sometimes regulators are inclinedto the latter course. They argue that institutions that cannot be supervisedare not safe, and therefore should not be allowed to take small depositors’savings. After all, are not small and poor customers just as entitled to safetyas large and better-off customers?

But this analysis is too simple if it does not consider the actual alterna-tives available to the depositor. Poor people can and do save. If formaldeposit accounts are not available, they have to fall back on savings toolslike currency under the mattress, livestock, building materials, or informalarrangements like rotating savings and credit clubs. All of these vehicles are

100 Microfinance

risky, and in many if not most cases, they are more risky than a formalaccount in a small unsupervised intermediary. Closing down the localsavings and loan cooperative may in fact raise, not lower, the risk faced bylocal savers by forcing them back to less satisfactory forms of savings.Because of these considerations, most regulators facing the issue havechosen to exempt community-based intermediaries below a certain sizefrom requirements for prudential regulation and supervision. The sizelimits are determined by number of members, amount of assets, or both.(Sometimes the exemption is available only to “closed bond” institutionswhose services are available only to members of a pre-existing group suchas employees of a company.) Once the size limits are exceeded, the institu-tion must comply with prudential regulation and be supervised. If smallintermediaries are allowed to take deposits without prudential supervision,a good argument can be made that their customers should be clearlyadvised that no government agency is monitoring the health of the institu-tion, and thus that they need to form their own conclusions based mainlyon their knowledge of the individuals running the institution.

Supervisory Tools and their Limitations

Some standard tools for examining banks’ loan portfolios are ineffective formicro-credit. As noted earlier, loan-file documentation is a weak indicatorof micro-credit risk. In a commercial bank, one can often capture most ofthe portfolio risk by examining a small number of large loans, but this isnot true in a micro-credit portfolio consisting of thousands of tiny loans.Sending out confirmation letters to verify account balances is usuallyimpractical for micro-credit, especially where client literacy is low. Instead,the examiner must rely more on analysis of the institution’s lending systemsand their historical performance. Analysis of these systems requires knowl-edge of microfinance methods and operations, and drawing practical con-clusions from such analysis calls for experienced interpretation andjudgment. Supervisory staff are unlikely to monitor MFIs effectively unlessthey are trained and to some extent specialized.

When an MFI gets in trouble and the supervisor issues a capital call,many MFI owners are not well-positioned to respond to it. NGOs who ownshares may not have enough liquid capital available. Development agenciesand development-oriented investors usually have plenty of money, but theirinternal procedures for disbursing it sometimes take so long that a timelyresponse to a capital call is impractical. Thus, when a problem surfaces ina supervised MFI the supervisor may not be able to get it solved by a timelyinjection of new capital, as the Colombian banking supervisor found outwhen the MFI FinanSol ran into trouble.


Another common tool that supervisors use to deal with a bank in troubleis the stop-lending order, which prevents the bank from taking on furthercredit risk until its problems have been sorted out. A commercial bank’sloans are usually collateralized, and most of the bank’s customers do notnecessarily expect an automatic follow-on loan when they pay off their exist-ing loan. Therefore, a commercial bank may be able to stop new lending fora period without destroying its ability to collect its existing loans. The sameis not true of most MFIs. Immediate follow-on loans are the norm for mostmicro-credit. If an MFI stops issuing repeat loans for very long, customerslose their primary incentive to repay, which is their confidence that they willhave timely access to future loans when they want them. When an MFI stopsnew lending, many of its existing borrowers will usually stop repaying. Thismakes the stop-lending order a weapon too powerful to use, at least if thereis any hope of salvaging the MFI’s portfolio.

A typical MFI’s close relationship with its clients may mean that loanassets have little value in the hands of a different management team.Therefore, a supervisor’s option of encouraging the transfer of loan assetsto a stronger institution may not be as effective as in the case of collateral-ized commercial bank loans.

The fact that some key supervisory tools do not work very well formicrofinance certainly does not mean that MFIs cannot be supervised.However, regulators should weigh this fact when they decide how many newlicenses to issue, and how conservative to be in setting capital standards orrequired levels of past performance for transforming MFIs.

Where to Locate Microfinance Supervision

Given the problem of budgeting scarce supervisory resources, alternativesto the conventional supervisory mechanisms used for commercial banksare frequently proposed for depository MFIs.

Within the existing supervisory authority?The default option for MFI supervision would normally be the supervisoryauthority responsible for commercial banks. Using this agency to supervisemicrofinance takes advantage of existing skills and lowers the incentive forregulatory arbitrage. If this option is chosen, the next question is whetherto create a separate department of that agency. The answer will vary fromcountry to country, but at a minimum, a specially trained supervisory staffis needed, given the differing risk characteristics and supervisory tech-niques in the case of MFIs and microfinance portfolios.

The location of microfinance supervision becomes a more compli -cated question when both non-depository micro-lending institutions and

102 Microfinance

depository MFIs are to be addressed within a single, comprehensive regula-tory scheme. The tasks involved in issuing permits to non-depository micro-lending institutions have relatively little to do with the prudential regulationand supervision of licensed depository institutions. In some contexts,lodging both of these disparate functions within the same regulatory bodymight be justified on pragmatic grounds—such as the absence of any otherappropriate body, or the likelihood that the permit-issuing function wouldbe more susceptible to political manipulation and abuse if carried out byanother body. In other cases, non-depository MFIs are required to report tothe banking supervisor in order to make it easier for them to move eventu-ally into more services and more demanding prudential regulation. Often,however, the risks of consolidating prudential and non-prudential regulationof microfinance within the banking authority will outweigh the benefits.These risks include the possibility of confusion on the part of supervisors asto the appropriate treatment of non-depository institutions, and the possi-bility that the public will see the supervisory authority as vouching for thefinancial health of the non-depository institutions, even though it is not (andshould not be) monitoring the health of these institutions closely.

A separate and independent agency?In some countries the banking authorities’ reluctance to take responsibilityfor microfinance leads to plans or decisions to lodge microfinance super -vision in an independent agency. Building skills and experience in a differentbody can be time-consuming, and the new MFI supervisor may not be aspolitically independent as the banking supervisor. One approach to short-ening the learning curve is to entrust prudential supervision of deposit-taking MFIs to an apex agency that is already making wholesale loans toMFIs. This structure can present conflicts of interest. If such an apex super-visor has large loans outstanding to a troubled MFI, will the agency betempted to drag its feet when depositors’ interests are best served by shut-ting the MFI down? On the other hand, central banks frequently deal withsimilar conflicts of interest.

In developing countries, regulation and supervision of credit unions hasusually been lodged outside the banking authority, often in the governmentdepartment that is responsible for cooperatives of all sorts. The experiencewith this arrangement has usually been very disappointing.

Delegated supervision?Sometimes the government financial supervisor delegates part or all of thetasks of direct supervision to an outside body, while monitoring and con-trolling that body’s work. This seems to have worked, for a time at least, insome cases where the government financial supervisor closely monitored


the quality of the delegated supervisor’s oversight, although it is not clearthat this model reduces total supervision costs. Where this model is beingconsidered, it is important to have clear answers to three questions: (1) whowill pay the substantial costs of the delegated supervision and the govern-ment supervisor’s oversight of it? (2) if the delegated supervisor provesunreliable and its delegated authority must be withdrawn, is there a realis-tic fallback option available to the government supervisor? and (3) whena supervised institution fails, which body will have the authority andresources to clean up the situation by intervention, liquidation, or merger?

Because many MFIs are relatively small, there is some temptation tothink that their supervision, or at least on-site inspection, can be safely del-egated to external audit firms. Unfortunately, experience has been thatexternal audits of MFIs, even by internationally affiliated audit firms, veryseldom include testing that is adequate to provide a reasonable assuranceas to the soundness of the MFI’s loan assets, which is by far the largest riskarea for micro-lenders. If reliance is to be placed on auditors, the super visormust require microfinance-specific audit protocols that are more effective,and more expensive, than the ones now in general use, and must regularlytest the auditors’ work.

Self-regulation and supervisionWhen regulators decide that it is not cost-effective for the banking author-ity to provide direct oversight of large numbers of MFIs, self-regulation issometimes suggested as an alternative. Discussion of self-regulation tendsto be confused because people use the term to mean different things. In thisdiscussion, “self-regulation” refers to regulation (and/or supervision) bysome body that is effectively controlled by the regulated entities.

This is one point on which historical evidence seems clear. Self- regulationof non-bank financial intermediaries in developing countries has been triedmany times, and has virtually never been effective in protecting the sound-ness of the regulated organizations. One cannot assert that effective self-regulation in these settings is impossible in principle, but it can be assertedthat such self-regulation is almost always an unwise gamble against verylong odds, at least if it is expected that the regulation and supervision actu-ally enforce financial discipline and prudent risk management. Sometimesregulators have required certain small intermediaries to be self-regulated,not because they expect the regulation and supervision to be effective, butbecause this is politically more palatable than saying that these deposit-takers will be unsupervised. This can be a sensible accommodation in somesettings. While self-regulation probably will not keep financial intermedi-aries healthy, it may have some benefits in getting institutions to begin areporting process, or in articulating basic standards of good practice.

104 Microfinance

AFTERWORD: DOES PRUDENTIAL REGULATIONWORK?

There has long been a respectable minority of academic opinion that is skep-tical about the conventional wisdom that government prudential regulationand supervision are effective in limiting bank failures and financial systemcrises. Barth, Caprio and Levine (2006) conducted a cross-country analysisbased on a 150-country database, and drew some jarring conclusions:

In terms of what works best, our analyses raise a cautionary flag regarding thefoundations of current international best practice recommendations. In par -ticular, our results question the efficacy of Basel II’s first two pillars on capitalregulations and official supervision . . .

Across the different statistical approaches, we find that empowering directofficial supervision of banks and strengthening capital standards do not boostbank development, improve bank efficiency, reduce corruption in lending, orlower banking system fragility. Indeed, the evidence suggests that fortifyingofficial supervisory oversight and disciplinary powers actually impedes theefficient operation of banks, increases corruption in lending, and therefore hurtsthe effectiveness of capital allocation without any corresponding improvementin bank stability . . .

In contrast to these findings on capital regulations and direct supervisory over-sight of banks . . . supervisory and regulatory policies that facilitate privatesector monitoring of banks improve bank operations, which endorses Basel II’sthird pillar on market discipline. One mechanism for fostering private monitor-ing of banks is by requiring the disclosure of reliable, comprehensive, and timelyinformation. Countries that enact and implement these pro-private monitoringregulations enjoy more efficient banks and suffer from less corruption in lending.Furthermore, laws that strengthen the rights of private investors enhancethe corporate governance of banks. In contract, policies (for example, depositinsurance) that weaken market monitoring of banks tend to have adverseramifications on the banking system [including reducing system stability ratherthan strengthening it] . . .

We recognize, of course, that many countries do not have the legal and politicalinstitutions necessary to support effective market monitoring of banks. Con -sequently, many readers may conclude that a practical approach involvesempowering official supervisors until countries develop the institutional foun-dations for market monitoring. The cross-country results thus far, however, donot support this conclusion. The results instead indicate that regulatory restric-tions on bank activities, impediments to the entry of new banks, governmentownership of banks, and reliance on powerful official supervisors to overseebanks have adverse effects on the operation of banks. Moreover, it is exactly incountries with weak political and legal institutions that empowering officialsupervisors is likely to be most detrimental. (Barth, Caprio and Levine, 2006,pp. 10–16)


Other econometricians disagree about interpretation of the data, andthis author has neither the competence nor the courage to offer any judg-ment about the issue. In any event, hardly any of the world’s countries havebeen willing to take the advice of those who favor eliminating official super-vision of banks. Whatever the merits of the issue, it would be politicallyimpossible for most governments to back away from protecting depositors.

Then why raise the question in a discussion about regulating micro -finance? The reason is that in crafting a regulatory regime for microfinance,many decisions about detail and degree depend on one’s underlyingassumptions about the appropriateness and effectiveness of governmentprudential oversight. To the extent that one has doubts about such over-sight, or is concerned that regulatory power put into the hands of a gov-ernment supervisor may be used for personal rather than public benefit,then one might be more cautious in deciding particular issues related to thescope of such power.

NOTES

1. Author’s note: this chapter draws heavily on a set of consensus guidelines developed inconsultation with a wide range of experienced regulators, policy advisors, and industryanalysts (Christen, Lyman and Rosenberg, 2003) as well as the working experiences of theauthor and many colleagues. Citation of sources will be limited. The discussion is meantto apply to developing and transition economies only: the dynamics of microfinance inrich countries, and the regulatory issues they present, can be substantially different.

2. It is important to note that if Compartamos priced its loans to produce no profit at all, itwould still have to charge interest of about 77 percent, which would shock most con-sciences despite being completely driven by costs of lending. Compartamos’s loans areextremely small in relation to the Mexican context, so its administrative costs alone arevery high (Rosenberg, 2007).

3. In most countries, “micro-credit” is understood to refer mainly to small uncollateralizedloans to people who have informal micro-businesses rather than jobs in the formal sector.

4. This section draws heavily on Isern, Porteus, Hernandez-Coss and Egwuagu (2005).5. www.mixmarket.org/en/demand/demand.show.profile.asp?ett=1784&.

REFERENCES

Barth, J.R., G. Caprio and R. Levine (2006), Rethinking Bank Regulation: TillAngels Govern, New York: Cambridge University Press.

Christen, R.P., T.R. Lyman and R. Rosenberg (2003), “Guiding principles on regulation and supervision of microfinance”, CGAP, accessed at www.cgap.org/portal/binary/com.epicentric.contentmanagement.servlet.ContentDeliveryServlet/Documents/Guideline_RegSup.pdf.

Isern, J., D. Porteus, R. Hernandez-Coss, and C. Egwuagu (2005), “AML/CFT regulation: implications for financial service providers that serve low-incomepeople”, CGAP Focus Note No. 29, accessed at. www.cgap.org/portal/binary/

106 Microfinance

com.epicentric.contentmanagement.servlet.ContentDeliveryServlet/Publications/html_pubs/FocusNote_29.html.

Rosenberg, R. (2007), “CGAP reflections on the Compartamos initial publicoffering: a case study on microfinance interest rates and profits”, CGAP,Focus Note No. 42, accessed at www.cgap.org/portal/binary/com.epicentric.con-tentmanagement.servlet.ContentDeliveryServlet/Documents/FocusNote_42.pdf.


6. Gender empowerment inmicrofinanceBeatriz Armendáriz1 and Nigel Roome

INTRODUCTION

Ever since microfinance was popularized in the mid-1970s in Bangladesh,one of its salient features has been the overwhelming representation ofwomen. The trend has increased steadily, particularly during the 1980s.According to 2006 Microcredit Summit Campaign report, seven out often microfinance clients are women.2 Millions of these women are marriedor live with a partner, and many have children. Relative to initial lendingpractices by the Grameen Bank in Bangladesh, the bias in favor of loansto women in microfinance has been accompanied by an increasing trendto exclude men from microfinance services, particularly at very lowincome levels. The practice of exclusion might however prove to be coun-terproductive, for it can generate frictions within households, as men feelincreasingly threatened in their role as primary breadwinners within thehousehold.3

In this chapter we argue that the promotion of women in microfinanceinitiatives and the bias against men is taking place in the absence of solidempirical evidence on the effects of this strategy, and on the balance ofpower in households and on the health, education, and well-being of allhousehold members, which we hold to be key aspects of development. Wefurther argue that this issue deserves research given the possibility ofunforeseen outcomes and adverse consequences that run counter to thegoal of microfinance initiatives to promote development.

To clarify the central issues, on the one hand, higher household incomein the hands of women might increase health and education for women andtheir household members—we call this the women-empowerment effect.On the other hand, the exclusion of men from access to subsidized financemight create frictions, and rebound effects, that diminish the supportiverole women play for their spouses and wider household members in the pro-duction of health and education—we call this the women-disempoweringeffect. In the event that the latter effect dominates over the former, then

108

subsidized microfinance for women might have no impact, or even worse, anegative impact on health and education. An even more challenging ques-tion is what social and institutional conditions most strongly influencethese empowerment and disempowerment effects and their outcomes.

This chapter is structured as follows. First, it provides an overview ofwhat we currently know about microfinance, gender, health, and educationin the context of Bangladesh, where most research has been conducted.Second, some anecdotal evidence from Bangladesh and Africa on thenotion of microfinance empowerment is presented and discussed. Thisraises questions about the structures on the enhanced capacity of womento assert their role as the main providers of health and education, mainlyarising from the fact that the empowerment of women generates frictionswith their partners, which in turn leads to a potential disempowermenteffects. Third, anecdotal evidence from Chiapas in southern Mexico is out-lined, which provided the basis for empirical research on new approachesto microfinance now being undertaken in the region. Fourth, the chapteroutlines this experimental intervention in southern Mexico, where thewomen borrowers in a microfinance initiative invited their spouses to bepart of women-only solidarity groups as borrowers, in order to see whetherpotential frictions could be eliminated as a way to enhance women empow-erment and provide for better access to health and education at the house-hold level. The main challenges of implementing this type of intervention,which were revealed through this empirical study so far are described. Thefinal section spells out some concluding remarks.

CURRENT KNOWLEDGE OF MICROFINANCE,FINANCIAL RESOURCES, AND GENDER AS A BASISFOR THE PRO-WOMEN BIAS4

The most influential empirical study on microfinance and gender can befound in an article published by the Journal of Political Economy in 1998by Mark Pitt and Shahidur Khandker. In their study, Pitt and Khandkerdevelop a framework for estimating the impact of microfinance using cross-section data from Bangladesh for 1991–92. The paper pins down thepotential sources of bias, in identifying and estimating the impact ofmicrofinance initiatives alone on outcomes such as household expenditureson health and education.

For example, Pitt and Khandker addressed the bias that might arisebecause the individuals who self-select into microfinance programs may bethe least poor and most entrepreneurial members of their community. Thisbias would lead to an overestimate of the overall potential of microfinance

Gender empowerment in microfinance 109

on poverty reduction as and when microfinance projects were expanded toinclude poorer and less entrepreneurial individuals. Pitt and Khandkerfaced the well-known endogeneity problem that entrepreneurial individ -uals may deliver for themselves better incomes, which then enable them toqualify for further loans, which in turn increase their incomes. Typical waysto resolve this problem of estimating is through the use of an independentvariable, which correlates with entrepreneurial activity but not with out-comes. Pitt and Khandker used land-ownership as an instrumental vari-able: to qualify for a microfinance loan, individuals (both men and women)had to be poor as proxied by their holdings of land not being more thanhalf an acre. They used in this instrumental variable in studies that compared villages with microfinance opportunities and control villageswithout. This approach meant that those who received loans in treatmentvillages did so because they were landless poor, with the same entrepre-neurial abilities as those in the control villages where the landless poordid not access microfinance loans because there were not that manymicrofinance providers. Once village characteristics were controlled for, Pittand Khandker extended their analysis to the role of gender.

In particular, taking advantage of the fact that Bangladeshi microfinanceenterprises at the time were lending to men as well as women, Pitt andKhandker conducted a study to assess the relative impact of lending to menhead-of-households compared with their women counterparts on develop-ment outcomes in terms of health and education. Their results regarding rel-ative provision for health and education by both types of household headswere, at best, unclear. However, the main findings of the study of expendi-tures by both types of households in relation to household expenditures (par-ticularly on food and tools for expanding their respective businesses) are wellknown. These results have been highly influential, particularly in shaping thetrend in focusing donors’ aid through subsidized loans toward women.5

In particular, Pitt and Khandker showed that when a loan of 100 takawas extended to men it translated into 11 taka going into household expen-ditures (for food/nutrition/working tools), while the same amount lent towomen household heads led to 18 taka being spent on household expendi-tures (for food/nutrition/working tools).

While it would be too bold to claim that the findings of Pitt andKhandker alone have influenced the bias to women in recent microfinanceinitiatives. It is our conjecture that in the absence of any countervailingempirical evidence, Pitt and Khandker’s findings contributed to the normsand operational practices of CGAP (Consultative Group to Assist thePoor) World Bank, as well as many other multilateral organizationsengaged in providing subsidized microfinance. Their priority has been todirect subsidized loans to women.

110 Microfinance

The common practice in favor of subsidized micro-loans to women alsoflows from the research on the practice of delivering aid to women. Forexample, food stamps in the United Kingdom and Sri Lanka, and staplefood and cash deliveries under the PROGRESA (now called OPORTU-NIDADES) program in Mexico were directed to women rather than theirspouses. This was done for fear that if such aid was given to men, theymight sell the food stamps and mis-spend the resources, possibly wastingmoney on gambling, tobacco, and alcohol.

There are a number of empirical studies on the practice of targeting aidto women. Emmanuel Skoufias (2001) reports that the OPORTU-NIDADES project aimed at women in rural Mexico led to sharp socialimprovements: poverty decreased by 10 percent, school enrollmentincreased by 4 percent, food expenditures increased by 11 percent, andadults’ health (as measured by the number of unproductive days due toillness) also improved considerably.6

Duncan Thomas (1990) reports that child health in Brazil (as measuredby survival probabilities, height-for-age, and weight-for-height) along withhousehold nutrient intakes, tended to rise if additional non-labor incomewas in the hands of women rather than men. He observed that income in thehands of a mother had, on average, 20 times the impact of the same incomein the hands of a father with respect to children’s survival probabilities. In asubsequent study, also on Brazil, Thomas (1994) reports that increasing thebargaining power of women is associated with increases in the share of thehousehold budget spent on health, education, and housing as well asimprovements in child health. Patrice Engle (1993) similarly studies the rela-tionship between a mother’s and father’s income on child nutritional status(height-for-age, weight-for-age, and weight-for-height) for hundreds ofhouseholds in Guatemala, and reports that children’s welfare improves aswomen’s earning power increases relative to their husbands’. Paul Schultz(1990) finds that in Thailand non-labor income in the hands of women tendsto reduce fertility more than non-labor income possessed by men. He alsofinds that the impact of non-labor income has different effects on laborsupply, depending on which household member controls that income.7

Anderson and Baland’s (2002)’s article on Rotating Savings and CreditAssociations (ROSCAs) reports on a survey of hundreds of women inKenya. An overwhelming majority of the women responded that the prin-cipal objective for joining a ROSCA was to save money, and nearly all ofthe respondents were married. Anderson and Baland conclude that animportant motive for women joining ROSCAs is thus the desire to keepmoney away from their husbands. Other studies, not necessarily confinedto ROSCAs, suggest that savings motives (and the protection of assets) alsoapply to women’s involvement in microfinance institutions.


Christopher Udry’s (1996) research on agricultural practices in BurkinaFaso provides evidence on the ways men and women invest in agriculture.Using panel data, controlled for soil quality and other variables, he findsthat agricultural productivity is higher in plots cultivated by men thanwomen. He also finds that compared with plots cultivated by women, thehigher yields of plots cultivated by men are due to a greater intensity ofproductive inputs (including fertilizer and child labor). He thus concludesthat productivity differentials are attributed to the intensity of productionbetween plots cultivated by men and women and not to inherent skilldifferentials: an outcome he regards as inefficient since there are sharplydiminishing returns to the use of fertilizer. Not only are resources notfully shared, they are allocated in ways that diminish total householdincome. Udry suggests that by reallocating inputs to plots cultivatedby women can thus enhance efficiency. Another solution (that is, themicrofinance solution) is to provide women with credit sufficient to pur-chase additional inputs. A second way that microfinance can potentiallyaddress problems like this is by tackling the social norms that preventwomen from having adequate access to inputs and marketing facilities inthe first place. This might be done through demonstration effects andfrom pressure created by the micro-lender to ensure higher returns to bor-rowers’ investments.

From the point of view of practice, a field loan officer would see womenas better customers for loans compared with men for at least four reasons.First, repayment rates on loans by women are higher, because women aremore risk-averse and therefore more conservative in their investment strat-egy. Also, women are more vulnerable to peer pressure and the threat ofpublic humiliation with regards to failure in the repayments on their loans,women have less opportunities than men to access alternative sources ofcredit, which in turn reduces the scope for moral hazard.8 Moreover, fieldpractitioners in microfinance argue that women are less argumentative,which reduces the transaction costs of the loan, both for their peers and thebank. Women also lower the agency costs of bank officers because asgroups they are more punctual at repayment meetings, which avoids thebank officer having to devote time looking for them at their homes/businesses. Last but not least, women loan officers cost less than men, andin many instances women are more efficient at granting and collecting repayments.9

Taken together, the findings of empirical investigations, the perspectivesof donors, and experience of practitioners, have led to an establishedwisdom in favor of lending to women. Moreover, the conventional wisdomhas been that excluding men from microfinance has no significant or im -portant detrimental outcomes. However, more recent views from the field

112 Microfinance

expressed at the recent Microfinance Forum in Beijing (2006) suggest otherwise:

male exclusion can lead to negative consequences for women who join financialservices: they may meet resistance from men who see their exclusive participa-tion as unfair and threatening; their loans may be hijacked . . . A family whoseadult members all have access to financial services is better off than one wherehalf are ineligible. (Hugh Allen at the Microfinance Forum, 2006)

While the experiential knowledge of people like Hugh Allen should not beaccepted without detailed investigation, his views have been a considerationfor social scientists and anthropologists voicing similar concerns for sometime now. Their observation, which run counter to conventional wisdomare reviewed in the following section.

ANECDOTAL EVIDENCE FROM BANGLADESH ANDAFRICA

In this section, we argue that there are potential dangers in excluding menfrom subsidized microfinance as this may lead to frictions between house-hold heads, leading to lower quality and quantity of health and educationprovision within the overall household. At this stage the evidence for thisposition is anecdotal, deriving from Bangladesh and Africa. It suggests thatthere is a need to take into account the potential danger of excluding themale head of household from microfinance, as their exclusion can over-burden women and lower health and education outcomes.

Long before the 2006 Nobel Peace Prize was awarded to the creator ofGrameen Bank, Muhammad Yunus, for his work in microfinance, house-hold surveys from Bangladesh, dating back to 1999, documented evidencethat microfinance was increasing frictions between husbands and wives, ashusbands often felt threatened in their role as primarily income earners(Rahman, 1999). Moreover, well-known evidence, also from Bangladesh,suggests that microfinance does not increase women’s bargaining powerentirely, because on average, women borrowers surrender nearly 40 percentof their control over the investment decisions they make. More alarmingly,over 90 percent of the returns these women realize from their investmentsare handled by their husbands (Goetz and Sen Gupta, 1996).

In Africa, Linda Mayoux (1999), reports on a survey of 15 differentmicrofinance programs. She finds that the degree of women’s empowermentis household- and region-specific, with women’s empowerment dependenton inflexible social norms and traditions. These findings have to be weighedagainst the fact that impact on empowerment will also depend on how well


particular programs were designed. The issue of context factors andprogram design leads to the exceedingly preliminary observations from afield experiment undertaken in southern Mexico.

ANECDOTAL EVIDENCE FROM FIELDEXPERIMENTS IN SOUTHERN MEXICO10

Grameen Trust Chiapas, AC (henceforth called GTC) is one of the firstreplications of the Grameen model of microfinance in Latin America. Theproject is located in the highlands of southern Mexico. It deploys fundsfrom the Deutsche Gesellschaft für Technische Zusammenarbeit (GTZ) viaGrameen Trust Bangladesh. The replication in southern Mexico started bylending to women-only groups in 1997.11 In 2003, in sharp contrast to theoriginal Grameen model, GTC took the risk of lending to men of someotherwise women-only groups. Since then the organization grew rapidly,and it now has over 12 000 borrowers in different groups, a large majorityof those mixed groups of women and men.

When branch managers in different geographical locations in the south-ern Mexican replication are asked why they have accepted men into women-only groups, four explanations are offered. The first relates to informationalasymmetries between men and women. One loan officer argues that even ifloan disbursements and repayments are publicly known in women-onlygroups, men tend to overestimate the amount of money that women arehandling, and they therefore contribute less to overall household expendi-ture, which often creates frictions within the household. This has dynamiceffects. In many instances, women under these conditions are no longerusing their loans for investment but for normal household expenditures onfood, health, and education (particularly in the month of August when theacademic year starts). They also often quarrel with their husbands who areno longer providing as much for these expenditures as they used to. Invitingsome men to join the group allows them to have a more accurate estimateof women’s real investments and their realized returns. With this informa-tion they are less likely to reduce their contributions to household expen-ditures. In those groups that became mixed, women borrowers invest more,and there were increased repayment rates by both men and women in suchgroups.

A second explanation relies on the potential workload externalities ofhaving women as the only recipients of loans within the household. In par-ticular, another loan officer argues that when women contract a loan fromGTC, they become busier, and that the quality of the services that womentraditionally provide to the household such as meals, and household

114 Microfinance

chores, decreases in quantity or quality or both. This, the bank managerargues, irritates men and creates a “tense” atmosphere within the house-hold. This family tension causes women to default more often or preventsthem from making their repayments on time. When men are invited to joingroups, they seem to internalize the negative workload externalities createdby GTC micro-loans to women. In loan officer Regis Ernesto Figueroa’sown words: “invited men help more their spouses in their businesses and inhousehold chores, which in turn, reduces tensions, and enables women torepay on time, as men become de-facto business partners of women.”

A third explanation relates to the absence of secure places for women tohide money while they save for two consecutive weeks or more in order tomake their repayments to GTC. In particular, another loan officer arguesthat women cannot open bank accounts in commercial banks as thesebanks do not accept their very small savings, as the transactions costs forthe commercial banks are too high relative to the amounts deposited.Women borrowers of GTC therefore hide money away from their husbandsin different places, generally in the house, because husbands steal themoney and use it on alcohol and tobacco. When men are invited to join thegroup, this particular loan officer argues, the situation changes becauseunder the “Grameen rules,” he becomes responsible for the debt of theother male and women members of the group. “Women become happier.They no longer complain about their husbands or men in general. Thehousehold heads work harmoniously together,” the loan officer explains.

A loan officer at the headquarters of GTC offers a fourth, and last, expla-nation of why men are invited into women’s groups. She argues that theinclusion of men brings more women clients into the scheme, in particularmore single women. She explains that the reason is because women gener-ally face a trade-off between being financially independent via a micro-loanfrom GTC, or, getting married. The argument goes that since GTC acceptsmen, women no longer face this trade-off, and they are therefore more likelyto become clients. Moreover, the inclusion of men, according to this loanofficer, has increased marriage rates!

ATTEMPTS TO MEASURE EMPOWERMENT ANDDISEMPOWERMENT12

The anecdotal evidence set out above suggests a substantive need to explorein greater depth the relationship between microfinance structures and theissues of gender in development and empowerment around microfinance.This calls for experiments designed to test the effects of the inclusionof male heads of households into women-only solidarity groups. Such


experiments are exceedingly demanding. Nevertheless they are importantgiven the challenges to the conventional wisdom, that women are increas-ingly empowered by microfinance that enables them to expand their busi-nesses, earn a higher return so that their spouses would value them better,which translates into higher health and education for the household.

In designing such studies it is recognized that cultural and institutionaldifferences may impact the results. Ideally, any test of the empowering- disempowering hypothesis should therefore take place in Bangladesh,Africa, and Latin America to establish whether the results are culturallyand institutionally robust. However, finding partner microfinance institu-tions that would allow researchers to conduct scientific experiments of thiskind, is difficult enough; to do so in three continents is yet more difficult.

We report here progress to date with a pioneering study developed byHarvard and Yale researchers from the Innovations for Poverty Action(IPA) in their continuing study on the impact of gender issues ofmicrofinance and health and education outcomes in association with theGrameen Trust Chiapas, AC in southern Mexico. The elements of thisstudy are reported below.

Innovations for Poverty Action (IPA) researchers designed a survey anda follow-up random experiment, using a sample of approximately 2000borrowers in women-only solidarity groups in 2006. In this experimentmarried-women-only solidarity groups were (randomly) selected into treat-ments and controls. Control solidarity groups were not subject to any kindof “intervention” while the treatments were.

The treatments were divided into four different sub-groups. Interventionin the first sub-group consisted of allowing women to voluntarily invitetheir partners/husbands to join the Grameen-style solidarity group in orderto acquire a micro-loan. The study sought to take into account the poss -ible “network effects” that might follow as invited male spouses joined soli -darity groups, increasing the synergies through the span of group members.IPA researchers therefore allowed for a sub-group of women who couldinvite other women friends to join their group. Similarly, it was recognizedthat as partners/spouses were invited to join, so household income wasincreased. This factor was taken into account by extending larger loan sizesto a sub-group consisting of women-only clients. Last but not least, a treat-ment sub-group consisting of women who could invite their partners/husbands via providing them high and low monetary incentives to betterproxy women’s marginal benefit from being financially independent.

The IPA researchers are using a follow-up survey of the four sub- treatment groups and the control group to assess and evaluate any behav-ioral changes at household level. Some of the questions that will be resolvedare: did women borrowers decide to invite their male spouses, and if so, did

116 Microfinance

their willingness to do so increase as they were provided with incentives? Inwhat way did the inclusion of male spouses alter outcomes in terms ofhealth, education, and child labor, for example? It is recognized that thereare many other possible dimensions of change to address, but it is not pos-sible to develop a comprehensive questionnaire until the initial results fromthe 2008 follow-up survey of behavioral changes has been processed and analyzed.

In the meantime, IPA researchers have been able to detect some interest-ing idiosyncrasies in the sample of borrowers, which include five geo-graphical areas, quite distant from one another. These preliminary findingsshow that most of the decisions regarding investment and the expendituresfrom returns realized from micro-loans are taken by men heads of house-holds, not by women, in the bank branch with the highest proportion ofindigenous population. In addition, Grameen Trust Chiapas, as well asother group-lending institutions such as AlSol, another Grameen replica-tion founded by Beatriz Armendáriz, have been serving borrowers in thesebranches/regions for a long time. Health and educational expenditures inthese two branches do not differ considerably. If microfinance to womenhad not already empowered women in such traditional societies, andassuming the head of household relationship is basically frictionless aswives systematically defer decision-making to their husbands, it is difficultto imagine what an intervention of the sort we undertook in that regioncould actually bring about in terms of changes in behavioral patterns.

However, when the women in these branches are given the power to invitetheir spouses into the group this provides an empowering tool in its ownright, and, household income is expected to increase when women actuallydecide to include their spouses. On the other hand we might expect that ifany anticipated increase in household income, once partners are includedas microfinance clients, is then controlled purely by men, there would notbe the expected changes in outcomes, particularly with respect to healthand education. Anticipating this, women might decide not to invite theirspouses in the first place. It is, however, much too early in the experimentto make any predictions of substantive outcomes from the project.

A somewhat similar scenario seems to prevail in two more affluentbranches and regions. Interestingly, in at least one of the two branches,women borrowers have remained with the Grameen Trust Chiapas for amuch longer period of time compared with the other four branches. At thistime their higher income and expenditure, on average, might be due to thiscontinued microfinance activity, nevertheless, educational levels seem to bejust as low as in the poorest branch, while it appears that health expendi-tures are somewhat higher. Whether women will opt for bringing theirspouses in to the project, and whether this will translate into higher


outcomes in terms of health and education remains an open question atthis stage in the program.

An interesting situation exists in the other two branches and regionswhere the income of the borrowers is the highest. Women in both theseregions are not just wealthier, but are also more educated and their healthexpenditures appear to be higher. Women seem to be more empowered inthat they often declared themselves as being the main household head, andtake most of the household decisions. Their spouses seem to be more sup-portive of their micro-businesses. In this situation these women might valuetheir financial independence, and this might be leading to a completely fric-tionless relationship, in which case we should probably not expect thosealready empowered women to actually invite their partners to join theirproject either. Again, it is too early to tell as we do not yet know the levelof take-up of male membership, and if the invitation of spouses to join willlead to higher outcomes in terms of expenditures on health or education.

CONCLUDING REMARKS

At present the baseline survey in Chiapas indicates that the degree ofwomen empowerment is in line with Linda Mayoux’s (1999) findings in 15different microfinance programs in Africa. That is to say, expenditure ishousehold- and region-specific and inflexible as a consequence of socialnorms that seem exceedingly difficult to change.

However, empowering women via an additional tool, namely by givingthem the right to voluntarily invite their partners, might help to acceler-ate the process for change in those social norms. This might, however,prove more difficult in poorer regions where household heads seem mostlyto be men. The question then is why should subsidized loans that makewomen responsible for repayment, but do not give them power overcrucial decisions regarding their business and household expenditures inhealth and education seem to be endorsed by donors? Moreover, as themicrofinance industry becomes increasingly commercial, micro-creditbecomes increasingly burdensome on women. Why should women takeon the responsibility over higher repayments in the first place? This viewaccords well with for-profit microfinance enterprises in Latin Americawhere men are increasingly self-selecting into programs offered by suchenterprises. In the absence of subsidies, Grameen Trust Chiapas as well asother organizations in the region might be increasingly attracting men,not women. And the interest rates charged could be “friendlier” towomen, if only because women are the main brokers of health and edu-cation within the household.

118 Microfinance

As far as the more affluent clients served by Grameen Trust Chiapasare concerned, it might be that the whole idea of excluding husbands canbe counterproductive, because of informational asymmetries that appearto lead to mistrust, frictions within the couple, and worse, a decreasedparticipation by men in household expenditures. This outcome is notwhat we understand as a preferred outcome for women. However, suchdisempowerment effects should be weighed against the value that womenattach to their financial independence. We see an important balancebetween greater financial independence on one side, and more demandson time and loss of money for established levels of expenditure on theother.

Given these scenarios women might be reluctant to invite their part-ners into their groups, but for different reasons. In the case of less affluenthouseholds in very traditional societies, because having partners join theproject would not change anything. And in the case of relatively moreaffluent households because women attach too much value to their financial independence.

Final results on take-up as well as potential behavioral changes fromthis experiment on gender should further clarify the questions and shedlight on implications of microfinance structures around the exceedinglyimportant issues of gender in the field of microfinance, development, and empowerment.

NOTES

1. I gratefully acknowledge the support of Alissa Fishman, Randall Blair, and Julio Lunafrom IPA in Mexico. Co-authors Dean Karlan and Sendhil Mullainathan have beenincredibly patient in teaching me how to conduct field work and without their guidance,my part in this chapter would never have been written. Valuable comments from seminarparticipants at Harvard, Columbia, Solvay Business School, and CERMi in Brussels arealso greatly appreciated. Finally, the collaboration of Grameen Trust Chiapas manage-ment, and in particular, the support from Ruben Armendáriz, Maricela Gamboa,branch managers, and loan officers from Tuxtla, Ocosingo, Comitan, and LasMargaritas, have been exceedingly valuable. Nigel and I are solely responsible for theviews and errors in this chapter.

2. Daley-Harris, Sam (2003).3. Some evidence on this and follow-up debate is found in Mayoux (1999) and Rahman

(2001), among others.4. This section borrows from Armendáriz’s joint work with Jonathan Morduch (2005). For

a more general survey on gender issues in economic development, see Duflo (2005).5. Pitt and Khandker’s econometric estimations and results are, however, exceedingly con-

troversial. For a more comprehensive critique, see Pit (1999) and Armendáriz andMorduch (2005).

6. Promoting women to powerful positions in villages and regions may, by the same token,bring social benefits. In a recent paper on India, Raghabendra Chattopadhyay andEsther Duflo (2003) show that by empowering women, and, in particular, by allowing


them to be elected to local councils, spending on public goods most closely linked towomen’s concerns increased.

7. Evidence from India also shows that there is a positive correlation between the relativesize of a mother’s assets (notably jewelry) and children’s school attendance and medicalattention (Duraisamy and Malathy, 1991; Duraisamy, 1992).

8. Hossain (1988): 81 percent of women had no repayment problems versus 74 percent ofmen.

9. Khandker, Khalily, and Kahn (1995): 15.3 percent of male borrowers were “struggling”in 1991 versus 12.4 percent of female (missing some payments before the final due date).

10. This section draws from current field work with Dean Karlan and Sendhil Mullainathanin southern Mexico.

11. A year later, and under the auspices of Grameen Foundation USA, some of theGrameen Trust Chiapas’s managers founded AlSol, which currently serves approxi-mately 3000 borrowers.

12. For an explanation on random experiments, see Duflo, Glennester and Kremer (2006).

REFERENCES

Anderson, Siwan and Jean-Marie Baland (2002), “The economics of ROSCAs andintrahousehold allocation”, Quarterly Journal of Economics, 117 (3), 983–95.

Armendáriz, Beatriz and Jonathan Morduch (2005), The Economics ofMicrofinance, Cambridge, MA: MIT Press.

Chattopadhyay, Raghabendra, and Esther Duflo (2003), “Women as policy makers:evidence from an India-wide randomized experiment”, typescript, Cambridge,MA: MIT Economics Department.

Daley-Harris, Sam (2003), State of Microcredit Campaign Report 2003, November,Washington, D.C.: Microcredit Summit, accessed at www.microcreditsummit.org/pubs/reports/socr/2003/socr 03_en.pdf.

Duflo, Esther (2005), “Gender equality in development”, Massachusetts Instituteof Technology Poverty Action Lab working paper, Cambridge, MA.

Duflo, Esther, Rachel Glennester, and Michael Kremer (2006), “Using randomiza-tion in development economics research: a toolkit”, Massachusetts Institute ofTechnology Poverty Action Lab working paper, Cambridge, MA.

Duraisamy, Paul (1992), “Gender, intrafamily allocation of resources, and childschooling in South India”, Yale University Economic Growth Center, workingpaper no. 667, New Haven, CT.

Durisamy, Malathy (1998), “Children’s schooling in rural Tamil Nadu: gender dis-parity and the role of access, parental and household factors,” Journal ofEducational Planning and Administration, XII (2), 131–54.

Engle, Patrice (1993), “Influences of mothers’ and fathers’ income on children’snutritional status in Guatemala”, Social Sciences and Medicine, 37 (11), 1303–12.

Goetz, Anne Marie and Rina Sen Gupta (1996), “Gender, power, and control inrural credit programs in Bangladesh”, World Development, 24 (1), 45–63.

Hossain, Mahabub (1988), “Credit for alleviation of rural poverty: institute researchreport 65”, February, Washington, DC: International Food Policy Research.

Khandker, Shahidur R., Baqui Khalily and Zahed Kahn (1995), “Grameen Bank:performance and sustainability”, World Bank Discussion Paper no. 306,Washington, DC.

Mayoux, Linda (1999), “Questioning virtuous spirals: microfinance and women’sempowerment in Africa”, Journal of International Development, 11 (7), 957–84.

120 Microfinance

Pitt, Mark (1999), “Reply to Jonathan Morduch’s: Does microfinance really helpthe poor? New evidence from flagship programs in Bangladesh”, typescript,Providence, RI: Department of Economics, Brown University.

Pitt, Mark and Shahidur Khandker (1998), “The impact of group-based credit pro-grams on poor households in Bangladesh: does the gender of participantsmatter?”, Journal of Political Economy, 106 (5), 958–96.

Rahman, Aminur (1999), “Microcredit initiatives for equitable and sustainabledevelopment: who pays?” World Development, 26 (12) December.

Rahman, Aminur (2001), Women and Microcredit in Rural Bangladesh: AnAnthropological Study of Grameen Bank Lending, Boulder, CO: Westview Press.

Shultz, T. Paul (1990), “Testing the neoclassical model of family labor supply andfertility”, Journal of Human Resources, 25 (4), 599–634.

Skoufias, Emmanuel (2001), “Is PROGRESA working? Summary of the results byan evaluation by International Food Policy Research Institute (IFPRI)”, IFPRIFood Consumption and Nutrition Division discussion paper no. 118,Washington, DC.

Thomas, Duncan (1990), “Intrahousehold allocation: an inferential approach”,Journal of Human Resources, 25 (4), 635–64.

Thomas, Duncan (1994), “Like father like son, or, like mother like daughter: parentaleducation and child health”, Journal of Human Resources, 29 (4), 950–88.

Udry, Christopher (1996), “Gender, agricultural production, and the theory of thehousehold”, Journal of Political Economy, 104 (5), 1010–46.


Index

A Little World 79administrative costs 87, 90Africa 18, 20, 21, 113–14, 118

see also individual countriesagencies see aid agencies; development

agencies; NGOs (non-governmental organizations)

agency costs 112agricultural loan initiatives 4agricultural pricing information 13agricultural production 112aid 111aid agencies 5, 92airtime credits 76, 77, 78Allen, Hugh 113AML (anti-money laundering) 16–17,

77–8, 82, 89–90Anderson, Siwan 111Asia 74

see also Asia Pacific; East Asia;South Asia; individual countries

Asia Pacific 78see also individual countries

asset class 40–41assets, median total of lending

institutions 20, 21assortative matching 11–12, 19

Baland, Jean-Marie 111BancoSol 92Bangladesh 4, 100, 109–10, 113

see also Bangladesh Bank; BRAC;BRAC securitization; EasternBank Limited (EBL); GrameenBank; PKSF (Palli KarmaSahayak Foundation); Proshika

Bangladesh Bank 49–50bank failures 86, 93–4, 98, 105banks

and CDOs 31, 32costs of access to 7and equity 9, 37

India 72, 73, 74and Internet 20and mobile phone technology 13–14,

20, 79, 80, 82, 83, 98prudential regulation 86, 94, 95, 96,

97–8, 105and securitization of micro-loans

35supervision 98–100, 102, 103, 105women’s empowerment 18, 19, 20see also bank failures; commercial

banks; overbanking; ruralbanks; rural branches ofnationalized banks; villagebanks; individual banks

barcode-reading point-of-sale (POS)terminals 12–13

Barth, J.R. 105Beam 74, 80–83BlueOrchard 27, 28, 32, 46Bolivia 89, 91–2, 93–4BOMSI (BlueOrchard Microfinance

Securities I) 27–32bonds 28, 30, 38, 50, 53–5, 62–8borrower characteristics’ information

59, 69–70, 73, 90, 95borrowers 3, 7–8, 19, 26, 47, 48

see also borrower characteristics’information; consumerprotection; men; self-helpgroups (SHGs); solidaritygroups; women; individualborrowers

BRAC 47–9BRAC securitization

currency risk 46future prospects 69–70performance 46, 61–8political economy considerations 49–

53structure overview 35, 53–60transaction rationale 47–9

123

branching requirements 96branchless banking 97–8

see also mobile phone technologyBrazil 12–13, 111BRI (Bank Rakyat Indonesia) 94broad borrowers 8BTMs (biometric teller machines) 13Burkina Faso 112

capital adequacy 93–4capital calls 101capital markets 5, 9–11

see also domestic capital markets;international capital markets

capital structure 8–9Caprio, G. 105cash 14, 75, 76, 77, 111, 113, 114, 115cash flow 60–61, 62, 63, 65, 68, 70CDOs (collateralized debt obligations)

10, 27–34, 46CFT (combating the financing of

terrorism) 17, 82, 89–90CGAP (Consultative Group to Assist

the Poor) 6–7child health 111China 78Citibank 53Citigroup 10co-signers as borrowers’ restrictions 95collateralized loans 94–5commercial banks 26, 38, 39–40, 72,

74, 93–4commercial investors 29–31, 35–9Compartamos 6, 10, 37, 88competition, commercial and non-

commercial investors 38–9consumer protection 17, 88–9contracts see loan contracts; putscooperatives 18, 20

see also credit unions; self-helpgroups (SHGs); solidaritygroups

correspondent banking 12–13corruption 77, 105costs

access to credit and financial services6–9, 14

cash handling 75, 76electronic matching of borrowers

and lenders 11

m-commerce 78micro-loans 87–8prudential regulation 87women as microfinance customers

112see also administrative costs; fees;

transaction costsCRAB (Credit Rating Agency of

Bangladesh) 50, 53credit enhancements, BRAC

securitization 55–6, 62–8credit history 4, 8, 112credit ratings 6, 11, 35, 38, 50, 62, 70, 89

see also CRAB (Credit RatingAgency of Bangladesh);MicroRate

credit rationing 3credit supply 4credit unions 4, 7, 9, 18, 19, 20, 21, 94,

103culture 20–21

see also social normscurrency 33

see also currency risk; local currencycurrency risk 33, 46, 49–50

data 56–60, 69–70debt/equity ratio 9default insurance 3default risk 11, 15, 19defaults, loan see loan defaultsdelegated supervision 103–4delinquent loans

and BRAC securitization 55, 56, 57,59–60, 61–2, 63–7, 68

collateralized versus uncollateralized94–5

demography, India 71–2deposit insurance 16, 97deposits/savings

BRAC microfinance funding 48in domestic capital markets 26India 71, 72, 74regulation 16, 86–7, 91, 92, 106unsupervised small community-

based intermediaries 100–101and women 111, 115

development agencies 26, 27, 30, 35,38–9, 40, 101

see also FMO; KfW

124 Index

Dexia Microcredit Fund 27domestic capital markets 10, 26, 40Done Card 79donor funding 48, 92DSRAs (debt service reserve accounts)

56, 69, 70DWM (Developing World Markets)

26, 27, 28, 29–30, 31, 38, 41

East Asia 18, 20, 21see also individual countries

Eastern Bank Limited (EBL) 47, 53eChoupals 13economic downturns 31, 41

see also bank failures; sub-primemortgage crisis

economy, India 71–2education 108–9, 110, 111, 114, 116,

117–18see also financial education; literacy

levels; primary educationelectronic matching of borrowers and

lenders 11–12Engle, Patrice 111entrepreneurship 4, 19, 110, 116, 118equity 9, 28, 35–7Equity Bank (Kenya) 10, 37equity capital 10Europe 35, 37, 38

see also FMO; KfWexchange rates 50exits 36, 37

FATF (Financial Action Task Force)89–90

fees 34, 50financial controls 42

see also regulation; supervisionfinancial education 88–9, 101financial return 34, 36, 41–2financial services 5, 7–9Fino 79–80FMO 35, 40, 49–50, 53food expenditures 110, 111, 114formal credit markets 2–3

see also banksformal markets 2fraud 77friction, marital 108, 113, 114, 115,

119

funding sources 48, 51funds 27–9, 32–4, 38

G-cash 14GCMF (Global Commercial

Microfinance Facility) 38gender see men; women; women-only

solidarity groups; women’sdisempowerment; women’sempowerment

geographic variation, women’sempowerment 18, 19–21, 117–18

Gonzales, A. 7–8, 9, 18, 19governments 5, 9, 48, 51–3

see also prudential regulation;supervision

Grameen Bank 4, 20, 87, 108, 113Grameen Trust Chiapas, AC (GTC)

114–15, 116, 117–19group lending see self-help groups

(SHGs); solidarity groups;women-only solidarity groups

GSMA 74–8Guatemala 111

health 2, 22, 108–9, 110, 111, 114, 116,117–18

hedging 38, 39–40, 44, 55–6high-end borrowers 8high-risk borrowers 3HNWIs (high net worth individuals)

32, 38household expenditure 110, 111, 114,

117–18household income 1, 108, 110, 111,

116, 117–18, 119housework 114–15

ICICI 13, 46, 80IFC Washington 10, 75–8IFIs (international financial

institutions) 10see also individual IFIs

illiquidity 34, 41income 1

see also household incomeIndia 4, 13, 15–16, 71–5, 78

see also ICICI; mobile phonetechnology in India

individual loans 8, 9, 18, 19

Index 125

individual responsibility 42, 74Indonesia 87, 94, 99–100informal credit markets 3–4informal markets 3information see agricultural pricing

information; barcode-readingpoint-of-sale (POS) terminals;borrower characteristics’information; data; eChoupals;information asymmetry;information availability;information disclosure;information sharing; Internet;loan files; MIS (managementinformation systems); reporting;transparency

information asymmetry 2–3, 114, 119information availability 3information disclosure 14–15, 88–9information sharing 70insider lending 96insiders, and equity 36interest payment frequencies 3interest rate caps 88interest rates 4, 6–7, 11, 14–16, 19, 48–

9, 52, 87–8, 91international capital markets

amount of investment 25asset class, on the path to 40–41CDOs versus funds 32–4commercial investors, growing

participation in CDOs 30–32commercial investors, introducing to

CDOs 29–30and concerns about microfinance

41–3equity 35–7from funds to CDO 27–9importance 25–7and local currency 33, 38, 39–40non-commercial investors 26, 27,

37–9and securitization of micro-loans

34–5international money transfer services

81, 82, 83Internet 11–12, 13, 20, 70, 79intimidation 89IPA (Innovations for Poverty Action)

116–18

IPOs (initial public offerings) 10, 36, 37ITC (India Tobacco Company) 13ITZ Cash 79

JiGrahak 79joint liability 3, 8, 74

Kenya 10, 77, 98, 111KfW 10, 35, 50, 53Khandker, Shahidur 109–10KIVA 12

land ownership 110Latin America 4, 18, 19, 20, 21, 36, 87

see also Compartamos; individualcountries

legal vehicles see SPVs (special-purpose vehicles)

Lending Club 12Levine, R. 105licensing 91–2, 99–100, 102liquidations 28, 36liquidity 34, 40, 44, 62

see also illiquidityliteracy levels 12, 101Little World, A 79loan contracts 3, 8, 9, 11, 35loan defaults 15–16, 29, 42, 59, 89, 94–5

see also default insurance; defaultrisk; delinquent loans

loan files 42, 95, 101loan loss provision 95loan maturity 48

see also short-term loan contractsloan repayments

and consumer protection 89frequencies 3, 48and inclusion of men in women’s

solidarity groups 114and m-commerce 14, 76, 77self-help groups (SHGs) in India 74women 4, 112

loan size 19–20, 26see also broad borrowers; high-end

borrowers; low-end borrowers;small borrowers

local commercial banks 38, 39–40, 53see also individual banks

local currency 33, 38, 39–40, 46, 49–50logistics 57–8

126 Index

long-term investment 27–9low-end borrowers 7, 8low-risk borrowers 3

m-commerce 75–8, 79marriage rates 115Mayoux, Linda 113–14, 118MChek 79men 108, 110, 111–13, 114–18, 119Mexico 111, 114–15, 116–19MF Analytics 53, 56MFBA (Microfinance Bank of

Azerbaijan) 38MFIs (microfinance institutions)

borrowers, global number of 26capital adequacy 93–4costs 87–8India 72–4loan size 19–20, 26mission drift 42–3ownership requirements 97regulation see regulationsize 100stock exchange listing 10supervision 99–104sustainability 92and women’s empowerment 18, 19,

20–21see also banks; NBFIs (non-bank

financial institutions);individual MFIs

MFRC (Micro Finance RegulatoryCouncil) 89

micro-credit backed securities 5micro-insurance 81, 82, 83micro-investment 10, 81, 82, 83

see also international capital marketsmicro-payments 82, 83microfinance, evolution 4–6, 9, 108Microfinance Securities XXEB 26, 31MicroRate 6, 38–9minimum capital requirements 93, 99,

100MIS (management information

systems) 51, 53, 56–8, 69, 70mission drift 42–3MIX data 6, 7–8, 9, 18–21, 87–8, 92mobile phone technology

and banks 13–14, 20, 79, 80, 82, 83, 98and BRAC securitization 70

India see mobile phone technologyin India

and m-commerce 75–8mobile phone technology in India

future and challenges 81–3industry 79–80and microfinance 74–5, 78products 80–81usage 72

money see AML (anti-moneylaundering); cash; cash flow; CFT(combating the financing ofterrorism); currency; fees; loanrepayments; loan size; moneylaundering

money laundering 77see also AML (anti-money

laundering)Morgan Stanley 32

NBFIs (non-bank financialinstitutions) 5, 7, 9, 18, 19, 20, 21

see also credit unions; PACS(primary agriculture creditsocieties); ROSCAs (RotatingSavings and CreditAssociations); self-help groups(SHGs); village banks

NGOs (non-governmentalorganizations)

and capital calls 101and capital markets 10–11costs of access to financial services

7, 9in evolution of microfinance 5, 9ownership requirements 97and women’s empowerment 18, 19,

20, 21see also individual NGOs

non-commercial investors 26, 27, 30,32, 37–9

non-profit organizations 26, 33–4see also aid agencies; credit unions;

development agencies; NGOs(non-governmentalorganizations); PACS (primaryagriculture credit societies);ROSCAs (Rotating Savings andCredit Associations); self-helpgroups (SHGs); village banks

Index 127

non-prudential regulation 17, 86, 87–90, 103

Omidyar, Pierre 5OPIC (Overseas Private Investors

Corporation) 30OPORTUNIDADES 111origination risk 35over-collateralization 63–8, 69–70overbanking 33, 41–2ownership requirements 97

PACS (primary agriculture creditsocieties) 72

Pakistan 91pawnbrokers 4pay-down structure 54–5payday lending 4Paymate 79payment systems 74–5peer monitoring 14, 15, 18peer pressure 15, 74, 112performance, BRAC securitization 46,

61–8Philippines 14, 76–8, 98, 99–100physical security 96Pitt, Mark 109–10PKSF (Palli Karma Sahayak

Foundation) 48, 51, 53, 54political economy 49–53, 70, 92, 99,

104, 105pool maintenance, BRAC

securitization 53, 56poor persons 2–3, 5, 22, 100–101, 110

see also poverty alleviation; povertymeasurement

Portio Research 78poverty alleviation 7–8, 110, 111poverty measurement 1prepayment risks 55–6, 60–61, 69prepayments, and BRAC securitization

55–6, 57, 60–61, 62–3, 64–7, 68,69

primary education 1, 22ProCredit Bank Bulgaria 35ProCredit Holding AG 5, 35ProFund Internacional SA 36promotion of microfinance 90–91Proshika 52–3Prosper.com 11, 12

prudential regulationadjusting regulations to fit products

and institutions 92–8critique 105–6versus non-prudential regulation

86–7regulation as promotion 90–91regulation that follows the market

91–2puts 36PWC (PriceWaterhouseCoopers) 53,

69

RBI (Reserve Bank of India) 72regulation

deposits/saving 16, 86–7, 91, 92, 106formal markets 2infrastructure problems 35interest rates 14–16and m-commerce 77–8and mobile phone technology in

India 82non-prudential regulation 17, 86,

87–90, 103prudential regulation (see prudential

regulation)self-regulation 104see also peer monitoring; supervision

repayment capacity 89reporting 53, 56, 70, 96, 103, 104risk

BRAC securitization 69, 70CDOs (collateralized debt

obligations) 27–30, 31, 32, 33and costs of access to financial

services 8–9group lending and self-help groups

(SHGs) 3non-commercial investors as shield

for commercial investors 37–8and overbanking 33, 41–2see also currency risk; default risk;

high-risk borrowers; low-riskborrowers; origination risk;prepayment risks; risk analysis;risk aversion; risk/return ratios;security risks

risk analysis 58–61, 70risk aversion 26, 70, 112risk/return ratios 29–30

128 Index

ROSCAs (Rotating Savings and CreditAssociations) 111

RRBs (regional rural banks) 72, 73, 74rural areas 71, 72

see also RRBs (regional ruralbanks); rural banks; ruralbranches of nationalized banks;village banks; village Internet

rural banks 99–100see also RRBs (regional rural

banks); rural branches ofnationalized banks; villagebanks

rural branches of nationalized banks 5,7, 9, 72, 73

sachet purchasing 76–7savings see deposits/savingsSCBs (scheduled commercial banks)

72, 74Schultz, Paul 111secondary markets 32, 43Securities and Exchange Commission

(SEC) 50securitization of micro-loans 5, 10,

34–5, 46see also BRAC securitization; CDOs

(collateralized debt obligations)security risks 74, 90, 91, 100–101

see also physical securityself-employment 74self-help groups (SHGs) 3, 14, 19,

73–4see also solidarity groups; village

banks; women-only solidaritygroups

self-regulation and supervision 104servicers 35

see also BRAC securitizationshort-term investment 27, 50, 53,

54–5short-term loan contracts 3, 11, 35Skoufias, Emmanuel 111small borrowers 73, 87, 90, 91,

100–101see also low-end borrowers

small community-based intermediaries100–101

SMS (short message service) 78, 79,80–81

social norms 112, 113–14, 118social pressure 15, 74, 112socially responsible investors see non-

commercial investorssoft capital 5solidarity groups 3, 8, 9, 14, 15, 18, 19

21see also women-only solidarity

groupssolvency 93–4South Africa 13–14, 77, 89, 90, 91, 98South Asia 18, 19, 20, 21

see also individual countriesSPVs (special-purpose vehicles) 28, 53,

54, 55, 62–8stand-alone loans see individual loansStephens, B. 10–11stock exchange listing 10stop-lending order 102sub-prime mortgage crisis 32substitution, and BRAC securitization

55, 56, 63, 64–7, 68Superintendency of Banks (Bolivia)

89, 92supervision 96, 98–106sustainability of MFIs 92Suvidha 74, 80, 81–3swaps 40SWIFT 74, 80, 81syndication 48, 49

Tanzania 91taxation 7, 82technological innovation 5, 11–14,

15–17technology 2, 8, 22

see also Internet; mobile phonetechnology; technologicalinnovation

Thailand 111Thomas, Duncan 111TIAA-CREF 5transaction costs 2, 7, 73, 78, 112transparency 30, 33trustees 28, 30, 47, 53

Udry, Christopher 112uncollateralized loans 94–5underclass 76United Kingdom 11, 12, 111

Index 129

United States 4, 11, 12, 32see also IFC Washington;

MicroRate; OPIC (OverseasPrivate Investors Corporation)

unsecured lending limits 94USAID 14usury limits 87–8

village banks 5, 8, 9, 18–19see also rural banks; rural branches

of nationalized banksvillage Internet 13violence 89volatility 41, 48–9

Wallet 365 79women 4, 111, 112, 115women-only solidarity groups 114–19womens’ disempowerment 108–9, 113,

115–19women’s empowerment

agricultural inputs and productivity112

assets 20, 21and education 108, 110, 111, 114,

116, 117–18entrepreneurship 4, 19, 110, 116, 118in evolution of microfinance 4, 108

geographic variation 18, 19–21,117–18

and health 108, 110, 111, 114, 116,117–18

and household expenditure 110, 111,114, 117–18, 119

and household income 108, 111, 116,117–18

and housework 114–15individual versus solidarity group

lending 18–19, 21and lending institutions 18, 19,

20–21, 108and level of aggregation 20–21loan size 19–20and marital friction 108, 113, 114,

115, 119measurement 115–19microfinance in India 72SHG (self-help group) lending in

Asia 74and social norms 112, 113–14, 118

World Bank 1, 74, 110

Yanus, Muhammad 4, 113

Zambia 92Zopa.com 11, 12

130 Index

Download - Micro Finance - Emerging Trends and Challenges

Top Related