SCALING UP POVERTY REDUCTION Case Studies in Microfinance Consultative Group to Assist the Poor World Bank Financial Sector Network Washington, D.C. Global Learning Process for Scaling Up Poverty Reduction and Conference in Shanghai, May 25-27, 2004 Published 2004 by CGAP, the Consultative Group to Assist the Poor Copyright © 2004 CGAP / The World Bank Group 1818 H Street, NW Washington, DC 20433 USA All right reserved Manufactured in the United States of America First printing May 2004 The findings, interpretations, and conclusions expressed in this report are entirely those of the authors and should not be attributed in any manner to their employers, nor to CGAP, the World Bank, their affiliated organizations, any members of their Boards of Directors, or the countries they represent. The World Bank does not guarantee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. CGAP offices: 1919 Pennsylvania Avenue, NW Washington, D.C. 20006 tel: 202-473-9594 fax: 202-522-3744 email: cgap@wlrldbank.org www.cgap.org Cover photo: Two women holding money © 1999-2004 Getty Images, Inc. All rights reserved. Case Studies Running Head Contents Foreword ................................................................................................................v -- Marilou Uy and Carlos Cuevas, World Bank Financial Sector Network -- Elizabeth Littlefield, Director and CEO, CGAP Authors ....................................................................................................................ix Struggling Through the "Growth versus Best Practice"Tradeoff:..........................................1 The CrediAmigo Program of the Banco do Nordeste, Brazil Robert Peck Christen Equity Building Society: A Domestic Financial Institution..........................................................15 Scales up Microfinance Tamara Cook The Agricultural Bank of Mongolia................................................................................................27 Jay Dyer, J. Peter Morrow, and Robin Young Microfinance in Bangladesh: Growth, Achievements, and Lessons ..........................................47 Hassan Zaman Madagascar: Credit with Education Program in the TIAVO..........................................................65 Savings and Loan Associations Network Herminia Martinez Managing Scaling Up Challenges of a Program for the Poorest:................................................77 Case Study of BRAC's IGVGD Program Imran Matin and Rabeya Yasmin Bank Rakyat Indonesia:Twenty Years of Large-Scale Microfinance ..........................................95 Klaus Maurer Integrating the Poor into the Mainstream Financial System: ....................................................107 The BANSEFI and SAGARPA Programs in Mexico Lisa Taber The Kazakhstan Small Business Program: Commercial Banks ..............................................125 Entering Micro and Small Business Finance Eva Terberger and Anja Lepp The Case of K-Rep--Naairobi, Kenya ..........................................................................................139 John Nyerere Scaling Up Poverty Reduction iii For ward FOREWORD MARILOU UY AND CARLOS CUEVAS WORLD BANK FINANCIAL SECTOR NETWORK The international development community has rallied around the Millennium Development Goals in an unprecedented way. Microfinance supports the MDGs for the simple but powerful reason that many of the MDGs--especially those for improved nutrition, health care, and edu- cation--are the priorities of poor people themselves all over the world. Study after study con- firms that poor families with access to financial services eat better, keep their children in school longer, receive better medical care, and live in safer housing than those who do not have such access, other factors being equal. Access to financial services hands poor people the tools to solve their own problems and to chart their own paths out of poverty. The problem, then, is not establishing whether microfinance works. The problem is how to increase its scale exponentially. Today microfinance is only reaching about four percent of the estimated demand for financial services by poor households. Annual rates of increase are also stuck in single digits. We cannot reach the hundreds of millions of poor people who urgently need access to finance by replicating nonprofit microfinance institutions one branch at a time. Rather, to achieve massive scale and fulfill its huge potential, microfinance must become fully integrated into developing countries' mainstream financial systems, rather than being isolated in a niche of the development community. After all, poor people are the overwhelming major- ity in those countries. It is illogical, as well as unjust, that any country's formal financial system would serve a handful of elites while excluding most of its citizens. The cases prepared for the "Scaling Up Solutions to Poverty Reduction" conference in Shang- hai (May 2004) represent some of the most powerful examples of scaling up in the history of the modern microfinance industry. The cases provide notable diversity in terms of both geog- raphy and institutional type. They range from Central Asia, South Asia, Latin America, and Africa, and include traditional non-governmental organizations, commercial banks, agricultural banks, and building societies. The commonalities, however, are perhaps even more striking than the differences. All the institutions made a conscious management decision to pursue scale while serving poor clienteles. All demonstrated creativity and a willingness to take calculated risk. And all operate in accordance with commercial business principles, regardless of governance struc- ture or legal status. Scaling Up Poverty Reduction v For ward ELIZABETH LITTLEFIELD, DIRECTOR AND CEO, CGAP Last year in CGAP's annual report, I wrote about CGAP's vision of the day when "[t]he term `microfinance' will no longer be necessary as we remove the walls--real and imaginary -- that separate the microfinance community from the much broader world of financial systems, markets, and development." Today we already hear the term microfinance much less often. All over the world, it is being replaced by the phrase "financial systems for the poor." The difference is more than semantic. "Microfinance" suggests that there is something in- herently different about poor people's financial needs, something that is-- and should be--seg- regated from the financial mainstream. "Financial systems for the poor" describes a self-evident truth: that it is neither just nor economically logical that any country's financial system should exclude the majority of its population. In partnership with likeminded financial institutions, donors, service providers, policymakers, and opinion-shapers, CGAP is working to help build democratic, inclusive financial systems all over the entire developing world that serve the ma- jority--the poor. We are finally beginning to see promising examples of a financial democratization that was hard to imagine even a few years ago. In Kazakhstan, commercial and state savings banks are competing to deliver microfinance products to the poor. In Mongolia, the state agricultural bank restructured, privatized, and moved into microfinance. It now serves half of all the rural households in Mongolia through 350 branches. And it is making money. India's second-largest bank is building a network of thousands of village internet kiosks with low-cost card readers, point of sale devices and automatic teller machines to deliver banking and insurance services to poor families throughout rural India. Similar experiments are underway in Latin America and Central Asia. Working partnerships are springing up between institutions who would never even have en- tered dialogues a few years before. In places as different as Haiti, Georgia, and Mexico, part- nerships between banks and microfinance institutions enable the microfinance providers to use banks' branch networks to transfer funds and collect payments. Microfinance institutions are working with insurance carriers to develop and deliver new lines of pro-poor insurance products. Information technology providers and software developers are waking up to the huge market for automation in the microfinance sector--and microfinance practitioners are waking up to the huge efficiency gains they can achieve through technology. Credit bureaux are being established to track--and potentially reward--the financial behaviors of clients who a few years ago were considered beneath notice. Mainstream socially responsible investors are stepping up their investments in microfinance. Ten new funds will be launched in 2004, and the total amount invested is expected to more than double. Governments, too, are keeping pace with this rapid evolution. More than vi Case Studies Foreword 50 countries are developing or implementing new microfinance regulatory frameworks to en- sure that their poor citizens' financial assets are protected. CGAP works to increase the capacity of existing microfinance institutions, to encourage new entrants into the sector, and to broker partnerships that graft the collective wisdom of the mi- crofinance industry--how to make uncollateralized loans and get paid back--onto institutional models that can reach the poor on a much more massive scale. CGAP's flexible approach--strengthening what needs to be strengthened, building what needs to be built--is informed by a sense of urgency. More than a billion people still lack access to basic financial services. Our approach is systemic. In developing countries, we believe the en- tire financial system should serve the poor majority. Our aim is to strengthen a wide range of complementary and competing providers that might serve the poor. In developed countries we aim to help improve aid effectiveness, addressing systemic problems that hinder donor effec- tiveness, especially in supporting the financial and private sectors. And at a global level, we aim to help build the architecture that will underpin large scale, ro- bust and healthy financial systems for the poor--a supportive legal and policy environment and ready availability of high quality financial information. The next phase of microfinance will succeed as the best of different worlds are integrated: the mission focus from the development origins of microfinance with the efficiencies of the banking sector and the force multiplier of commercial capital. It is deeply exciting to see so much of CGAP's vision--the dismantling of the walls--starting to be realized. But as with any wall that finally comes down, it's not so much the final blow that does the job as the accumu-lated impact of all the blows that came before it. Many groups have shared and continue to share CGAP's vision of financial democrati-za- tion: microfinance practitioners, donors, service providers, governments, and not least, the poor themselves. It is CGAP's honor to serve them all. Scaling Up Poverty Reduction vii AUTHORS AND CONTRIBUTORS The writers of the cases are microfinance experts affiliated with or employed by leading practitioner or support institutions. The views expressed are the authors' own and may not necessarily reflect those of their clients or employers. Robert P. Christen is Senior Adviser to the Consultative Group to Assist the Poor (CGAP). He works on issues related to commercialization and regulation and supervi- sion. He is also the Director of Microfinance Training Program (at Naropa University in Boulder, Colorado) and is the Chair of the editorial board of The Microbanking Bulletin, an industry publication devoted to financial sustainability and benchmarking. Before joining CGAP in 1998, Mr. Christen advised commercial banks interested in microfi- nance, central banks and bank superintendencies interested in the regulatory framework for microfinance, and donors interested in performance standards. Mr. Christen also worked for ACCION International. He is the author of several publications related to sustainable microfinance. Mr. Christen received a master's degree from The Ohio State University. Tamara Cook is a Microfinance Analyst with the Consultative Group to Assist the Poor (CGAP). Her work at CGAP over the last seven years includes monitoring CGAP's investments, conducting institutional appraisals, contributing to CGAP's network initia- tive, research for CGAP publications and programs, writing numerous CGAP reports, and working with CGAP senior management on corporate priorities. In October 2003, Ms. Cook was seconded to Equity Building Society in Kenya for a five-month assignment in the credit department. Ms. Cook is currently working on CGAP's scaling up and diver- sification strategies. Ms. Cook graduated from The George Washington University where she studied international development and business administration. She has also taken substantial training courses in microfinance and related topics. Jay Dyer is Senior Development Specialist with Development Alternatives, Inc. (DAI). He has 16 years of experience in commercial banking and financial institution develop- ment, and currently oversees banking projects and contracts for DAI's Finance, Banking and Enterprise Group. He also provides ongoing technical assistance to financial institu- tions. Mr. Dyer has implemented and managed projects that facilitate lending to microentrepreneurs, small and medium-sized businesses, and large companies. Prior to joining DAI, Mr. Dyer spent eight years working in U.S. commercial banking. He received his undergraduate degree in business administration from the University of Richmond. Scaling Up Poverty Reduction ix Authors and Contributors Anja Lepp is Managing Director of Internationale Projekt Consult, a German consult- ing firm that provides long-term management and consulting services to microfinance institutions and commercial banks. Dr. Lepp, who has twenty years of experience in microfinance, has implemented and managed numerous long-term institution-building projects. Among them are some of the most successful microfinance projects in the world. She has also conducted financial sector studies in various countries in Africa, Asia, Latin America, and Eastern Europe. Dr. Lepp currently serves as General Manager of ProCredit Bank Ukraine. She received her doctorate in economics from Johann Wolfgang Goethe University, Frankfurt. Herminia Martinez has been with the World Bank in various capacities since 1970 where her work has focused on banking and finance, privatization and regulation, and business environments. She currently serves as a consultant to the Bank's Africa Private Sector and Financial Sector Units and to the Financial Sector Operations and Policy Department. Prior to joining the World Bank, Ms. Martinez was an assistant professor of economics at Georgetown University. She received her doctorate in economics from Georgetown University. Imran Matin is director of the Research and Evaluation division at BRAC in Bangladesh. His work focuses on developing and managing research activities to help BRAC identify and effectively serve extremely poor clients. Prior to joining BRAC, Dr. Matin worked with the Consultative Group to Assist the Poor (CGAP) and served a research fellowship in the Poverty Research Unit at the University of Sussex. He has published extensively on microfinance issues. Dr. Matin received his doctorate in economics from the University of Sussex. Klaus Maurer is an independent consultant with almost 20 years experience in the fields of rural and microfinance and financial systems development. His clients include multi- lateral and bilateral development organizations such as the Asian Development Bank (AsDB); the International Fund for Agricultural Developmen (IFAD)t; the Consultative Group to Assist the Poor (CGAP); the Swiss Agency for Development and Cooperation (SDC); the German Agency for Technical Cooperation (GTZ); the German Bank for Reconstruction (KfW) and others. Dr. Maurer's work focuses on research, technical assis- tance, training, and designing major projects in support of financial sector development, with a particular emphasis on microfinance. Prior to his work as an independent consult- ant, he was a staff member of KfW and later served GTZ as an advisor in Indonesia. He received a doctorate in Financial Economics from the Free University Berlin. J. Peter Morrow has been chief executive officer since 2000of the Agricultural Bank of Mongolia ("Khan Bank") which the Financial Times Group, London named "Mongolian Bank of the Year" in September 2003. He has 34 years of experience as a banker and financial consultant in the United States and developing countries, overseeing insolven- cies and reorganizations, complex debt restructurings, and troubled company workouts.. In addition to his responsibilities at Khan Bank, Mr. Morrow currently serves as Vice Chair of the Arts Council of Mongolia and as a board member of the Mongolian Bankers Association and the Police-Public Review Board. He was honored as "Entrepreneur of the Year" in 2001 by the Mongolian National Chamber of Commerce; as "Banker of the x Case Studies Authors and Contributors Year" in 2002 by Mongolia's Central Bank; and as "Investment Envoy of the Year" in 2003 by the Foreign Investment and Foreign Trade Agency (FIFTA). He speaks fre- quently to international groups on development and financial reform issues. Lisa Taber is a Technical Specialist with the World Bank's Financial Sector Operations and Policy group. She is responsible for helping develop and execute strategies for improving permanent access to financial services for poor people in World Bank-client countries, with a current focus on Mexico, Nicaragua and Colombia. She also provides policy and technical advice in the areas of microfinance regulation and supervision. Prior to her current position, Ms. Taber served the World Bank's Latin American and Caribbean Region unit as a research analyst and later as an operations analyst. She received an undergraduate degree in economics from Stanford University and is complet- ing work on a Masters of Business Administration from the University of Maryland (degree expected 2006). Eva Terberger is Professor of Banking and Finance at the University of Heidelberg, Germany. She received a doctorate in economics from the University of Frankfurt, Germany. Dr. Terberger's main research interests are development finance with a special focus on microfinance, development of financial markets, corporate finance, and corpo- rate governance. She has participated in numerous conferences and studies on micro- and development finance and has worked as a short-term expert consultant evaluating proj- ects in the financial sector in several countries. She has been a member of the Scientific Advisory Board to the Federal Ministry for Economic Cooperation and Development since 1998, and has chaired that body since 2000. Dr. Terberger has published extensive- ly on key issues of development policy. Rebeya Yasmin has been working with BRAC Development Programme since 1995. She is currently a program manager responsible for all BRAC programs for the ultra-poor. Robin Young is senior microfinance specialist in the Finance, Agriculture and Business Group of Development Alternatives, Inc.(DAI). She provides technical assistance and training to micro and small enterprise finance programs in policy, strategic planning, institutional development, organizational restructuring, financial management and microfinance credit technology. In addition, she conducts designs, evaluations, assess- ments, and financial analyses of microfinance programs. Currently she is the research leader for a large commercial banks in microfinance study. She has written several publi- cations on the regulation and commercialization of microfinance. Before joining DAI, Ms. Young worked as a consultant to the Microenterprise Unit of the Inter-American Development Bank, conducting research on guarantee funds for microfinance institu- tions, and as an Associate of Resource Development for ACCION International. Ms. Young holds a master's degree in business administration from Georgetown University. Hassan Zaman is a Senior Economist with the Finance and Private Sector Development group in the South Asia region of the World Bank. During the last five years at the World Bank, he has worked on various issues, ranging from structural reforms to microfinance, in Malawi, Pakistan, Nepal, and Bangladesh. Prior to joining the World Bank, Dr. Zaman worked for four years on microfinance issues for BRAC in Bangladesh. He holds a doc- Scaling Up Poverty Reduction xi Authors and Contributors torate in economics from the University of Sussex with earlier degrees in economics from the London School of Economics. Dr. Zaman's doctoral research focused on the impact of microcredit on poverty in Bangladesh, and he is the author of numerous publications on microfinance issues. xii Case Studies Chapter 1 Struggling Through the "Growth versus Best Practice" Tradeoff: The CrediAmigo Program of the Banco do Nordeste, Brazil ROBERT PECK CHRISTEN, WITH STEVEN SCHONBERGER AND RICHARD ROSENBERG Executive Summary The CrediAmigo microfinance program mounted by Brazil's Banco do Nordeste (BN) shows how an international financial institution like the World Bank can be a useful cata- lyst in the development of microfinance retail capacity. The World Bank's patient, phased support to BN as it designed, launched, and nurtured CrediAmigo goes against the com- mon perception that multilateral banks always focus on large near-term disbursements to the detriment of longer-term capacity building. In five years, the CrediAmigo program has provided microcredit to over 300,000 of Brazil's working poor and is financially sustainable. Most of these families live on less than two dollars per day, and work in the cities and towns of the northeast region. Even more importantly, CrediAmigo has demonstrated to the development community throughout Brazil that microcredit can be delivered sustainably, on a large scale, and with great impact on low income families. Prior to the success of CrediAmigo, no microcredit program in Brazil over the past 30 years reached more than a few thousand clients. Today, although there are a large number of well-funded public and private initiatives throughout Brazil, CrediAmigo stands alone in the scale and effectiveness of its achievement. Progress so far suggests some lessons for multilateral donors in microfinance: Freedom from dogmatic presuppositions (for instance, "large state-owned banks can never do good microfinance") allows an opportunistic approach that is more likely to yield results. After proper pilot work, a bank with a large pre-existing branch network can roll out mi- crofinance much more rapidly than a new microfinance-only institution. In the case of many public development financial institutions (DFIs), they also have the mandate to do so. Scaling Up Poverty Reduction 1 Struggling Through the "Growth versus best Practice" Tradeoff Outcomes may be better when large negative operational and financial trends. amounts of lending follow, rather than Overall, until now, CrediAmigo has largely precede, the development of proven stayed the course and, as a result, is consid- retail capacity. ered to be a world-class microenterprise credit program. This standing, however, is Donors must help DFIs stay focused threatened by the Brazilian government's on best practice finance principles, es- recent interest rate policy, that could de- pecially when the short term interests prive CrediAmigo of important financial re- of both donor and DFI lie in disburs- sources necessary to re-invest in its future ing large amounts of credit with little growth and stay in keeping with its mission. regard for loan recovery. Donors MUST counter the ultimate Implementation Process tendency of DFIs to grow too quickly Brazil has long been considered one of the and with too little financial discipline. world's great untapped microfinance mar- Donors can be effective with a limited kets. Because of the country's large popula- technical role--setting benchmarks con- tion, high rate of poverty, and open econ- sistent with international best practice, omy, it has the largest concentration of and putting the client institution in con- microenterprises in Latin America--esti- tact with top microfinance practitioners. mated at more than 9 million, with at least 2 million in the Northeast Region alone.1 To accomplish this, generalist donor Despite this large potential market and scant staff working on microfinance activities outreach by the banking sector, in 1998 no should get a basic grounding in the el- Brazilian microfinance program had more ements of sustainable microfinance, than 5,000 clients. And only two programs, preferably through training or, at a both nongovernmental organizations minimum, close work with specialists. (NGOs), could even be considered to be on Throughout the five years that Banco do a path to full sustainability. Nordeste developed its CrediAmigo pro- In 1996 the World Bank decided to ex- gram, it has struggled to balance its desire plore the development of microfinance as to expand the program quickly and reach a part of its poverty reduction efforts in maximum number of clients with the con- Brazil's Northeast Region, the poorest in strictions imposed by maintaining high lev- the country. Since the Inter-American De- els of loan repayment and full cost recovery. velopment Bank was planning a US $150- This tension has been characterized by re- million microfinance apex to finance the few peated episodes of rapid expansion, deteri- NGO MFIs operating in Brazil, the World oration in the quality of its loan portfolio, Bank task team decided to pursue a comple- and low staff productivity, followed by pe- mentary approach focused on developing riods of consolidation and reversal of these the commercial bank model. 2 Case Studies The CrediAmigo Program of the Banco do Nordeste, Brazil The task team spoke with private and gram at BN through an existing loan for public banks operating in the Northeast Re- technical assistance and training. gion to gauge their interest and capacity. Private banks viewed microfinance as char- The World Bank Trains Its Staff ity work rather than as a commercial oppor- It is clear in retrospect that the effectiveness tunity. Public banks were more interested, of the World Bank's engagement with BN given their social mission, but seemed to depended on the development of World provide a weak basis for a financially sustain- Bank staff skills and the close involvement able program. In November 1996, a World of successful microfinance practitioners. Bank team met with senior management at Once the Bank decided to support the Cre- BN's headquarters in Fortaleza to discuss diAmigo pilot, both the task manager and microfinance experience and was impressed his division chief attended the three-week with the commitment shown for obtaining Microfinance Training Program in Boulder, and supporting best-practice microfinance Colorado (USA). These officers say that the operating principles. training helped them focus the dialogue At an early stage in the discussions, the with BN on key elements of success, and to World Bank sought technical help from ex- bring to the table specific technical advice perts on the staff of CGAP. In February from successful, experienced practitioners 1997, a joint World Bank-CGAP mission they contacted through their training in visited BN to evaluate it as a potential mi- Boulder. A number of these practitioners crofinance platform. The team was well were later recruited by the World Bank task aware that microcredit programs in state team and BN. banks seldom succeed. But they found that BN was unusually business oriented and World Bank Technical Involvement seemed relatively free from external politi- is Limited cal interference. The bank was reorganizing During the pilot stage, World Bank and to improve its efficiency through better staff CGAP assistance was limited to helping BN incentives, information systems, client fo- find high-quality international expertise and cus, and flexibility--all key elements for suc- learn from similar experiences in other coun- cessful microfinance. The mission con- tries. BN used World Bank funding to send cluded that these factors, along with the senior managers to study successful MFIs in strong commitment of BN senior manage- Bolivia, Chile, Colombia, and Indonesia. ment and BN's already significant outreach Based on these visits, BN was able to con- in the Northeast Region, outweighed the sider different technical approaches and de- risk of political interference inherent in a velop a short list of consultants. BN chose public bank. Based on the mission's assess- ACCIÓN International, a group with strong ment, the World Bank agreed to provide experience in solidarity group lending. minor funding for a pilot microfinance pro- With ACCIÓN's assistance, BN surveyed Scaling Up Poverty Reduction 3 Struggling Through the "Growth versus best Practice" Tradeoff informal enterprises and developed pilot bank. In view of the high quality of the pro- loan products. It also prepared training ma- gram's management and design, the World terials and selection criteria for its microfi- Bank decided to arrange a US $900,000 nance loan officers. The bank chose to out- Japanese grant to support loan officer train- source the microfinance loan-officer ing, information system development, and function because of the inflexible qualifica- further technical assistance, all in preparation tions and salary levels of its unionized work- for a possible World Bank loan in support of force. Throughout this process the World CrediAmigo's later expansion. Bank's role was limited to administrative as- sistance in managing the funds. CGAP's Overconfidence Leads to Disaster . . . role was limited to helping identify the re- The pilot seemed to go well for the first four quirements for program development, or- months. BN's management was excited by ganizing study tours, and finding potential the program's potential and the positive re- technical assistance providers. Ultimately, sponse in high quarters of the government. however, through its periodic presence, In their understandable enthusiasm, BN's CGAP provided a constant reminder to all senior managers decided to speed up imple- of the best practice operating principles that mentation, expanding CrediAmigo from would ultimately be the key to Credi- five branches to 50. They announced pub- Aamigo's success. licly that CrediAmigo would have 100,000 In December 1997, BN initiated Credi- clients by the end of its first year of opera- Amigo in five of its branches. The pilot in- tion. The World Bank, CGAP, and the pro- corporated lessons from the study tours, a gram's own technical assistance advisors all market study, and technical assistance. Test- warned that four months was far too short ing was restricted to a single loan product: to test the repayment performance of the 90-day loans to individual clients organized new loan product, because defaults typically in "solidarity groups" of about five borrow- rose in later loan cycles. But BN manage- ers who cross-guaranteed each others' ment felt committed to the expansion in or- loans. Payments fell due every 15 days. The der to gain a presence in all regions covered interest rates were considerably higher than by its development mandate. The World BN's rates to its conventional borrowers (as Bank then indicated that further support, high as 6% per month), but far below the including the Japanese grant, would be con- rates of informal money-lenders. Prompt re- tingent on maintaining good portfolio qual- payment was encouraged by offering inter- ity, with 30-day portfolio at risk no more est rebates and new loans within 24 hours than 5 percent.2 for groups that consistently repaid on time. As predicted, the expansion resulted in BN's cabinet chief was named general co- rapid deterioration of portfolio quality and ordinator of the program, with the freedom heavy loan losses. Poorly selected and to recruit top staff from throughout the trained loan officers were given quantity- 4 Case Studies The CrediAmigo Program of the Banco do Nordeste, Brazil based performance targets, so they rushed managers at all levels that microfinance to lend without sufficient focus on repay- portfolio quality was challenging and ment capacity and follow-up. The rapidly volatile, and that they needed to manage it mounting loan losses took BN's managers much more carefully. Since this initial mis- by surprise.3 After two months of expan- step, CrediAmigo has focused consistently sion, the President of BN told all regional on growth with quality. This commitment managers to slow or stop new lending and has been evident in a variety of ways: concentrate almost exclusively on loan re- Growth in the number of branches covery. ACCIÓN worked with BN on loan and loan officers has been carefully recovery strategies, and on retraining loan controlled.4 officers and branch managers. Despite these As a state-owned bank, BN has main- efforts, the episode eventually cost BN tained a commitment to profitability more than US $2 million in loan losses. It in the design and management of Cre- took another six months before portfolio at diAmigo. The program was initiated risk fell back within the agreed limits. with a 5-percent flat monthly rate, translating to a 6.9-percent effective . . . But Strong BN Commitment Gets CrediAmigo Back on Track monthly rate after adjusting for infla- tion.5 Since then, the interest rate has It would have been easy for the World Bank declined in proportion to the cost of to walk away at this point----there were no funds in Brazil, but it has remained at deep or longstanding commitments. BN levels consistent with achieving prof- had done precisely what naysayers had pre- itability.6 dicted at the outset of the project. Few in the informed microfinance community Sustainability required a high-produc- thought the relationship was worth contin- tivity model with low costs and an in- uing. But the task manager and key mem- stitutional culture that would be diffi- bers of the CGAP team sensed that BN cult in a large public development management was still committed to making bank. BN created a "bank within a CrediAmigo sustainable, so they persevered bank," first by outsourcing loan offi- with what from the World Bank's perspec- cers, and then by replacing BN staff tive had become a far riskier proposition. branch managers with coordinators Indeed, it is not unusual for large banks drawn from the loan officer pool. to underestimate the complexity of consol- idating a microfinance program. BN's expe- Finally, the World Bank Makes rience shows that the resulting problems do the Big Loan not have to be fatal, if the focus returns to Because of the World Bank's historical sustainability. The problems and costs of knowledge of the CrediAmigo program, their premature expansion convinced BN and the high quality of CrediAmigo's infor- Scaling Up Poverty Reduction 5 Struggling Through the "Growth versus best Practice" Tradeoff mation systems, the Bank needed only a sin- of these, such as sponsoring the develop- gle preparation/appraisal mission to prepare ment of point-of-service technology a US $50 million loan to support the pro- throughout extensive retail and commercial gram's expansion over the next five years.7 infrastructure have the potential to posi- The mission agreed with BN management tively transform the entire financial land- on performance targets that kept the loan scape of the country. and its disbursement tied to portfolio qual- Others, such as the stated, well-meant ity, efficiency, and sustainability. The World desire of the government to reduce what it Bank's involvement with BN had looked views as excessively high prevailing spreads very risky two years earlier; but by the time in the pre-existing community of microcre- the large loan was prepared, the risk had dit operators threaten the very survival of lessened dramatically because BN now had most of the leading NGO microlenders. a proven microfinance business and had The government has decided to impose a 2 demonstrated the management commit- percent cap on the difference between funds ment needed to keep that business sound as it makes available to MFIs, and the effective it expanded. The World Bank's Board of Di- rate they can charge per month to end rectors approved the loan in May 2000. clients with small loans. These MFIs exhibit Subsequently, BN expanded its outreach a significant dependence on government by increasing the number of outlets. Cred- funds in their liability structure and cur- itAmigo operates in 165 of 175 branch of- rently require a far greater spread to cover fices and through a number of additional their operating costs. points of service in poor neighborhoods. It While CrediAmigo does not utilize funds has negotiated additional funding with for- from the government, and therefore is tech- eign agencies. And BN leads discussions on nically exempt from this cap, BN is a public microfinance in many forums throughout bank whose president was named by the cur- Brazil. Its most enduring problem remains rent government of Brazil. The bank and its the high number of clients that take out one microcredit program CrediAmigo have a very or two loans and drop out. This situation is high profile in the microfinance community, common across microfinance in many coun- but it can not be seen to diverge from the tries and in many institutions, and is a core government position on interest rate spreads. issue that needs to be addressed by the in- If BN accepts those same spread caps for its dustry as a whole. CrediAamigo program, the program may In recent months, CrediAamigo has not be able to cover its fully allocated costs faced its most difficult challenge to date. out of its operating revenue without an inter- The newly installed government of Brazil est subsidy on its sources of funds, much less has announced a significant number of ini- invest the required resources to take on tiatives designed to broaden the access of enduring problems, such as high drop-out the working class to financial services. Some rates, or increase market penetration through 6 Case Studies The CrediAmigo Program of the Banco do Nordeste, Brazil technological innovation and product in the world, in relation to GNP per capita. re-engineering. BN had already been steadily The average outstanding loan balance was reducing its initial interest rates to microbor- R$541.47 (US $270), less than 6 percent of rowers before the government decree, Brazil's per-capita GNP.8 By and large, the though not to the levels that would be con- clients of CrediAamigo have incomes that sistent with the new cap. permit levels of consumption below two In other countries with relatively more dollars a day, while many families are living robust microcredit sectors, interest rates on less than one dollar per day. Firm statis- have steadily fallen as a result of competi- tics are not yet available, although the bank tion, a situation that has not yet arrived in is now commissioning an impact study to be Brazil due to extremely low levels of market carried out with the support from the World penetration. Longstanding experience in Bank to understand more fully the nature of development financial institutions (DFIs) its clients and their household finances. around the world suggests that, in spite of CrediAmigo's portfolio quality and staff their outreach mission, unprofitable pro- productivity are at international best-prac- grams end up starved of resources and, ul- tice levels. Only 3.5 percent of its loans are timately, fail to serve the poor with high late, using a 90-day portfolio-at-risk meas- quality financial products. In today's mod- ure. Its annualized loan loss rate has been ern financial sectors, unprofitable DFIs are steadily increasing from its historical level of ready targets for closure. 2.5 percent, after fully provisioning all loans with any payment 90 days or more overdue, Impact Analysis though it is still just within acceptable pa- rameters. Loan officers with nine months or Within only three years of operation, more of experience are each handling more CrediAmigo was already among the top mi- than 300 clients, on average. crofinance institutions (MFIs) in Latin About 85 percent of CrediAmigo's 108 America, in terms of geographical penetra- branches became operationally sustainable tion, numbers of clients, and depth of out- within three years. The program as a whole reach. As of November 30, 2000, the pro- reached full financial sustainability in mid- gram had over 55,000 active clients in 358 2001.9 By 2002 it had recuperated all of its municipalities throughout the Northeast accumulated operating losses. CrediAmigo's Region of Brazil. Three years later, Credi- returns on assets peaked at 4.7 percent for Amigo has served close to 300,000 clients 2002, and have steadily fallen ever sincebe- with loans through its 175 branch offices, cause it has lowered its interest rates, while and ranks first on the continent in number still operating in positive territory. Thus, of active clients. CrediAmigo is demonstrating that a "down- But most notably, BN has one of the market" focus can be consistent with sustain- lowest average loan balances of any program ability in commercial banking in Brazil. Scaling Up Poverty Reduction 7 Struggling Through the "Growth versus best Practice" Tradeoff To maintain program sustainability in the acquiring and implementing best practice face of the mandate to decrease interest microcredit principles, regardless of the fact rates, CrediAamigo has undertaken to re- that this was a significant departure from duce costs and increase productivity--­twin normal operating practices in the bank. tasks that were desirable under any circum- Above-market interest rates are of- stances as the program has reached greater fered to clients, in order to cover the scale. It reduced operating expenses from relatively high cost of administering 30 percent of total loans to 20 percent over very small loans sustainably. the course of the year 2003. This brings the Compensation of microcredit staff is program within international best-practice based on the results they achieve (per- efficiency ratios. It did so by increasing loan sonal accountability). officer productivity, adjusting the staff in- centives package, and controlling expenses Management information systems give in middle management. microcredit staff immediate access to The latest round of interest rate decreases accurate transaction history and cur- (to 2 percent a month) would make the rent repayment status for all clients. program fall below sustainable levels if com- Credit decisions are decentralized and mercial costs of funds were applied to the li- backed up by ex-post quality controls. ability structure of CrediAamigo, even if some fees were allowed. At 2 percent a The bank is committed to very high month maximum interest and a 4 percent levels of loan recovery. up-front fee, the effective interest rate for Microcredit lending operations are Crediamigo is 38 percent. With an operat- specifically separated from BN's gov- ing cost of 20 percent, loan losses of 5 per- ernment-directed lending programs. cent, and a commercial cost of funds at around 16 percent (1990 CD rate), for a to- The World Bank considered the fact tal cost of 41 percent, the program requires that BN President Queiroz was intend- an interest-rate subsidy to break even. While ing to stay an additional 5 years critical the immediate level of interest-rate subsidy to the successful launch of the program. is not great, it does raise a concern about whether cost cutting measures will unduly Commitment and Political Economy impair the operations of the program over for Change the long term. In September 1996, BN's president Costa de Queiroz told the World Bank that he was Driving Factors interested in developing a world-class mi- crofinance program. Emerging from a ma- Several factors led to the success of Credi- jor reform, BN had R$6 billion (US $3 bil- Amigo. The single most important was the lion) in assets and 176 branches throughout commitment of the President of BN to 8 Case Studies The CrediAmigo Program of the Banco do Nordeste, Brazil the Northeast Region. To satisfy BN's re- BN invested heavily in staff training gional development mandate, Queiroz was and information systems. looking for ways of reaching the poor that It was into this environment of far-reach- were more effective than the bank's directed ing institutional reform that CrediAmigo lines of credit had been. He was particularly was placed. interested in the informal sector. Dr. Costa de Queiroz was appointed Institutional Innovation president of BN in March 1995, following The program was ultimately put under the a very successful career as a private business- direct management of Queiroz's chief of man in the state of Ceara. Dr. Queiroz im- staff, which ensured that it would receive mediately began a major reform of the the unconditional and full support of the bank, with the objective of making it more President as it was deployed throughout the modern, efficient, and responsive to the de- branch network and operating divisions of velopment needs of the Northeast Region. the bank. CrediAmigo has been managed Loan assets grew from R$2.6 billion to from the start by a small group of fewer more than R$6.0 billion; the number of than 25 full-time headquarters staff. Line loans grew from 68,000 to more than managers were drawn from bank staff in 404,000. branches. Banco do Nordeste outsourced BN's market share increased from 35 the contracting of its cadre of loan officers percent of all loans granted in the and their immediate supervisors in order to Northeast Region to 78 percent. circumvent restrictions imposed by collec- Administrative expenses as a percent- tive bargaining agreements that would have age of assets were less than half their made microcredit unprofitable and unsus- 1995 level (3.4 percent in 2000 versus tainable within the normal staffing patterns 7.9 percent in 1995). of the bank. These outsourced loan officers and coordinators represent 85 percent of Staff were reduced from 5,468 to total staff of the program. fewer than 4,000. As the program was developed and im- Lag time between loan request and ap- portant issues arose, CrediAamigo drew on proval was reduced from an average of senior managers from all departments in the 217 days to a mandated limit of 21 to bank. This guaranteed a higher level re- 60 days depending on loan size. sponse when compared to the traditional bank downscaling approach, which has been BN reorganized its structure and to build and staff a separate division within processes to make the client its central the bank wholly dedicated to microfinance. focus and reduce barriers to staff inter- The added advantage of this approach has action across functional areas. been the far deeper levels of integration of Scaling Up Poverty Reduction 9 Struggling Through the "Growth versus best Practice" Tradeoff the program and its culture into the bank. Both occasions reveal the fundamental co- President Quieroz always saw in microfi- nundrum faced by large, public institutions nance an opportunity to reinforce broader when they confront the undeniable fact that reforms within the bank. This has been ac- they have not been successful in meeting complished. BN managers have repeatedly the full demand for high-quality financial said that the example of CrediAmigo is hav- services in their target market. They con- ing a catalytic effect on the rest of the bank. stantly face the trade-off between just push- BN is using the experience gained under ing services out and incurring the higher CrediAmigo in areas such as the use of staff costs that ensue from high loan losses, and incentives and the development of a low- trying to stick with best practice methodolo- delinquency loan culture. gies and incurring the higher costs of staff training that are required to keep loan losses Learning and Experimentation low. Thus far, BN has struggled through this On at least two occasions, Banco do dilemna ultimately favoring best practice in Nordeste has attempted to grow too its decisions, at the clear expense of program quickly, which has led to excessive losses and growth. It has not been easy, and BN is the need to re-establish core operating pro- constantly looking for ways to grow more cedures and principles. In each case, BN has quickly and increase its impact. This is, after been able to pull back and consolidate its all, its clearly stated mission. operating procedures and re-establish best practice performance. The first occasion External Catalysts was triggered by the need felt by the bank Successive Brazilian governments have to establish the presence of the program demonstrated a longstanding interest in throughout the Northeast Region, where it promoting microenterprise credit through operates and has a development mandate. a series of public-sector initiatives. Access to This decision was costly in terms of extraor- microcredit has frequently been an impor- dinary loan losses, but the bank has since tant issue in political campaigns. While the recuperated all lost funds through profits government's initiatives can typically be generated by the program. characterized as leaning toward subsidized The second occasion was prompted by the and targeted development credit, the Cre- desire to grow the program more quickly, in diAmigo program resulted from a fortunate part due to the excess liquidity represented coincidence between the ongoing reform by the additional funding negotiated with process in BN and the interest and ability of development organizations and by the desire the World Bank staff member to pursue a of BN to expand its outreach and the impact radically different approach. Certainly, BN of the program. This decision also led to a felt the mandate from the highest govern- sharp increase in loan losses, although not as ment authorities to engage in microcredit. severe as in the first occasion. This continues to the present day, and the 10 Case Studies The CrediAmigo Program of the Banco do Nordeste, Brazil current administration has launched a num- activities. It spent about US $5 million of ber of new initiatives designed to broaden its own funds for the salaries of the BN per- access to finance, some of which fall into the sonnel who designed the program, training classic subsidized credit paradigm while oth- costs, full financing of the loan portfolio, ers, are quite forward looking. and accumulated losses, including the US $2 million write-off of bad loans.10 The BN experience suggests that donors Lessons Learned may need to be opportunistic rather than The activities most critical to CrediAmigo's dogmatic in their approach to microfinance. development into a world-class microfi- At first the World Bank was reluctant to nance program included conceptualization, support a large state-owned development design, piloting, and initial consolidation. bank, even one that had undertaken signif- The most notable aspect of these activities icant internal reforms. But the demon- is that they required limited external fund- strated strength of the commitment by BN ing and occurred before any disbursements senior management, along with a pressing from the World Bank loan. Although these need to develop substantial microfinance re- processes required significant investment tail capacity in the Northeast Region, and from BN over almost three years, external the Inter-American Development Bank's assistance costs were relatively low. "preemption" of the NGO arena, all com- From 1996 to the loan appraisal in 1999, bined to persuade the World Bank to take a the World Bank spent about US $150,000 risk and support CrediAmigo. Although the of its own budget, including the costs of Bank was advised to withdraw its support, participating in the initial seminar, the particularly after the ill-advised expansions observation missions (usually two-days), the in 1998, the clear commitment of BN's Microfinance Training Program in Boulder, senior managers to develop a world-class one identification mission, and the project program kept the World Bank engaged. appraisal mission. CGAP provided about In comparison with other recipients of US $50,000 in technical support, including donor support, managers of competent staff time and external consulting. In all, banks are particularly skeptical of, and resist- BN received US $1.2 million in external ant to, technical advice from donors or from financial assistance (US $300,000 of consultants selected by donors. Recognizing reprogrammed funds from a prior World this, the World Bank team limited its direct Bank loan and US $900,000 from the role to providing focus rather than technical Japanese), although it spent less than half advice. Nevertheless, it took the unusual step of that on technical assistance related to of investing in a three-week training program program operations. for two of its key staff. In retrospect, the WB Other funds were spent on impact team members felt that this training was evaluation assessments and client training critical to the success of the engagement. Scaling Up Poverty Reduction 11 Struggling Through the "Growth versus best Practice" Tradeoff Recruiting outside experts is important but resolving its early problems and developing cannot substitute for the role of donor staff a sense of how to manage the trade-off as the authoritative interlocutors for their in- between growth and portfolio quality with- stitution. Donor staff who lackbasic techni- out the pressure of disbursing a World Bank cal literacy in microfinance will not usually loan has brought the program to a level be able to identify good consultants or use where neither BN nor the World Bank see them well, and may not have the credibility the US $50 million loan as an incentive to that is essential for an effective dialogue with "supply-led growth." the implementing institution. In recent years, BN management has The most important contribution of negotiated further lines of financing from the World Bank management was its patience in Interamerican Development Bank and the allowing development to proceed at its own "Germans." These funds raise the total fi- pace. The process of developing, piloting, nancing available for CrediAmigo to and consolidating the CrediAmigo pro- close to US $100 million, and seem to gram, particularly with the need to recover have provided some of the impetus for the from the early rapid expansion, took three mid-2003 push to grow the loan portfolio. years before a loan was appraised. This re- As with the initial push in 1999, this push led quired exceptional patience from the World to a deterioration in the quality of the loan Bank's country management in Brazil, portfolio, resulting in a spike in non-recov- which was under pressure from various erable loans, a decrease in loan-officer pro- sources to move more quickly with the loan. ductivity, and a drop in overall profitability Although there was some consideration of for the year. Profits for 2003 were consider- moving more quickly with a smaller loan, ably down from their high in 2002. While management continued to support the task BN management has not allowed this team's assessment of the program's readi- degraded portfolio quality to persist, this ness for a major World Bank loan. episode reveals how the constant pressure As a result of this patient approach, a applied by the donor community to support larger loan was provided with less risk of successful programming, combined with undermining the program's focus on the temptation on the part of government sustainability. The initial development agencies to grow the same, can combine to process allowed BN to develop its expertise blow up initiatives that, left on their own, and confidence in managing a large micro- would prosper at a more natural pace. finance program. Even the early missteps, The CrediAmigo experience suggests based completely on BN decisions with that the World Bank and other multilateral consequences financed by BN resources, donors can play a catalytic role in microfi- were an important element in establishing nance development if they: the program's low arrears culture in a Pursue the best available opportunities public bank environment. BN's success in in a country rather than impose a 12 Case Studies The CrediAmigo Program of the Banco do Nordeste, Brazil universal model. (All capacity-building Encourage donors to take advantage approaches involve substantial risks.) of their technical role to set bench- marks consistent with international Allow programs to develop their ca- best practice and put institutions in pacity to manage growth with portfo- contact with top microfinance practi- lio quality before providing significant tioners, which can be the critical ingre- funding for program expansion, even dients that keep a DFI on track. when this may delay lending targets. (The biggest constraint to the spread Ensure that donor staff working with of microfinance services is a shortage, MFIs are grounded in the basic ele- not of funding, but rather of compe- ments of sustainable microfinance, tent retail capacity.) preferably through training or, at least, close work with high-quality specialists. End Notes 1 For background on Brazil's microenterprises, finan- 6 When BN transfers funds to a CrediAmigo branch, it cial sector, and microfinance industry, see Steven charges an internal transfer price equal to the prevailing Schonberger, Microfinance Prospects in Brazil, Latin rate on Interbank Certificates of Deposit. America and the Caribbean Region and Economic and 7 In fact, BN could have funded the five-year expan- Socially Sustainable Development Unit (Washington, sion from its own resources. The World Bank loan was D.C.: World Bank, 2000). still attractive to the Brazilian government because it 2 "Portfolio at risk" is a measure that divides the brought in needed foreign exchange. remaining balance of loans that are late by the remain- 8 At the end of 2000, BancoSol in Bolivia had 61,000 ing balance of the whole loan portfolio. loan clients, after 13 years of operation. Compartamos 3 Because microcredit is unsecured, repayment disci- in Mexico had 64,000 clients, after about 10 years. The pline can collapse very rapidly. When micro-borrowers median outstanding loan balance for Latin American sees many of their peers defaulting on their loans, moti- MFIs was 45 percent of per-capita GNP, according to vation to repay plummets. The main reason borrowers the April 2001 issue of MicroBanking Bulletin. In repay their loans is the expectation of future services. Bangladesh, Grameen Bank's average outstanding loan They know that once people stop repaying, the service is US $140, or about 40 percent of per-capita GDP. will not be available very long, and they do not want to 9 "Operational sustainability" is the ability to pay all be the last one aboard the sinking ship. operating costs except the cost of funds from interest 4 To increase its geographic outreach cost-effectively, income on loans. "Financial sustainability" is the ability CrediAmigo began establishing "individual branches" to pay all costs, including financial costs, from interest consisting of a single loan officer. These branches income. account for most of the growth in branch numbers since 10One can argue that BN did not need to spend this August 1999. amount of money, that it could have achieved the same 5 In comparison with normal loans, microloans are tiny, results with a third of the up-front investment. But per- yet staff intensive. Thus, administrative costs are high haps these types of "mistakes" are normal for working when measured as a percentage of the loan amounts. It through very large institutions. The alternative might takes a very high interest rate to recoup these costs, have been heavy handed (and expensive) technical especially if management is committed to breaking even assistance, which the Brazilians would not have accept- at an early stage. Experience has shown that clients ed. There is evidence of banks that have spent far less, value the loans so highly that they are willing to pay ele- such as the Banco del Estado in Chile, which made vir- vated interest rates; in fact, many of them have been tually no mistakes after spending a few years looking for paying much higher rates to informal money-lenders. the right model, and banks that have spent far more, such as Bank Rakyat Indonesia, which had to turn around the performance of 3,500 branches. Scaling Up Poverty Reduction 13 Struggling Through the "Growth versus best Practice" Tradeoff Bibliography This Case Study was prepared by Robert Peck Christen http://www.bndes.gov.br/conhecimento/microfin/02gol based on an earlier publication, A Multilateral Donor dm.pdf Triumphs over Disbursement Pressure: The Story of Kumar, Anjali. Brazil: Access to Financial Services. Microfinance at Banco do Nordeste in Brazil, CGAP Washington, D.C.: World Bank, forthcoming July 2004. Focus Note No. 23 (Washington, D.C.: CGAP, December 2001), http://www.cgap.org/docs/FocusNote_23.pdf, Nichter, Simeon, Lara Goldmark, and Anita Fiori. written with Steven Schonberger and Richard "Understanding Microfinance in the Brazilian Context." Rosenberg. Mr. Schonberger is Senior Operations Rio de Janeiro, Brazil: Banco Nacional de Officer working in the World Bank's Cambodia Country Desenvolvimento Econômico e Social, Programa de Office. He previously worked on rural development and Desenvolvimento Institucional, July 2002. micro and rural finance issues as part of the http://www.bndes.gov.br/english/studies/03livreto_eng- Environmental and Socially Sustainable Development lish.pdf. Department in the Latin America and Caribbean Region of the World Bank, based in Washington D.C. Mr. Parente, Silvana. "Microcredit Policy as a Financial Rosenberg is a Senior Advisor at CGAP. Although Market Mechanism to Reduce Poverty: The Experience reordered, most sections are taken verbatim from that of the Banco do Nordeste in Brazil." Photocopy, Banco earlier CGAP publication. Other more recent references Nacional de Desenvolvimento Econômico e Social, Rio include: de Janeiro, Brazil. Unpublished paper prepared for the Special Program for Urban and Regional Studies of Brusky, Bonnie. "From Skepticism to Success: The Developing Areas, Department of Urban Studies and World Bank and Banco do Nordeste." CGAP Case Planning, Massachusetts Institute of Technology, June Studies in Donor Good Practices, No. 3, Washington, 2002. D.C., April 2003. http://www.cgap.org/direct/docs/ case_studies/BancoDoNordeste.pdf Sanchez, Susana M., Sophie Sirtaine, and Rita Valente. "Bringing Microfinance Services to the Poor: Goldmark, Lara, Steve Pockross, and Daniele Vechina. CrediAmigo in Brazil." World Bank En Breve 7 (August "A situação das microfinanças no Brasil." Rio de 2002). http://inweb18.worldbank.org/ Janeiro, Brazil: Banco Nacional de Desenvolvimento External/lac/lac.nsf/0/1FE8954E8345086D85256C0800 Econômico e Social, Programa de Desenvolvimento 4FEA3E?OpenDocument. Institucional, 2000. 14 Case Studies Chapter 2 Equity Building Society: A Domestic Financial Institution Scales up Microfinance BY TAMARA COOK, CGAP Executive Summary Equity Building Society (Equity) is a homegrown success story that has only recently attracted international attention. Equity's beginning, 20 years ago, was inspired by an entrepreneurial vision of a potential demand for financial services in the underserved, low-income population of Kenya. Despite high hopes for serving this market, the first 10 years were characterized by a difficult environment, fierce competition, and a lack of institutional knowledge of how to operate a profitable financial institution. On the verge of collapse in 1993, Equity brought in outside experts and committed to radical steps to turn the institution around. This commitment led to improved financial performance and increased outreach. Since 1993, Equity has grown from 12,000 depositors to more than 250,000 depositors (as of December 2003), making Equity Building Society home for more than 13 percent of all bank accounts in Kenya. This growth can be attributed to committed staff and lead- ership, high-quality customer service, effective marketing, low barriers to access (especially compared to traditional commercial banks), appropriate product design, an acceptable en- abling environment, and more recently to external support from donors. Equity also offers a wide range of loan products and other financial services, such as money transfers, banker's checks, and bid bonds. In the last few years, Equity has gained recognition within Kenya, and throughout Africa and the world, for its dramatic growth and professional operations. Of course, success brings its own challenges. In the coming years, Equity will need to manage the sustainability of this impressive growth by incorporating new staff, supervising the expanding branch network, maintaining the quality of customer care, and strengthening the quality of its loan portfolio. Implementation Process Equity Building Society opened its doors in 1984 to bring financial services to ordinary Kenyans, especially to those who had no access to the formal banking sector. The founders hailed from senior positions in government and business, and shaped the institution based on their indepth understanding of the socio-political circumstances, their commitment to the Scaling Up Poverty Reduction 15 A Domestic Financial Institution Scales Up Microfinance government's rural development policy, and stood at 5.8 percent, far below the required their ability to turn a commercial profit 20 percent. However, the Central Bank in Kenya. Taking advantage of new rules of Kenya did not request closure of that enabled Kenyans to open formal, Equity, and the Society was given the licensed financial institutions, Equity was chance to turn around for the sake of its registered as a building society, which was an existing clients who were still committed affordable option in terms of license fees to the institution. and capitalization. The building society Realizing the need for a radically new and legislation influenced Equity to offer savings professional approach to retail banking, the services and mortgage loans. Equity realized, Equity Building Society board recruited however, that it was servicing a microfinance two experts, a promising banker (who market for low minimum-balance deposits later became the finance director) and an and loans that were rarely used for housing. experienced trainer, to help staff meet the In its early years, Equity faced fierce challenges of the new environment, expand competition from a multitude of locally- its marketing, and on the whole enhance its owned financial institutions, which also were efforts. With the help of the new manage- set up during this time. Equity resisted the ment team, Equity focused on staff training temptation to attract deposits from govern- and marketing. The original culture of good ment and parastatal organizations in contrast client service and teamwork was revived, to other financial institutions. Instead, and the number of depositors grew. Equity mobilized individual customers by The efforts to turn Equity around were one-on-one marketing, entrenching a also assisted by the new political atmosphere culture of service excellence from the start. that took hold in the early 1990s. The For its first decade of operations, Equity country was moving away from a one-party struggled to maintain clients and cover costs, political system, and more freedom was evi- while other small institutions closed and dent. Equity succeeded in mobilizing deposits confidence in the sector eroded. from churches, and the government allowed Despite Equity's dedicated efforts in a employees to bank at financial institutions not difficult environment, a report issued by the under government control. Many foreign Central Bank of Kenya in 1993 confirmed banks converted from retail to corporate that the institution was technically insolvent banks, leaving a vacuum in the retail sector. At and had poor board supervision and inade- the same time, banks were closing their rural quate management. Non-performing loans branches and rationalizing urban branches, were 54 percent of the portfolio, and attracting a flood of new deposits. accumulated losses totaled KSh 33 million Equity also formalized its commitment against a paid-up capital of KSsh 3 million. to smaller clients in a new mission statement At this stage, deposits were being used to which recognized how financial services meet operating expenses. Its liquidity ratio contribute to the welfare of clients as well as 16 Case Studies Equity Building Society, Kenya the national economy. Deposit mobilization turnover and portfolio growth. In 2003, increased by 40­60 percent per year, and Africap, a regional microfinance investment profits grew. More recently, a special fund of international financial institutions, marketing and public relations campaign in chose Equity for its first African investment of August 2003 celebrated Equity's 20th US $1.54 million, accompanied by technical anniversary of "Providing Financial assistance of US $244,000. Solutions to Kenyans," which yielded A fitting illustration of the turnaround spectacular increases in the number of new of Equity is the improved ratings by the savings accounts. Central Bank of Kenya. The overall rating During its first 15 years, Equity relied (shortly after the 1993 rating) rose to entirely on its own resources and had marginal, then to fair, and in 2002 to neither solicited nor been offered assistance satisfactory. The Equity board and manage- from international partners. This began to ment were deemed proficient to govern and change in 1999, when Equity undertook manage a financial institution. Capital two small projects with EU-MESP and adequacy and asset quality were satisfactory. UNDP-MicroStart. Deeper partnerships Management, earnings, and liquidity were were developed with Swisscontact and rated as strong. MicroSave-Africa. Both relationships have The 2003 report from the Central Bank focused on providing local and regional has just been finalized, and its overall finan- technical expertise and on jointly develop- cial condition has been rated as satisfactory ing new products and services. MicroSave- again. Capital adequacy, earnings, and Africa, in particular, brought a focus on liquidity were all rated as strong. However, services and product design based on client asset quality dropped to marginal (19.2 demand. This led Equity to redesign its percent portfolio at risk) and management products, reflecting its change to focusing was rated as satisfactory. The senior on clients. The marketing department was management team has taken this report very restructured, and eventually included a seriously and is revamping the credit great deal more time on market research. operations and methodologies to improve DFID's Financial Deepening Challenge credit risk management and administration. Fund supported the launch of Equity's The report also noted deficiencies in mobile banking program. DFID currently the management information system and provides technical assistance for specific improvements are already underway. targets, and Equity also receives more The profile of Equity, as of March 2004, general technical assistance from Africap. shows an institution that is consistently prof- The UNDP-MicroStart program also itable, operates through 15 branches and 25 played a role in upgrading and computerizing mobile units, has 297,000 depositors with the management information system, which deposits of KSh 3.8 billion (US $50 million), has had a tremendous impact on Equity's and a loan portfolio of KSh 2.1 billion (US Scaling Up Poverty Reduction 17 A Domestic Financial Institution Scales Up Microfinance $28 million). It is fully computerized and sizes can point to income and accumulated has 384 staff members, 8 directors, and wealth of the type of clients being reached. 2,470 shareholders. As of December 2003, 73 percent of Equity's savings account balances were less than US $70, not including the 50,000 new Impact Analysis accounts that have not yet received their Financial services help poor and low- first deposit.2 Although most of Equity's income households increase their incomes savings accounts have very low balances, 63 and build the assets that allow them to percent of Equity's total deposit volume is mitigate risk, plan for the future, increase made up of corporate savings accounts and food consumption, and invest in education, fixed deposits over US $1,000 (see Figure health, housing, water, and sanitation. 2). Given the volume of clients with small Equity made a deliberate decision to focus accounts, Equity staff spend most of their on poor and low-income clients to help time handling small depositors. However, them improve their lives. Although Equity this does not preclude the relatively large reaches a wide range of clients, the major- deposits that help provide stability and of- ity of its clients are small-holder farmers, ten longer-term funding to the institution. low-end salaried workers, and micro On the credit side, about 80 percent of and small businesses. Figure 1 shows the outstanding loans have outstanding balances breakdown of loans according to these of less than US $400, although they only categories1 as well as a summary of deposits account for about 18 percent of the total by account type. outstanding loan balance. (See Figure 3 Although it is not a perfect indicator for showing average outstanding balances poverty level of clients, average account on Equity's loan categories.) On average, Figure 1: Distribution of Clients Types of Depositors Types of Borrowers Ordinary Farmers Business Salaried Fixed Dep Small/Micro 5% 3% Development 4% Mobile 1% Other 3% Corporate 9% 16% 29% 81% 49% Sources: EBS data, own calculations 18 Case Studies Equity Building Society, Kenya Figure 2: Distribution of Deposit Size Numbers of Depositers u $1-30 Balance of Deposits $1-30 $30-70 $30-70 $70-270 $70-270 $270-$1000 7% 1% $270-$1000 2% $1000-2700 9% 16% $2700-13,300 Over $1,000 Over 13,300 32% 11% 18% 9% 64% 20% 11% Sources: EBS data, own calculations *Overdrafts and new accounts with 0 balances have been excluded Figure 3: Average Loan Sizes (US$) 1,400 1,279 1,200 1,000 $922 800 600 400 $201 200 $132 0 Farmers Salaried Small/Micro Development **Corporate loans for medium businesses average $9,500 small-scale farmers take out the smallest average outstanding balances (US $201). loans (US $132), which are usually evaluated Loan repayments are deducted automatically on the basis of predicted remittances from from salaries that are directly deposited the Kenya Tea Development Authority or into clients' savings account at Equity. other farm produce marketing agencies. Equity's Corporate Branch in Nairobi caters Farm input loans are also provided for dairy, to a higher-end market, where business loans coffee, and other crops on a smaller-scale. average US $9,500. However, in other Salary-based loans (including education branches, loans to smaller businesses average and medical loans) have slightly higher US $1,300. Equity's business loans have Scaling Up Poverty Reduction 19 A Domestic Financial Institution Scales Up Microfinance been based on revenues demonstrated by medical provider. Equity encourages saving turnover in savings accounts and secured with for education through its "Super Junior" collateral such as land and vehicles. A new account for children that includes free microcredit product targeting even smaller banker's checks for school fees--a unique entrepreneurs is slated for pilot testing in service in the Kenyan market. It also offers 2004. It will require reorientation of credit favorable pricing on issuing banker's methodology (including more flexible cheques required for the payment of school security requirements with a strict emphasis fees each term. The Jijenge savings account, on the analysis of regular cash flows) and a monthly contractual savings product, substantial training for credit officers. helps clients develop the discipline to In mid-2003, a qualitative impact assess- save regularly for education and medical ment3 was conducted in conjunction with expenses (among other reasons) and also Equity's market research team (using allows clients to borrow against their savings MicroSave-Africa tools), and examined the at a favorable rate. nature, scope, and depth of client impact at Equity's impact can also be viewed the individual, household and enterprise level. through the lens of its financial return. By The research team? conducted 23 focus group the beginning of 1994, Equity had accumu- discussions from eight rural and urban lated losses of KSh 33 million (US branches. The findings provided insight into $440,000). Given that it held only KSh how clients use Equity's services to manage 31 million in deposits, this represented a risk and meet household and business cash serious problem. After the dramatic shift in flow demands. Equity's clients report using its operations in 1994, Equity reported its savings and loans to invest in business activi- first profitable year, and by 1997 Equity ties (inventory, salaries); save for the future had accumulated a profit of US $80,000, (education, medical costs); manage household which grew to US $5.2 million by the end cash flow needs (rent); invest in assets (land, of 2003 (see Figure 4). In the 1980s and housing, appliances); and cope with crises or 1990s, Equity's growth was financed life events (illness, marriages, funerals). In the entirely by mobilized savings (domestic) focus group discussions, clients reported and the personal investment of the founders growth in their incomes and assets over the and shareholders. Survival meant they had past ten years, and Equity appears to be an to use their resources efficiently, especially important part of that growth. to dig themselves out from the crisis in Equity's commitment to increasing edu- 1993. The recent technical assistance cational opportunities and addressing and financial investments in Equity are healthcare needs for Kenyans permeates the beginning to reap results, but it is too early institution. Education and medical loans are to measure the full impact. offered at lower rates if the funds are paid directly to the educational institution or 20 Case Studies Equity Building Society, Kenya Figure 4: Accumulated Profit/Loss (US $ in thousands) 6,000 5,000 4,000 3,000 2,000 1,000 0 93 94 95 96 97 98 99 2000 2001 2002 2003 -1,000 Driving Factors to the target market. This helped develop Equity's success can be attributed to the institutional capacities without overstretch- dedication of its management and staff, as ing management, operations, and internal well as to Equity's ability to innovate and controls. Equity's recent growth is built on maintain extremely high customer service. the foundation of high quality services and Equity's success attracted limited external strong institutional capacities. support in 1999, and more substantial Equity's founding and turn-around support after 2002, but support followed benefited from external reforms catalyzed by and reinforced the "homegrown" success key decision makers who shared Equity's rather than creating it. vision. Equity was established under new rules in 1980s to create locally-owned Commitment and Political Economy formal financial institutions with the hope for Change that such institutions would serve the Internal and external commitment to unbanked populations neglected by the change pushed Equity in the right direction. commercial banks. In the early 1990s, Internally, Equity's management and board when Equity was finding its footing again, governed the institution effectively during multi-party democracy and financial and a period of generally weak governance in market liberalization helped Equity attract Kenya. The deliberate decision to concen- new clients, such as government employees trate on the smaller end of the market who were (finally) allowed to direct deposit shielded the organization from political their salaries in non-government controlled influence. In addition, Equity's leadership financial institutions. Influenced by focused operations in one single region World Bank/IMF financial sector reform (Central region) to provide the best services requirements and lobbying by building Scaling Up Poverty Reduction 21 A Domestic Financial Institution Scales Up Microfinance societies (especially Equity), Parliament the selection of staff is not so much to bring revised the Building Societies Act to widen skills into the institution as to mold inexperi- the scope of activities almost in line with the enced graduates to its operational norms and Banking Act. For example, the 1999 revision customer-service philosophy. This has worked regularized deposit taking. And last, but not exceptionally well and is reflected in the work least, regulators who were sympathetic to ethic and culture emanating from the activi- Equity's mission granted it one last chance ties of the entire staff complement. Staff are to shape up after it was rated "technically encouraged to see their work as a "calling," insolvent" in 1993 by the Central Bank of not just a job. This view unifies staff around Kenya rather than shut it down. Equity's vision and has the added benefit of inducing staff to work long hours to achieve Institutional Innovation this vision when needed. Arguably, Equity's most essential institutional Equity is committed to bringing financial innovation is motivating and sensitizing its services to rural clients, which make up 68 staff to embody the company pledge: "...take percent of Equity's clients but only 28 pride in the noble responsibility of offering percent of total deposit volume, given the financial services and solutions, to empower relatively lower average savings account size our customers to face the future with dignity of rural clients (see Figure 5). Building on and realize our full potential." Equity has a its existing branches in small towns in tradition of recruiting young people at entry Central Province, in 2002 Equity launched points, who are well educated but have little mobile banking (with support from DFID) or no experience. The emphasis in Equity is to deliver services deeper into rural areas. on instilling the corporate culture. In a way, The mobile units are designed as stand- Figure 5: Distribution of Deposit Size Distribution of deposits Rural-branch Distribution of deposit (Ksh) accounts (#) Rural-mobile Urban 26% 37% 2% 58% 2% 5% 22 Case Studies Equity Building Society, Kenya alone mini-branches, wholly contained in a difficulties of a manual information system, fortified sport utility vehicle with satellite which were amplified at every level of connection to the branch, solar power, and growth. Both customers and staff members a fold-down teller window. The mobile felt the strain of the manual system as the mini-branch drives up near factories, for volume of Equity's business expanded over example, when salaries are disbursed and the years. Equity launched its computerized serves clients directly from the vehicle. management information system in June More often, however, the vehicles are 2000 (with support from UNDP- used to transport 2­3 staff over rough rural MicroStart), completing the installation in roads to rented offices where Equity opens a record four months. Equity's efficiency in temporary mobile branches at least one day collecting and reporting data and its a week. These minimally- equipped rented service delivery to customers improved locations generally have a counter for staff greatly thereafter. With the new system, and an area for clients to wait for service. The Equity managed to improve its customer mobile branch downloads the information turnaround time from 30­40 minutes to for their clients from the branch onto a about five minutes at the counter. laptop (but also brings a hardcopy of the Although Equity's growth is partly at- data in case there is a shortage of electricity tributed to its marketing and customer-fo- or they lose the satellite connection with the cused efforts, it is clear from its high growth branch). All transactions by the mobile office spurt in 2001 that the new computerized are uploaded at the branch before running system has been a major factor. This, of the close-of-day report. The mobile branches course, demonstrates the importance of are a cost-effective way to provide better technology to banking in general and to access to rural clients without the significant high volume, low-value microfinance in investment needed to open a full-scale particular. Currently, Equity is investing branch. As of December 2003, Equity had additional resources in a wide area network 25 mobile units serving 12,161 clients to connect all the branches on a real-time (5 percent of existing clients) with an basis, in data warehousing to enable better average deposit size of about US $100, reporting and to prepare for potential (a significantly lower average than for other credit scoring, and in upgrades to its savings products). The mobile units earn a information system to gain more function- profit, in part by charging a small fee to ality and reliability. clients, which is less cost than transportation to the nearest branch. Learning and Experimentation Innovations in information technology Since its inception, Equity has been can dramatically affect an institution's abil- relentless in its constant focus on clients. ity to serve its clients well. For over 16 Of late, Equity has marketed itself as "the years, Equity survived despite the growing listening, caring financial partner" in an Scaling Up Poverty Reduction 23 A Domestic Financial Institution Scales Up Microfinance aggressive marketing campaign including 90 percent of their accumulated savings radio, newspaper, and television, not to account for urgent needs. The product was mention the often more important role designed using client input, pilot-tested, re- of informal marketing by key members fined through discussions between head of- of society. Far from just a marketing cam- fice and branch management, and success- paign, the slogan represents how Equity fully rolled out with support from the runs its business. marketing team. Learning and experimentation play key roles in Equity's ability to attract and retain External Catalysts clients. Equity "listens" to its clients through Since 1999, Equity has selectively accessed various mechanisms for their suggestions external resources that propel it towards its about current products, needs for new serv- mission. Key medium- to long-term donors ices, and perceptions of Equity. Focus group and investors formed a steering committee discussions have been an effective method to provide more coordinated support and for listening to clients, and have resulted in guidance to Equity. This harmonized improved service as well as demonstrated approach encourages complementary inputs Equity's "caring" to clients in a tangible way. by donors and investors and lessens the Equity's bright and energetic marketing transaction costs for Equity in managing team received training from MicroSave on relationships with external supporters. designing and conducting these discussions Equity has received technical support and for maximum impact. (MicroSave has also funding from Africap, British Department conducted "mystery shopping," after the for International Development, European launch of new products to determine how Union, MicroSave, SwissContact, and well staff know and deliver the products, as UNDP-MicroStart. a valuable feedback loop.) Equity managers have an open-door policy at the branches Lessons Learned and head office, and welcome hearing about Equity's 20-plus years of ups and downs of- clients' experiences with Equity--another fer lessons for institutions, donors, and gov- way of listening to clients. ernments committed to scaling up financial Equity's approach to learning from services to the poor in other countries. clients contributed to the successful launch of the Jijenge savings account, a contractual Lessons for Financial Institutions savings product to help clients "realize their · Understanding poor and low-income dreams." It was designed to meet the needs clients is essential for identifying "niche" of lower-end clients seeking discipline markets, designing appropriate products to save for needs, such as school fees, and services, and developing and weddings, and household items. Jijenge maintaining client loyalty. clients can also access cheaper loans of up to · Appropriate technology can increase 24 Case Studies Equity Building Society, Kenya efficiency (i.e., computerizing informa- and expansion of indigenous financial in- tion systems) and outreach (mobile bank- stitutions with the potential to reach scale. ing for better penetration and delivery in · Non-bank financial institutions can play rural areas.) an important role in reaching unbanked · Investing in staff training and developing clients with appropriate services. In meaningful incentives helps staff to inter- Kenya, building societies were given a nalize and be motivated by the vision mandate to operate like banks. But in of the institution, which in turn leads to other countries in Africa, building soci- tangible results. eties have been limited in their ability to · Visionary leadership and top management, scale up. while vital to the success of an institution, must be supported by strong and techni- Equity's Future: Managing the cally competent middle management. Sustainability of Scaling Up At the end of Equity's first decade of oper- Lessons for Donor Aagencies and ations, it was a defunct building society. Service Providers During its second decade, Equity emerged · Donors should seek a consortium approach as an organization with great potential. As to funding with a medium- to long-term it enters its third decade, Equity will commitment and emphasize results. need to transform into a high-performance · Donors can help institutions manage institution. The main challenges will growth by offering targeted and timely be managing growth, transforming the technical assistance that is consistent with institutional culture of what was a family- the institution's own vision and objec- like business, and balancing the enormous tives. Donors should focus on strategic success in mobilizing deposits by expanding support for optimal impact. and improving the quality of its lending · Domestic financial institutions, even non- operations. The continuing growth and banks, exhibit great potential as a mecha- expansion will require significant enhance- nism for massive outreach. Donors should ments of Equity's organizational architec- explore relationships with local partners ture, especially at the governance level (e.g., who have demonstrated capacity to grow advancing the executive oversight function through visionary leadership, existing in- of the board of directors). Planned frastructure and client base, and brand upgrades to the lending methodology will recognition in local markets. need to be rolled out including a more serious emphasis on recovery and portfolio Lessons for Governments and quality. Internal controls need to be Policy Makers strengthened and an effective asset liability · Government can create enabling environ- management system developed. One of ments for the creation, proper regulation, the essential ingredients in sustaining the Scaling Up Poverty Reduction 25 A Domestic Financial Institution Scales Up Microfinance institutional growth is building up the ade- of managing these challenges and should quate internal capacity of appropriate continue to scale-up its success in the years staffing and middle management capacity. to come. Equity has demonstrated that it is capable Figure 6: Number of Deposit Accounts (in thousands) 350 300 250 200 150 100 50 0 94 95 96 97 98 99 2000 2001 2002 2003 Mar- 04 End Notes 1 Development loans are for longer-term construction and land development provided to all three categories of clients. 2 Equity reduces its barriers to entry by allowing clients to open accounts without a minimum deposit. 3 UNCDF Impact Assessment report. Bibliography Coetzee, Gerhard, Kamau Kabbucho, and Andrew erated reports from Banker's Realm, marketing materi- Mnjama. Understanding the Rebirth of Equity Building als, and other Equity documents. Nairobi, Kenya. Society in Kenya. Nairobi, Kenya: MicroSave-Africa, Negre, Alice, and Josephat Mboya. PlaNet Rating: August 2002. Equity Building Society, Nairobi, Kenya. Paris, France: Craig, Kim, and Ruth Goodwin-Groen. "Donors as PlaNet Rating, June 2001. Silent Partners in MFI Product Development: Wright, Graham A.N. and James Mwangi. Equity MicroSave-Africa and the Equity Building Society in Building Society's Market-Led Approach to Kenya." CGAP Case Studies in Donor Good Practices, Microfinance. Nairobi, Kenya: Forthcoming, 2004. No. 8. July 2003, http://www.cgap.org/direct/ resources/case_studies.html United Nations Development Program. Client Impact Assessment Report. New York: 2003. Equity Building Society. Annual Report and Accounts, 1999, 2000, 2001, 2002, and draft 2003. Internally gen- 26 Case Studies Chapter 3 The Agricultural Bank of Mongolia BY JAY DYER, J. PETER MORROW, AND ROBIN YOUNG Executive Summary The Agricultural Bank of Mongolia (Ag Bank or XAAH) is the main provider of financial services in the rural areas of Mongolia. It has the largest branch network in the country, with 379 locations (93 percent in rural areas), and provides deposit and loan products at each point of service. Although it was in receivership in 1999 and faced possible liquida- tion, this former state bank has been completely turned around and was privatized through international tender to a major Japanese company in March 2003. Turnaround efforts have resulted in Ag Bank disbursing 878,000 loans between late 2000 and February 2004, while maintaining an arrears rate consistently below 2 percent and becoming the most profitable bank in Mongolia, with a return on equity of 44.19 percent in 2003. As of February 2004, 128,227 loans were outstanding for a portfolio of almost US $50 million with US $75.5 million in 377,424 deposit accounts. The 15,433 domestic transfers totaled US $260,000 for the month. The average outstanding loan balance is US $382, deposit accounts average US $200, transfers US $17, and half of Mongolia's households do business with Ag Bank. This turnaround identified and mobilized the strengths of an existing institution to rapidly disseminate desperately needed financial services to the rural areas and protect access to the few existing financial services upon which many rural Mongolians rely. After many years of operating deficits, loan losses, and a failed attempt at privatization, Ag Bank was placed in receivership in 1999, and many in the international community felt that it could never operate sustainably and should be closed. However, the importance of Ag Bank to Mongolia's rural sector can not be overstated. Although one of the smaller financial institutions in the country, it was the only bank with branches throughout Mongolia's vast territory through which to transfer money, make government pension and salary payments, and accept deposits. Closing it would have had a devastating impact on the rural economy. The World Bank made reforming Ag Bank a condition of its Financial Sector Adjustment Credit Program for Mongolia, and the US Agency for International Development (USAID) Scaling Up Poverty Reduction 27 The Agricultural Bank of Mongolia agreed to provide funds for an outside The rural areas of Mongolia have been the management contract. The Government bank's main focus, and its successes there have of Mongolia agreed to provide this outside laid the foundation for substantial growth and manager with full authority to manage new services for all Mongolians. Ag Bank's the institution, free from political or other current mission statement is "to be the prin- interference. cipal nationwide financial services company in A contract was signed with Development Mongolia by delivering first-class products Alternatives, Inc., (DAI) of Bethesda, Mary- with the highest level of customer service." land, USA, to manage the bank. In July Key to the turnaround of the institution was 2000, a team led by Chief Executive Officer developing and implementing products J. Peter Morrow arrived in Mongolia. throughout the country profitably. Ag Bank DAI staff and advisors provided technical has shown that financial products can be support as the team set the agenda for the created and delivered gainfully in even scarcely turnaround and transformed Ag Bank. By populated and poor areas if they truly meet agreement among all parties, the mission of the needs of customers. The institution has the turnaround was to: (1) restore financial proven that it can sustain itself and, in fact, soundness to the bank; (2) bring financial contribute significantly to Mongolia's overall services to the country's rural population; economic development. Aside from the large and (3) prepare Ag Bank to operate impact of its loan and deposit activity, the independently and be privatized. opening of 110 new branches and creation of The management team developed a new more than 1,000 good paying jobs boosted lending program, converted payment services the economies of many communities. In into deposits, created an extensive marketing addition, Ag Bank is now one of the largest program to improve the bank's image and taxpayers in Mongolia where before it was a attract clients, implemented strong controls cash drain on the government. through new policies and procedures, estab- lished a more effective management structure, Implementation Process and significantly increased training activities. Following almost a decade of political interfer- In January 2003, the Government of ence and mismanagement that led to Mongolia received three viable bids for Ag losses and insolvency, a new governance struc- Bank, all from strong and fully qualified pri- ture combined with an independent and vate sector buyers. H.S. Securities of Japan professional management team turned Ag was the highest bidder, and at US $6.85 Bank into a profitable and rapidly growing million became the successful purchaser of private bank. Following is a timeline of the key Ag Bank. The sale closed on March 25, events in Ag Bank's history: 2003. The new owners have hired DAI to 1991. The bank was founded as the continue managing Ag Bank, a contract that Agricultural Cooperative Bank. The is fully paid for out of the bank's income. government arranged for agricultural 28 Case Studies The Agricultural Bank of Mongolia cooperatives, herders, and farmers to 2000. Development Alternatives, Inc., capitalize this new bank, which took of Bethesda, Maryland, USA, won an over 326 rural branches and settlement international competition to manage centers. It had 2,600 employees, held a the Ag Bank. After the contract was portfolio of approximately US $2 mil- signed in July 2000, a team led by lion mostly in non-performing loans Chief Executive Officer J. Peter Mor- and deposits whose interest rates were row and Chief Operating Officer De- set by the Central Bank, and bra Boyer arrived in Mongolia to man- assumed the activities of the former age the Ag Bank. A new operating monopoly State Bank. strategy was implemented and prof- itability achieved in 2001. By agree- 1992. The deterioration of Ag Bank's ment among all parties, the mission of business and the movement toward a the turnaround was to: market economy in Mongolia forced a reorganization of the institution. The - Restore financial soundness to bank expanded its lending to larger loans the bank because of government interference. - Bring financial services to the country's rural population 1996. The Ag Bank's liquidity and finan- - Prepare the bank to operate inde- cial position deteriorated to a level where pendently and be privatized the Central Bank appointed a receiver. 2003. H.S. Securities was selected in an 1999. Existing shareholders' interests international bid as the buyer of Ag were eliminated, and the Central Bank Bank and retained DAI on a manage- put together a restructuring plan. The ment contract. Growth and profitability government became the new sole owner continue to increase. through a capital infusion from govern- ment restructuring bonds, converting debt to equity, and injections of cash. Objectives: Sustainability and Privatization 1999. The World Bank made reforming At the beginning of the management the Ag Bank a condition of its contract, the team knew it would be crucial to Financial Sector Adjustment Credit (1) develop internal staff capacities, (2) Program for Mongolia, and USAID develop proven products for the market, and agreed to provide funds for an outside (3) find a successful resolution to the political management contract. The govern- influence that dominated Ag Bank for years. ment of Mongolia agreed to provide The initial project goal was privatization. this outside manager with full author- By stabilizing the institution and creating ity to manage the institution, free from transparency, a buyer would clearly understand political or other interference. what was being bought. By developing and Scaling Up Poverty Reduction 29 The Agricultural Bank of Mongolia expanding the product and service offerings in Thus, a critical part of the restructuring the market, the Ag Bank would have the prof- plan was to insulate the bank from this itability to sustain itself. As a measure of suc- political interference. The agreements cess, the bank was privatized in March 2003. between the government of Mongolia, H.S. Securities, the buyer of Ag Bank, Central Bank, World Bank, and USAID has stated that it wants to continue to required the government not to interfere expand based on the bank's target in any Ag Bank operations. Other than markets. It will be investing millions of continuing Central Bank supervision and dollars more into the bank to continue monitoring, no governmental entity could the outreach and market penetration. take any steps to influence Ag Bank. Since privatization, Ag Bank has continued In effect, normal corporate governance to grow rapidly, practically doubling was suspended for the remediation period. the loan portfolio from US $24.7 to An independent Board of Directors (made US $49 million during the first year. up of two members from the government of This was supported by a 110 percent Mongolia, two nominated by USAID, and capital increase, from MNT 3.9 billion an unaffiliated professional Chairman) was to MNT 8.2 billion (US $3.32­$6.98 appointed to monitor the remediation million at actual exchange rates), through process and ensure the government did not reinvesting all earnings plus a fresh capital interfere. The Board, a semi-annual meet- injection from H.S. Securities of US $2 ing between the management team and million. This commitment and financial donors, an annual independent audit, and capacity is the best assurance that access to Central Bank supervision formed the ad these financial services throughout the hoc corporate governance structure for the country will continue. restructuring period. This arrangement was essential to break the pattern of interference Political Issues that had contributed significantly to the Before the intervention of outside manage- bank's demise. ment, Ag Bank had never escaped the close Shortly after the project started in July supervision of political authorities, particu- 2000, there was a complete change in gov- larly at the local level. Throughout the ernment from the Democratic Party back to 1990s, soum (county) and aimag the former Communist Party. Although ini- (province) governors continued to appoint tially suspicious of the agreements made for the bank's local managers and considered Ag Bank's remediation, the new govern- them part of the local municipal manage- ment eventually supported it fully. ment team. Loans were directed or heavily influenced by local political needs. The Credit Culture head office had relatively little control over No bank in Mongolia had previously Ag Bank's offices and their operations. attempted to develop and implement a fully 30 Case Studies The Agricultural Bank of Mongolia transparent lending program available to all essary to turn the bank around to a position who qualify. Some donors suggested that of growth, efficiency, and profitability. Al- Mongolia's loan history was so bad that though the employees were well educated Ag Bank should be remediated without and adept at performing most day­to­day lending. However, the team felt that it banking tasks, it was clear that a new organi- was critical to make loans, otherwise, the zational culture would not develop bank would not recover financially or overnight and would be an ongoing process. restore services to the public. Through The key was to harness the capabilities of the the appropriate design of financial products staff right away and then expand on their and the right staff motivation, a new credit skills. Management wrote new policies and culture was developed. procedures for all areas of the bank and Previously, most people in Mongolia had developed a training program. Given the to rely on pawnbrokers and family members regional and branch operating structure of for loans. Thus, Ag Bank's new lending Ag Bank, training-of-trainers programs were products were truly welcome as they gave used. The program focused on training at people access to larger sums of money at the aimag level, which served as provincial or a lower cost. In the past, loans that were regional hubs, and then the aimag staff secured through banks were associated with would train the rural locations that reported the government. Hence, institutional pres- into that hub. Thus, a fast and efficient way sure to repay on time was often lax, and the of disseminating new products and other culture of timely repayment among poten- communication was developed. Because of tial borrowers was weak or nonexistent. the "state culture" to follow the rules, For successful growth and profitability, people caught on quickly and responded to this culture had to change, from the level of the changes. the branch manager to the customers. The In addition, each aimag center branch first step in meeting this challenge was manager comes to the head office in Ulaan- to hold the branch managers completely baatar once a quarter and is given substantial accountable for the decisions they made. Ab- one-on-one time with the senior management solute zero tolerance was given to politicians of the bank. Targets for lending and deposit locally or nationally who tried to influence generation are set quarterly. Administrative the lending. Because an outside management issues also are handled at this time. Thus, the team was running Ag Bank, it was easier to targets are set mutually by the regional build and sustain this firewall. branches and the home office so they are realistic yet meet the bank's overall strategy Staff Culture and objectives. Some felt Ag Bank had too much of a Achievement of targets is reinforced "state sector" mentality and that the staff through the incentive system that has would not be able to accept the changes nec- become a critical part of the success of the Scaling Up Poverty Reduction 31 The Agricultural Bank of Mongolia institution. Every employee receives incen- market, ensuring diversification in product tive compensation on a quarterly basis that is and geographic area. The idea, from the a significant percentage of his or her salary. beginning, was to pilot test products quickly Fully changing the culture from one fo- and then to expand delivery rapidly via loca- cused on tasks to one where priorities are tions countrywide, while always maintaining customers and revenues still is evolving, but a focus on quality lending. This is different is clearly much improved. from the strategy of a slower national rollout and a focus on quick saturation at each Brand and Image of the Ag Bank branch. All Ag Bank products were designed A major obstacle was changing the public's to be integrated into the branches by existing perception of Ag Bank. Although the insti- staff who are responsible for delivering a wide tution had many advantages, the bank still array of these products. Combined with suffered from its reputation as a politically setting quarterly goals mentioned above, this driven, untrustworthy, and insolvent organ- strategy ensures that appropriate product ization. The first approach to change the offerings and combinations are tailored to public's perception was to provide valuable each market. products and services to customers as soon as possible. The one-percent withdrawal Loans fee was immediately dropped (which had Ag Bank continues to identify market op- been a major impediment to mobilizing portunities for new loan products. Starting bank deposits), and rates on time deposits with working capital loans for micro and were raised to the top of the market, quickly small businesses just four months after the giving Ag Bank the liquidity it needed. management team took over, they have ex- The management team brought in panded to loans for medium enterprises, external marketing expertise and launched pensioners, and herders, plus payroll deduc- a major public relations campaign to an- tion loans and agricultural loans. Currently, nounce throughout Mongolia that Ag Bank the bank is piloting mortgage loans. was serious about business and was open to Although most of the loans are in local all. Likewise, the bank rebranded itself currency, some new loans are available in by switching its name to an acronym, US dollars. Prudent credit policies ensure "XAAH." This translates to khan or king, a that lending decisions are based primarily name with strong local appeal because on a client's cash flow followed by collateral of Genghis Khan. Television and radio to ensure low default rates. Following are campaigns soon followed. descriptions of key loan products. Micro and small business loans. This Product Strategy product is the mainstay of the lending The strategy has been to develop products program in all markets. The original prod- that meet the needs of a large segment of the uct, launched in November 2000, was the 32 Case Studies The Agricultural Bank of Mongolia "small trader loan," which evolved into the have substantial cash at certain times of the "small business loan" as the market for the year, penetration of the herder market with product was widened to include small serv- comprehensive training and banking services ice, production, and mixed purpose busi- is a priority. The average outstanding herder nesses--essentially any kind of micro or loan is US $722. small business. Although initially only for Agricultural production loans. Intro- short-term working capital, qualified and duced in May 2002, these loans have experienced borrowers can now access term garnered interest primarily from vegetable loans. Current interest rates are 2.2­4.0 growers, small private wheat farmers, and percent per month and are decreasing be- hay producers. The average outstanding cause of competition and efficiency gains in agricultural production loan is US $604. underwriting from growing economies of Payroll loans. Borrowers can borrow scale. With micro and small business loans up to seven times their monthly salary for a starting at US $80, the average outstanding variety of purposes for a term up to one loan size for this product is US $1,419. year. Loan payments are made directly Small and medium-size enterprise through a deduction from their regular (SME) loans. This product was developed salary. These loans provide significant for medium-size production companies. Be- cross-marketing opportunities for Ag Bank. cause these loans are larger in size and term, Employers realize that the bank can handle they are inherently more risky. They are be- their entire payroll under a mutually ing made on a limited basis now while the beneficial agreement: employers save on lending team for SME loans is trained and administrative costs and bring employer de- the products are tested in the market. The posits and fees to the bank. The average collateral on some of these larger and outstanding payroll loan is US $218. longer-term business loans consists of per- Pensioner loans. Pensioners were coming sonal residences or business assets. A signif- into Ag Bank branches monthly to with- icant expansion of this product is expected draw their payments, an indication that this in 2004. The average outstanding small market was underserved for loans. There business loan is US $5,011. were also enormous transaction costs for Herder loans. These loans were specifically paying so many small pensions per month. designed to meet some of the unique needs of As part of the pension disbursement reform nomadic Mongolian herders, while taking effort in 2001, a loan product was devel- their cultural and business differences into oped for this market. These small loans, consideration. Available on terms of up to one which can be up to six times the value of a year, they help cover the gap between living monthly pension, are repayable by assign- and operating expenses in the months when ment and automatic deduction from future herders are not generating income or wish to pension payments. These loans empower purchase herd-related goods. Because herders pensioners to control their cash flow, enable Scaling Up Poverty Reduction 33 The Agricultural Bank of Mongolia them to borrow money to help their family out fees. The account pays interest. members venture into microbusinesses, and Time deposits. This product pays a are an alternative to the traditional practice higher interest rate than a savings of borrowing from pawnshops and the account. Deposits are encouraged to be informal sector. The average outstanding held for at least three months, and pensioner loan is US $69. incentives are given to savers to add to their deposits over the term. Deposits The deposit strategy since DAI took over Pension direct deposit. This program was management of Ag Bank in 2000 has been contractually agreed to with the Social threefold: (1) to increase the deposit base to Insurance Fund in early 2001. This fund the bank's growth; (2) to attract a meant opening accounts for each pen- large and diverse number of customers sioner, which provided them with an (business and consumer) to solidify the additional service and raised the deposit bank's franchise and strengthen the net- base of Ag Bank by opening hundreds of work; and (3) to lessen the dependence on thousands of new accounts. government sources of funds. Payroll direct deposit. Similar to the By creating products that meet the needs pension direct deposit, this product is of the market, dropping withdrawal marketed to large employers that want fees, and providing good customer service, to eliminate the administrative expense Ag Bank has been able to increase its of paying employees in cash. In addi- deposit base 740 percent since July 2000. tion to increased fee income for the This growth has in turn propelled bank, the bank also benefits from the the bank's expansion. The dependence on many new individual deposit accounts government deposits has dropped from and cross-selling opportunities. 52 percent of all deposits at the end of December 2000 to 6.7 percent at the end of January 2004. Money Transfer Products All deposit products are offered in the All transfer products are available to anyone, local currency, tugrug, or US dollars. but Ag Bank customers receive reduced Following is a summary of deposit products: fees. Following is a summary of the bank's Personal or business current accounts. transfer products: These accounts are designed for those Quick Pay This franchise, which has who are making regular financial transac- been developed throughout Mongolia, tions and who need regular statements. has been key to the success of Ag Bank. This product guarantees fast de- Savings accounts. Depositors may add livery (in three hours or less) of cash and withdraw money at any time with- transfers between offices in the capital, 34 Case Studies The Agricultural Bank of Mongolia Ulaanbaatar, and any one of the 77 on- The total portfolio outstanding is US line locations throughout the country. $49 million. The average loan size is US $382. Money transfer. Money can be trans- Ninety percent of all lending is made in ferred from one Ag Bank location to rural areas, across different population any other Ag Bank location in the sectors. country. A transfer is delivered the next Deposits grew 740 percent from about day to a regional center and within US $9 million to US $75.5 million; three days to a rural center. the 377,424 deposits have average Western Union. Money transferred balances of US $200. through Western Union can be received There were 15,433 monthly domestic at any Ag Bank in Mongolia, and sent to transfers, totaling US $259,796, and more than 180 countries worldwide. the average transfer was US $16.83. Monthly pre-tax profits grew from a loss to a current average of US Impact Analysis $300,000, with a return on assets of Ag Bank operates a network with 379 2.96 percent, and a return on equity of points of service throughout Mongolia, 44.19 percent in 2003. much greater than any of the other 16 Ag Bank paid US $2.9 million (at banks operating in the country. With 354 of current exchange rates) in income taxes its offices in the countryside, the bank in 2001, 2002, and 2003 combined. reaches 98 percent of the rural communities The management team developed a new in Mongolia. lending program, converted payment serv- The results from the new products intro- ices into deposits, inaugurated an extensive duced and other restructuring initiatives marketing program to increase deposits, have been impressive. Below are some of established strong controls through new the remarkable statistics for the period July policies and procedures, structured manage- 2000­February 2004: ment more effectively, and significantly Bank offices grew from 269 to 379 increased training activities. Products include points of service. loans, deposits, domestic and international The number of employees increased transfer instruments, and government pay- from 803 to 1,833. ment services. Clients are micro and small One of every two Mongolian house- businesses, herders and farmers, consumers, holds uses Ag Bank. and government organizations. Almost 900,000 loans were made with Ag Bank has designed and offered mi- arrears consistently under 2 percent. croenterprise loans, SME loans, crop and (Portfolio at risk over one day was 1.83 herder credits, and pension and salaried- percent as of December 31, 2003.) Scaling Up Poverty Reduction 35 The Agricultural Bank of Mongolia based loans (see Table 1). Individuals who deposit accounts. Experience at Ag Bank do not qualify for the bank's small business shows that one-third of all government loans often use a consumer loan to get payments made through a deposit account started in a new venture. Ag Bank has stay in the account for an extended period. created a new class of bank borrowers. By This conversion brought more deposits to recognizing the informal lending sector as a the bank and helped build relationships with real competitor, Ag Bank has developed a new customers who could then take advan- service that encourages many borrowers to tage of other financial services and loans. successfully move into the formal financial Money also has moved from under mat- system. Ag Bank's branches are quickly re- tresses to the bank, witnessed by the growth placing pawnshops, store owners, and rela- of individual deposits from just over US $2 tives for business and consumer borrowers. million to US $54 million between Decem- As a result, families and businesses are able ber 2000 and February 2004. The average to borrow more money at better terms and deposit account at Ag Bank is US$168. Most lower costs. As of February 2004, 878,976 of the new depositors are people who previ- loans had been disbursed and 128,227 ously did not have accounts in banks.(See were outstanding. Tables 2 and 3 for a breakdown in the distri- Another example of an important change bution and evolution of deposits.) has been the conversion of 200,000 govern- Ag Bank's aggressive growth of deposits ment social security and salary payments into and loans, and the subsequent expansion of Table 1: Ag Bank Loan Porfolio (as of February 2004) Total Total Value Total Total Value Average Date Number of of Loans Number of of Loans Loan Loan Product Product Loans Disbursed Loans Outstanding Oustanding Launched Disbursed (US $000s)* Outstanding (US $000s)* (US$) Micro and Small Nov 2000 80,835 98,780 13,485 19,135 1,419 Small & Medium May 2001 3,300 15,555 1,633 8,184 5,011 Pensioners May 2001 628,431 36,912 72,277 5,003 69 Herders Aug 2001 31,373 20,753 9,449 6,819 722 Payroll Based Oct 2001 132,784 34,177 30,539 6,654 218 Crop May 2002 1,483 746 148 89 604 Mortgage Apr 2003 770 3,752 696 3,144 4,518 Loan Product 878,976 210,674 128,227 49,028 382 36 Case Studies The Agricultural Bank of Mongolia other banks' operations throughout Mongo- it provides internal resources for increased lia, have had an impact on the intermediation investment. of Mongolia's money supply. In 2000, more Another measure of Ag Bank's impact is than half the country's money supply was in the growth of the small businesses that currency held outside of banks; today, less dominate rural Mongolia. Average small than one-third is held as currency. The business loan sizes have increased steadily long-term benefit of this intermediation for over the past three years as increased inven- Mongolia's development is considerable as tory levels have translated into increased Table 2: Ag Bank Deposit Porfolio (as of February 2004) Average Outstanding % of Total Deposit Number of % of Total Accound Balance Outstanding Category Depositers Depositers Balance (US $000s) Blanace (US$) Organizations 15,849 20.99 14,354 3.8 1,104 Current 12,024 15.92 14,318 3.79 840 Time 2,823 3.74 22 0.01 128,347 Demand 1,001 1.33 14 0.00 71,471 Individuals 59,667 79.01 363,070 96.20 164 Current 2,044 2.71 266,875 70.71 8 Time 47, 265 62.59 39,971 10.59 1,182 Demand 10,358 13.72 56,224 14.90 184 TOTAL 75,515 100.0 377,424 100.0 200 Table 3: Ag Bank Deposit Evolution (in US $000s)* Deposit Category 31 Dec 2003 31 Dec 2002 31 Dec 2001 31 Dec 2000 Business 9,730 13.2% 9,798 22% 8,117 33% 3,663 28% Government 4,293 6.7% 8,207 18% 8,570 35% 6,712 52% Individuals 50,225 80.1% 27,028 60% 7,752 32% 2,520 20% Total 64,248 100% 45,034 100% 24,439 100% 12,896 100% *Note: Most loans and deposits are in the local currency, tugrug, and have been converted here to US dollars at the current exchange rate of 1,174 Scaling Up Poverty Reduction 37 The Agricultural Bank of Mongolia sales. Today, with loans to micro, small, and Commitment and Political Economy medium enterprises, Ag Bank's average out- for Change standing business loan is US $1,807. Management's Independence. From a cost recovery perspective of the A critical part of the turnaround was getting original work that began in July 2000, the assurances from the government that the new value of Ag Bank has increased so much management team could operate free from over the period that it could have easily paid political influence. The new management for the cost of the turnaround by the value team was given the authority over personnel that was created--witness the bank's sale issues, credit policies, and expenditures price and the ongoing management con- through a memorandum of understanding. It tract between HS Securities and DAI. The also was given capital forbearance (that is, nor- earnings in 2003 were close to the entire mal minimum capital and capital adequacy cost of the three-year, donor-financed turn- standards were suspended) for the period of around. Today, Ag Bank is the second most the restructuring agreements. profitable bank in Mongolia, with average monthly net income of US $300,000 Central Bank Guidelines (including the cost of the DAI management Early on, the turnaround team reached contract that is paid out of earnings). In agreement with the Central Bank as to when 2003, Ag Bank ranked number one of all Ag Bank's financial ratios would comply with Mongolian banks in terms of return on prudential standards, and a conservative assets (2.96 percent) and return on equity timetable was put in place. Management met (44.19 percent). all Central Bank guidelines a full year before the end of remediation and has continued to Driving Factors exceed all requirements. The government of Mongolia's commitment Privatization to change Ag Bank, as part of its overall Although the outside management had the privatization and economic modernization authority and responsibility to run Ag Bank, program, was key to success. Its strategy the concern was that problems could arise of hiring international consultants and priva- when they left, when the "firewall" came tizing the bank was the right recipe. The down. This is why it was so critical that experienced management team, coupled one of the pre-conditions previously dis- with the commitment of the local staff and cussed--government commitment to priva- strong business strategy, provided the skills tization of the bank--was held to. The gov- to achieve the projected outcomes. Finally, ernment of Mongolia agreed to turn the the support and pressure from external institution over to private investors for forces encouraged political fortitude and accountability, and the initial management kept the project on track. contract ran until the bank was privatized. 38 Case Studies The Agricultural Bank of Mongolia The government of Mongolia carried out its when they are asked to circumvent policy; privatization agreement by organizing an and (3) staff's willingness to use the foreign international competitive tender, on the ad- managers as a way not to comply with inap- vice of KPMG Barents (now BearingPoint), propriate requests. The first encounter with another USAID contractor. Most observers an official who wants his way and cannot get found the process fully transparent and it may be difficult, but when people learn successful. that the new management operates trans- parently, they usually will not ask for favors Local Staff or bribes. Taking control of the decision process did not end with ministry and senior Institutional Innovation management. Previously, local branches An experienced management team using functioned as part of the province or county a sound business strategy, upgraded and administration. Branch managers were retailored those products, services, and appointed in most cases by the local gover- operations that took advantage of the vast nors of the ruling parties. Appointment as a branch network. branch manager was a source of jobs and loans for people who desperately needed Managerial Leadership. them. Based on the culture, they were more The skill sets and competency of the senior likely to first accommodate the need than to management were critical. The team understand the implications from a business brought together experiences, knowledge, perspective. Managers at Ag Bank branches and management skills gained from working now understand the importance that the in countries with similar economic situations, institution be strong, which will increase as well as indepth knowledge and experience outreach to even more viable clients and, of US banking. In addition, some of thus, stimulate the overall economy. They the most experienced Mongolians were also know that the senior management team recruited for the executive team, who pro- holds them accountable for their actions, vided critical technical and cultural knowl- but that if they perform well, they will be edge for the transformation of the bank. The compensated fairly and receive proper management team set the tone and led the incentives and bonuses for performance. organization to develop a culture based on The foreign managers also recognized operating discipline and service innovation. that the national staff needed insulation Management featuring clear lines of from local pressures, as they would become authority was the joint effort of expatriate the next logical targets. The key to over- and local professionals. The starting point of coming these obstacles and temptations has Ag Bank's turnaround was precarious. Deci- been: (1) clear policies that require trans- sions had to be reached and implemented parency; (2) personnel who will speak up quickly, and it was critical to deploy individ- Scaling Up Poverty Reduction 39 The Agricultural Bank of Mongolia uals with relevant outside experience who At the beginning, more than 50 percent of were not beholden to domestic political deposits were from the government. This re- interests. Local departmental managers liance put Ag Bank's liquidity at potential risk hired or promoted to the executive level should the government withdraw significant were highly valued for their institutional funds. The bank sought and attracted deposits memories as well as their keen understand- from private businesses and individuals. The ing of the domestic market and culture. result has been a shift from 52 percent govern- The project could not have been successful ment deposits to the current 6.7 percent. without this synergy between outside and The private sector deposits are made up of inside management. hundreds of thousands of individual deposi- tors, which created greater diversification. Products and Services Before restructuring a bank and introducing Operational Efficiency new financial services, it must be confirmed If faced with challenging financial targets, the that an underserved market exists with first reaction of many new managers is to start untapped demand sufficient to support the re- cutting costs. A common problem at state- structured bank. If the prospective market is owned banks, however, is that not enough is being adequately served by others, the cost spent to generate required revenues. There and effort of a turnaround are probably not are costs associated with delivering products justified and could crowd out existing finan- and offering an acceptable level of customer cial institutions. In Mongolia, studies were service that are critical in a competitive conducted to understand the general market market. It was found that small additional ex- potential before the turnaround contract penditures (primarily staff training and minor began. The fact that a strong, informal improvements to the banks themselves, such market offered loan rates of 12­15 percent a as roofs and purchases of furniture and com- month revealed a demand that banks were not puters), combined with prudent lending at serving. Every project needs a golden goose, the initial 269 Ag Bank points of service, a competitive advantage. In the case of could lead to exponential growth in revenues. Ag Bank, a real need for sustainable financial With only US $300 profit from each small services existed in the rural areas of the point of service and the contributions from country. The bank was able to leverage exist- larger regional centers, it was not long ing branches and staff to deliver services before US $300,000 was earned each month. quickly and efficiently. These numbers were substantial for a bank without capital and with initial assets of Diversification only US $10 million. Nonetheless, with In addition to offering new products, one im- client-focused services, prudent lending, and portant shift in the product and client mix was professional management, the results were diversification away from government reliance. quite achievable. 40 Case Studies The Agricultural Bank of Mongolia To generate profitability, however, the in- ularly unsavory reputation in the eyes of the crease in revenues needed to be met by cut- public, but it was trusted because of its im- ting unnecessary expenses. For example, al- plicit government backing. Through efforts though new staff were hired to improve the to develop brand identities, better customer services or technical aspects of the bank, service, and overall public confidence, Ag other staff had to be cut because they were Bank now commands much more favorable less competent or were redundant. In the opinions. Staff buy-in is critical to the success case of Ag Bank, most of these cuts took of an institution. Within the bank, motivating place while it was under Central Bank re- the staff with training, fair and performance- ceivership. One of the first steps new man- based compensation, and recognition for agement took was cutting the right costs. good work has created better and more efficient working environments. Learning and Experimentation In the two and half years of the USAID- The most important change for the new funded remediation, training was an integral organization was focusing on client service part of reform for staff at all levels. The and a healthy bottom line. This meant all training program focused on the practical staff had to learn new ways of doing busi- skills and information that employees ness, and the organization had to experi- needed to deliver bank products and per- ment with products and systems that form their jobs. Because of the importance matched the institution with its market. training played in the turnaround of Ag Bank, Deutsche Gesellschaft für Technische Staff Training Zusammenarbeit GmbH (GTZ) provided One of the clear advantages at Ag Bank was assistance to establish a professional training a well-trained and disciplined staff. Al- department at the bank and significantly up- though they were not used to thinking grade the skills of trainers. strategically, they quickly grasped new products and marketing approaches and im- Product Development plemented them. In the countryside, staff In addition to staff training, product devel- and their families are imbedded in the local opment and rollout was based on a rapid but communities, and they know who, where, integral learning and experimentation and how to find people to borrow money process. As described above, products were who would pay it back. identified by managers at various levels. State-owned banks often have negative Once standard policies were drafted and public images. It is important to capitalize on approved, the products were pilot tested, the existing positive aspects of the brand, but refined, and then rolled out nationally. the institution also needs to be seen as "new Training for product rollout went from the and improved" from both a customer and an national to the regional and branch levels, employee perspective. Ag Bank had a partic- and decisions on how much to focus on a Scaling Up Poverty Reduction 41 The Agricultural Bank of Mongolia particular product were decided jointly at the turnaround in Mongolia would not have local and national levels. In this way, staff was occurred without participation by the World able to experiment and learn which products Bank and USAID. Other bilateral and were most appropriate in each market. multilateral entities also made significant contributions. International donors helped Management Information Systems ensure the decision-making authority of the One common reason why a business fails is turnaround teams and provided critically uncontrolled growth. In a state-owned bank needed funds for capital improvements turnaround, it is imperative that revenue and technical assistance. The World Bank growth be balanced by a quality portfolio was instrumental in reaching consensus and by systems and procedures to detect and with the Mongolian government, while correct problems. Systems to monitor portfo- USAID funded the DAI team that led Ag lio performance and to detect fraud are both Bank's turnaround. crucially important. These systems do not Consultants and Managers. As an expa- have to be complex--at remote branches, triate on a management team, operating they need not always be computerized--but under a clear sense of corporate governance they must be timely and complete. At Ag and of protecting Ag Bank's assets, the Bank, the lack of computers and networked CEO and his team had clear power and the communication systems was not a hindrance authority to say "no." Foreign managers to timely and accurate reporting. Managers working on the bank's turnaround did developed simple paper reporting systems not have the societal imperatives that through which local, regional, and national would create expectations and temptations data were tracked and used for ongoing analy- contrary to the turnaround efforts. There- sis and decision making. The data are critical fore, it was critical that the new senior to the learning and experimentation process managers come from outside of Mongolia as they provide the evidence of which innova- for the rapid turnaround. Likewise, the tions work and which require refinement. external consultants were important for managing a transparent valuation, bid, and External Catalysts selection process. The support and pressure provided by external agents--including donors and Private Investors private firms--were key to securing initial The private investors that purchased the bank and ongoing commitment for change and were the final link in ensuring the turnaround achieving sustaining results at Ag Bank. efforts created economic value and that the bank would continue to offer services to ru- Donor Agencies ral Mongolians in a sustainable manner. Their Although a turnaround is conceptually international links have brought new ideas feasible without international support, the and resources to the institution. 42 Case Studies The Agricultural Bank of Mongolia Lessons Learned injected as cash or as development bonds, It is challenging to turn around a large depending on the institution's liquidity state-owned bank. Success rests on a conflu- position. If bonds, the interest should be ence of political and financial circumstances, paid in cash to provide adequate cash flow. both internal and external. The driving There should be adequate funds to cover factors summarized above should be consid- initial operating costs and to purchase ered as requirements for any successful turn- immediately required fixed assets. A long- around effort. The most critical lessons term cash flow stream needs to be identified learned are summarized below. to support the ongoing operations. Management must be politically Staff requires training, incentives and independent and qualified protection from political pressures Management was able to pursue successful Staff requires ongoing training in skills, turnaround strategies at Ag Bank because of policies, products, and systems, especially dur- the unique design of the remediation and ing a time of significant organizational the support of the bank's stakeholders. As change. Transparent and performance-based the historically state--owned bank in the incentive systems help ensure that priorities countryside, Ag Bank was under very strong and behavior are aligned with the institution's pressure to serve the government's political best interests. When staff is threatened or needs. Because the USAID-funded project pressured by political forces, senior manage- team was given full authority for the bank, ment should provide protection in the form a unique approach for USAID, a new credit of being used as an "excuse" for not giving in. and operating culture could develop that has resulted in a very profitable and high Marketing is essential quality credit portfolio. The carefully struc- A key lesson learned by Ag Bank is the impor- tured independence, buttressed with donor tance of marketing- focused research, strong conditions, was essential to rebuilding a brand promotion, and products responsive to sound bank. customer demand to provide an income stream to sustain the organization. Ag Bank's The starting balance sheet must be clear managers used a combination of local Before beginning an assignment, the turn- knowledge from their branch managers, around team must understand the true po- formal market surveys, and experience from sition of the bank's balance sheet and nego- working in other developing countries. tiate accordingly with the government. All earnings and fixed assets must be properly Financial intermediaries can profitably evaluated, and any needed capital must be service low-income markets put into the institution to raise the capital Some doubt whether low-income popula- base to no less than zero. Capital can be tions can or will pay for the financial Scaling Up Poverty Reduction 43 The Agricultural Bank of Mongolia products and services they need. The expe- up to significant net income. Nonetheless, rience in Mongolia proved different--that operating costs should remain under control. where per capita incomes are low, there is a State-owned banks can usually meet the large market for the right kind of deposit needs of their market segment with relatively and credit products, even if the interest low-cost operations. The intent is not to rates and fees are relatively high. Low-in- compete with international banks going come market segments will pay for the right after high-end customers, but to appropri- products and good service. ately serve rural and low-income market seg- ments. Money should be spent to make the Meeting client's financial services needs branches adequate and comfortable, but they has a positive economic impact. do not have to be the best in town. Occasion- Often, small businesses will borrow lower ally, only two people are needed in a location, amounts of money than they need, which and hand ledgers and a calculator may be ad- can be valuable in building a credit history. equate technology. However, if the lending products do not adequately meet the needs of the business State-owned banks can be owners, and match their ability to service turned around. certain levels of debt, a large dropout rate With a low operating costs and good will occur. Thus, the products (including returns from lending, Ag Bank was able to the delivery and services attached to them) generate the profits to reinvest in physical must grow and develop to reflect accurately and human infrastructure and make the the needs of the market. The most bank sustainable, while providing the expensive part of a borrower relationship is needed services to the underserved market. acquisition--bringing the borrower into the The process evolved into a "virtuous cycle" institution for the first time. Profitability driven by the strong demand, as well as and sustainability will come from loan unmet need, for the new services. The Ag renewals and cross-selling other products. Bank experience stands in contrast to the commonly held view that state-owned Penny wise may be pound foolish when it "dinosaur" banks cannot be turned around comes to operating efficiency. and successfully privatized. There may be Although competitive and profitable banks other cases where, despite a poor history must operate efficiently, the focus should not and state sector culture, latent but strong be exclusively on cutting costs. Rather than franchise value in a branch network or closing branches, the focus should be on rev- customer flow can be capitalized into enue-generating services that take advantage profits and sustainability. of the network. Incremental revenue can add 44 Case Studies The Agricultural Bank of Mongolia Bibliography Loan, deposit, and profitability data were provided by Dressen, Robert, Jay Dyer, and Zan Northrip. "Turning Ag Bank CEO Peter Morrow. Other information on Ag Around State-Owned Banks in Underserved Markets." Bank can be found on its website, www.agbank.mn. Small Enterprise Development 13, no. 4 (December 2002). Boyer, Debra, and Jay Dyer. "The Agricultural Bank of Mongolia: From Insolvent State Bank to Thriving Private Morrow, J. Peter. "Marketing for Profit in the Land of the Bank." Paper prepared for "Paving the Way Forward for Khans." DAI Developments, Spring 2002. Rural Finance: An International Conference on Best Practices," Washington, D.C., June 2003. http://www.basis.wisc.edu/live/rfc/cs_12a.pdf Scaling Up Poverty Reduction 45 46 Case Studies Chapter 4 Microfinance in Bangladesh: Growth, Achievements, and Lessons BY HASSAN ZAMAN Executive Summary It is truly remarkable that the microfinance industry in Bangladesh has been able to provide access to credit to around 13 million poor households. There are hundreds of organizations offering microcredit, although the bulk of the clients borrow from a handful of large organizations--Grameen Bank, BRAC, ASA, and Proshika. This growth in access took place during several distinct phases over the last three decades. The origins of the current microcredit model can be traced back to action research in the late 1970s, carried out by academics as well as practioners in organizations that were created to deal with the relief and rehabilitation needs of post-independence Bangladesh. The 1980s witnessed a growing number of non-governmental organizations (NGOs) experimenting with different modalities of delivering credit to the poor. The various models converged around the beginning of the 1990s toward a fairly uniform "Grameen-model" of delivering microcredit. This last decade, especially, saw a sharp increase in access to microcredit. And in recent years, the standard Grameen-model has undergone greater refinement in order to cater to different niche markets as well as to different life-cycle circumstances. Looking at the Bangladesh experience in perspective, one can argue that the current, remarkable scale of access is attributable to specific factors. First is visionary leadership within the pioneering microfinance organizations. The founders and leaders of Grameen Bank and BRAC, in particular, created decentralized structures with appropriate incentives that encouraged high staff performance, which in turn underpinned rational expansion based on existing capacity and client demand. Second, the government of Bangladesh cre- ated a conducive macro-environment and implemented a "hands-off" regulatory policy. Third, donors played a constructive role by providing resources at the appropriate time. This included funding the initial expansion phase of several microfinance institutions and then building the institutional capacity and systems needed to ensure sustainability. Fourth, high population density and relative ethnic, social, and cultural homogeneity made "fran- chising" the microcredit model less difficult, and significantly propelled its expansion. Fifth, the public-private microcredit "wholesaler," PKSF, was far-seeing enough to take Scaling Up Poverty Reduction 47 Growth, Achievements, and Lessons advantage of already-established retail ca- excessively regulating the sector with un- pacity to scale up the microcredit industry, necessary red tape. as well as demand professional standards A second lesson is that microcredit may be and a focus on sustainability. a more effective remedy against poverty and The consensus in the literature holds that vulnerability if it is complemented by other in- access to these micro-loans has considerably terventions. These interventions may be espe- reduced the vulnerability of poor households cially important for the poorest households, in Bangladesh. Poor households are able to which face the greatest risk of income fluctu- smooth their consumption more depend- ations and have the greatest need for a range ably, thereby limiting the hardships arising of financial and non-financial services. from seasonal shortfalls of income. Unantic- Third, there is a role for donor financial ipated shocks such as natural disasters can assistance in expanding the capital base of be better absorbed by building up assets. emerging microfinance institutions, as well Female borrowers are less vulnerable and as developing the technical capacity neces- more empowered within their households sary for organizational sustainability. and the wider community. Hence, subsidies can be justified to support The availability of microcredit has indirectly microfinance institutions in their earliest affected social conditions--for instance, chil- stages, as long as there is a viable route to dren of borrowers are more likely to go to institutional sustainability. school, have better sanitation facilities, and A fourth lesson is that, while visionary better nutrition. These impacts are due to the leadership cannot simply be "franchised," "increased income effect" of microcredit the systems and formal rules that govern as well as the "social mobilization effect" of the successful microfinance industry in borrower group meetings. Bangladesh can to an extent be replicated. This paper proposes five lessons from These may vary according to the size of the Bangladesh that are relevant to microfi- organization, but by and large, the success- nance growth and impact in other ful organizations delegated significant countries. First, an "enabling environment" decision-making authority away from for microfinance is critical, especially main- head-offices, monitored individual staff taining a stable macro-environment where performance, and linked staff incentives to both interest rates and inflation are kept at program targets. Client feedback and reasonable levels. Government regulations program monitoring are also crucial. As and policies are needed to create an organizations grow, the willingness to appropriate environment for the growth of change products based on client need and the sector, where regulatory policies strike demand and to create products tailored to a balance among protecting the interests niche markets is crucial for success. of depositors, supervising microfinance Fifth, one of the lessons unique to the institutions that collect savings, and not Bangladesh experience was the critical role 48 Case Studies Microfinance in Bangladesh played by a microfinance wholesaler, in rent microcredit model can be traced back expanding access and developing profes- to action-research in the late 1970s, carried sional standards. However, apex bodies are out by academics as well as practioners in not a panacea, and a rigorous analysis of the organizations that were created to deal underlying retail capacity and demand for with the relief and rehabilitation needs of funds must be carried out before they post-independence Bangladesh. The 1980s are established. witnessed a growing number of non- governmental organizations (NGOs) which Introduction experimented with different modalities of delivering credit to the poor. The various The fact that the microfinance industry models converged in the beginning of has been able to provide access to credit, the 1990s toward a fairly uniform currently, to nearly thirteen million poor "Grameen-model" of delivering microcre- households in Bangladesh is truly remark- dit. It sparked a sharp growth of access to able. There are around twelve hundred microcredit during this decade. In recent microfinance institutions (MFIs) operating years, the standard Grameen-model has in Bangladesh,1 but the industry is undergone more refinements in order to dominated by by four large MFIs--BRAC, cater to different niche markets as well as to Grameen, ASA (Association for Social different life-cycle circumstances. Advancement), and Proshika--that serve around 11.5 million, or 90 percent of all The 1970s MFI clients.2 After the "big four," the next Experimentation in providing credit to largest NGO, Swarnivar Bangladesh, households considered "unbankable" by has 0.7 million clients, and then there are the formal financial system originated a few probably only ten NGOs that have more years after Bangladesh's war for independ- than 100,000 borrowers. The bottom line ence in 1971. The independence move- is that the majority of the MFIs are small ment gave rise to a new generation of (less than 5,000 borrowers), and that the young activists who were keen on bulk of the access to microcredit is supplied contributing to the reconstruction of this by four MFIs. As such, the experiences of war-ravaged country. The new government scaling up discussed here draw primarily and a myriad of aid agencies that arrived on upon these large MFIs. the scene were unable to cope with the scale of destitution, and non-governmental The Evolution of the Microfinance organizations emerged to meet the Industry in Bangladesh challenges. The early years of the NGO The growth in the poor's access to credit movement in Bangladesh focused on relief took place in several distinct phases over the and rehabilitation with an emphasis on last three decades. The origins of the cur- community development. However, by the Scaling Up Poverty Reduction 49 Growth, Achievements, and Lessons mid-1970s, two of the NGOs that would establishment of the Grameen Bank under subsequently expand in scale, BRAC and a special ordinance in 1983. Proshika, found that "elite capture" was a serious impediment to their development The 1980s objectives. As a result, a separate focus In the early 1980s, several NGOs experi- on the poor through a "target-group" mented with different ways of delivering approach was introduced. Moreover, an credit. One important mode tested was ideological debate within both these organ- the efficacy of providing loans for group izations began to brew, between those who projects compared to offering loans to favoured economic tools (credit, savings, individuals with peer monitoring. The etc.) to support poverty reduction and broad lesson was that the latter was more those who believed that social mobilization effective because of the incentives and it against existing injustices would suffice and lacked the "free-rider" problems seen in financial services were unnecessary. lending to a group. Hence, by the late Around the same time, a team of 1980s, the predominant model became researchers at Chittagong University, led by providing individual loans to a target group Professor Yunus, began an action-research of poor households, with peer monitoring program that provided loans to poor and strong MFI staff follow-up. households in a few villages. Borrowers The Association for Social Advancement were mobilized in "peer groups" composed (ASA) is a classic example of this shift. Its of four to five individuals who were jointly initial emphasis was on forming "peoples' responsible for each others repayment. Sev- organizations," mobilized for social action eral of these small "peer monitoring against oppression. It changed to target groups" would be organized together into groups and then to provision of financial a larger unit which would meet weekly with services in the late 1980s. Now ASA is the the primary purpose of repaying loan install- fourth largest MFI in Bangladesh in num- ments. The process of trial and error initially ber of clients, and its unique low-cost credit combined males and females in the same delivery mechanism is being replicated in credit group, but then changed to separate several other countries. ASA keeps paper- gender groups. It also included "occupa- work requirements to a minimum, has de- tional groups," but this was dropped in centralized most decision making to the favor of village-based groups. The demand field, and overall has a very lean operation.3 for loans grew rapidly and Professor Yunus The 1980s and early 1990s were also enlisted the support of the Bangladesh Bank important to the development of manage- and other commercial banks to provide ment capacity within several of the large the Grameen Project--as it was then MFIs, which allowed them to expand their called--with resources. The success of microcredit programs. What is particularly this experiment paved the way for the interesting is that the development of the 50 Case Studies Microfinance in Bangladesh know-how and confidence to implement Bangladesh now have microcredit services, large programs arose, in some cases, from although there are many smaller pockets the experience of scaling up programs not with little or no coverage (e.g., Chittagong related to microcredit. For instance, in the Hill Tracts). A closer look shows that there case of BRAC, its first major experience is somewhat greater coverage of poor with a nationwide program came when it households in the central and western dis- implemented an oral rehydration program tricts. The southeast and pockets of the to combat diarrheal disease. Thirteen mil- northeast still have room for expansion of lion women were trained to use a simple but coverage.6 effective rehydration solution, and BRAC staff were paid based on how many of their Mid-1990s Onward trainees used and retained this knowledge.4 Feedback from the field, academic research, and international experience contributed to Early to Mid-1990s an increasing emphasis on providing The early 1990s was the period of rapid ex- diversified financial services for different pansion of the Grameen-style microcredit groups of households from the mid-1990s approach.5 The growth was fueled largely onwards. The benefits of a narrow focus on by "franchising," whereby new branches microcredit during the expansion phase was replicated the procedures and norms that that it kept costs low, operations transpar- prevailed in existing branches. It was clearly ent, and management oversight relatively aided by the high population density straightforward. However, it became and relative ethnic, social, and cultural clear that the standard Grameen model of homogeneity in Bangladesh. A notable shift providing microcredit with fixed repayment occurred during this expansion phase to schedules, and standard floors and ceilings placing a greater emphasis on individual on loans sizes, was not sufficient to meet the borrower accountability for loan repayment needs of the extreme poor or the vulnerable and less reliance on peer monitoring. non-poor. Staff follow-up of loans became more Moreover, existing microcredit borrow- rigorous and professional with the use of ers also required complementary financial computerized management information and non-financial services. The standard systems.Donor funds helped in varying practice for MFIs until the late 1990s degrees to expand the revolving loan funds was to collect compulsory weekly savings for MFIs, particularly during expansion from their clients, holding the money phases of the various institutions. Moreover, as a de facto lump sum "pension," which PKSF emerged during this period as a was returned when a client left the organi- wholesale financing institution. Following zation. Access to these deposits was this expansion, a geographical mapping of otherwise limited, which curtailed a microfinance suggests that all districts in potentially important source for smoothing Scaling Up Poverty Reduction 51 Growth, Achievements, and Lessons consumption. separate financial institution, BRAC Bank, Recognizing these limitations, an in- that focuses on lending to the "smaller end" creasing number of MFIs in Bangladesh of the small enterprise sector, with loans av- have offered savings accounts that clients eraging 400,000 taka. can withdraw from more freely, in addition Moreover, evaluation studies pointed to the fixed deposit scheme. Moreover, out that extremely poor households were many MFIs have life insurance products, struggling to benefit from the standard mi- whereby outstanding microcredit debts are crocredit model, even if they joined the written off and other benefits are paid programs. There were a number of factors following the death of a borrower. Non- that kept the extreme poor from borrowing credit services can also take the form of or from benefiting from loans if they input supply, skills training, and marketing obtained them. Minimum loan floors for a support for micro-enterpreneurs.7 A com- first loan sometimes exceeded what clients plementary package to microcredit can also perceived they needed. Fixed weekly loan take the form of providing education for repayments could be difficult to commit to the children of borrowers. Grameen Bank, in light of seasonal income. Other members for instance, has a scholarship program of peer-monitored groups sometimes do for secondary education for girls, and not wish to guarantee loans for extreme a student loan program for tertiary poor households. Residing in remote or education. Similarly many MFIs have depressed areas can also complicate access. community health programs, legal literacy Programs have been developed to make training, and information on how to access these constraints more manageable. ASA's local resources. Flexible Loan Program introduced more flex- MFIs began to experiment with new ible repayment schedules. Mimimum loan niche markets as the traditional microcredit floors for first loans were lowered so that business became standardized (and horizon- amounts as small as 500 taka ($9) could be tal expansion slowed) and required less borrowed. Grameen's program offers zero attention. For instance, several NGOs began interest loans to beggars. The Resource Inte- providing larger loans to "graduate" micro- gration Center's program specializes in offer- credit borrowers, and in some cases to ing loans specifically to the elderly poor, an households which were not part of the unserved vulnerable group. Various programs microcredit system but which wanted a also combine food aid with microcredit and micro-enterprise loan. These loans typically training, like BRAC's IGVGD program. ASA range from 20000 taka (around US $320) has targeted remote areas, offering services to 200,000 taka (US $3,200). Innovative through its cost-effective mini-branch system, solutions are also emerging to address the and Integrated Development Foundations problem of access for the small enterprise work in the Chittagong Hill Tracts. sector. For instance, BRAC has established a 52 Case Studies Microfinance in Bangladesh Factors That Led to the Scaling Up Staff motivation is also enhanced by de- centralizing significant responsibility to the Institution Building--Leadership, Staff lower tiers of the administrative structure. Incentives, and Learning by Doing ASA is the best example of a lean credit de- It is unquestionable that the vision and per- livery structure with high levels of decision- sistence of the leaders of the NGO/MFI making authority given to field offices, from movement are key factors behind the loan sanctioning decisions to staff human success of the microfinance industry in resource issues. Moreover, the structure Bangladesh. Leadership skills were instru- within field offices is relatively horizontal mental at initial stages in persuading a with a branch manager who works with in- skeptical public that providing credit dividuals fieldworkers to resolve problems to the poor could become a viable and and typically shares living quarters with replicable proposition. These skills were other field staff.8 equally important during the process of Effective internal controls are also impor- scaling up--skills such as being able to tant in ensuring effective staff performance. recruit and motivate staff, decentralizing The fact that financial transactions are han- authority away from the center, building dled openly, in the weekly meetings and in management information systems and the branch offices, is a major deterrent to internal controls, as well as having the any form of discretionary behavior by field humility to learn from mistakes. workers. Many NGOs, particularly the ones Staff recruitment, motivation, and reten- that have successfully expanded in scale, tion are particularly important for large have developed measures that include organizations. BRAC, for instance, employs frequently rotating staff within and between around 28,000 staff in its various programs; branches, scheduling regular field visits by Grameen has around 12,000 in its microcre- senior management, developing a strong dit program; and ASA's microcredit program internal audit team, and contracting annual employs around 8,000 staff. A critical element external audits. in this process is an objective performance A fundamental part of the scaling up of evaluation system for staff that is linked to Bangladesh's NGOs, and more specifically career mobility and other incentives for staff the microfinance movement, has been the to perform well both individually and in ability to learn from experiences and adapt teams. Grameen Bank, for example, has programs accordingly. This learning process introduced a system for rating branch offices takes place both through informal feedback on the achievement of specific targets, which by field staff during regular interactions not only include standard loan recovery but with management, as well as through a also factor in social indicators, such as the formal monitoring and evaluation process. proportion of children of Grameen clients BRAC's Research and Evaluation Division going to school. has around 20 professionals whose key Scaling Up Poverty Reduction 53 Growth, Achievements, and Lessons function is to evaluate BRAC's multi-di- enterprises. Two facets of these trends are mensional programs and give timely feed- worth highlighting. back to program staff and management. First, the decisions to subsidize these This feedback process occurs in longer term operations were not free from controversy. research as well as assessments with quick Advocates for funding these loan funds turnaround. The shift to more flexible fi- had to argue their case with officials nancial services, that took place in recent within their own agencies who believed that years, was largely based on client feedback the capital base for loan operations ought and analysis of the limitations of a uniform to be enhanced only by savings mobiliza- microcredit model. tion or borrowing from commercial sources. In retrospect, these decisions A Constructive Donor-Client Relationship to contribute to MFI loan funds were by External resources played an important part and large correct, as almost all of the MFIs in the experimentation, subsequent growth that received this support have either in outreach, and institutional strengthening attained financial self-sufficiency or are well of the microfinance industry. At the same on their way to doing so. Donors also time, the large microfinance institutions invested in organizational systems and have been successful in "managing donors." MFI staff training in order to strengthen International NGOs, such as the Ford the capacity to administer these growing Foundation, Oxfam, and the Aga Khan programs. Foundation, played an important role in the Second, large NGOs in particular have initial stages of the NGO-MFI industry in been reasonably successful in managing Bangladesh. The subsequent expansion and donors. BRAC, with its large multi-faceted consolidation was funded largely by official programs, has a long history of working bilateral agencies, and later by multilateral with donors, and the evolution of this agencies, when international NGOs could relationship is worth highlighting. Donors, not match the growing resource require- who have their own incentives to commit ments of the larger MFIs. The 1990s have resources and demonstrate results on seen dependence on donor resources the ground, have been eager to provide progressively decline for the large MFIs. resources to organizations with proven track Grameen Bank, ASA, and BRAC do not records. Hence, the likes of BRAC have had receive any grant financing for their micro- to deal with multiple donors who each credit operations. Moreover, out of BRAC's wanted to fund specific projects. These total $160 million expenditure on develop- uncoordinated donor missions and disparate ment programs in 2002, more than 80 disbursement and reporting arrangements percent was financed from its own resources, taxed BRAC's internal capacity and led to through the interest income on microcredit its management proposing changes for how as well as surplus from its commercial donors ought to operate. 54 Case Studies Microfinance in Bangladesh In the early 1990s, donors shifted their long relationship has not been free from approach from financing specific BRAC tensions on both sides, the government of projects to financing BRAC programs. Bangladesh has thus far been able to place Donors also formed a "consortium" that the interests of the poor foremost when pooled funds, negotiated jointly with dealing with NGO issues. A recent example BRAC, and agreed to common reporting is the decision to release donor funds ear- requirements. An important part of the marked for Proshika, even though it was consortium funding arrangement and the being investigated for alleged irregularities move toward program funding has been in the use of its funds. an improvement in the predictability of Ultimately, though, the relationship resource flows. For instance, BRAC secured between the government and the NGOs financing for its Rural Development Pro- depends on individual personalities and gram for a five-year period from the donor social ties9 as there have always been widely consortium. Moreover, the establishment of varying individual views regarding NGOs a donor liason office for BRAC also acts as within the civil service and the Cabinet. a buffer between BRAC staff and the Individuals in key positions within the various visitors, consultants, and evaluators. government have time and again been in- strumental in facilitating the growth of the A Progressive Government Stance microcredit sector. The early development The appropriate "enabling environment" of the Grameen project, its registration as a that existed in Bangladesh greatly aided the bank, and the decision to grant it manage- early experimentation and later scaling-up of rial autonomy are clear examples,10 as was the microfinance industry. The macro-econ- the establishment of PKSF with a strong omy of Bangladesh has, by and large, been autonomous board. However, the prevail- soundly managed, and the significance of ing consensus is supportive of NGOs, this should not be underestimated. The rate although accusations of involvement in of inflation has been kept to single digits, party politics by a handful of NGOs have and economic growth over the past decade strained the overall government-NGO has averaged around 5 percent per annum, relationship of late. thereby creating economic opportunities for Looking forward, it is clear that the reg- microcredit-financed investments. ulatory framework for microfinance needs It is also significant that the government to be strengthened, particularly in light of of Bangladesh has thus far maintained a the large amounts of deposits mobilized for balanced approach towards regulating and the poor. The Central Bank, PKSF, and rep- supervising the activities of the NGO sec- resentatives of MFIs are currently working tor. This has been critical in ensuring the to produce a set of guidelines and standards operational flexibility that is the cornerstone to strengthen the regulatory framework. of service delivery by NGOs. While this Scaling Up Poverty Reduction 55 Growth, Achievements, and Lessons A Professional Apex Body for One of the fundamental factors behind the Microfinance success or failure of an apex is the underly- The Palli Karma Sahayak Foundation ing retail capacity in a particular country. (PKSF) was created in 1990, and is gov- The overall strength of the MFIs in erned by a board composed of both public Bangladesh has been key to PKSF's success. and private sector representatives. It is a Overestimating the capacity to absorb funds public-private apex body that channels by the MFIs on the ground can lead an funds for microfinance to MFIs, and has apex body to fail. However, if a realistic as- been critical to the expansion and improved sessment of the underlying retail capacity is professionalism of the microcredit industry made, then apexes offer many benefits, such in Bangladesh. PKSF's core functions as the ability to screen MFIs on standard include (i) lending money to MFIs, which criteria and creating a "level playing field." meet certain eligibility criteria, to expand their microfinance operations; (ii) building The Impact of Microfinance in capacity and giving hands-on assistance to Bangladesh strengthen MFIs and move them towards The evidence of the impact of microcredit can financial sustainability; (iii) advocating be assessed from two interrelated angles. First, microfinance issues and helping develop who does credit reach, and second, how does an appropriate regulatory framework for it affect the welfare of different groups of the industry. individuals and households? PKSF played an instrumental role in Land ownership, occupational criteria, contributing to the sharp increase of access and asset valuations are standard targeting to microcredit that took place in the 1990s tools used by microcredit providers in by expanding the capital base for MFIs to Bangladesh in order to direct resources to onlend to the poor. For instance, as of the rural poor. These indicators have been December 2003, PKSF loans constitute shown to be relatively accurate correlates of around 30 percent of ASA's current revolv- poverty by program administrators who do ing loan fund. PKSF is also widely credited not have the time, resources, or expertise for sharpening the focus of many MFIs to carry out more sophisticated calculations on financial sustainability and in setting of poverty for each household in their appropriate standards to pave the way to targeted area. a strengthened regulatory structure for In practice, the land criterion is the one microfinance. that is more closely adhered to in the field. There is a growing experience with Several studies show that between 15­30 setting up apex institutions worldwide, e.g., percent of members of microcredit pro- PPAF in Pakistan, RMDC in Nepal, grams are from "non-target" households as FONCAP in Argentina, LID in Bosnia- measured in terms of land.11 However, they Herzogovina, and MISFA in Afghanistan. typically are marginal farmers and can still 56 Case Studies Microfinance in Bangladesh be considered part of the vulnerable non- landholding criterion tend to be marginally poor, prone to transient bouts of poverty.12 above the poverty line and are susceptible On the other hand, there is also evidence to transient poverty in certain years. that a large proportion of extremely poor The literature broadly supports the households join microcredit programs.13 hypothesis that access to microcredit For instance in Khandker's sample, 65 contributes to poverty reduction in percent of BRAC households had no Bangladesh, although the evidence is not agricultural land, compared to 55 percent entirely clear-cut.18 For instance, data for Grameen members, and 58 percent collected by the World Bank in 1992 have for a comparable government-run microcre- been used to show widely varying results dit program. depending on the methodology chosen to Not only do the poorest join BRAC's assess impact. Khandker estimates that for credit program, but their borrowing pat- every 100 taka lent to a woman, household tern is similar to better-off members.14 In consumption increases by 18 taka; interest- other words, the presence of wealthier ingly, the figure is 11 taka if the same households does not appear to affect the amount was lent to a man.19 Moderate credit supply to poor households; however, poverty falls by around 15 percent, and there is evidence to suggest that poorer ultra-poverty by 25 percent, for households households use a larger share of their loans who have been BRAC members for up for consumption purposes, compared to to three years (controlling for other better-off households.15 Noting that the factors), according to the author. Similar poorest join BRAC's credit program and results are found for Grameen Bank and that they also actively borrow after they Bangladesh Rural Development Board join, it must also be mentioned that there is (BRDB) members. evidence which suggests that households On the other hand, using the same data who join microcredit programs a few years and a different way of correcting for selec- after the village group has been established tivity bias, Morduch finds that microcredit tend to be less poor, compared to the does not have a significant impact on con- members who join at the start of the sumption levels and therefore on income program.16 This feature of better-off poverty.20 Consumption data from 1,072 households joining over time has also been households in one district of Bangladesh is noted as a general rule of thumb in used to show that the largest effect on many targeted anti-poverty programs poverty occurs when a moderate-poor worldwide.17 The bottom line is that BRAC client borrows more than 10,000 the literature on targeting suggests that taka ($200) in cumulative loans.21 In other microfinance programs are reasonably suc- words, there may be a threshold level of cessful at reaching the poor, and that those credit above which a household gains most households who fall above the stipulated in terms of increases in income. The Scaling Up Poverty Reduction 57 Growth, Achievements, and Lessons Bangladesh Institute of Development eligible BRAC households, and 51 percent Studies (BIDS) carried out an extensive lower for eligible BRDB households, com- study of the impact of PKSF PO's microcre- pared to a control group.27 This consump- dit program using longitudinal data of tion smoothing is driven by income 3,000 households between 1997­2000. smoothing as evidenced by the significantly One of the key findings was that "microcre- lower labor supply variability experienced by dit has a positive and significant effect microcredit members compared to the on poverty status of the program control group.28 The importance of this households...."22 The study also finds that result cannot be over-emphasized, given members of microcredit programsare the fact that seasonal deficits play a key part less vulnerable when faced with crises. in the poverty process in Bangladesh.29 Es- Moreover, improvements in other social sentially, Morduch's results indicate that indicators (child immunization, use of san- program participants do not benefit in itary latrines, prevalence of contraception) terms of greater consumption levels, but are also more noticeable for microcredit they participate because they benefit from program members compared to non- risk reduction. members. Asset creation is important to reduce The literature also suggests that moder- household vulnerability to various liveli- ately poor microcredit borrowers benefit hood risks. The findings of an impact assess- more than extremely poor borrowers, in ment of ASA borrowers, conducted in terms of reduction in income (consumption) 2003, suggests that the average value of poverty. The basic premise is that the physical assets increased by 127 percent in poorest have a number of constraints (fewer rural areas, and grew by about 150 percent income sources, worse health and educa- in urban areas over a five-year period. tion, etc.) which prevent them from invest- Moreover, the average increase in cash sav- ing the loan in a high-return activity. This ings rose by 133 percent and 111 percent in could be due to the higher risk associated rural and urban areas, respectively, over this with a high-return activity or because of a same five-year period. Similar evidence is long gestation period for the returns to found in studies of BRAC, Grameen, and accrue.23 This is borne out by detailed PKSF's partner organizations. case-study evidence24 and by comparing Another pathway by which microfinance participants of credit programs who cater to appears to reduce vulnerability is through different socio-economic groups.25 the emergency assistance provided by many There is strong evidence that microcredit microfinance organizations during acute contributes to reducing household vulnera- natural disasters, such as the recent floods bility. Morduch shows that consumption in Bangladesh. The fact that these organiza- variability is 47 percent lower for eligible26 tions turn into de facto relief agencies is Grameen households, 54 percent lower for crucial to sustaining these households in the 58 Case Studies Microfinance in Bangladesh immediate aftermath of a natural disaster. to the fact that they were seen as income Moreover, post-disaster rehabilitation earners for the family because of their access assistance, in terms of both financial and to credit.34 other services, is also highly valued by Hashemi et al developed an "empower- microcredit clients. ment index" based on eight empowerment The pathways by which microcredit re- indicators. Their analysis establishes that duces vulnerability, that have been discussed contributing to her household's income is a here, relate to income and consumption significant factor contributing to a woman's smoothing and asset building. However, the own empowerment. However, Hashemi et impact of credit on women's empower- al also show that credit programs can em- ment, or reducing female vulnerability, has power women independently of whether also received considerable attention. they contribute to family income or not, Empowerment of women in Bangladesh can after controlling for other factors.35 be viewed against the backdrop of Those who are skeptical about the em- patriarchy, defined by Cain et al as a "set of powering effect of microcredit have focused social relations with a material base that on the issue of women's control over loans. enables men to dominate women."30 Goetz et al used a sample of 253 female Hence, it can be thought of as an improve- borrowers from four rural credit providers ment in intra-household gender relations.31 in Bangladesh. Their investigation of loan Moreover, given the institution of purdah histories led the authors to conclude that (loosely translated as "veil"), a pervasive "about 63 percent of the cases fall into the social construct which restricts the three categories of partial, very limited, or female sphere within a typical Bangladeshi no control, indicating a fairly significant pat- household, empowerment can also be tern of loss of direct control over credit." viewed in terms of a woman's interactions The authors disaggregated their data in outside the homestead and the acquisition terms of loan activity and concluded that of skills, knowledge, and confidence that investing in traditional women's work in- such interactions can bring.32 creased their chances of being able to con- The work by Amin et al in 36 villages trol the loan.36 Montgomery et al also have in Bangladesh showed that membership in reservations about the empowering effect of microcredit programs positively affected a microcredit. Their argument is based largely woman's decision-making role, her marital on secondary sources and a small field stability, her control over resources, and her survey focusing on the issue of control over mobility, but had less impact on her attitude loans.37 While the authors admit that their regarding marriage and education of sample is small, they on balance support daughters.33 Naved finds that the women Goetz et al that microcredit reinforces participants in credit programs, in her existing gender patterns and inequalities by sample, felt their status had improved due promoting traditional income generation Scaling Up Poverty Reduction 59 Growth, Achievements, and Lessons activities,38 which they believe do little to regulating the sector excessively (i.e., stran- alter the social status quo. gling it with unnecessary red tape). On the whole, the evidence presented by those who argue that microcredit improves Microcredit may be a more effective the status of females within a household remedy against poverty and vulnerability if it is complemented with other interventions. appears more convincing that that argued These interventions may be particularly ap- by the skeptics' camp. There are two main propriate for the poorest households, which reasons for this contention. First, the under- face the greatest risk of income fluctuations lying thread of the argument, that access to and have the greatest need for a range of fi- an important household resource (credit) nancial and non-financial services. More- enhances a female's status within the house- over, while the provision of microcredit can hold, is both intuitively appealing and enhance a woman's status in the eyes of resonates with the theoretical literature on other household members, social mobiliza- bargaining models of the household.39 tion and legal education interventions in Second, the focus on female control over conjunction with credit are likely to have a loans, as a key component of the skeptics' more significant effect than credit alone. argument, fails to recognize that credit However, this does not imply that microfi- enters the overall household income pool nance institutions ought to provide these and that household members jointly partic- services. In many cases, organizations may ipate in the loan investment. prefer to specialize in providing microfi- nance and facilitate linkages to providers of Lessons Learned other non-credit interventions. The importance of an enabling There is a role for donor financial assis- environment for microfinance cannot tance in expanding the capital base in be underestimated. emerging microfinance institutions, as A critical part is maintaining a stable macro- well as in developing technical capacity economic environment with both interest that leads to organizational sustainability. rates and inflation kept at reasonable levels. Hence, subsidies can be justified to support The lack of macro-stability has seriously "infant" microfinance institutions, as long constrained the growth of microfinance in as there is a viable route to institutional sus- several countries, e.g., Malawi. Government tainability. The duration of these subsidies regulations and policies are also crucial in would vary according to local conditions creating the appropriate environment for and level of poverty of the clients. the growth of the sector. These policies need to strike a balance between protecting the interests of depositors, in microfinance institutions that collect savings, and not 60 Case Studies Microfinance in Bangladesh The systems and formal rules that govern to tailor or create products for niche mar- the successful microfinance industry in kets is critical for success. Bangladesh can, to an extent, be replicated. These vary according to the size of the or- The creation of a microfinance whole- ganization, but by and large, these organi- saler, like PKSF in Bangladesh, has the zations delegate significant decision-making potential to play an important role in authority away from head offices, are able expanding access and developing profes- to monitor individual staff performance, sional standards. and have linked staff incentives with However, apex bodies are not a panacea, program targets. Client feedback and and a rigorous analysis of the underlying re- program monitoring are also crucial. As tail capacity and demand for funds needs to organizations grow, the willingness to be carried out before they are established. change products based on this feedback and End Notes 1 Credit and Development Forum, "CDF Microfinance 9 N. Hossain, "Elites and Poverty in Bangladesh," DPhil Statistics," (Dhaka: CDF, 2002). thesis, University of Sussex, 2003. 2 The latest figures indicate that BRAC has 3.5 million 10 M. Yunus, Banker to the Poor (Dhaka: UPL, 1998). borrowers, Grameen Bank has 3.1 million clients, 11 It is interesting to compare this figure with Proshika 2.9 million, and ASA 2.1 million clients. Copestake's (1992) evaluation of India's Integrated 3 S.H Choudhury, "Financing the Poor: ASA Rural Development Project, where the proportion of Experience," The Daily Star (Dhaka), March 13, 2003. non-poor households ranged upto 36 percent. See S. Mustafa, et al, "Beacon of Hope: An Impact 4 In addition to this innovative staff incentive system, a Assessment Study of BRAC's RDP," Research and detailed evaluation of the oral rehydration experience Evaluation Division, BRAC, Dhaka, 1996; R. also points to a number of other success factors: sys- Montgomery et al, "Credit for the Poor in Bangladesh: tematic recruitment and training of staff, an effective The BRAC Rural Development Programme and the feedback loop, the willingness of senior management to Government Thana Resource Development and learn from the lessons from the field, and support from Employment Programme," in Finance against Poverty, the government, donors, and professional experts. See ed. D. Hulme and P. Mosely, vol. 2 (London: Routledge, M. Chowdhury and R. Cash, A Simple Solution: 1996); H. Zaman, "Assessing the Poverty and Teaching Millions to Treat Diarrhoea at Home (Dhaka: Vulnerability Impact of Microcredit in Bangladesh: A UPL, 1996). Case Study of BRAC," Policy Reseach Working Paper Series 2145, World Bank, 1999; and S. Khandker, 5 S. Ahmed, "Microcredit and Poverty: New Realities Fighting Poverty with Microcredit, (Oxford: Oxford and Strategic Issues," in Attacking Poverty with Micro- University Press, 1998). credit, (Dhaka: UPL, 2003). 12 Zaman, "Assessing the Poverty and Vulnerability 6 PKSF, "Maps on Microcredit: Coverage in Impact," 1999. Bangladesh," (Dhaka: PKSF, 2003). 13 Khandker, Fighting Poverty with Microcredit; and 7 For instance, to support sericulture, BRAC trains the Zaman, "Assessing the Poverty and Vulnerability entrepreneur in silkworm farming, supplies the silkworm Impact." eggs, plants the mulberry trees, arranges for extension services by a BRAC specialist, purchases the cocoons 14 Zaman, "Assessing the Poverty and Vulnerability from the farmers at their homesteads, and supplies Impact"; and S. Halder and A.M. Husain, "Indentification cocoons to a BRAC-run silk-reeling center. of the Poorest and the Impact of Credit on Them: The Case of BRAC," mimeo, BRAC Research and Evaluation 8 P. Jain, and M. Moore, "What Makes Microcredit Division, Dhaka, 1999. Programmes Effective? Fashionable Fallacies and Workable Realities," IDS Working Paper 177, University 15 Halder and Husain,"Indentification of the Poorest." of Sussex, 2003. Scaling Up Poverty Reduction 61 Growth, Achievements, and Lessons 16 I. Matin, "Mis-targeting by the Grameen Bank: A 27 These results are statistically significant at the 95 Possible Explanation," IDS Bulletin 29, no. 4 (1998). percent level. 17 Lipton, M. "Successes in Anti-poverty." Issues in 28 Morduch's estimates of labor supply variability is 39, Development Discussion Paper, No. 8, International which is 46 percent lower for microcredit members Labour Office, Geneva, 1996. compared to a control group. 18 The methodological problems associated with 29 H. Rahman, "Mora Kartik: Seasonal Deficits and the assessing the impact of microcredit are complex. The Vulnerability of the Rural Poor," in Rethinking Rural literature typically uses "control groups," usually "eligi- Poverty: Bangladesh as a Case Study, ed. H. Rahman ble non-members," or "recently joined members" in and M. Hossain (Dhaka: UPL, 1995). order to address the problem of the counter-factual. 30 M. Cain, S.R. Khanum, and S. Nahar, "Class, There have been attempts to account for selectivity bias Patriarchy and Women's Work in Bangladesh," with varying degrees of success. Population and Development Review 5, no. 3 (1979): 19 Khandker, Fighting Poverty with Microcredit. 405-38. 20 J. Morduch, "Does Microfinance Really Help the 31 R. Naved, "Empowerment of Women: Listening to Poor: New Evidence from Flagship Programs in the Voices of Women," Special Issue on Women, Bangladesh," Department of Economics and HIID, Development, and Change, Bangladesh Development Harvard University, and Hoover Institution, Stanford Studies 22 (2 & 3): 155-79; N. Kabeer, Reversed University, 1998. Realities: Gender Hierarchies in Development Thought (Dhaka: UPL, 1994); S. Hashemi, S. Schuler, and I. 21 Zaman, "Assessing the Poverty and Vulnerability Riley, "Rural Credit Programs and Women's Impact." Empowerment in Bangladesh," World Development 24, 22 BIDS, "Monitoring and Evlauation of Microfinance no. 4 (1996):635-53. Institutions," Bangladesh Institute of Development 32 S. Amin and A. Pebley, "Gender Inequality within Studies, Dhaka, 2001. Households: The Impact of a Women's Development 23 G. Wood and I. Sharif, eds., Who Needs Credit? Programme in 36 Bangladeshi Villages," Special Issue Poverty and Finance in Bangladesh ( London: Zed on Women, Development, and Change, Bangladesh Books, 1997). Development Studies 22 (2 & 3): 121-54; S.C. White, Arguing with the Crocodile: Gender and Class in 24 F. Farashuddin and N. Amin, Poverty Alleviation and Bangladesh (London: Zed Books, 1992); and S. Empowerment: An Impact Assessment Study of BRAC's Mahmud, "From Women's Status to Empowerment: RDP--Ten Qualitative Case Studies, mimeo, BRAC The Shift in the Population Policy Debate," Bangladesh Research and Evaluation Division, 1998. Development Studies 22, no. 4 (1994): 77-99. 25 R. Montgomery, D. Bhattacharya, and D. Hulme, 33 Amin and Pebley, "Gender Inequality within "Credit for the Poor in Bangladesh: The BRAC Rural Households." Development Programme and the Government Thana Resource Development and Employment Programme," 34 Naved, "Empowerment of Women." Naved uses in Finance against Poverty, vol. 2, ed. D. Hulme and P. Participatory Rural Appraisal (PRA) techniques to iden- Mosely (London: Routledge, 1996). Montgomery et al tify the effect of participation in Save The Children's sav- compare the performance of BRAC borrowers with the ings and credit program in Manikanj. borrowers from a government-run microcredit scheme, 35 Hashemi et al, "Rural Credit Programs and Women's the Thana Resource Development and Employment Empowerment." Programme (TRDEP). The initial endowment conditions of TRDEP's borrowers are higher than BRAC's (average 36 A. Goetz, and R. Sen Gupta, "Who Takes the Credit? pre-loan landholding is 46 and 30 decimals* for TRDEP Gender, Power, and Control over Loan Use in Rural and BRAC members, respectively, and the percentage Credit Programmes in Bangladesh." World of income derived from daily labour is 5 percent and 32 Development 24, no. 1 (1996): 49. percent, respectively), while the credit-delivery mecha- 37 Montgomery et al, "Credit for the Poor in nism and average loan size are, broadly speaking, very Bangladesh." similar. The typical TRDEP borrower's increase in assets and income during the course of the most recent loan is 38 Goetz, "Who Takes the Credit?" higher than for BRAC clients, giving rise to the author's contention that better-off borrowers benefit more than 39 S. Lundberg and R. Pollak, "Separate Spheres poorer borrowers. (*Note: A decimal is 1/100 of an acre.) Bargaining and the Marriage Market," Journal of Political Economy 101, no. 6 (1993): 988-1010. 26 Morduch only includes households who fulfill the tar- geting criteria of the three organizations and labels them "eligible households." 62 Case Studies Microfinance in Bangladesh Bibliography Ahmed, S. "Microcredit and Poverty: New Realities and Strategic Issues." In Attacking Poverty with Micro-cred- Lipton, M. "Successes in Anti-poverty." Issues in it. Dhaka: UPL, 2003. Development Discussion Paper, No. 8, International Labour Office, Geneva, 1996. Amin, S., and A. Pebley. "Gender Inequality within Households: The Impact of a Women's Development Lundberg, S., and R. Pollak. "Separate Spheres Programme in 36 Bangladeshi Villages." Special issue Bargaining and the Marriage Market." Journal of on Women, Development, and Change, Bangladesh Political Economy 101, no. 6 (1993):988-1010. Development Studies 22 (2 & 3): 121-54. Mahmud, S. "From Women's Status to Empowerment: Bangladesh Institute of Development Studies. The Shift in the Population Policy Debate." Bangladesh "Monitoring and Evlauation of Microfinance Development Studies 22, no. 4 (1994): 77-99. Institutions." BIDS, Dhaka, 2001. Matin, I. "Mis-targeting by the Grameen Bank: A Cain, M., S.R. Khanum, and S. Nahar. "Class, Patriarchy Possible Explanation." IDS Bulletin 29, no. 4 (1998). and Women's Work in Bangladesh." Population and Development Review 5, no. 3 (1979): 405-38. Montgomery, R., D. Bhattacharya, and D. Hulme. "Credit for the Poor in Bangladesh: The BRAC Rural Choudhury, S.H. "Financing the Poor: ASA Development Programme and the Government Thana Experience." The Daily Star (Dhaka), March 13, 2003. Resource Development and Employment Programme." In Finance against Poverty, edited by D. Hulme and P. Chowdhury, M., and R. Cash. A Simple Solution: Mosely. Vol. 2. London: Routledge, 1996. Teaching Millions to Treat Diarrhoea at Home. Dhaka: UPL, 1996. Morduch, J. "Does Microfinance Really Help the Poor: New Evidence from Flagship Programs in Bangladesh." Credit and Development Forum. "CDF Microfinance Department of Economics and HIID, Harvard University, Statistics." Dhaka: CDF, 2002. and Hoover Institution, Stanford University, 1998. Farashuddin, F., and N. Amin. Poverty Alleviation and Mustafa, S., I. Ara, D. Banu, A. Kabir, M. Mohsin, A. Empowerment: An Impact Assessment Study of BRAC's Yusuf, and S. Jahan. "Beacon of Hope: An Impact RDP--Ten Qualitative Case Studies. Mimeo, BRAC Assessment Study of BRAC's RDP." Research and Research and Evaluation Division,1998. Evaluation Division, BRAC, Dhaka, 1996. Goetz, A., and R. Sen Gupta. "Who Takes the Credit? Naved, R. (1994) "Empowerment of Women: Listening Gender, Power, and Control over Loan Use in Rural to the Voices of Women." Special issue on Women, Credit Programmes in Bangladesh." World Development, and Change, Bangladesh Development Development 24, no. 1 (1996): 45-63. Studies 22 (2 & 3): 155-79. PKSF. "Maps on Microcredit: Coverage in Bangladesh." Halder, S., and Husain A.M. "Indentification of the Dhaka: PKSF, 2003. Poorest and the Impact of Credit on Them: The Case of BRAC." Mimeo, BRAC Research and Evaluation Rahman, H. "Mora Kartik: Seasonal Deficits and the Division, Dhaka, 1999. Vulnerability of the Rural Poor." In Rethinking Rural Poverty: Bangladesh as a Case Study, edited by H. Hashemi, S., S. Schuler, and I. Riley. "Rural Credit Rahman and M. Hossain. Dhaka: UPL, 1995. Programs and Women's Empowerment in Bangladesh." World Development 24, no. 4 (1996):635-53. White, S.C. Arguing with the Crocodile: Gender and Class in Bangladesh. London: Zed Books, 1992. Hossain, N. "Elites and Poverty in Bangladesh." DPhil thesis, University of Sussex, 2003. Wood, G., and I. Sharif, eds. Who Needs Credit? Poverty and Finance in Bangladesh. London: Zed Jain, P., and M. Moore. "What Makes Microcredit Books, 1997. Programmes Effective? Fashionable Fallacies and Workable Realities." IDS Working Paper 177, University Yunus, M. Banker to the Poor. Dhaka: UPL, 1998. of Sussex, 2003. Zaman, H. "Assessing the Poverty and Vulnerability Kabeer N. Reversed Realities: Gender Hierarchies in Impact of Microcredit in Bangladesh: A Case Study of Development Thought. Dhaka: UPL, 1994. BRAC." Policy Reseach Working Paper Series 2145, World Bank, 1999. Khandker. Fighting Poverty with Microcredit. Oxford: Oxford University Press,1998. Scaling Up Poverty Reduction 63 64 Case Studies Chapter 5 Madagascar : Credit with Education Program in the TIAVO Savings and Loan Associations Network BY HERMINIA MARTINEZ Executive Summary The network of Savings and Loan Associations (SLA), known as the TIAVO network, be- gan as a pilot program in rural microfinance in 1996, serving Fianarantsoa, the poorest province in Madagascar. The TIAVO network has impressively expanded its services to the poor, especially to the very poor, by integrating the SLA and Credit with Education programs. It achieved these gains despite political turmoil in Madagascar in 2002 and other severe constraints. Fianarantsoa--like all of Madagascar--faces inherent disadvantages in providing financial services for the poor: it has a low per-capita income, the population is dispersed, and communications are inadequate, in part because of the country's rugged topography. The expansion of the TIAVO network is part of a nation-wide program supported by the government of Madagascar with donor financing. It aims to assist local groups to develop self-sustaining financial institutions, in four of Madagascar's six provinces (initially), to ad- dress the virtual lack of financial services for the poorest of the population. Proponents of the program recognized at the outset that the microfinance institutions created or expanded under this microfinance program might not reach the very poor without specific efforts. Consequently, they tested and implemented special programs for socially excluded groups-- typically women and the poor living in isolated areas. The Credit with Education (CWE) program was introduced in the TIAVO network in 1999, using Freedom from Hunger's methodology and financed through MicroStart (United Nations Development Programme). Under CWE, poor women are encouraged to form groups, or credit associations (CAs), that join the SLA and thus gain access to credit from the TIAVO network. The TIAVO SLAs are member-owned microfinance institutions that serve low-income members. Members must have savings and provide physical guaran- tees in order to get access to credit--requirements that individual women could not meet, but now can through CWE and the CAs. CWE program loans are small and are guaranteed by the CA members; repayments are weekly or biweekly. Loan recipients are selected by local women's committees. The CWE Scaling Up Poverty Reduction 65 TIAVO Savings and Loan Associations Network program integrates training with the credit with its rugged topography. The TIAVO program. Promoters (field agents) visit network in Fianarantsoa, the poorest the CAs in their communities every week. province in Madagascar (83 percent of its During these visits, they also provide train- population is poor), was chosen for the case ing on family health, child nutrition, and study because both the network and a tar- business practices; and disburse loans and geted program for the very poor have made collect repayments. impressive progress over the past four years, From 2000­03, loans outstanding in the despite the odds against. In addition, the TIAVO network increased more than sev- program targeted at the poor integrated enfold, to around US $850,200, and mem- two other initiatives, from which there are bership more than tripled to 14,000. In the lessons to be learned. three and one-half years since the CWE program started, 129 CAs were established, Objectives and Features with close to 2000 women members. CA The TIAVO network is one of four net- members received a total of 11,000 loans works supported under the 15-year Micro- by the end of 2003. Initially an independ- finance Program which is financed by the ent program within the TIAVO network, government and donors, including the CWE is now being integrated into the net- World Bank through its Adaptable Program work operations. Credit. The program aims to encourage the The experience of the program offers one development of sustainable microfinance more example of how successfully microfi- institutions, primarily savings and loan nance institutions can target programs to associations (SLAs), by providing technical the very poor, and more evidence that these support to local groups in the four programs do increase the depth of outreach provinces where most of the population to the very poor. lives.2 The program also will begin putting a legal framework for microfinance into Implementation Process place. Microfinance networks supported by the government program expanded from Background and Rationale 9,500 members in 1999, to close to 100,000 members at the end of 2003. In Madagascar, providing services to the The Microfinance Program includes a poor, particularly financial services, is diffi- specific initiative--the Credit with Educa- cult and costly. The country is one of the tion program--to provide services to the poorest in Africa (per capita income was very poor (the socially excluded) in each of US $240 in 20021), it is very large (582 the networks. The architects of the program thousand square kilometers), and the recognized that large segments of the pop- population is dispersed. The country has ulation, who were not already clients of the inadequate transportation links associated participating institutions, would initially be 66 Case Studies Credit with Education Program, Madagascar missed. The special initiative actively sought local women's committees based on their out this target population and enlisted personal knowledge of applicants. knowledgeable network field staff (or agents The program integrates training with the from a private agency contracted on a fee credit program. The promoters visit the CA basis) to educate them about the advantage groups in their communities every week and of the services offered. Performance of the give training on family health, child nutri- program for the socially excluded is meas- tion, and business practices. Instruction is ured by how efficiently and how many also offered on financial matters, such beneficiaries are served. as managing the family budget, saving, A specialized institution within the making deposits, and requesting loans. The Microfinance Program provides support to promoters disburse loans and collect repay- the SLA network on technical aspects, in- ments during their weekly visit to each CA. cluding the program for the socially CWE managers approve the loans, which excluded groups, which will decline over are made by the TIAVO network with funds time. Contracts with the specialized institu- obtained for this purpose from UNDP and tions providing technical assistance3 include World Bank, and are disbursed through the performance benchmarks that are moni- local SLA. Since 2002, some SLAs have tored. The participating SLA networks are been dealing directly with the groups, required to comply with prudential norms approving the loans that they disburse. set by the Madagascar Banking Commission Features of the savings and loan transac- and to meet performance targets associated tions in the TIAVO SLAs and CAs are with the technical support they receive. summarized in Table 1. The TIAVO savings and loan associations are member-owned microfinance institu- Political and Regional Context tions serving low-income clients. The SLAs Madagascar today is a democracy with a encourage member savings, which is a pre- generally liberal economic system. The requisite to credit. Under the Credit with country has undergone a process of politi- Education program (CWE), poor women cal and economic liberalization since the are helped by promotersof the program, or late 1980s, which has entailed periods of field agents, to organize into credit associa- civil unrest. Regionalism is strong, although tions (CAs). The women choose the mem- until the late 1990s when reforms began, bers of their CA, and the CA becomes a the government was highly centralized. member of an SLA. Women are eligible for Fianarantsoa, located in central Madagascar, loans from the network without meeting has traditionally carried considerable politi- savings or physical guarantee requirements. cal weight. The population in the province The loans are small and guaranteed by the is dispersed and lives in two distinct CA members, and repayments are weekly or regions--the highlands (cattle grazing and biweekly. Loan recipients are selected by subsistence farming) and the coastal plain Scaling Up Poverty Reduction 67 TIAVO Savings and Loan Associations Network (historically coffee exports, but increasingly clients through their large branch network. subsistence agriculture). Communications The Microfinance Program was intended to between the two regions are weak, which partially address the vacuum created by the makes reaching the poor difficult. In the privatization of the state banks. The pro- 1970s and 1980s, Fianarantsoa benefited gram is aligned with the objectives of the from a large number of government-funded government's Interim Poverty Reduction programs, including credit programs. The Strategy of December 2000 and the Poverty latter contributed to a tradition of non-re- Reduction Strategy of July 2003. payment of loans in Fianarantsoa, a problem Initial commitment to the Microfinance the TIAVO network has had to address. Program by the government and other stake- The Microfinance Program was devel- holders was strong. The TIAVO network oped by the government, in the second half leadership was interested in the program of the 1990s, to bring about economic because it could increase its membership growth and poverty reduction through am- and influence in the region. A number of bitious policy reform program. An element non-governmental organizations, including of the reforms was government divestiture Freedom from Hunger, had discussed credit from ownership of two banks. The state programs with local leaders in Fianarantsoa, banks were bankrupt and had abandoned where the main bank, the Agricultural Bank their mandate of financing small and poorer of Madagascar, was reducing its branches in Table 1: Features of Savings and Credit Transactions Feature TIAVO SLA TIAVO MEMBER CA Members Persons/groups in the Low-income women in the municipality municipality of the SLA of the SLA; members of a credit group Membership fees Contribution to SLA capital: Fee: US $0.50­US $0.80 US$1.7; fee: US$0.8 Credit eligibility SLA member; savings deposit in Member of a credit group SLA, 10-30% of credit; physical guarantees, 100-150% of credit Loan size Varies: Average loan for low- Loan maximum: US$23 (first loan); income members, US$50; rural, US$152 (last loan) US$473; urban, US$1,097 Interest Rate Equivalent of 3% per month for Equivalent of 4.5% per month for amounts due amounts due Maturity of loans Depends on type of loan Initial loans 4 months; subsequent loans 5 months Repayment Depends on client Weekly or semi-monthly, incorporating Arrangements member training Sources: Assouline. 2003, TIAVO 2003 Annual Report. 68 Case Studies Credit with Education Program, Madagascar preparation for privatization. The government Preliminary Results and Comparison with saw support for the microfinance program, Original Objectives and the programs for the most disadvantaged, The TIAVO network began in 1996 as a pi- as an action that could help alleviate the lot rural microfinance program, then in 2000 pervasive lack of financial services for the significantly expanded and strengthened its poor, which had been aggravated by the state systems. Over the past three years, its bank failures. operations have increased substantially, with The TIAVO network is the principal in- membership more than tripling to 14,000. stitution in the microfinance program. The Loans outstanding increased more than CWE component of the network was seven-fold during the three-year period, funded for the first three years by the to US $850,200 by the end of 2003, partly UNDP as part of its MicroStart program us- reflecting pent-up demand after the political ing the methodology of Freedom From difficulties in 2002 (see Table 2). Arrears, Hunger. Initially, implementation of CWE which were a serious problem in the was assigned to the Agency for the Execu- network, have been declining for the past tion of the Microfinance Project three years, from 16 percent (arrears of more (AGEPMF), which was specifically created than thee months) in 2001 to a more accept- in 1999 and charged with the coordination able 3 percent in 2001. and follow-up of the 15-year Microfinance The network surpassed its self-financing Program.4 The CWE program was later objectives, covering about 77 percent of all moved to TIAVO to pre-empt a similar costs (except depreciation). The govern- (and redundant) pilot program that ment finances TIAVO operations on a TIAVO intended to initiate. declining basis under the Microfinance Program. At the end of 2003, 36 of the 130 Table 2: Evolution of TIAVO Network Indicators 1999 2000 2001 2002 2003 No. of Members 4452 4820 6223 8408 14000 No. of SLAs 29 24 23 27 36 Total Number of Active Loans 1078 1898 897 1302 2834 Loans Outstanding (US $000) 112.0 157.2 133.8 257.3 850.2 CWE/Total Loans Outstanding (%) 14.1 13.4 31.6 18.6 9.1 Arrears (more than 30 days) * N.A. 12 16 6 3 Operational Self-Sufficiency ** N.A. 50 54 64 77 Sources: AGEPMF database. * Figure at end of September of each year. ** Operational Self-sufficiency covers all costs except depreciation. Scaling Up Poverty Reduction 69 TIAVO Savings and Loan Associations Network savings-based SLAs in the country were make it more reliable and began rotating members of the TIAVO network. the promoters. The CWE program, started in 1999, was The CWE program exceeded the origi- initially a separate activity within the nally planned objectives in terms of mem- TIAVO network with its own accounts and bers and loans approved (see Table 3). The independent management. The CWE costs of the program (US $450,000) program was gradually integrated into the were higher than estimated at the start of TIAVO network as the latter was expanded, the Microfinance Program. These higher beginning in 2000 with the transfer of costs can be attributed to the fact that the CWE accounts from the TIAVO head office program is larger (more members, more to the individual SLAs.5 (The transfer loans, more promoters) than anticipated, should be completed in 2005.) Over the and because it was set up as a stand-alone past three years, CWE activities have been program rather than integrated into the consolidated, including transferring CA su- network, which would have reduced costs. pervision to TIAVO, and incorporating CWE operations into the regular TIAVO Impact Analysis accounting system. CWE increased its The impact of the program on benefici- efficiency by raising the ceiling for loans in aries, the network, and the region cannot urban areas and introducing incentive pay- yet be quantified because base line data was ments for promoters. (The incentives are not gathered as planned, and the results of linked to the quality of the portfolio and an impact study are not yet available. The number of members served.) TIAVO also average loan to the members of the CAs tightened control of CWE operations to is US $37, while the averge loan size of Table 3: Actual and Projected Development of Credit Associations in TIAVO Network Indicators 1999 2000 2001 2002 2003 Projected No. of Members 634 1189 1610 1746 1922 960 No. of Savings Groups 26 61 105 105 129 160 Avg. Size of Group 15 6 No. of Promoters 6 4 No. of Active Loans 581 1015 1409 1443 1657 960 Loans Outstanding (US $ 000) 16 21 42 48 78 Sources: Background documents for World bank Appraisal of Microfinance Project, 1999. ABEPM and TIAVO databases 70 Case Studies Credit with Education Program, Madagascar members of the TIAVO SLAs is US $222 particularly along the coast. Analysis shows (see Table 4). It seems the program is reach- that the focus on lower income clients ing the poor, even those who live in remote has had an impact on the operations of the areas. About half of the CAs are on the network.7 Visits by CA members to the SLAs coast, a region which is particularly isolated has increased the familiarity of the women and where there are few financial interme- with the activities of a savings and loan diaries. Figures for another network in association, and has helped the SLA leadership the Microfinance Program, which is located understand issues and needs of this clientele. in a more urban area, show that close As a response to the program, the TIAVO to 50 percent of the members of the groups network is planning to develop new credit in the socially excluded program (the CWE products geared to the CA members. To have program of TIAVO) are among the very access to credit, the CA women will need to poor (the poorest one third among join the SLA as individuals, pointing to the in- the poor).6 terest that SLAs have in this potential market. The impact on the development of the TIAVO network itself will also be signifi- Cost and Efficiency in the Use of cant. At the end of 2003, loans outstanding Resources under CWE program accounted for some 9 The US $450,000-grant from MicroStart percent of total loans outstanding in the covered the operating costs of the network, and 40 percent of all TIAVO SLAs program--salaries of promoters and related had CWE programs. The importance of the expenses such as transportation, equipment, CWE program in the individual SLAs varies specialized training, and administration considerably: CWE program loans out- expenses. Equipment and administration costs standing range from 2­65 percent of total were high in part because the program was set loans outstanding in the SLAs where the up as an independent activity and did not ben- program operates. efit from the TIAVO systems. Costs were also It is expected that the CWE program will affected by the turnover of promoters at the increase the impact of the TIAVO network, beginning of the program. This partly reflects Table 4: Comparison of TIAVO and CA Indicators 12/31/03 Indicator TIAVO CAs No. of Members 1385* 1922 No. of SLAs/CAs 36 129 Loans Outstanding (US$ `000) 850 78 Average Loan Size (US$) 222 37 Deposits/Member (US$) 429 0.1 Sources: TIAVO 2003 Annual Report, AGEPMF Database. Assouline, 2003. *Including 129 CAs. Scaling Up Poverty Reduction 71 TIAVO Savings and Loan Associations Network the very difficult environment in Fianarantsoa restrictions on a number of activities, (extreme poverty, poor transportation) which including the creation of cooperatives. makes the task of promoters unusually diffi- Successive governments have recognized cult. (The TIAVO network also suffered from the need to improve the conditions of the the new system of personnel rotation). poor, particularly in the rural areas, and The costs of operating the program were in supported programs, including microfi- line with what had been anticipated. By 2002, nance, designed to assist them. Programs to income from operations covered some 70 address poverty directly were emphasized percent of direct operating costs of the CWE in the late 1990s when macroeconomic program (excluding indirect costs such as the reforms were advanced. The government TIAVO supervision and general TIAVO prepared an interim poverty reduction expenses). Efficiency indicators, such as the strategy in 2000, followed by a full strategy number of beneficiaries served by each in 2003, which focuses on governance, promoter, are consistent with those estimated inclusive growth, and service delivery for at the beginning of the Microfinance Program. human and material security. An analysis by IRAM has estimated that in Microfinance initiatives supported by order to cover costs, the number of CAs the government, international (official and served by each promoter will have to increase private), and bilateral organizations began from an average of 15 (in 2003) to an average in the early 1990s, in response to the of 20.8 Efficiency of promoters should increase chronic shortage of financial services in ru- as administrative functions are transferred to ral areas. The design of the Microfinance the network. However, as measures to increase Program involved extensive consultations CWE efficiency are introduced, it is critical with stakeholders, and particularly among that the training aspects of the program not be organizations in this field working in sacrificed. In fact, there is a concern over an ex- Madagascar. The Microfinance Program, of cessive preoccupation with ensuring quick which the TIAVO network is part, focused profitability of this type of program. The push on providing technical support to local toward early profitability could lead to groups, and its costs were small compared shifting the program towards higher income to other government programs which beneficiaries who can absorb larger loans at the involve large transfers. Even though the same cost to the CA or SLA. government obtained financing for the technical assistance being provided to the Driving Factors four networks, it has not, on the whole, been involved in their operations. Commitment and Political Economy for Change Institutional Innovation Madagascar's political liberalization of The CWE program in the TIAVO net- the early 1990s saw the government lift work confirms the view that microfinance 72 Case Studies Credit with Education Program, Madagascar institutions can provide financial services favored the TIAVO head office initially, and and training for the very poor. The program now will fall from 13­5 percent of the credit. also confirms the importance of including a strong training component in programs Lessons Learned aiming at increasing availability of financial Programs targeted to the very poor can services to the very poor. benefit both microfinance institutions and clients. Learning and Experimentation The experience in Fianarantsoa points to Testing new approaches was the objective the benefits of introducing programs for the of the CWE program for the socially very poor in microfinance institutions. The excluded by the Microfinance Program. In approach is beneficial both to the institution addition, the integration of the CWE and to the client: it generates new clients, program into the SLA network came about especially women, for the microfinance in- gradually, and involved testing different stitutions, it helps microfinance institutions approaches in four SLAs and replicating the understand the problems of the lower end successful ones. Increasing the role of the of the market, and thus develop appropriate SLAs in CWE collections and loan approval products for this market, and it educates the (after the fourth loan to a CA member), and very poor in the use of a financial institu- including CA members in the SLA credit tion. Direct involvement of the SLAs in the committees were both tests. The expanded CWE program, which requires the members SLA functions has contributed to the sense of of the CAs to visit the SLAs, has ownership of the program among CA and contributed to the sense of appropriation of SLA members. It has helped improve collec- the program by the SLAs and the CAs; and tions under the CWE program, as CA mem- has improved loan recovery in the CWE bers do not want to to be in arrears with the program, as women want to establish a SLA. CA member participation in the SLA good record with the SLAs. Creating a pro- committees (an ad hoc arrangement not pro- gram parallel to the microfinance institution vided for under the SLA regulations) did not increases inefficiencies through increased prove particularly useful because the CA costs, and may also lead to competition members were not able to represent their among programs, which in this case is not groups effectively. However, other forms of desirable as the programs in support of the CA participation in the SLAs will be tested. poor are subsidized. Promoter performance improved when incentive payments were introduced. Also, it Microfinance institutions require financial is expected that individual SLAs will play a and technical support to establish pro- more active role in the CWE program as their grams for the very poor. compensation is increased from 2­10 It cannot be expected that microfinance in- percent of the credit. Program remuneration stitutions will be able to finance targeted Scaling Up Poverty Reduction 73 TIAVO Savings and Loan Associations Network programs from their own resources. The Monitoring impact on poverty is difficult. criteria for assessing the program should be It is critical to have good base line efficiency in service delivery and not short- indicators to measure progress and to be term sustainability, although sustainability able to design improvements in programs. should be a medium-term objective. Exces- However, the time it takes to design and, sive emphasis on profitability may have the more importantly, put in place, monitoring effect of changing the type of client served. tools should not be underestimated. For instance, despite considerable work at the Programs that reach the poor need time time the Microfinance Program was to yield results. designed, the base year data to measure the Microfinance programs in general take time impact of the program was not gathered in to yield results because they entail changes time. A qualified agency could provide in behavior. This is particularly true of pro- support in the introduction of an appropri- grams aimed at the very poor; such pro- ate impact assessment methodology in grams should have strong training compo- programs that require it. nents such as that included in the TIAVO program. The time it takes for programs to Government has a role in supervising the yield results should be taken into account performance of microfinance institutions. when the programs are designed. The government has a role to play in super- vising the performance of microfinance Staff incentives work. institutions, as it safeguards the financial The experience with the CWE program system and depositors in a manner akin to confirms what is generally known of other the role it performs in supervising the bank- microfinance programs: incentives, particu- ing sector. However, governments can be larly for staff, matter. Performance of pro- tempted to interfere with the operations of moters in the CWE program improved af- microfinance institutions, which is not ter incentive payments were introduced (it useful, when it obtains financing for them. was also true of the TIAVO staff). Similarly, The program in Madagascar has generally it is expected that the SLAs will play a more had little government interference over the active role in the program as they are better years, which has contributed to its progress, remunerated for the work involved. even in the midst of a very difficult political period. 74 Case Studies Credit with Education Program, Madagascar End Notes 1 Gross National Income per capita, World Bank Atlas 5 The process of integration of the CWE program in the Methodology. TIAVO network is summarized in Assouline, "Processus d'intégration." 2 The process of integration of the CWE program in the TIAVO network is summarized in Nathalie Assouline, 6 Anton Simanowitz, "Appraising the Poverty Outreach "Processus d'intégration de l'Activité Crédit avec of Microfinance: A Review of the CGAP Poverty Education au Sein du Réseau TIAVO," prepared for Assessment Tool," Imp-Act Occasional Paper 1, AGEPMF, IRAM, August 2003. (Brighton, U.K.: Imp-Act, Institute of Development Studies, University of Sussex, 2003). 3 The specialist institution providing technical support to the TIAVO network is Institut de Recherche et 7 Assouline, "Processus d'intégration." D'Application de Méthodes de Développement (IRAM). 8 Assouline, "Processus d'intégration." 4 AGEPMF continues to monitor the performance con- tracts with the specialized agencies providing technical support to the networks. Bibliography AGEPMF [Agency for the Execution of the Microfinance Morin, Gilles, Faisabilité du Projet Microfinance dans la Project]. Rapport d'Activités. 2001, 2002. Région de Fianarantsoa. Report prepared for AGEPMF. June 1998. AGEPMF. Rapport D'évaluation a Mi-Parcours de la Première Phase du Programme Microfinance. February Réseau TIAVO. Rapport de Présentation de la Situation 2002. a la Fin 2003 (2003 Annual Report). January 2004. Assouline, Nathalie. Processus d'intégration de World Bank. Memorandum and Recommendation of the l'Activité Crédit avec Education au Sein du Réseau President for a Rural Finance Technical Assistance TIAVO. IRAM report prepared for AGEPMF. August Project (Report No. P-5807-MAG), January 1993. 2003. World Bank. Project Appraisal Document on a pro- Simanowitz, Anton. Appraising the Poverty Outreach of posed Adaptable Program Credit to the Republic of Microfinance. A Review of the CGAP Poverty Madagascar for a Microfinance Project. (Report No. Assessment Tool. Impact Occasional Paper 1. Brighton, 18959-MAG) April 1999. U.K.: Imp-Act, Institute of Development Studies, World Bank. Madagascar ­ Poverty Reduction Strategy University of Sussex, 2003 Paper and Joint Assessment. (Report No. 27335-MAG), November 2003. Scaling Up Poverty Reduction 75 76 Case Studies Chapter 6 Managing Scaling Up Challenges of a Program for the Poorest: Case Study of BRAC s IGVGD Program BY IMRAN MATIN Executive Summary BRAC approaches microfinance as a key instrument to build ladders of opportunity for the poorest people, who tend to be left out. BRAC's main point of departure from conventional thinking is that, although the poorest do need subsidy-based programs to supply their imme- diate food needs, microfinance can play a fundamental role in constructing a long-term, sustainable foundation for improving food security and livelihoods. However, this is unlikely to happen automatically. BRAC's experiences suggest that creating a strategic linkage between grant-based and market-based microfinance programs requires careful planning, and solid and committed management. Scaling up this approach to reach significant numbers of the poorest requires constant learning and innovation, and ongoing negotiation with partners based on practical field experience. In particular, it requires an appetite for tackling the larger challenge of developing markets that can open up new opportunities for the very poor. Most important of all, it requires vision and commitment to include the poorest. BRAC's experiences suggest that carefully designed strategic linkages, which include grants with a central role for microfinance, can work for the poorest. There certainly will be many differ- ent models and approaches for including the poorest, which will vary according to country contexts. However, the starting point has to be reversing the trend of apathy--which either excludes the poorest or treats them as "relief cases" to be dealt with by "others." BRAC believes that the poorest are, can, and must be central to the vision and commitment of mi- crofinance institutions. Only then will the search for possibilities and opportunities to in- clude the poorest begin and develop. Implementation Process BRAC's Microfinance Canvas BRAC was originally set up as a relief and rehabilitation committee in 1971 to address the immediate needs of the refugees returning home after the nine-month war for independence. Scaling Up Poverty Reduction 77 Managing Scaling-Up Challenges of a Program for the Poorest Today it is today one of the largest micr onto BRAC's different microfinance pro- ofinance NGOs in the world, providing grams, one gets a picture like Figure 1. financial services to over 3.5 million There are two principles of BRAC's con- poor women throughout Bangladesh. The cept of microfinance worth highlighting. One distinctive feature of BRAC's microfinance is is the belief that microfinance can work for a a perspective on the poor as a diverse group diverse group of the poor, and two is the with diverse livelihoods, needs, and poten- importance of creating deliberate linkages to tial, which change over time in response to support continuous progression in the liveli- life events, new opportunities, and external hoods of the poor, including the poorest. shocks. This diversity and dynamism of poor BRAC's microfinance programs are thus peoples' lives is the canvas on which BRAC about building ladders of opportunity for the conceptualizes and designs its repertoire of extreme poor--those who tend to be left out development programs, where microfinance of conventional microfinance programs. The is a core element. If one maps the various idea is to design subsidies in ways that categories of poverty, described in the strengthen the initiatives for the extreme Bangladesh literature that profiles poverty, poor, so that they, too, over a time can build Figure 1: BRAC's Microfinance Canvas Target group Urban (20%) Rural (80%) Better-off (27%) Vulnerable MELA non-poor (20%) EDP Moderate Poor Microfinance (17%) Extreme Poor (31%) IGVGD CFPRP/TUP Destitute (5%) Sources: Background documents for World bank Appraisal of Microfinance Project, 1999. ABEPM and TIAVO databases 78 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh the capacities to benefit from microfinance IGVDG faced in scaling up, what it learned and other mainstream development pro- from the dealing with those challenges, and grams. The Income Generation for Vulnera- the steps taken to address them. One of the ble Group Development (IGVGD) and the primary outcomes gained from the IGVGD new BRAC program for the ultra poor, experience has been the design of a new "Challenging the Frontiers of Poverty Reduc- BRAC program for the poorest, "Challeng- tion/Targeting the Ultra Poor" (CFPR/ ing the Frontiers of Poverty Reduction-- TUP), are examples of this approach. Targeting the Ultra Poor" (CFPR/TUP), It is also about combining microfinance which is also discussed in brief. strategically with other interventions to cre- ate new livelihoods for people facing sudden IGVGD Program and Ladders of vulnerabilities. Recently BRAC began pilot Opportunity for the Poorest testing a microfinance program that offered In 1985 BRAC approached the World Food training in new skills, for retrenched workers Program (WFP), which was providing time- from the ready-made garments factories and limited food assistance to the extreme poor state-owned enterprises, and counseling. through its Vulnerable Group Feeding Finally, BRAC's microfinance vision (VGF) initiative, to pilot a new program supports growth by using the knowledge and model. The results were impressive. A embodied in its institutional networks to BRAC study found that the income of provide financial services to new market women who participated in the pilot segments. The Microenterprise Lending increased significantly, and that their addi- Program (MELA) and the Enterprise tional income was more than the wheat Development Program (EDP) both serve donations from the VGF program. Around growth segments of the market with finan- 80 percent of the women had also entered cial services. (See Table 1 for key informa- BRAC's Rural Development Program and tion on BRAC's microfinance programs.) gained access to its microcredit and social The main idea behind BRAC's microfi- development services. A separate assessment nance programs for the poorest combines of the VGF, by contrast, found that many of basic social service interventions (such as its participants were no better off when they food aid),which are based on grants, with left the VGF than when they joined. promotional ones (uch as training, savings, and credit). IGVGD, which hs been in op- Impact Analysis eration since 1985, is a good example of such an approach. Even more interesting, it Targeting the Poorest is a partnership between a donor (World Reviews of the IGVGD have been favor- Food Program), the government of able and a study commissioned by the WFP Bangladesh, and an NGO (BRAC). This found evidence that the program reached paper highlights the main challenges that Scaling Up Poverty Reduction 79 Managing Scaling-Up Challenges of a Program for the Poorest Table 1: BRAC's Microfinace Programs Program Target Group Term and Conditions Product Details MELA Larger loans provided to · Must have good entrepre- · Loan size range US BRAC and non-BRAC neurial skills $400­$4,000 microentrepreneurs to · Must not have any outstand- · 15 % flat interest rate scale up their enterprises ing loans from BRAC or · 12-, 18-, and 24- other microfinance month loan products institutions repayable in monthly · Must open a bank account installments to receive loan Microfinance · Less than 50 deci- · Must be a member of · Loan size range US mals* of land owned, BRAC VO $50­$350 live in slums, and earn · Must save · 15 % flat interest rate a living by manual · Must not have a loan with · Loans repayable in labor other NGOs weekly installments · Households headed by over a year women and vulnerable poor households (tar- geted specifically through IGVGD) IGVGD · Households headed by · To be eligible for loans: · Initial loan size about women, who own no · Must be a VO member US $50 more than 10 decimals · Must save · Other conditions of land. similar to · Women divorced, Microfinance separ-ated, or have a disabled husband. CFPR/TUP · No more than 10 dec- · Must not be members of · Distribution of income imals of land any government or NGO earning assets · No adult earning development program · Subsistence allowance member · Must have at least one for a specified period · No productive assets adult woman who is physi- · Employment and · School-age children cally able enterprise develop- working Adult women ment training and in manual labor technical support · Essential health care support * A decimal is 1/100 of an acre. the very poor, that the economic position of with key poverty indicators for rural IGVGD recipient households improved, Bangladesh (see Figure 2) found that and that access to NGO microfinance serv- IGVGD attracted members who had signif- ices was greatly enhanced. Hashemi's com- icantly higher levels of absolute landlessness, parison1 of the 1994 WFP baseline survey were functionally landlessness (owned less 80 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh than a half acre of land), owned two sarees engaged in begging dropped dramatically or less, and lacked winter clothing, than the (see Table 2). extremely poor identified by Rahman and A cross-sectional study that compared Hossain.2 While 8 percent of rural house- IGVGD participants at the various stages of holds and around 10 percent of extremely the program cycle found significant positive poor households overall were headed by impacts over a range of social and economic widowed, divorced, or abandoned women, dimensions as the participants progress approximately 44 percent of households en- through the IGVGD cycle.3 [Inst. Tb 2-3] tering the IGVGD program in 1994 were Another performance indicator, which from this social category. This indicates that examined how successful the IGVGD pro- the program reached a substantial number gram is at "graduating" very poor house- of the population for whom poverty is likely holds to regular microfinance programs, to be persistent. provided evidence of improvement. At the beginning of the program, only 15 percent Impacts and Graduation of the IGVGD participants were MFI In terms of economic indicators, the clients. By the program end in 1996, this 1994 WFP survey found that, on average, had increased to 28 percent, and by 2000 incomes of IGVGD clients rose signifi- had reached 66 percent. Although access to cantly, material assets (ownership of home- microfinance increased across Bangladesh stead plots, land, beds, and blankets) during the late 1990s, a 440 percent in- increased, and the percentage of households crease in MFI membership for such a cohort Figure 2: Targeting Performance in IGVGD Key Poverty Indicators 80% 40% 0% widowed, landless functionally no sandals 2 sanees no winter divorced, landless or less clothing abandoned IGVGD clients Hard-core Poor Rural Bangladesh Scaling Up Poverty Reduction 81 Managing Scaling-Up Challenges of a Program for the Poorest Table 2: IGVGD Economic Impacts Variables Time 1994 1996 1999 (pre-program) (end of program) (3 years after program) Monthly income (Taka) 75 717 415 Percentage of households earning more than 7 64 31 TK300 per month Percentage of households with homstead land 73 87 na Percentageof funtionally landless households 94 72 na Percentage of households with beds 58 60 64 Percentage of households with blankets 14 na na Percentage of households begging 18 2 0 Source: Adapted from Hashemi et al. (2001, p. 9) Table 3: IGVGD Social Impacts Food Aid + Food Aid + Savings First Loan Savings + Training % Attending most VO1 meetings 2 19 83 % Attending Gram Shobha meetings2 2 17 40 % Reporting positive difference due to 38 55 83 meeting participation % Reporting that they wanted to start 42 53 82 an IGA after training % Reporting greater levels of 62 78 91 confidence % Reporting that they aspired to be a 6 16 25 VO leader % Reporting that life now is better 24 32 54 1The Village Organization (VO) is the gateway of BRAC's development programs. About 30 BRAC mem- bers from a village form a VO. 2The Gram Shobhas are monthly meetings on specific issues, where both the women members and their spouses attend. These provide a fairly regular forum in which VO members are exposed to and discuss various social and economic problems, and potential solutions. of very low-income, very low-asset people Cost Effectiveness represents massive improvement in the There are no comprehensive cost-effective- numbers of the poor gaining access to fi- ness or cost-benefit analyses of the IGVGD nancial services. available, but Hashemi estimates that the 82 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh amount of subsidy per household was US government of Bangladesh (specifically the $135 per cycle for the year 2000, which he Ministry of Women's and Children's Affairs, argues was a reasonable cost for the im- Directorate of Relief and Rehabilitation, provements that have been recorded.4 This and local government representatives), and evidence, plus other reviews, has certainly a development organization (BRAC). The convinced aid agencies that IGVGD can main role of the various important actors in successfully reduce poverty for sections of the IGVGD program is shown in Table 4. the population that few other programs can reach. Over the last few years, donors and Commitment and Political Economy MFIs have been keen to build on the for Change IGVGD experience and expand programs BRAC's commitment to bring the most for "those left behind." vulnerable into its development program in ways that are cost effective and sustainable in Driving Factors the long run has been the main driver for the The IGVGD program is a partnership be- IGVGD program. Concerns over food secu- tween a donor (World Food Program), the rity, safety nets, and improved nutrition for Table 4 Major Actors and Their Role in the IGVGD Program Partners Main role Ministry of Women and Children Affairs - IGVGD household selection - Arrange funds for training - Extend administrative support Monitoring programme progress Directorate of Relief and Rahabilitation - Allocating and distributing food aid - Extending administrative support World Food Programme - Provide food aid - Arrange funds for training - Monitoring the programme progress - Reseach and Evaluation - Coordinate with GoB and BRAC PKSF and other banking institutions - Provide credit funds to IGVGD programme BRAC - Development and implementation of the programme which includes: - Arrange income generating activities (IGA) and social awarness training - Provide credit and other sector support - Savings managment - Follow, supervision and monitoring - Mobilize donor funds for training - Research and Evaluation Scaling Up Poverty Reduction 83 Managing Scaling-Up Challenges of a Program for the Poorest the most vulnerable women attracted the VGD, program. They also reached an agree- government of Bangladesh to the Vulnera- ment with BRAC to expand the pilot ble Group Feeding (VGF) program. scheme into the IGVGD program. Because Although these needs are central to the lives the architects of this expansion did not of the extreme poor, BRAC felt that without grow complacent, the IGVGD continued to a more strategic approach, that linked these evolve.5 For example, in 1989, field staff vulnerable women to development activities, pointed out that even though they were the food aid-based program on its own members of IGVGD, many women could would not be able to create sustainable, only buy and raise a single chicken at a time positive change in the lives of these women. because of lack of capital. Why not provide The IGVGD approach, therefore, grew out loans to program recipients as soon as they of BRAC's determination to leverage the ex- completed their training? This led to the isting commitment of the government of addition of a third element to the Bangladesh towards the most vulnerable IGVGD­­microcredit­­with the aim of women of rural Bangladesh. speeding up client adoption of more pro- ductive livelihoods and graduation to Institutional Innovation BRAC's programs for the moderate poor. In 1987, the government of Bangladesh This three-pronged approach (food grant, and WFP transformed the VGF program skills training, and microcredit) has been the into the Vulnerable Group Development, or basis of IGVGD throughout the 1990s. [Insert Figure 3] Figure 3: The Idea Behind the Strategic Linkage Approach Household income/ consumption Upper Poverty Line Join BRAC VO (Moderate Poor) Skills-training Microcredit Lower Poverty Line (Hardcore Poor) Food Aid Time 84 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh Learning and Experimentation areas. These areas generally had weak infra- structure and poor communication with the Challenges Faced in Scaling Up IGVGD. town in the thana where the AOs were During the scaling-up phase, the program typically located. This made organizing train- expanded into areas with poor infrastruc- ing, providing credit, and managing savings ture and weak communications. The new for VGD women difficult at best. Follow up focal ministry did not have offices in all the and monitoring was also less than optimal, thanas where BRAC had expanded its which led to repayment irregularities and IGVGD program.6 Accordingly, per the eventual dropout of VGD members. agreement with the government of The current VGD cycle had been reduced Bangladesh, in thanas where the new focal to 18 months, from the 2-year cycle in 1996. ministry did not have presence, the previous The logic behind shortening the time frame ministry responsible for the program would was that more VGD beneficiaries could be continue to fill in the gap. This was a com- helped with the resources at hand. The plex and challenging arrangement because reduced cycle hampered the effectiveness of it required coordination between the old the overall objective, which was further and new VGD focal ministries--a far from affected by administrative delays in the easy task. beginning of the program. Scaling up the IGVGD program, which The life of the program lost 3­4 months was based on BRAC's existing area office to finalizing the VGD beneficiaries and infrastructure, also posed challenges for getting the contract between the govern- BRAC's microfinance program. The main ment of Bangladesh and BRAC signed. component of the IGVGD program was ul- Almost another two months passed before timately to incorporate the VGD benefici- the government circular to the local officials aries into BRAC's mainstream development was issued--without which BRAC could not programs that operated through Village start its work on the development package. Organizations (VOs)--the gateway of This in effect meant that BRAC had only BRAC's development programs. However, a year to implement the activities of the once BRAC agreed to implement the development package, which was exacer- IGVGD program in a thana, all of the bated by the difficulties of the poor cover- existing VGD beneficiaries in the thana had age, mentioned above. BRAC was also to be covered. acutely aware that the period when VGD It was not easy to provide the develop- beneficiaries could benefit the most from a ment package to VGD beneficiaries who loan was when they had the food security lived in areas where BRAC had not extended offered by the VGD card. Their experience its microfinance infrastructure. Setting up suggested that the VGD beneficiaries, who new VOs was not always feasible either, if took out a loan early and completed the loan there were too few VGD beneficiaries in the cycle while they had food assistance, would Scaling Up Poverty Reduction 85 Managing Scaling-Up Challenges of a Program for the Poorest be much more likely to stay in the microfi- mechanism that gets the poor to arrange nance program, compared to those who got their imaginations for tomorrow and be- loans late in the cycle and had to repay the yond. It also is important for building loan after VGD support ended. relationship between the savings institution A lack of understanding about the main and the people it serves."7 purpose of the IGVGD program between However, the other partners of IGVGD the various partners was a major challenge. did not (and do not) readily understand The IGVGD program has a strong develop- the larger role of savings in development. mental focus in contrast to the earlier This lack of conceptual congruence over VGF program, which principally focused on savings becomes most evident when the relief. What this fundamental change in food aid ends and there is pressure from focus meant in terms of program design other partners in IGVGD to return the and approach was not adequately discussed savings to the VGD members. Returning with the most important partners--the their savings, for BRAC, signals the end local government officials. This gap led of the relationship with the VGD members. to major difficulties for BRAC as the imple- This is inconsistent with the central logic menter of the developmental component of the IGVGD program--that of creating of the program. strategic linkages between a relief and One area which best captures this chal- a development program to include the lenge is the issue of returning savings. Sav- poorest. ings is not only a financial product, it is an Another area of contest among the important part of the process for VGD partners is targeting. The development members who are moving from living focus of IGVGD targets the poorest women hand-to-mouth to planning for the future. who are physically and mentally able to ben- For development staff, too, the act of efit from the income generating training saving regularly by the VGD members is and who can manage an income-generating important. On one hand, it challenges the project with microcredit. This condition ex- image of the members as beneficiaries and cludes certain categories of the extreme relief recipients. On the other hand, poor for whom the IGVGD approach is not it provides a space for engagement. This the most appropriate.8 Yet, because the is why savings constitutes such an impor- VGD food aid is one of the scarce resources tant place in development programs. Abed that the local government representatives aptly captured the whole idea behind can allocate among the poorest, they face a savings as a developmental concept when lot of pressure to include different he said: "In Bangla, we have a very categories of the poorest, including the old apt word for planning, porikolpona, mean- and the disabled. This compromises the idea ing "arranging imagination." Regular sav- behind the IGVGD program. ings, however meager, is a very powerful 86 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh Managing the Challenges was scaling up. From the 1999 cycle, den- The main challenge to scaling up the sity gradually increased, related to a delib- IGVGD program was providing effective erate BRAC policy to focus on thanas with coverage to areas that were outside the op- good VGD-card density. erating area of an Area Office--the lowest Several steps were also taken to stream- unit of BRAC management. Branch Offices line the internal management system of the were set up to provide better follow up and IGVGD program. First, in management monitoring to members residing in areas meetings at all levels, separate sessions on beyond the operating area of an Area Of- the IGVGD program were held as a matter fice. The expansion of the Branch Office of routine. This had not been done system- network is shown in Figure 4 below. atically before. Second, the monitoring The density of VGD card holders in a function was separated from management; thana is an important operational variable it had been combined earlier. Separate mon- for the IGVGD program. If the density is itors for IGVGD were hired, who regularly too low, organizing the VGD women be- monitor the program on specific issues. comes difficult, and the whole purpose of Third, a more intensive follow-up system the program suffers. Figure 4 also shows the was put in place by deploying staff at the re- average number of VGD cards per thana for gional (roughly district) rather than the di- the VGD cycles in which BRAC's IGVGD visional level. (Table 5 summarizes some of program was in operation. The average den- the steps that BRAC took in response to the sity of VGD cardholders decreased sharply challenges faced in scaling up the IGVGD during the time that the IGVGD program program.) [Insert Figure 4] Figure 4: IGVGD Over Time 3000 2500 800 700 2000 600 1500 500 400 1000 300 200 500 100 0 0 1994-96 19971998-991999-002001-02 1990 1992 1994 1996 1998 2000 2002 Scaling Up Poverty Reduction 87 Managing Scaling-Up Challenges of a Program for the Poorest As the biggest implementing partner of a emphasizing this as an important constraint flagship program for the poorest, BRAC for the program gradually led the focal min- also had an important voice and role. BRAC istry to take more initiative, such as promot- repeatedly highlighted the importance of ing attendance at the BRAC workshops, ex- the government of Bangladesh and its local panding their infrastructure in the thanas, representatives to this program, especially and strengthening their monitoring of the the need to have a common understanding program. [Insert Table 5] of the objectives and program activities. To support this, BRAC initiated regular work- External Catalysts shops in the thanas for local government The IGVGD idea and the subsequent representatives. Coordination with local program has largely been driven by BRAC, government representatives was the weak- especially during the scaling up phase of the est in thanas where the focal ministry did program. However, during the initial pilot- not have any offices or presence. Repeatedly ing of the idea, WFP played a pivotal role. Table 5 Key Challenges, Consequences, and Steps Taken Challenges Consequences Steps Taken Inadequate Area Office · Poor follow up · Branch offices set up to coverage · Training program hampered cover areas far from the Area · Credit disbursement Office hampered · Separate sessions dedicated to IGVGD issues in all BRAC management meet- ings · IGVGD monitors deployed · Workshops held in the thanas to coordinate and get better support from local government · Infrastructure of focal Ministry increased in thanas Time for Implementing · Credit not be given to all · From the 2003 cycle, the Development Package Limited during the food aid cycle VGD cycle increased to 24 by Procedural Matters · Adequate time not available months. to provide technical support and supervision for sustain- able graduation Lack of Common · Pressure for savings return · Workshops in thanas with Understanding · Poor motivation local government officials · Drop out · Closer dialogues among the partners · Field level exposure 88 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh As a matter of fact, when WFP decided to "savings" and "training" as the develop- transform its Vulnerable Group Feeding ment components of the program. The program into the Vulnerable Group Devel- government of Bangladesh, also without opment program, the government was consulting the VGD guidelines, carried out running the program and was expected to an investigation into the matter and found implement this new approach. In practice, BRAC "guilty." In response, the WFP however, the change of the program name conducted its own investigation and found from "Feeding" to "Development" was in that BRAC was very effectively following name only. Skeptics called it a "poverty con- the guidelines. This brought the matter out tainment program" rather than a "poverty into the open, and the relevant government alleviating" one. ministry started discussing the VGD In 1985, the founder of BRAC, F.H. approach more closely with BRAC. It was a Abed, had a meeting with the deputy turning point for BRAC as a major, credi- executive director of WFP who told him ble actor in developing the IGVGD idea. that that the government of Bangladesh was running the VGD program just like VGF, Lessons Learned and that if it continued as such, it would be difficult to maintain support for the pro- The Complexity of Including the Poorest: gram. In terms of outreach, VGD was the Poultry Sector Development and IGVGD largest safety net program in Bangladesh, The story of scaling up IGVGD is closely and its withdrawal or contraction would be linked to BRAC's entry and its poultry pro- disastrous for the extremely poor. The mat- gram. BRAC wanted to find appropriate ter was extensively discussed within BRAC products for the VGD women, that had to assess the pros and cons of BRAC partic- large-scale potential; were relatively easy to ipating in the program as a development market with existing skills; were home partner along with the government. based, and could generate quick cash flow. BRAC asked WFP to allocate 700 VGD BRAC also had to be able to provide or ration cards especially to test the VGD idea facilitate provision of the required support in one sub-district of Manikgonj where services. Poultry seemed to satisfy all of BRAC had significant presence. However, these conditions. there was a hiccup during this pilot phase, The human and physical infrastructure which turned out to be a blessing in needed to support the development and disguise. An influential national daily ran scaling up of a poultry sector had to be a story on the pilot, accusing BRAC of created incrementally. This process needs "taking 25 taka every month" and "making elaboration to show that developing pro- the VGD women work" for the "free card." grams useful to the extreme poor is a com- Yet the VGD guidelines, under which plex undertaking that requires coordinated BRAC was operating, clearly mentioned action at various levels. Scaling Up Poverty Reduction 89 Managing Scaling-Up Challenges of a Program for the Poorest BRAC observed that the various linkages satisfactory, and supply could not be deliv- that made the poultry sector viable was ered on time or in the right quantity. In absent, especially for the poor living in rural 1996, BRAC modernized its small poultry areas. Poultry raising was not considered a farm. Today, BRAC has six such poultry business for the poor. BRAC realized that farms which supply over 1 million day-old the crucial challenge for success would be chicks every month. controlling poultry mortality. Poultry exten- Ensuring quality feed was another chal- sion services were non-existent, and there lenge that that had to be addressed. Hybrid were many bottlenecks in the supply chain maize, the main ingredient of poultry feed, for poultry vaccination. The other big chal- was not produced in Bangladesh, so BRAC lenge was to improve the quality of the breed imported five tons of hybrid maize seeds. to increase yield. A local bird lays only 60 To attract farmers to cultivate hybrid eggs a year. The government had a HYV maize, a guaranteed floor price was set. In (high yield variety) poultry setup, but 1999, BRAC set up a poultry feed mill to because of the severe supply and support guarantee a supply of high quality and timely bottlenecks, effective demand was low. The feed for the expanding poultry sector. government was very supportive when Recognizing that importing enough hybrid BRAC began to work on relieving these seed would be expensive and would translate bottlenecks, because demand would increase into higher feed prices, BRAC entered all over Bangladesh. into a joint venture with an Australian seed BRAC used its local presence and grass- company. Today, BRAC's Maize Seed roots knowledge to create a cadre of poul- Production Program produces 400 tons of try workers from the VGD women. They hybrid maize seeds, which in turn produces were trained in basic poultry diseases and an annual maize crop of 100,000 tons. provided medicine and vaccination services for a modest charge. BRAC facilitated the Taking the Challenge Further: BRAC's networking of these poultry workers with New Program for the Ultra Poor the government's local poultry and livestock BRAC's IGVGD experiences demonstrated infrastructure. Today BRAC has more than the possibility that opportunities could 40,000 such poultry workers working be created from safety nets for those throughout rural Bangladesh. who are left behind by conventional micro- BRAC started by buying day-old chick finance. This made BRAC experiment from government-run poultry farms, but even more boldly with the concept of there were several problems. The largest strategic linkages. unit was in Dhaka and transporting large BRAC noticed that for a great majority numbers of day-old chicks to rural areas of the poorest, the IGVGD increased their was both expensive and hard on the chicks. ability to benefit from regular microfinance The quality of the day-old chicks was not programs, but for a significant minority, this 90 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh did not happen. It was more worrisome that 2006) of this experimental program, 25,000 those who failed to make it were among the ultra-poor women will be targeted. poorest and most vulnerable. There were Program components for the ultra poor several reasons for this. are selected by using a careful targeting BRAC was at times dissatisfied with the methodology, that combines participatory targeting that local government representa- approaches with a simple survey-based tool. tives carried out, which was sometimes based The program also includes a special invest- on political and other motives. More impor- ment program that grants assets and a tantly, the women often failed to get the full stipend, an enterprise specific skills develop- benefits of the VGD food aid. Often, one ment training program, a program of essen- VGD card was unofficially shared between tial healthcare, and a social development two or more women. Sometimes, women program. The program aims to cover had to "buy" the VGD cards by bribing the 70,000 ultra poor from 2002­2006. government official. The destitute women The whole idea behind TUP is to enable would borrow the money for the card from the ultra poor develop new and better op- wheat dealers, who would repay themselves tions for sustainable livelihoods. This re- by charging the women more than market quires a combination of approaches, some rates for wheat bought with the VGD card. promotional, such as asset grants and skills BRAC wanted a program where it had more training; and some protective, such as control over the processes, and which was stipends and healthcare services. It also re- specifically designed to build a solid founda- quires addressing constraints at various lev- tion from which the extreme poor could els, household and the wider environments move forward. of institutions, structures, and policies. In January 2002, BRAC started a new ex- The results from the pilot phase suggest perimental program with these challenges in that the targeting of the TUP has been suc- mind, called "Challenging the Frontiers of cessful. Comparison with the general Poverty Reduction: Targeting the Ultra IGVGD membership profile suggests that Poor,"or TUP, for short. The program was TUP is targeting people more poor than designed to address the various interlocking VGD members. [Insert Table 6] constraints affecting rural women who live in The changes that are taking place in the severe poverty. The strategy is to help the lives of the ultra poor targeted by the pro- ultra poor build a solid physical and socio- gram also show promising results. The political asset base. The first two years of the program, according to a recently conducted program were the pilot phase, during which mid-term review, has produced noticeable 5,000 ultra-poor households were selected gains in social and human capital. A each year. In 2004, the program will be scaled consolidated view of the profiles indicates a up, and 10,000 more ultra poor will partici- majority have escaped their ultra-poor status pate. In the remaining two years (2005 and and acquired a poor but improved standard Scaling Up Poverty Reduction 91 Managing Scaling-Up Challenges of a Program for the Poorest Table 6 Key Differences between VGD and TUP Members Variables TUP VGD Average owning land 2.13 4.72 % of households owning no 93 87 cultivable land % of households not owning the 54 43 land they live on % of households reporting out- 2.13 36 standing loan from any source % of households reporting they 2.13 61 eat at least two meals a day % of households reporting deterioring economic condi- 44 35 tions over the last year of living. About 40 percent are moving Consortium, 2004). This very early result toward the standards of households targeted suggest that the TUP model may actually be by microfinance institutions. Although quite cost effective. predictions are premature, about 75 percent In the long run, the most important have some prospect for sustaining these variable will be the proportion of the ultra improvements into the future. poor who manage to build sustainable At $291 per member, financial allocation livelihoods. Joining mainstream microfi- for a 18-month cycle of the TUP program nance programs, especially being able to is more expensive than IGVGD, which has participate over a long time, can be an a financial allocation of $225. However, the important marker. This has been widely two programs target different groups of the used to assess graduation in IGVGD. extreme poor. TUP focuses on those who Although this will certainly be an important fail to benefit from the VGD card or do not indicator for monitoring graduation, it may get or it. The TUP approach is different and not be the only one, and the challenge will requires a more comprehensive, and more be to find more appropriate ones for the expensive, set of instruments for the types ultra poor targeted by the TUP program. of ultra poor the program targets. The main question is the extent to which the benefits Bringing in the Poorest provided by the program are sustainable. BRAC's conviction, underlying its micro- It may be too early to assess sustainabil- finance programs for the poorest, has ity, but the mid-term review found that been simple yet unconventional--that the the mean increased net income for TUP poorest can and should be included. The members is actually significantly higher than strategy for BRAC was not to force the that of IGVGD members (CFPR Donor existing model of microfinance, as it was 92 Case Studies Case Study of BRAC's IGVGD Program, Bangladesh clear that the model itself was part of the only bypassed the poorest, but mainstream problem. BRAC knew from experiences microfinance discourse did not consider that the interlocking constraints and depri- them to be an area of concern or relevance vations, within which extreme poverty ex- for microfinance. ists, are too complex to tackled by microfi- BRAC's experiences suggest that nance alone as a strategic intervention. carefully designed strategic linkages that The IGVGD program used the period of combine grants with a central role for security provided by the food aid to develop microfinance can work for the poorest. the skills and confidence of the poorest, and There will surely be many different models finally provide them with credit to build on. and approaches for including the poorest, The TUP program was based on the expe- which will vary according to country con- rience of IGVGD--that there are people texts. However, the starting point will have too poor to benefit from IGVGD, who still to be reversing the trend of apathy, that should not be ignored as an eventual target either excludes the poorest or treats them as group for microfinance. "relief cases" to be dealt by "others." The The idea of building inclusive markets poorest are, can, and must be central to the and systems, especially focusing on those vision and commitment of microfinance who tend to be left out, has been central to institutions. Only then will the search for the microfinance movement. Yet, ironically, possibilities and opportunities to include mainstream microfinance models have not the poorest begin and develop. End Notes 1 S. Hashemi, Including the Poorest: Linking 5 Imran Matin and D. Hulme. "Programs for the Poorest: Microfinance and Safety Net Programs. CGAP Focus Learning from the IGVGD Program in Bangladesh," Note No. 20, (Washington, D.C.: CGAP, 2001). World Development 31, no. 3 (2003): 647-65. 2 H. Rahman and M. Hossain, eds., Rethinking Rural 6 In 1998, the focal ministry for the VGD program was Poverty: Bangladesh as a Case Study, (Dhaka: UPL, changed from Ministry of Relief and Rehabilitation to the 1995). Ministry of Women's and Children's Affairs. 3 P. Webb, et al, "Expectations of Success and 7 F.H. Abed, "Programs for the Extreme Poor: BRAC Constraints among IGVGD Women," report prepared for Experiences So Far," paper presented at the interna- World Food Program, Bangladesh, 2001. tional conference, "Staying Poor: Chronic Poverty and Development Policy," Chronic Poverty Research Centre, 4 Hashemi, Including the Poorest, p. 11. IDPM, University of Manchester, UK, April 7-9, 2003. 8 Matin and Hulme, "Programs for the Poorest." Bibliography Abed, F.H. "Programs for the Extreme Poor: BRAC CFPR Donor Consortium. "Review of the CFPR/TUP Experiences So Far." Paper presented at the interna- Specially-Targeted Ultra Poor (STUP) Program." Mission tional conference, "Staying Poor: Chronic Poverty and Report, 2004. Development Policy," Chronic Poverty Research Centre, IDPM, University of Manchester, UK, April 7­9, 2003. Scaling Up Poverty Reduction 93 Managing Scaling-Up Challenges of a Program for the Poorest Hashemi, S. "Those Left Behind: A Note on Targeting Rahman, H., and M. Hossain, eds. Rethinking Rural the Hardcore Poor." In Who Needs Credit? Poverty and Poverty: Bangladesh as a Case Study. Dhaka: UPL, Finance in Development, ed. G. Wood and I. Sharif. 1995. Dhaka. Bangladesh: UPL, 1997. Sattar, M.G., N.S. Chowdhury, and M. Hossain. "Food Hashemi, S. Including the Poorest: Linking Microfinance Aid and Sustainable Livelihoods: BRAC's Innovations and Safety Net Programs. CGAP Focus Note No. 20. against Hunger. Research and Evaluation Division." Washington, D.C.: CGAP, 2001. Report prepared for BRAC, July 1999. Matin, I., and D. Hulme. Programs for the Poorest: Webb, P., et al. "Expectations of Success and Learning from the IGVGD Program in Bangladesh. Constraints among IGVGD Women." Report prepared World Development 31, no. 3 (2003): 647-65. for World Food Program, Bangladesh, 2001. 94 Case Studies Chapter 7 Bank Rakyat Indonesia (BRI): Twenty Years of Large-Scale Microfinance BY KLAUS MAURER Executive Summary Established 20 years ago, the Unit system of Bank Rakyat Indonesia (BRI) today is the largest and one of the most successful microfinance institutions in the world. The 3,855 Units--small outlets mostly in rural areas with around six staff members each--are scat- tered all over Indonesia and provide services to almost 30 million small savers (the average account is US $108) and 3.1 million small borrowers (average loan outstanding is US $540). The BRI Units have followed a profitable, sustainable approach to microfinance on a large scale, based on locally-mobilized savings without subsidies and funds from govern- ment or donors. The commercially-based provision of credit and savings services has had a powerful positive impact on the lives of millions of poor and low-income households. Several factors have driven the reform and implementation of the BRI Unit system. Effective leadership, strong commitment, and political support were crucial at the initial reform stage but also throughout the development process. The institutional design of the BRI Unit as the nucleus of the entire system combined standardization and flexibility in a unique way. The BRI experience drew extensively on the lessons and experimentation of other initiatives and from BRI's own trials and pilots. External factors gave the impetus for the initial reform. Later on, stable macroeconomic conditions and a series of financial sector reforms provided a conducive environment in which the new Unit system could develop and prosper. When the Indonesian banking system collapsed in 1998, BRI's Unit system remained profitable, loan repayment rate stayed high, and the deposit volume more than doubled. The BRI Units emerged from the crisis stronger and even more robust than before. Implementation Process Origins of the BRI Units Bank Rakyat Indonesia (BRI) is a state-owned commercial bank that has historically concen- trated on providing banking services to the agricultural sector and rural areas of Indonesia. BRI Scaling Up Poverty Reduction 95 Twenty Years of Large-Scale Microfinance established 3,600 village units in the early separate balance sheets and profit and loss 1970s as conduits for channeling subsidized statements for each Unit. Almost one third credit to farmers under the BIMAS program. of the Units was identified as having low po- When the government decided to terminate tential and were downsized to village service the BIMAS program, and with it the massive posts. Many Units were physically relocated subsidies to unit operations, BRI was faced from out-of-town field sites to central loca- with a difficult choice: either introduce dras- tions close to the markets. tic measures to make the village units viable, New products were introduced. A pair or close them down. With the encouragement of products, KUPEDES (for credit) and of the Ministry of Finance, BRI decided SIMPEDES (for savings), have become the to convert the units into a rural banking backbone and trademark of the new Unit network that would meet a wide range of system. KUPEDES is a single loan product financial needs of rural households in a sus- for general rural credit. It is non-targeted tainable manner. and is available to any creditworthy cus- The timing and conditions for a tomer for any kind of productive enterprise. "big-bang reform" of the BRI Unit system KUPEDES interest rates were set at 1.5 were favorable as the collapse of BIMAS percent flat per month (an annual effective coincided with important events: the rate of 33 percent). KUPEDES borrowers collapse of oil prices and the sharp decline must provide sufficient collateral to cover in oil revenues which put heavy strains on the value of the loan, usually in the form of the government budget, a new cabinet and land titles, but also can pledge buildings, key personnel changes in the government motorcycles or other property. KUPEDES (finance minister), and the launching of also has a timely repayment incentive, fundamental economic reforms. A major equivalent to a refund of 25 percent of the financial sector deregulation package was interest paid on the loan. announced in June 1983 allowing banks to SIMPEDES is a simple passbook savings set their own interest rates; this created product that was introduced after pilot-test- opportunities and the enabling environ- ing and experimentation. Savings have been ment for a viable rural banking operation.1 an integral part of the Unit banking philos- ophy and strategy from the outset. Because Establishment of the New Unit System more people in rural areas tend to be savers (1984­86) than borrowers at any one time, providing Viability guided the transformation of the better savings services was seen to be more Units from conduits for BIMAS (which was effective in achieving an equitable distribu- not profitable) to full-service rural banking tion of banking services than providing units. A massive restructuring took place. cheap credit. In addition to SIMPEDES, the One of the first steps taken was reorganizing main product, the BRI Units offer demand the Units into profit centers and the creating and time deposits, as well as another savings 96 Case Studies Bank Rakyat Indonesia product called TABANAS, which is mainly easily absorbed the Units' excess liquidity targeted at school children and students. and channeled the funds into large corpo- KUPEDES and SIMPEDES had a suc- rate loans that were rapidly expanded dur- cessful start, and after three years it became ing the 1990s. clear that the BRI Unit system could be fi- In 1997, the Unit Banking System nancially viable. In 1984, only 14 percent of recorded more than 18 million savers and the Units were profitable. Two years later, 2.6 million borrowers (see Table 1). This this had increased to 72 percent. The BRI dimension of outreach clearly established Unit system as a whole achieved profitabil- the BRI Unit system as the major institution ity in 1986, the third year of operations. in rural finance in Indonesia. Expansion and Scaling Up (1987­97) Financial and Economic Crisis (1998­99) The initial phase of transformation and es- In late 1997, Indonesia was hit by a severe tablishment of the new BRI Unit system financial and economic crisis. Within a few was followed by a decade of rapid expansion months, the country's currency--the and scaling up. The expansion was fueled by rupiah--experienced a dramatic 80 percent continued deregulation and reforms of the plunge in its value against the US dollar, financial sector and by macroeconomic followed by sharp increases in inflation (77 growth and stability. Every year, the BRI percent in 1998) and interest rates. The Unit system added an average of 100,000 ensuing political crisis mounted in May borrowers and 1.5 million depositors. 1998 with the downfall of President In the sixth year of operation in 1989, Suharto. The banking system was on the the BRI Unit system achieved self- brink of collapse, forcing the government to sufficiency in funding when the volume of step in and provide a blanket guarantee deposits was equal to the loan amount cover for all bank deposits. Many private outstanding. Since then, deposits continued banks were closed down, major state banks to outstrip loans both in terms of number merged, and non-performing assets trans- of accounts and volume. In terms of ferred to the newly-created Indonesian volume, the deposit-to-loan ratio has been Bank Restructuring Agency (IBRA). about 2:1, with the result that only half The BRI Units weathered the crisis pe- of the funds mobilized by the Units were riod remarkably well. On the one hand, the recycled as KUPEDES loans, and the other BRI Unit system experienced an enormous half were transferred to the branch system. influx of deposits because it benefited from For the Units, the fund transfer was attrac- its status as a state-owned bank and safe tive because they were paid an interest haven for depositors. In 1998 alone, more rate--the so-called transfer price--which than three million new deposit accounts was normally set slightly above the interest were opened, and the volume of deposits in rate paid on time deposits. The branches Rupiah doubled. On the other hand, the Scaling Up Poverty Reduction 97 Twenty Years of Large-Scale Microfinance Table 1: Deposits and Loans in BRI Units 1984-2003 DEPOSITS LOANS OUTSTANDING Year Number Amount Amount Number Amount Amount End (in thousands) (Rupiah billion) (US$ million) (in thousands) (Rupiah billion) (US$ million) 1984 0.003 42 39 641 111 103 1985 0.04 85 76 1,035 229 204 1986 419 176 107 1,232 334 204 1987 4,184 288 174 1,315 430 260 1988 4,184 288 174 1,386 542 313 1989 6,262 959 533 1,644 847 471 1990 7,263 1,695 892 1,893 1,382 727 1991 8,588 2,541 1,276 1,838 1,456 731 1992 9,953 3,399 1,644 1,832 1,649 798 1993 11,431 4,325 2,058 1,896 1,957 931 1994 13,067 5,232 2,379 2,054 2,458 1,118 1995 14,483 6,016 2,631 2,264 3,191 1,395 1996 16,147 7,092 3,026 2,488 4,076 1,739 1997 18,143 8,837 2,424 2,616 4,685 1,285 1998 21,699 16,146 2,044 2,458 4,697 595 1999 24,236 17,061 2,419 2,474 5,957 845 2000 25,823 19,115 1,986 2,716 7,827 813 2001 27,045 21,991 2,105 2,790 9,873 945 2002 28,262 23,480 2,627 3,056 12,011 1,344 2003 29,869 27,429 3,240 3,100 14,183 1,675 number of KUPEDES loans stagnated dur- to honor their obligations despite the ing the two years of crisis, and the portfolio economic hardships they faced. Microenter- level declined in real terms. prises primarily employed family labor and Most surprisingly, however, loan repay- their own capital, and were able to adjust ment suffered only marginally. Contrary to more flexibly to the external shock. Further- the massive defaults of large and corporate more, the impact of the crisis on the poor customers in the Indonesian banking sector, was generally more severe in urban areas the KUPEDES borrowers maintained a than in the countryside.2 A key factor was strong repayment discipline and continued also the long-term banking relationship that 98 Case Studies Bank Rakyat Indonesia had developed between the Units and their million people, had fallen below the poverty customers.3 Amid a general credit crunch in line, from around 15 percent before the the banking sector, KUPEDES loans were crisis. The Poverty Alleviation Committee, continuously available to existing customers established in 2001, has launched a ten-year during the crisis years, although Unit man- program of poverty alleviation until 2010 agers did not seek out new clients during and has set an intermediate target to reduce that period. Borrowers were particularly poverty to 14 percent of the population, or anxious to continue having access to BRI's 26.8 million people, by the end of 2004. credit facilities because such credit availabil- The Committee has realized the important ity represented a form of insurance for role of the banking sector and has obtained dealing with external shocks. Hence, they the commitment of major banks, including strived to keep their borrowing history BRI, to expand their lending to micro and good and placed a high priority on repaying small enterprises. their loans to BRI. In an attempt to provide access to com- The impact of the crisis on the viability of mercial credit for the enterprising poor in In- BRI Unit operations was thus only mar- donesia on a massive scale, in January 2000 ginal. The BRI Units were able to maintain BRI introduced a small-scale KUPEDES their profitability, albeit at slightly lower lev- product with simplified administrative proce- els. The BRI Units came out of the crisis dures and flexible collateral requirements for stable and robust, contrary to the rest of loans under Rp 1 million (US $120). BRI. Due to heavy loan losses in the corpo- Although BRI has developed a significant rate business unit, BRI almost collapsed and portfolio of small-scale loans, it has not ex- was recapitalized by the government. panded to full capacity in the small-scale credit market, and there remains considerable Post-crisis Period (2000­Present) scope for financing a significant portion of mi- In the four years since the crisis, the BRI cro-entrepreneurs below the poverty line. But Unit system continued to expand in scale. there are also limits to the provision of very At the end of 2003, the Units recorded small loans in a financially sustainable way. 30 million deposits and a volume of US Despite the simplified administrative require- $3.1 billion. The number of borrowers ments, the break-even point for small-scale increased to 3.1 million, with a total loan lending is Rp 1.2 million (US $143) when ac- amount outstanding of close to US $1.7 counting for the full cost of lending.4 billion. With less than five percent of the The latest development has been the portfolio at risk, the quality of loans privatization of BRI. The sale of 41 remained excellent. percent of BRI shares to the public through The crisis had brought a considerable rise an IPO was completed in November in poverty. In 1998, over 24 percent of 2003. The IPO was successful and drew Indonesia's population, or almost 50 strong interest from investors--its shares Scaling Up Poverty Reduction 99 Twenty Years of Large-Scale Microfinance oversubscribed by 16 times. The IPO was within the system. This has formed the the largest equity deal since the financial cri- basis for transparency, accountability, and sis. Total proceeds from the IPO amounted efficiency in the use of resources. BRI has to Rp 4.1 trillion (US $486 million), of set clear benchmarks for staff productivity: which around Rp 2.5 trillion (US $297 one credit officer for every 400 borrowers, million) is to be transferred to the state one teller for every 200 daily cash transac- budget and the remaining is to increase tions, and one bookkeeper for 150 daily BRI's capital.5 What impact this partial pri- transactions. Nowadays, a credit officer is vatization will have on the microbanking responsible for more than 500 borrowers business and the Unit system is yet to be and a teller handles an average of 6,000 seen. However, with the BRI IPO, microfi- deposit accounts. The computerization of nance has made a successful entry into the BRI Units in the course of the 1990s private equity markets, and this will send a has significantly contributed to the gains in powerful message to policy makers and staff productivity. As a result, cost efficiency bankers in Asia and the developing world. of the Unit system has increased consider- ably over the years. In 2000, administrative costs as a percentage of the average loan Impact Analysis portfolio came down to about 8 percent, Looking back on 20 years of implementa- which is very efficient by microfinance tion, the BRI Unit system has had a strong standards. impact on rural microfinance in Indonesia and beyond. The immediate impact was felt Effectiveness by millions of microenterprises and poor Judging from the BRI Units' success in households who benefited as savers and/or building and maintaining a large, stable, and borrowers from the financial services offered growing customer base, it is clear that BRI by the BRI Units. Moreover, the BRI expe- has been highly effective in meeting the rience was a powerful demonstration to rural population's demand for savings and policy makers and other financial institutions credit services at a reasonable cost. Beyond in Indonesia and in many other countries. Its these direct effects, the performance of the impact can be measured in terms of effi- BRI Units had a significant impact on finan- ciency, effectiveness, and poverty outreach. cial sector policy in Indonesia. The design of the 1988 decree (so-called PAKTO) for Efficiency in the Use of Resources setting up private rural banks was strongly Efficiency and productivity have been key influenced by the initial success of BRI's operating principles in the BRI Unit commercial approach to rural microfinance. system. The Unit system functions as an The BRI Unit system has emerged as the independent profit center within BRI, and leading institution in rural areas and has set each small BRI Unit is a profit center the benchmark for rural microfinance in 100 Case Studies Bank Rakyat Indonesia Indonesia. Moreover, BRI has gained inter- in terms of income or wealth.7 It remains to national recognition as the most prominent be seen whether the small-scale KUPEDES showcase of large-scale microfinance and product introduced in 2000 can address has become a learning ground for policy these issues and successfully expand lending makers and practitioners from many coun- to poorer segments of the population. tries. The BRI experience served as a model Much of the debate on poverty outreach of for institutional design in Cambodia, for the BRI Unit system (and of other microfi- savings product development in Thailand, nance institutions) has focused on loans and and as input for microfinance legislation in on the poor as potential borrowers. One may Tanzania, to name a few examples. argue, however, that poor people are more likely to be found among the BRI Units' 30 Poverty Outreach million savers (average deposit US $108) than Over the past 20 years, the BRI Units have among the three million borrowers. provided financial services to millions of Savings are a cushion against emergencies microenterprises and rural households, for poor households. They help reduce their most low-income in the general population vulnerability and provide them with a tool and many among the poorer of the rural for managing uncertainty and risks. For population. BRI's KUPEDES lending many poor people, credit may not be appro- program does not specifically target the very priate as it may not lower but rather increase poor below the poverty line but rather the their risk. Findings of the recent impact working poor who have viable economic evaluation seem to confirm this hypothesis activities and sufficient repayment capacity. for rural Indonesia. When potential borrow- Poverty data for BRI Unit borrowers are ers with viable enterprises were asked why not available, but proxy indicators (such they didn't borrow from formal financial in- as the number of small KUPEDES loans) stitutions, two thirds said they did not want provide an indication of depth of outreach. to be indebted.8 It would be worthwhile to Presently, the average outstanding loans study the characteristics of BRI Units' small are US $540, which is about half of the per savers in more detail, as well as the impor- capita income in Indonesia. In 2001, 60 tance of savings services for the poor, and percent of KUPEDES loans were below US the impact of savings services in terms of $300.6 However, KUPEDES lending has poverty reduction. not reached much of the population below the poverty line or the very poor, but rather Driving Factors those near the poverty line. A recent impact evaluation found that regular BRI Unit Commitment and Political Economy borrowers are relatively better off than for Change other respondents, including BRI savers Effective leadership, strong commitment, only and non-customers, whether measured and political support were crucial factors Scaling Up Poverty Reduction 101 Twenty Years of Large-Scale Microfinance not only at the initial reform stage but also sectors, BRI continued to implement subsi- throughout the development process of the dized credit programs for priority sectors BRI Unit system. In 1983, several events and specific target groups. These programs converged, to create the political economy were shifted to the BRI branches at the dis- and climate for change. The government trict level, strictly separate from Unit oper- launched overall economic and financial ations. This way, BRI was able to accommo- reforms, which prepared the groundwork date specific requests and special programs for a radical reform of the BRI Units. The from the government and from donors, blueprint for the new BRI Unit system was without disturbing the development of the perfectly in line with the financial sector Unit system. reforms announced in June 1983, thus winning the support of the architects of the Institutional Innovation reform in the Ministry of Finance and the The BRI Unit--as the nucleus of the entire Central Bank. system--is itself an institutional innovation. Commitment and leadership within BRI Much of its success may be attributed to the have been essential. BRI's president organizational set-up of the single BRI Unit director from 1983 to 1992 took personal as a highly decentralized and semi-au- initiative and responsibility for the develop- tonomous financial entity. The BRI Unit is ment of the Unit banking system. Jointly commonly found in a central location of the with the other members of the Board, he sub-district town, often near the market protected the Units from interference and place. It typically rents a one-room office, in led the development of what they called order to keep overhead costs low. A Unit BRI's new institutional culture and Indone- covers about 16­18 villages at the sub-dis- sia's new rural banking. system.9 The change trict, and nowadays serves an average of in culture was associated with the shift from 10,000 savers and a little over 1,000 subsidized farm credit to commercial borrowers. The individual Unit was pur- microbanking. Starting from the top man- posely kept small, by limiting the number of agement, it triggered a major transformation staff and focusing its operations. The four of the entire institution by changing the staff--a manager, a loan officer, a teller, and mindset of its 14,000 employees. a desk officer--have clear job descriptions Despite being part of a government- and division of responsibilities. Personnel owned bank, the BRI Units were able to responsibilities and performance-based maintain their operational autonomy and to standards are harmonized across the Unit stay free from interventions, such as credit system. As volume of operations increase, up targeting, interest rate restrictions, provi- to 11 additional staff are posted to a Unit. If sion of cheap funds, or from interference in the business of a Unit expands beyond the lending decisions. However, as the desig- maximum staff limit, the Unit is split to keep nated bank for the rural and agricultural the operation small and focused. 102 Case Studies Bank Rakyat Indonesia The accounting system allowed each learned how to analyze the viability of Unit's performance to be evaluated as profit informal microenterprises. The design of center. Unit managers and staff are account- the KUPEDES product, the analysis of able, which has instilled a high degree of borrowers' repayment capacity, and the responsibility among them. A standardized system of monitoring and collection were management information system (MIS), all built on this early experience. centered on a few key performance indica- Pilot testing became standard for BRI. tors, provided timely information to Until 1983, BRI had virtually no experience managers and supervisors at all levels. On with rural savings mobilization. Following this foundation the staff incentive system an initial demand study, a first version was built. of SIMPEDES was introduced as a pilot Overall, the Unit as the institutional nu- project in November 1984, and quickly cleus combines standardization and flexibil- showed evidence of massive demand for a ity; it can be easily replicated and easily liquid, convenient, and safe deposit facility. adapted to the scale of operations in a par- After some modifications and refinements, ticular area, providing an ideal institutional the facility was expanded to all Units by solution for expansion and scaling up. September 1986. The BRI Units also learned from mis- Learning and Experimentation takes. For example, continuous access to The BRI experience is based on extensive credit is an important incentive for borrow- learning and experimentation. Learning ers to repay their loans. When tight mone- from others dominated the period before tary policy caused a liquidity shortage in implementation while learning by doing BRI and the management imposed a halt in (including pilot testing) were the main lending in 1991­92, loan repayment dete- features during implementation. riorated because borrowers perceived the BRI did not invent all the features that availability of future loans to be at stake. characterize the Unit system. It pointedly During the recent financial and economic studied the experiences of others before crisis (AUTHOR: which crisis, or when?), setting up its own system, such as, Bank the KUPEDES lending window remained Dagang Bali, a private bank founded in open to credit-worthy borrowers. 1969; the Badan Kredit Kecamatan (BKK), a community-based institution in External Catalysts Central Java; and informal moneylenders, External factors played an important role specifically that they collected valuable for the initial reform. The oil price collapse insider information on prospective borrow- in 1983 and the decline in oil revenues ers from input suppliers and buyers. BRI forced the government to impose austerity had experimented with the Kredit Mini and on budgetary expenses and subsidies. Kredit Midi products for several years and Economic pressures made politicians adopt Scaling Up Poverty Reduction 103 Twenty Years of Large-Scale Microfinance a commercial approach to rural microfi- practitioners from all over the world. BRI nance. During the scaling-up period, stable has established an International Visitors macroeconomic conditions and a series of Program (IVP) to facilitate international financial sector reforms provided a exchange and learning, and every year, conducive environment in which the new about 20 international delegations visit and Unit system could develop and prosper. A study the BRI Unit system. decade later, the BRI Unit system was stable and robust enough to weather the Reforming a state-owned bank and utiliz- severe economic crisis. ing existing infrastructure is possible within a short period of time. External assistance was crucial, especially For scaling up poverty reduction, the major in the early years. Technical assistance was lesson to be learned is perhaps that reform- provided by the Harvard Advisory Group, ing a state-owned bank, and utilizing the partially funded by the World Bank and existing infrastructure and human resources USAID. Together with local bankers, the to implement a sustainable approach of experts developed the initial design and large-scale microfinance, is possible within principles of the new BRI Unit system. a short period of time. Commercially-based Financial resources were only required in provision of credit and savings services has the very beginning. With the know-how had a powerful positive impact on the lives and the systems in place, the BRI Units of millions of poor and low-income house- were soon able to generate their own re- holds, based on locally-mobilized savings sources locally. A World Bank loan of US without subsidies and funds from govern- $97 million for KUPEDES onlending was ment or donors. This required a change of disbursed in 1990, at a time when, strictly culture: treating the poor no longer as speaking, the BRI Unit system no longer beneficiaries but as customers who can save, needed outside funds. who are able and willing to pay market prices for good services, and who honor Lessons Learned their obligations and repay their loans Many lessons can be learned from the BRI despite economic hardships. experience. Some of its key principles have been replicated and adopted by the interna- Expand microbanking services to ensure tional microfinance industry as best prac- sustainability. tices. Some of the features are unique to the Some challenges remain and actions should specific context of rural Indonesia and/or be taken to ensure the sustainability of pres- to BRI as an institution, but others can be ent achievements. First, lending should be generalized and applied in other countries. further expanded. Even though the BRI The BRI experience has become a learning Units have done a remarkable job in extend- ground for policy makers and microfinance ing savings and credit services throughout 104 Case Studies Bank Rakyat Indonesia Indonesia, there is considerable scope for ture in order to improve the services expanding these microbanking services, to those who enabled the profits, Unit especially to poorer segments of the rural borrowers and savers. population. Most Indonesians still do not make use of formal banking services, so the Microfinance providers need competition. challenge for BRI is to address the needs of BRI Units need competition, which they this "unbanked majority."10 However, an have not had, and they hold a quasi-mo- expansion of borrower outreach can only be nopolistic position in the rural areas of achieved with additional staff, especially Indonesia with market shares of 74 percent loan officers who serve more than 500 (deposits) and 39 percent (loans). Policy borrowers on average. Management must makers should create a conducive policy ensure that the increase in staff--and thus environment, and donors should provide cost--is balanced by an increase in income support to other microfinance operators from lending to additional borrowers, in or- (for example, private rural banks) to der to maintain the profitability and sustain- encourage healthy competition. Competi- ability of Unit operations. tion is the driving force behind innovations, expanding outreach, and improving services Re-invest profits to ensure sustainability. to the poor. All these efforts are necessary BRI should not use the Units' profits to to maximize the impact of large-scale cross-subsidize the non-profitable parts microfinance on the reduction of poverty of BRI as has been done. Profits should and on the achievement of the Millennium rather be re-invested in the Unit infrastruc- Development Goals. End Notes 1 K. Maurer and H.D. Seibel, Agricultural Development 5 World Bank, Financial Sector Monthly Report-- Bank Reform: The Case of Unit Banking System of Bank October 2003, internal document, World Bank, Rakyat Indonesia (BRI), Rural Finance Working Paper, Washington, D.C., 2003. No. B5 (Rome: International Fund for Agricultural 6 The MIXMarket. 2004. http://www.mixmarket.org Development [IFAD], 2001). 7 Bank Rakyat Indonesia (BRI), BRI Microbanking 2 M.S. Robinson, The Microfinance Revolution, Volume Services: Development Impact and Future Growth 2: Lessons from Indonesia (Washington D.C.: World Potential (Jakarta: BRI; and Cambridge, Mass.: Bank, 2001). Harvard University, Center for Business and 3 R.H. Patten, J.K. Rosengard, and D. Johnston, Jr., Government, John F. Kennedy School of Government, "Microfinance Success amidst Macroeconomic Failure: 2001). The Experience of Bank Rakyat Indonesia during the 8 BRI Microbanking Services: Development Impact and East Asian Crisis," World Development 29, no. 6 (June Future Growth. 2001): 1957­69. 9 M.S. Robinson, "Savings Mobilization and 4 J. Marquez and J. Seward, Bank Rakyat Indonesia: Microenterprise Finance: The Indonesian Experience," Financially Sustainable Lending to the Enterprising Poor in The New World of Microenterprise Finance, ed. M. (Cambridge, Mass.: Harvard University, John F. Kennedy Otero and E. Rhyne (West Hartford, Conn.: Kumarian School of Government, 2002) Press, 1994). Scaling Up Poverty Reduction 105 Twenty Years of Large-Scale Microfinance Endnotes Continued 10 BRI Microbanking Services: Development Impact and Future Growth. Bibliography Afwan, I., and S. Charitonenko. Commercialization of Maurer, K., and H.D. Seibel. Agricultural Development Microfinance: Indonesia. Manila: Asian Development Bank Reform: The Case of Unit Banking System of Bank Bank, 2003. Rakyat Indonesia (BRI). Rural Finance Working Paper No. B5. Rome: International Fund for Agricultural Asian Development Bank. Technical Assistance No. Development (IFAD), 2001. 3810--Indonesia: Rural Microfinance. Frankfurt: Bankakademie International, 2003. The MIXMarket. 2004. http://www.mixmarket.org Bank Rakyat Indonesia (BRI). Introduction to BRI's Unit Robinson, M.S. "Savings Mobilization and Banking System. Jakarta: BRI, 1997. Microenterprise Finance: The Indonesian Experience." In The New World of Microenterprise Finance, ed. M. Bank Rakyat Indonesia (BRI). BRI Microbanking Otero and E. Rhyne. West Hartford, Conn: Kumarian Services: Development Impact and Future Growth Press, 1994. Potential. Jakarta: BRI; and Cambridge, Mass.: Harvard University, Center for Business and Robinson, M.S. 2001. The Microfinance Revolution. Vol. Government, John F Kennedy School of Government, 2, Lessons from Indonesia. Washington D.C.: World 2001 Bank. Charitonenko, S., R.H. Patten, and J. Yaron.. Indonesia: Patten, R.H., and J.K. Rosengard. 1991. Progress with Bank Rakyat Indonesia--Unit Desa 1970­1996. Profits: The Development of Rural Banking in Indonesia. Sustainable Banking with the Poor, Case Studies in San Francisco: International Center for Economic Microfinance. Washington, D.C.: World Bank, 1998. Growth. Hieman, W. Case Study: Bank Rakyat Indonesia. In The Patten, R.H., J.K. Rosengard, and D. Johnston, Jr. Challenge of Sustainable Outreach: How Can Public "Microfinance Success amidst Macroeconomic Failure: Banks Contribute to Outreach in Asia? Five Case The Experience of Bank Rakyat Indonesia during the Studies from Asia, ed. D. Steinwand and M. Wiedmaier- East Asian Crisis." World Development. 29, no.6 (June Pfister. Eschborn, Germany: German Agency for 2001): 1957­69. Technical Cooperation (GTZ), 2003. World Bank. KUPEDES: Indonesia's Model Small Credit Holloh, D. ProFi Microfinance Institutions Study. Program. Operations Evaluation Department, OED Denpasar, Indonesia: Bank Indonesia and German Précis, No.104. Washington, D.C.: World Bank, 1996. Agency for Technical Cooperation (GTZ), 2001. World Bank. Financial Sector Monthly Report­October Marquez, J., and J. Seward. Bank Rakyat Indonesia: 2003. Internal document, World Bank, Washington, Financially Sustainable Lending to the Enterprising Poor. D.C., 2003. Cambridge, Mass.: Harvard University, John F. Kennedy Yaron, J. Successful Rural Financial Institutions. World School of Government, 2002. Bank Discussion Paper 150. Washington, D.C.: World Maurer, K. Bank Rakyat Indonesia (BRI): Indonesia. Case Bank, 1992. study for CGAP Working Group on Savings Mobilization. Washington, D.C.: CGAP, 1999. 106 Case Studies Chapter 8 Integrating the Poor into the Mainstream Financial System: The BANSEFI and SAGARPA Programs in Mexico BY LISA TABER, WITH CARLOS CUEVAS AND INPUTS FROM JUAN NAVARRETE AND GABRIELA ZAPATA Executive Summary With the passage of the Popular Savings and Credit Act in 2001 and the subsequent launch- ing of a US $150 million program to strengthen savings and credit institutions and expand their outreach in marginal rural areas, the Mexican government embarked on a pioneering effort to alleviate poverty and increase income-generating potential by massively scaling-up access to safe and efficient financial services for the poor. The Savings and Credit Sector Strengthening Program, BANSEFI The Savings and Credit Sector Strengthening Program, implemented by the National Sav- ings and Financial Services Bank (BANSEFI), is building the capacity of more than 400 "popular savings and credit institutions" (or EACPs by their Spanish acronym) to meet new legal and regulatory standards and offer safer, more efficient financial services. The EACPs target almost 3 million users who are generally among the poor and lower-income groups. These clients typically have no access to the commercial banking sector, which currently reaches only about 25 percent of the adult population in Mexico's urban areas and has a much smaller presence in the rural sector. BANSEFI wants to ensure that the clients of every savings and credit institution gain access to safe and efficient financial services via its outreach expansion programs. Hence, BANSEFI is developing an information system to link EACPs (those that elect to participate) to its network federations and confederations, the banking supervisor, and BANSEFI itself. This shared technological platform will offer retail financial service advanced software and services--such as treasury management, funds compensation, and portfolio risk analysis--at lower costs available to a network of financial institutions than to individual institutions. With this technological platform (currently in its pilot stage), BANSEFI also spearheaded the development of a voluntary internet-based network of savings and credit institutions Scaling Up Poverty Reduction 107 Integrating the Poor into the Mainstream Financial System called "L@Red de la Gente," or "the Peo- government subsidies through L@Red de la ple's Network." L@Red links BANSEFI's Gente. This has not only made Oportu- 551 branches to more than 180 offices of nidades more transparent, but has had 19 participating institutions, and creates an the effect of coaxed the poorest Mexican expansive network of more than 730 families into using financial products. As branches. As of 2005, clients of any partic- these poor families earn interest on their ipating branch or institution will be able to new accounts, they learn how savings can make transactions on their accounts, for open access to other types of financial serv- most of the products offered, from any ices. This is the major permanent impact of other network participant.1 the program. As of December 2003, for More importantly, these individual insti- example, 78 percent of the 633,974 people tutions and branches have been incorpo- who received transfers through L@Red rated as nodes in the network of the from Oportunidades had savings accounts national payments system. As a result, with positive balances in branches of emigrants and public institutions can send L@Red. This suggests that this strategy is cash transfers securely and cost-effectively introducing savings services into the welfare to recipients in some of Mexico's poorest dynamics of poor families, which can be the and most remote locations. Federal basis of future asset accumulation. programs providing payments to targeted In addition to the transfers from govern- groups for health, education, and agricul- ment programs, the network also channeled tural production are already tapping the more than US $53.6 million in remittance potential of L@Red, which will expand to payments, sent from abroad to friends and include approximately 2,000 branches when family members. It is worth noting that the Savings and Credit Sector Strengthen- L@Red de la Gente offers accounts that link ing program is completed in 2008. by to. As different savings and investment products of December 31, 2003, L@Red de la Gente (standard savings accounts, housing-savings had distributed payments from these accounts, etc.). This feature is particularly programs to more than 1.5 beneficiaries, useful for remittances, since emigrants and this number is expected to grow to 3.3 can chose the proportions of the money million by the end of 2004. sent to be cashed or invested in different Among these programs, Oportunidades financial products. (a program that provides cash subsidies for health and education expenses to the Reaching the Rural Marginal Areas, poorest of Mexican families) is worth SAGARPA noting. This is one of the most successful As part of the same government strategy, anti-poverty programs in Mexico. Approx- the Secretariat of Agriculture, Livestock, imately 750,000 savings accounts have Rural Development, Fisheries, and Nutri- been opened as a new way to transfer tion (SAGARPA) is now working in 13 108 Case Studies The BANSEFI and SAGARPA Programs in Mexico states to set up new savings and credit insti- the new law requiring minimum perform- tutions, build the capacity of existing ance standards. EACPs. SAGARPA aims to bridge the As of January 2004, PATMIR had gaps in knowledge, understanding, and strengthened and increased the outreach ca- interaction between these financial service pacity of 41 EACPs in Chiapas, Huasteca, providers and the communities of poor and Guerrero, and Veracruz, improving the fi- mostly indigenous people living in the areas nancial institutions upon which nearly they serve. 24,000 people rely. In addition, more than A survey of Mexico's marginal rural areas 10,000 clients gained access to financial serv- in 2000 showed that just 2.5 percent of ices at new branches or institutions.Clients households had access to credit from a have more secure savings and can cash financial institution, and less than 6 percent checks at savings-driven institutions that used formal financial savings instruments. follow sound and prudent financial practices. Poor households in marginalized rural areas These institutions also offer other services, stand to gain the most from increased access such as credit funded by member savings and to financial institutions and services--where access to the payments system to receive they can safely build up liquid savings--be- remittances and government transfers (via cause currently they rely on informal savings L@Red de la Gente). mechanisms and physical assets, like small The program has now expanded to three livestock, that have low yields and high loss more regions of the country and is expected rates. They generally pay high transaction to sustainably integrate more than 80,000 costs and commissions to receive remittance people from Mexico's poorest, most mar- payments, and are typically charged about ginalized groups into the financial system 120 percent2 per year in interest if they before its completion in 2007. To achieve borrow money. this goal, PATMIR works extensively To combat these trends, SAGARPA be- with individuals and groups in these rural gan to implement the Rural Microfinance communities to educate them about the Technical Assistance Project (or PATMIR benefits of formal savings, how to use in Spanish) in 2001 to expand the network savings and other financial services to of EACPs into Mexico's poor and isolated their advantage, and the rights and respon- rural communities. PATMIR provides sibilities of membership in a cooperative EACPs with technical assistance and limited financial institution. Rather than quickly start-up funding to develop the outreach adding names to a membership roster, the capacity of their existing institutions, program focuses on cultivating conciencia, including opening new branches or savings confianza y compromiso--awareness, trust, and credit institutions. In each case, and commitment--between the new mem- technical assistance providers also ensure bers (who are typically from indigenous that participating institutions comply with groups and include everyone (women, Scaling Up Poverty Reduction 109 Integrating the Poor into the Mainstream Financial System youth, and the aged, as well) and their services in the future to alleviate poverty financial institutions, so that each under- and stimulate equity-enhancing growth. stands its investment in the other. The main lessons the BANSEFI and In addition, PATMIR helps the EACPs SAGARPA programs have revealed so far in- to incorporate new ways of attracting and clude the following: responding to marginalized clients. They Government intervention can effectively can hire multi-lingual staff to promote their increase access to financial services for services and introduce personal digital assis- low income households and businesses tants and mobile banking to make savings through getting the regulatory frame- more convenient and attractive. work right, building the institutional In 2003, more than 4.2 million families capacity of financial institutions, and received transfers through one federal safety subsidizing the deployment of modern net program alone, and US$13.2 billion in technology to increase the efficiency of remittances flowed into the country from service providers. Mexicans living and working abroad. Most of these funds went to the marginal areas Packaging the development of ade- where PATMIR is working. By incorporat- quate regulations and supervision, ing poor and marginalized households into institutional capacity-building, and the financial system and providing them technological infrastructure is more with safe savings facilities, families an likely to yield sustainable results and develop the capacity to better manage the cost less over time than traditional funds received and capitalize on these funds interventions which focus on increasing to generate more on their own. credit flows. Outreach expansion efforts among the Main Lessons poor in marginalized rural areas have Bringing the poor into the financial much better long-term prospects if system, especially those at the margins of they focus on building the stake that society in geographically isolated regions, these clients have in their savings and teaching them to better manage the cash credit institutions, as savers, rather than resources they have, and giving them access building financial relationships based to new sources of finance at lower costs is on the expectation of getting a loan. the root objective of the massive and ambi- tious programs undertaken since 2001 by BANSEFI and SAGARPA. While it is Implementation Process premature to assess the programs' overall impact, results to date indicate that they are Context likely to have far-reaching effects on the In 2000, Mexico was in the midst of a scope and direction of sustainable financial remarkable economic comeback and 110 Case Studies The BANSEFI and SAGARPA Programs in Mexico profound political transformation. After ten completely unregulated financial institu- years of gains in poverty reduction were tions to safeguard their savings or provide undone by the severe macroeconomic and loans, including a variety of savings and financial crisis of 1994­95, the financial credit associations, cooperatives, credit system had stabilized, credit to the private unions and NGOs. While some of these sector was growing, and the economy was instituciones de ahorro y crédito popular, or expanding at a rate of nearly 7 percent per popular savings and credit institutions,3 had annum. President Vicente Fox's electoral built a tradition of solid financial intermedi- victory in July ended 71 years of domina- ation over more than 50 years, the lack of tion by a single political party. It brought effective regulation and supervision of energy and hope into the policy-making and the sector as a whole put millions of people legislative processes with the widespread at risk of losing hard-wrought savings and expectation of more open, participatory, investments in poorly or unscrupulously and transparent government. managed institutions. These risks became Yet about 58 million Mexican people--50 more apparent with the failure of numerous percent of the total population--still lived in institutions due to fraudulent activities poverty in 2000. While the commercial bank- in 1998­2000. The vast majority of these ing sector had recovered from the collapse of institutions were small, community-based 1994­95, about three-quarters of all adults in organizations, moreover, with no links to metropolitan areas--and 85 percent of indi- the national payment system, limited prod- vidual entrepreneurs--had no access to these uct offerings, and low levels of technology financial outlets. Bank density ratios were low, and efficiency. with a country-wide average of one branch Despite these weaknesses, in the late per 12,000 people. In the southern states of 1990s policy-makers began to perceive Oaxaca, Chiapas, and Guerrero, bank branch that popular savings and credit institutions densities ranged from one for every 25,000 to could potentially play a key role in 30,000 people. In rural areas, where about alleviating poverty and stimulating half the population live in extreme poverty, inclusive economic growth. In 1999, the participation in the formal financial system Secretariat of Finance and Public Credit was predictably low. A survey of rural areas in (SHCP), the National Banking and Securi- the Oaxaca and Huasteca regions in 2000 ties Commission (CNBV), the Central showed that just 2.5 percent of households Bank of Mexico (BANXICO), and the had access to credit from a financial institu- National Savings Bank (PAHNAL) initiated tion, and less than 6 percent used formal a process of consultation with legislators, financial savings instruments. sector institutions, and the World Bank on About 3 million people countrywide, or the development of a new legal and regula- approximately 7 percent of the economi- tory framework for popular savings and cally active population, relied on loosely or credit institutions. Scaling Up Poverty Reduction 111 Integrating the Poor into the Mainstream Financial System Both the outgoing and incoming services through savings and loans, credit administrations employed a two-pronged unions, cooperatives and savings associa- development strategy to combat poverty: tions, which were operating about 1800 efficient and well-targeted public spending offices--including in some of the most combined with sound macroeconomic remote corners of the country--became a management. Clearly the lack of access to top public policy priority for Fox's adminis- safe, efficient financial services was affecting tration in 2000. The Consejo Mexicano de the ability of poorer Mexicans to get out of Ahorro y Crédito Popular (COMACREP), poverty through their own savings, the help a council representing at least 80 percent of of emigrant family members, or productive the sector, was formed to participate in the investment in their own enterprises. The ac- dialogue with government officials, thus cess problem was particularly acute in rural giving about 1.5 million clients or members parts of the country, where the government from communities throughout Mexico a had recently put forward major initiatives-- voice in the future of their organizations. such as dissolving the state grain purchasing After more than 16 months of analysis and institution--to curtail market distortions negotiations, the Popular Savings and and encourage the pursuit of new business Credit Act was passed in April 2001. opportunities in the NAFTA environment. The law, which became effective on June Agricultural growth picked up after 1998 4, 2001, provides for the gradual incorpo- and aggregate poverty trends declined, but ration of all non-bank savings and credit in- productivity gains were still low and one of stitutions into a new legal and regulatory every two Mexicans living in rural areas was framework over a four-year period, and extremely poor. Maintaining a stable macro- forms the basis of what is perhaps the most economic environment, funding social ambitious effort to massively scale-up access safety net programs, and investing in criti- to financial services for poor and marginal- cal infrastructure were key elements of the ized people in the world. Under the new Fox administration's poverty alleviation law, two types of popular savings and credit strategy. In addition, the new government institutions (referred to in Spanish as Enti- focused on enhancing competitiveness, par- dades de Ahorro y Crédito Popular, or ticularly of small-scale enterprises, by better EACPs), will be authorized by the CNBV integrating markets and targeting capitaliza- to collect deposits from the public and con- tion programs, especially at poorer groups duct a range of intermediation activities. in the rural sector. The scope of intermediation activities and prudential standards to which institutions Overhauling the Legal, Regulatory, and are held depends on the size, geographic Institutional Framework coverage, and demonstrated operational The potential to provide millions of people capacity of each entity, as distinguished by with efficient and competitive financial four distinct categories of operations. 112 Case Studies The BANSEFI and SAGARPA Programs in Mexico The CNBV is also in charge of licensing The Scope of the Initiative federations (voluntary groupings of retail The US$150 million program, which EACPs) and confederations (voluntary BANSEFI initiated in 2002 in coordination groups of federations), which play pivotal with other public sector agencies, including roles in the new regulatory system. To be the CNBV and the Secretariat of Agricul- licensed by the CNBV, an institution must ture, Livestock, Rural Development, Fish- first receive a favorable rating from the eries and Food (SAGARPA), encompasses Supervision Committee of its federation, the following : which evaluates candidates and subse- Consolidating and strengthening ap- quently supervises its licensed members--in proximately 400 popular savings and an "auxiliary" capacity--according to credit institutions to bring them into standards established by the banking au- line with new regulatory standards and thority.4 The decisions of the supervision improve their efficiency and outreach committees with respect to the entrance or capacity continued operation of a member EACP Strengthening EACP federations and can be overruled by the CNBV, however, confederations which maintains ultimate responsibility for the prudential oversight of the individ- Developing an information technology ual retail institutions. system linking EACPs, federations, con- A complementary law, enacted on June 1, federations, regulators, and BANSEFI, 2001 transformed PAHNAL into the which will help the retail outlets reduce National Savings and Financial Services Bank operating costs, improve the efficiency (Banco del Ahorro Nacional y Servicios of their administrative processes, and Financieros), or BANSEFI. PAHNAL had a facilitate regulatory compliance through 50-year history and network of 590 branches automated reporting focused exclusively on capturing deposits, Developing the capacity of BANSEFI mainly among small-scale clients, investing to provide second-tier central banking their funds primarily in government debt services to the sector, such as back-of- securities, and proving nodes in the payment fice functions, accounting, reporting system often used to distribute payments by and financial management, and liquid- the federal government. Under the new in- ity brokerage stitutional framework, BANSEFI retained PAHNAL's role of promoting savings habits, Strengthening the CNBV's capacity to and gained an important new set of man- adequately supervise hundreds of small, dates as a bank for the popular savings and dispersed financial intermediaries, the credit institutions and the coordinator of majority of which are engaged in an ambitious one-time program of public microfinance activities, through the support to the sector. auxiliary supervision structure Scaling Up Poverty Reduction 113 Integrating the Poor into the Mainstream Financial System Evaluating the social and economic im- start-up funding to develop new savings and pact of the program and the sector on credit institutions, or for existing institu- an on-going basis, to gauge its success tions to open new branches. In each case, and allow for adjustments as needed to technical assistance providers also assist meet poverty-alleviation objectives participating institutions in complying with Most importantly, the program will the new legal and regulatory standards. incorporate millions of new clients into the formal financial intermediation system. Implementing BANSEFI and Progress Retaining recipients of government transfer to Date payments as new account holders in a The government faced formidable challenges process called bancarization is already in developing and implementing the program rapidly integrating previously unserved to transform the savings and credit sector, as segments of the population into the approximately 90 percent of the existing insti- economy through their access to financial tutions lacked the experience and capacity services. needed to meet the standards of sound regu- BANSEFI and SAGARPA are also coor- lation and supervision. The great majority of dinating efforts through an innovative part- institutions, moreover, are very small, with ner project (implemented by SAGARPA) to single branches located in dispersed commu- expand the outreach of popular savings and nities throughout the country. credit institutions particularly among A census of institutions was initiated women and indigenous groups, in marginal in September of 2001 to determine more rural areas of Chiapas, Hidalgo, San Luis precisely the number and size of entities Potosí, Veracruz, Oaxaca, Guerrero, More- operating in the sector, their locations, and los, Michoacán, the State of Mexico, affiliations with other such organizations, if Puebla, and Tlaxcala. any. A more in-depth evaluation of a subset The Rural Microfinance Technical Assis- of those institutions provided more infor- tance Project, or PATMIR, provides train- mation on the financial health and techno- ing in basic principles of household finance logical capacity of popular savings and credit and participation in financial institutions for institutions, as well as their operational, groups and individuals in marginal rural management and governance capacity. communities. It also assists savings and This information was used to help gauge credit institutions in increasing their the extent--and cost--of the program of outreach to these clients by tailoring technical assistance needed to strengthen products, promotional materials, and the intermediaries. outreach methods to the needs of commu- Technical assistance and training are nity members. In marginal areas where no being provided to Supervision Committees intermediaries exist, PATMIR provides and EACPs countrywide, so they can evaluate professional technical assistance and limited every savings and credit institution which opts 114 Case Studies The BANSEFI and SAGARPA Programs in Mexico to participate, and strengthen or restructur- benefited from on-site training with expert ing those institutions which do not yet meet consultants and participated in seminars and the operational or solvency standards of the study tours to improve their capacity to ef- new legal and regulatory framework. fectively supervise the savings and credit To meet these objectives, BANSEFI sector through the auxiliary mechanism. has recruited international organizations The CNBV has also developed a number of specialized in providing technical assistance tools, including the Supervision Guidelines to savings and credit institutions and their for the federations' supervision committees, sector organizations through a competitive the Integrated Supervision Manual for in- bidding process that began in 2001. Action ternal use, and a specialized financial analy- plans have now been formalized for 381 sis system. institutions, of a total of 400 that are BANSEFI has undertaken initiatives to expected to participate in the program. disseminate the requirements of the new Those remaining have either chosen to stop law, and accompanying supervision and reg- mobilizing deposits or will be merged, ulatory framework. Seven workshops have acquired, or dissolved. been conducted jointly with the CNBV, In addition to their individual action each in a different region of the country, to plans, popular savings and credit institutions explain the changes, encourage institutions have the opportunity to attend training to participate in the sector consolidation sessions organized by BANSEFI. So far, and strengthening program. courses on risk management, credit analysis, Key to BANSEFI's role, however, are accounting, and governance have been the actions it has taken with respect to attended by 2,900 people from about 400 developing a technological platform which institutions. Training for federations and sector institutions can use to access back- confederations has emphasized the prepara- office or liquidity services, to file mandated tion of supervision committee members. As reports to their supervision committee and of November 30, 2003, seven auxiliary su- to the CNBV, and also to network with pervision committees had received intensive other institutions. The technology training in the law and applications of platform is expected to reduce the transac- the norms and guidelines developed for tion costs and expand the range of services evaluating and classifying institutions, and provided by the sector, in addition to have been certified by the CNBV. facilitating the accreditation of the entities The CNBV now has 27 people to by the CNBV. provide oversight for the sector through a The organizational basis of the commer- specialized unit, the General Directorate for cially-oriented network has already been the Supervision of Popular Savings and formed with the establishment of "L@Red Credit Institutions (or DGSEACP, by its de la Gente" ("The People's Network"), a Spanish acronym). Each of these staff have voluntary internet-based alliance of popular Scaling Up Poverty Reduction 115 Integrating the Poor into the Mainstream Financial System savings and credit institutions and BANSEFI. financial institutions, so that each under- L@Red has already linked BANSEFI's 554 stands its investment in the other. branches with more than 180 offices of 19 In addition, PATMIR helps the EACPs participating institutions to provide a geo- incorporate new ways of attracting and re- graphically expansive network of more than sponding to marginalized clients--by hiring 730 branches of coverage. multi-lingual staff to promote their services and introducing personal digital assistants Implementing SAGARPA and Progress and mobile banking to make savings more to Date convenient and attractive, for instance. In As of January 2004, PATMIR had so far 2003, more than 4.2 million families strengthened and increased the outreach ca- received transfers through one federal safety pacity of 41 savings and credit institutions net program alone, and US $13.2 billion in and their branches in Chiapas, Huasteca, remittances flowed into the country from Guerrero, and Veracruz, providing more Mexicans living and working abroad. Much than 10,000 clients with access to financial of these funds went to the marginal areas services from new branches or institutions, where PATMIR is working. By incorporat- and improving the financial institutions ing poor and marginalized households upon which nearly 24,000 people rely. into the financial system and providing The program now has operations in three them with safe savings facilities, families are more regions of the country and is expected developing the capacity to better manage to sustainably integrate more than 80,000 and capitalize on funds they generate and people from Mexico's poorest, most margin- receive from multiple sources. alized groups into the financial system before it finishes in 2007. To achieve this goal, Impact Analysis PATMIR works extensively with individuals The cost of the popular savings and credit and groups in these rural communities to ed- sector strengthening program, including the ucate them about the benefits of formal sav- Rural Microfinance Technical Assistance ings, how to use savings and other financial Project administered by SAGARPA and services to their advantage, and the rights the comprehensive sector strengthening and responsibilities of membership in a program underway through BANSEFI, cooperative financial institution. Rather than totals about US $150 million. The objective quickly adding names to a membership of this program is to develop a vibrant roster, the program focuses on cultivating and viable savings and credit sector with conciencia, confianza y compromiso--aware- dramatically increased financial services out- ness, trust, and commitment--between new reach, especially to poor and marginalized members (who are typically from indigenous rural populations. It is premature to assess groups and include everyone--women, the program's overall impact, given that youth, and the aged, as well) and their it is still being implemented and ex-ante 116 Case Studies The BANSEFI and SAGARPA Programs in Mexico cost-benefit analyses of this type of institu- EACPs by BANSEFI relates to liquidity tion-building initiative are problematic. The management. BANSEFI consolidates the degree to which program results translate liquidity of EACPs and invests it on their into poverty reduction will be carefully behalf in the commercial banking system, monitored, however, using panel data from thus securing a higher interest rate than 6,000 households that will be surveyed what the entities could obtain individually. annually over the five-year period beginning By late December 2003, a total of in 2004. 93 EACPs had already invested 626.2 mil- Implementation progress and indicators lion Mexican pesos--or US$55.8 million-- of impact, to date, indicate that the pro- in BANSEFI. gram is likely to achieve its objective of The technological platform for EACPs increasing the number of people served by currently being piloted by BANSEFI will participating financial institutions from 2 contribute significantly to their operational million to 5.5 million by 2008, and show efficiency and income-generating potential. substantial progress in poverty alleviation. Rather than each individual institution The benefits to members and clients of acquiring the technology components EACPs are reaped directly through lower needed to manipulate, transfer, and store costs of credit and payment services, the information needed to operate and positive yields and increased security on communicate with other sector institutions, savings deposits, and greater access to an EACPs can subscribe to these services array of government programs, including selectively and share the cost with other social insurance. participants. The creation of this common technology platform will allow for shared Benefits for Institutions operating, technological, and marketing When financial institutions become more processes, and a common trademark. efficient and financially sound, clients can In addition to compatible or shared hard- benefit from the lower costs and risks asso- ware and software, the alliance BANSEFI ciated with their products and greater access has created with sector institutions through to services. In addition to the advantages of L@Red de la Gente (which currently adequate supervision and regulation, the includes more than 730 participating economic benefits for the institutions branch locations) is expected to include participating in the program are clear. more 3,000 points of service by 2008. Efficiency gains will accrue to the EACPs L@Red aims at breaking a key barrier to from staff training, upgrading internal developing a viable system of financial serv- controls, and building credit appraisal and ices for the poor: gaining sufficient scale risk management capacity, and through the and scope to cost-effectively process large services provided by BANSEFI and network numbers of small-volume transactions. An- organizations. One such benefit provided to other important result of this technological Scaling Up Poverty Reduction 117 Integrating the Poor into the Mainstream Financial System interlinking is the ability to organize partic- of these assets into financial savings would ipating institutions as conduits for govern- benefit the poorest segments of the popula- mental support programs, which will make tion by increasing returns and providing it possible for institutions to increase their greater liquidity. revenues through fees for the distribution By providing poor and remotely located of financial services, and more importantly, households with access to the national pay- by retaining users and increasing their mar- ments system through the L@Red de la ket penetration. Gente, moreover, they can more securely and cost-effectively receive remittance Benefits for Clients payments from family members working Providing access to sound, efficient finan- elsewhere in Mexico or abroad, as well as the cial services clearly contributes to income benefits of social safety net and development generation and poverty alleviation at the programs. As of December 31, 2003, household level. This is particularly true L@Red de la Gente had distributed payments for the poorer rural households, which are from federal health, nutrition, education, expected to gain from access to adequate and agricultural production programs to deposit instruments and an improved more than 1.5 million beneficiaries, and the ability to take advantage of productive number is expected to grow to 3.3 million investment opportunities. by the end of 2004. By receiving transfer Economic analysis5 has shown that payments through L@Red, poor beneficiar- transforming informal, non-earning finan- ies are introduced to financial institutions cial assets, and monetizing even a fraction and encouraged to participate as members of the savings held in physical form by rural or clients. households, would have substantial eco- About 750,000 new savings accounts have nomic benefits in terms of safety and return. been opened at L@Red institutions through Informal means of savings that dominate one such program, Oportunidades, which the households' portfolios are, in principle, provides targeted cash subsidies for health and inferior to financial instruments in terms of education. This transfer mechanism has safety and returns. For example, tandas, or helped to introduce some of the poorest fam- rotating savings clubs, report a 6 percent ilies in Mexico to the financial system. As these rate of noncompliance (group members clients earn interest on their new accounts, who cease to contribute once they have they learn how savings can open access to taken a turn); and the most common forms other types of financial services. This is the of livestock holding, pigs and chickens, have major permanent impact of the program. As mortality rates above 40 percent. Since of December 2003, for example, about 80 holding livestock as savings is more preva- percent of those beneficiaries of Oportu- lent in poorer areas and among indigenous nidades who had savings accounts in branches people than elsewhere, the transformation of L@Red were maintaining a positive account 118 Case Studies The BANSEFI and SAGARPA Programs in Mexico balance. This suggests that this strategy is was in part the result of a historical evolu- introducing savings services into the welfare tion in Mexican politics. The election dynamics of poor families, which can be the of President Fox in July of 2000 was the basis of future asset accumulation. culmination of many years of work to create In addition to the transfer of government a political system characterized by greater programs, L@Red had also channeled more participation, openness, and accountability than US $53.6 million in remittance on the part of Mexicans from diverse social payments to recipients as of December 31, and economic groups and political affilia- 2003. L@Red de la Gente offers accounts tions. These same forces have arguably that link different savings and investment helped foster the political will needed to products--such as standard savings ac- embark on an ambitious and comprehensive counts and housing investment funds, for program to integrate into the formal instance--so that emigrants can chose how financial system those institutions that serve to allocate remittances among different people on the margins of society. types of accounts. The degree of economic stability Mexico SAGARPA is working to ensure that the regained over the last decade has also played most marginalized people living in largely an important part in providing the resources indigenous rural communities benefit needed to reform the EACP sector. from access to L@Red, as well as the other The results of increased economic integra- financial services that savings and credit tion with the United States through institutions provide. Through the PATMIR NAFTA, sound domestic macroeconomic project, more than 41 savings and credit management, and high oil prices since 1999 institutions and branches have received spe- have all contributed fundamentally to the cialized technical assistance to strengthen prosperity and stability that have clearly and increase their outreach capacity among facilitated the EACP reform. marginalized clients, so far improving or generating access to financial services for Institutional Innovation more than 34,000 poor people. Prior to directing its attention to the popu- lar savings and credit sector, the DRIVING FACTORS government had previously attempted to provide access to financial services primarily Commitment and Political Economy through first and second-tier development for Change banks and trusts. This approach generally The determination to reform and transform relied on directed credit at subsidized the popular savings and credit sector, as a interest rates, subsidized credit guarantees, means of alleviating poverty and integrating and debt forgiveness and restructuring. millions of poor and lower-middle class These interventions were unsuccessful in Mexicans into the formal financial system, providing sustained access to financial serv- Scaling Up Poverty Reduction 119 Integrating the Poor into the Mainstream Financial System ices for their target populations and re- both the government and the sector initiated quired large fiscal outlays, estimated to cost a dialogue that led to the design and approximately US$28.5 billion over the approval of the new law which governs the 1983­92 period. sector's activities. The government began introducing re- At the same time, international develop- forms in rural credit policies in the early ment agencies were putting increasing focus 1990s, reducing interest rate subsidies, on the potential of popular savings and making transfers to development banks credit institutions as a means for sustainable more transparent, and experimenting with increasing broad-based access to financial alternatives to promote sustainable, com- services. A substantial number of these types mercial intermediation. A project was of institutions had been offering financial launched in 1995 to provide subsidies for services in Mexico for decades--without commercial banks to establish services in any regulation or supervision. Low levels of small cities and rural towns where there was efficiency and a lack of widespread confi- no banking presence. Banks were reluctant dence on the part of depositors, however, to participate and retreated from the severely limited their outreach capacity. program and rural areas altogether with the While it became clear that savings and exchange rate crisis and rapid economic credit institutions represented great prom- decline of 1995­96. ise in achieving the outreach and sustain- In addition to these policy failures,the lack ability goals that so many public sector in- of an adequate legal and regulatory frame- stitutions and interventions had not, there work for savings and credit institutions com- were many difficult questions to answer pleft low-income depositors exposed to the with regard to regulatory framework and risk of losing their savings and constrained the nature and scope of public investment the growth of the sector. Numerous savings needed to shore up the sector. and credit institutions were thus supporting It was clear that the sector, which was the development of an appropriate legal and is dominated by small institutions-- framework for their activities. At the same only three of which have more than 50 time, the need for institutional change was branches--would likely take many years to reinforced by the detection of fraudulent consolidate and organize on its own. activities committed by managers of some BANSEFI was thus created to accelerate popular savings and credit institutions in the and enhance the process, through the late 1990s. Not only did thousands of benefits of networking, based on successful low-income depositors lose their savings in experiences in other countries. BANSEFI is these cases, but some municipal govern- not an ordinary development bank, as its ments that managed their funds through mission and methods exclude lending. To savings and credit institutions lost millions of avoid shifting risk to the public sector, dollars in scarce public resources. As a result, BANSEFI has concentrated on providing 120 Case Studies The BANSEFI and SAGARPA Programs in Mexico fee-based services without exposing itself to tions involved. Some institutions have resis- the credit risk of the EACPs or their clients. ted the changes, preferring to remain out- The value that BANSEFI provides to the side of the new legal and regulatory system. EACP network institutions is likely to be Some of them have tried to influence legis- transferred to them in the future when it lators in order to boycott the transition of puts itself up for sale to these clients. While cooperatives into either of the two types of designed from the experiences of successful savings and credit institutions considered in savings and credit sector institutions world- the Popular Savings and Credit Act.6 The wide, an institutional transformation of the inclusiveness that characterized the design nature and scope being undertaken by of the new law7 and its general acceptance BANSEFI is undoubtedly a first in the has prevented this resistance from garnering annals of development banking. much support from legislators or other At the same time, PATMIR was launched sector institutions, however. BANSEFI has in order to make sure that sound, efficient also organized an ongoing campaign to savings and credit services are delivered to inform and facilitate participation in the the poor in rural areas, which does not new system through greater understanding normally happen through a "trickle down of how and what it will require. effect." Large intermediaries have seldom BANSEFI has not only emphasized the shown any interest in the marginal rural relevance and practical aspects of imple- sector; they tend to focus on an urban and menting the new law in its communications semi-urban market. PATMIR therefore and outreach, but has had to focus on provides technical assistance and small explaining its own role in the sector as well. in-kind grants of support to institutions Indeed, its traditional mandate of promot- interested in developing tentacles to rural ing popular savings together with its new communities and also creates intermediaries role of coordinating public support for adept at responding to the needs of the the sector seem contradictory to some insti- people in these communities. SAGARPA's tutions. A few see BANSEFI more as a innovative program is to trying to demon- competitor rather than as an agent coordi- strate that the provision of financial nating support. As a result, BANSEFI has services--based on a cooperative savings- had to reiterate its intention to be sold to driven model--can be profitable in marginal the sector and shift its focus to providing rural areas. second-tier central banking services to popular savings and credit institutions.8 Learning and Experimentation In some cases, such as with the transfer The implementation of this ambitious new of government programs through L@ Red regulatory and development framework for de la Gente, branches of the sector have the savings and credit sector has been a been given priority when a branch of challenging process for many of the institu- BANSEFI is located in the same commu- Scaling Up Poverty Reduction 121 Integrating the Poor into the Mainstream Financial System nity. Communicating its mission has not External Catalysts been easy, nor have the effects of BANSEFI's The government's program of support for efforts to demonstrate its role as a partner the EACP sector, and for financial educa- to popular savings and credit institutions tion and the expansion of savings groups in been immediate. The perceptions of the marginal rural areas, was developed and is bank as a "competitor" are diminishing, and being implemented with the energetic and participation in the services and networks it proactive participation of numerous supports are increasing. development agencies and professional SAGARPA's efforts to expand the out- savings and credit institutions networks reach of financial institutions in marginal from many countries. rural areas have met the difficult challenge of USAID has sponsored US $2 million in educating people about the use and benefits technical assistance work to upgrade the of financial services. New promotional tools capacity of one of Mexico's primary savings and methods have been developed to and credit networks since 1999. The US explain how financial products work and Agency for International Development what improvements they offer over tradi- has also been active in helping forge tional savings instruments are. Technical alliances between financial institutions in assistance providers work with institutions to the United States and the savings and credit develop products tailored to the particular sector in Mexico to help lower the costs to needs of different marginalized communities Mexican emigrants sending funds back to in order to attract greater participation. their families. Some of these products include savings plans Through its Multi-lateral Investment and accounts tailored to specific groups, Fund, the Inter-American Development including savings plans for children, and Bank (IDB) has provided US $3 million in accounts set aside for school fees and Christ- technical cooperation for components mas expenses. Group savings plans have been of BANSEFI's information system that is created, called "tandahorro," which are being upgraded. IDB has also provided based on traditional rotating savings support to the CNBV for training in auxil- clubs; and specialized credit products, such iary supervision. The German government as credimujer for women and crediapuros for has also supported the sector with US $ 1.2 emergencies, have been designed to intro- million for federation restructuring. duce new users to the formal financial The World Bank has been a major spon- system. Many of these services are provided sor of the government's initiative, having directly at the community level by "promo- provided a loan of US $64.6 million in tores" or, in the case of a savings and credit 2002 to consolidate and strengthen savings institutions in Chiapas, via mobile banking and credit institutions, build supervision services which operate on market days on a capacity, pilot the technology platform for weekly or fortnightly basis. the sector, and expand access to financial 122 Case Studies The BANSEFI and SAGARPA Programs in Mexico services in rural marginal areas. The World Government intervention can effectively Bank is also preparing a new loan, worth ap- increase access to financial services for low proximately US $70 million, to help the income households and businesses by government finance the national roll out of getting the regulatory framework right, the technology platform. building the institutional capacity of financial institutions, and subsidizing the Lessons Learned deployment of modern technology to increase the efficiency of service providers. Bringing the poor into the financial system, Packaging the development of adequate especially those at the margins of society in regulations and supervision, institution geographically isolated regions, teaching capacity building, and technological them to better manage the cash resources infrastructure is more likely to yield sus- they have, and providing access to new tainable results and cost less over time than sources of finance at lower costs is the root traditional interventions which focus on in- objective of the massive and ambitious pro- creasing credit flows. grams undertaken since 2001 by BANSEFI Outreach expansion efforts among the and SAGARPA. While it is premature to as- poor in marginalized rural areas have sess the overall impact of the two programs, much better long-term prospects if they focus results to date indicate that they are likely on building the stake that these clients have to have far-reaching effects on the scope and in their savings and credit institutions as direction of sustainable financial services in savers, rather than building financial the future to alleviate poverty and stimulate relationships based on the expectation of equity-enhancing growth. The main lessons getting a loan. the BANSEFI and SAGARPA programs have revealed so far include the following: End Notes 1 L@Red de la Gente represents, for many financial governmental organizations (NGOs). Over 90 percent of intermediaries, an additional incentive to accelerate these intermediaries are CAPs. their formalization process, since complying with estab- 4 It is not obligatory for candidates to join a Federation. lished performance standards is a requisite to join this Nonetheless, they must obtain an opinion from a feder- strategic alliance. ation and submit to that federation's auxiliary supervi- 2 World Bank, Mexico Rural Finance: Savings sion. Any institution that does not wish to join a federa- Mobilization Potential and Deposit Instruments in tion may submit its application directly to the CNBV. The Marginal Areas, Report No. 21286-ME, Washington CNBV will then determine which federation will issue the D.C., World Bank, 2001. opinion and, if applicable, exercise auxiliary supervision over the institution, provided that the authorization is 3 "Popular savings and credit institutions" or "savings granted. Unaffiliated institutions must apply to a confed- and credit institutions" are used throughout this paper eration to participate in its depositor protection fund, to denote a variety of non-bank financial intermediaries whereupon they would enter into an auxiliary supervi- that serve middle- and low-income segments of the sion agreement with that body's member federations. population, such as Sociedades de Ahorro y Préstamo (SAPs), Cooperativas de Ahorro y Préstamo (CAPs), 5 World Bank, Mexico Rural Finance, Report No. 21286- Uniones de Crédito (UC), Cajas Solidarias (CS) or non- ME. Scaling Up Poverty Reduction 123 Integrating the Poor into the Mainstream Financial System 6 They have sought immunity from the new law under 8 The term Caja de Cajas is commonly used to denote the umbrella of the so-called Ley General de this role. Sociedades Cooperativas. 7 Most notably, through organizations (like COMACREP) created to represent the sector in the dialogue that was initiated to design the new law. Bibliography This case study was prepared by Lisa Taber, based on World Bank. Mexico: Country Assistance Strategy internal World Bank, BANSEFI, and SAGARPA reports, Progress Report. Report No. 22147-ME. Washington and written inputs from Carlos E. Cuevas, Lead D.C: World Bank, 2001. Financial Economist for the Financial Sector Operations World Bank. Mexico Rural Finance: Savings Mobilization and Policy Department of the World Bank; Juan Jose Potential and Deposit Instruments in Marginal Areas. Navarrete Luna from BANSEFI; and Gabriela Zapata, Report No. 21286-ME. Washington D.C: World Bank, Director of the PATMIR program in SAGARPA. 2001. World Bank. Savings and Credit Sector Strengthening World Bank. Mexico: Country Assistance Strategy, and Rural Microfinance Capacity Building Project 1996. Report No. 16056-ME. Washington D.C: World Appraisal Document. Report No. 23868-ME. Bank, 1996. Washington D.C: World Bank, 2002. World Bank. Mexico: Country Assistance Strategy, 2002. Report No. 23849-ME. Washington D.C: World Bank, 2002. 124 Case Studies Chapter 9 The Kazakhstan Small Business Program: Commercial Banks Entering Micro and Small Business Finance BY EVA TERBERGER AND ANJA LEPP EXECUTIVE SUMMARY The Kazakhstan Small Business Programme (KSBP) is the second program to try to "down- scale" commercial banks as a means of delivering micro and small enterprise (MSE) loans in transition countries. Since its beginning in April 1998, KSBP has worked with seven private commercial banks (including the biggest and strongest local banks), and has greatly outperformed expectations. It serves as a model for expanding the outreach of commercial banks to poorer clients. The Kazakhstan Small Business Programme, supported by EBRD, was established at the request of the Kazakh government to succeed a failed credit line for small and medium en- terprise loans. KSBP's principal objectives are "to provide finance to MSEs which currently have insufficient access to the formal sector finance; and to build up the credit capabilities of Kazakhstan's financial sector so that local banks are able to provide MSEs with access to finance on a permanent basis." KSBP received a sovereign, guaranteed EBRD credit line of US $77.6 million as a refinancing facility. Furthermore, considerable funding was also pro- vided by the government of Kazakhstan, EBRD, USAID, and the EU for a technical assis- tance package to support institution building in partner banks. According to its objectives, KSBP was not designed to fight poverty directly, but to pro- mote financial market development that would have long-term, but indirect, impact on poverty reduction by creating sustainable access to formal loan financing for micro and small entrepreneurs. These small businesses are a major and growing source of employment both in developing and transition countries. Measured against its objectives, KSBP can show im- pressive achievements. Outreach. By February 2004, KSBP was established in all urban centers in Kazakhstan. Partner banks opened MSE departments in 135 branches, with an additional 50 loan outlets in 39 different cities, some of them as small as 20,000 inhabitants. The out- standing MSE portfolio has grown to over US $162 million, with more than 35,000 Scaling Up Poverty Reduction 125 Commercial Banks Entering Micro and Small Business Finance loans. The MSE departments employ and outlets in new locations. MSE 535 loan officers, who disburse over lending, so it seems, has become a 4,000 micro and small loans every sustainable business which private com- month. The program is still showing mercial banks will stick with, even high growth rates. when the program staff have gone and Target group. At least 90 percent of KSBP itself has been wound up. KSBP clients had never had access to Spill over. KSBP has positively affected formal bank loans, and 85 percent of the institutional set up of the partner the outstanding loans are micro-loans. banks and their competitors as well. Over 50 percent of the newly disbursed For example, the program's training loans are "express micro loans," a and recruitment methods are copied, newly developed product for very small its incentive-based pay schemes are customers needing uncollateralized adapted for use in other departments, working capital loans that are disbursed and MSE loan officers are frequently almost instantly. promoted within the banks to senior Sustainability. Although the program positions or are hired by competitors. has been moving far down in the mar- ket to serve the smallest traders, and at The KSBP owes its success to sound, the same time growing very fast, arrears professional institution building. As a first have stayed extremely low. (Arrears > step, a limited number of smoothly-func- 30 days stood at 0.25% of portfolio vol- tioning, independent MSE departments ume in February 2004.) Loan-officer were established within each partner bank. efficiency has continuously increased. A Further expansion--into other regions, into newly developed profit-center account- poorer market segments, and into the orga- ing method, currently being tested, is nizational structure of the partner banks-- yielding strong evidence that the MSE was not attempted until the program could business has not only passed the rely on a solid base of successful MSE profitability threshold, but is just as, or lending, and human resource capacity had even more, profitable than the average been strengthened by the first generation of alternative business in the Kazakh local consultants trained by the program. banking sector. The partner banks are KSBP also had a favorable environment responding accordingly. They are using in which to thrive. Kazakhstan's macroeco- considerable amounts of their own nomic situation was stable, and the govern- funds to expand the business--over 40 ment showed a strong political commitment percent of the portfolio is now being to developing the SME sector. It supported financed with the banks' own funds-- KSBP from the beginning, while at the same and are continually training new loan time resisting the temptation to directly officers and opening MSE departments intervene in the program. Furthermore, the 126 Case Studies The Kazakhstan Small Business Program Kazakh financial market had been liberalized, lion via an apex structure, but less than one and was well regulated and supervised. third had been disbursed by 1997. Further- Strong competitive pressure on banks, as more, portfolio quality was rather poor, and the sector consolidated, made even the big one partner bank had completely failed, private Kazakh banks interested in participat- forcing the government to honor its guar- ing in KSBP and developing the new antee for a lost portfolio of US $9.5 million. MSE business. Confronted with this failure, but To replicate KSBP's downscaling success nonetheless still convinced that the highest in other countries, sound institution build- political priority should be the development ing needs to be paired with favorable polit- of the SME sector, the Kazakh government ical and economic conditions. If these occur requested that the unused funds (US $77.5 together, downscaling for-profit commercial million)--and the sovereign guarantee--be banks should be tried first, before any other transferred to a new program modeled on alternative for microfinance, in case the fi- the Russian Small Business Fund. This re- nancial services market just needs a jump quest was approved by the EBRD board in start in order to become a powerful instru- 1997, and KSBP became operational in late ment in the fight against poverty. spring of 1998. Its history as the successor of a failed pro- Implementation Process gram had a positive effect on the KSBP in two respects. First of all, there could be no Background and Project Design doubt about the strong commitment of the Kazakh government to support the SME The history of the Kazakhstan Small Busi- sector and the new program. Secondly, the ness Program, teaches an important lesson: lack of success of the SME line had given Never underestimate the significance of clear hints as to what went wrong. It was program design and institution building. unlikely that lack of demand was the main Quite unusually for a project as innova- cause of failure. Even though there had tive as KSBP, the initial impetus to create been no reliable numbers on the size of the the program did not come from the EBRD, SME sector in the mid-1990s, there could but rather in response to an urgent request be no doubt that the sector was growing from the Kazakh government. It needed a and that it had hardly any access to the for- replacement for an underperforming pro- mal banking sector. Almost certainly the gram (or more precisely, a failure), consist- failure of the first program had been caused ing of an SME credit line provided by by a deficiency on the supply side. Partner EBRD and guaranteed by the Kazakh gov- banks lacked the interest and/or know-how ernment. This SME line, which had been to issue and monitor SME loans. There was approved in 1993, offered partner banks a no mechanism in the old program to over- financing facility of up to US $122.6 mil- come this deficiency--a structural fault in Scaling Up Poverty Reduction 127 Commercial Banks Entering Micro and Small Business Finance design which the new program had to payment capacity, close monitoring, gradu- avoid. The mission of KSBP was not only to ation principle, etc.). The experts became provide funds to onlend, but equally, or acting heads of the new MSE loan depart- even more importantly, it was to transfer ments, which opened their doors in eight know-how and to create incentives to build branches of the five banks, in the two main the capacity of the partner banks to go after Kazakh cities, Almaty and Astana. They MSE business. Accordingly, the Kazakh offered micro and small business loans in government and several donor organiza- tenge (local currency) and in US dollars as tions1 financed a technical cooperation their basic products. package to support institution building. KSBP concentrated on the organizational structure of these new departments, intro- The Start-Up Phase, 1998-99 ducing IT-based management information KSBP went into operation in May 1998. systems (MIS), preparing MSE lending Five partner banks which met the eligibility guidelines, and implementing an incentive- criteria were selected beforehand (namely, based pay scheme for loan officers which a full banking license, approval by the took into account all aspects of their per- National Bank of Kazakhstan, an IAS audit, formance, from disbursement level to port- a program-compatible strategy, bank man- folio quality. agement committed to MSE business, a The aim of the first phase was to build up strategically interesting geographic location, specialized loan departments as stable and and financial stability as defined by bank independent entities within each partner regulation standards). Four of these banks, bank. Once this was achieved, the KSBP Almaty Merchant Bank (AMB), Center extended its activities in two dimensions: Credit Bank (CCB), Kazkommertsbank Horizontally, by building MSE depart- (KKB, the biggest Kazakh bank and one of ments in the new banks joining the the strongest in Eastern Europe and Cen- program and in more large cities. By tral Asia), and Tsesna Bank (TB) were pri- the end of 1999, KSBP had reached six vately owned. The fifth bank, Halyk Savings more cities--Aktubinsk, Karaganda, Bank (HSB) initially was state-owned, but Shymkent, Atyrau, Ust-Kamenogorsk, was privatized in 2001. Two more private and Pavlodar. banks joined the program later: Bank Vertically, within the banks by training Turan Alem (BTA) in November 1998, and staff who worked in lending operations Temir Bank (TB) in September 1999. other than the MSE department; and Responsibility for implementing the outside the banks through efforts to in- first steps lay in the hands of eight experi- fluence the regulatory environment. enced foreign experts.2 They hired new, Such activities included discussions local staff to be loan officers, and trained with local authorities to streamline them in microfinance (assessment of client's collateral registration procedures and 128 Case Studies The Kazakhstan Small Business Program raise the level of understanding of the Regional expansion KSBP's undertakings, as well as inter- From 2000 on, KSBP attempted to capture action with government agencies to re- the urban MSE loan market throughout move legal and administrative obstacles Kazakhstan. MSE departments were set up in to MSE development in general, and banks in all major cities, and some small cities MSE finance in particular. (of 20,000 inhabitants) were covered by sub- Furthermore, KSBP improved its own branches. By December 2002, 23 more cities organizational capacity by requiring regular had been included in the program. The num- internal audits of the program's portfolio in ber of lending outlets had risen to 117, and all partner banks and all regions where the the number of MSE loan officers to 283. program was being implemented. The institution building efforts began to Product expansion and moving show results. The loan portfolio grew steadily, down market and arrears stayed low. Only a few months The two basic products of the MSE after its launch, KSBP passed its first test departments were targeted to the needs of successfully. In April 1999, in the aftermath of established micro and small businesses, but the Russian financial crisis, the Kazakh gov- were not really suitable for market traders ernment stopped supporting the local and other potential clients who needed currency, and the value of the tenge dropped small working capital loans that could be from KZT 85 to KZT 120 to the dollar. Yet disbursed quickly. To be able to offer reli- this did not undermine the quality of the able services to this customer group, KSBP MSE portfolio, and the portfolio at risk developed a new product, the "express mi- (PAR > 30 days) remained under 4 percent. cro loan." In contrast to micro and small By the end of 1999, the first generation loans, express loans can be secured with of local consultants had joined the project moveable assets as (unregistered) collateral. team, taking over training and setting up Disbursement is extremely fast (within 24 lending operations in new locations. The hours), and all client contact is handled by KSBP was now ready to tackle the second special MSE outlets located at the markets. phase of implementation. Expansion within the partner banks The Expansion Phase, 2000­02 As the MSE business became increasingly Being able to rely on a core of smoothly important in every partner bank, the banks' functioning MSE departments and having managements realized that it was necessary tapped local sources to create new consult- to firmly integrate these departments into ing capacity, KSBP now moved to expand its the banks' organizational structure. MSE operations without danger of destabilization. management positions were created for Expansion took place in several dimensions. experienced MSE staff; and a head office Scaling Up Poverty Reduction 129 Commercial Banks Entering Micro and Small Business Finance department, responsible for coordinating Institutionalization, Preparing for the and monitoring the MSE business, was es- Hand-Over, and New Challenges, 2003­04 tablished, step by step, in all partner banks. To date, KSBP's lending operations have spread to every city in Kazakhstan with a Expansion of training and monitoring. population over 100,000. The lower end As training new loan officers became rou- of the market has been penetrated to a tine, and standardly contracted to local degree which hardly seemed possible for consultants, KSBP concentrated on further commercial banks. In February 2004, developing its training program. Seminars express loans accounted for 50 percent of for experienced loan officers and back-office all loans disbursed. Partner banks finance personnel, as well as advanced training around 40 percent of the MSE business courses for local consultants, were intro- with their own resources. MSE products duced. Furthermore, first steps were taken are marketed regularly. MSE managers in to develop performance benchmarks which the banks are taking increasing responsibil- would allow banks to monitor the efficiency ity for planning and implementing new of their MSE operations. lending operations, as well as for budget By the end of 2002, another milestone control and human resource development. had been reached: MSE lending was no Full responsibility for all basic training longer a new business, but was firmly estab- activities will be handed over to the part- lished in the market and within program ner banks in the near future. bank. Banks were training loan officers and Accordingly, KSBP can now concentrate opening new outlets on their own initiative; on product development, advanced training they even began using their own funds to (management, personnel development, con- expand the business. The skills acquired by flict resolution, etc.), and audits. It also will the MSE personnel were highly appreci- streamline the partner banks' institutional ated. The partner banks--and their procedures to prepare them to gradually competitors--were now eager to hire and assume more and more responsibility. promote KSBP-trained staff to fill senior Handing over audit functions to the partner positions in lending departments, in risk banks is currently being discussed, and management, and even in the management the banks have started to adapt their of the bank itself. They reduced the internal structures accordingly. Furthermore, number of foreign experts working for KSBP is actively developing new reporting the KSBP, and those who stayed on concen- and monitoring instruments and further trated on the remaining program task-- standardizing the express loans. assuring the sustainability the MSE business Last but not least, as urban MSE lending and preparing to hand over more and more via commercial banks no longer poses a responsibility. "real" challenge in Kazakhstan, KSBP is looking for new challenges. The first pilot 130 Case Studies The Kazakhstan Small Business Program projects to test the new "agricultural loan" 1999, the number of branches doubled, the product have already been set in motion. outstanding loan portfolio grew almost 100 percent in volume and almost 120 percent Impact Analysis in number of loans. Since then, the number of branches has grown steadily but at a The Impact of Downscaling in Financial slower pace, picking up new momentum Market Development when smaller outlets were added and when Establishing MSE lending in (private) com- the express loan was introduced. mercial banks, so-called "downscaling" in The number of outstanding loans more microfinance, offers clients (especially or less doubled every year of the program's micro and small entrepreneurs) not served existence, rising to roughly 35,000 MSE by the formal financial sector, sustainable loans outstanding in February 2004. Port- access to loan financing. Impact assessments folio volume grew at a rate of 170 percent which concentrate solely on whether loans in 2000. It slowed the following two years, actually reach the targeted group of the to approximately 85 percent in 2001 and 23 poor, and whether their incomes and living percent in 2002, only to explode again in conditions improve, would be misleading 2003, with a growth rate of more than 100 for a downscaling project because its effect percent. As of February 2004, the outstand- on poverty reduction is indirect. The suc- ing portfolio came to US $162 million. cess or failure of a downscaling program Growth of KSBP's portfolio was not does not hinge on direct poverty reduction. bought at the price of moving up-market Rather, success is determined by whether and abandoning the original target group of the program can build and maintain sustain- micro and small business. On the contrary, able institutional structures, after donors' average loan amounts (which may serve as support and monitoring are gone. Private an indicator for target group orientation) commercial banks can only be expected declined considerably during the first two to continue lending to micro and small years of the program. After the introduction businesses if it stays profitable. of the express loan, the average loan amounts stayed relatively stable at about US KSBP's Outreach and Client Structure $5,000 per disbursed loan, US $1,800 for The dynamics of KSBP have been remark- outstanding micro loans, and US $20,000 able from the beginning. During the first for outstanding small business loans. The eight months of its existence (1998), 14 median client had an outstanding loan of MSE departments were set up in five differ- around US $2,000. ent cities, 842 loans totaling almost Data on KSBP loans confirm that banks US $7.7 million were disbursed, and the have overcome their former reluctance to outstanding portfolio consisted of 764 loans cater to micro and small business clients. In with a total volume of US $5.8 million. In February 2004, over 65 percent of all loan Scaling Up Poverty Reduction 131 Commercial Banks Entering Micro and Small Business Finance clients (more than 50 percent of the out- Is Lending to MSEs Sustainable? standing portfolio), were individuals with Straightforward data about the profitability an unregistered business. Women hold 61 of MSE departments is difficult to obtain percent of all loans (42 percent of the out- because their cost and revenue structure is standing portfolio), and 65 percent of mi- embedded in the financial accounting sys- cro-loans (59 percent of the outstanding tem of the whole bank. Only now is a spe- micro portfolio). Working capital loans ac- cial reporting system for MSE departments count for almost 80 percent of all KSBP as profit centers being developed. Looking loans, and 82 percent of micro-loans (65 at the main costs associated with the new percent and 75 percent, respectively, of the MSE departments (i.e., loan defaults, ad- portfolios outstanding). ministrative costs, and cost of funds, as well KSBP's market penetration is remarkably as the revenues generated by MSE lending advanced: it is represented in all Kazakh and preliminary profit calculations) allows cities with over 100,000 inhabitants, in 64 some preliminary conclusions about their percent of the smaller cities with popula- profitability to be drawn. tions of 50,000­100,000, and in 25 percent of cities even smaller. Credit Risk and Loan Loss Provisions. KSBP's performance and outreach to its Like other best practice microfinance proj- target group has surpassed expectations, and ects, KSBP has been extremely successful in it still shows extremely high growth rates. controlling credit risk. Even in 1999, in the There were 535 loan officers in the MSE de- aftermath of the Russian financial crisis, partments in February 2004--newly created portfolio at risk (PAR > 30 days of arrears) jobs in the banking sector. The loans they is- never exceeded 4 percent of portfolio sued have supported or created around volume. During the last 3 years, which were more 200,000 jobs for their clients' busi- characterized by financial stability and a nesses in less than 6 years. Tables 1-4 booming economy, PAR always stayed well under 1 percent. Considering that Table 1: Portfolio Development Outstanding Portfolio Number Volume (US$ million) Month Annual Annual Mirco Small Total Mirco Small Total growth growth Jun 1998 50 4 54 0,424 0,123 0,547 Dec 2002 16,433 978 17, 411 78% 41,778 31,609 73,386 46% Dec 2003 27,724 4,880 32,604 87% 51,507 97,486 148,993 103% Feb 2004 29,814 5,424 35,238 54,907 107,791 162,698 132 Case Studies The Kazakhstan Small Business Program Table 2: Distribution of Loans by Loan Amount (February 2004) Loan Amount in US$ Number Percent Volume Percent < 1000 6,904 19.59 3,039,038 1.87 1001-2000 9,959 28.26 10,191,287 6.26 2001-5000 9,543 27.08 21,941,387 13.49 5001-10000 4,346 12.33 24,495,695 15.06 10001-20000 2,384 6.77 27,404,825 16.84 20001-30000 933 2.65 18,686,632 11.49 30001-50000 652 1.85 20,831,178 12.80 50001-70000 199 0.56 9,370,367 5.76 70001-100000 194 0.55 13,232,331 8.13 > 100000 124 0.35 13,505,302 8.30 Total 35,238 100.00 162,698,041 100.00 Table 3: Loan and Customer Characteristics - February 2004 Micro Small Total Number Volume Number Volume Number % Volume % Working capital 24,469 41,167,524 3,443 64,714,780 27,912 79.21% 105,882,304 65.08% Fixed assets 3,746 10,838,584 1,760 37,905,268 5,506 15.63% 48,743,852 29.96% Others 1,599 2,900,928 221 5,170,955 1,820 5.16% 8,071,883 4.96% Total 29,814 54,907,036 5,424 107,791,003 35,238 100.00% 162,698,039 100.00% Male 10,282 21,869,2342, 407 43,779,604 12,689 36.01% 65,648,838 40.35% Female 19,424 32,588,6102, 202 36,048,176 21,626 61.37% 68,636,786 42.19% Not applicable 108 449,194 815 27,963,222 923 2.62% 28,412,416 17.46% Total 29,814 54,907,038 5,424 107,791,002 35,238 100.00% 162,698,040 100.00% Trade 25,633 45,021,792 3,709 66,138,744 29,342 83.27% 111,160,536 68.32% Production 789 1,985,455 502 12,716,110 1,291 3.66% 14,701,565 9.04% Service 3,362 7,826,454 1,170 27,877,748 4,532 12.86% 35,704,202 21.95% Agriculture 16 31,512 13 169,918 29 0.08% 201,429 0.12% Others 14 41,824 30 888,484 44 0.12% 930,308 0.57% Total 29,814 54,907,036 5,424 107,791,003 35,238 100.00% 162,698,040100.00% Scaling Up Poverty Reduction 133 Commercial Banks Entering Micro and Small Business Finance Table 4 Relative Market Share of KSBP Loans Small Aggregate KSBP MSE KSBP Loan KSBP Loan Business Loan Loan Portfolio Loan Portfolio : Portfolio : Portfolio of All Banks Portfolio Aggregate SME Portfolio of All Banks ( US$ million) (US$ million) Portfolio of All Banks ( US$ million) Dec 2000 1,899 510 28 1.46% 5.42% Dec 2001 3,327 828 50 1.51% 6.05% Dec 2002 4,370 952 73 1.68% 7.71% Dec 2003 6,385 1,343 149 2.33% 11.09% Table 5: Portfolio at Risk and Loan Officer Efficiency Loan Officer Effciency (each value per loan officer in corresponding month No. of Loan Month Officers Volume of No. of Loan Volumes of No. of Loans Loans Granted Loans Granted Outstanding Outstanding Jun 1998 31 1.29 11,520 1.74 17,640 Dec 2002 283 7.28 37,797 61.52 259,315 Dec 2003 527 7.80 40,842 61.87 282,719 Feb 2004 535 7.77 42,299 65.87 304,108 Capital at Risk No. of Loan Month Number Volume Officers Arrears Arrears >30 Arrears >1 Arrears >1 Arrears >30 >30 days day (US$ day (%) day (%) days (%) (%) million) Jun 1998 31 0.00% 0.00% 0.00% 0.00% 0.000 Dec 2002 283 0.46% 0.28% 0.89% 0.64% 0.469 Dec 2003 527 0.32% 0.22% 0.32% 0.26% 0.395 Feb 2004 535 0.57% 0.25% 0.62% 0.25% 0.410 almost 20 percent of all bank loans in Administrative Costs Kazakhstan were doubtful at the end of Lending to MSEs is staff intensive and 2002 (according to the National Bank of personnel costs account for most of the Kazakhstan), the low loan default and write- administrative costs. MSE lending requires off rates for MSE loans are remarkable. thorough credit analysis and close monitor- [Insert Table 5] ing of the portfolio to keep arrears 134 Case Studies The Kazakhstan Small Business Program low. Furthermore, MSE loans are small and branches usually achieve a much better per- have higher administrative costs per formance than new ones, due to greater ef- dollar lent. Therefore, loan officer efficiency ficiency and experience. Compared to the is strategically important in managing average return on assets (RoA) and return profitability of the MSE business. In on equity (RoE) in the Kazakh commercial KSBP, loan officer efficiency has increased banking system (reported by National Bank continually over time; further efficiency of Kazakhstan for 2002, as 1.8 percent and improvements and cost reductions can be 12.8 percent, respectively), there can be no expected once all the loan officers hired by doubt that MSE lending is already a prof- the participating banks have become fully itable business which can compete with experienced lenders. other businesses of the banks. Cost of Funds Spill-Over Effects of the KSBP Funding costs vary considerably among KSBP has undoubtedly had a positive effect KSBP partner banks. While the former on the whole organizational structure of its state-owned savings bank has access to rel- partner banks, as well as on their competi- atively cheap savings as a source of loanable tors and on the financial system in general. funds (incurring 5.5% p.a. interest expenses Credit assessment and monitoring tech- in 2003), other banks had to pay between 9 niques, administrative processes (credit percent and 13.5 percent for their funds, guidelines, auditing procedures), as well as depending on the currency (tenge or US$) incentive-based pay schemes, were originally and depending on which bank acquired introduced into the banks by KSBP and the funds. Drawing on the funding line have been adopted in other departments provided by the EBRD, the interest costs of the banks. KSBP's methods of human were just under 8.5 percent. resource development and standard recruit- ment procedures, introduced for MSE Gross Margin, Net Margin, and staff, are now used by some partner banks Profitability throughout their institutions. The average gross interest rate earned on Training courses have become an integral the MSE portfolio is just over 20 percent, part of human resource management in all giving banks a margin ranging from partner banks. A considerable number of loan 7.5­14.5 percent, depending on cost of officers trained by KSBP have been promoted funds. Deducting 1 percent for loan loss within their banks or hired by competitors. provisions and 2.8 percent for administra- KSBP has successfully been able to dissemi- tive costs, a net margin of 3.7­10 percent nate knowledge and build capacity, thereby goes to bank profits or other overhead ex- helping develop the financial services market. penses. Furthermore, the pilot profit-center KSBP has decisively increased provision accounting method clearly shows that older of loan finance to thousands of micro and Scaling Up Poverty Reduction 135 Commercial Banks Entering Micro and Small Business Finance small entrepreneurs in Kazakhstan who pre- development bank owned by the govern- viously had little or no access to financial ment. Just as important, the government services. What is more important, however, refrained from any directed lending. is that MSE finance has been well estab- Thanks to this policy environment, the lished as a sustainable and profitable busi- Kazakh government not only was commit- ness for Kazakh commercial banks, which ted to supporting the KSBP, but made a will continue without KSBP's support. contractual commitment to abstain from interfering in the program. Driving Factors It was significant, too, that a very efficient banking supervisory authority had been Commitment to Political and Economic established at the National Bank of Kaza- Change khstan. A legal framework in accordance Without doubt, the political and macro- with international supervisory standards was economic situation in Kazakhstan has to enacted as early as 1995, and the NBK put be a critcal factor in KSBP's development pressure on the banks to fulfill its require- and success. Kazakhstan is one of the most ments. Many of the more than 200 banks advanced countries in Central Asia in terms which sprang up in the early 1990s due to of transition and economic development, lax licensing were closed down. In 1997 the mainly because as early as 1993­94 the gov- number of banks was reduced to 101, and by ernment firmly committed to pursuing a 2003, just 34 banks were left. policy of liberalization, privatization, The consolidation of the Kazakh banking and structural reform. In the mid-1990s, sector is continuing, and competition has the reforms began to show measurable been rather fierce in recent years (and effects. As a result, Kazakhstan has had remains so). Fighting for market shares, positive growth rates (except in the after- banks were eager to improve their compet- math of the Russian financial crisis), it has itive position by acquiring knowledge and almost balanced its national budget, and it attracting new business. This was an ideal has successfully brought down its inflation field in which to downscale banks, and it en- rate to under 10 percent. abled KSBP to win the strongest Kazakh Reform of the financial sector had al- commercial banks as committed partners. ready reached an advanced stage when KSBP was launched in 1998. Interest rate Institutional Innovation ceilings had been abandoned, a two-tier The most important institutional innovation banking system had been established in was the downscaling approach itself. Numer- 1993, and the government was pushing ous institutional design features were newly forward with privatization. The last state- introduced into the Kazakh financial sector, owned commercial bank, Halyk Savings such as MSE loan technology based on the Bank, was privatized in 2001, leaving one assessment of cash flow; performance-based 136 Case Studies The Kazakhstan Small Business Program pay schemes; and recruitment and training number of MSE departments within part- methods, to name just a few. ner banks. These departments will be the Because KSBP had no experience in base from which to expand. The second downscaling, some new institutional com- phase should focus on promoting regional ponents had to be newly developed. The expansion by opening new branches, mov- most fundamental was a method for ing down-market by offering smaller loans, smoothly integrating MSE departments and firmly integrating the MSE depart- into the banks' organizational structures. ments into the organizational structures of KSBP also implemented a profit-center ac- the banks. The third phase should prepare counting system, specifically to monitor the the hand-over of responsibility step by step. MSE departments. Last but not least, it ventured into pilot testing and launching Downscaling private commercial banks to new products (like the express micro loan serve MSE loan clients can be successful if the conditions are favorable. and the forthcoming agricultural loan) as The macro-economy is fairly stable and MSE lending moved further down market. the government shows a strong com- Another institutional innovation deserves mitment to SME development. mention. KSBP used the competition between partner banks to spur outreach, and The financial market is sufficiently lib- at the same time to impel exemplary coordi- eralized (most importantly, there are nation, cooperation, and mutual learning via no interest rate ceilings) and prudently the structure provided by its program. The regulated, and financial institutions are Kazakh experience needs to be analyzed privatized. more thoroughly to identify optimal levels of The banking sector is characterized by competition and coordination. But regard- fairly strong competition, forcing com- less, when prime conditions for downscaling mercial banks to look for new business. are present, the KSBP model deserves to seriously considered for replication. The market for MSE loans is not swamped by non-profit institutions Lessons Learned offering the same product on more attractive terms and conditions than Intelligent project design is the key to the for-profit players can afford. success of any downscaling project. Institution building in downscaling projects Ideally, a profit center accounting method should be undertaken in several steps or should be deployed right from the start. phases, which can of course overlap in cer- Monitor profitability, with profit center tain cases. In the initial phase, institution accounting is a must if MSE lending is building should concentrate on selecting to become a sustainable business for the partner banks, training, and creating a small partner banks. Scaling Up Poverty Reduction 137 Commercial Banks Entering Micro and Small Business Finance Downscaling should be tried first. mercial microfinance business, the financial If conditions are favorable, downscaling services market can work towardpoverty re- should be tried first--before other institu- duction. Donors do not need to support tion building projects are tested--because non-profit institutions to provide micro- downscaling private commercial banks is loans. Instead, donors can channel their the touchstone for microfinance as a "win- scarce resources into those projects that win" solution. If subsidized institution fight poverty, where the market--left to its building in for-profit partner banks is the own devices--would fail. only support necessary to kick-start a com- End Notes 1 Specifically, EBRD, USAID, and TACIS. 2 The project was, and still is, managed by Internationale Project Consult (IPC) GmbH, Frankfurt, Germany. Today the program is run by 30 local experts and three foreign experts. Bibliography Baydas, M.M., D.H. Graham, and L. Valenzuela. Republic of Kazakhstan National Bank Supervisory Commercial Banks in Microfinance: New Actors in the Board. Regulations on Prudential Standards for Second- Microfinance World. Economics and Sociology tier Banks. Resolution No. 213. June 30, 2002. Occasional Paper No. 2372. Columbus, Ohio: The Ohio http://www.nationalbank.kz State University, 1997. Schor, G. "Commercial Financial Institutions as Micro- Krahnen, J.P., and R.H. Schmidt. "Development Finance Lending Partners. Some Lessons of the Micro Global as Institution Building: A New Approach to Poverty- Program in Paraguay." IPC Working Paper, Frankfurt, Oriented Banking." Journal of Development Economics Germany, 1997. 50:392-95. Terberger, Eva. "Instituciones de microfinanciación en el Morduch, J. "The Microfinance Promise." Journal of desarrollo de mercados financieros." Revista de la Economic Literature 37 (1999): 1569-1614. CEPAL 81 (2003): 195­211. (Unidos Naciones Comisíon Económica para America Latina y el Caribe). Morduch, J. "The Microfinance Schism." World Development 28, no. 4 (2000): 617-29. ["Microfinance Institutions and the Development of Financial Markets." CEPAL Review 81 (2003): 195-21. National Bank of Kazakhstan. Current Development of (UN Economic Commission on Latin America and the the Banking Sector. 2003. Caribbean) ]. National Bank of Kazakhstan. Annual Report 2002, Wenner, M. D., and S. Campos. "Lessons in 2001, 2000. Microfinance Downscaling: The Case of Banco de la Empresa, S.A." IDB Working Paper, Washington, D.C., 1998. 138 Case Studies Chapter 10 The Case of K-Rep Nairobi, Kenya BY JOHN NYERERE WITH INPUT FROM KIMANTHI MUTUA, WILLIAM F. STEELE, ALEKE DONDO, AND JOHN KASHANGAKI Executive Summary K-Rep was started in 1984 as a USAID project applying an integrated methodology to in- termediate funds for non-governmental organizations (NGOs) that were addressing poverty alleviation in Kenya through microfinance and enterprise assistance. K-Rep has undergone two major transformations, the most recent and most significant during 1997-2001, when a commercial bank (K-Rep Bank, Ltd.), a development agency (K-Rep Development Agency), and a consulting firm (K-Rep Advisory Services [Africa], Ltd.) merged to fulfill the mission of the K-Rep Group: to empower low-income people, promote their participa- tion in the development process, and enhance their quality of life. K-Rep's first transition in the late 1980s was characterized by changing its institutional status from a project to an NGO, and by changing its leadership from international man- agers to indigenous microfinance specialists--thereby establishing a Kenyan NGO. K-Rep's second transition, from an NGO to a commercial microfinance bank, was in the late 1990s, where the main goal continued to be the pursuit of self-sustainability as a basis for long- term scaling up of outreach to the poor. To this end, K-Rep sought sustainable funding sources from the financial markets and financial independence from donors. K-Rep also wanted to influence financial sector policy to be more favourable toward microfinance for low-income people, and to change perceptions of the public, clients, and players in the for- mal financial economy through its acceptance, legitimacy, and recognition as a licensed fi- nancial institution. K-Rep's need to find external investors for funding and expertise posed a risk to preserv- ing its core mission. The K-Rep Board of Directors developed guidelines for selecting in- vestment partners with the right combination of desirable characteristics: strong financial and commercial discipline; sufficient resources for additional future investments; sufficient clout to influence Kenyan authorities and to protect K-Rep from political interference; and willingness to divest some of their equity holdings when K-Rep went public. K-Rep's Scaling Up Poverty Reduction 139 The Case of K-Rep leadership also had to lobby and educate the K-Rep's leadership has remained focused regulators of the industry on the benefits on the poor while pursuing sustainability of an NGO transforming into a regulated by building stakeholders' confidence microfinance commercial bank. that large-scale microfinance business can Despite intensive preparations and focus on the poor and still be profitable. training, the actual changeover to a Not to be ignored is the for-profit consult- commercial bank presented a cultural ing arm of K-Rep, KAS, which has indirectly shock for personnel and, as a result, oper- contributed to scaling up and transferring ating systems. Management had to bring knowledge by providing advice and in new staff to cope with regulatory technical assistance in over 15 countries requirements, introduce special training to across Africa. resolve a culture clash between new Besides the need for a conducive regula- and old staff, and restore morale and tory environment to successfully scale up commitment by creating an employee and transform, an important lesson from stock ownership program. the evolution of K-Rep is that leaders After the expected initial teething prob- have to be ready to continually respond lems, transformation yielded very positive to challenges, to be innovative, and to results. In just four years, the combined modify institutional structures in order to outreach of K-Rep institutions rose to over successfully adapt to changing environ- 90,000 borrowers and savers (including the ments. In an industry that is still considered autonomous financial service associations-- new, emerging, and informal, continuous FSAs--created and managed by K-Rep learning and experimentation is essential Development Agency), from slightly over to respond to needs of clients and 15,000 before transformation. Most of stakeholders in microfinance. Perhaps the these clients are poor, and new products most significant challenge in institutional have further deepened outreach to provide transformation and scaling up is developing active poor with opportunities to improve the human capacity and commitment. their incomes by maintaining and growing K-Rep's success has provided a platform their businesses. Although non-poor for legitimizing microfinance and improv- clients were added in the search for deposit ing awareness among the public about mobilization as a base for scaling up, the the immense growth opportunities in ability to reach the poorest of the poor was the industry in Kenya. As a result of also enhanced. Out of 45,000 active K-Rep's demonstrated success and its borrowers and 62,000 active savers of lobbying efforts, a bill institutionalising K-Rep Bank in December 2003, approxi- a microfinance category among financial mately 6,000 are considered very poor; institutions, licensed by the Central Bank, is and over 70 percent of the 48,000 FSA awaiting parliamentary discussion. shareholders are considered poor. 140 Case Studies Nairobi, Kenya Background intervention strategies to alleviate poverty The issue of transformationfrom an NGO, through the delivery of micro credit and here a donor-funded project, to a licensed other financial services. In 1987, the proj- and regulated commercial bank is critical in ect was incorporated as WEREP Ltd., a the microfinance field. There is immense company limited by guarantee with no share unmet demand from the poor for access capital, thus assuming the status of an to appropriate financial services, and trans- NGO. By 1987, K-Rep started seeking formation holds promise for institutions to other donors to broaden its funding base increase their outreach by tapping into and changed its strategy from being solely a commercial sources of funding. K-Rep is service provider to other NGOs to also de- the only microfinance bank in Kenya veloping its own loan portfolio, assuming and in Africa that has gone through this direct lending responsibilities in 1989. process, and its success can guide other African microfinance institutions (MFIs) Development of a Modified Credit and regulators. Methodology: Juhudi and Chikola World Education, Inc., a private Ameri- After exchange visits with other MFIs in can- voluntary organization launched the Bangladesh and Latin America, K-Rep in- Kenya Rural Enterprise Programme troduced a group-based lending approach (K-Rep) in 1984 as a five-year project among its partner NGOs and launched its funded by the United States Agency for In- own lending program in September 1990. ternational Development (USAID). Its mis- The program, known as Juhudi, was mod- sion was to wholesale services by providing elled after the Grameen Bank group-based grants, training, and technical assistance to lending method and modified to the address the financial, management and Kenyan environment. Juhudi is based on technical needs of NGOs involved in micro the co-guarantees of peer groups of five to and small enterprise development. An seven members (called watanos) within evaluation in 1986 by USAID concluded larger groups of five to six watanos (known that the project had limited development as a kiwas). These groups receive an initial impact, was not cost-effective, and should two months of training on group dynamics be terminated at the end of the five years. and methodology that emphasizes the im- The report prompted K-Rep founders to portance of savings, before receiving loans.1 question the future sustainability of the In 1991, in response to demands and in project. It also raised issues of reliance on a an effort to upscale outreach, K-Rep tar- single donor, and sub-grantees, for results. geted the pre-existing indigenous rotating This crisis--deciding to terminate or savings and credit schemes primarily located continue dependent on one donor--planted in rural areas and created a new product the seed for creating a sustainable institu- known as Chikola. Lending to Chikolas was tion that would focus on long-term initially a highly cost-effective method of in- Scaling Up Poverty Reduction 141 The Case of K-Rep creasing outreach since only one check was began to address long-term issues of issued to each group, which was then financial sustainability, growth, continued responsible for allocation to individual dependence on donors, and appropriateness members. In 1994­95, however, repayment of the NGO institutional form. rates fell to 90 percent due to a lack of K-Rep reorganized its operations into group cohesiveness among the larger two divisions: the financial services Chikolas. As a result, K-Rep changed many division oversaw the Juhudi and Chikola features of the scheme and began disburs- group credit methodologies, and the ing loans to individual members within the non-financial services division handled group, and moving Chikolas toward smaller research and evaluation of K-Rep's groups and weekly repayment in line with programs and the NGOs it supported. The the Juhudi lending approach. research helped develop new products, understand the needs of their customers, Transformation to NGO improve methodologies, and develop In 1993, K-Rep received its NGO certifi- new information for the sector. Although cate of registration, in compliance with the the financial services division was NGO Co-ordination Act of 1990, as Kenya designated as the profit center and provided Rural Enterprise Program (K-Rep), Ltd. In credit services, the non-financial services 1996 and 2000, it became K-Rep Holdings, division gradually engaged in activities Ltd., and K-Rep Group, Ltd, respectively, that were not only strategic but also solidifying its move from a project to an profitable, and helped build outreach institution and self-sufficiency. capacity through consulting services. Then followed a period of rapid expansion K-Rep's transformation from a project to an with larger loans, which unfortunately re- NGO with two divisions yielded dividends sulted in increased delinquency and in- in the breadth and depth of outreach, creased desertion rates because many clients shown in Table 1. were uncomfortable co-guaranteeing large In 1997, K-Rep decided to cut back loans. Neither were credit officers able to ef- on lending for three reasons: 1) the loan fectively manage their portfolios and provide delinquency was increasing and K-Rep the necessary client follow-up in this period feared that continued rapid growth would of rapid growth. In 1994, to turn these hide the deteriorating quality of the loan problems around, all wholesale lending to portfolio; 2) it had become difficult to raise NGOs was stopped, due to their increasing funds for lending; and 3) K-Rep needed to arrears to K-Rep. K-Rep began to focus on consolidate its lending activities in prepara- direct lending activities to low-income com- tion for its transformation to a commercial munities and merged the administration of microfinance bank. It wrote off some old the Juhudi and Chikola lending methodolo- NGO loans, and implemented a "Back to gies. K-Rep's institutional strategic plan Basics" program. 142 Case Studies Nairobi, Kenya Table 1 Outreach and Performance, 1991-98 Outstanding No. of Active Average Loan Total Assets Year Loan Portfolio Loan Clients Size (US$) (US$ million) (US$) 1991 1,253 580,607 463 0.9 1992 2,852 982,991 344 1.2 1993 4,331 1,087,100 251 2.1 1994 5,149 3,514,797 682 5.0 1995 11,137 4,601,441 413 5.0 1996 12,885 4,534,323 351 5.0 1997 10,958 3,622,043 330 5.8 1998 13,150 3,816,639 290 6.3 The Process of Transforming to a Under "Back to Basics," K-Rep retrained Licensed Institution all of its credit officers in its original philoso- K-Rep's vision of transformation to a licensed phy and the fundamentals of microfinance, institution emerged in 1994 from a staff con- and re-emphasized the institution's commit- cept paper on possible transformation, and a ment to the microenterprise sector. Service feasibility study funded by the Ford Founda- delivery changes were also introduced to help tion in 1995. The decision to transform was clients gradually learn the habit of weekly based on mitigating the following challenges: deposits, and collateral requirements were K-Rep's NGO structure prevented it revised.2 The loan portfolio was reduced from attracting funds from investors and to US $3.6 million in 1997, then increased inhibited potential benefits of private gradually to just under US $4 million in ownership. Accessing additional sources 1998. By late 1998, total member savings of capital, particularly from client sav- had risen to over US $1.37 million, and ings (by mobilizing deposits3), would delinquency rates were drastically reduced, permit sustained scaling up of credit to with arrears over one day past due down from the target population. 20.9 percent in 1997 to 9.7 percent. The number of active borrowers rose from 11,000 Cross-subsidizing the non-financial in 1995, to 13,150 in 1998, while 4,149 services from lending operations was members held savings but had no loans. It impeding the growth its lending activ- became clear that increasing savings was both ities. In addition, the energy and focus desirable as a service to clients and necessary required to oversee adequately the to help finance further scaling up and long- micro-lending program was overshad- term self-sufficiency of K-Rep itself. owing the potential for new product Scaling Up Poverty Reduction 143 The Case of K-Rep development and expansion of the commercial bank required creating three non-financial activities of the NGO. legal entities under the umbrella of K-Rep Group Ltd.,5 which is a holding company K-Rep expected that transformation to with the largest equity holding in K-Rep a regulated financial institution would Bank. K-Rep Bank, Ltd., is a private, for- enable it to right the inequity of client profit company, owned jointly by others, savings being onlent to wealthier pursuing microfinance. K-Rep Advisory clients of formal banks. K-Rep client Services (Africa), Ltd., or KAS, is a private, savings were deposited in these banks, for-profit consulting firm, owned 100 but neither K-Rep nor the clients could percent by the K-Rep Group. K-Rep, the access loans from the same banks. NGO, was not created anew, but was Transformation would help ensure in- reincarnated as the K-Rep Development stitutional permanence of K-Rep's mi- Agency, with substantial changes in its in- crocredit program through improved ternal operations. governance and increased profitability. K-Rep Group transferred the financial as- sets, liabilities, and activities of the financial Initially, some board members had reser- services division to K-Rep Bank. The assets, li- vations about K-Rep's transformation to a abilities, and activities of the non-financial serv- commercial bank. They especially feared ices division remained with KDA, which was losing control of K-Rep's mission to new then divided into microfinance research and investors with different views, ideas, and innovations; and microfinance capacity build- missions. This challenge of attracting invest- ing. In 2001, the microfinance capacity-build- ment partners who shared K-Rep's twin ing division was spun off to K-Rep Advisory goals of providing financial services to low- Services, which was incorporated to provide income communities on a commercial basis fee-based microfinance consulting services. while also pursuing social development The new structure was intended to protect objectives was daunting. Once initial reser- K-Rep's original vision of providing both fi- vations were overcome, a nine-member nancial and non-financial services to the poor. advisory team4 was formed to prepare a Each institution provides different services business plan, present it to potential within the microfinance and micro enterprise investors and the Central Bank of Kenya, sector. The institutions within the K-Rep educate Central Bank officials about micro- Group are separate legal identities. Each has finance as a commercial proposition, and its own board of directors and own mission, negotiate the licensing requirements for the vision, core values, and organizational culture. proposed K-Rep Bank. K-Rep Group, Ltd. K-Rep's New Organizational Structure K-Rep Group, as the holding company, has The transformation of K-Rep into a the largest equity participation in K-Rep 144 Case Studies Nairobi, Kenya Bank (28.8 percent), and wholly owns both KDA with the understanding that its subsidiaries, KDA and KAS. The board of services are valuable resources for microen- directors of K-Rep Group are the same terprise development in Africa. members who oversaw K-Rep's transforma- tion from an NGO to commercial bank, K-Rep Advisory Services (Africa), Ltd. with the addition of the managing director Incorporated in January 2001, KAS is an of KAS. This board continues to be the mis- independent profit center with its own sion bearer for the K-Rep Group and en- board of directors and capital structure. As sures that, although the entities are pursu- the advisory and consulting arm of the ing different strategies in the African K-Rep Group, it provides a wide range of microfinance industry, they remain united services that include institution capacity in their mission. building and training, project management, business, and financial and strategic plan- K-Rep Development Agency. ning for MFIs. It also manages a public As the research and development arm, KDA relations program, which hosts visitors from focuses on product research and innova- all over the world and introduces them to tion, and dissemination of results to the K-Rep Group activities and other leading broader microfinance industry. KDA's MFIs in Kenya. All KAS services are strategy is to identify, test, and develop new fee-based. KAS continues the management microfinance products to help the poor and technical support to NGOs that K-Rep organize their financial lives, as well as to was initially formed to provide. facilitate partnerships with institutions that can deliver these products and services to K-Rep Bank, Ltd. poor clients. Some of these innovations K-Rep Bank, Ltd., began operations as a include financial service associations (au- fully-fledged commercial bank in December tonomous user-owned, -managed, and 1999, and to date is the only commercial self-financed savings and credit schemes), bank catering to low-income communities healthcare financing and insurance for the that is licensed and regulated by the Central poor, microloans for smallholder dairy Bank. It is also the first microfinance NGO farms, low-cost housing finance, and rural to transform itself into a regulated banking savings mobilization (rural communities institution in Kenya and, indeed, in Africa. pool savings to access higher returns, for example, through treasury bills). While Issues and Challenges of K-Rep Bank benefits from the research Transformation findings, KDA shares its findings with all K-Rep's transformation took four years, re- interested parties, and helps implement new quiring difficult decisions of institutional product lines within other development transformation and persistence in mobiliz- organizations. Donors continue to support ing the right investment partners who could Scaling Up Poverty Reduction 145 The Case of K-Rep assist in overcoming the Central Bank's The biggest issue was grappling with the concerns with the proposal. Other issues institutional options for K-Rep Bank. Though arose, including the danger of mission the final decision was a commercial bank, drift, the appropriate institutional form for management came to this after considering a K-Rep Bank, the Central Bank's mispercep- finance company (non-banking financial insti- tion of microfinance, concern over meeting tution), a co-operative society, and a building regulatory requirements, and the search society. Staff, board members, and external for investment partners were other transfor- stakeholders, such as customers, government, mation issues. the Central Bank, and other players in the Despite extensive staff training and microfinance sector, were canvassed regarding systems development and simulation, nu- transformation. Internal operations were merous problems were encountered in data assessed to determine if systems were capable migration and staff morale. The new of operating as a commercial financial institu- requirements exacerbated a culture conflict tion and were supported by adequate staff between "old" and "new" staff. The tension capacity and strong leadership. The finance between different shareholders was not company format initially appeared to offer the always creative, which affected the gover- best fit, with a lower capital and liquidity re- nance of K-Rep Bank. The commitment quirement than a bank. However, this option of management and staff to resolving was closed by the Central Bank as it pursued these challenges of transformation was a universal banking policy. The co-operative severely tested over a transition period that society presented a good option for including stretched throughout the first year of K-Rep clients in the ownership structure, but had a Bank's operations. weak regulatory framework. Hence the final preferred option was a commercial bank. Internal Issues The overriding concern of K-Rep's board External Issues regarding transformation was "mission Sceptics and doubters were many, the most drift"--the risk that commercial banking critical being the Central Bank, as the considerations would drive K-Rep Bank up regulators of Kenya's financial industry. They market to serve higher-income clients at doubted the viability of microfinance, given the expense of scaling up their mission of its unconventional lending practices and the serving low-income and poor people. fact that hitherto it was a donor-funded activ- The second was the need for partners that ity. Whether an NGO could own a bank was shared the original vision and objectives. also as an issue, given that NGOs have no real A third concern was the K-Rep's prepara- owners. The Central Bank worried that no tions to comply with rigors of supervision one would be responsible if things went and prudential guidelines of a regulatory wrong. The situation was exacerbated by the authority. fact that five indigenous banks had recently 146 Case Studies Nairobi, Kenya been placed under Central Bank management months after commencement rated K-Rep due to lack of liquidity, and in 1998 the Na- Bank as strong, largely due to its capital tional Bank of Kenya (the fourth largest) base, but also cautioned K-Rep Bank about nearly collapsed. The Central Bank therefore its data migration difficulties, high operat- decided not to license any new banks, and ing costs, slow growth of deposits, and low K-Rep's application was placed on hold. The earnings. Most surprising, the Central Bank reluctance of the Central Bank was finally took issue with non-compliance with the overcome only through lengthy lobbying, regulatory requirement limiting each share- donor support to help the Central Bank to holder to a maximum of 25 percent of total understand how microfinance was regulated shares, even though K-Rep Group had in other countries, investment by international negotiated its 28.8 percent shareholding financial institutions in K-Rep Bank, and the with the licensing department of the requirement that K-Rep meet a minimum Central Bank. The Central Bank also ques- capital investment6 of KSh 500 million. tioned K-Rep's field offices because they Whereas K-Rep had sufficient loans, were not licensed as bank branches and did cash, and other assets to meet the required not meet the minimum requirements in minimum paid-up capital of US $6.3 mil- security and infrastructure. lion, it needed at least three other investors Furthermore, the inspectors advised K-Rep to comply with a 25 percent ownership Bank to book large loans and attract large limit per single investor. K-Rep restricted its deposits to increase earning and reduce costs, subscription to US $1.82 million in equity in contradiction to K-Rep Bank's fundamen- and invested US $2.06 million in income tal objectives. Clearly, more work needed to notes,7 while US $4.48 million was offered be done to help the Central Bank fully under- to other investors. A search for local in- stand microfinance. K-Rep Bank's manage- vestors was not successful, as the few that ment formally applied for a waiver of the 25 were identified sought returns of over 20 percent ownership limit; it redefined field percent. Donor agencies helped search for offices as "marketing offices" rather than foreign investors, which finally bore fruit. formal branches, and secured approval; After many months of lobbying and net- lobbied for a Microfinance Regulation Bill working by the Advisory Team, six investors (approved by the Cabinet of Ministers in were found that satisfied the desired com- 2004); and lobbied for a microfinance unit bination of financial expertise and commit- within the supervision department of the ment to K-Rep Bank's social mission. Central Bank (set up in early 2004). The Challenge of Regulation Data Migration and Harmonizing K-Rep Bank came from an NGO back- Operations ground with no stringent external regulator. Staff members were overwhelmed by The first inspection by the Central Bank six new products (such as deposits) and Scaling Up Poverty Reduction 147 The Case of K-Rep procedures--new savings products doubled policy, assisting management, and directing transaction volumes. The accuracy and the institution. The resulting tensions led to timeliness of data were adversely affected by mistrust between management and the the centralization of the system. Initially, K- board, and diverted energy from productive Rep Bank was licensed to operate only one activities; and confidence dropped. branch, so the head office bank branch took Some changes in the individuals repre- over the accounting and record-keeping senting investors on the board were made. that previously had been spread over 17 The board sought individuals who under- field offices. The distance between many stood microfinance, and brought in a hybrid field offices and the head office slowed data of professionals that represented broad in- transferof the transactions. terests but common objectives. It also ad- Urgent problems like these diverted dressed tensions at the board level by ensur- management's attention away from business ing that other directors were trained and development. Management hired consult- exposed to the dynamics of microfinance. ants to address the problem, train staff, and restructure data migration at extra cost The Challenges in Human Resources which had not been anticipated. While the Difficult decisions were made as to which problems were solved within six months, board members and staff members would the higher costs of operation and reduced move to the new K-Rep Bank or continue profitability below the projections in the with the NGO (K-Rep Development business plan and past performance con- Agency) after transformation. On one hand, cerned the Central Bank and K-Rep Bank's staff who did not move to K-Rep Bank felt board of directors. insecure and were not convinced that NGO would survive. On the other hand, some of The Challenge of Governance those who moved to the bank were not con- The new regulatory environment was more vinced of its viability and were further un- stringent in matters of governance. Board settled by the difficulties encountered with members represented diverse shareholder data migration and new procedures. At the interests, and the individual members board level, only two directors moved to the changed frequently, had limited under- K-Rep Bank board, and the rest felt left out standing of microfinance, and mistook the of an institution that they helped establish. transformation difficulties to mean that mi- During the first three months of bank crofinance was not profitable. They sup- operations, productivity actually decreased. ported the Central Bank's recommenda- K-Rep Bank hired traditional bankers to tions for alternative (non-microfinance) run certain functions (front and back office products to improve profitability. Some operations, clearing functions, regulatory board members became involved in details compliance, and treasury management), of management rather than formulating at higher salaries for specific skills and 148 Case Studies Nairobi, Kenya experience not available among existing and board members of the group, has staff. The commercial bankers were less helped K-Rep remain united with one de- sympathetic to microfinance clients and velopment vision, which is pursued through focused more on conventional banking different objectives and corporate entities. practice and adherence to strict procedural rules and regulations. In contrast, the NGO The Challenge of Mission Drift staff was more sympathetic to their poor To demonstrate its commitment to serving clients and embraced microfinance best the poor, K-Rep Bank located its head practices. The resulting culture conflicts and quarters in one of the largest slums in personality clashes slowed operations and Nairobi, Kawangware. Nonetheless, K-Rep polarized the new bank. The tension, how- Bank encountered potential mission drift, ever, presented an excellent opportunity to as regulators and some board members create a new culture that embraced the best urged the bank to move into more prof- of both values. itable lending activities (especially after the K-Rep Bank brought the staff together Central Bank's first inspection). Some to discuss the issues that affected them. of the new staff from the banking sector The professional bankers went to the tried hard to instill conventional lending field to learn about microfinance, while practices that would exclude poor clients. It former NGO staff members worked in the would have been difficult to overcome banking hall as tellers and clerks to give these challenges if the board composition them a feel of traditional banking. Training and top management did not include key activities were carried out across cadres proponents of microfinance. rather than across disciplines. Any staff The characteristics of K-Rep Bank's (old and new) who could not adjust to product did not change after transforma- the new environment had to be let go tion, even though new loan and savings or was transferred back to the NGO products were introduced. The bank (KDA). It took time for the two groups continues to target poor and low-income to work harmoniously. people, as did the NGO. The average loan To overcome feelings of being left size is one indicator of the bank's orienta- out, staff and board members who did tion toward the poor, which has generally not move to K-Rep Bank were allowed to remained at less than US $500 both before buy equity shares through the Employee and after transformation (with the exception Stock Ownership Program (ESOP). The of 1994 and 2001). The average size of new organization structure (K-Rep Group, first loans is lower at US $263; the higher KDA, and KAS) offered new opportunities average loan size is due to larger repeat for board members who did not move loans. The average customer deposit size is to the K-Rep Bank Board. This, coupled US $247. with a bi-annual get-together of all staff Scaling Up Poverty Reduction 149 The Case of K-Rep Impact of Transformation on Scaling K-Rep Group's outreach has also been UP, Services, and Clients enhanced by KDA's support of developing The primary objective of scaling up, to in- and spreading the FSA model. By late 2003, crease outreach, was quickly achieved. K- 67 FSAs had been established (half in arid Rep's initial 1,253 loan clients grew steadily areas), with some 48,000 shareholders, in 1991 to 11,137 in 1995, but then most of whom use the FSA to save and growth slowed considerably, reaching only withdraw when needed. They have an out- 13,636 by 1999, the year of transformation standing loan portfolio of US $0.5 million. (see Table 2). Other projects supported by KDA, for Teething problems limited growth in the example, in HIV/AIDS, agriculture, first two years after transformation. But healthcare, and low-cost housing, reach an from 2001 to 2003, loan clients doubled additional 9,000 poor people, largely in from 22,659 to 45,379. The loan portfolio rural areas. increased from more than US $4 million in 2000 to US$ 20 million in 2003. Although Impact on Products and Services some of this increase came from non-poor Transformation led to a number of new clients, a survey of the income status of banking products. New loan products con- clients indicates that the number of poor sisted of individual loans, wholesale loans, clients in 2003 is triple the total number of bank overdrafts, consumer loans, health clients before transformation. Furthermore, loans and others. This has permitted K-Rep the ability to take savings has brought in Bank to accommodate more group-based new clients, many of them poor. The clients, as well as new small and medium number of customers demanding savings enterprise (SME) clients and others who deposits has greatly outstripped those were targeted as a source of new deposits. demanding loan products, reaching 62,643 The deposit instruments include passbook in 2003. Table 2 Client Outreach Before and After Transformation 1991-2003 No. of Loan Avg. Avg. No. of Total Year Savings Portfolio Loan Size Deposits Deposit Loans Assests Accounts (US $ 000) (US $) Size As an 1991 1,253 0 581 463 0 0 0.9 Ngo 1999 13,636 0 3,228 237 0 0 6.3 2000 17,139 19,863 4,178 275 3,512 177 12.64 As a 2001 22,659 28,584 9,458 417 5,377 188 15.42 Bank 2002 38,739 46,969 14,535 375 10,779 230 22.00 2003 45,379 62,643 20,389 449 15,481 247 28.58 150 Case Studies Nairobi, Kenya savings (attractive to the poor) and time clients. In 2000, loans and advances deposits (to mobilize savings from SMEs increased by 76 percent. K-Rep Bank also and others). developed a credit product, known as Kati-Kati, for individual entrepreneurs with Loan Products high-potential businesses and credit needs K-Rep Bank continued traditional group- that exceed US $1,250. Overdraft facilities based lending to individual members for a are available for current account holders-- maximum of 24 months, at an interest rate a standard banking product positioned higher than standard commercial bank loans, competitively with other commercial banks with options for repayment frequency and a and aimed at attracting small business ac- graduated loan size. Prior to transformation, counts as a source of funds. After transfor- K-Rep offered only the group-based loans mation, the outstanding loan portfolio grew (the Juhudi and Chikola products had been by nearly 200 percent in 2001, from US merged). The group loan product did not $4.2 million to US $9.03 million, then in- change after transformation, although the creased again by over 120 percent in just six proportion of the group-based loans was months, to US $10.9 million in June 2002, deliberately reduced to diversify risks. despite a poorly performing economy. The Proportionately it has decreased from 81 primary focus remains on the poor: 80 per- percent to 49 percent of the total portfolio; cent of the loan portfolio continues to rep- volume grew from US $3.7 million in 2000 resent loans advanced to micro-enterprises to US $10.1 million in 2003. through the group-based loan products. As a commercial bank, K-Rep has been able to expand its credit products to cater Savings Products to both conventional clients and its core One key objective of transformation was the clients who have outgrown the group based mobilization of deposits. In 1996, clients method. Individual loans are issued to and non-clients within the target population clients within the group structure, but are were surveyed to determine the level of not co-guaranteed by group members as interest in opening savings accounts. The with group-based loans. K-Rep Bank also majority of the respondents (91.7 percent of offers individual loans to clients who have entrepreneurs and 86.4 percent of non- grown their businesses, have substantial sav- entrepreneurs) expressed their willingness to ings, and have a minimum of three years open a savings account. Their interests, in successful repayment experience with K-Rep order of importance, were access to credit, Bank. Previously, such clients would have low minimum balance, proximity, good cus- had to seek loans from commercial financial tomer relations, fast service, and competitive institutions, with little chance of success. interest rates. Prior to its transformation, K-Rep Bank also began traditional collat- K-Rep could only use compulsory savings as eral-based individual lending with new security for its loans. After its transformation, Scaling Up Poverty Reduction 151 The Case of K-Rep K-Rep Bank offers five main kinds of savings works in partnership with a community products: group savings,8 standard savings phone distribution company in structuring accounts (voluntary), Msingi children's the finance, training and marketing of the accounts,9current/checking accounts; and product, and K-Rep Bank provides loans of term/fixed deposit accounts. a minimum of US $1000 to a maximum of During 2000, the first year after transfor- US $1400, repayable in six months. This mation, K-Rep Bank mobilized deposits to- particular initiative began in earnest in Oc- talling US $3.5 million, largely from group tober 2003. As at March 2004, 421 units savings held in other banks. By the end of have been sold, with K-Rep Bank clients 2003, deposits had grown substantially to forming the vast majority of owners. US $15.5 million from a wide spectrum of clients. Despite this growth, K-Rep Bank Interest Rates has a funding gap, which it covers with lines In the 1990s, K-Rep Bank was able to of credit. The majority (96 percent) of offer its group-based loans at an average in- K-Rep Bank's depositors are micro-savers, terest rate of 33 percent per annum--com- whose savings balances average less than US parable to commercial bank lending rates-- $650. The total amount of savings for this despite the relatively high operational costs group accounts for 38 percent of the de- of microfinance, due to its grant-based posits. Depositors with large balances of source of funds. Despite the higher cost of over US $13,000 contributed 39 percent commercial sources of funds after transfor- of the deposits as of the end of 2003. mation, it was able to reduce the interest rate on the same loans to 31 percent in Other Services 2001 and to 30 percent in 2002 (although Together with KAS, K-Rep Bank is devel- this was higher than commercial bank rates, oping additional structured financing serv- which averaged around 14.5 percent in ices. One such initiative helps clients to start 2004). Before 2001, K-Rep did not offer a community telephone business. KAS wholesale or term loans because it did not Table 3 Average Deposit Sizes for K-Rep Bank Clients (as of December 2003) Amount US$ No of Clients % of Total % of Total (US$ 000) < $650 60,267 96% 5,861 38% $650-$1,300 1,605 3% 1,480 10% $1,300-$6,500 634 1% 1,601 10% $6,500-$13,00 58 0% 508 3% > $13,000 79 0% 6,054 39% Total 62,643 100% 15,504 100% 152 Case Studies Nairobi, Kenya fall under its mandate as an NGO. As a focus on the poor provides some lessons re- commercial bank, it has been able to offer garding the factors leading to successful these products, with the average interest scaling up and implementation of institu- rate falling from 25 percent in 2000 to 14 tional change. Transformation that seeks an percent in 2003 (from slightly above to appropriate business model has been under- comparable commercial bank rates).10 taken with the objectives of extending out- Thus, the cost of finance to customers has reach, in both breadth and depth, and been declining (at least in nominal terms) as achieving institutional permanence. Each access to credit has been expanding. major transformation stage presented im- portant challenges of implementation. This Impact on Clients section analyzes lessons from how these K-Rep Bank has been able to maintain its fo- challenges were met and key success factors cus on providing financial services to poor for institutional change, including: institu- clients, especially women, who would not tional commitment, political economy for normally be able to access commercialsources change, innovation and adaptation, a learn- of finance. Although non-poor clients were ing culture, and external catalysts. added in the search for deposit mobilization as a base for scaling up, the overwhelming ma- Commitment and Political Economy jority still use the group-based products for Change designed for low-income clients. In 2002, Transformation posed risks to K-Rep's 11,000 first-time borrowers participated in poverty focus by forcing it to deal with ex- Juhudi and Chikola groups. Approximately ternal investors and the Central Bank, both 5,000 (over 12 percent) customers surveyed with different objectives than K-Rep, and in 2003 were categorized as very poor. to hire staff members who came from a Women, who are largely in trading and small- purely commercial banking culture. A key scale manufacturing, account for 52 precent success factor was management's readiness of K-Rep Bank clients. With KDA, over 70 to recognize problems and treat them as precent of the 48,000 FSA shareholders are challenges to be overcome, and determina- considered poor, 49 percent of them women. tion of its leadership to maintain the insti- Clients have testified not only to the impact tution's original commitment to serving the of K-Rep Bank's financial services on their poor. This is reflected in the "Back to businesses and their lives, but also to the Basics" program of staff training and the positive results of transformation. continued predominance of women among K-Rep Bank's current loan clients (over Implementation and Success Factors 60 percent). External factors helped create a political K-Rep's experience in searching for a model economy for change. The government's of sustainability while retaining a consistent growing focus on developing a strategy for Scaling Up Poverty Reduction 153 The Case of K-Rep poverty reduction helped to raise the recep- critical to the success of transformation. tiveness of officials to the role that MFIs They built a solid working relationship, could play. International partners played a confidence, and trust with the governor and catalytic role, through donor-supported simultaneously educated Central Bank offi- activities such as sponsored trips to success- cers on microfinance. ful MFIs in South America and Asia and Management continued these efforts through the willingness of international after transformation to a bank to help financial institutions and investors to pro- regulators better understand the mechanics vide capital and help educate and persuade of microfinance. They successfully the Central Bank. Public and private responded to issues raised by Central Bank institutions worked together to hold regulators without requesting special ex- seminars to try to develop guidelines that emptions, thus instilling the credibility of would enable other MFIs to join the microfinance. The result has been not only financial sector with less stringent hurdles the establishment of a microfinance unit in than those that K-Rep had to overcome. Central Bank, but also the development of Commitment to poverty focus is further legislation that will make it easier for future demonstrated by the decision to appoint a NGO MFIs to transform and for CBK to seasoned leader from within (rather than an regulate the industry more appropriately. outside banking professional) to undertake Besides contributing to preparation of the the launch and management of K-Rep new bill, other MFIs are also developing an Bank, in order to ensure consistency and institutional framework for self-regulation commitment to the mission. K-Rep Group through which they can ensure that has also demonstrated the commitment of non-genuine players do not bring the indus- its leadership to poverty focus through col- try into disrepute. laborative efforts of its different subsidiaries to develop and introduce new financial Institutional Innovation products for the poor, piloting financial K-Rep's innovativeness has positioned it as service associations in low-income commu- a market leader and created an environment nities, and working with other organizations that has helped other MFIs and encouraged outside as well as inside Kenya to share K- other stakeholders to support the industry. Rep's experience in poverty alleviation. From the beginning, K-Rep established a tradition of experimenting with credit Success with the Central Bank methodologies, such as Juhudi and Chikola, The regulatory environment was a signifi- to adapt international best practices to the cant barrier to transformation efforts. Kenyan context. The board chairman and the managing K-Rep's second transformation was far director identified and targeted the more complex in externalising different governor of the Central Bank as a person functions into separate but inter-related 154 Case Studies Nairobi, Kenya organizations. The holding company orga- traditional banking techniques. Fourth, nizational design not only provided legal they included K-Rep pioneers and existing ownership, it was crafted to create distinct staff in the shareholding of the new venture structures (KDA, KAS, and K-Rep bank) through an employee stock ownership pro- with sufficient synergies to depend on each gram which helped reinforce their commit- other in pursuit of common objectives. The ment. The result was a blended workforce structure of three separate institutions was that enables K-Rep Bank to straddle the an innovative approach to resolving the middle way of serving low-income commu- potentially divergent objectives (profit vs. nities in a profitable, commercialized way. social) by allowing each to specialize within the overall K-Rep Group mission of serving Conclusions and lessons the poor. K-Rep has been modelled as a "learning or- Governance of the new organizational ganization." It learned very quickly that its structure was complicated by having to initial integrated model of packaging non- bring in external investors to meet regula- financial services with finance might not at- tory requirements. An important success tain the objectives of outreach and sustain- factor was the establishment of guidelines ability, and shifted to a minimalist model by board and management for selection of focusing on developing financial products suitable partners. Further efforts were appropriate for its target population of the needed to ensure that the individuals repre- poor. It then experimented with new prod- senting these partners on the board had the ucts, such as the Juhudi credit scheme (an right skills and understanding of microfi- adaptation of Grameen Bank methodol- nance to help the new institution fulfill its ogy), which had greater impact on low-in- dual mission of financial sustainability and come customers by tapping on their group outreach to the poor. strengths and focusing on unity, commit- Institutional change such as K-Rep went ment and trust by members. through can be devastating to staff morale. The results of a study in K-Rep's initial Management's humility and readiness to phase as a service wholesaler indicated that take corrective action were critical with the objectives of the MFIs it assisted were respect to new staff. First, they recognized not in concert with its objectives of outreach the need to recruit seasoned bankers in (both in breadth and depth) and sustainabil- order to quickly meet Central Bank ity. It then embarked on direct financial requirements. Second, they recognized the service provision to the poor to improve on conflict of cultures and brought staff face to outreach, efficiency, and effectiveness. face to discuss it. Third, they recognized Observing that donor interest and the need for special training to educate new funding were waning, K-Rep embarked on staff in microfinance methodologies further transformation. It had learned from and mission, and to train "old" staff in Scaling Up Poverty Reduction 155 The Case of K-Rep experience that microfinance was not bold enough and dynamic enough to let go only viable, but also sustainable if the of projects that are now incubated in the right models were implemented. It sought K-Rep Development Agency. KDA has a to combine the drive for profit as a com- number of projects that it is piloting, with mercial bank with its mission to serve the its biggest and most promising being the poor is its latest venture in learning financial services associations. KAS is incu- and experimentation. bating a community phone project. K-Rep's Learning was institutionalised by the structure is designed in such a way that creation of a research and development when the board feels that the time is right, department, which could test products and financial services associations or the com- adapt them to customer requirements. munity phone project, for example, could Several products adapted from various be institutionalised and hived off to become geographical locations have been incorpo- K-Rep entities in their own right. rated into the K-Rep portfolio of products, e.g., Kati-Kati, Juhudi, Chikola, and FSA. Implications for Replication in This department laid the foundation for Other Settings the eventual emergence of KDA, whose The leadership of institutions seeking to scale mandate largely continues to be experimen- up and transform need first to understand tation, leaving K-Rep Bank to focus on and foresee the implications that scaling-up making profits while the other subsidiaries may have, particularly with respect to the undertake the risks of experimentation and structures needed to maintain their core pass on the benefits. This model has facili- mission and objectives. In Kenya, the Co- tated outreach in terms of learning: the operative Bank is scaling up its microfinance advisory and consulting services provided activities successfully as a result of better by KAS have enabled the K-Rep Group to awareness and a regulatory climate more extend their experience to over 15 countries conducive to microfinance, as championed in Africa. by K-Rep and its partners. K-Rep benefitted A key lesson is the need for an institution from establishing a research and develop- to continually remain dynamic and relevant ment department that would keep it to the environment in which it operates. informed of trends in the industry and exper- K-Rep was initially one entity, but when the iment. The importance of testing new time was right, the Board made the some- methodologies locally and adapting them to what painful decision to spin off the suit the circumstances requires the creation financial services division into what is now of an organization capable of learning both K-Rep Bank. A second decision was made internally and externally. to spin off the microfinance capacity- Government support is critical in estab- building division into what is now K-Rep lishing a regulatory climate that will allow Advisory Services. K-Rep's leadership was sustainable microfinance institutions to 156 Case Studies Nairobi, Kenya service poor clients. Donors should be un- stages support institutions by assuming the derstanding, flexible and accommodative of risks in investments for social reasons, so as the unique situations institutions operate in, to build confidence with other partners. providing opportunities that seek to Finally, leaders must continuously transform empower the leadership and linkages to institutions and develop their human international networks to expose managers resource capacities in response to client to best practices and nurture leadership demands and to the changing environment. capabilities. Donors also could in the initial Table 4 Select K-Rep Bank Performance Indicators Overall Financial Performance K-Rep Bank K-Rep Bank Average for Average for (2003) (2002) African MFIs all MFIs Adjusted Return to Assets (AROA) 2.3% 2.5% 3.7% 0.1% Adjusted Return to Equity (AROE) 6.9% 6.1% 14.1% 2.3% Operational Self Suffciency (OSS) 132.0% 133% 148% 115% Financial Self Suffciency (FSS) 126.0% 127% 133% 104% Productivity Borrowers Per Staff Members 172 199 138 121 Risk And Liqudity Portfolio At Risk >30 Days 6.7% 2.3% 2.1% 2.8% End Notes 1 Savings are collected and loans are disbursed at 3 In Kenya, only financial institutions licensed by the weekly meetings with two members of each watanos Central Bank of Kenya (CBK) can mobilize deposits from eligible for loans in the first month, two more in the sec- the public. The "forced savings" that K-Rep held for its ond month, and the rest in the following month. This clients as a guarantee for loans was not considered loan disbursement pattern may vary depending on other deposit taking from the public as understood by the factors. Each member pays a membership fee and buys CBK; it was considered part of the credit methodology a passbook for $3.70 combined. Each borrower pays and not for onlending. 1.5 percent of the loan amount to cover the application 4 The chairman of the board, managing director, and fee (1 percent) and insurance (0.5 percent). All member another board member were included as three members savings are compiled in a group savings account at a on the advisory team. formal (commercial) bank with a credit officer and two members as co-signatories. Client savings serve as an 5 By law, K-Rep the NGO was not permitted to own all additional guarantee against loan default, but are with- or part of a commercial bank. drawn only as a last resort. 6 The minimum capital requirement at that time was US 2 In the past, K-Rep required collateral security in the $2.5 million, but there was a proposal to increase this to form of savings equivalent to 10 percent of each Chikola US $6.3 million over the next two years. loan and 20 percent of each Juhudi loan. In January 1998, K-Rep changed this loan guarantee amount for 7 A convertible debt redeemable after five years. new clients to 5 percent of the loan amount, incremen- 8 Each individual in a group has their own account with tally increasing to 20 percent for loans of US$ 862 and all the group member accounts electronically linked to higher. each other to cover group guarantees. Scaling Up Poverty Reduction 157 The Case of K-Rep End Notes continued 9 This account encourages parents and guardians to 10 Both inflation and treasury bill rates fluctuated during start a savings account for their children under 18 years the early 2000s, but on an overall declining trend, with of age. A minimum balance of US $6 is required to open 90-day treasury bills falling from 13% in 1999 to 1.6% in the account and the account earns interest at a mini- 2003. mum balance of US $60. 158 Case Studies