Document of The World Bank Report No.: 86803 PROJECT PERFORMANCE ASSESSMENT REPORT BRAZIL NORTHEAST MICRO FINANCE DEVELOPMENT PROJECT (IBRD-45540) April 15, 2014 IEG Public Sector Evaluation Independent Evaluation Group ii Currency Equivalents (annual averages) Currency Unit =Brazilian Real (R$) 2001 US$1.00 $2.31 2002 US$1.00 $3.56 2003 US$1.00 $2.90 2004 US$1.00 $2.69 2005 US$1.00 $2.33 2006 US$1.00 $2.15 2007 US$1.00 $1.76 Abbreviations and Acronyms ABCRED Associao Brasileira de Entidades Operadoras de Microcredito e Microfinanas (Brazilian Microfinance Association) BNB Bank of the Northeast of Brazil (Banco do Nordeste do Brasil) BNDES National Bank for Economic and Social Development CAS Country Assistance Strategy CGAP Consultative Group to Assist the Poor CMN National Monetary Council CRESCER Federal Government's National Microcredit Program ETENA Technical Office for Economic Studies of the Northeast FAT Fundo de Amparo ao Trabalhodor (Government Social Security and Unemployment fund) FGV Fundaao Getulio Vargas (Brazilian Research Institute) GDP Growth Domestic Product ICR Implementation Completion and Results report IEG Independent Evaluation Group IEGPS IEG Public Sector Evaluation NGO Non Government Organization OSCIPs Organizao da Sociedade Civil de Interesse Pblico (Civil Society and Public Interest Organizations) PAD Project Appraisal Document PHRD Policy and Human Resources Development Program PPAR Project Performance Assessment Report ROA Return on Assets ROE Return on Equity SDI Subsidy Dependence Index SEBRAE Servio Brasileiro de Apoio s Micro e Pequenas Empresas (Brazilian Micro and Small Business Support Service) SELIC Special Clearance and Escrow System Fiscal Year Government: January 01 - December 31 Director-General, Independent Evaluation Ms. Caroline Heider Director, IEG Public Sector Evaluation Mr. Emmanuel Jimenez Manager, IEG Public Sector Evaluation Mr. Mark Sundberg Task Manager Ms. Anjali Kumar 111 Contents Principal Ratings................ . ....................... ........ v Key Staff Responsible..............................................v Preface........................................................ vii Summary ....................................................... ix 1. Background and Context...................1.... .................1 Macroeconomic Context .................1.... ..................1 Project Context........................................... ....... 5 2. Objectives, Design, and their Relevance .................7.... ...........7 Objectives ..................................................... 7 Relevance of Objectives .....................8...... ...............8 Design ....................................................... 9 Relevance of Design ................................................. 10 M&E Design ........................................ .......... 11 3. Implementation ......................................... ........ 13 Project Cost and Financing .................. .................. 14 Implementation and Project Management Arrangements. ............... ..... 14 Changes to the Project during Implementation...................... 14 Factors that Affected Implementation .................................. 16 Safeguard Compliance........................ ................ 17 Fiduciary Compliance............................................. 18 4. Achievement of the Objectives ...................................... 18 5. Efficiency .................................................... 25 6. Ratings ...................................................... 25 Outcome..................................................... 25 Risk to Development Outcome......................................26 Bank Performance...............................................26 Quality at Entry..............................................26 Quality of Supervision..........................................27 Borrower Performance...........................................28 Government Performance ........................................... 29 This report was prepared jointly by Manuel Hinds, who visited Brazil for assessing the project in April, 2012, and by Sushma Narain, who undertook its write up and completion. The report was peer reviewed by Stephen F. Rasmussen, and panel reviewed by Robert M. Lacey. Anjali Kumar led the preparation of the overall report. Yezena Z. Yimer provided administrative support iv The Implementing Agency.......................................29 Monitoring and Evaluation .........................................30 7. Lessons......................................................30 References......................................................33 Annex A. Basic Data Sheet..........................................35 Annex B. List of Person Met.........................................39 Annex C. Recent Developments in Microfinance in Brazil....................41 Annex D. Borrower Comments....................................... 43 Tables Table 1. Project Cost (Appraisal & Actuals) by Component (US$m equivalent)......... 13 Table 2. Project Cost (Appraisal & Actual) by Source of Financing (US$m equivalent) 13 Table 3. Key Indicators of Expansion of the CrediAmigo Program ...............20 Table 4. Access to Credit (Percentage of Population: 1997 and 2003)............. 22 Table 5. CrediAmigo: Indicators of Sustainability..........................22 Figures Figure 1. Brazil: Fiscal Deficit, Inflation and Current Account Deficitsv..... .......2 Figure 2. Commodity Prices and Real GDP growth in Brazil and Latin America...... 2 Figure 3. Credit as a Percentage of GDP ..................................3 Figure 4. Exchange Rates, Equity Prices and Interest Rates ........... ........... 4 V Principal Ratings Brazil Northeast Microfinance Development Project (P050776, IBRD- 45540) ICR* ICR Review* PPAR Outcome Satisfactory Satisfactory Moderately Satisfactory Risk to Low or Negligible Negligible to Low Significant Development Outcome Bank Satisfactory Satisfactory Moderately Performance Satisfactory Borrower Satisfactory Satisfactory Moderately Performance Satisfactory * The Implementation Completion Report (ICR) is a self-evaluation by the responsible Bank department. The ICR Review is an intermediate IEGWB product that seeks to independently verify the findings of the ICR. Key Staff Responsible Task Manager/ Division Chief! Proj ect Leader Sector Director Country Director Appraisal Steven Schonberger John Redwood Gobind T. Nankani Completion Susan Sanchez Ethel Sennhauser John Briscoe Vi IEG Mission: Improving World Bank Group development results through excellence in evaluation. About this Report The Independent Evaluation Group assesses the programs and activities of the World Bank for two purposes: first, to ensure the integrity of the Bank's self-evaluation process and to verify that the Bank's work is producing the expected results, and second, to help develop improved directions, policies, and procedures through the dissemination of lessons drawn from experience. As part of this work, IEG annually assesses 20-25 percent of the Bank's lending operations through field work. In selecting operations for assessment, preference is given to those that are innovative, large, or complex; those that are relevant to upcoming studies or country evaluations; those for which Executive Directors or Bank management have requested assessments; and those that are likely to generate important lessons. To prepare a Project Performance Assessment Report (PPAR), IEG staff examine project files and other documents, visit the borrowing country to discuss the operation with the government, and other in-country stakeholders, and interview Bank staff and other donor agency staff both at headquarters and in local offices as appropriate. Each PPAR is subject to internal IEG peer review, Panel review, and management approval. Once cleared internally, the PPAR is commented on by the responsible Bank department. The PPAR is also sent to the borrower for review. IEG incorporates both Bank and borrower comments as appropriate, and the borrowers' comments are attached to the document that is sent to the Bank's Board of Executive Directors. After an assessment report has been sent to the Board, it is disclosed to the public. About the IEG Rating System for Public Sector Evaluations IEG's use of multiple evaluation methods offers both rigor and a necessary level of flexibility to adapt to lending instrument, project design, or sectoral approach. IEG evaluators all apply the same basic method to arrive at their project ratings. Following is the definition and rating scale used for each evaluation criterion (additional information is available on the IEG website: http://worldbank.org/ieg). Outcome: The extent to which the operation's major relevant objectives were achieved, or are expected to be achieved, efficiently. The rating has three dimensions: relevance, efficacy, and efficiency. Relevance includes relevance of objectives and relevance of design. Relevance of objectives is the extent to which the project's objectives are consistent with the country's current development priorities and with current Bank country and sectoral assistance strategies and corporate goals (expressed in Poverty Reduction Strategy Papers, Country Assistance Strategies, Sector Strategy Papers, Operational Policies). Relevance of design is the extent to which the project's design is consistent with the stated objectives. Efficacy is the extent to which the project's objectives were achieved, or are expected to be achieved, taking into account their relative importance. Efficiency is the extent to which the project achieved, or is expected to achieve, a return higher than the opportunity cost of capital and benefits at least cost compared to alternatives. The efficiency dimension generally is not applied to adjustment operations. Possible ratings for Outcome: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Risk to Development Outcome: The risk, at the time of evaluation, that development outcomes (or expected outcomes) will not be maintained (or realized). Possible ratings for Risk to Development Outcome: High, Significant, Moderate, Negligible to Low, Not Evaluable. Bank Performance: The extent to which services provided by the Bank ensured quality at entry of the operation and supported effective implementation through appropriate supervision (including ensuring adequate transition arrangements for regular operation of supported activities after loan/credit closing, toward the achievement of development outcomes. The rating has two dimensions: quality at entry and quality of supervision. Possible ratings for Bank Performance: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Borrower Performance: The extent to which the borrower (including the government and implementing agency or agencies) ensured quality of preparation and implementation, and complied with covenants and agreements, toward the achievement of development outcomes. The rating has two dimensions: government performance and implementing agency(ies) performance. Possible ratings for Borrower Performance: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. vii Preface This PPAR assesses the Brazil Northeast Microfinance Development Project, approved by the Board on May 30, 2000. The project closed on March 3 1st, 2007. A total of US$50 million was committed by the Bank for the project. Total project cost was US$100 million. The rest of the funds were financed by the borrower. At project closure, US$38.5 million had been disbursed. US$10,233 was cancelled and US$1,277 was not disbursed. The project was extended from May 31, 2005 to March 3 1st, 2007 when it closed, 15 months behind schedule This report is based on a field visit to Brazil that took place in April 2012. The field visit included discussions in Brazil with CrediAmigo - a specialized unit for microfinance within the Bank of the Northeast of Brazil, its beneficiaries, Bank staff, government officials, non-governmental organizations, institutions, banks, donors, and private sector managers. It also includes a review of the project's appraisal report, legal documents, Implementation Completion Report and Implementation Completion and Results Report, and other relevant material. The cooperation and assistance of all stakeholders as well as the support of World Bank Country Office in Brasilia is gratefully acknowledged. The PPAR was undertaken for two reasons: first, to reevaluate the project in light of long term evidence six years following closing and second, as part of a larger cross country study on Bank-financed support for financial inclusion, covering Mozambique, Brazil, Mexico, and Ghana. As part of this larger exercise, the document includes some considerations and assessments that normally are not contained in PPARs, mostly, the evaluation of the performance of the project after the end of the Bank's involvement as well as the inclusion of evaluation criteria within a broader, more global perspective, than those normally used to assess specific projects. Following standard IEG procedures, copies of the draft PPAR was sent to relevant Government officials and agencies for their review and comments. Comments received are included in Annex D.  ix Summary This PPAR assesses the Brazil Northeast Microfinance Development Project, approved by the Board on May 30, 2000 and closed on March 3 1st, 2007. The project's objectives were to improve the access of microenterprises throughout the Northeast region to sustainable, formal financial services by supporting the expansion of CrediAmigo- a pilot microfinance program in the Bank of the Northeast of Brazil, a regional development bank with branches throughout nine Northeastern states and the northern areas of Minas Gerais. The objectives were highly relevant to the development of microenterprises in the Northeast of Brazil, which had a high priority in the Federal Government's efforts to eradicate poverty. The intention was to demonstrate the effectiveness of sustainable microfinance and develop microcredit as a market-driven activity, eliminating subsidies and thus encouraging private participation. The project's objectives were highly relevant to the strategic goals of both the Borrower and the World Bank Group in Brazil. The project components were well-designed and there was a logical causal chain between the activities to be funded and the intended outcomes. Outcome of the project is rated moderately satisfactory due to partial achievement of project objectives. Three unanticipated factors affected implementation. First, the new government administration that was inaugurated in January 2003 changed the orientation of public policy towards the financial sector, from moving toward broad based liberalization to increased state presence in some areas of the financial system. Second, CrediAmigo, which was created within a large public sector development bank, the Bank of the Northeast of Brazil, was supposed to be spun off once its operations had stabilized. The spin-off did not occur. Third, the project included a study to evaluate the impact of access to CrediAmigo's services on microenterprise development, income growth and client household welfare. This study did not take place within the scope of the project, although later efforts were made to evaluate the impact of CrediAmigo by external researchers, and by the Bank. CrediAmigo became an important provider of microfinance to the region, based on sound lending practices. However due to the significant government policy changes, CrediAmigo, which was intended originally to lead a move toward market oriented allocation of resources for microfinance, did not move in this direction. It benefitted from the government programs that earmarked funds for microfinance. Bank and Borrower Performance are rated moderately satisfactory. On the one hand, the project achieved an increase, through CrediAmigo, in the outreach of microfinance to north east Brazil, and with a positive impact on beneficiaries. On the other, the provision of these services remained dominated by the public sector, benefitting from changes in sector policy. The cautious approach adopted by the Bank team during the mid-term review in March 2003, and reluctance to undertake formal restructuring reflected tension between the objectives of expanded outreach and overall sustainable growth of microfinance. Lessons can be drawn from the early pilot of CrediAmigo and its implementation that are useful for other microfinance institutions, the Bank, and other donors for the design of microfinance operations. Lessons can be described in two areas: those applicable to all projects, and those applicable to microfinance. x Lessons Applicable to All Projects * Flexibility in supervision to adapt to changing scenarios is essential for success, both when the changes are marginal and when they are fundamental to the point of rendering the attainment of the operation's objectives impossible. In the former case, some details may have to be sacrificed for the sake of achieving the ultimate objectives of the operation. In the latter case, this may require courageous decisions from the project team to the point of significantly restructuring the operation, or revising the project development objectives. * The Bank can help in building lasting reforms by limiting the scope of projects to what can be accomplished within a given political and economic environment, even if that environment is not fully market-oriented. Nevertheless it must be cognizant of the fact that the environment is fragile and may depart further from market based principles. An expansion of outreach, per se, even to a desired clientele, is not necessarily an indicator of sustainability of a program. * Donors such as the Bank can add value by setting benchmarks consistent with international best practice. Lessons Applicable to Microfinance Projects * In principle, a viable and broadly independent microfinance institution can be set up within an existing public sector bank and there may be benefits to doing so. However, this leaves it open to influence by the prevailing political climate. In an environment where segments of public financial intermediation benefit from state subsidies, it may be difficult to isolate an 'independent' microfinance program. Even if ring fenced it remains vulnerable to future recapture by the parent institution. Tensions persist about the extent to which public or private sector principles apply. * There can also be tensions between the achievement of expanded microfinance outreach and the achievement of financial sustainability. It may be tempting in some political environments to expand outreach at the expense of sustainability. Yet expansion of outreach to a desired clientele is not necessarily an indicator of sustainability. * Transparency can be a powerful aid to sustainability. In principle, subsidies can be explicitly taken into account through a Subsidy Dependence Index provided it is well defined. However, it is difficult to incorporate measures of subsidized access to funds. * While institutional aspects are crucial for the delivery of financial services to microenterprises, national economic policies are equally crucial. The operations of well managed institutions can reflect financial unsustainability from the point of view of the economy as a whole if the enabling environment reflects distortions in financial markets. Caroline Heider Director-General Evaluation 1 1. Background and Context 1.1 In 2000, the year when the loan was approved, Brazil had just emerged from a period of hyperinflation and was still under the grip of a serious recession that started in the late 1990s.1 The average GDP growth rate in 1998-2000 was 1.5 percent in real terms and the poverty rate of 17.4 percent was well above the norm for middle-income countries. According to the World Bank Poverty Assessment (1995) as per capita GDP stopped growing and inflation increased in the early 1980s, the poverty headcount index increased by 40 percent and stagnated thereafter. 1.2 There were wide disparities in the incidence of poverty within the country. The Northeast accounted for only about 30 percent of Brazil's population (45 million) but more than half of the country's poor. While the country's overall poverty rate was 17.4 percent of the population, the poverty rate of the Northeastern states ranged between 19.4 percent (Sergipe) and 50.9 percent (Piaui).2 Further, the poverty rate estimates for Northeast Brazil in 2002 ranged from 34 to 38 percent in urban areas and from 55 to 57 percent in rural areas.3 Macroeconomic Context 1.3 Macroeconomic conditions greatly improved during the life of the project. By the time the loan was closed on March 31, 2007, GDP was growing at 6.1 percent in real terms while the fiscal deficit over the project period averaged 2.5 percent of GDP, a little lower than in the previous seven years (see the left-hand panel of Figure 1). Inflation had fallen from 7.0 percent in 2000 to 3.6 percent in 2007 and the current account balance went from a deficit of 4.0 percent of GDP to a small surplus (see the right-hand panel of Figure 1) during the same period. The economy was booming with stable prices by Latin American standards. The incidence of poverty had fallen not just in the Northeast but also in Brazil as a whole. People living with less than US$2 per day halved from 22 to 10 percent of the population. 1 During the 1980s and early 1990s, median annual inflation in Brazil was 500 percent. This was primarily a result of persistently weak macroeconomic fundamentals. However, with the implementation of the Real Plan in 1994, inflation decelerated markedly, averaging about 9 percent per year since then (IMF, 2005). 2 The report estimated a poverty line based on the cost of a food basket meeting recommended caloric requirements and took into account cost of living differences within Brazil. The resulting poverty line (in September 1990 prices) ranged from $200/yr in rural areas, to $450/yr in the cities of Sao Paulo and Rio de Janeiro. The latter is equivalent to 18 percent of GDP per capita Skoufias, Emmanuel. 2007. "Brazil: Measuring Poverty Using Household Consumption." Report 36358-BR, Poverty Reduction and Economic Management Sector Unit, Latin America and the Caribbean Region, World Bank, Washington, D.C. An updated version of this paper, using a refined methodology, has since been published in an academic journal. 2 Figure 1. Brazil: Fiscal Deficit, Inflation and Current Account Deficits Brazil: Central Government cash surplus/deficit Inflation and Current Account Deficit % GDP 15% 5% r-~ 00 0 C 8C14 M ln 10 r- 00 0 13% 4% 0% 11% 3%5 -1% -1% 9% /'2% -2% / .7% 1 -3%5 -4% -5% 3% -1% -6% 1% -2% -7% -% -3% c -8%I Inflation ------- Current Account Deficit % of GDP Source: International Financial Statistics, IMF. 1.4 The improved performance of the Brazilian economy has been widely attributed to the prudent macroeconomic policies in place since the early 1990s. However, the economy also benefited from the commodity boom that started in 2004 and has lasted up to the time of this writing, with just an interruption in 2009. As shown in Figure 2, the fluctuations in GDP in both Brazil and Latin America at large were closely associated with movements in the rate of change of commodity prices. Both the 2000-2003 recessions and the subsequent boom were prompted by changes in those prices. Figure 2. Commodity Prices and Real GDP growth in Brazil and Latin America Brazil: Real GDP and Real Commodity Prices Latin America: Changes in Real Commodity 30% 8% 30% Prices and Real GDP per Capita 6 30%- -;8% S 206- -'06% 20% 10 %~4 -'ell 4 a 10%- -C2% 1 -20% -% 0% 4 --0% 0% -40% -20%% ~ -40%20%- -2- 4% 100- , -20%-2% -40% 10-% C- cc- \ c 000 cc~ C>C4 \c cC -20 - - -2%--- ON ON ON ON Cl C C C Cl C ------- World Bank Index of Non-Oil Commoditt Real commodity prices Prices GDP (constant 2000 US$) GDP per capita (constant 2000 US$) Source: World Development Indicators, World Bank. 1.5 Most of the country's financial indicators improved during the years of loan implementation. Credit grew fast when the commodity price boom began. From January 2005 to October 2011, total credit grew at 16.6 percent per year, and credit to the private sector grew at 22.5 percent per year (on a compound basis). At those rates, the ratio of the banking system's credit to GDP (including both to the private and the public sector) 3 increased substantially, from 72 percent in 2000 to 92 percent in 2008. This ratio was by far the highest among the largest economies in Latin America. Credit to the private sector increased from 32 percent of GDP in 2000 to 48 percent in 2007 (Figure 3). Figure 3. Credit as a Percentage of GDP Banking Credit % of GDP Brazil: Credit by Sector % of GDP 100% 100% 80% 80% 60% 6 0%0 40% ~40% ------------------ 20% t20 0%00% 80% 0 a,0 C> C_ C= _= Or-00a,C 0000000 000-, 01 00000,00 0000 C C - - - - - - - - - C 4 "' C] 4 r C] 4 " C] C 4 " C],C Argentina Brazil - Total credit - --Private sector -----Colombia - -Mexico - ---Public sector Source: International Financial Statistics, IMF. 1.6 The rapid monetary and financial growth of those years was largely attributable to inflows of foreign exchange caused by booming commodity exports. These led to a rapid real and nominal appreciation of the currency (shown in the top left-hand panel in Figure 4), which in turn led to an also rapid increase in non-tradable asset prices. These attracted large capital inflows-particularly because interest rates in the developed countries were very low in those years. These inflows, in turn, led to even higher asset prices (see top right panel in Figure 4). 4 Figure 4. Exchange Rates, Equity Prices and Interest Rates Exchange rates Real equity share prices 180 4.5 2.50 160 4 140 3.52.00 to0 2.0 80 2 t 1.50 60 1.5 o 40 1 20 0.5 0 0 > > ~ C- 0 0 00.50 - - Real Effective Exchange Rate O-~a 0 Z] oo o00o ___Reais per US Dollar C C Brazil: Real Interest Rates Interest rates, dollar equivalent 140% 60% 120% 50% 100% A 40% A80%A 60% A 30% 40% 20% 20% - 0% 10% -20% 0% -40% So : I a F Stais t i I 004 0q Cq 000 " ] 04 CC C q q 2 C C ] 0 C' C]C CC] C 1CC C C 4 C]4 C]CC] -o-Real Lending Rates -Real Deposit Rates - eoi ae,$euvln Real TB3 rates -Lending rates, $ equivalent Source: International Financial Statistics, IMF. 1.7 These financial developments took place in spite of the high lending rates that prevail in Brazil-or maybe as a result of them. As shown in the lower panels of Figure 4, real lending rates in the last few years have been considerably higher than deposit rates or Treasury bill rates; in the order of 30-50 percent. Their dollar equivalents have fluctuated at around 60 percent as a result of the appreciation of the real. These rates, plus the rapid appreciation of assets, elicited enormous inflows of capital, which in turn led to further appreciation of the currency and asset values. Asset and commodity prices fell as a result of the 2009 global crisis. Although both recovered in 2009 and early 2010, by late 2011, commodity prices were faltering and both the real price of equity and dollar equivalent interest rates were also falling. 5 Project Context 1.8 Financial services in the Northeast region. At the time of appraisal, access to formal financial services in Brazil was commensurate with that of other upper middle income countries in Latin America. However, there were substantial regional differences. The Northeast had less access to banks and greater dependence on public banks than the rest of the country. In 2000 there were 12 bank branches per 100,000 people in the South but only 4 per 100,000 in the North and 5 per 100,000 in the Northeast. About 75-86 percent of the formal credit outstanding in the Northeast Region originated from public banks. The disparity in coverage and lack of incentives for increased coverage by private banks was one of the primary justifications for the Government's continued support of publicly-owned banks (which accounted for 46 percent of bank branch infrastructure and 55 percent of financial sector assets in Brazil) and directed credit lines, despite generally high default rates and administrative costs. 1.9 Access of microenterprises to financial services. As in most Latin American countries, microenterprises are quite common among the poorest regions of Brazil. Also as in the rest of Latin America, these involve high rates of self-employment. Almost 40 percent of all poor heads of households in the Northeast were self-employed. Statistics prepared by the Brazilian Micro and Small Business Support Service (SEBRAE) estimated that there were about 9.5 million microenterprises nationwide in 1997, with more than 2 million located in the Northeast Region alone. Most of those microenterprises were composed of self-employed individuals and accounted for about 61 percent of the labor force in the Northeast compared with 50 percent nationally. 1.10 Survey data from SEBRAE and the Technical Office for Economic Studies of the Northeast (ETENE) indicated that, at the time of project preparation, few microenterprises had access to financial services, particularly lending services, from either public or private banks, despite the high interest rates they paid to moneylenders and suppliers, which were several times those offered by directed credit lines; and despite a demonstrated willingness of the population to utilize savings services. While access to savings and deposit services was more common, this occurred mostly in large, urban centers and even in those areas it reached less than a quarter of microentrepreneurs. 1.11 Similarly, a 1997 survey of microenterprises in five different cities in the Northeast, conducted by Banco Do Nordeste (Bank of the Northeast of Brazil) found that only 17 to 42 percent of microenterprises received any type of credit services and that access to bank loans was only around 3.5 percent. Suppliers and moneylenders provided most of the credit to microenterprises. As a result, the average monthly interest rate paid was about 15.2 percent, compared to the usually very low annual interest rates of directed credit lines offered by public banks. 1.12 The PAD attributed the lack of financial services, including microfinance services, where Brazil lagged behind regional comparators, to three main reasons: (i) the long period of high inflation that prevailed before the Real Plan pegged the Brazilian real to the dollar, (ii) restrictions on the ability of NGOs to serve as financial intermediaries and to mobilize 6 external financing, and (iii) a strict usury law that placed unsustainably low ceilings on the interest rates charged by NGOs. 1.13 Government programs to improve services to the poor. During the late 1990s, the government piloted the use of market-based mechanisms to improve the quality and efficiency of its services targeted to the poor. While the introduction of demand-driven programs improved the efficiency and effectiveness of some social investments targeted to the poor, programs to facilitate access to land and credit were less successful, requiring high, ongoing public subsidies while reaching a limited number of beneficiaries. By the time of appraisal the Government was seeking alternative mechanisms for deepening access to financial services, which required limited and declining subsidies to initiate, and which could potentially be provided by the private sector on commercial basis, reducing or eliminating the need for public interventions such as directed credit lines. 1.14 The CrediAmigo pilot program. In November 1997, the Bank of the Northeast of Brazil, a development bank with majority government ownership, launched a microfinance pilot program called CrediAmigo, based on loans secured by solidarity group guarantees- that is, loans to groups in which each person guaranteed his or her own loan and those of the other members of the group. The Bank provided financing for technical assistance for designing and launching this pilot as part of project preparation. It reallocated funds from the Northeast Rural Poverty Alleviation Program, obtained a Japanese PHRD Grant, and involved the Consultative Group to Assist the Poor (CGAP). CrediAmigo grew into a large- scale program that dominated the microfinance market in Brazil during the late 1990s. By December 1999 it had more than 37,000 active clients and an outstanding portfolio of almost US$10 million. Portfolio quality was good, measured as 30-day portfolio-at-risk, at 2.5 percent of the total portfolio. 1.15 The original decision to focus the development of the new system on a public bank, with its inherent risk of political intervention, was based on: (a) the commitment of Banco do Nordeste's management to the development of a sustainable microfinance program; and (b) the potential support provided by its 176 branches throughout the Northeast Region. Banco do Nordeste's financial and institutional support systems and banking know-how provided an important platform for CrediAmigo's initial development. 1.16 Though initially focused on lending services in urban areas and market towns, the program was designed to also introduce savings and other financial services, consistent with the high demand indicated by official surveys and their potential as a source of sustainable funding for the program. Consistent with international experience, savings services would be introduced following thorough legal, regulatory and institutional evaluation, and market surveys would be carried out to ensure savings products were designed for the target clientele. The project would also pilot mechanisms to cost-effectively reach isolated clients through the use of post offices as service windows, mobile branches, smart cards and ATMs. 1.17 By demonstrating the viability of a sustainable microfinance program, based on international best practice, CrediAmigo was expected to have an important spillover effect on the development of microfinance in Brazil. Through CrediAmigo, Banco do Nordeste was the first bank in Brazil to provide unsecured loans based on solidarity group guarantees. In 7 addition, the program emphasized sustainability and was demonstrating the willingness of micro entrepreneurs to borrow and repay at interest rates substantially higher than those charged under directed credit lines. By the end of the project period it was expected that further financing for CrediAmigo would be obtained commercially, at least for its more profitable market segments, while any further Bank support would be limited to technical and financial assistance to extend the program's outreach to increasingly isolated, under-served communities or to pilot financial products for special categories of activities. Once the profitability of CrediAmigo was well established, it was expected that private banks would begin to enter the microfinance market as has occurred in other countries. 1.18 Changes in political economy during project implementation. There were significant changes in the political economy affecting the project, during the project implementation period. On January 1, 2003, Luiz Ignacio Lula da Silva replaced Fernando Henrique Cardoso as President of Brazil. The new administration had ambitious social priorities, and different ideas as to the means towards their realization. The previous administration had been the architect of the newly conservative macroeconomic policies that had stabilized the country in the late 1990s and early 2000s. The previous regime had also supported the introduction of a new institutional context for sound microfinance, as evidenced by a series of measures introduced in 1999. The new regime after 2003 kept the prudent macroeconomic policies in place. However, while the previous regime had moved toward market liberalization, particularly in the financial markets, the new administration moved toward increased government intervention to support social policies. In the financial sector, this was manifest in terms of earmarked funds for microfinance lending, at subsidized interest rates, and niche protection for public sector financial institutions, with the intention of better supporting the implementation of social policy. Under the new policies, the availability for funding to programs such as CrediAmigo increased, through increased access to subsidized funds to the Banco do Nordeste, which however undermined the program goal of financial sustainability for private microfinance as a whole. Going forward, the policy shift had substantial effects on the allocation of financial resources, the sustainability of such allocations and the implementation of the project. 2. Objectives, Design, and their Relevance Objectives 2.1 According to the PAD, the project's objective was "to improve the access of microenterprises throughout the Northeast region to sustainable, formal financial services by supporting the expansion of the program known as "CrediAmigo in Brazil." According to the Loan Agreement, the objective was to: (a) expand the availability of financial services for Microenterprises within the Project Area; and (b) ensure the financial viability of such services. This PPAR uses the PAD definition as it is aligned with the project's objective as it was conceived and implemented. It not only specifically mentions access to both formal and sustainable financial services but also states that this will be accomplished through 8 CrediAmigo, a formal financial institution. The loan agreement does not make a reference to "formal financial services." 4 Relevance of Objectives 2.2 The relevance of objectives is rated High. The project was highly relevant to the country and Bank development objectives at the time of project preparation and continued to be so throughout the project implementation period. At the time of project preparation, the government believed that increasing credit to microenterprises would have a positive impact on the reduction of poverty; that the financial discipline necessary for sustainability of financial services would have a positive effect on the fiscal adjustment; and that the project would be a factor in renewing growth and in improving the effectiveness of public institutions. In particular, the government was concerned with regional inequalities and imbalances. The project's target area, of the Northeast of Brazil, was among the poorest regions of the country. The CrediAmigo model had already proven to be effective as a Banco do Nordeste pilot project, and the Government wanted to build on that experience to further increase access to financial services by microenterprises. The project objective continues to be relevant for Brazil's economic development, given the important role that microenterprises continue to have in employment generation and poverty reduction particularly in underdeveloped areas of Brazil. Eventually, the project was also expected to provide an example, through the experience of CrediAmigo, to 'crowd in' other private microfinance institutions to provide financial services to microenterprises in a sustainable form with minimum government subsidies. 2.3 The project was in line with several of the 1998 and 2000 Country Assistance Strategy objectives and the government goals at the time. The project supported the government's recognition of the importance of microfinance as part of its strategy to reduce poverty and develop the financial sector (Brasil em Acqjo, Ministry of Planning). CrediAmigo, Banco do Nordeste's pilot project, was designed using best practices in microfinance with a focus on sustainability while targeting low-income microentrepreneurs. The project's objectives remain relevant to current conditions in Brazil, where the Northeast remains the least developed area of the country and microenterprises play a crucial role in generating employment and reducing poverty. 2.4 The project objectives were well aligned with the goals set out in the June 12, 1997 Country Assistance Strategy (World Bank Report 16582-BR) to reduce rural and urban poverty in the Northeast Region and to increase the sustainability of poverty reduction activities. The relevance of the project objectives to the Bank's assistance strategy is exemplified by the Country Assistance Strategy (27043-BR, December 9, 2003) and Progress Report (36116-BR, June 6, 2006). Building on the Brazilian government's multi-year plan, Plano Brasil de Todos: Participaqdo e Inclusdo (A Brazil Plan for All: Participation and Inclusion), the Country Assistance Strategy objective for the period2003-07 was also to improve the well-being of Brazilians-especially the poorest-by fostering greater equity, sustainability, competitiveness, and macroeconomic stability. The Country Assistance 4 While there is no mention of CrediAmigo in the objective statement of the Loan Agreement, quoted above, Section 1.02 (i) of the loan agreement states that "Microfinance Program" means the Borrower's program, currently known as CrediAmigo." 9 Strategy sets the goal of improved access by the poor to financial services as a key medium- term outcome in the effort toward building a broader and more efficient financial sector, which in turn is a central element for making Brazil more competitive. The Country Assistance Strategy Progress Report of June 6, 2006 singles out increasing CrediAmigo's loan and client base as key indicators of progress in the World Bank Group's microfinance activities and reiterates the importance of microfinance lending to achieving competitiveness results. The Bank's Country Partnership Strategy (CPS) for Brazil 2012-2015 does not specifically focus on microenterprise finance but proposes continued assistance to the government's efforts to further promote financial inclusion and private sector development by providing financing to specialist banks, non-banks, and insurance carriers that focus on MSMEs, low-income communities and underserved peri-urban and frontier regions. The CPS also continues to prioritize addressing regional equalities by supporting development projects in the Northeastern region of Brazil Design 2.5 The project included three components with a total cost of US$100 million. The Bank financed 50 percent of total project cost: 2.6 Component 1. Expanded Microfinance Loan Portfolio (Estimated at Appraisal: US$85.5 million, with Bank financing of US$47 million - Actual: US$ 67.3 million, with a cancellation of US$10,000 of the line of credit). This component supported an increased portfolio of microfinance sub loans. The component was designed to provide funds for CrediAmigo to make loans for working capital and investment to solidarity groups and individual solidarity group clients who "graduated" from solidarity loans by virtue of having a good loan repayment record. Demand was expected to be high given the limited availability of formal loan products designed for the microenterprise sector in the Northeast. Competition from NGO lenders was restricted to a few large cities and these institutions had not pursued aggressive growth strategies. Informal lenders provided loans whose characteristics were equally or more convenient for microenterprise borrowers but at interest rates that were at least twice as high as CrediAmigo. As a result of the component, CrediAmigo's loan portfolio was expected to expand from about US$10 million in December 1999 to US$110 million and its active clients from 37,000 to 150,000, through the project. 2.7 Component 2. Strengthened Capacity for Sustainable Growth of CrediAmigo (At Appraisal: US$13.7 million, with a Bank financed share of US$2.2 million; Actual: US$12.3 million). This component supported a set of activities to strengthen CrediAmigo's capacity to expand its operations in a sustainable manner. The subcomponents were: * Marketing (US$1,130,000). To support marketing studies and help develop new promotional mechanisms for financial services. * Risk Management (US$120,000). To improve portfolio quality as well as variation in arrears by supporting the creation of a Risk Management Unit (with risk model operational and staff training in use of system). * Loan Officer Productivity (US$230,000). To develop mechanisms to enhance the productivity of loan officers in terms of gaining new clients and better managing a larger number of active clients. 10 * Geographic Coverage (US$11,170,000). To support identification of mechanisms to reach clients in low-density areas and the financial and regulatory requirements for implementation of such mechanisms. * New Financial Products (US$330,000). To support development of new lending and financial products to address the needs of microenterprises and the financial, legal, and regulatory requirements for implementation of such products. * Environmental Management (US$110,000). To support an environmental Management Program (with an effective screening process and follow-up client orientation program) that identifies risky microenterprises and mitigates identified risks). * Institutional Independence (US$310,000). To establish the basic components needed to increase the independence of CrediAmigo within Bank of the Northeast of Brazil (legal studies, due diligence reviews, dissemination activities, etc.). * Loan Administration (US$300,000). To support the coordination unit team for loan administration, including annual financial audits, due diligence, maintenance of project reporting system, project midterm review, procurement assistance for bidding documents and contracts, and annual reports. 2.8 Component 3. Impact Evaluation (at Appraisal US$0.3 million; Actual US$ 0.06 million): This component included financing to support a Socioeconomic Impact Evaluation (technical assistance for survey design, field testing, and training, fieldwork, data analysis, and interpretation) to identify the impact of access to CrediAmigo services on microenterprise development, income growth and client household welfare. Such impact evaluation would help the government determine the public welfare aspects of microfinance in relation to other poverty reduction tools. 2.9 The project had two loan amendments to modify the definition for portfolio-at-risk and loan losses. The first one was an amendment to the minutes of negotiations and the second one (July 2004) was to update the definitions of the key indicators according to guidelines from the Central Bank. Other changes were a 15-month extension of the closing date from December 31, 2005 to March 31, 2007. Third, in December 2005, at CrediAmigo's request, the Bank canceled US$10,233,876 from the loan. Of this, the cancellation of US$10,000,000 referred to the line of credit. By that time BNB and thus CrediAmigo had additional and lower cost internal resources through microfinance deposits from commercial banks, and there was limited rationale for the drawdown of Bank funds. The rest was cancelled because the component it would finance, the impact evaluation, was not executed,. Finally, about US$1.2 million remained undisbursed at the time of loan closing. Relevance of Design 2.10 The relevance of project design was Substantial. The project was conceived in the context of Brazilian government's historically proactive government-funded or guaranteed and subsidized public credit and a repressive microfinance regime, at a juncture when it was 5 The Bank project supervision team, together with BNB, made efforts to explore alternative avenues for the use of the Bank funds, through a municipal finance project, but these proposals were not approved by the authorities. 11 prepared to move towards a more market oriented microfinance regime. The CrediAmigo pilot was intended to demonstrate that a microfinance program can provide financial services effectively to target groups and without subsidy. Building on the CrediAmigo pilot, the project was to provide the Government with an alternative to market distortionary credit instruments. 2.11 The proposed causal chain with respect to activities, outputs and intermediate outcomes was clear and convincing. The proposed actions were expected to contribute to achieving the Project Development Objectives of improving "the access of microenterprises throughout the Northeast region to sustainable, formal financial services by supporting the expansion of the program known as CrediAmigo in Brazil." 2.12 The three components as designed were all relevant to the achievement of the Project Development Objectives. The component Expanded Microfinance Loan Portfolio of CrediAmigo was designed to have a geographical reach as well as depth of outreach that were expected to improve the access of microenterprises throughout the Northeast region to formal finance. The Strengthened Capacity for Sustainable Growth of CrediAmigo supported actions to strengthen this financial institution in order to ensure that micro enterprises have access to sustainable finance. The component Impact Evaluation was to support a socioeconomic impact evaluation to identify the impact of access to CrediAmigo services on microenterprise development, income growth and client household welfare. Such an impact evaluation was included to measure the project's impact on the beneficiaries, to help the government determine the public welfare aspects of microfinance in relation to other poverty reduction tools. The impact evaluation was also expected to demonstrate the sustainability of microfinance i.e. the willingness of micro-entrepreneurs to borrow and repay at interest rates substantially higher than those charged under directed credit lines. M&E Design 2.13 The project's results framework was designed to largely measure the project's intermediate and final outcome, in terms of the sustainability of CrediAmigo itself. CrediAmigo's technical unit was in charge of M&E implementation. The institutional indicators were appropriate and measured financial sustainability (return on assets, portfolio quality, risk management, subsidy dependence), efficiency (administrative cost ratio, loan officer productivity, client generation and retention), achievement of development objectives (breadth and depth of outreach) and impact on client business development. The key indicators for monitoring progress toward achievement of the project objectives were: i. Extent of outreach: At least 150,000 additional active clients. ii. Depth of outreach: Average outstanding loan balance of less than 65 percent of Northeast per capita GDP. iii. Breadth of outreach: Number of municipalities in Northeast with CrediAmigo clients. This indicator did not have a quantified target. iv. Financial sustainability: the Subsidy Dependence Index (SDI) must be zero and the adjusted return on assets must be greater than 3 percent. The Subsidy Dependence Index indicates the interest rate that the program would have to charge, after adjustment for inflation and all explicit and implicit subsidies, in order to break-even. 12 A Subsidy Dependence Index at or below zero indicates that the program is subsidy independent, and as such does not depend on public funding to survive. In principle, the SDI estimated at CrediAmigo covered implicit subsidies from concessional liabilities, implicit subsidies related to the opportunity cost of equity, and any explicit donations for operating expenses. Yet reflecting all these elements in the estimated SDI has been difficult. Even at the time of appraisal, it was admitted that some 6 privileges of CrediAmigo were not reflected in the SDI. The Subsidy Dependence Index does not reflect reduced funding costs to the Banco do Nordeste, following new government regulations. The estimation of the SDI was adjusted after it became clear that CrediAmigo would not be spun off from BNB, reflecting a lower benchmark than the prevailing market interest rate. Partly because of this, CrediAmigo appeared to exceed the Subsidy Dependence Index targets. v. Operational efficiency: Operational costs (annualized) less than 24 percent of average net outstanding portfolio. vi. Institutional independence: Spin-off plan approved by Bank of the Northeast of Brazil Board of Directors. Equity to earning assets ratio of at least 15 percent. 2.14 Additionally, the PAD embedded in the project two relevant Country Assistance Strategy indicators: i. A declining poverty gap for poor client households in the Northeast; and, ii. A declining subsidy requirement per beneficiary. 2.15 In addition to demonstrating to the government and the central bank that sound and sustainable microfinance works, it was expected that once the profitability of CrediAmigo would be well established, private banks would begin to enter the microfinance market as has occurred in other countries. The project design however did not include any indicators to track the impact of CrediAmigo's demonstration effect on private banks' entry into the microfinance market. 6 CrediAmigo, as a branch of BNB, was exempted from new equity requirements for opening branches. This constitutes a subsidy not reflected in the SDI. 7 At appraisal, the market interest rate was the benchmark, as it was assumed that CrediAmigo would be funded from increased savings mobilization. However once CrediAmigo was fully incorporated into the structure of BNB, the interbank rate was used as a measure of opportunity cost. Thus the 2003 microfinance policies. had a direct influence on the pricing, profitability and client composition of CrediAmigo, and BNB started using the newly available sources of funding to support the expansion of CrediAmigo 13 Table 1. Project Cost (Appraisal & Actuals) by Component (US$m equivalent) Components Appraisal Actual/latest Percentage of estimate (US$ estimate (US$ appraisalb million) million)' Expanded Microfinance Loan 85.5 67.3 78.7 Portfolio Strengthened Capacity to Support Sustainable Growth of 13.7 12.3 89.5 CrediAmigo Impact Evaluation 0.3 0.1 1.6 Total Baseline Cost 99.5 79.6 80.0 Physical contingencies 0.0 0.0 0.0 Price contingencies 0.0 0.0 0.0 Total Project Costs 99.5 79.6 80.0 Project Preparation Fund 0.0 0.0 0.0 Front-end fee IBRD 0.5 0.5 100.0 Total Financing Required 100.0 80.1 80.0 a US$10.233 million was cancelled from the Bank loan and US$1.277 million was undisbursed. b Percentages calculated from actual amounts before rounding. Source: PAD & ICR. Table 2. Project Cost (Appraisal & Actual) by Source of Financing (US$m equivalent) Type of Appraisal Actual/latest Percentage of Source of Funds cofinancing Estimate estimate appraisal (US$ (US$ million) million) International Bank for Reconstruction and 50.0 38.5 77.0 Development Borrower; Banco do 50.0 42.6 85.2 Nordeste Source: ICR. 3. Implementation 3.1 The loan was approved on May 30, 2000. The project became effective on September 6, 2000. The Mid-term review took place on March 31, 2003 and the project closed on March 31, 2007 14 Project Cost and Financing 3.2 The project's a total cost was US$100 million. The Bank financed 50 percent of total project costs. The rest was financed by the borrower, Banco do Nordeste. The actual cost was US$ 80.1 million, of which IDA financed US$38.5 million, and the borrower financed $42.6 million. US$10.233 million was cancelled from the Bank loan and US$1.277 million was undisbursed (Table land Table 2). Implementation and Project Management Arrangements 3.3 The CrediAmigo Program in Banco do Nordeste would implement the project. The CrediAmigo Technical Unit would have overall responsibility for implementation of project activities, working in conjunction with relevant areas of Banco do Nordeste. Individual branches would be responsible for lending under the Microfinance Loan Portfolio Component. The CrediAmigo Technical Unit would have direct responsibility for implementation of the Institutional Development Component of the project, and ETENE would manage implementation of the Impact Evaluation Component. 3.4 Due to the importance of the expected demonstration and learning aspects of the program, the project included a number of mechanisms to promote review and evaluation by both primary (Banco do Nordeste, the Government, CrediAmigo clients, Bank) as well as secondary (Central Bank, other microfinance institutions, other commercial banks) stakeholders. The program included an annual due diligence review that would be conducted by an outside agency to thoroughly evaluate the program's progress towards commercial viability according to several financial sector criteria. Banco do Nordeste, Government and the Bank would review the due diligence jointly each year. A report would be prepared for general dissemination. 3.5 The program Management Information System would provide a framework for on- going analysis of client benefits at the enterprise level. A mid-term review would be carried out no later than 30 months after loan effectiveness, during which Banco do Nordeste, the Government and the Bank would review, inter alia, progress toward the program's financial sustainability objectives, institutional evolution, results of the impact evaluation, environmental mitigation program, and the effectiveness of dissemination activities. Changes to the Project during Implementation 3.6 Amendments, cancellations and extension of the Closing Date. The project had two loan amendments to modify the definition for portfolio-at-risk and loan losses. The first one was an amendment to the minutes of negotiations and the second one (July 2004) was to update the definitions of the above indicators according to guidelines from the Central Bank. Other changes were the cancellation of US$10,233,876 from the loan reflecting the reduced value added to CrediAmigo from Bank funding, given the availability of increased funding to BNB. There was also a 15-month extension of the closing date from December 31, 2005 to March 31, 2007. About US$1.277 million remained undisbursed at the time of closing. 15 3.7 The Bank extended the closing date of the loan by 15 months, to March 31, 2007, to allow the project to design a new model for providing sustainable financial services to microenterprises through Civil Society and Public Interest Organizations (OSCIPs) and to allow more time to complete the impact evaluation of the program. For reasons explained below this evaluation was not carried out. 3.8 Cancellation of the Evaluation Component. One significant change to the project during implementation was the borrower's decision not to carry out the contractual evaluation that was financed with the third component of the project. In its place, on the exclusive initiative of CrediAmigo, Fundaqao Getulio Vargas (FGV), a leading Brazilian think tank, analyzed the impact of CrediAmigo on financial access and poverty with quite favorable results.8 The Bank, however, did not accept this as a fulfillment of project obligations. For this reason, the component that would finance it was cancelled. 3.9 Differences arose between the Bank and CrediAmigo regarding the method that should be used to carry out the evaluation. The point of contention was the inclusion of a meaningful control group in the evaluation. CrediAmigo, while recognizing that a control group was essential for a meaningful statistical analysis, pointed out that the only way to make sure that the people included in the sample would be comparable with those receiving financing would be to include them in the process of credit evaluation, approve them and then deny them the credit. Only in this way would the control group be seen as identical to the group receiving credit. CrediAmigo considered it unethical to deny credit to people who met all the requisites to receive it. The Bank tried to offer CrediAmigo alternative proposals for control groups, based on potential customers meeting all the CrediAmigo requisites, or on credit scoring. However early baseline surveys were unsuccessful, and CrediAmigo's staff preferred to outsource this work to the Fundaqao Getulio Vargas (FGV), which undertook a form of evaluation based primarily upon existing data; an approach not supported by the Bank There was difficulty in finding common ground. The evaluation, as a part of the loan, was cancelled. Later, the Bank undertook its own impact evaluation of the impact of CrediAmigo (Skoufias et. al; 2009, 2013). 3.10 Introduction of further subsidies. In October 2011 the government launched the CRESCER program. Among its measures there is a further subsidy to microfinance, given to public sector institutions exclusively, to reduce the rates of interest charged for microcredit, provided that they are used to finance solidary loans. This subsidy allowed CrediAmigo to reduce its rate of interest from 1.2 percent per month (15.4 percent annually) to 0.64 per month (8 percent annually). This boosted CrediAmigo's strength in the microcredit market. However, it puts the interest rate well below the costs for private banks and other private financial institutions, and the spread between market rates for non-privileged customers (42 percent). A more detailed discussion of the evolution of microfinance in Brazil subsequent to loan closure is provided in Annex C of the PPAR which discusses the impact of the government's new policies on the development of private microfinance, and the large role played by CrediAmigo and other public sector banks. 8 Marcelo Neri, Microcridito, 0 Mistrio Nordestino E 0 Grameen Brasileiro, Perfil and Performance dos Clientes do CrediAmigo. Rio de Janeiro: Banco do Nordeste, CrediAmigo, FGV and FGV editora, 2008; and Marcelo Neri (2010) in brasilinfocus.com. 16 Factors that Affected Implementation 3.11 Changes in Macro Environment. There were three substantial changes in the macro environment during the life of the project that affected implementation. * First, the economy slowed down in 2002 and the Real was devalued. This reduced the rate of growth of the loan portfolio in U.S. dollars (because CrediAmigo's clients work mainly in the non-tradable sector, showing limited sensitivity to foreign exchange fluctuations). CrediAmigo's management dealt with the impact of this problem on the institution's profitability by tightening expenditure controls and enhancing loan officer productivity. Both of these actions were consistent with the project's development objectives of improving access to sustainable, formal financial services. * Second, as discussed in paragraph 1.18, the new regime after 2003 increased the interventions of the government in the economy, reversing, to some degree, the move towards liberalization that had prevailed during the previous administration. * Third, towards the end of the implementation of the loan the economy experienced a positive turnaround caused by a commodity price boom that started by 2004. By the mid-2000s, the economy was growing at accelerating rates and within a stable macroeconomic environment. * Changes in the regulatory environment. As discussed in section 1, the regulatory environment during the 1990s was not conducive to the development of microcredit.9 Market oriented reforms introduced from 1997 established a category of microfinance that could circumvent some prevailing regulatory restrictions. These reforms were eventually of limited effectiveness, as the resulting organizations, the OSCIPs (Organizaqjo da Sociedade Civil de Interesse Piblico), did not have the time to mature before the policy reversals of the new administration were introduced and remained weak institutions.10 The project was designed to support the early market oriented reforms. The reversals that came with the 2003 change in administration, while also well intended, had their own adverse consequences in some regards, and had a substantial impact on the project. Many of the policy innovations introduced in 2003 significantly contradicted what was expected during the design of the loan, and marked a reversal of earlier government policy efforts towards sound microfinance within the context of liberalized financial markets. The new measures introduced after 2003 included the following: * National Monetary Council Resolution 3109, which forced banks to direct credit equivalent to 2 percent of their sight deposits towards microcredit operations. If not directed for this purpose, an equivalent amount was to be held as unremunerated reserves. Earmarked funds were estimated to amount to 3.0 billion reais (US$1.7 billion) by the end of 2007. Given that lending rates averaged over 51 percent during 9 In particular, the limited authorization to NGOs to mobilize funding, the usury law that restricted nominal interest rates to 12 percent per year, and relatively high minimum capital requirements. 1o In 1997, a working group was set up in the office of the President of Brazil (Community Solidarity program), which defined, through Central Bank regulations passed in 1999, a new group of 'Civil Society Organizations' that could circumvent the funding difficulties described above, Some exemptions from the usury law were also approved (July 1999), and the issue of minimum capital requirements was being explored at project Appraisal. 17 the implementation of the project, this imposed a huge opportunity cost on the banking sector, if it did not lend to microenterprises, estimated at 1,528 million reais (US$863 million) per year, in that year. This was a strong but artificial incentive for banks to lend to microenterprises. Some banks that did not have in-house microfinance lending programs offered their funds to the Banco do Nordeste for deployment in the CrediAmigo program. This lowered the funding costs of the Banco do Nordeste. * The regulations also established interest rates caps on credit based on such funds." CrediAmigo began to use the newly earmarked funds, which resulted in average interest rates well below those prevailing in the unregulated market (26.8 percent annual rate against 40-45 percent). While free market rates did not fall perceptibly in subsequent years, CrediAmigo reduced its interest rates further by using cheaper official funds. By 2011, its lending rate was 1.64 percent per month. And in 2012 it was further reduced to 0.6 percent per month, reflecting a mix of products at market rates as well as CRESCER rates.12 These measures fundamentally changed the project setting. It was no longer possible to believe that the operation of microcredit would be transferred to the free market, a basic assumption in the design of the project. CrediAmigo's loan yields fell from 2004 to 2011 though they were partially offset, especially from 2011, by high loan origination fees of 20 percent. 3.12 Nevertheless, not all the reforms introduced after 2003 contradicted those implemented before that date. Brazil achieved strides in other areas, especially formal financial access, for example through the introduction of basic bank accounts (conta simplificada) correspondent banking, and improved credit information services.13 Safeguard Compliance 3.13 The Project was classified as Category B for the purposes of environmental assessment. The project design included an environmental mitigation program for at-risk microenterprise clients and training of CrediAmigo's business managers and environmental officers in the branches. This was the result of a cautious approach towards expanding CrediAmigo's clientele in recognition of the possibility that some clients may pose potential environmental impacts. However, more than 97 percent of CrediAmigo clients are from sectors without significant environmental impacts, such as commerce. Therefore, the subcomponent's scope was refocused to provide entrepreneurs with environmental guidance on business practices. A firm was contracted to develop an educational plan that emphasized such environmental guidance for entrepreneurs, with 15 seminars expected to take place in 2007. Furthermore, over 100,000 pamphlets on environmental issues were distributed during 1 These caps were set at 26.8 percent per year with loans originating fees of 2 to 4 percent; loan amounts would be limited to R$600 for individuals and R$1,000 for small business owners; and loan maturity may not be less than 120 days (ICR Footnote 6). 12 At the same time, the measures also introduced the creation of simplified deposit accounts (a basic package of bank services, free of charge up to a certain transaction limit) and simplified conditions for opening a bank account. The cost of providing these services were charged to other services, which favored access on the one hand, while deepening the distortions that increase the costs of intermediation, which in turn become an obstacle to access, on the other hand. 13 As noted in World Bank (2013) Financial Inclusion in Brazil - Building on Success. Financial Sector Assessment Program Update - Brazil. Washington DC, May. 18 field visits by CrediAmigo loan officers. IEG found that the environmental impact proved to be a non-issue in CrediAmigo operations, and that as a result, they proposed to refocus the subcomponent as discussed above. This was done. The seminars were conducted and the pamphlets were distributed. 3.14 Activities financed were very small-scale, focused on service activities without any impact on the environment, and all were on-going. Based on international experience with microfinance, this profile is not expected to change significantly as the program expands. As some on-going activities which would be supported could have environmental impacts (tanneries, mechanical shops) which could be fairly easily mitigated, these activities were identified in advance and were visited by the training agent to prepare a simple mitigation plan with the micro-entrepreneur in the context of improved, overall business management. Businesses which failed to implement simple mitigation activities would not be eligible for follow-up loans. No such cases were reported. Fiduciary Compliance 3.15 With regard to fiduciary compliance, all financial management aspects related to implementation were rated as Satisfactory, with internal controls representing best practice within the Brazil portfolio. The PAD did not include a financial management supervision plan supported by audit report reviews and follow-ups. However, the financial management team identified a misunderstanding in the consolidated/accumulated PMRs. As a result, the canceled loan amounts, and a reallocation requested on December 28, 2005, were not reflected. However, CrediAmigo updated the PMRs accordingly. IEG found that while this mistake was a manifestation of negligence, it did not have any material consequence. The legal validity of these changes was enforced with other documents 4. Achievement of the Objectives 4.1 The project's objective was "to improve the access of microenterprises throughout the Northeast region to sustainable, formal financial services by supporting the expansion of the program known as "CrediAmigo in Brazil." The overall objective is examined from the perspective of two sub- objectives: * Improving the access of microenterprises throughout the Northeast Region to formal financial services; and * Improving the access of microenterprises throughout the Northeast Region to sustainable financial services. * Improve the access of microenterprises throughout the Northeast Region to formal financial services. Outputs 4.2 Output targets to measure expansion of access to formal finance were all achieved. By the time of project closure, based on the ICR of September 2007, CrediAmigo expanded its geographical operations to 170 branches and 23 sub-branches, with clients in about 1,420 19 municipalities (about 83 percent of all municipalities in the Northeast). CrediAmigo's new group loan product, 'CrediAmigo Comunidade', expanded access in low density areas, targeting low-income borrowers. CrediAmigo Comunidade had about 11,000 clients with a loan portfolio of R$2.5 million. Overall, CrediAmigo's loan portfolio expanded to R$235 million, and its active clients to 240,000, far exceeding the targets. Six years later, the Bank's FSAP mission report of 2013 reports further expansion, with more than a million active clients at CrediAmigo and another 650,000 in a parallel program it established, AgroAmigo.14 4.3 The achievement of the project's objectives were to be measured, according to the Project Appraisal Document, using indicators for extent, depth, and breadth of CrediAmigo's outreach (Table 3). i. Extent of outreach. CrediAmigo had almost 236,000 active clients as of December 2006, up from about 58,000 in 2000 (and 266 in December 1997). The initial goal of 150,000 by the end of 2005 was surpassed. ii. Depth of outreach. The ratio of average loan size to GDP per capita in Northeast Brazil reached a high of 27 percent in 1999 and has since fallen incrementally to 19 percent in 2006. The PAD projected a ratio of 35 percent in 2000 that would grow to 65 percent by 2002 then level off, but the actual figures were substantially lower. The primary reason has been the shift in focus by CrediAmigo toward lending to the lower income population of the Northeast, with particular emphasis on reaching those without access to bank services. This reorientation in CrediAmigo's client profile resulted in a smaller average loan size (US$350 in 2006) than originally projected in the PAD (US$518 in 2004). Indicative of the client base reached, about one-third of CrediAmigo's clients are beneficiaries of Bolsa Familia, a conditional cash transfer program targeting the poor. iii. Breadth of outreach. The number of municipalities covered by CrediAmigo grew from 358 in 2000 to 1,420 in 2006, increasing coverage from 21 percent to 83 percent of municipalities in the Northeast. This dramatically exceeded the projections in the PAD, which anticipated reaching only 150 municipalities by 2004. Based on a market estimate using data from the Brazilian Geography and Statistics Institute's (IBGE) 2000 National Survey ofHouseholds (PNAD), CrediAmigo serves about 12 percent of its estimated potential market (2 million microenterprises). 14 World Bank 2013. Statistics at end 2013 for BNB echo these findings, citing 1.6 million customers, with a portfolio balance of R$ 2 billion in microcredit loans, as of November 30, 2013, with 13,170 credit transactions per day and an average interest rate of 5% per annum, under the Crescer program. See Agencia Brasil (2013): Microcridito 4 recorde no pals, mas atende apenas parte dos pequenos empreendedores. 0 Globo 2013 Dec 22. 20 Table 3. Key Indicators of Expansion of the CrediAmigo Program Target 1999 2000 2003 2004 2005 2006 Indicator 1. Extent of Outrach: 150000 Outreach: additional 37,976 57,967 138,497 162,868 195,378 235,729 Number of active clients Indicator 2: Breath of Outreach: number of municipalitie No target n.a. 358 1172 1165 1200 1420 s in NE Brazil with CrediAmigo Clients: Intermediat e Indicator 1: Depth of outreach: No target 17,613,634 32,816,435 91,516,126 111,541,109 141,317,421 177,621,238 Outstanding Loan Portfolio Reais Intermediat e Indicator 2: Average loan size 65% 27% 25% 20% 18% 19% 19% over NE Brazil per capita GNP Source: CrediAmigo. 4.4 One key output, however - the impact evaluation- was cancelled. Differences arose between the Bank and CrediAmigo regarding the method that should be used to carry out the evaluation. The point of contention was the inclusion of a meaningful control group in the evaluation.. CrediAmigo, while recognizing that a control group was essential for a meaningful statistical analysis, argued that the only way to make sure that the sample would be comparable with those receiving financing would be to include them in the process of credit evaluation, approve them and then deny them the credit. This CrediAmigo considered unethical.15 The impact evaluation was included to demonstrate to the government the public welfare aspects of sustainable microfinance in relation to other poverty reduction tools and its cancellation implied a loss of this potential benefit. The Bank tried to offer CrediAmigo alternative proposals for control groups, based on potential customers meeting all the CrediAmigo requisites, or on credit scoring. However early baseline surveys were unsuccessful, and CrediAmigo's staff preferred to outsource this work to the Funda(ao Getulio Vargas (FGV), which undertook a form of evaluation based primarily upon existing data; an approach not supported by the Bank There was difficulty in finding common ground. The evaluation, as a part of the loan, was cancelled. Later, the Bank undertook its own impact evaluation of the impact of CrediAmigo (Skoufias et. al; 2009, 2013). 21 Outcome 4.5 In the absence of results from the planned impact evaluation under the CrediAmigo project, other attempts were made to assess program impact, both by the Bank and by researchers in Brazil. In the Bank study based on CrediAmigo, the impact of expanded microcredit on microenterprise clients in the northeast was evaluated using a quasi- experimental approach.16 It suggests some beneficial results: an increase in the utilization of credit from any source, over the CrediAmigo period up to 2003 (i.e, before the changes in national regulations) (1.2 percent) as well as an increase in formal credit as a primary source (1.5 percent). The study also showed that formal credit use was associated with improved profitability and increased employment of user enterprises. 4.6 Another evaluation of the CrediAmigo program was conducted by the Fundaqao Getulio Vargas (FGV), a prestigious Brazilian think tank. The FGV team used the entire set of microenterprises in the same areas where CrediAmigo operates as a control group. While these aggregates cannot be formally taken as a control group, they give an approximation to a control group by enabling the comparison of the behavior of the CrediAmigo customers with the population of all microenterprises in the same areas. 4.7 The results of the FGV study show that access to credit increased rapidly in the Northeast, faster than in the rest of the country and poverty declined in this region (as described in Marcelo Neri et al, 2008 (Table 4) over CrediAmigo's lifespan. Regarding the impact on poverty, the study found that people who have borrowed from CrediAmigo for five years are between 36 and 41 percent more likely to leave poverty than those who just have started to borrow from the institution. Even though these results cannot be directly attributed to CrediAmigo in the absence of a robust impact evaluation, they suggest that CrediAmigo helped to fuel the certainly expanded access to credit in the Northeast, which exceeded all the planned targets. And by the end of 2013, rapid growth in microfinance in Brazil was widely recognized. An increase of 25 percent was recorded at the end of 2013 compared to 2012, especially due to BNB and other public banks.19 16 Emmanuel Skoufias, Philippe Leite and Fransisco Haimovich (2009). 'Expanding Access to Microfinance: Evidence from the CrediAmigo Natural Experiment in the NE of Brazil'. World Bank, October. While the study was careful to point out the difficulties of assessing the impact of increased access to credit (in general, or through CrediAmigo), it sought to explain the impacts of increased supply of credit on credit utilization and thus on firm performance. 17 Marcelo Neri, Microcridito, 0 Mistrio Nordestino E 0 Grameen Brasileiro, Perfil and Performance dos Clientes do CrediAmigo. Rio de Janeiro: Banco do Nordeste, CrediAmigo, FGVand FGVeditora, 2008. 18 However, there was some double counting. 19 Agencia Brasil (2013). Microcridito 4 recorde no pals, mas atende apenas parte dos pequenos empreendedores. 0 Globo 2013 Dec 22. 22 Table 4. Access to Credit (Percentage of Population: 1997 and 2003) Year Northeast Elsewhere Growth 1997 3.97 5.34 1.3 2003 6.27 5.99 -0.28 Advantage to the Northeast 1.65 Source: Marcelo Neri, Microcredito, 0 Misterio Nordestino E 0 Grameen Brasileiro, Perfil and Performance dos Clientes do CrediAmigo. Rio de Janeiro: Banco do Nordeste, CrediAmigo, FGV and FGV editora, 2008. 4.8 Notwithstanding the limitations of the two studies cited above, and based largely on the achieved expansion of access to microfinance, with noted beneficial results on microenterprise profitability, employment, and regional poverty, the achievement of the objective of expanded availability of formal financial services is rated Substantial. Improve the access of microenterprises throughout the Northeast Region to sustainable financial services Outputs 4.9 Measures of the sustainability of CrediAmigo itself, as an institution, were largely achieved. These included improved client retention and loan officer productivity measured in terms of numbers of active and new clients. These were achieved through better marketing as well as better risk management and training, career development plans and an improved performance-based compensation system. Lending methodology improved. CrediAmigo developed operational routines for loan officers and harmonized procedures across branches. 4.10 Thus targets for the key indicators of sustainability- Operational Efficiency; Financial Sustainability (Adjusted Return on Assets before taxes) and Subsidy Dependence Index, were all met (Table 5). Table 5. CrediAmigo: Indicators of Sustainability Indicatorl: Operational Efficiency: Administrative Eficecy dmnita.v <24% 51% 43% 29% 25% 23% 22% Costs over average outstanding loan portfolio (%) Indicator 2: Financial Sustainability: Adjusted Return >3% 2% 1% 3% 4% 4% 8% on Assets before taxes (%) I I Subsidy Dependence Index Zero -11% -8% -4% -4% -1% -10% Equity to earning assets >15% -35.30% -17.20% 7.60% 14.10% 21.40% 31.10% Source: ICR 23 Outcomes 4.11 It is, however, critical to note that within the project's monitoring framework, the achievement of sustainability focused on the financial sustainability of CrediAmigo itself. However, the Bank had conceived of the project as an intermediate step towards broadening the use of sound microfinance to the entire financial system. As expressed in the PAD: "While supporting the expansion and institutional development of the CrediAmigo program, emphasis would be placed on progress towards full financial sustainability of service provision in order to permit eventual phase-out of public or donor assistance" (PAD, Pg 5). CrediAmigo was supposed to be an example that could have been reproduced by any financial institution. 4.12 The principle of lending on market terms which was the basis of the project design was reversed when the new administration's Resolution 3109 of the National Monetary Council introduced the obligatory earmarking of a percentage of sight deposits for microcredit, with the alternative of holding them as unremunerated reserves at the Central Bank. Attempting to earn returns on these deposits, commercial banks channeled their funds to institutions such as BNB, which consequentially gained considerable access to subsidized funding for its flagship microcredit program, CrediAmigo. 20As pointed out by the ICR, "BNB started using the newly available sources of funding to support the expansion of CrediAmigo, which together with the devaluation of the real, slowed Bank disbursements" 4.13 CrediAmigo was not spun off but remained within BNB. Partly also reflecting its improved operational efficiency, the Subsidy Dependence Index, as measured, was greatly reduced (-10) and eventually exceeded its target (zero) and CrediAmigo became a profit center. Thus while the sustainability targets were all met they do not reflect the impact of the new policies on the availability of funding from Banco do Nordeste which inter alia contributed to the decline in the need for World Bank funds, and the subsequent cancellation of around a quarter of the loan proceeds. These issues were exacerbated by the recent policy decisions introducing the CRESCER program (2011) that lowered the cap on microfinance lending rates even further, to 0.64 percent per month (around 8 percent a year) at a time when 21 Brazil's lending rates averaged around 44 percent per annum. 4.14 The generous access to funding through the government's new special programs facilitated the expansion of microcredit through CrediAmigo. CrediAmigo positioned itself well to reap the benefits of the new programs, taking steps to improve and strengthen its lending methods. The FGV study gave a reasonable estimate of the impact of CrediAmigo on microcredit provision and poverty reduction within the context of the prevailing conditions in the country's financial markets. But such studies were not designed to capture the counterfactual - i.e., the extent to which the availability of formal finance from other 20 The ICR adds: "The 2003 microfinance policies introduced by the Central Bank of Brazil had a direct influence on the pricing, profitability, and client composition of CrediAmigo, as well as in the use of alternative funding, namely funds from the Fundo de Amparo ao Trabalhador (FAT) and funds originating from National Monetary Council (CMN) Resolution 3109." The FAT fund is a workmen's compensation fund, largely for unemployment. Resources from these reserves were available for investment in alternative priority government programs. 21 Source: IMF, International Financial Statistics, 2013. 24 sources would have expanded, in the absence of the CrediAmigo program, which leaves a veil of uncertainty on their results. 4.15 Another measure of long term sustainability planned by the project was CrediAmigo's spin-off as a fully functional and separate institution. This was to help in transitioning microcredit activities to the free markets, build long term sustainability of CrediAmigo and demonstrate that microfinance can be profitable and sustainable. However the spin-off was not implemented. Eventually staying within Banco do Nordeste was beneficial for CrediAmigo through the sharing of premises with the bank, the cost sharing of certain value-added services (current accounts, ATM machines, life insurance and credit default insurance) which would otherwise distract large amounts of resources from an independent CrediAmigo, and critically, access to funding through the government's special programs. Banco do Nordeste argued that CrediAmigo could thus operate with a very small staff and concentrate on its core business, while assuring CrediAmigo's autonomy could be ensured through appropriate steps. 4.16 The arguments for keeping CrediAmigo within Banco do Nordeste however were well known when the loan was designed (PAD, Pg. 9). Yet, the decision at the time of appraisal was to spin it off because it was intended to transition CrediAmigo from the public to the private sector, after achieving long term sustainability and demonstrating the success of sustainable microcredit. It was also considered important and central to its access to market funding and long term sustainability both by the Bank and by Banco do Nordeste senior management (PAD, pg.58). 4.17 There were also clear advantages for Banco do Nordeste of keeping CrediAmigo within Banco do Nordeste. As discussed in paragraph 6.13, this included, first, a financial advantage. Under the new schemes introduced for microfinance in Brazil in 2003, banks were required to earmark a portion of deposits for microfinance, or transfer them to the Central bank as unremunerated reserves. Since many banks did not have the infrastructure or the experience to undertake microfinance lending, Banco do Nordeste saw this as an opportunity to gain deposits (remunerated at potentially low rates) from such institutions, to undertake microfinance lending on their behalf. Banco do Nordeste thus had access to a new line of business, and the undoubted expertise acquired by CreditAmigo in this area became a valuable source of business.22 4.18 Nevertheless it should be noted that quite apart from CrediAmigo, parallel and sound expansion of access to finance in Brazil for underserved consumers was also made, across both administrations, through a range of other measures. This includes, notably, the creation of correspondent banking, which has helped overcome the restrictions imposed by the high costs of bank branches and provided electronic access to the national payments system through retail locations. Other achievements after 2003 included the introduction of basic bank accounts. Recent measures include the availability of improved creditor information and strengthened financial consumer protection. The expansion in the provision of credit 22 However, uneasy tensions remained as a result of Banco do Nordeste effectively taking the best of both positions, in terms of the integration versus the spinoff of CrediAmigo. In 2003, employees at CrediAmigo protested that while they were effectively employees of Banco do Nordeste, they were not being paid accordingly in terms of rates and benefits for public bank employees. 25 cards, and consumer credit has been notable, though opportunities for enhanced savings have been scarce. 4.19 On balance, given all these considerations, the PPAR rates the achievement of the sub-objective of sustainability as Modest. 5. Efficiency 5.1 There is no cost effectiveness analysis to properly assess the efficiency of this project, and there is no information in the ICR to compare this project to similar projects in other countries. The ICR rates project efficiency based on the fact that during project implementation the operational efficiency of CrediAmigo improved substantially. However the measure discussed in the ICR is a measure of CrediAmigo efficiency and not of overall project efficiency, including economy wide objectives. As noted above, CrediAmigo gained funding through BNB to funding under special schemes at privileged rates, which detracted from the efficiency of microfinance for the economy as a whole. The project had a 15 month extension of the closing date to carry out activities that were not fully completed (to design a new model for providing sustainable financial services to microenterprises through Civil Society and Public Interest Organizations and to allow more time to complete the impact evaluation of the program), and one component of the project, the evaluation component, was not implemented. Moreover, almost a quarter of the project's funds were eventually cancelled, as the new policy regime implied that there was limited need for the Bank's line of credit, in comparison to more favorable funding. Based on these facts, efficiency is rated as Modest. 6. Ratings Outcome 6.1 Project outcome is Moderately Satisfactory. The objectives were highly relevant to the development of microenterprises in the Northeast of Brazil, which had a high priority in the Federal Government's efforts to eradicate poverty. The Relevance of Design was Substantial. Efficacy of the objective of increasing access to formal finance is rated as Substantial, however, achievement of access to sustainable finance is rated as Modest. Even though CrediAmigo expanded access beyond the stipulated target, and microfinance in the north east of Brazil grew substantially, this expansion was aided by the government's special programs for microfinance at interest rates below market rates, achieved through public banks. Efficiency too is rated Modest based on the fact that the project had a 15 month extension of the closing date to carry out activities that were not fully completed (to design a new model for providing sustainable financial services to microenterprises through Civil Society and Public Interest Organizations and to allow more time to complete the impact evaluation of the program), that one component of the project, the evaluation component, was not implemented, and the cancellation of a part of the project's proceeds. 26 Risk to Development Outcome 6.2 CrediAmigo has expanded access for microenterprises through solidarity networks and has solidified its operational and financial management. Other players, notably a handful of public banks, have also entered the arena and access to formal microfinance in Brazil is growing. However, as noted, policy changes brought about in 2003 (and reinforced by later changes in 2011) made it impossible to achieve the project's overall development objective, of increasing the provision of sustainable microfinance. It appears likely that there has been some crowding out of sound microfinance as a result of the new programs. Furthermore, CrediAmigo's spin- off as an independent institution, as planned in the PAD to facilitate its access to market funds and long term sustainability, was abandoned. CrediAmigo was institutionalized and absorbed within Banco do Nordeste. It was moved within the institution from a special program of the BNB's president to a formal Superintendency. As a part of BNB, it continues to benefit from special government programs. 6.3 These changes are contrary to Banco do Nordeste 's own stated goal of developing CrediAmigo as a sustainable institution with, according to Banco do Nordeste, a sustainable interest rate and direct access to financial markets. As stated, it intended to "develop a sustainable institution as indicated by its decision to charge an interest rate which is higher than most NGO programs in order to accelerate achievement ofprofitability and the ability to mobilize funds through savings and direct access to financial markets which is an essential element of longer-term sustainability." (PAD, Pg 21). However at the end of 2013, the average interest rate on BNB's microfinance lending, at 5 percent per annum, was considerably below prevailing market rates of over 40 percent. Based on these considerations, Risk to Development Outcome is rated as Significant. Bank Performance 6.4 Bank Performance is rated Moderately Satisfactory QUALITY AT ENTRY 6.5 Quality at Entry is rated Satisfactory. The project's objectives were well aligned with the country's priorities and the Bank's assistance strategy. The Bank team had worked closely with CGAP and other specialists in microfinance (ACCION International) since CrediAmigo's pilot phase and built the project directly on the lessons learned from the pilot phase at Banco do Nordeste. The project team had undertaken a huge amount of work in terms of mapping out the microfinance landscape in Brazil for the first time, benchmarking Brazil with regional comparators, and setting parameters for sustainability. It also correctly assessed the implementing agency's commitment and capacity. The project design had a strong monitoring mechanism. At the same time there were notable weaknesses at entry, discussed below, that impacted the achievement of the project's objectives. 6.6 First and foremost, although it would have been difficult for the project team to foresee such major changes in policy as the new government administration brought about in 2003, the project design failed to identify "continuity in policy framework" as an assumption for the successful achievement of its objectives. 27 6.7 The project design recognized the "moderate" risk of resistance, and the controversial nature of the proposed spin-off of CrediAmigo from Banco do Nordeste into a legally and financially independent financial entity (PAD, Pg. 24). However, its strategy to address this risk was not effective as the spin-off was not implemented. While this may have been partly due to the changed policy environment, the spin-off was by no means a done deal at the design stage and was dependent on the outcome of the review of the regulatory framework by the central bank and the approval of the Banco do Nordeste board. The design also failed to clearly lay out whether CrediAmigo would be a "partial or completely independent entity" (PAD Pg 13). Thus, while the changed political environment after 2003 may have stalled the proposed spin-off, the ambiguities at the design stage were also equally reflected in the cancellation of the spin-off. The project team may have been too optimistic and not sufficiently realistic, especially given the strongly entrenched roles of the major public sector banks in Brazil. QUALITY OF SUPERVISION 6.8 Quality of Supervision is rated Moderately Satisfactory. 6.9 It is noted, first, that throughout project implementation, supervision was intensive, particularly in terms of (i) regular monitoring of a wealth of indicators, (ii) 11 supervision missions, and (iii) a close working relationship and excellent cooperation with the implementing agency. 6.10 The Bank team attempted to proactively identify threats to the achievement of the Project Development Objectives and to respond fast to the policy changes in 2003. However, it did not adequately recognize that the policy framework and the borrower's decisions were changing reality in ways that turned the attainment of the loan's original objective of sustainability into an impossibility, notwithstanding good progress in terms of outreach. First, the 2003 change in interest rate policies and directed credit regulations affected the overall access to funds and profitability of CrediAmigo, and was a key factor leading to the request by the government for the cancellation of US$10.2 million. Second, the decision to keep CrediAmigo inside Banco do Nordeste also went against the principle of demonstrating sustainable and market based access to microfinance, as evidenced by the continued dominance of this market by CrediAmigo in particular, with recent support from a couple of other public banks. Tensions faced by the Bank's supervision team illustrate the difficulty faced in achieving both objectives. 6.11 As discussed, the policy framework on the basis of which the project had been designed had changed significantly, and it was no longer possible to attain core elements of the development objectives as originally conceived. Under these circumstances, either restructuring (to change the development objectives) could have been considered. The Bank did not undertake this although it was apparent that the project was not going to achieve its objective of financial sustainability in some business areas which may have been cross- subsidized by other areas of lending. The position adopted by the Bank team appears to have reflected the dilemma due to the fact that CrediAmigo's operations did indeed expand microenterprises' access to formal finance, and efforts by the Bank to maintain a constructive dialogue with a new administration. Thus the supervision team sought to take the measures it 28 could to protect sound operations at CrediAmigo, and try to ensure its functional independence, despite its absorption into BNB, and the access to BNB funding reflected in the calculation of its subsidy dependence. These factors may partly explain why the Bank did not proceed with formal restructuration with changed Project Development Objectives. 6.12 The project team's response to the decision to keep CrediAmigo within Banco do Nordeste was to strengthen its autonomy within Banco do Nordeste. The project team cited a number of advantages of remaining within Banco do Nordeste for CrediAmigo such as access to management information systems, facilitation of financial transactions and treasury services, routine personnel management, security, purchases, ATM machines, life insurance and credit default insurance, which would distract large amounts of resources from an independent CrediAmigo and which Banco do Nordeste cross subsidized. However, the advantages of staying on within CrediAmigo were well known even at the design stage (PAD, Pg 9) and yet CrediAmigo's independence was considered important and central to its access to market funding and long term sustainability both by the bank and the Banco do Nordeste senior management (PAD, Pg.58). 6.13 As noted above, there were also clear advantages for Banco do Nordeste of keeping CrediAmigo within its operations (para 4.18). Under the new schemes introduced for microfinance in Brazil, all banks were required to earmark a portion of deposits for microfinance, or, transfer them to the Central bank as unremunerated reserves. Since many banks did not have the infrastructure or the experience to undertake microfinance lending, the expertise acquired by CrediAmigo in this regard provided Banco do Nordeste with an opportunity to gain deposits (remunerated at potentially low rates) from such institutions, to undertake microfinance lending on their behalf. This added to the reasons why the more expensive Bank line of credit was no longer needed. The PAD cites potential tensions on that account to Banco do Nordeste's spin off and identified this as a controversial aspect on that account. However, in its justification for supporting CrediAmigo's autonomy within Banco do Nordeste, the project team does not mention the new advantages of retaining CrediAmigo's line of business to Banco do Nordeste. 6.14 The estimation of the indicator of Subsidy Independence (or the Subsidy Dependence Index,) continued to be benchmarked to the interbank rate, however once it was clear that the CrediAmigo program would remain within BNB, CrediAmigo could also benefit from the volumes of low funds made available by the government for microfinance programs. 6.15 Another notable important element was the cancellation of the impact evaluation. The impact evaluation was intended to be an important mechanism to demonstrate the effectiveness of sustainable microfinance as a poverty reduction tool to the government. Its cancellation was therefore a missed opportunity. The project team could have tried to persuade CrediAmigo to go ahead with the evaluation using the control group on the basis suggested by the World Bank. Borrower Performance 6.16 Borrower performance is rated Moderately Satisfactory. 29 GOVERNMENT PERFORMANCE 6.17 The government performance is rated Moderately Satisfactory. 6.18 The government kept in place a stable macroeconomic environment conducive to the success of the project throughout its implementation and beyond, and was committed to increasing microenterprises' access to finance throughout the life of the project. It undertook a number of initiatives to increase access to the unbanked such as through simplified deposit accounts and correspondent banking. Yet, the nature of many of the policies used to foster the sustainable provision of microcredit after 2003, although well-intentioned, increased the potential of crowding out private microfinance. While the project was designed to help in the facilitation of a transition toward freer markets, policy changes introduced in 2003 added distortions to the pricing of microfinance and discouraged the provision of funds through the free market. 6.19 The policy reforms in 2011 further aggravated this problem. The launch of CRESCER program gives a further subsidy for public sector institutions exclusively to reduce the rates of interest charged for microcredit solidarity loans. This subsidy allowed CrediAmigo to reduce its rate of interest from 1.2 percent per month (15.4 percent annually before compounding) to 0.64 per month (8.0 percent annually before compounding). This has boosted CrediAmigo's strength in the microcredit market. However, it puts the interest rate well below the cost of funding for private banks. These policies run contrary to the project objectives as well as to the development of sustainable microfinance. THE IMPLEMENTING AGENCY 6.20 The implementing agency performance is rated Moderately Satisfactory. 6.21 Banco do Nordeste was actively involved in the project design phase and remained highly committed through project preparation and implementation. There were no problems in financial management and procurement. Monitoring arrangements were well designed and implemented efficiently. Monitoring data was promptly produced and utilized in decision making. Decision-making and resource allocation were based on the M&E system. However, there were some weaknesses: Banco do Nordeste did not carry out two important contractual activities: CrediAmigo's spin-off and the impact evaluation. With respect to the former the Bank team revised its strategy and worked to establish CrediAmigo's autonomy within Banco do Nordeste , while with respect to the latter, it cancelled the impact evaluation as it could not come to an agreement on the design of the evaluation methodology with the implementing agency. 6.22 The changes in the political environment likely influenced the Banco do Nordeste's decision regarding CrediAmigo's spin-off as an independent institution as it began to be funded by government subsidies. Additionally, keeping CrediAmigo within its fold became more advantageous to Banco do Nordeste including through the access to CrediAmigo's captive subsidized credit line. The PAD also refers to potential tensions on account of CrediAmigo's independence as Banco do Nordeste stood to gain in prestige by having this successful microfinance model within its fold (PAD, Pg 24). An impact evaluation of 30 CrediAmigo operation was later conducted independent of the Bank. However its results are not robust. Quasi experimental analyses undertaken within the Bank demonstrate the beneficial results of microfinance for users, but do not address the question of the benefits of the CrediAmigo approach, compared to other potential providers of microfinance. 6.23 CrediAmigo's attempt to shield itself from political influence in its operations was also not successful. CrediAmigo began operations outsourcing credit promotion and analysis as a way to reduce or eliminate potential political influences in these activities. This procedure, however, was successfully challenged in the court of law. The court decided that, since CrediAmigo is a public sector institution, public sector employees should conduct its main activities. Monitoring and Evaluation 6.24 Monitoring and evaluation is rated Modest 6.25 Design: The project had a robust monitoring and evaluation system, in terms of measures for financial sustainability of CrediAmigo itself (return on assets, portfolio quality, risk management, subsidy dependence), efficiency (administrative cost ratio, loan officer productivity, client generation and retention), achieving development objectives (breadth and depth of outreach) and impact on client business development. However there is no discussion of indicators that could track overall microfinance availability and cost including that from institutions other than CrediAmigo. 6.26 Implementation: Available measureable indicators and good project management allowed the tracking of results at CrediAmigo. Implementation, however, had flaws. As noted earlier, the Subsidy Dependence Index used the interbank borrowing rate as a reference cost. And the Subsidy Dependence Index did not reflect the changing availability of funding to BNB itself. Second, the final impact evaluation, a project component, was not carried out as planned, within the scope of the project. 6.27 Utilization: M&E was used to make managerial decisions. An annual due diligence review to provide perspective on past performance rounded out the information system. Data collection, analysis and monthly reporting were as expected and the information was used to evaluate progress and make adjustments to operations. The annual due diligence reviews were delayed early in the program, but this issue was resolved by replacing the contractor. On balance, and reflecting all these factors collectively, the M&E system is rated Modest. 7. Lessons 7.1 A number of lessons can be drawn from the early pilot of CrediAmigo and the project's implementation that are useful for other microfinance institutions, the Bank, and other donors for the design of microfinance operations. The lessons can be considered under two rubrics: those applicable to any project and those applicable specifically to microfinance. 31 Lessons Applicable to All Projects 7.2 Flexibility in supervision to adapt to changing scenarios is essential for success, both when the changes are marginal and when they are fundamental to the point of rendering the attainment of the operation objectives impossible. In the former case, flexibility must be manifest in the willingness to sacrifice details for the sake of achieving the ultimate objectives of the operation. In the latter case, this requires courageous decisions from the project team, to the point of significantly restructuring the operation, should it be necessary to remain true to original objectives. 7.3 The Bank can help in building lasting reforms by limiting the scope of the projects to what can be accomplished within a certain political and economic environment, even if not fully market-oriented. Nevertheless it must be cognizant of the fact that the environment is fragile and may depart further from market based principles. 7.4 Donors such as the Bank can add value by setting benchmarks consistent with international best practice and putting the client institution in contact with top practitioners in each field. Lessons Applicable to Microfinance Projects 7.5 In principle, a viable and broadly independent microfinance institution can be set up within an existing public sector bank and there may be benefits to doing so. However, this leaves it open to influence by the prevailing political climate. In an environment where public sector financial intermediation, or some segments of it, may be subsidized by the state, it may be difficult to isolate the 'independent' microfinance program. Even if ring fenced today, unless spun off, the micro-finance institution remains vulnerable to future recapture by the parent institution. Tensions about the extent to which public or private sector principles apply persist. 7.6 There can be tensions between the achievement of expanded microfinance outreach and the achievement of financial sustainability. It may be tempting in some political environments to expand outreach at the expense of sustainability. The Bank's decision to remain engaged was anchored on real successes achieved in terms of outreach, which illustrates dilemmas about outreach versus sustainability. Yet expansion of outreach to a desired clientele is not necessarily an indicator of sustainability. 7.7 Transparency can be a powerful aid to sustainability. Subsidies can be explicitly taken into account through a Subsidy Dependence Index provided it is well defined. However, it is difficult to incorporate the measure of subsidies in funding. 7.8 While institutional aspects are crucial for the delivery of financial services to microenterprises, national economic policies are equally crucial. The operations of well managed institutions can reflect financial unsustainability from the point of view of the economy as a whole if the enabling environment reflects distortions in financial markets  33 References ABCRED 2012. (Associaqdo Brasileira de Entidades Operadoras de Microcredito e Microfmanqas - Brazilian Microfinance Association). Interview with Almir da Costa Pereira, President April 17th. As reported in 'guaracui.com'. Responsible Finance and Sustainable Development Agencia Brasil (2013). Microcr6dito 6 recorde no pais, mas atende apenas parte dos pequenos empreendedores. 0 Globo 2013 Dec 22. Belaisch, Agnes. 2003. "Do Brazilian Banks Compete?" IMF Working Paper No. 03/113, Washington, DC. de Mariz, Frederic 2012. 'Brazilian Microfmance: discussion with ABCRED. Private Players Small but Growing' At http://guaracui.com/. April 27 Dollar, D. and Kraay, A. 2002. "Growth is Good for the Poor," Journal of Economic Growth, vol. 7, p 195- 225, Kluwer Academic Publishers. Economist Intelligence Unit 2012 Global Microscope on the Microfinance Business Environment 2012. An Index and Study. Eliana Cardoso. 2003. "Seigniorage, Reserve Requirements and Bank Spreads in Brazil," Patrick Honahan (Ed), Taxation of Financial Intermediation Theory and Practice for Emerging Economies (p 70- 92). New York: Oxford University Press. Govemo Federal do Brasil. 2011. CRESCER. Programa Nacional de Microcr6dito. Brasilia, 24 de agosto. Kumar, A., Access to Financial Services in Brazil. 2004. World Bank Washington, DC. , Thorsten Beck, Christine Campos and Soumya Chattopadhyay. 2005 "Assessing Financial Access in Brazil," Working Paper No. 50, World Bank Washington, DC. Neri, Marcelo (ed.). Microcr6dito 0 Mist6rio Nordestino e o Grameen Brasileiro: Perfil e performance dos clientes do CrediAmigo. Rio de Janeiro: Editora FGV, 2008. . 2010. Brazilian Microcredit Takes Off. Article published in Brazilinfocus.com. 30 November. OECD 2010. Tackling Inequalities in Brazil, China, India and South Africa. The Role of Labour Market and Social Policies. . 2013. Brazil:Combining Growth and Social Inclusion. The Contribution of Regional Policies. OECD Territorial Reviews. Olsen, Tricia 2013. Rethinking Collective Action:The Case of Microfinance in Brazil and Mexico.Paper Prepared for the Si6 Research Seminar Series, Joseph Korbel School of Intemational Studies, University of Denver. May 17, SEBRAE. 2011 Perfil das Instituiq6es de Microfinanqas (OSCIP) no Brasil Annual Report. Segura-Ubiergo, Alex. 2012. "The Puzzle of Brazil's High Interest Rates." IMF Working Paper No. 12/62, Washington, DC. Skoufias, Emmanuel. 2007. "Brazil: Measuring Poverty Using Household Consumption." Report 36358- BR, Poverty Reduction and Economic Management Sector Unit, Latin America and the Caribbean Region, World Bank, Washington, D.C. , Phillippe Leite, Fransisco Haimovich. 2009. Expanding Access to Microfmance:Evidence from the CrediAmigo Natural Experiment in the NE of Brazil. World Bank October. Powerpoint. ., P. Leite, and R. Narita. 2013. "Expanding Microfmance in Brazil: Credit Utilisation and Performance of Small Firms." Journal of Development Studies Vol. 49, No. 9, pp. 1256-1269. http://www.tandfonline.com/eprint/uR3CvzcqqzrU7IgizteT/full). Weisman, Ethan, and Femando Blanco 2007. BRAZIL. Improving Fiscal Circumstances for Growth. World Bank Latin America Region. 'En Breve' March 2007. Number 104. World Bank. 1995. "Brazil: A Poverty Assessment." Human Resources Operations Division, Country Department 1, Report No 14323-BR. Washington DC. 34 . 2006. "Brazil: Interest Rates and Intermediation Spreads." Brazil Country Management Unit, PREM Unit, Report No. 36628-BR, Washington, DC. 2007. Brazil Improving Fiscal Circumstances for Growth*Ethan Weisman and Fernando Blanco. 'En Breve' March 2007 * Number 104 2013, Financial Inclusion in Brazil - Building on Success. Financial Sector Assessment Program Update-Brazil. Washington DC, May. 35 Annex A. Basic Data Sheet Brazil Northeast Microfinance Development Project (P050776, IBRD- 45540) Key Project Data (amounts in US$ million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Total project costs 100.0 80.1 80.1 Loan amount 50.0 38.5 76.2 Cancellation 00.00 11.5 00.0 Cumulative Estimated and Actual Disbursements FY01 FY02 FY03 FY04 FY05 FY06 FY07 Appraisal estimate (US$M) 3.7 9.8 19.6 33.0 45.1 50.0 50.0 Actual (US$M) 7.9 12.7 16.8 19.8 28.8 36.5 38.5 Actual as % of appraisal 213.5 129.5 85.7 60.0 63.8 73.0 77.0 Date of final disbursement: December 20, 2006 Project Dates Original Actual Initiating memorandum 09/18/1998 09/18/1998 Negotiations 04/03/2000 03/22/2000 Board approval 05/30/2000 05/30/2000 Effectiveness 09/06/2000 09/06/2000 Closing date 12/31/2005 03/31/2007 ANNEX A 36 Task Team Members Names Titles Unit Responsibility Lending Steven N. Schonberger Senior Rural Development Specialist LCSES Team Leader Irani Escolano Procurement Specialist LCSPT Susana Sanchez Financial Specialist LCSFP Carlos Cuevas Lead Financial Economist FSD Robert Christen Senior Microfinance Advisor Fred Levy Principal Economist Tulio Correa Financial Management Specialist LCSFM Marta Molares-Hoberg Senior Counsel LEGLA Morag van Praag Senior Disbursement Officer LOA Luis Coirolo Lead Specialist LCSAR Peer Reviewer Jacob Yaron Peer Reviewer Khalid Siraj Consultant Peer Reviewer Arie Chupak Richard Rosenberg Consultant PSD Peer Reviewer Beatriz Iraheta Language Program Assistant LCSEN Supervision/ICR Susana M. Sanchez Senior Financial Economist LCSSD Team Leader Jennifer Sara Sector Leader LCSSD Susana Amaral Financial Management Analyst LCSFM Musa S. C. Asad Senior Financial Analyst LCSEN Tulio Correa Financial Management Analyst LCSFM Jose C. Janeiro Senior Financial Management Speciali LCSFM Pedro Olinto Senior Economist LCSPP Emmanuel Skoufias Lead Economist PRMPR Ilias Skamnelos E T Consultant ECSPF Luciano Wuerzius Procurement Analyst LCSPT 37 Staff Time and Cost Staff Time and Cost (Bank Budget Only) Stage of Project Cycle US$ Thousands (including No. of staff weeks travel and consultants costs) Lending FY97 10.49 FY98 0.19 FY99 40.97 FY00 37 138.59 FY01 9 35.32 FY02 0.00 FY03 0.16 Total: 46 225.72 Supervision/ICR FY00 0.79 FY01 6 39.40 FY02 13 39.89 FY03 15 84.89 FY04 17 80.60 FY05 17 83.25 FY06 16 58.81 FY07 7 25.90 Total: 91 413.53  39 Annex B. List of Person Met In Washington D.C. Ms. Susana Sanchez Mr. Fernando Montes Ms. Margaret Miller Mr. Ilias Skamnelos In Sao Paulo Mr. Loy Pires, IFC, In Bank headquarters Mr. Magnus Lideloa, HD Sector Leader In the Central Bank Mr. Aloisio Tupinamba and his team, Chief of Staff of Financial System department. Mr. Paulo Sampaio, Financial System Department In the Banco de Brazil Mr. Alvaro Scepious, Small Scale Enterprises and Micro Credit, Brasilia. Mr. Alvaro Toscetto, Agribusiness Credit, Brasilia. In CrediAmigo Ms. Annette Torres, Superintendent CrediAmigo, Banco do Nordeste. Marcelo Azevedo, Gerente de Ambiente do Microfinanqa Urbana, Banco do Nordeste Ten agents in a CrediAmigo branch in Fortaleza. In VivaCred (Rio de Janeiro) Mr. Teofilo Cavalcanti, Director, VivaCred, Rio de Janeiro. 10 agents in VivaCred branch. In the private sector Mr. Marcelo Cortes Neri, Chief Economist for Center for Social Policies (CPS) at Getulio Vargas Foundation (FGV). Valter Alves de Macedo, Banco Santander Vanessa Cabral, Banco Santander  41 Annex C. Recent Developments in Microfinance in Brazil An expected outcome (albeit untracked in the results framework) of the success of the CrediAmigo model was that other banks and financial institutions would enter the microfinance market encouraged by CrediAmigo's success. While the project's focus was on the institutional development of CrediAmigo, the Bank saw this emphasis as an intermediate step towards broadening the scope to the entire financial system. As expressed in the PAD: "While supporting the expansion and institutional development of the CrediAmigo program, emphasis would be placed on progress towards full financial sustainability of service provision in order to permit eventual phase-out of public or donor assistance" (p 5). CrediAmigo was supposed to be an example that could have been reproduced by any other financial institution. Thus, even if no explicit numbers were given for the overall increase in credit, it is clear that its expansion would be larger than that which would take place in CrediAmigo itself. This did not happen. The substitution of sustainable private sector credit for government supported public sector credit, which appeared feasible at the time of the appraisal, became difficult after the introduction of new government policies in 2003 and became less feasible with subsequent policy changes. The private sector's participation in microcredit expansion was eventually very small. The government's increasing subsidization of government-owned financial institutions continued crowding out the scant private participation that remained in this market. Though there were no targets to track this, the private sector did enter the microfinance market, mainly in the shape of NGOs. One large bank, the Banco Real of Sao Paulo, entered the field as well, mainly in its Northeast operations (this bank was subsequently bought by Banco De Santander, which continued the program). It is difficult to attribute this to the demonstration effect of CrediAmigo, given that lending for microfinance overall was rendered more attractive by the government's change of policy. Leading microfinance personae and observers in Brazil note that in recent years private sector participation in expanded microcredit has been small. As noted in an interview with the president of ABCRED (Associaqio Brasileira de Entidades Operadoras de Microcredito e Microfinanqas- the Brazilian Microfinance Association) Brazilian private sector microfinance continues to suffer from scarce and expensive funding, despite sound operational performance data. Restrictions on deposit taking by NBFIs remain.23 And in its annual report on the profile of microfinance institutions, SEBRAE notes that over the past decade the number of institutions registered under the productive microfinance credit program was reduced from 152 to 103.24 Other authors have also pointed towards the continued domination of the BNB in terms of support for socially oriented programs in the Northeast,25 and deleterious effects of some of the 2003 administration's policies on sound microfinance.26 23 Frederic de Mariz April 27 2012. 'Brazilian Microfinance: discussion with ABCRED. Private Players Small but Growing' At http://guaracui.com/. The article cites: "The Brazilian microfinance sector unfortunately lags far behind its Latin American peers as a result of its inappropriate regulatory environment. MFIs have a tough time raising funding, suffer from caps on interest rates and from the crowding out effects of public programs. Brazil doesn't have to re-invent the wheel, all it has to do is look at its neighbors to design a friendly regulatory environment that would enable this giant country to release its potential on the microfinance sector." 24 SEBRAE (2012) Perfil das Institui(5es de Microfinan(as (OSCIP) no Brasil Annual Report. January. 25 OECD 2013. Brazil:Combining Growth and Social Inclusion. The Contribution of Regional Policies. OECD Territorial Reviews. BNB's social and microfinance programs also benefitted from central government support through its Constitutional Funds, as well as from the proceeds from the FAT (Fondo de Amparo do Trabalhador), the national ANNEX C 42 In its annual Global Microscope on the microfinance business environment (2012), the Economist Intelligence Unit (EIU) notes, in its profile on Brazil, that ."The government has boosted microcredit to serve social purposes. Public institutions are being mobilized to extend microfinance loans at below market rates whereby the Treasury transfers funds mainly to public sector banks" In its overall microfinance business environment rankings for 55 countries, Brazil ranks 16th, behind most regional comparators. EIU acknowledges positive aspects of the microfinance environment in Brazil for example in terms of transparency and credit information but the overall rankings are pulled down by the regulatory environment, where Brazil ranks 21s among comparators. According to the World Bank / IMF financial sector assessment of Brazil in 2013, in 2012 BNB remained by far the largest microcredit provider in the country, with a market share of over 70 percent, more than a million active clients in the CrediAmigo Program and another 650,000 in a parallel program it established, AgroAmigo. Three state-owned banks continue to be the main source for financial products and services for low income consumers.28 One private bank, Banco De Santander, has entered the field but it is difficult to attribute this to the demonstration effect of CrediAmigo, or to the effect of the availability of earmarked funding through specialized programs. As of today, the crowding out impact of the government's programs appears to be dominating any potential crowing in from the success of the CrediAmigo example. A recent review cites that the recent increase in microfinance can be almost entirely attributed to the entry of two additional public sector banks, Banco do Brasil and the Caixa Economica Federal.29 Researchers in Brazil point out that despite recent growth; microfinance in Brazil still reaches only a fraction of its potential clients, pending more universal participation by private sector banks. According to Lauro Gonzalez, coordinator of the Center for Microfinance Studies in the School of Business Administration at Sdo Paulo's Fundagio Getulio Vargas (FGV), there needs to be investment and innovation at private banks, that need to develop appropriate technologies and products for the public - but private bank participation is effectively inhibited by subsidies offered through the Crescer Program.30 workmen's support and unemployment fund, as noted in OECD(2010): 'Tackling Inequalities in Brazil, China, India and South Africa. The Role of Labour Market and Social Policies'. 26 Olsen, Tricia 2013. Rethinking Collective Action: The Case of Microfinance in Brazil and Mexico. University of Denver. May 17. The paper discusses the disappointed reaction of Brazil's microfinance sector to the changes of 2003, noting that the changes tended to further increase the role of the government, instead of expanding sound private microfinance. 27 Including Peru, Bolivia, El Salvador, Colombia, Mexico, Panama, Ecuador, Paraguay and Chile. Brazil also ranked behind Uganda, Brazil, Rwanda and Tanzania. 28 World Bank (2013), 'Financial Inclusion in Brazil - Building on Success'. Financial Sector Assessment Program Update - Brazil. Washington DC, May. 29 As cited in Agencia Brasil (2013):Microcridito 4 recorde no pais, mas atende apenas parte dos pequenos empreendedores. 0 Globo 2013 Dec 22. 30 Ibid. Agencia Brasil (2013):Microcridito 4 recorde no pals, mas atende apenas parte dos pequenos empreendedores. 0 Globo 2013 Dec 22. 43 Annex D. Borrower Comments CrediAmigo SBanco doNordeste Ref. 2014-00580/0049 Fortaleza, 08 de Abril de 2014 Mr. Pia Schneider Acting Manager, Public Sector Evaluation Independent Evaluation Group The World Bank |BRAZIL: Northeast Micro Finance Development Project (11BRD-45540), Draft Project Performance Assessment Report Senhor Schneider, O Banco do Nordeste firmou o acordo de empréstimo com o Banco Mundial em novembro de 2000, com os seguintes objetivos: a) Expandir a disponibilidade de serviços financeiros para as Microempresas na área do projeto. b) Garantir a viabilidade financeira de tais serviços. Referido acordo foi constituido das seguintes partes: Parte a) Subempréstimos Parte b) Fortalecimento Institucional do Programa de Microfinanciamentos Parte c) Avaliação de Impacto Indicadores de monitoramento do projeto foram definidos para cada um dos objetivos e seguem abaixo na posição de fechamento do projeto (2006) e atualizados até 31.12.2013. a) Indicadores de expansão: (objetivo a) Indicadores de Expansão Esperado 2006 2013 Clientes Ativos 150,000 235.729 1.659,699 Municipios Atendidos No targeti 1.420 1 995 Carteira Ativa (R$ em Mii No target 177 621.238 2.156.811,77 Média de Empréslmos sobre o Pb Per Capta da eg ão Nordeste 65,00% 19% 16% Av. Pedro Ramalho, 5700 Passare 60743-902 - Fortaleza-CE - Brasil For e: 0800 72830 30 E-masi: wwigr!albts@n.Ro,br Hfomepage: www.bnb.gov.br ANNEX D 44 pág. 2/3 b) Indicadores de produtividade: (objetivo b) Indicadores de Sustentabilidade Esperado 2006 2013 Custos Administrafivos 24,00% 22% 12,80% ReonoAusaoSobre AU vos 3,00% 8% 16,86% Ínie:: eedê sde Sutbsidios 0,0%-0%rd. Aetorno sobre Patrirrôrio Líquido 15,00% 31,10%J 33.8a Como se depreende dos resultados demonstrados acima o Crediamigo - programa de microfinanças do Banco do Nordeste se consolidou como a maior e uma das mais bem sucedidas experiências de microfinanças da América do Sul. Os objetivos estabelecidos no projeto tanto no que se refere ao alcance, como à sustentabilidade foram alcançados de forma mais que satisfatória. O apoio técnico e financeiro do Banco Mundial foram de fundamental importância para a implantação e consolidação do Programa, em bases sólidas e em consonância com as melhores práticas de microfinanças existentes à época. Em que pese a avaliação de impacto não haver sido realizada no âmbito do Acordo de Empréstimos, o Banco do Nordeste viabilizou a realização do estudo de monitoramento dos resultados do Programa Crediamigo junto à Fundação Getúlio Vargas1, centro de pesquisa de excelência no Brasil. Referida Fundação identificou, dentre outros, os seguintes principais resultados: a) Impacto do acesso ao crédito como impulsionador da superação da condição de pobreza, principalmente para clientes com mais de 05 anos no Crediamigo. b) Aumento médio da renda para clientes de cerca de 35%. c) Elevação do consumo das famílias em aproximadamente 28%. Adicionalmente, a atuação do Programa Crediamigo no segmento das microfinanças trouxe os seguintes benefícios para este setor no Brasil: a) Contribuição para criação do marco regulatório das microfinanças no país. Tal situação permitiu a entrada de novos atores no setor e promoveu melhorias na forma de atuação das instituições. b) Existência de um modelo bem sucedido, com escala e resultados satisfatórios no Nordeste do Brasil, que se coloca como padrão para o setor. O Banco do Nordeste recebe visitantes do Brasil e do exterior interessados em conhecer seu modelo de atuação nas microfinanças. Os resultados dos estudos da Fundação Getúlio vargas sobre o Crediamigo estão disponíveis em http)://lcps.fgv, br/crediamigo. 45 ANNEX D pág. 2/3 c) Expansão interna do modelo de atuação com a criação, em 2005, do Agroamigo - programa de microcrédito do Banco do Nordeste, voltado para o setor rural. O Agroamigo possuía 734.265 clientes ativos em 31.12.2013. d) Disseminação da experiência do Crediamigo e Agroamigo em fóruns e eventos nacionais e internacionais promovendo a consolidação da visão estratégica das microfinanças como fonte de redução da pobreza e melhoria da qualidade de vida das populações de baixa renda. Oportuno o registro de que a busca constante pela excelência de seus resultados se mantém como direcionamento para o Programa Crediamigo. Após 14 anos da formalização do Acordo de Empréstimos o desempenho alcançado demonstra o compromisso do Banco do Nordeste em conduzir o Crediamigo em consonãncia com as melhores práticas de gestão existentes e o foco de atuação junto às populações de baixa renda. Os resultados foram consubstanciados pela elaboração de ratings financeiro e social em 2012, que avaliaram as práticas de gestão do Crediamigo e as classificaram como compatíveis com as melhores práticas das instituições do setor. Diante dos resultados alcançados, avaliando-os em comparação com os objetivos do Acordo de Empréstimos, a posição do 3anco do Nordeste é de que os resultados alcançados pelo projeto foram Satisfatórios. CoLocamo-nos à disposição para quaisquer esclarecimentos julgados necessários. Cordialmente, DIRETOR DE DESENVOLVIM NTO SUSTENTÁVEL E MICROFINANÇAS ANNEX D 46 Translation CrediAmigo Banco do Nordeste Av. Pedro Ramalho, 5700 Passar6 60743-902 Fortaleza CE Brazil Telephone: 0800 7283030 Email: clienteconsulta(,bnb.gov.br Website: www.bnb.gov.br Ref. 2014-00580/0049 Fortaleza, April 8, 2014 Mr. Pia Schneider Acting Manager, Public Sector Evaluation Independent Evaluation Group The World Bank Brazil: Northeast Microfinance Development Project (IBRD-45540), Draft Project Performance Assessment Report Dear Ms. Schneider: The Banco do Nordeste signed a loan agreement with the World Bank in November 2000, based on the following objectives: (a) Expanding the financial services available to microenterprises in the project area. (b) Guaranteeing the financial viability of these services. The following were the components of this agreement: Part (a) Subloans Part (b) Institutional Strengthening of the Microfinance Program Part (c)Impact Evaluation The project monitoring indicators were identified for each objective and the status on the date of project closing (2006) and an update as of December 31, 2013 are provided below: (a) Expansion Indicators (objective a) Expansion Indicators Projected 2006 2013 Active clients 150,000 235,729 1,659,699 Municipalities covered No target 1,420 1,995 Active portfolio (in millions of R$) No target 177,621,238 2,156,811.77 Average loan size to per capita GDP in the 65% 19% 16% Northeast Region 47 ANNEX D (b) Productivity indicators (objective b) Sustainability Indicators Projected 2006 2013 Administrative costs 24% 22% 12.88% Adjusted return on assets 3% 8% 16.86% Subsidy dependence index 0% -10% N/D Return on liquid assets 15% 31.10% 33.88% As the above results indicate, CrediAmigo, the microfinance program of Banco do Nordeste, became the biggest and one of the most successful microfinance experiments in South America. The project objectives established in terms of scope and sustainability were more than satisfactorily achieved. The technical and financial support received from the World Bank were critically important for establishing and strengthening the program on a solid basis and in keeping with the microfinance best practices existing at that time. Despite the fact that the impact evaluation was not conducted in the context of the loan agreement, Banco do Nordeste facilitated the study to monitor the results of the CrediAmigo Program in conjunction with the Getulio Vargas Foundation,31 Brazil's research center of excellence. This Foundation identified the following main results, among others: (a) The impact of access to credit as the impetus for overcoming poverty, primarily among clients that have belonged to the CrediAmigo program for more than five years. (b) Average increase in the income of clients of close to 35 percent. (c) Higher consumption rate of families (approximately 28 percent). In addition, in terms of the microfinance sector in Brazil, the work of the CrediAmigo Program has: (a) Contributed to the establishment of a microfinance regulatory framework in Brazil, which has facilitated the entry of new actors into the sector and led to improvements in the operations of institutions. (b) Created a highly successful model, which has become the standard for the sector, at the scale needed and with satisfactory results in Northeast Brazil. Banco do Nordeste has received visitors, from within and outside Brazil, who are interested in learning about its microfinance operating model. 31 The findings of the studies done by the Getfflio Vargas Foundation on CrediAmigo are available at http://cps.fgv.br/crediamigo. ANNEX D 48 (c) Led to the internal expansion of the operating model, with the establishment in 2005 of AgroAmigo- the Banco do Nordeste microcredit program geared toward the rural sector. AgroAmigo had 734,265 active clients as of December 31, 2013. (d) Sharing of experiences related to CrediAmigo and AgroAmigo at national and international forums and events, promoting the strengthening of the strategic vision of microfinance as a mechanism for reducing poverty and improving the quality of life of low-income population groups. It should be noted that the ongoing quest for excellent results is the central focus of the CrediAmigo Program. Fourteen years after the conclusion of the loan agreement, the performance achieved demonstrates the commitment of Banco do Nordeste to carry out the CrediAmigo Program in keeping with current best practices in the area of management and to focus its work on low-income population groups. The results were corroborated by the 2012 financial and social ratings, which evaluated CrediAmigo's management practices and rated them as being compatible with the best practices of the institutions in the sector. In light of the results obtained and based on their evaluation in relation to the loan agreement objectives, it is the view of Banco do Nordeste that the project results achieved were satisfactory. We are available to provide any clarifications deemed necessary. Sincerely, /s/ Stelio Gamy Lyra Junior Director, Sustainable Development and Microfinance