Report No. 17689-ES El Salvador Rural Finance: Performance, Issues, and Options April 10, 1998 Finance, Private Sector & Infrastructure Central America Department Latin America and the Caribbean Region CURRENCY EQUIVALENTS Currency Unit Col6n (0) $1 = 8.7 (May, 1996) ABBREVIATIONS AND ACRONYMS AMPES Association of Salvadoran Small and Medium Enterprises (Asociaci6n de Pequen7os y Medianos Empresarios Salvadorenos) BMI Multisectoral Investment Bank (Banco Mulitsectoral de Inversiones) BCR Central Bank of El Salvador (Banco Central de Reserva de El Salvador) BFA Agricultural Development Bank (Banco de Fomento Agropecuario) CAM Microenterprise Support Center (Centro de Apoyo a la Microempresa) CEDEAGRO Agricultural Sector Time Deposits (Certificado de Dep6sitos a Plazo Agropecuario) CC Commercial Code (Codigo Comercial) EAP Economically Active Population FAO United Natios Food and Agriculture Organization FEDA Agricultural Development Trust Fund (Fideicomiso Especial de Desarrollo Agropecuario) FEDECACES Federation of Salvadoran Credit Unions (Federacion de Asociaciones Cooperativas de A.horro y Credito de El Salvador) FEDECREDITO Federation of Credit Banks (Federaci6n de Cajas de Credito) FIGAPE Guarantee Fund for Small Enterprises (Fondo de Financiamiento y Garantiapara la Pequeha Empresa) FIS Social Investment Fund (Fondo de Inversi6n Social) FOCAM Credit Fund for Environmental Activities (Fondo de CrMdito para el Medio Ambiente) FOGAPE Guarantee Fund for Small-Scale Entrepreneurs (Fondo de Garantias para Pequenos Empresarios) FOGARA Agricultural Guarantee Fund (Fondo de GarantiaAgropecuaria) FOSAFFI Financial Sector Recapitalization Fund (Fondo de Saneamiento y Fortalecimiento Financiero) FUSADES Salvadoran Foundation for Economic and Social Development (Fundaci6n Salvadoreila para el Desarrollo Econ6mico y Social) IDB Inter-American Development Bank INSAFOCOOP Cooperative Institute (Instituto Salvadoreflo de Fomento Cooperativo) ISSS Salvadoran Social Security Institute (Instituto Salvadoreho de Seguridad Social) ISTA Salvadoran Institute of Agricultural Transformation (Instituto Salvadoreno de Transformaci6n Agraria) MAG Ministry of Agriculture and Livestock (Ministerio de Agricultura y Ganaderia) NGO Non-governmental Organization NIT Tax Identification Number (Nu2mero de Identificaci6n Tributaria) PROPEMI Program for the Promotion of Small and Microenterprises (Proyecto de Promoci6n a la Pequeha y Micro Empresa) PRN National Reconstruction Program (Programa de Reconstruccidn Nacional) SCC Savings and Credit Cooperative (Asociaciones Cooperativas de Ahorro y Credito) SSF Superintendency of the Financial System (Superintendencia del Sistema Financiero) TIBP Basic Deposit Interest Rate (Tasa de Interes Bdsica Pasiva) Managers and Staff Responsible Vice Presidendt: Shahid Javed Burki Director: Donna Dowsett-Coirolo Sector Leader: Mark Cackler and Ian Bannon Task Managers: Rodrigo Chaves and Susana Sanchez EL SALVADOR RURAL FINANCE: PERFORMANCE, ISSUES, AND OPTIONS Table of Contents EXCUTIVE SUMMARY: MAIN FINDINGS AND RECOMMENDATIONS ............................................. i CHAPTER 1. INTRODUCTION ........................................................... I BACKGROUND AND OBJECTIVES ........................................................... 1 DATA ............................................................1 ORGANIZATION OF THE REPORT ........................................................... I CHAPTER 2. FINANCIAL SECTOR REFORM, POLICIES, AND INTERMEDIARIES............................2 FINANCIAL SECTOR REFORM FROM THE 1980s To THE I 990S ............................................................ 2 Regulation, Legislation, Supervision ............................................................ 3 Privatization of Financial Institutions ........................................................... 3 Liberalization of Interest Rates ........................................................... 4 Reserve Requirements and Sectioral Allocation of Credit ............................................................4 Debt Refinancing or Restructuring Programs ........................................................... 5 FINANCIAL INTERMEDIARIES AND LENDERS ............................................................ 6 Specialized Financial Institutions ............................................................ 6 Private Commercial Banking ........................................................... 8 Semi-Formal Institutions ........................................................... 8 MARKET SHARES OF FORMAL AND SEMI-FORMAL LENDERS ........................................................ 9 IS AGRICULTURAL CREDIT SCARCE? ....................................................... 9 CHAPTER 3. EL SALVADOR: RURAL CREDIT MARKETS................................................................12 THE RURAL HOUSEHOLDS ....................................................... 12 LENDERS AND TECHNOLOGIES ....................................................... 13 THE RELATIVE IMPORTANCE OF FORMAL, SEMI-FORMAL, AND INFORMAL LENDERS .................................................... 14 CHARACTERISTICS OF LOAN AND DEBT CONTRACTS ....................................................... 15 PERFORMANCE OF RURAL CREDIT MARKETS ....................................................... 18 The Participation of Rural Households in RCMs ....................................................... 18 The Matching of Debtors and Lenders ....................................................... 20 CHAPTER 4. PERFORMANCE OF RURAL FINANCIAL INTERMEDIARIES.........................................21 OUTREACH ........................................................ 22 FINANCIAL PERFORMANCE ........................................................ 23 Performance of BFA ........................................................ 24 Performance of Client-Owned Organizations ........................................................ 28 MAIN ISSUES ......................................................... ................................ 29 CHAPTER 5. THE ENVIRONMENT FOR FINANCIAL TRANSACTIONS, AND NEW FINANCING INSTRUMENTS .................................30 ENVIRONMENT FOR FINANCIAL TRANSACTIONS ............................... 30 NEW F INANCING INSTRUMENTS ............................... 3 1 CHAPTER 6. RECOMMENDAT'IONS ............................. 35 ECONOMIC ENVIRONMENT ................................ 35 DiSTRIBurioN NETWORK ............................. 36 Governance, Regulation and Supervision of BFA, FEDECREDITO and FEDECACES ....................................... 36 Technology Transfer: Private Banks and Financieras .................................... . ,39 BIBLIOGRAPHY ................................................................. 40 Annex A Regulation, Legislation, and Supervision .45 Annex B Financial System and Economic Indicators .48 Annex C Rural Household Survey .52 Annex D Specialized Financial Intertnediaries: Financial Indicators .55 Annex E Debt Refinancing or Restructuring Programs .62 LIST OF TABLES Table 2-1 Summary Statistics of the Financial System for 1989 and 1995 ..................................................................3 Table 2-2 Comparison of CEDEAGRO/CEDEVIV and other Term Deposits ............................................................5 Table 3-1 Annual Household Income (colones) ................................................................. 13 Table 3-2 Debtors versus Non-Debtors .................................................................. 13 Table 3-3 Characteristics of Loan Contracts by Lender: Amount, Maturity, and Interest Rates ................................ 16 Table 34 Debt Contracts by Lenders: Amount and Nominal Interest Rates (% p.a.) ................................................ 17 Table 4-l Outreach Indicators: 1995 ................................................................. 22 Table 4-2 BFA: Government Assistance and Interventions .................................................................. 25 LIST OF FIGURES Figure 2-1 Real Annual Lending and Deposit Rates ................................................................... 4 Figure 2-2 Agriculture Share in GDP, Debt Balances, and Disbursed Credit ..............................................1............. ao Figure 3-1 Loans Received: Market Shares ................................................................. 15 Figure 3-2 Rural Household Debt: Market Shares ................................................................. 15 Figure 3-3 Participation of Rural Households in Credit Markets ...................................... ............................ 19 Figure 3-4 Participation of Rural Entrepreneurs as Debtors ................................................................. 19 Figure 4-1 BFA: Distribution of Loans by Economic Sector, 9/95 ................................................................. 26 LIST OF BOXES Box 2-1 Innovative Finance: Financiera Calpia ................................................................. II Box 4-]L Risk of Financial Intermediation .................................................................. 23 Box 4-2 Similarities and Differences between FEDECREDITO and FEDECACES Systems ................................... 28 Box 4-3 FEDECACES ................................................................. 29 Thi report wa rte ySsn v.Snhe LSP,Cro .Cea.AW.),adRdiCays(CR EXECUTIVE SUMMARY MAIN FINDINGS AND RECOMMENDATIONS 1. This report analyzes trends in agricultural credit and savings mobilization, examines the performance of main rural financial intermediaries, provides a diagnosis of rural household participation in credit markets, and reviews the prospects and constraints for the adoption of improved financial mechanisms. Aggregate trends in the past decade suggests that a substantial decline in profitability of Salvadoran rural sector, especially agriculture, may be at the root of the problems identified in rural financial markets. The steady decline in relative agricultural prices may have made the agricultural sector less creditworthy than other sectors while probably increasing its demand for external liquidity. Sectoral shares in domestic credit have declined faster than sectoral contributions to GDP. 2. After a period of substantial stress and lack of confidence in the financial system, a number of corrective measures were adopted beginning in 1989: privatization and recapitalization of banks and finance companies (financieras); liberalization of interest rates; removal of barriers to entry; and reorganization of the supervisory agency (SSF). While these measures represent significant progress in strengthening the financial system, several shortcomings remain: (i) despite the entry of new institutions, three large banks continue to dominate the banking system; (ii) further improvements are needed in banking supervision, and the legal and regulatory framework for secured transactions; and (iii) various Government decisions have undermined financial discipline, notably debt forgiveness programs (open or disguised). 3 The Government may have worsened the performance of financial markets through its participation in the sector, notably by: (i) excessive administrative costs and delinquency within the public development bank; (ii) strategic defaults promoted by debt forgiveness programs; (iii) interest rate subsidies; and (iv) potential regulatory problems affecting private organizations (e.g., credit cooperatives). Finally, the informal financial sector was found to be quite underdeveloped. The scant supply of informal credit may result from the difficulty of enforcing credit contracts in rural areas. The country's recent conflict and the instability in conflict areas certainly weakened traditional enforcement mechanisms. Agricultural Credit and Savings 4. Agricultural Credit. The share of agriculture in formal sector credit flows has declined significantly over the last five years. Its share in total outstanding debt has also declined, although slightly less (see Chapter 2). Successive periods of low returns should make these units less profitable and less creditworthy and may induce lenders to allocate resources to other sectors. At the same time, this pattern could have increased units' demand for external financing or on the contrary reduced their demand because of lower returns. 5. Private banks provide three-fourths of total formal credit to agriculture, while specialized financial intermediaries (especially BFA., the agriculture development bank) account for one-fourth of the total, in spite of their relative specialization (about one-half of their portfolio in agriculture compared to less than 10% in private banks). In terms of number of clients, however, the agricultural clientele of private banks is limited to less than 2,000 clients, while the agricultural clients of BFA are estimated at about 20,000. Adding the agricultural clients of all banks and financieras, including the BFA, we reach a figure of about 25,000 clients, or less than 10% of all agricultural enterprises. This limited access is confirmed by the survey findings (see below). The existence of important under-served sectors in rural areas may explain, for example, the rapid expansion in agricultural lending by a new innovative provider (Financiera Calpia) which has been able to develop a portfolio of more than 2,000 small-scale farmer clients (micro- and small rural entrepreneurs) in less than 2 years, in addition to its established clientele of about 10,000 urban micro- and small enterprises. 6. Savings Mobilization. Deposit services are primarily provided by banks both in urban and rural areas. Outside San Salvador banks mobilized 32.9% of all savings in 1995. Bank density ratios have improved both in urban and rural areas since 1992, although the proportion of bank branches outside San Salvador has decreased. i ii El Salvador: Rural Finance Performance of Rural Financial Intermediaries 7. The financial analysis of BFA, FEDECREDITO and FEDECACES systems, and Financiera Calpia -- intermediaries active in rural financial markets-- reveals, among other findings, that: * BFA's most significant limitations are: a governance structure that allows for political intervention, and goals tlat are inconsistent with its financial viability, a hybrid role as an instrument of the State held to the rules and standards of commercial banking, a de facto constraint on BFA's loan pricing in an environment of liberalized interest rates, and substantial losses due to high operating costs and poor loan recovery. Reorganization efforts have been inconsistent, closing non-bank activities to then re-open them, and managing a number of governuent programs which essentially burden the bank with unprofitable activities. Debt forgiveness programs mandated by the Government have affected BFA the most, as the ensuing loss of repayment discipline further contaminates its portfolio with bad debt, while its institutional mandate forces it to continue lending to a high-risk borrower class. * Recent improvements in management at FEDECREDITO have not yet brought it up to commercial standards. The credit-union sector affiliated in FEDECACES is also in a process of slow recovery.' Finally, Financiera Calpia, with innovative techniques of service provision, state-of-the-art information management and market pricing has been able to develop a rural portfolio of small-scale clients with high recovery rates. Access to Credit Services in Rural Areas 8. Rural credit markets (RCMs) in El Salvador are shallow because rural households have limited access to credit services-less than 12% of rural households received a loan in 1995 (see Chapter 3). About 20% of rural households had outstanding debt balances from formal or informal sources. Furthernore, rural households do not have access to a variety of lenders, as only 0.3 % reported debt balances with more than one type of lender. The weak participation of rural households in credit markets is explained by both demand- and supply-side factors. On the demand side, 870/o of rural households did not apply for loans in 1995. Although information regarding the weak demand for credit could not be fully assessed, the findings suggest that households find it difficult to use movable goods as collateral, and that access to formal lenders is limited to those rural households with titles for real estate property. Furthermore, the terms and conditions of financial products traditionally available in RCMs may be inadequate for rural entrepreneurs seeking to diversify away from depressed and unstable activities. 9. A combination of factors likely hinder the supply of credit: (i) an underdeveloped institutional infrastructure; (ii) Government interventions which crowd out private lenders by allowing weak public sector institutions to lend with poor recovery and low interest rates; (iii) debt-forgiveness programs which have created serious credibility problems by promoting strategic defaults; (iv) previous interventions in agricultural marketing that have prevented the development of informal financing sources such as crop purchase credits; and (v) the recent conflict and the resulting insecurity in rural areas, which impeded the development of informal markets by weakening the traditional enforcement mechanisms used by informal lenders, and which have not been replaced by formal legal structures or new informal mechanisms. 10. Among the rural households that did report loan transactions, formal sector lenders provided three-fourths of the total amount borrowed, with BFA accounting for 61% of this total (see Chapter 3). These findings are at odds with previous studies, where the bulk of loan amounts was accounted for by informal sources. The share of informal loens in the total number of loan transactions, 49%, is also much lower than that registered in previous surveys, where informal loans represented about 80% of the loan transactions effectively performed by rural households. Il. Loan contracts with formal and informal lenders showed different interest-rate, loan-size and maturity characteristics. Informal sector lenders charged nominal interest rates equivalent to an annual rate of almost Currently, there are proposals to make FEDECACES and FEDECREDIT() formal financial intermediaries under SSF supervision and regulated by a special financial cooperative law. The potential shortcomings of this approach are discussed below. Executive Summary: Main Findings and Recommendations iii 100%, while nominal average rates on formal loans were slightly over 18%.2 High interest rates differentials between formal and informal lenders suggests that albeit at relatively high lending transaction costs compared to urban areas, there is room for provision of viable credit services in rural areas. Main Issues and Recommendations 12. El Salvador has made significant progress since the early 1990s in creating adequate conditions for private sector growth. In particular, the substantive reforms and liberalization of the financial sector have enabled a conducive environment for financial intermediation where institutions can function without the burden of excessive regulation and pricing restrictions. The issues that remain, and the recommendations to address them, are presented here in a hierarchy of significance, under four headings: (i) restructuring of the main rural financial intermediaries (BFA, FEDECREDITO, FEDECACES) and of their governance, regulation, and supervision; (ii) financial sector competitiveness and supervision; (iii) technological transfer; and (iv) modernization of legal framework for securing financial transactions. 13. There is no single proven institutional model for El Salvador to follow in order to deepen its rural financial markets. Any type of institutional re-structuring requires, among other factors, an adoption of cost- effective financial innovations to provide financial services appropriate to rural areas, and measures that enable the use of new financing instnrments. Provided appropriate conditions are present, three non-exclusive institutional models have proven successful in other countries. The recommendations below are based on these three non- exclusive alternatives: * The full privatization of a public bank, or a clearly defined split of a public financial institution, akin to the Unit Desa system of the Bank Rakyat Indonesia (BRI) * Thorough reform in the regulation and legal structure of the credit cooperative system to eliminate the governance problems that have plagued such organizational forms * The down-scaling of operations among bank and non-bank commercial intermediaries, following the path of banks such as Colombia's Caja Social, and non-banks such as El Salvador's own Financiera Calpia. Financial Sector Competitiveness and Supervision 14. To improve financial sector competitiveness the Government should: • Consolidate the restructuring and liberalization of the banking system, especially through: consolidation of ownership restrictions; disclosure of interest rate information; implementation of a rating system for financial institutions; and strengthening of SSF through training and modernization. * Remove from the menu of policy interventions debt refinancing/forgiveness programs as an effort to restore repayment discipline in rural areas. Governance, Regulation and Supervision of Rural Financial Intermediaries: BFA 15. As presented above, the central issue dominating the BFA's operations are governance, political interference and its imposition of special social mandates. Although there is ample room for improving managerial practices and information systems, and for adopting better lending and savings mobilization practices, serious doubts will always remain with respect to the incentives of BFA's management to implement such improvements under the existing government ommership and governance structure. 16. The immediate step should be to eliminate all non-financial operations assigned to BFA, and transfer away from the bank all welfare-related mandates. The ensuing actions would be a rigorous audit to determine the market value of BFA's equity and a detailed analysis of the spectrum of options for its divestiture-including the possibility framework in order to privatize some of BFA's branch network as independent local unit banks. The regulatory implications of this possibility are analyzed below. An implication of the proposed strategy is to 2The average interest rate differentials between formal and informal lender for loans and outstanding debt balances are 81 percentage points and 22 percentage points, respectively. iv El ,alvador: Rural Finance suspend BFA's current efforts and investments in information technology. At this stage it is not known, for example, whether any investments made to integrate the branch network's information systems will be consistent with an optimal strategy for BFA's divestiture (e.g., independent unit banks). 17. Is it possible to refonn BFA to make it a sustainable public development bank? To achieve sustainability, BFA would have to be compelled to adopt general principles that are necessary conditions in the sustainable provision of financial services to small and medium entrepreneurs in rural areas. The political cost of reforming BFA along this way may be fairly close to that of divesting the institution altogether, because any meaningful reform requires rreducing or eliminating the rents extracted by three constituencies: bureaucrats, larger borrowers, and politicians. These groups would probably fight off refonn with the same strength with which they would oppose divestiture. Reform is a gradual, long-tern and reversible process that puts time on the side of these constituencies to organize opposition. In contrast, divestiture may be achieved swiftly and irreversibly. 18. The following are examples of change control actions that could be pursued until the time in which BFA is completely divested to the private sector. These actions are, in any event, required intermediate steps for divestiture:3 * Organizational control: the objective should be to provide control of BFA to individuals who are not subject to political pressures and who are not representative of the bank's borrowers. This objective would require: (i) reforrn of the law and charter of the bank because the Board of Directors is currently integrated by political appointees (Ministers) and representatives of the beneficiaries (farmer groups). In the meantime, the Government's political appointees should vote (in block) for measures and policies aimed at improving the bank's performance (e.g., suspend current investments in modernizing BFA). * Distribution network: BFA branches are characterized by a large fixed-cost structure and are mostly located in highly populated urban centers. The fixed costs of operating branches could be reduced by reducing branch staff, and by selling buildings and reallocating branches in smaller rented offices- preferaebly in locations not served by commercial banks. * Credit policies and targeting: BFA should change its credit policies to increase its outreach to small rural entrepreneurs and, more importantly, to protect itself from rent-seeking organized groups of borrowers. In particular, it should significantly reduce the maximum amount of credit granted to a single borrower (e.g., maximum of 4 times the country's per capita income). It should increase interest rates significantly and finance all types of rural entrepreneurs (not only farmers). This policy, combined with collection incentives on the part of branch managers (mostly the threat of closing problem branches), would reduce the amount of rents an individual borrower may extract and, hence, would also reduce the incentives to establish and organize opposition to the reforms. * Credible sink-or-swim strategy: the Government should make credible the fact that the resources currently at the disposal of the organization represent a once-in-a-lifetime capital endowment. It should be made clear to these groups that branches that are not sustainable will be closed within 6 months. FINANCIAL COOPERATIVES (FEDECREDITO AND FEDECACES): 19. Financial Cooperatives (FEDECACES and FEDECREDITO). The Govemment is considering unification of the regulation, governance, charter and status of the savings and credit cooperatives (cooperativas de ahorro y crddito), most of them affiliated to FEDECACES, and the cajas de credito affiliated to FEDECREDITO. Such unification would imply the emergence of a single organizational form for savings and credit cooperatives (SCCs), and the subjection of the resulting unified systems to SSF supervision, which may include deposit insurance. Moreover, the Government may chose to include this regulatory reform in a broader body of legislation to cover non-bank financial intermediaries, with or without cooperative ownership. Although 3A complete package of damage control actions (including quantitative goals) and divestiture steps should result from the proposed detailed analysis of BFA's operations and financial situation. Such analysis should be completed as soon as possible. Executive Summary: Main Findings and Recommendations v there may be clear advantages to unifying the regulatory and supervisory framework for SCCs, the proposed reforms need to take into account the potential risks involved in chartering these new intermediaries. 20. The main governance features of the organization that would result from the proposed new structure are: ownership by both depositors and borrowers, one-member-one-vote rule for decision making by general assemblies, and voluntary association and withdrawal of members. Potentially negative implications of these governance features on the capital structure of the organization need to be explicitly regulated upon by the new legislation. Specifically, minimum capital requirements would need to be complemented with rules governing the withdrawal of share capital, and the accumulation of institutional capital. 21. International experience with credit cooperatives. The international experience of financial cooperatives is rather mixecd in terms of perfornance and viability. Supply-led cooperatives such as the uniones de credito in Mexico have resulted in financial collapse, and other networks such as the Costa Rica SCCs have required govermment bail-out after short-lived success. On the other hand, successful experiences of credit-union reform and modernization have been documented in Guatemala and the Dominican Republic, aside from the African examples of Benin, Cameroon and Togo.4 22. Fiscal liabilities. The Government may assume contingent liabilities from the reforms being proposed, as it does by chartering any type of financial intermediary. Even if explicit deposit insurance is not provided, the Government may be unable to avoid bailing-out SCCs in case of financial difficulty as they would be under SSF supervision and their depositors would be numerous and unsophisticated. Protecting large numbers of unsophisticated depositors, on the other hand, ought not to deter Govermment from diversifying the financial system, especially when the amounts involved represent a minimal share of total deposits in the financial system. 23. The new framework. In the new framework, the legal ability to carry out specific financial intermediation activities needs to be directly contingent upon the fulfillment of requirements and conditions set forth by the new body of legislation. Most importantly, minimum capital requirements for entering the regulated system and rules for the accumulation of institutional (non-withdrawable) capital should ensure the rapid build-up of a stable capital base. In addition, risk-weighted capital adequacy rules should at a minimum be equivalent to those required from other deposit-taking institutions. 24. Specific rules preventing "borrower-domination" (extraction of benefits by coalitions of members at the expense of the organization's financial viability) should be made part of the framework (e.g., prohibiting dividend allocation on the basis of borrower-patronage) to reinforce internal rules modern SCC organizations typically follow. 25. The new framework should establish reserves and allowances created from earnings to maintain capital levels adequate to cover potential losses from loans and investments. The legislation should also require provisions for loan losses, guided by portfolio classification rules similar to those applied to other regulated institutions. Likewise, norms on portfolio concentration and insider lending, and reserve requirements (encaje) comparable to those of other deposit-taking institutions should apply to the new institutional charter. 26. In the case of the current affiliates of FEDECACES and FEDECREDITO, minimum capital requirements will have obvious consequences on the magnitude of the consolidation process that will likely follow the regulatory reforms. From about 36 and 50 affiliates, respectively, the networks are likely to dwindle in number of affiliates to about a dozen each, primarily through mergers (much like the consolidation observed in the Mexico cajas populares after the 1992 reforms). Hence, the total amount of deposits now falling under SSF supervision may initially decrease, to then recover as the newly regulated institutions gain public confidence. 4 In Africa, SCCs were not exposecl to the donor-induced subsidized-credit trajectory most SCCs networks in Latin America experienced in the 60s and 70s. vi El Salvador: Rural Finance 27. Prudential supervision mechanisms. Building SSF capacity and establishing prudential supervision mechanisms need to accompany any regulatory reforms implemented by the Government. Regulatory functions should not be allocated to the federations, as it is unlikely that it would be possible for them to credibly supervise their affiliate-owners. Instead, mechanisms that supplement the SSF capacity with private external audits and examination of the SCCs should be established, to function under the guidance of a small high-quality team of SSF specialized supervisors. 28. A transitional period to adopt and adjust to the new legislation will be required. In addition, the Government should carefully consider the future of existing financial cooperatives and other credit-granting organizations that will not meet the new requirements. Specifically, access to BMI funds by non-regulated intermediaries should be strictly ruled by the creditor as no explicit or implied guarantee will apply to those entities. Increasing Access through Technology Transfer and Modernization 29. Technology transfer. The down-scaling of operations among bank and non-bank commercial intermediaries can be achieved through research and development of new financial technologies, and the promotion of their adoption. The Government should familiarize private financial intermediaries, as well as the BCR and SSF officials, with successful technologies in lending and deposit mobilization in rural areas. For example, private banks in Indonesia operate successfully in rural finance. Furthermore, the Government should design incentive packages to encourage the adoption of these innovative financial techniques and methods. In Mexico, the WVorld Bank is currently sponsoring a technical assistance program to provide incentives for private banks to offer financial services to rural communities. 30. New financing instruments. Because the Government has done little to promote the use of new financing instruments, the general guidelines proposed in the 1995 Country Economic Memorandum (CEM) would be appropriated. The main recommendations to expand use of new financing instruments are summarized below: * Improve the environment and legal framework for new financing instnuments such as leasing and factoring. * Consolidate the reforms currently in progress, and the study of new ones and amendments to the Financial Institutions Law that clarify the rules that govern leasing should be adopted. * Set guidelines and uniform treatment for leasing, factoring and discounting of commercial paper. * Develop an enabling framework for asset-backed securities, reducing notary costs and eliminating mulliple taxes. 31. Colllateral security issues. The Government could increase access to credit services by reducing the costs of using collateral to secure credit transactions. To reduce the costs of collateral use: (i) improve legal framework to facilitate the creation, perfection, and enforcement of security interests; and (ii) modernize registries to reduce transaction costs and delays of registering security interests such as mortgages and liens. Furthermore, consider CEM recommendations to expand use of liens on inventory and commercial equipment. Chapter 1. INTRODUCTION BACKGROUND AND OBJECTIVES 1.1 The study addresses the question of access to finance by rural households, and provides a diagnosis of the performance and overall competitiveness of rural financial markets. It also examines the prospects and constraints for improved and innovative financial mechanisms. 1.2 The importance of moire effective financial intermediation has been highlighted in previous Bank documents in connection with, among other issues, the lack of opportunities for financial savings. This, along with other factors, explains the continued low domestic savings levels observed in spite of an increase in national savings fueled by worker remittances.5 Since a more than proportionate share of migrants originated in the rural sector, the need for increased depth and efficiency is even more pronounced in rural financial markets than elsewhere.6 1.3 Although El Salvador has long abandoned administered credit allocation policies, the "project-oriented, collateral-based" traditional lending practices of commercial banks, as opposed to the client-oriented performance- based untargeted modem methods, has tended to increase the concentration of commercial credit in an elite of relatively successfil, and large, farmer and agribusiness clients. 1.4 Four main issues are typically associated with efficient financial intermediation, namely: (a) the macroeconomic environment, (b) sectoral policy context, (c) financial market constraints, and (d) legal and regulatory barriers.7 The study focuses on the last two of these problems. We keep in mind and remind the reader, however, that ultimately the key shortcomings in sectoral growth and competitiveness may not be found in the fmancial market but in sectoral or macro policies. Rural finance does have a role to play, nevertheless, and large room for improvement exists foir rural finance to help the sector adjust to the challenge of globalization. DATA 1.5 The report relies on secondary and primary sources of data. The secondary sources come from reports and bulletins from the BCR and the SSF, on published and unpublished documents prepared by FUSADES, and on recent studies carried out primarily under the auspices of USAID and IDB. In addition, numerous and extensive interviews were conducted at public and private institutions and organizations during a Bank mission in May 1996. The primary source of data is the 1996 FUSADES survey of rural households. This survey contains information about several aspecis of the productive activities of 628 households from the rural areas of El Salvador. ORGANIZATION OF THE REPORT 1.6 The report comprises six chapters and five annexes. Chapter 2 reviews the financial sector reforms and policies of the 1980s and 1990s, highlighting the areas in which further adjustments seem necessary. It also analyzes the extent to which the rural sector is adequately served with financial services. Sectoral patterns of financial services provision, and the institutional structure of the financial sector are critically reviewed. The question of access to financial services, especially credit, by rural households is addressed in Chapter 3, on the basis of the FUSADES survey. Rural household participation in financial transactions, and the factors that influence this access are analyzed there. The performance of three major rural financial intermediaries (BFA, 'World Bank (1995). 6 It has been estimated that 25% of the rural households in El Salvador receive remittances from abroad, accounting on average for as much as 11% of their income (L6pez, 1996) 7Yaron and Piprek (1997). 2 El Salvador: FEDECREDITO, IFEDECACES) relative to that of commercial banks is analyzed in Chapter 4. The environment for financial transactions and new financing instruments, and the potential for these new instruments to be effective in rural aueas are discussed in Chapter 5. Key conclusions and recommendations to develop financial markets are offerecd in the final chapter. Chapter 2. FINANCIAL SECTOR REFORM, POLICIES, AND INTERMEDIARIES 2.1 During the 1980s, the financial system El Salvador experienced a loss of public confidence due to unfavorable investment conditions, market inefficiencies, and delinquency problems. The problems of delinquent debt originated from the nationalization of the banking system; the conflict; inadequate supervision of banks, and concentrated loan portfolios. The portfolio of problem loans of the nationalized financial system mounted during the 1980s as more producers were unable or unwilling to meet debt service requirements. At the end of 1989, banks were left with inadequate capital positions. The threat to the stability and viability of the financial system and the deterioration of the loan portfolio forced the Government to implement drastic measures, starting in 1989, including but not limited to the privatization of banks and financieras and the adoption of major changes in the regulation, legislation and supervision of the financial system. 2.2 This chapter presents the main elements and the effects of the financial sector reforms of the 1990s. While significant progress has been made toward strengthening the financial system in recent years, the following shortcomings call for attention by policy makers: the banking system remains concentrated in three banks; improvements are still needed in the supervisory and regulatory framework; and debt refinancing programs have created negative effects for private lenders and induced strategic default behavior by borrowers. The chapter then analyzes the institutional structure of the financial sector. Brief profiles of the institutions active in rural finance are presented here. More extensive analyses of BFA, FEDECREDITO and FEDECACES are carried out in Chapter 4. 2.3 Next, the key question addressed in this chapter is whether the agricultural sector is under-served with lending services vis a vis its inherent profitability and potential, and relative to the access to financial services in other sectors of tlhe economy. The findings reported here indicate that the share of agriculture in formal sector credit flows has declined significantly over the last five years, while the sector's share of total outstanding debt has not declined as much. Sectoral credit participation seems to reflect the relatively poor sectoral performance associated with its deteriorating terms of trade, rather than serious constraints or barriers in the banking sector. However, Financilera Calpia which started in 1995 with agricultural lending has rapidly developed a portfolio of more than 2,000 farmer clients with loan sizes averaging less than 0600 in less than 2 years. FINANCIAL SECTOR REFORM FROM THE 1980s To THE 1990s 2.4 By the end of 1989, after 10 years of a nationalized financial system and an environment of financial repression, the financial system (banks, financieras, Banco Hipotecario and SFIs) had so many delinquent loans that it was technically bankrupt. Delinquent loans amounted to %3,861 million or 37.4% of'the loan portfolio and the negatiye equity of these institutions was 02,212 million or 26% of deposits (see Table 2-1). 'The insolvency of the banking system prompted Government intervention in order to restore public confidence in the system, and avoid a bank run that would have compounded the country's already critical situation in a Ieriod of civil-war and economic instabillity. 2.5 The nationalized system was used to finance the Government's budget deficit, priority segments of the private sector, especially export agriculture, and agrarian reform beneficiaries (Danby, 1995). The Government's reluctance to foreclose loans to these favored groups lead to the practice of refinancing loans on concessional terms that made it advantageous to delay repaying. During 1987-91 an average of 26% of the outstanding loan Rural Finance 3 balances of commercial banks were in the refinanced category (Abt, 1993). In addition to refinancing practices, banks lacked crucial information on financial health and indebtedness of clients, and corresponding registry of the real-estate collateral of the loams were signs of a deficient information management system. 2.6 To avoid the potential collapse of the financial system the Government started a process of financial liberalization. Since 1989, this process has restored public confidence in the system, supported the efficient performance of financial institutions and promoted economic development. The main elements of this process included: reforming the regulatory framework for financial institutions and for the SSF; privatizing commercial banks and Savings & Loan Associations; liquidating three state-owned banks; establishing a special fund to purchase the worst performing loans from commercial banks (FOSAFFI); liberalizing interest rates; eliminating directed credit programs; and creating uniform reserve requirements across financial intermediaries. As a result of these reforms, at the end of 1995, lending and deposit rates were positive in real terms, privatized banks and fmancieras had 06,066 million of net capital, and the number of financial institutions had almost doubled with respect to 1989 (Table 2-1). Table 2-1 Summary Statistics of the Financial System for 1989 and 1995. Indicator 1989 1995 Quasi4noney/GDP' 16.0 27.3 Real Interest Rates (% per annum) Lending 0.7 8.2 Deposit -1.2 3.9 Banks and Financieras b No. of Branches 230 270 No. of Institutions 11 21 No. of Employees 7,681 12,635 Capital of privatized Banks and Financieras -¢2,212 million 06,066 million % of loans in category 0 and E 36.7 % 3.4% H-H Index for Assetsd 1,690 1,569 Sources: * Intemational Financial Statistics, various issues. BCR (1996, Cuadro 7) and SSF Bulletin, 1995. Includes: Agricola Comercial, De Comercio, Cuscatlin, Desarrollo, Salvadorehio, Hipotecario, Ahorromet, Atlacati, Casa, and Credisa. c BCR (1996. Cuadro 10). d SSF Bulletin, 1995, and BCR Bulletin, 1991 p. 32 and p.48. The H-H index is equal to the sum of squares of the market share of assets of each firm. Regulation, Legislation, Supervision 2.7 In the early 1990s, the Government passed a number of laws that partially corrected the regulatory framework for financial institutions. However, the enhancement of certain aspects of the regulatory framework is needed to fully consolidate the reform. Strengthening prudential supervision, and rules for new financing instruments are areas for new reforms of particular importance. Improvements are still needed in the rules that govern non-lending financial mechanisms such as leasing, factoring, and warehouse receipts that do not appear to be fully covered in the upcoming legislation, but the effect of adequate regulation may be significant (Chapter 5). 2.8 The modification of the banking legislation and regulation of 1996 represented an important and decisive step in improving the soundness of the financial system. These modifications resulted in a tightening of capital adequacy requirements from 8% to 10%; reduced loans to insiders to 15% of stock and capital reserves; and the establishment of a system for information sharing. A detailed account of current regulations covering entry, ownership, capital requirements, portfolio concentration and inside lending, supervision, loan risk classification and information sharing is presented in Annex A. Privatization of Financial Institutions 2.9 Based on a privatization law passed in November 1990, the Government privatized five banks and four financieras from 1991 to 1994. Prior to the privatization process the Government began to strengthen commercial banks and financieras to attract private investors by exchanging bad loans with Government bonds issued by FOSAFFI and by direct capital injections. Investor bought a total of , 1,143 million of stock, out of which 0 135.4 corresponded to employees. Even though the privatization process limited equity participation to 5% for single investors, this requirement was bypassed, as employees and small investors bought stock for large investors. 4 El Salvador: 2.10 The privatization of banks and the removal of barriers to entry into the banking system have had a minimal effect on the market structure of the system.8 Although the number of financial institutions has almost doubled, new entrants held only 9% of total system assets in 1995. In fact, the H-H index for assets, an indicator for industry concentration,9 decreased slightly from 1,690 to 1,569 between 1991 and 1995. Assets were concentrated in three banks (Banco Agnicola, Salvadorenlo, and Cuscatlan) which in 1995 had the same share of total banking assels as in 1991 (61%). Liberalization of Interest Rates 10 ___ 2.11 Beginning November 1990, BCR liberalized interest rates. As s ------------------------------------------------------ .shown in Figure 2-1, this measure resulted in positive real deposit rates o - \ ' ' E | > 1 M / r / | which apparently attracted a large 5 ' S \-o co C: Y 3 number of depositors. The volume of quasi-money (time deposits, current -10 ----------------- - ~ ----------------------------- ---------------------------savings, and foreign currency deposits) as percentage of GDP has increased -15 reflecting public confidence in the -- Deposit - Lending financial system. In 1994, quasi-money and M2 reached levels of 29.9 and Source: International Financial Statistics 40.5% of GDP, respectively, the Figure 2-1 Real Annual Lending and Deposit Rates highest levels since 1980. Finally, the liberalization of interest rates together with other measures have created conditions which have attracted new financial intermediaries into the market through increased profitability of financial activities. 2.12 Because banks must now pay higher interest rates for their deposits, they will need to charge higher lending rates in order to remnain profitable. However, higher lending rates may have an adverse selection effect as low credit risk borrowers exit the market and leave banks with riskier loan portfolios. Danby (1995) argues that recent concentration of loan portfolios in the construction sector and high levels of excess liquidity suggest difficulties findinlg good clients. Reserve Requirements and Sectoral Allocation of Credit 2.13 The BCR currently uses selective reductions in reserve requirements as an instrument to stimulate long- term sources of funds for the financial sector and channel funds to housing and agriculture. The funds raised with these savings instruments have a reserve requirement of 15%, and all the reserves are remunerated. These savings instruments are CEDEVIV which targets housing and CEDEAGRO (introduced in 1996) which targets agricultural production. 2.14 As incentive instruments, CEDEVIV and CEDEAGRO give banks and financieras substantial flexibility since their participation is voluntary and there is no restriction on the lending rate of those funds. Banks and financieras may be willing to offer CEDEVIV or CEDEAGRO deposits after considering costs of funds, liquidity See BCR (1990) for a description of the privatization process. 9 This index is equal to the sum of the squared market share of assets of each firm. Higher values of this index imply a greater concentration of the industry. For reference, if all 13 banks had equal market shares, the H-H would be 769. Rural Finance 5 management, the interest rate elasticity of the demand for savings, and the demand for credit from eligible housing and agricultural activities.'° 2.15 During 1992-1995 depositors showed an increasing preference for term deposits. For example, CEDEVIV increased its share from 2.8% to 11% of savings balances and reported real growth rates in excess of 50% (Table B-5). The aggressiveness of banks in promoting this long-term deposit instrument is matched by a strong allocation of credit to housing construction. 2.16 CEDEVIV has been very popular in the department of San Salvador. However it is unlikely that CEDEAGRO will be able to replicate Table 2-2 Comparison of CEDEAGRO/CEDEVIV and other Term CEDEVIV's success in mobilizing funds Deposits for various reasons. First, the instrument Concept Cost of Funds Deposit Rate assumes that banks do not lend to CEDEAGRO (two years) 16.92% 15.50% OneDyear TD 15.19% 12.50% agriculture due to a lack of funds and high Sxmonths TD 16.44% 13.50% deposit rates. Second, since most 30idays TD 14.56% 12.00% agricultural lending for working capital is 30 days TD 14.56% 12.00% ~~~~~short-term, the instrument forces banks to Source: Calculations of the authors. The deposit rates were taken from the publication of Banco Cuscatlin on April 1996. lt was assumed that the TIBP have a maturity mismatch. Third, an equals 11.71% and that CEDEAGRO will be offered at CEDEVIV's rate. important weakness of CEDEAGRO is that land purchases are not included as a possible use of the funds. Hence, the narrow targeting of these funds to selected agricultural activities may deter conservative banks from using this instrument as extensively as would be desirable. Finally, after adjusling for differences in reserve requirements and remuneration of these reserves, CEDEAGRO has the highest cost of funds among term deposits (Table 2-2). Debt Refinancing or Restructuring Programs 2.17 Refinancing delinquent debt raises two concerns. First, if the restructuring of the debt reduces the net present value of the original contract, then borrowers in good standing will have incentives to stop paying their loans because they face less favorable terms than delinquent borrowers. Second, these programs may encourage economy wide strategic default and create a culture of no repayment as borrowers simply default and wait for the next refinancing program. Despite the potential negative effects, in 1992 the Government started a debt forgiveness program to compensate producers affected by the conflict and in 1996 it passed two new refinancing programs to promote agricultural production. A sunmary of these three debt forgiveness programs are explained below; a more detailed explanaition is presented in Annex E. 2.18 War-related Debt Reifinancing Program. The Ley para la Rehabilitaci6n de Sectores Productivos Directanente Afectados por el Conflicto (D.L. 292) attempts to renegotiate, on concessional terms, the delinquent debt of producers directly affected by the conflict. The Decree, approved in July 1992, had two main objectives: refinance the debt to allow borTowers to pay back; and restore the creditworthiness of delinquent borrowers to increase their access to credit. However, a restrictive definition of "those directly affected by the conflict"; high transaction costs for potential beneficiaries, and conservative banking criteria applied to each specific case at the level of branch bank agencies have limited the use of the Decree by potential beneficiaries. In order to overcome some of these limitations, the D.L. 292 has had two major modifications. In February 1994, the Decree recalculated back interest at 6% for beneficiaries whose original debt was below lt100,000 (Decree 814). Furthermore, banks were not allowed to charge commission or fees for processing applications. In August 1994, the debt of beneficiaries with original loans below 05,000 from BFA or FEDECREDITO was forgiven (Decree 114). 10 CEDEAGRO could finance: (a) working capital loans for agriculture with the exception of coffee and sugar cane; (b) agricultural investments as delimited by the BMI; and (c) agribusiness investments, excluding investments on sugar mills or beneficios de cafe. On the other hand, CEDEVIV finances: (a) acquisition and construction of housing by low and medium income families; and (b) acquisition of long-term mortgage loans that were originally financed with CEDEVIV funds. 6 El Salvador: 2.19 As of April 1996, 17,753 applications had been received for a total of s1,609.1 million. The total amount of interest reduced is estimated at ¢413.7 million. Although BFA received the largest number of applications (60% of applications), they accounted for only 23% of total funds requested, while private banks and financieras received applications for 70% of this total. This reflects the different clientele served by these intermediaries. 2.20 Debt Forgiveness for Agriculture. In June 1996, the Government issued two debt forgiveness programs which target the agricultural and agrarian debt in an effort to stimulate agricultural production. Since 1989, the agricultural sector of El Salvador has been greatly affected by the policy reforms introduced. As a result of these reforms and the end of the conflict, rural and agricultural producers have been forced to adjust their economic activities. 2.21 The sources of delinquency in agriculture are related to: difficulties created by conflict; the effects of droughts; rent-seeking behavior induced by Government intervention in credit markets; the impact of a strong currency on agricultural profitability; and the diversion of credit funds to unproductive activities (Abt. 1993). Indeed, many authors have concluded that the environment created by the conflict was the most important cause of agricultural delinquency (Abt., 1993; Fusades, 1994; MAG, 1995). The war affected agricultural production through various channels such as payment of war taxes, invasions and kidnappings. Uncertainties caused by the conflict may have forced farmers to operate below efficient production levels causing low returns. 2.22 The agrarian reform process originated the so-called deuda agraria (agrarian debt). As reported by MAG (1995) the agrarian debt resulted from: the low productivity of agrarian reform beneficiaries due to poor land quality and inexperience; lack of a recovery strategy by the Government; the conflict; and lack of working capital. 2.23 Ley de Apoyo a la Reactivaci6n del Sector Agropecuario (D.L. 698). The Law targets all agricultural producers, but agricultural producers directly affected by the conflict (those that qualify under D.L. 292) received special treatment. It is estimated that a total of l,152.7 million will be forgiven of which $486.1 million correspond to agricultural loans that qualified under D.L. 292. 2.24 Ley de Restructuraci6n de la Deuda Agraria (D.L. 699). The Law allows for 60% discount for prompt payment of the original agrarian debt, the so-called adelantos ISTA-BFA, and FFRAP debt. The beneficiaries are agricultural cooperatives, communal associations, and beneficiaries of Decrees 842, 207, 839 and 713 of the agrarian reform. The D.L. 699 resembles the failure of the agrarian reform process and the inadequacies of the land market in El Salvador. It has negative consequences on the repayment culture because it penalizes those borrowers that were in good standing at the time of the Decree while loan defaulters are rewarded. FINANCIAL INTERMEDIARIES AND LENDERS 2.25 El Salvador has a diversity of formal and semi-formal institutions providing financial services. These institutions can be classified into six categories: specialized financial institutions, second-tier bank, private banks, fmancieras, credit cooperatives, and non-governmental organizations. Specialized Financial Institutions 2.26 The Government's key rural finance institutions are: specialized financial institutions such as the Agricultural Development Bank (BFA), the Federation of Credit Banks (FEDECREDITO) and the Small Investment Guarantee Fund (FIGAPE); specialized guarantee funds such as the Agricultural Guarantee Fund (FOGARA) and the Guarantee Fund for Small Enterprises (FOGAPE); the Multisectoral Investment Bank (BMI); and the Agricultural Trust Fund (FEDA). The BFA, FEDECREDITO and FIGAPE target rural producers and microenterprises and their loan portfolio accounted for 4.8% of total loans of formal and semi-formal lenders (Table B-7). Rural Finance 7 2.27 Banco de Fomento Agropecuario. The BFA was created in 1973 as an autonomous, public sector specialized financial institution after reorganization of the Peasant Welfare Administration (Administracion Bienestar Campesino), which operated a credit program since 1962. The bank has operated under the same regulations and supervision as private commercial banks, although it has clearly different objectives. The BFA has a network of branches and semi-agencies with the most substantial rural presence in El Salvador. It has a total of 27 branches -located mainly outside San Salvador- and 15 semi-agencies which mainly offer deposits facilities. 2.28 As presented in Chapter 4, the BFA has many problems that undermine its financial performance: a governance structure that allows for political intervention, and goals that are inconsistent with its financial viability, a hybrid role as an instrument of the State held to the rules and standards of commercial banking, a de facto constraint on BFA's loan pricing in an environment of liberalized interest rates, and substantial losses due to high operating costs and poor loan recovery. Reorganization efforts have been inconsistent, closing non-bank activities to then re-open them, and managing a number of government programs which essentially burden the bank with unprofitable activities. Debt forgiveness programs mandated by govermment have affected BFA the most, as the ensuing loss of repayment discipline further contaminates its portfolio with bad debt, while its institutional mandate forces it to continue lending to a high-risk borrower class. Finally, the bank is overstaffed and 30% of its 1,463 employees are located at headquarters. 2.29 Federacion de Caias de, Credito (FEDECREDITO). FEDECREDITO is a federation of cajas de credito (rural credit funds) and bancos de trabajadores (workers' banks) which started operation in 1943 as an autonomous specialized financial institution. FEDECREDITO was founded under the Rural Credit Law which was revised in 1991 (Law of Credit Funds and Workers' Banks). As of December 1995, FEDECREDITO's assets amounted to $401 million, and had 242 employees. 2.30 Although it is an autonomous institution owned by the members of the credit funds and workers banks, FEDECREDITO has historically been subject to strong governnent intervention. The rural credit funds and the workers' banks are borrower-ovmed organizations that provide loans only to their members. These loans finance a range of expenditures such as purchases of vehicles and livestock, merchant- and trade-related expenditures, as well as famnily expenses (i.e., medical). The loans are provided through special credit programs: agricultural credit; credito popular; credito de hienestar familiar; credit for microenterprises; and special lines of credit such as FEDA and FOCAM (El Diario de Hoy, 5/15/96). 2.31 Loan Guarantee Programs. Although loan guarantee programs have not proven effective in improving credit access and to support small enterprises, the Government has established two guarantee schemes to support these enterprises: the Agricultural Guarantee Fund (FOGARA) and the Guarantee Fund for Small Enterprises (FOGAPE), which emerged from the Peace Accords. In May 1990, BCR created FOGARA with an initial capital of 100 million to provide loan guarantees for small farmers and agrarian reform beneficiaries. This fund is administered by Banco Cuscatalin, a private commercial bank. FOGARA guarantees up to 70% of the loan amount and the borrower must pay 2% of the covered amount. 2.32 Banco Multisectorial de Inversiones (BMI). The Banco Multisectorial de Inversiones (BMI) is a public second-tier bank founded in 1994 to relieve the BCR from the management of multiple credit lines and allow it to focus on monetary policy. Its main goals are to help increase the amount of resources available to the private sector for medium- and long-term investments, to promote economic growth, and to increase the competitiveness of micro- and small-enterprises among others. As of December 1995, the BMI had total assets of 03,312 million and outstanding loan balances of '2,916 million which represent a 22% increase over December 1994. 2.33 FEDA (Fideicomiso Especial de Desarrollo Axropecuario). The Government recently established a trust fund for agricultural activities (FEDA) with an initial assignment of 0 100 million, managed by BMI. As of April 1996, this trust fund had an outstanding balance of 021.4 million. Since the start of operation a total of p24.3 million has been granted. The objective of FEDA is to promote modernization of the agricultural sector through long-term financing to increase production, employment and the competitiveness of the sector. FEDA offers credit 8 El Salvador: for livestock; aquaculture and fisheries; agricultural development; conservation, distribution and marketing of agricultural products and their industrial derivatives; and non-traditional agro-industrial enterprises. Private Commercial Banking 2.34 The privalte commercial banking sector includes commercial banks andfinancieras which are not allowed to issue checking accounts and offer credit transactions related to international trade. At the end of 1995, El Salvador had licensed 12 private banks and 8 financieras with a total of 178 and 74 branches respectively. Although the performance of private banks andfinancieras is beyond the scope of this report, an issue of concern is the growth of the loan portfolio. As explained earlier, during 1991-1995 the real loan portfolio of banks grew very rapidly at an average of 30% per year. This could be a sign of potential credit risk problems. For example, the outstanding real portfolio going to the construction sector grew at a rate of 66% per year during the same period. 2.35 Banks and financieras have an increasing urban focus as their branches are concentrated in the San Salvador Department with little presence in the rest of the country. In 1995, 57% percent of private bank branches and 59% of financiera branches are located in the capital. In San Salvador, banks and financieras served on average 15,121 people per branch, while the rest of the country they served 37,180 people per branch. The geographical coverage of banks and financieras outside the San Salvador implies a greater distance between lenders and their clients that is translated into higher transaction costs for users that may induce potential clients (depositors and borrowers) to withdraw out of the market. 2.36 This urban focus impacts the geographical distribution of savings. Term deposit and passbook deposit accounts and balances are concentrated in San Salvador, even though this department has less than one-third of the total population (Table B-6). In 1995, 62.8% of savings balances of these two instruments were mobilized in the department of San Salvador and the remaining 37.2% outside San Salvador." This gap in savings was greater 3 years before, with the urban areas providing 65.5% of savings and rural areas 34.5%. Semi-Formal Institutions 2.37 FEDECACES. The Federation of Savings and Credit Cooperatives (FEDECACES) is a legally constituted organization of savings and credit cooperatives (cooperativas de ahorro y credito) founded June 1966. It operates as a second-tier organization that provides financial services to affiliated and non-affiliated savings and credit cooperatives (FEDECACES, 1995). 2.38 A total of 352 credit unions are registered win INSAFOCOOP, of which 150 are active. Thirty-eight of these credit unions are affiliated to FEDECACES, with more than 38,000 members. The savings and credit cooperatives occujpy a prominent place in the non-regulated semi-formal financial sector. Members of these cooperatives are primarily microentrepreneurs, mostly men, engaged non-agricultural activities, although 30% of the affiliates are in rural areas. 2.39 Non-Governmental Organizations (NGOs). NGOs are heterogeneous in terms of the services offered and of their objectives. The principal NGOs established El Salvador include: the Centro de Apoyo a la Microempresa (CAM), the Programa para la Pequenia y Microempresa (Program for the Promotion of Small and Micro- enterprises (PROPEMI), Fundaci6n Salvadorenia de Apoyo Integral (FUSAI), the Integrated Community Foundation of El Salvador (FINCA/EL SALVADOR), and the Segundo Montes and Vecindad Community Bank (BANCOMO). 2.40 In terms of volume of loans, the NGOs sector is relatively small. The loan portfolio of CAM, FUSAI, and PROPEMI accounted for 0.3% of the loan portfolio of the system. NGOs concentrate in credit services, and are not legally allowed to mobilize savings from the general public. In 1992, the UNDP carried out an inventory of " These figures include BFA deposits which accounted for 1.2% of deposits in 1995. Rural Finance 9 NGOs in El Salvador. It was found that out of 186 NGOs, 49 NGOs provide credit services and 5 of them are completely specialized in credit services. MARKET SHARES OP FORMAL AND SEMI-FORMAL LENDERS 2.41 In 1995, the market shares of formal and semi-formal lenders reveals that private banks and financieras are the main players in financial markets in El Salvador (see Table B-7). However, SEIs account for an important share of agricultural loans and of equity. The discussion of agricultural lending is presented in the next section. 2.42 The most relevant featuires shown in Table B-7 are: * In terms of assets, the cooperatives affiliated to FEDECACES and FEDECREDITO had almost the same share of total assets of the system, 0.9 and 0.8%, respectively. * In terms of equity, SFIs have a disproportional share of 16.5% mainly due to BFA having 14.8%. As discussed in Chapter 4, the BFA has received large capital injections. * In terms of deposits, the two systems of credit cooperatives (FEDECACES and FEDECREDITO) have a marginal participation in savings mobilization, 0.7% of the savings balance in 1995. In terms of volume, SFis played a proportionally very small role in savings mobilization given their size and particularly given their country-wide presence. Commercial banks have increased their share in savings mobilization. Of the three savings instruments used in El Salvador, Banks accounted for 81.3% of their balance in 1995, as compared to 76.8% in 1992. These savings instruments are passbook deposits accounts, term deposits, and Housing Certificates (CEDEVIV). The increasing participation of commercial banks comes from their aggressive promotion of CEDEVIV and from the transformation of some financieras into banks. - In terms of branches, the SEIs have a wide presence in the country as its branches accounted for 26.5%, with FEDECREDITO having 17.3% and BFA 6.9%. Private banks have thus been able to capture a larger share of savings volume relative to the size of their branch network. Part of this phenomenon is, of course, explained by the smaller size of demanid deposits and savings accounts offered by BFA branches -mainly located in rural areas- as opposed to private commercial banks which operate mainly in urban areas. IS AGRICULTURAL CREDIT SCARCE? 2.43 The agricultural sector has lagged behind other sectors in the Salvadoran economy in terms of its GDP share and its credit share. After accounting for more than one-fourth of total GDP 1980, its share declined to 16.5% of GDP in 1991, and continued steadily decreasing to 13.7% in 1995 (Figure 2-2). Various factors have induced the slowdown of agriculture. A striking decline in sectoral terms of trade and in apparent sectoral profitability; the large incidence of export crops in overall agricultural output; the stagnation in international prices for these crops; and a fixed exchange rate policy are among these factors. 2.44 Despite govermnent efforts to accelerate agriculture growth with credit, the banking sector has reduced agricultural lending relative to olher sectors, resulting in a decline of the share of agriculture in total banking credit even more pronounced than its fall in GDP share. The increase in real lending rates has forced investors to search for projects with better returns, amd to face higher financial charges. From this perspective, the sectoral allocation of credit has moved away from the agricultural sector which has lower returns, to manufacturing and commercial projects which can better absorb higher interest payments. 2.45 The magnitude of the decline of share of agriculture in total banking credit can be measured with two indicators: outstanding loan balances (stock) and disbursed credit (flow). Real credit of commercial banks to all sectors has expanded rapidly in recent years. The outstanding loan balance of the commercial banking system grew on average at 31 % per year in real terms, from 06,460 million June 1991 to 0 18,978 million in June 1995 (colones of 1990). However, the agricultural real loan balances expanded at 22% per year in that same period (Table B-2). As a result, the share of agriculture the stock of bank credit diminished from 14.3% in 1991 to 10.6% in 1995, thus increasing its sliglit "deficit" with respect to its share in GDP from 2.2 to 3.1 percentage points. 10 El Salvador: However, these rates do not include refinanced balances, as refinancing (stock) data for 1995 are not reported by sector. Using an estimate of agricultural refinance for 1995, the decline in participation in the stock of formal debt would be from 22.8% in 1991 to 15.2% 1995. 2.46 Disbursed credit (flow) to agriculture, including refinancing, decreased from 24.9% of total disbursements in 1991 to 13.4% in 1995. New disbursements (i.e., excluding refinance) show an even more striking decline, from 21.4 to 9.9% of total credit disbursements over the same period. Hence, the amount of refinanced-loans has increased in agriculture faster than in other sectors. For example, refinanced disbursed loans to agriculture grew 4 times faster than that of non-agriculture between 1991 and 1995 (24.9% and 6.4% p.a., respectively). The increase of refinanced disbursed loans to agriculture reflects increasing repayment problems of the sector . The amount of agriculture refinanced disbursed credit accounted for 14% of the total disbursed credit to agriculture in 1991, and increased to 27% of the total in 1995. Meanwhile, the share of non-agricultural refinancing out of a total credit disbursed to non-agriculture (including refinancing) declined firom 9.5% in 1991 to 4.4% in 1995. 2.47 The increase repayment problems of the 30 1 agricultural sector partially explains the relative 25- decline of the share of agriculture in credit allocation. It also explains why the decline the agriculture share 20 - . of the stock of crediit is not as marked as that reflected - - --_ by the flow measures, since overdues disguised as refinanced balances continue accumulate in the end- 'o - . of-year figures reported by the banking system. - .-. 2.48 In sumnnary, the share of agriculture 1991 .9,2 I993 1994 1 995 commercial banks credit flows has declined r .. GDP .b .-.b.r.-... significantly over the last five years, while the Figure 2-2 Agriculture Share in GDP, Debt Balances, and Disbursed sector's share total outstanding debt has not declined credit as much. These trends have manifested during a period of overall banking sector growth and financial reforms, and dynamic performance of non-agricultural sectors such as commerce. These facts collectively suggest that sectoral credit participation reflects the relatively poor sectoral economic performance, rather than constraints or barriers in the banking sector to agricultural lending. The relative decline in formal credit flows occurs at a time when required changes product mix and changes technology are important measures to increase agricultural sector profitability. Financial markets can facilitate those changes. 2.49 Who is providing agricultural credit? Looking at the sources of credit for the agricultural sector as a whole in September 1995, private banks provided 74% of total credit granted to agricultural enterprises, while SFIs accounted for 26% and financieras for 0.5%. Private banks and SFIs clearly have different economic sectors of specialization. As of September 1995, the SFIs loan portfolio was almost equally split between agriculture and non-agriculture sectors (47% and 53% respectively), while only 10% of private bank and 0.4% of financiera credit is extended to agriculture (Annex B, Table B-8). 2.50 In terms of the number of clients, therefore, SFIs served approximately 90% of totail agricultural clients, while private banks and financieras served the remaining 10%. In 1995 BFA served approximately 32,000 clients (agricultural and non-agricultural combined) with about 31,700 loans. Given a 60% portfolio composition of agricultural credit, we can estimate that the number of clients with agricultural loans was about 19,200. In December 1995, FEDECREDITO had 11,145 agricultural loans in its portfolio which can be used as a good approximation of its number of clients. In contrast, three of the largest banks of El Salvador (Salvadoreiio, Agricola, and Comercio) and Financiera Calpia lend to a total of 3,391 agricultural clients, of which 2,151 are clients of Calpia, a's estimated through interviews with credit officials from those institutions (see Box 2-1). Rural Finance 11 2.51 The banking system concentrated agricultural credit in two traditional export crops: coffee and sugar cane. Sixty-four percent of the total credit disbursed to agriculture in 1995 went to these crops. Traditional export crops such as sugar cane and coffee are attractive due to well-developed infrastructure for their marketing, high profit margins, and low cost of enforcing credit contracts. Banks can reduce their credit risk by requiring borrowers to provide an orden irrevocable de pago from the coffee beneficios and sugar mills. In other instances, banks lend directly to these beneficios and mills, which in turn may give advances to crop growers towards upcoming harvests. 2.52 The difference in the market shares in terms of loan portfolio and number of clients suggests that private banks and SFIs serve different market niches within the agricultural sector. Banks specialize in large loans to coffee beneficios, sugar mills, and large-scale farmers. In contrast, SFIs serve mostly small agricultural producers and agrarian reform beneficiaries. BX 2-1 Innati Fice: Financiera Calpid . ~ ~ ~ ~ ~ ~ ~ .... :: :. : - - : ... . . -: - ... .... ...-- Calpi& wasced 'in 1l995_ with. the trnformation of Servicio Crediticio de AES,'a'se' i-f nc intenediy,. mito a- finafciera, wth - teni'1cal assistance from 'GTZ and fuding from 1DB., 'CAEI o spo . I mission. is to provide loans to microenterprises in the infonnal sector at real interestprates which provide for vcos recoery and ........... .. .... . ... . . ...-_ allow for. nancial and institutional sustainability. The importance of micro and small bui'nesses in El Silvd is reflec in Ihe ,fi-t tha.m're' than, .9% of productive units and about 50% of the economically active popltinarefound pin th informal sector-accoutg. fior- an- estimnted 40% of.GDP. In March 1995. Calpia had disbursed 31,477loas' f abo .1' 9 n oper,ted-one ,cenWt unit an'd braches with a team of 68 professionals. The key to Calpia approah t .i. ,,leding i thor credit risk analysis. It does not allocate any of its budget to supervision. with analysts' time spent on risk .anysis an .e.ati loan disbursement, and recovery of funds. Curr.en'tly, Calpit serves 12,300 clients, of which about 1,800 are engaged 'nagicul activities The majo - of clients r '.wom''.. .(F.ud go, 1995). It extends agricultural and non-agricultua credit, ba o a i s .e ... Fo a of 6 brahich ovide: loans to agriculture. According to its officials, Calpin started 1995 - aagricu 'of 4 ill-n, .uhich had risen to al11 million by year-end. Although agricultural lans greww 175%i. :12 monts te delinquency rate remains at 1-1.5%. AgricuLtural loans account for about: 18% of the total loan portfolio, despite the fact that agricultural cedit was introduced .y i late 194. Calpia. proes credit tofarers who do not have access to othersources, with loan amountsas small as 05O. MoJstloa,nsare for working capital, withian average term of 6-9 months, for a maximum--of 1-8 months lnterest rates charged vary by sector,.Accordingto its officials, Calpia subsidizes interest rates on agricultural credit, whiclh:are offered at 3%/ per mouth, althou tcost 5.6% per month Loans to trade and services, cost 3.25% and are offered at 4%.per Moth. Loans are classified in two categories. (a) A for loans with an-ers of less than 30 days- and (b) E for loans with arr'''s omore. than 30 days. he. delinquency rate in. 1995 (amount of delinquent loans over total loan portfolio) was 2.22%. Source. Miion inte ihvarious professionals at Financiera pi y 16 and 22, 996.. 12 El Salvador: Chapter 3. EL SALVADOR: RURAL CREDIT MARKETS 3.1 This chapter analyzes the availability of credit services to rural households in El Salvador. The main objectives are the evaluation of the performance of rural credit markets (RCMs) and the knowledge of "who receives credit services, under what conditions, and from whom". The data set used in this chapter come from the 1996 FUSADES survey of rural households. The survey contains information about several aspects of the productive activities of 628 rural households from all regions of El Salvador. These households were stratified according to the main economic activities: (a) farmers; (b) landless agricultural households; (c) landless non-agricultural households; and (d) landless households with mixed sources of income. 3.2 The evidence presented in this chapter suggests that the perfornance of RCMs is poor. The rural credit markets (RCMs) in El Salvador are shallow because rural households have limited access to credit services: less than 12% of rural households received a loan transaction and about 20% had outstanding debt balance from formal or informal sources. Furthermore, rural households do not have access to a variety of lenders, as only 0.3% reported debt balances with more than one type of lender. Rural households have little opportunity to participate in credit markets because of the weak supply of ancd weak demand for credit. On the demand side, 870/o of rural households did not apply for loans in 1995. Although information regarding the weak demand for credit could not be assessed, the findings suggest that households find it cdifficult to use movable goods as collateral. Furthermore, access to formal lenders is limited to those rwual households that could offer real estate properties as collateral. 3.3 The supply of credit may be impeded by a combination of factors: (a) an underdeveloped institutional infrastructure; (b) government interventions which croNvd out private lenders by allowing weak government-owned institutions to lend with poor recovery rates and low interest rates; (c) debt-forgiveness programs which have created serious credibility problems by promoting strategic defaults; and (d) previous interventions in the marketing of agricultural products have prevented the development of informal sources of funding such as crop purchaser credits. 3.4 The rural economy may have difficulties in adjusting to the policy reforms implemented since 1989. As a result of these reforms, rural households may need to change their productive activities, adjust their scale of operations, or invest in new technologies. The performance of financial markets is crucial to the success of rual household's alttmpts to undertake those changes. Shallow credit markets may force rural households to rely on internal sources of funds which are often limited. Furthermore, assortive matching patterns imply that rural credit markets are broken down into small clusters in which each type of lender serves a particular clientele. The survey shows that informal lenders rely on idiosyncratic information generated through personal relationships to screen their borrowers, a practice that limits the size of the market. 3.5 This chapter is organized as follows. First, the characteristics of rural households, lenders and credit market shares are presented. The conditions of loan and debt contracts are provided next. Then, the performance of rural credit markets is addressed from the perspective of participation and interest rates differentials. THE RURAL HOUSEHOLDS 3.6 Rural households are varied in their levels of productive income. Table 3-1 shows the quartile distribution of the annual household income that rural households earn from their economic activities.12 The average income of the fourth quartile is almost 10 times larger than the average income within the first income quartile. Furthermore, in real terms the income of the first and second quartiles are below the real GDP per capita for 1995. 12 Annual householid income includes the monetary income and the imputed value of non-monetary revenues received by the household. It does not include agricultural production costs. Rural Finance 13 Table 3-1 Annual Household Income (colones) Numbber Average S. D. Minimum Maxkmem Averag hI Annual Household Income Obs ¢ of lef 1I quartile 157 5,459.9 1,982.1 610.0 8,640.0 3,899.9 2ndquartile 157 11,621.0 2,014.1 8,700.0 15,000.0 8,300.7 3equartile 157 20,158.9 2,946.7 15,002.4 25,715.0 14,399.2 4th quartIle 157 52,732.0 69,810.1 25,744.0 812,800.0 37,665.7 TOTAL 628 22,492.9 39,360.5 610.0 812,800.0 16,066.4 Source: Encueesa Desarrollo Rural, 1996. The GDP per capita in 1990 colones fbr 1995 was c8,254. 3.7 Comparative statistics on selected variables were calculated for rural households. These households were classified according to their participation in credit markets as debtors. Debtors are those households dtat reported an outstanding debt balance at the time of the survey. As shown in Table 3-2 debtors and non-debtors are different in various aspects. On average debtors live in smaller municipalities, are younger, have larger households, higher share of income from wage employment in non-agriculture, higher non-agricultural income, and lower income from other sources and from wage employment in agriculture. However, borrowers and non-boffowers differ in a smaller number of variables thzn debtors and non-debtors. On average, borrowers are younger, participate more in groups and associations, and have higher levels of non-agricultural income (rable C-i). Table 3-2 Debtors versus Non-Debtors All Households Debtors Non-ebtors Variables Mean S. D.. Mean S. D.. Mean S. D. Age of household head 45.7 14.3 42.7 13.6 ' 46.5 14.4 Average age of household members 26.3 11.7 23.8 8.8 26.9 12.3 Number of people in household 5.9 2.5 6.5 2.6 ^ 5.7 2.5 Participation in groups (% households) 9.7 29.6 14.1 35.0 8.6 28.0 Income from: non-agricultural activities 890.5 9,373.1 2,357.7 20,055.9 ** 518.6 2849.6 wage employment in agriculture 4,477.2 6,712.0 3,609.1 5,123.4 4,697.3 7,045.5 other sources 1,628.5 4,703.3 828.5 4,243.0 ** 1,831.3 4,795.7 % income from wage employment in 36.8 44.0 47.9 52.5 * 34.0 41.1 non-agriculture Population in the household municipalkty 38,298.1 43,550.8 32,485.3 30,195.7 39,771.6 46,237.4 of residence Source: Encuesta Desarrollo Rural, 1993 IrPr differences between means significant at the 90/95/99 percent confidence intervals, respecvely. LENDERS AND TECWMOLOGIES 3.8 The financial system includes three type of intermediaries: (a) formal, regulated by financial authorities; (b) semi-formal depending on whether they have a government-granted license to provide financial services or operate under a specific regulation; and (c) informal. The formal sector includes private commercial banks, specialized financial institutions and financieras, supervised by SSF. 3.9 The semi-fornal sector includes non-bank financial intermediaries and NGOs. For the purpose of this study, the informal sector includes commercial credit providers, moneylenders, friends and relatives, and traders. Commercial credit providers are enterprises or individuals that exchange goods or services today for the promise of repayment in the future. These lenders provide the loans in-kind and the effective interest rate charged is implicit in the credit price of the goods or services sold.'3 Moneylenders are individuals who provide loans at explicit interest payments. Friends and relatives are individuals that give credit for altruistic and reciprocity reasons without a profit motive, usually at no interest. Traders provide cash or input advances in exchange for a promise to deliver certain outputs in the future. The main example of this category are compradores de cosecha or coyotes. 13Unfortunately, the survey does not contain information regarding the differential between the cash and credit price, and the amount of downpayment requested by these lenders. 14 El Salvador: 3.10 Lending Technologv. Any lender must develop a method to grant and recover loans. This method has two components: operational and information. The operational component is related to the organizational structure, management information systems, physical infrastructure, and internal control mechanisms of lenders. The information component of the lending technology reflects a particular solution to the problems of adverse selection, moral hazard, and enforcement that characterize RCMs. These problems arise from differences in information between borrowers and lenders and the separation of disbursement and repayment over time. In order to reduce the degree of information asymmetry, lenders spend resources to screen their loan applicants, monitor borrower actions and enforce loan contracts. The costs of acquiring or buying information depends on the nature of the information-gathering technology used by lenders, the heterogeneity of rural households, and lenders' a priori stock of knowledge. 3.11 Lenders use standard or idiosyncratic information generated by borrowers and/or lenders to screen and monitor borrowers and to enforce credit contracts. The ability to read and use standard information is not affected by the distance between borrowers and lenders. Its value can be easily transmitted across distance. However, the value of idiosyncratic information declines with the distance between lenders and borrowers. Examples of standard information are: audited financial statements, real estate property appraisals, credit reports, and stock price quotations. Idiosyncratic information such as reputation is generated by interactions among local residents, or among agents in the same market. 3.12 The household survey provides indirect evidence regarding the relative importance of these two types of information. Infonnal lenders rely mainly on idiosyncratic infonmation such as borrower attributes and actions to screen their clients and enforce loan contracts. Rural households reported that the most common requirements of informal lenders are personal attributes such as reputation in the locality. For example, 17.8% of informal lenders require a good reputation from their loan applicants. However, the same households reported that formal lenders use standard information to screen their borrowers by looking at the ability and willingness of rural households to pledge real collateral. For example, as reported by rural households, the most important requirement of formal lenders is real estalte collateral (70.5% of transactions). THE REILATIVE IMPORTANCE OF FORMAL, SEMI-FORMAL, AND INFoRMAL LENDERS 3.13 The markcet shares for each type of lenders are measured in terms of the total number of individual transactions and the total volume of transactions as reported in the household survey. Figure 3-1 summarizes the market shares (amounts and number of transactions) of the different types of lenders for 1995.14 During 1995, formal sector lenders provided 76.1% of the total loan amounts borrowed by rural households, with the BFA accounting for 61.2% of this total. This shows that BFA specializes in servicing rural households, in contrast to private banks which provided three-fourth of total credit to agriculture (Chapter 2) but represent only 4% of the total loan amount reported by the households in the survey. The share of informal lenders in terms of total amount borrowed, 17.7%, differ substantially from the one reported by Cuevas et al. (1991) and by Moll et al. (1997) where informal loans accounted for 66% and 10% of credit received during 1991 and during 1992-1994 respectively. 14 Although the suntey recorded all loans received during 1991-95, the analysis concentrates on loans received in 1995 because it seems that households had difficulties in remembering loans received in previous years. Rural Finance 15 3.14 In sharp contrast to other countries and Tansactions 111111 Amount with previous studies for El Salvador, the survey 80.0- 70.0 - findings suggest that informal credit m ahets are 60.0- -relatively underdeveloped. The informoal sector 50.0- -accounts for 48.8% of loan transactions, in contrast 40.0- -with previous studies, where informal loans 30.0 - represented between 64% and 80% of the loan 20.0 - _ transactions performed by rural households. Loan '10_0 ---- --------- M ---------- ---- services are provided mainly by BEFA (30%); 0.0o _ |111 || _L; friends and relatives (16.3%); and comeercial Formal Semi-formal Informal credit (15%). Government intervention in the Figure 3-1 Loans Received: Market Shares marketing of agricultural products may have contributed to the scanty representation of interlinked credit contracts. For example, some agricultural products were marketed through goverrmnent-owned agencies (e.g., INCAFE, INAZUCAR, and IRA). I X*% Transactions 0% Balance X 3.15 The relative importance of formal, 7D.0 semi-formal and informal lenders in terms of 60.0 - outstanding volume of debt is su sta tially so.o 4I.-i- .I- -i ldifferent from the shares of ount 4 0.0- 3__:7= [= ...borrowed in 1995. The shares of the 30.0 informal sector in of volume of outstading .... Y ........ .~~~~~... .... . ... ... 20.0 -.... > . ; : - ; ... ......... : .... debt and transactions increased, While that 10.0 for formal and semi-fom . sectors 0.0 decreased. The data show that 3 .1 1% of the Formal Semi-formal Informal total volume of debt and 62.3% of the Figure 3-2 Rural Household Debt: Market Shares number of outstanding debt contracts were from informal lenders. These appear more realistic estimates of the relative importance of informal lenders. CHARACTERISTICS OF LOAN AND DEBT CONTRACTS 3.16 The characteristics of credit contracts prevalent in rural credit markets are presented in Table 3-3 and Table 3-4 and derive from the features of contracts reported by rural households during 1995, and from the outstanding debt balances. Five elements of the loan contract are considered: loan amount, maturity, interest rate, repayment problems, and collateral and guarantee. The main findings are: (a) formal and informal lenders provide loan contracts with different characteristics (e.g., interest rates) and exist side by side; (b) the characteristics of loan contracts received by different borrowers depend on the lenders to which they have access; and (c) private banks provide larger loan sizes that allow them to charge lower effective interest rates than other types of lenders. 3. 17 Amount. While it varied significantly, the average loan size across lenders was 04,040. rhe formal sector provided the largest loans by far in rural credit markets in El Salvador with an average loan of 07,725. Loans between p5,000 and 0 10,000 represented 42% of the number of loans and 37.1% of the total amount disbursed by formal lenders. The semi-formal sector ranked second with an average loan of ¢2,439, followed by the informal sector which provided loans with an average size a little over half of that ( 1,441). 3.18 BFA provides the second largest loans in rural credit markets (p8,020) which is almost twice as large as the average loan granted by FEDECREDITO (,4,833). Moneylenders loans placed sixth in rank and averaged p1,645, while NGOs had the smallest average loan size (p707). Friends and relatives grant verv small loans with 53.8% amounting to less than 1,000 and 19% of the total amount disbursed. 16 El Salvador: Table 3-3 Characteristics of Loan Contracts by Lender: Amount, Maturity, and Interest Rates Average % Transactions of Each Lender Type of Lender No. Loan Maturity Effective Interest Land Verbal Repayment Obs. Arount (Months) Rate (% p.a.) Collateral Proriises Problems Fonnal 31 7,725.0 16.3 18.3 57.6 7.0 33.3 BFA 24 8,019.8 12.8 18.2 58.3 8.7 33.3 FEDECREDITO 3 4,833.3 18.0 19.7 66.7 - 66.7 Other Formal 1 13,000.0 72.0 66.7 - Private Banks 3 6,500.0 22.0 17.7 33.3 - 33.3 Seml-formal 8 2,439.1 17.9 51.3 25.0 18.5 12.5 NGOs 2 706.5 10.5 59.1 - 57.5 - Bancos Comunalb 6 3,016.7 20.8 48.2 33.3 15.5 16.7 Cooperatives Informal 39 1,441.0 7.8 99.3 10.3 60.6 28.2 Comrnmercial Credit 1 2 1,491.7 9.0 28.3 8.3 36.3 16.7 Moneylender 11 1,645.5 7.3 199.5 18.2 69.1 45.5 Friends/Relatives 13 1,092.3 6.6 0.0 - 84.9 23.1 Interlinked Credit 3 2,000.0 8.7 324.0 33.3 50.0 33.3 TOTAL 78 4,040.9 12.2 58.0 31.3 17.3 28.8 Source. Encuesta Desarrollo Rural, 1996 The number of observations varies for each variable because of missing data. 3.19 The average loan amount granted to different groups of rural households vary considerably. For example, farm households receivedL loans averaging ,6,154 almost three times larger than landless households working in agricultural activities (02,194). Landless mixed-income households received very small loans with 50% amounting to less than 1,000 and accounting for 4.7% of the total amounts received. 3.20 The average outstanding debt balances held with different lenders show different values than that of loans granted. However, the same pattern of market specialization is evident. For example, the average debt with formal lenders is almost 4 times higher than that of informal lenders. The average debt held with private banks was 2.2 times higher than the average loan size granted by infornal lenders in 1995. 3.21 Maturitvy Formal, semi-formal, and informal lenders provide loans with different maturities. Semi-formal and formal sector loans had a significantly longer repayment period, averaging 17.9 and 16.3 months respectively, as compared to informal loans (7.8 months). Formal sector loans are usually granted with a maturity between 6 to 12 months. Sixty-four percent of loans and 40.6% of the amount disbursed by formal lenders were in that category. As might be expecled, the informal sector required shorter repayment period than formal and semi-formal lenders. Fifty-seven percent of the loans and 47.4% of the amounted disbursed by informal lenders had a maturity of less than 6 months. 3.22 The maturity of loans vary with the type of household because they have different access to alternative lenders. For mixed-income landless households the average duration of loans is 2.5 times of agricultural landless households. The average maturity of loans received by farm households is 10 months, and the corresponding average of non-agricultural landless household is 15.5 months. As a whole, rural households received small loans vith a short term repayment period. For example, 25.7% of the loans had a maturity between 6 and 12 months and a size between 0500 and 02,000. 3.23 Nominal Interest Rates. Loan transactions in RCMs have a wide range of nominal interest rates. The average interest rate ranges from 17.7% p.a. in the case of private banks to 324% in the case of interlinked credit transactions. Infoimal lenders charged the highest interest rate with an average of 99.4% p.a.. Semi-formal intermediaries followed with an average rate of 51.3% p.a., while formal lenders required a much smaller return from rural households, or 18.3%. The very high financial cost of informal loans may be explained by the short maturity and small amounts of the transactions. The nominal interest rates charged by lenders within the same group varied significantly more for informal than for formal lenders. The coefficient of variation (CV) is the highest in the case of infomral lenders, equal to 243. 1%.' 5 The coefficient of variation equals the standard deviation of the variable divided by its mean. Rural Finance 17 3.24 The interest rates charged to different groups of rural households varied. Rural households that belong to the fourth quartile of income paid an average nominal interest rate of 15.1%, which is 8 times the average charged to rural households that belong to the first quartile of income. The average interest rate charged to farm households was 7 times that of landless households working in the non-agricultural sector. 3.25 The nominal interest rates on outstanding debt balances are generally lower than that of loans granted in 1995. The nominal interest rates charged on loans correspond to the cost of acquiring additional debt, (marginal cost) while the interest rates on debt balances represent the average cost of debt (average cost). Table 3-4 Debt Contracts by Lenders: Amount and Nominal 3.26 Repayment Problems. The survey Interest Rates (% p.a.) collected information to calculate two proxy Mean Type of Lender N Debt Amount Nominal Interest indicators of loan default: the percentage of Rate (% p.a.) loan transactions and the percentage total Fonnal 44 7,138.5 17.6 amount borrowed that experienced repayment BFA 28 6,12.62 17.4 problems. Within each approach, the FEDECREDITO 4 2,810.3 18.5 Other Formal 7 8,502.9 18.5 percentage of repayment problem occurrence Private Banks 5 14,360.0 18.8 was computed by institution and as a Semi-formal 8 3,350.0 16.3 NGOs 2 5,200.0 20.5 percentage of all nstitutions These proxies Bancos comunales/ 6 2,425.0 14.3 of delinquency have the following cooperatives drawbacks: (a) the information was collected Commerial Credit 87 2,2959. 306 from borrowers who may under-report Moneylender 13 3,728.9 145.3 delinquency; (b) the amounts used to proxy Friends/Relatives 38 695.7 - arrears correspond to the loan size and not to Interlinked Credit 4 2,175.0 200.4 Other lnformal 3 3,300.0 33.2 the outstandmg balance at the time of the TOTAL 134 3,728.9 31 3 arrears; and (c) the delinquency problems of Source: Encuesta Desarrollo Rural, 1996 informal sector loans may be overestimated because the questionnaire ignores that these loans are usually granted with flexible repayment terms. 3.27 Twenty-nine percent of all loans received by rural households experienced repayment problems. Formal and informal sectors have the same proportion of delinquent loans (13.8% of all loans). However, formal sector loans accounted for 17.5% of the total amount. Delinquent loans by BFA alone accounted for 10% of total number granted, followed by moneylenders 6.3%, friends and relatives 3.8% and FEDECREDITO 2.5%. 3.28 Repayment problems are more severe for formal lenders than for their informal and semi-informal counterpart in terms of the number of individual loans. The very poor recovery rates of the formal sector are due to the dismal performance of FEDECREDITO with two-thirds of the number of loans granted in delinquency and BFA with one third of loans granted in delinquency. Interestingly, private banks displayed the same record as BFA. Formal sector loans in arrears equaled 23% of the total amount granted by this sector. The corresponding percentages for loans from the semi-formal and informal sector were 25.6 and 37.6%, respectively. In terms of amount borrowed, informal lenders have more repayment problems than fbrmal internediaries. 3.29 Landless agricultural households fell in arrears more frequently (36%) than farm households (30.4%), non- agricultural landless households (26.9%), and mixed landless households (0%). Hence, agricultural landless households seem to pose the highest credit risk. The loans that fell in arrears amounted to 3 7.4% of the total amount disbursed to such borrowers, and to 36% of the total number of loans. In contrast; mixed landless households did not report repayment problems.'6 It may be the case that loans granted to mixed landless households were not yet due at the time of the interview. 18 El Salvador: 3.30 Cllate al and Guarantees on Loans. Lenders in El Salvador rely mainly on land as collateral to guarantee loan repayment. Lender groups can be ranked by their use of real collateral as follows: * Fonnal lenders have the strictest collateral requirements. Land collateral are mainly used by formal lenders who required it for 57.6% of their loans. These loans amounted to 63.3% of the total amount disbursed by formal lenders to rural households. Furthermore, formal lenders used animals and other property as collateral for 18.4% of their total amount disbursed and for 18.2% of their loans. * The semi-formal sector mainly relied on fiduciary contracts, which were used to guarantee 50.2% of loan amount granted by the sector and 12.5% of loans. This was a result of its wide-spread use with Bancos Comunales and Cooperativas. NGOs relied heavily on verbal promises (57.5% of total loan amount) and crop guarantees: (42.5% of this total). e As expected, informal sector lenders had to rely mainly on verbal promises and informal agreements, which were used to guarantee 74.3% of transactions and 60.6% of loan amount granted by the sector. Moneylenders accepted four types of collateral, with verbal promises backing the majority of them (69% of loan amount and 63.6% of transactions they granted). Only 6.1% of moneylenders' disbursements to rural households had land as collateral. * Friends and family use the least stringent collateral requirement. Seventy-seven percent of loans disbursed were backed by verbal promises, which also accounted for 84.9% of the amount disbursed. 3.31 The reliance on fiduciary contracts and informal arrangements to guarantee loans indicates that rural credit markets in El Salvador may be segmented. On average, about 59.7% of loans received by rural households were guaranteed widtout real collateral. These types of credit contracts make the use of local knowledge and idiosyncratic information about potential borrowers an important ingredient in the provision of credit. The increase mobility that resulted from the conflict has weakened the flow of local information between informal lenders and their clients. Consequently, the lending transaction costs have increased resulting in a reduction of the supply of credit of informal lendevrs. These lenders may have lost some of their comparative advantage in servicing low income clients. The pool of already know creditworthy clients was suddenly reduced as they migrated to other areas or the profitability o:f their activities was hampered. An efficient use of collateral may reduce both lender and borrower transaction costs. Collateral acts as a signaling device to show borrower commitment and incentive to repay the loan. PE, RFORMANCE OF RURAL CREDIT MARKETS 3.32 Credit markets in El Salvador are very similar to those in Mexico but differ from those in the rural areas of other developing countries. These two countries have experienced similar events such as macroeconomic and stabilization lprograms initiated in the late 1980s, nationalization and privatization of commercial banks, and govermment intervention in output and input markets. As in Mexico, Salvadoran RCMs are: (a) very shallow as few rural households have access to credit services from any source (e.g., formal or informal); and (b) characterized by a matching system in which households are matched with a specific sector (e.g., formal or informal). The Participation of Rural Households in RCMs 3.33 The participation in RCMs of El Salvador is different from that of other developing countries. First, Salvadoran RCMs presenlt very low rates of participation as measured by the number of borrowers and debtors. According to these findings, 11.8% of rural households had received a loan from any source (e.g., formal, semi-informal, or informal) in 1995 and 20.2% had outstanding debt balances at the time of the survey. These unusually low rates of participation in credit markets differ from previous studies that reported 9% to 15% of rural households receiving loans from formal sources, while observed rates of access to informal finance ranged from 15% to 61 %. The reported rates of participation in the rural areas of Bolivia, Costa Rica, Honduras, Mexico, Nigeria, The Philippiries, and Thailand range from 45% to 96%. Rural Finance 19 3.34 Second, almost the same number of rural households received loans from fonnal or informal lenders, in contrast to other developing countries Formal 1.1% Informal where the number of rural households receiving 5 ..4% loans from informal sources is 2 to 16 times higher that those from foarmal lenders. Only 5.3% of rural households borrowed from formal lenders, 1. 1% borrowed from semi-formal Non- lenders, and 5.4% received loans from informal borrowers lenders (Figure 3-3). However, 6.7% of rural 88.2% households had outstanding debt from formal lenders, 0.8% from semi-formal sources, and Figure 3-3 Participation of Rural Households in Credit Markets 12.4% from informal lenders (Figure 3-4). 3.35 The participation of rural households in credit transactions is related to their land endowments and the economic sector in which they operate. Farmers have the lowest participation rate in credit markets, as 16.7% had any outstanding debt balance. Access to credit services increases for landless households, as 21.3% of them had positive level of debt. Furthermore, non-agricultural landless households had the highest participation rate, as 24.7% of them were debtors. 3.36 The type of lenders to which rural 80.0 - . .households have access is also associated with 70.0 . - ---- the economic sector in which they operate and 600 .0---------.----.-.-.-.-.-- - - ----- -.- their land endowments. A larger share of 500 0- farmers (9.4%) had outstanding debt balances 40.0 - with formal lenders than corresponding landless 30.0 --....... _ households (5.5%). In addition, non-agricultural 20.0.- . --- landless households had the highest access to 100. - _ _ infonnal lenders as 17.2% of them reported debt 00oInformal o balances. Veiy similar percentages of male-headed o and female-headed households reported outstanding Figure 3-4 Participation of Rural Entrepreneurs as Debtors debt balances, close to 20%. Seven percent of male headed households were debtors from formal lenders, while the corresponding r ate for female headed households was only 4%. 3.37 Participation rates and The type of lender to which rural households have access is not associated with their levels of income. Similar percentages of first quartile and the fourth quartile reported debt balances with formal or informal lenders, close to 8.2% and 11.5% respectively.)7 In other developing countries, households with low levels of income are more likely to have access to informal lenders, especially friends and relatives. In El Salvador, BFA (a formal lender) is the main provider of credit services in the rural areas and targets a wide range of clients. Even though participation rates are not affected by household income, the characteristics of loans and debt are associated with it. As expected, rural households that are richer, as proxied by their income, received larger loans and had larger outstanding debt balances than their poorest counterparts. 3.38 The weak participation of rural households in credit markets results from the interaction of lenders and borrowers characteristics and of the economic environment in which these agents interact. The supply of financial services may be impeded, among other things, by the following reasons: 17 The Chi-square test for the association between participation and income levels yields that the null hypothesis of no association can not be rejected. 20 El Salvador: * Previous government interventions in output and input markets hampered the development of informal sources of finance, such as supplier and crop purchasers.'8 * The civil war and the induced migration have weakened the availability of idiosyncratic information and the cost advantage of informal lenders in screening potential borrowers. * Debt-forgiveness programs have had a negative spillover effect by promoting strategic default. For example, Financiera Calpia reported that after the most recent debt forgiveness program many borrowers wanted to receive the same benefits from the financiera. * Underdeveloped institutional infrastructure that result in poorly defined property rights and a weak and inefficient legal system. These conditions give borrowers opportunity to default in their loans, * Reliance on traditional lending technologies that entail large fixed costs which makes rural lending unprofitable with small loans and a limited volume of transactions. 3.39 The effeetive demand for credit is weak as 87.2% of rural households did not apply for loans during 1995. Unfortunately, the survey did not record why rural households did not request loans. The demand for credit is affected by: (a) interest rates; and (b) transaction costs of borrowing such as timeliness, number of trips, travel expenses, and fees and commission. These transaction costs are affected by: (a) lending technologies and their associated screening requirements; and (b) the ability of rural households to signal creditworthiness to lenders. 3.40 Access to formal loans is limited to those rural households that have adequate levels of real collateral and that demand larger loans that can support the high costs of using real collateral. As reported by rural households, 70%/o of forma!l sector lenders require the title of the house or land to grant a loan. Even though informal sector lenders rely on local information and applicant's reputation to grant a loan, the conflict may have reduced the availability of local information which is generated through daily interactions in other markets such as output, input, and labor. The Matching of Debtors and Lenders 3.41 Only two rural households in El Salvador reported outstanding debt balances with different types of lenders (e.g., formal, semi-formal or informal). As in other developing countries, rural credit markets exhibit assortive matching patterns that sort and link borrowing households and lenders according to classes. This regularity resuts fiom cost complementarities between lending technologies and borrower abilities to convey information about their risk types to lenders (Sanchez-Schwarz, 1996). 3.42 A multinomial logit model is fitted to the participation of rural households as debtors in RCMs. The results of the model indicate that rural households do not have uniform access to formal and informal lenders.'9 Rural households that have debt balances with formal lenders participate in groups and associations, and have land endowments. In contrast, debtors of informal lenders have larger household size, and a higher share of off-farm non-agricultural labor income. Rural households that are more likely to be non-debtors have smaller household size and a lower share of off-farm non-agricultural labor income. Although, various lenders provide loan contracts with different size, interest rates, maturities, and collateral requirements, the clientele of these lenders are not differentiated by size as measured by household income. 8 Adverse selection, incentive, and enforcement problems can be solved through linkages between credit market and other markets (i.e., input and commodity markets). 9 The multinomial logit model is a discrete regression model in which the dependent variable is classified into m categories. The participation of rral households in credit markets was classified into three categories: non-debtors, debtors from formal lenders, and debtors from informal lenders. The categories of infornal and semi-formal lenders were merged. Rural Finance 21 Chapter 4. PERFORMANCE OF RURAL FINANCIAL INTERMEDIARIES 4.1 This chapter summarizes the overall performance of three financial intermediaries (Fls) operating in El Salvador: the Banco de Fomnento Agropecuario (BFA), the Federaci6n de Cajas de Credito y de Bancos de los Trabajadores (FEDECREDITO); and the Federaci6n de Asociaciones Cooperativas de Ahorro y Cr6dito (FEDECACES). The BFA was created in 1973 to assist small farmers and rural producers, and now serves about 32,000 clients. The FEDECREDITO system began operations in 1943, currently offer credit services to microentrepreneurs and has 57 affiliates with 201,316 members. The FEDECACES system, founded in 1966, has 38 affiliates with 38,406 members. The three FIs differ in two aspects: clients and financial products. The BFA targets small agricultural producers, while FEDECREDITO and FEDECACES systems serve mainly micro, small and medium enterprises. The institutions differ in termns of the financial services offered. BFA and FEDECACES affiliates offer savings and credit services, while FEDECREDITO affiliates focus on credit products. 4.2 These FIs have been singled out for analysis because they are perceived to be an important instrument to deliver financial services (credit and savings) to under-served sectors of the population. In 1995, these intermediaries accounted for 6.1% of the net loans of the formal financial sectors, however the field survey reported that these FIs amounted for 65.7% of the total amount disbursed in rural areas.20 An assessment of the performance of these intermediaries is timely because the Government is looking for the identification of possible roles and policy refortns and instruments that could assist these intermediaries in providing financial services to the rural population. The review of these institutions draws upon previous study, available data, including financial statements and field visits. 4.3 A comprehensive comrparative analysis of these FIs cannot be done because of gaps and inconsistencies in the available data. In particular, consolidated data for the two federations and their corresponding affiliates were not available. As a result, the, financial performance of each federation and its affiliates is presented separately. This highlights the difference between the FEDECACES and FEDECREDITO systems: while the latter operates as a vertical structure with most funding for the affiliates being provided by the federation, FEDECACES affiliates are essentially independent units relying primarily upon locally mobilized funds. As a consequence, for the FEDECACES system, one should focus more on the affiliates than on the federation, while the opposite is true for the FEDECREDITO system. For the latter, there is an inherent risk of double-counting in the figures reported for the federation and the affiliates, a risk that one cannot be certain to have fully under control. Throughout the chapter FEDECACES and FEDECREDITO refer to the federations, while their networks (federation and affiliates) are denoted as FED]ECREDITO and FEDECACES systems. 4.4 The presence of these rural financial intermediaries in El Salvador have been costly for the Government. Recapitalization measures, and government transfers due to foreign exchange losses have drained the budget and deprived other sectors of the economy. During 1990-94, the Government transferred ,2,361.4 million to private and autonomous financial institutions (FUSADES, 1994). In general, these intermediaries have not mobilized deposits efficiently. They are characterized by inadequate provisions for bad debts, poor financial reporting, and deficient regulatory and legal frameworks. 4.5 Following Yaron (1992), two criteria for the performance of RFIs are used: (a) the level of outreach achieved; and (b) self-sustainability. Self-sustainability is defined as the ability of the financial intermediaries to obtain a return on equity at least as high as the opportunity cost of equity. The subsidy dependence index is used to measure the level of self-sustainability. Outreach is measured by number of loans and savings accounts, the average size of loans and savings accounts, and the real growth of assets. Detailed financial statements of the FIs are presented in Annex B. 20The relative importance of the affiliates of FEDECREDITO and FEDECACES in the market of loans has remained stable since 1991. However, BFA's share of total net loans has declined from 5.2% in 1991 to 3.5% in 1995. 22 El Salvador: OUTREACH 4.6 The clients of the institutions are engaged in different economic activities. BFA targets agricultural producers, while the two cooperative systems serve microentrepreneurs. The clientele of the two federations have different gender profiles. The clients of FEDECACES affiliates are mostly men, while the clients of FEDECREDITO affiliates are mostly women (especially the program of Credito Popular). 4.7 FEDEICREDITO can provide its services only to rural credit funds and workers' banks. In contrast, FEDECACES provides financial services and technical assistance to credit cooperatives that are not affiliated to the system. FEDECREDITO and its affiliates grant loans under individual and group lending technologies, while the BFA and FEDECACES affiliates use individual loans. Table 4-1 Outreach Indicators: 1995 FEDECREDITO FEDECACES BFA Federation Affiliates Federation Affiliates No. of clients or members 32,000 201,316 - 38,406 No. of clients in rural areas 19,200 12,595 n/a nta n/a Real growth rate of assets (1994-1995) 6.3 -4.4 12.6 -27.4 4.0 Number of affiliates or branches 27 branches 51 rural credit funds 38 credit cooperatives and 6 workers' bank Number of employees 1,463 242 704 51 Loan Outreach No. of loans 30,977 68,345 87,025 nta n/a Loan portfolio (thousands of colones) 1,189,700 340,938 516,758 39,385 334,718 Average Loan Size (colones) 38,406 4,988 5,938 nta n1a b Number of Loans per staff 21 282 124 nta n/a Savings Outreach No. of accounts 57,259 n/a n/a n/a n/a Savings Portfolio (thousand of colones) 369,736 23,123 20,589 4,222 222,855 Average Account Size (colones) 6,457 Source: Annual Reports of BFA, FEDECREDITO, and FEDECACES; Fundaungo (1995); and Interviews * BFA: 60% of the number of agricultural loans; FEDECREDITO: number of loans to agriculture plus 58% of the number of loans of the program Credto Fopular. Fundaungo (1995) reports that 58 % of the clients of that credit program lived in rural areas. Fundaungo (1995) reports for two SCCs that loan size ranges between one to 350 thousands of colones. Correspond to passbook savings and certificate of deposits 4.8 Loan Outreach. The data indicate that the BFA and FEDECREDITO have a limited loan outreach as they served about 16% of all agricultural enterprises. However, as shown in Chapter 2 BFA granted 30% of all loans received by rural households in 1995. According to BFA officials, the bank serves 32,000 clients on a regular basis and 38,000 clients as a result of the Peace Agreements.2' About 30,000 of BFA's current clients receive loans of less than 0 100,000. 4.9 Comparing the average outstanding loan size among these institutions with the average loan size reported in the rural household survey reveals that BFA serve a more diverse clientele than FEDECREDITO. The average loan size of the BFA is p38,406, however, the rural household survey shows that loans granted by BFA had an average size of 08,019. The average loan size of BFA may seem disproportionally high but it is 21 times lower than that of the average agricultural loan of two banks that account for 37.5% of the agricultural loans. 4.10 In 1995, the real growth rate of the loan portfolio of the BFA and FEDECACES have been moderate as compared to FEDECREDITO and FEDECACES affiliates, and the average Salvadoran bank. The real growth 21 In 1992, BFA had 9,024 clients (Wenner and Utunafla, 1993) Rural Finance 23 rate of the loan portfolio was 8.5% for the BFA, 22.9% for FEDECACES affiliates, 17.3 percent for FEDECREDITO affiliates, 4.5 for FEDECACES, and 17% for the average bank. 4.11 Savines Outreach. The FIs are allowed by law to use different mechanisms to mobilize savsings. FEDECREDITO is restricted to offer forced savings services that are attached to credit programs. In contrast, BFA and FEDECACES can offer voluntary savings and in recent years have initiated strong savings mobilization programs. High rates of real growth of deposits shows that savings mobilization efforts have been successful for BFA and for FEDECACES affiliates. In 1995, the real growth rate of deposits was 43.1% for the BFA, 23.7% for FEDECACES affiliates, against -22.2% for FEDECREDITO affiliates (see Table D-1 1). The relative success of the BFA in savings mobilization during the past three years, suggest that the demand for savings in the rural areas have been overlooked in El Salvador.22 The average savings account for BFA was of 06,457, or 18% lower than the average savings account of commercial banks outside the San Salvador department. 4.12 Savings mobilization efforts at the BFA, however, are far from satisfactory. Mobilized savings accounted for 31% of the BFA's loan portfolio and for 29% of assets in 1995 (Table 5-1), a factor that contributes to its continuing dependence on outside funds and subsidies. In contrast, FEDECACES affiliates finance two-thirds of their loan portfolios with deposits mobilized from their members and the general public, a share much closer to that shown by commercial banks (see Table D-12). FINANCIAL PERFORMANCE 4.13 The financial performance of the BFA was compared against the average Salvadoran bank. While comparing a development bank to the average commercial bank may have some drawbacks due to the differences in clientele and business practices, this type of analysis can be useful in revealing general industry trends and highlighting the strengths and weaknesses of this institution. The FEDECACES and FEDECREDITO systems were compared against each other when possible. Trends in the financial performance of these intermediaries were also summarized for 1993-95.23 Box 4-1 Risk of Financial Intermediation Credit Risk. Credit rsk refers to the- potential variation in net Interest:Rate Risk. Interest rate ris=k is the in incorne'that results from non-repayment ordelayed payment. interest rates will adversely affect net interest margin due'to a mismatch of rate-sensitive assets and liabilities when interest rates: fluctuate.. capital .or- Solvency Risk. Capital risk refers to the potential Liquidity risk. Liquiidity of an .assets: refers to the ability of-an- decrease in tnet,assets.yvalueleading to insolvency. An insolvent owner to convert the asset to cash withminimal. loss from price and financial.intermediary.can notmeetr cash outflows:due to operating depreciation. Liquidity risk then refers to hvariation in net expe 4 deposit s 'wthdrawals, and maturtdebt obligations. income:causes by aJfinancial.intenediaries' dfficul to obtain cash at a reasonable cost. .O-erational Risk. O .peraiarisk is the variation in net income Subsidy-dependence Risk. Reductions in the-annua fl..ow Of - ark-et vaue of euity due to :the ossibility that operating subsidies may -.endanger the stability of the intermediary and. eense .. tvary significanty fom what is expected due to therefore, the holdihgs of ts depositors if any. More importadly operaing co6ts, ,employeeso errors, incidence of employee and the Instability of services discourages repayrent. and deteriorates ..customer theft, and other factors, customer relationships. S$ou'r.c ..Koch1(1990) and Chaves and Gonzalez (1992) -Net intee tn44uaIi the rato of net interest income to eaming assets. 4.14 An important caveat needs to be introduced when analyzing the FEDECACES system: the "locus of sustainability" is not the federation (as it is in FEDECREDITO given its vertical structure) but the individual affiliates, for which we did not have sufficient data. This means that the SDI calculated for the federation would not have the same meaning (or level) as that estimated for the affiliates had data been available, and can hardly reflect the subsidy dependence of the system since most of the network savings mobilization is recorded at the 22 Unfortunately, the household survey did not ask for financial savings. 23 The data used for the analysis of the BFA and FEDECREDITO were derived from the financial statements published by the SSF while for FEDECACES financial statements published in its annual reports were used. 24 El Salvador: affiliates, not the federation. For example, the SDI estimated for FEDECACES is much higher than the level calculated for the BFA, even though the FEDECACES affiliates relative reliance on external funds is less than one-half that of the BFA (see below). 4.15 The financial analysis is organized by types of fundamental risk in financial intermediation. These risks are: (a) credit iisk; (b) interest rate risk; (c) liquidity risk; (d) operational risk; and (e) capital or solvency risk (Box 4-1). Each type of risk is associated with the uncertainty in the realization of net returns on assets. In addition, development banks face subsidy-dependence risk, which is associated with the possibility of reduction in the annual flow of subsidies that may endanger the stability of the intennediary and, if they mobilize deposits, the savings of its depositors (Chaves and Gonzalez-Vega, 1992). Financial indicators are presented in Table D- 11 and Table D-12. Performance of BFA 4.16 The BFA is a stated-owned bank founded in 1973 that competes with private commercial banks and other financial institutions in offering financial services such as loans, savings, wire transfers, checking accounts and money orders. In December 1995, BFA's assets amounted to 0 1,635 million which accounted for 4.1% of the total assets owned by commercial banks. 4.17 Besides its regular banking activities, BFA manages other programs according to government mandates. These programs are: (a) Programa para Reserva Estrategica de Alimentos Basicos; and (b) Government trust funds. The combination of banking activities with these other programs weakens the reputation of BFA as a commercial bank. In 1989, BFA entered a period of reorganization, by initiating the closing of its non-banking activities, such as marketing of fertilizer and grains. Nevertheless, BFA now plans to create a almacen general dep6sito, an inmobiliaria, and a brokerage firm (puesto de bolsa). 4.18 Significant government transfers have been necessary to keep BFA afloat due to its very low recovery rates, subsidized interest rates, directed credit programs, and the exertion of political influence to the detriment of sound business policies and banking practices (IPC, 1993). Table 4-2 shows assistance from Government and other measures undertaken in recent years. 4.19 Profitability. The performance of BFA in termns of return on assets (ROA) is not satisfactory and has been declining since 1993. BFA's return on assets (0.2%) fell short of the average private bank (1.7%) and used less debt-financing so that its equity multiplier was almost 8 times smaller. 4.20 The three main factors affecting the profitability of BFA are: administrative expenses; quality of the loan portfolio; and high proportion of fixed assets. Operating expenses were well above that of the average bank most likely because BFA grants smaller loans and targets small agricultural producers. BFA was less efficient in controlling expenses as compared to the average bank since it had a profit margin of 0.9% (against 12.1% of the average bank). BFA's asset utilization (operating income / average assets) of 17% was insufficient to offset its lower profit margin, resulting in a low ROA. Provision for loan losses relative to operating income is four times that of the average bank because BFA's management recognizes that the bank is exposed to large losses from problem loans and thus reports an amount of interest income which overstates what will actually be earned. Rural Finance 25 Table 4-2 BFA: Government Assistance and Interventions. Year Measures 1969 * BFA exchanged lbad loans for FFRAP bonds for 0176.8 million." 1990 * 8FA received government bonds (bearing 14% per year) for ¢227.9 million as a payment for bad loans from agrarian reform beneficlaries (adelantos-ISTA-BFA) and In exchange of ¢91.1 million of FFRAP bonds.eb 1991 * Government assumed BFA's debt to foreign creditors for ¢533 million of which ¢400 million went to BFA's capital. The remaining were refinanced at 4% per year. a * BFA bought back $11 million worth of loans which were sold to FFRAP.b 1992 * BFA pay back debt related to forelgn creditors for ¢122.3 million of which ¢107.7 paid with Bonos de Conversi6n y Consoldacdin de l DVeuda Pabica * D.L. 292 1994 * The debt of beneliciaries of D.L. 292 of less than ¢5,OOO was forgiven (Decree 114). According to BFA's official this debt forgiveness represented ¢25 million 1996 * D.L. 698 and D.L. 699 (Chapter 2). Sources: Carrandi (1992). IPC (19D93), and 'Memorias BFA, 1992 4.21 Fixed assets represented 7.8% of total assets, and are comprised mostly of warehouses and agricultural processing equipment for grcnos bdsicos that BFA has been unable to sell. The bank started a program to increase the utilization of its warehouses in August 1996 and undertook a feasibility study to operate its four warehouses from one subsidiary company. However, the business prospects of what is being proposed are not good, the facilities are technically in excellent repair. Lacroix (1997) proposes to lease the facilities to private investors as an alternative to circumvent privatization and to avoid recognizing a big capital loss. 4.22 Credit Risk Management. Despite government interventions to rehabilitate BFA's loan portfolio, its credit risk surpasses that of the average bank for all financial ratios. Past due loans were four times higher than the average bank accounting for 22.5% of loans in 1995. BFA's annual provision for loan losses, measured as a percentage of the average loan portfolio, was 12.7% in 1993, 3.2% in 1994, and 4.4% (3.4 times higher than the average bank) in 1995. Even though BFA is exposed to higher credit risk, its annual lending interest rate averages 309 basis points lower than ithat of the average bank. BFA has increased its lending rate in real terms, reaching positive values in 1995. Its interest rate policy is not consistent with its credit risks.24 4.23 The bank was able to improve the quality of its loan portfolio over the last three years, after a long history of delinquency problems (IPC, 1993; Abt., 1993). Reserve for loan losses as a percentage of the average annual loan portfolio serves as a proxy of the riskiness of the loan portfolio and the adequacy of loan loss reserves when compared to the risk classification of loans. This indicator was 34.9% of loans in 1993, 30.6% in 1994, and 26.1% in 1995. Loans classified in risk categories D and E represented 27.9% of risk assets, or 8.2 times than the corresponding figure for the bianking system as a whole (3.4%). 4.24 BFA has implemented the following actions to increase loan repayment: (a) creation of a "gerencia de recuperaci6n"; (b) installation of new software to keep track of defaulted loans, classify loans according to risk, and monitor loan disbursement; and (c) creation of an incentive system for prompt payment (BFA Memorias, 1995). 4.25 BFA's delinquency piroblems can be seen as a by-product of the agrarian reform that started in 1980. The bank was then forced to grant loans to recently-formed cooperatives without experience in farm management, generating in this way, the so-called "adelantos-ISTA-BFA" delinquent portfolio. When Phase II of the agrarian reform started in 1980, BFA's policy of granting loans to small beneficiaries further inflated the delinquent portfolio and generated a pattern of repayment problems which worsened the viability prospects of the bank. 24 IPC (1993) summarizes the major obstacles for BFA to raise its lending rates. 26 El.Salvador. 4.26 The composition of BFA's portfolio reflects its driven mandates and Manufacturing Other the political interference it has suffered Comnmerce 5% 5% over the years. Figure 4-1 indicates that 10% , - 54.6% of total outstanding loan balances were used to finance agriculture activities. The other activities financed Agricuture by BFA are: other sectors (25.7%) which Other ( kedi_ 54% included refinanced loans, commerce (Refinancing) (9.9%), manufacturing (4.6%), and 26% services (2.9%). A close look at the "other" category may uncover diverse forms of politically motivated lending. Figure 4-1 BFA: Distribution of Loans by Economic Sector, 9/95 4.27 Capital Adeqiuacv. Government transfers to restore BFA's capital resulted in a strong capital position that is consistent with the risky profile of its loan portfolio. However, the real value of its capital has been declining since 1991. During the last three years equity averaged real growth of -8.2%. BFA's decapitalization results from low levels of net income that are insufficient to compensate for inflation. 4.28 The bank had an unadjusted equity to assets ratio of 39% and a risk-adjusted capital to assets ratio of 46% 1995-the second highest among commercial banks.25 Total capital was 63.4% of total liabilities and off balance sheet items, which is almost 16 times higher than the minimun required by law and 6.9 times higher than the average bank. 4.29 Operational Management. BFA has improved management of its administrative expenses during 1993-95, but continues to have lower productivity indicators than the average bank. As a percentage of its average loan portfolio, administrative expenses of BFA reached 9.2% in 1995, almost two times higher than the average bank. BFA's administrative costs were equivalent to 12.4% of its average annual net loan portfolio, as compared to 3.5% for banks. Given its extensive network of rural branches, centralized operations, and small loans it would be expected that the BFA would have high overhead expenses. Yaron (1992) explains that operational costs of 5 to 6% of the average annual loan portfolio are reasonable for an institution that targets low-income rural producers, while in practice operation costs of 8 to 10% arc commonly observed in development banks in Latin America (Cuevas and Poyo, 1989). 4.30 The productivity of BFA's employees is lower than that of the average bank. BFA employed almost four times more people relative to its assets than the average bank. The BFA had 1,463 employees, of which about 500 are located at the main office, with administrative costs of 121.5 million. The number of employees increased in 1995 to handle activities related to the Plan de Reconstrucci6n Nacional (PRN). Another indicator of the low productivity of BFA is the number of loans per employees. At the end of 1995, each BFA employee serviced on average 21 loans, less than one-sixth of the average for Financiera Calpia which also serves low income clients. 4.31 Ownership and Organizational Structure. An important factor that affect the performance of the BFA is its governanoe structure. BFA is governed by a Board of Governors which is composed of the President of BFA- who, along with the vice-president, is named by the President of the Republic-representatives from the Ministry of Economy, officials from the cooperative sectors and the President of the Central Bank. The current governance structure pennits members of the Board to use BFA for political purposes instead of focusing on its financial viability. The Board of Governors sets policies that are implemented by the Board of Directors and the top management of the bank. 25 The banking law requires a proportion between equity and weighted assets of 8%. Rural Finance 27 4.32 Borrowed Funds. Borrowed funds from BMI and foreign institutions, as a source of liquidity, increased from 18.8% of assets in 1993 to 28.7% in 1995. BMI loans to BFA represented 99.3% of borrowed funds in 1994, while in 1995 they accounted for only 67.8%. The relative decline in BMI funds has been brought about by an increase in borrowings from foreign institutions. 4.33 Savinns Mobilization. The "savings idea" has been around BFA for a long time, and in recent years it has resulted in significant changes, of its liability structure. This change is the result of BFA's official policy to change the image of the bank from a lending window to a financial intermediary that depends on savings mobilization to support its credit operations. The savings mobilization efforts have increased deposits from 13.8% of assets in 1993 to 29% in 1995. Even though deposits relative to assets in the average bank are 2.5 times larger than in BFA, they have been growing at BFA in real terms 15.3 times faster than deposits at the average bank. According to BFA officials, the bank has opened 15 sub-agencias that offer savings opportunities in rural areas. Women are an important target of the molbilization program, as reflected by the regular raffles of equipment such as sewing machines and planchas de pup7usas used by the bank to promote savings. 4.34 Savings mobilization always create liquidity risks due to unexpected withdrawals of funds. BFA is in a better position than the average bank to cover cash requests from depositors and borrowers. It had 44.2% of its deposits in commercial banks and BCR, while the average bank had only a third of its deposits available as liquid assets. Increased savings mobilization has resulted in an increase of interest expenses with respect to interest income. This ratio increased from 26.4% in 1993 to 32.6% in 1995, due to the reduction of the relative importance of subsidized borrowed funds ifrom BMI. 4.35 Index of Subsidy Dependence. The subsidy dependence index was calculated for the period 1993-95 to obtain an indication of whether BFA has been making progress in reducing its dependence in subsidies. The calculated SDI makes the foll,owing assumptions: (a) miscellaneous subsidies are zero; (b) reserve requirements are not remunerated; and (c) loans from foreign institutions are not acquired at concessional rates. It is expected that the total amount of subsi,dies is underestimated because it is likely that loans from foreign institutions are borrowed at concessional rates, and that BFA has received direct or indirect subsidies from the Government. In any case, the estimation of SDI can be useful in revealing general trends of subsidy dependence and should be considered as a lower bound. 4.36 The SDI calculations indicate that BFA has worsened its subsidy dependence performance during 1993- 1995 and had a moderate degree of subsidy dependence in 1995 (100.8%). This means that BFA would have had to increase its average lending interest rate by 100.8%, from 20% to 40.7%, to eliminate subsidies. BFA increased its dependence on subsidies due to low levels of net income, and an increase of concessional borrowed funds. 4.37 Trust Funds Administration. The administration of these trust funds increases overhead expenses for BFA. Among the four trust funds administrated by BFA, the Plan de Reconstrucci6n Nacional (PRN) accounts for 92.1% of the loans granted from these funds (BFA Memorias 1995). The number of beneficiaries since the beginning of the PRN reached 25,277 of which 15,889 received disbursements during 1995. According to BFA's officials, in 1995, the bank opened 9 branches and expended 40% of total operating expense to serve these beneficiaries. It needs to be determined whether BFA receives any monetary benefits from this activity. In any event, linking BFA to PRN has perverse reputation effects on BFA's loan recovery rates. 4.38 Summary and Evaluation. In 1995, BFA faced the same structural problems that it did in 1992. These problems are: (a) increasing SDI; (b) significant Government transfers have been necessary to keep BFA afloat; (c) unsatisfactory and declining rate of return on assets; (d) high operating expenses which are well above those of the average private bank; (e) despite government interventions to rehabilitate BFA's portfolio, its credit risk surpasses that of the average bank for all financial ratios; (f) while management has cut administrative costs during 1993-95, BFA continues to have lower productivity indicators than the average bank; (g) the current governance structure is amenable to politicization of decisions; and (h) savings mobilization efforts have increased deposits from 14% of assets in 1993 to 29% in 1995. 28 El Salvador. Performance of Client-Owned Organizations 4.39 Two independent systems of credit cooperatives developed independently in El Salvador. These systems are organized under the umbrella of their corresponding federations: FEDECREDITO and FEDECACES. The similarities and differences between these two systems are presented in Box 4-2. 4.40 The performance of these two systems in terms of return on assets (ROA) and return on equity (ROE) is not satisfactoiy and has been declining for FEDECREDITO and FEDECACES affiliates. The common factors affecting the lprofitability of these intermediaries are high administrative expenses and credit risk management which are documented below. After adjusting FEDECREDITO's income statement, it had a ROE of -0.7% the lowest among all FIs.26 4.41 Credit Risk Management. The credit risk of FEDECREDITO, and FEDECACES surpasses that of the average bank. The reserve for loan losses accounted for 7.5% of its gross outstanding loans for FEDECREDITO, 4.8% for FEDECACES, and 5.8% for FEDECREDITO affiliates. The two federations have implemented actions to increase loan repayments. FEDECACES's measures include: supervision to the savings and credit cooperatives (SCCs), technical assistance to SCCs that have shown deficiencies in credit risk management, and refinancing. The actual credit risk faced by the two systems (the federations and their affiliates) and the adequacy of the reserve for loan losses can not be properly evaluated. There is a need to improve the accounting procedures of the institutions by providing information on rescheduled loans. The reporting of arrears is generally poor. 4.42 Capital Adequacy. In 1995, the two federations had negative real growth rate of equity, however their affiliates showed positive real growth rate. The federations are eroding their capital base due to low levels of net income and high operating expenses. The unadjusted equity to assets ratio of FEDECACES, FEDECREDITO and FEDECREDITO affiliates increased during 1993-95. 4.43 Operational Manatement. The operational expenses percentages calculated for these institutions are high, however, they show promising trends.27 These estimates indicate a lower bound because some of the institutions have received subsidized technical assistance. The data show a wide range of operational costs as measured against average annual assets. For 1995, these operational costs varied from 5.6% (FEDECACES affiliates) to 8.7% (FEDECACES federation) of total assets. Box 4-2 Similarities and Differences between FEDECREDITO and FEDECACES Systems ffi ffi E . . . . ... . EE E i * = t ^ ....................ii .........Eii ..E.hREnsafiAense. . . . . . . . . EE EEE i vlrEson' :E:iEiE : e iii i ........................-.................. . ......................I........... . . ................E REE lZw; X Savings Mobilization. The FEDECACES.systm ha. Cifent-Owned. The: pr-blemns. of o ..10 own l-:n-stIlon- - n-shown a greater ephas7s on savings mbili2ation than theS f n which are increased by-the rule of oe-an-oevote are f :EDECREDITO system. present in both federations-and their affiliates which result in.. * hAflits In toFEDECREDITO low le o4 io:Anternal monin (a 19. ai r ..... mu ie dential supertvisionisu ................is.required anFEDECACS affiliates arecreatedby local initiativef mn do e Unfficato E Te tw ooperative systems of EI no .need appa cI FEDEACES. Sa0000:0\lvador00 ar X tadng tsimilar stetors th ntgrtiont ofa ..n '.-......r.n: T e is no minimum t.................f............. - ;filiate, .ep -ar (a)t .............-1 u n torat a savins and creditcorae acountingsof tirafilte: (b).Imp su..r but:the ----i..u.num number.:f: f of members I; fteen, activities; .(c)raft.ing of newa leva iframewr f- the .................DE CES iI t i s (d) traI ...................ning acivities of enl and(s) ..t--o.ofeach -meme. -i nt, at, 2.O s retuired * tIA. tot h . e :to . .. . ;. ; ; . . ... ...;. .. .. .. .. t...,,:,. .create.a c.. o crDi.l Thne 8CR drarr law proposed to otain..delegat. superv.so. y role frorn the .. te miore rAis hslnit 0 000 -eroranimfuheinafiiaes ....... ...... .. ...... ..i p ow. -0-00 --- ;"n ;; 0* tWHIM -" *W - E:- r :l t ;i;j; 4.44 FEDECREDITO, and the affiliates of FEDECREDITO and FEDECACES improved management of thieir administrative expenses during 1993-95. While steadily improving, the productivity of FEDECREDITO's employees was lower than that of the average bank. FEDECACES affiliates presented the lower level of 26 FEI)ECREDlTO has the policy of accounting as revenue the reversion of reserves. 27Provision for loan losses are not included as operational expenses. Rural Finance 29 administrative expenses as a percentage of assets (5.6%) than FEDECREDITO affiliates (7.6%). FEDECACES activities such as staff training contributes to lower operational expenses of its affiliates, but its ability to control administrative expenses at the federation deteriorated during 1993-95. As a share of average annual assets, these costs went from 6.8% in 1993 to 8.7% in 1995. The ratio of operational expenses and assets of FEDECREDITO affiliates (7.6%) is higher than their federation (5. 1%). 4.45 Index of Subsidy Dependence. The Subsidy Dependence Index (SDI) was only computed for the federations because of the shortage of available data. The data indicate that the two federations differed in their level of subsidy dependence. Furthermore, FEDECREDITO increased its level of subsidy dependence during 1993-95. FEDECACES at ithe federation level had a high degree of subsidy dependence in 1995 (234.5%). This means that FEDECACES would have had to increase its average lending interest rate by 234.5, from 11.9 to 39.7%, to eliminate subsidies its dependence of subsidies. The structure of the FEDECACES network and its functioning makes this indicator misleading, however. While the unit of operation and performance is the affiliate credit union, the federation is the official recipient of technical assistance grants and soft loans. A more accurate, and meaningful, analysis of the subsidy dependence of the network would require further detail in the data available. Box 4-3 FEDECACES The Fe,eratlon of Savns: and Cr'di' Cooperati'es (FEDECACES) Is a legjally constituted d''e '4 of sa lnd credit. ...oope.ti.e (as.c, clones cooperatives de ahorro y.cr6ofto, SCCs) founded ..in..Ju 9..ne ,19 6. It.. seco-a n-ier organat atrovie... nnl.. s..t affiliatedandd non-afliatedsavings and cd erats i A ,1 I : .tol f 352 -credit unions. are registered with INSAFOCOOP, of which 150 are active. .(Fundaungo, 1.995). mlyioh t SCCs ar 4'f.0- e wianh;FEQECACES ad credit cooperatives affiliated wth FEDECACES a r p n reguated 42m0ormalfinanci s t with tal ets of,¢ 42.4 millIon n 1995 these coo prativ had '38,406 mbs Afte pass t h te se of erng Xw anempis on. savings mobili FE AS hs as is objeties a}ocandenoic:1nte§ratlon f .the siaving ad crdit coopratives of El Salvador; (b) finiancial strengtheningoth l oeratrves thro h. h est ishmert of systems which allow them t ' a 'ttrt inve't and effiently m anag an a i resources, and p:) promotionof the organiation, affiliation and development of new saving and credit cooperatives. F EDECACES has set as its new role ^.the.* esta ment tof a sytem of cooperatv financial intermediation." Howve theaffilits e ve dive'i s ad a t lgre .wth thnew :ro. of FEDECACES. Consequentty, the achievement of the proposed new role may be:thten. In an event the inana viab"`b of FEDECACES affliates must be strengthen before any measures are taken towards integration. 4.46 Ownership and Organizational Structure. The current governance structure of FEDECREDITO allows political interference in the financial decisions of this system of client-owned institutions. For example, almost half of the members governing board which is responsible for setting financial policies represent the Ministries of Economy and Agriculture, and BCR. The president of this board (named by the President of the Republic) along with managers for administration, finance, credit and operations are responsible for planning, implementing and monitoring daily operations and compliance with laws and financial regulations (Abt., 1993). The owners' lack of control and management of FEDECREDITO has weakened its financial viability. The proposed project of reform of the FEDECREDITO will address this deficiency by restoring the control of the organization to its members. MAIN ISSUES 4.47 The FIs reviewed have many similarities and differences. All three institutions started as supply-led FIs focusing in the delivery of credit services rather than providing savings facilities. The BFA and FEDECREDITO system have been subject to strong political interference and the imposition of special "social" mandates clearly dominates the institutional philosophy and their operations. 4.48 Remilatorv Framework and On-going Reforms. The legal frameworks of the Fls is fragmented into three laws, one for each Fl. Even though, the BFA is regulated as a commercial bank, its governance structure and mandates are clearly defined by the Ley Orgdnica del BFA. Due to the similarities between the FEDECACES and FEDECREDITO systems the Government is considering the unification of the regulation, governance, charter, and status of these systems. Such unification would imply the emergence of a single organizational form for the country's credit cooperatives and would include them under the supervision of the SSF. The implications of such unification are addressed in Chapter 6. 30 El Salvador: 4.49 Decapitalization. The BFA and the two federations of cooperatives are eroding their capital in real terms. Even though FElDECREDITO's capital declined in real terms last year, the capital of its affiliates grew 16.7% in real terms in the same year. FEDECACES (federation) experienced the largest declined of capital in real terms but its affiliates show the lowest decline of all Fls. 4.50 Operational Expenses. The performance of FIs in terms of administrative costs measured against the FI's assets varied suabstantially. The difference in administrative costs between FEDECACES and FEDECREDITO affiliates can be explained, in part, by their different average loan size. In general, credit programs which involved the handling of small loans are more likely to incur higher administrative costs. The BFA has improved the management of its administrative expenses during 1993-95, but continues to have lower productivity indicators than other FIs, but FEDECACES. The productivity of all FIs' employees is lower than that of the average bank. However, FEDECREDITO had a higher value of assets per employees than BFA and FEDECACES. Chapter 5. THE ENVIRONMENT FOR FINANCIAL TRANSACTIONS, AND NEW FINANCING INSTRUMENTS 5.1 This chapter turns to the question of what are some of the more appropriate vehicle agencies and products to deliver quality financial services to under-served sectors of the population. What financial products, including non-bank credit products, offer possibilities for expanding rural credit, reduce its cost, or alleviate information asymmetry problems? Are there legal, regulatory, or policy barriers to the development of such products? If so, what can be done to remove them? 5.2 The chapter reviews first the status of the overall business environment for financial transactions, drawing upon the 1995 CEM and its background material, especially Hanson (1993), mission interviews and observations. Furthermore, specific barriers affecting new, especially non-credit, financial products are discussed and reforms are proposed. We identify opportunities for additional reforms in leasing, housing and mortgage finance, collateral registries, and the use of electronic cards. Among these, the modernization of registries, with private sector participation, and the integration of technical standards and networks for electronic transfers appear as the most promising. ENVI1RONMENT FOR FINANCIAL TRANSACTIONS 5.3 The business environment for financial transactions may have a discriminating effect against the poor (Fleisig, 1995). Examples of laws with discriminatory effects are: limits or barriers to using movable property, accounts receivables, and loans as collateral; and requirements that loan contracts must be in writing and signed by the borrower. Barriers to using movable property limit access to credit for those individuals or enterprises without real estate properties. 5.4 Compared to many other developing countries, El Salvador has done relatively well in creating the preconditions required for growth in the private sector. Its laws and regulations have been found generally adequate for commercial transactions, including the borrowing of investment funds from formal institutions. No fixed obstacles exist in the Commercial Code (CC) or other laws that actively prohibit basic, contemporary financing arrangements. Rather, other obstacles, such as lack of information networks, and lack of explicit legal guidelines for advanced financing, suppress much financing activity. A variety of new laws formalizing leasing and the factoring of accounts receivable, as well as slight modifications to the Commercial Code, are needed to Rural Finance 31 enable new financing opportunities. Enforcement on a variety of levels, however, remains weak and will still require improvement Table 6-1 El Salvador: Type of Collateral and Legislation Type of Collateral Financial Contract El Salvador Movable property Lcoans, Leasing Collateral registration is lengthy. Registries need to be improved. The CC does not have an explicit provision for leasin. ~~~~~~~~~~............ .......................................I............................................. ,,..., ,,., , ..., .................., :.. ........... Accounts receivable Factoring, chaKtel paper Accounts receivable alone do not have legal standing; they must financing be formalized as pagares. Receivables may not be held and Warehousing receipts Supplier-trade credit Receipts can be endorsed and used as collateral. There are .p~~~~~~~~~~~~~~~~~roblems of qualitv, standarizatlon and renutation. ............................................ .......................................... ,p ..................... Inventories Floating liens on inventory The CC does allow liens on inventory, but the provisions are limited (Art. 839). It is costly to create a secunty interest on inventory because it requires registering every item in the inventory to secure it. The borrower and the lender must check the accuracy of the list1 which is expensive and cumbersome. Negotiable instruments: Bill of Trade Credit It is relatively easy to create a security interest exchange (letra de cambio), promissory notes (pagar6s), and checks (cheques). Source: Hanson (1993) 5.5 In El Salvador, acceptable collateral may be registered at their appropriate registries. This point is emphasized because previous reports have claimed that certain types of security cannot be registered and that the laws of El Salvador should be changed to include them, for example bills of exchange (or checks), letters of credit, and warehouse receipts. While the law allows creditors to take these instruments as security for credit, it is also true that they may not be registered in the Commercial Registry under Article 4 of the Regulations. However, it should be noted that not all documentary security interests need to be filed at the registry in order to be secure, and that it is in fact inefficient to do so, as possession of the original document in some cases is required to collect on the pledged assets. In fact, when lawyers in El Salvador are asked about the limitations on registering bills of exchange, none saw it as a problem, as such instruments wouldn't be registered anyway. Table 6-1 summarizes the status of regulations pertainin,g to different types of collateral and financial contracts. NEW FINANCING INSTRUMENTS 5.6 Governments have an important role to play in creating appropriate incentives for the introduction of modem, innovative methods of financial services provision in rural areas. Local expertise exist and local skills can be further developed to provide required human capital. Furthermore, in the overall picture of donor activity, promoting the transfer of financial technology and improved management information systems, and financing start- up costs, represent relatively small amounts of grant financing or soft funding with high payoffs in terms of enlarged, effective outreach to under-served clients. 5.7 In addition to capacily building alternatives, it is necessary to consider alternative approaches based on two observations. First, since credit is fungible, the problem of financing rural small-scale enterprises must be thought of in the larger context of all types of credits available to small business owners-considering both consumer and business credits, and credits from the banking and non-banking sectors. Such alternatives may well involve consideration of some financial instruments and markets that are not particularly focused on and are certainly not restricted to rural or microenterprise-oriented credits. Second, financial products and technologies that permit reduction in the large transaction costs of making small bank credit should be considered. 5.8 Leasing. The leasing industry can be an important provider of microenterprise credit in addition to the commercial banking system. (Commercial banks can themselves be important participants in the leasing industry either through on-balance-sheet leasing or through leasing subsidiaries.) Leasing can provide point of purchase credit, up to 100% financing, greater assurance against technical obsolescence, and reduce debt to equity ratios compared to straight debt financing. These factors make leasing a particularly attractive form of financing for 32 El Salvador: microenterprises, farmers, or rural industries. In turn, lease receivables are an attractive asset to securitize and leasing companies can be important commercial bank clients. Leasing accounts for between 15-35% of all business equipment financing in many countries of the world. Yet, the leasing industry remains quite underdeveloped in El Salvador. 5.9 There are some relatively minor problems with leasing in El Salvador. First, the Salvadoran commercial code does not explicitly deal with leasing28. While the current legal situation does not bar leasing transactions, and some do occur, many participants do remain somewhat concerned (perhaps needlessly or excessively) about executing lease transactions. Second, the market for leasing, particularly operating leases, would remain rather limited given the small size of the Salvadoran market and the difficulties of trading used goods. Third, it is necessary to clarify the relevant tax treatment and accord a tax treatment that encourages leasing without subsidizing it. 5.10 Home and Real Estate Loan Markets. The development of a home mortgage market offers an indirect but very important possibility of reducing microenterprise credit constraints in general. In less developed home finance markets, a large part, possibly the whole, of the home investment is equity-financed. The proportion of personal equity financing is typically higher at lower income levels. Thus, the development of a home mortgage market and home equity loan market can free up a large part of personal savings for possible business investment. Such loans entail much lower information asymmetries and risks for lenders, and are thus possible to make far faster and less expensively than loans to a small business. Clearly, such loans may finance investments as well as greater consumption. But because home stocks are so large (2-4 times GNP), even a modest debt financing of, say, 20% of the aggregate housing stocks can expand aggregate credits by an order of magnitude. Such an increase cannot but relieve some of the financing constraints on business investment. In some developed countries, home mortgage loans equal or well exceed the total assets of the banking system. 5.11 There is considerable potential for liberating the personal equity of a large fraction of the population through home mortgages in Salvador. However, the extent to which such additional credit will trickle down to the lower income rural or agricultural borrowers depends on several factors including the unit value, formality, registration, ease and cost of mortgage documentation, notarizing, etc.. Clearly, the more valuable the home units are, the easier to finance. Nonetheless, mortgage loans are being successfully made for units worth $3,000-5,000 in Peru. There are significant problems in recording properties and liens thereon. El Salvador is divided into 14 departments, which are covered by 10 property registries that constitute Registro Propiedad de Raiz. These registries are not automated, have serious record-keeping and retrieval problems, and currently it can take as much as 12 months to get a property registered. Some efforts are being made to develop modern registries with the collaboration of private sector agencies. For instance, Ahorromet, a finance company with strong mortgage interest, has collaborated in developing a modem computerized registry. There is a need for modernizing registries, both for real estate as well as for commercial transactions. 5.12 Possibility of an Asset Backed Securities Market. Housing finance cannot develop adequately only through the banking system, since financing long term housing loans with bank deposits would leave the banks vulnerable to serious maturity mismatch. El Salvador has attempted to overcome this problem by offering special concessions (lower legal reserve requirements and higher interest rates on such reserves) on long term deposits which are utilized for housing finance (Chapter 2). While such attempts are understandable, they create segmented financial markets based on specific subsidies. Ideally, banks should originate and package the mortgage loans, selling them off to bond market investors as mortgage-backed securities. Such a development would be highly desirable for the development of bond markets and institutional investors. However, mortgage- and asset-backed securities market is a relatively new financial innovation which requires a facilitating legal framework which has not been developed in Salvador. Countries with Napoleonic law traditions typically encounter severe problems in 2 The Law of Financial Institutions (passed in 1989) has long been slated for a substantial overhaul since 1991 when the draft of an extensive revision of this law was prepared. The amendment has not beeni passed for a variety of reasons. The draft provides for a much more complete regulation of the leasing industry, thereby further deepening concerns regarding the legal status of transactions entered into prior to passing the amendment. Rural Finance 33 developing such markets due to the notary system, which is usually a very restrictive and expensive franchise. In these countries, the creation of a special purpose vehicle is needed for issuance of mortgage-backed securities, and problems of multiple taxation of the same transaction arise. It would be highly desirable to identify the extent of similar problems in El Salvador and take remedial action. 5.13 Supplier-trade Credit. Suppliers and trading institutions such as equipment suppliers, feedlots, abattoirs, packing, refrigerating and other processing plants could have an important role in granting credit to rural and small business. These institutions who have natural business relationships with rural entrepreneurs have greater information about their clients and a better ability to deter default and possess and resell collateral than commercial banks. 5.14 The development of su]pplier-trade credit requires financing of warehouse receipts, of inventory, and of accounts receivables. Licensed warehouses are authorized to issue warehouse receipts and warrants. Warehouse receipts and warrants can be endorsed, thus providing the endorsee the collateral of underlying inventory, and creating the possibilities of financing. However, the current warehousing system has several drawbacks for efficient storage of grains and their use as collateral which has hindered the acceptance of warehouse receipts as collateral (Lacroix, 1997). The current law does not contain penalties for forging, altering or counterfeiting warehouse receipts and warrants or for the unlawful use of commodities stores. Hence, there are problems of quality, standardization, and reputation that may restrict acceptability and financing of receipts of specific warehouses. 5.15 The commercial code allows use of inventory as collateral, but recording requirements are difficult and can be improved upon. A commercially registered lien on inventory requires specific listing of each item. A floating lien requires a notarized and court-authorized public document (escriturapilblica) following an expensive process. On the other hand, accounts receivable do not have a legal standing, unless the underlying claim is reduced to a pagarM (a promissory note), which is typically not the case for trade credits. 5.16 Factoring. Factoring is of particular importance to rural and agricultural microenterprises, especially those who supply to large retailers, sugar mills, or other processing plants. Because factoring provides financing based on the creditworthiness of the larger client rather than the microenterprise supplier, even suppliers with poor creditworthiness, accounting, or lack of collateral can obtain such credit. Salvadoran commercial code does not disallow factoring. There are at. least two commercial factoring agencies: Fadesa and Rapicash. The proposed amendment to the Financial Institutions Law referred to earlier also provides for fuller regulation of factoring industry. When implemented, such regulation will remove any lingering uncertainty about the status of factoring companies. 5.17 Use of Credit and Smart Cards. Use of smart cards and credit cards are expected to produce phenomenal benefits for small and rural lending. First, the cards can drastically reduce the administrative costs of making a loan for the lenders.29 Second, the increased use of smart cards and credit cards (which may be thought of as single loan, not so smart, electronic cards) reduce the system-wide cash float with individuals and increase the level of bank deposits. Third, it allows banks to forge profitable business ties with suppliers to their card-holder clients (with the suppliers effectively acting as disbursement agents for the bank), increasing their willingness to finance such suppliers' accounts receivables. But most important benefits are likely to emerge as the banks start systematically analyzing the considerably more informative histories of withdrawals, credit use, and repayment patterns generated by the cards. Over time, intelligent analysis of these data can dramatically reduce the cost and problems of information asymmetry, permitting banks to identify more responsible and regular customers. Credit cards permit the lenders (issuing banks) to make loans based on very inexpensive, but limited and thus high risk, credit evaluations. The risks are controlled through tiny credit extension initially, perhaps smaller even than the 29 Financiera Calpia reported that smart cards reduced its administrative costs of small loans by as much as 6% of the loan amount annually, in the first nine months of its introduction. Cards are used to make loans varying between ¢500 to 0200,000 and they can be programmed to process several loans (Calpia uses them to make three working capital loans, one investment loan, and one seasonal working capital loan), with complex disbursement criteria. 34 El Salvador: cost of a thorough credit evaluation itself, with higher credits possible only through a track record of responsible payments. 5.18 Electronic cards benefit rural and other microenterprises in many ways. As businesses, increased use of card payments allows microenterprises to extend relatively inexpensive credit to their own clientele and raise their sales. Second, they benefit from savings in bank administrative costs that are passed on to them. Third, the cards can dramatically reduce the borrowers' own transaction costs; including the cost of unnecessary float by providing borrowers with point-of-purchase loans, the costs of visiting the bank, and some of the costs of providing information to the lender. 5.19 The benefits of this modem technology are limited by short-sighted commercial strategies used by Salvadoran institutions. Credit cards, smart cards, ATMs share a substantial common infrastructure. Thus sharing of common standards and networks would greatly increase the efficacy of the whole system and its economies of scale. There are about 200,000 electronic cards in El Salvador now. There are several competing proprietary ATM networks in El Salvador using incompatible standards, and thus limiting access of customers. There are no technological problems in integrating them. Some financial institutions are issuing electronic cards that can only be used within their own branch network. Others limit issuance of smart cards to their best clients (due to prestige connotations). These practices seriously limit the use and cost savings potential of these payment methods. There is thus a need for the Central Bank to take the lead in the development of integrated standards and networks, including perhaps development of screens that would facilitate transactions by illiterate users. 5.20 Further development of this industry, and reducing its costs, would require appropriate legal structures for accurate identification of borrowers, sharing credit history information across lenders and borrowing instruments, development of credit rating bureaus, cooperation among lenders to avoid wasteful duplication of products, and lowering the cosis of underlying technology. These developmental measures are necessary to deepen financial markets generally. 5.21 In summary, opportunities for additional profitable reforms exist in leasing, housing and mortgage finance, collateral registries, supplier credits, and the use of electronic cards: * In leasing, the principal reform is clarification of the relevant enabling framework and tax treatment, perhaps through passage of the amendment of the Financial Institutions Law. * Housing finamce needs to be further developed by appropriate study and development of a facilitating framework for asset-backed securities, reduction of notary costs, facilitating a special purpose vehicle, and avoiding multiple taxes. This market has traditionally faced severe problems in most Napoleonic code countries, although the needed reforms have not been studied in El Salvador. * Real asset registries and commercial registries need to be modernized to reduce the existing large costs and delays of registering homes, mortgages, liens and other security interests. Promising reforms involving private sector participation need to be replicated elsewhere in the country. * In terms of supplier credits, principal reforms would be easing recording requirements for using stocks as collateral, and reforms of the commercial code to permit accounts receivables to be used as collateral. * There is considerable potential for reducing costs of small credits economy-wide by expanding the use of electronic payments, credit cards and smart cards. The Central Bank should take the lead in integrating relevant technical standards and current competing and incompatible networks. There is also a greater need for education of intermediaries in exploiting the full potential of these technologies by not limiting its judicious application to a wider clientele. Rural Finance 35 Chapter 6. RECOM[MENDATIONS 6.1 El Salvador has made significant progress since the early l990s creating adequate conditions for private sector growth. In particular, the substantive reforms and liberalization of the financial sector have enabled a conducive environment for financial intermediation where institutions can function without the burden of excessive regulation and pricing restrictions. However, the exclusion of entire borrower classes is apparent in the commercial bank (non-BFA) loan portfolio. As shown in Chapter 3, RCMs in El Salvador are shallow because rural households have limited access to credit services-less than 12% of rural households received a loan in 1995 and about 20%h had outstanding debt balances from formal or informal sources. The weak participation of rural households in credit markets is explained by both demand- and supply-side factors. 6.2 Some form of govermnent intervention may be desirable, but traditional solutions to increase the supply of credit in rural areas have had tudesirable outcomes. Traditional solutions in Asia, Africa, and Latin America have attempted to augment the supplly of credit through a variety of means. Administrative or fiscal initiatives include directed credits, usury laws, interest rate ceilings, explicit or implicit subsidies, tax breaks for borrowings used for small-enterprise loans, among others. Directed credit programs have misallocated resources and tend to compromise prudential superAision of banks involved and increase pressures for regulatory forbearance. Specialized banks have suffered from a myriad of problems of misdirected credits, highly non-diversified portfolio, non-businesslike operations of the specialized institutions, bad loans, and perpetual dependence on specific government or external funding. 6.3 There is no single proven model for El Salvador to follow in order to deepen its rural financial markets. The failure of past governnent intervention to improve the performance of financial markets may be overcome if the Government formulates adequate strategies to deal with the issues at hand. The development of rural financial markets should have two main goals: increasing the availability of viable, competitively priced, and untargeted deposit and credit services from formal financial intermediaries and promoting competition in the sector. This objective is tantamount to increasing the outreach and the sustainability of formal financial intermediaries rural areas. Such an effort would increase overall access to rural financial markets and change the composition of that market niche by including entrepreneurs whom the formal sector does not currently serve. A combination of strengthening the economic environment and the distribution network can help us meet both goals. The recommendations are organizeci into these two areas. ECONOMIC ENVIRONMENT 6.4 An economic environment in which property rights, contracts, and financial services can prosper is the Government's primary responsibility. Such an environment consist of improved legal framework for enforcing contracts and making use of collateral; networks for sharing information between lenders (e.g., credit checks); regulatory framework; and financial policies and competition. Specific recommendations are: 6.5 Collateral security issues. The legal framework affects the cost of transactions, the characteristics of credit contracts, and the access of small clients to financial services. The Government could increase access to credit services by reducing the costs of using collateral to secure credit transactions. To reduce cost of collateral use it is necessary to: (a) improve legal framework to facilitate the creation, perfeetion, and enforcement of security interests; and (b) modernize registries to reduce transaction costs and delays of registering security interests such as mortgages and liens. The restructuring of the public registries should be coordinated with the corresponding legal reforms, which would create the necessary pre-conditions for well functioning credit markets. Furthermore, consider 1995 GEM recommendations to expand use of liens on inventory and commercial equipment. 36 El Salvador: 6.6 New fmancinR instruments. Because the Government has done little to promote the use of new financing instruments, the general guidelines proposed in the 1995 CEM would be appropriated. The three main recommendations to expand use of new financing instruments are: (a) improve the environment and legal framework for new instruments such as leasing and factoring; (b) consolidate the reforms currently in progress, and the study of new ones and amendments to the Financial Institutions Law that clarify the rules that govern leasing should be adopted; and (c) set guidelines and uniform treatment for leasing, factoring and discount of commercial paper. 6.7 Financial sector supervision and competitiveness. To improve financial sector competitiveness the Government should implement two strategies. First, the liberalization of the banking system could be consolidated especially through: consolidation of ownership restrictions; implementation of a rating system for financial institutions; and istrengthening of SSF through training and modernization. Furthermore, the recommendations of the 1995 CEM can be used as a checklist. Second, the Government should remove from the menu of policy interventions debt refinancing/forgiveness programs as an effort to restore repayment discipline in rural areas. Also effective prudential regulation and supervision is particularly important in El Salvador due to the entry of new financial institutions and the rapid expansion of credit during 1991-1995. DISTRIBUTION NETWORK 6.9 The secmnd area of policy intervention relates to the restructuring of the distribution network of financial services. Any type of institutional re-structuring requires, among other factors, an adoption of cost-effective financial innovations to provide financial services appropriate to rural markets, and measures that enable the use of new financing instruments. Provided appropriate conditions are present, three non-exclusive institutional models have proven successful in other countries. The recommendations below are based on these three non-exclusive alternatives detailed below: * The full privatization of a public bank, or a clearly defined split of a public financial institutions, akin to the Unit Desa system of the Bank Rakyat Indonesia (BRI). * Thorough reform of the regulation and legal structure of the credit cooperative system to eliminate the governance problems that have plagued such organizational forms. * The down-scaling of operations among bank and non-bank commercial intermediaries, following the path of banks such as Colombia's Caja,Social, and non-banks such as El Salvador's own Financiera Calpia. Governance, Regulation and Supervision of BFA, FEDECREDITO and FEDECACES 6.10 BFA. As presented Chapter 4, the central issue dominating the BFA's operations are governance, political interference and its imposition of special social mandates. Although there is ample room for improving managerial practices and information systems, and for adopting better lending and savings mobilization practices, serious doubts will always remain with respect to the incentives of BFA's management to implement such improvements under the existing government ownership and governance structure. 6.11 The imrediate step should be to eliminate all non-financial operations assigned to BFA, and transfer away from the bank all welfare-related mandates. The ensuing actions would be a rigorous audit to determine the market value of BFA's equity and a detailed analysis of the spectrum of options for its divestiture -including the possibility framework in order to privatize some of BFA's branch network as independent local unit banks. The regulatory implications of this possibility are analyzed below. An implication of the proposed strategy is to suspend BFA's current efforts and investments in information technology. At this stage it is not known, for example, whether any investments made to integrate the branch network's information-nation systems will be consistent with an optimal strategy for BFA's divestiture (e.g., independent unit banks). Rural Finance 37 6.12 Is it possible to reform BFA to make it a sustainable public development bank? To achieve sustainability, BFA would have to be compellled to adopt general principles that are necessary conditions in the sustainable provision of financial services to small and medium entrepreneurs in rural areas. The political cost of reforming BFA along this way may be fairly close to that of divesting the institution altogether, because any meaningful reform requires reducing or eliminating the rents extracted by three constituencies: bureaucrats, larger borrowers, and politicians. These groups would probably fight off reform with the same strength with which they would oppose divestiture. Reform is a gradual, long-tenn and reversible process that puts time on the side of these constituencies to organize opposition. In contrast, divestiture may be achieved swiftly and irreversibly. 6.13 The following are examples of change control actions that could be pursued until the time in which BFA is completely divested to the private sector. These actions are, in any event, required intermediate steps for divestiture: * Organizafional control: the objective should be to provide control of BFA to individuals who are not subject to political pressures and who are not representative of the bank's borrowers. This objective would require: (1) reform of the law and charler of the bank because the Board of Directors is currently integrated by political appointees (Ministers) ancl representatives of the beneficiaries (farner groups). In the meantime, the Government's political appointees should vote (in block) for measures and policies aimed at improving the bank's performance (e.g., suispend current investments in modernizing BFA). * Distribution network: BFA branches are characterized by a high fixed-cost structure and are mostly located highly populated urban centers. The fixed costs of operating branches could be reduced by reducing branch staff, and by selling buildings and reallocating branches in smaller rented offices-preferably in locations not served by commercial banks. * Credit policies and targeting: BFA should change its credit policies to increase its outreach to small rural entrepreneurs and, more importantly, to protect itself from rent-seeking organized groups of borrowers. In particular, it should significantly reduce the maximum amount of credit granted to a single borrower (e.g., maximum of 4 times the coumtry's per capita income). It should increase interest rates significantly and finance all types of rural entrepreneuars (not only farmers). This policy, combined with collection incentives on the part of branch managers (mostly the threat of closing problem branches), would reduce the amount of rents an individual borrower may extract and, hence, would also reduce the incentives to establish and organize opposition to the reforms. - Credible sink-or-swim straiegy: the Government should make credible the fact that the resources currently at the disposal of the organization represent a once-in-a-lifetime capital endowment. It should be made clear to these groups that branches that are not sustainable will be closed within 6 months. 6.14 Financial Cooperatives (FEDECACES and FEDECREDITO). The Government is considering unification of the regulation, governance, charter and status of the savings and credit cooperatives (cooperativas de ahorro y credito), most of them affiliated to FEDECACES, and the cajas de credito affiliated to FEDECREDITO. Such unification would imply the emergence of a single organizational form for savings and credit cooperatives (SCCs), ancl the subjection of the resulting unified systems to SSF supervision, which may include deposit insurance. Moreover, the Government may chose to include this regulatory reform in a broader body of legislation to cover non-bank financial intermediaries, with or without cooperative ownership. Although there may be clear advantages to unifying the regulatory and supervisory framework for SCCs, the proposed reforms need to take into account the potential risks involved in chartering these new intermediaries. 6.15 The main governance features of the organization that would result from the proposed new structure are: ownership by both depositors and borrowers, one-member-one-vote rule for decision making by general assemblies, and voluntary association and withdrawal of members. Potentially negative implications of these governance features on the capital structure of the organization need to be explicitly regulated upon by the new legislation. Specifically, minimum capital requirements would need to be complemented with rules governing the withdrawal of share capital, and the accumulation of institutional capital. 38 El Salvador: 6.16 International experience with credit cooperatives. The international experience of financial cooperatives is rather mixed in terms of performance and viability. Supply-led cooperatives such as the uniones de credito in Mexico have resulted in financial collapse, and other networks such as the Costa Rica SCCs have required governnent bail-out after short-lived success. On the other hand, successful experiences of credit-union reform and modernization have been documented in Guatemala and the Dominican Republic, aside from the African examples of Benin, Cameroon and Togo.30 6.17 Fiscal liabilities. The Government may assume contingent liabilities from the reforms being proposed, as it does by chartering any type of financial intermediary. Even if explicit deposit insurance is not provided, the Government may be unable to avoid bailing-out SCCs in case of financial difficulty as they would be under SSF supervision and their depositors would be numerous and unsophisticated. Protecting large numbers of unsophisticated depositors, on the other hand, ought not to deter Government from diversifying the financial system, especially when the amounts involved represent a minimal share of total deposits in the financial system. 6.18 The newframework. In the new framework, the legal ability to carry out specific financial intermediation activities needs to be directly contingent upon the fulfillment of requirements and conditions set forth by the new body of legislation. Most importantly, minimum capital requirements for entering the regulated system and rules for the accumulation of institutional (non-withdrawable) capital should ensure the rapid build-up of a stable capital base. In addition, risk-weighted capital adequacy rules should at a minimum be equivalent to those required from other deposit-taking institutions. 6.19 Specific rules preventing "borrower-domination" (extraction of benefits by coalitions of members at the expense of the oirganization's financial viability) should be made part of the framework (e.g., prohibiting dividend allocation on the basis of borrower-patronage), to reinforce internal rules modern SCC organizations typically follow. 6.20 The neNv framework should establish reserves and allowances created from earnings to maintain capital levels adequate to cover potential losses from loans and investments. The legislation should also require provisions for loan losses, guided by portfolio classification rules similar to those applied to other regulated institutions. Likewise, norms on portfolio concentration and insider lending, and reserve requirements (encaje) comparable to those of other deposit-taking institutions should apply to the new institutional charter. 6.21 In the case of the current affiliates of FEDECACES and FEDECREDITO, minimum capital requirements will have obvious consequences on the magnitude of the consolidation process that will likely follow the regulatory refonns. From about 36 and 50 affiliates, respectively, the networks are likely to dwindle in number of affiliates to about a dozen each, primarily through mergers (much like the consolidation observed in the Mexico cajas populares after the 1992 reforms). Hence, the total amount of deposits now falling under SSF supervision may initially decrease, to then recover as the newly regulated institutions gain public confidence. 6.22 Prudential supervision mechanisms. Building SSF capacity and establishing prudential supervision mechanisms need to accompany any regulatory reforms implemented by the Government. Regulatory functions should not be alllocated to the federations, as it is unlikely that it would be possible for them to credibly supervise their affiliate-owners. Instead, mechanisms that supplement the SSF capacity with private external audits and examination of the SCCs should be established, to function under the guidance of a small high-quality team of SSF specialized supervisors. 6.23 A transitional period to adopt and adjust to the new legislation will be required. In addition, Government should carefully consider the future of existing financial cooperatives and other credit-granting organizations that 30 In Africa, SCCs were not exposed to the donor-induced subsidized-credit trajectory most SCCs networks in Latin America experienced in the 60s and 70s. Rural Finance 39 will not meet the new requirements. Specifically, access to BMI funds by non-regulated intermediaries should be strictly ruled by the creditor as rio explicit or implied guarantee will apply to those entities. Technology Transfer: Private Banks and Financieras 6.24 The successful experience of Financiera Calpia demonstrates that realistic loan pricing, appropriate management information systems, and innovative lending techniques can be used to effectively expand outreach to rural small-scale producers. As described in Chapter 2, Financiera Calpia developed a rural client base of almost 2,000 farmers (about 15% of its total clients in 1995) in less than 2 years, with a loan balance at the end of 1995 of 11 million colones, about 18% of Calpia's total loan portfolio. Delinquency rates in its agricultural portfolio remain at 1 to 1.5%, comparable to their urban micro and small loans. 6.25 Key to Calpia's succesis, urban and rural, have been its innovative lending techniques, and its modem information systems for general management and portfolio control. It is important to highlight the features of Calpii's lending as they radically differ from the practices of conmmercial banks in El Salvador. Indications of this striking contrast are the much hligher productivity of loan officers in Calpia, about six times that of an average bank officer, and the better quality of its portfolio, with a delinquency rate about one-fourth that of the average commercial bank.3' While bank loan officers in their agricultural units typically carry 45 to 70 clients, a Calpia credit officer works closely with 200 to 400 clients, depending upon the proportion of new borrowers included in the workload. 6.26 At present, Financiera iCalpia is the only institution in El Salvador using modem, innovative techniques and practices, along with realistic pricing of financial services and effective management information systems. It is certainly not alone, however, in Latin America. Other non-bank financial institutions successfully apply a similar approach, such as Caja de los Andes in Bolivia (as Calpia, also preceded by an NGO), and the Cajas Municipales de Ahorro y Credito in Peru. Moreover, leading commercial banks involved in microfinance, notably the Caja Social in Colombia, are adopting comparable methods to carry a large portfolio of small loans in a sustainable manner. 6.27 The question arises as to why other financial institutions in El Salvador have not adopted these modern lending methods and management practices to expand their outreach to under-served sectors. An important part of the answer to this question is the start-up investment required in technical assistance, training, and cautious subsidization. How best to implement innovative financial technologies is difficult without local trials and experimentation. Given the nature of research and development as a public good, governments could fund these trials to develop sustainable technologies for delivering financial services to small rural entrepreneurs. For example, Calpia emerged as a regulated financial intermediary after its predecessor NGO, AMPES/Servicio Crediticio had been in operations for 6 years with technical and financial assistance provided primarily by GTZ and IDB. On the other hand, large, established commercial banks (e.g., Caja Social) are able to gradually adopt these innovations with less of a need for infant-industry subsidization. 6.28 By promoting and supporting pilot programs that familiarize local financial intermediaries with successful innovations and encourage their adoption or adaptation, the Govermment could demonstrate that it is possible to supply financial services to small and micro-entrepreneurs in a sustainable matter. Such evidence will guide decision makers in the adoption of an adequate policy towards the sector, will persuade interest groups to accept such reforms, and will induce increased private sector participation to serve traditional excluded populations. This experimentation is essential because mistakes in the design and implementation of large scale reforms in rural financial markets may be extrernely expensive. The design of such an experimentation research program should consider four elements: clients and financial products, financial intermediation technologies, organizational forms, and managerial incentives.O 31 Calpia's delinquency is strictly measured as portfolio at risk. 40 El Salvador: BIBLIOGRAPHY ABANSA. Directorao y Servicios de Instituciones del Sistema Financiero de El Salvador. San Salvador. 1994. Abt. (Wenner and Umania). "Agricultural Credit Market: Assessment in El Salvador" APAP, Phase III, Technical Report No. 130". 1993. Banco Central de Reserva de El Salvador. "Leyes, Reglamentos e Instructivos del Sistema Financiero." 1990. Banco Central de Reserva, BMI. Resumen General de Solicitudes de Credito Amparadas a la Ley para la Rehabilitaci6n de los Sectores Productivos Directamente Afectados por el Conflicto, Presentadas al Sistema F-inanciero y FOSAFF1 y Solicitudes en Apelaci6n en BCR. April 1996. Banco Central de Reserva. Anteproyecto de Decreto de Reformas a la Ley de Fedecred&to. San Salvador. Julio 1995. Banco Central de Reserva. Fortalecimiento y Privatizaci6n del Sistema Financiero, San Salvador. April, 1990. Banco Central de Reserva. Programa de Reformas del ,istema Financiero. San Salvador. 1996 Banco Central de Reserva. Revista Trimestral. San Salvador. Various issues. Banco de Fomentci Agricola. Memorias de Labores. San Salvador. 1995. Benito, Carlos. "Debt Overhang and Other Barriers to Growth of Agriculture in El Salvador" Agricultural Policy Analysis Project, Phase II. Technical Report No 134, USAID. 1993 Carrandi, Alnunfo R. "Evaluaci6n de la Actividad Crediticia y Definici6n del Rol del Banco de Fomento Agropecuario en el Financiamiento Agricola." Arizona State University. December 1992. Chaves, Rodrigo and Claudio Gonzalez-Vega. "The Design of Successful Rural Financial Intermediaries: Evidence from Indonesia." World Development. 24: 65-78. Cuevas, Carlos E'. and Jeffrey Poyo. "Costos de Intermediaci6n en el Banco Agricola de la Republica Dominicana. Ohio State University. 1989. Cuevas, Carlos E. "Costs of Financial Intermediation in the Banking System of El Salvador." Report to the Center for Latin American Studies. Arizona State University. 1990. Cuevas, Carlos E., Douglas H. Graham and Julia A. Paxton. "The Informal Financial Sector in El Salvador." Report prepared for USAID El Salvador. Columbus Ohio: The Ohio State University. 1991. Danby, C. "Challenges and Opportunities in El Salvador's Financial Sector." World Development. Vol. 23. No. 12. pp: 2133-2152. 1995. FAO/WB. "El Salvador: Proyecto de Tierras y Servicios al Agro. Documento de Trabajo: Mejoramiento del Sector Financiero Rural." November, 1994. FOSAFFI. Memoria de Labores. San Salvador. 1992. FUNDAUNGO. "'Microfinanzas en El Salvador: Lecciones y Perspectivas." December 1995. Rural Finance 41 FUSADES. "Una Estrategia de Desarrollo Agricola para El Salvador, 1994-2000." Documento de Trabajo No. 37. San Salvador. 1994. Hanson, Rebecca. "An Assessment of the Legal and Regulatory Framework for Enterprises in El Salvador." Washington, D.C.: IDB. October, 1993. Inter-American Development Bank. "El Salvador: Global Credit Program for Microenterprises. Loan Proposal. October, 1993. IPC. "El Salvador: Evaluation of the Financial and Institutional Viability of the Banco de Fomento Agropecuario. June 1993. Lacroix, Richard L. "El Salvador: The Warehouses of Banco de Fomento Agropecuario." unpublished manuscript. May, 1997. L6pez, Ram6n. "Rural Poverty iin El Salvador; A Quantitative Analysis." in Ramon L6pez and Alberto Vald6s (eds) Rural Poverty in Latin America, forthcoming. 1997. Ministerio de Agricultura y Ganaderia, Grupo de Opciones de Politica Agraria. "Consolidaci6n y Restructuraci6n de la Deuda Agraria." San Salvador. November, 1995. Moll, H.A.J., r. Ruben, E.W.G. Mol and A.A. Sanders. "Segmentation of Rural Financial Markets: An Exploration of the Borderlines in El Salvador." manuscript, Wageningen Agricultural University Rioseco, German. CentralAmerican Rural Credit Study. Unidad Regional de Asistencia T6cnica. RUTA. 1994. Sanchez-Schwarz, Susana M. "Assortive Matching of Borrowers and Lenders: Evidence from Rural Mexico." Unpublished Ph.D. dissertation. The Ohio State University. 1996. Sorto, Francisco and Alexander Segovia. "La Reforma Financiera de ARENA: Hacia D6nde se dirige la Privatizaci6n de la Banca?," Politica Econ6mica, 1: 1-25. 1992. Superintendencia del Sector Financiero. Normas y Reglamentos. Superintendencia del Sistema Financiero. Boletin Estadistico. San Salvador. December, 1992 -1995. World Bank. El Salvador. Meeting the Challenge of Globalization. Report No. 14109-ES, October 19, 1995. World Bank. El Salvador. The Challenge of Poverty Alleviation. Report No. 12315-ES, June 9, 1994. World Bank. "El Salvador: Competitiveness Enhancement Technical Assistance Project. Technical Annex." Report No. T-6679-ES. Washington, D.C.: World Bank. September, 1995. Yaron, J., M. Benjamin and G. Piprek. Rural Finance: Issues, Design and Best Practices. World Bank Discussion Paper, 1997. Yaron, Jacob. "Assesing Development Finance Institutions." World Bank Disucssion Papers No. 174. 1992. Yaron, Jacob. "Successful Rural Finance Institutions." World Bank Disucssion Papers No. 150. 1992. ANNEXES Annex A. REGULATION, LEGISLATION, AND SUPERVISION REGULATION, LEGISLATION, SUPERVISION 1. Regulation and Legislation. During 1990 and 1991, the Government undertook several reforms to the regulatory framework of financial institutions to increase their performance. New laws were drafted and approved for the BCR and banks and financieras. The new law for the BCR increased its administrative autonomy, while prohibiting the BCR to directly or indirectly finance central Govemrnment or public enterprises and set interest or exchange rates. The law for Banks and Financieras introduced more stringent prudential norms such as capital adequacy requirements. 2. The legislation introduces market segmentation by limiting the scope of financial activities of banks and financieras. Financieras are limited to the intermediation of non-demand deposits, while commercial banks are allowed to provide all modem banking services. A drawback of the legislation is that it ignores non-bank financial mechanisms such as leasing, factoring, and warehouse receipts. 3. At present, there is a. number of draft laws under review by the authorities. These are: Ley de instituciones Auxiliares de Crdito, Ley de las Instituciones Oficiales de Credito, Ley de las Instituciones de Seguros, Ley de Fondos de lnversi6n, Fondo de Garantia de Dep6sitos, Ley de la Superintendencia, Ley de FEDECACESy del Sistema Cooperativo. 4. The Superintendencv of the Financial System. The reorganization of SSF was initiated with the approval of the Ley Organica7 de la Superintendencia del Sistema Financiero in November, 1990. The elements of this reorganization targeted its weak authority and lack of autonomy from BCR and the executive power, and its obsolete supervision system. To perform its new responsibilities, SSF was reorganized into four divisions: (a) bank and finance companies; (b) insurance and pension fimds; (c) securities exchange; and (d) administration and finance. It supervises 140 institutions with about 215 employees. An ongoing reform of the Superintendency calls for its division into three independent superintendencies: one for the mercado de valores, one for the pension fund administration, and one for banks and financieras. The last one will have three intendencies: banks and financieras, insurance, and credit cooperatives. 5. Limitations of ownership. The banking law imposes limits on ownership based on nationality and' on a 1% restriction on capital ownership which can be lifted subject to SSF approval. The legislation assigned to Salvadoran nationals the majority ownership of financial intermediaries, i.e., foreigners could not own more than 50% of total capital. However, the last modification to the banking law allows foreign banks (Central America and other countries) to own 75% of stocks. This measure should foster the participation of foreign banks in the financial system through the purchasing of existing banks or through the incorporation of new institutions. At the moment, the Government is evaluating a proposal to sell Banco HipotecarIo to a foreign bank. 6. Capital Requirements, Minimum capital and capital adequacy are the two main restrictions regarding capital requirements in El Salvador.' The minimum capital requirement to enter the market is 050 million for banks and 025 million for financieras. The capital adequacy ratio, the ratio of equity to risk weighted assets, was raised to 10% in January 1996, from 8%. SSF established a transitional period of six The main functions of capital are: (a) to serve as a cushion which supports a finn's normal risks; (b) to convince depositors that adequate coinsurance protection exists; and (c) to absorb losses with enough margin to inspire confidence to depositors. 45 46 El Salvador: Rural Finance years to allow banks and financieras to meet the new capital adequacy ratio. For the purpose of this ratio equity consists of: (a) primary capital such as stocks and legal reserves; and (b) secondary capital such as retained earnings and 50% of net income. Furthermore, any investments in other institutions are subtracted to prevent artificial infusions of capital or the so-called double leveraging. 7. Assets are classified into three categories that receive a weight of 100% (loans), 50% (lines of credit), 20% (government bonds) and 0% (cash and due from banks) according to risk. The main merits of the risk-based calpital requirement are the recognition of a risk differential among assets, and the inclusion of off-balance sheet items. However, the main drawbacks are that it ignores other types of risk, such as liquidity and interest rate risk, correlation among assets or between liabilities, establishes arbitrary categories, and its does not use the market value of assets. In addition, it is difficult to appraise the value of off-balance sheet items. 8. Diversification Rules. The legislation also restricts the concentration of outstanding loan balances around a single customer (Art. 104). This limit is set to up to 15% of the equity capital of the financial intermnediary. However, this limit can be raised up to: (a) 30% if the loan is guaranteed with real collateral; (b) 50% if the ollateral consists of deposits; and (c) more than 50% if guarantee with line of credit of foreign banks, or deposits. 9. Limits on Loans to Insiders. The amount of outstanding loan balances that may be granted to bank insiders was reduced from 50 to 15% of stock and capital reserves in January 1996 (Art. 110). SSF gave establishing banks a period of two years to adjust to the new ratio requirement. Insiders are those employees involved in the administration of the institution as directors or managers and those stockholders that own more than 3% of the stocks. 10. Limits on loans to insiders is an important aspect of prudential regulation since the intermediary's owners may use these loans to recapture a proportion of their equity contribution. Under the new limit, a stockholder that owned 3% of the stocks can receive an average loan size up to 6.4 times his/her actual ownership.2 The previous 50% requirement implied that the same stockholder could receive 21.2 times his/her actual ownership. 11. Supervision. The supervision of financial intermediaries in El Salvador may be characterized as a situation in which SSF performs these functions with respect to banks, financieras and other institutions, but it has implicitly delegated to FEDECREDITO the supervisory responsibilities of its affiliates. There are some costs and benefits of the implicit delegation of supervisory activities to FEDECREDITO which are analyzed in Chapter 5. 12. The supervisory activity of SSF can be described as a two-tier monitoring system. The first-tier is called off-site supervision an consists of a series of reports that the financial intermediaries have to send to SSF. The seconi-tier of SSF supervision system is on-site supervision, which takes place only when off- site supervision data show that a particular internediary is experiencing problems. Since the 1990 reform program, bank supervision has increased substantially. 13. The department of financial analysis of SSF executes off-site supervision by doing a monthly analysis of financial statements with a completely computerized system. The main goal of this off-site supervision is to provide an early-warning mechanism about changes in the financial performance of an intermediary that may become insolvent, illiquid or both. SSF requires from banks and financieras the 2 The calculation was done using the formula: [Ratio of Loals to insiders / % ownershipl l I + (reserve capital / stocks)]. Annexes 47 following: (a) balance sheets and income statements should be prepared and sent to the SSF monthly; (b) financial statements should be published in widely-read newspapers quarterly; and (c) monthly submission of risk classification of borrovvers. 14. Loan Risk Classificaltion. The main objective of loan risk classification is to make risk exposure explicit and to require lending institutions to protect their solvency by taking adequate provision for loan losses. SSF classifies loans to enterprises, consumption loans, and mortgage loans with three different criteria. For enterprise loans, there are five credit risk categories (A, B, C, D, and E)with an attached required provision for expected loan losses. Table A-1 Loan Risk Categories Risk Borrower Characteristics Required Type Loan Reserves A Excellent capacity to repay and solid project proposal. 0% B Good project, however technical and marketing risk exist 1% C Projects with weak financial conditions. Delinquent less than 60 days. Collateral value becomes 10% important. D Dubious repayment possibilities. Delinquent less than 180 days on previous loans. 50% E Insolvent clients. Delinquent more than 180 days on previous loans, no guarantee. 100% Source: Wenner and Umafha (1993, Table 3.5). 15. Consumption and mortgage loans are classified into categories A to E depending on the days of arrears of the loan. These loans are classified as type A if they have been less than 30 days late. Consumption loans are categorized as type E if the number of days in arrears is more than 120 days. In contrast, mortgage loans must be classified as type E when they are late for more than 360 days. 16. Each bank gives a credit rating to their clients according to the instructions of SSF. It supervises the risk classification of loans performed by banks and financieras. The departmnent of risk assets of SSF samples 35% of risk assets of the system three or four times per year. SSF requires banks to sample up to 80% of their risk assets balances. 17. Credit Bureaus and Information Sharing. Current legislation requires SSF to manage a risk classification system about the borrowers of banks and financieras. The institutions that report to the credit bureau are banks and financieras, insurance companies, FOSAFFI, FEDECREDITO, FIGAPE, FONAVIPO, Fondo Social para la Vivienda, Instituto Nacional de Pensiones de los Empleados Publicos, Instituto de Previsi6n Social de la Fuerza Armada (SSF). 18. Access to borrower credit risk classification reduces borrowers and lenders transaction costs. Currently, lending officials use the database operated by the SSF as a preliminary screening device of potential clients. Banks have a clear preference for borrowers classified as type A or B because their lower reserve requirements imply a hiigher rate of return. Banks and financieras must provide SSF with the credit rating of borrowers when the loan is approved. In addition to credit rating, banks and financieras must provide other information such as: sector of economic activity, date of the loan, outstanding debt, and type of collateral. 19. The SSF gives banks and financieras a monthly report with the credit ratings of borrowers. Although the outstanding debt of borrowers or the number of loans are not shared, banks and financieras can explicitly request that iniformation to SSF. Borrowers are identified through the Numero de Identification Tributaria (NIT). The upgrading of the credit rating of a borrower must by approved by the SSF if the total debt of the borrower is more than 0500,000. SSF uses the information provided by banks to calculate delinquency problems by sector of economic activity. 48 El Salvador: Rural Finance Annex B. FINANCIAL SYSTEM AND ECONOMIC INDICATORS Table B-I El Salvador: GDP by Sector of Economic Activity (%) Economic Sector 1991 1992 1993 1994 1995 Agriculttre 16.5 16.5 16.0 13.8 13.7 Coffee 4.4 4.5 3.8 3.3 3.0 Cotton 0.1 0.1 0.1 0.1 0.0 Basic Grains 3.1 3.7 3.4 2.8 3.2 Sugar Cane 0.7 0.7 0.7 0.6 0.6 Ottwr crops 2.4 2.1 2.2 2.2 2.2 LIvestock 2.6 2.4 2.1 2.0 2.0 Oher 3.0 2.9 2.7 2.8 2.7 Non-Agriculture 83.5 83.6 85.0 86.2 86.3 Total 100.0 100.0 100.0 100.0 100.0 Source: BCR Bulletin, January-March 1996 Table B-2 Commercial Banks: Distribution of Outstanding Loan Balances and Average Real Growth Rate by Sector of Economic Activity (Millions of Colones of 1990) Economic Sector Avg Real Growth Jure 1991 June 1995 Rate p.a. S Agriculture 925 14.3 2,014 10.6 21.5 Mining 6 0.1 18 0.1 30.0 Manufacturing 1,479 22.9 3,360 17.7 22.8 Construction 363 5.6 2,729 14.4 65.6 Electricity 7 0.1 15 0.1 20.4 Cornmerce 1,725 26.7 7,373 38.9 43.8 Transpcrt 81 1.3 408 2.2 49.8 ServIces 245 3.8 1,145 6.0 47.0 Other Credit 25 0.4 - - - Refinancing 1,604 24.8 1,915 10.1 4.5 Agriculture 548 8.5 n/a n/a n/a Non-AglrIcukure 1,056 16.3 nla n/a n/a TOTAL 6,460 100.0 18,978 100.0 30.9 Sources: For 1991: BCR, Boletin Trimestrul, Octubre-Dkhembre 1993. For 1995: Unpublished data of BCR 'The awVrag, real growth rate equals to ((Be/8E1)25 - 1)1 00 Annexes 49 Table B-3 Commercial Banks: Credit Disbursed by Sector of Economic Activity (millions of colones) Economic Sector 1991 1992 1993 1994 1995 Total Agriculture 2,272 3,383 3,796 2,424 2,933 Agriculture 1,951 2,878 3,290 1,997 2,151 Agriculture Refinancing 321 505 495 427 782 Total Non-Agriculture 6,839 12,699 15,149 16,88 18,883 Mining 5 2 13 17 6 Manufacturing 2,589 4,486 4,641 4,618 4,868 Constructon 387 1,610 2,855 2,989 2,190 Electricity 5 5 8 19 30 Commerce 2,809 3,948 5,086 5,285 7,868 Transport 165 256 426 562 654 Services 197 486 567 773 798 Other Credit 34 400 27 119 228 Consumer Loans - 983 1,047 1,500 1,411 Non-Agriculture Refinancirng 648 524 478 606 829 TOTAL 9,110 16,082 18,934 18,912 21,816 Source: BCR Bulletin Various Issues Table B-4 Commercial Banks: Credit Disbursed by Sector of Economic Activity (% of total) Economic Sector 1991 1992 1993 194 1996 Total Agriculure 24.9 21.0 20.0 12.8 13.4 Agriculture 21.4 17.9 17.4 10.6 9.9 Agriculture Refinancing 3.5 3.1 2.6 2.3 3.6 Total Non-Agrlculture 75.1 79.0 80.0 87.2 86.6 Mining 0.1 0.0 0.1 0.1 0.0 Manufacturing 28.4 27.9 24.5 24.4 22.3 Construction 4.3 10.0 15.1 15.8 10.0 Electricity 0.1 0.0 0.0 0.1 0.1 Commerce 30.8 24.5 26.9 27.9 36.1 Transport 1.8 1.6 2.3 3.0 3.0 Services 2.2 3.0 3.0 4.1 3.7 Other Credit 0.4 2.5 0.1 0.6 1.0 Consumer Loans - 6.1 5.5 7.9 6.5 Non-Agriculure Refinancing 7.1 3.3 2.5 3.2 3.8 TOTAL 100.0 100.0 10.0 100.0 100.0 Source: 8CR Bulletin Various Issues 50 El Salvador: Rural Finance Table B-5 Saving Instruments by Type of Institution (millions of colones) 1995 1994 1993 1992 1991 ¢rn % ¢rn % Om % Om % ¢m % Passbook Deposits 9,707.9 34.2 9,213.2 37.7 8,268.0 41.0 7,471.4 47.8 3,771.0 42.5 Commercial Banks 8,367.8 29.5 n/a n/a 6,572.1 32.6 5,809.1 37.1 3,771.0 42.5 Financleras 1,340.1 4.7 n/a n/a 1,695.9 8.4 1,662.3 10.6 n/a n/a Term Deposits 15,280.6 53.9 13,051.7 53.5 10,922.7 54.2 7,625.0 48.7 4,781.6 53.9 Commercial Banks 12,152.3 42.9 nia n/a 8,596.5 42.6 5,903.0 37.7 4,781.6 53.9 Financheras 3,128.2 11.0 n/a n/a 2,326.2 11.5 1,722.0 11.0 n/a n/a CEDEVIV m 3,114.3 11.0 1,924.4 7.9 833.6 4.1 440.3 2.8 235.8 2.7 Commercial Banks 2,526.0 8.9 n/a n/a 567.8 2.8 302.4 1.9 235.8 2.7 Flnancheras 588.3 2.1 n/a n/a 265.8 1.3 138.0 0.9 n/a n/a FEDECREDITO 23.1 0.1 39.2 0.2 n/a n/a n/a n/a n/a n/a SCC affiliated with 222.9 0.8 185.1 0.8 138.8 0.7 107.4 0.7 83.8 0.9 FEDECACES TOTAL 28,348.7 100.0 24,413.6 100.0 20,163.1 100.0 15,644.2 100.0 8,872.2 100.0 Sources: For 1995 and 1992, BCR, Boletin Trmestral, 1995 and 1992: Classification of Deposits by Geographical Zones; for 1991, 1993 and 1994: unpublished data provided by the BCR * CEDEVIV begins being issued in 1992 by Banks and 1993 by Financieras. Prior to that, three-year certificates were issued.. Table B-6 Passbook Deposits and Term Deposits In and Outside San Salvador. Selected Features, 1992, 1995 1992 1995 Dpt of San Outside San Dpt of San Outside San Salvador Salvador TOTAL Salvador Salvador TOTAL Average Account Size (0) 9,504 6,491 7,590 12,771 7,533 10,149 Banks 10,288 5,805 8,027 12,120 7,844 9,948 Financleras 7,719 4,408 6,388 15,853 5,775 11,186 Average Accotnt Size (¢ of 1990) 8,272 4,779 6,605 9,051 5,339 7,193 Banks 8,954 5,052 6,986 8,590 5,559 7,050 Flnancheras 6,718 3,836 5,559 11,235 4,093 7,927 % of Accounts 52.3 47.7 100.0 49.9 50.1 100.0 Banks 49.6 50.4 100.0 49.2 50.8 100.0 Fhnancieras 59.8 40.2 100.0 53.7 46.3 100.0 % of Balance 66.6 34.5 100.0 62.8 37.2 100.0 Banks 63.5 36.5 100.0 60.0 40.0 100.0 Financieras 72.3 27.7 100.0 76.1 23.9 100.0 % of Balance within each area 100.0 100.0 100.0 100.0 100.0 100.0 Banks 75.2 82.0 77.6 78.3 88.5 82.1 Financieras 24.8 18.0 22.4 21.7 11.5 17.9 % of Branches (%) 46.4 53.6 100.0 63.3 46.7 100.0 Banks 46.4 53.6 100 50.8 49.2 100.0 Financieras n/a n/a n/a 60.3 39.7 100.0 % of Population 29.5 70.5 100.0 % of Rural Population 11.4 88.6 100.0 % of Rural Population In each Deparbtent 19.1 62.3 49.6 Nunber of People per Branch 10,501 28,623 18,958 Nunber of People per Bank Branch 23,263 48,086 36,561 15,121 37,180 25,983 Source: BCR and SSF bulletin Annexes 51 Table B-7 Formal and Non-bank Lenders in El Salvador, December 1995 (millions of colones) %. Net Agric. Type of Institution Assets % Equity % Deposits Loans % Loans % Branches % I 1~~~~~I(9196)I Specialized Financial Institutions 2,147 4.5 664 16.5 498 1.4 1,478 4.8 774 25.6 104 26.5 BFA 1,635 3.4 592 14.8 474 1.4 1,067 3.5 719 23.8 27 6.9 FEDECREDITO 401 0.8 48 1.2 23 0.1 341 1.1 55 1.8 68 17.3 FIGAPE 111 0.2 24 0.6 0 0.0 71 0.2 - - 9 2.3 Private Banksb 38,150 80.2 2,636 65.7 28,508 82.0 23,858 77.9 2,229 73.8 178 45.3 Financierasb 6,852 14.4 569 14.2 5,557 16.0 4,915 16.1 16 0.5 82 20.9 SCC affiliated to FEDECACES 420 0.9 145 3.6 223 0.6 260 0.8 29 7.4 NGOs n/a n/a n/a 0 0.0 105 0.3 n/a Centro de Apoyo ala n/a n/a n/a 0 0.0 24 0.1 n/a Microempresa 9/95 d Fundaci6n Salvadorefla de n/a n/a n/a 0 0.0 4 0.0 n/a Apoyo Integral (FUSAI) 12/94 d PROPEMI 12/94 d n/a n/a n/a 0 0.0 76 0.2 n/a TOTAL 47,670 100.0 4,015 100.0 34,786 100.0 30,616 100.0 3,019 100.0 393 100.0 Sources: *Unpublished data of SSF. b Bulletin of SSF December 1995 FEDECACES, Memorias 1995. Data is as of June 1995 and correspond to 29 credit cooperative d Fundaungo (1995 Data on loans granted by FEDECACES and NGOs by sector of economic activity was not available. Hence, although they do finance agricultural activities their shares could not be calculated. As a result, shares in agricultural lending of other intermediaries are overestimated. Table B-8 Sectoral Distribution of Loan Portfolios of Banks, Financieras and Specialized Financial Institutions (millions of colones, 9/95) Economic Sector SFI a % Private % Financieras % Total % Banksb Agriculture 774 46.7 2,229 9.9 16 0.4 3,019 10.6 (% of total agriculture) 25.6 73.8 0.5 100.0 Non-Agriculture d 883 53.3 20,239 90.1 4,390 99.6 25,513 89.4 (% of total non-agriculture) 3.5 79.3 17.2 100.0 Total 1,657 100.0 22,468 100.0 4,406 100.0 28,532 100.0 (% of total) 5.8 78.7 15.4 100.0 Sources: Unpublished Data of BCR and Memorias of FEDECREDITO and Calpia Calpia: According to Memorias 1995, 23% of the loan portfolio is agriculture d Includes 'other activities (refinancing) 52 El Salvador: Rural Finance Annex C. RURAL HOUSEHOLD SURVEY Table C-1 Debtors versus Non-debtors All Households Debtors Non-Debtors Mean S. D. Mean S. D. Mean S. D. Age of household head 45.7 14.3 42.7 13.6 ... 46.5 14.4 Male household head (Dummy) 0.9 0.3 0.9 0.3 0.9 0.3 Years of Edlucation of household Head 3.0 3.2 3.0 3.2 3.0 3.2 Average agie of household members 26.3 11.7 23.8 8.8 26.9 12.3 Average education household members 3.7 2.5 3.4 2.1 3.7 2.7 older than 12 years-old Household size 5.9 2.5 6.5 2.6 ... 5.7 2.5 Household members that migrated 0.7 1.2 0.8 1.3 0.6 1.2 Participation in Groups (Dummy) 0.1 0.3 0.1 0.4 0.1 0.3 Farmers (Dummy) 0.3 0.5 0.3 0.4 0.3 0.5 Household Annual Income 22,492.9 39,360.5 20,772.5 28,172.4 22,929.1 41,735.0 Agriculural Income 6,465.7 35,632.4 4,193.6 13,203.6 7,041.6 39,326.6 Nonagricuftural Income 890.5 9,373.1 2,357.7 20,055.9 518.6 2,849.6 Wage Inicome in agriculture 4,477.2 6,712.0 3,609.1 5,123.4 * 4,697.3 7,045.5 Wage Income in non-agriculture 9,031.0 14,047.6 9,783.6 11,917.2 8,840.3 14,542.6 Other Sources 1,628.5 4,703.3 828.5 4,243.0 1,831.3 4,795.7 Share income from wage employment in 36.8 44.0 47.9 52.5 ... 34.0 41.1 non-agriculture Distance to Secondary School (Kms) 6.1 9.9 5.7 10.0 6.2 9.8 Distance to Pavement Road (Kms) 5.9 6.7 6.1 7.5 5.8 6.5 Population iin the household municipality 38,298.1 43,550.8 32,485.3 30,195.7 ... 39,771.6 46,237.4 of residence Percentage of rural population in the 68.0 20.2 67.5 19.5 68.1 20.4 household municipality of residence Municipality Population Density 230.0 165.2 2359 190.2 228.4 158.4 Source: Encuesta Desarrollo Rural, 1996 * significant at the 95% and 99% confidence intervals, respectively Annexes 53 Table C-2 Borrowers versus Non-borrowers All Households Borrowers Non-borrowers Mean S. D. Mean S. D. Mean S. D. Age of Household Head 45.7 14.3 43.4 13.5 46.1 14.4 Male Household head (Dummy) 0.9 0.3 0.9 0.3 0.9 0.3 Years of Education of Household Head 3.0 3.2 3.3 3.3 2.9 3.2 Average age of household members 26.3 11.7 23.5 10.0 26.6 11.9 Average education of household 3.7 2.5 3.7 2.4 3.7 2.6 members older than 12 years-old Household size 5.9 2.5 6.2 2.6 5.8 2.5 Household members that migrated 0.7 1.2 0.6 1.0 0.7 1.2 Years in Locality 70.3 18.7 71.0 18.8 70.2 18.7 Participation in Groups (Dummy) 0.1 0.3 0.2 0.4 0.1 0.3 Farmers (Dummy) 0.3 0.5 0.3 0.5 0.3 0.5 Household Annual Income 22,492.9 39,360.5 23,921.5 39,519.4 22,302.1 39,371.2 Agricultural Income 6,465.7 35,632.4 6,231.9 16,653.6 6,496.9 37,455.9 Nonagricultural Income 890.5 9,373.1 4,090.7 26,310.9 463.1 2,583.5 Wage Income in agriculure 4,477.2 6,712.0 4,966.8 8,650.7 4,411.8 6,415.9 Wage Income in non-agriculture 9,031.0 14,047.6 7,614.3 16,004.1 9,220.3 13,770.4 Other Sources 1,628.5 4,703.3 1,017.8 4,749.8 1,710.1 4,695.3 Share income from wage employment in 36.8 44.0 34.1 48.7 37.2 43.3 non-agriculture Distance to Secondary School (Kms) 6.1 9.9 6.3 12.0 6.1 9.5 Distance to Pavement Road (Kms) 5.9 6.7 5.9 7.0 5.9 6.7 Population in the household municipality 38,298.1 43,550.8 36,792.0 36,132.8 38,499.3 44,472.3 of residence Percentage of rural population in the 68.0 20.2 64.8 20.3 68.4 20.2 household municipality of residence Municipality Population Density 230.0 165.2 223.9 126.3 230.8 169.8 Source: Encuesta Desarrollo Rural, 1996 I/'** difference in means significant at the 95 percent and 99 percent confidence intervals, respectively. Table C-3 Relative Importance of Formal, Semi-formal, and Informal Sectors in Rural Areas Loans Received in 1991 Outstanding Debt Balances Sector No. %No. %Anount No. %No. % Balance Formal 33 41.3 76.0 46 33.3 64.7 BFA 24 30.0 61.1 29 21.0 35.3 FEDECREDITO 3 3.8 4.6 4 2.9 2.3 Other Formal 3 3.8 4.1 8 5.8 12.3 Private Banks 3 3.8 6.2 5 3.6 14.8 Semi-formal 8 10.0 6.2 6 4.3 4.1 NGOs 2 2.5 0.4 2 1.4 2.1 Cajas/bancocomu/coope 6 7.5 5.7 4 2.9 2.0 Informal 39 48.8 17.8 86 62.3 31.1 Commercial Credit 12 15.0 5.7 26 18.8 11.9 Moneylender 11 13.8 5.7 15 10.9 10.0 Friends/Relatives 13 16.3 4.5 38 27.5 5.4 Interlinked Credit 3 3.8 1.9 4 2.9 1.8 Other Informal 3 2.2 2.0 TOTAL 80 100.0 100.0 138 100.0 100.0 Source: Encuesta DesaITolIo Rural, 1996 54 El Salvador: Rural Finance Table C-4 Estimated Coefficients of Multinomial Logit Model of Access to Credit Services PFormal I Non-Debtor Pinfmal I Non-bMor Variable Coeffici Std. Error Coefficle Std. Efror ent nt Constant 0.6664 3.0980 2.2100 2.2800 Log (age of household head) -0.5507 0.6010 -0.5943 0.4205 Average education of household members older than 12 years -0.0974 0.0781 -0.0982 0.0587 old Male household head (Dummy) 0.3869 0.7605 -0.1477 0.4400 Household size 0.1177 0.0643 0.1468 0.0502- Number of extended family that migrated -0.0400 0.1564 0.2376 0.0892 ... Number of years In the locality 0.0083 0.0099 0.0121 0.0079 Log of annual household income -0.2147 0.2119 -0.2736 0.1737 Share of wage income in non-agricultural activities 0.0066 0.0042 0.0087 0.0031 Participation In groups (Dummy: 1= Yes 0=No) 1.1770 0.4275 ... 0.2233 0.4131 Population in the household municipality of residence -0.0103 0.0061 -0.0064 0.0040 Percentage of the population in the rural areas in household's -0.0102 0.0102 -0.0103 0.0071 municipality of residence Farm-household (Dummy) 0.9416 0.3890 -0.3419 0.3445 L8.... 9 ............................................................................................................................................................................... Likelihood Ratio Test: Chi-Squared (df.-24) = 64.89 Number ol Observations: 620 Source: Multinomial Logit Estimation /I- significant at the 90% 1 95% 1 99% confidence interval respecively Table C-S Marginal Effects of Explanatory Variables on the Probability of Having Debt P Nan-O.bto Pronns P ubmi Variable Coeff. Std. Error Coeff. Std. Error Coeff. Std. Error Constant -0.2381 0.4349 0.0217 0.1781 0.2165 0.2966 Log (age of household head) 0.0817 0.0839 -0.0258 0.0341 -0.0558 0.0573 Average education of household members older than 12 years 0.0138 0.0110 -0.0046 0.0046 -0.0092 0.0076 old Male household head (Dummy) -0.0045 0.1070 0.0218 0.0413 -0.0172 0.0721 Household size -0.0193 0.0090 0.0054 0.0039 0.0139 0.0069 Number of extended family that migrated -0.0203 0.0222 -0.0037 0.0083 0.0240 0.0171 Number of years in the locality -0.0015 0.0014 0.0004 0.0006 0.0012 0.0010 Log of annual household Income 0.0357 0.0296 -0.0098 0.0126 -0.0259 0.0212 Share of wage incomne in non-agricultural activities -0.0011 0.0006 * 0.0003 0.0002 0.0008 0.0004* Participation in groups (Dummy: 1= Yes O=No) -0.0767 0.0583 0.0619 0.0311 0.0147 0.0396 Population in the household municipality of residence 0.0011 0.0008 -0.0005 0.0003 -0.0006 0.0006 Percentage of the population In the rural areas in householdcs 0.0014 0.0014 -0.0005 0.0006 -0.0010 0.0010 municipality of residence Farm-household (Dummy) -0.0127 0.0550 0.0529 0.0279 -0.0402 0.0422 Source: Multinomial Logit Estimation ...* significant at the 90% 1 95% 1 99% confidence interval respectively Annexes 55 Annex D. SPECIALIZED FINANCIAL INTERMEDIARIES: FINANCIAL INDICATORS Table D-1 BFA: Profitability Measures 1993-95 Ratio 1993 1994 1996 ROE = Net Incoe / Average Equity 10.0 2.6 0.4 ROA = Net Income I Average Assets 4.8 1.2 0.2 Equity Multiplier = Average Assets I Average Equity 2.1 2.2 2.6 Asset Utilization = Total Operating Income I Average Assets 28.4 16.9 17.0 Proft Margin = Net IncomelTotal Operating Income 17.0 7.0 0.9 Net Interest Spread = Yield on Assets - Cost of Liabilities 3.6 4.6 4.8 Profit Margin Coinponents Interest Expense I Total Operating Income 12.6 24.7 29.0 Operational Expenses / Total Operating Income 28.8 51.6 47.3 Provision for Loan Losses/ Total Operating Income 37.9 16.7 22.7 Average Interest Cost of Liabilities 8.6 8.6 8.8 Deposits - 7.9 7.2 Borrowed Funds - 8.7 9.1 Average Yield on Assets Gross Loans - 13.9 15.0 Net Loans - 20.0 20.3 Average Real Grcoss Yield on Assets Gross Loans - 5.6 4.1 Net Loans 11.3 8.8 Market Share of Total Assets 4.1 Real Growth Rate! Gross Loans 5.5 9.8 2.2 Net Loans 12.1 17.5 8.5 Past-Due Loans -27.3 22.4 -14.3 Borrowed Funds 4.7 13.9 13.1 Deposits 19.8 24.3 43.1 Equity 3.4 -7.6 -8.8 LIabilities 8.2 14.6 17.3 Assets 5.9 4.0 6.3 Source: Audited Financial Statements Published by the Superintendency of Banks, and the SFA. 56 El Salvador: Rural Finance Table D-2 BFA Risk Indicators 1993-95 Ratio 1993 1994 1996 Credit Risk Provsion for Loan Losses / Average Gross Loans 12.7 3.2 4.4 Averag Past Due Loans a / Average Gross Loans 26.8 23.5 22.5 Average Reserve for Loan Losses/ Average Gross Loans 34.9 30.6 26.1 D and E / Risky Assets b 32.0 32.0 27.9 Capital Risk Average Equity / Average Assets 48.1 44.7 39.0 Equiy / Weighted Assets b 55.0 55.0 46.0 Average Equity I Average Uability and Contingencies 92.6 80.1 63.4 Operational Managenent Average Assets / Number of Employees (In 0000s) 924.3 1,034.4 Operational Expenss /Average Net Loans 14.8 14.4 12.4 Operational Expenses / Average Eaming Assets 9.9 10.1 9.1 Operational Expenses I Average Assets 8.2 8.7 8.0 Uquidity Risk Average Borrowed Funds / Average Assets 24.8 25.9 28.0 Average Cash and Due from Banks / Average Deposits 56.1 51.0 44.2 Subsidy Dependence Index 93.5 97.6 100.8 Comnposition of Assets and Liabilities Fixed Assets (% Assets) 10.1 9.4 7.8 DeposKs (% Assets) 18.0 21.6 29.0 Borrowed Funds (% Assets) 24.7 27.0 28.7 Equity (% Assets) 47.5 42.2 36.2 Deposits (% Net Loans) 31.9 33.7 44.5 Source: Audited Financial Statements Published by the Superintendency of Banks, and the BFA. a More than 90 days past due b SSF Bulletin Table D-3 FEDECREDITO: Profitability Measures 1994-95 Ratio 1994 1995 ROE = Net Income / Average Equity -5.0 -0.7 ROA = Net Income / Average Assets -0.6 -0.1 Equiy Mukiplier = Average Assets / Average Equity 8.6 8.5 Assd Utilization = Total Operating Income / Assets 14.9 16.6 Profit Margin = Net Income/Total Operating Income -3.9 -0.5 Proft Margin Components Interest Expense / Total Operating Income 65.9 64.2 Administrative Expense / Total Operating Income 35.1 30.7 Provision for Loan Losses/ Total Operating Income 2.9 5.6 Average Interest Cost of Uabilities 12.0 12.9 Market Share of Total Assets Real Growth Rate Gross Loans 6.1 14.8 Net Loans 12.7 16.8 Past-due loans 42.8 -26.0 Borrowed Funds 7.6 9.0 DeposKts -14.9 -46.6 EquKy 7.0 -2.2 Labilitles 6.0 -4.7 Assets 6.2 -4.4 ' The reversion of resov of loan loses, which FEDECREDITO reports as income, were subtracted from the net-non-interest income. Annexes 57 Table D-4 FEDECREDITO: Risk Measures 1994-95 Ratio 1994 1995 Credit Risk Provwsion for Loan Losses / Average Gross Loans 0.6 1.2 Average Past Due Loans a / Average Gross Loans 4.6 4.2 Average Reserve for Loan Losses/ Average Gross Loans 10.8 7.5 Capital Risk Average Equity / Average Assets 11.6 11.8 Average Equity / Average Liability and off-balance sheet 9.2 9.3 items Operational Manageumient Average Assets / Number of Employees (in 0000s) 1,322.4 1,552.5 Operational Expenses I Average Net Loans 8.3 7.0 Operational Expenses / Average Eaming Assets 5.9 5.6 Operational Expenses / Average Assets 5.2 5.1 Uquidity Risk Borrowed F'jnds I Total Assets (%) 63.4 68.4 Cash and Due from Banks / Deposits (%) 29.7 35.0 Subsidy Dependence Index 86.5 89.0 Comnposition of Assets and Liabilities Net Loans (* assets) 64.9 79.2 Fixed Assets (% assets) 4.8 4.3 Deposis (% assets) 10.3 5.8 Borrowed Funds (% assets) 63.8 72.7 Equity (% assets) 11.6 11.9 Deposits (% net loans) 15.9 7.3 Source: Financial Statements published by FEDECREDITO 58 El Salvador: Rural Finance Table 0-5 Cajas Rurales and Banco de Trabajadores: Profitability Measures 1994-95 Ratio 1994 1996 ROE = Net Income / Average Equity 11.1 8.1 ROA = Net Incorne /Average Assets 2.0 1.6 Equtty Multiplier = Average Assets / Average Equity 5.7 5.2 Asset Utilization = Total Operating Income / Assets 25.4 25.7 Profit Margin = Net Income/Total Operating Income 7.7 6.1 Promft Margin Cornponents Interest Expense / Total Operating Income 51.9 57.5 Administrative Expense / Total Operating Income 31.0 29.6 Provision for Loan Losses/ Total Operating Income n/a n/a Average Interest Cost of Liabilities 17.8 20.1 Real Growth Rate Gross Loans 23.9 19.8 Net Loans 26.7 17.3 Non-performing loans -25.5 -2.5 Borrowed Funds 22.3 13.1 Deposits 42.4 -22.2 Equity 47.4 16.7 Liabilities 18.1 11.6 Assets 22.7 12.6 Source: Financial Statements published by FEDECREDITO Table D-6 Cajas Rurales and Banco de Trabajadores: Risk Measures 1994-95 Ratio 1994 1995 Credit Risk Provision for Loan Losses / Average Gross Loans 0.0 0.0 Average Past Due Loans a / Average Gross Loans 5.0 3.5 Average Reserve for Loan Losses/ Average Gross Loans 5.6 5.8 Capital Risk Average Equity / Average Assets 17.6 19.4 Average Equity / Average Liability and Contingencies 18.0 20.9 Operational Managemnent Operational Expenses I Average Net Loans 10.1 9.4 Operational Expenses / Average Assets 7.9 7.6 Liquidity Risk Borrosved Funds / Total Assets (%) 63.0 63.1 Cash and Due from Banks / Deposits (%) 127.7 172.1 Subsidy Dependence Index n/a n/a Composition of Assets and Liabilities Net Loans (% assets) 79.3 82.6 Fixed Assets (% assets) 6.1 5.3 Deposits (% assets) 4.4 3.0 Borrowed Funds (% assets) 63.0 63.2 Equity (% assets) 19.0 19.7 Deposits (% Net Loans) 5.5 3.7 Sourcie: Financial Statements published by FEDECREDITO Annexes 59 Table D-7 FEDECACES: Profitability Measures 1993-95 Ratio 1993 1994 1995 ROE = Net Income / Average Equity 4.4 3.5 4.5 ROA = Net Income / Average Assets 1.0 0.9 1.3 Equity Multiplier = Average Assets I Average Equity 4.4 3.9 3.4 Asset Utilization = Total Operating Income / Average Assets 11.1 12.8 14.5 Profit Margin = Net Income/rotal Operating Income 9.2 7.0 9.2 Profit Margin Components Interest Expense / Total Operating Income 21.8 21.8 18.8 Operational Expenses / Total Operating Income 61.3 60.9 60.2 Provision for Loan Losses/l Total Operating Income 1.7 1.7 2.1 Expenses for Reserve / Total Operating Income 6.1 8.5 9.7 Average Interest Cost of Liabilifies 3.4 3.8 3.9 Average Gross Yield on Akssets Gross Loans 11.7 11.6 11.3 Net Loans 12.1 12.1 11.9 Average Real Gross Yielid on Assets Gross Loans 3.9 3.5 0.7 Net Loans 4.3 3.9 1.3 Real Growth Rate Gross Loans 12.2 -3.1 4.4 Net Loans 12.8 -4.0 4.1 Past-Due Loans -21.1 7.4 -25.4 Borrowed Funds -11.0 -7.3 -36.1 Deposits 400.5 -37.4 24.5 Equity 4.9 2.6 -13.5 Liabilities -10.5 -11.1 -32.5 Assets -7.1 -7.8 -27.4 Source: Memorias de FEDECACES Table D-8 FEDECACES: Risk Measures 1993-95 Ratio 1993 1994 1996 Credit Risk Provision for Loan Losses /'Avg Gross Loans 0.4 0.4 0.4 Avg Past Due Loans a / AvgS Gross Loans 12.8 11.4 10.1 Average Reserve for Loan Losses/ Average Gross Loans 3.8 4.1 4.8 Avg Refinanced Loans / Avg Gross Loans 11.6 10.3 10.9 Capital Risk Avg Equity / Avg Assets 22.9 25.7 29.4 Avg Equity / Avg Liability anid Contingencies 13.5 14.1 14.6 Operational Management Operational Expenses / Average Gross Loans 13.0 13.2 12.1 Operational Expenses I Average Net Loans 13.5 13.8 12.7 Operational Expenses / Average Net Eaming Assets 10.7 11.6 11.2 Operational Expenses / Average Assets 6.8 7.8 8.7 Liquidity Risk Average Borrowed Funds / Average Assets 67.6 66.4 63.0 Average Cash and Due from Banks / Average Deposits 81.3 48.7 40.1 Subsidy Dependence Index 285.2 239.2 234.5 Composition of Assets and Liabilities Net Loans (% assets) 55.4 57.7 82.7 Fixed Assets (% assets) 3.7 4.0 5.8 Deposits (% assets) 7.5 5.1 8.7 Borrowed Funds (% assets) 66.2 66.6 58.6 Equity (% assets) 24.3 27.1 32.2 Deposits (% net loans) 13.5 8.8 10.5 Source: Calculated with Financial Statements published by FEDECACES 60 El Salvador: Rural Finance Table D-9 FEDECACES Affiliates: Profitability Measures 1993-95 Ratio 1993 1994 199S ROE = Net Income I Average Equity 9.7 13.7 7.3 ROA = Net Income I Average Assets 3.6 5.0 2.6 Equity Multiplier = Average Assets / Average Equity 2.7 2.7 2.8 Asset Utilization = Total Operating Income I Average Assets 17.4 18.6 9.7 Proft Margin = Net Income/Total Operating Income 20.6 26.9 26.5 Profit Margin Comnponents Interest Expense / Total Operating Income 43.7 43.1 42.4 Operational Expenses/ Total Operating Income 35.7 30.0 31.1 Provision for Loan Losses/ Total Operating Income n/a n/a n/a Average Interest Cost of UabilUtles 16.2 16.2 7.9 Average Gross Yield on Assets 20.4 20.9 10.6 Average Real Gross Yield on Assets 12.0 12.1 0.1 Real Growth Rate Gross; Loans 17.0 22.9 5.8 Deposits 20.2 23.7 8.9 Equity 12.9 15.7 -1.1 Liabilities 17.7 17.6 7.0 Assets 15.9 16.9 4.0 Source: Calculated with Financial Statements published by FEDECACES S Figuires correspond to June 1995 Table D-10 FEDECACES Affiliates: Risk Measures 1993-95 Ratio 1993 1994 19S5' Captal Risk Avg Equity / Avg Assets 37.1 36.5 35.3 Operational Management Operational Expenses / Average Gross Loans 8.4 7.3 3.8 Operational Expenses / Average Assets 6.2 5.6 3.0 Composition of Assets and Liabilities Gross Loans (% assets) 74.5 78.3 79.6 Deposits (% assets) 47.9 50.6 53.0 Equity (% assets) 36.7 36.3 34.5 Deposits (% loans) 64.3 64.6 66.6 Source: Calculated with Financial Statements published by FEDECACES 'Figures correspond to June 1995 Annexes 61 Table D-1i Profitability Measures of Selected Financial Intermediaries, 1995 Fedecredito Systema Fedecaces System Ratio BFA Banks Federation Affiliates Federation Affiliates D ROE = Net Income l Average Equity 0.4 27.2 -0.7 8.1 4.5 1 3.7 ROA = Net Income I Average Assets 0.2 1.7 -0.1 1.6 1.3 5.0 Equity Multiplier = Average Assets / Average Equity 2.6 15.1 8.5 5.2 3.4 2.7 Asset Utilization = Total Operating Income I Average Assets 17.0 14.5 16.6 25.7 14.5 18.6 Profit Margin = Net Income/Total Operating Income 0.9 12.1 -0.5 6.1 9.2 26.9 Profit Margin Components Interest Expense I Total Operating Income 29.0 59.3 64.2 57.5 18.8 43.1 Operational Expenses I Total Operating Income 47.3 22.5 30.7 29.6 60.2 30.0 Provision for Loan Losses Total Operating Income 22.7 5.3 5.6 n/a 2.1 n/a Average Interest Cost of Liabilities 8.8 9.5 12.9 20.1 3.9 16.2 Average Gross Yields on Assets Gross Loans 15.0 18.9 - - 11.3 - Net Loans 20.3 0.0 - - 11.9 - Average Real Gross Yields on Assets Gross Loans 4.1 7.6 - - 0.7 - Net Loans 8.8 - - - 1.3 - Real Growth Rate (1994-1996) Gross Loans 2.2 17.0 14.8 19.8 4.4 - Net Loans 8.5 17.0 16.8 17.3 4.1 22.9 Past-Due Loans -14.3 -26.0 -2.5 -25.4 - Borrowed Funds 13.1 9.0 13.1 -36.1 - Deposits 43.1 2.7 -46.6 -22.2 24.5 23.7 Equity -8.8 14.5 -2.2 16.7 -13.5 15.7 Liabilities 17.3 8.0 -4.7 11.6 -32.5 17.6 Assets 6.3 7.2 -4.4 12.6 -27.4 16.9 Source: Financial Statements = The reported net income of Fedecredito was adjusted by subtracting the reversion for the reserve for loan losses. Correspond to 1994 Table D-12 Risk Measures of Selected Financial Intermediaries, 1995 Fedecredito a Fedecaces Ratio BFA Banks Federation Affiliates Federation Affiliatesb Crednt Risk Provision for Loan Losses / Average Gross Loans 4.4 1.3 1.2 n/a 0.4 n/a Average Past Due Loans a l Average Gross Loans 0.0 0.0 4.2 3.5 10.1 n/a Average Reserve for Loan Losses/ Average Gross Loans 26.1 2.9 7.5 5.8 4.8 n/a D and E / Risky Assets b 0.0 0.0 - n/a n/a Avg Refinanced Loans / Avg Gross Loans n/a n/a n/a nla 10.9 n/a Capital Risk Average Equity / Average Assets 39.0 8.0 11.8 19.4 29.4 36.5 Equity I Weighted Assets c 0.0 0.0 n/a nla n/a n/a Average Equity / Average Liability and Contingencies 63.4 9.2 9.3 20.9 14.6 n/a Operational Managemnent Average Assets / Number of Employees (in ¢000s) 1034.4 3987.5 1,552.5 n/a 1,068.3 n/a Operational Expenses / Average Net Loans 12.4 3.5 7.0 9.4 12.7 n/a Operational Expenses / Average Assets 8.0 3.3 5.1 7.6 8.7 5.6 Liquidity Risk Average Borrowed Funds / Average Assets 28.0 12.9 68.4 63.1 63.0 n/a Average Cash and Due from Banks I Average Deposits 44.2 32.6 35.0 172.1 40.1 nla Subsidy Dependence Index 100.8 n/a 89.0 n/a 234.5 n/a Comnposition of Assets and Liabilities Fixed Assets (% Assets) 7.8 3.4 4.3 5.3 5.8 n/a Deposits (% Assets) 29.0 72.8 5.8 3.0 8.7 50.6 Borrowed Funds (% Assets) 28.7 14.7 72.7 63.2 58.6 n/a Equity (% Assets) 36.2 8.1 11.9 19.7 32.2 36.3 Deposits (% Net Loans) 44.5 116.3 7.3 3.7 10.5 64.6 Source: Financial Statements of the Institutions a More than 90 days past due b Correspond to 1994 ' SSF Bulletin 62 El Salvador: Rural Finance Annex E. DEBT REFINANCING OR RESTRUCTURING PROGRAMS DEBT REFINANCING OR RESTRUCTURING PROGRAMS 1. Debt refinancing programs are the main strategy used by the Government to deal with debt delinquency. These programs raise two main concerns: (a) creation of incentives for borrowers in good standing to stop paying their loans with less favorable terms than delinquent borrowers; and (b) encouragement of strategic default. The three most recent debt forgiveness programs are explained below. 2. Ley para la Rehabilitaci6n de Sectores Productivos Directamente Afectados por el Conflicto (D.L. 292). D.L. 292 is a process to renegotiate on concessional terms the delinquent debt of producers directly affected by the conflict (agriculture and non-agriculture). The decree, approved in July 1992, has two main objectives: (a) refinance the debt to allow borrowers to pay back; and (b) restore the creditworthiness of delinquent borrowers to increase their access to credit. 3. The Decree created the following measures for qualified delinquent debtors. First, their back interest was recalculated at 8% and refinanced at the same interest for 20 years with two years of grace. Second, the it implemented a procedure to terminate any credit embargoes during the processing period of the Decree. These debtors continued with their risk classification of C, D, or E with respect to their delinquent debt and were classified again for new loans. They regained their status as credit-worthy, but banks were required to establish a reserve against each loan. Third, the Government compensated banks for the interest forgiven to rehabilitated debtors.3 This procedure was applied to both the banks that transferred their debt to FOSAFFI, as well as other banks, and in particular BFA. Fourth, a guarantee fund for small enterprises (Fondo de Garantia para Pequefios Empresarios) was created to back new loans granted to rehabilitated debtors. Benito (1993) argued that banks are discouraged from providing new credit to rehabilitated debtors due to the costly procedures involved in accessing these guarantee fuinds. 4. D.L. 292 has had two major modifications. In February 1994, the Decree recalculated back interest at 6% for beneficiaries whose original debt was below ¢ 100,000 (Decree 814). Furthermore, banks were not allowed to charge commission or fees for processing applications. In August 1994, the debt of beneficiaries with original loans below ¢5,000 from BFA or FEDECREDITO was forgiven (Decree 114). 5. The main limitations of D.L. 292 are: (a) restrictive definition of "those directly affected by the conflict"; (b) high transaction costs for potential beneficiaries; and (c) conservative banking criteria applied to each specific case at the level of branch bank agencies. 6. As of April 1996, 17,753 applications had been received for a total of ¢1,609.1 million. The total amount of interest reduced is estimated at ¢413.7 million. BFA and FEDECREDITO accounted for more than 95% of the number of applications received and approved. About 60% of the number of applications were received by BFA, while 35% were received by FEDECREDITO. Although BFA received the largest number of applications, they accounted for only 23% of total funds requested, while private banks and financieras received applications for 70% of this total. The average loan amount the applications varied among lenders. Applications from private banks averaged ¢1.5 million, while the average for BFA was 034,500. FEDECREDITO had the smallest average loan application of all institutions at ¢5,600. 3 Difference between interest calculated at the original rates and those recalculated at 8% not subject to reserves. Annexes 63 7. Private banks and financieras approved 72% of the total number of applications they received, granting 60% of the loan arnounts requested. FEDECREDITO and BFA, on the other hand, approved 99 and 95% of total number of applications they received and granted 92 and 90% of the amount requested, respectively. This reflects the different clientele served by these intermediaries or less stringent evaluation criteria. FIGAPE approved 59% of total applications it received, granting 92% of amount requested. Agriculture and Debt Delinquency 8. In June 1996, the Oiovernment issued two debt forgiveness programs that target the agricultural and agrarian debt as a mechanism to stimulate agricultural production. The rural sector of El Salvador has been greatly affected by policy reforms introduced since 1989. As a result of these reforms and the end of the conflict, rural and agricultural producers are being forced to adjust their economic activities. These reforms included: (a) liberalization of basic grain prices; (b) elimination or privatization of agricultural parastatals; (c) elimination of subsidized interest rates; and (d) trade reforms (FUSADES, 1994). 9. The sources of agriculture delinquency are related to: (a) difficulties created by conflict ; (b) the effects of droughts; (c) rent-seeking behavior induced by Government intervention in credit markets; (d) the impact of a strong currency on agricultural profitability; and (e) the diversion of credit funds to unproductive activities (Abt. 1993). In addition, the inadequate supervision of production programs on which credit is based and the attitude created by the subsidized credit organizations need to be emphasized. However, many authors haLve concluded that the environment created by the conflict was the most important cause of agricultural delinquency (Abt., 1993; Fusades, 1994; MAG, 1995). 10. The war affected agricultural production through various channels such as payment of war taxes, invasions and kidnappings. lLJncertainties caused by the conflict forced farmers to operate below efficient production levels causing low returns. As reported by Abt. (1993) many farmers argued that exchange policies were the major cause of low profitability and loan defaults. Furthermore, the conflict created the opportunity for rent-seeking behavior, through lobbying for debt restructuring or forgiveness programs. I1. The agrarian reform process originated the so-called deuda agraria (agrarian debt). As reported by MAG (1995) the agrarian debt resulted from: (a) the low productivity of agrarian reform beneficiaries due to poor land quality and inexperience; (b) lack of a recovery strategy by the Government; (c) the conflict; and (d) lack of working capital. 12. Ley de Apovo a la Reactivaci6n del Sector Agropecuario (D.L. 698). The Law targets all agricultural producers, but agricultural producers directly affected by the conflict (that qualify under D.L. 292) receive special treatment. It is estimated that a total of 0 1,152.7 million will be forgiven of which p486.1 million correspond to agricultural loans that qualified under D.L. 292. Key measures are: * It allows for a 70% discount for prompt payment to agricultural producers who qualified under Decree 292. If qualified producers pay 30% of their debt held at BFA, FOSAFFI, FEDECREDITO, private banks or financieras, thei remaining 70% is forgiven. However, debt is totally forgiven for balances below 05,000 of the debt owed by agricultural producers to those same creditors. * The grace period for beneficiaries of D.L. 292 is extended to June 1997. As a result of this decree, FOSAFFI and the BFA will have to divide proportionally the interest accumulated during the extended grace period among installments of the remaining years. * Authorizes FOSAFFI to apply a 40% discount for prompt payment of agricultural loans that do not qualify under D.L. 292. BFA and FEDECREDITO should apply the same discount to D or E loans. * Its allows financial institutions to include as secondary capital the value of the loss generated as a result of the reserve for loan losses of those agricultural loans that qualify under D.L. 292. 64 El Salvador: Rural Finance 13. Ley de Restructuraci6n de la Deuda Agraria (D.L. 699). The Law considers the original agrarian debt, the so-called adelantos ISTA-BFA, and FFRAP debt. Lenders involved are ISTA and Banco de Tierras. The beneficiaries are agricultural cooperatives, communal associations, and beneficiaries of Decrees 842, 207, 839 and 713 of the agrarian refonn. 14. D.L. 699 has negative consequences on the repayment culture because it penalizes those borrowers that were in good standing at the time of the Decree while loan defaulters are rewarded. 15. The main measures of this Decree are: (a) a 60% discount for prompt payment of total debt (capital and interest) for the agriculture cooperatives of the agrarian reform and communal associations of farmers. Individual (personas naturales) agrarian reform beneficiaries will receive a 60% discount for prompt payment plus an additional 05,000 write-off. When the total debt is less than 016,665, it is fully forgiven; ancd (b) ISTA and Banco de Tierras should also apply the 60% discount for prompt payment. IBRD 25454 LSS,AN .I'~~~~~~~~~~~~. - ~~~~RUSS ArJ 60' \ FED J FEDERATION AZERBAIJAN KAZAKHSTAN A. 'f, 4 1 '- > MONGOLIA - KY. RGYZ SLAM C REP 1 \ REP K A Z A K H S T A N OF IRAN TAUISTAN-L AFGsHANISTAN r . 2S >CH N 6 77n NAus KAZAKHSTAN U Z B E K I S T A N KYRGYZ d chkoduk - - REPUBLIC UrgesT Uct-t a\\ A V 25,yAtn ,{ E Adi.ho.n The boundaries- colors, ergo | W Ikrd detononotions and on,y T U R K M E N I S T A N 40 on this nroo do not - Burar - '/@ws rmply, on the port of The World Bank Group\ any udg_ient an the legol status of any territory, S k k \ S or any endorsemenat Ch sr\ s/zso.ydry or acceptnce of suchar boundaries. / /: ~ ~K-rht T TAJ I K I S TA N ISLAMIC REP D/- Sri.s OF IRAN ShN.h U Z B E K I S T A N arku,g.n Termez 7-si In * National Capital o Selected Cities Railroads . ~~~~~~~~AFGHANISTAN Main Roads Rivers 0_________LES I 2': PAKISTAN International Boundaries HIMOETOTFV, - 20' 60 70 DECEMBER 1993