Document of The World Bank Report No.: 40139 PROJECT PERFORMANCE ASSESSMENT REPORT REPUBLIC OF COLOMBIA FINANCIAL MARKETS DEVELOPMENT TECHNICAL ASSISTANCE PROJECT (TF-22479; SCL-41960) FINANCIAL SECTOR ADJUSTMENT LOAN (FSLT- 70000) FIRST PROGRAMMATIC FINANCIAL SECTOR ADJUSTMENT LOAN (Loan No. 7167-CO) SECOND PROGRAMMATIC FINANCIAL SECTOR ADJUSTMENT LOAN (TF-52531; FSL T-72580) June 26,2007 County Evaluation and Regional Relations Independent Evaluation Group Currency Equivalents (annual averages) (Exchange Rate Effective February 2,2006) US$1 = 2,268.60 COP FISCAL YEAR January 1 - December 3 1 Abbreviations and Acronyms CAF Corporation Andina de Fomento EFF Extended Fund Facility FMDTA Financial Markets Development Technical Assistance Loan FOGAFIN Deposit Guarantee Fund FIRST Financial Sector Reform and Strengthening FSAL Financial Sector Adjustment Loan GDP Gross Domestic Product ICR Implementation Completion Report IDB Inter-American Development Bank IDF Institutional Development Fund IEG Independent Evaluation Group IEPS Initial Executive Project Summary IFC International Finance Cooperation IF1 International Finance Institution IMF International Monetary Fund MoF Ministry of Finance PFSAL Programmatic Financial Sector Adjustment Loan PPAR Project Performance Assessment Report PSR Project Status Report SB Superintendencia de Bancos (Superintendency of Banks) SBC Superintendency of Banks sv SuperintendenciaJdevalores (Superintendency of Securities) TA Technical Assistance FOR OFFTCTAL USE ONLY IEG Mission: Enhancing development effectiveness through excellence and independence i n evaluation. About this Report The Independent Evaluation Group assesses the programs and activities of the World Bank for two purposes: first, to ensure the integrity of the Bank's self-evaluation process and to verify that the Bank's work is producing the expected results, and second, to help develop improved directions, policies, and procedures through the dissemination of lessons drawn from experience. As part of this work, IEG annually assesses about 25 percent of the Bank's lending operations. In selecting operations for assessment, preference is given to those that are innovative, large, or complex; those that are relevant to upcoming studies or country evaluations; those for which Executive Directors or Bank management have requested assessments; and those that are likely to generate important lessons. The projects, topics, and analytical approaches selected for assessment support larger evaluation studies. A Project Performance Assessment Report (PPAR) is based on a review of the Implementation Completion Report (a self-evaluation by the responsible Bank department) and fieldwork conducted by IEG. To prepare PPARs, IEG staff examine project files and other documents, interview operational staff, and in most cases visit the borrowing country for onsite discussions with project staff and beneficiaries. The PPAR thereby seeks to validate and augment the information provided in the ICR, as well as examine issues of special interest to broader IEG studies. Each PPAR is subject to a peer review process and IEG management approval. Once cleared internally, the PPAR is reviewed by the responsible Bank department and amended as necessary. The completed PPAR is then sent to the borrower for review; the borrowers' comments are attached to the document that is sent to the Bank's Board of Executive Directors. After an assessment report has been sent to the Board, it is disclosed to the public. About the IEG Rating System The time-tested evaluation methods used by IEG are suited to the broad range of the World Bank's work. The methods offer both rigor and a necessary level of flexibility to adapt to lending instrument, project design, or sectoral approach. IEG evaluators all apply the same basic method to arrive at their project ratings. Following is the definition and rating scale used for each evaluation criterion (more information is available on the IEG website: http://worldbank.org/oed/eta-mainpage.html). Relevance o f Objectives: The extent to which the project's objectives are consistent with the country's current development priorities and with current Bank country and sectoral assistance strategies and corporate goals (expressed in Poverty Reduction Strategy Papers, Country Assistance Strategies, Sector Strategy Papers, Operational Policies). Possible ratings: High, Substantial, Modest, Negligible. Efficacy: The extent to which the project's objectives were achieved, or expected to be achieved, taking into account their relative importance. Possible ratings: High, Substantial, Modest, Negligible. Efficiency: The extent to which the project achieved, or is expected to achieve, a return higher than the opportunity cost of capital and benefits at least cost compared to alternatives. Possible ratings: High. Substantial, Modest, Negligible. This rating is not generally applied to adjustment operations. Sustainability: The resilience to risk of net benefits flows over time. Possible ratings: Highly Likely, Likely, Unlikely, Highly Unlikely, Not Evaluable. lnstitutional Development Impact: The extent to which a project improves the ability of a country or region to make more efficient, equitable and sustainable use of its human, financial, and natural resources through: (a) better definition, stability, transparency, enforceability, and predictabilityof institutional arrangements and/or (b) better alignment of the mission and capacity of an organization with its mandate, which derives from these institutionalarrangements. Institutional Development Impact includes both intended and unintended effects of a project. Possible ratings: High, Substantial, Modest, Negligible. Outcome: The extent to which the project's major relevant objectives were achieved, or are expected to be achieved, efficiently. Possible ratings: Highly Satisfactory, Satisfactory, Moderately Satisfactory. Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Bank Performance: The extent to which services provided by the Bank ensured quality at entry and supported implementation through appropriate supervision (including ensuring adequate transition arrangements for regular operation of the project). Possible ratings: Highly Satisfactory. Satisfactory. Unsatisfactory, Highly Unsatisfactory. Borrower Performance: The extent to which the borrower assumed ownership and responsibility to ensure quality of preparation and implementation, and complied with covenants and agreements, towards the achievement of development objectives and sustainability. Possible ratings: Highly Satisfactory, Satisfactory, Unsatisfactory, Highly Unsatisfactory. This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not be otherwise disclosed without World Bank authorization. Contents Principal Ratings ................................................................................................................v Key Staff Responsible ..................................................................................................... vn .. Preface ...............................................................................................................................ix Summary........................................................................................................................... xi 1. Introduction.................................................................................................................. 1 The Fiscal Roots of the Crisis ...................................................................... 1 Colombia's Financial Sector in the Early 1990s.......................................... 2 . 2 The Financial Crisis: 1998-99 ..................................................................................... 3 Onset of the Crisis ........................................................................................ 3 Government Response ................................................................................ -3 Bank Response ............................................................................................. 5 . 3 The Financial Markets Development Technical Assistance Project .......................7 Analytical and Financial Support. 1993-97 ................................................. 7 Project Preparation. Objectives and Design................................................. 7 Implementation .......................................................................................... 10 Assessment and Outcomes......................................................................... 10 . 4 The Financial Sector Adjustment Loan (FYOO) ......................................................17 Preparation. Objectives and Design .......................................................... 17 Implementation ......................................................................................... -20 Assessment and Outcomes......................................................................... 21 5 . The Programmatic Financial Sector Adjustment Loans I and I1 .........................26 Introduction................................................................................................ 26 PFSAL I: Preparation. Design and Implementation .................................. 28 PFSAL 1 1: Preparation. Design and Implementation................................. 31 Assessment of the Program ........................................................................34 . 6 Lessons ........................................................................................................................ 36 . Annex A Basic Data Sheet ..............................................................................................39 Annex B. Key Objectives ................................................................................................47 Annex C. List of People Met .......................................................................................... 53 Tables Table 5.1 : Main Economic Indicators............................................................................... 30 Table 5.2. Banking Sector Indicators................................................................................ 30 Principal Ratings Financial Markets Development Technical Assistance (TF-22479; SCL-41960) ICR* ICR Review** PPAR Outcome Satisfactory Satisfactory Moderately Unsatisfactory Sustainability Likely Likely Likely Institutional Development Substantial Substantial Modest lmpact Bank Performance Satisfactory Satisfactory Moderately Unsatisfactory Borrower Performance Satisfactory Satisfactory Moderately Satisfactory Financial Sector Adjustment Loan (FSLT- 70000) ICR* ICR Review** PPAR Outcome Highly Satisfactory Moderately Satisfactory Satisfactory Sustainability Likely Non-evaluable Likely Institutional Development High Modest Substantial lmpact Bank Performance Satisfactory Satisfactory Moderately Unsatisfactory Borrower Performance Highly Satisfactory Satisfactory Satisfactory First Programmatic Financial Sector Adjustment Loan (Loan No. 71674'0) ICR* ICR Review** PPAR Outcome Satisfactory Satisfactory Satisfactory Sustainability Likely Likely Likely Institutional Development Substantial Substantial Substantial lmpact Bank Performance Highly Satisfactory Satisfactory Satisfactory Borrower Performance Highly Satisfactory Satisfactory Satisfactory Second Programmatic Financial Sector Adjustment Loan (TF-52531; FSL T- 72580) ICR* ICR Review* PPAR Outcome Satisfactory Moderately Moderately Satisfactory Satisfactory Sustainability Likely Likely Likely Institutional Development Substantial Substantial Substantial lmpact Bank Performance Satisfactory Satisfactory Satisfactory Borrower Performance Highly Satisfactory Satisfactory Satisfactory * The Implementation Completion Report (ICR) is a self-evaluation by the responsible operational division of the Bank. '*The ICR Review is an intermediate Independent Evaluation Group (IEG) product that seeks to independently verify the findings of the ICR. Key Staff Responsible Financial Markets Development Technical Assistance (TF-22479; SCL-41960) Project Task ManagedLeader Sector Manager Country Director Appraisal Andres D. Jaime Krishna Challa Paul lsenman Completion John Pollner John Pollner (Acting) Isabel M. Guerrero Financial Sector Adjustment Loan (FSLT- 70000) Project Task ManagedLeader Sector Manager Country Director Appraisal Krishna Challa Augusto de la Torre* Olivier Lafourcade Completion Krishna Challa & Fernando Montes- Isabel Guerrero John Pollner Negret * Lead Specialist First Programmatic Financial Sector Adjustment Loan (Loan No. 716 7 4 0 ) Project Task ManagedLeader Sector Manager Country Director Appraisal John Pollner Fernando Montes- Isabel M. Guerrero Negret Completion John Pollner Fernando Montes- Isabel M. Guerrero Negret Second Programmatic Financial Sector Adjustment Loan (TF-52531; FSL T-72580) Project Task Manager/Leader Sector Manager Country Director Appraisal John Pollner Danny Leipziger Isabel M. Guerrero Completion Constantinos A. Makhtar Diop Isabel M. Guerrero Stephanou Preface Attached is the Project Performance Assessment Report for four loans to Colombia: the Financial Markets Development Technical Assistance Loan (FMDTA), for $15 million, approved in June 1997, the Financial Sector Adjustment Loan (FSAL) for $500 million, approved in November 1999, and the Programmatic Financial Sector Adjustment Loans I and 1 1(PFSAL I and PFSAL II), for $150 and $100 million, approved in April 2003 and September 2004, respectively. An Independent Evaluation Group (IEG) mission visited Colombia in November 2005 to discuss the effectiveness of Bank assistance with government officials, donors and other stakeholders. Their cooperation and assistance in preparing this report is gratefully acknowledged Comments from the Bank's Regional Management were incorporated in the report. Following standard IEG procedures, copies of the draft PPAR was sent to government officials and agencies for their review and comments. No comments were received. This report was prepared by Mr. Manuel Penalver (Consultant) under the supervision of Ms. Laurie Effron and Chad Leechor. Roziah Baba, Janice Joshi and Agnes Santos provided administrative support. Summary 1. Attached is the Project Performance Assessment Report for four loans to Colombia: the Financial Markets Development Technical Assistance Loan (FMDTA), for $15 million, approved in June 1997, the Financial Sector Adjustment Loan (FSAL) for $500 million, approved in November 1999, and the Programmatic Financial Sector Adjustment Loans I and II (PFSAL I and PFSAL II), for $150 and $100 million, approved in April 2003 and September 2004, respectively. The FSAL, PFSAL I and PFSAL I1 were all designed to support Colombia's financial sector in the aftermath of the economic and financial crisis that affected the country in 1998-99. 2. Colombia experienced a rollercoaster of events on the economic front during the decade covered by the four loans. At the beginning of the 1990s, Colombia was recovering from economic difficulties in the 1980s, which had included a banking sector crisis. By 1993, the economy was growing; the banking sector was profitable and appeared strong, though the financial sector and the capital markets in particular were less developed than in other countries at a similar level of development. At the same time, important risks to the soundness of the financial sector were developing. One was changes in the indexation system for housing loans starting in the early 1990s which put borrower's and the lenders' portfolios at risk. Another was the increasingly poor performance of (and corruption in) the government-owned banks. A third was the increasing fiscal deficits since 1990. 3. Bank analytical work in the financial sector focused on the development of capital markets (including preparation of the FMDTA). By mid-1998, growing fiscal and current account deficits were aggravated by the changed international environment that prevailed following the East Asia crisis in mid-1997. Attempts by the Central Bank to support the exchange rate aggravated the crisis: interest rates mounted to extremely high levels, and the fiscal and external difficulties were translated into a fiscal crisis as borrowers were unable to service their loans and the banks' portfolios, liquidity and, eventually, solvency deteriorated. This resulted in a full-scale economic and financial crisis in 1998-99. - 4. Design of the FMDTA was based on an optimistic diagnostic of country and sector conditions that had already changed by the time of approval. The growing risks to the financial sector during the 1990s were absent from the Bank's analysis: relevance was negligible when the crisis broke out. Implementation was slow, lasting six years, and nearly half of the loan amount was cancelled. Relevance during the last two years of the loan was higher (after some restructuring) as some of the components were used to support recovery efforts from the crisis. The outcome of FMDTA is rated moderately unsatisfactory. 5. The Colombian government requested Bank assistance in the form of emergency adjustment lending immediately after a new administration took office in August 1998. But lack of preparedness on the Bank's side, and the rapidly deteriorating situation in late 1998 and early 1999, made the preparation of the FSAL a longer process than the government had anticipated in late 1998. In early 1999 the Bank decided that its support should not be through an emergency SAL but through a more conventional, two tranche FSAL, and the Bank and Fund agreed to carry out their first joint FSAP in Colombia. 6. The FSAL was approved in late 1999. By that time, the worst of the crisis was over, but the damage to the financial sector and the economy had already been substantial. The government had introduced several rounds of measures from November 1998 to late 1999 that were successful in avoiding a systemic crisis, although many financial institutions were closed down, and others merged or were taken over by the government. The deposit guarantee fund (FOGAFIN) became also a banking resolution agency. The several sub-sectors in the financial sector were affected differently: the mortgage banks, credit cooperatives and public sector banks were the hardest hit by the crisis while private sector commercial banks remained comparatively strong. Although the Bank's approval of the FSAL was not timely in terms of containing the worst of the crisis, it was nevertheless important in the recovery from the crisis and in laying the foundations to reduce risks and vulnerabilities that could cause a repeat of the problems. By the time of the FSAL tranche second release (late 2000), the international markets had improved, and the government requested that the second tranche of the FSAL be converted into a Policy Based Guarantee (PBG) to facilitate borrowing in the international markets. The Bank agreed and a PBG was approved in February 2001. In spite of the initial delays and underestimation of the importance of the crisis in the early period of preparation, relevance of FSAL was substantial and its outcome is rated satisfactory. 7. The 2002 country assistance strategy discussed the financial sector in detail, including a proposed support program to be provided through two PFSALs, for $150 million each (in FY03 and FY04). The program addressed the "mopping up" of issues still pending from the earlier crisis, including financial distress by some institutions, resolution and privatization of public and "officialized" banks, a new legislative framework for the banking sector, as well for the rest of the financial sector and the securities market, new prudential regulations, supervision and surveillance, and institutional support and development for the public agencies in the sector. 8. PFSAL I was approved in April 2003 and achieved most of its objectives in the context of a rapidly improving economic environment: GDP growth in 2003 was well above the country strategy and other projections (4 percent instead of the 2 percent projected) and the fiscal deficit was reduced. The improving economic conditions facilitated substantially the achievement of the "mopping up" component of the PFSAL I, whereas other areas progressed at different speed: improvements in housing finance and in private capital markets development fell short of expectations. In spite of these shortcomings, relevance of PFSAL I was substantial and its outcome is rated as satisfactory. 9. Country economic and financial sector conditions had improved further by the time of approval of PFSAL I1 in September 2004: the program document notes that high capital inflows in early 2004 were causing an appreciation of the peso in spite of the Central Bank attempts to prevent it by lowering interest rates and accumulating,reserves. With the financial sector recovery well under way as shown by its indicators of profitability and solvency, PFSAL I1 focused on legislative, regulatory and institutional development issues as well as remaining vulnerabilities and the lack of recovery of housing finance, particularly for low income housing. Given the medium term nature of many of these objectives (and the availability of other capital inflows) a single-tranche adjustment operation was not an obvious choice to support their achievement. In the event, improvements in the regulatory environment of the banking system took place, but improvements in housing finance failed once again to materialize and were very limited in the private capital markets area. The relevance of PFSAL I1 was modest and its outcome moderately satisfactory. 10. The medium term aspects of the financial sector support program contemplated in the country assistance strategy would have been better served by a programmatic approach covering a longer period of time, (e.g., five years instead of the two contemplated by the FSALs), and including not only quick disbursing but also accompanying institutional development support through other instruments (TA loans, IDF or FIRST grants). A longer framework for the program could also have provided more flexibility to adjust the later interventions to the changing country conditions. 11. Overall, the Bank's support to Colombia's financial sector has been positive and the outcome of the three adjustment operations was satisfactory, in spite of the relevance issues related to PFSAL 11. The outcome of the earlier FMDTA, however, was unsatisfactory because the project design was inadequate to address the issues facing the sector at the time, and achievement of project objectives was hampered by the financial sector crisis: Colombia's capital markets are arguably less developed now than at the time of the approval of the FMDTA. Yet, even in this case, the last few years of the FMDTA provided support for institutional development of the sector, in parallel to the FSAL. 12. Bank performance was satisfactory, except in the cases of the FMDTA and FSAL. The re-emergence of a potential crisis situation during 2002, and Bank work on the financial sector which found a significant agenda of unfinished business, warranted substantial Bank assistance. It justified the decision to support further reforms through programmatic adjustment lending in 200312004. On the other hand, the 1997 FMDTA was poorly designed, without reference to the emerging internal risks, including macro- economic risks, and lacked relevance. In the early stages of preparation of the 1999 FSAL, the Bank likewise underestimated the extent of the crisis, while, partly because of a lack of preparedness to deal with financial crisis situations, its assistance was delayed until after the worst of the crisis was over. 13. Finally, throughout the implementation of the three adjustment operations, one policy issue has remained pending: in the midst of the 1998-99 crisis, the government introduced a temporary tax on financial transactions that was intended to apply for 18 months and to contribute to pay the costs incurred during the financial crisis. Several years later, however, the financial transactions tax and the distortions it introduces in the sector are still in place (at a higher rate that when originally introduced) and the issue needs to be addressed. PFSAL 11, at a time when the fiscal situation was improving rapidly, seems to have been a major missed opportunity to eliminate the financial transactions tax. 14. The main lessons from these loans are: Accurate, up-to-date sector information and analytical knowledge is critical for an adequate support to thefinancial sector: an outdated assessment of the sector by 1997 made the FMDTA irrelevant and blindsided the Bank to the vulnerabilities of the financial sector. It also slowed down the Bank's response after the onset of the crisis. Agility and quick reaction to crises requires adequate technical resources and streamlined decision-making: the Bank compounded its inadequate prior knowledge of the sector in 1998 by allocating insufficient technicalhuman resources until the joint FSAP was launched, and by slow and bureaucratic decision-making in the face of urgent client demands. Agility, flexibility, and openness to change are also crucial during implementation of a project orprogram: during implementation of FMDTA, the inadequate starting point could have been partly remedied by earlier and deeper restructuring to face the new conditions on the ground; during implementation of the FSALs, improved economic conditions and slow progress in some areas (housing finance, capital markets) could have been better addressed (even in the absence of a TA project) through a change in the second phase of the program (e.g. by splitting PFSAL I1 into two smaller, more spaced operations implemented over a longer period of time). Exploiting the synergies among different Bank instruments is crucial: technical assistance operations and adjustment lending can reinforce each other when they are part of a coordinated sector intervention (and based on adequate analytical work). In Colombia, an outdated TA project was only belatedly and marginally linked to the FSAL; during the short period of the PFSALs (and after them) there was no TA vehicle on the ground to continue the support. A programmatic approach tofinancial sector reforms is likely to require a time horizon of several years to allowfor policy and institutional reforms to be translated into outcomes and results: the two-phase single year approach of the PFSALs was too short to address the complex second-and third-generation issues facing Colombia's financial sector. Vinod Thomas Director-General Evaluation 1. Introduction 1.1 Colombia had a long history of prudent macroeconomic management until the early 1990s, even in the face of periodic internal upheavals leading to civil strife and violence. One indicator of the success of this management is that in the entire twentieth century there were only two years with negative growth of GDP (1933 and 1999: the first episode was a result of the Great Depression; the second was part of the 1998-99 financial crisis). Another indicator is that Colombia did not request IMF support between the mid-1 960s and 1999: and even during the most recent period, the Government of Colombia has treated the Extended Fund Facility of 1999 and two subsequent Stand-By Arrangements as precautionary, not drawing financial resources from any of them. 1.2 Colombia's economic growth, however, has only been modest, hovering for many years at about 4 percent per year, with some episodes of substantially lower growth. At the same time, the financial sector has been less developed than that of other middle- income countries (e.g. in terms of monetary aggregates as a share of GDP), and has experienced several episodes of crisis during the last two decades. The modest economic growth is an indication of the lost opportunities of the Colombian economy that, today, is still hovering in the mid-range of the middle-income countries and has been unable to provide its population with a substantial and sustained increase in its standard of living. 1.3 During the last ten years, Colombia has experienced several episodes of economic difficulties. Although parallel in timing to the crises in East Asia, Russia and other Latin American countries, the Colombian economic difficulties have been at least as much the result of internal factors as of external shocks in the mid- and late-1990s. In the most recent years, the impact of external factors has been largely positive (particularly the increases in oil prices) which has facilitated the resolution of the crises without the need to recourse to IMF financing. These positive external factors (which included high inflows of foreign capital starting in the early 1990s, and again in the last two years), however, have reduced the urgency to address some of the root causes of the difficulties (particularly the structural fiscal deficit) and may have left the Colombian economy exposed to further difficulties after the windfall effects of high oil prices disappear, together with the expected decline in Colombia's oil production in the coming years. The Fiscal Roots of the Crisis 1.4 The main source of Colombia's economic difficulties in the 1990s and early 2000s was fiscal. During the two successive administrations of Presidents Gaviria (1990- 94) and Samper (1994-98) there was a continuing increase in public sector expenditures as a share of GDP, which increased from 21 percent to 28 percent in 1994 and 32 percent in 1998. The public sector accounts, which were largely in balance by the early 1990s, were showing a deficit of 2.6 percent of GDP in 1996 and 4 percent in 1997. An important source of the fiscal difficulties, in addition to the increase in expenditures, was the constitutional changes of 1991, which allocated increasing amounts of fiscal revenues to sub-national,entities, thus limiting the resources available to the central government. 1.5 The twin increases in public expenditure and in the fiscal deficit were easily financed in the expansionary climate that characterized the early and mid-1990s, with large increases in foreign capital inflows to developing countries. Colombia, with a reputation for good financial management, had little difficulty in attracting the large capital inflows it needed to finance its fiscal deficits. The Asian crisis in 1997 followed by the Russian one in 1998 led to an overall drop in capital flows to developing countries as well as higher spreads on international lending to developing countries. These factors combined with declining prices for Colombia's exports put pressure on the peso. In this context, the large and growing fiscal deficits led to strong pressures on domestic interest rates and were one of the major causes of the resulting financial sector crisis. Colombia's Financial Sector in the Early 1990s 1.6 Colombia's financial sector had experienced a banking crisis in the 1980s, but was generally considered fully recovered, although still small and highly concentrated, by the early 1990s. Financial sector policies after the early crisis had introduced a number of improvements, including recapitalization of banks, elimination of (or substantial reductions in) forced investment requirements by the banks, and promotion of competition (including entry of foreign banks). Other improvements took place during the 1990-94 period, including a pension reform in 1993, institutional improvements in the Superintendency of Banks (SBC) and privatization of some banks previously taken over by the government during the 1980s banking crisis. 1.7 By the mid- 1990s, the financial sector was considered, both in Colombia and in the Bank, a sector moving towards further strengthening and deepening, which only required assistance to consolidate previous gains and to introduce greater diversification of financial services. The small size and high concentration were expected to be addressed by the twin demand and supply factors of increasing demand for financial assets, as confidence in the sector continued to increase, and by the entry of new institutions (including foreign banks). The still relatively large share of the public sector in the financial sector was considered a minor cause for concern, to be gradually addressed. 1.8 The above assessment of the financial sector is reflected in the main Bank documents of the period, including the Financial Sector Review of 1993, the Capital Markets Report of 1996, the Financial Markets Development TA Project of 1997 (the first of the four projects included in this PPAR), and the mid-1990s country assistance strategy. Yet, this view of the financial sector as a basically sound sector that required only "high end" assistance on the more sophisticated aspects of instruments, institutions and regulation was to be shattered within less than a year of the approval of the FMDTA project. In addition to the prevailing fiscal problems, two other internal sources of risk to financial sector soundness were: the increasingly poorer performance of (and corruption in) the government-owned banks; and, changes in the indexation system for housing loans starting in the early 1990s which put borrowers and the lenders' portfolios at risk. 2. The Financial Crisis: 1998-99 Onset of the Crisis 2.1 The gradual deterioration of Colombia's economic performance since the mid- 1990s (fiscal deficits above 4 percent of GDP, current account deficits exceeding 5 percent of GDP and growth slowing down to about 2 percent) came to a head in 1998, aggravated by internal (political) and external factors (declines in oil and coffee prices, tightening of international credit markets). In the financial sector, rising interest rates were weakening the portfolio of the mortgage financing institutions, and poor lending practices in the publicly owned banks added to the vulnerabilities of the sector. The tightening of international credit markets, and the initial government responses, particularly regarding exchange rate policy, came as an additional shock in 1998. 2.2 In early 1998 economic activity was slowing down, the banking portfolios were showing some signs of distress, the fiscal deficit was rising, and the Colombian peso was under pressure: the Central Bank's attempts to support it (with only minor adjustments to the exchange rate band) in the run up to the May presidential elections led to a loss of reserves of $250 million. More importantly, interest rates increased to very high levels in real terms. In addition, although Colombia had initially maintained a good international credit rating, following the East Asian crisis and the Russian default of August 1998 (when the new Colombian government took office), the spreads on international loans grew to more than 900 basis points (bsp). 2.3 In the financial sector, the combination of high fiscal deficits, credit demand from sub-national levels of government (including municipalities) and the private sector, and the tightening of international capital markets led to a rapid increase in (real) interest rates on loans from about 14 percent in late 1997 to around 26 percent in late 1998. The combination of high interest rates and slowing economic activity led to a deterioration of the loan portfolio of many financial institutions, as borrowers were increasingly unable to service their loans promptly. The resulting increase in the arrears reported by the financial institutions, however, was underestimated because of weak financial sector supervision and the reluctance of financial institutions to fully report the volume of their poor performing loans. Government Response 2.4 In August 1998, when President Pastrana's government took office, the extent of the country's economic difficulties was clear in terms of the slowing down in economic activity, increasing unemployment (nearly 17 percent in urban areas in late 1998), a growing fiscal deficit and a continuing large current account deficit. The extent of the financial crisis was less evident because of the underestimation of arrears by the financial institutions, although an ongoing review of banking supervision practices had raised serious alarms and the incoming government was alerted to the risks. 2.5 Shortly after taking office, the government introduced a package of fiscal measures, including a reduction of (budgeted) expenditures for the year of 0.5 percent of GDP, and a fiscal package to increase revenues by about 1 percent of GDP. Revenue measures, however, required congressional approval and were to take effect mostly in 1999. At about the same time (September 1998) the Central Bank widened the exchange rate band, which resulted in a devaluation of about 6 percent. 2.6 In parallel, the new government took a new approach in its relations with the international financial institutions and requested large increases in financing by the World Bank, the Inter-American Development Bank (IDB) and the Corporation Andina de Fomento (CAF), while remaining opposed to entering into an agreement with the IMF (the IMF was providing "enhanced surveillance" through semi-annual reviews, but Colombia was keen on staying away from formal IMF programs, the last of yhich had taken place in the 1960s).' The first requests for increased assistance by the Bank took place during a programming mission in September 1998, at the Annual Meetings in early October, and again during a meeting between President Pastrana and Mr. Wolfensohn in late October 1998. 2.7 The seriousness of the financial sector situation became increasingly evident as financial sector arrears mounted and a large mortgage-financing institution was intervened by the government (in October 1998). This prompted the government to prepare a package of emergency measures, approved by Decree, in November 1998, to deal with the growing financial c r i ~ i s . ~ 2.8 The November 1998 emergency measures included relief measures for low income mortgage borrowers (refinancing at preferential interest rates) and small depositors, and other measures to help with problems related to problem mortgage loans, as well as support measures for public banks, financial cooperatives and private banks. All of these measures, however, were subject to ratification by the Constitutional Court. To provide financing for the relief measures, the government introduced a temporary financial transactions tax, effective through December 3 1, 1999. During the second half of 1998, the government also had to intervene and take over two medium size private credit institutions (an S&L institution and a cooperative), as well as several other smaller cooperatives, an early indication that further such interventions would become necessary. 2.9 In early 1999 the crisis continued and additional government measures became necessary. Following the ratification by the Constitutional Court (in March 1999) of some, but not all, the November emergency measures, the government announced a new program for recapitalizing distressed banks in April: the program went into effect in June 1999. By that time, the financial sector distress was widespread, total losses in the financial sector were mounting (although largely concentrated in the mortgage banks and ' The Pastrana government was also keen to be seen as not influenced by the IMF because of political reasons, and particularly as it embarked into an ambitious peace initiative with the guerrillas. By that time, the discussions with the World Bank had already focused on the preparation of a Financial Sector Adjustment Loan (FSAL), which the government requested on a priority basis and expected to have available in early 1999 (and possibly as a "Special SAY). Eventually this became the FSAL approved in November 1999. publicly owned banks), and Standard and Poor's had downgraded the ratings of the four leading private banks. 2.10 By mid- 1999 the country's sovereign risk had also been downgraded, the Central Bank had introduced yet another small discrete adjustment in the exchange rate band, and the government had started discussions with the IMF on an EFF program. The government also introduced (and Parliament approved) a new Financial System Framework Law, and closed one of the public banks (Caja Agraria). Later in the year (September) the Central Bank abandoned the exchange rate band and moved to a free float. 2.1 1 During 1999, economic activity slowed down substantially, with GDP registering a decline of more than 4 percent. The recession helped reduce the level of interest rates from their peak in early 1998 and also helped reduce inflation from its historical levels of more than 20 percent during the previous three decades. The worst of the financial crisis was over by early 2000 as confirmed by an FSAL supervision mission. Bank Response 2.12 During the early 1990s, in an environment of easily available private financing-in international markets, Colombia had reduced its borrowings fiom the World Bank, and was gradually reducing total outstanding loans from the bank. The mid-1 990s country assistance strategy assumed these trends would continue, and missed the vulnerabilities arising from the large fiscal and current account deficits in a different external credit situation. The strategy (four months after the devaluation of the Thai baht, that signaled the end of the 1990-97 period of growing private capital flows to developing countries), was written at a time when Colombia's growth had declined, fiom 5.4 percent in 1995 to just over 2 percent in 1996, the current account deficit was about 5 percent in the previous 4 years and the fiscal deficit turned from about 0 in the 1991-94 period, to nearly 3 percent of GDP in 1996. 2.13 The mid- 1990s country assistance strategy discussed the potential risks of the above trends, but reached an optimistic assessment, based on expected sharp increase in oil production and the turn around in growth after the second trimester of 1997. In the event, GDP growth slowed down to less than 1 percent in 1998 and turned strongly negative in 1999. In other words, the Bank's vision of Colombia in late 1997 was that, in spite of a continuing deterioration in the fiscal performance, the economy would rebound strongly and that Colombia's previous access to external capital markets would continue at a pace that had made borrowing from the Bank unne~essary.~ 2.14 The financial sector was largely absent from the mid-1990s country assistance strategy. The FMDTA project had been approved a few months before the country strategy discussion and, with no other financial operations in the pipeline, the financial sector was only mentioned in passing when discussing the IFC program. The six The mid-1990s country assistance strategy noted that net Bank disbursements to Colombia had turned negative and debt outstanding and disbursed (DOD) had declined and mentions a "catalytic role" for the Bank and small-staff-intense projects. priorities in the country assistance strategy (peace, rural development, human capital, public sector efficiency, infrastructure and sustainable development) covered a wide terrain but paid little attention to the financial sector. 2.15 When the new government realized that it needed additional international financing that could only be obtained from the IFIs, and approached the World Bank, IDB and CAF, the Bank was unprepared. After several rounds of discussions in September and October of 1998, by early November discussions had focused on a possible FSAL for about $500 million and an overall lending program considerably above the levels discussed in the mid-1990s country assistance strategy. Yet, for a variety of reasons (discussed in Section lV below) the FSAL was only approved a h l l year later, in November 1999. A country assistance strategy progress report was also discussed at that time. Also during 1999 the first joint Bank-IMF FSAP took place in Colombia, and the government reached agreement with the IMF on an EFF. 3. The Financial Markets Development Technical Assistance Project Analytical and Financial Support, 1993-97 3.1 The Bank's analytical work on the financial sector during the early and mid- ~, 1990s, included a Financial Sector ~ e v i e wfollowed by a Capital Markets Report, in 1996,' and complemented by a review of (and assistance on) Banking Supervision in 1997. The different components of this analytical work were part of the Bank's assistance program set out in the early 1990s country assistance strategy, and guided the preparation of the Financial Markets Development Technical Assistance project of 1997. 3.2 The Financial Sector Review and the Capital Markets report were prepared during a period when the domestic financial sector had recovered from the banking crisis of the 1980s, and broader reforms, such as improvements in banking supervision (strengthening of the Superintendencia), privatization of banks previously taken over during the earlier crisis, and pension reform (approval of the Pension Funds Law in 1993) were being implemented. Their optimistic assessment was influenced also by an international environment in which foreign investment directed to developing countries was increasing at a fast pace. Project Preparation, Objectives and Design 3.3 The FMDTA project preparation started in 1992 in parallel with (and based on the analysis and conclusions of) the Financial Sector Review of 1993,6but the preparation suffered from many delays, was put on hold in 1994 and re-started in 1996.' The project approved in 1997 (see below) was substantially different from the one described in the ambitious initial design which included policy reforms, technical assistance for capital markets development, and financing for a backstop facility (to extend maturities of newly issued securities) and from the one described in the project summary, which contemplated two separate operations: a technical assistance project and a separate, larger loan or guarantee for a backstop facility. Both designs were largely Bank- (supply-) driven designs, based on the analysis of the Financial Sector review, but with limited ownership by the government that was, at the time, reducing its borrowings from the Bank and other IFIS.' "Colombia: Financial Sector Reform" 1992. "Mission de Estudios del Mercado de capitals" July 1996. The mission took place in 1994. The team visited Colombia in September 1991 and September 1992. ' The ICR for the project (Report No. 28928, of June 20,2004) presents a good description of the complex rroject preparation (in Section 3.3, pages 4 to 6). Bank staff commented that the project took longer than average to prepare because it spanned an election period during which an opposition party with different priorities was elected. Only after completion of a new government study on local capital markets in 1996 could the project be realigned. Thus, rather than constituting a "supply-driven" approach, changes in project design reflected a responsiveness to client priorities. 3.4 After a new government took office in August 1994, project preparation was put on hold, partly because of the availability of a large Japanese grant approved in 1994, which, in the government's view, provided ample financing for the technical assistance requirements for the capital markets. The government remained cool to the ambitious proposals put forward by the Bank, and particularly to the idea of a Bank loan to finance a backstop facility. The newly approved Japanese grant was also used to fund the Capital Markets mission of October 1994 (whose final report was issued in 1996). 3.5 Appraisal of FMDTA took place in November 1996 (more than four and a half years after the initial project summary), with the benefit of the new capital markets work, and resulted in a new project design, supported by a loan of $15 million, that was finally approved in June 1997. The project was a self-standing technical assistance project with a total cost of $20 million, covering a variety of issues and making it a complex operation, particularly in terms of implementation arrangements: the project included a total of six components with 11 implementing agencies. At least some of the components, conceived in the original project design, had become dated: this is the case for the component to support Social Security reform, a reform that had been ongoing since the approval of the new social security system in 1993.' The Bank loan was to finance mostly consultant services ($1 1 million) as well as some training ($2.3 million) and computer and office equipment ($1.7 million) 3.6 Components: The six original components of the project, their share in total project costs and the main implementing agencies were: Develop the financial markets into a more eficient channel for mobilizing savings andfinancing investment. This component accounted for about seven percent of total project costs, and included support for the Superintendency of Securities (Superintendencia de valores-SV), the borrowing department (Credito Publico) of the Ministry of Finance, and also the department of financial regulation (Direccion General de Regulation Financiera-DGRF in the MoF) Improve the institutional capacitiesfor financial sector policy formulation, and strengthen regulatory and supervision framework. This component, accounting for 24 percent of total project costs, provided support to four agencies: DRGF, SV, Superintendency of Cooperatives (Superintendencia de Economia Solidaria- SES) and Superintendency of Banks (Superintendencia de Bancos-SB). Increase the supply of securities byfacilitating the introduction of newfinancial instruments. This component accounted for 14.5 percent of total project costs and provided support to the SV. A key aspect of this component is that it identified, at an early stage, the securitization of mortgages as a significant issue that would later become even more important. Bank staff disagreed, arguing that quantifying the public pension liability was key to securing future fiscal sustainability. Consolidate the Social Security Reform. This component, accounting for 24.5 percent of the project cost changed its rationale during the long gestation of the project, as the social security reforms of 1993 developed. In its final form it supported the estimation of the costs of the government's pension liabilities, and provided support to a department in the Ministry of Finance and the pensions Directorate in the SB. Deter corrupt or fraudulent practices such as money laundering. This component, for 15 percent of project costs, supported the establishment of a new agency (the UIAF) and development of a system of financial intelligence. Foster more eficient operation of markets by promoting competition and deterring anti-competitivepractices. This component, for 10 percent of proj ect costs, supported the development of an anti-trust policy capacity, providing support for the Superintendency of Industry and Commerce. 3.7 In addition to the original six components, new components were included as part of an amendment and project restructuring in October 2000, to include the Deposit Guarantee Fund (FOGAFIN), and support to the municipality of Bogota on its debt issuance. 3.8 By the time the above project design was approved, in June 1997, the underlying conditions in the country's financial sector had changed substantially from those prevailing at the time of the original project design: the weaknesses in the capital markets remained a major feature of the financial sector, and the analysis in the Capital Markets Report was able to identify specific recommendations for improving the functioning of the capital markets. But the overall economic conditions in the country, the changing international environment, and the increasing financial sector distress, resulting from internal financial sector risks which had been ignored in project design, reduced the relevance of the FMDTA project and had a negative impact on its implementation immediately after it was approved.1° 'O Bank staff commented that two unforeseen exogenous events - the emerging financial markets crisis which started in Asia, and a Colombian Constitutional Court ruling which shifted interest rate risk massively to domestic credit institutions - were the underlying causes of the Colombian financial crisis. They point out that the country assistance strategy was approved in 1997, the same year in which the Asian crisis broke out but that the contagion effects did not reach Latin America until 1998, and Colombia until early 1999 Therefore, it would be unrealistic to expect, as IEG does, that the Bank team could anticipate a crisis which neither the Bank, the Fund, international rating agencies, or Wall Street analysts had forecast in 1997 when this technical assistance loan was being prepared. However, the Colombian authorities did state in their 1999 Letter of Intent to the IMF that, by end-1996, the health of the financial sector had already deteriorated markedly, particularly affecting financial cooperatives and the public banks. Moreover, spreads on Colombian paper had begun widening markedly during the second half of 1997, a reflection of diminished confidence and anxiety about the financial outlook. And the Constitutional Court ruling cited by Bank staff was not issued until mid-1999, almost a year after the crisis started. Implementation 3.9 Initial implementation of the FMDTA project proceeded slowly and in isolation from the broader events developing in the country and the financial sector. This was partly a result of the project design, with its multiple components and implementing agencies operating independently, and partly a result of the implementation arrangements, including a project implementation unit (PIU) in the Ministry of Finance, which focused solely on the project, and was not integrated into the discussion of the broader issues that the government and the individual implementing agencies were starting to face. 3.10 Disbursements and cancellations. Disbursements were slow throughout the implementation period. During the mid-tern review, in early 2000, there was an initial discussion of reallocations and cancellation but no cancellations took place until more than a year later. By mid-2001, when the peak of the financial crisis was long passed (and the second tranche conditions of the FSAL were being met), total disbursements had reached only $5 million, or a third of the total loan over a four-year period. Cancellation of about a third of the loan was discussed with the government again during the May 2001 supervision mission and a cancellation of $5.5 million took place in November 2001. The project continued to languish during the second half of 2001 and early 2002 (disbursements by April 2002 had reached only $6.8 million) well after the worst of the financial crisis was over, and the FSAL had closed. 3.1 1 After a new government took office in August 2002, there was a sense of renewed relevance for the FMDTA project as the government started preparation of a new round of financial sector adjustment operations (eventually, the PFSAL I and 11). The government requested a second extension of the closing date (to December 2003), which was supported by the Bank, and some activities supported by the project took on greater importance because of their links to the proposed PFSAL (e.g. support of debt workouts for companies). 3.12 By late 2003, however, only $8.2 million was disbursed and the project closed with a second cancellation of about $1.2 million. By that time, the 1999 FSAL was closed, and a subsequent PFSAL had been approved. Thus, six years after approval of the FMDTA, and nearly 12 years after the IEPS, the project had disbursed less than 55 percent of the loan amount. Assessment and Outcomes 3.13 Quality at Entry: The project design was based on a diagnostic of country and sector conditions that had already changed by time of approval (and even appraisal). The diagnostic of the sector was based on substantive policy reform and improved banking supervision undertaken after the crisis of mid-1980s. These reforms included varying interest rates, reducing segments in the financial system, greater autonomy for Central Bank and improving banking supervision. This diagnostic, in May 1997, barely a year before a major banking crisis erupted and exposed the fragilities of the banking sector, was seriously off track and diverted the attention of the project away from more important issues in the sector. Thus, quality at entry is rated as unsatisfactory. 3.14 In addition, the project risks were rated as low, including the possibility of loss of political commitment. Yet the project was presented to the Board less than a year before national elections, towards the end of an administration that was under accusations for corruption in the public sector banks, and for alleged links to drug funds in financing its campaign. Most importantly, it put the Bank-Government dialogue on the financial sector on the wrong footing at the beginning of the crisis, and put the Bank at a disadvantage to recognize at an early stage, and help deal with, the financial sector crisis. 3.15 Some project components were of questionable relevance, particularly those dealing with Social Security: a major reform of the Pension System had already taken place between initial project design and project approval. The final project design was largely supply-driven by the Bank, and included too many components and implementing agencies. On the positive side, several of the components were well identified and designed, using the lessons and recommendations of the Capital Markets report, and the project supported early work on the issues of mortgage securitization and establishment of the UIAF. 3.16 Relevance: For the reasons mentioned above, the project relevance was unsatisfactory at entrance. Also, it was unsatisfactory during the first two years of implementation and, at least until the mid-term review (early 2000), because of the absence of links with the work on FSAL, FSAP, and the government measures to deal with the financial crisis. 3.17 Relevance improved somewhat after the mid-term review but this review was a missed opportunity to have a major restructuring of the project." The second half of 2000 and 2001 saw also small improvements in relevance (particularly after the amendment adding some components-yet, no component was ever dropped during the long project life). After 2002, and during preparation of the PFSAL I, there were also some small improvements in relevance but they were limited by insufficient restructuring. Thus project relevance is rated as negligible. 3.18 Outcomes: Assessment by Objective/Component. Achievement of the six original objectives (and corresponding components) incorporated in the Matrix of key objectives and performance indicators, (as well as of the added components in 2000, included in items 1 and 2) is summarized below: II Bank staff commented that Bank efforts dating back to April 1999 to better align the project with crisis needs were delayed by opposition from substantial vested interests within the Government's project implementation units, which were reluctant to forego the budgetary allocations they had already received under the original project framework. Moreover, high turnover in the Government-appointed Project Director significantly disrupted the decision-making process. Finally, in the late 1990s, formally restructuring a project was cumbersome within the Bank, requiring lengthy Board and Borrower approvals. So, instead, Bank staff opted for a quicker, but less drastic restructuring, shifting expenditures to areas more relevant to the crisis, such as bank resolution, deposit insurance, and improved banking risk supervision. Develop the financial markets into a more efficient channel for mobilizing savings andfinancing investment. The Matrix lists six expected outcomes (improvements in securities trading, custody, settlement and clearing; well hnctioning secondary market for public sector securities; more efficient financial management and borrowing strategy of the Treasury; increased number of market makers and entities trading in government securities; improved capacity of institutional investors for risk evaluation and portfolio management; and, increased size and depth of the financial markets) and four performance indicators (total yearly transactions volume on the exchanges1GDP; total yearly transaction volume -in exchanges plus OTCIGDP; yearly volume of TES traded in the public markets1GDP and yearly average of total assets of institutional investors1GDP). At best, only two of the original outcomes (secondary market for public sector securities and Treasury financial management and borrowing strategy), plus the added support to the Bogota Municipality, and one of the monitoring indicators (volume of TES traded in the marketlGDP) were achieved.12 Thus, the outcome of this component is rated as unsatisfactory. Improve the institutional capacitiesfor financial sector policy formulation, and strengthen regulatory and supervision framework. The Matrix lists six outcomes (more effective supervision of the money and capital markets; more effective supervision of financial intermediaries; better financial sector policy formulation capacity, improved financial sector policies; effective supervision and control of the financial cooperatives system; and effective supervision of the cooperatives sector in general) and two monitoring indicators (number of persons formally trained in supervision of creditldeposit institutions and number of persons formally trained in supervision of securities market). The findings of the 1999 FSAP and its 2005 update, as well as recent reviews of the sector (e.g. in the ICR for PFSAL 11), show all of the six outcomes were achieved, to different degrees, during the project period, even if attribution is difficult and the project contribution is likely to have been very small compared to the impact of the crisis, the FSAL, the two PFSALs, and other factors. The amendment in 2000, although too late and too small, also contributed to achievement of these objectives. The outcome of this component is rated as satisfactory. Increase the supply of securities byfacilitating the introduction of newfinancial instruments. The matrix has four outcomes (increased supply of securities and availability of other financial instruments; improved access to owned-housing for low income families; improved access to tertiary education for low income students; and improved risk management for economic agents facing uncertainty related to exchange rates, interest rates, stock market level) and two performance indicators (yearly average of equity market capitalizationIGDP and number of 12 The ICR notes that the quantitative indicators for the project were too broad and ambitious, and that "the onslaught of the 1998-1999 economic/financial crisis made these indicators somewhat irrelevant" (ICR, page 10). Although recent improvements in the country's economic performance have resulted in increased securities market capitalization, the FSAP update of early 2005 found private capital markets slow, due to scarcity of local investment securities and weaknesses in clearance and settlement systems. securities andlor issuers rated per year). Only the second part of the first outcome ("Increased supply of securities and availability of otherfinancial instruments") was achieved because of the progress in securitization of mortgages. The economic and financial sector crisis meant a substantial reversal to overall financial markets in Colombia and the housing market in particular. On the positive side, a Securities Law was drafted (and subsequently approved, in 2005), largely under the support of the PFSAL I and I1 (though it was approved after PFSAL I1 closing and cannot be attributed to this project). Moreover, the ICR for PFSAL I1 notes that the new law will not have a significant impact in the short run (see Section V below). Because of the early identification of the importance of mortgage securitization, the outcome of this component is rated as moderately unsatisfactory. Consolidate the Social Security Reform. The matrix lists six outcomes (determination of the value of pension bonds and liabilities of decentralized entities; issuance of bonds by decentralized entities to the pension funds, better knowledge of the expected impact of pension liabilities on public sector finances; improved investment performance of pension fund; better capacity of Instituto de Seguros Sociales (ISS) to serve its affiliates; and, improved financial performance of the new pension fund system) and one monitoring indicator (yearly volume of pension liabilities actually determined-by number of affiliates). Two of the outcomes were fully achieved (determination of value of liabilities and better knowledge of the impact of pension liabilities on public sector finances), and some institutional strengthening took place (of the SS department in MoF and Pensions Directorate in the SB-although neither was listed in the Matrix). As noted above (and in the ICR) this was a low relevance component. Its outcome is rated as moderately satisfactory. Deter corrupt orfraudulent practices such as money laundering. The matrix lists two outcomes (detection of irregular, fraudulent or corrupt practices in financial transactions; and, practices identified and relevant authorities receive necessary information to take action) and one monitoring indicator (number of persons formally trained in [detection ofl money laundering and other fraudulent practices). The main output (establishment of the UIAF) can be considered as contributing substantially to the expected outcomes. Thus, it is rated satisfactory. Foster more efficient operation of markets by promoting competition and deterring anti-competitive practices. The Matrix includes two expected outcomes (improved quality and faster decision-making in anti-trust cases; and, reduced welfare losses from non-competitive market arrangements) and two monitoring indicators (yearly number of cases of anti-competitive practices investigated by the Superintendency of Industry and Commerce--SIC; and yearly number of anti- competitive practices decided by the SIC). Although the project provided funding to the SIC (inter alia, to improve its IT system), and the backlog of pending cases was reduced, only one third of the amount of this small component was disbursed and there is no evidence of its impact on market efficiency. Its relevance and the link with financial market efficiency are also low. The outcome of this component is rated unsatisfactory. 3.19 In addition, most of the monitoring indicators were poorly designed, and neither the ICR nor the 17 internal project assessment reports include any quantification of these monitoring indicators: all the relevant tables to these reports are empty, and the ICR includes only a small table (Annex 1) with five indicators. As noted above (footnote 12) the ICR has a similar assessment of the indicators. Data available suggests that most of the indicators in the matrix were lower at the end of the project than they were at the beginning; this was a result of the financial sector crisis and emphasizes the low relevance and quality at entry of the project. 3.20 Given the above, the outcome of the project is rated as moderately unsatisfactory. The project would have been rated unsatisfactory (or lower) except for its contributions under the second objective, particularly after the 2000 amendment.13 3.21 Institutional development is rated modest, mainly because of its contributions to several institutions supported by the project, including the UIAF, the SES (Superintendency of Cooperatives), the DGRF in the Ministry of Finance, the DSS, also in MoF, as well as the SB and FOGAFIN. 3.22 Sustainability of the achievements, particularly in the area of institutional development, is rated as likely, mainly because of the subsequent experience under the PFSAL program and recent developments. 3.23 Bankperformance during preparation and appraisal is rated unsatisfactory, because of the misjudgments regarding the financial sector condition and most relevant issues at the time, the low relevance of the overall project design, and of some components in particular, and the "supply driven" approach exhibited during the long preparation. 3.24 Early Bank supervision of the project was also unsatisfactory and reflected its isolation from the developments in the rest of the sector: throughout 1998 all components were rated as satisfactory, and the internal project assessments made no mention of the financial crisis erupting in the country, except for repeated references to problems with counterpart funds (e.g. in the supervision report of December 1998). In May and June 1998, as conditions in the financial sector were deteriorating, two successive internal project assessment reports highlighted the strengthening of the capability of participating agencies. The first mention of the financial crisis appeared in the an internal project assessment in 1999 as several implementing agencies requested re-direction of the funds to priority areas and more flexibility in the use of resources. Still, this assessment, written at the height of the financial crisis, rated the promotion of financial markets as highly satisfactory. Later in the year, the next supervision report was based on a desk review because the government requested a postponement of the supervision visit. 13 Bank staff disagreed with this characterization of the outcome, and specifically with this rating. 3.25 The project's mid-term review of the FMDTA took place in March 2000 (four months after the FSAL was approved) and was conducted together with the Task Team Leader for the FSAL. This appears to be the first link between the two operations, more than a year and a half after the government first requested emergency quick disbursing assistance to address the financial sector distress. Disbursements at the time were a very low $2.7 million, and the report mentions, once again, the issue of "counterpart funds". The mid-term review's discussion of implementation progress and development objectives was more aligned with the financial sector issues addressed by the FSAL, but the repeated references to problems with counterpart funds indicate a poor assessment of priorities by the Bank.I4The mid-term review initiated discussions of possible reallocations and cancellations but fell short of addressing the drastic restructuring that the financial sector conditions required.15 3.26 Throughout this period, the internal project assessment reports did not include any management comments, suggesting also low priority and attention level within the Bank. The assessments started to include management comments with the May 2001 report. The reports continued to rate DO as satisfactory on the basis of the project's contribution to maintaining an engagement in the sector and identifying second-generation reforms, but noting that the project's quantitative indicators for the capital markets would not be achieved. The final assessment notes, once again, the inability of the Project Unit to get sufficient budget throughout the project, but does not link this issue with the question of project relevance as seen by the government (and particularly by the Ministry of Finance, responsible for both the budgetary allocations as well as the P U . 3.27 Given the above, Bank performance during the first two years of implementation is also rated as unsatisfactory because of the lack of linkages with evolving events in the country as well as the lack of management input into the supervision reports. Subsequent Bank performance is rated as moderately satisfactory because of its gradual, but only partial, adaptation to changing country conditions (e.g. in the mid-term review) and gradually increasing management inputs in the supervision process. In the final years of the project Bank performance was satisfactory. Total project costs to the Bank were high, at over $400,000 for a project with total disbursements of about $8.2 million, and could have been better used in other areas of support to the financial crisis. For the entire project, Bank performance is rated as moderately un~atisfactory.'~ 3.28 Borrower performance is rated as moderately satisfactory. The performance of most implementing agencies was satisfactory, but overall project coordination was poor and lacking in a strategic vision to restructure the project at an early stage. The government managed its priorities through the indirect mechanism of its budgetary l 4 Bank staff commented that the shortage of counterpart funding affected all public investment projects equally at this time, due to a deepening fiscal crisis, and in no way reflected any diminished Borrower commitment to this particular project. IS Staff noted that formal project restructuring was a cumbersome process at the time, and there were no institutional incentives for task teams to propose them. While this may explain the approach taken by the team, it does not alter the (institutional) rating of Bank performance. 16 Bank staff disagreed with the negative characterization of Bank performance, and this rating specifically. allocations of the project (which led to the frequent and misguided Bank complaints about "counterpart funds"), instead of engaging the Bank in a more substantive discussion on the need to restructure the project. Administration costs of the project, to the government, are unclear as the ICR lists it at $556,000 in one table (Project Cost by Component) but only $213,000 in another (Project Financing by Source). Nevertheless, the Government did introduce a number of reforms in the sector in the aftermath of the crisis, so that on balance, its performance is considered moderately satisfactory. 4. The Financial Sector Adjustment Loan (FYOO) 4.1 Overview. As mentioned, in addition to the underlying macroeconomic causes of Colombia's financial sector crisis in 1998-99, discussed in Chapter 1, two other sources of the crisis can be found within the financial sector itself: one was the increasingly poor performance of Colombia's publicly owned banks, because of poor lending practices, political interference and corruption. The other internal source of the crisis was linked to the apparently most successful component of the financial sector: the large mortgage lending sector that had developed, since the early 1970s, based on an indexation system for mortgage loans (the WAC-Unidad de Poder Adquisitivo Constante-linked to the inflation index). 4.2 In spite of some three decades of double-digit inflation, the W A C had allowed for a huge development of mortgage lending, and the growth of financial institutions (the CAVs or Cooperativas de Ahorro y Vivienda) with a portfolio specialized in mortgage lending. Since the early 1990s, the W A C indexation system was changed to include not only the inflation rate but also a link to short term interest rates. As interest rates started to increase, the value of the UPAC-denominated mortgages also increased rapidly and, when a housing bubble burst, in 1997-98, mortgage holders found themselves with debts that exceeded the value of their houses: not surprisingly arrears increased and the portfolio of the CAVs deteriorated. 4.3 In late 1997 and early 1998, the above trends became a much stronger source of instability due to two external shocks (a tighter international financial environment, and a decline in Colombia's terms of trade), and the Central Bank's attempts to defend the exchange rate (through small adjustments in the exchange rate "band") in the face of growing pressures on the peso. The results were several large losses of reserves (with every unsuccess~l attempt to defend the "band") and very large increases in interest rates that became the detonating factor for the financial sector crisis. 4.4 The Bank was largely unprepared to support Colombia when the crisis developed and, to some extent, so was the government, in spite of several early warning signals (some financial institutions failing in early 1998, the onset of a recession, and the findings of some consultants working with the SB in strengthening banking supervision). In the event, the government that took office in August 1998 reacted much faster than the Bank, introducing successive policy packages to address the financial crisis in November 1998, and March, June and September 1999. The Bank's financial support arrived in November 1999 with the Financial Sector Adjustment Loan. Preparation, Objectives and Design 4.5 As noted in Chapter 2, the government that took office in August 1998 was confronted with growing economic and financial problems and moved on two fronts, focusing on the fiscal problems and requesting a large increase in (quick disbursing) financing from the international financial institutions.I7 The initial discussions with the Bank took place in the context of a Programming Mission (September 11-12, 1998). An internal Bank correspondence of the mission mentions three areas for quick disbursement loans: (i) the financial sector (to capitalize FOGAFIN), (ii) public sector reform (fiscal) and (iii) the peace process. The government also noted that many ongoing (and planned) investment projects were of low priority and should be cut or restructured. The government request was for disbursements of about $500 million per year in 1999 and 2o00.'8 4.6 The Bank failed to respond in a timely way to increasingly urgent requests for assistance from the government, including a letter from the Minister of Finance to President Wolfensohn on October 2, 1998. It took the Bank two months (until December 1998) to respond that it could provide a $500 million loan, but that it would require revising the country assistance strategy, and that the new strategy incorporating the new loan would be scheduled for early 1999. In the event, the loan was not approved until November 1999, well over a year after the initial discussions. 4.7 Initiating Memorandum. The initiating memorandum for the FSAL was discussed by the Operations Committee (OC) in early 1999. The memorandum mentioned the November Emergency Measures, assessing them positively, but did not discuss the need for additional measures to address the crisis. The memorandum focused instead on structural reforms. The financing needs were not presented as urgent as the operation was seen as a precautionary step to minimize risk of future crisis.19 The OC meeting endorsed the FSAL, emphasizing more the structural, not emergency aspects, and asking for back-loading of tranches. It also asked for more specific, tighter focused conditionality, sharper presentation of costs, and an explanation of how government actions would help reduce interest rates. 4.8 Between late March and mid-April, a string of announcements suggested the still growing magnitude of the crisis: the Superintendency of Banks announced that losses in financial sector doubled from January to February, with the state banks accounting for 70 percent of the total; Standard and Poor's downgraded the ratings of the four leading private banks, and FOGAFIN announced that it needed up to $800 million to shore up the 17 With a strong pressure on the peso, the Central Bank moved again the exchange rate "band" and allowed a small depreciation but pressure on peso continued; it lost more reserves (over $400 million) and interest rates continued high. 18 Only a few days later, (September 17) the Minister of Finance visited Washington: and requested increased financing, and disbursements, to $700 million per year, with half of the 1999 amount in the first quarter (meeting with Regional Vice President). l9 Earlier in February 1999 a Colombian delegation visited Washington (to discuss emergency financing for the January earthquake) and learned that the FSAL would not go to the Board in March as implied by the Bank's December letter. This prompted an angry letter from the Minister of Finance to Mr. Wolfensohn on agreed terms noting that he was disappointed at this news, but that he hoped the impasse could be resolved soon. The letter was answered by Mr. Wolfensohn expressing his expectations of this operations going forward in the first-half of the year. financial sector.20FOGAFIN also announced details of the recapitalization program for private banks, and other measures for state banks, mortgage holders and cooperatives. 4.9 FSAP and furtherpreparation. Preparation of the FSAL took yet another turn in late April when the Bank and the Fund launched their first (pilot) FSAP in Colombia. In the meantime, the continuing deterioration of the financial sector's portfolio and downgrading of commercial banks and of Colombia's sovereign risk led the government to start discussions on the preparation of a program with the IMF. The FSAP work led to several missions between late April and early June 1999. Although these developments resulted in a further delay in the calendar for preparation of the FSAL, they also brought a much greater level of attention and of human and financial resources to the process. On the downside, however, this increased attention arrived after the government had had to manage by itself in addressing the more pressing aspects of the crisis." 4.10 Between June and September 1999 FOGAFIN was engaged in its new program for recapitalizing the banks, a public sector bank (Caja Agraria) was closed, and a Financial System Framework Law (Law 5 10, or Ley Organica) was approved. Law 5 10 formalized the expanded role of FOGAFIN that it had already performed, defacto, since early in the year. At the same time, however, the Central Bank launched yet another attempt to keep the exchange rate band by a discrete devaluation (with a further loss in reserves). Finally, in September, the Central Bank abandoned the exchange rate band and moved to a free float, while FOGAFIN announced a third round of relief measures for the mortgage sector." 4.1 1 Objectives and Components. The FSAL, for $500 million to be disbursed in two equal tranches, was approved on November 18 and became effective on December 7, 1999. The main objectives of the program supported by the FSAL, were to support: (i) required changes in financial sector legislation and regulations to facilitate strengthening of bank supervision and deposit insurance and to promote timely, fair and efficient resolution of banks facing financial solvency problems; (ii) to strengthen the institutional capacity of the Superintendency of banks and FOGAFIN; (iii) restructuring, recapitalization, privatization and, if necessary, liquidation of state owned and "officialized" banks under public sector management, and (iv) effective implementation of special measures to address the immediate impact on the financial system stemming from economic and other emergencies. The matrix of reform actions presented a slightly different list by adding the achievement of a satisfactory macro-economic framework and 20 FOGAFIN announced that the transactions tax was not available for private banks (because of the Constitutional Court decision) and that its total amount was expected to be insufficient even for others. The CAF had provided $100 million and more was expected, before June, from others including the WB. 21 Bank staff commented that the FSAL design benefited significantly from the FSAP assessment, which incorporated one of the earliest frameworks for modernized bank resolution methods, matching assets and liabilities from good and bad banks in lieu of fiscal subsidies for deposit insurance and liquidations, providing virtual, rather than real, loans to bank shareholders which were used to fund a Government bond asset, and other innovations, all intended to minimize fiscal outlays, which it successfully accomplished. However, this approach required legislative remedies, which were the primary reason for delays in loan effectiveness, rather than any slowness or lack of responsiveness by the Bank task team. 22 To comply with an August ruling of the Constitutional Court regarding the UPAC indexation, and in advance of a new Housing Law (that was passed on December 23). a separate mention of the reform of the cooperatives system. The resulting areas of reform in the matrix are: Ensuring a satisfactory macroeconomic framework Strengthening the legal, regulatory and institutional framework of the financial system and problem bank resolution. Restructuring and resolution of state-owned and "officialized" banks. Reform of the cooperatives system Recent special programs being implemented by FOGAFIN. 4.12 The matrix of reform actions also includes the actions completed prior to Board presentation, and the second tranche release conditions in each of the five areas above, (see Annex B). Implementation 4.13 With all conditions for first tranche release being also conditions for Board presentation, the first tranche was disbursed immediately after effectiveness (December 7). Shortly afterwards, in March 2000, two parallel FSAL supervision and FSAP follow-up missions found that "The most acute phase of the banking crisis has passed and conditions in theprivate banking system appear to be easing". Non-performing loans were down from their peak in November 1999, and interest rates were also down. The FSAP follow-up note raised its concern about the financial transactions tax, but this was not reflected in the FSAL supervision report. 4.14 By December 2000, the Bank found that all conditions for second tranche release had been met, and notified the Board (on February 2,2001) that it had informed the government of the availability of the second tranche. In early 2001, Colombia's economy was already on its way to recovery: GDP had grown by 3 percent in 2000 (following the decline of more than 4 percent in 1999), the fiscal deficit had been reduced because of higher oil revenues (but was a still high 3.6 percent for 2000), and the current account was roughly in balance. At the same time, however, the country's security situation was deteriorating because of guerrilla (and paramilitary) activity, foreign investment was down and external debt increased. 4.15 With elections about a year ahead, and fiscal deficits still high, the government decided to capitalize on its improved performance (and improved international financial markets) to return to the international capital markets and requested the Bank to convert the FSAL second tranche into a Policy Based Guarantee (PBG) to support its international borrowings. The Bank agreed and the government raised a total of $1 billion in two steps, in April and May 2001. The total amount of the PBG was about $160 million, and the remainder of the second tranche (about $62 million) was disbursed. The FSAL closed in December 2001 (but the PBG continued open and remains in effect today). Assessment and Outcomes 4.16 Quality at entry: The identification and preparation of the FSAL was marked by a systematic underestimation of the financial crisis and of the systemic risks to the financial sector. Bank support was insufficient (in quantity and quality) during a long period in 1998 and the first half of 1999, and internal decision-making was poor. The delays put an additional burden on the government and reduced the effectiveness of Bank advice. At the same time, it appears that the government also underestimated the risks and vulnerabilities during the early stages of the crisis. In addition, the timing and details of the government measures was highly limited by judicial involvement (by the Constitutional Court). The quality of Bank assistance improved after the (joint Bank- Fund) FSAP started in late April 1999. 4.17 By the time of FSAL approval, the peak of the crisis was over, most corrective policies were already in place (some for about a year), and others well under preparation. he prior actions (conditions of Board presentation) were based on the policies already implemented, which were adequate, and the conditions for second tranche release, focusing on continuations of the reforms were also adequate. Therefore, quality at entry is rated as satisfactory, in spite of the delays and mixed signals during preparation. 4.18 Relevance. Although the bulk of the credit for avoiding a systemic crisis goes to the government (with only a supporting role by the Bank, other IFIs and the IMF), the Bank's financial and technical support was useful, particularly in the later period of the crisis, and the FSAL's letter of development policy outlined a useful path for subsequent government reforms in the sector. The Bank's support for (and insistence on) privatization of public banks was found particularly useful by the reformers in the government, even if at times they complained about what they considered excessively ambitious timing proposed by the Bank. Thus, relevance of the FSAL is rated as substantial. 4.19 In the worse moments of the crisis (late 1998 to mid-1999), however, the govemment had to fend for itself. Unintentionally, the Bank's early view (in the IM) that the operation was not addressed to the crisis (paragraph 4.7 above) turned out to be accurate: there was a crisis at the time, but the FSAL support was provided only after the peak of the crisis was over, and was more useful to help prevent the possibility of later ones. 4.20 Outcomes: Assessment by Objective/Component. Achievement of the outcomes in the five areas of reform in the Matrix of Reform Actions is summarized below: Ensuring a satisfactory macroeconomic framework. By the time of Board approval, the government had reached agreement with the IMF on a program (three-year EFF, formally approved on December 20), and subsequently complied with all the program targets (fiscal, monetary and external sector),23 although with some difficulties in the later part of the period. In the words of the IMF review of the EFF in January 2002: "The momentum of Colombia's economic recovery slowed in 2001 in the context of a drop in oil exports related to guerrilla sabotage, a weakening in external demand as world growth slowed, and a fall in coffee prices to historically low levels. Nonetheless, the external position remained strong and Colombia maintained access to international financial markets". The IMF review in January 2002 also noted that the performance criteria for mid-2001 were met with the exception of the combined public sector deficit, and that criteria for September and December 2001 were not set because of the delays in completing the review.24 Although with the above limitations in the fiscal area, the overall outcome of this policy area is rated as satisfactory. Strengthening the legal, regulatory and institutionalframework of the financial system andproblem bank resolution. The main measures in this area were taken in late 1998 and throughout 1999, prior to FSAL approval. As a result of these measures, and particularly the bank recapitalization started by FOGAFIN in mid- 1999, a systemic crisis was avoided. By March 2000, an FSAL supervision mission (and FSAP follow-up mission) noted 'The most acute phase of the banking crisis has passed and conditions in the private banking system appear to be easing'. Non-performing loans were down from their peak in November 1999 (before FSAL approval), and interest rates were down from their peak in early 1998. The FSAP note (but not the FSAL supervision report) mentions concerns about the financial transactions tax. The conditions for second tranche release were formalizing the mechanisms already established in 1999, and strengthening the two main institutions implementing them: the Superintendency of Banks and FOGAFIN. These measures were taken and remained in place. The outcome of this component, mainly achieved by the prior actions in 1999 is rated as satisfactory. Restructuring and resolution of state-owned and "ofJicialized" banks. At the time of Board approval the government had already closed Caja Agraria, and during implementation, three more (BCH, BanEstado and Uconal) stopped operating, while some of their bad assets were transferred to a newly established asset resolution company (CISA). Another institution, FES, had been intervened in 1999 and was liquidated. The two remaining public banks (Granahorrar and Banco Cafetero) were prepared for privatization, but the process was not -- - - 23 The main economic indicators for the period are summarized below: 1998 ~~~~ GDP Growth (%) 0.6 -4.2 2.9 1.4 1.6 Public Sector Deficit (% of GDP) 4.6 6.5 3.5 3.5 4.2 Inflation (CPI) 16.7 9.2 8.7 7.6 7.0 Interest rate (90 day deposits) 34.1 15.8 13.4 11.5 7.7 International Reserves (months of imports) 6.6 6.7 6.8 8.0 7.8 Total External Debt (% of GDP) 36.3 42.5 46.1 47.5 52.3 24 In spite of this poor assessment by the IMF, the Bank approved a large ($400 million) Structural Fiscal Adjustment Loan (SFAL) in December 2001. completed. The second tranche conditions, however, were only to "bring to the point of sale" or to "initiate liquidation": therefore, the formal objectives were exceeded and the output of this component is rated highly satisfactory in spite of the delays regarding Granahorrar and Banco Cafetero. Reform of the cooperatives system. The component was based on actions already taken by the government prior to FSAL approval. The second tranche condition was "continued satisfactory implementation of the action plans and agreed resolution process for the financial cooperatives". Support for this component was also provided by the FMDTA project. Of the more than 6000 cooperatives identified in a census, only 54 were classified as "financial cooperatives" and their supervision was transferred to the Superintendency of Banks. The remainder was left under SES (created in June 1998). Of the 54,22 became "closed" cooperatives and returned to SES supervision, and 5 became "social" cooperatives under another agency, 11 were liquidated, 3 merged with other institutions and 13 remained under the SB. The outcome of this component is rated satisfactory. Help effective implementation of special FOGAFINprograms. From late 1998 to late 1999, FOGAFIN had transformed from a small deposit insurance institution into a large and powerful agency with multiple areas of activity in addressing the financial crisis. This had happened well before the formulation of the details of the FSAL components. The programs designed by FOGAFIN had evolved as the size of the crisis was revised upwards, and its interventions were sanctioned or curtailed by the Constitutional Court. In addition, it had also been given a role in providing support to the victims of the January earthquake. The component was based on establishing a clear difference between its regular and crisis-related programs, and between its own funds and those handled on behalf of the government (these objectives were achieved already prior to approval). Second tranche conditions were continued compliance with this distinction, and provision of additional resources by government. These conditions were fully met as the accounting was kept separate, and the government provided additional resources to FOGAFIN ($197 million in addition to the deposit insurance premiums collected by FOGAFIN). The outcome of this component is rated as satisfactory. 4.21 Most of the resources of the FSAL second tranche were used to fund a PBG in March and May 2001. This assessment does not cover the outcomes (nor Bank and Borrower performance) under the PBG, as the PBG has not yet closed and remains active today." Based on the all of the above, the outcome of FSAL is rated satisfactory. 4.22 The Institutional development impact of the FSAL is rated as substantial, although the question of attribution is even more difficult to assess than is normally the case. In any event, the institutional development impact of the program implemented with FSAL support was clearly substantial, particularly regarding the changes in the public 25 The government, however, is re-purchasing the bonds issued under the guarantee, prior to their due date, in order to reduce the outstanding amount of the PBG and shorten its term. sector component of the banking sector. Institutional development in SB and FOGAFIN was also substantial. 4.23 Sustainability of the reforms is rated as likely. Four years after the closing of FSAL, all of the reforms remain in place and further progress has been made in many areas, with the support of two additional adjustment operations (PFSAL I and 11). 4.24 Assessment of Bank performance raises some difficult issues. In terms of timeliness of response, the Bank gets a very low rating. During the early stages of preparation, the Bank underestimated the extent of the crisis (as did the government). The Bank also gave conflicting signals in response to the early and increasingly urgent government requests for financial support. On financing, as well as on technical support, the government had to manage the crisis largely by itself for nearly a year. The several letters from the Minister of Finance to the Bank raising these concerns are an indicator of this lack of support. For example, the letter from Minister of Finance to Mr.Wolfensohn, in late 1998 to reiterate a request made to Regional Vice President notes his concern that not enough human and financial resources were available. As noted earlier, this letter was answered by the Country Director two months later, saying that the Bank expected a new country assistance strategy incorporating the new loan to be endorsed in the first quarter of 1999. In the event, the FSAL and the country assistance strategy progress report went to the Board in November 1999.26 4.25 The Bank's early underestimation of the extent of the crisis is clear from the Annual Meetings Brief of October 2, 1998, which does not mention the financial sector among the economic issues disc~ssed.~' Even in late February, the initiating memorandum for the FSAL argues that there is no liquidity crisis in Colombia, and that the proposed operation is precautionary. At about the same time the Minister wrote another angry letter to Mr. Wolfensohn asking for faster processing of the operation. - 26 Bank staff argued that IEG had misinterpreted the meaning of these exchanges of correspondence which indicated not government impatience with the Bank's slowness in getting the FSAL approved, but rather that these letters were meant to accelerate an internal Bank decision to significantly alter the resource allocation pattern approved so recently in the mid-1990's country assistance strategy so as to more effectively address the unanticipated financial sector crisis. A more important source of delay resided in the fact that the FSAL was never intended to provide solely emergency finance, but rather quick-disbursing assistance within a supporting framework. This framework involved passage of extensive legislation to implement state-of-the-art bank resolution measures, which took some time. Nonetheless, the delivery of the loan hnds was timely in that it provided liquidity to both Government and FOGAFIN, the deposit insurance and resolution agency, to meet depositor claims and fund bank capitalization and resolution. 27 Bank staff commented that the omission of any reference to the condition of the financial sector in the Meetings Brief was a deliberate concession to the Government, which wanted to keep information about the deteriorating situation secret during 1998, so as to avoid capital flight and further exchange rate deterioration. However, the Pastrana Administration publicly discussed the banking crisis as early as May 1998, immediately after the elections and five months before the Annual Meetings. It also closed several coops and public financial institutions during the first few months of its rule, and issued a comprehensive package dealing with the crisis in November 1998, only one month after the Annual Meetings. 4.26 The relations with the IMF were also a source of confusion as the Bank's initial attitude was that no IMF program was needed. When the government realized that, in fact, they needed one, this added to the length of preparation. 4.27 It should be noted, however, that these were difficult times for all concerned: the Bank was trying to address crisis situations in multiple countries, for which it was not prepared, the relations with the IMF were strained because of different views regarding the adequate response to the crises and the contagion and the internal arrangements in the Bank created strains between the Regions and the newly created Special Financial Operations Unit, which had autonomy from the regional managements regarding where to direct its financial and technical resources. In addition, differing views within the Region, with the sector unit arguing for early support to the sector and the country unit initially reluctant to depart so drastically from the program presented in the country assistance strategy of a year earlier, contributed to the initial delays and confusion. The result was that the Bank was unable to offer adequate and consistent support and advice until late April 1999. After that time, the Bank was able to provide increasing technical support, particularly regarding regulatory reforms, institutional strengthening of the main agencies involved, and focusing attention to the medium-term issues. 4.28 As a result, Bank performance during preparation was unsatisfactory, but it became satisfactory at a later stage and continued so during supervision, at times jointly with FSAP updates. Overall, Bank performance is rated as moderately unsatisfact~ry.~~ 4.29 Borrower performance was satisfactory, overall and in terms of compliance with FSAL conditions. In at least the area of addressing the issues of the public sector banks it went beyond the agreed plans and was highly satisfactory. 28 Bank staff disagreed with this rating downgrade of Bank performance, which it considered h l l y satisfactory. 5. The Programmatic Financial Sector Adjustment Loans I and I1 Introduction 5.1 The closing of the FSAL, in December 2001, coincided with a period of renewed political and economic difficulties in Colombia, and a negative impact on the still recovering financial sector. At the political level, peace negotiations with the guerrilla groups had broken down in February 2002, followed by escalating violence. 5.2 On the economic front, the IMF summarized the conditions: "The economic situation turned adverse in the first half of 2002. Activity remained sluggish as the civil conflict intensified; consumption remained subdued by persistently high private sector debt levels and uncertainty associated with the continued high unemployment levels. At the same time, fiscal consolidation went significantly off track. As the year progressed, the peso came under pressure and sovereign bond spreads widened sharply, reflecting increased contagion from neighboring countries, rising market concerns about fiscal solvency and political events".29 5.3 This situation was cause for concern because it echoed exactly the conditions prevailing four years earlier: it could have been written in mid-1998. There were two major differences, however: one was that the levels of the relevant economic indicators were considerably better in 2002 than in 1998, even if they were worsening. More importantly, the policy and institutional environment was better, after the reforms of 1999. At the macro level, the exchange rate was floating and thus had cut the earlier damaging link to interest rates; in the financial sector there was a stronger legal, regulatory and institutional framework in place, and many of the weaker institutions had already disappeared, merged or contracted substantially, although there were still problems remaining with some public banks. 5.4 After taking office in August 2002, the new government took steps on several fronts, including the internal conflict and security situation and the fiscal and macroeconomic performance, and negotiated a two-year Stand-by agreement with the IMF (agreed in January 2003, at the end of the previous EFF).30 5.5 The Bank's Program. The Bank's response to the new economic and political challenges faced by Colombia in 2002 was articulated in the 2002 country assistance strategy (covering the period of the 2002-2006 government). The 2002 country assistance strategy proposed a much larger lending program, of about $1 billion per year (twice the level of the previous country assistance strategy), about half through fast disbursing adjustment operations. These consisted of four groups of Programmatic operations 29 "Colombia: 2002 Article ZV Consultation and Request for Stand-by Agreement"; IMF Country Report No. 3/19 of January 2003, page 5. 30 Colombia did not request any drawings from the 1999-2002 EFF, which turned out, expost, to be precautionary. The Stand-by (2003-04) was already explicitly discussed as precautionary during preparation. covering fiscal, financial, social and environmental areas. The financial sector would receive two of the ten programmatic operations, amounting to 15 percent of the total, while the bulk of adjustment lending focused on fiscal adjustment. 5.6 The 2002 country assistance strategy included a detailed discussion of the financial sector issues and objectives, including: (i) building on the financial sector reforms of 1999-01 (improving public debt management, governance and regulatory issues, consolidating state controlled second tier banking entities, and improving regulation for bank resolution and bad asset disposition); (ii) addressing the continuing problems in the housing and rural finance areas; (iii) completing the reform of the pension system; and (iv) supporting a deeper, more liquid capital market.3' 5.7 The country assistance strategy listed eleven specific objectives for the financial sector: (i) restructuring andlor privatization of public banks (particularly Banco Cafetero and Granahorrar); (ii) institutional strengthening of CISA (dealing with management and disposition of repossessed assets); (iii) increasing the effectiveness and efficiency of second-tier public development banks;32(iv) capital market development (capital markets law, regulations, institutional strengthening, new instruments); (v) strengthening/ restructuring existing financial groups; (vi) reforming the pension and social security system; (vii) improving the framework for problem bank resolution and strengthening risk management; (viii) reforming judicial system procedures affecting the financial sector (including foreclosure procedures); (ix) improving solvency standards and improving supervision of the insurance sector; (x) developing sustainable financial services to the rural areas; and (xi) strengthening the framework for combating money laundering and the financing of terrorism. The otherwise comprehensive program did not mention elimination of the financial transactions tax, introduced as a one-year emergency measure during the 1999 crisis and still in operation today.33 The strategy's objectives for the financial sector provide the context to assess the relevance and effectiveness of the two Programmatic FSALs approved in FY03 and FY04. 5.8 The PFSALs. The country assistance strategy objectives in the financial sector were to be achieved through two programmatic single-tranche operations of $150 million each, in the first two years of the country assistance strategy (FY03 and FY04). Given the substantial amount of unfinished business in the financial sector at the time of the 2002 strategy (the eleven areas listed above), addressing them through two closely-spaced single tranche operations was an ambitious approach. A longer program, covering at least five years, and spacing the proposed FSALs over a longer period would have been more realistic. 31 This had already been the main objective of the FMDTA project, approved five years earlier. 32IFI, FINDETER, FEN, FINAGRO and BANCOLDEX. 33 The distortions caused by the tax were already noted in 2000 by the FSAP follow up missions. The tax rate today is twice as high as when it was introduced. PFSAL I: Preparation, Design and Implementation 5.9 Preparation and Design. Preparation of PFSAL I took place in late 2002 and early 2003, at a time still marked by serious concerns about Colombia's fiscal and external conditions, following the poor fiscal and economic performance in 2002. At the time of appraisal, the new government had been in office for only a few months, and concluded negotiations with the IMF on a new two-year Stand By (in January 2003). The Bank's work on the financial sector, influenced by the experience of the 1998-99 crisis and its aftermath, recognized that there was significant unfinished business, not only in the areas of legislation, regulation and institutional strengthening, but also in terms of distressed portfolios and institutions, publicly owned and "officialized" banks and in the housing finance institutions. 5.10 The PFSAL I objective was defined as "to complete the clean up of the banking system as a result of the 1999-2001 crisis and to strengthen the government's capacity to manage and mitigate weaknesses in the financial system. At the same time the objective is to prevent crisis contagion effects from neighboring countries or other exogenous events, by utilizing strong regulatory incentives for prudential management, market risk- sharing mechanisms and optimal supervisory powers to expeditiously contain any emerging risks in the financial system."34 Other objectives included improvements in the housing mortgage market, and in non-bank financial services and securities institutions, and to strengthen the legislative framework and capabilities of supervisory institutions 5.1 1 As expected in the first phase of a programmatic approach, PFSAL I includes a description of the different areas to be addressed by the full program, including the reforms to be supported under each of the two operations. These areas are: Banking system: strengthen corporate governance (under the new financial system law 795, approved in January 2003), improve enforcement powers of the SB, regulatory infrastructure and tracking of sensitive funds (AMLICTF), improve mechanisms for resolution of problem banks, restructuring and resolution of State Owned Banks, prudential regulation and supervision, and institutional development of the SB. Housingfinance: improve mechanisms for mortgage foreclosure, management of recovered assets, bank privatization or closure, encouraging micro-finance, and improving FOGAFIN7ssupport for hedging and securitization (through FRECH); a broader set of improvements were expected for the second operation. Insurance sector and non-bankfinancial services: capital requirements, solvency standards, reserves, investment norms and enforcement regime for insurance sector, improve SB supervision of the trust industry; regulations on most of the above areas, and for the private pension funds, were expected for the second operation. Capital markets: establishing asset valuation standards, joint supervision by SB and SV of financial groups, regulations on capital requirements and conflict of interests, and institutional strengthening of the SV; under the second operation, a 34 PFSAL I Program Document, paragraph 142. new Securities Law and further regulatory and institutional improvements were planned, together with improved market infrastructure (e.g., electronic trading); the securities law was expected to cover also investment banking. Government Debt and money markets: increase the liquidity of Treasury securities, improve regulation of brokersldealers, establish a valuation framework; under the second phase further actions were expected on debt market development, strengthening of the money market and improving risk measurement and management by market participants. 5.12 The "Policy Matrix for Programmatic Reform Program" (Annex 1 of the Program Document), lists the prior actions taken (before PFSAL I approval) for each of the areas above, adding the maintenance of a sound macroeconomic policy framework, the future actions to be taken prior to PFSAL I1 approval, as well as the key expected results at the end of the program. 5.13 The list of areas indicate the mixed character of the program: a combination of "mopping up" problems remaining from the previous crisis (mainly housing finance and the public banks), plus legislative, regulatory and institutional development activities intended to develop a deeper and more sophisticated financial sector, many of them centered around the implementation of the "Ley de Reforma Financiera" (Financial Reform Law, Law 795 of January 2003). The objective of hrther deepening and development of the financial sector is also related to the objectives set a full decade earlier, in 1993, when preparation for the FMDTA project started. 5.14 Implementation and Outcomes. PFSAL I, for $150 million, was approved in April 2003, declared effective (and disbursed) in June 2003, and closed also in the same month. As a single tranche operation with all conditions met at approval, implementation of FSAL I is better seen in terms of the initial outcomes and related performance of the sector.35Economic and financial sector performance were positively affected by a macroeconomic environment that turned out more favorable than anticipated at appraisal (see tables 5.1 and 5.2 below). GDP growth in 2003 was close to 4 percent, well above the 2.5 percent projection in the project document (and another 4 percent in 2004). The fiscal deficit was within the expected range in 2003 (about 3 percent of GDP) and a much lower 1.7 percent in 2004 and inflation continued to decline. The external sector was strong, with a current account deficit of 1.5 percent of GDP in 2003 and 1 percent in 2004, and large capital inflows that led to an appreciation of the peso and improved terms for external government financing. At the policy level, the main accomplishment was the approval of Law 795, passed prior to approval of PFSAL I. 35Although PFSAL I formally closed in March 2003 and the ICR was issued in December 2003, this discussion of implementation covers the period until approval of PFSAL I1 in September 2004. Table 5.1: Main Economic Indicators 2002 2003 2004 2005 (projected GDP Growth (%) 1.6 3.8 4.0 4.0 Public Sector Deficit (% of GDP) 4.2 3.2 1.7 2.5 Inflation (CPI) 7.0 6.5 5.5 5.0 Interest rates (90 day deposits) 7.7 7.9 7.7 -- International Reserves (months of imports) 7.8 6.6 7.2 6.5 External Debt (% of GDP) 52.3 46.0 37.1 35.9 Source: Internal World Bank data. 5.15 The improved economic performance translated into a better performance of the banking sector and facilitated achievement of the objectives of this component. Banking sector profitability increased in 2003 and the level of arrears declined noticeably (though it remained high) with substantial differences among different categories of institutions: the housing banks and publicly owned banks continued to perform worse than the rest of the sector. Privatization of the public banks was delayed: there were no bidders for BANCAFE, and there were delays in the preparation (to point of sale) of GRANAHORRAR.36 More substantial improvements took place on the restructuring and privatization of second tier banks ( c l o s i n g b f ~ ~ part of its activities merged with with BANCOLDEX). Table 5.2: Banking Sector Indicators 2000 2001 2002 2003 2004 Return on Assets -2.3 0.1 1.1 1.9 2.7 Return on Equity -20.7 1.1 9.6 17.1 23.0 Capital Adequacy Ration 13.2 13.3 12.5 13.1 13.8 Past due1Total Loans 11.0 9.7 8.7 6.8 3.3 Gross LoansIGDP 27.0 25.3 24.6 23.1 24.5 Source: Internal World Bank data. 5.16 A review of the improvements in the financial sector carried out after the closing of PFSAL I and before approval of PFSAL 113' noted that there remained vulnerabilities in the banking sector, particularly regarding the mortgage banks (poor credit quality, interest rate mismatches), excessive concentration of assets on public sector debt (crowding out, income volatility risks), and need for improved &eillance, for coordination among supervisory agencies, and for consolidated supervision of financial groups. Some of these issues were taken up under PFSAL 11. 5.17 Improvements in the housing finance area fell short of expectations. New procedures were put in place to facilitate foreclosure, but in terms of results on the ground there was little progress: as noted above, the health of the housing finance sub- 36 Granahorrar has been recently sold (October 3 1,2005) after the closing of PFSAL 11. 37 "The Recovery of Colombia's Financial System and Remaining Vulnerabilities", WB and Ministry of Finance (Colombia), July 2004. sector remained precarious and the expected increase in mortgage lending, particularly to low income segments of the population, failed to materiali~e.~' 5.18 Achievements in the non-bank financial institutions, capital markets and debt and money markets were, at best, moderate. New regulations were instituted, SB and SV signed an MOU to cany out joint supervision and surveillance, and the debt management unit worked on the construction of a yield curve, but it is difficult to identi@ actual improvements on the ground. For example, the size and variety of the capital market experienced little or no impr~vernent,~~liquidity of government securities suffered from the market turbulence in late 2002 and did not improve substantially until 2004, and the insurance industry had only moderate improvements in terms of capitalization and activity. 5.19 Overall, the PFSAL I period saw major improvements in economic performance and the recovery of the banking sector, but improvements in the broader financial sector were of a more subtle, institutional development type.40 Ratings for PFSAL I are discussed later, in parallel to those of PFSAL 11. PFSAL 11: Preparation, Design and Implementation. 5.20 Preparation and design. Preparation of PFSAL I1 took place in a very different situation than the one prevailing when PFSAL I, and the overall program, were prepared. As a Bank documentpoints out, GDP growth continued to be strong in 2004, and capital inflows in the first two months of 2004 were some $600 million (compared with an outflow of $80 million during the same period of 2003). The strong capital inflows had led to an appreciation of the peso (and accumulation of reserves) in spite of efforts by the Central Bank to limit the appreciation (and growing monetary base) by lowering interest rates. In this environment, and with the consolidated fiscal deficit also continuing to decline (after dropping to 2.7 percent in 2003 it fell further to 1.3 percent in 2004)41the choice of another quick disbursing operation to support medium term regulatory reform and institutional development in the financial sector was driven more by the commitment to the programmatic approach than by macroeconomic conditions. A small TA operation -- 38 The decline in mortgage lending continued also during the PFSAL I1 period. The FSAP Update acknowledged that the Colombian mortgage portfolio had shrunk from what it was before the crisis at percent GDP to about half. On the positive side, the ICR for PFSAL I1 reports that housing finance started to slowly expand again in 2005. 39 The FSAP Update internal memorandum found private capital markets slow, due to scarcity of local investment securities and weaknesses in clearance and settlement systems The ICR for PFSAL I1 shows a continuous decline in the number of new issues of equities and private sector bonds between 2000 (21 issues in total) and 2004 (1 1 issues). 40 This raises the question as to whether these aspects of the PFSALs could not have been better attended to through a TA or similar project more specifically focused on institutional and capacity building (see later discussion in the Assessment section). 41 The projection in the program document, however, was for a higher deficit, of 2.4 percent of GDP. It should also be noted that the large reduction in the consolidated public sector deficit was largely as a result of surpluses in the sub-national entities; the central government deficit remained around 5 percent. (of the type that had been approved just before the 1998 crisis) would have been at least as appropriate. 5.21 In spite of the changed country conditions, the momentum of the Programmatic approach and the country assistance strategy lending targets led to the preparation and approval of PFSAL I1 in 2004,42 with only two small adjustments: the loan size was reduced to $100 million (instead of the $150 million planned in the country assistance strategy and mentioned in the Program Document of PFSAL I) and the approval date was moved back by a few months (to September 2004), thus making it an FY05 operation instead of FY04 as previously planned. The PFSAL I1 became effective, and was hlly disbursed, in October 2004. 5.22 Design of PFSAL I1 followed closely the program outlined under PFSAL I, covering the same areas of: (i) banking regulation and resolution (regulatory framework, supervision of consolidated financial groups, supervision, risk management and sanctions, assessment of remaining vulnerabilities and further liquidation of state-owned banks); (ii) housingfinance (micro-housing credit, swap options to mortgage banks from risks emanating from Constitutional court decisions, and mortgage securitization); (iii) insurance, pension and Trust industries (regulatory framework including risk analysis and investments diversification); (iv) capital markets (Securities law, mutual funds regulation, valuation methodology), and (v) government debt and money markets (to improve the functioning of the market for the very large volume of treasury bills). Once again, in this list of activities, legislation and regulations already issued, or acted upon, by the government, there is no discussion of the need for, or possibility of, elimination of the financial transactions tax. 5.23 Implementation and outcomes. PFSAL 11, for $100 million, was approved on September 23, 2004 (somewhat later than anticipated). It was declared effective (and disbursed) in October 2004 and closed on March 3 1,2005. As in the case of the PFSAL I, PFSAL I1 was a single tranche operation with all conditions met at approval. Thus implementation of PFSAL I1 can better be seen in terms of the initial outcomes and related performance of the sector. The PFSAL I1 program document included a set of "quantifiable results" and sector performance indicators at the time of approval. The list, however, lumped a disparate set of program outputs (e.g., divestment of banking assets by the government, FOGAFIN inflation hedge for mortgage borrowers), issuance of new regulations (e.g., in the insurance sector), and sector performance indicators (e.g., banking sector profitability, non-performing assets of the banking sector, volume of treasury bills issued). For these performance indicators, the question of attribution is important: for example, the health of the banking sector is also highly correlated with the improved economic situation discussed above. 5.24 A more relevant assessment of the outputs and outcomes of the program, and of each of the two PFSALs, is presented in the ICR for PFSAL I1 (of 2005). The conclusions of this very good ICR can be summarized as follows: 42 Another, larger quick disbursement loan was also approved two months later: The Programmatic Labor Adjustment Loan, for $200 million, approved in November 2004. The overall outcome of the program was favorable, although the program includes many medium term goals that have not yet been fully realized; The banking sector performance continued to improve through 2004, indicating that recovery from the crisis is on a sustainable path; the recovery is largely due to the resumption of growth, lower interest rates (but stable margins), the exit or restructuring of non-performing assets in previous years, and the bank recapitalization program; The track record on bad loan recovery, managing weak banks and downsizing and privatizing state-owned banks43was good and the fiscal liabilities were reduced, but some vulnerabilities remain; The health of the housing finance portfolio increased but origination of new mortgages remains limited; interest rates on mortgages (including for social priority housing) have started to decline, and housing banks have started to diversify their assets; The outcome in capital markets has been mixed to date, even after the new Securities Law was approved in July 2005 (after closing of PFSAL II);44 The insurance industry shows only a marginally improving trend in terms of capitalization and level of activity; and The domestic debt market has benefited greatly from the improved fiscal policy, a more favorable macroeconomic environment and better institutional arrangements. 5.25 Another relevant assessment of the financial sector after the completion of PFSAL I1 is presented in the final report of the Joint Bank-IMF FSAP Update of 2005.45 The update concludes: "Prudent macroeconomic policies, structural reforms, and a favorable external environment have permitted the economy to recover gradually but steadily from the 1999 crisis", but it also notes a number of remaining vulnerabilities that casts doubt on the appropriateness of the Bank's approach, in the 2002 country assistance strategy, based on the two one tranche PFSALs over a two year period when many of the vulnerabilities may persist for a longer time. 5.26 The main issues highlighted in the FSAP update include: (i) the replacement of private sector loans with government bonds in bank balance sheets, which has altered the nature of the risks confronting the banking sector and represents a lingering legacy of the crisis; (ii) remaining weaknesses in the legal and regulatory framework (including lack of sufficient autonomy by the Superintendency of Banks); (iii) a corporate insolvency regime introduced during the 1999 crisis (Law 550) that has not yet been replaced by new instruments more favorable to financial intermediation; (iv) the distortionary impact of the financial transactions tax; (v) lack of recovery of mortgage lending; (vi) weaknesses 43 One of the remaining two public banks slated for privatization (Granahorrar) was eventually privatized after PFSAL I1 closure, on October 31, 2005. Privatization of the other (Bancafe-now GranBanco) failed twice because of lack of bidders. The government plans to offer it again for sale, after some fk-ther restructuring in 2007. 44 The ICR notes, correctly, that "The new Securities law, which was recently approved, will likely not have a measurable effect on market activiq in the short run, although it is expected to eflectively reduce some of the existing barriers to market entry over the medium-term "(ICR, paragraph 57). 45 IMF Country Report No. 051287 of August 2005. in the new risk-based approach to bank provisioning (potential underestimation of provisioning and capital requirements is still of concern); (vii) remaining issues with the pension system; and (viii) low liquidity and insufficient depth of the capital markets.46 5.27 While the conclusions of the FSAP Update suggest that the remaining agenda is still considerable, the country assistance strategy progress report of September 2005 is silent about financial sector issues. The "Recent Economic Development and Prospects" section has no mention of the financial sector, and the "Political Developments" section has a passing reference to recent legislative activity (approval of the Securities Law).47 Assessment of the Program 5.28 Quality at entry. The re-emergence of a potential crisis situation at the time of the preparation of the program (second half of 2002) justified the decision to support further reforms through programmatic adjustment lending. Yet, the need to address a combination of "mopping up" issues from the earlier crisis, other short term measures for crisis prevention, and longer term institutional development objectives would have been better served with a longer program (either including more than two PFSALs or spacing them over a longer period of time-e.g., five years). Bank support would have also benefited from a parallel TA operation andlor a combination of analytical work and institutional development assistance to continue the work of the FSAP, and to support the medium term reforms in a more sustained manner.48 5.29 In addition, the very ambitious program for the financial sector put forward in the 2002 country assistance strategy was constrained because of the proposed timing of the two operations, and the program failed to address the distortions caused by the financial transactions tax. Overall, however, the quality at entry for PFSAL I is rated as satisfactory. For PFSAL 11 the design was adapted to benefit fiom updated analysis of the sector and quality at entry is also rated as satisfactory. 5.30 Relevance. As noted above, the conditions at the time of preparation of PFSAL I, and the analysis in the 2002 country assistance strategy made this program an important component of the Bank's assistance to Colombia. Thus, relevance of the PFSAL I is rated as substantial. By the time of the preparation of PFSAL 11, however, the changed conditions of 2004 required a reassessment of the program, at a time when high capital inflows were creating problems of appreciation of the peso. A change in the program, by delaying the PFSAL 11, converting it into several smaller operations, more spaced in time, and/or adding a TA-type project to continue the technical support for the medium term reforms in the sector would have been much more appropriate. Thus relevance of PFSAL 11 is rated as modest. 46 The FSAP update also includes a detailed set of recommendations to deal with the pending issues. 47 The section on IFC activities in the CAS PR has some references to the sector as context for the specific IFC interventions. 48 Other possible options to support institutional development in the financial sector include grants from the Institutional Development Fund (IDF), and from the FIRST program, intended to follow up on the work of the FSAP. 5.3 1 Outcome. Outcome of the different components has been summarized above, showing a similar pattern across components and for the two operations: it was satisfactory for the banking sector component in both operations, including their regulatory framework and supervision, overall improvement and liquidation and privatization of public banks (although with delays); only moderately satisfactory for the housing finance sector, particularly for PFSAL 11, because, although the performance of the sector improved, mortgage financing continued to decrease and the expected increase in mortgage lending for low income housing, (an important expected outcome of PFSAL 11) did not materialize, satisfactory for the other financial intermediaries and the domestic debt market, and only moderately satisfactory for the capital markets component (as new issues continued to decline during the PFSAL I1 period). 5.32 Although the pattern of outcomes of the different components is similar under each of the two operations, the achievements were higher under PFSAL I than PFSAL 11: for example, major improvements in the banking sector took place already under PFSAL I and the expected improvements in micro-housing credit under PFSAL I1 were not achieved. In addition, as noted above, the relevance of PFSAL I was higher than that of PFSAL 11. Because of the above, the outcome of PFSAL I is rated as satisfactory and the outcome for PFSAL 11 is rated as moderately satisfactory. 5.33 Institutional development is rated substantial for both, PFSAL I and PFSAL 11, and sustainability of the achievements under the program is rated likely for the two operations. 5.34 Bankperformance for both PFSAL I, and PFSAL 11, is rated as satisfactory. Bank performance during preparation of PFSAL 11, however, failed to address more substantively the changed country conditions at the time. 5.35 Borrower performance was satisfactory during PFSAL I and PFSAL 11. There were delays in implementation in at least two areas (approval of the Securities law and privatization of public banks), but the delays were largely due to delays in the parliamentary discussion and approval in one case and to lack of market demand in the other. 6. Lessons 6.1 The experience with the four operations reviewed in this PPAR shows a remarkable turnaround in both the government's approach, policy response, and regulatory and institutional framework in the financial sector, and in the Bank's analytical knowledge of the sector. Mainly thanks to the early and effective government response, a systemic financial crisis was avoided and the financial sector today, although still smaller in number of institutions and total assets than before the crisis, is much stronger and well regulated. In addition, the Colombian economy has recovered and the economic policies of the government have managed to put an end to the persistent double-digit inflation that was a characteristic of the Colombian economy for three decades. The success in lowering inflation has eliminated an important vulnerability of the financial sector. Unfortunately, however, the other major macroeconomic source of vulnerability, i.e., the fiscal deficit, remains a problem that needs to be addressed. 6.2 On the Bank's side, the progress in its analytical understanding of the sector, and in its contribution to addressing its problems has also been substantial. Both analytical understanding and support to addressing the problems were low at the onset and during the first year of the crisis but improved considerably after mid-1999. In the most recent years, however, the narrow definition of the programmatic approach adopted in 2002 has reduced the opportunities for continuing support to the sector (as well as reducing the relevance of PFSAL 11). In addition to the many specific lessons derived from the operations, and from the analytical work during the period (including under the FSAP), five more general lessons emerge from this review: Accurate, up-to-date sector information and analytical knowledge is critical for an adequate support to theJinancial sector: an outdated assessment of the sector by 1997 made the FMDTA irrelevant and blindsided the Bank to the vulnerabilities of the financial sector. It also slowed down the Bank's response after the onset of the crisis. Agility and quick reaction to crises requires adequate technical resources and streamlined decision-making: the Bank compounded its inadequate prior knowledge of the sector in 1998 by allocating insufficient technical/human resources until the joint FSAP was launched, and by slow and bureaucratic decision-making in the face of urgent client demands. Agility, flexibility, and openness to change are also crucial during implementation of a project orprogram: during implementation of FMDTA, the inadequate starting point could have been partly remedied by earlier, and deeper restructuring to face the new conditions on the ground; during implementation of the FSALs, improved economic conditions and slow progress in some areas (housing finance, capital markets) could have been better addressed (even in the absence of a TA project) through a change in the second phase of the program (e.g. by splitting PFSAL I1 into two smaller, more spaced operations implemented over a longer period of time). Exploiting the synergies among different Bank instruments is crucial: technical assistance operations and adjustment lending can reinforce each other when they are part of a coordinated sector intervention (and based on adequate analytical work). In Colombia, an outdated TA project was only belatedly and marginally linked to the FSAL; during the short period of the PFSALs (and after them) there was no TA vehicle on the ground to continue the support. A programmatic approach tojinancial sector reforms i s likely to require a time horizon of several years to allowfor policy and institutional reforms to be translated into outcomes and results: the two-phase single year approach of the PFSALs was too short to address the complex second- and third-generation issues facing Colombia's financial sector. Annex A. Basic Data Sheet Financial Markets Development Technical Assistance (TF-224 79; SCL-41960) Key Project Data (amounts in US$ million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Original commitment 15.00 8.19 54.6 Total cancellation -- 6.81 -- Total project cost 20 11.7 58.5 Cancellation -- 6.81 - Proiect Dates Original Actual Departure of Appraisal Mission 11/15/1996 1111511996 Board approval 0611911997 0611911997 Signing 07/29/1997 07/29/1997 Effectiveness 06/30/1996 11/21/1997 Closing date 02/28/2002 12/31/2003 Staff Inputs (staff weeks) Actual/Latest Estimate P f ' Staff weeks US$US$(OOO) IdentificationI Preparation 45.3 124.0 Appraisal / Negotiations 45.2 88.6 Supervision 72.6 188.0 ICR 4.5 19.2 Total 167.6 419.8 Mission Data Performance Date No. of Staff days Specializations Ratings (month/year) persons in field represented Rating Types of trend ~roblems Identification1 411994 2 -- Task Manager (1); Preparation Capital Markets Specialist (1) Task Manager (1); Banking Consultant (1); Capital Markets Specialist (1) Task Manager (1); Financial Specialist (1); Procurement Specialist (1); Bank Supervision Specialist (1) Appraisal 1 511997 5 Task Manager (1); Legal Negotiations (1); Financial Specialist (1); DisbursementOfficer (1); Procurement Specialist (1) Supervision 0611911998 2 -- Task Manager (1); Consultant (1) 0411711999 1 -- Financial Sector (1) 0211012000 4 -- Task Manager (1); Banking (I); Financial Sector (1); Operations (1) 0611012000 2 -- Task Manager1 Financial Sector (1); Financial Sector (1) 0811012000 1 -- Task Manager1 Financial Sector (1) 04/06/2001 2 -- Task Manager1Financial Sector (1); Financial Sector (1) 0112512002 2 -- Task Manager1Financial Sector (1); Banking Specialist (1) 1211012003 2 -- Financial Sector (1) ; Capital Markets (1) Completion 05/02/2003 1 - Task Manager1Financial Sector (1) 12/15/2003 1 - Task Manager1Financial Sector (1); Banking Consultant Financial Sector Adjustment Loan (FSL T- 70000) Kev Proiect Data famounts in US$ million) % of estimate current estimate appraisal estimate Original commitment 500 500 100 Total cancellation -- -- -- Total project cost -- -- -- Cancellation -- -- -- Project Dates Oriainal Actual Departure of Appraisal Mission Board approval Signing Effectiveness Closing date Staff Inputs (staff weeks) Actual/Latest Estimate Staff weeks USSUSS('OO0J IdentificationI Preparation 17.0 Appraisal INegotiations 16.0 Supervision 20.6 ICR 6.0 Mission Data Performance Date No. of Staff days Specializations Ratings (month/year) persons in field represented Rating Types of trend p~oblems Identification1 111999 1 -- Task Manager (I)/ -- -- Preparation Financial Sector1 Economist 311999 3 Task Manager (I)/ -- -- Financial Sectorl Economist; Banking (1); Housing Finance (1) 511999 5 Banking Restructuring -- -- (1); Task Manager (I)/ Financial Sector; Banking Supervision (1); Housing Finance (1); Securities Market (1) Appraisal I 1011999 4 Task Manager (I)/ -- -- Negotiations Financial Sectorl Economist; Treasury Operations (1); Financial Management (1); Legal (1) Supervision 1211999 1 Task Manager (I)/ S S Financial Sectorl Economist 512000 5 Banking (1); Housing (1); S S Finance (1); Capital Markets (1); Task Manager (1)1 Financial Sectorl Economist Task Manager (I)/ HS HS Financial Sector1 Economist; Housing Finance (1); Banking Restructuring (1); Banking Supervision (1) 612001 3 Task Manager (I)/ HS HS Financial Sectorl Economist; Banking (1); Housing Finance (1) Completion 2 -- BankinglEwnomist (1); HS S Financial Sector Specialist (1) First Programmatic Financial Sector Adjustrn ent Loan (Loan No. 7167-CO) Key Project Data (amounts in US$ million) Ao~raisal Actual or Actual as % of esiirnate current estimate appraisal estimate Original commitment 150 150 100 Total cancellation Total project cost Cancellation Project Dates Original Actual Departure of Appraisal Mission 1111212002 03/27/2003 Board approval -- 04/24/2003 Signing -- 05/28/2003 Effectiveness 06/04/2003 06/04/2003 Closing date 06/30/2003 06/30/2003 Staff Inputs (staff weeks)* Actual/Latest Estimate N" Staff weeks USJUSS('OO0) IdentificationIPreparation Appraisal I Negotiations Supervision ICR Total * Not Available Date No. of Staff days Specializations Performance Ratings (month/year) persons in field represented Rating trend Types of problems Identification1 Preparation Appraisal I Negotiations Supervision Completion * Not Available Second Programmatic Financial Sector Adjustment Loan (TF-52531; FSLT-72580) Key Project Data (amounts in US$ million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Original commitment 100 100 100 Total cancellation -- -- -- Total project cost Cancellation Project Dates Original Actual Departure of Appraisal Mission -- 0811812004 Board approval 09/23/2004 09/23/2004 Signing 09/24/2004 09/27/2004 Effectiveness 10/04/2004 10/04/2004 Closing date 03/31/2005 03/31/2005 Staff Inputs (staff weeks) Actual/Latest Estimate ' Staff weeks N US$US$(‘OOO) Identification/ Preparation 44.12 180.7 Appraisal I Negotiations 1 4.5 Supe~ision 5.67 18.5 ICR 10.7 31.5 Total 61.49 235.6 Mission Data Performance Date No. of Staff days Specializations Ratings (month/year) persons in field represented Rating Types of trend problems Identification1 0812003 4 -- Team Leader (1); - - Preparation Financial Officer (1); Insurance Specialist (1); Housing Finance Specialist (1) 1212003 3 -- Team Leader (1); -- - Financial Officer (1); Housing Finance Specialist (1) 0212004 3 -- Team Leader (1); -- - Financial Economist (2) Appraisal 1 0812004 3 -- Team Leader (1); -- - Negotiations Financial Economist (1); Lawyer (1) Supervision 1112004 3 -- Team Leader (1); Team S S Member (1); Legal Specialist (1) 1212004 3 -- Team Leader (1); Team S S Member (1); Consultant (1 0112005 2 -- Team Leader (1); Sr. S S Financial Economist ( I ) Completion 0812005 1 - Consultant -- - 0912005 1 -- Team Member -- -- 1012005 1 - Team Leader -- - Annex B. Key Objectives FINANCIAL SECTOR ADJUSTMENT LOAN (FSLT-70000) Indicator ActuaVLatest Estimate L I. Maintain a satisfactory Macroeconomic Government's macroeconomic program on target; Framework inflation and interest rates down significantly, contributing to a more stable financial system. 11. Strengthen the Legal, Regulatory and Amended Law passed in 1999: enhances ability of Institutional Framework of the Financial regulators to act quickly and take corrective System measures to prevent systemic problems and support sustainable growth. 1 Capital adequacy ratios of the banking system increased significantly, thus improving the system's ability to withstand economic shocks and downturns. "institutional recovery programs" contributed to reduced incidence of bank failures and avoidance of systemic crisis. Greater depth of regulatory norms and increased effectiveness in their implementation and in supervision efforts by SBC and FOGAFIN which has reduced systemic risk. Market based burden sharing mechanisms implemented to absorb risks via sale of mechanisms implemented to absorb risks via sale of assets. 111. Restructuring and Resolution of State- New Banco, Agrariohas improved the quality of the Owned Banks loan portfolio of public sector banks. has substantially reduced the presence of the Government in the financial sector and increased the efficiency of the system and prospects for sustainable growth. has contributed to greater efficiency in financial intermediation by passing on greater participation to s private sector banks. IV. Reform of the Cooperatives System By applying same rules to financial cooperatives as those applied to other intermediaries, inherent structural weakness of this sub-sector has been largely diminished. The current seven financial coo eratives are in the rocess of diversif in their Indicator ActuaVLatest Estimate business into other asset / liability products in line with their competitive strengths. V. Rationalization of Special Programs FOGAFIN has been able to increase its efficiency as Implemented by FOGAFIN it has increased the number of financial intermediaries participating in special programs without detracting from its basic deposit insurance functions. The special programs have provided temporary support while financial institutions become more strongly capitalized. FIRST PROGRAMMATIC FINANCIAL SECTOR ADJUSTMENT LOAN (Loan No. 716 7 4 0 ) SECOND PROGRAMMATIC FINANCIAL SECTOR ADJUSTMENT LOAN (TF-52531; FSLT- 72580) Objectives ActualILatest Estimates of PFSAL 1 ActualILatest Estimates of PFSAL 2 I I I. Bank Regulation and Resolution Implementation of a Law 795 approved in January 2003. Specific The net return on average earning assets of reformed law to ensure measures as follows: (a) Article 13, and Circular the banking system increased from 1.1% at sustainability of the 007; (b) Articles 12 and 13; (c) Articles 12 and end-2002 to 2.7% by May 2004. financial system 45: (d) Articles 81 and 82; (e) Articles 6 and 26; Policy Objective: and (0Articles 32, 33, and 45. Some 30 decrees Upgrading the Banking and regulations have been issued by SB and Regulatory, MOF to proceed with the implementation of the Supervisory, and new rules established by Law 795 including the Institutional procedures for consolidated supervision. New Framework credit risk management system (SARC) based on External Circular 01 1,2002 in process of implementation. Extensive regulations set forth concerning the The non performing loans ratio decreased reporting of financial transactions as specified in from 13.6% to 9.2% same period. During External Circular 25 as per Law 795. the same period. As per Law 795 the specified areas were The capital adequacy ratio (capital as a % of covered as follows: (a) Articles 11, 12, 28, and risk weighted assets) increased from 12.5% 29; (b) Article 28 f. g, h, and 1:and (c) Article to 14.5% during a period in which stricter 28. capital and loan classification standards I were implemented. In addition to the carve-out procedures, Article I The Banking Superintendency, between 7,29,30, and 31 and SB ~ e c r e e1335 address May 2003 and June 2004, applies fifty three the winding down of a financial institution. sanction actions and penalties to credit institutions that were not complying with I legal and prudential norms. Commitment letter to the Bank confirmed, and I The Banking Superintendent has approved a study currently underway. modified regulatory circular to incorporate procedures for monitoring, assuring oversight, and conducting consolidated supervision of financial conglomerates related party businesses, and computation of effective capital of financial institutions within conglomerates, and has subjected the draft regulation for review by the financial I industry and the SV. Transfer of IF1 assets to BANCOLDEX I The SB has developed its internal risk completed as per shareholder assembly. assessment procedures for evaluating vulnerabilities of supervised financial institutions and prioritizing the SB's inspection program commensurate with SB's human and budgetary resource base and capacity. Privatization process of BANCAFE has The SB has developed an action plan proceeded with public bid requests issued specifying the human, budgetary and time including: (i) sale of 51 % of the bank to a resources required to complete technical strategic investor; (ii) sale of the remaining 49% reviews of banks' internal credit risk of shares to the public: and (iii) strategic management models which incorporate investor to buy unsold shares. Base1 I1 orientations, allowing for full implementation in a prescribed period commensurate with SB's human and budgetary resource base and capacity. 11. Housing Finance Strengthening the (a) Law 794 enacted in January 2003 including The FOGAFIN administered inflation Housing Finance Sector the relevant Articles: 124, 300,315,424, and hedge for mortgage borrowers, was and Expansion of the 501; (b) Decree 710 issued stating procedures implemented with 15,000 subscribers Mortgage Market for micro-credit from commercial banks through covering loans up to US$135 million Policy Objective: finance companies; (c) Decrees 775, 777, and equivalent. Upgrading the 779 and External Circular 22 of 2003 provide RegulatO@perational the basis for housing lease finance. The Framework of the streamline of banking foreclosure procedures is Mortgage Market and Improving Access to expected to significantly reduce delays in Credit executing a foreclosure. (a) Law 795, Article 96 and Decree 66 establish Thirty issues of mortgage backed securities the guidelines for the FOGAFIN swap. were completed in 2003 for a total of US$ FOGAFIN has issued External Circular 001 for 700 million equivalent. In June 2004, the 2003 providing detailed guidelines for financial first issue of non performing asset backed intermediaries. In addition, informational securities (US$ 64 million) took place. literature has been developed for consumers explaining the swap protection program; (b) the Central Bank has also adopted and published procedures whereby financial intermediaries can utilize mortgage backed securities in REPO operations. A comprehensive assessment of the tax incentives affecting the overall housing finance sector has been completed by the MOF, including sequenced proposals to improve their consistency, efficiency and targeting, which will constitute elements of policy actions for future reforms. The MoF and the Central Bank restructure the FRECH swap hedge into an interest rate options product, after an agreement on the actuarial methodology and pricing parameters reached with the mortgage finance industry, and contracts are subscribed with mortgage banks to protect against interest rate risk. The FRECH liquidity fund to permit mortgage securities access to REPO markets is fully approved and operational. 111. Non-Bank Financial Institutions (Insurance, Pension, Trusts) Strengthening Law 795 adopted the appropriate language as Regulations on solvency margins and Regulatory Framework follows: (a) Articles 17 and 43: (b) Articles 5 technical reserves for the insurance and Financial and 16 and Decree number 1222: and (c) Article industry, by branch of business, have been Soundness of the New capital standards are expected to be drafted by the Insurance arm of the SB, and Insurance, Trusts, and enforced as of January 2004. approved by the Bank Superintendent Pension Industries subject to the final review of the insurance Policy Objective: industry and actuarial experts. Upgrading the Regulatory The SB has provided a policy response on Framework for Risk each core principle with respect to the Management and Capital evaluation report of insurance sector Adequacy ojthe Non Bank Financial Service regulation and its adherence to core Institutions principles for supervision. IV. Capital Markets Modernizing the Both SB and SV have issued regulations Twenty-eight public offerings of securities Legislative Framework forgoverning valuation of securities, including the were made in 2003 for an amount of the Securities Market classification of holdings and the treatment of US$1.8 billion equivalent, an increase of Policy Objective: gains and losses. The Borrower has also 19% with respect to the previous year. Upgrading the Regulatory Framework for the developed a methodology to establish a yield Governance, curve used as a basis for valuation of other non- T~~~~~~~~~~~ operations liquid securities. The Colombian Stock and Liquidity of the Exchange publishes a daily price vector which is Securities Market to be used by financial intermediaries in the valuation of their investment portfolios. Both SB and SV have signed a memorandum of Specific reporting requirements pertaining understanding for mutual collaboration. The to both credit rating and market risk ratings, procedures for joint supervision are being have been established for the mutual funds worked out. and collective investment scheme industries. Uniform asset valuation standards and methodologies have been established by regulation for private sector fixed income securities. V. The Government Debt and Money Markets Upgrading the Policy (a) As per Law 795, Article 16, and External A program of weekly issuance of 90 days and Operational Circular issued by the Central Bank regarding T-bills has been implemented with issuance Framework for the the purchase of on-the-run securities; (b) authorization expanded from COP 5 billion Government Securities resolution regulating primary dealers was (US$ 1.9 mn. equivalent) in January 2003 to Market issued; and (c) Resolutions 964 and 1047 were COP 80 billion (US$ 3 1 mn equivalent). 90 Policy Objective: issued to regulate the valuation of domestic day T-bills issued in 2003 amounted to US$ Implementation of a treasury securities based on the methodology 238 mn. and 180 day T-bills were US$ 34 Public Debt Issuance used in estimating a yield curve that is to apply mn. In 2004 through July, 90 day T-bill Strate&?'to Improve the in the case of non-liquid segments of the market. issuance was US$661 mn. Functioning ojthe Fixed Income Market Responsible Government Agencies: DGCP, SV, BR Provisions legislating legal finality for repurchase, short term collateralized loan instruments and securities swap transactions, have been incorporated in the new Securities Law presented to Congress. Annex C. List of People Met Colombian Officials Jorge Castellanos Ex-President of FOGAFIN Sergio Clavijo Ex-Vice Minister of Finance Sara Ordonez Ex-Superintendent of Banks Jorge Pinzon Ex-Superintendent of Companies and Ex-Superintendent of Banks Leonardo Villar Director, Banco de la Republica Hernando Vargas Deputy Governor, Banco de la Republica Jose Leibovitch Director of Research, Banco de la Republica Alejandro Gamboa Director Credito Publico, Ministry of Finance Javier Arias Acting President, CISA Enrique Valderrama Superintendent of Cooperatives (Economia Solidaria) Private Sector Patricia Cardenas President, Banking Association Alejandro Figueroa President, Banco de Bogota Juan Pablo Cordoba President, Colombia Stock Exchange Javier Fernandez Economic Consultant Bank Group Staff Fernando Montes-Negret Sector Director Guillerrno Perry Chief Economist Krishna Challa Sector Leader and Task Team Leader Alberto Chueca-Mora Operations Adviser John D. Pollner Task Team Leader Mariluz Cortes Consultant