Mexico - Financial Sector Restructuring Loan Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No.: 18012 IMPLEMENTAT'ION COMPLETION REPORT MEXICO FINANCIAL SECTOR RESTRUCTURING LOAN [Loan 3911-ME] May 26, 1998 Finance, Private Sector and Infrastructure SMU Country Management Unit 1 Latin America and the Caribbean Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENTS Currency Unit = New Mexico Peso (MexNP) US $1.00 MexNP8.48 (May 1, 1998) FISCAL YEAR OF BORROWER January 1st - December 31st ABBREVIATIONS AND ACRONYMS ADE Acuerdo de Apoyo Immediato a los Deudores de la Banca (Debtor Relief Program) BBV Banco Bilbao Vizcaya (Spanish Bank) CAMEL Evaluation of Capital, Asset Quality, Management, Earnings and Liquidity CNB Comision Nacional Bancaria (National Banking Commission) CNBV Comisi6n Nacional Bancaria de Valores (National Banking and Securities Commission) (Until March 1995, CNB, Comisi6n Nacional Bancaria) FOBAPROA Fondo Bancario de Protecci6n al Ahorro (Bank Fund for Savings Protection) FSRL Financial Sector Restructuring Loan FTAL Financial Sector Technical Assistance Loan (3838-ME) HSBC HSBC Holdings plc (Parent of Hong Kong and Shanghai Banking Corporation) IDB Inter-American Development Bank IFC International Finance Corporation IMF International Monetary Fund NAFIN Nacional Financiera, S.N.C. OECD Organization for Economic Cooperation & Development PROCAPTE Programa de Capitalizaci6n Temporal (Temporary Capitalization Program) SHCP Secretaria de Hacienda y Credito Pdblico (Ministry of Finance and Public Credit) SPEUA Sistema de Pagos Electr6nicos de Uso Ampliado (Amplified Use Electronic Payments System) UCABE Unidad Cordinadora para el Acuerdo Bancario Empresarial (Bank Enterprise Coordination Unit) UDI Unidad de Inversi6n (Investment Unit) VVA Valuaci6n y Venta de Activos : 0 00 0 00 Vice 0President ; 0 Shahid Javed Burki t; 0S;0 Country Managemernt UJit Director Olivier Lafourcade : v~~Sector Management Unit Director Danny Leipziger 0 : ~Task Manager .Michele Lubrano  FOR OFFICIAL USE ONLY MEXICO FINANCIAL SECT'OR RESTRUCTURING LOAN Loan 3911-ME TABLE OF CONTENTS PREFACE EVALUATION SUMMARY.............................................................................. PART I: INTRODUCTION ........................................................I PART II: BRIEF HISTORICAL BACKGROUND ON THE SECTOR A. Conditions Prior to the Crisis..............3 B. The Government's Response ..................................................3 C. Constraints on the Government's Response ..................................................5 PART III: STATEMENT/EVALUATION OF OBJECTIVES ................................ 7 PART IV: ACHIEVEMENT OF STATED OBJECTIVES A. Solvency and Soundness of Mdexico's Banks; Bank Restructuring; Confidence in the Financial System .................................................. 9 B. Improved Accounting Standards; Prudential Regulations; Strengthened Supervision; Prevention of Recurrence ................................. 18 C. Improved Discipline in the Provision of Liquidity by the Banco de Mexico ................................................. 22 D. Initiate Reforms in Development Bank Lending, Accounting Practices and Regulation of Financial Groups, and the Deposit Insurance System ................................................. 22 PART V: MAJOR FACTORS AFFEiCTING THE PROJECT A. Macroeconomic Environment ................................................. 25 B. Socio-Political Environment ................................................. 28 This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  PART VI: PROJECT SUSTAINABILITY ....................................... 29 PART VII: BANK PERFORMANCE A. Preparation and Negotiation ..................................... 31 B. Supervision ........................................... 31 C. Inputs to Government Response; Policy Dialogue .................................... 33 PART VIII: BORROWER PERFORMANCE ................................... 37 PART IX: ASSESSMENT OF OUTCOME ...................................... 39 PART X: LESSONS LEARNED A. Responding to a Crisis ........................................... 41 B. Lessons for Operations ........................................... 41 C. Lessons for the Financial Sector ........................................... 42 PART XI. ANNEXES ........................................... 43 A. Chronology of Important Events in the Mexican Banking Sector 1994-1997 B. Capital Increases and Purchased Loans by FOBAPROA C. Government Estimate of Fiscal Cost of Banking System Rescue D. Changing Composition of Mexican Banking System, 1994-1998 E. ICR Borrower's Report  IMPLEMENTATION COMPLETION REPORT MEXICO FINANCIAL SECTPOR RESTRUCTURING LOAN LOAN 3911-ME PREFACE This is the Implementation Completion Report (ICR) for the Financial Sector Restructuring Program, for which Loan 3911-ME in the amount of US$1,000 million was approved on June 22, 1995 and made effective on July 21, 1995. The loan was closed on December 31, 1996. It was fully disbursed, and the second and final disbursement was approved on July 12, 1996. This ICR was prepared by Mr. Michele Lubrano (CCMSD) and is based on information available as of February 1998. It was reviewed by Mr. Fernando Montes- Negret, Sector Leader, Finance, Private Sector and Infrastructure, LCC 1 C and Mr. Sri- Ram Aiyer, Task Manager of the FSRL and former Director, LAT (now Director, EACKO). Mr. Carl Dahlman (Staff Director, DECWD), Mr. Monte-Negret's predecessor as Sector Leader, also contributed to the preparation of the report. A draft of this ICR was circulated to the staff of LAC, FSD, DEC, LEGLA and EXC involved in the Mexican financial sector operations, discussions were held, and comments received have been incorporated into this final version. Part IV. A. ("Major Factors Affecting the Project-Macroeconomic Environment") was prepared by Mr. Ulrich Lachler, Principal Economist, LCC1C. Ms. Andrea Semaan assisted in the preparation of Annexes A through D to this report. The Borrower contributed to the ICR by preparing his own evaluation of project execution, which is included as Annex E.  IMPLEMENTATION COMPLETION REPORT MEXICO FINANCIAL SECT'OR RESTRUCTURING LOAN LOAN 3911-ME EVALIJATION SUMMARY Introduction 1. The Financial Sector Restructuring Loan (FSRL) was conceived as part of a set of bi-lateral and multi-lateral measures (including assistance from the United States government, the International Monetary Fund (IMF) and the Inter-American Development Bank (IDB) collectively referred to as the "Mexican Rescue Package" intended to reassure financial markets and restore confidence in the Mexican financial system. The Mexican Authorities, with the help of the Bank team, developed a set of objectives and guidelines for action. The loan was structured as a two tranche operation-US$500 million disbursed almost immediately and a second tranche conditional on completion of diagnost:ic audits of intervened and troubled banks, thorough examinations of all major banks, implementation of new accounting rules and completion of studies for the improvement of the legal/regulatory framework for private debt restructuring, secured transactions and asset-backed securities. Developing options for a roll-back of Mexico's over-extensive deposit protection system was included as a dated covenant. Project Objectives 2. The stated objectives of the program supported by the FSRL were to: (a) restore the solvency and soundness of Mexico's banking system, restructure banks and improve confidence in the financial system; (b) reform accounting standards and prudential regulation for banks, and strengthen supervision to prevent future recurrence of systemic problems; (c) improve discipline in the provision of liquidity by the Banco de Mexico; and (d) initiate reforms in development bank lending, accounting practices and regulation of financial groups, and the deposit insurance system. Although, as discussed below, objective (a) was probably an over-ambitious one and achievement of the other objectives was not complete, the issues addressed by the stated objectives of the loan-adequacy of bank capital, restructuring of banks, accounting standards, prudential regulation, the incentive regime, supervision and the quality of information on banks-have proven over time to be the critical ones and they remain the focus of discussion both among the Mexican Authorities and between the Mexican Authorities and Bank staff. Implementation Experience and Results 3. The degree to which each of the stated objectives of the program has been achieved has been varied:  ii a) While it was not expected that a rapid and full restoration of the banking system's solvency and soundness would be achieved in the immediate aftermath of the crisis, more rapid progress in this direction was anticipated than in fact occurred. This slower than expected progress is largely explained by the fact that the condition of the banks was in reality much worse than the information available to the Mexican Authorities indicated when the loan was prepared. The disposition of intervened banks, while initially proceeding at a slower than expected pace, accelerated toward the end of 1996. However, these interventions and the open-bank assistance program that became a central element of the Authorities' strategy have created an overhang of assets in Government hands that has ballooned to US$42 billion. A comprehensive strategy for resolving these assets still needs to be developed along with a strategy for financing the additional fiscal costs that are likely to be revealed in the process. b) The FSRL can be scored positively on achievement of its objectives in respect of accounting standards, prudential regulations and supervision. When fully implemented and applied, the new accounting standards adopted by the Mexican Authorities will be a big step forward for transparency. Prudential regulation has also been importantly strengthened and those areas that require further strengthening (in particular, classification and provisioning rules and capital composition) have been identified as areas to be addressed as soon as practicable. The FSRL can also be scored positively on helping the National Banking and Securities Commission (CNBV) to establish procedures to improve internal controls in banks and to increase the quality of information on the condition of banks now available. However, evidence suggests that internal controls of banks necessary for accurate reporting to the CNBV continue to be in need of significant improvement. c) The objectives of the FSRL in respect of liquidity provision by the Banco de Mexico were substantially achieved prior to Board approval of the operation. In March 1995 the Banco de Mexico began requiring pricing of central bank intra-day credit on a real time basis, pledging of collateral by banks that maintain large and frequent overdrafts with the central bank and debit caps on the portion of daylight overdrafts not collateralized. The Banco de Mexico also established the Amplified Use Electronic Payments System (Sistema de Pagos Electr6nicos de Uso Ampliado - SPEUA), a large value payments system for Mexican commercial banks and their foreign correspondents. d) Progress toward reform of development bank lending practices, better accounting and regulation of financial groups and a more economically rational system for deposit insurance has clearly been made under the program supported by the FSRL. However, greater progress was anticipated to have followed, albeit after second tranche release, in the areas of development bank lending practices and particularly with respect to the roll-back of FOBAPROA liability coverage.  iii Summary of Findings, Assessment of Outcome and Lessons Learned 4. A number of important factors had an important impact on the Mexican Authorities' ability to respond to the crisis in the banking system and the achievement of the objectives of the program supported by the FSRL. These included: macroeconomic and fiscal constraints (generating a strong bias against actions that might result in large fiscal outlays in the short-term); institutional limitations and incomplete information on the financial condition of banks; distortions in the incentives facing bank managers and limited managerial capacity in the financial sector; and socio-political considerations. Each of these factors affected the design and implementation of the actions taken by the Mexican Authorities in response to the crisis in the banking system. 5. The actions taken by the Mexican Authorities in the aftermath of the peso devaluation of December 1994 succeeded in restoring a degree of stability and confidence in the Mexican banking system. Significant consolidation has taken place and strong foreign institutions have now entered the market. The surviving banks in the system are clearly less vulnerable than they were in early 1995. However, the system as a whole remains fragile and does not yet appear to be on a reasonably clear and sustainable path to full recovery. Some, partial, elements of a complete strategy for full recovery of the financial sector are clearly in place. But these, by themselves, appear insufficient to take the system the rest of the way towards full solvency and soundness. Accordingly, project sustainability should be regarded as uncertain. The principal concerns about project sustainability include: likely undercapiitalization of banks and continued doubts about asset quality; the continued presence of excessive bank liability protection and the distortions it creates in the incentive framework for banks; a very large overhang of impaired assets in the hands of the Government's resolution agency; remaining unaddressed issues in the legal/regulatory infrastructure for banking operations; and the need for a well-articulated strategy for financing the considerable costs incurred in rescuing the banking system. 6. The FSRL was prepared in record time. It demonstrated the Bank's ability to effectively respond to the needs of the client and permitted the Bank to remain engaged in the resolution of a crisis whose implications went well beyond Mexico. With the assistance of staff from other regions, the Financial Sector Development Department (FSD), Development Economics (DEC) and International Finance Corporation (IFC), and in cooperation with the Inter-American Development Bank (IDB), the Region's team was able to build on the work done in connection with the Financial Sector Technical Assistance Loan (prepared before the onset of the crisis) and to work with the Authorities to design and prepare an operation that identified and addressed the key policy issues that are still the focus of both the Government's and the Bank's efforts in the sector. 7. The Bank's initial supervision efforts were fruitful and reasonably effective. During the preparation of the FSRL, Bank staff and consultants contributed importantly to the Government's efforts to gather accurate information on the condition of banks,  iv develop its strategy for bank resolution and implement other key policies described in the Memorandum of the President for the operation. After approval of the FSRL, the Bank staff regularly met with its Mexican counterparts to discuss the progress of the various programs and actions implemented to address the crisis. They provided periodic inputs on the longer-term goals for the financial sector, including roll-back of FOBAPROA liability coverage, greater transparency and disclosure, consolidation of the system, resolution of intervened banks and disposition of assets and steps to encourage greater reliance on market supervision. 8. However, important shortcomings in the effectiveness of the Bank's supervision efforts arose in the period after October/November 1995. Supervision of the FSRL absorbed a great deal of the time of the limited number of Region staff with financial sector experience. While the Region received important assistance from FSD and experienced outside consultants, these resources were rarely available on a full time basis. The CNBV commented on several occasions that World Bank staff turnover and the Bank's use of consultants on a short-term basis limited-the Bank's potential contribution. In addition, differences arose between the Bank staff and the Mexican Authorities. These principally revolved around the aggressiveness and timing of implementation of policy actions. The productivity of the dialogue between Mexico and the Bank on these points was complicated by disagreements over the need to share non-public information on a timely basis or to consult with the Bank prior to taking major steps to address the crisis. 9. Two additional factors complicated efforts to maintain a fruitful dialogue during the course of supervision. First, the World Bank's financial contribution to the Mexican Rescue Package was much smaller than that of the IMF or the United States government, but came with what appeared to some of the Bank's counterparts as much more taxing conditions. Also, those Mexican officials most often charged with interfacing with the Bank were the overworked staff of the CNBV. This agency was not the immediate recipient of the resources disbursed under the FSRL. 10. The Mexican Authorities responsible for the financial sector can take credit for a number of critical steps whose successful achievement was by no means a certainty at the time the FSRL was under preparation, including: averting a collapse of the banking system at a time of economic contraction and great uncertainty; stabilizing the system through mechanisms that avoided large current fiscal outlays and that attracted additional private capital; increasing their capacity to monitor the condition of individual banks and to analyze the strengths and vulnerabilities of individual institutions and the system as a whole; and restructuring, liquidating and disposing of intervened banks. Areas where borrower performance could have been better include: coordination among working level staff of different agencies; assignment of clear responsibility among agencies (or at least departments of agencies) and avoidance of conflict of interest; transparency; assessment of longer-term costs and risks; and sensitivity to, and avoidance of, negative incentives. 11. Overall, the outcome of the FSRL is assessed as unsatisfactory. A number of the operation's major objectives have only partially been achieved. Progress toward  v accomplishing the longer-term objectives of the program supported by the FSRL has been slow. More importantly, the operation's sustainability remains in question. The strategy followed thus far has prevented a generalized collapse of the financial system and restored a degree of confidence, but at high cost and without removing the major policy (deposit insurance) and structural (transparency and financial infrastructure) impediments to a market-driven and sound financial system. A clearly articulated strategy to take the Mexican financial system the rest of the way toward contributing to economic growth on the basis of efficient and well-capitalized intermediaries has yet to emerge. 12. The experience under the FSRL generated important lessons for future Bank operations in response to a crisis, sector adjustment operations generally, and Bank projects in the financial sector: Responding to a Crisis. The Bank should develop special standards for identification, content, preparation, negotiation and supervision of emergency loans like the FSRL. By definition, such loans do not benefit from the long identification and preparation periods that characterize most other Bank operations. In preparing and supervising operations in sensitive sectors such as the financial s:ystem during a crisis, the Bank needs to have intelligence on the socio-political context and how it is likely to constrain implementation of policy actions. The Bank must also be realistic about the degree of access to sensitive information it can expect its counterparts to provide. The Bank staff must make clear to their counterparts early on in a sensitive operation what types of information will be required to be shared. Finally, the Ban]k's ability to contribute to the resolution of crises like the one that occurred in Mexico would strongly benefit from a well-developed financial sector strategy at the Global and Regional level. Lessons for Operations. The Bank needs to develop appropriate Bank-wide strategies for sector adjustment and other operations whose objectives can be fully achieved only years after disbursements are complete(l. Better mechanisms also need to be developed for maintaining continuity of staff working on large and complex operations. Undue staff turnover hinders effective supervision, generates inefficiencies, is a burden on counterparts and hurts client confidence in the Bank's ability to offer effective advice. Lessons for the Financial Sector. The Mexico experience demonstrates the need for policymakers to be sensitive to the negative incentives created by emergency programs for the financial sector. In particular, it highlights the importance of ordinarily conditioning relief to troubled banks on replacement of existing management. A country's financial sector infrastructure., in particular the legal/regulatory/institutional framework for making and enforcing credit transactions can be a critical constraint on developing sounder banking systems and resolving banking problems. The experience of Mexico highlights the need for greater Bank effort in this area. The Bank has comparative advantage in bringing to bear a great deal of knowledge and expertise to address emerging problems in the financial sector. Finally, the Bank's efforts in one area  vi of the financial sector need to be placed in the context of an over-all strategy for policy actions in the sector.  IMPLEMENTATION COMPLETION REPORT MEXICO FINANCIAL SECTOR RESTRUCTURING LOAN LOAN 3911-ME PART I. INTRODUCTION 1. The Financial Sector Restructuring Loan (FSRL) was conceived as part of a set of bi-lateral and multi-lateral measures (including assistance from the United States government, the International Monetary Fund (IMF) and the Inter-American Development Bank (IDB)) collectively referred to as the "Mexican Rescue Package" intended to reassure financial markets and restore confidence in the Mexican financial system. On February 9, 1995, following discussions with the Mexican Finance Minister, the World Bank issued a press release welcoming the international package of loan guarantees announced the previous week by the President of the United States and committing the Bank to defining a program of support for Mexico in three areas: commercial banking and financial sector reform; health, nutrition and educational services for the poorest regions of the country; and privatization. Preparation of the Financial Sector Restructuring Loan and other emergency operations began in earnest following this announcement.' 2. The Mexican Authorities, with the help of the Bank team, developed a set of objectives and guidelines for action. However, because of the dynamic nature of the crisis they left open the precise design Emd timing of a number of the measures that might ultimately be employed. The loan was structured as a two tranche operation-US$500 million disbursed almost immediately and a second tranche conditional on completion of diagnostic audits of intervened and troubled banks, thorough examinations of all major banks, implementation of new accounting rules and completion of studies for the improvement of the legal/regulatory framework for private debt restructuring, secured transactions and asset-backed securities. Developing options for a roll-back of Mexico's over-extensive deposit protection system was included as a dated covenant. From the start, the Bank and the Mexican Goverrnment planned to negotiate a follow-on operation to support further progress in the areas of troubled bank resolution, regulation and supervision, the incentive regime and particularly the roll-back of Mexico's over- extensive bank liability protection scheme. The Board of the World Bank approved the FSRL on June 22, 1995.2 IDB approved a companion US $750 million operation on At the request of EXC, an analysis of the World Bank's response to the Mexico crisis, "Note on the Lessons for the World Bank of the Mexican Crisis," was prepared by the Office of the Senior Vice President, Development Economics, and presented to the Board of the Bank on December 17, 1996. 2 At the same meeting, the Board approved an amendment to the Financial Sector Technical Assistance Loan (FTAL; originally approved in January 1995) to provide additional funds for diagnostic audits of banks, consultants, seconded superviso:rs, and other efforts to shore-up supervision and provide the  2 substantially the same terns. The release of the second and final tranche of the loan was approved on July 12, 1996. National Banking and Securities Commission (CNBV) with the information until then lacking on the condition of banks and the system as a whole.  3 PART II. BRIEF HISTORICAL BACKGROUND ON THE SECTOR A. Conditions Prior to the Crisis 3. The rapid devaluation of the peso, and the period of economic recession and high interest rates that followed, hit a newly-privatized commercial banking system. In 1990, constitutional amendments and new banking legislation substantially liberalized the business of banking and set the stage for the return of the commercial banks to private hands after a period of almost ten years of state ownership. During the period of state ownership (1982-1991), lending by the banks to the private sector was limited, with the bulk of bank assets invested in Government securities and loans to parastatal entities. Between June 1991 and July 1992, all of the commercial banks then in state hands (18) were privatized, most of which became the largest subsidiaries of newly-established financial holding companies. Purchasers paid high multiples (an average of 3.1 times) of book value and, not surprisingly, quickly began to compete fiercely for market share in the provision of loans, both by increasing the volume of existing lines of business and by entering the markets for relatively new financial services, such as home mortgages and credit cards. Rapid growth in lending 'an annual real growth rate of 23.7% between 1991 and 1994) and the proliferation of new lending products strained the capacity of banks to maintain proper controls. Reported aggregate past due loans under Mexican accounting climbed from 4.09% at year-end 1991 to 8.3% at the end of the first half of 1994. 4. The banking system was severely shaken by the peso devaluation and the recession that followed. The three immediate adverse effects were: (i) foreign exchange losses incurred by the banks on account of their unhedged open positions; (ii) decline in the value of the banks' non-loan assets (consisting mainly of fixed-income securities) resulting from the sharp rise in domestic interest rates following the collapse of the peso; and (iii) a sharp decline in capital ratios due to the increase in the peso value of the banks' dollar-denominated assets and liabilities. In addition to these direct effects, banks faced significant liquidity pressures, increased funding costs and even more seriously, an accelerated deterioration in the quality of their loan portfolios attributable to high interest rates and economic contraction. B. The Government's Response 5. The Mexican Government implemented a series of actions to prevent bank runs and a general collapse of the banking s,ystem as well as to encourage restructuring of problem loans, debtor performance and recapitalization of banks. The main elements of the Government's program (several of which were designed and implemented subsequent to negotiation of the FSRL) were:  4 * Establishment of a special liquidity facility (through the bank support fund - FOBAPROA3) to provide dollar funding to banks. This facility closed towards the fall of 1995 as confidence returned and banks were able to return to private money markets. * Requiring all banks to increase reserves for loan losses to the highest of: (i) 4% of loan portfolio; or (ii) 60% of past due loans; or (iii) as required by quarterly loan classifications under existing rules. In practice, the 60% of past dues requirement has been the binding one. (This requirement was reduced to 45% after the new accounting rules went into effect on January 1, 1997.) * The Temporary Capitalization Program (PROCAPTE), under which FOBAPROA purchased subordinated convertible debentures from undercapitalized banks, with the proceeds redeposited with the Banco de Mexico. Banks were permitted to count the debentures for purposes of meeting regulatory capital requirements. The program was wound up within a year as participating banks recapitalized from other sources or were intervened. * An inflation-indexed (Unidad de Inversi6n - UDI) funding program to permit banks to carry troubled commercial, mortgage, agricultural, state and municipal and consumer loans restructured into long-term UDI-denominated obligations. Additional programs to provide direct subsidies to mortgage, agricultural and other borrowers who remained current on their payments ultimately augmented the initial program. * An immediate debtor support program (ADE) that provided certain classes of borrowers (principally consumers and mortgage debtors) with a direct interest rate subsidy for one year. A number of other support programs for different classes of debtors were established over the following year or so. * A loan purchase/recapitalization program under which FOBAPROA purchased a portion of a bank's loan portfolio in return for a commitment on the part of the bank to raise additional capital. The additional capital was raised from both existing and new shareholders and resulted in a dilution of the interest in the bank represented by its original shares. As payment for the purchased loans, the bank received 1 0-year non-transferable zero-coupon Government bonds. There were three waves of transactions involving all of the surviving privatized banks with a total of approximately US $20 billion face value of assets purchased. 3 FOBAPROA (Fondo Bancario de Proteccion al Ahorro - Bank Fund for Savings Protection) is a trust managed by the Banco de Mexico, but whose oversight committee is composed of representatives of the Ministry of Finance and Public Credit (SHCP), CNBV and the Banco de Mexico. FOBAPROA provides a blanket guarantee of all unsubordinated liabilities of the commercial banks. A fuller description of the governance and activities of FOBAPROA is included in Annex IX of the Memorandum of the President.  5 * Interventions of seriously decapitalized banks and assistance to stronger banks willing to merge with weaker institutions. Merger assistance was usually provided in the form of purchases of troubled assets of the weaker bank by FOBAPROA. Interventions and mergers typically wiped out all or most of the original interest of shareholders in intervened or weak banks. A chronology of important eve nts in the Mexican banking sector is provided in Annex A. C. Constraints on the Government's Response 6. A number of factors were regarded by officials as constraints on the Government's ability to respond to the crisis in the banking system. These perceived constraints explain some of the differences between what was anticipated during the preparation of the FSRL and the actual steps undertaken: * Macroeconomic/Fiscal Constraints. Subduing inflation levels after the devaluation was a priority. Therefore, Government actions which caused monetary expansion were strongly disfavored. To assure the credibility of its fiscal program, the Government sought to postpone actions that would result in large fiscal outlays in 1995 (and 1996) even at the risk of possibly much larger costs in the long run. - Institutional/Information Constraints. The demands of the crisis required more trained staff than the Authorities were able to mobilize. Bank interventions in particular severely strained resources. When the crisis erupted, the Mexican Authorities did not yet have the supervisory systems in place to fully assess the condition of the banks and the system as a whole. This made it difficult to differentiate between the conditiorns of individual banks and tailor responses. * Incentives/Bank Management Constraints. The Goverrnment guarantee of all unsubordinated bank liabilities distorted the incentives for the banking system and discouraged market discipline. (It also meant that in an individual case, intervention could be a more costly option for the Government than would have been the case if large depositors and bondholders could have been made to share in losses.) After ten years of Government ownership, the newly-privatized banks lacked adequate management, systems and trained personnel. Weak management is usually regarded as justifying greater supervisory intervention. However, because the banking system had only recently been privatized, the Government was not anxious to take more aggressive action against bank shareholders and managers since this might be regarded by the public as renationadization and would call into question the execution of the still recent privatization process. * Socio-political Considerations. Even prior to the onset of the crisis, an active debtor movement was militating for some sort of Government relief for over-indebted firms  6 and individuals. In the aftermnath of the peso devaluation, much of the public placed the responsibility for the economic crisis with the Government and the debtor movement gained nationwide prominence.  7 PART III. STATEMENT/EVALUATION OF OBJECTIVES 7. The stated objectives of the program supported by the FSRL were to: (a) restore the solvency and soundness of Mexico's banking system, restructure banks and improve confidence in the financial system; (b) reform accounting standards and prudential regulation for banks, and strengthen supervision to prevent future recurrence of systemic problems; (c) improve discipline in the provision of liquidity by the Banco de Mexico; and (d) initiate reforms in development banks lending, accounting practices and regulation of financial groups, and the deposit insurance system. Objectives (b), (c) and (d) were not original to the program. Most of their elements had been discussed as longer-term policy goals during preparation of the FTAL. Given that a crisis was unfolding and that the information then available to the Authorities and the Bank on the condition of the banking system was known to be incomplete, objective (a) was probably an over-ambitious one. However, one of the goals of the Authorities (as well as the Bank and other participants in the Mexican Rescue Package) was to reinforce confidence in the banking system and demonstrate that immediate and effective steps could be taken to address the crisis. In any case, the issues addressed by the stated objectives of the loan-adequacy of bank capital, restructuring of banks, accounting standards, prudential regulation, the incentive regime, supervision and the quality of information on banks-have proven over time to be the critical ones and these remain the focus of discussion both among the Mexican Authorities and between the Mexican Authorities and Bank staff.   9 PART IV. ACHIEVEMENT OF STATED OBJECTIVES 8. Although most of the concrete elements of the Government's program supported by the loan have now been implemented, some of the program's objectives were not expected to be fuilly achieved for several years. The degree to which the short- and medium-term goals of the FSRL have been accomplished can be adequately evaluated at this time. It is also possible to examine current indications and reach some early conclusions about whether the longer-term objectives are likely to be achieved. Such an interim assessment is essential as a guide for future World Bank operations in the sector. A. Solvency and Soundness of Mexiico's Banks; Bank Restructuring; Confidence in the Financial System 1. Solvency and Soundness of Mexico's Banks. 9. Before Mexico's banking system can be said to be solvent and sound, adequate levels of capitalization must be restored, a legal/regulatory environment that sets the correct incentives for prudent and effective management should be in place, and a financial infrastructure for lending based on sound practices and adequate information is required. Important progress in each of these areas has been made by the Authorities with the Bank's assistance, but much still needs to be done before it can be said that Mexico is firmly on its way to restoring solvency and soundness. a. Condition of the System. At the end of December 1997, Mexico's commercial banking system excluding intervened banks (and other bank assets now under FOBAPROA ownership) was comprised of 40 banks with assets of P$1,018 billion. These 40 included 11 banks that were privatized in 1991 and 1992, 10 newly-chartered domestic banks and 19 primarily newly-chartered subsidiaries of foreign banks.4 The banks most affected by the crisis and the principal beneficiaries of the Government's support have been the original privatized banks. Eleven of the original 18 have survived. At December 31, 1997, these 11 banks reported assets of P$947 billion representing 93% of the banking system (again excluding intervened banks and assets now under FOBAPROA ownership). The banking system overall rernains heavily burdened by nonperforming and low- earning assets and still inadequate reserves for loan losses. In January 1997, Mexico formally implemented new accounting rules that materially changed the financial statement reporting by Mexican banks and financial groups. (See Part IV, B, 1.) The area most affected by the changes was reporting of past due loans, which, according to CNBV figures, increased from P$48 billion or 6.4% of loans in December 1996 to P$101 billion or 13.2% at the end of March 1997. Prior to that time, past dues had been 4 Citibank is the only foreign bank operating in Mexico that is not newly chartered or primarily the result of the acquisition by a foreign bank of a Mexican-owned bank.  10 increasing, but at a declining rate under the old accounting rules. The more than doubling in past due loans primarily occurred because the previous accounting rules included only missed payments and not outstanding loan principal-unless the loan had fully or partially matured. In anticipation of the accounting change, most of the privatized banks worked aggressively to build reserves for loan losses, assisted in this effort by the FOBAPROA loan purchase/recapitalization program which relieved them of troubled assets and created incentives for the infusion of P$63.3 billion in new and committed capital by shareholders through the FOBAPROA loan purchase/recapitalization program. (See Part IV, 1, D.) Commercial bank and CNBV officials now indicate that the deterioration of bank portfolios (with the possible exception of mortgage loans) has abated. However, it will be several quarters before the extent of past dues is fully recognized under the new accounting rules.5 b. Capitalization of Mexican Banks. Capital adequacy remains of paramount concern for the Mexican banking system. However, the lack of transparency in bank accounting and reporting under the old rules and the as yet not fully consistent application by banks of the new rules make estimating the actual capital deficiency of the system imprecise if not speculative. Despite continuing serious problems in their loan portfolios, Mexico's banks report relatively high capital levels. At June 30, 1997, the system's average capitalization (excluding intervened banks) was reported at 16% of risk- weighted assets. Factors that contribute to this apparent anomaly include: (i) less stringent capital regulation, accounting standards and classification rules than in most OECD countries (See Part IV, B, 2); (ii) recognition in prior periods of substantial amounts of capitalized interest; and (iii) recognition of accrued interest from doubtful assets.6 However, in the aftermath of the crisis new rules were implemented to improve the quality of capital by preventing the financing of bank shares with loans from the financial group or using such shares as guarantees for loans. c. Foreign Participation. To foster the entry of new capital, the Mexican Congress, at the urging of the Authorities, modified bank ownership laws in April 1995 to permit foreign financial institutions to take control of troubled domestic banks. This opened up Mexico's banking system to foreign participation far more quickly than had 5 Salomon Brothers Inc, Latin America Equity Research, "Mexican Banking System - 1997 Will Be a Transition Year," May 13, 1997, pg. 8. 6 Government securities received by the banks under both the FOBAPROA loan purchase/recapitalization and UDI programs represent about half of the performing loans of the non- intervened banks. The bulk of the income on these securities is not paid in cash and the ultimate realization of principal and accrued interest is dependent either wholly or partially on the ultimate performance of the underlying loans. Because of incomplete access to data, the Bank staff (and probably the Mexican Authorities as well) cannot determine the precise extent of net income dependent on these securities nor can it assess the performance of loans sold or restructured under these programs. If income on these securities is indeed an important source of net profits and the underlying loans are ultimately impaired, then capital increases resulting from retained earnings may, in the end, be partially reversed.  11 been required under the North American Free Trade Agreement. As a result, a significant amount of foreign-sourced financial and human capital has been attracted to the Mexican banking system. Spanish banks, Banco Bilbao Vizcaya (BBV) and Santander, now control two formerly troubled privatized banks (Probursa and Mexicano, respectively) and Canada's Bank of Nova Scotia manages the operations of a third (Inverlat) with an option to assume majority ownership in the year 2000. Citibank has purchased Confia. Each of these transactions was made possible by significant Government assistance, most of it in the form of FOBAPROA purchases of troubled loans. Foreign investors have also taken important positions in major banks that remain controlled by Mexicans. In the second quarter 1996, Bank of Montreal acquired a 20% stake in Bancomer, Mexico's second largest bank. Mexico's third largest bank, Serfin, is now operating under an arrangement in which HSBC (the parent of Hong Kong Shanghai Bank) has agreed to purchase a 20% shareholding7 and is taking an active role in management of the bank. Bital also has a significant part of its ownership held by a foreign financial institution. Foreign-controlled banks now account for about 20% of the system's assets.8 If Serfin were to be included, this percentage would approach 30%. d FOBAPROA Loan Purchase/Recapitalization Program. As an incentive to recapitalize banks, beginning with the initial recapitalizations of Probursa and Serfin in June 1995, the Authorities launched the loan purchase/recapitalization program under which FOBAPROA agreed to purchase loans from banks in an amount equal to roughly two times the amount of new equity (iincluding mandatory convertible subordinated debt) that shareholders were willing to commit. FOBAPROA paid for the loans it purchased with 10 year zero-coupon Government securities at a purchase price equal to book value less initial provisions for loan losses. In addition, the selling bank remained responsible for an average of 25% of any losses on the portfolio realized over the 10-year life of the FOBAPROA security. To be eligible for the program, a participating bank had to be deemed by the CNBV to have satisfactory management-but actual changes in management and control were required only in exceptional cases. Shareholders had to demonstrate substantial financial commitment, and the bank had to undergo a management audit by an independent accounting firm and commit to strengthen systems, management and internal controls. All of the privatized banks that had not been intervened participated in three waves of transactions ending in December 1996. Over the life of the program, new and comrmitted capital in the form of equity and subordinated debt totaled P$63.3 billion (largely from existing shareholders and foreign financial institutions) and gross loan purchases by FOBAPROA totaled P$ 142 billion (P$ 119 billion net of reserves for loan losses). 7 Conditioned on the successful completion of a public offering for $300 million in the second half of 1997. 8 This figure includes BBV/Probursa, Santander/Mexicano and Inverlat as foreign-controlled, but excludes Serfin, in which HSBC has only a minority interest.  12 Like most open bank assistance (particularly methods not strictly conditioned on change in management), the FOBAPROA loan purchase/recapitalization program carried with it considerable risks, including moral hazard risks. One risk was that despite the efforts of the CNBV, inadequate, ineffective, or otherwise deficient management teams might remain in the banking organizations after the transactions. Another risk was that if the capital deficiency of banks was initially underestimated and the program kept open, repeated recourse would create moral hazard among bankers and the public and discourage more aggressive debt collection efforts. Some of these risks became realities (the first round of loan purchases did prove inadequate and two additional rounds followed) and are discussed in the following section. (Part of the explanation given for the repeated rounds was that it was felt that shareholders were not in a financial position to make large once-and-for-all infusions in the early stages of the crisis.) e. Incentive Framework. Mexico's broad system of bank liability coverage and the unintended side-effects of the Government's support programs created and continue to raise important concerns about distorted incentives and their impact on bank management. Mexico's traditional Government guarantee of all bank liabilities has long been recognized as an obstacle to market-based supervision of banks and a competitive and efficient financial sector. Although some differentiation in pricing in the inter-bank market exists, it is not commensurate with differences in risk. The private sector, and in particular the interbank market, continues to operate on the basis of the full Government guarantee-permitting even very poor institutions to raise deposits. The Bank staff and the Mexican officials that prepared the FSRL recognized that there might be serious negative consequences to significantly modify the liability protection scheme in the midst of a crisis.9 Accordingly, the conditions of the loan required the Authorities to take only the first steps in the development of a strategy for achieving a roll-back of coverage without undue risk to the system. Even so, as discussed in Part IV, D, 3, progress toward enunciating clear criteria for scaling back the guarantee has been slower than hoped. The larger banks that proved highly insolvent have now for the most part been intervened, merged with other institutions or otherwise restructured. However, a number of smaller banks remain largely dependent on interbank funding. Although detailed information on the portfolio condition of such banks is not available to the Bank or the public, recent portfolio growth may indicate a combination of capitalization of interest and possibly imprudent lending growth. The long-term viability of such banks in terms of maintenance of a sound funding base and the ability to make quality loans is at least questionable. The costs of resolving such banks, if they prove to be insolvent, can only be expected to increase if they continue to operate under the current incentives. The chronology set out in Annex A shows the progression of Government responses to the banking crisis, beginning in the first half of 1995 with the PROCAPTE program, the UDI scheme and the first loan purchase/recapitalization transactions. In the 9 Memorandum of the President, p. 26.  13 latter half of that year the ADE debtor support program was announced along with additional UDI debtor assistance programs and the first loan purchase transactions. Two more rounds of loan purchase/recapitalizations were carried out in 1996 and the debtor relief programs progressively expanded (first for mortgages (May), then agriculture (July) and then small and medium enterprise (August). The potentially negative incentive effects of the Government's progressive expansion of the original programs was for much of 1995 and 1996 the subject of some debate among the Mexican Authorities, the World Bank staff and the IDB. It was argued by the staff of the multi-lateral agencies that such an approach encouraged both banks and borrowers to believe that the Government could be counted on come up with new incentives for recapitalization and provide borrowers with additional relief, so long as the imamediate cash outlays are manageable. A once- and-for-all, "drawing a line in the sand" approach-stating after the initial emergency responses that no further debtor relief programs would be forthcoming and that open Bank assistance would be at the cost of extreme dilution of shareholders and complete replacement of management-might hiave been more effective at maintaining the proper incentives. Socio-political concerns may explain this apparent reluctance to take a more aggressive approach. A tougher line vwould also likely have resulted in more bank interventions which would have required larger current fiscal outlays, strain institutional capacity and appear to the public as a reversal of the privatization process. f Financial Infrastructure. The rapid deterioration in bank portfolios after the December 1994 devaluation highlighted the critical weaknesses of Mexico's legal/regulatory framework for lending. The costs to Mexico's financial system of not having a better developed infrastructure for structuring and enforcing financial sector operations have been dramatic. An eff ective bankruptcy system is lacking and there remain numerous legal/regulatory obstacles to private debt restructuring."0 Perhaps most significantly, important non-bank institutions such as credit bureaus and rating agencies are in their infancy and better business accounting and appraisal standards are needed. Mexico's current legal framework and the poor condition of its property registries prevents the widespread use of moveable property as collateral for loans. Legal, tax and regulatory obstacles impede more widespread use of securitization techniques. The underdeveloped financial infrastructure in these areas prejudices the soundness of the asset base of banks upon which the solvency of the system as a whole necessarily depends. When the FSRL was under preparation, the Bank staff encouraged the Mexican Authorities to take advantage of the opportunity to undertake concrete policy actions on secured transactions, asset-backed securities and obstacles to private debt restructuring, including Mexico's barely functioning bankruptcy system (problem areas identified in preparation of the FTAL). Their Mexican counterparts preferred to first build greater awareness, support and coordination within different parts of the Government. While 10 With financing provided under the FTAL, the CNBV recently conducted a training seminar on the topic of bankruptcy which recognized the poor state of the existing system and the need for reform.  14 they agreed that such matters were important to resolve in the medium to longer-term, the immediate priority was rescuing the existing system."' The Authorities produced official studies on these topics proposing in many cases concrete reforms that might be supported by a future Bank operation. Significant steps were taken in the area of mortgage transferability and securitization, but there have not been similar efforts on secured transactions and bankruptcy. 2. Restructuring of Banks 10. The need for considerable restructuring and consolidation of the banking system was foreseen during the preparation of the FSRL. However, the final resolution of intervened banks and the restructuring of other troubled entities proceeded very slowly at first. While specialized consultants (investment banks and accounting firms) were hired (as required under the FSRL) to assist in developing strategies for the disposition of intervened banks, the depressed market for banks and bank assets, the Authorities' decision not to sell into such a market, Mexico's lack of experience in resolving banks, the limited institutional capacity of the Authorities, their focus on saving the system as a whole and an unwillingness to incur immediate fiscal costs all acted to delay final resolutions of these entities. However, the Authorities deserve a great deal of credit for the pace with which resolutions were carried out once the economic recovery was underway in 1996. Today, the landscape of Mexico's banking system is significantly different from what it was at year-end 1994. Foreign shareholders have taken control of about 20% of the banking system and other troubled banks have been merged with less impaired institutions (with substantial Government assistance). It is anticipated that additional consolidation will occur in the near future, with some of the remaining weak banks merging with the stronger ones, with additional financial support of the Authorities. 11. The Operating Guidelines for Bank Restructuring agreed between the Bank and the Mexican Authorities'2 set out general conditions and methodologies for open-bank assistance. However, the principal mechanism ultimately employed by the Authorities (the FOBAPROA loan purchase/recapitalization program) was not explicitly contemplated by the guidelines, nor did the Government seek the Bank's direct input into its design or consult with the Bank in advance of structuring the initial transactions. The key differences between the types of open bank assistance contemplated in the Guidelines and the structure of the transactions that have actually occurred were: (i) the Guidelines l The Government found it necessary to sponsor a special organism (the Bank-Enterprise Coordination Unit or "UCABE") to develop informal arrangements for restructuring the debts of the largest borrowers to partially compensate for the inadequacy of the legal and institutional framework. This experience, and the Government's experience with the portfolios of intervened banks and those purchased under the FOBAPROA loan purchase/recapitalization program have fostered a much greater awareness of the pressing need to address these problems. 12 Attachment II to the Government's Letter of Financial Sector Policy (Annex I of the Memorandum of the President).  15 did not contemplate as systematic and general a program (open to all banks) as the FOBAPROA loan purchase/recapitalization program, but rather individualized responses; (ii) the Guidelines stressed the need to ordinarily condition open-bank assistance on a change of management, and this was not imposed in most cases; and (iii) the Guidelines were designed as an outline for a much smaller government intervention in the banking system, not a program that would ultimately result in the Government purchasing a quarter of the system's total loans (in addition to those acquired as part of bank interventions). Throughout 1995 and 1996 there were discussions between the Authorities and the Bank on the effec'tiveness of the program and ways in which transactions might be better tailored to achieve the desired goals. The Mexican Authorities' rationale for diverging from the Guidelines was essentially fiscal and practical. The diagnostic studies and bank examinations revealed loan quality problems much larger than was evident at the time the FSRL was negotiated. The loan purchase/recapitalization program permitted the Government to pay for troubled bank assets with zero-coupon non-tradable government securities, assuring minimal fiscal outlays at the onset. An approach more demanding on shareholders and management or that precluded future assistance might have resulted in the intervention of a larger segment of the banking system, requiring more immediate recognition of losses and larger short-term fiscal outlays. As a practical matter, more bank interventions would also have further strained the institutional capacity of the Authorities during a time of crisis. a. Overhang of Intervened Assets. FOBAPROA is now the largest holder of financial assets in Mexico and one that provides no financial intermediation. The huge overhang of impaired assets now under its ownership impedes the effective restructuring of the banking system in at least three ways: (i) many borrowers under the roughly one million loans under FOBAPROA controls are unlikely to return to credit markets until they have restructured their existing debts; (ii) the continued overhang of assets acts to depress the market value of bank assets generally; and (iii) continued Government control of such a large share of total indebtedness encourages continued politicization of the asset restructuring process. The evolution of FOBAPROA into a vehicle to acquire, hold and ultimately dispose of a large share of the impaired assets of the banking system was not anticipated at the time the FSRL was negotiated, but it was understood that the intervention process would leave the Government in possession of a significant amount of bank assets. As required under the FSRL, the Authorities developed plans for the restructuring and disposing of banks under FOBAPROA control with assistance from investment banking and accounting firms. The Authorities were optimistic that there would be appetite for the purchase of whole banks and accordingly FOBAPROA continued to operate such banks (including Cremi, Union, Oriente and Banpais) as going concerns despite continued losses well into 1996. The Bank staff urged early disposition, noting in the Guidelines that "experience shows that delays usually increase the ultimate restructuring costs." In the end, liquidation strategies and sales after dramatic portfolio  16 clean-ups were the options selected by the Authorities, but only after significant time had passed. In early 1996, the Authorities established a new agency, Valuaci6n y Venta de Activos (VVA), to assist FOBAPROA in managing and disposing of assets under its control."3 However progress in actually carrying out sales was very slow. In addition to the lack of institutional capacity, a number of factors discouraged greater progress: (i) sales of assets at less than their acquisition prices would begin to reveal the fiscal cost of the FOBAPROA loan purchase/recapitalization program; (ii) asset sales would also reveal the likely amount of loss sharing for the banks and force their recognition; and (iii) "fire sales" of VVA assets might create benchmarks for valuing the assets remaining on banks' books, with possible negative implications for capital adequacy. After conducting the first small (P$132 million face amount) trial auction of FOBAPROA loans in July 1997, the functions of VVA were absorbed directly by FOBAPROA. In the aftermath of the dissolution of VVA, FOBAPROA's strategy is to retain title to the largest share of the assets it has come to own. Consumer and residential mortgages and loans of commercial borrowers under US $7 million will continue to be managed by the banks for the account of FOBAPROA. A Corporate Assets Division has been established to more actively manage FOBAPROA's larger assets. Outside investment banking and management consultants have been hired to advise on how best to maximize the recovery of FOBAPROA's assets through such mechanisms as contracting management and recovery sharing. New strategies are clearly called for, but the results of the work of the consultants have yet to be publicized. 3. Confidence in the Financial System: Prospects for New Lending 12. The steps taken under the Government's program clearly restored a degree of confidence to the Mexican banking system. The FOBAPROA guarantee of all unsubordinated bank obligations is no doubt an underpinning of this confidence, but there is also a sense in the investor and depositor communities that the steps taken have stabilized the system and that risks are now manageable, at least for the short- to medium- term (maturities on bank obligations continue to average below one year). However, the intermediation function of the banks remains limited. Lending has not grown in a manner commensurate with the recovery of the economy, reflecting the continuing low demand for new credit given the still relatively high interest rates and concerns by the banks about 13 After the negotiation and effectiveness of the FSRL, both the Region's staff and the Financial Sector Development Department (FSD) provided the Government with input into possible mechanisms for disposing of assets of intervened banks and the assets acquired by FOBAPROA through its loan purchase/recapitalization program. Bank staff, consultants and the International Finance Corporation (IFC) discussed with the Authorities the possible structure and powers of VVA beginning in early 1996.  17 the creditworthiness of borrowers.'4 Much loan growth appears to be the result of the capitalization of interest from negative amortization on mortgage loans, loan restructuring and the appreciation on inflation-indexed loans (i.e., UDIs). Many bank managers are now anticipating loan growth starting in the second half of 1997. The extent to which this occurs will be affected by the conitinuing overhang resulting from the over US$42 billion book value of loans and other assets held by FOBAPROA. Evaluation. The first stated objective of the FSRL proved overly ambitious and progress has accordingly been unsatisfactory. While a full restoration of solvency and soundness was not expected to have occurred so soon after the crisis, more rapid progress in this direction was anticipated. This slower than expected progress is largely explained by the fact that the condition of the banks was in reality much worse than the information available to the Mexican Authorities indicated when the loan was prepared. The economic recession was also deeper Emd the ultimate recovery less balanced than anticipated. Although much additional capital was attracted to the system to compensate for some of the large losses realized over the past three years, Mexico's banks are still in need of additional infusions. Much of the system is fragile and the incentives facing a number of banks overly dependent on inter-bank funding remain in question. The Mexican Authorities' reliance on open-bank assistance and the form such assistance has taken has created unintended negative incentives and generated substantial, but as yet indeterminate, costs. The disposition of intervened banks, while initially proceeding at a slower than expected pace, accelerated toward the end of 1996. However, these interventions and the open-bank assistance program have created an overhang of assets in Government hands that has balloonedL to over US$42 billion. As discussed above, a comprehensive strategy for resolving these assets still needs to be developed along with a strategy for financing the additional fiscal costs that are likely to be revealed in the process. 15 14 Because of the unfavorable environment for increasing the level of performing assets, banks have concentrated on reducing their overall cost of funds, which historically was dependent upon high- cost interbank borrowings. Of late, this has materialized in a heightened competition for lower cost deposit funding and the opening of new bank branches. While the overall level of deposits in the banking system has likely increased as new deposits have been attracted to the banking system by increased services and higher rates paid on savings, there has also been redistribution of deposits among existing banks as depositors have chased after higher rates of interest. 15 Estimates of the ultimate costs of the steps taken by the Mexican Authorities to date are necessarily imprecise because of the continued lack of reliable data on the condition and performance of assets acquired by FOBAPROA under the loan purchase/recapitalization program, bank intervention and the assistance provided in bank mergers ancl acquisitions. Most private analysts expect the costs to exceed current Government estimates. See Standard & Poor's Credit Week, "Mexican Banks: The 12% Solution," August 7, 1996, pp. 9-12.  18 B. Improved Accounting Standards; Prudential Regulations; Strengthened Supervision; Prevention of Recurrence 13. The Bank's work in preparing the FTAL and the evident shortage of reliable information available to the Authorities and the markets at the onset of the crisis made it clear to Bank staff that Mexico needed to bring its bank accounting standards and prudential regulations more into line with OECD standards and further strengthen the supervisory function. The need for improvements in these areas was recognized by the new leadership of the CNBV in 1994. Greater transparency, more rational prudential norms and a stronger CNBV (along with increased exposure to market discipline through a roll-back of deposit insurance) would go a long way toward putting in place checks and balances to minimize the prospects for a recurrence of a banking crisis. The events following the peso devaluation provided impetus to accelerate needed improvements and program supported by the FSRL has clearly moved Mexico more quickly toward international standards in all three areas. 1. Improved Accounting Standards 14. Conditioning the release of the second tranche of the FSRL on new accounting rules for banks compatible with international standards was perhaps the greatest contribution of the FSRL to modernizing Mexico's banking system and minimizing the likelihood of a recurrence of a generalized crisis. During preparation of the FSRL, the Bank staff made it clear that imposition of new rules should not be delayed further and had to be a central part of the operation. Accounting rules and practices at the outbreak of the crisis obscured the true condition of banks, making effective supervision very difficult, and precluded market analysts from reaching informed judgments. While implementation of new rules by the banks is stretching their institutional capacity, the Bank staff and the Authorities agree that they are a necessary condition for introducing greater market discipline, maintaining public confidence and improving effective supervision. 15. The new rules, however, continue to differ from U.S. (and OECD) accounting practices, principally in the following areas: (i) treatment of restructured loans; (ii) inflation accounting; and (iii) the inclusion of transitional measures to permit banks to gradually recognize the negative effects of the new accounting on certain portfolios (in particular, their mortgage portfolios)."6 Although all banks have now released accounts 16 Under the transitional provisions, previously accrued but unpaid income derived from the capitalization of interest on past due loans will be treated as principal if recorded as a result of a loan restructuring before year-end 1996 and banks can continue to accrue income on 30% of the total past due mortgage loans for an initial six month period in anticipation that they will be able to encourage borrowers to restructure their mortgage loans under existing government support programs and thereby eventually restore the loan to current status. However, mortgage loans that either were past due at year-end 1996 or become past due in 1997 will require a reserve of 35%, of which 10% must  19 under the new rules, questions remaini about the accuracy and completeness of application and comparability across banks. Hovvever, while differences from US accounting remain, and the CNBV and the banks have much work ahead to assure correct and uniform application by banks, there is general recognition in the markets that the new rules are reasonably consistent with international standards and when fully applied by the banks provide the markets with a degree of transparency that was lacking in the past.17 The CNBV is currently preparing a uniform set of accounts for presentation of balance sheet and income statement data that is expected to further improve the comparability of statements under the new rules across institutions. 2. Prudential Regulations 16. The review of existing prudential norms for banks carried out by the CNBV and the Bank in the aftermath of the 1994 devaluation reaffirmed the need to bring regulation in a number of areas into line with international standards. The crisis swiftly made apparent the shortcomings of the Mexican regulatory framework and bank practice in the areas of classification and provisioning. As part of the program of reforms negotiated with the Bank under the FSRL, the CNIBV issued new capital rules for banks and brokerage firms that require financial intermediaries to assign risk-weights to assets taking into account both credit and market risk, incorporating the January 1996 amendment to Basle Capital Accord for market risk. The CNBV has also drafted, but not yet implemented, a new methodology for loan classification based exclusively on the application of a matrix of criteria covering payment history, documented capacity of payment and the borrower's overall financial situation. This would, once put in place, eliminate much of the ambiguity that has permitted banks to maintain unduly generous classifications for underperforming loars. 17. The methods for computing Tier 1 and Tier 2 capital under SHCP regulations remain significantly more generous than international standards and this has contributed to banks continuing to report higher capital ratios. Banks include in Tier 1 capital the full amount of any outstanding subordinated debt convertible within three years issued as part of a recapitalization agreement negotiated before year-end 1996. In the context of the newly adopted inflation accounting, fixed assets are now adjusted according to changes in an official price index and the full amount of the adjustment is reflected in the equity accounts along with the net gain or loss arising from the appreciation of monetary assets and liabilities. Previously, based on appraisals, banks included 80% of the revaluation surplus for fixed assets and subsidiaries as Tier 1 capital with the remainder eligible as Tier 2. Under the Basle Capital Accord, both subordinated debt and revaluation surpluses be recognized immediately and the remaining 25% will be amortized over eight years beginning in 1997. 17 Moody's Investors Service, "Changes in Mexican Bank Accounting Principles - Transparency Triumphs, Yet Regulatory Forbearance is Required to Ease Transition," Global Credit Research, January 1997.  20 are only included in Tier 2. (Further, subordinated debt is subject to a general limit of 50% of Tier 1 capital and revaluation surpluses are subject to a 55% discount.) In the case of several Mexican banks the contribution to capital from one or the other of these items is over 50%. The Mexican Authorities should make it a policy priority to bring standards for composition of Tier 1 and Tier 2 capital up to OECD standards. 3. Strengthened Supervision 18. The CNBV has made considerable progress over the past three years in improving its effectiveness as a supervisor.'8 An institutional development plan was initiated by the CNBV in mid-1994 under the CNBV's new management team, with financing of various components provided under the FTAL.'9 At the time of the December 1994 devaluation, the CNBV had only recently established a basic framework for implementing new policies and procedures for bank inspections and off-site surveillance of the banking system. There had not been sufficient time to fully prepare the CNBV's examiners and analysts for the difficulties that were to follow.20 The CNBV's ability to supervise the banking system at the time of the crisis was further impaired by outmoded and non- transparent accounting and loan classification practices among Mexican banks. 19. Developmental efforts at the CNBV have emphasized supervision instruction through courses, seminars and on-the-job training. From 1995 through mid-year 1997, more than 200 supervisory staff from the CNBV and the Banco de Mexico participated in varied courses and seminars on supervision.2' Such courses range from basic introduction to bank inspections and loan analysis to more advanced and specialized supervision subjects such as capital markets, electronic data processing, risk management and trust companies. Specialized on-the-job training has included participation with U.S. bank supervisors in the examinations of Mexican subsidiaries of U.S. financial institutions and the U.S. subsidiaries and branches of Mexican financial institutions. With the assistance of six seconded U.S. bank examiners (Federal Reserve, Federal Deposit Insurance Corporation and Office of the Comptroller of the Currency) financed 18 A summary of the principal financial sector legislation and the powers of the financial sector supervisors was included in Annex VII of the Memorandum of the President. 19 The Government reorganized the National Banking Commission (Comisi6n Nacional Bancaria - CNB) in April 1994 and installed new management. In August 1994, FSD conducted a retreat for high level staff of the CNB to discuss regulatory and supervisory issues with experienced US and Canadian regulators. Throughout that summer and the following fall, the Bank staff worked closely with the new management of the CNB (as well as with the Securities, Insurance and Pensions Commissions) to identify those areas of policy, regulation and supervision that were most in need of attention and to work out a plan of action and assistance. In March 1995, the CNB was merged with the principal securities regulator to form the National Banking and Securities Commission (CNBV). 20 Accordingly, at the same time as the FSRL was under preparation, a supplemental FTAL was prepared to provide the CNBV with additional assistance in key areas of supervision. 21 The Bank's participation has included the annual Seminar on Banking Supervision in Mexico sponsored jointly with the U.S. Federal Reserve and CEMLA.  21 under the supplemental FTAL, the CNBV conducted detailed CAMEL-style inspections of all 18 privatized banks by year-end 1995 thus initiating a program of annual examinations of each major bank.22 Executive summaries of each inspection report were transmitted to and reviewed by the Bank. The principal revelations emerging from these inspections, applicable to most banks, were: (i) deficiencies in internal controls; (ii) inadequate credit analysis and loan processing procedures and poor documentation; and (iii) poor risk and liquidity management. 20. In August 1995, the Banco de Mexico in consultation with the CNBV issued minimum standards required of financial institutions prior to their participation in Mexico's various foreign exchange, money and capital markets. The certification process involved outside accounting firms who participated to ensure the development of a proper internal control environment. Concurrent with this effort, the CNBV instituted a new program for assessing risk management practices at Mexican financial institutions associated with trading and other market activities. The CNBV routinely collects risk management reports and position information from the banks in order to evaluate overall risk exposures. Moreover, as part of the CNBV's permanent supervision program, it regularly meets with risk managers with the objective of discussing regulatory issues, commonly faced problems with the implementation of their risk models, availability of information and instrument valuations among others. Evaluation. The FSRL can be scored as satisfactory on achievement of its objectives in respect of accounting standards, prudential regulations and supervision. When fully implemented and applied, the new accounting standards will be a big step forward for transparency. Prudential regulation has also been importantly strengthened and those areas that require further strengthening (in particular, classification and provisioning rules and capital composilion) have been identified as areas to be addressed as soon as practicable. Altough shortcomings in the capacity of the CNBV remain, the improvements from the state of the institution in 1994 are substantial. The FSRL can be scored positively on helping the CNBV to establish procedures to improve internal controls in banks. The information on the condition of banks now available to the CNBV is significantly greater, and the ability to the CNBV to analyze this data is much improved. However, evidence suggests that in practice there remain areas where internal controls of banks are far from adequate. Since systems similar to those necessary for effective internal controls are necessary for accurate reporting to the supervisor, the CNBV needs to take steps to see that the banking community makes substantial improvements in this area. 22 CAMEL - Evaluation of Capital, Asset quality, Management, Earnings and Liquidity  22 C. Improved Discipline in the Provision of Liquidity by the Banco de Mexico 21. The objectives of the FSRL in respect of liquidity provision by the Banco de Mexico were substantially achieved prior to Board approval of the operation. As recounted in the Memorandum of the President,23 the Banco de Mexico in March 1995 began requiring pricing of central bank intra-day credit on a real time basis, pledging of collateral by banks that maintain large and frequent overdrafts with the central bank and debit caps on the portion of daylight overdrafts not collateralized. The Banco de Mexico also established the Amplified Use Electronic Payments System (Sistema de Pagos Electr6nicos de Uso Ampliado-SPEUA), a large value payments system for Mexican commercial banks and their foreign correspondents. This system employs multilateral netting and bilateral credit limits, which together substantially reduced the risks of contagion in the payments system. These efforts substantially reduced the systemic risk to the banking system arising from inter-bank transactions. D. Initiate Reforms in Development Bank Lending, Accounting Practices and Regulation of Financial Groups, and the Deposit Insurance System 1. Development Bank Lending 22. As a condition of second tranche release, experienced consultants carried out detailed studies of risk-based exposure limits, eligibility criteria and interest rate policy for lending by Mexico's principal development banks to private commercial banks. The results of these studies confirmed the need to increase the transparency, efficiency and risk sensitivity of transactions between the development and commercial banks through the establishment of appropriate criteria for eligibility, exposure limits and market sensitive interest rates linked to risk. 23. Based in part on these studies (and on other work conducted with the assistance of the IDB), reforms were in fact initiated in several areas. Prior to second tranche release, SHCP presented an action plan to the Bank and the IDB for reforms in the areas of eligibility and exposure limits. The following steps have already been carried out at Mexico's largest development banks (Nafin and Bancomext): (i) removal of limits on the rate of interest that may be charged by first tier commercial banks to the final customer in domestic currency (Nafin and Bancomext); (ii) implementation of differing interest rates based on the degree of intermediary risk (Bancomext); (iii) consolidation of six second tier lending products into one single program, substituting the rediscounting mechanism with a secured line of credit (Nafin); (iv) establishment of a rating system for intermediary banks (Bancomext); and (v) establishment of risk-based exposure limits for each intermediary bank (Nafin). SHCP has undertaken to complete the remainder of the 23 Memorandum of the President, p. 9.  23 action plan, principally the establishment of a risk analysis system and implementation of global exposure limits (taking into account loans to a single intermediary from all development banks), by 1998. 24. Notwithstanding the above, the ultimate economic utility of these and other changes in development bank lending policy outside the context of a coherent Government strategy for development banking is questionable. A well-defined strategy for the role of development banks within the present system of private commercial banks and increasing foreign participation has not been articulated. While conception of such a strategy was well beyond the parameters of the operation under review, it remains essential to fully coherent Government policies in the financial sector. 2. Accounting Practices and Regulation of Financial Groups 25. In March 1996, as required under the FSRL, the CNBV issued revised accounting principles to be applied by brokerage firms beginning January 1, 1997. Later that year, the CNBV issued revised accounting principles for other non-bank financial institutions under its jurisdiction. These new principles, generally consistent with accepted international standards and with the corresponding provisions of the new bank accounting rules, permit better consolidation of financial group accounting. (Insurance company accounting has not yet been revised to accommodate the changes in accounting for other financial sector entities, most notably in respect of inflation accounting.) Similarly, in June 1996 the CNBV issued new prudential regulations with respect to capital adequacy standards for brokerage firms. Such rules are substantially similar to those applicable to banks, applying the same methodology for credit and market risk. As in the case of banks, the new standards represent an important improvement from previous regulations and once fully implemented and correctly and consistently applied will greatly facilitate consolidated supervision. 26. The package of financial sector legislation approved by the Mexican Congress in February 1995 importantly extended the CNBV's powers of administrative and management intervention of banks and brokerage companies to financial holding companies as well. The CNBV may now immediately take control of a financial holding company if in its judgment there are irregularities affecting its stability or solvency which put in danger the interests of the public, or its creditors. Prior to the amendments, the CNBV's powers over holding companies were more limited, critically complicating the resolution of banks. The new legislation has facilitated the seizure and disposition of both banks and non-bank subsidiaries of financial holding companies. 3. Deposit Insurance System 27. The FSRL committed the Government to begin to formulate a plan to reduce the Government guarantee of bank and brokerage liabilities. Mexico's system is extreme by international standards, providing an implicit Govermment guarantee of all unsubordinated bank liabilities. Even interbank loans and traded bonds are covered by  24 the FOBAPROA guarantee. As discussed in Part IV, A, 1, e, it was not anticipated that an actual roll-back of Mexico's broad liability coverage system would occur during the period before or soon following the second tranche release. Rather, implementation was expected to be a key component of the expected follow-up World Bank operation in the sector. The objective of the deposit insurance system component of the FSRL was more limited-to identify the relevant issues and commit the Government to the principle of a phased roll-back. 28. The Savings Protection Report prepared by the Banco de M6xico reviewed alternative savings protection systems and examined the options for reforming the current Mexican system. In preparing the report, SHCP, Banco de Mexico, and the CNBV conducted an open dialogue with the World Bank. The Bank organized a conference at the Banco de Mexico on reforming Mexico's savings protection program that included presentations from acknowledged international experts and that was attend by high-level officials from all three Government agencies. The major point of disagreement was the speed with which reform could be carried out. The Bank staff argued that Mexico could and should move faster to reduce the extent and level of coverage. Notwithstanding differences on the timing of implementation, Mexico's dialogue with the Bank has clearly and publicly placed reform of Mexico's bank liability protection scheme on the policy agenda. SHCP, Banco de Mexico and the CNBV are all now publicly committed to a gradual roll-back. The President of Mexico noted this in his address to the 1996 meeting of the Mexican Bankers Association. However, to date the work of the Authorities and the assistance provided by the Bank have not yielded concrete proposals for implementation. Evaluation. Progress toward reform of development bank lending practices, better accounting and regulation of financial groups and a more economically rational system for deposit insurance has clearly been made under the program supported by the FSRL. Accordingly, there has been generally satisfactory achievement of this objective. However, greater progress was anticipated to have followed, albeit after second tranche release, in the areas of development banks lending practices and particularly with respect to the pivotal roll-back of FOBAPROA liability coverage.  25 PART V. MAJOR FACTORS AFFECTING THE PROJECT A. Macroeconomic Environment 29. The President's Report for the FSRL clearly recognized that one of the principal risks to the success of this operation would be the failure to maintain a satisfactory macroeconomic framework and a continuation of high inflation and interest rates.24 30. The table below describes the macroeconomic variables most often considered relevant in influencing the solvency of the banking system,25 together with the evolution of those variables during 1995-96. Table 2: Macroeconomic Variables and Bank Insolvency Expected Effect Macro-Variables on Bank Solvency Observed Values _________=___ 1994 1995 1996 1997 Real GDP growth (%) + 4.4 -6.2 5.2 7.0 CPI Inflation (Dec/Dec; %) + 7.1 52.0 27.7 15.7 Interest rates-- nominal (28-day cetes;ann. %) + or - 14.1 48.4 31.4 19.8 --real + or - 8.2 10.0 9 2 6.3 Bolsa US$ Index (% change prev. Dec.) + -- -22.9 18.0 52.1 Terms of Trade (Px/Pm, index) + 104 112 114 n.a. BoP Current Account (% of GDP) + or - -7.4 -0.6 -0.6 -1.8 Exchange rate change (%ave chge) - nominal -7.7 -47.4 -15.8 -3.8 -real -3.8 -33.2 10.2 13.7 Fiscal balance (% of GDP) + -0.4 0.0 0.0 -0.8 31. From the right-hand column of Table 2 it is clear that most macroeconomic variables deteriorated dramatically in 1995 (except the Terms of Trade and Fiscal Balances). This would certainly have undermined the soundness of the banking system. The current account deficit of the balance of payments also improved significantly from the values in preceding years. While part of this improvement reflected greater export dynamism, it also reflected considerable import compression brought on by the economic recession as well as a drastic reductiorn in external borrowing opportunities. The latter would generally be expected to increase the risk of bank insolvency. 32. The deterioration in the econornic situation in 1995 was more profound than was anticipated at the time that the FSRL was prepared. However, when the economic situation turned around in 1996 and 1997, growth was more robust than expected. According to the FSRL President's Report, real GDP was projected to decline by 4.8 percent in 1995 and then recover gradually, by 2 percent, in 1996. The recession turned 24 Memorandum of the President, pp. iv, 36. 25 See Lindgren, Carl-Johann, Gillian Garcia and Matthew I. Saal, Bank Soundness and Macroeconomic Policy, IMF, 1996; page 15  26 out to be much deeper in 1995, with GDP falling by 6.9 percent (or 6.2% under the 1993- based accounts), while the recovery in 1996 and 1997 also was much stronger than predicted, yielding growth rates of 5.1 and 7.0 percent.26 33. The greater-than-anticipated deterioration of the economy in 1995 was clearly an important factor discouraging the Mexican Authorities from taking more aggressive actions in restructuring the banking system while it encouraged the expansion of debtor support programs. The Authorities may have felt that they had already done as much as they prudently could. There was also a perception that any disruption caused by the closure of banks and depreciation of asset prices from an accelerated sale of the intervened banks' assets could have exacerbated the economic contraction. However, even through the economy began to exhibit clear signs of expansion by mid-1996, and the resolution of intervened banks began to be accelerated, new debtor-support programs were introduced. 34. The economic recovery in 1996 and 1997 does not appear to have had the expected positive effect on bank portfolios. Past dues continued to increase, albeit at a declining rate. In 1996, this could be partly explained by the initially unbalanced nature of the recovery (concentrated in export industries and largely financed from abroad) and continued low real incomes. Although real exports grew by 19 percent in 1996, real GDP grew by only 5.2%, which means that economic activity not oriented toward exports grew much more slowly than GDP. Also, domestic incomes, which in 1995 suffered an 8.1 percent decline in per-capita terms, did not recover to their pre-crisis levels until about mid- 1997. As the economic recovery continued into 1997, it became much more broadly based, encompassing the construction and service sectors in addition to exports, yet bank portfolios performance apparently failed to improve concomitantly. It should be noted that banks increased significantly their provisions during 1997. 35. The evolution of inflation turned out to be less positive than initially envisaged during negotiation of the FSRL. Inflation was clearly expected to increase in 1995, in response to the December 1994 depreciation of the peso. In October/November of 1995, a bout of financial market turbulence led to a depreciation of the peso significantly beyond the projected values. This unanticipated depreciation partly reflected a slow reaction of the monetary authorities to several adverse shocks. The financial markets eventually stabilized at the end of the year after the authorities tightened their monetary stance and the Banco de Mexico temporarily intervened in the exchange market. 36. The deterioration of several macroeconomic variables in 1995 clearly impinged on the soundness of the banks, primarily through the insolvency of their borrowers. Most important in this regard appear to be: (i) the decline in domestic demand reflected in the 26 The GDP growth figures in the FSRL report are based on the 1980-base National Income Accounts. They tended to be about 0.7 percentage points lower than the now more commonly used 1993-base accounts, which are shown in Table 2.  27 fall of GDP; (ii) the increase of inflation and interest rates (which are jointly determined); and (iii) the rapid adjustment of exchange rates. The rapid exchange rate depreciation in December 1994 and in 1995 had a detrimental effect on some banks, which had open foreign exchange positions. While significant, this source of loss does not appear to have been a major contributor to bank distress. Rather, a more important consequence of the real exchange rate adjustment that resulted from these nominal adjustments was that it drastically altered the relative price of tradables versus nontradable goods. The period of rapid expansion of the banking system, 1991-1993, was accompanied by a significant appreciation of the real exchange rate, encouraging the investment of resources into the production of non-tradable goods. The real depreciation in 1995 rendered many of those activities unprofitable. Borrowers whose investments were tied up in these activities, together with the banks that extended credit to them, suffered dearly as a consequence. 37. There also appears to have been an important linkage in the other direction. That is, the weakness of the banking system contributed significantly to the deterioration of the economy-independent of the banking, sector's share of GDP. The macroeconomic deterioration in 1995 may have exacerbated solvency problems in the banking sector, but those problems appear to have preceded the onset of the economic crisis. (The proportion of non-performing/overdue loans in the banking system was already increasing at a fast clip in 1994, in spite of stable inflation and rising GDP.) Furthermore, the weak state of the banking system most likely inhibited the authorities from taking more concerted policy measures to contain the economic crisis. This was the case both during the onset of the devaluation in December 1994 as well as during the period of financial market turbulence in October/November 1995. In both cases, more concerted action to stem the depreciation of the peso would have required a tighter monetary contraction, which would have translated into higher real interest rates, and greater distress among borrowers. In other words, the fear of dLamaging the banks further may have led the authorities to adopt less decisive macroeconomic countermeasures, thereby permitting a greater macroeconomic deterioration than would otherwise have been the case. Eventually, more stringent counter-measures were taken and interest rates rose accordingly. It is not clear that the banking system benefited much from the policy delay. 38. In sum, the economic contraction that occurred in the aftermath of the December 1994 devaluation proved to be much deeper than expected at the time of preparation of the FSRL, while the recovery in 1996 and 1997, though much stronger than expected, did not lead to a proportional improvement in the quality of bank portfolios. As discussed above, these portfolios continued to deteriorate, albeit at a declining rate, throughout 1996 and 1997. This was initially expliained by the uneven nature of the recovery and a failure of real incomes to return to 1994 levels. The rapid and better balanced growth performance in 1997, however, cast doubts on that explanation, giving more credibility to the hypothesis of lagging real incomes and unwillingness to repay loans by bank borrowers. Instead, several other factors are likely to have been more important in limiting the positive effect on loan performance of the economic upturn: (i) the initially unbalanced state of the banking system and poor loan origination practices; (ii) the  28 existing overhang of unresolved debts (largely held by FOBAPROA) and the high loan to value ratios characteristic of many segments of bank portfolios; and (iii) continued availability of foreign financing on attractive terms to many of the industry's best clients (particularly exporters). B. Socio-Political Environment 39. The crisis in the banking sector occurred at a very difficult socio-political juncture. The incoming administration was in office less than a month when the crisis erupted. Important segments of the population (particularly debtors and the financial community) placed responsibility for the problems of the banking system with the mismanagement of monetary, exchange rate and economic policy by the outgoing and incoming administrations. Adversely affected sectors consequently looked primarily to the Government for relief. An organized debtor movement that existed prior to the devaluation ("El Barz6n" or "the yoke") became an active force in the country. It promoted a "culture of non-payment," carried out regular and well-publicized demonstrations and organized payment boycotts-taking advantage of the legal and practical difficulties of enforcing obligations in Mexico. The ADE program, which provided immediate relief to many troubled debtors beginning in August 1995, successfully stemmed the growing popularity of this movement. However, El Barz6n remained an active force, especially in the Federal District and major urban centers and it still commands a considerable amount of media attention. 40. Despite economic difficulties, the President of Mexico retained his commitment to opening up the political system. Predictably, the increasing political openness during a time of economic distress brought with it a resurgent political opposition and more independent behavior on the part of Congress (the ruling party lost its 68-year majority in the lower house in the July 1997 elections). There has been heated debate of financial sector policy and representatives of the executive have been periodically called to report to Congress on conditions in the financial sector. El Barz6n maintained a high profile in the run-up to the July Congressional elections and the election of the Governor of the Federal District (which went to a 1994 opposition Presidential candidate). These sorts of developments are likely to continue to challenge the maneuverability of policymakers.  29 PART VI. PROJECT SUSTAINABILITY 41. It was implicit in the objectives of the FSRL that the program supported by the operation should place the Mexican system on a reasonably clear and sustainable path to full eventual recovery. It is difficult ait this time to conclude that the system is unquestionably on such a trajectory. While the surviving banks in the system are clearly less vulnerable today than they were in early 1995, the system as a whole remains fragile. There is little prospect of a sudden retrogression, absent another large exogenous shock. However, the longer the current fragility persists the more adversely affected the system will be by exogenous shocks should they occur. 42. Some, partial, elements of a complete strategy for full recovery of the financial sector are clearly already in place. Butt these, by themselves, appear insufficient to take the system the rest of the way towards full solvency and soundness. Accordingly, project sustainability should be regarded as uncertain. The principal concerns about project sustainability can be summarized as follows: * Condition of the System. The FOIBAPROA loan purchase/recapitalization program attracted significant but still insufficient amounts of new foreign and domestic capital. UDI restructured loans (particular]y mortgage credits) appear to have reached loan to value ratios that make repayment on near-contractual terms unlikely in most cases. Banks appears to be highly dependent on accrued income from government securities issued under the UDI and FOBAPROA loan purchase/recapitalization programs. This constrains real bank liquidity and under the worst of circumstances may lead to phantom income (and capital). * Incentive Framework. The continued presence of excessive bank liability protection and the relative shortage of crediblie financial data on bank portfolios continues to limit market discipline in the system, permitting weak banks to remain in operation. The FOBAPROA loan purchase/recapitalization program appears to have missed some opportunities to eliminate institutions that are not viable. Loan growth appears greatest in the weakest banks, which may indicate either capitalization of large amounts of unpaid interest or possibly imprudent growth in new lending. * Overhang of Unrestructured Assets. The likely pace of dispositions of FOBAPROA assets under the current strategy is insufficient to remove much of the overhang and an effective strategy for effecting rnore rapid restructuring of FOBAPROA loans by banks has yet to be formulated. Failure to work out these loans will impede economic recovery and reactivation of the intermediation function of banks. Economic recovery and growth is unlikely, by itself, to sufficiently improve portfolio quality in the absence of a workout of the existing debt overhang and the prospect of new lending.  30 Unfinished Agenda/Fiscal Costs. The Government's efforts thus far have not effected the kind of fundamental restructuring of the financial system and realignment of incentives necessary for self-sustaining growth in the financial sector on the basis of well-capitalized institutions and sound practices. The remaining issues not addressed in the legal/regulatory infrastructure for banking transactions, particularly in the areas of bankruptcy and secured transactions, raise doubts that new lending, when it occurs, will be on any firmer bases that prior to the crisis. Finally, the fiscal costs (and redistribution effects) of the bank rescue, put off and to some extent obscured by the strategy followed to date, are likely to become more apparent as time passes, and a well-articulated strategy for financing these costs will have to be developed. 43. Addressing these issues will require substantial human and financial resources. But it should be possible to make reasonably rapid progress to fill in the gaps of a truly sustainable strategy. Identifying such gaps and determining where the World Bank can assist should be the goal of continuing collaboration with the Authorities.  31 PART VII. BANK PERFORMANCE A. Preparation and Negotiation 44. The Bank's performance in preparing the FSRL should receive high marks. The operation was prepared in record time. It demonstrated the Bank's ability to effectively respond to the needs of the client and permitted the Bank to remain engaged in the resolution of a crisis whose implications went well beyond Mexico. Prior to the banking crisis, the Bank's dialogue with the Mexican Government in the financial sector was limited. It revolved around the preparation of the FTAL, an operation aimed at improving the legal/regulatory environment and institutional capacity of the banking, securities, insurance and pension fund regulators. The IFC, however, had prepared a partial analysis of the overall banking system that showed that the system was under- provisioned and that capital was therefore overstated. With the assistance of staff from other regions, FSD, Development Economics (DEC) and the IFC, and in cooperation with the IDB, the Region's team was able to build on the FTAL and to work with the Authorities to design and prepare an operation that identified and addressed the key policy issues that are still the focus of lboth the Government's and the Bank's efforts in the sector. B. Supervision 45. In the immediate aftermath of the Board's approval of the FSRL, the Bank staff's supervision efforts were fruitful. Although CNBV staff occasionally appeared reluctant to transmit information in a timely fashion (particularly if such information indicated a worsening of conditions), this was attributed to concerns for confidentiality and the need for complicated internal clearances before the CNBV staff could discuss sensitive information with outsiders. Relations between the Task Manager for the operation and the highest level officials at the CNBV were cordial and reasonably open, particularly given the crisis conditions. The Bank staff remained in regular contact with the CNBV officials in charge of conducting the initial round of Bank examinations and overseeing the diagnostic studies of troubled banks. However, citing confidentiality concerns, the CNBV would not share the full results of the examinations and studies with the Bank. Nevertheless, between the summary information received directly from the CNBV and from meetings with CNBV supervisory staff and the seconded U.S. bank examiners, the Bank staff felt confident that adequate progress was being made. 46. In the early stages of supervision, the Bank assisted the Authorities in four important areas: (i) exposing the CNB'V to experts in banking crisis management with experiences in other countries; (ii) highlighting the need for greater coordination among responsible agencies (CNBV, SHCP and the Banco de Mexico); (iii) providing advice to the Banco de Mexico on the roll-back of bank liability protection; and (iv) developing a strategy with SHCP for possible future Bank assistance in the financial sector, including  32 in the areas of pensions and non-bank financial services. The Bank was able to bring in experts with experience in the United States, Spanish, Chilean and Swedish banking crises to provide advice to the Authorities. Some of these experts were eventually contracted directly by the Mexican Government.27 At the Bank's urging, a coordinating committee of high level SHCP, CNBV, Banco de Mexico and FOBAPROA officials was established to assure more efficient sharing of information and coordination of execution of tasks. A DEC-organized seminar and periodic consultations between Bank staff and Banco de Mexico officials provided critical inputs into that institution's understanding of the issues and options for roll-back of deposit insurance.28 FSRL supervision efforts also led to much of the work on preparation of the Contractual Savings Reform Loan-which addressed the low domestic savings rate, a root cause of the financial sector's fragility. Finally, supervision of the FSRL was also the impetus for the development with SHCP of a strategy matrix for financial sector reforms and the role of the Bank in providing assistance to the Mexican Government. This work was spearheaded by the Task Manager for the FSRL who also served as task manager for the pension fund operation. 47. However, important shortcomings in the effectiveness of the Bank's supervision efforts arose in the period after October/November 1995 (when a further "mini- devaluation" occurred, putting additional strains on the banking system and the Authorities). Supervision of the FSRL absorbed a great deal of the time of the limited number of Region staff with financial sector experience. As recounted in the DEC report cited in Part 1, prior to the crisis insufficient resources had been allocated to financial sector work in Mexico. The financial economist in charge of preparation of the FTAL left the region towards the end of 1995 and there was a great deal of general staff turnover in the Mexico department as well. Accordingly, there were few experienced staff available to conduct or supervise production of detailed economic and sector work and the first such post-devaluation work was not produced until mid- 1996.29 While the Region received important assistance from FSD and experienced outside consultants, these were rarely available on a full time basis. The CNBV commented on several occasions that World Bank staff turnover and the Bank's use of consultants on a short- tern basis limited the Bank's potential contribution. 27 Most notable of these were the U.S. bank supervisors (OCC, Federal Reserve and FDIC) referred to above and Mr. Arist6bulo de Juan, forrner Head of Bank Supervision for the Bank of Spain. Mr. de Juan was hired in early 1997 by the CNBV as a special consultant. 28 Bank staff also made significant contributions to the study of asset-backed securities required as a condition of second tranche release, although Bank input was not sought during preparation of the study of the secured transactions. 29 Hemant Shah, Mike Lubrano and Laura Mecagni, "Mexico: Role of Financial Markets in Promoting Savings: Regulatory and Development Agenda," July 10, 1996. This report was prepared as part of the 1996 Country Economic Memorandum on Savings and Growth.  33 48. The DEC study cited in Part I notes some shortcomings in coordination between the Region and the central units (FSD and DEC).30 Most of this occurred in the aftermath of the October/November 1995 mini-devaluation. At that time, the Region requested additional assistance from FSD to reexamine the effect of the deeper than expected recession on the condition of the Mexican banking system. A mission was conducted in January 1996 with the participation of a number of outside experts. However, lack of access to detailed and reliable information from the Mexican Authorities limited the ultimate utility of this exercise. As a result, a divergence of opinion developed over the true condition of the Mexican banking sector and the Bank's proper response. These views were debated in a series of meetings involving all areas of the Bank working on the Mexican financial sector operations. As conditions in Mexico stabilized and as the dialogue between the Bank and the Authorities became more clearly channeled through the Finance Sector Leader in the decentralized country department, it became easier to coordinate the activities of the different areas of the Bank involved in the financial sector, build consensus and make more effective and efficient use of their varied expertise. The progress in developing the role of the EBank-wide professional networks has also been instrumental in this regard. C. Inputs to Government Response; Policy Dialogue 49. Prior to the crisis, the Bank's efforts in the Mexican financial sector were directed toward assisting the Authorities to impirove the regulatory and supervisory framework in connection with designing and implementing the FTAL. This useful work did not require the Authorities to share with Bank staff sensitive information on the condition of individual banks or the system as a whole. During the preparation of the FSRL a very productive relationship continued, with Bank staff and consultants contributing importantly to the Government's efforts to gather accurate information on the condition of banks, develop its strategy for bank resolution and implement other key policies described in the Memorandum of the President for the operation. 50. After approval of the FSRL, the Bank staff regularly met with its Mexican counterparts to discuss the progress of the various programs and actions implemented to address the crisis. The Bank staff provided periodic inputs on the longer-term goals for the financial sector, including roll-back of FOBAPROA liability coverage, greater transparency and disclosure, consolidation of the system, resolution of intervened banks and disposition of assets and steps to encourage greater reliance on market supervision. The principal differences between the Bank staff and the Mexican Authorities during the course of supervision revolved around the aggressiveness and timing of implementation of policy actions: * Conditions for Open-Bank Assistance/Bank Intervention. The Bank staff advocated ordinarily conditioning all open-bank assistance on changes in management. It also 30 DEC, p. iv.  34 encouraged earlier interventions of troubled banks on the grounds that their continued operation under the existing incentive framework would result in greater long-term losses and fiscal costs. * Attention to Negative Incentives. The Bank staff favored early ternination of bank support programs and debtor relief schemes, even if this meant larger up front costs. It was felt that repeated recourse to the FOBAPROA loan purchase/recapitalization program and gradual expansion of the debtor relief programs would encourage both banks and borrowers to expect future handouts. * Disposition of Intervened Banks. In 1995 and 1996, the Bank staff argued that the Government should accelerate its plans to sell intervened banks and their assets, even if that require clean-up of their portfolios and recognition of heavy losses. Two banks intervened in 1994, just prior to the onset of the crisis, remained under government control for more than two years. * Liability Protection. Early announcement of the intention to roll-back Mexico's broad liability protection scheme and development of a policy to implement such a measure were promoted by the Bank as critical steps to introducing market discipline to the banking system. * Disclosure. The Bank staff encouraged the Mexican Authorities to be more forthcoming with information on the banking system and to require banks to make greater public disclosure of their true condition. This would promote the beginnings of market discipline and over time increase the credibility of the Authorities and the reliability of information in the markets. * Capital Rules and Regulatory Forbearance. While applauding the implementation of better prudential regulation and accounting standards, the Bank argued that strict and short deadlines should be established for regulatory forbearance and that the effects of such forbearance on bank financial statement should be as transparent as possible. The Bank supported bringing the Mexican definitions of Tier 1 and Tier 2 capital into line with international (OECD) standards. * Financial Infrastructure. The Bank staff encouraged the Mexican Authorities to give greater priority to legal, regulatory and institutional changes needed to facilitate sound credit transactions and debt restructuring. 51. The productivity of the dialogue between Mexico and the Bank on these points was complicated by the reluctance, at times, of the Mexican Authorities to share non- public information on a timely basis or to consult with the Bank staff prior to taking major steps to address the crisis. This reluctance prevented the Bank from serving as a full partner in the efforts supported by the FSRL. It greatly complicated the Bank staff's efforts to make its own assessment of the condition of the banking system and provide  35 useful inputs into the decision making process. The Mexican Authorities' reticence to share information on a timely basis appears to have involved: (i) concern for confidentiality and leaking of sensitive information; (ii) internal and inter-agency differences of opinion over the accuracy of data; (iii) socio-economic conditions that limited the options available to the Authorities and a reluctance to communicate this to outsiders; (iv) a perception that the Bank's advice would not be timely or sensitive to practical realities; and (v) concern that too much openness would translate into additional conditionality in future World Bank lending. Whatever the causes, this situation at times left the Bank staff in the position of communicating its comments on policy actions to the Mexican Authorities only after such actions had already been announced to the public. This was sometimes apparently perceived by the Authorities as unnecessary, unproductive and non-collegial criticism. 52. Two additional factors complicated efforts to maintain a fruitful dialogue. First, the World Bank's financial contribution to the Mexican Rescue Package was much smaller than that of the IMF or the United States government, but came with what appeared to some of the Bank's counterparts as much more taxing conditions. Also, those Mexican officials most often charged with interfacing with the Bank were the overworked staff of the CNBV. This agency was not the immediate recipient of the resources disbursed under the FSRL. Given their time constraints, CNBV staff may simply have decided that the Bank was not a priority.   37 PART VIII. BORROWER PERFORMANCE 53. As discussed earlier, the Mexican Authorities responsible for the financial sector can take credit for a number of critical steps whose successful achievements was by no means a certainty at the time the FSRL was under preparation: * A collapse of the banking system was prevented at a time of economic contraction and great uncertainty.3" * Stabilization of the banking system was accomplished through mechanisms that avoided large current fiscal outlays and that attracted a significant sum of additional private capital to the system. * The Authorities, and the CNBV in particular, were able to greatly improve capacity to monitor the condition of individual banks and analyze the strengths and vulnerabilities of individual institutions and the system as a whole. * The CNBV, SHCP and the Banco de Mexico were effective in getting needed supervisory powers and foreign investment legislation passed by Congress, strengthening accounting and prudential norms and carrying out the other elements of the Financial Sector Restructuring Program. * Although progress was initially slower than expected in resolving intervened banks and disposing of assets that have come into the hands of the Government, important strides in restructuring, liquidating and disposing of intervened banks have been made in the last year. 54. In assessing borrower performance it needs to be kept in mind that design of the various debtor and bank relief progranis, negotiation of each of the loan purchase/ .recapitalization transactions and the numerous steps to improve accounting, reporting and prudential norms were in fact directed and carried out by a very small number of initially fairly inexperienced staff of the CNBVI, Banco de Mexico and SHCP. 55. Areas where borrower performance could have been better include: (i) coordination among working level staff of different agencies; (ii) assignment of clear responsibility among agencies (or at least departments of agencies) and avoidance of conflict of interest; (iii) transparency; (iv) assessment of longer-term costs and risks; and (v) sensitivity to and avoidance of negative incentives. These last two points have already been addressed in Part VI and Part IV, A, 1, e. 31 Finance Secretary Guillermo Ortiz appeared on the cover of the 1997 "Finance Man of the Year" issue of Latin Finance in recognition of his team's efforts to rescue the Mexican economy and banking system in the aftermnath of the December 1994 devaluation.  38 56. At the outset it appeared to the Bank staff (and perhaps to the Mexican Authorities as well) that although coordination at the highest levels was excellent,32 lower down there was some duplication of effort and failure to share key information. This was an area where the Bank suggested that greater efforts be exerted, finally resulting in the Financial Sector Working Group being organized to address the problem. The Mexican response to the banking crisis was characterized by a high degree of consensus building across the various agencies. The need to develop consensus sometimes led to slowness of response. In part this was due to a certain blurring of responsibility that developed among the responsible agencies and importantly within the CNBV. Shortage of staff particularly burdened the CNBV and created apparent conflicts of responsibility. Staff working on bank supervision also worked on bank relief programs and, more importantly, the terms of the individual recapitalizations. These staff found themselves in the awkward position of having to criticize the practices and accounting of banks under their supervisory authority while also publicizing the successes of the recapitalization program. This mixing-up of functions (supervisor and also rescuer and promoter of the industry) continues to affect the credibility of the CNBV. 57. The continued presence of important gaps in transparency created further credibility problems for the Authorities. An absence of transparent accounting and financial reporting was one of the serious shortcomings of the system going into the crisis, but is now being addressed through implementation of the new accounting rules. However, there remain other key areas where further information could and should be provided to the public: (i) financial condition of intervened banks; (ii) performance of UDI and FOBAPROA portfolios; (iii) composition of regulatory capital of banks; and (iv) details on the terms of some bank rescue packages. A number of factors contributed to this reticence: (i) lack of reliable data; (ii) fear of causing bank runs; (iii) confidentiality rules; and (iv) fear that information would be improperly understood or used by the debtors movement or opponents of the Government. This has led to market discounting of published inforrnation on the Mexican banking system (particularly as regards capitalization) and may be leading to greater market uncertainty than would otherwise be the case.33 32 High level officials of all three agencies served on the ruling committee of FOBAPROA as well as the oversight boards of intervened banks. 33 See Salomon, p. 7, citing continued information disclosure problems with Banamex and Bancomer limiting the ability to conduct accurate assessments of their condition.  39 PART IX. ASSESSMENT OF OUTCOME 58. Overall, the outcome of the FSRL is assessed as unsatisfactory. As discussed above, a number of the operation's major objectives have only partially been achieved. Progress toward accomplishing the longer-term objectives of the FSRL has been slow. Most importantly, the sustainability of the program supported by the operation remains in question. The strategy followed by the Government thus far has prevented a generalized collapse of the financial system and restored a degree of confidence, but at high cost and without removing the major policy (deposit insurance) and structural (transparency and financial infrastructure) impediments to a market-driven and sound financial system. The outlines of a strategy to take the Mexican financial system the rest of the way toward contributing to economic growth on thie basis of efficient and well-capitalized intermediaries have yet to emerge. The Bank and the Mexican Authorities have thus far not reached agreement on the components of the expected follow-on World Bank operation for the financial sector.   41 PART X. LESSONS LEARNED 59. The experience under the FSR1, generated important lessons for future Bank operation in response to a crisis, sector adjustment operations generally and Bank projects in the financial sector specifically. A. Responding to a Crisis * The DEC study cited in Part I highlighted that "The Bank needs to clarify its role in crisis situations." 34 As part of this effort, the Bank should develop special standards for identification, content, preparation, negotiation and supervision of emergency loans like the FSRL. This should be a key goal of the newly-established Special Financial Operations Unit. By definition, such loans do not benefit from the long identification and preparation periods that characterize most other Bank operations. Critical information may be absent and future economic performance may be difficult to predict. Operations carried out umder these exceptional circumstances are a breed apart that merit special attention. * In preparing and supervising operations in sensitive sectors such as the financial system during a crisis, the Bank needs to have intelligence on the socio-political context and how it is likely to impact implementation of policy actions. This will in turn provide input into the design and content of Bank operations in the sector. * The Bank must be realistic about the degree of access to sensitive information it can expect its counterparts to grant it. This requires a thorough understanding of the political, institutional and legal obstacles to information sharing. These should be taken into account in designing, irrmplementing and supervising operations. Indeed, it should figure into the decision of wvhether or not to lend. The Bank staff need to make clear to their counterparts early on in a sensitive operation what types of information will be required to be shared. * The Bank's ability to contribute to the resolution of crises like the one that occurred in Mexico would strongly benefit lrom a well-developed financial sector strategy at the Global and Regional level. Such a pre-agreed strategy within the Bank would give working staff a better vision of where the Bank's efforts should be focused in circumstances similar to those encountered in Mexico. B. Lessons for Operations * The Bank needs to develop appropriate Bank-wide strategies for sector adjustment and other operations whose objectives can be fully achieved only years after 34 DEC, p. v.  42 disbursements are completed. How should the objectives of such operations be stated? What bearing should evaluations have on further efforts in the sector? * Better mechanisms need to be developed for maintaining continuity of staff working on large and complex operations. Undue staff turnover hinders effective supervision, generates inefficiencies, is a burden on counterparts and hurts client confidence in the Bank's ability to offer effective advice. C. Lessons for the Financial Sector * The Mexico experience demonstrates the need for policymakers to be sensitive to the negative incentives created by emergency programs for the financial sector. In particular, it highlights the importance of ordinarily conditioning relief to troubled banks on replacement of existing management. * A country's financial sector infrastructure, in particular the legal/regulatory /institutional framework for making and enforcing credit transactions can be a critical constraint on developing sounder banking systems and resolving banking problems. The experience of Mexico highlights the need for greater Bank effort in this area. * The Bank has comparative advantage in bringing to bear a great deal of knowledge and expertise to address emerging problems in the financial sector. The new professional networks can contribute to better practices for coordinating the resources available from the Regions, the central vice-presidencies, the IFC and outside agencies to make more effective use of their different expertise. * The Bank's efforts in one area of the financial sector need to be placed in the context of an over-all strategy for policy actions in the sector. The financial sector policy matrix developed in the spring of 1996 and revised repeatedly since then has served as an instrument for assisting the Bank and its counterparts to develop a more consistent and coherent strategy in such related areas as banking, pension reform, housing, capital markets development and rural financial markets.  43 PART XI. ANNEXES A. Chronology of Important Events in the Mexican Banking Sector 1994-1997 B. Capital Increases and Purchased Loans by FOBAPROA C. Government Estimate of Fiscal Cost of Banking System Rescue D. Changing Composition of Mexican Banking System, 1994-1998 E. ICR Borrower's Report  Annex A Page 1 of 3 Chronology of Important Events in the Mexican Banking Sector 1994-1997 1994 April Appointment of Eduarcdo Fernandez as new President - beginning of reorganization of CNB. August Seminar conducted for CNB staff by Bank's FSD. September Management intervention of Banco Uni6n. Management intervention of Banca Cremi. December Peso devalues. Eduardo Fernandez narned President of National Securities Commission. 1995 January Opening of FOBAPROA special dollar window. March Banco Oriente placed on special situation. Management intervention of Banpais. Establishment of new minimum provisioning requirements. Establishment of temporary capitalization program (PROCAPTE). Amendments to finance sector legislation. April Establishment of UDIS restructuring programs. Merger of the Banking (CNB) and Securities (CNV) Commissions. May Beginning of FOBAPROA portfolio purchase/recapitalization program. Official announcement of Probursa financial group affiliation to Banco Bilbao Vizcaya (BBV). June After this quarter inforrnation on Banco Obrero ceases to be published due to being considered in special situation FOBAPROA portfolio purchase/recapitalization agreement with Banca Serfin. Management intervention of three credit unions and one savings and loan association. August Announcement of ADE, (immediate debtor relief) program. September After this quarter information on Banco del Centro ceases to be published due to being considered in special situation. FOBAPROA portfolio purchase/recapitalization agreement with Atlantico and Promex. Management intervention of Banco Interestatal. FOBAPROA portfolio purchase/recapitalization agreement with Bancrecer-Banoro. October FOBAPROA portfolio purchase/recapitalization agreement with Banorte. Management intervention of Estrategia Bursatil as well as Leasing and Factoring Estrategia. November Period of extreme currency and interest rate volatility.  Annex A Page 2 of 3 November Announcement of the UDI restructuring program for dollars debts. December FOBAPROA portfolio purchase/recapitalization agreement with Banamex. 1996 January FOBAPROA portfolio purchase/recapitalization agreement with Probursa (II and III). FOBAPROA portfolio purchase/recapitalization agreement with Bancomer. February Announcement of agreement by which Bank of Nova Scotia will acquire 55% of Grupo Financiero Inverlat. March FOBAPROA portfolio purchase/recapitalization agreement with Confia. April FOBAPROA portfolio purchase/recapitalization agreement with Serfin (II). May Announcement of Additional Benefits Program for housing credit debtors (Debtor Relief Program) Management Intervention of Banco Capital. Management Intervention of Casa de Bolsa Bursamex. Management Intervention of Banco del Sureste. June FOBAPROA portfolio purchase/recapitalization agreement with Bancorecer-Banoro (II). FOBAPROA portfolio purchase/recapitalization agreement with Bital (II) July FOBAPROA portfolio purchase/recapitalization agreement with Banorte (II) Announcement of Agreement for the Financing of the Agriculture and Fishing Sector (FINAPE) (Debtor Relief Program). FOBAPROA portfolio purchase/recapitalization agreement with BBV/Probursa. August Signing of Agreement by which BBV/Probursa will acquire the branch networks of Banca Cremi and Banco del Oriente. Announcement of Program of Financial and Development support for the micro, small and medium enterprise (FOPYME) (Debtor Relief Program). FOBAPROA capitalizes Banpais and the bank separates from the Asemex-Banpais Financial Group. October Signing of Agreement by which Banca Promex will acquire the branch network of Banco Uni6n. Announcement of Santander's take over of Invermexico (Banco Mexicano) Financial Group. BBV/Probursa purchases the branch network of Banca Cremi. Banorte acquires 100% of the stock shares of Banco del Centro. November Management intervention of Anahuac Financial Group and CBI Casa de Cambio  Annex A Page 3 of 3 1997 January FOBAPROA portfolio purchase/recapitalization agreements with Banamex and Bancomer in respect of mortgage portfolios. April Announcement of HSBC acquisition of 20% share in Banca Serfin. FOBAPROA has taken about $200 billion pesos which represent about 25% of the bank's portfolio (it includes the guarantee of Mexicano and Serfin, as well as the credits of audited banks). June Citibank announces its interest on acquiring Confia. July The overdue bank's portfolio reaches $100,322 million pesos. August Overdue portfolio increased 2%. VVA merged to FOBAPROA. Only one assets' auction sale completed. Citibank presents its proposal for acquiring 60 - 70% of Grupo Financiero ABACO, Confia's parent. Management intervention of Banca Confia until the conclusion of its sale to Citibank. Citibank takes 100% control of Confia,-which implies a capitalization of between $150 to 200 million dollars. The government assumes a loss of US $1 billion in the sale of Confia. Management intervention of another exchange house. Only 8 out of 18 are not intervened. September HSBC signs understanding to acquire 19.9% of Grupo Financiero Serfin. Restructured credits in UDI or in hands of FOBAPROA account for over $44 billion pesos. Managerial intervention of CBI foreign exchange house has lifted. October Banamex agrees to absorb Banco del Atlantico. December Banorte pays FOBAPROA $680 million pesos for Banpais.  Annex B Page 1 of 1 Capital Increases and. Purchased Loans by FOBAPROA (billion of pesos) Capital Increase Purchased Loans Net Gross Probursa 3.8 7.6 9.3 Serfin 12.8 24.9 28.5 Atlantico 3.2 6.0 7.1 Prornex 1.7 3.0 3.4 Bital 3.8 7.5 8.8 Bancrecer/Banoro 7.4 14.0 15.4 Banorte 2.1 3.6 4.1 Banamex 12.5 23.0 28.6 Bancomer 12.7 23.6 30.1 Confia 3.3 6.0 6.6 TOTAL 63.3 119.2 142.0 Source: CNBV  Annex C Page 1 of I Government Estimate of Fiscal Cost of Banking System Rescue % GNP' AMOUNT (Million pesos) Debtors Program 3.0 96,300 Interventions and Betterment 5.9 187,300 Capitalizations Schemes 2.4 77,400 Toll Roads 0.6 18,800 Total 11.9 379,800 I/Estimated GNP for 1997: 3,201.8 billion pesos Source: CNBV  Annex D Page I of 3 Changing Composition of Banking System Structure of Mexican Commercial Banks (millions of pesos) Year-end 1994 Year-end 1995 Year-end 1996 Year-end 1997 ~~~~~~ B.1~ . .1'. k Bes Privatized 17 700,502 87.3 11 845,238 76.9 9 806,065 55.1 9 796,028 79.0 Domlestically Owned 12 i5,581 1.9 15 36,510 3.3 13 43,071 2.9 14 88,359 8.8 ForeignOwned 2 11,331 1.4 16 60,561 5.5 21 317,688 1 21.7 19 163,610 16.2 Intervened 2 74,872 9.3 7 108,855 9.9 10 126,160 8.6 12 n.a. l Assets Acquired 0 48,660 4.4 170,2631 11.61 | 0| Banking System 33 802,286 100.0 49 1,099,824 100.0 53 1,463,247 100.0 54 | 1,007,153 100.0 'Foreign bank assets have been reduced by $28.6 billion pesos to reflect the Mexicano assets which were acquired by FOBAPROA when Santander acquired Mexicano. 2Not including Bancrecer/Banoro for being under capitalization process and Atlantico for being under merging process. 3Information on these banks ceased to be published due to being considered in special situation. Includes: Union, Obrero, Cremi, Interestatal, Capital, Promotor del Norte, Sureste, Oriente, Anahuac, Inverlat, Confia, Industrial. Source: CNBV  Annex D Page 2 of 3 Mexican Commercial Bank System Year-end 1994 Year-end 1995 Year-end 1996 Year-end 1997 'GrQup N~af No,of =-n,e Privatized/ 17 Banamex Mexicano 11 Banamex Mexicano 10 Banamex Bital 9 Banamex Bital Domestically Owned Confia Oriente Confia Bital Confia Promex Bancomer Promex (Large banks) Bancomer Bital Bancomer Promex Bancomer Banorte Banoro I Banorte Banoro Comermex Banoro Banorte Banoro Bancrecer Serfin Atlantico I Serfin -Inverlat Serfin Serfin Atlantico Bancrecer2 Bancrecer Promex Bancrecer Atlantico Banorte Atlantico Centro Obrero M.Probursa Banpais Domestically Owned 12 Interaccione Mifel 15 Interacciones Mifel 13 Interacciones Promotor N 14 Interacciones Mifel (Small banks) s Invex IXE Invex IXE Invex Mifel Invex IXE Inbursa Quadrum Inbursa Quadrum Inbursa IXE Inbursa Quadrum Sureste Capital Sureste Afirme Industrial Quadrum Banregio Afirme Industrial Industrial Capital Banregio Afirme Del Bajio Bansi Banregio Banregio Bansi Del Bajio Bansi Banpais 3 Alianza Promotor N Promotor N Anahuac Alianza Centro 3 Del Bajio Del Bajio Foreign Owned 2 Citibank 16 Citibank Tokyo 19 Citibank ABN 19 Citibank ABN Santander Boston MProbursa Boston NY Boston NY Santander Bank of Am Santander Chicago Santander Chicago BNP Comermex -Mexicano AmEx -Mexicano AmEx JPMorgan - Inverlat BNP Nations BNP Nations Dresdner ABN JPMorgan BNS/ JPMorgan BBV/ Chase NY Dresdner Inverlat Dresdner MProbursa Societe Chase BBV/ Chase Tokyo Fuji Societe MProbursa Societe ING ING Fuji Tokyo Fuji Comerica Bank of Am ING Bank of Am Intervened 2 Union 7 Union Banpais 10 Union Banpais 12 Union Sureste Cremi Obrero Oriente Obrero Sureste Obrero Oriente Cremi Cremi Oriente Cremi Anahuac Interestatal Interestatal Anahuac Interestatal Inverlat Centro Centro Capital Confia Capital PromotorN Industrial Banking System 33 49 53 4 54 1/ Planned mergers. 2/ Foreign partner being sought. 3/ Merged with Banorte.  Annex D Page 3 of 3 Mexican Financial System 1995-1998* Financial Institutions 1994 1998 Privatized Banks 19 5 (1) NewBanks 16 10 Foreign Banks 8 22 (2) Investment Banks n.d. 33 Storage Warehouses 28 28 Leasing 63 53 Factoring 62 44 Foreign Exchange Houses 49 36 Credit Unions 360 98 (3) Savings and Loans n.d. 16 Sofoles n.d. 33 Mutual Funds n.d. 277 Afores (Pension Trust 17 Funds) * Data as of February 27, 1998 (1) Includes banks still belonging to original investors. Excludes Bancen and Banpais being merged with Banorte, as well as Atlantico and Promex, the first one recently merged with Bital and the second one, currently for sale, or merged with Bancrecer. (2) Includes Inverlat and Bancrecer, together with Santander, BBV and Confia given that its control will be soon transfered to foreign institutions. (3) CNBV registers 359 authorized unions, nevertheless, more than 250 are on the process of liquidation. Source: CNBV  Annex E Page 1 of 21 ICR BORROWER'S REPORT Executive Summary * Almost three years have elapsed since the onset of the Mexican financial crisis. Initial results indicate that the precarious financial situation of an important number of banks and debtors has begun to improve. * As of June 1997, the banking capitaLlization ratio reached 16%', without considering the remaining capital committed as a result of the Capitalization Loan Purchase Mechanism (CLPM) for the rest of 1997. * The provisions created by the 12 banks that remain under private control from the beginning of 1995 to June 1997, represent 415% of the outstanding provisions as of December 1994. These provisions were equivalent to 213% of their equity as of December 1994.