Argentina - Provincial Bank Privatization Loan Project and Bank Reform Loan Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 19467 Implementation Completion Report Argentina: Provincial Bank Privatization Loan (Loan 3878-AR) and Argentina: Bank Reform Loan (Loan 3926-AR) June 29, 1999 Finance, Private Sector & Infrastructure Country Management Unit 7 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.  ii CURRENCY EQUIVALENTS Currency Unit = Argentinean Peso EXCHANGE RATE US$1.00 Dollar= 1 Peso WEIGHTS AND MEASURES Metric System GOVERNMENT OF ARGENTINA FISCAL YEAR January 1 -December 31 ABBREVIATIONS AND ACRONYMS BCRA Central Bank of Argentina CAMEL Standard Bank Rating System (Capital, Assets, Management, Earnings, Liquidity) DECRG Development Research Group GDP Gross Domestic Product ICR Implementation Completion Report IMF International Monetary Fund LCSFP Finance, Private Sector & Infrastructure Department PR President's Report ROA Return on Assets ROE Return on Equity RWA Risk-weighted Assets SEF Superintendency of Financial Entities Vice President Mr. Shahid Javed Buwk Director, Country Management Unit Ms. Myrna Alexander Director, Finance, Private Sector & Infrastructure Unit Mr. Danny M. Leipzi;.er Task Manager Mr. Stefan Alber  iii FOR OFFICIAL USE ONLY IMPLEMENTATION COMPLETION REPORT Argentina: Provincial Bank Privatization Loan (Loan 3878-AR) and Argentina: Bank Reform Loan (Loan 3926-AR) TABLE OF CONTENTS Page No. EVALUATION SUMMARY .................................................................... VI A. INTRODUCTION .................................................................... VI B. PROJECT OBJECTIVES .................................................................... VI C. MAJOR FACTORS AFFECTING THE PROJECTS .................................................................... VI D. ACHIEVEMENT OF OBJECTIVES .................................................................... vil E. SUSTAINABILITY OF PROJECTS ..................................................................... VII F. BANK PERFORMANCE ..................................................................... VII G. BORROWER PERFORMANCE ..................................................................... VII H. ASSESSMENT OF OUTCOMES .................................................................... VII 1. LESSONS LEARNED .................................................................... VIII J. FUTURE OPERATIONS .................................................................... IX PART I: PROJECT IMPLEMENTATION ASSESSMENT ................................................................. I A. INTRODUCTION: LOANS 3878 AND 3926 IN THE CONTEXT OF THE CONVERTIBILITY PLAN ................. 1 B. OBJECTIVES .................................................................... 3 C. MAJOR FACTORS AFFECTING THE PROJECT ............. ........................................................ 5 D. ACHIEVEMENT OF OBJECTIVES ..................................................................... 6 E. SUSTAINABILITY OF RESULTS .................................................................... 25 F. WORLD BANK PERFORMANCE .................................................................... 26 G. BORROWER PERFORMANCE .................................................................... 28 H. LESSONS LEARNED .................................................................... 29 ANNEXES ANNEX 1: SUMMARY OF ASSESSMENTS .................................................................. 30 Part 1. Loan 3878, Provincial Bank Privatization .................................................................. 30 Part 2. Loan 3926, Bank Reform Loan ................... ................................................ 31 ANNEX 2: SUMMARY OF TRANSACTIONS SUPPORTED BY FIDUCIARY FUND FOR BANK REFORM/CONSOLIDATION ................ 32 ANNEX 3. TABLES .. ...... 34 Table 1: World Bank Involvement in Argentine Financial Sector, 1985-1998 ............. ...................... 34 Table 2: Characteristics of Purchasers of Privatized Entities ........................................... ................ 34 Table 3: Terms of the Provincial Bank Privatizations ...................................................................... 35 Table 4: Sizes of Privatized and Residual Entities ...................................................................... 36 Table 5: Creating Privatized and Residual Entities, Consolidated Balance Sheet for all Privatizations .................... ...................................................................................................................................... 37 Table 6: Liabilities Retired with Fiduciary Fund Loans ..................................................................... 38 Table 7: Total Disbursements and Losses of Fiduciary Fundfor Bank Capitalization ...................... 38 Table 8: Acquisition/Purchaser Comparisons, First Quarter, 1995 ................................................... 39 Table 9: Post Privatization Performance ....................................................................... 40 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.  iv Table 10: Loans by public provincial banks* ....................................................................... 41 Table II: Summary of Performance, Banks Involved with Fiduciary Fundfor Bank Recapitalization ...............................................................................................................1 ............................................. 41 Table 12: Pre and Post-Transaction Information, FFfor Bank Reform/Consolid . ............................ 42 Table 13: Lending to the public sector and change in employmentfor public andprivatized provincial banks (Q3 1998) .........................:.. . ANNEX 4. FIGURES ..6...................... 46 PART I I: STATISTICAL ANNEX .......................... 48 PROVINCIAL BANK PRIVATIZATION - LOAN 3878-AR TABLE 1: LOAN DISBURSEMENTS: CUMULATIVE ESTIMATED AND ACTUAL .......................... ............... 4.8 TABLE 2: PROJECT TIMETABLE ...................................................... 4.8 TABLE 3 : BANK RESOURCES: STAFF INPUTS ................................, 49 TABLE 4: RELATED BANK LOANS/CREDITS ............................... 4 9 TABLE 5: MISSIONS .............................................................................. .... C 0 TABLE 6: STATUS OF LEGAL COVENANTS ............................. S 2 BANK REFORM LOAN - LOAN 3936-AR TABLE 1: LOAN DISBURSEMENTS: CUMULATIVE ESTIMATED AND ACTUAL .......................... ............... 53 TABLE 2: PROJECT TIMETABLE ....................................................... 53 TABLE 3 : BANK RESOURCES: STAFF INPUTS ...................................................... 54 TABLE 4: RELATED BANK LOANS/CREDITS ...................................................... 54 TABLE 5 : MISSIONS ....................................................... 55 TABLE 6: STATUS OF LEGAL COVENANTS ...................................................... 57  v IMPLEMENTATION COMPLETION REPORT ARGENTINA Provincial Bank Privatization (Loan 3878-AR) and Bank Reform Loan (Loan 3926-AR) PREFACE 1. This is the Implementation Completion Report (ICR) for two operations in Argentina, Provincial Bank Privatization (Loan 3878-AR) and the Bank Reform (Loan 3926-AR). Because of their twin nature - their timing was essentially identical; their purpose was to support macro policies and bank restructuring - 3878 focused on public sector bank reform, 3926 focused on private bank reform - the two loans are analyzed together to gain a better understanding of the cross-relationships. Loan 3878 was approved May 4, 1995, became effective May 5, 1995 and closed on April 30, 1997. Loan 3926 was approved July 25, 1995, became effective December 1, 1995 and closed on December 31, 1998. Both loans were fully disbursed. 2. The ICR was prepared by Robert Cull (DECRG) and Stefan Alber (LCSFP). Because of the prospect of more bank restructuring loans in the future, a more extensive evaluation and analytical support are being presented. All data in this report are drawn from the public record. 3. Preparation of the ICR was begun one month after closure of Loan 3926, during January 1999. It is based on information in the project files and from detailed data on individual banks from the Central Bank of Argentina (BCRA), and on discussions with the staff of the implementation authorities - the Trust Fund for Provincial Development for Loan 3878; the Trust Fund for Bank Recapitalization for Loan 3926. Discussions with others familiar with Argentina's banking sector, including managers of the banks that were re-structured under these loans, were also very helpful. These discussions were carried out in Buenos Aires and the province of Mendoza during January 1999. 4. Argentine authorities reviewed this ICR and agreed with its focus and conclusions. The ICR was considered of high quality.   vi EVALUATION SUMMARY A. INTRODUCTION 1. Although Bank practice calls for separate ICRs for each loan, this ICR covers both loans given the confluence of their timing and objectives. A joint discussion should enhance understanding because (1) the two loans trace their roots to the same source, the Tequila Crisis of 1994-5, (2) in addition to supporting macro policies, they pursued the same purpose, structural reform of Argentina's banking sector (Loan 3878 focusing on public banks, 3926 on private), (3) their identical timing meant that they were supervised jointly, (4) there was cross-use of conditionality between the two loans, and (5) some banks received funds under both loans. These loans are a part of the Bank's long- standing support for financial sector development in Argentina, dating from the late 1 980s and extending to the latest operation, US$3.0 billion (the Special Structural Adjustment Loan approved in November 1998 - Loan 4405-AR). These two loans, therefore, were important phases of a comprehensive sustained support strategy. B. PROJECT OBJECTIVES 2. Both loans took advantage of the opportunities for change that often accompany crisis. Both were attempts to restore confidence in the system by eliminating long- recognized structural problems in banking. The President's Report (PR) for loan 3878 noted that, due to their poor performance, "provincial banks have been a major and chronic issue, and the current financial crisis is providing a historic opportunity to bring about structural and difficult-to-reverse change." Similarly, the PR for 3926 noted that it would contribute to the restoration of confidence in the banking system by, "hasten[ing] the process of consolidation ... and improv[ing] the financial structure of a distressed banking sector in the context of acquisitions, mergers, and restructurings." C. MAJOR FACTORS AFFECTING THE PROJECTS 3. The two loans attempted serious restructuring of some of the weakest banks in a fragmented system. At the same time, regulators permitted an increasingly strong foreign presence in the sector, and regulation and supervision of banks was substantially improved. As competitive pressure increases and regulatory forbearance declines, weak banks are more likely to fail. To create viable banks out of those involved in these transactions was a more difficult challenge under these circumstances.  vii D. ACHIEVEMENT OF OBJECTIVES 4. Each loan helped facilitate a substantial number of transactions in a relatively short time period. Given the high number of financial institutions that were part of the se transactions, the two loans contributed to re-establishing confidence in the banking sector after the Tequila Crisis. These structural changes were, moreover, accomplished at relatively little cost. For provincial bank privatizations, US$1.3 billion were disbursed to achieve ownership changes at sixteen banks whose pre-Tequila assets totaled about US $6 billion. US$767 million in disbursements contributed to acquisitions, mergers, and restructurings at banks whose pre-Tequila assets totaled US$14.6 billion. The banks involved in these transactions, therefore, comprised nearly a quarter of the banking system's pre-Tequila assets. E. SUSTAINABILITY OF PROJECTS 5. Although market conditions may dictate that some of the banks created under these two loans will experience further restructuring or even failure, the majority appear to be sound. More generally, the increased systemic stability now enjoyed by Argentina's banking sector is linked to the restructuring undertaken under these loans. These systemic benefits, which were the key objective of the loans, appear to be sustainable. F. BANK PERFORMANCE 6. Bank performance was satisfactory in terms of timeliness and willingness to assume the risks associated with adopting untried schemes to achieve structural change in banking. Staff also provided constructive supervision and analytical work that aided the borrower. The continuity in staff, and the resulting familiarity with the Argentine situation, also contributed to the success of the projects. G. BORROWER PERFORMANCE 7. Comments regarding the timeliness of the World Bank response apply equally t: the Argentine authorities. Trust funds were charged with the management of the privatizations and consolidations/mergers. Relatively simple operating guidelines, lean management structures and high caliber staff were charged with re-shaping Argentina's banking sector for some time to come. The continuity in personnel on both the Bank ar, I borrower side further fortified that trust. H. ASSESSMENT OF OUTCOMES 8. Project performance for both loans was satisfactory due to the substantial contributions of each loan in re-shaping Argentine banking during a turbulent period, to substantial provincial fiscal savings due to the privatization of provincial banks combined with a sharp redefinition of the role of the provincial state. The Special Trust Funds which were created have demonstrated their value as crisis tools. Demand for Trust Furmd  viii assistance was marked by a clear slowdown when conditions of normalcy in the banking system set-in starting late 1996. I. LESSONS LEARNED 9. Among the lessons applicable to both loans are: * Financial/fiscal crises offer opportunities for structural change in banking. * However, structural/ownership change in banking requires resources to re- structure past losses. These financial resources facilitate acceleration of structural changes that would otherwise not occur, or occur at a more modest speed. * Even when sufficient financial resources are available, there will be selection bias in the roster of potential acquirers. Troubled banks tend not to attract prime-rated buyers. It is unclear whether program design features alone (such as more incentives) can overcome that bias. Neither loan analyzed here completely resolved this problem. * Funds were disbursed only after the careful review of the individual transactions. The approach may be useful in other countries and contexts. * Some of the banks created through these transactions may require additional restructuring as market conditions change. Some failures need to be anticipated. * A strong borrower commitment is indispensable in such complicated, time- intensive transactions. 10. Lessons specific to Loan 3878 (Provincial Bank Privatization) include: * Post-privatization credit market development is likely a long-tern proposition in geographically isolated regions like Argentina's provinces. Moreover, given the selection bias alluded to above, the new private owners can be expected to remain cautious in adapting to retail credit/banking in these areas. * Preparatory work to facilitate the sale/recovery of residual assets (e.g., inventory of assets and assessment of their quality) should begin as early as possible, earlier than has occurred in many cases to date. Experience in Argentina and elsewhere has shown this to be among the most intractable problems associated with privatization. Early private sector involvement in the recovery process is the most promising solution. * Provinces that retain a relatively large ownership stake in their bank decrease the probability of attracting a high-quality buyer. Sales of 100% of the shares of the privatized entity should be encouraged. * Even under difficult market conditions, privatization can proceed. 11. Two final lessons specific to Loan 3926 (Bank Reform) is that: * In an effort to attract high-quality buyers of troubled banks, the Bank might explore more flexible lending conditions. For example, it may have been better to allow interest rates to vary with quality of the acquirer in the transactions supported by Loan 3926.  ix Many of these transactions involved weak institutions. Even if the consolidatecl entities later fail, these types of lending operations can buy time for an overworked bank failure resolution mechanism. J. FUTURE OPERATIONS 12. Some of the restructured banks may well fail in the future but, with the ever improving failure resolution capacity of the Central Bank, market solutions, rather than public sector bail-outs, are likely to prevail. The most recent example is the failure of Banco Mendoza and the subsequent sale to private sector buyers. 13. While both Trust Funds were created with charters of limited duration, it is likely that these mechanisms will only be abandoned once it is clear that they are no longer needed. For instance, in provincial bank privatizations, there are still several small provincial banks and the large Cordoba bank which may enter a privatization program given the by-now proven fiscal benefits. The continued value of the Bank Capitalization Trust Fund is more in doubt but will be determined with the context of a broader review of the adequacy of bank failure resolution mechanism as part of the loan conditionality of the recent 1998 US$3.0 billion loan (Special Structural Adjustment Loan - Loan 4405). 14. The Bank's continued involvement in the restructuring of the financial sector is likely to be maintained through Loan 4405 and possibly further operations. Promising new initiatives are those of the new Governor of the Province of Cordoba who expressed a strong interest in privatizing the provincial bank of Cordoba. This would complete the privatization of all major provincial banks, other than that of the Provincial Bank of Buenos Aires.  PART I : PROJECT IMPLEMENTATION ASSESSMENT IMPLEMENTATION EXPERIENCE AND RESULTS 1. Although Bank practice calls for separate ICRs for each loan, this ICR covers both loans given the confluence of their timing and objectives. A joint discussion should enhance understanding because (1) the two loans trace their roots to the same source, the Tequila Crisis of 1994-5, (2) in addition to the support for macro policies, they pursued the same purpose, structural reform of Argentina's banking sector (Loan 3878 focusing on public banks, 3926 on private), (3) their identical timing meant that they were supervised jointly, (4) there was cross-use of tranches between the two loans, and (5) some banks received funds under both loans. These loans are a part of the Bank's long- standing support for financial sector development in Argentina, dating from the late 1980s. These two loans, therefore, were important building blocks of a larger, more sustained support package. A. INTRODUCTION: LOANS 3878 AND 3926 IN THE CONTEXT OF THE CONVERTIBILITY PLAN' 2. The nature of the Convertibility Law, which made the Central Bank into a currency board by mandating a 100 percent international reserve requirement for high- powered money, creates special challenges for the financial system. Actually, the challenge points in both directions, given the symbiotic relationship of the Convertibility Law with the financial system. While external shocks (which may be amplified as a consequence of adopting the Convertibility Law) are first manifested in the financial system, the health of the financial system on its own may also challenge the viability of the Convertibility Plan. Whichever is the nature of the shock, it is critical for the sake of the Convertibility Plan to have a strong financial system, as other countries, developed and developing, are currently painfully realizing. 3. The inherent rigidity of a currency board, which denies the use of discretionary monetary or exchange rate policy, has contributed to building the credibility of Argentina's stabilization program. At the same time, such rigidity reduces the ability of the Central Bank to provide the functions of lender-of-last-resort and other monetary operations, which usually, in times of crisis, facilitate the stabilization of the banking system. Furthermore, unlike Hong Kong's currency board, which counts on significant excess reserves for use in case of crisis, Argentina's access to such reserves has I This sub-section is taken from, "Argentina: Financial Sector Review," World Bank, Report 17864-AR, Sept. 28, 1998 (hereafter, "Financial Sector Review").  2 historically been quite limited. As a result, with a monetary base that is very susceptible to movements in capital flows, interest rates become a major adjustment variable. 4. Large fluctuations in interest rates have a direct impact on the banking system through the changed valuation of its assets and liabilities, and indirectly through their impact on real economic activity. Prolonged rises in interest rates punish more severely weakest banks first, leading often to insolvency and contagion of other banks which under other arrangements would not be as susceptible to a crisis. While both the exchange rate and monetary policy rigidities of the Convertibility Law, and the resulting limitations to the lender-of-last-resort function, have played a constructive role in preventing crises by instilling discipline on individual banks and the Argentine banking system as a whole, once a crisis occurs, these very qualities become serious impediments to the stabilization of the banking system. The 1994/95 Challenge 5. The challenge of the currency board to the financial system became evident during the Tequila Crisis. Argentina's crisis in 1994/95 began as an exchange rate crisis, not a banking crisis. Initially, Argentine banks experienced outflows from peso- denominated accounts and inflows into dollar-denominated accounts, reflecting a concern about the maintenance of convertibility. Such concerns were related to political uncertainty leading to the May 1995 Presidential elections, recent slippages in fiscal anct current account balances, rapid growth in credit, and the absence of an IMF program.2 As long as confidence in the financial system was maintained, the Government could deal with the currency run through the dollarization of the financial system. In that sense, the presence of a bimonetary system played a constructive role during the crisis, by providing an alternative to outright capital flight. 6. By March of 1995, the shrinkage in domestic money related to outflows of capitnl (and lack of access to international credit) had produced significant increases in credit risk within the banking system, and the character of the crisis changed from a run on the peso to a run on both the peso and the domestic banks. The structural conditions of the banking sector--which were characterized by the absence of deposit insurance, limited nature of a lender of last resort, and a segmented and inefficient financial system, as we] L as memories of many previous financial crises that ended in asset confiscation--fueled a systemic run on deposits (mostly by domestic residents). 7. For the first four months of the crisis prior to the May Presidential elections, Argentina suffered a massive liquidity shock, with bank deposits declining by 18 perceni; (Annex 4, Figure 2), and liquid international reserves by 30 percent, as access to the international financial markets was cut off (with the exception of multilateral lending). Under the Convertibility Plan, where the monetary base has to be fully backed by international reserves, capital outflows result in the demonetization of the economy. However, the credit squeeze in the economy was moderated by a reduction in reserve requirements, and the increased use of dollar denominated bonds as part of international 2 See "Maintaining Financial Stability in Global Economy" Remarks by Pedro Pou in a Conference sponsored by the Federal Reserve Bank of Kansas, 1997.  3 reserves. This demonetization affected both the performance of the financial system, but also, as expected, real economic activity, leading to a sharp recession. From a policy perspective, given that bank reserve requirements were less than 100 percent, M2 at that time was about 3.5 times the stock of international reserves, under the Convertibility Law the Central Bank had more limited resources than other countries to confront a run against bank deposits (being a weak "lender of last resort"). 8. By the end of May, 1995, forceful action both by the Government (cutting the fiscal deficit and obtaining support for its new economic program by multilateral organizations) and the Central Bank (through skillful management of liquidity to the financial system, including lowering remunerated reserve requirements, and the extension of credit through swaps and rediscounts) reinstated confidence in the peso and the banking system's solvency. Additional measures included the establishment of a privately managed small-depositor deposit insurance scheme and the establishment of trust funds to facilitate the privatization of provincial banks (partly funded by Loan 3878) and the restructuring of private sector banks (partly funded by Loan 3926). These measures facilitated the return of deposits to the banking system, which, by early 1996, reached their pre-crisis levels (Annex 4, Figure 2). The two loans analyzed here, moreover, provided foreign reserves to Argentine authorities at a crucial time. Dollars deposited in the foreign account of the Central Bank were converted to pesos and then used to capitalize the two trust funds charged with achieving the structural changes outlined in the loans' objectives. B. OBJECTIVES Common Objectives 9. While focusing on somewhat different specific issues, the two loans shared common objectives and provided reserves bolstering the convertibility plan. Both were attempts to restore confidence in the system by eliminating long-recognized structural problems in banking. The President's Report (PR) for loan 3878 noted that, due to their poor performance, "provincial banks have been a major and chronic issue, and the current financial crisis is providing a historic opportunity to bring about structural and difficult- to-reverse change." Similarly, the PR for 3926 noted that it would contribute to the restoration of confidence in the banking system by, "hasten[ing] the process of consolidation ...and improv[ing] the financial structure of a distressed banking sector in the context of acquisitions, mergers, and restructurings." 10. The similarity between the loans extended to institutional design. In each case, the loans were used to partially capitalize Trust Funds. The Inter-American Development Bank also played a key co-financing role in capitalizing the Trust Fund for Provincial Development, and provided US$750 million in support. These two funds then on-lent to individual banks or to provinces to facilitate an ownership change at a troubled bank (either through acquisition or merger). Both the provincial banks targeted under loan 3878 and, to a lesser extent, the troubled private banks covered under loan 3926 were insolvent. Although the methods to facilitate transactions under the two loans varied, the funds supplied were, therefore, an important part of the restructuring process. Indeed, a  4 strong case can be made that, without these funds, far fewer transactions would have taken place. 11. A final reason to analyze the two loans in the same document is simply that sorne banks accessed both funds - that is, a private bank bought and later merged with a provincial bank3. Objectives Specific to Loan 3878 (Provincial Bank Privatization) 12. Although the two loans shared common objectives, each focused on a different part of the banking sector, and each, therefore, had its own more narrowly tailored goals. First, the Loan 3878 was designed to support the privatization of up to fifteen provincial banks comprising about ten percent of total banking assets (forty percent of provincial banking assets). Second, because the provincial banks were insolvent, their balance sheets would be divided into a so-called good bank (to be assumed by the private purchaser) and a residual entity, which would remain the responsibility of the province. The loan proceeds would be used to support the orderly and timely resolution of residual liabilities. Third, in the event of failed privatization, the loans would provide resources to the provinces to facilitate bank closures. The second and third objectives lead naturally to the fourth which was to improve the fiscal situation in provinces, both in the short-term through the immediate injection of funds, and in the longer term by severing ties with banks that had long been fiscal drains. A fifth, and final, objective was to bring additional banks into the monitoring and supervisory umbrella of the Central Bank (BCRA), as many provincial banks had not complied fully with regulatory requirements; in the past. Objectives Specific to Loan 3926 (Bank Reform Loan) 13. The private bank consolidation process was, by its private sector nature, more dynamic and thus timing and volumes were harder to predict than was the privatization process. In privatization, the actors -- the provinces and their banks - had a clear objective to sell and as rapidly as possible. By contrast, private consolidation was a fast moving process where business opportunities appeared sporadically, often after long negotiations. Out of necessity, this was reflected in the objectives and the operating guidelines for Loan 3926. The Loan was designed, first, to stimulate and hasten the process of consolidation of a fragmented private banking sector. In so doing, it would, second, improve the financial structure of a distressed banking sector by encouraging acquisitions and mergers that matched weaker partners with stronger ones. 14. In what follows, each loan is evaluated based both on its contribution to the shared, broad objectives and with respect to its own more narrowly tailored objectives. 3Each loan, however, is treated separately in the accompanying statistical appendices.  5 C. MAJOR FACTORS AFFECTING THE PROJECT Context Surrounding World Bank Involvement 15. Hyperinflation in the late 1 980s severely undermined the ability of banks and other financial institutions to serve as effective intermediaries between savers and investors. The price stability resulting from the 1991 Convertibility Plan provided an opportunity for sustained financial sector development. The Argentine government placed a high priority on restructuring many financial institutions to ensure that such development was not ultimately de-stabilizing. The World Bank was clearly an important partner in this process, and the two loans reviewed here are, perhaps, the clearest illustration of the Bank's efforts to facilitate structural change in the Argentine financial sector. 16. Substantial recent involvement in the sector put Bank staff well along the learning curve in identifying which structural difficulties could feasibly be addressed. Beginning with the 1986 report, "Banking Sector: The Need for Reform," Bank staff had completed at least four substantial pieces of economic and sector work and three loans aimed at financial sector reform prior to the two loans discussed here (Annex 3, Table 1). In addition, the President's Report (PR) for loan 3926 noted that the capabilities of the Superintendency of Financial Entities (SEF) were improved during the 1 990s under the Bank's Public Sector Reform Technical Assistance Loan (3362-AR). These repeated interactions helped forge strong relationships between Bank staff and Argentine officials. The high levels of trust between the two parties likely increased the probabilities of success for the loans analyzed here. Evolution of Banking System4 17. Since Convertibility, the Argentine financial system had undergone a series of fundamental changes. Hyperinflation in the 1980s had reduced the financial system to levels below that found in the lowest income countries, and in a certain sense these past crises, coupled with effective policies in recent years, have helped make the system more robust today. Once credibility of the Convertibility Law was established, inflation decelerated and the re-intermediation of the financial system began in earnest. As seen in Annex 4, Figure 1, credit, both total and to the private sector, has grown significantly in recent years, though financial depth (M3/GDP), while growing rapidly, is still somewhat modest relative to other countries at comparable levels of per capita income. 18. The 'Tequila' crisis of 1994-95 marked a turning point for the authorities, with a run on some of the smaller banks and then on the system. Deposits declined sharply (far left, Annex 4, Figure 2), and interest rates rose by as much as 1500 basis points on peso 4 This sub-section also draws heavily on the, "Financial Sector Review." The figures that follow show developments by type of bank from Tequila through late 1997. Robert Cull, co-author of this report, also constructed those figures. Separating domestically-owned from foreign-owned private banks proved difficult given the available data sources. That identification was not, therefore, updated for 1998 data. However, the most important developments occurred by late 1997 and the data presented are sufficient to understand general trends. Most importantly, for the banks involved in the transactions covered by these two loans, all data and analysis are extended through third quarter, 1998, the last quarter for which BCRA data were available.  6 loans and 900 basis points on dollar loans. The dramatic 18% decline in total banking sector deposits in first quarter, 1995, masks much steeper declines in deposits at weaker banks, many of them owned by the provinces. The crisis was halted by the assembly of a support package by the multilateral institutions, convincing support by the governmern for convertibility, and a credible program of bank restructuring and closure, of which the two loans reviewed here were a key part. 19. Annex 4, Figure 2 also describes the evolution of deposits by type of bank by quarter from the fourth quarter of 1994 to third quarter, 1997. In nominal terms, total deposits have been increasing more or less steadily since second quarter, 1995. Total deposits in both private domestic and public banks floated around the 20 billion peso mark, although a recent spate of foreign acquisitions coincided with a dip in private domestic deposits late in the period. The increase in total deposits is accounted for by foreign banks and, to a lesser extent, privatized banks. 20. To some extent, the increase for foreign banks reflects not only acquisitions late .n the period, but also depositor flight to quality, especially during the Tequila crisis. In 1994-95, for example, the number of foreign-owned banks did not increase (Annex 4, Figure 4) but their deposits, both in nominal terms and as a share of total deposits, climbed steadily if not dramatically. The share figures for foreign-owned banks did, however, stabilize by mid 1995 (Annex 4, Figure 3). The only type to lose nominal deposits (relative to the December, 1994 level) was the private domestic banks, and losses occurred in two waves. Early in the period, private domestic deposits declined during the flight to quality. There were also a number of domestic bank closures at this time. The end of the period saw a decline in deposits that coincided with aforementioned acquisitions by foreign-owned banks, and some of those transactions were financed by loan 3926. These figures provide some evidence that the two loans contributed to the restoration of confidence in the sector - foreign and privatized banks led the recovery i:l deposits and these were among the banks that benefited most from the assistance of the Trust Funds. 21. In terms of the major factors affecting loan outcomes, the overall development c f the sector is important to keep in mind. The two loans attempted serious restructuring 4)f some of the weakest banks in a fragmented system. At the same time, regulators permitted an increasingly strong foreign presence in the sector, and regulation and supervision of banks was substantially improved.5 D. ACHIEVEMENT OF OBJECTIVES 1. The Transactions Common Elements 22. Though the details of the transactions underwritten by the two loans differed, andI thus will be discussed separately, they did share certain similarities. First, because these 5 See "Financial Sector Review."  7 transactions were focused on changing the governance structure at faltering banks, there were substantial difficulties in attracting prime-rated buyers. In many cases, the only banks willing to undertake these acquisitions or mergers were themselves in a weakened position. The injection of funds for transactions covered by 3926 and the opportunity to enter provincial banking for acquirers covered by 3878 were likely most attractive to banks in need of change regarding their core business. In addition, certain restrictions placed on recipients of 3926 funds (described below) also limited the attractiveness of that type of assistance and thus may have affected the roster of potential acquirers. 23. The low quality of the banks on the sales block meant that the two loans shared a second feature -- the funds disbursed were used to fill holes either in bank capital or in provincial finances that were brought about by poor past performance. Without these resources it would have been difficult, perhaps impossible, to face up to the "sins of the past" and thus achieve ownership changes. Another, more minor, feature shared by these transactions was that the troubled banks not only had weak financial assets, their physical assets were also unattractive. For the provincial banks, their branch networks were quite extensive, which meant relatively high shares of fixed to total assets. For some of the banks acquired in the 3926 transactions, their existing branch coverage overlapped with those of potential purchasers. In both cases, this may have made these transactions more difficult. These "unfavorable initial conditions" need to be kept in mind when evaluating these loans. Loan 3878: Provincial Bank Privatizations Purchasers 24. Provincial banks were not particularly attractive acquisition targets - they had been losing money over an extended period and the quality of their assets was low. In addition, these banks were mostly small and their asset portfolios were not well diversified (at least geographically). This is likely to have influenced who was willing to take these banks over. Annex 3, Table 2 indicates that the purchasers were not the best performing banks in Argentina. They were mostly small - none had over US$1.2 billion (and most had less than US$500 million) in assets at the time of the acquisition, and none held more than one percent of total banking assets in Argentina. 25. Appraisal evaluations indicated grave doubts about finding buyers for these banks. It comes as little surprise, therefore, that the eventual purchasers were not large, well-functioning banks. The return on assets (ROA) and return on equity (ROE) figures in Annex 3, Table 2 show that in all cases except Mendoza, Prevision Social, Jujuy, and Santa Cruz, the provincial banks were bought by banks with ROA and ROE below the sector average (and often well below). Further, the Jujuy and Santa Cruz cases are aberrations because both were bought by former provincial banks (Jujuy by Banco Macro-Misiones and Santa Cruz by Banco San Juan). Due to the transfer of bad assets to the residual entity, these two newly privatized entities tended to enjoy great initial success and, therefore, their performance, at that time, probably did not reflect their (likely) long-term performance.  8 26. Annex 3, Table 2 also shows that the transactions varied widely in the share of total capital transferred to private owners (5 0.4% to 100%), and in the share of capital held by the largest private owner (34.5% to 100%). These ownership stakes might have had implications for the corporate governance of the privatized entities and might, therefore, have affected the performance of the privatized banks. Given the short post- privatization time-series available, however, the governance-performance links that we discuss below should be interpreted with caution. Sellers (The Provinces) 27. One feature shared by the public provincial banks was their penchant for losing money. This provided the underlying fiscal rationale for privatization.6 Recent research has demonstrated that, based on the observed loss rates at these banks, the present valuw of future re-capitalization required by Argentine law exceeded the maximum short-termL costs associated with privatization. As will be described below in more detail, these short-term costs came from realizing and cleaning up the past losses of the provincial banks. The main conclusion is that, even if the provinces receive nothing for their privatized entities and recover no residual assets, it made sense from a fiscal perspective to privatize to avoid the large future losses of the provincial bank. 28. Although the fiscal benefit is not likely to be the only, let alone the most important, gain from privatization, these results suggest that the fiscal gains alone may have justified privatization. In what follows, we discuss a handful of ways in which the privatization process and residual asset recovery might have been improved. We shoulc. not, however, lose sight of the fact that Loan 3878 produced a fundamental fiscal benefit to the provinces that privatized. Below, we also provide evidence on the post- privatization performance gains and the overall effect that privatization had on the provincial banking sectors. Privatization Method: The Details 29. A key institutional innovation allowed the provinces to meet the fiscal and political costs of privatization more easily. In all but one case, the purchaser of the privatized entity did not assume ownership of all pre-privatization assets and liabilities. Although the individual cases varied, the basic strategy was to first place attractive assetls in the privatized entity and then to match those assets with liabilities. The remaining assets and liabilities were then transferred into a residual entity. 30. Residual asset recovery would have been neither quick enough, nor on such advantageous terms to liquidators, that it would have covered most residual liabilities. Consequently, provinces needed some way to meet a substantial portion of their residual obligations immediately. To address this, the Argentine Government, the Inter-America.-, Development Bank and the World Bank developed the Trust Fund, a part of the federal governnent that extended loans to provinces that planned to privatize their provincial 6 In this sub-section, we briefly summarize results found in George R.G. Clarke and Robert Cull, "Why Privatize? Th;! Case of Argentina's Public Provincial Banks," World Development, 27(5), May, 1999 (a), pp. 867-888.  9 banks. The provinces then used the loan proceeds to pay off part of the obligations of the residual entity. In this way, short-term obligations were converted to longer terms. From a political perspective, financing obligations in this way was presumably also beneficial, as the yearly loan payments due to the Fund are less eye-catching than short-term obligation payments would have been. More fundamentally, because the provinces lacked the fiscal resources to meet these short-term obligations, it was imperative that some re-scheduling be done. The Trust Fund, therefore, enabled provincial governments to design privatization arrangements that they could afford and that mitigated political objections. Contract Terns 31. In an implementation completion report, one might expect to focus more narrowly on whether objectives were met rather than on how they were met. Here, however, we discuss the contract features in detail for two reasons. First, it helps underscore the difficulties in structuring these transactions. Second, analyzing the trade-offs made by both parties to these transactions may yield some lessons that deepen our understanding regarding privatization. In addition to price paid, there were other non-price features of the privatization agreements that appear to have been important that are summarized in Annex 3, Table 3. 32. In the Argentine context, opponents to bank privatization were bought off through agreements to limit the number of layoffs or to compensate laid-off workers and to maintain branches in certain cities. Based upon a review of the requirements imposed on the purchasers of the privatized banks, it appears that limits on branch closings and layoffs were the rule rather than the exception (Annex 3, Table 3). Of the sixteen contracts summarized, half had some restriction on the number of employees that could be dismissed. Another purchaser agreed to implement a job re-training program. Three of the contracts stipulated that the private purchasers maintain the existing branch network and ten permitted the closure of branches, but required that the purchaser maintain service provision in all locations served at the time of privatization. These contract features strongly suggest that the political buyoffs present in other privatizations were also evident in Argentina. 33. As noted above, the public provincial banks were chronic money losers frequently in need of frequent re-capitalizations. To pass a substantial share of their low-quality assets onto a private purchaser while, at the same time, imposing branching and labor restrictions on that purchaser would have been difficult, if not impossible, without concessions on other dimensions. The most attractive of these were the fiscal agency contracts that were awarded to purchasers to provide banking services to the provinces and guarantees as to the quality of the acquired assets. The fiscal agency contracts, which, among other things, provide income to the private owners for coordinating the payments activities of the provincial government, varied in duration from five to twenty years. Ten of the sixteen agreements provided fiscal agency contracts of at least ten years. In interviews, the new private owners confirmed that these contracts were of vital importance, especially initially, as an abnormally high share of the privatized banks' income is generated from services (described in more detail below).  10 34. In many cases, however, the lure of the fiscal agency contract appears to have been insufficient to entice a private bank to acquire assets of dubious quality. Rather than verify the quality of each individual asset, which proved time-intensive, many provinces ultimately took to guaranteeing a substantial share of the assets transferred to the privatized entity. World Bank staff were quick to recognize asset quality verification as a potential stumbling block in these transactions. Given that the privatization program supported by the Trust Fund was scheduled to close in 1998, these privatizations had to be completed quickly. Despite the potential fiscal risks, the decision to guarantee assets was probably a good one. Without it, fewer privatizations probably would have occurred. 35. In six cases, the province guaranteed assets up to either a fixed limit or a certain share of the total assets acquired. In two other cases, private owners were able to substitute assets from the residual entity for privatized assets during some trial period; i:A another, the buyer could shift low-quality assets to the residual for a period. Finally, in two cases the guarantee was set as a fraction of the residual assets recovered. Presumably, since the owner of the privatized entity was also charged with managing the residual entity, the idea was to increase incentives to recover residual assets. Only four cases did not guarantee the privatized asset portfolio at all. 36. Annex 3, Table 4 shows the size of the privatized and residual entities for the sixteen completed privatizations that relied on Trust Fund assistance.7 The most striking feature of the data is the size of the residual entities. In only four of fifteen cases for which data are available did the province manage to transfer more than half of the pre- privatization assets to the purchaser and one of these (Entre Rios) is somewhat misleading. Both Entre Rios and Chaco had nearly finalized their privatizations before the Fiduciary Fund became operational. A desire to provide some fiscal relief to these early privatizers enabled them to enter the program after the fact. Because most details :f these two privatizations had been worked out, hovvever, neither transaction was typical cf those that followed. In particular, Entre Rios transferred all pre-privatization assets to the purchaser, and later used Fiduciary Fund assistance to guarantee some of them.8 Residual Asset Recovery 37. Experience in Argentina and elsewhere indicates that residual asset recovery is exceedingly difficult, perhaps the most intractable of the problems associated with privatization. Problems are severe when the public sector remains centrally involved in the recovery effort. The most promising solution is to begin the process early and to 7 We refer to the cases listed in Table as provincial bank privatizations throughout the paper. One of the sixteen, Municipal de Tucuman, is referred to as a municipal rather than a provincial bank in Argentina. Because municipal bank privatizations were also eligible for Fiduciary Fund assistance, we include that case in our analysis. Data for another, the Santa Fe privatization, were not yet available as of our last visit to Buenos Aires (January 1999). s Correspondence between World Bank staff and Argentine officials makes it clear that the Chaco and Entre Rios case, were different than the others. On the one hand, Argentine officials wanted to provide fiscal relief for two provinces that had privatized early. On the other, Bank staff were concerned about funding transactions that did not meet the criteria set forth in loan, and doing so in an 'after the fact' manner. In the end, Bank staff decided that privatizations undertaken prior to the approval of the loan could be supported, provided they met the criteria set forth. On that basis, Corrientes was excluded from receiving support, while Chaco and Entre Rios were deemed eligible. Neither Chaco nc- Entre Rios have enjoyed large post-privatization performance improvements, at least in comparison with the other cases (see below). One cannot help but wonder whether it had something to do with the way they entered the program.  I1 involve the private sector as early as possible to overcome the public sector's political and incentives' constraints. 38. Annex 3, Table 4 summarizes the provinces' experience with residual asset recovery. So far, results have not been impressive. In most cases, provincial governments realized that they were ill equipped to handle the recovery effort, and they therefore initiated incentive contracts to encourage the purchasers of the privatized entities to try to recover these assets. The provinces granted the privatized entity a higher share of the face value of recovered assets for the items that were most difficult to recover (as indicated primarily by the length of time that the asset had been deemed non- performing). While the underlying intuition may have been sound, the figures in Annex 3, Table 4 suggest that these relatively simple incentive contracts have failed to produce substantial asset recovery. 39. Mendoza's experience is instructive because the residual entity created from its two provincial banks (Banco de la Provincia de Mendoza and Banco de Prevision Social) is large and the recovery process has been rife with hiccups.9 The province has, so far, recovered only one percent of its residual assets, and the owners of the privatized entity (one private bank was created from the two provincial banks) are no longer involved in the process. Provincial policy makers decided that it would be best to acquire the services of private experts skilled in bank asset recovery. That would be difficult to do, however, without providing very detailed information on the assets to be recovered - information on the delinquency of credit re-payment did not adequately describe the prospects for asset recovery. 40. To improve the available information, the province has employed about one hundred people (many of them ex-employees of the provincial banks) to make a complete catalogue of the residual assets. The ultimate goal is a database that can be searched by credit (or by creditor), which provides complete information on any guarantees or collateral attached to each loan. Initial interest in the residual assets has been high, with over twenty companies reviewing the existing database. Mendoza intends to award the recovery contract to the company that displays the most technical expertise. A subset of those that have reviewed the database will pay a fee for the right to make a presentation that outlines their asset recovery capability. 41. Determining which company is best suited to recover assets appears to be as much art as science. The only certainty is that the private companies generally will only manage but not bid money for the right to recover assets. Strategy for Separating Residualfrom Privatized Assets 42. Annex 3, Table 5, which presents a consolidated balance sheet for the privatized and residual entities created to date, provides more detail on the "anatomy" of these transactions. On the assets side of the balance sheet, 92% of the available liquid assets 9 Bank staff were also quick to recognize that residual asset recovery was not going well and required a substantial re- thinking regarding strategy. A special paper was commissioned to analyze these issues. Many of the suggestions from that paper were eventually incorporated into Mendoza's asset recovery strategy.  12 (which excludes public bonds) were transferred to the privatized entity. Among the public bonds, 76% of those that were publicly quoted were transferred to the privatized entity. Only 17% of those that were not publicly quoted were transferred to privatized entities. These figures should not be surprising, since private purchasers had a strong preference for liquid, easily valued assets. In contrast, they assumed only about a third of the loans and the fixed assets of the provincial banks. Presumably, the private purchasers have been able to meet their geographic service requirements (Annex 3, Table 3) with far fewer fixed assets than those of the old public provincial banks, an indication that the public banks had inefficient branch networks and poor information technology. The figures also suggest that the provinces will face an uphill battle selling many illiquid fixed assets and recovering loans of dubious quality. 43. On the liabilities side, provinces transferred over ninety percent of their private deposits to the privatized entity. This reflects a desire to maintain the confidence of depositors. Indeed, all of the owners of the privatized entities chose to keep the name of the provincial bank that they acquired. In fact, in the case of Misiones and Rio Negro, the acquirer chose to merge with the privatized bank, under the name of the provincial bank.' These owners must have recognized the value of the provincial bank's name in attracting, and holding onto, deposits. That the provinces were able to shift ownership of their faltering banks to private entities without undermining the confidence of depositors is a clear success of Loan 3878. 44. Having transferred most private deposits to the privatized entities, the residual entities were left with public sector deposits and obligations to BCRA and Banco de le Naci6n, the two public banks that provided liquidity to faltering banks during the Tequila Crisis. Over ninety percent of the BCRA/Naci6n obligations and almost eighty percent of the public sector deposits of the provincial banks went into residual entities. 45. Another important issue was whether the funds used to facilitate transactions weie sufficient. Conversations with staff at each of the Trust Funds suggest that the authorithk s were concerned about the moral hazard problems associated with 100% financing of liabilities. In the case of provincial bank privatizations, it is clear that the strategy was to cover some, but not all, residual liabilities. Annex 3, Table 6, presents a summary of the funds received from the Trust Fund to facilitate the individual transactions, and the type of liabilities that were retired with those funds. One potential problem with the Trust Fund program was that maximum loan amounts were determined based on the assets of the provincial bank rather than on the quality of those assets. The eight provinces that received loans of US$78-80 million had pre-privatization assets ranging from US$314 to US$523 million and shares of non-performing credits ranging from fourteen to sixty percent. Within a given asset class, therefore, the assistance received by the weakest banks might have been inadequate. 46. Annex 3, Table 5, indicates, however, that residual obligations to financial intermediaries (most to BCRA and Banco de la Naci6n) totaled US$1.8 billion while Annex 3, Table 6, indicates that total Trust Fund disbursements were US$1.3 billion (US$841 million going to BCRA and Naci6n). This suggests that, despite the limits, 1