Argentina - Special Repurchase Facility Support Loan Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No: 25548 IMPLEMENTATION COMPLETION REPORT (SCL-44060) ONA LOAN IN THE AMOUNT OF US$ 505.05 MILLION TO THE ARGENTINE REPUBLIC FOR A SPECIAL REPURCHASE FACILITY SUPPORT LOAN 03/27/2003 Argentina, Chile, Paraguay and Uruguay Country Management Unit Poverty Reduction and Economic Maniagement 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 (Exchange Rate Effective) Currency Unit = Argentine Peso US$1.00 = ARG$3.1 (March 2003) FISCAL YEAR January 1 - December 31 ABBREVIATIONS AND ACRONYMS IBCRA Banco Central de la Repiblica Argentina CAN Core Accountability Implementation Completion Report CAS Country Assistance Strategy GOA Government of Argentina ,Dt bInter American Development Bak IllFSL Repurchase Facility Support Loan SA Special Dollar Account SSAL Special Structural Adjustment Loan Vice President: David de Ferranti Country Director: Axel van Trotsnburg Sector Director: Ernesto May Task Team Leader/Task Manager: Paul Levy  ARGENTINA REPURCHASE FACILITY SUPPORT LOAN (RFSL) (4406-AR) PREFACE This is the Core Accountability Implementation Completion Report (CAI) for the Repurchase Facility Support Loan (Loan No. 4406-AR) in the amount of US$505.05 million. The Loan provided additional enhancement to the Repurchase Facility established by the Central Bank of the Argentine Republic (BCRA) in 1996. The Facility provided access to a liquidity pool financed by a group of intemational commercial banks. The Bank Loan was granted to the Republic of Argentina (the Borrower) but Loan resources were under the management of the BCRA. It was approved by the World Bank Board on November 10, 1998 and made effective on September 13, 1999. The Loan was closed on January 22, 2003, prior to the original Closing Date of September 15, 2003. The Loan was fully disbursed by the time of its closing. Besides the Bank Loan, the Inter American Development Bank (LADB) cofinanced this Project and contributed a US$500 million loan of similar characteristics to the one provided by the Bank. The support provided to the Repurchase Facility was an integral part of a broader emergency structural adjustment operation. The Special Structural Adjustment Loan (Loan No. 4405) that provided the developmental underpinnings was approved by the Board the same day and in conjunction with the operation being assessed here. Preparation of this Core ICR benefited from a mission to Argentina that took place in mid-March 2002. Also, it is based on material in the project files, several interviews with participants during the preparation and execution of this operation, and on perspectives gained by the Region during years of involvement in Argentina's financial reform efforts. This ICR is complemented by the ICR of the SSAL, a loan to which the RFSL was intricately linked and presented jointly to the Board. The BCRA contributed to the preparation of the ICR by preparing its own evaluation of this operation, a copy of which has been included unedited in Section I.   ARGENTINA AR SPEC REPURCHASE CONTENTS Page No. 1. Project Data I 2. Principal Performance Ratings 1 3. Assessment of Development Objective and Design, and of Quality at Entry 2 4. Achievement of Objective and Outputs 8 5. Major Factors Affecting Implementation and Outcome 3 6. Sustainability 15 7. Bank and Borrower Performance 16 8. Lessons Learned 19 9. Partner Comments 20 10. Additional Information Annex 1. Key Performance Indicators/Log Frame Matrix 23 Annex 2. Project Costs and Financing 24 Annex 3. Economic Costs and Benefits 25 Annex 4. Bank Inputs 26 Annex 5. Ratings for Achievement of Objectives/Outputs of Components 27 Annex 6. Ratings of Bank and Borrower Performance 28 Annex 7. List of Supporting Documents 29  Project ID: P062992 IProject Name: AR SPEC REPURCHASE Team Leader: Paul Levy TL Unit: LCC7A ICR Type: Core ICR Report Date: March 31, 2003 1. Project XData Name: AR SPEC REPURCHASE LIC/TF Number: SCL-44060 Countty/Department: ARGENTINA Region Latin America and Caribbean Region Sector/subsector: Central govemment administration (100%) KEY DATES Original RevisedlActual PCD: 09/12/1998 Effective: 09/13/1999 Appraisal: 09/10/1998 MTR: Approval: 11/10/1998 Closing 09/15/2003 01/22/2003 Borrower/lmplementing Agencv: GOVERNMENT OF ARGENTINA/BANCO CENTRAL DE LA REPUBLIC Other Partners: STAFF Current At Appraisal Jlice President: David De Fenranti Shahid Javed Burki Counny Manager: Axel van Trotsenburg Myrna Alexander Sector Manager: Ernesto May Guillermo Perry Team Leader at ICR: Paul Levy Paul Levy ICR Primarv Author: Claudio Pardo 2. Principal Performance Ratings (HS=Highly Satisfactory, S=Satisfactory, U=Unsatisfactory, HL=Highly Likely, L=Likely, UN=Unlikely, HUN=Highly Unlikely, HU=Highly Unsatisfactory, H=High, SU=Substantial, M=Modest, N=Negligible) Outicome: U Sustainability: HUN Institutional Development Impact: M Bank Performance: S Borrower Performance: U QAG (if available) ICR Quality at Entry: S Project at Risk at Any Time: Yes Bol r,-ow'er s Pei lorniamn. c Vote: Pei lorrnimzi e o/u'thl Borurtm er's (C,,nn al Bank ix I ated Swilqa tori  3. Assessment of Development Objective and Design, and of Quality at Entry 3.1 Original Objective From a strictly financial perspective, the stated objective of Loan 4406-AR was to enhance the Repurchase Facility (repo facility) established in 1996 by the BCRA to insulate domestic banks from the illiquidity of their asset portfolios for brief-although sufficiently long-periods during international liquidity crises. The repo facility was established to provide a new line of defense to preserve the liquidity of the financial system. It gave the BCRA the option to sell to a syndicate of international commercial banks tradable dollar-denominated fixed-income securities issued by the. Government of Argentina (GOA). The sale was tied to a short-term repurchasing commitment. (When the Loan was declared effective in September 1999, there were 16 international commercial banks participating in the repo facility, to which they had committed US$6.35 billion for locked-in periods of 2 to 5 years carrying a quarterly "green-shoot" renewal clause and an embedded implicit interest rate averaging some 200 basis points over LIBOR. Specific terms and conditions varied somewhat from bank to bank.) However, in a more fundamental way, the RFSL operation was an integral part of the conditionality of the US$2.53 billion SSAL approved by the Bank's Board in November 1998 in conjunction with the operation being assessed in this ICR. Thus, from a structural perspective, the RFSL supported the second objective under the Banking Sector Reform component of the SSAL operation, that is: "Increase resilience of the banking system to external shocks, and facilitate the exit of weak banks from the system". From the perspective of the SSAL, the direct support of the repo facility provided by the RFSL was crucial since it added to the lines of defense of the banking system against potential liquidity shocks, particularly those originating from abroad. It was also important to the objectives of the SSAL since the RFSL gave further assurances of substantial private sector contribution to the effort, as it encouraged private banks to participate in the repo facility. Both the financial and structural objectives were clear, realistic, reasonably matched with the objectives of the BCRA and its capacity as implementing agency. They were important for the country as well as for the work on financial sector structural reform that the Bank was actively supporting at the time in the general context of the Convertibility Plan. Furthermore, there was a consonance,between the objectives of this operation and those in the Bank's Country Assistance Strategy (CAS), particularly given the external threats that Argentina was facing at the time and to which the Bank, jointly with the Inter American Development Bank, responded with a timely and convincing emergency support financial package. From this perspective, the objectives anchoring the financial support given to the repo facility were quite responsive to the borrower's circumstances and fully in line with its developmental priorities, given the paramount interest of the authorities in enhancing public confidence in the country's banking system and preserving its deposit base and access to additional liquidity during periods of turmoil in international financial markets. The Tequila crisis a few years earlier had demonstrated the need to reinforce the arsenal of options available to the BCRA to preserve systemic liquidity during periods of international crisis, particularly in a country like Argentina that under the Convertibility Plan operated with a currency-board scheme, which greatly limited the options of the BCRA in its role as lender-of-last resort to the banking system. -2 -  3.2 Revisedl Objecti ve The Central Bank had negotiated a "Repo" facility with commercial banks to provide access to liquidity in times of crisis. Liquidity at times of crisis was a structural issue for Argentina under the currency board arrangement. In conjunction with the SSAL, this contingent loan in support of the Central Bank's Repurchase "Repo" facility was intended to strengthen the financial system's safety net and similarly safeguard economic and social achievements. 3 3 Originial Components: The RFSL included only one original component to the arrangement in existence with commercial banks, which was designed to provide direct financial support to the already mentioned Repurchase Facility at the BCRA. However, the Loan also included US$5.05 million to cover the payment of the up-front 1% fee due to the Bank by the Borrower once the Loan became effective. Effectiveness of the RFSL was made conditional to the effectiveness of the SSAL, the consistency of the Government's macroeconomic policy framework and satisfactory progress by the Government in carrying out the economic program stated in its Development Letter dated November 2, 1998. From this perspective, the RFSL was dependent on the release of the first tranche of the SSAL, effectively making it a fourth "floating" tranche of that special structural operation. Disbursements to support the repo facility could only take place after June 1999 and after the BCRA had entered into a specific repo transaction activating the facility. This single component was not revised during implementation. While the loan was made to the Republic of Argentina, the GOA had to open and maintain at all times a special dollar account (SA) at the BCRA, where all disbursements from the Loan had to be deposited. This account had no.other purpose than to serve as the vehicle for providing the intended support to the repo facility, so its proceeds could be used exclusively by the Central Bank to meet three basic cash needs: o To repurchase securities it had sold previously to international banks participating in the repo facility; o To meet margin calls it received from banks that had purchased securities under the repo facility either by delivering cash or additional eligible securities purchased with SA funds; and o To buy securities during periods of no margin calls and when the BCRA thought it advisable to replenish its stock of eligible securities. There was also an obligation of the BCRA to deposit into the SA: a) cash received from participating banks making margin refunds, b) maturity receipts from securities previously purchased with SA funds and c) receipts from the sale of such securities. Clearly, this single component was directly related to achieving the objective of enhancing the repo facility, particularly in the eyes of participating and potential candidates among international banks. It was thought that this would encourage more banks to participate or, if turmoil in international financial markets or the perception of Argentine sovereign risk increased, fewer participating banks would be tempted to quit the repo facility or reduce their commitments to it. The stated goal of the BCRA was to maintain a repo facility able to provide additional liquidity during times of financial distress equivalent to some 10% of the deposit base of the banking system. The almost US$1 billion enhancement provided by the RFSL and a similar loan - 3 -  from the Inter American Development Bank was aimed at achieving this target, which at the time of Loan approval was close to 9% of the deposit base. While the financial support provided by the two multilateral organizations could not be used directly to increase the size of the repo facility, the cash enhancements it provided reduced the counter-party risk faced by participating banks and encouraged participation in the facility. The repo facility clearly improved the Central Bank's capacity as a lender-of-last-resort under the Convertibility Plan and complemented the high liquidity requirements it already had imposed on local commercial banks following the Tequila crisis, when large sums were drawn out of Argentine banks. Thus, the primary purpose behind the strategy of the BCRA, which also followed a conservative policy of 100% backing of the monetary base with international reserves, was to have a banking system with access to a liquidity cushion large enough to deter a run on banks, even under extremely stressful systemic liquidity circumstances. The crucial point of the BCRA's policy was then to establish a liquidity shield, of which the repo facility was one of several components, robust enough to discourage massive deposit withdrawals from local banks. This meant that the repo facility and consequently the RFSL operation were of a preventive nature, having a small chance of actually being disbursed provided that the perceived strength of Argentina's fundamentals continued into the foreseeable future. The design of the RFSL single component was clearly in line with the objective pursued by the BCRA, which had moved aggressively to secure altemative sources of liquidity in case of eventual crises. The Central Bank also clearly had the necessary administrative and financial management capabilities as the implementing agency of the repo facility. Besides having a strong professional team, it was the creator of the facility and had a solid commitment to its success. (Thus, although the World Bank did not have experience with similar operations in the past, from a purely financial viewpoint the single component design of the RFSL was adequate, it supported the structural objectives of the SSAL and clearly met the Borrower's needs at the time it was granted. For a more In-depth analysis of the structural problems faced by Argentina at the time of entry in 1998 and general background information, please see the ICR prepared for that operation (Report No. 22619).) 3 4 Revised Components N/A 3 5 Onality tit Entrv Despite a significant degree of controversy over this operation, the performance rating for guality of entry of the RFSL is considered to be satisfactory. There are several reasons for this. To start with, at the time of Board approval, the objectives of the RFSL responded to a high priority of the BCRA and the GOA, which were trying to avert a major financial crisis in the midst of serious turmoil affecting emerging economies. The Bank also shared this priority and in fact the repo facility was seen as a central component to the authorities' strategy to defend the Convertibility Plan and maintain the economic stability that it had brought to Argentina. A convincing repo facility was seen by all parties as a critical tool in the Central Bank's effort to mitigate the contagion impact on the local banking system from extemal shocks. The Repurchase Facility was previously evaluated by the Bank in the."Argentina Financial Sector Review" (1998). Support to the facility had been discussed in the past and had been resisted by the Bank, for reasons described below. However, when the liquidity crisis menaced Argentina in 1998 from - 4 -  abroad, the request of the BCRA became more focused and clearly stated. The country's access to global markets was practically closed at the time, in spite of significant improvements in financial supervision and the regulatory framework. At the same time, the Bank was willing to explore more innovative products to meet the needs of countries like Argentina, which had an impressive and proven track record of policy improvement and successful adjustment borrowing, particularly in the financial sector. All this provided a strong rationale for the Bank wanting to support the repo facility, particularly after the Board had approved the guidelines for Special Structural Adjustment loans as a new type of instrument in the Bank's arsenal to address market shocks as the one Argentina was facing. Besides, the objectives of the proposed operation were in line with the CAS and the Bank's safeguard policies. There was nonetheless significant reluctance on the part of the Bank to act as a lender-of-last-resort, mostly because of the perception that this could be seen as a reversal of roles vis-a-vis the IMF, besides the difficulties for any such operation to meet the Bank's "productive purpose" test. Sentiments were also quite strong in the Bank's Board, where the approval of the RFSL actually confronted the opposition of five Chairs, something quite unusual and to a great degree due to the view that this operation was beyond the Bank's statutory mandate. But the fact was that the Argentine authorities had not been successful in convincing the IMF-which was facing a potential liquidity problem because its US$90 billion capital increase had not yet been approved-to provide ex-ante assurances for access to emergency liquidity support, such as that needed to enhance the repo facility. Besides, Argentina had agreed and publicly stated that the Fund's US$2.8 billion Stand-by Arrangement approved in February 1998 was precautionary. So the IMF funding was available only for reacting to specific events and circumstances and its use by Argentina to support the repo facility could have been easily misconstrued as a sign of weakness by the markets. Thus, the Argentine authorities turned to the Bank in search of an innovative solution to the dilemma they were facing. The Bank staff also saw clearly the institutional constraints of pledging ex-ante financial support to a precautionary liquidity facility, which basically implied that legally speaking support for the repo facility could not stand on its own. It was decided that the way to proceed with the proposed operation was to subject it to the same policy conditionality of the SSAL that was also under preparation (e.g., satisfactory macroeconomic policy framework, the Development Policy letter and the first tranche conditionality). After all, the enhancement of the repo facility reinforced the liquidity cushion developed by the BCRA and, consequently, protected the integrity of the banking system. By becoming a floating "fourth tranche" to the companion SSAL, albeit under a separate Loan Agreement, the RFSL actually became an adjustment loan operation supporting the reform package set out in the matrix of the 1998 SSAL. This allowed the Bank to keep its developmental focus and still provide support to the repo facility. This is not the place to assess the quality of entry of the SSAL, which in any event was found to have been satisfactory in the ICR prepared for such operation. Furthermore, in early 1999, a few months after Board approval of the SSAL and the companion RFSL, a high-level panel of experts reviewed five rapidly processed adjustment operations approved in calendar 1998 for their quality of entry and depth of assessment. It found that the Argentina's SSAL was highly satisfactory and the strongest of all both in terms of product and process. The panel felt that: "It - 5 -  built on strong domestic ownership and an excellent stock of accumulated country and sector knowledge." From Office Memorandum in the project files by Joe Wood, Panel Leader, to Prem C. Garg, Director, MDOQA, dated February 3, 1999. A copy of this memo is in the project files. The panel also felt that there was "strong professional consensus both within the Bank and with the Borrower and partner institutions on the nature of the problems faced and the appropriate policy responses." They also judged the RFSL to have been satisfactory overall. In the view of the Bank's management the emergency adjustment operation put together late in 1998 contained a reform package sufficiently robust to justify the full US$3 billion finally committed under the two related loans: the SSAL and the RFSL. However, from the viewpoint of design, there were practical reasons for preferring to have a separate loan agreement for the RFSL. The RFSL had to be a stand-by loan whose disbursement was contingent on the actual use of the repo facility by the BCRA. Moreover, this was a transaction-based operation that had to become nonetheless irrevocably committed to disbursement on the part of Bank once the SSAL became effective. This was an essential feature to make the Bank support a credible enhancement of the repo facility in the eyes of financial markets and the international banks participating in it. Contrary to the funding provided by the SSAL, the RFSL was considered unlikely to disburse, particularly if the adjustment program was successful and bank deposit levels did not have a severe reversal. The loan design also took into account the need to incorporate a fast disbursing feature so as to meet the short deadlines implicit in margin calls. This spoke in favor of allowing disbursements following the activation of the repo facility by the BCRA. There was no need to disburse prior to activation of the facility.' Also, it would have been simply too costly for the Borrower since funds had to be kept in the special account at the Central Bank Because the money was deposited in a special account and it was not at the free disposal of the BCRA, it could not be accounted as part of the international reserves. with a small chance of actually being used, given the low probability assigned to an activation of the facility. Bank guarantees were also impractical since they normally make payments only after an event of default is declared, which was too late for an effective enhancement of the repo facility. At the time of entry, Loan disbursement was seen as an event in the future, if at all, and likely only in connection with a major systemic liquidity crisis in the local banking sector, as it unfortunately happened in the end. The key assumption in this respect was that the repo facility was helping to avoid or mitigate a threat from abroad, that is, the contagion effects of a global financial crisis on Argentina's financial system. The question is if, with today's hindsight, it was reasonable to ignore a different scenario where the severe economic crisis that Argentina was facing had more domestic origins. After all, Argentina had not been successful in increasing a small export base and its external debt was growing rapidly in an already highly indebted economy, while its fiscal performance continued to be a drag on the economy. Furthermore, its rigid exchange rate system, although instrumental in bringing stability, made of Argentina an economy vulnerable to external shocks. But these were weaknesses the structural reform package supported by the SSAL was trying to address. Moreover, the substantial international financial package put together at the end of 1998, of which the SSAL and the RFSL were part, aimed at securing a block of time sufficient to give the reforms a chance to take a hold while the - 6 -  international financial environment regained a degree of normalcy. The central issue remained nonetheless of whether too much emphasis was being placed in the existing framework, one of continuity in the same approach to Argentina's existing challenges and trust in the ability of the GOA to implement extremely taxing political decisions less than one year prior to presidential elections. The answer to this dilemma impinges in the judgment of the quality of entry of the RFSL and they have been satisfactorily addressed in the parallel ICR for the SSAL (Report No. 22619). From the narrower perspective of this report, it can be said that once the Bank decided to continue to defend the reforms in the policy matrix of the SSAL, and the Convertibility Law underpinning it, the RFSL became appropriate, consistent with and supportive of financial stability and the country's developmental program. If nothing else, by contributing to reduce the likelihood of a systemic liquidity constraint, which gravely threatened economic activity with an additional blow at a time when an economic recession was starting to be felt. In the President's report, there was recognition that the reform process could be delayed by the upcoming presidential elections, but there was the positive previous experience with actually a deepening of structural reforms following the severe impact of the Tequila crisis, which also had coincided with national elections. There was also the risk of a more severe and prolonged intemational crisis or delays in Argentina regaining a more normal access to global capital markets. All this led to state in the President's Report to the Board that the risks of a worsening international scenario and its impact on the local economy could end up being much more severe than anticipated, as it eventually happened. However, the conclusion at the time of entry was that made the proposed reforms and the supportive measures under the SSAL and RFSL all the more necessary. Also relevant when assessing the quality at entry is that these two operations were made in the context of the Bank's new and recently formalized guidelines for programmatic and emergency adjustment lending. See "Programmatic and Emergency Adjustment Lending: World Bank Guidelines", Operations Policy and Strategy, September 29, 1998, which was considered by the Board on October 22, 1998. Under these policy guidelines, existing lending instruments were adapted so that the Bank could have a better set of tools to assist countries at risk of going into crises or, as a last resort, to help those in crisis to better manage their responses. Crisis prevention under these guidelines was seen as "an extraordinary last-resort effort". The financial consequence was that this lending had to be made under terms and conditions that took into account the higher risks and costs involved. As a result, the two operations for Argentina received the minimum emergency pricing of 400 basis points considered in the guidelines and an accelerated 5-year repayment schedule, including 3 years of grace for the principal. The emergency pricing for the RFSL included in the final design was reasonable, given that this loan was a candidate for disbursement only under the higher risk conditions of a crisis and severe systemic liquidity constraints. In hindsight, the support provided to the repo facility, the objectives sought and the assumptions made at the time of entry were quite reasonable. In fact, the repo facility played a constructive role as a deterrent to panic-driven liquidity crises for about two years prior to its actual disbursement, so it was a useful preventive tool for the BCRA that acted by its mere presence. This was prior to the alarming liquidity problems confronted by Argentina's financial sector starting in 2001 and that created the conditions that led to the activation of the repo - 7 -  facility, which also finally triggered the disbursement of the RFSL by the BCRA in September 2001 (long after the policy measures associated with the SSAL had been implemented and those funds had been disbursed). 4. Achievement of Objective and Outputs 4 1 Outcon7/achievement of objective: The Bank's ICR guidelines require assessing the degree to which this operation achieved its objectives considering today's perspective and evaluating their consistency with current country and sectoral assistance strategies. In this light, the RFSL had an unsatisfactory general outcome, in particular since the SSAL's overall objectives were not met even though its specific output were all achieved. In this respect, the ICR for the SSAL states on page 9 in connection with its achievements: "It is clear, in retrospect, that the package of policy reforms supported by the SSAL were necessary but ultimately insufficient to redress the cumulative impact of the series of shocks that confronted Argentina in 1999 and 2000." That being said, however, the RFSL was successful in supporting the repo facility as a preventive tool for some two years following Board approval. At the time of approval of the emergency structural adjustment program in November 1998, world financial markets were already in turmoil following the uncertainties triggered by the Russian crisis. The months that followed saw further deterioration in confidence on emerging economies, as illustrated by the crisis in Brazil, which resulted in a drastic devaluation of the real. Brazil is Argentina's neighbor and a leading trading partner; so all this had negative repercussions on Argentina's balance of payments and its capacity to reverse the economic downturn. Fortunately, the good timing of the financial/adjustment package of November 1998, which received strong support from several official creditors, provided Argentina with a powerful countervailing weight. In fact, by February 1999 Argentina was able to regain market'access by placing a new bond issue in international capital markets. Perhaps more relevant to the restoration of confidence in Argentina's financial good standing were the continued efforts by the authorities to steadily improve the financial regulatory framework and the supervision of the domestic bank system, and many of the implemented policy measures were directly assisted by the SSAL. Thus, a credible set of policy decisions and skillful managemert on the part of the BCRA were major factors behind the sustained gains shown by overall bank deposits during 1999 and 2000, despite weak economic activity at the time. Chart I below illustrates the behavior of bank deposits, which show a steady tendency to grow in the aftermath of the Tequila crisis; a trend that lasted until late 2000. Deposit levels then started to go down suddenly and sharply. -8 -  Chart i Argontino: Total Dopoofto Bonting Syotom 90,000.000 80.000.000 70.000,000 ,,.--' Pnrte deposxts 60,000,000 50.000,000 g 40,000,000 30,000,000 - 20.000.000 - 10,000,000 I0 Pubc sector depooits Jon 1094 to Occ 2001 Source: BCRA In this context and prior to 2001, the repo facility supported by the RFSL was part of a successful set of instruments used by the BCRA as preventives against systemic liquidity crises in the local banking sector. Perhaps what is important to highlight is that, independently of the quality of other government policies, market confidence in financial policies, particularly those under direct control of the BCRA, appear to have generated confidence in the system and encouraged growth in the deposit base. This took place in an adverse overall economic environment, where other government programs, such as those trying to bring the fiscal deficit under control, had a much harder time making credible inroads. The deepening in financial intermediation that lasted until the end of 2000 and that was supported by the expansion in domestic bank deposits provided one of the few pillars in which to sustain an economic recovery, which unfortunately did not materialize. In the end, a combination of complex factors that it is still too early to disentangle, but that certainly included a dose of poor economic policies and a profound political crisis, also brought the country's banking system into a downspin, adding a whole new dimension to the current and unprecedented hardships of Argentina. Today, the problems of the banking system go well beyond temporary liquidity shortcomings as those the repo facility was trying to alleviate. Now the difficulties have the characteristics of a full-blown solvency crisis across the banking system. The current crisis has meant a complete turnaround in the assumptions underpinning the repo facility and the loss of its purpose. The currency board is gone and the Central Bank has regained its authority to conduct a much more proactive monetary policy, including the function -9-  to act as a lender-of-last resort. Nonetheless, the main stumbling block for the BCRA today is first and foremost a deep crisis of confidence, which impedes the Central Bank to effectively exercise its new powers in the marketplace. However, the assessment of the RFSL should not be excessively distracted by current events, since even in today's economic environment there is still relevance in its original objectives. In fact, it is highly probable that insulating domestic banks from international liquidity crises and increasing the system's resilience to extemal shocks will form part of the objectives of any new financial policy consensus likely to crystallize once the current economic and financial chaos prevailing in Argentina subsides. Despite the repo facility's initial contributions as part of the arsenal of Central Bank's liquidity-support policies, there are in hindsight serious reservations about the outcome of this opeiation derived mainly from the decision of the BCRA to go ahead and activate the repo facility for the wrong reasons and when market conditions were extremely unfavorable for doing so. Indeed, in September 2001 Argentina was not in fear of a contagion from an extemal crisis; on the contrary, it was clearly seen as a source of problems for the global market, as illustrated by the huge and widening spread shown by its sovereign debt vis-a-vis the market benchmark. By September 2001, JP Morgan's EMBI data showed that a wide gap had opened up in the course of the year between Argentina's spread and that of other countries in the EMBI index (excluding Turkey), strongly suggesting that investors in emerging countries debt were actively discriminating against Argentina. As it will be explained in more detail in the next section, the activation was not only expensive for the BCRA but also futile, since the funds from the facility went directly to finance capital flight with practically no chance to reverse the process, accomplishing no practical developmental impact. At the time of the repo facility activation, the Bank had little saying on the convenience of doing so. More importantly, the RFSL loan had long been committed with no string attached besides the effectiveness of the SSAL-that triggered the disbursement of its first tranche-an operation that had been fully disbursed and had been closed by the end of 1999. The point is that, given the prevailing adverse circumstances at the time of the facility activation, the outcome of the RFSL would have been quite satisfactory overall had it not been disbursed by the BCRA. 4 2 Ouitptuts bv components. The RFSL had a single component designed to provide an enhancement for the repo facility. This enhancement was at its best when perceived as reducing systemic liquidity risk, thus encouraging private deposits to continue flowing into the banking system. It was argued earlier that the repo facility played a satisfactory dissuasive role while the perceived dominant risk was contagion from extemal financial turmoil, a period that lasted roughly until the end of 2000. This section will concentrate on providing an assessment of the period following the activation of this facility, when the disbursement of the RFSL took place. By disbursing, the RFSL took on the characteristics of a traditional transaction-based operation, even tough was conceived as a stand-by loan. Because of this and the fact that the RFSL was seen as the fourth "floating" tranche of the SSAL, the Bank basically accepted the design of the repo facility as negotiated by the BCRA with the syndicate of intemational banks at its inception in 1996. However significant, changes in the repo facility during implementation did not have a bearing on the nature of the Bank's commitment to it. In particular, Bank disbursements were not tied to the size of the facility or the amounts disbursed by participating banks. - 10 -  When the repo facility was activated by the BCRA in September 2001, bank deposits in Argentina already had been dropping for some months. Chart I shows that private deposits after reaching a peak of $75.8 billion in February 2001 had gone down 13% by the end of August. The drop in overall deposits was even more pronounced at 14.2%, given that public sector deposits also were decreasing rapidly. This relative decline was slightly smaller than the one experienced between September 1994 and April 1995 as a consequence of the Tequila crisis, although in absolute terms there was no comparison-$7.2 billion in seven months during the Tequila episode versus $12.1 billion in the six months up to August 2001. Also, a major difference with the 1995 event was that after recovering slightly in September 2001, overall deposits in Argentina's banking system resumed their sharp decline, a trend that then continued unabated. In mid-2001 Argentina presented a recessionary and deflationary environment, where domestic factors and not contagion from abroad, were dominant. The CAS Progress Report of October 1, 2001 summanzed the situation at the time as follows: "Argentina remains in the midst of a three year recession and is confronted with problems of competitiveness, fiscal vulnerability and uncertainty, along with the related social consequences." Clearly, there was a growing resistance from capital markets to continue to finance the fiscal deficit and the country's more general needs for external fmancing. This market exhaustion had taken a hold despite the fact that the GOA had met all its first semester 2001 performance criteria under its Stand-by Arrangement with the IMF. By mid-year, it had become evident to the authorities that in order to trigger a revival of investor confidence they needed to bring on a profound reversal in economic fundamentals. It was in this context that the economic team in the national govemrnment led by Minister Cavallo proposed the draconian policy of the "zero fiscal deficit" of July 2001. Its enactment with Congressional backing facilitated the GOA efforts to secure additional extemal financing from the IMF for some US$8 billion under the Stand-by Arrangement. The activation of the repo facility was part of this effort to restore Argentina's financial sustainability and credibility. The use of the repo facility, however, was mostly a cosmetic move by the GOA aimed at emphasizing the solidarity of leading intemational banks with the new agreement negotiated with the IMF an4 the "zero fiscal deficits" policy. At the time, the Central Bank still had a reasonable level of international reserves and other less onerous ways to assist domestic banks with their liquidity difficulties. Furthermore, according to its stated policy, the activation of the facility was not the BCRA's first line of defense to raise the level of cash available in the banking system. This notwithstanding, the first repo operation took place in mid-September 2001 and generated US$1.27 billion to the BCRA. The transaction required the sale by the BCRA of senior dollar-denominated securities issued by the GOA either from its own portfolio or in its custody. The face value of the securities sold by the Central Bank under this short-term repurchase agreement was US$US4.78 billion (for an implicit over-collateralization of 25% at market value). A second repo transaction was executed six weeks later for a cash value to the BCRA of US$500 million. All told, for these two transactions adding up to US$1.77 billion, the BCRA had to sell to participating banks tradable securities with a face value of some US$5.83 billion-that is, each dollar in cash received by the BCRA was backed by GOA-issued bonds with a face value of US$3.29. In these transactions, the main limiting factor was the shortage of eligible dollar securities, which had become quite scarce in the local market following the "mega-bond" exchange of June 2001. (This exchange had enabled the GOA to swap US$29.5 billion of its own -1 1-  paper, largely in the hands of domestic financial institutions, for new bonds carrying more favorable terms and conditions.) At the time, commitments under the facility by participating intemational banks were still substantial at US$4.75 billion, although they were markedly lower than the US$6.35 billion available at Loan effectiveness in September 1999. Thus, given prevailing prices for GOA securities, a full utilization of the funding available by the BCRA would have required the sale of bonds with a face value of some US$15.63 billion-a sum which simply was not available in the portfolio of domestic financial institutions. The argument can be made that if the BCRA really wanted to cash in its portfolio of senior and highly tradable government bonds, it always had the option of going for an orderly sale of these securities in the marketplace, which had global reach. While it is true that this necessarily would have caused further downward pressure on an already weak market, the sale of bonds with a buy-back clause did not eliminate the risk that participating banks could sell them in the market once the bonds were in their hands. The risk that this could happen was even higher in a market where prices were spiraling downward. Besides, banks also were getting 25% more bonds because of the agreement to over-collateralize these repo transactions. As a matter of fact, many participating banks did sell the securities they received. The incentive was there to make a quick profit, if they took care of buying them back later on in the marketplace at a cheaper price and prior to the repurchasing date with the BCRA. There was even a couple of participating banks that were not able to deliver all of the promised securities to the Central Bank, simply because they were not able to find them when the time came to resell them to the BCRA. The repo facility had an embedded average pricing of 200 basis points over LIBOR, which looked quite attractive in view of the high spread commanded by Argentina's sovereign risk in the global market. However, the perception of this pricing being low is deceiving, given that repo transactions were over-collateralized with liquid assets that participating banks could sell. Besides, participating banks were protected from reductions in the market value of the bonds, since any drop in bond prices of more than 5% triggered a margin call, and these calls had the benefit of the combined US$1 billion enhancement provided by the two stand-by loans pledged by multilateral institutions. All this meant that the banks' exposure to counter-party risk with the BCRA was quite low and the small spread was well justified. The price of Argentine bonds did continue to fall after the repo transactions were made. In fact, margin calls were being made to the BCRA on a daily basis, which created a myriad of logistic and management problems. Nonetheless, the BCRA was able to cope reasonably well and its performance met expectations at the time of entry. A major inconvenience took place in December 2001, following a speech by the newly appointed President Rodriguez Saa that touched upon Argentina's external debt. This was interpreted by a couple of participating banks as meaning that Argentina had entered into an event of default. Since this also meant that the repo facility was in a technical default as a result of cross-default clauses in the master agreement, these banks claimed that they had the right to settle existing repo transactions and the price of collateral bonds practically at their own will. The issue was resolved favorably soon thereafter, but the BCRA decided to accelerate the two existing transactions. In the end, the market value of the bonds included in the two repos had dropped some 40% by the time the final buy-back took place on January 10, 2002. While the repo facility was active, cash deliveries in an amount of some - 12 -  US$660 million met three-fourth of the margin calls. Other margin calls were met by delivering additional bonds to the banks. Today, the repo facility has been phased out and its master agreement cancelled. In summary, the activation of the repo facility was ineffective in stopping deposit withdrawals from the local banking system and all this for the benefit of a minor and short-lived cosmetic impact. Furthermore, it ended up being an expensive exercise for the Central Bank, which had to assume the opportunity cost of not selling the bonds directly in the marketplace. The BCRA also has assumed the liabilities involved in servicing the two loans used for the facility enhancement and which were granted under the unusually onerous terms and conditions of the Bank's emergency operations. If nothing else, the activation of the facility under conditions that went much beyond a temporary liquidity squeeze from foreign contagion may have violated the original policy commitment of the BCRA. Under the conditions the facility was used, it also meant that the local banks were accelerating their solvency problems. Indeed, a not insignificant portion of the bonds used in the repo transactions came from local banks and were among the best quality assets in their portfolios, so their sale to finance loss of deposits necessarily meant a deterioration not only of the banks' liquidity position but also and perhaps more crucially a fall in their solvency ratios. 4.3 Net Present Value/Econornic rate of retuirn. N/A 4.4 Financial rate of return: N/A 4.5 Instittional development impact: (Rating of Institutional Development Impact is not being implemented by the Bank at this time pending further analysis to permit meaningful and consistent guidance for its application.) 5. Major Factors Affecting Implementation and Outcome 5 I Factors ouitside the control of government or implementing agencv: The main factor outside the control of the GOA and the BCRA affecting the outcome of this operation was clearly the state of global financial markets. The repo facility was designed precisely to counterbalance the negative impact on domestic deposits brought about by contagion from abroad. The Argentine economy has been for some time heavily dependent on a continuous access to international capital markets since this is a country with a heavy debt burden. So the market turmoil that resulted from the Mexican, Asian and Russian crises all brought about a real danger of serious contagion for Argentina. But Argentina was able to deal effectively with contagion symptoms in all those occasions, particularly once it deepened financial reforms and the BCRA established the liquidity shield already commented upon and of which the repo facility was one of the components. The global economic and financial environment confronted by Argentina in 2001, however, was quite different. As already mentioned, Argentina's overall economic situation in mid-2001 had changed for the worse. To start with, the devaluation in Brazil had weakened - 13-  further the competitiveness of Argentina and in 1999 real GDP fell sharply. While in early 2000 the economy showed incipient signs of recovery, these were short-lived and real GDP registered another decline for the year. By 2001 the world economy was amid the most severe slowdown since the early 1990s, something that had a serious negative impact on activity in emerging markets. Moreover, emerging economies were facing a climate of increased investor risk aversion, which resulted in rising bond spreads for these countries, although on average the degree of contagion witnessed was not nearly as dramatic as those experienced following the Tequila and Asian crises. The tragic events in the U.S. on September 11, 2001 accentuated all these tendencies in international financial markets. All this was very bad news for Argentina, the largest sovereign debtor among emerging countries in the global market. Besides, the country confronted heavy international financing requirements in 2001 and 2002. The international financial markets had started to discrimiinate against Argentina in early 2001 and the spread on its sovereign bonds had started to show a wide gap vis-a-vis fixed income securities from other emerging economies, with the exception of Turkey. With its economic and political fundamentals deteriorating, contagion from abroad was not the main problem for Argentina, and in the flight to quality that took place in the gloomy global atmosphere of 2001, Argentina was singled out. At that point, with domestic economic and political issues the dominating factors shaping up the prospects for the Argentine economy, the market perception was that a sharp increase in the degree of default risk was in the making. All this was promptly reflected in the pricing of Argentine debt in the international market. 5.2 Factors generally subject to government control: Together with the SSAL, the RFSL was providing financing to assist the GOA and its agencies with the implementation of the Letter of Development Policy dated November 2, 1998. Thus, all of the policy actions summarized in the Policy Matrix of the SSAL were factors generally subject to government control. This was an expanded structural reform agenda that expected to consolidate the progress already achieved and move the reform process to a next stage. Social, fmnancial, regulatory, institutional and fiscal issues were included in the Policy Matrix. The single most important item that could not be fundamentally altered was the ConvertibiliVy Plan, which was a constant for the SSAL and the RFSL. In fact, the Bank's support was trying to assist in the efforts to preserve and sustain the Convertibility Plan, which was seen as the main pillar behind the macroeconomic stability that Argentina had enjoyed for almost a decade. Thus, practically all other domestic factors under the government's control were dispensable in the larger effort to preserve the structural progress made previously under the far-reaching macroeconomic program anchored by the Convertibility Plan. From an operational standpoint, the GOA was the Borrower and had the right to cancel the RFSL, independently of any decision that the BCRA could take with respect to the repo facility. 5 3 Factors generally subject to iinplernenting agencv control. Although the RFSL was made to the Republic, the implementing agency and beneficiary for all practical purposes was the BCRA, which also was the manager of the repo facility. Within the context of the Convertibility Plan, the BCRA had a great deal of latitude as an autonomous agency to negotiate with the syndicate of international banks the details and any changes in the repo facility. It was responsible for negotiating terms, financial conditions and amounts with each - 14 -  participating bank and for seeking new participating banks. In this respect, the RFSL gave the BCRA the benefit of the doubt and imposed very few contractual constraints on the actual structure of the repo facility. The use of the financial enhancement provided by the RFSL was much more closely specified in the Loan agreement, but the Bank could not withdraw its financial support once the Loan became effective. In hindsight, a factor that perhaps had a significant bearing on the decision of the BCRA to activate the repo facility and disburse the RFSL when they finally happened was the change in the Presidency of the BCRA that took place in April of 2001. The outgoing President had played an active role in the design and implementation of the repo facility and had been instrumental in convincing the GOA and the Bank of the need to count with the enhancement provided by the RFSL. The change in the Presidency of the BCRA came about in the middle of a strong controversy over the direction of macroeconomic policy. The fact that the President of the BCRA had to leave prior to the end of its scheduled mandate was widely interpreted in the market as a loss of independence of the Central Bank to conduct monetary policy. 5.4 Costs andfinancing: In the end, the total cost of the enhancement for the repo facility amounted to some US$1,005.05 million, of which the Loan 4406-AR contributed slightly more than half. This Loan of US$505.5 million was entirely disbursed. The other costs of the enhancement were covered by a US$500 million loan provided by the IADB as co-financing and which was also entirely disbursed. No government counterpart funds were used under this operation. The financial resources provided by the enhancement were used to meet margin calls and buy additional eligible government bonds. A portion of the funds was not used and remained deposited in the Special Account maintained at the BCRA by the Borrower for the specific purpose of enhancing the repo facility. There were no delays in the implementation of this operation, which had been fully disbursed by Septemberl7, 2001. The original Closing Date for it had been scheduled for September 15, 2003, but since the repo facility and its purpose had disappeared long ago, the closing of the Loan account was formally carried out on January 22, 2003. 6. Sustainability 6.1 Rationale for ststainabilitv rating The sustainability of any achievements generated by this operation that provided a direct enhancement for the repo facility of the BCRA is highlv unlikely in the new macroeconomic scenario prevailing in Argentina. This is particularly so given the abandonment of the Convertibility Plan, the default status of the national government, and the different role of the Central Bank in the new monetary policy framework adopted by the Argentine authorities. A repo facility like the one supported by the RFSL does not have a clear role to play in the context of the adoption of a floating exchange rate system and a Central Bank that has regained an active capacity as lender-of-last-resort. In the current macroeconomic policy environment the priorities seem to be in other directions. From the Bank's perspective, there are no plans for an operation similar to the RFSL in the foreseeable future. All that being said, the success of the enhancement - 15 -  provided to Argentina's repo facility cannot be judged entirely by the lack of sustainability of the early achievements this operation had in helping to prevent a deposit run in the local banking system. Rather, success perhaps should also be assessed in light of the contribution that a Bank operation such as the RFSL, in the form of a stand-by or contingent operation, can contribute to the tool kit of the Bank. This type of lending modality has been rarely used by the Bank in the past, but it could very well be a model worth considering in other situations around the world when standard transaction-oriented loans or partial guarantees could not entirely fit the specific needs of a Bank borrower. But perhaps prior to this the Bank should try to produce a set of clear guidelines to that effect. In the case of the RFSL, with few precedents at hand, the final design was the result of educated ad-hoc decisions. 6 2 Transitioni arrangemenit to regular operations N/A 7. Bank and Borrower Performance Banhk 7. 1 Lending The Bank's performance in the identification, preparation assistance and appraisal of the Loan was satisfactory. The original objectives of the RFSL were consistent with the agenda of the National Government for financial reform and the Bank's CAS. Also, the Loan design relied on the already established repo facility of the BCRA and the proven management capabilities of a highly professional Central Bank. Furthermore, the Bank knew well and had contributed heavily already to the financial reform in Argentinza. The Bank's experience was well documented and it was familiar to the staff that worked on the Loan preparation. The Bank staff presented a comprehensive report (No. 17864-AR) entitled " Argentina Financial Sector Review" on September 28, 1998, which contained a thorough review and assessment of the role of the repo facility in the Central Bank's policy at the time. This was an operation that was prepared in record time, something that was possible due to the quality of the dialogue between the Bank and the Argentine authorities, and to the thorough sector knowledge the staff already possessed on Argentina's financial system and its macroeconomic framework. As already mentioned, the Bank entered into the SSAL and the accompanying RFSL under emergency conditions that were recognized in the pricing and short maturity schedule of these two operations. The Bank felt that lending under the conditions prevailing in Argentina at the time, which were aggravated by unfavorable conditions in the international picture, was a risk worth taking. This was particularly so since the Bank had a lot at stake in trying to defend many years of sound and profound reform work by the GOA, which had brought about an unsurpassed long period of economic growth and stability to the country. Perhaps the main shortcoming in the design of the RFSL was the absence of a direct link between the amount of enhancement available under the RFSL and the actual size committed or disbursed by international banks under the repo facility. Since the concept of leverage was not included in the design of the Bank's support, in the end the repo facility could have become insignificant and the BCRA still could have disbursed the US$1 billion available for enhancements from the Bank and the IADB loans. As it happened, the enhancement package was fully disbursed but the funding generated by the repo facility was just 1.7 times the size of the combined enhancement provided. This was far from the maximum - 16-  potential leverage of 6.4 times available at the time of Loan effectiveness or the 4.8 times at the moment of the facility's activation. 7.2 Stupervision: Supervision of the RFSL was conducted mostly in the context of the SSAL supervision-which was judged to have been highly satisfactory, particularly in light of complexity and broad coverage of that structural adjustment operation-and so the supervision of the RFSL is considered to have been quite satisfactory as well. It was mentioned earlier that from a legal perspective the RFSL as a stand-alone operation deviated from the Bank standard policies (e.g., Bank as a lender-of-last-resort) and procedures (e.g., accepted procurement practices), so the case was made for this operation to be part and justified in the context of the special structural adjustment operation that was approved simultaneously with it. In general, the Supervision Forms 590 were realistic and adequately reported implementation progress. Supervision missions took place every six months. The critical risk rating of this operation was considered to be modest until the June 2001 supervision mission when the perceived risk was raised to substantial, as the prospects for the repo facility from directly averting the potential consequences of a large shock to the banking system were deemed to have diminished. At that time, the mission discussed with the BCRA the convenience of considering a cancellation of the RFSL. The mission estimated that the Central Bank might not be able to engage in repo transactions worth more than a US$1 billion, due to the scarcity of eligible government bonds. The mission also came to the conclusion that repo transactions at that point were hardly worth the effort since the estimated costs were higher than the benefits for the BCRA. The recommendation in mid-2001 was to discuss the facility's prospect with the Minister of Economy (the Borrower) if adverse circumstances continued, as the latter could rightly feel that the repo facility was taking up space in intemational banks' credit risk allocation for the country. The staff was responsive to the Central Bank's changing needs to build a more effective shield to protect local banks against short-lived systemic liquidity crises. Following informal discussions with the BCRA during supervision missions in 2000 on ways to improve the enhancement the Bank was providing to the repo facility, the Bank staff explored back at headquarters the possibility of: a) using the Bank's new guarantee instruments; b) providing a more direct Bank commitment to participating intemational banks, and c) finding ways to extend the tight amortization schedule of the RFSL. In the end, despite the efforts made to accommodate the evolving needs of the BCRA, the Bank enhancement to the facility kept its original design. 7 3 Overall Bank perJbormnance. Given the satisfactory lending and supervision ratings, the Bank's overall performance can also be considered satisfactory Borrower 7.4 Preparatuon. The GOA preparatory work for the RFSL was satisfactorv and took place mostly in the context of the negotiations of the SSAL and the corresponding Letter of Development Policy, which faithfully reflected a broad degree of "ownership" of the proposed measures assisted by - 17 -  that structural adjustment operation. The preparation and establishment of the repo facility had already been done by the BCRA in late 1996, so there was a strong stakeholder commitment from the beginning. The Bank basically took the work already done by the BCRA on this matter and from there it negotiated a support operation that could best meet the Bank's requirements and constraints. The BCRA made strenuous efforts to try to get the standard terms and conditions for the RFSL and in this respect had the sympathies of the Region, although in the end emergency terms and conditions prevailed given that the RFSL was preventive and disbursements were expected only if there was a severe financial crisis. 7 5 Government uinplementation performance: The direct participation of the GOA in the implementation of the repo facility was minimal until the time of its activation. Its relationship with the repo facility and the RFSL had to do with the macro picture, that is, with delivering the correct set of economic policies so as to reduce the perceived Argentine risk in global financial markets. Without meeting that precondition, the repo facility and the enhancement provided by the RFSL could not do a good job at smoothing out short-lived systemic liquidity crises. In hindsight, it is clear that the GOA failed to maintain a viable macroeconomic and political framework able to keep Argentina's debt spread from skyrocketing. Among its failings was pushing for the use of the repo facility under the wrong circumstances and for the wrong reasons. Even prior to that, it had compromised the market perception of an independent and technically guided Central Bank policy. It is from this perspective that the National Government's implementation performance under this operation can be considered unsatisfactory. 7.6 Implementing Agency. From a practical standpoint, the BCRA was the implementing agency for this operation, even though the Republic was the Borrower and the owner of the Special Account. Positive comments have already been made about the Central Bank's management and financial effectiveness. From this perspective, its implementation performance under this operation is considered to have been satisfactory. In order to comply with Section 3.02 of the Loan Agreement, the BCRA requested from the General Auditor of the Nation (AGN) a formal audit of the Deposit Account opened at the central bank for the purpose of transactions financed with money from the IDB and World Bank loans and that went to support the Repurchase Facility. The audit of the AGN covered the active period of the Repurchased Facility (Sept 18, 2001 to March 19, 2002). The AGN issued an unqualified opinion, which was acceptable to the Bank in form and substance. The AGN report concluded that the financial statements of the Deposit Account for the mentioned penod reflected reasonably the actual balance and transactions undertaken, being in accordance with normal auditing principles and meeting the requirements established in the Loan Agreement No. 4406-AR of July 22, 1999. As to the actual decision to activate the repo facility, this is considered to have escaped from the effective domain of the Central Bank since when it was made it became in the context of a broader political decision, where technical considerations on the convenience of activating the facility became secondary. While it remains hard to entirely separate responsibilities, it appears in hindsight that the key decision originated from the national economic authorities. 7 7 Overall Bon7 owe, petformance: A persistent track record of deficient public action and policy inaction on crucial but politically intractable matters like the fiscal issues, including those with the provincial governments biased the macro performance of the GOA and contributed heavily to the final - 18 -  untenability of the Convertibility Plan. One of the consequences of its abandonment was the loss in sustainability of the early achievement brought about by the policies sponsored by the BCRA to invigorate its weak functions as a lender-of-last-resort, of which the repo facility was one of the principal components. It is in this context that in hindsight the overall performance of the Borrower is considered to have been unsatisfactory. 8. Lessons Learned Several of the lessons derived from the execution of the RFSL are implicit in the earlier text. This section summarizes them and highlights those of more general applicability and relevance for similar projects. o As in the case of the preparation of the SSAL, the RFSL shows that the Bank has developed a solid capacity to respond rapidly and effectively to the needs of its client countries in times of distress due to unexpected turbulence in international financial markets. o There is room for stand-by loans as crisis-preventing tools in the arsenal of financial instruments of the Bank. Standby lending operations can be better suited than crisis-reactive loans under some circumstances, particularly when the purpose of the former is to prevent crises from developing in the first place, as the case was with the RFSL. However, as with other forms of financial support, contingent lending by the Bank would certainly benefit from formal, explicit guidelines, which currently do not exist. The RFSL was prepared under a severe time constraint and could have benefited from such guidelines. o While emergency lending. such as that provided by the RFSL. should include terms and conditions that do reflect the riskier credit conditions under which is made, it should also take carefully into account the future ability of the receiving country to service such debt. Indeed, as currently prevailing macro conditions in Argentina illustrate, the higher pricing and shorter amortization schedules of such emergency lending can contribute heavily to the debt-burden problem faced by the borrower if the crisis lasts longer or ends up being more severe than originally expected. This speaks in favor of the Bank staff conducting a detailed and formal assessment, as part of the loan appraisal process-including a thorough sensitive analysis, considering altemative likely scenarios-to determine the ability of the country recipient of emergency lending to timely service its public debt in the event of an actual disbursement of these shorter and 'more expensive Bank credits. o Altemative disbursement scenarios of transaction-based standbv/contingent loans require to be carefully analyzed. Not only should their contribution as a preventive instrument be assessed during appraisal. Indeed, even if such standby operations are a complement to policy-based loans-such as the RFSL was in its preventing function-the appraisal of standby operations should include a careful analysis of likely transaction scenarios, including an assessment of the Bank's marginal contribution under each one of them. Such an exercise could help with improving the final design of the proposed operation, introducing palliatives for preventing some of the major risks involved, particularly if the standby operation is likely to be disbursed under critical or emergency conditions. o Financial enhancements provided by the Bank have to be commensurate with the financial leverage is likely to generate. In the case of the RFSL, the lesson is that the amount disbursed by the Bank should have been probably tied to the size of the repo facility and/or to the amounts effectively disbursed by participating intemational banks. While it was important for the Bank to make a firm contractual commitment that the RFSL was going to be available to support the repo - 19 -  facility and its participating international banks, the Bank commitment could have been conditioned by specific financial covenants that limited the amounts it was willing to disburse under different activation scenarios. Such financial covenants could have, for example, introduced the concept of a leveraged contribution by the Bank and protected it from any ex-post unexpected or undesirable reformulation in the terms and conditions of the repo facility. 9. Partner Comments (a) Borrower/inmplementing agency The BCRA prepared its own final evaluation report of the Project, an unedited copy of which appears below: BCRA Final ReDort - Loan 4406 AR Background The loan provided by The World Bank was an important enhancement for the Repo Facility and the analysis can be made in two steps: before and after the execution of the facility. When the loan was approved by the Board of The World Bank, and at the same time a similar loan was provided by the IADB, the Repo Facility was strongly improved. Before having the support of The World Bank most of the counterparties of the Repo Facility had expressed the intention to terminate their commitment. That situation changed when the loan was approved and BCRA was enabling to continue with the Repo Facility and also to replace part of the commitments that had been already terminated. The enhancement provided by the loans of the IADB and WB reduced the risk borne by each participating bank in the syndicate, and as a consequence of that the capital allocated to this facility was lower, allowing them to continue rolling over their participation in the facility with the same level of fees agreed in 1996. Therefore, the total amount of the facility remained stable in a level of USD 4.75 billion during the last two years. Execution of the Repo Facility and the Contingent Loan During the first semester of 2001 the Argentine financial system started to face liquidity problems. At the beginning of January the total of deposits was $ 87.7 billion, this amount dropped to $ 85.1 billion at the end of June and then it decreased dramatically. At the middle of August the deposits were $ 75.8 billion. In a month and a half the deposits had dropped $ 9.3 billion, equivalent to 11% of the total deposit base. That situation forced the Central Bank to provide liquidity to the banks by taking different decisions. First, it decided a reduction of the reserve requirements from 20 % to 15 % for time deposits. At the same time, it provided liquidity to the Banks through repo and rediscounts transactions using the available flexibility of the convertibility -20-  system. Nevertheless, these measures were not enough to stop the deterioration of the liquidity of the argentine financial system. Under this situation the Argentine Government agreed with the IMF a support program. At the same time, it was considered convenient to get the supplementary support from the international private banks. Due to the fact that the international capital markets were virtually closed for Argentina, the only instrument available was the Contingent Repo Facility through which the private international financial entities could provide funds to be used only to meet the liquidity needs of the banks and not to cover the fiscal deficit. Therefore, it was expected a change in the expectations about Argentina, by a liquidity injection through de execution of the Repo Facility and the funds provided by the IMF. So, on September 2001, the Board of Directors of The Central Bank of The Republic of Argentina, decided to call on the facility which was executed on September 12, 2001. In order to pay the margin calls, The Central Bank decided to use the contingent loan of the IDRB and the IADB for US$ 1,0 bn. The execution of the Repo Facility allowed the Central Bank to get funds at an interest rate of approximately 5.5 % p.a. that was very low, considering the fact that the credit markets were closed for Argentina. Nevertheless, it should be pointed out that the execution of the Repo Facility overlapped with the terrorist attack on the United States. This dramatic event determined an increase of the risk aversion in the financial markets that prevented the expected change of perception about Argentina, not to happen. The maximum amount executed was $ 1,772.2 millions, part of it was rolled over on December Ilth for $ 1,350 millions. Since the first transaction was executed until the abovementioned roll over the market value of the securities felt about 40%. To keep the market value of the bonds, including the required 20% haircut, $ 495 millions of face value of additional bonds and $ 651.6 million of cash were delivered in margin calls. According to the legal documentation signed with the Banks, the BCRA had to deliver only cash when the market value of the original securities sold was lower than the amount of the received funds. The reason of using only a portion of the Repo Facility, was that the market value of the securities available by the Central Bank had dropped dramatically in the last months, and as a consequence of that the capability of this Institution to get a higher funding from the facility was strongly reduced. The support of World Bank and IADB to the Repo Facility was very important to allow BCRA to fulfill its obligations avoiding to commit any event of default At the beginning of January 2002, some of the participating banks expressed their understanding that an event of default had occurred because of the address of the President of the country delivered in the Congress at the end of December 2001. According to the legal documentation there could be a cross default with the Republica - 21 -  and, as consequence of that the participating banks had the right to close the transaction asking the Central Bank an early prepayment. In case that the Central Bank could not pay the amounts disbursed the Banks had the right to sell the delivered securities in the market at the price they could get. In order-to avoid the important looses arising out of a massive sell of bonds (USD 4,800 millions of face amount) that would have caused a dramatically drop in their prices, this BCRA made clear its intention to close the outstanding transactions. Conclusions The Repo Facility worked as a very important tool that allowed the financial sector to face the turmoil of the economy. As the chart attached shows, the decrease of the level of deposit was lower than the dramatically increase of the country risk. Anyway, the political and economic conditions in Argentina continued deteriorating and the facility had to be executed as we described above, providing funding until a cross default event was declared by some participating banks. Once executed, the funds disbursed by World Bank and IADB allowed the BCRA to cover the margin calls and to repurchase transactions minimizing the impact in the/ intemational reserves. Furthermore, the fulfillment of all its obligations allowed to the BCRA not to commit any event of default allowing, after the floating exchange rate policy was adopted, to issue short-term bonds (the "Lebac") to carry out an open-market monetary policy, being one of the few financial instruments in the Argentine economy highly accepted by local investors. (b) Cofinanciers IDB. No comments were received to an earlier draft of this document. (c) Other partners (N! GOs/private sector): N.A. 10. Additional Information - 22 -  Annex 1. IKey Performance Rndicators/Log IFrame Matrix Not applicable in this case. However, a detailed list of the performance indicators tracked by the sister SSAL operation and their outcome and impact, is included in Annex I of that operation's ICR. - 23 -  Annex 2. Project Costs and Financing Project Cost by Component (in US$ million equivalent) Appraisal Actual(Latest Percentage of Estimate Estimate Appraisal Component US$ million US$ million 1. Support for the Repo Facility at the BCRA 500.00 500.00 100 2. Capitalized Up-front Bank Fee (1%) 5.05 Total Baseline Cost 500.00 505.05 Total Project Costs 500.00 505.05 Total Financing Required 500.00 505.05 Project Financing by Component (in USS million equivalent) Percentage of Appraisal Component Appraisal Estimate ActualLatest Estimate Bank Govt. CoF. Bank Govt. CoF. Bank Govt. CoF. Support for the Repo 500.00 495.00 Facility at the BCRA Capitalized Up-front Bank 5.05 5.00 Fee (1%) Total Project Costs 505.05 500.00 Co-financing by the Inter American Development Bank under their Loan AR-0254 - 24 -  Annex 3. Economic Costs and Beneflts Not applicable - 25-  Annex 4. Bank Inputs (a) Missions: Stage of Project Cycle No. of Persons and Specialty Performance Rating g (e.g. 2 Economists, I FMS, etc.) Implementation Development Month/Year Count Specialty Progress Objective Identiflcation/Preparatlon 09/22/1998 1 Task Manager* S S Supervision 05/07/1999 1 Task Manager S S 11/19/1999 1 Task Manager S S 06/02/2000 1 Task Manager S S 06/05/2001 1 Task Manager S S ICR 03/10/2002 1 Consultant *Identification/Preparation. Identification/Preparation performed jointly with the SSAL (b) Staff: Stage of Project Cycle Actual/Latest Estimate No. Staff weeks US$ ('000) Identification/Preparation 1.9 5.1 Supervision 4.5 23.7 ICR - 1.4 7.0 Total 7.8 35.8 - 26 -  Annex 5. Ratings for Achievement of Objectives/Outputs of Components (H=High, SU=Substantial, M=Modest, N=Neghgible, NA=Not Applicable) Rating Macro policies O H OSUOM O N O NA 3Sector Policies O H OSUOM ON O NA F Physical O H OSUOM O N O NA ? Financial O H OSUOM O N O NA s Institutional Development 0 H O SU O M 0 N 0 NA EEnvironmental O H OSUOM ON ONA Social S Poverty Reduction O H OSUOM O N O NA 21Gender OH OSUOM O N ONA Oi Other (Please specify) O H OSUOM O N O NA