Document of The World Bank FOR OFFICIAL USE ONLY Report No. 20688 PERFORMANCE AUDIT REPORT PHILIPPINES FIRST AND SECOND RURAL FINANCE PROJECTS (LOANS 3356, 3938, 3939 and 3940-PI) June 30, 2000 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 (annual averages) Currency Unit = Peso (P) Abbreviations and Acronyms ACPC Agricultural Credit Policy Council ALF Agricultural Loan Fund BSP Bangko Sentral ng Pilipinas (the Central Bank) CLF Countryside Loan Fund DOF Department of Finance ERR Economic rate of return FRR Financial rate of return GOP Government of the Philippines ISAP Institutional Strengthening Action Plan LBP Land Bank of the Philippines OED Operations Evaluation Department PCIC Philippine Crop Insurance Corporation PCR Project Completion Report PFI Participating financial institution QGFB Quedan Guarantee Fund Board QUEDANCOR Quedan and Rural Guarantee Corporation SAR Staff Appraisal Report Fiscal Year Government: January 1- December 31 Director-General, Operations Evaluation Mr. Robert Picciotto Director, Operations Evaluation Department : Mr. Gregory K. Ingram Manager, Sector and Thematic Evaluation : Mr. Ridley Nelson (Acting) Task Manager : Mr. Madhur Gautam FOR OFFICIAL USE ONLY The World Bank Washington, D.C. 20433 U.S.A. ROBERT PICCIOTTO Director-General Operations Evaluation June 30, 2000 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: PERFORMANCE AUDIT REPORT ON PHILIPPINES FIRST AND SECOND RURAL FINANCE PROJECTS Loans 3356, 3938, 3939 and 3940-PH Attached is a combined Performance Audit Report prepared by the Operations Evaluation Department (OED) for the above two projects. The First Rural Finance Project (RF-I) was supported by Loan 3356-PH for US$150 million. The Second Rural Finance Project (RF-1I) was supported by Loans 3938-PH, 3939-PH and 3940-PH for a total of US$150 million. RF-I was approved on June 21, 1991, and fully disbursed on June 8, 1994, substantially ahead of the original closing date of March 31, 1997. RF-II was approved on September 14, 1995, and almost fully disbursed by the end of 1999, well before the anticipated closing date of June 30, 2002. Both projects had the same basic objective: to expand the volume of longer-term commercial credit for agricultural and rural development. Neither project had poverty alleviation as an explicit objective, although both anticipated poverty reduction impact indirectly through job creation in rural areas. Both projects also had institutional development objectives, but RF-I was broadly focused and included a number of institutions involved in rural/agricultural finance, while RF-II was more narrowly focused on the Land Bank of the Philippines (LBP), the borrower and the executing agency for both loans. The projects largely met their major objective of providing longer-term credit for productive investments in rural areas (i.e., anything outside of Metro Manila). Under both projects, small amounts had to be reprogrammed because they were not disbursing. Under RF-I, credit intended for cooperatives was undercut by cheaper funds from other donor sources. Under RF-II, funds for LBP retail lending disbursed only very slowly, even after terms were modified. In general, however, credit disbursement was very rapid and well ahead of schedule under both projects - so rapid that concerns were raised about whether interest rates on Bank funds were in fact at market levels and whether Bank funds were simply substituting for funds from other sources. Patterns of disbursements, especially under RF-I, suggest that substitution may have occurred when rates on Bank funds deviated from market rates due to lags in the adjustment of rates on Bank funds, but this problem was largely eliminated under RF-II as the adjustment formula was modified to reduce lags. The OED audit devoted specific attention to the question of additionality, but the findings were inconclusive. Implementation of non-credit components of the two projects was more mixed. While the institutional strengthening of LBP under both projects was successful, under RF-I the strengthening of government institutions providing insurance for agriculture and guarantees for loans failed, while some required studies were not carried out. However, these failures did not significantly impede the main accomplishment of meeting the credit objective, suggesting that these components were in fact peripheral, and, indeed, RF-II focused more narrowly on LBP. 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. -2- June 30, 2000 On the other hand, continuing concerns with non-credit components related to conditionalities designed to further financial liberalization and to assure the long-run viability of LBP appear more serious. These problems have led the audit to be less optimistic than the ICR for RF-II about institutional development and project sustainability. The audit sees sustainability as likely, rather than highly likely as seen by the ICR. While the government has taken several steps that are beneficial to the financial and institutional health of LBP, there remain potential concerns with adverse policies such as below market interest rates on competing funds and the Government's failure to take formal steps to permanently remove the unfairly burdensome dividend payments. The institutional development of LBP was clearly substantial, but the same cannot be said for other Government agencies, leading the audit to give a modest rating rather than the substantial rating given by the ICR. In fact, the audit is specifically concerned by the apparent lack of appreciation by the Government that financial liberalization extends beyond simply removing controls over interest rates to removing barriers to a flexible and competitive financial sector, and especially to opposing subsidized interest rates that can disrupt financial markets and interfere with a level playing field for financial institutions. The PCR for RF-I and the ICR for RF-II find both Bank and borrower performance to be satisfactory, and the OED audit agrees. Nonetheless, behind these overall ratings, the OED audit finds wide variations. Specifically, the Bank deserves especially high praise for continuity in project staffing and for the ability of Bank staff to work very effectively with LBP staff in promoting project objectives and LBP development. On the other hand, the Bank was far less effective in working with other institutions to attain project objectives or, more importantly, in focusing on some policy issues that can ultimately have a major affect on sustainability. In addition, the Bank's handling of the monitoring of job creation and environmental protection could have been better. Similarly, borrower performance was highly satisfactory in reference to LBP but quite mixed with respect to other government agencies. The failures in performance by peripheral agencies are of little concern, but basic issues with respect to interest rates and promoting the long-run viability of LBP, not to mention concern with foreign exchange losses, may be important indications of the lack of a sufficiently broad view of financial liberalization that could ultimately endanger institutional development and project sustainability. Whatever shortcomings the two projects may have had, the main objective of increasing access to longer-term credit on a market basis for private-sector investments in areas outside of Metro Manila was clearly facilitated by RF-I and II, along with substantial strengthening of LBP's operational capabilities. The OED audit thus concurs with the PCR's and ICR's outcome ratings of satisfactory for both projects, with some reservations as noted in the audit. The main lessons from RF-I and II follow directly from the foregoing. Project focus should be as narrow as possible to achieve project objectives avoiding extraneous and difficult to implement components, but broad enough to include elements that may be important for long-run sustainability. Financial liberalization is not just decontrolling interest rates but also involves attention to policies fostering competition and innovation and promoting a level playing field for financial institutions. Such objectives as poverty alleviation and, in the absence an independent and effective monitoring system, environmental protection may exceed the reach of a project designed to meet market-based, demand-driven credit needs. Attachment 1 Contents Principal R atings...........................................................................................................................iii Preface.............................................................................................................................................v 1. Introduction/Background ................................................................................................... 1 2. T he Projects ............................................................................................................................. 1 Objectives......................................................1 Design of Components.... ..................2....... ................2 Costs..... ........ ................................................ 4 3. Im plem entation........................................................................................................................ 5 4. Findings .................................................................................................................................... 7 Credit Components ........................................ ........ 7 Non-Credit Components ....................................... ..... 10 Additionality..................... ............................... 11 5. R atings....................................................................................................................................13 Outcome...................................................... 13 Institutional Development Impact ......................................... 14 Sustainability ............................................ ............ 15 Bank Performance ...................... .................... 16 Borrower Performance ..............I.....7.. .................. 6. Lessons.................................................................................................................................... 18 T ables ............................................................................................................................................ 21 Annex A: Basic Data Sheets ..................................................................................................... 27 Annex B: PFIs and Sub-borrowers ....................................................................................... 31 Annex C: Land Bank of the Philippines: Detailed Balance Sheet.......................................35 Annex D: Comments from the East Asia Region......................................................................36 Annex E: OED response to East Asia Region's comments..................................................38 Annex F: Borrower's Comments............................................................................................ 39 Annex G: OED response to Borrower's comments.............................................................48 This report was prepared by Robert C. Vogel (consultant). Madhur Gautam was the task manager. William B. Hurlbut edited the report. Marcia Bailey provided administrative support. 111 Principal Ratings Rural Finance I PCR Audit Outcome Satisfactory Satisfactory Sustainability Likely Likely Institutional Development Modest Modest Borrower Performance Satisfactory Satisfactory Bank Performance Satisfactory Satisfactory Rural Finance 11 ICR Audit Outcome Satisfactory Satisfactory Sustainability Highly Likely Likely Institutional Development Substantial Modest Borrower Performance Satisfactory Satisfactory Bank Performance Satisfactory Satisfactory Key Staff Responsible Rural Finance I Task Manager Division Chief Country Director Appraisal A. Chupak D.M. Dowsett-Coirolo G.S. Kaji Midterm A. Chupak Completion A. Chupak J. Gutman C.E. Madavo Rural Finance II Task Manager Division Chief Country Director Appraisal A. Chupak J. Gutman C.E. Madavo Midterm A. Chupak Completion A. Chupak G. Fox V.K. Bhargava V Preface This is a combined Performance Audit Report (PAR) for the Rural Finance Project and the Second Rural Finance Project (henceforth referred to as Rural Finance I and II projects, RF-I and II) in the Philippines. Loan 3356-PH in the amount of US$150 million was approved for RF-I on June 21, 1991. The loan was fully disbursed on June 8, 1994, substantially ahead of the original closing date of March 31, 1997. Loans 3938, 3939, and 3940-PH for a total amount of $150 million were approved September 14, 1995, for RF-II. The project will close by June 30, 2002, and has been almost fully disbursed since early 1999 (only part of the reduced retail component remains). The PAR is based on a review of the Staff Appraisal Reports; the Loan and Guarantee Agreements; Bank Supervision reports; correspondence between the Bank and the borrower; internal Bank memoranda; and field work by an OED mission in March, 1999. OED is grateful to the management and staff of the Land Bank of the Philippines for their excellent cooperation and assistance in preparing this PAR. Following standard OED practice, a draft of this report was sent to the World Bank's East Asia Region and the borrower for comments. The report has been modified in the light of these comments. However, some areas of disagreement remain between OED and the East Asia Region. The Region's views, in the form of comments on the final draft of the report, are attached as Annex D, and OED's response is attached as Annex E. The borrower's comments are attached as Annex F and a detailed OED response to these comments is attached as Annex G. 1 1. Introduction/Background 1.1 The First and Second Rural Finance Projects (RF-I and II) are part of a long series of such projects in the Philippines, directly preceded by the Agricultural Credit Project in the late 1980s and followed by the Third Rural Finance Project, approved in late 1998. Justifications for all projects have remained remarkably similar. The importance of the rural and agriculture sector is consistently emphasized: for overall economic growth; for export enhancement (and import reduction); and to generate employment and thereby reduce poverty in rural areas where most of the Philippines' poorest people are located.' 1.2 The Staff Appraisal Reports (SARs) for each project cite a number of macroeconomic problems that have impeded the performance of the Philippine economy, together with problems specific to the agricultural sector, with particular attention to limitations involving rural finance. The reports point out that the participation of the banking system in rural finance has been limited, especially in long-term financing for investment. They attribute this to such causes as high risks in agriculture, lack of collateral acceptable to banks by all but the largest rural economic units, the weakness of the Philippine rural banking system, and even to government and donor programs of subsidized and targeted credit that have limited the ability of commercial banks to compete on a sustainable basis. 1.3 Each ensuing SAR states that the project will carry on with the achievements of the preceding projects but, at the same time, learn from the shortcomings of those projects. Nevertheless, there have been significant changes from one project to the next, some due to changing Philippine circumstances and others to the Bank's changing policies governing financial intermediary lending. For example, RF-I coincided with a major shift from an Agricultural Loan Fund (ALF) within the Central Bank of the Philippines (now Bangko Sentral ng Pilipinas, or BSP) to a Countryside Loan Fund (CLF) within the Land Bank of the Philippines (LBP). From RF-I to RF-Il, the focus narrowed sharply to deal almost exclusively with the LBP in terms of both operations and conditionalities. 2. The Projects Objectives 2.1 The overriding objectives for both RF-I and II are to expand the volume of commercial credit for agricultural and rural development, with some attention to job creation in rural areas. Reducing rural poverty is not an explicit objective, although it is used to justify the projects' design in two instances: * For both projects, the same argument is used to justify using Bank loan funds to finance a substantial portion of local currency costs (57% in RF-I and 35% in RF-II): "because of the Philippines' current economic situation and the poverty alleviation objectives of the project." 1. For the Third Rural Finance Project, distress caused by the Asian crisis has been added as a major justification. 2. ALF was also moved to LBP. 2 * In discussions of the SAR for RF-I (May 13, 1991), it was noted that poverty alleviation would be addressed, in general, by job creation in rural areas through activities financed under the project and, more specifically, by improvements in the operations of the Philippine Crop Insurance Corporation (PCIC) and the Quedan Guarantee Fund Board (QGFB), the rehabilitation of rural banks, and the upgrading of agricultural credit cooperatives through credit and training. 2.2 Discussions with LBP officials during the PAR mission revealed that poverty alleviation was not considered a major objective for RF-I or II, and project implementation deviated significantly from any implied poverty alleviation objectives. Thus, while LBP regularly collected and reported the number of jobs created from each sub-loan, including through an end- use survey for a sample of the sub-loans, the audit mission was informed that little effort was expended on verifying the validity of the numbers reported by the sub-borrowers or the PFIs. 2.3 Both projects also have secondary institutional development objectives with related conditionalities, but with a narrower focus in RF-II. RF-I includes support for PCIC and QGFB to enhance loan guarantees, and a special credit line and training to support the development of agricultural/rural credit cooperatives. It also includes BSP/Monetary Board policy reforms to liberalize bank branching and merging, and especially to promote rural bank rehabilitation, and requires that two policy studies be carried out by the Agricultural Credit Policy Council (ACPC) in the Department of Agriculture. 2.4 Perhaps because of the difficulties in achieving these broader secondary objectives in RF- I, RF-II is more narrowly focused on LBP credit lines and on reforms and institutional development related specifically to LBP, the executing agency for both projects. The policy reform objectives are related primarily to protecting LBP from potential losses stemming from the requirement of LBP to pay dividends to the Government of the Philippines (GOP) based on stated profits rather than real profits (adjusted for proper loan loss provisioning and for inflation); from its responsibilities for agrarian reform bonds; and from political pressure to finance subsidized credit lines for agrarian reform beneficiaries.4 The institutional strengthening objectives for LBP cover a wide range of elements, but with a central focus on reducing administrative costs and improving recovery on loans to agrarian reform beneficiaries. Design of Components 2.5 The central component of both projects is the credit line, which is designed to be undirected with the only restriction that the activities financed be in rural areas, defined as any geographic location outside of Metro-Manila. The sub-loans were also designed to be at market- determined rates, although the lags in adjusting the rates created some temporary disconnects between the project fund rates and the true market rate. 3. Most of the specifically mentioned poverty alleviation activities did not occur. Of the institutions to be strengthened for the benefit of the poor, only the rural bank rehabilitation took place, but in a manner and with results that were less than anticipated. 4. As indicated below, the LBP dividend issue is resolved only on an ad-hoc year to year basis. As noted in LBP's comments (Annex F), the specific payments for 1998 and 1999 have only recently been resolved (June, 2000) between DOF and LBP in a meeting that included World Bank staff. 3 2.6 While the design of the credit components of the projects is generally sound, three issues deserve some comment. The most important of these is achieving interest rates that are continuously at market levels, given that rates are fixed for a period of time by a formula that has a lag (adjustments in the formula have reduced the lag; and the appropriate reference rate, which was a problem earlier, appears to have been resolved satisfactorily). This design issue is reflected in a rapid but erratic disbursement of funds, which appears to be related to the relationship between market rates and project rates, with a parallel concentration of credit in relatively large loans.s 2.7 Other design issues with respect to the credit line are who should assume the foreign exchange risk on Bank funds and through what mechanism (which initially seemed unimportant but has now become an important cost issue, see paras. 2.10-2.11) and whether LBP should retail as well as wholesale funds provided by the Bank. Under RF-I there was no LBP retailing, but in RF-H, at the insistence of the GOP, the Bank designed a retailing component for LBP, but with conditions so unattractive that almost no funds disbursed for almost two years6 Under RF-I, the pilot credit line for agricultural/rural cooperatives never disbursed and was reprogrammed to the general credit line, with the apparent design failure attributed to the availability of cheaper funds for these cooperatives from alternative sources (e.g., GOP and international donors). As noted above and discussed in paras. 2.10-2.11, the underlying interest rate policy issue might have been given greater attention. 2.8 Several secondary components in RF-I were not completed (e.g., PCIC and QGFB strengthening), and the ACPC studies were undertaken several years after the closing of RF I; similar components were not included in RF-II. This raises two questions: first, were they not completed because of design flaws or because of faulty implementation and, second, were the components important for achieving the main objectives of the projects? On the first question, the failure to implement the components under RF-I reflects faulty design: LBP had no leverage on the other institutions to ensure implementation, nor did the other institutions have an incentive to pursue them. The underlying policy issues should have been subject to clear and enforceable conditionalities, or other mechanisms included in the project design to ensure progress towards project objectives. On the second question, considering that poverty alleviation was clearly not the top priority, these secondary components do not seem to have significantly impeded the implementation of RF-I. Further, their exclusion from RF-II implies that they should not have been included in the first place. 2.9 Institutional strengthening of LBP and the training for it and the participating financial institutions (PFIs) are given greater attention in RF-II than in RF-I. It is not clear from the SAR for RF-II whether this is due to the shift in focus away from trying to deal with a range of institutions (not very successfully under RF-I) toward what is clearly the key institution for project success, or to shortcomings detected in LBP performance under RF-I, or to some vision of an enhanced role for LBP in Philippine rural finance. The training and technical assistance areas enumerated for LBP and for the PFIs (and for the agricultural cooperatives that participate in LBP 5. It is difficult to resolve the interest rate design issue perfectly given the objectives of pricing funds at market rates, yet disbursing quickly, combined with the costs to the institutions involved of selecting, monitoring, and evaluating in even the most "untargeted" credit projects (as those under consideration are). In the present case, the more substantial issue is what to do about interest rate design (and policy) when the GOP and certain international donor agencies are pricing funds below market on potentially competing lines. 6. RF-II would have closed ahead of schedule except that the decision on restructuring the LBP retail component took some time. However, decisions on reprogramming were made in time to ensure that the project remains on schedule. 4 agrarian reform lending but not in Bank-funded lending) seem well thought out and appropriate. In this respect, it is also worth noting the emphasis given in the SAR for RF-Il on the need to protect LBP not only from the policy threats already mentioned (e.g., its potentially costly role in the agrarian reform process and in agrarian lending) but also from a law requiring state-owned banks to pay half of their profits to the GOP as dividends without adjustments for inflation or for adequate provisions for problem loans. Thus, the greater emphasis on institutional strengthening for LBP in the design of RF-Il may be seen as a substitute for broader policy reform, in addition to being beneficial in itself. Costs 2.10 For both projects, costs are almost entirely related to implementing the credit (see the key project data tables in Annex A). PFIs are expected to contribute 10% to the costs of the sub- projects financed, with sub-borrowers to contribute 15%. Both the PFIs and the sub-borrowers substantially exceeded their expected contributions in RF-I and RF-Il. This signifies greater leveraging of Bank funds, as more (larger) sub-borrower activities are being carried out under both projects. Nonetheless, there are also negative cost aspects, specifically the increased costs (especially for the GOP) stemming from having assumed foreign exchange risks given the impact of the Asian crisis. Under both projects, foreign exchange costs are assumed almost entirely by the GOP, through the Department of Finance (DOF). These were to be covered by a foreign exchange risk cover fee, as determined by a market-based formula, paid by LBP to DOF for the Peso denominated sub-loans. The underlying argument being that, with market-determined interest rates and a capital account that is essentially open, such a market-based formula provides reasonable foreign exchange risk cover.7 This has not turned out to be the case as the exchange rate has depreciated by more than the spread is likely to cover (and LBP and DOF seemed unconcerned/unaware that this involved heavy costs to be borne by the government).8 The ICR for RF-II considers the impact of the Asian financial crisis and notes that the foreign exchange risk cover fee will not be sufficient to cover the foreign exchange losses to GOP. However, the net financial impact on the GOP, taking into account the guarantee fee and the gross receipts tax income (from LBP and other PFIs) is likely to be still positive, but significantly lower than anticipated. 2.11 Costs to LBP for training under the projects are thoroughly detailed in its quarterly reports and, in line with project estimates, are small relative to the amounts of credit involved and to the revenues earned by LBP from the projects. Nonetheless, there has been no analysis of costs relative to the different training activities and to their possible benefits? Another neglected aspect 7. The argument is more carefully "hedged" in the SAR for the RF-II than for RF-I (while the PCR for RF-I is silent on the issue, having been written before the crisis). Nonetheless, knowledgeable Filipinos had been arguing since the mid- 1990s that the Philippine foreign exchange rate was seriously overvalued due to faulty monetary and exchange rate management in the face of substantial capital inflows. 8. Both LBP and DOF were asked during the PAR mission to prepare estimates of future foreign exchange costs, using the current exchange rate for the estimate (based on developing a simple model that can vary the estimate as the exchange rate changes). 9. For example, the number of people trained relative to costs of each training activity could be measured, and attempts could be made to measure benefits by comparing performance of different PFIs relative to their participation in different training activities. Furthermore, there is no concern for the costs incurred by the PFIs in having their employees participate, which could provide a "market test" of the value of training and perhaps lead toward post- project training sustainability. Instead, LBP staff lamented that they could not cover all participant costs. In addition, a significant percentage of training was for LBP staff on environmental issues, resulting from the failure of Philippine agencies charged with environmental responsibilities to fulfill them. 5 of costs is that, in calculating the profitability of the two projects for LBP and the PFIs, the full impact of the project on LBP's operating costs for the credit lines, or to the added administrative costs that PFIs may have incurred in using the credit lines, has not as yet been considered.' The profitability of the two projects is likely to be reduced once full cost accounting is in place (LBP is currently in the process of putting in place a transfer pricing mechanism). 3. Implementation 3.1 Implementation of the credit components of both projects has been highly satisfactory, while implementation of the non-credit components has been mixed. Most shortcomings in implementation of the non-credit components can be attributed to design flaws, in particular, unrealistic assumptions about the willingness or ability of institutions other than LBP to fulfill responsibilities ascribed to them, especially in RF-I. Supervision reports for both projects only infrequently make negative comments or give lower ratings on the implementation aspects of the projects. 3.2 The Project Completion Report (PCR) for RF-I mentions a gap between signing and effectiveness due to delays in the Monetary Board's approval of the liberalization of bank branching in rural areas and the rural bank recapitalization program, but disbursements under RF- I were subsequently well ahead of schedule. The PCR for RF-I also notes the need to reallocate funds originally programmed for lending to cooperatives because of competition from cheaper (subsidized) funds available from other sources, but this can be attributed more to design than to implementation. Similarly, disbursements under RF-II were also well ahead of schedule for the wholesaling component of the credit line, although scarcely any funds programmed for retail lending by LBP were disbursed until early 1997. Subsequently, the retailing component was restructured by reallocating the majority of the funds earmarked for retailing to the wholesale component and significantly modifying the conditions for retail lending. At the close of RF-II, the reduced retail lending component was almost fully disbursed. 3.3 Table 1 (all tables referred to in the main text can be found at the end of the report, starting on page 19) provides information about interest rates, inflation rates, and disbursements under the two projects. In addition to rapid disbursement, the table shows a tendency for disbursements to be larger during quarters in which the pass-on interest rate to PFIs was below the T-Bill rate, especially under RF-I, but to be less closely related to variations in the pass-on rate relative to the rate of inflation (the "real" interest rate)." This suggests that PFIs were especially eager to access funds from LBP when these were a cheaper source for lending than selling T-Bills, whereas the timing preferences of sub-borrowers, who might be expected to be more affected by real interest rates, appear to be a less important influence. Nonetheless, there is also a tendency for sub-loan sizes to be larger during periods of favorable interest rate gaps and 10. Costs are available for the project unit within LBP, but they have not been analyzed by LBP, nor have the costs incurred by other LBP units in handling the credit lines. 11. Adjustments in the reference rate and in the lag in the formula have tended to reduce divergence between the pass- on rate and the T-Bill rate, especially under RF-II. Specifically, from the beginning of RF-I in January 1992 until June 1993, bank time deposits provided the reference rate. In July 1993, the reference rate was changed to the 91-day T-Bill rate, the lag was increased slightly, and the premium was increased. In October 1995, the lag was reduced slightly, and in July 1998 there was a change from quarterly to monthly repricing with a consequent major reduction in the lag. 6 larger disbursements, but this may be due to the ability of PFIs to "rush through" a smaller number of larger loans at advantageous moments. 3.4 The end-use of funds deviated from expectations at appraisal. According to the PCR for RF-I, although 83% of project loans were intended for small businesses (assets under P20 million before financing), their actual share is only 22% of the amounts lent, while medium businesses (assets between P20 and 200 million) received 30%, and large businesses (assets over P200 million) received 48%.12 Also, less than 20% of project loans were for new undertakings, with by far the major portion for expansion of existing businesses (this pattern had been even more pronounced under the preceding credit project). Lending under RF-I was also highly concentrated in the more developed regions and among PFIs, with 20 participating commercial banks accounting for 72% of amounts lent and 20 rural banks accounting for only 3%. 3.5 Most performance targets for RF-II were substantially met or exceeded, except for the number of loans (77% of target), largely because of performance of the retail lending component. Many of the patterns noted in RF-I are still present in RF-II, but tend to be less pronounced. For example, rural banks have so far accounted for about 45% of loans and about 8% of amounts disbursed, with commercial banks accounting for 33% and 59%, respectively, and thrift banks accounting for the remaining 21% and 33%. Moreover, because of attempts to restrain large loans to large firms, only 32% of amounts lent have been in the large asset category, with 47% in the small category, and 21% in medium. Also, the majority of lending has been for new undertakings as compared to expansion of existing businesses, unlike preceding projects. Nonetheless, lending has continued to be highly concentrated in more developed regions. Deviations in credit allocation patterns from expectations at appraisal, although clearly at odds with the stated poverty alleviation objectives of the projects, may be justified by the demand-driven basis of the projects. 3.6 Implementation of the non-credit components of the projects has not fared as well, especially those in RF-I that were not under direct control of LBP: the ACPC studies were not satisfactorily completed in the intended time frame,3 and GOP policy changes prevented PCIC and QGFB from carrying out their development programs.14 Even conditions to protect the solvency of LBP were not been fully resolved, especially those not under direct control of LBP: under RF-I, to shift responsibility away from LBP for agrarian reform bonds (although subsequently, after the project closed, the covenant requiring the government to assume the responsibility for the bonds from LBP was complied with); and, under RF-II, to apply the requirement to pay 50% of profits as a dividend to the GOP only to "real" profits (i.e., adjusted for inflation and adequate loan loss provisions). The implementation of training programs for LBP and PFI staff was fully satisfactory, although little attention was paid to measuring costs and benefits. Finally, there was little implementation of measures to improve the performance of cooperatives: the pilot credit line for cooperatives under RF-I had to be reallocated; and LBP 12 About 14% of the number of loans went to medium sized firms and only 3% to large firms. 13. The lack of commitment to these studies seems to have been due not only to a lack of funding but also to a lack of clarity as to how the results of these studies were to be of use to the Bank or to the GOP agencies involved. 14. Policy reversals prevented these institutions from charging high enough insurance premiums to cover their costs and to carry out strengthening programs (until these reversals, both institutions were given high marks in Bank supervision reports). In discussions during the PAR Mission, DOF did not seem particularly concerned about the lack of performance by QUEDANCOR (the successor institution to QGFB) or PCIC. Moreover, subsequent studies of public sector loan guarantee institutions and mechanisms, in both the Philippines and numerous other countries, suggest high costs and few benefits (little additionality) anywhere in the world. 7 assistance to cooperatives under RF-II has not significantly improved their performance in lending to agrarian reform beneficiaries. 4. Findings Credit Components 4.1 The completion reports for both RF-I and II infer from reviewing a sample of sub-loans for which ex-ante economic rates of return (ERRs) were calculated that the sub-projects financed should have been quite profitable, with median ERRs of 35% for RF-I and 30% for RF-II. A better indicator that loans were not only put to good use but also were of generally high quality might be repayment rates, since repayment requires that the borrower be both willing and able to repay.16 The PCR reports in passing that only 0.02% of the LBP portfolio funded under RF-I was overdue, but does not mention whether sub-borrowers were overdue on repayments to PFIs. Under RF-II, however, LBP quarterly progress reports show not only amounts overdue from PFIs to LBP (0.5 2% of total amounts outstanding as of end 1998)8Y but also a detailed breakdown of loan arrearages from sub-borrowers to PFIs. As of end 1998, 21.4% of sub-loans were in arrears, but this amounted to only 3.7% of the total value of the portfolio, and 70% of this was less than six months overdue.'8 Arrearages are highly concentrated in smaller sub-loans: 28.3% and 8.2%, respectively, in the small loan category; and 24.0% and 7.8%, respectively, in the small asset category. Arrearages also tend to be concentrated in the more developed regions of Luzon, and in lending by rural banks (33.6% and 9.4%, respectively) and by thrift banks (21.5% and 7.3%, respectively) as compared to commercial banks (5.8% and 0.7%, respectively).'9 4.2 Under RF-II in particular, attention was placed on the repayment performance of LBP's agrarian reform loan portfolio, as this was seen to be a threat to the profitability and even to the future solvency of LBP.20 In policy discussions preceding both projects, the threat to LBP from agrarian lending was seen to have three aspects: the subsidized interest rates on these loans, their high administrative costs, and their high arrearage rates. However, because of politicization of the interest rate issue, no significant progress was seen possible in this area, so RF-Il focused on 15. Reductions in delinquency rates in LBP's agrarian reform loan portfolio reflect increased prudence in credit administration. The establishment of cooperative accreditation criteria has led to a shift in the portfolio composition away from cooperatives, in favor of rural banks. Even among cooperatives, the LBP lending has shifted in favor of the better performing savings and loans cooperatives relative to the agricultural cooperatives. However, LBP's administrative costs for agrarian lending have tended to continue to rise in spite of reductions in the number of personnel involved. 16. Since interest rates were not subsidized, this suggests that real earnings were sufficient to repay and that the borrower found the relationship with the lender to be attractive enough to want to repay. 17. Arrearages to LBP were concentrated in only three rural banks. 18. The end of 1998 is a particularly useful date not only because it represents the most recent data available at the time of the PAR mission but also because it reflects relatively high arrearages due to the impact of the Asian crisis and the El Nifio weather phenomenon. It should be noted, however, that the total portfolio past due for LBP was lower than the overall figure for the banking sector as a whole. 19. It could be useful to investigate the reasons behind these patterns (if they are in fact statistically significant) and also what portions of these loans are ultimately repaid (most it would seem given the relatively small amounts [0.8%] and numbers [5.0%] of sub-loans more than one year overdue), but this is beyond the scope and resources of this audit. 20. The issue of the agrarian reform bonds had finally been resolved by removing them as a potential liability of LBP. 8 monitoring and reducing administrative costs and arrearage rates, especially the latter.21 LBP reports on its Institutional Strengthening Action Plan (ISAP) provide data on agrarian loan disbursements and collections from 1995 through 1998,2 and the pattern is quite consistent over time: rural bank collections are about half of their disbursements by the end of the first year and 98% by the end of the second year; credit cooperative collections are about a quarter in the first year but ultimately over 90%; and agricultural producer and marketing cooperative collections are about one third in the first year but ultimately only about three-quarters. To improve overall collection rates, LBP has logically more than doubled its agrarian lending through rural banks (to P7,024 million in 1998), shown an even greater percentage increase in lending through credit cooperatives (but only reaching P2,106 million by 1998), while lending through agricultural cooperatives has held fairly steady (at about P3,000 million).23 4.3 Other important aspects for evaluating the outcomes of the credit components of the two projects are the impacts on profitability and on certain balance sheet indicators (e.g., funding sources and loan maturities) for LBP and the PFIs. The PCR for RF-I estimates that the project's credit activities added substantially to LBP profits (P69 million in 1993) but this does not seem to take LBP administrative costs fully into account. Under RF-II, the ICR reports profits from CLF II of P12 million for 1996, P26 million for 1997, and P79 million for 1998. Overall, because of the requirement for LBP to pay a dividend of 50% of its profits to the GOP, measuring LBP's total profits has received much greater attention under RF-II24 Table 2 shows the major impact of the inflation adjustment on LBP profitability, most dramatically in 1998 when stated nominal profit is P1,003 million and adjusted real profit is -P203 million. 4.4 The reason for this dramatic decline in LBP profitability in 1998 is not just the inflation adjustment, nor is it Bank-funded credit operations (ALF and CLF I and II), nor does it appear to be a deterioration in agrarian reform lending, about which the Bank has been so concerned. 25 Rather, it appears to be a dramatic decline in the profitability of LBP's commercial loan portfolio, something LBP had looked on in the past as a source of profits to compensate for losses on agrarian reform loans. In fact, contemporaneous with the Asian crisis, the past due rate on LBP commercial loans has risen from less than 5% at the beginning of 1998 to more than 12% at the end. As a result of increased past dues and conservative provisioning policies, LBP increased its loan loss provisions by P5,300 million in 1998. While this largely explains the dramatic decline 21. During RF-I these issues were not so pressing because most of the losses, those due to low interest rates in particular, were covered by external subsidies. Under RF-II with respect to administrative costs, it is interesting to note that staff devoted to agrarian reform operations have been reduced from over 3,300 in 1995 to less than 2,500 in 1998, but that costs relative to net loans have continued to rise, from 7.1% in 1994 to 13.8% in 1998. The rise is costs is main attributable to the fact that in 1994 a significantly greater proportion of the agrarian reform operating costs were charged to the Agrarian Reform Fund. 22. Collection rates need to be interpreted carefully: they include only collections in a given year against that year's disbursements and then add each subsequent year, so that collection rates at first glance seem to be declining over time because fewer years have passed for more recent years. 23. The rapid increase in agrarian lending through credit cooperatives appears to have adversely affected these collection rates, while the stability in amounts lent through agricultural cooperatives has apparently allowed LBP to improve its selection and control procedures, thereby somewhat improving these collection rates. Logically, an optimizing lender should want to equalize recovery rates for different categories at the margin (other costs being equal). 24. The question of the (non) payment of the LBP dividend in conformance with conditionality for RF-II for 1998 and 1999 was finally resolved in June, 2000. 25. CLF II profits rose in 1998, while agrarian loan recovery has not deteriorated, and the increase in its administrative costs is relatively minor. 9 in LBP profitability in 1998, and LBP's conservative provisioning policies are certainly praiseworthy, warnings are nonetheless raised about LBP's long-run viability." 4.5 The behavior of certain balance sheet items also indicates how project outcomes have impacted LBP (see Table 3 for selected balance sheet items; a more detailed balance sheet for LBP is given in Annex C). Most notably, ALF/CLF loans have consistently amounted to only 3.5% to 4.0% of total LBP assets.? Meanwhile, agrarian loans have fallen from almost 20% to less than 10% of total assets, as commercial loans have risen from as low as 13% to more than 40%. Thus, while the direct effects of project credit activities on LBP cannot be large, the indirect positive effects on LBP profitability (at least until 1998) due to the shift from agrarian to commercial lending may be quite large. With respect to funding sources (one of the projects' goals is to increase market-based, long-term funding), LBP deposits have risen from just over 35% of total assets in the early 1990s to about 70% in more recent years. However, more than 60% of deposits are still from GOP sources, while time deposits continue to account for less than 15%. Moreover, almost 90% of LBP deposits are accounted for by large deposits (over P200,000), which are only 3.4% of the total number of accounts, raising some concerns about the potential stability of LBP's funding sources. . 4.6 An attempt has also been made to evaluate the impact of the projects on PFIs by comparing, for the 1994-98 period when the most data are available, the behavior of PFIs with institutions that did not participate (and also by separating PFIs into those that have continued to participate and those that have had their access to CLF funds suspended, cancelled, or not renewed). Three aspects were examined for each category of bank (commercial banks, thrift banks, and rural banks): past due loan ratios, loan maturities, and deposits as a source of funds. Findings are presented in Tables 4-6 and discussed in detail in Annex B." Relatively little can be said about thrift banks or commercial banks, the former because of the very small size of the sample and the latter because LBP data for PFIs is not sufficiently comprehensive and BSP data could not be made available for individual institutions as would be required to separate other institutions from the PFIs.29 4.7 Some very interesting patterns nonetheless emerge for rural banks. First, few participating rural banks were eliminated because they failed to meet LBP project criteria, but many were eliminated because they failed to use the CLF credit line. Second, even rural banks that were eliminated had lower past due ratios than rural banks in general. Third, participating rural banks had substantially lower deposit to loan ratios than other rural banks. Fourth, participating rural banks are much larger than both rural banks in general and those that have been approved but are not using the CLF credit line. Thus, LBP appears to have been quite successful in attracting the best (and largest) rural banks to the CLF credit line, but with some danger that excessive dependence on the credit line as a source of funding relative to deposits could become a problem (cf., Massagna 99).30 The issue of the widespread failure of approved 26. LBP has recently increased its capitalization to improve its capital adequacy ratios. 27. Share of ALF/CLF loans to gross loans for 1998 and 1999 was 6.4% and 6.2%, respectively. 28. The annex also notes the projects' modest impact on PFI profitability. 29. If evaluations of PFI behavior and performance are to be important in the future, LBP might collect potentially pertinent data in formats consistent with BSP data at little cost to itself or the PFIs. 30. Massagna 99 was an earlier subsidized government sponsored credit program to promote rice production. Many rural banks became excessively dependent on government funds and subsequently failed due to poor recovery of their government-funded loans. 10 rural banks to use the CLF credit line was explored during the PAR mission through one-on-one discussions with rural bankers at a conference in Mindanao, and the high minimum size of sub- loans under the projects appears to be the main impediment to use." Many of the rural bankers interviewed there stated that they used other LBP rediscount lines and were thus persuaded to apply for the CLF credit line, but later did not use it because minimum sub-loan sizes were too large for their operations and for their clients' needs.2 This may help explain why project sub- loans appear not to have been focused on poverty alleviation. Non-Credit Components 4.8 The non-credit components of the two projects have not been as successful. The PCR for RF-I states that objectives for QGFB (now QUEDANCOR) and PCIC were not achieved, primarily because neither institution was allowed by the GOP to charge premiums adequate to cover expenses and reimbursements for covered losses, which had been part of project conditionality, especially for PCIC. Audited financial statements for PCIC covering the period 1992 through 1998 show mounting losses as gross income has declined while operating expenses have risen, although equity has not declined as much as might be expected because of external infusions of capital. For QUEDANCOR, financial statements for 1997 (only year received and with no indication that they were audited) show that the very modest net income is due mainly to income from investments (guarantee fees accounted for less than 10% of gross income). 4.9 The two policy studies that were the responsibility of ACPC were not satisfactorily completed, ostensibly because of lack of agreement on funding, but perhaps more basically because neither LBP nor other GOP agencies saw the potential usefulness of the studies. Improvements in rural bank performance, ascribed in the PCR to the recapitalization program and other policy measures supported under RF-I, may in fact be due mainly to improvements in the economy from 1991 to 1993 and to the closure of failing rural banks, especially since as of mid- 1999 a rural bank recapitalization program was still in effect. Moreover, in contrast to the intent of these policy measures, opening new bank branches still requires feasibility studies and the "assignment" of a substantial amount of bank capital to each new branch, both of which make it relatively unattractive to open small branches, as new rural branches would likely be. Furthermore, in response to the Asian crisis, new norms have been issued and changes in the banking law are being contemplated that could further restrict bank branching, especially in rural areas.33 4.10 In discussing policy impacts, the PCR for RF-I details training carried out under the project (e.g., number of courses, numbers of participants) - on environmental issues and on CLF policies and procedures for LBP and PFI staff, and for strengthening rural banks and cooperatives. Training costs and benefits are not analyzed, however, except to mention that the quality of LBP's cooperative loan portfolio deteriorated despite all the training devoted to 31. The interviewees included rural bankers who were: participating in CLF; accredited but not participating; and not accredited. 32. Complaints were also registered about approval and monitoring requirements for sub-loans. The rural bankers were mostly concerned with the high fixed costs of accessing CLF funding which in turn required making larger loans. 33. It is unclear how laws and norms that interfere with the internal operations of otherwise solvent and profitable banks can be seen to enhance safety and soundness in the system. 11 cooperatives.34 3 Training for LBP staff is also reported to be a significant component of RF-II, as discussed in LBP's ISAP progress reports. Project supported training is also reported for PFI staff and for cooperatives, with considerable attention to environmental issues for PFI and LBP staff, but still without analysis of the costs and benefits of this training. In addition, ISAP progress reports suggest that progress has been made on the major goal of protecting LBP viability, including especially its overall financial condition, its agrarian loan portfolio, and strengthening the cooperatives that serve as its intermediaries. Nonetheless, it is unclear if the underlying policy conditionalities to protect the future of LBP are in fact firmly in place. Additionality" 4.11 Would the activities undertaken with project funding have been undertaken in any case? This needs to be examined at two levels: the sub-borrower and the PFI. Would the sub-borrower have found another source of funding (including possibly own funds) in order to undertake the activity? Would the PFI have made the loan available from another source of funds? 4.12 Interviews with four randomly selected sub-borrowers during the PAR mission are described in detail in Annex B and, although the sample is very small, the results are revealing: except for one sub-borrower, the others could have undertaken their activities without loans funded by CLF and they could even have obtained term financing elsewhere if it were needed. They took CLF-funded loans simply because the terms were more attractive (i.e., lower interest rates). Moreover, in two cases these were large businesses, and their activities substituted capital for labor and relied largely on imports, thereby providing few benefits to the local (rural) community. These observations are consistent with observations in LBP's own project completion report for RF-I, which noted that most of the new projects were not labor intensive, as well as with the relationship with the patterns of disbursement and the lag in interest rate adjustments noted earlier. 4.13 At the PFI level, the data that could be obtained to compare the behavior of PFIs with institutions that did not participate have already been discussed (see above) and provide some tentative conclusions that are not favorable to additionality. Loan maturities do not appear to have been lengthened appreciably, and it is possible that CLF funding may have substituted for deposit mobilization, notwithstanding the project goal of enhancing the mobilization of funds, especially longer term. Two other conclusions emerge from the data: (a) LBP was able to select better financial institutions but not necessarily able to make them better; and (b) the LBP's CLF loan 34. In addition, but not mentioned in this section of the PCR, the pilot CLF credit line for cooperatives did not disburse and the funds were reallocated. 35. The performance of the cooperatives has improved somewhat in the years after RF I closed, as discussed in footnote 15. 36. Enhancement of LBP's accounting and management information systems has also become a significant element of strengthening LBP under RF-II. 37. The World Bank's East Asia Region maintains a substantial difference between it's view on the issue of additionality and that of this audit, as discussed in the Region's comments given in Annex D. OED response to these comments is given in Annex E. 12 product was not appropriate for the average-sized rural bank, which led to higher spreads and possibly also to greater risks.38 4.14 CLF disbursements appear to have been significantly affected by the relation of project interest rates to market interest rates (i.e., the difference between the pass-on rate to PFIs and the T-Bill rate), but less by the real interest rate (i.e., the difference between the rate to PFIs and the inflation rate). Such interest rate sensitivity suggests that PFIs were prepared to capture whatever subsidies might become even momentarily available by substituting among sources of funds - another indication of limited additionality. The apparently weaker relationship between disbursements and real interest rates (which would matter more to sub-borrowers) may show that CLF-funded investments were so profitable that they would have been undertaken and financed in any case so that small fluctuations in real interest rates simply did not matter. 4.15 The concentration of sub-loans in larger loans to larger sub-borrowers in more developed regions further raises concerns about additionality. The image of rural/agricultural activities that comes to mind (especially given the Bank's overarching objective of poverty alleviation) is far from the types of activities that received most project funding. In the Philippines, "rural" is often defined as anything outside of Metro Manila, and this was clearly the definition used by LBP and accepted by the Bank. 4.16 Despite these concerns, however, it is difficult to establish additionality, or the lack of it, conclusively. For RF-I and RF-II the issue is open to debate, as the ICR for RF-II also notes. On the one hand, the limited evidence above suggests little additionality; on the other, as the ICR points out, the issue is whether there was an overall increase in investments and whether the project funds "crowded out" deposit mobilization. There is little evidence to establish a credible counterfactual either way (LBP's commercial and agrarian lending has gone up over the life span of the projects, as have private and total deposits, including time deposits)" Given the fungibility of funds, it is difficult to say how much the projects helped incremental lending or substituted for incremental deposit mobilization.0 Both the medium- and long-term investment financing and time-deposits increased significantly over the projects' lifespan, and their growth has outpaced that of short-term lending and deposits. Further, the issue is complicated by the positive institutional development outcomes of the projects, and how the projects may have contributed to the increased lending and deposit mobilization, as well as qualitative shift in LBP's rural lending portfolio. 38. These conclusions are supported by agrarian loan collection patterns and by the negligible impact of LBP training and technical assistance on rural banks and cooperatives. 39. PFIs offer their best clients for attractive CLF funding, clients who would have received credit anyway. To establish additionality would require knowing whether some new clients were served because CLF funding released funds for them, or if instead funding from other sources was reduced. Such an exercise would be too costly and was not attempted by the audit. 40 The LBP's contribution to the ICR (p.15) takes a more negative view of additionality: "It could not be claimed, however, that these project investments would not have pushed through without CFL-II funding as these viable projects could have otherwise tapped alternative sources of financing." 13 5. Ratings Outcome 5.1 The projects' outcome is evaluated applying the standard OED methodological criteria of relevance, efficacy and efficiency. Both RF-I and II are clearly focused on the Bank's broad goal of fostering growth in the private sector, specifically through enhancing the availability of longer- term credit to finance private sector investments. Development of human resources was also promoted in that skills in LBP and the PFIs were upgraded, and there also appear to have been some improvements in the human resource base of some financial institutions serving the agrarian reform sector. Neither project had poverty alleviation as an explicit objective, although both anticipated poverty reduction impact indirectly through job creation and by focusing on areas outside of Metro Manila. However, the projects did not provide for any special incentives for job creation or provisions for effectively monitoring job creation, while areas outside of Metro Manila receiving most project loan funds were among the relatively more prosperous and developed areas of the Philippines. With respect to the environment, there were requirements under both projects to comply with environmental regulations, but such compliance was not effectively monitored. The projects' relevance thus hinges primarily on fostering private sector growth through increasing the availability of term financing for investment. 5.2 The Bank's 1998 Country Assistance Review for the Philippines supports the relevance of RF-I and RF-Il through giving high priority to channeling savings more efficiently into investment through the financial sector, especially banks. Developing the market for long-term finance and rehabilitating the rural credit system are also specifically mentioned as important goals for the Philippines. Thus, overall the relevance of both RF-I and RF-II can be rated as satisfactory. 5.3 The efficacy of RF-I and RF-II in meeting their stated objectives was also satisfactory. While certain non-credit components of RF-I did not perform well, the key element of project implementation was clearly very rapid and largely trouble free disbursement of the credit line. Because RF-II was more narrowly focused on the role of the LBP in disbursing credit, efficacy could have been even higher, except that LBP's retail credit component proved problematic. Efficacy was also hindered by the Government's lack of total commitment to assuring a policy environment fully supportive of LBP's viability and the long-run sustainability of project outcomes. 5.4 Judging from estimated ERRs for investments financed with project funds, both RF-I and RF-II also can be rated satisfactory in terms of efficiency. However, it must be noted that these ERRs were calculated ex-ante and only for those loans that were sufficiently large that such calculations were required for approval. While there was no ex-post analysis of ERRs, the fact that loan repayment rates from PFIs to LBP were almost 100%, and also very high from sub- borrowers to PFIs, suggests the FRRs must have been adequate ex-post for repayment. 5.5 Thus, with respect to the credit components of the two projects, the disbursements have been well ahead of schedule, have financed economically viable investments, and the recovery rate has been satisfactory. These positive elements are counterbalanced by the lingering questions about the additionality of the funds provided by the two projects and their limited poverty 14 alleviation impact, especially in areas that are normally considered to be "rural" in nature" For RF-I, an issue that may concern some observers is the large share of large sub-projects financed. However, this is the result of a consciously designed undirected credit line and independent decisions made by institutions striving to become a thriving commercial entities. As such, the audit feels that these decisions made by LBP and PFIs must be respected. The outcomes of the non-credit components, essentially in RF-I, were unsatisfactory. Overall, in terms of the stated objectives of the two projects of financing productive investments and the institutional development, as discussed in detail below, the outcome for both projects is rated satisfactory. Institutional Development Impact 5.6 Viewed from the perspective of LBP's ability to implement the credit components of the two projects, specifically the capacity for wholesaling, institutional development has been substantial. Minor criticisms have been made in certain reports about LBP's monitoring of PFIs and sub-borrowers, but these can be highly costly activities for all involved if carried too far, and loan collection rates clearly indicate that whatever monitoring was carried out was highly effective in maintaining viability.42 LBP and the PFIs (and possibly the cooperatives involved in LBP's agrarian lending) also appear to have been strengthened by the training and technical assistance provided under the two projects, but the evidence is not conclusive because neither the Bank nor LBP designed or implemented monitoring systems to evaluate the costs and benefits of this training and technical assistance. 5.7 Overall, however, the institutional development impact of the projects may be only modest. In particular, while ALF/CLF credit activities have undoubtedly been highly profitable, achievement of the goal of cutting losses in LBP's agrarian reform loan portfolio has not been substantial. Moreover, the long-run viability of LBP has not been assured, first, because of the marked deterioration during 1998 in LBP's commercial loan portfolio 4 and, more importantly, because a number of policy threats continue despite project conditionalities. Specifically, the LBP dividend on profits payable to the GOP has not been reduced, on a formal basis, to account for the impact of inflation on equity and for fully adequate loan loss provisioning, but rather on an ad- hoc basis each year. In addition, no significant advance has been made in dealing with interest rates on LBP's agrarian loan portfolio, and GOP is currently negotiating credit lines with other donors for other Philippine financial institutions with interest rate conditions that could greatly weaken LBP's competitive position. In short, although LBP is well able to handle Bank credit lines, its institutional development may be stunted by policies outside LBP control. 5.8 With respect to institutional development of other Philippine entities involved, particularly in RF-I, the results are clearly negligible. While project training and technical 41. See Annex D on the East Asia Region's views on the issue of additionality, and Annex E on OED's response to the Region's comments. 42. LBP and PFI monitoring of environmental impacts is a clear example of a potentially costly activity for all involved and one that will tend to reduce rates of return. No one would disagree about the importance of environmental protection, at least to the point where costs and benefits are equated at the margin, but this should be the responsibility of environmental protection agencies, and putting the responsibility on lenders not only raises their costs and reduces their return but also fails to place the responsibility - and the institutional strengthening - where it belongs. 43. The fact that Philippine private commercial banks have also suffered since the onset of the Asian crisis (though apparently not as much as LBP's commercial loan portfolio) does not make the LBP more robust with respect to external shocks. 15 assistance for rural banks and agricultural cooperatives and the recapitalization program for rural banks may have had positive impacts on some of these institutions, there are still relatively few strong cooperatives and many rural banks are still in need of recapitalization. Neither PCIC or QGFB became stronger but, rather, continued to deteriorate when the GOP failed to adjust insurance and guarantee premiums to the levels needed for sustainability. ACPC studies were produced several years later than expected, and BSP has never fully liberalized bank branching. While interest rates are decontrolled, the GOP has been willing to accept donor money conditioned on subsidized interest rates and to devote its own funds to targeted lending at below- market interest rates when political pressures are great enough. Finally, neither DOF nor LBP seem preoccupied about the potential losses on project loans stemming from the dramatic depreciation of the Philippine exchange rate, perhaps (correctly) because these are sunk costs, but at the same time continuing to accept similarly "covered" arrangements on new loans from bilateral and multilateral donor agencies, presumably on the assumption that an Asian financial crisis or something similar will not happen again. Thus, from a broader sub-sectoral perspective, institutional development impact is considered modest. Sustainability" 5.9 LBP and almost all PFIs are solvent and have been profitable, even in 1998, and the credit lines provided under both projects have added to profitability. Reviews of a sample of sub- project FRRs and ERRs reported in the PCR for RF-I and ICR for RF-II indicate high rates of return on the activities financed and, more importantly, under both projects loan repayment rates have been excellent from PFIs to LBP and good from sub-borrowers to PFIs. Thus, the investments financed are clearly productive and sustainable. 5.10 Nonetheless, there are three areas of concern which will affect the sustainability of projects' achievements: * Losses for the GOP are implied by current Philippine foreign exchange rates because on- lending under both projects has been almost entirely in local currency and the amounts allocated to cover GOP foreign exchange risks appear to be inadequate in the wake of the Asian crisis. However, the ICR for RF-II argues that the net impact of including the fees and taxes on income generated by the sub-project portfolio is likely to be smaller than for RF-I, but still positive. However, the recent experience shows that unless better formulas are used to cover foreign exchanges risks, or currencies never devalue so substantially over the long repayment periods, such lines of credit are likely to have future negative impacts, requiring government subsidies. * Term lending such as that carried out under RF-I and II could be maintained if LBP and/or the PFIs secured sufficient long-term resources apart from external funding, and this was in fact a goal of both projects. However, despite some efforts (prior to the financial crisis) by LBP to secure long-term resources in the international markets, it needs to make significantly greater progress in this direction than it has so far, as shown by the continuing predominance of sight deposits and government deposits among its deposit liabilities. It is also unclear if PFIs have moved to develop more substantial long-term funding sources. 44. The East Asia Region also maintains a different view of the sustainability of RF II's outcomes. See Annex D for the Region's comments and Annex E for OED's response. 16 * While interest rates are no longer controlled in the Philippines, there are nonetheless a large number of directed credit lines with below-market interest rates, some subsidized by the GOP and others by donor agencies. The continuation of such programs (and the policies thereby implied) suggests that GOP commitment to financial liberalization is not total. Furthermore, introducing credit lines with subsidized interest rates for institutions competing with LBP and the PFIs could jeopardize the sustainability of project credit lines and adversely impact the viability of LBP and even some PFIs. 5.11 Of these areas of concern, the third is the most serious in introducing uncertainty in the prospects of sustainability of LBP as an institution. On deposit mobilization, LBP has made efforts to mobilize additional deposits, and indicated to the PAR mission that it is committed to continue to do so. On balance, the audit concludes that the outcomes of the two projects are likely to be sustained. 5.12 Continuing concerns with non-credit components related to conditionalities designed to further financial liberalization and to assure the long-run viability of LBP are potentially serious and have led the OED audit to be less optimistic than the ICR for RF-II about sustainability. The audit sees sustainability as only likely, rather than highly likely as seen by the ICR. While the government and LBP management have taken important steps that are beneficial for the financial and institutional health of LBP, there remain potential threats from below market interest rates on competing funds and of the Government's failure to take formal steps to permanently remove the unfairly burdensome dividends payments. Bank Performance 5.13 Bank performance under both projects could be rated highly satisfactory,45 except for two elements: the question of additionality with respect to the credit components (basically identification and appraisal issues); and the failure to obtain full compliance with conditionalities that have relied on institutions other than LBP (largely supervision issues but also involving identification and appraisal). 5.14 As discussed under project outcomes, additionality can be neither proved nor disproved for most credit projects and for these two projects in particular. A further question of the nature of additionality that might have been expected (e.g., just more term credit for any legitimate activity outside of Metro Manila, or term credit that would provide support for agricultural and rural activities with benefits for rural communities) raises issues of Bank performance. The Philippine agencies involved clearly saw this as demand-determined (untargeted) credit for any legitimate activity outside of Metro Manila, and, given the speed and patterns of disbursements, the PFIs clearly found project interest rates highly attractive, especially when they were below T-Bill rates. Did the Bank's Board, senior management and staff have a similar view? Discussion of poverty alleviation and the rural and agricultural activities described in the SARs for the projects were clearly not consistent with many of the activities financed. Did the Bank at its different levels have a clear and consistent view of what it wanted to see financed, compared to what was being financed, that might have been communicated to the Philippines (but apparently was not)? 45. One of the most unusual and positive aspects of Bank performance on these two projects, as well as on the preceding and succeeding ones, is the extremely high degree of continuity in Bank staff that have had the main roles in these projects. 17 5.15 If non-credit components were really important to the achievement of project goals, then supervision should have tried harder to achieve compliance with loan conditionalities. If these conditionalities were not really so important, then they should have been left out. On one hand, what PCIC, QGFB, and ACPC have done or not done is probably largely irrelevant to the achievement of main goals of the projects. On the other hand, the less than total support for certain BSP and DOF related policy conditionalities and project goals may have reduced the likelihood of project sustainability and the effectiveness of institutional strengthening. 5.16 Because of these two elements, bank performance is rated satisfactory rather than highly satisfactory. The audit ratings of satisfactory bank performance are the same as the PCR for RF I and ICR for RF HI. Borrower Performance 5.17 Borrower performance parallels Bank performance in being highly satisfactory for LBP and quite mixed for the other institutions involved. LBP performance was highly satisfactory in project preparation, and especially implementation, as reflected by rapid loan disbursement and good loan recovery in particular. The only elements of LBP performance that can be faulted were on the retailing of project funds and the fact that LBP staff might have analyzed more what it was implementing (e.g., cost-benefit analysis of training activities and monitoring of PFI behavior as compared to non-participating institutions with respect to certain key indicators such as deposit mobilization, loan maturities, loan past due ratios, etc.). Not surprisingly, LBP was also unable to ensure fulfillment of project conditionalities that affected it but were beyond its direct control (e.g., rationalizing the dividend potentially payable to the GOP and the continuing below-market interest rates on agrarian reform loans), nor could it always achieve project results involving other institutions (e.g., strengthening the performance of cooperatives involved in agrarian lending). 5.18 Performance of the other institutions involved was much more mixed. Apparently, little attention was paid to the willingness or ability of PCIC, QGFB, and ACPC to carry out the responsibilities assigned to them, and such deficiencies became clear during implementation. In addition, and more importantly, DOF has demonstrated little commitment to the issue of LBP's potential dividend payment for 1998 that would result in a substantial reduction in LBP's real equity, and little concern about the foreign exchange losses that are likely to be incurred by the GOP on the projects. At the same time, BSP has shown a lack of commitment to leaving bank branching to market forces (as it should be, at least for banks in conformance with requirements for solvency, liquidity, etc.) and to discouraging subsidized credit lines funded by donors and the GOP that continue to create widespread distortions in Philippine financial markets. Because of the mixed performance of these other institutions, borrower performance is rated satisfactory rather than highly satisfactory. These ratings for RF I and RF II are the same as in the PCR and ICR, respectively. 18 6. Lessons 6.1 Project focus should be as narrow as possible to achieve the objectives of the project, specifically avoiding extraneous and difficult to implement components, but broad enough to include elements that may be important for long-run sustainability. In the case of the Philippine rural finance projects, the narrowing of focus that took place between RF-I and RF- II clearly illustrates both points. Narrowing was good in that it has been more realistic about what can be accomplished if a variety of institutions are involved, some of which do not receive project funding. Realistically, unfunded institutions will only collaborate if they see it as clearly in their own best interest to do so. Therefore, if the reform or participation of an institution is crucial for project success, funding should be provided; otherwise, the institution should be left out as it will just complicate the tasks of project implementation and supervision. On the other hand, narrowing from RF-I to RF-II was bad to the extent that it avoided dealing with policy conditionalities potentially important for project success that will, if not dealt with, tend to weaken institutional development impact and reduce the chances of project sustainability. Thus, where policy conditionalities or institutional changes are important for advancing long-term project objectives, they should be included, appropriately funded in project design, and pursued diligently during project implementation. 6.2 Financial liberalization is more than just decontrolling interest rates. Government subsidizing of interest rates for certain classes of borrowers or types of activities - or permitting or encouraging donor agencies to do so - disrupts the financial system and adversely affects the viability of financial institutions that play by market rules. In the case of Philippine rural finance, this lack of a "level playing field" has adversely affected participation in RF-I and II and, if such policies continue, may even threaten the viability of LBP in the longer run. Financial liberalization also includes eliminating restrictions that limit competition and outreach without effectively promoting "safety and soundness" of financial institutions. A clear example in the case of Philippine rural finance is BSP's continuing intervention in branching decisions by banks that satisfy all regulatory requirements for safety and soundness. The requirement to assign capital to new branches limits the attractiveness of opening small, rural branches and thereby limits outreach to the projects' target clientele. 6.3 Sustainability of credit operations requires on-going access to funding, and credit projects should include specific incentives and adequate monitoring to promote this. Since access to external credit lines is a substitute for deposits, especially if the interest rate is attractive, deposit mobilization is unlikely to improve unless specific incentives are created and the results are monitored. In the case of Philippine rural finance, LBP started from a low base of mainly government deposits and has improved its deposit mobilization performance only modestly. Moreover, the Bank did not require LBP either to monitor or reward deposit mobilization by PFIs, and disbursement patterns under RF-II and especially RF-I suggest that PFIs were highly sensitive to interest rates on project funds relative to other sources. 6.4 Finally, such objectives as poverty alleviation and, in the absence an independent and effective monitoring system, environmental protection may exceed the reach of a project designed to meet market-based, demand-driven credit needs. The experience with the Philippine rural finance projects suggests that these objectives can be difficult to reconcile. 19 Funds disbursed rapidly under both RF-I and II and in ways that suggest that market demand was the key element in credit allocation. However, any achievement of environmental protection and poverty alleviation is much less clear. Although poverty alleviation was used to justify project support for local costs, neither project included mechanisms that focused directly on alleviating poverty. Instead, there was only monitoring of job creation consisting of reports based on statements by sub-borrowers. Given that PFIs and sub-borrowers are certainly aware that job creation is considered good, it is not surprising that, despite overall reports of substantial job creation, PAR mission visits to some of the sub-borrowers found negative job creation, and the PCR for RF I indicates that job creation was less than anticipated. Moreover, the definition of rural restricted project credit only to anywhere outside Metro Manila, and the result was credit allocated heavily in favor of other relatively developed and urbanized areas of the Philippines. With respect to environmental protection, both projects asked PFIs to enforce environmental protection measures. Although the requirement for environmental clearance was enforced at the time of sub-project appraisal, as the ICR admits (p. 11), project performance on environmental protection has had significant shortcomings. In the absence of proper enforcement, monitoring compliance with environmental regulations presents a moral hazard for PFIs, who view this activity as costly distraction from their primary business of financing activities that provide the highest likelihood of repayment.46 For PFIs and LBP, enforcing their regulations only for Bank financed sub-projects also presented a conflict of interest, as it raises their costs relative to their competitors. To create a level playing field, and to ensure uniform and objective implementation of the country's environmental laws, enforcement might better be assigned to institutions devoted primarily to this task. With weak environmental institutions, as in the Philippine according to project documents, trying to use PFIs as a substitute will not resolve their problems either. 46. PFIs also responded rationally in the case of environmental training activities which were consistently under- subscribed by PFIs and then filled by LBP staff who were apparently required to attend. 21 Table 1: Land Bank of the Philippines: Interest Rates, Inflation Rate, and Onlending under CLF 1993 994 1995 1996 1997 Q1 Q2 03 Q4 Q1 Q2 03 Q4 0Q4OQ1 Q2 Q3 1 Q2 3 2 Q4 Q1 Q2 Q3 04 Q 0 ,- Q4 1nterest Rate toPFs(% 12.9 10.7 11.0 11.2 19 10.0 11.0 13.4 10.6 11.6 10.6 10.7 14.7 16A 17AL5. 14,6 T-Bill Rate (%) 1 .3 10.9 10.9 14.7 16 l4 11.7 14.2 10.1 11.0 127 12,2 C6 10.5 10.4 14.2 16.9 13 Difference between PFI Rate and -1 0.03 -1,9 -0.4 -0.2 0.1 -3.5 0. 0.4 1,0 1 3 -1.7 -3.2 3.3 -04 -.9 0.0 V 0,2 1.1 0.2 -3.5 -1.5 2,9 1.1 1,2 T-Bili Rate Quarterly Inflation (>6369861. Qatrynfaon()7.9 12.0 9,1 3,9 5.4 9.2 1.3 7.9 9,4 9~4 &.2 6.5 6 .3 6.9 8.6 10. 11.0 10.4 8.2 6.7 5-.3 5.3 5.9 7.9 ',9 10.4 10.6 Difference between PFi Rate and Iffereonc bewea PRte an 9.7 3,4 4.6 9.9 7.5 1.5 9.7 3.3 6 1 5,7 6,6 4,4 3.7 4.1 4.8 0 8. 235 4-16 5 1 6.3 5.3 4.8 758.5 7.5 4.8 4.0 Inflation Rate New Releases (million pesos) CLF I 477 44 63 83 404 575 340 543 219 2745 225 500 31 214 2 484 133 258 334 429 4 174 95 73 13 CLF IlI -1- 18- -5-2-8119 181175 248 190 224 289 404 387 Number of Loans CLFI 216 51 64 49 81 43 676 8 9 CLF IlI -Z- 39 38 36 53 43 47 57 51 22 Table 2: Land Bank of the Philippines: Detailed Balance Sheet PROFIT IN NOMINAL AND REAL* TERMS (In Million Pesos) 1994 1995 1996 1997 1998 Nominal Net Income 1,220 1,314 1,601 1,700 1,003 Less: Net Income Required to Adjust Equity in Real Terms 505 952 477 700 1,206 Real Net Income 715 362 1,124 1,000 (203) * Adjusted for Inflation. 23 Table 3: LAND BANK OF THE PHILIPPINES SELECTED BALANCE SHEET ITEMS (In Million Pesos and % of Total Assets) 1990 1991 1992 1993 1994 1995 1996 1997 1998 TOTAL % TOTAL % TOTAL % TOTAL % TOTAL % TOTAL % TOTAL % TOTAL % TOTAL % ASSETS Cash & Due From Banks 4,400 19.3 3,404 11.0 8,169 16.5 18,344 20.5 13,687 14.1 14,437 13.5 18,367 13.7 19,953 12.5 19,505 11.0 Investments 6,980 30.6 9,692 31.4 14,901 30.1 31,448 35.1 37,382 38.5 29,027 27.1 29,398 21.9 36,122 22.6 39,679 22.4 Commercial Loans 3,301 14.5 4,250 13.6 6,413 12.9 12,819 14.3 16,753 19.3 34,892 32.6 55,585 41.5 69,099 43.1 76,692 43.2 ALF/CLF Loans 1,873 3.8 3,304 3.7 3,334 3.4 4,057 3.8 4,913 3.7 6,475 4.0 6,981 3.9 Agr. Reform/DtherAgri. Loans 3,119 13.7 5,985 19.4 7,694 15.5 10,713 12.0 11,142 11.5 11,245 10.5 13,471 10.1 14,759 9.2 14,820 8.4 Allowance for probable Losses (613) -2.7 (792) -2.6 (854) -1.7 (1,417) -1.6 (1,790) -1.8 (1,847) -1.7 (2,781) -2.1 (4,839) -3.0 (7,368) -4.2 Sub-Total Loans & Discounts 9,073 39.8 12,850 41.7 20,027 40.4 32,289 36.1 36,625 37.7 58,380 54.6 80,075 59.8 93,659 58.5 104,467 58.9 TOTAL ASSETS 22,823 100.0 30,837 100.0 49,546 100.0 89,508 100.0 97,207 100.0 106,920 100.0 133,955 100.0 160,149 100.0 177,479 100.0 LIABILITIES AND CAPITAL Deposits (Pesos) 7,899 34.6 11,282 36.6 23,737 47.9 54,401 60.8 52,404 53.9 70,836 66.3 83,389 62.3 97,597 60.9 105,712 59.6 Deposits - Foreign Currency 361 1.6 632 2.0 946 1.9 3,366 3.8 2,609 2.7 4,280 4.0 7,736 5.8 15,590 9.7 18,249 10.3 TOTAL LIABILITIES 16,548 72.5 23,586 76.5 41,434 83.6 80,499 89.9 86,986 89.5 95,720 89.5 120,529 90.0 145,359 90.8 162,223 91.4 TOTAL CAPITAL 6,275 27.5 7,251 23.5 8,110 16.4 9,009 10.1 10,221 10.5 11,199 10.5 13,428 10.0 14,791 9.2 15,256 8.6 24 TABLE 4: PHILIPPINE RURAL BANKS Past Due Ratios Average Amount of Loan Portfolio (%) (in Thousand Pesos) 1995 1996 1997 1998 1995 1996 1997 1998 All Rural Banks 17.0 14.1 15.3 na All Rural Banks 41,768 55,467 67,315 na Currently Participating na 8.3 10.2 12.4 Currently Participating na 203,101 307,332 297,635 Approved but Not Using na 9.6 11.5 14.8 Approved but Not Using na 72,635 90,942 92,039 Suspended/Cancelled na 10.3 13.7 20.8 Suspended/Cancelled na 203,800 239,637 222,690 Average Amount of Deposit Liabilities Deposits/Loans (in Thousand Pesos) (%) 1995 1996 1997 1998 1995 1996 1997 1998 All Rural Banks 39,235 50,450 60,600 na All Rural Banks 93.9 91.0 90.0 na Currently Participating na 160,814 227,637 232,569 Currently Participating na 79.2 74.1 78.2 Approved but Not Using na 69,483 76,196 86,558 Approved but Not Using na 96.7 83.8 94.0 Suspended/Cancelled na 256,107 246,486 214,122 Suspended/Cancelled na 125.7 102.9 96.2 25 TABLE 5: PHILIPPINE THRIFT BANKS Past Due Ratios Deposits/Total Fund 1994 1995 1996 1997 1998 1994 1995 1996 1997 1998 All Thrift Banks 7.8 10.1 10.2 13.5 na All Thrift Banks 61.2 64.8 66.8 66.4 na Currently Participating na 5.2 7.4 6.8 7.2 Currently Participating na 60.0 46.8 52.9 52.6 Suspended/Cancelled na 15.2 26.2 49.3 na Suspended/Cancelled na 60.0 76.3 72.0 61.5 Loan Maturities for All Thrift Banks Deposit Composition for All Thrift Banks 1994 1995 1996 1997 1998 1994 1995 1996 1997 1998 Short 36.3 31.3 32.3 41.5 na Demand 2.1 2.7 3.6 3.8 3.8 Medium 48.3 51.0 45.2 34.0 na Saving 84.1 85.2 84.3 85.1 86.1 Long 15.4 17.7 22.5 24.5 na Time 13.8 12.1 12.1 11.1 10.1 26 TABLE 6: PHILIPPINE COMMERCIAL BANKS (Y%) Past Due Ratios Deposits/Total Fund 1994 1995 1996 1997 1998 1994 1995 1996 1997 1998 All Commercial Banks 3.9 3.2 2.8 4.7 10.4 All Commercial Banks 64.6 63.2 60.4 59.9 63.3 Currently Participating na 2.5 2.4 4.3 10.8 Currently Participating na 66.5 63.9 64.5 68.0 Suspended/Cancelled na 4.9 3.0 9.5 15.5 Suspended/Cance!led na 74.0 73.0 75.0 78.0 Loan Maturities for All Commercial Banks Deposit Composition for All Commercial Banks 1994 1995 1996 1997 1998 1994 1995 1996 1997 1998 Short 74.7 75.9 67.3 71.3 69.7 Demand 7.5 7.9 8.4 7.0 8.4 Medium 16.7 20.2 22.3 18.5 17.6 Saving 75.0 74.5 69.8 69.8 68.8 Long 8.6 3.9 10.4 10.2 12.7 Time 17.5 17.6 21.8 23.2 22.8 Annex A 27 Annex A: Basic Data Sheets Rural Finance I Key Project Data (amounts in US$ million) Item Appraisal Actual Actual as % of Estimate Appraisal Total Project Costs 203.5 263.5 129 Loan Amount 150 150 100 Cancellation N/A 0 0 Date Physical Components Completed N/A N/A N/A Economic Rate of Return (%) Not Estimated 35 N/A Cumulative Estimated and Actual Disbursements (in US$ million) FY92 FY93 FY94 FY95 FY96 FY97 Appraisal Estimate 36.0 63.0 90.0 117.0 135.0 150.0 Actual 28.3 112.5 150.0 Actual as % of Estimate 78.6 178.6 166.6 Project Dates Planned Actual Identification 11/89 11/06/89 Preparation 02/90 02/18/90 Pre-appraisal 09/90 10122/90 Appraisal 01/91 02/24/91 Negotiations 05/91 02/24/92 Board Approval 07/91 06/21/91 Loan Signature 07/11/91 07/11/91 Effectiveness 10/01/91 11107/91 Loan Closing 03/31/97 06/08/94 Project Completion 09/30/96 06/08/94 Staff Inputs (staff weeks) Stage of Project Cycle Planned Actual Weeks US$ Weeks US$ To Appraisal 181.0 175.2 Appraisal through Board Approval 61.0 15.8 Supervision 33.0 47.4 Completion 8.0 4.0 Annex A 28 Mission Data Stage of Project Month/Year Number of Days in Field Staff Performance Rating Cycle Persons Specialties Identification 11/89 5 85 B,C,A,D,E Preparation 02/90 4 80 B,C,A,E,D Pre-appraisal 06/90 4 80 B,G,E,C Pre-appraisal 10/90 5 100 B,E,A,D,G Appraisal 02/91 3 30 B,E,G Supervision 9/21/91 1 15 B,F Supervision 2/24/92 1 21 B,F Supervision 9/4/92 1 10 B Supervision 1/6/93 1 6 B Supervision 6/15/93 1 6 B Specialty: A = Agricultural Economist B = Financial Analyst; C = Agricultural Credit Specialist; D = Financial Management Specialist; E = Economist; F = Environmental Specialist; G = Legal Counsel Other Project Data Follow-on Operations Operation Credit Number Amount Board Date (US$ million) Second Rural Finance Project Loan 3938-40-PH 150.0 10/10/95 Third Rural Finance Project Loan-PH 150.0 12/98 Annex A 29 Annex A: Basic Data Sheets Rural Finance II Key Project Data (amounts in US$ million) Project Component Appraisal Estimate Latest Estimate Actual Percentage of Appraisal CLF 11 a/ 120.0 222.9 186% RCF al 141.2 38.5 27% TA & Training 1.5 1.5 100% Total 262.7 262.9 100% Cumulative Estimated and Actual Disbursements (in US$ million) FY96 FY97 FY98 FY99 FY00 FY01 Appraisal Estimate n/a n/a n/a n/a n/a n/a Actual n/a n/a n/a n/a n/a n/a Actual as % of Estimate n/a n/a n/a n/a n/a n/a Project Dates Planned Actual Identification 05/12/93 Preparation Pre-appraisal Appraisal 04/28/94 Negotiations Board Approval 09/14/95 Loan Signature 10/10/95 Effectiveness 01/08/96 04/23/96 Loan Closing 06/30/02 06/30/02 Project Completion 06/30/02 Annex A 30 Staff Inputs (Staff Weeks) Stage of Project Cycle Actual/Latest Estimate No. Staff weeks uS$ (,000) Identification/Preparation 35.6 150.0 Appraisal/Negotiation 35.2 119.8 Supervision 49.6* 180.0" ICR Total * Because there were other ongoing activities during the project period involving the same task manager and consultants, close contact was maintained with the project than the one per year supervision missions indicated here. ** According to the ICR, this Information is as of June 1999. No further report was available under the SAP. Mission Data Stage of Project Cycle No. of Persons and Specialty Performance Rating (e.g. 2 Economists, I FMS, etc.) Implementation Development Month/Year Count Specialty Pors betv Progress Objective Identification/Preparation 5/93 5 A,B,D,F S S Appraisal/Negotiation Appraisal: 4/94 5 A,B,D,F S S Pre-Approval: 1/94 Negotiation.: 6/95 4 A,B,D,F Board: 9/95 3 B,G Signing: 10/95 Effectiveness: 4/96 Supervision 7/96 2* A,B S S 4/97 2* A,B 2/98 3- A,B,F ICR 6/99 4 A,B,F S S Specialty: A: Agricultural Economist B: Financial Analyst C: Agricultural Credit D: Financial Management Specialist E: Economist F: Environmental Specialist G: Legal Counsel 31 Annex B Annex B: PFIs and Sub-borrowers PFI Profitability The PCR for the First Project states that the impact on the profitability of individual PFIs was positive, but modest, because funds were spread among so many different PFIs. For eleven PFIs specifically reviewed, the estimated impact on net income ranged from less than 1% in the case of commercial banks to as much as 8% for thrift banks. Under the Second Project, fund use has continued to be widespread among PFIs (and interest-rate spreads have also been similar for both projects), so it is doubtful if a different conclusion will be reached. PFIs with Access to CLF Cancelled, Suspended or not Renewed It is interesting to note the numbers of PFIs that have had their access to CLF fund cancelled, suspended or not renewed and the reasons for this as given in LBP progress reports under the Second Project. Of the twelve commercial or thrift banks that were ineligible as of end 1998 (compared to 20 commercial banks and 14 thrift banks that continued to be accredited), most failed to meet either the CLF profitability or capital adequacy criteria, or both, and in all but two cases this happened during 1998.47 Among rural banks, the five that were suspended as of end 1998 failed to meet a variety of criteria (liquidity, capital adequacy, profitability, etc.), but for the thirty that had their CLF lines cancelled, by far the main reason was failure to use the line 4 PFIs Compared to Non-Participating Institutions To have a more complete picture of the projects' impacts on PFIs, additional data were requested from LBP and BSP in anticipation of the PAR Mission, focusing especially on past due loan ratios, the maturity profiles of loan portfolios, and deposits as funding sources, in order to compare the behavior of PFIs with institutions that did not participate and to note any significant changes over time. Because of the potential importance and perceived weaknesses of rural banks, and the fact that rural bank strengthening was a goal of both projects, it is useful to focus first on rural banks. As of end 1998 there were 38 rural banks with cancelled or suspended ALF/CLF credit lines, 24 that were accredited but did not use the lines during 1998, and 43 that actually participated during 1998. Table 4 shows average past due ratios, loans, deposits, and deposit to loan ratios for each category of LBP rural bank for 1996 through 1998, and for 1995 through 1997 for the universe of rural banks.49 None of the past due loan ratios is surprising for the different categories of LBP rural banks, as active PFIs are consistently lowest (best), followed by those not using, with those suspended or cancelled the highest (worst, and with marked deterioration in 1998). Moreover, all categories are better than the universe of rural banks for the 47. Past due loan ratios reached a peak in 1998, likely due to the impact of the Asian crisis and possibly the El Niflo weather phenomenon. 48. In addition to these 35 rural banks, there were 55 accredited rural banks of which 34 were actively participating as of end 1998. Meetings were held with a number of rural bankers at a conference in Surigao (in Mindanao) during the PAR Mission to the Philippines in order to explore reasons for the widespread failure to access CLF credit lines at LBP (see below). 49. Rural banks are slow in reporting to BSP so that overall data are not yet available for 1998. It has also been noted by observers that rural bank data may not be as accurate as desirable, but this should not be the case for rural banks receiving funds from LBP which should be carefully monitored. 32 Annex B two overlapping years of 1996 and 1997, which can be attributed to the effectiveness of LBP selection criteria even for rural banks not currently participating. Nonetheless, the deterioration in 1998 for all categories of LBP clients is troublesome, but unfortunately comparisons are not possible with the universe of rural banks, as overall BSP rural bank data are not yet available for end 1998. Deposit to loan ratios are consistently by far the lowest for actively participating rural banks, undoubtedly attributable to access to ALF/CLF funds. While this shows the effectiveness of the projects, it may also recall experiences with earlier credit programs when many rural banks failed, especially those highly dependent on external funds. Loans stagnated or fell for all categories of rural banks reporting in 1998, but deposits continued to rise except for rural banks that had their LBP accreditation suspended or cancelled where they fell sharply (after a modest fall the year before when all other categories were rising). These patterns reinforce the view that people continue to save and deposit during difficult economic times but carefully avoid making deposits in banks that may appear risky. It is also interesting to note that the average size of rural bank (measured by loans or deposits) is much smaller overall and for the "not using" category than for rural banks that are currently participating or have been cancelled or suspended. Conversations at a rural bank conference in Mindanao attended during the PAR Mission shed light on this interesting pattern.so The typical rural bank is much smaller than the average participating rural bank, and even those of good quality among small rural banks (having passed ALF/CLF criteria and having relatively low past due ratios) do not find ALF/CLF credit lines attractive, even though all accessed other LBP credit lines. The problem, all of them said, were the approval and monitoring requirements imposed on ALF/CLF-funded loans, which translated into high fixed costs, first, for setting up a unit to handle such loans and, then, for administering each individual loan. As result, accessing ALF/CLF funds was only worthwhile for a large volume of loans, and especially for loans of relatively large size. However, this implied risks of an excessively concentrated and undiversified loan portfolio for the typical small rural bank." Similar data for Philippine thrift banks are presented in Table 5, but these must be treated with great caution because there are only six thrift banks participating in ALF/CLF in 1998 and only three that have been suspended or cancelled (and data are reported for only one of these three). In any case, the past due ratios reflect the ability of LBP to have the better thrift banks among its active clients as it did for rural banks, and likewise the share of deposits in total funding sources is lower for participating than for all thrift banks, presumably due to their greater reliance on external funding from ALF/CLF. With respect to loan maturities and deposit composition, it was not possible to obtain comparable data for LBP clients, so that only the total of all thrift banks can be analyzed, of which the PFIs are only a very small proportion. Nonetheless, it is interesting to note that long-term lending has tended to expand relatively from 1994 to 1997, but mainly at the expense of medium-term loans, while the share of time deposits has fallen in favor of both demand deposits and savings deposits. This suggests that thrift banks have been engaging in more term transformation in recent years which may have added to their risk exposure (note the increase in past due ratios in 1997, with data for 1998 not yet available). 50. More than 20 rural bankers were interviewed individually on a variety of topics, but with special attention to use of ALF/CLF credit lines as compared to other LBP rediscount lines. 51. The fact that interest rate spreads on ALF/CLF funds are much larger on average for rural banks than for either commercial banks or thrift banks has led many observers to wonder if rural banks may not be highly inefficient. Rather, the spreads may be required to cover the costs and especially the risks of delivering a product designed by the Bank and LBP that is inappropriate for average size rural banks. 33 Annex B The data for Philippine commercial banks are presented in Table 6.52 For commercial banks, unlike rural and thrift banks, differences between the universe and the PFIs for past due ratios are negligible, including the marked increase in 1998, while those suspended or cancelled as usual show consistently much worse performance. Deposits relative to total funding sources appear stable over time for each category of commercial banks, but somewhat surprisingly the PFIs have a larger share of deposits in funding sources than all commercial banks (although this may be due to discrepancies in the way total sources are measured). Loan maturities show some tendency to lengthen over time, as they did for thrift banks, but short-term loans are still heavily predominant at 70%. Unlike thrift banks, the share of time deposits in total deposits has risen slightly over time, mostly at the expense of savings deposits. Commercial banks are thus tending to lengthen their maturity structures somewhat, but it should be noted that this is occurring largely in dollar deposits (while for loans information by currency was not available). Visits to Selected Sub-borrowers As part of the PAR Mission, visits were made to four LBP sub-borrowers. The selection of sub- borrowers to be visited was made from a complete listing of all sub-borrowers. In the selection process, exceptionally large sub-borrowers like Reynolds and San Miguel were omitted, and the search was limited to areas within three hours of Manila. The selection was difficult because the Mission wanted to identify agricultural or rural enterprises, but most sub-borrower investments were located in largely urban areas (for project purposes, rural was anything outside Metro Manila). In addition, most original CLF-I sub-borrowers had already repaid their loans, and LBP staff felt it would be difficult to get agreement for a visit. In the end, two sub-borrowers visited were from CLF-I, who had repaid their loans but were still willing to meet with the Mission, and two were from CLF-II. The investments financed were a mix: a biscuit (cookie) factory; a plastic bag manufacturer; a piggery; and a transportation enterprise (Jeepney for hauling goods). Except for the piggery case, none of the sub-borrowers felt that the availability of funding was a constraint to their businesses, nor was it prior to borrowing under the CLF program. All were able to borrow as and when needed from different sources, and long-term financing specifically was reported not to be a problem. Even for the relatively small investment in a Jeepney, long-term financing was said to be readily available from local finance companies, and the sub-borrower did not feel constrained. The piggery investor was the only one who noted a problem in obtaining term financing for the initial expansion of the piggery. However, the sub-borrower noted that the investment in sows is highly profitable and expected it to pay for itself within a year, questioning the rationale for borrowing for five years. Two of the operations, biscuit making and plastic bag manufacturing, are in no way small businesses. The biscuit manufacturer is among the largest in the Philippines and is heavily into exports. Although it is an agro-processing firm in theory, its benefits to the local agricultural economy are negligible. It imports all its machinery and equipment, as well as the milk, sugar and flour used to manufacture the biscuits (because of the better quality of imported raw material). In terms of employment, the sub-loan had a negative impact as it was used to import capital- intensive, labor-saving equipment to improve company profits. The plastic bag manufacturer is 52. Unlike rural banks where delays in reporting to BSP is the main reason for incomplete data, for commercial and thrift banks it is the lack of a breakdown into categories of deposits for PFIs and only partial data for PFIs on loan maturities, which are consequently unusable for making comparisons. Because about half of Philippine commercial banks are PFIs, contrasts between PFIs and the universe of commercial banks are not so revealing, while data over time for all commercial banks may be more revealing about PFI behavior than in other cases. 34 Annex B also located in an urban area and also used the sub-loan to import improved technology that cuts labor costs and processing time. In neither case are there any apparent benefits to the local rural economy. Why did the sub-borrowers take CLF sub-loans if financing is not a problem? The answer in all cases was better terms, that is, the sub-loan carried a lower rate of interest than was available elsewhere. This was confirmed by representatives (loan officers) of the PFIs who accompanied the visit. The cost of CLF loans to the PFIs was lower than the cost of funds from other sources, and these concessions were passed on to sub-borrowers for, among other reasons, fast disbursements to take advantage of rates lower than T-Bill rates 35 Annex C ANNEX C: LAND BANK OF THE PHILIPPINES BALANCE SHEET (In Million Pesos) 1990 1991 1992 1993 1994 1995 1996 1997 1998 TOTAL % TOTAL % TOTAL % TOTAL % TOTAL % TOTAL / TOTAL % TOTAL % TOTAL % ASSETS Cash & Due From Banks 4,400 19.3% 3,404 11.0% 8,169 16.5% 18,344 20.5% 13,687 14.1% 14,437 13.5% 18,367 13.7% 19,953 12.5% 19.505 11.0% Investments 6,980 30.6% 9,692 31.4% 14,901 30.1% 31,448 35.1% 37,382 38.5% 29,027 27.1% 29,398 21.9% 36,122 22.6% 39,679 22.4% Loans & Discounts LandAmorization(ARR) 2,013 8.8% 2,054 6.7% 2,00 4.2% 2,118 2.4% 2,065 2.1% - 0.0% - 0.0% - 0.0% - 0.0% Interbank Loans 532 23% 770 2.5% 1,470 3.0% 2,877 3.2% 1,363 1.4% 6,706 6.3% 2875 2.1% 2830 1.8% 3,841 2.2% Commercial Loans 3,301 14.5% 4,250 13.8% 6,413 129% 12,819 14.3% 18,753 19.3% 34,892 32.6% 55,585 41.5% 69,099 43.1% 76,692 43.2% ALFICLF Loans 1,873 3.8% 3,304 3.7% 3,334 3.4% 4,057 3.8% 4,913 3.7% 6,475 4.0% 6,981 3.9% Agr. RefornlOtherAgri. Loans 3,119 13.7% 5,985 19.4% 7,694 15.5% 10,713 12.0% 11,142 11.5% 11,245 10.5% 13,471 10.1% 14,759 9.2% 14,820 8.4% Bills Purchased Etc. 721 3.2% 583 1.9% 1,341 2.7% 1,875 2.1% 1,7568 1.8% 3,327 3.1% 6,012 4.5% 5,335 3.3% 9,521 5.4% Allowance for probable Losses (613) -2.7% (792) -2.6% (854) -1.7% (1,417) -1.6% (1,790) -1.8% (1,847) -1.7% (2,781) -2.1% (4,839) -3.0% (7,388) -4.2% Sub-Total Loans & Discounts 9,073 39.8% 12,850 41.7% 20,027 40.4% 32,289 36.1% 36,625 37.7% 58,380 54.6% 80,075 59.8% 93,659 58.5% 104,467 58.9% Bond Sinking Fund 160 0.7% 190 0.6% 220 0.4% 259 0.3% 298 0.3% 0.0% - 0.0% - 0.0% - 0.0% Land for Resale 131 0.6% 856 2.8% 2,077 4.2% 3,461 3.9% 5,383 5.5% - 0.0% - 0.0% - 0.0% - 0.0% Bank Premises, Other Real Property, Fum. and Equipt. 518 2.3% 774 2.5% 973 2.0% 1,135 1.3% 1,261 1.3% 1,387 1.3% 1,618 1.2% 3,425 2.1% 6.539 3.7% Other Assets 1,561 6.8% 3,071 10.0% 3,179 6.4% 2,572 2.9% 2,571 2.6% 3,689 3.5% 4,497 3.4% 6,990 4.4% 7,289 4.1% TOTAL ASSETS 100.0% 30,837 100.0% 49,546 100.0% 89,508 100.0% 97,207 100.0% 100.0% 133,955 100.0% 160,149 100.0% 177,479 100.0% UABILITIES AND CAPITAL Deposits (Pesos) 7,899 34.6% 11,282 36.6% 23,737 47.9% 54,401 60.8% 52.404 53.9% 70,836 66.3% 83,389 62.3% 97,597 60.9% 105,712 59.6% Deposits - Foreign Currency 361 1.6% 632 2.0% 946 1.9% 3,366 3.8% 2,609 2.7% 4,280 4.0% 7,736 5.8% 15,590 9.7% 18,249 10.3% Special Time Deposts/Financng 528 2.3% 101 0.3% 112 0.2% 144 0.2% 131 0.1% 138 0.1% 142 0.1% 186 0.1% 100 0.1% Bills Payable 1,653 7.2% 2,503 8.1% 5,341 10.8% 8,045 9.0% 12,694 13.1% 11.439 10.7% 15,676 11.7% 18,177 11.3% 20,700 11.7% Bonds Payable 1,412 6.2% 1,491 4.8% 1,725 3.5% 2,016 2.3% 2,497 2.6% - 0.0% - 0.0% - 0.0% - 0.0% Agrarian Reform Fund 1,919 8.4% 3,230 10.5% 3,507 7.1% 4,348 4.9% 5,367 5.5% - 0.0% - 0.0% - 0.0% - 0.0% Other Llabilftles 2,776 12.2% 4,347 14.1% 6,066 12.2% 8,179 9.1% 11,284 11.6% 9,027 8.4% 13,586 10.1% 13,809 8.6% 17,462 9.8% TOTAL LIAILTIES 16,548 72.5% 76.5% 41,434 83.6% 60,499 89.9% 86966 89.5% 95,720 89.5% 120529 90.0% 145,35 90.8% 162,223 91A% Paid In Capital - Common 1,800 7.9% 1,800 5.8% 1,800 3.6% 1,800 2.0% 1,800 1.9% 6,000 5.6% 6,650 5.0% 7,400 4.6% 7.800 4.4% Paid in Capital - Preferred - 0.0% - 0.0% - 0.0% - 0.0% - 0.0% - 0.0% 1.000 0.7% 1,200 0.7% 1,200 0.7% Paid in Surplus 57 0.2% 57 0.2% 57 0.1% 57 0.1% 57 0.1% 57 0.1% 57 0.0% 57 0.0% 57 0.0% Reserves - Alocated 1,833 8.0% 1,828 5.9% 3,262 6.6% 3,261 3.6% 3,854 4.0% 2,790 2.6% 2,793 2.1% 2,694 1.7% 2,696 1.5% UndMded Profits & Free Surplus 2t585 11.3% 3,566 11.6% 2,991 6.0% 3,891 4.3% 4,510 4.6% 2,352 2.2% 2,926 2.2% 3,440 2.1% 3,503 2.0% TOTAL CAPITAL 6,275 27.5% 7,251 23.5% 8,110 16.4% 9,009 10.1% 10,221 10.5% 11,199 10.5% 13,428 10.0% 14,791 9.2% 15,256 8.6% TOTAL LIABILIES AND CAPITAL 22,823 100.0% 30,837 100.0% 49,544 100.0% 89806 100.0% 100.0% 100.0% 133,957 100.0% 150150 100.0% 177,479 100.0% 36 Annex D Comments from the East Asia Region THE WORLD BANK/IFC/M.I.G.A. OFFICE MEMORANDUM DATE: June 26, 2000 TO: Ridley Nelson, Acting Manager, Sector and Thematic Evaluations Group FROM: Mark D. Wilson, Acting Director, EASRD EXTENSION: 39200 SUBJECT: Philippines: First and Second Rural Finance Project - Comments on June 22, 2000 Draft OED's Projects Performance Audit Report 1. Thanks for sending us the second draft PAR of June 22, 2000 of the above mentioned projects. While most of the report is reasonable and there are no major differences between the audit and the PCR of Rural Finance I, and the ICR for Rural Finance II, there are two key areas, where differences are substantial. These are related to the conceptual approach regarding additionality and sustainability. We, therefore, request that the Region's comments, as set out below, be clearly reflected in the final version of its PPAR. Additionality 2. The Region does not agree with OED's view that the project did not result in additionality on credit availability in the country. If a credit operation is to finance only viable investments, additionality can be measured only through macro indicators. Subject to availability of funds, viable investments can always be financed by other financial institutions. Under the project, commercial sub-project selection criteria were applied, and as a result, the investments supported were sound. Because of this, many of the specific investments financed under the project would probably also have been made 'without project' using alternative sources of funding. Viable sub-projects, by definition, may be financed by other financial institutions. 3. The project's additionality, in this regard, should be assessed not on whether specific investments were induced, but rather on whether it resulted in an increase in overall investments, and further that it did not 'crowd out' deposits, particularly term deposits. Macro data for the Philippines banking system as a whole indicates that over the project period, there was on aggregate a substantial expansion in intermediate and long term loans. Between 1995 and June 1998, intermediate and long-term loans by commercial banks as a whole (including non-rural areas) grew by about P364 billion or 320%, while the corresponding growth of short- term and demand loans was 208%. Project contribution to this growth was 1%. Project 37 Annex D contribution in terms of rural investments is likely to be much larger than 1%, as not all of the P364 billion total growth in lending went to rural areas. 4. Over the period between December 1995 and December 1998, deposits in the banking system grew by 72%, while time deposits alone grew by about 120%. In absolute terms the growth in the amount of time deposits was P280 billion, compared with project's funds of P3.6 billion. Furthermore, over this period, Philippine banks, on average (based on figures of 30 banks out of 54 commercial banks) had an adequate risk assets to equity ratio; consequently, limitation to lending was availability of resources. A reasonable conclusion to be drawn from this data is that the project contributed to the overall expansion of the availability of longer term resources to finance investments. There is no evidence of 'crowding out'. 5. The audit approach, however, deals with additionality in a very narrow manner by checking whether or not a specific sub-project would have been financed by other financial institutions, in the absence of the Bank finance projects. Under this approach, the project and the Bank will always be wrong. Either the project finance viable sub-projects and wins on sub- loan collection but loses on additionality, or the project finance non-viable sub-projects, that no other financial institution is willing to touch, and wins on additionality but loses on loan collection. Sustainability 6. A major difference between the PAR and the two completion reports is the perception regarding CLF's sustainability. The ICR and the PCR approach is that the Projects would be 'sustainable' if there were a continuation of significant wholesaling of LBP's CLF funds (or replacement). The ICR rated it as highly likely because of the followings reasons: (i) the CLF is profitable to all players (the apex bank, PFIs, and sub-borrowers); (ii) CLF operation is efficient and effective with excellent collection rates, which indicates that accreditation of PFIs and selection of sub-projects were done properly, in professional manners; (iii) following these two loans, the Bank has introduced a third loan (RF IHI), using the same wholesale operation to finance viable rural investments, as well as to alleviate the credit crunch that resulted from the Asian financial crisis. Certainly as far as RF I is concerned, the program will be sustainable. CLF will continue for at least another five to ten years (which is 14-19 years since RF I was initiated). Even if Land Bank were to get into some financial trouble, it is unlikely that this wholesaling operation, which has low risk, would be shut down. Furthermore, before the Asian financial crisis, LBP was able and actually issued its own bonds to mobilise US$250 million for five years. While the Asian crisis interrupted this process of mobilisation of medium term lending resources, this process is planned to be continued in 2001. cc: Madhur Gautam (OEDDI) 38 Annex E OED response to East Asia Region's comments Additionality The audit clearly states, para 4.16, that the issue of additionality is open to debate. The audit does not maintain that there was no additionality; rather it arrives at the conclusion that it is not possible to establish the additionality of project funds, or the lack of it. The reason is the lack of data which would allow developing an appropriate counterfactual: what would have happened in the absence of the projects? Trends in the macro aggregates for the entire commercial banking sector by themselves do not provide a sufficient basis to establish additionality. As the region notes, the projects added only 1% to the growth of intermediate and long-term lending by commercial banks between 1995 and 1998. However, there is no evidence that the project had either a positive or a negative impact on the growth rate. Similarly for deposits, there is no evidence that links the project to the outcomes, especially the outcomes for the entire sector. The Region argues that there is no evidence of "crowding out". The audit notes (pare 4.16) that there is also no evidence to suggest that deposits would not have grown faster without the projects. The audit approach, as presented in paras 4.11 to 4.16, is quite different than simply looking at the possibility of financing viable sub-projects from other sources. As noted in footnote 37, the issue is one of what the projects' contribution was at the margin. The audit tried to compare the performance of institutions that participated with those that did not participate in the projects. With the data in hand, however, this cannot be conclusively established. The limited information that is available, however, is not favorable to additionality. In this respect, it should also be noted that the project made no provisions to collect the appropriate data to demonstrate this important aspect of the projects' contributions. Sustainabiity First, it should be noted that the audit rates sustainability for both RF I and II as likely. The only difference is between the ICR for RF II, which rates sustainability as highly likely (the PCR for RF I rates sustainability as likely). The audit raises some concerns about future impact of certain policies that are beyond the control of LBP (paras 5.10 to 5.12). Since these problems could become potentially serious, the audit is a bit more guarded than the ICR in its rating of sustainability for RF II. Borrower's Comments 9 Annex r LAND BANK OF THE PHILIPPINES FLORIDO P. CASUELA 319 Sen Gi Pjva Ave Er President Makat City 120C ohi11ppines and Te4 943a Nos 843-8824 e450P Cief Executive oicer June 20, 2000 MR. RIDLEY NELSON Acting Manager Sector and Thematic Evaluations Group Operations Evaluation Department The World Bank Dear Mr. Nelson: This refers to the draft Performance Audit Report on the Rural Finance Projects I & 11 (Loans 3356, 3938, 3939 and 3940). Generally, we find the audit report acceptable. However, we are enclosing our comments to some audit observations which need to be corrected or clarified. We have also included updates on significant developments related to some of your audit findings. We did not anymore include in our comments those which have been raised by the WB Task Manager, Mr. Arie Chupak, as Mr. Madhur Gautam advised our Program Lending Group that those have been addressed already in the revised text. With my best regards. Very truly yours, FLORIDO . CASUELA President & CEO cc: Mr. Madhur Gautam Sector and Thematic Evaluations Group Operations Evaluation Department Mr. Are Chupak Rural Development & Natural Resources Sector Unit Mr. Vinay Bhargava WB Manila Office k LANDBANK 40 Annex F LBP Comments on the WB-OED Performance Audit Report 1. Para. 2.2 "... Thus, while LBP collected estimates of the number of jobs created by each loan from the sub-borrowers, and reported these numbers regularly to the Bank as part of project monitoring requirements, there was no attempt to verify the numbers reported." This is not accurate. LBP conducts sub-project impact assessment on random basis (as we do not have the resources to conduct 100% checking) and one of the main areas verified is the number of employment generated vis a viz projections. We have observed that the actual number of employment generated approximated the projections. Based on a total of 120 RF I & II projects which we have visited, the actual employment generation which was reported at 3,792 exceeded the total expected employment generation of 3,177. 2. Para. 2.4 - LBP Dividends to the Government of the Philippines (GOP) The cash dividend payment for CY 1998 & 1999 was recently discussed during the meeting last 5 June 2000 with the representatives from World Bank (Mr. A.Chupak, P. Harrison & A. Ordu), LBP ( Pres. F. Casuela, FVP N. Veran & J. Climaco Jr.) and the Department of Finance (Sec. J. Pardo, Usec. F. Tarriela, Usec. L. Gruba & L. Magdamo). The World Bank and the LBP representatives presented the computation of net income available for cash dividends. Following the computation based on the loan and guarantee agreements, LBP is not required to pay cash dividends to the national government for CY 1998 & 1999. Sec. Pardo has agreed in principle that LBP will not pay cash dividends for 1998 & 1999. With these positive developments, we are confident that in the future the DOF will follow the provisions in the loan and guarantee agreements in computing for the cash dividends to the National Government. 3. Para. 2.8 - Several secondary components in RF-l were not completed (e.g. PCIC and QGFB strengthening and ACPC studies). Pursuant to the program covenants with the World Bank under RF I, LBP commissioned the ACPC per Memorandum of Agreement (MOA) signed on October 26, 1994 for the conduct of studies on rural finance. The MOA stipulated that the conduct of the studies would be accomplished within twenty-one (21) months commencing on day of the issuance of a Notice to Proceed by LBP (issued on June 15, 1995). However, due to the realignment of functions and personnel movement within ACPC, the completion of the studies was delayed by more than two(2) years as ACPC came up with the final drafts of the research only in July 1999. The results of the studies are now for publication. Modules covered increased from the required two (2) studies to include Extent of Agricultural Lending of Non-Government Organizations and Cooperatives and Evaluation of LBP Group Lending and Rediscounting Programs. 41 Annex F Also, consonant with the secondary components of the subject program, LBP, in collaboration with BSP and PDIC implemented the Countryside Financial Institutions Enhancement Program (CFIEP) in 1991 consisting of three modules, namely: Conversion of RB's arearages into equity (Module 1), Provision of Equity Counterpart by LBP (Module 11); and Financial Incentives of PDIC for merging countryside financial institutions (CFIs) (Module Ill). A total of 142 CFIs participated in the program involving total equity infusion/conversion of P532.513 Million. These modules greatly contributed in the rehabilitation of the ailing CFIs in the countryside. In addition to the capital strengthening program, LBP, through its subsidiary, the LANDBANK Countryside Development Foundation Inc. (LCDFI), implemented the CFI Management Development Programs I & 11 which basically involves training and systems installation depending on the specific requirements of the CFIs. Funds for these programs were also taken from the gains of the debt-buy back scheme. Under the CFI-MDP I conducted from 1993 to 1995, three (3) training modules were participated in by 190 CFIs. In the subsequent program (CFI-MDP II) which ran from 1995 to 1998, seven (7) training modules were conducted and participated in by 201 CFIs. As regards the impact /benefits of these training modules, the LCDFI reported that the financial conditions of the program participants have generally improved with sustained growth in capital, deposits, resources and income coupled with enhanced systems in credit process and internal control. Currently, the LCDFI is requesting additional financing assistance from the gains of the debt-buy back to pursue a third tranche of CFI-MDP where three (3) new training modules will be introduced and focused assistance to cooperative rural banks will be given. 4. Para. 2.10 - "Under both projects, foreign exchange costs are assumed almost entirely by the GOP, to be covered by the spread between the interest rate at which Bank funds are received by the Department of Finance (DOF) and on-lent to LBP as determined by a market-based formula....(LBP and DOF seemed unconcerned/unaware that this involved heavy costs to be borne by the government)." The statement that WB funds are received by the DOF and onlent to LBP is incorrect. Under both RF I & II, the funds are directly lent by WB to LBP and notthru the DOF. The GOP (as represented by DOF) acts as guarantor and provides foreign exchange risk cover on peso sub-loans. The FX risk on Dollar sub-loans under RF 11 is borne entirely by the sub-borrowers. Based on the standard arrangement among the government financial institutions (LBP and the Development Bank of the Philippines) and the government (thru DOF), a market-based formula is being adopted to cover for the FX risk. There is a self- compensating mechanism in this arrangement as GOP receives and is able to accumulate huge amount of fees during high interest rate scenario as what generally occurred in 1991 to 1998. Thus, heavy FX losses as what happened during the financial crisis in 1997 can be compensated by the fees accumulated by the GOP. As gathered from the DOF, the GOP has maintained a positive position in its provision of the FX risk cover. 42 Annex F Also, we wish to note that in addition to the foreign exchange risk cover fee directly remitted to the GOP, the GOP also earned the Guarantee Fee and incidental taxes related to the project which can be considered as additional buffer for the foreign exchange rate differential. 5. Para. 2.11 - "Another neglected aspect of cost is that, in calculating the profitability of the two projects for LBP and thePFIs, no attention has been paid to the operating costs incurred by LBP in operating the credit lines or to the added administrative costs that PFIs may have incurred in using the credit lines. The two projects might not have appeared as highly profitable if these costs had been included." The Wholesale Lending Department (WLD) within LBP-Program Lending Group is mainly reponsible for the implementation of the credit components of RF I & II. All costs directly identifiable in running the program are fully accounted. However, there are other units within LBP which provide operational support to WLD. But since these support units also provide services to other LBP units, costs should be properly allocated. Thus, LBP Management has recognized the need to put in place a transfer pricing mechanism to be able to account for the all the cost incidental in implementing the program. As soon as this system is in place, all costs shall be considered in determining the programs' profitability. Nevertheless, the manpower and operating expenses of the Wholesale Lending Department (the RFI & II program manager) for CY 1998 was reported at P7.763 Million. Assuming that the cost of services of the various support units will be the same, the administrative costs of RF I & II will be estimated at P 16.000 Million. Thus, taking into account the administrative expenses, the total RFI & II net income for the year ended December 31, 1998 will result to a minimal decrease from P403.53 Million to P387.53 Million. As regards the operating expenses of the PFIs, it is not feasible for LBP to conduct a periodic analysis on the operating expenses of the PFI considering the number of the PFIs and the relevance of this information to LBP. The cost benefit analysis of trainings would be very difficult to conduct considering the highly qualitative impact of trainings and the different variables affecting PFIs' overall performance. Nonetheless, LBP shall try to institute a feedback mechanism by conducting a survey (say, after one year of training) to determine the impact of the training to the PFIs and LBP staff and what other trainings are still needed. Implementation 6. Para. 3.4 - This para. discussed RF I implementation results based on loan amounts. In assessing the impact of RFI & II, the audit should not only focus on the amount of loan disbursed to the small and medium enterprises. Impact based on the number of sub-projects assisted should likewise be considered. RF I was able to assist 677 in the small and medium asset size group (defined at the time of PCR preparation 43 Annex F as those with assets up to P50.0 Million) which represented 98% of the number of sub-projects. Of all the PFI categories, the rural banks led in terms on number of sub-projects assisted with 242 out of the 687 sub-projects. Also, we wish to correct the data that there were 16 rural banks which participated under RF I. It should be reported as 20 (OED data was perhaps based on the number of rural banks with outstanding CLF I loans as of PCR date and thus was not able to capture rural banks with fully paid accounts). 7. Para. 3.5 - This para. discussed number of loans versus target and lending concentration in more developed regions Regarding the number of loans under the RFP II, the inability to exceed the target was a result of the delay in the disbursement of the retail component which caters to smaller asset size sub-borrowers with lesser funding requirements than those under the wholesale component. We expect that upon full disbursement of RF II, the number of loans will meet the targeted level. We wish to note that during our marketing activities, we gave emphasis in promoting lending to small and medium enterprises and in areas which are less developed. However, given the market-driven approach of the projects, the PFIs ultimately decide on the sub-projects to be financed based on their evaluation and credit judgment. Moreover, the regional volume of loan availments is greatly affected by the level of progress or economic activity in each area. 8. Para. 3.6 - Improvements in the performance of cooperatives. We measure the cost and benefits from training programs implemented for the cooperatives through their performance. Agrarian loan releases (loans to small farmers and fisherfolks) more than doubled to P14.05 billion in 1999 from P 6.78 billion in 1995 (Refer to Table 1). Agrarian loan portfolio quality also improved as the past due rate declined to 20.9% in 1999 from 29.7% in 1996 (Refer to Table 2). Valuation reserves to past due loans also improved to 92.6% in 1999 from 76.5% in 1996. Moreover, agrarian unit operating cost declined to 10.6% in 1999 from a high 13.7% in 1996. 9. Footnote 14 - On the reduction of delinquency rate of LBP Agrarian Portfolio The improvement in delinquency ratio in Agrarian loans cannot be accurately attributed to the shift in composition from coops to rural banks as there has not really been a drastic change in the agrarian loan portfolio mix. Rather, the reduction in delinquency rate is attributable to prudence in credit administration which is partly a result of the establishment of the Cooperative Accreditation Criteria. 10. Para. 4.3 - "...For 1997 and 1998, however, the quarterly progress reports indicate that more than half of the profits are accounted for by income from investments rather than from sub-loans." 44 Annex F This statement is inaccurate. The quarterly reports explained that the present accounting system did not fully reflect the cost of unrelent funds. The income on investments of unrelent funds represented net income after deducting the 20% withholding tax on the investment earnings only . All the other costs such as the borrowing cost from World Bank, guarantee fee, gross receipts tax and foreign exchange risk cover fee were absorbed by the relending operation. If the financial cost of the funds invested will be proportionately allocated to relending and investment operations, net income on investments will be negligible. (After the estimated allocation of cost to relending and investment operations, income on CLF II as as of December 31, 1998 was reported at P58.339 Million while net income on investments amounted to (P9.151 Million). 11. Para. 4.4 - Warnings are nonetheless raised about LBP's long run viability. LBP increased its capitalization by issuing P 2.0 billion preferred shares to the National Development Company on May 31, 2000. The strengthening of the capitalization will improve LBP's capital adequacy ratios, resulting to stable long-term financial position. 12. Para. 4.5 - Share of ALC/CLF Loan to Total Loans & Share of Large Deposits The percent share of ALF/CLF loans to gross loans was at 6.2% and 6.4% in 1999 and 1998 respectively (Refer to table 4). It is more appropriate to compare the share of ALF/CLF loans to gross loans rather than total assets because the growth in total assets is mainly driven by the growth in deposits. Private deposits sustained a significant growth rate since 1994. The share of private deposits to total deposits increased to 38.3% in 1998 from only 16.9% in 1994. The level of private deposits also increased to P 47.5 billion from P 9.3 billion for the same period. Various deposit generation and retention campaigns such as "You make the Difference I & II", "Todo Bigay Bagong Bahay" and the latest "Salubong sa Milenyo" initiated by the LBP triggered the growth in private deposits. The large deposit accounts (P200,000 and above) of LBP are only 3.4% of the total number of accounts (Refer to Table 5). More than 96% ofLBP's depositors have deposit balances of P 200,000 and below. While it is true that the amount of P 200,000 and above represents 90% of total deposits, this is brought about by the bulk of deposits that come from the government sector, LBP being the main government depository bank. 13. Para. 4.7 - "The issue of widespread failure of approved rural banks to use the CLF credit line was explored in Mindanao, and the high minimum size of sub-loans under the projects appears to be the main impediment to use.... Minimum sub-loan sizes were too large for their operations and for their clients' needs. The average CLF availments of rural banks for RF I & II amounted to PO.483 Million and P 1.018 Million (as of December 31, 1998), respectively . Thus, the minimum loan size under RF I & 11 which is P100,000.00 ($2,500) cannot be said to be too large for RB operations. Also, we understand that the RBs in Mindanao 45 Annex F that were interviewed by the OED Audit Team were mostly not accredited under the CLF Program or those which have not participated under the program. 14. Para. 4.10 - Quality of LBP's cooperative loan portfolio deteriorated despite all trainings devoted to cooperatives. Please see our similar comments to Para 3.6 Additionality 15. Para. 4.11 - "Would the activities undertaken with the project undertaken in any case? Would the sub-borrower have found alternative source of funding? Would the PFI have made the loan available from another source?" The sub-projects funded by the programs have been evaluated as feasible. As such, it is true that they can find alternative sources of financing had there been no RF I or II. If the auditors will gauge additionality by financing only sub- projects which cannot avail other financing options, then the project should have targetted specific beneficiary and should not be market driven. Moreover, loan collection and sustainabiity of the project will suffer if we will fund sub-projects wherein there are no other willing lenders. 16. Para. 4.12 Discussion on results of audit interview with four (4) sub-borrowers The sample of sub-projects taken during the OED review (4 out of 1337) was indeed very small to make sweeping conclusions. The projects are market driven and while employment generation are encouraged, sub-projects which would like to increase their efficiency thru mechanization are not discriminated. What should be considered is the overall impact of the project given the increase in local investments, direct and indirect employment opportunities and incremental taxes and government revenues. Same comment applies to Annex B re: Visits to selected sub- borrowers. 17. Para. 4.13 - On the two conclusions made (a) LBP was able to select better financial institutions but not necessarily able to make them better (b) LBP's CLF loan product was not appropriate for the average-size rural bank, which led to higher spread and possibly also to greater risks. We cannot confidently claim that financial institutions will become better-off when they participate in the CLF Programs as there are lot of factors to consider in the improvement of the PFIs' operations, such as the quality of management, level of capitalization, etc. Also, if the P100,000.00 ($2,500) minimum CLF loan size (as mentioned in our comments to Para. 4.7) is too big for average-size rural banks, it is nevertheless a good alternative source of long term funds for bigger projects of rural banks. Borrower Performance 18. Footnote No. 38. "In addition, while LBP did well what was expected of it, Bank 46 Annex F staff could have expected more of LBP staff in terms of thinking about what they are doing (underscoring supplied) (e.g., effective cost -benefit analysis of training programs in addition to effective implementation). We suggest that this be either deleted or rephrased. We find this statement unnecessary as the content has already been reflected in para. 5.17 and the language quite offending. Annex G OED response to Borrower's comments The responses below correspond to the comments by number. 1. The text has been modified to reflect the end-use sample monitoring. However, during the audit mission, discussions with LBP staff indicated that they relied on the numbers reported by the sub-borrowers but there was no further follow up. 2. The issue of dividend payments has still not be permanently resolved. The new information provided in LBP comments has been incorporated in footnote 4. 3. The text has been modified to reflect that the studies were eventually completed. The audit is also more concerned with the project objectives which applied to PCIC and QGFB. 4. Para 2.10 has been modified accordingly. On foreign exchange losses, the audit mission was unable to obtain appropriate figures either from LBP or the government. The ICR mission subsequently conducted an analysis and concluded that the foreign exchange risk cover fee will be inadequate. The financial impact of the whole project on the GOP may be still positive as a result of guarantee fees and taxes, but it should be noted that protecting against foreign exchange fluctuations is not the purpose of the guarantee fees. Further, although the net positive impact may hold ex-post for the past losses, there is no guarantee this will continue to so in the future. 5. The audit is not implying improper accounting practices. The issue is a proper analysis of the benefits relative to costs. However, as noted later in LBP's comments, certain costs have not as yet been fully accounted for and new systems are being put in place to correct for this. The audit has been revised to reflect this. 6. The basic argument the audit is making remains the same. However, footnote 14 has been added to include the additional information and the figure on the number of rural banks has been corrected. 7. The text has been modified to include the reason for not meeting the target. 8. The arguments in the text are not inconsistent with the new data. 9. Footnote 15 (previously footnote 14) has been modified. 10. The text has been modified. 11. A new footnote (26) has been added to reflect the new information. 12. The additional information is presented in a new footnote (27) and the main text modified. However, the essential argument of the audit remains the same. 13. The comments on the size of the minimum CLF loan size are those made by the rural bankers. As noted in the audit, several of the interviewees did apply for the CLF but later did not use it. Annex G 48 14. The audit discussion refers to the RF I period whereas the data presented in the comments refer to the years after RF I. Nevertheless, a footnote has been added to note the improved performance in subsequent years. 15. The audit agrees with the statement. 16. The audit is not making any sweeping conclusions. It is reporting the information gathered for what it is worth. However, the sub-projects were chosen at random from the full list of projects, in full consultation with LBP staff. The difficulties in visiting additional sites, noted by LBP staff, are discussed in Annex B. Nevertheless, the projects are indicative, and some of the findings, e.g., on the relative attractiveness of CLF funds, as verified by the accompanying LBP and PFI staff members, were representative of general trends. Irrespective of their lack of statistical robustness, the information cannot be simply ignored. Further, the information collected is consistent with statements on the low incrementatility of RF I on jobs created in LBP's own project completion report. 17. The audit agrees with the first comment. However, on the second, while it may be true, it is apparently not a view shared by the rural banks who chose not to access CLF funds. 18. The audit agrees and the footnote has been dropped.
Groupe de la Banque mondiale · Project Performance Assessment Report
Philippines - First and Second Rural Finance Projects
Voir le document original
Le texte intégral est hébergé par l’organisation qui le publie. lawenc.com indexe les métadonnées et renvoie vers la source officielle.
Texte intégral
Informations clés
Organisation
Groupe de la Banque mondiale
Type de document
Project Performance Assessment Report
Pays
Philippines
Source
Banque mondiale