Philippines - Agricultural Credit Project

Document of The World Bank FOR OFFICIL USE ONLY Report No. 10969 PERFORMANCE AUDIT REPORT PHILIPPINES AGRICULTURAL CREDIT PROJECT (LOAN 2570-PH) JULY 28, 1992 Report No: 10969 Type(PPAR) Cov.F AGRICULTURAL CREDIT PROJECT RICE, E.B./ X31738 / T9059/ OEDDI Date Ent.: 19920724 By: MAA Operations Evaluation Department 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 Name of Currency: Peso (P) Rate of Exchange: End of Year 1984 US$1.00 = P20.00 1985 = P18.61 1986 = P20.39 1987 - P20.57 1988 = P21.10 1989 = P21.74 The US$ loan amounts in this report should be understood to mean US$ equivalent. ABBREVIATIONS AND ACRONYMS ACPC - Agricultural Credit Policy Council ALF - Agricultural Loan Fund ALFU - Agricultural Loan Fund Policy Unit ARF - Agrarian Reform Fund Bank - World Bank Group CALF - Comprehensive Agricultural Loan Fund CARL - Comprehensive Agrarian Reform Law CARP - Comprehensive Agrarian Reform Program CBI Central Bank Institute CBP - Central Bank of the Philippines CLF - Countryside Loan Fund DA Department of Agriculture DBP - Development Bank of the Philippines B Institution Building IMF - International Monetary Fund KB - Commercial Bank LBP Land Bank of the Philippines M&E - Monitoring and Evaluation PCR Project Completion Report PFI - Participating Financial Institution PAR - Performance Audit Report RB Rural Bank SAR - Staff Appraisal Report SES - Supervision and Examination Section STD - Special Time Deposit TA - Technical Assistance TB - Thrift Bank TBAC - Technical Board for Agricultural Credit U.S.AID - United States Agency for International Development FISCAL YEAR January 1 - December 31 WEIGHTS AND MEASURES Metric System FOR OFFICIAL USE ONLY THE WORLD BANK Wasingof, D.C. 20433 U.SA Office of Director-General Operations Evaluation July 28, 1992 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Performance Audit Report on PHILIPPINES Agricultural Credit Project (Loan 2570-EH) Attached, for information, is a copy of a report entitled "Performance Audit Report on the Philippines: Agricultural Credit Project (Loan 2570-PH)" prepared by the Operations Evaluation Department. Attachment 2 This document has a shstictd distribedon and mmy be =se by edplaenm onAsi th peformace of thei *elda dales. its comnt may not odlwise* be disld withea Worl Bank authorisila FOR OFFICIAL USE ONLY PERFORMANCE AUDIT REPORT PHILIPPINES AGRICULTURAL CREDIT PROJECT (Loan 2570-PH) TABLE OF CONTENTS Page No. Preface .i Basic Data Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . ift Evaluation Summary . . . . . . . . . . . . . . . . . . . . . . . . . v I. BACKGROU1ND * * * * * * * * * * * * * * * * * * . . . .I I ACKPReparation . . . . . . . . . . . . . . . . . . . . . . . . 1 A. Preparation . . . . . . . . . . . . . . . . . . . . . . . . 3 C. Project Design . . . . . . . . . . . . . . . . . . . . . . 5 II. IMPLEMENTATION EXPERIENCE . .N P . . . . . . . * * * *. . . .o 7 A. Project Costs and Loan Disbursements . . . . . . . . . . . 7 B. Onlending Component . . . . . . . . . . . . . . . . . . . . 8 C. Technical Assistance Component . . . . . . . . . . . . . . 11 D. Transfer of ALF . . . . . . . . . . . . . . . . . . . . . . 11 B. Bank Performance . . . . . . . . . . . . . . . . . . . . . 12 III. PROJECT OUTCOME . . . . . . . . . . . . . . . . . . . . . . . . 13 A. Policy Impact . . . . . . . . . . . . . . . . . . . . . . . 13 B. Institutional Strengthening . . . . * . . . . . . . . . . . 15 C. Deposit Mobilization . . . . . . . . . . . . . . . . . . . 16 D. On-Farm Recovery and Development . . . . . . . . . . . . . 16 E. Technical Assistance Component . . . . . . . . . . . . . . 19 F. Balance of Payments Support..... . . . . . . . . . . . 20 IV. FINDINGS AND ISSUES . . . . . . . . . . . . . . . . . . . . . . 20 A. Participation by Small and Medium Scale Farmers . . . . . . 20 B. Role of the Rural Banks . . . . . . . . . . . . . . . . . . 21 C. ALF Interest Rates . . . . . . . . . . . . . . . . . . . . 22 D. Policy Reform . . . . . . . . . . . . . . . . . . . . . . . 23 E. Were the Results of the Project Satisfactory? . . . . . . . 24 F. Lessons for Credit Policy . . . . . . . . . . . . . . . . . 25 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. Pano No. 1. AFL Lending Patterns by Subsector . . . . . . . . . . . . . . . 9 Anne 1. Agricultural Production Loans Granted, by Institution 1985-89 * * * * * * * * * * * * * . . .0. * 29 2. Characteristice of AFL Loan Availments . . . . . . . . . . . . 30 3a. ALF Interest Rates . . . . . . * * * * . * * * * * * * . . . 31 3b. Currency Devaluation, Inflation and Interest Rates under ALF * * * * * * * * * * * * * * * &.. 32 4. Regional Office Memorandum on the Draft PAR . . . . . . . . . . 33 5a. CBP letter on the Draft PAR . . . . . . . . . . . . . . . . . . 40 5b. LBP letter on the Drfat PAR . ............ 43 Map IBRD 22431 PERFORMANCE AUDIT REPORT PHILIPPINES AGRICULTURAL CREDIT PROJECT (Loan 2570-PH) PREFACE This is a Performance Audit Report (PAR) on the Philippines Agricultural Credit Project, involving an IBRD Loan equivalent to US$100.0 million. The objectives were to reorient credit policy, improve rural credit institutions, stimulate agricultural recovery, and provide balance of payments support. The Loan was approved on June 6, 1985. It was closed, fully disbursed, on October 31, 1989, fourteen months before the original Closing Date. The PAR is based on the Project Completion Report (PCR) prepared by the Asia Regional Office and submitted to the Board on September 30, 1991, the Borrower's PCR submitted to the Bank in 1990, the Staff Appraisal Report (SAR), the President's Report, the Loan documents, a report of the Executive Directors' meeting at which the project was considered, a study of project files, and discussion with Bank staff. An OED mission visited the Philippines in October 1991 and discussed the effectiveness of the Bank's assistance and project execution with the Central Bank of the Philippines (CBP) and other relevant agencies. CBP's kind cooperation and valuable assistance in the preparation of this report are gratefully acknowledged. The PCR provides a good account and assessment of the project experience, and the performance of the Bank and the project executing authority. The PAR elaborates on particular aspects of the overall lending operation, which is of special importance in the evolution of the Bank's own rural credit policy. Following standard OED procedures, copies of the draft were sent to the Government for comment in April, 1992. Responding letters from CBP and the Land Bank of the Philippines are included in the Annexes. A memorandum of the Regional Office, giving additional background and disputing several audit conclusions, is also annexed. PERFORMANCE AUDIT REPOPI PHILIPPINES AGRICULTURAL CREDIT PROJECT (Loan 2570-PH) BASIC DATA SHEET KeT Project Data Appraisal Actual Estimate Performance % Project Costs (equiv.US$m) 183.0 217.4 119 Loan Amount (equiv.US$m) Disbursed 100.0 100.0 100 Short Term Credit 45.5 56.0 123 Medium and Long Term Credit 54.0 44.0 82 Number of Sub-loans 10,000+ 2,013 PFI Lending in Project Period 175 ' 41 Number of Policy Studies 5 4 80 Economic Rates of Return 1 37-84% 15-30% Financial Rates of Return 1i 25-46% 15-30% 11 Accredited at time of appraisal. gi Range for indicative models of sub-projects. The common figures for ERR and FRR actuals are taken from the PCR. This is an unusual relationship and may involve error. Cumulative Actual and Estimated Disbursements FY86 FY87 FY88 Eng FY90 Appraisal Estimate (equiv.US$m) 46.00 20.00 14.00 16.00 4.00 Appraisal Cumulative 46.00 66.00 80.00 96.00 100.00 Actual 22.92 8.51 19.19 39.67 9.71 Actual Cumulative 22.92 31.43 50.62 90.29 100.00 Actual as % Appraisal 50 48 63 94 100 Date of Last Disbursement: October 10, 1989 -iv- Project Tim.table Item Date Planned Date Actual Identification - February 1982 Preparation - 1983/1984 Appraisal - October 1984 Negotiations - April 24, 1985 Board Approval - June 6, 1985 Loan Signature - June 10, 1985 Effectiveness September 9, 1985 August 16, 1985 Project Completion December 31, 1989 October 31, 1989 Closing Date December 31, 1990 October 31, 1989 Staff Inputs Total FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 Ident./Preparation 62.7 0.6 50.2 11.9 Appraisal 64.5 64.5 Negotiations/Board 1.8 1.8 Supervision 39.9 0.6 10.4 14.7 10.1 2.5 1.5 Completion 15.0 15.0 Total 183.9 0.6 50.2 78.8 10.4 14.7 10.1 2.5 1.6 15.0 Missiun Data T.O.R. month/ Number of Staff Days Speciality Performance Type of Date Year Persons in Field Rating Problem Before Appraisal 1) Ident./Preparation 02/83 3 21 A/B 2) Ident./Preparation 05/16/83 06/83 1 7 A 3) Preparation 08/24/83 09/83 1 10 A 4) Preparation 01/06/84 01/84 3 25 A/B/C 5) Pre-appraisal 04/23/84 05/84 6 115 A/B/C/D/E Appraisal 09/24/84 10/84 4 100 A/B/C/D Supervision 1 06/12/85 06/85 1 15 B I n.a. Supervision 2 10/23/85 10/85 1 10 B 1 n.a. Supervision 3 06/13/86 06/86 1 7 B 1 n.a. Supervision 4 02/02/87 02/87 1 12 B 2 n.a. Supervision 5 01/13/88 01/88 1 8 B I U.9. Completion 01/10/91 03/91 1 40 F n.a. Specialty: A - Agricultural economist E - Rural development specialist B - Financial analyst F - Management specialist C - Agricultural credit specialist G - Forestry specialist D - Financial management specialist -v - PERFORMANCE AUDIT REPORT PHILIPPINES AGRICULTURAL CREDIT PROJECT (LOAN 2570-PH) EVALUATION SUMMARY Introduction policy, and its credit experts now articulate the case for deregulation, 1. The first "agricultural credit discipline and Lull-cost pricing. But project" vas the 14th Bank-funded in terms of the types of banks and agricultural project with a credit farmers who benefitted, the results component in the Philippines. Four of differed from the indicative plan these were "rural credit projects" and presented at appraisal. Participation also primarily credit, but targeted at by the smaller commercial farmers was the "Rural Banks" (RB) and the smaller minimal, and the line of refinance commercial farmers which this network came to be dominated by the KBs and of local, private banks services their traditional larger commercial throughout the islands. The overriding clients. There is evidence that at objective of the new project, prepared least for this group of clients the by the Central Bank of the Philippines project funds were partly substituting (CBP) in 1983/84 and appraised in 1985, for other available finance. The was a reform of policy: to encourage a project offers useful early lessons shift from targeted - and invariably from experience about a shift in poli- subsidized - credit to an open line of cy for agricultural credit operations CBP refinance with market rates of which has subsequently spread through- interest for any agricultural activity. out the Bank. The refinance would be available to all banks and their clients that met the Obiectives eligibility conditions. It implied that, for the first time in a Bank- 3. Apart from the first objective - supported credit project in the Philip- (1) reorienting agricultural credit pines, the much larger network of policy away from targeted subsidies - private "commercial banks" (KB) would four others can be identified. The enter. The smaller commercial farmers SAR emphasized (2) improvements in the were expected to access the fund, but institutional apparatus and procedures their participation was not essential for rural lending, (3) incentives to to project design. the banks for raising local deposits, and (4) encouraging the recovery and 2. This was the first Bank project expansion of agricultural enterprise, anywhere to reorient lending terms for which had been repressed by the tight agricultural credit to the financial liquidity position that had accompa- market-place (para 1.2). The Govern- nied a severe recession in 1983-85. ment accepted the major shift in credit Governmeiat gave considerable priority - vi- also to (5) the support for the balance so that this small amount was eventu- of payments offered by the US$100.0 ally disbursed as well into ALF. million Bank Loan and a parallel grant U.S.AID had allocated US$2.6 million of US$20.0 million from the US Agency of its grant to training and studiess for InternatiLnal Development (U.S. about US$500,000 of this was later AID). When the project was initially canceled. conceived it was intended to provide funds exclusively for medium- and long- 6. The CBP was not the original term lending, a facility which was choice to implement the ALF. Two almost completely missing in both other public development lending agen- formal and informal farm finance in the cies which might have been more natu- Philippines. At appraisal, however, ral homes had suffered from mismanage- almost half of the onlending funds were ment and high arrears in the early redirected to increasing the supply 1980s and were disqualified. One was also of seasonal credit, which had the Bank's earlier favorite implemen- earlier been considered adequate but tor of projects with credit compo- was cut sharply by the recession. nents, the Development Bank of the Philippines (DBP). The CBP project Implementation Experience tranagement unit thus became the focus of institution building efforts by the 4. A special unit was established Bank and U.S.AID. This part of the within the CBP to administer the Agri- project had a poor ending in 1989 when cultural Loan Fund (ALF and ALFU). Government took the decision - sup- Disbursement from the ALF, and from the ported by the Bank - to transfer re- Bank's Loan, were slowed in the first sponsibility for administering ALF to two years by continuing effects of the a third development lending agency, recession. Recovery can be dated by the Land Bank of the Philippines the change in Government following the (LBP). LBP is presently handling both departure of President Marcos in 1986, the reflows from ALF and the new although the new Government's platform availabilities from the follow-on for land reform acted as a further US$150 million Bank Loan, under a disincentive to rural lending. However newly constituted Countryside Loan a surge of ALF lending starting in mid- Fund (CLF). ALF and CLF operate side 1988 accelerated the rate of disburse- by side, although they will be merged ments, and the Bank Loan was fully eventually. The staff of the CBP unit utilized before the end of 1989 and a has turned to other non-project tasks, year before the original closing date. while LBP has established its own The U.S.AID onlending component of implementation team with new personnel US$17.4 million was also fully dis- (paras 2.10-2.11). bursed. The CBP is credited with having done a competent job in adminis- Results tering the ALF within the conditions on bank and borrower eligibility estab- 7. The SAR does not specify ar even lished at appraisal. speculate on the proportions of the ALF to be used by the different banks 5. From the original Loan, US$ or by different type and scale of 500,000 was allocated to finance five farmer. That neutrality is consiotent identified studies and preparation of with the central feature of the ALF a follow-on project. Four of the stud- design, which was to eliminate prefer- ies were completed, but financed by CBP ential targeting. U.S.AID had warned that application of the ALF eligibility toward established KB customers for and interest rate formulas as then maintaining or rehabilitating their drafted might cut out the RA and their enterprisest but that in many cases smaller commercial farmers. The Bank those borrowers would have financed responded with an assurance that that the activities with other funds had %.as not so (paras 1.11 and 2.13). The the ALF refinancing window not exiev- results in terms of actual use are ed. Subsequent work by the same firm otherwise. The KB received 722 of the for OE showed that this was not the disbursements from ALF in the project case for the smaller size enterprise period 1985-89. Thrift banks (TB), a loans provided by the TB (pars 3.14- network of smaller savings and develop- 3.21). ment-oriented banks with a clientele generally more substantial than those 9. The SAR had expected the project of the RB, took 25Z. The RB took 1% to expand the number of branch banks (paras 2.2-2.5). The KB borrowers were operating in rural areas, to bring mostly traditional customers, and an into the formal credit system a larger important share of those funds went not number of rural borrowers, and to to individuals but to large and some- spread the funds over a wider geo- times dominating agro-industrial firms graphic area. None of this happened. in the economy. The SAR showed an The concentration on short term fund- average#sub-loan size of US$10,000 in ing and working capital, coupled with its indicative plan for medium and long the evidence of some substitution of term sub-loans. The actual average was ALF funds for other sources, suggests US$123,000 for all sub-loans. Only that the project objective to promote about 5% of the ALF was used for new the recovery and expansion of the enterprise: the rest was for expansion agricultural economy - implying incre- of existing enterprise or for working mentality - was substantially under- capital. In fact 75% of the funds were shot. The U.S.AID consultant also used for either seasonal credit or for concluded that the ALP had little or short term or permanent working capi- no impact on increasing the level of tal. Even the "long term" finance deposits by rural households, though supporting investment was shorter than that finding has been disputed by CBP anticipated, averaging three to four (para3.10). years, reflecting the participating banks' continuing efforts to stay as 10. The main reason the smaller com- liquid as possible in the uncertain mercial farmers were left out was that economy. Formal Bank field supervision few of the RB met the eligibility stopped in early 1988, before the surge criteria. Ultimately only 13 RB made of lending, and throughout the period use of ALP, out of a total number of seems to have been unconcerned about about 800 RBs that were still operat- the evolving profile of ALF clientele ing in the late 1980s. Nevertheless, (para 2.12). CBP figures show that the category it identifies as "small" sub-loans in- 8. U.S.AID commissioned a consultant cluded 861 of the 2,013 loans made firm in late 1990 to evaluate the during the project, taking 482 of the results of the U.S.AID project - and by funds disbursed. Thus there was some implication of the Bank's Loan. This "small" farm activity in the project, Impact Evaluation determined during its and the ALP credit line to this cate- extensive interview program that not gory represented exactly the type of only was the ALP clientele skewed reformed rural financial system that - viii - the Bank had hoped would replace the the KB and even the TB - are leading previous, targeted, subsidized, and farmers by the standards of their poorly performing operations. Repay- communities. Thu.p the fact that the ment rates across all categories of ALF ALF ultimately served the larger farms lending are above 952. and companies should not of itself attract criticism from the perspective 11. There is not much information of rural equity (para 4.1-4.4). available on the on-farm impact of the ALF loans. However the U.S.Aid Impact 14. Role of the Rural Banks Never- Evaluation and the Bank's Project theless a fundamental part of the Completion Report (PCR) are satisfied reform program was to create the con- that the funds were generally put to ditions for a complete overhaul of the good use (para 3.12-3.13). Also, CBP RB system, setting them up as rural controls and selecti-e end-use surveys financial intermediaries that could suggest that diversion or other misuse efficientlv raise, lend *.d recover of funds was minimal, funds. This was a dominant position in the Bank and U.S.AID9 although Sustainability there were officers in both who were skeptical about the possibility of RB 12. The CBP structure for implement- reform and would not be surprised or ing the project disappeared. But the concerned that their participqxion in LBP is clearly serious about making the ALF was insignificant. The ALF pro- CLF work and performance during 1990 ject was not designed to reconstruct and 1991 was good. Thus the ALF model by itself the RB system: the Bank and seems to have been institutionalized at U.S.AID expected that a separate pro- the level of the refinance, and the ject or public initiative would take philosophy behind it has taken root. shape for that purpose, based on rec- On the policy level, then, the new ommendations of the project-financed approach to rural finance seems to be study program, and in time to bring a widely accepted (one qualification is large number of the RB to ALP eligi- given in para 16 below). The follow-on bility. For several reasons that project provides the funding needed to complementar; effort never gathered continue the program. At the farm speed. The RB arrears position con- level, the borrowers already had prov- tinued to deteriorate, the eligibility en, profitable enterprises and the even of many of those RB that initial- investment and shorter term applica- ly qualitied was withdrawn, many of tions of the ALF funds appear to have the eligible RB refused to use ALF, made sense and helped sustain and and the KB and their scale of borrower expand those businesses. took over. The LBP is promoting yet another program for restructuring the Findings and Issues RB in association with the CLP (para 4.5-4.7). 13. Small Farmers The project was never meant to reach the "real" small 15. ALP Interest Rate Central to the farmer, that vast majority of 5 hectare project design was a formula for set- corn and rice households who are the ting an ALP onlending rate that would target population of the agrarian match market rates and encourage the reform program. The clientele of the banks to seek to raise more deposits. RB - though much smaller in total The "floating" formula was lagged over assets than the typical borrower from nine months. This cushion was provid- - ix - ed to avoid unsettling borrowers, who simultaneously managing another fund, were accustomed to fixed rates, with for many more farmers, that for the rapid and large fluctuations. But the time being remains outside the reform. formula was not appropriate for periods The extent of Government's conversion of sustained, rising inflation. Then, to the new concept of rural finance the built-in lag would open a substan- would appear to have been compromised, tial differential below market rates in though it is argued in the main body favor of the participating borrowers of this report that the compromise may and bankers. That is what happened in this case make sense (paras 4.9- starting in mid-1988 for the last year- 4.129 4.22). and-a-half of the project. It was profitable for both the larger farmers 17. Were the Results of the Proiect and their commercial banks to switch Satisfactory? The audit concludes their traditional line of finance to they were, though that is a very close ALF, sharing the savings. Perhaps half call. It is based largely on the of all project lending was influenced impressive shift in policy, and the by this predictable response by borrow- fact that most public agricultural ers and bankers to adjust their portfo- credit programs are now unsubsidized. lios. This rational behavior, and the Government can also be satisfied that surge of ALF disbursement that fol- the transfer of resources for balance lowed, is a consequence of an interest of payment support took place. But rate formula that overemphasized the one must be concerned also about the cushion (para 4.8). concentration of ALF funds on large farm clientele, to finance not invest- 16. Policy Reform As part of the ments but their requirements for work- agrarian reform program, another fund ing capitai, even though those charac- was put under the control of LBP to teristics of the project outcome are disburse as credit to the "real" small not inconsistent with project objec- farmers, the actual and intended bene- tives and working capital is a legiti- ficiaries of the agrarian reform pro- mate use of public funds (provided gram. LBP determined that most of this they do not simply substitute for would be lent to farmer cooperatives, other sources). Other tests of suc- for onlending to members or for collec- cess are whether the funding mechanism tive activities. This fund is not only that was put into play helped targeted but subsidized. LBP manage- strengthen both the recovery of agri- ment does not accept that its primary cultural enterprise and the capability target group, which includes most of of banks lending in rural areas to the country's farmers, should be made carry out effectively the job of fi- to accept floating, market-oriented nancial intermediation. These imply interest rates, or that they should be in turn substantial incremantality made to pay the full costs of credit above what would have been financed administration. Whether managed by the anyway, an impact on deposit mobiliza- coops or another agency, small farm tion, and some progress on the implic- credit is expensive and part of the it agenda of RB reform. None of these burden is seen to be a legitimate appear to be prominent features of developmental expense. Thus the same project results (para 4.13-4.16). organization that has been selected to manage the CLF, under the disciplined 18. Lessons for Credit Policy The terms that are associated with the new project is the first to test the prac- policy position discussed above, is ticality of the new policy. But four circumstances warn against any whole- profit maximizers and risk averters, sale generalization from this outcome. that they prefer to remain on short First, the economic and political exposure, especially in periods of environments in which the project was inflation and economic uncertainty, implemented between 1985 and 1989 were that they will seek secure loans, that unsettled, and inhospitable to aggres- they never like to make small loans to sive private banking. Second, the RB small farmers# that they are slow to network that the project formulators open rural branches, and that they expected to balance out the large farm cannot be expected to respond enthusi- activity was already in a downhill astically to the development objec- slide from arrears previously contract- tives of Government and the Bank. ed. The project under audit was not Philippine bankers are no different in fortified with the equipment it would these respects than their counterparts have needed to arrest that decline. in other countries. Open lines of Third, the inflation of 1988 ambushed credit that are not targeted will the project's floating interest rate concentrate rather on more secure formula and encouraged the big banks objectives: unless, that is, project and borrowers to grab a profitable design can harness those profit making opportunity. In the new credit policy instincts of commercial bankers better model, a large differential should not than this project did (paras 4.20- have emerged. Fourth, the agrarian 4.24). reform program of the present govern- ment has severely reduced the collater- 20. Thus the policy reorientation can al value of farm land, the basis for be described as a necessary condition most lending under ALF in the project for reform of rural credit, but it is period. It is important now to watch not sufficient to build a better and performance under the follow-on project broader banking system. For that, the and the CLF. RB reform, an interest policy changes must be matched by rate formula that keeps closer to the equally important improvements in the market, and some restrictions on loan institutional apparatus, rural infra- size and working capital are put into structure, and the profitability of the new design. small farmer enterprise (para 4.25). 19. But the clearest lesson of this experience is that the KB and TB are PERFORMANCE AUDIT REPORT PHILIPPINES AGRICULTURAL CREDIT PROJECT (Loan 2570-PH) 1. BACKGROUND A. Preparation 1.1 Up until the early 1980s the Bank's agricultural program in the Philippines had given priority to irrigation and credit. The first Project Brief for the project under audit was issued in 1982, and at that time the Bank had participated in 13 projects with credit components. Four of these were in a series labelled "rural credit projects". They were aimed at a network of Rural Banks (RB) and had been executed (the fourth was then under implementation) by a special project unit in the Rural Bank Department of the Central Bank of the Philippines (CBP). The RB were private, unit banks that catered to the smaller commercial farmers resident in rural areas. The medium and larger scale farmers sought bigger loans and typically used the "commercial banks" (KB), which thus far had not participated in any of the Bank-financed projects. Eight of the latter had been executed by the Development Bank of the Philippines (DBP): these were commodity or area specific projects with credit components. The thirteenth of the Bank-supported projects was then being executed by the Land Bank of the Philippines (LBP) and supported beneficiaries of the Government's agrarian reform program. 1.2 Two characteristics of all of these projects were, first, that they were directed toward one or another group of farmers and/or network of individual banks or bank branches, and, second, that the interest rates on farmer loans were subsidized. These characteristics were common to all of the Government's rural credit programs, which were part of a highly regulated credit system that focused on the needs and potential of the targeted farmer populations rather than the requirements of successful banking. The pattern was familiar in many of the Bank's agricultural credit portfolios in Asia and elsewhere. What distinguished the Philippine portfolio in the early 1980s was that in this case the Bank staff decided to shift its support decisively to another model of rural finance - from a "programmatic" approach to a market determined, "laissez faire" approach. This was to be the first experience in a general shift in the Bank that accelerated. later in the decade toward market-oriented rural credit projects. 1.3 The Bank wanted to narrow and consolidate its future rural credit activities in the Philippines, and in the process turn from targeted, subsidized programs to an open line of unsubsidized wholesale credit available to all retail-level banks and their rural borrowers who met stricter eligibility criteria. One of the reasons behind this posture was the rising level of arrears in the ongoing projects. At this point the Bank's interest in reorientation was shared only in part by Government. In 1981 it had taken some important steps to loosen the regulations on rural banking. But there was no widespread agreement -2- with the Bank that scattered, subsidized, targeting was part of the explanation of the arrears. In fact the Technical Board for Agricultural Credit (TBAC) - the credit research and policy advisory unit attached to CBP - was ambivalent on this issue, with some Board members still supporting the older portfolio and some research staff supporting reform. Nevertheless, the Bank could count on the unequivocal support of the Governor of CBP. 1.4 Preparation of the first Project Brief in 1982 was followed by a Bank mission to study the agricultural credit sector and determine whether DBP was capable of managing a consolidated line of credit. DBP's financial position had been declining in recent years, also due to mounting arrears. The mission ruled out DBP from such a program, and, in its Agricultural Credit Sector Review issued in 1983, set the stage for CBP to prepare and execute the proposed apex refinance facility (large balances in the two ongoing Bank Loans to DBP, as well as the one to LBP, were later to be canceled). Preparation carried on through 1984, with several Bank missions joining the Philippine task force in support. An "Agricultural Loan Fund" (ALF) was proposed and formally approved by CBP's Monetary Board in late 1984. CBP was already managing a much larger, short term rediscount operation for the country's banking systems, also subsidized. One of the key conditions agreed with Government for the new project was that CBP would eliminate, in a relatively short time span, the subsidy from this traditional rediscounting program as well. 1.5 The Bank had sought other donor support, and the United States Agency for International Development (U.S.AID) decided to add a grant of US$20.0 million to the US$100.0 million proposed for the Bank Loan. The U.S.AID project, called the Rural Financial Services Project, would run parallel to the Bank's project. U.S. disbursements were not cross-conditioned on Bank disbursements, however, nor included (at that time) in the Bank's cost estimates. Of the grant, US$2.6 million was allocated to training, studies and other technical assistance. The rest would be placed in ALF. The Americans were expecting that the program would lead to a result that for them was a key component of their strategy for rural credit in the Philippines, which was the transformation of the RB from what they described as cashiers for Government's subsidized lines of credit into proper financial intermediaries. But they concurred with the Bank that rehabilitation should precede disbursement to the weaker RBs. 1.6 The Agricultural Credit Project, with the already approved ALF at its core, was appraised in late 1984, and agreed and brought rapidly to effectiveness in 1985. The timing was not propitious. The economy had moved into a severe recession in 1983, accompanied by a particularly deep crisis of confidence in the banking sector and a significant contraction of the financial system starting the next year. This had forced all banks to adopt a conservative position with respect to long-term lending and to maintain portfolios in as short maturities as possible. Ninety-eight percent of all formal lending to agriculture was short term, though some of that was routinely rolled over and served the purpose of longer term maturities. The liquidity crisis accompanying the recession, however, led also to a decline in short term finance itself. Whereas the Bank had originally intended to restrict its contribution to the pioneer medium and -v - PERFORMANCE AUDIT REPORT PHILIPPINES AGRICULTURAL CREDIT PROJECT (LOAN 2570-PH) EVALUATION SUMMARY Introduction policy, and its credit expert. now articulate the case for deregulation, 1. The first "agricultural credit discipline and full-cost pricing. But project" was the 14th Bank-funded in terms of the types of bank. and agricultural project with a credit farmers vho benefitted, the results component in the Philippines. Four of differed from the indicative plan these were "rural credit projects" and presented at appraisal. Participation also primarily credit, but targeted at by the smaller commercial farmers was the "Rural Banks" (RB) and the smaller minimal, and the line of refinance commercial farmers which this network came to be dominated by the KB. and of local, private banks services their traditional larger commercial throughout the islands. The overriding clients. There is evidence that at objective of the new project, prepared least for this group of clients the by the Central Bank of the Philippines project funds were partly substituting (CBP) in 1983/84 and appraised in 1985, for other available finance. The was a reform of policy: to encourage a project offers useful early lessons shift from targeted - and invariably from experience about a shift in poli- subsidized - credit to an open line of cy for agricultural credit operations CBP refinance with market rates of which has subsequently spread through- interest for any agricultural activity. out the Bank. The refinance would be available to all banks and their clients that met the Objectives eligibility conditions. It implied that, for the first time in a Bank- 3. Apart from the first objective - supported credit project in the Philip- (1) reorienting agricultural credit pines, the much larger network of policy away from targeted subsidies - private "commercial banks" (KB) would four others can be identified. The enter. The smaller commercial farmers SAR emphasized (2) improvements in the were expected to access the fund, but institutional apparatus and procedures their participation was not essential for rural lending, (3) incentives to to project design. the banks for raising local deposits, and (4) encouraging the recovery and 2. This was the first Bank project expansion of agricultural enterprise, anywhere to reorient lending terms for which had been repressed by the tight agricultural credit to the financial liquidity position that had accompa- market-place (para 1.2). The Govern- nied a severe recession in 1983-85. ment accepted the major shift in credit Government gave considerable priority - vi - also to (5) the support for the balance so that this small amount we eventu- of payments offered by the US$100.0 ally disbursed as weil into ALF. million Bank Loan and a parallel grant U.S.AID had allocated US$2.6 million of US$20.0 million from the US Agency of its grant to training and studiess for International Development (U.S. about US$500,000 of this was later AID). When the project was initially canceled. conceived it was intended to provide funds exclusively for medium- and long- 6. The CBP was not the original term lending, a facility which was choice to implement the ALF. TWO almost completely missing in both other public development lending agen- formal and informal farm finance in the cies which might have been more natu- Philippines. At appraisal, however, ral homes had suffered from mismanage- almost half of the onlending funds were ment and high arrears in the early redirected to increasing the supply 1980a and were disqualified. One was also of seasonal credit, which had the Bank's earlier favorite implemen- earlier been considered adequate but tor of projects with credit compo- was cut sharply by the recession. nents# the Development Bank of the Philippines (DBP). The CSP project Implementation Experience vanagement unit thus became the focus of institution building efforts by the 4. A special unit was established Bank and U.S.AID. This part of the within the CBP to administer the Agri- project had a poor ending in 1989 when cultural Loan Fund (ALF and ALFU). Government took the decision - sup- Disbursement from the ALF, and from the ported by the Bank - to transfer re- Bank's Loan, were slowed in the first sponsibility for administering ALF to two years by continuing effects of the a third development lending agency, recession. Recovery can be dated by the Land Bank of the Philippines the change in Government following the (LBP). LBP is presently handling both departure of President Marcos in 1986, the reflows from ALF and the new although the new Government's platform availabilities from the follow-on for land reform acted as a further US$150 million Bank Loan, under a disincentive to rural lending. However newly constituted Countryside Loan a surge of ALF lending starting in mid- Fund (CLF). ALF and CLF operate side 1988 accelerated the rate of disburse- by side, although they will be merged ments, and the Bank Loan was fully eventually. The staff of the CSP unit utilized before the end of 1989 and a has turned to other non-project tasks, year before the original closing date. while LBP has established its own The U.S.AID onlending component of implementation team with new personnel US$17.4 million was also fully dis- (pares 2.10-2.11). bursed. The CBP is credited with having done a competent job in adminis- Results tering the ALF within the conditions on bank and borrower eligibility estab- 7. The SAR does not specify or even lished at appraisal. speculate on the proportions of the ALF to be used by the different banks 5. From the original Loan, US$ or by different type and scale of 500,000 was allocated to finance five farmer. That neutrality is consistent identified studies and preparation of with the central feature of the ALP a follow-on project. Four of the stud- design, which was to eliminate prefer- ies were completed, but financed by CS? ential targeting. U.S.AID had warned - vii - that application of the ALF eligibility toward established KB customers for and interest rate formulas as then maintaining or rehabilitating their drafted might cut out the RBs and their enterprises, but that in many cases smaller commercial farmers. The Bank those borrowers would have financed responded with an assurance that that the activities with other funds had was not so (paras 1.11 and 2.13). The the ALF refinancing window not exist- results in terms of actual use are ed. Subsequent work by the same firm otherwise. The KB received 72% of the for OED showed that this was not the disbursements from ALF in the project case for the smaller size enterprise period 1985-89. Thrift banks (TB), a loans provided by the TB (paras 3.14- network of smaller savings and develop- 3.21). ment-oriented banks with a clientele generally more substantial than those 9. The SAR had expected the project of the RB, took 25%. The RB took 12 to expand the number of branch banks (paras 2.2-2.5). The KB borrowers were operating in rural areas, to bring mostly traditional customers, and an into the formal credit system a larger important share of those funds went not number of rural borrowers, and to to individuals but to large and some- spread the funds over a wider geo- times dominating agro-industrial firms graphic area. None of this happened. in the economy. The SAR showed an The concentration on short term fund- average sub-loan size of US$10,000 in ing and working capital, coupled with its indicative plan for medium and long the evidence of some substitution of term sub-loans. The actual average was ALF funds for other sources, suggests US$123,000 for all sub-loans. Only that the project objective to promote about 5% of the ALF was used for new the recovery and expansion of the enterprise: the rest was for expansion agricultural economy - implying incre- of existing enterprise or for working mentality - was substantially under- capital. In fact 75% of the funds were shot. The U.S.AID consultant also used for either seasonal credit or for concluded that the ALF had little or short term or permanent working capi- no impact on increasing the level of tal. Even the "long term" finance deposits by rural households, though supporting investment was shorter than that finding has been disputed by CBP anticipated, averaging three to four (pare 3.10). years, reflecting the participating banks' continuing efforts to stay as 10. The main reason the smaller com- liquid as possible in the uncertain mercial farmers were left out was that economy. Formal Bank field supervision few of the RB met the eligibility stopped in early 1988, before the surge criteria. Ultimately only 13 RB made of lending, and throughout the period use of ALP, out of a total number of seems to have been unconcerned about about 800 RBs that were still operat- the evolving profile of ALF clientele ing in the late 1980s. Nevertheless, (para 2.12). CBP figures show that the category it identifies as "small" sub-loans in- 8. U.S.AID commissioned a consultant cluded 862 of the 2,013 loans made firm in late 1990 to evaluate the during the project, taking 481 of the results of the U.S.AID project - and by funds disbursed. Thus there was some implication of the Bank's Loan. This "small" farm activity in the project, Impact Evaluation determined during its and the ALP credit line to this cate- extensive interview program that not gory represented exactly the type of only was the ALP clientele skewed reformed rural financial system that -viii - the Bank had hoped would replace the the KB and even the TB - are leading previous, targeted, subsidized, and farmers by the standards of their poorly performing operations. Repay- communities. Thus, the fact that the ment rates across all categories of ALF ALF ultimately served the larger farms lending are above 95%. and companies should not of itself attract criticism from the perspective 11. There is not much information of rural equity (para 4.1-4.4). available on the on-farm impact of the ALF loans. However the U.S.Aid Impact 14. Role of the Rural Banks Never- Evaluation and the Bank's Project theless a fundamental part of the Completion Report (PCR) are satisfied reform program was to create the con- that the funds were generally put to ditions for a complete overhaul of the good use (para 3.12-3.13). Also, CBP RB system, setting them up as rural controls and selective end-use surveys financial intermediaries that could suggest that diversion or other misuse efficiently raise, lend and recover of funds was minimal. funds. This was a dominant position in the Bank and U.S.AID, although Sustainability there were officers in both who were skeptical about the possibility of RB 12. The CBP structure for implement- reform and would not be surprised or ing the project disappeared. But the concerned that their participation in LBP is clearly serious about making the ALE was insignificant. The ALE pro- CLF work and performance during 1990 ject was not designed to reconstruct and 1991 was good. Thus the ALF model by itself the RB system: the Bank and seems to have been institutionalized at U.S.AID expected that a separate pro- the level of the refinance, and the ject or public initiative would take philosophy behind it has taken root. shape for that purpose, based on rec- On the policy level, then, the new oendations of the project-financed approach to rural finance seems to be study program, and in time to bring a widely accepted (one qualification is large number of the RB to ALE eligi- given in para 16 below). The follow-on bility. For several reasons that project provides the funding needed to complementary effort never gathered continue the program. At the farm speed. The RB arrears position con- level, the borrowers already had prov- tinued to deteriorate, the eligibility en, profitable enterprises and the even of many of those RB that initial- investment and shorter term applica- iy qualified was withdrawn, many of tions of the ALF funds appear to have the eligible RB refused to use ALF, made sense and helped sustain and and the KB and their scale of borrower expand those businesses. took over. The LBP is promoting yet another program for restructuring the Findings and Issues RB in association with the CLE (pare 4.5-4.7). 13. Small Farmers The project was never meant to reach the "real" small 15. ALE Interest Rate Central to the farmer, that vast majority of 5 hectare project design was a formula for set- corn and rice households who are the ting an ALE onlending rate that would target population of the agrarian match market rates and encourage the reform program. The clientele of the banks to seek to raise more deposits. RB - though much smaller in total The "floating" formula was lagged over assets than the typical borrower from nine months. This cushion was provid- - ix - ed to avoid unsettling borrowers, who simultaneously managing another fund, were accustomed to fixed rates, with for many more farmers, that for the rapid and large fluctuations. But the time being remains outside the reform. formula was not appropriate for periods The extent of Government's conversion of sustained, rising inflation. Then, to the new concept of rural finance the built-in lag would open a substan- would appear to have been compromised, tial differential below market rates in though it is argued in the main body favor of the participating borrowers of this report that the compromise may and bankers. That is what happened in thi. case make sense (paras 4.9- starting in mid-1988 for the last year- 4.12, 4.22). and-a-half of the project. It was profitable for both the larger farmers 17. Wer. the Result. of the Proiect and their commercial banks to switch Satisfactorl? The audit concludes their traditional line of finance to they were, though that is a very close ALF, sharing the savings. Perhaps half call. It is based largely on the of all project lending was influenced impressive shift in policy, and the by this predictable response by borrow- fact that most public agricultural ers and bankers to adjust their portfo- credit programs are now unsubsidized. lios. This rational behavior, and the Government can also be satisfied that surge of ALF disbursement that fol- the transfer of resources for balance lowed, is a consequence of an interest of payment support took place. But rate formula that overemphasized the one must be concerned also about the cushion (para 4.8). concentration of ALF funds on large farm clientele, to finance not invest- 16. Policy Reform As part of the ments but their requirements for work- agrarian reform program, another fund ing capitai, even though those charac- was put under the control of LBP to teristics of the project outcome are disburse as credit to the "real" small not inconsistent with project objec- farmers, the actual and intended bene- tives and working capital is a legiti- ficiaries of the agrarian reform pro- mate use of public funds (provided gram. LBP determined that most of this they do not simply substitute for would be lent to farmer cooperatives, other sources). Other tests of suc- for onlending to members or for collec- cess are whether the funding mechanism tive activities. This fund is not only that was put into play helped targeted but subsidized. LBP manage- strengthen both the recovery of agri- ment does not accept that its primary cultural enterprise and the capability target group, which includes most of of banks lending in rural areas to the country's farmers, should be made carry out effectively the job of fi- to accept floating, market-oriented nancial intermediation. These imply interest rates, or that they should be in turn substantial incrementality made to pay the full costs of credit above what would have been financed administration. Whether managed by the anyway, an impact on deposit mobiliza- coops or another agency, small farm tion, and some progress on the implic- credit is expensive and part of the it agenda of RB reform. None of these burden is seen to be a legitimate appear to be prominent features of developmental expense. Thus the same project results (para 4.13-4.16). organization that has been selected to manage the CLF, under the disciplined 18. Lessons for Credit Policy The terms that are associated with the new project is the first to test th prac- policy position discussed above, is ticality of the new policy. But four circumstances warn against any whole- profit maximizers and risk averters sale generalization from this outcome. that they prefer to remain on short First, the economic and political exposure, especially in periods of environments in which the project was inflation and economic uncertainty, implemented between 1985 and 1989 were that they will seek secure loans, that unsettled, and inhospitable to aggres- they never like to make email loans to sive private banking. Second, the RB small farmers, that they are slow to network that the project formulators open rural branches, and that they expected to balance out the large farm cannot be expected to respond enthusi- activity was already in a downhill astically to the development objec- slide from arrears previously contract- tives of Goverment and the Bank. ed. The project under audit was not Philippine bankers ar* no different in fortified with the equipment it would these respects than their counterparts have needed to arrest that decline. in other countries. Open lines of Third, the inflation of 1988 ambushed credit that are not targeted will the project's floating interest rate concentrate rather on more secure formula and encouraged the big banks objectivest unless, that is, project and borrowers to grab a profitable design can harness those profit making opportunity. In the new credit policy instincts of commercial bankers better model, a large differential should not than this project did (paras 4.20- have emerged. Fourth, the agrarian 4.24). reform program of the present govern- ment has severely reduced the collater- 20. Thus the policy reorientation can al value of farm land, the basis for be described as a necessarl condition most lending under ALF in the project for reform of rural credit, but it is period. It is important now to watch not sufficient to build a better and performance under the follow-on project broader banking system. For that, the and the CLF. RB reform, an interest policy changes must be matched by rate formula that keeps closer to the equally important improvements in the market, and some restrictions on loan institutional apparatus, rural infra- size and working capital are put into structure, and the profitability of the new design. small farmer enterprise (pare 4.25). 19. But the clearest lesson of this experience is that the KB and TB are PERFORMANCE AUDIT REPORT PHILIPPINES AGRICULTURAL CREDIT PROJECT (Loan 2570-PH) I. BACKGROUND A. Prepr1tion 1.1 Up until the early 1980s the Bank's agricultural program in the Philippines had given priority to irrigation and credit. The first Project Brief for the project under audit was issued in 1982, and at that time the Bank had participated in 13 projects with credit components. Four of these were in a series labelled "rural credit projects". They were aimed at a network of Rural Banks (RB) and had been executed (the fourth was then under implementation) by a special project unit in the Rural Bank Department of the Central Bank of the Philippines (CBP). The RB were private, unit banks that catered to the smaller commercial farmers resident in rural areas. The medium and larger scale farmers sought bigger loans and typically used the "commercial banks" (KB), which thus far had not participated in any of the Bank-financed projects. Eight of the latter had been executed by the Development Bank of the Philippines (DBP): these were commodity or area specific projects with credit components. The thirteenth of the Bank-supported projects was then being executed by the Land Bank of the Philippines (LBP) and supported beneficiaries of the Government's agrarian reform program. 1.2 Two characteristics of all of these projects were, first, that they were directed toward one or another group of farmers and/or network of individual banks or bank branches, and, second, that the interest rates on farmer loans were subsidized. These characteristics were common to all of the Government's rural credit programs, which were part of a highly regulated credit system that focused on the needs and potential of the targeted farmer populations rather than the requirements of successful banking. The pattern was familiar in many of the Bank's agricultural credit portfolios in Asia and elsewhere. What distinguished the Philippine portfolio in the early 1980s was that in this case the Bank staff decided to shift its support decisively to another model of rural finance - from a "programmatic" approach to a market determined, "laissez faire" approach. This was to be the first experience in a general shift in the Bank that accelerated later in the decade toward market-oriented rural credit projects. 1.3 The Bank wanted to narrow and consolidate its future rural credit activities in the Philippines, and in the process turn from targeted, subsidized programs to an open line of unsubsidized wholesale credit available to all retail-level banks and their rural borrowers who met stricter eligibility criteria. One of the reasons behind this posture was the rising level of arrears in the ongoing projects. At this point the Bank's interest in reorientation was shared only in part by Government. In 1981 it had taken some important steps to loosen the regulations on rural banking. But there was no widespread agreement -2- with the Bank that scattered, subsidized, targeting was part of the explanation of the arrears. In fact the Technical Board for Agricultural Credit (TBAC) - the credit research and policy advisory unit attached to CBP - was ambivalent on this issue, with some Board members still supporting the older portfolio and some research staff supporting reform. Nevertheless, the Bank could count on the unequivocal support of the Governor of CBP. 1.4 Preparation of the first Project Brief in 1982 was followed by a Bank mission to study the agricultural credit sector and determine whether DBP was capable of managing a consolidated line of credit. DBP's financial position had been declining in recent years, also due to mounting arrears. The mission ruled out DBP from such a program, and, in its Agricultural Credit Sector Review issued in 1983, set the stage for CBP to prepare and execute the proposed apex refinance facility (large balances in the two ongoing Bank Loans to DBP, as well as the one to LBP, were later to be canceled). Preparation carried on through 1984, with several Bank missions joining the Philippine task force in support. An "Agricultural Loan Fund" (ALF) was proposed and formally approved by CBP's Monetary Board in late 1984. CBP was already managing a much larger, short term rediscount operation for the country's banking systems, also subsidized. One of the key conditions agreed with Government for the new project was that CBP would eliminate, in a relatively short time span, the subsidy from this traditional rediscounting program as well. 1.5 The Bank had sought other donor support, and the United States Agency for International Development (U.S.AID) decided to add a grant of US$20.0 million to the US$100.0 million proposed for the Bank Loan. The U.S.AID project, called the Rural Financial Services Project, would run parallel to the Bank's project. U.S. disbursements were not cross-conditioned on Bank disbursements, however, nor included (at that time) in the Bank's cost estimates. Of the grant, US$2.6 million was allocated to training, studies and other technical assistance. The rest would be placed in ALF. The Americans were expecting that the program would lead to a result that for them was a key component of their strategy for rural credit in the Philippines, which was the transformation of the RB from what they described as cashiers for Government's subsidized lines of credit into proper financial intermediaries. But they concirred with the Bank that rehabilitation should precede disbursement to the weaker RBs. 1.6 The Agricultural Credit Project, with the already approved ALF at its core, was appraised in late 1984, and agreed and brought rapidly to effectiveness in 1985. The timing was not propitious. The economy had moved into a severe recession in 1983, accompanied by a particularly deep crisis of confidence in the banking sector and a significant contraction of the financial system starting the next year. This had forced all banks to adopt a conservative position with respect to long-term lending and to maintain portfolios in as short maturities as possible. Ninety-eight percent of all formal lending to agriculture was short term, though some of that was routinely rolled over and served the purpose of longer term maturities. The liquidity crisis accompanying the recession, however, led also to a decline in short term finance itself. Whereas the Bank had originally intended to restrict its contribution to the pioneer medium and -3- long term facility of ALF, by the time the project was appraised and approved almost half of the Bank funds were allocated to the ALF short term facility. 1.7 The Bank had to consider at appraisal whether the drop in demand for medium and long term rural credit compromised the justification for the Loan. The decision was that it did not, despite an IMP Standby Agreement and a Structural Adjustment Program initiated in late 1984 that worked at cross purposes with the project goal of stimulating the expansion of credit. Recovery of confidence in the banking sector was slow and spaced over the first three years of project execution. This, plus the growing availability of short term private funds at very low interest rates, and continuing aversion of the KB to longer term maturities, resulted in demand for ALF lagging behind appraisal forecasts. 1.8 Thus the hoped-for reform of the bases of rural finance had to contend with an inhospitable macro-economic climate that lingered much longer than had been anticipated. Later, in 1987, with recovery in full swing, the project was to be rocked by another unanticipated problem. The Government that replaced the Marcos regime in 1986 declared its commitment to accelerating the process of agrarian reform. The legislative measures that accompanied this policy (the Comprehensive Agrarian Reform Law - CARL) called for a ceiling of 5 hectares on land holdings, to be implemented over time but supposedly to have immediate effect with respect to any new land transactions. The land reform program (CARP), accompanied by the establishment of a large fund (ARF) to finance farmers and land purchase in the agrarian reform sector, started slowly, and has been delayed further by uncertainties related to the outcome of elections in May 1992. Nevertheless its impact on the bankers' assessment of the collateral value of land holdings offered in security for ALF and all other agricultural loans has been severe. In short, the innovative project design faced an uphill battle against external factors that could have been predicted only with superb foresight. B. Objectives 1.9 With respect to the Bank's position there is no doubt that the primary objective was to consolidate its future lending for agricultural credit within a program that focused on strengthening the system of rural finance - including raising as well as lending and recovering funds. No preference was expressed for the use of the sub-loans, provided certain general conditions on eligibility of the banks were respected. The uses of the fund would be determined by the operating norms and preferences of the banks themselves. Government agreed with this reform, though its conversion at that time to the new strategy was less intense, or at least less unanimous, than in the Bank. An overriding objective for the Government, which may have been almost as important as the credit reform, was to access the foreign exchange offered by a relatively fast disbursing line of rural refinance. The urgency of balance of payments support was tied to the recession and the IMF conditionality. The then Prime Minister told the audit mission that the project may not have been accepted by Government were it not for the added benefit of balance of payment relief. 1.10 Linked to the objective of policy reform was the expectation that the project would assist with, first, the restructuring and strengthening of all banks serving rural areas, as well as, second, the creation within CBP of a capability for running a refinance program that offered support and technical assistance to those participating banks that wanted it and, especially, to those non-participating banks that wanted to achieve eligibility. The attention here was not on the KB, which generally could help themselves, but on the RB and a third network of banks operating in rural areas called "thrift banks" (TB). These were savings banks, savings and loan associations and private development banks that occupied a niche between the RB and the KB. They competed for the better clients of the RB, and the smaller clients of the KB. Some TB were very big in assets and geographic spread, and resembled (and in a some cases were tied to) the KB. The relative shares of these private networks, and of the government banks, in agricultural lending is shown in annex 1. 1.11 The project was not intended to service the great majority of the country's farms, which were basically subsistence producers of rice and maize, averaged about 5 hectares in size, and included actual and intended beneficiaries of the old and new agrarian reform activities. But a reading of the Bank's files shows that the type of smaller commercial farmer who dominated most of the Bank's previous RB and DBP projects was expected at least to have access to the ALF. In fact, the indicative lending program presented in the Staff Appraisal Report (SAR) has a distinct small farm bias. On the seasonal credit side, a little over half (52%) of the funds were expected to be disbursed for rice and corn, suggesting smaller scale farms. On the medium and long term credit side, about 10,000 sub-loans were anticipated, with an average sub-project cost of US$10,000, again suggesting smaller scale farms. At one point during the appraisal process U.S.AID questioned whether all the smaller farmers and the RB themselves were not likely to be cut out by the stiff eligibility criteria. The Bank's response was that, although no targeting was intended, it fully expected that these communities would qualify. 1.12 Whether the participation of these smaller commercial farmers was an implicit objective of the project is a matter of importance, because in the end they were left out almost altogether. It seems clear that the appraisal mission and Bank headquarters staff assumed that many would participate, but at the same time did not consider that that participation was essential. The overriding objective was to put the new policy orientation in place. If the participating banks did not want to use the ALF facility to finance the smaller farmers, that was their prerogative and to try to interfere would override the primary policy goal. Discussions during the audit with members of the appraisal team suggests that they would have been surprised if more that between 5-15% of the ALF funds wotld be taken up by the RB for their smaller scale farmer clientele. While those low estimates may have been influenced by hindsight - the SAR referred to 125 RB approved for participation - the order of magnitude is probably correct. 1.13 But it also clear that project designers in the Bank, U.S.AID and the Government believed the project would accelerate the process whereby many more of the RB could be rehabilitated and brought up to the standards of ALF eligibility within the near future. Thus, although the RB weaknesses were seen -5- to preclude any massive participation of RB clientele, the rehabilitation program, if successful, might bring in a larger number of these clients before the end of the project period. In fact, some members of the Bank and U.S.AID, as well as the Governor of CBP, hoped that another project would be approved soon after the first that would be dedicated to this institution-building (IB) objective, targeting the weaker RB and TB. The importance of the IB effort was highlighted in the 1983 Agricultural Credit Sector Review. The Agriculture Credit Project under audit was not equipped with the instruments to effect such an IB program, though studies were to be financed under the project to determine the best methods for strengthening the rural banking networks, as well as to indicate the best ways to help the subsistence farmers who were left out of all of these credit programs. As mentioned above, CBP was expected to help the weaker banks to achieve eligibility and thus bring the smaller commercial farmers within the orbit of ALF. The leader of the appraisal mission was a strong supporter of the RB, and saw RB reform as an integral part of the new-style credit program of which this project formed the initial phase' 1.14 An additional objective prominent in the SAR was to increase the access of banks in rural areas to the savings of rural households. Deposit mobilization was another integral component of the reform program. On this there is no dispute. By setting the ALF pass-on rate at market levels, the participating banks would be discouraged from seeking ALF support as a substitute for their ordinary sources. The objactive was aimed at both the RB, which had made little effort previously to raise household savings, and the KB and TB, which in their majority depended on deposits but were likely to simply switch if the ALF rates were below market. For the RB, the project would permit them to shake off their dependence on government programmatic finance and become true financial intermediaries. 1.15 One final objective shared by both the Bank and Gov,rnment was that ALF should help finance the recovery and expansion of the agricultural sector. The production objective was not offered to justify the project. Xut it is clear in the SAR and other project descriptions that policy changes, IB and increased production vent hand-in-hand. The SAR says: "this project and the institutional and policy reforms supported by it are expected to play a maj r role in the economic recovery of the country". That support was to be coi :trywide. The successful implementation of the ALF was intended - and this language is also taken from the SAR - to facilitate a "wider" participation of financial intermediaries in agricultural lending, and "greater" geographic as well as farmer coverage. C. Proiect Design 1.16 Project costs were estimated at US$183.0 million, of which US$182.5 million were the costs of the sub-projects to be financed by ALF and US$500,000 11 Another Bank staffer involved with this project throughout its cycle says that the small farmers and the RB were never considered important to project design or implementation. OED disputes this !aterpretation, although opinions obviously varied during preparation. See Annex 4, paras 3,5 and 8-12. -6- was allocated to studies, training and other technical assistance (TA). The Bank Loan of US$100.0 million would cover the small TA component and the rest would be disbursed into ALF. Part of the CBP unit that had executed the series of four Rural Credit projects was assigned to the new project in a unit (ALFU) created within the Department of Loans and Credit (DLC). Staff in the Supervision and Examination Sections (SES I, II, and III) of another department, the permanent CBP offices that were responsible for CBP accreditation and examination of the RB, TB and KB respectively, would participate in the field work. A policy co-mmittee was established under the Deputy Governors of CBP - the ALF Policy Advisory Group. 1.17 The funds would be disbursed over a four year period. The short term component of the Loan would be disbursed into ALF in the first two years, and thereafter rolled over, while the medium and long term funds would be disbursed evenly over the whole project period. CBP would have to demonstrate that the short term finance was additional to the funds it was providing under its normal rediscounting program. There was no such conditioning on the medium and long term finance, since agricultural credit was rarely available for these terms. The ALF was a refinance operation. CBP would deposit funds ("subsidiary loans") using a mechanism called the Special Time Deposit (STD), to cover each "sub-loan" of the Participating Financial Institution (PFI) at the time the approved sub- loan was drawn by the farmer-borrower. CBP's subsidiary loan would finance up to 80% of the sub-loan. The PFI (RB, TB, KB or one of a few other governmental or quasi-banks that established eligibility) could draw against its sub-loans up to a preestablished total for each PFI, as determined in the process of accreditation. 1.18 The formula for fixing the interest rate on ALF loans was crucial to the objective of establishing a market-oriented credit wholesale operation. As mentioned, the rate had to be high enough so that the PFI would not be induced to abandon their customary sources of funds - presumably deposits - in favor of ALF. It was agreed that ALF would charge the higher of two bases, a break-even cost to CBP and a weighted average cost of deposits to PFI (including the cost of mobilizing deposits). The formula included averaging those bases over the previous three quarters to cushion the effect of any rapid movement. The rate would be reviewed semi-annually (in fact it was reviewed quarterly) and the Bank and U.S.AID would have to concur. An element of the foreign exchange risk was included in the break-even cost. PFI would be free to determine the spread on sub-loans above the ALF rate. At the time of appraisal, SES determined that 175 PFIs (15 KB, 35 TB and 125 RB) already met eligibility criteria, and others were expected to enter later. 1.19 The SAR did not provide indicative numbers for the proportion of sub- loans to be made by the different types of banks, or to be taken by different size of farm. That was in keeping with the principle of avoiding targeting. However the SAR did provide indicative tables showing a likely distribution of ALF sub-loans by commodity. As mentioned above, with respect to seasonal credit, the SAR forecast that just over half (52%) would be used to finance rice and maize. The rest was attributed mostly to sugar (33%). With respect to medium and long term credit, the SAR indicated that livestock (mostly poultry and -7- piggery), plantation crops, farm machinery (tillers and tractors) and agro- processing would be the popular lines. There was a controversy within the Bank over the extent to which agroindustrial projects should be i-cluded, since the Bank had a parallel line of industrial finance loans. It was agreed that they were eligible if tied to agricultural production. U.S.AID had some commodity restrictions on its loanable funds, but given the fungibility of the two donor packages these were unimportant. U.S.AID also originally put an upper limit of US$5,000 on the sub-loan size it would support, but this was later removed when experience showed it to be impractical. Altogether the Bank's project was expected to finance the 10,000 sub-projects for medium and long terms, and an unspecified number for short terms. 1.20 The TA component included finance for five studies: (1) strategies for expansion of banking services in rural areas, (2) alternative measures to support low income groups with inadequate access to institutional credit, (3) crop insurance and guarantee schemes, (4) review of CBP's rediscount policy and arrears problem, and (5) preparation of technical guidelines for lending along with a revision of CBP's policy manual for term lending. In addition the funds would support preparation of a follow-on project. The training component was to be executed by CBP's own training facility, the Central Bank Institute (CBI). U.S.AID came in with a larger allocation for TA, especially for training and more studies. 1.21 As part of the policy reforms, but outside the ALF program, Government also agreed to phase-out subsidies on other agricultural credit programs financed from CBP's own funds - including the rediscount operation - and to eliminate regulations on the spread on those sub-loans. These were some of the catalytic effects on financial reform expected from the project. II. IMPLEMENTATION EXPERIENCE A. Proiect Costs and Loan Disbursements 2.1 Actual project costs are estimated in the PCR at US$217.4 million, 19% above the SAR forecast. The difference includes the U.S.AID grant of US$20.0 million, which the PCR chose to include in project costs, plus larger contributions by the participating farmers. The Loan of US$100.0 million was fully disbursed, all of it into the ALF. The project studies were financed by CBP, and the training and other TA by the U.S.AID grant. Despite slow rates of ALF and Loan disbursements in the early years, a surge of lending started in mid- 1988 that overtook the original schedule and allowed the Loan to be fully disbursed two months ahead of the SAR disbursement profile and one year and two months before the Closing Date. -8- B. Onlending Component 2.2 Continuing surveillance by SES of bank financial performance and reasses&ment of eligibility led to the erosion of the list of accredited PFI. By 1986 the list of 175 eligible banks had dropped to 142, and by the last year of the project was down to 89. Not all of those eligible actually applied for refinance from ALP, though they may have continued to access CBP's normal rediscount program. The largest loss was in the ranks of the RB, where an apparently irreversible trend toward insolvency, due to bad and doubtful debt, precluded this favored network from participation. By project completion 46 RB were still eligible and only 25 were listed as having accessed ALF. Of those, 13 were active participants. The other 12 had their 1985 short term borrowings from CBP's rediscount window switched by CBP, after the fact, to project accounts, to satisfy CBP's obligation to co-finance. In addition to the 13 Rural Banks, 20 commercial banks, 13 thrift banks and 3 other institutions took ALF refinance, for a total of 49 participating PFI. Those figures can be compared with the numbers of banks existing at the time the SAR was prepared: 947 RB, 32 KB and 135 TB. Although much smaller than the number of participants anticipated, these PFI, especially the KB, consumed all the project funds available to ALF. 2.3 Data showing the breakdown of sub-loans by term of loan, by type of PFI, by loan purpose, and by subsector of investment are provided in the PCR (see annex 2 to this audit) and in a consultant "Impact Evaluation" study financed by U.S.AID, all based on CBP accounts and its own PCR. With respect to the term of loan, for the five year calendar period 1985-89, 51% of the value of total ALF disbursements was made against seasonal and other short term sub-loans, and the other 49% against medium term maturities (les than 5 years). Less than 1% was disbursed against long term maturities, underlining the PFI's continuing concern to lend as short as possible (the SAR had anticipated most of the "medium and long" term lending would be between 5 and 10 years). With respect to the type of PFI, the KB came to dominate the portfolio, ultimately taking 72% of the total with particularly heavy claims in the last two years. The TB took 25%, and much of that was taken by several thrift banks of quasi KB character (TB participation also increased steadily throughout the project period). The RB took 12. With respect to what CBP calls the purpose of the sub-loan, 21% was allocated exclusively to investment in fixed assets. Another 9% was for fixed assets mixed with permanent working capital. The rest, 70%, was lent for short or permanent working capital (41%) and short term production credit ("seasonal" and other - 292). CBP data shows that only 5% of all those "purposes" were associated with the establishment of new activities; the rest was for continuation or expansion of existing activities. 2.4 Finally, with respect to subsector allocation of ALF funds, table 1 shows the breakdown for the five-and-a-half year period 1985-mid/1990. The table compares actuals with the SAR indicative forecasts, and lumps together short and long term purposes. The dominating position of the category for agro-processing and post harvest facilities in the final project accounts is clear. Most of this was KB lending for working capital. The condition that it be linked directly to agricultural production was interpreted to include purchase of crops and - 9 - livestock from farmers. Even though short term sub-loans took 51% of project loan funds (para 2.3), "seasonal" credit took only 212 and virtually all of that was for estate sugar. Rice and maize took less than 0.5%. By contrast, the SAR had allocated 44% of the Loan to seasonal sub-loans, and over half of that to rice and maize. Another major difference between the indicative estimates at appraisal and the actual allocations of ALF finance was the disappearance of the farm machinery category. The sugar estates simply did not want to invest in new tractors, and the RB that participated did not want to lend long term credit for 7 years to finance the tillers. Paddy farmers who used tillers were too small to apply to the other banking networks. Table 1: ALF Lending Patterns by Sub-sector SAR Actual No. of No. of Projects Value Projects Value Livestock 2,090 16 559 21 Fisheries 410 5 403 15 Agro-processing and 700 6 372 36 Post Harvest Facilities Farm Mechanization 5,370 9 2 0 Plantation Crops 4,6601' 18 120 6 Seasonal Crops (n.a.) 46 (822) 21 13,230 100 2,278 100 Assumes 10 hectares per project. Sources: SAR - Pages 77-79. Actual - U.S.AID; Impact Evaluation; Main Report-Table 21, March 1991. 2.5 With respect to the geographic coverage of ALF, the expectation that the fund would lead to a "wider" pattern of lending was not met. This is a judgement call, since the SAR did not say "wider" than what. Central and South Luzon (including all of the National Capital and Southern Tagalog Regions) absorbed 59% of project funds and the Western Visayas region took another 19%. The latter was a result of the concentration of sugar in that area, and the explosive growth of prawn exports to Japan. The insignificant share of ALF taken by the RB helps explain the failure of the project to maintain, let alone expand, the map of lending under previous Bank credit projects. The entry of KB opened up a new network for Bank disbursement, suggesting good coverage of the regions of commercial agricultural activity, though not as widely spread as the RB. However, most of the KB and TB loans measured by value were applied for and - 10 - approved by these banks at their headquarters in Manila, turning their portfolios in the direction of those farmers and agribusinesses with easy access to the capital. This included some of the largest agribusinesses in the Philippines, for which loans were approved far in excess of any contemplated at appraisal. There were 11 loans of P100 million (about US$5 million) or more, including three to a major sugar company for a total of about US$28 million and one to a large animal feed, food and poultry corporation for about US$16 million. All but one of the 11, and all but three of the 61 "significant" ALF loans (above P15 million) listed in U.S.AID's Impact Evaluation, were made during and after 1988. The geographic spread is detailed in annex 2, which can be read in conjunction with the map attached to this audit report. 2.6 ALFU did a creditable job in executing the project. The financial performance of the ALF is discussed in some detail in the PCR and, especially, the Impact Evaluation. Repayment rates have been very good, a welcome fact compared with earlier Bank-supported credit projects in the Philippines. Repayment rates by the PFI to CBP are 100%, a perfect score explained by the fact that the CBP simply deducts the amortization from the banks' accounts at CBP and that only well performing banks were initially eligible. Repayment rates by farmers to PFI averaged 951. Most of the delinquency was concentrated in the prawn accounts. DBP met the eligibility criteria toward the end of the project. It made limited use of ALF, but the fact that its recovery rates were high was a major and welcome improvement over DBP's previous performance.' The Impact Evaluation shows that one financial problem for CBP that had not been appreciated at appraisal was the lose incurred from changes in the rate of exchange of the dollar vie a via the basket of currencies. The formula for adding the exchange risk to the ALF rate was in this sense inadequate. The analysis shows that substantial hidden subsidies were involved. The PCR also discusses this problem. ALF's financial ratios improved in the last two years, after CBP reversed itself and allowed ALF's idle funds to draw interest. If the foreign exchange losses were included, however, those profits would have disappeared. Diversion of sub- loans by borrowers from agreed purposes was negligible. ALFU inspected 211 sub- projects during the project period (a 10% sample) and found only two cases warranting recall of the funds. 2.7 The interest rate formula, apart from the foreign exchange risk factor, worked well during the first years (1985-87) of the project. At effectiveness, the basic rate was 16.7%, a relatively high rate reflecting the tight money market and, in part, the effective implementation of the IMF credit restraint conditionality. Rates declined thereafter, and, having reached a plateau at 10- 11% at the end of 1986, remained there for 30 months. When the economy started heating up again in 1988, the ALF rate lagged behind. The formula, with its roling average, guaranteed that market rates, fueled by rapid inflation, jumped well above the ALF rate. The margin between the time deposit rate - a weighted average of all maturities - and the ALF rate had fluctuated near zero from the 1 The Regional office points out that stopping the massive "hemorrhage" of public funds through unrecovered credit vas another implicit project objective and one of its best results. See Annex 4, para 5. - 11 - third quarter of 1986 to the end of 1987. By the end of 1989, however, a differential of 6% had appeareds 18.6% ve 12.5%. It was this margin that is suspected of having led to the rapid increase in availment of ALF, starting in 1988. In terms of real interest rates, these were very positive until 1988, when they settled closer to zero (Annex 3b shows real rates of 4.8%, 10.1%, 3.8%, 0.4%, and -0.6% for 1985-89 respectively. But these are based on annual averages and give too much importance to short intervals of highly negative rates. See also Annex 4, para 16). 2.8 The U.S.AID Impact Evaluation, in a good analysis of this phenomenon, points out that - as distinct from the ALF pass-on rate - PFI onlending rates stayed pretty close to market rates, so that the advantage of the differential was captured mostly by the PFI.1i They would have been persuading their traditional clients to allow their new loan accounts to be switched from customary financing, using the bank's ordinary funds, to ALF. Indeed during the audit mission the same behavior was detected at the RB level as well: RB officers had persuaded their larger-farmer associates in the neighborhood of the town to use ALF, accessed through the RB, to finance part of a proposed project that would otherwise have been financed by own funds or KBes outside the ALF network. We must imagine that the existence of that differential allowed the PFI to negotiate terms with the clients and share the savings. Annexes 3a and 3b give detail on interest and inflation rates. C. Technical Assistance Component 2.9 Four of the studies were carried out in 1987, a year behind the agreed schedule. The fifth - the CBP lending manual - was not started. Neither was preparation of the next project (at this time - see next para). Three of the studies were carried out by three separate government agencies, including CBP. The fourth, on the expansion of rural banking, was done by the same consortium of Philippine consulting firms that was later called upon by U.S.AID to write the Impact Evaluation. All four studies were financed by CBP. Two separate sets of studies, including the Impact Evaluation, were financed with US$1 million of the U.S.AID grant. D. Transfer of ALF 2.10 A decision was taken in November 1989 to transfer ALF from CBP to LBP. The main impulse was to take it out of CBP. Where it would go was debated. The reform Government that replaced President Marcos kept the sitting Governor in post at CBP. He now agreed with those who argued that CBP should get out of the business of administering credit programs. He felt it was spending its time trying to recover loans rather than administering monetary policy. This change was supported also by the Bank's Financial Sector Study mission which visited the Philippines in late 1987 and issued its report a year later. The separate 1 The Region points out that the "market" rate is not a good indicator of the rates at which KB were lending to favored customers. The volatility of the market rate was increased during the 1980s, by virtue of aggressive Treasury bill sales. The actual differential would have been lower than 6%. - 12 - consultant studies financed by U.S.AID challenged that position, but were unpersuasive. LBP was eventually selected. The transfer process took longer than anticipated, however, so that there was a hiatus of six months when CBP ran the subsidiary loans on behalf of LBP. The end-game in CBP was messy, with ALFU losing 13 of its staff of 24 before project completion and reducing its capacity for sub-loan review. Fortunately for the project, the bulk of the 1988/89 lending surge had already passed through the system. The ALF was formally taken over by LBP in July 1990. The protracted transfer explains the delay in preparing a follow-on project, which was postponed until LBP could undertake it. When it was appraised in October 1990, LBP was the corresponding agency. 2.11 The transfer process carried with it none of the institutional memory that had been created in CBP. Differences in terms of service between CBP and LBP, and LBP's refusal to absorb ALFU as a whole, explain why few staff transfered. Those who did in that period were not assigned by LBP to ALF (including, remarkably, the ex-assistant director of ALFU), so that none of the team formerly in ALFU are now associated with the program. Meanwhile, LBP has revised the ALF model and created a successor fund called the Countryside Loan Fund (CLF). This is the fund supported by the new Bank project, the Rural Finance Project, with a Loan of US$150.0 million effective in November 1991. Because LBP established a different operating system and a new manual for CLF, there has been a delay in consolidating the two funds. ALF and CLF are now running simultaneously, though they will eventually be integrated. The Bank Loan for ALF was all disbursed during the administration of CBP, however, so that the ALF under LBP is financed from reflows. On transfer, CBP retained the share of original finance of ALF that it had contributed from its own funds (para 2.2). E. Bank Performance 2.12 There are some disquieting elements of Bank supervision in this experience. They raise questions about how seriously the Bank took this innovative step in reform of an agricultural credit portfolio. The last formal supervision mission was in January 1988, missing the surge of lending that was to give the project its unique character. Thereafter the Bank dealt with CBP's submissions for approval of the large sub-loans - that were over the "free" limit and therefore required Bank review - from the desk in Washington. The Bank continued to watch the project, in the context of missions to other projects and in the process of identification and preparation of a successor. However the project files themselves are thin in this period, and no questions were being raised, at least that made it to the files, about the extraordinary concentration of ALF support on the large farmer and company clients of the KB and the disappearance of the RB clientele. The Bank's reorganization in 1987 explains part of this lose in focus. Regional staff told the audit it was well aware of the emerging borrower profile, was not particularly concerned that it was twisting away from the indicative profile anticipated at appraisal, but had been prepared to correct any glaring distortions in the follow-on project. The absence of any discussions of these issues in the project files nevertheless is disturbing to the audit (para 4.1). Moreover, Bank supervision ratings of project performance before and including the last supervision were mostly "1"s (four ls - minor problems, and one 2 - moderate problems), and this despite clear - 13 - signs, not only that project participation was concentrating in a few categories, but that the RB had virtually dropped out of the program, that the studies were late, and that their recommendations were not being implemented in parallel as part of the expected IB program. 2.13 When U.S.AID had queried the Bank during appraisal whether the project would lose the RB and the smaller commercial farmers, the Bank had not only answered "no" but, in a series of internal memoranda, had disparaged the intervention by U.S.AID and complained about that agency's lack of experience and understanding. That conceit was unwarranted. The Bank had miscalculated.!' The Bank's posture is partially recovered in the SAR of the follow-on project, by a good description of how the first project turned out. Measures are included there to see that the results are better the second time around. 2.14 Finally, the Bank ignored an important part of the three-phase monitoring and evaluation program (M&E) that had been agreed at appraisal. Phase (1) was for ALFU to set up a monitoring program to report on the use of ALF. The other phases were to be organized by CBP to (2) provide ongoing evaluations of the impact of ALF sub-loans at the levels of both the participating farms and the PFI, and (3) to prepare a completion report at the end of the project. ALFU carried out (1) and (3) satisfactorily. Phase (2) was also assigned to ALFU, but it was never carried out in any depth, and petered out when ALFU lost its staff. Supervision never mentions it. This fits the pattern of apparent indifference to sub-loan distribution and impact traced in the previous paragraphs. III. PROJECT OUTCOME A. Policy Impact 3.1 The change in Government's position on subsidized credit is remarkable. There is a qualification that is discussed in paragraph 3.4, but the overall expression of official policy on rural finance is now as strongly oriented as the Bank's toward market-based credit pricing and deregulation. CBP's rediscount program continues, and is also priced at market rates. That was not the case in 1984, and the shift must be attributed in a large measure to the project and the Bank-Government dialogue which surrounded it. The effective expression of this policy really begins with the change in government in February 1986. TBAC was dissolved. In December, Government created a new policy unit, the Agricultural Credit Policy Council (ACPC), which in a sense replaced TBAC. This time the unit it Refer to Footnote 1, para 1.13, and to the Regional Office Memorandum in Annex 4. It implies there was no miscalculation, and that the Bank's response to U.S.AID may have been an exercise in disinformation. ORD has concluded that miscalculation, at least by those responding to U.S.AID, is the appropriate word. - 14 - was firmly controlled by advocates of credit reform. ACPC is attached to the Department of Agriculture (DA), and the Secretary of Agriculture is also committed to reform. 3.2 ACPC is an executing as well as advisory unit. The same Executive Order that established ACPC also consolidated 17 of the then ongoing 39 targeted lines of subsidized credit administered by DA and CBP - and 70% of all the funds. A new fund - the Comprehensive Agricultural Loan Fund (CALF) - was created out of the majority of the balance in those 17 credit lines. CALF was established not as a successor lending operation but as an apex auarantee fund: helping to insure but no longer directly finance the retail credit activities. ACPC administers CALF. The body of reports issued by ACPC in the last three years is as comprehensive and articulate a set of documents arguing the case for credit reform - market rates, removal of subsidies, mobilization of deposits, strict observance of eligibility criteria, and true financial intermediation - that one can find in the Bank's world. Since some of the key actors in this reform establishment have passed through Ohio State University, the Land Grant College that U.S.AID has helped become a center of excellence for agricultural credit since the 1960s, and brought their professors back to the Philippines for technical support, the role of that faculty in the policy success story must also be acknowledged. 3.3 Success at the policy level with this project should be seen in the context of good ratings overall in policy improvements during the 1980s. These are mostly associated with the present Government, but important initiatives were also taken before. Three steps were made in that direction in 1984, with the approval of the Agenda for Action in Agriculture 1984-88, with the agreement with the Bank on its first sector adjustment loan with the Philippines (Agriculture Sector/Inputs Project, Loan 2469-PH), and with the IMF Standby Agreement. The PAR on the Agriculture Sector/Inputs Project describes the reforms that followed, which included partial liberalization of the import, export and internal trade regimes (OED Report No.10314). The SAR for the follow-on project to the one under audit claims that the Philippines ranks "near the top" in performance against structural adjustment targets. The agricultural credit reform fits that pattern. 3.4 The qualification referred to in para 3.1 is as follows. LBP is responsible not only for ALF and CLF, but for disbursements drawn from the agrarian reform fund, ARF, assigned for use for credit for beneficiaries of the land reform program, CARP. Total annual disbursements rose rapidly after ARF was created in 1986, and in 1991 alone were estimated at about US$170 million. Ninety-five percent of these loans are provided not to the farmers but to their cooperatives. The loans pass at a fixed rate of 12% mostly direct from LBP to the coops, the rest through the RB. They are almost all short term. They are not only targeted but subsidized. The target group is that large mass of small subsistence farmers who are not presently addressed by the formal rural credit institutions, including the RB. 3.5 The subsidy is neither as conspicuous nor as deliberate as the subsidies that were conveyed by the pre-1984 set of credit projects. The - 15 - onlending rate of the coops for their members at about 182 approximates market rates for commercial farmers. However the 122 return to LBP comes nowhere near covering total costs related to this agrarian reform lending line. LBP's subsidy to the coops allows the latter to cover their costs of organizing their members. The President of LBP is one of the most respected persons in the field of rural finance in the Philippines, having entered public service from a major private bank. He comfortably defends the subsidy on costs of administration, and the use of fixed interest rates on ARF loans, as farmer friendly, politically inevitable and non-destructive. He welcomes the day when these subsidies can be eliminated. But he says that he has no intention of adopting the reform platform of full cost and flexible rates as a matter of principle, at least at present. He insists among other points that his clientele are poor, not business oriented, and cannot be expected either to cover the costs of credit management or to understand and accept the floating rate. He feels also that neither ALF nor CLF can address the needs of these farmers, though, again, their farm economy may evolve over time in that direction. Thus, LBP now administers side by side the program which fully reflects the new credit discipline, and another important program that does not. ACPC would appear to accept that some farmers, and in this case the majority, are outside the scope of the reform. 3.6 The importance of this exception has been diminished in the last year by a decision from the Justice Department to limit LBP's access to ARF. Withdrawals both for onlending and for LBP's administrative expenses have been cut severely, in response to complaints that they were a misuse of monies originally assigned to land purchase. LBP is now forced to look elsewhere for substitute sources, including an increase in LBP's onlending rate. Nevertheless LBP's credit program for the subsistence farmers has been deliberately protected from the full thrust of the project reform package. B. Institutional Strengthening 3.7 Since the late 1970s, Government has initiated four rounds of remedial action aimed at financial recovery of the RB. The first two, in 1978 and 1982, were based on a restructuring and stretching out of the RB debt to CBP. The third, undertaken by the new Government in 1987, was based on offers to convert part of the RB debt to Government-held equity in return for fresh input of new equity by the RB owners themselves. This program reflected in part the proposals of the project-financed studies. As mentioned above, the results of this third round were unimpressive, and RB eligibility continued to decline. The fourth round is being implemented in tandem with the promotion of the CLF. It resembles the 1987 program, though it allows for the first time a write-off of part of the total RB debt. 3.8 At the level of the KB, the institutional impact of the project was also not strong (nor was it expected to be - para 1.10). This subject is treated in the U.S.AID Impact Evaluation. A few of the KB under the influence of the ALF did establish units specializing in agricultural loans. Others expanded rural lending above previous levels. But the overall pattern was one of business as usual. - 16 - 3.9 The dismantling of ALFU is another aspect of what must be assessed as a much more modest impact on IB than had been expected at appraisal. In CBP, the disappearance altogether of project implementation units established in the 1960s under the Bank's rural credit project line, followed by the ALFU, is a sad story of the ups and downs of collaboration with the Bank. There are real people and careers involved - 150 persons had to be reassigned in 1989 and 1990 and some may be released. Some also find the Bank's shifting and unpredictable loyalties intolerable. While the Bank had not intended that the rediscount facility remain with CBP forever, the transfer to LBP can only be described as badly managed. The PCR says the same thing, in gentler language. The PCR is concerned especially about the loss of control over the quality of CBP supervision of the ALP program during the transition. It says that the details of the transfer process warranted more attention at the planning stage than they got. C. Deposit Mobilization 3.10 The PCR is silent on whether there was any significant progress toward this objective. The U.S.AID Impact Evaluation does address this subject, based on interviews with selected PFI. It concludes that there is no evidence that the ALF program had any significant influence ("very slight") on the raising of rural deposits, in any of the banking networks. KB deposits were rising in that period, but the consultants were satisfied that that would have happened anyway. The Regional Officea' and CBP dispute this conclusion, as shown in Annexes 4 (para 18) and 5a (para 9). The subject is important, insofar as it is one of the keys to the strategy of retooling the apparatus of rural credit to fashion a fully developed system of financial intermediation. With respect to the RB, there is no dispute. The Bank had expected the participating RB to expand their deposits. They neither participated nor expanded deposits. 3.11 The SAR had talked about the project serving the purpose of "term transformation" - converting short term deposits into longer term loans. That aspect of the argument was also a disappointment. The KB used the ALF to replace the short term credits they were rolling over annually for several years to favored clients. But the ALF terms the KB extended were themselves relatively short - three or four years, instead of the 5-10 year package predicted at appraisal. The KB own-funds that comprised their 20% contribution to the sub- loan would have also been loaned for three or four years instead of a rolling one year agreement. These own-funds were mostly drawn from short term deposits, so the shift is in the right direction. But that is a much less impressive transformation process than had been expected. D. On-Farm Recovery and Development 3.12 There is also little information available to assess properly the impact of the sub-loans on production. At project completion CBP carried out a sample survey of 84 borrowers to determine the efficiency of the sub-loan 11 Though the Regional Office believes that part of this increase in KB deposits was attributable to the project, it says it did not say so in the PCR because it had no evidence. - 17 - delivery system, the uses of the funds, and their effect on farm sales, profits and employment. The Impact Evaluation provides some supporting information, and the PCR addresses the subject as well. The shift of ALF toward the KB and their large-scale borrowers, for maintaining and expanding their proven profitable ventures, may have guaranteed that the uses of the funds had a greater impact on production than it otherwise would have. For example, the ALP played an important role in the recovery of the large sugar estates, and the development of prawn farms, in the historically important commercial agricultural zone on the island of Negros in Western Visayas Region. Another example: the large loan to the animal feed/poultry corporation (para 2.5) helped finance the extension of its arrangement with the many small scale commercial poultry farmers contracted to supply broilers. These farmers also got separate ALF loans. The on-farm poultry enterprise was not only an important component of ALF, but one of the booming sub-sectors in Philippine agriculture. 3.13 The PCR provides reestimates of economic returns to the indicative investment projects that were evaluated at appraisal. The SAR shows an ex-ante range of economic rates of return between 37Z and 84Z. The PCR shows an ex-post range between 15% and 30%. The SAR for the follow-on project says the results of the previous project were "moderately" but not "spectacularly" profitable. In the view of the audit, that is an acceptable outcome of the project when judged against the difficult economic environment of the project period. The audit has no information to challenge those reestimates. 3.14 What the audit did find however was that the additionality in ALF lending was lower than expected: the incremental impact of project sub-loans over what would have occurred in the absence of ALF. This is an issue which attracted the attention of the consultants in the Impact Evaluation. In the summary narrative that report admits to some substitution of ALF funds for other farm finance; in the inside pages the criticism sharpens and the rate of substitution is alleged to be substantial; in private discussion with the authors during the audit mission it was clear that the persons who did the interviews of selected borrowers and bankers felt that it would be hard to demonstrate much additionality at all for the KB and part of the TB portfolios. This conclusion is easier to accept given (1) the size of the borrowers, (2) the relatively short maturities of most ALF loans and (3) the preponderance of working capital and other production loans, all of which suggest that alternative sources may have been available. 3.15 The audit mission considered this subject important to any overall evaluation of the project, and pursued it at first in visits to rural areas near Manila. These visits were made exclusively to borrowers from participating RB and TB. In their majority, it was clear that the ALP borrowers were traditional clients and that the banks would have been prepared to finance at least part of the activity in their "ordinary" portfolio. It was the farmer who was willing to wait out what was considered an overly-lengthy ALF approval and disbursement process that took ALF. If he was in a hurry, the bank would use its own funds, at a higher interest rate. There is an analytical problem of summing here, because although the RB and TB would have been able to support each of the loan - 18 - requests in their ordinary portfolios, they would not have been able to support them all. This was not the case with the KB (and some of the other TB). 3.16 The PCR admits that there was a substitution factor, but concludes that the level of incrementality achieved by the project was nevertheless substantial. The consultant who drafted the PCR was in Manila at a time when the other consultants' work on the Impact Evaluation was itself in draft, inaccessible, and not a factor in the Bank consultant's consideration. The audit does not accept the PCR conclusion of substantial incrementality, but neither is it prepared to minimize the significance of the additionality that did occur. 3.17 To get a better.fix on the incidence of this substitution phenomenon, the audit commissioned the same consultant firm to do a special study of additionality based on visits to borrowers of small and medium scale sub-loans from two of the TB that were highlighted in the consultant's Impact Evaluation. The TB were selected by the audit over the KB because it felt that the incidence of substitution would be higher among KB clients (the PCR makes the same assumption too). It was important to determine whether substitution was a characteristic also of TB lending under ALF (the RB were left out of the special study: their share was too small to give that group any significance in the total population of borrowers). 3.18 The results of the special study show that additionality was greater in the TB portfolio than is suggested in the consultants' 1991 Impact Evaluation. That discussion had been focused mainly on the large scale KB clients. Of 20 TB borrowers in the special survey sample, only 2 appear to have been prepared, at the time the sub-loan was made, to invest in the ALF financed enterprise in the absence of ALF, at least at the scale carried out. The rest would have financed a smaller operation, or none of it. The evidence came from interviews with the bankers as well as the farmers. This is an unrepresentative group of borrowers, since half of them were selected from the minority of ALF clients who invested in physical assets as well as or instead of working capital and seasonal credit. Thus the study cannot say anything about the behavior of the KB clients who borrowed large sums of ALF money for short terms or for permanent working capital. That is where the incidence of substitution is suspected to have been very high, a point made repeatedly in interviews during the audit mission, and highlighted in the 1991 Impact Evaluation.Y Nevertheless, the fact that at least some additionality seems to have been achieved for the small to medium scale clients who borrowed to invest in phys!aal assets provides support for the original project concept. This was almost certainly the class of farmer, and purpose of loan, that had been expected to predominate. The fact that the RB clients were largely left out, and that the big KB clients provided much less additionality, reflect implementation distortions at the bottom and top of the portfolio but r.ot an indictment of the project per se. 11 The Regional Office and CBP dispute this conclusion, and cite other information to support their claim that the level of additionality in KB use of ALF vas quite high. See Annexes 4 (para 17) and 5a (para 10). On the other hand, U.S.AID told the audit that its experiences vith supporting credit programs th-ough the Philippines banking system all demonstrate the phenomenon of low levels of additionalitZ. - 19 - 3.19 Nevertheless, the implications of substantial substitution effects are obvious. This project did not come as close to its objective to stimulate the recovery and expansion of the agricultural sector as the size of the lending program suggests. Some of the investments in new physical assets, and many of the disbursements against production costs and other uses of working capital, would have happened anyway. That means, for example, that although the ALF money can be closely associated with the recovery of sugar and the dramatic expansion of the prawn farms on Negros, these important developments in the sector would largely have proceeded regardless. 3.20 The discussion of substitution effects does not stop at that point. It must consider the uses of the funds that were substituted. In this case, ALF displaced: (1) the borrowers' own assets: (2) loanable funds that the PF1 already had available; and (3) funds that remained with the depositors who in the absence of ALF would have been approached by the PFI, with one or another incentive, including higher savings rates, to deposit additional funds needed to meet the demand. In each case any attempt to identify the uses and effects of those replaced funds requires a statement about how the farmers, bankers, and potential depositors would otherwise have used them. That is beyond the scope of the report. 3.21 In the opinion of the audit, the majority of those released funds would not have ended up in the agricultural sector. That is the basis for the assertion that the lack of additionality in ALF lending implies there was a proportional lack of incremental activity in the sector. The PCR takes a different position. It asserts that the "indirect" impact of ALF was substantial. By this it means that the released funds were shifted to another agricultural activity. Since most of the released funds were in the hands of the banks and potential depositors, and not the farmers, the audit does not agree with that conclusion. E. Technical Assistance Component 3.22 For the U.S.AID Impact Evaluation, the consultants reviewed in detail each of the recommendations put forth in the four reports completed in 1987 under the project, including its own report. The analysis is objective throughout, and the authors conclude that the quality of the reports was generally good but that the degree to which the recommendations were adopted by Government varied by study but was generally low. Part of the problem was that the recommendations were directed at a range of public institutions beyond the influence of CBP. The training program by CBI is also studied in detail in the Impact Evaluation. In this case, the report finds that a good-looking program, which reached over 1,000 in-service trainees in CBP and the PFIs, was offset by the lack of evidence of any long term impact on the trainees' performance back in post. For the CBP staff, that is accentuated by the transfer away of ALF. The PCR assessment is brighter than the Impact Evaluation, and concludes that the studies and training were, on balance, worthwhile. 3.23 However the question about the impact of the studies is more important that a simple count of recommendations put into effect. As mentioned above, the - 20 - study program was seen at appraisal to be an integral part of the reform strategy. The studies were to complete the reform plan, and their recommendations were to be put into operation soon enough to have an impact on the efficiency and spread of the ALF program. The parts of the study program that dealt with recovery and expansion of the rural banking structure, with the treatment of low income farmers, and with the handling of arrears with the CBP, if put into play, should have brought many more of the RB into the orbit of the project. It did not happen. Despite the 1987 rehabilitation program, Government never gathered the political will to take the decisive action that was required to eliminate the hopeless RB, force through a reorganization of the residual network, and manage the remaining RB debt. The network was in worse shape at the end of the project than at the beginning. F. Balance of Payments Support 3.24 The Bank Loan provided almost US$100 million of unconditioned foreign exchange, since the import content of the project sub-loans was small and Loan disbursement against foreign TA was zero. The Government had wanted to use these funds to help reduce the foreign exchange crisis of the mid-1980s. However, by the time Bank disbursements started in 1986, pressure on imports had abated. Almost half of the Loan disbursements were made after 1987, and served in relieving pressure during the next balance of payment crisis that developed in the overheated economy of the late 1980s. Government and the Bank urged the CBP then to accelerate use of the ALF. This promotional work partly explains the surge in disbursements starting in mid-1988. But whether in the mid- or late- 1980s, this objective of the project was met. IV. FINDINGS AND ISSUES A. Participation by Small and Medium Scale Farmers 4.1 The ALF was in no sense a small farm project. It was never intended to include the large mass of subsistence farms. It would be available to the smaller commercial farmers who were the principal beneficiaries of the 13 earlier Bank-supported agricultural credit projects. But these were not expected to consume a major share of the new fund, because their bankers - the RB - were in their majority not expected to qualify. Nevertheless, the audit cannot accept the assertion by some of the persons interviewed that the project design was indifferent to the participation of this group, and that an ex-post assessment of the project should be equally indifferent to the fact that the RB ended up w..th only 1% of the funds.' The indicative plan of the SAR, the project files 21 As is argued in Annex 4. - 21 - in the preparation period, and the intentions of the Bank's appraisal team and the U.S.AID officers who followed it all demonstrate an expectation that this group would continue to participate within the conditionality of the reform. To that extent, the absence of this group of farmers is disturbing and has wider implications (paras 4.20 ff). 4.2 It is worth pointing out that, while the Bank during this project appears to have abandoned the smallest commercial farm strata that were the targets of its earlier lending, this was neither intentional nor does it raise the kind of equity issues that would jump out if the shift in borrowers had indeed been from small subsistence to large commercial categories. 4.3 At the same time, one must not be overwhelmed by the very large loans that were made to the 61 "significant" farmers and companies (para 2.5). The US$28 million and US$16 million examples are of course exceptional. The vast majority of borrowers were in the size group that in other countries would probably be labelled medium scale farmers, and these were very much the target of the program. CBP statistics can be misleading in this sort of analysis. CBP project accounts distinguish between small, medium and large scale farms measured by pre-project assets, and also between small, medium and large scale ALF sub- loans made by the PFI. But these uses of "small" are relative, and the cut-off. between small and medium reveal that CBP's "small" categories include some very big actors. The asset definition classifies any enterprise less than P5 million (US$240,000) as small; the loan definition classifies any sub-loan less than PIO million (US$480,000) as small. Within the last classification, CBP records show that 48% of the value of ALF sub-loans, and 86% of the number of AFL sub-loans, were small. In the Philippine context, most of them are better addressed as medium. 4.4 In fact, 1,046 of the total 2,013 sub-loans (521) were under P1 million (US$48,000) in size, and averaged US$28,000 per sub-loan. This group absorbed only 121 of ALF funds. But, as seen above (para 3.18), the level of additionality in lending for the group seems to have been relatively high and it can be considered the implicit primary target of the project (irrespective of the SAR's indicative plan). It appears to be the intention of both the follow-on project and LBP to ensure that this group's share of funds from CLF, the now fund, will be substantially greater. Early reports on ALF lending (reflows) by LBP suggest that such a shift is underway (see LBP's letter in Annex 5B). B. Role of the Rural Banks 4.5 A fundamental part of the reform program was to create the conditions for a complete overhaul of the RB system that catered to the smaller commercial farmers. The objective was to set the RB up as rural financial intermediaries that could efficiently raise, lend and recover funds. This was a dominant position in the Bank and U.S.AID, although there were officers in both who were skeptical about the possibility of RB reform and would. not be surprised or concerned that their participation in ALF was insignificant. The ALP project by itself was not designed to reconstruct the RB system. As mentioned, the Bank and U.S.AID expected that a separate project or public initiative would take shape - 22 - for that purpose following the studies, and in time to bring a large number of the RB to ALF eligibility. For several reasons, that complementary effort never gathered speed. The RB arrears position continued to deteriorate, the eligibility even of many of those RB that initially qualified was withdrawn, and the KB and their scale of borrower took over. At the beginning of the project it was hoped that about 25% of the RB would eventually participate; at the end only 2% had. The LBP is promoting yet another program for restructuring the RB in association.with the CLF. That goal is still seen by many as distant, but LBP is already reporting some progress (Annex 5b). 4.6 The question arises whether the project may have accelerated the decline of RB viability, by shifting Bank funds within CBP from the rural credit projects - which directly supported the RB - to ALF. There is no basis for such an argument. The main reason the RB got into trouble was because they had been administering official, subsidized credit programs that had an inherent bias to default, without being able to shift the risk back to government. In fact the majority of accredited RB did not want to access ALF, partly to avoid further trouble.!' For some RB, and there are critics who claim it is the majority, this bias was compounded by weak or dishonest management. If ALF funds had been stuffed into the RB network through some preferential arrangement, the problems would have been aggravated. It is better to make the opposite assertion: that if the RB rehabilitation program had moved faster, the entry of these banks into ALF would have made for a more balanced portfolio and a higher incidence of incrementality. 4.7 But there may have been limits on the degree to which the RB system could have responded. Some recent papers issued by ACPC suggest that a part of the RB problem may be inherent in the structure of the system. These are small banks with localize clientele, in risky agricultural enterprise, with a high covariance among clients in production trends and banking requirements, and with no means to spread the risk. The implication is that the RB system may have to be consolidated. The leader of the appraisal mission for the project under audit rejects that view, and insists that it i3 the very decentralization of the RB system that gives it its great potential. Nevertheless, to the degree the ACPC point has validity, project design had an even more difficult uphill assignment. C. ALF Interest Rate 4.8 Central to the project design was the formula for setting an ALF onlending rate that would match market rates and encourage the banks to seek to raise more deposits. The "floating" formula was lagged to cushion the effects on borrowers - who were accustomed to fixed rates - of rapid and large fluctuations. But the formula was not geared to accommodate periods of a sustained ris or fall in inflation. Then, the built-in lag would open, on the upswing, a substantial differential below market rates in favor of the participating borrowers and bankers. That is what happened starting in mid-1988 for the last year-and-a-half of the project. It was profitable for both the See Annex Ua, pars 11. - 23 - larger farmers and their commercial banks to switch their traditional line of finance to ALF, sharing the savings. Perhaps half of all project lending was influenced by this predictable response by borrowers and bankers to adjust their portfolios. When inflation was falling, the reverse would occur. Farmers would avoid ALF in favor of market rates. This rational behavior, and the surge of ALF disbursements in 1988, is a consequence of an interest rate formula that proved in retrospect to have traded away too much of the indexing target in favor of the cushion. The follow-on project indexes the rate on a monthly basis. D. Policy Reform 4.9 As part of the agrarian reform program, a large fund was made available to LBP to disburse as credit to the "real" small farmers, the actual and intended beneficiaries of the agrarian reform program. LBP determined that most of this would be lent to farmer cooperatives, for onlending to members or for collective activities. This substantial credit fund, five times the size of the combined ALF/CLF in 1991, is not only targeted but subsidized. LBP management does not accept that its primary target group, which includes most of the country's farmers, should be subject to a floating, market-oriented interest rate, or be expected to pay the full costs of administering the program. Thus the same organization that has been selected to manage the CLF, under the disciplined terms that are associated with the new policy position discussed above, is simultaneously managing another fund, for many more farmers, that for the time being appears to remain outside the reform. 4.10 It is worth contemplating the extent of this exception. In orders of magnitude, KB lending to agriculture during the project period was financed mostly with its own funds (95%), and the rest from CBP rediscounts (4%) and ALF (1%). Since the CBP rediscounts are now deregulated, the SAR for the follow-on project is correct in saying that the bulk of farm finance, including KB use of its own funds, is market-oriented. But in 1991, again in orders of magnitude, the combined CBP and ALF/CLF flows were about equal to the funds being disbursed by LBP under ARF (unfortunately, this narrative and its readers must bear with many acronyms, including the look-alikes - ALF and ARF). This means that last year about half of these official lines of rural credit conformed to the new model and half did not. Also, over 80% of the farm families would be eligible for ARF, swamping the numbers with access to ALF. 4.11 The SAR for the first project allowed that there would be some exceptions to the new credit policy: "while CB and Government have agreed to phase out subsidies on agricultural credit, there may be situations when the Government would need to extend financial assistance to less privileged, low income groups which currently do not have adequate access to institutional credit". But that exception for a "safety net" was never intended to refer to all the small farmers and half the official flow of farm credit. The Ohio State model does not allow that kind of exclusion. In fact a key argument for the model is - 24 - that the small farmers themselves will be served better by deregulation and the removal of subsidy. 4.12 The SAR for the follow-on project recognizes this anomaly, but is not concerned by it. The SAR says the margin of the subsidy in the interest rate is not large, and the total value of the subsidy conferred by ARF is also small.!' Both points are correct. And, as noted above, LBP's access to ARF has recently been curtailed, so that in both respects these margins are going to shrink further (para 3.6). Thus the issue raised here may disappear from view, as LBP is forced to move cl )ser to the ALF model. But it moves of necessity and not of conviction. The point remains. The extent of Government's conversion to the new concept of rural credit remains in doubt. The CBP position in favor of reform is firm. But it appears that the agrarian reform movement has a momentum of its own, and can turn aside the credit discipline. The President of LBP has twice in the last year requested the President of the country to permit LBP to raise interest rates to the coops from 122 to 16%, to cover its costs, and twice was refused. The letter recently received from LBP indicates that that adjustment has now been made. Nevertheless, LBP continues to defend separate treatment for subsistence farmer (Annex 5b). The U.S.AID Impact Evaluation report stated in March 1991 that "Government appears consistently moving towards a deregulated rural credit system". The audit is not convinced (but see para 4.22). E. Were the Results of the Proiect Satisfactory? 4.13 The audit concludes they were, but this is a very close call. 4.14 One must be concerned about the concentration of ALF funds on large farm clientele, to finance mostly their requirements for working capital. That characteristic of the project is not in itself inconsistent with project objectives. And uses such as for working capital and expansion can have as great an impact on production as for fixed assets and new activities. Also, the fact that agroindustry came to dominate the ALF portfolio was not unhealthy. Through its arrangments with contract farmers agroindustry can have a significant impact on smaller farm enterprise. But the extent of these shifts was clearly unexpected, and the character of ALF changed from the appraisal model. 4.15 The appropriate tests of success are whether the policy reorientation is durable, whether the funding mechanism that was put into play helped strengthen the capability of banks lending in rural areas to carry out effectively the job of financial intermediation, and whether the fund contributed to the recovery of agricultural enterprise. These imply as well an impact on deposit mobilization, some progress on the implicit agenda of RB reform, and substantial incrementality in farm production above what would have been financed anyway. None of these appear to be prominent features of project results. Nevertheless, the reorientation of the entire CBP portfolio to the new credit disciplines and the experience gained in running the unsubsidized refinance I The Regional Office makee this point again in Annex 4, para 21. - 25 - operation are significant achievements that make the project worthwhile - provided the reform movement does not stop there. 4.16 In the judgement of the audit, the project fell far short of fully demonstrating the alleged advantages of the policy reform. The SAR for the new project says, in its section on lessons learned from the experience that preceded the project under audit, that: "directed credit by crop or type of investments, at subsidized interest rates, did not sustain economic growth, improve the credit delivery system in the rural areas, or increase the access of rural borrowers to formal credit facilities". The project under audit did not improve on that record. The audit also cannot accept the conclusion in the PCR that: "the project has made a substantial contribution to the institutional development of the rural credit system in the Philippines". It was precisely on the institutional side where lack of progress was most conspicuous. It is essential that this first step be followed by improvements that put this rediscount facility to better service. If the new project, the Rural Finance Project, ends up repeating the experience of the first, the value of both phases will have to be questioned. Adjustments in the design are intended to provide just those improvements. F. Lessons for Credit Policy 4.17 The project is the first to provide evidence of the practicality of the new policy. From one perspective the results were alarming: the Bank succeeded only in abandoning its preferred, traditional target populations and putting its credit funds instead in the accounts of companies that did not need them. This is an unfair characterization of the project, but a quick reading of the results leads the cynic in that direction. Certainly the project made less progress than expected. 4.18 However, three circumstances warn against any generalization from this experience. First, the RB network that the project formulators expected to balance out the large farm activity was already on a downhill slide from arrears previously contracted. The project under audit was not fortified with the equipment it would have needed to arrest that decline. That had been left to the study program, and to the effective implementation of its recommendations - which did not happen. Second, the inflation of 1988 ambushed the project's floating interest rate formula and encouraged the big banks and borrowers to grab a profitable opportunity. In the new credit policy model, a large differential is not permitted and should not have emerged. And third, the project was overtaken by the land tenure legislation of the new agrarian reform program, which removed - 26 - the collateral value of land and undermined the traditional basis for most bank lending to farmers. All of these problems may be correctable. 4.19 It is important now to watch performance under the follow-on project and the CLF. RB reform, an interest rate formula that keeps closer to the market, and some restrictions on loan size and working capital are put into the new design. And LBP is intent on playing a more activist role than CBP, and reorient the CLF toward smaller clientele and investment projects (and do this without "targeting"t). Even before the CLF went into operation in late 1990, LBP had been able to start shifting the ALF portfolio in those directions. Its letter to OED shows that that shift continues (Annex 5b). 4.20 But the clearest lesson of this experience is that the KB and TB are profit maximizers and risk averters, that especially in periods of inflation and economic uncertainty they prefer to remain on short exposure, that they will seek secure loans, that they never like to make small loans to small farmers, that they are slow to open new rural branches, and that they cannot be expected to respond spontaneously to the development objectives of Government and the Bank. Philippine bankers are no different than their counterparts in other countries. Open lines of credit that are not targeted will concentrate rather on more secure objectives: unless, that is, project design can harness those profit making instincts of commercial bankers better than this project did. 4.21 There seems to be a consolidation of opinion in the Philippines around the proposition that most of the smaller farmers - including commercial as well as subsistence farmers - will be squeezed out of credit programs that leave lending deci.sions entirely to the discretion of the bankers. Although the underlying development rationale of the project remains valid, it appears that some accommodation must be made to bring bankers and this clientele together. Thus the Impact Evaluation concludes: "it seems an institutional rural credit program such as the ALF cannot be left entirely at the mercies of the current formal financial system dominated by commercial banks. There must still be specific forms of intervention to ensure credit flows to small enterprises", and an ACPC report says: "the logic of this argument is that by removing interest rate restrictions or letting interest rates reflect market risks and administrative costs, banks or financial institutions will be induced to lend to rural areas consequently increasing the flow of credit. Indeed while the above argument appears logical, its assumptions are questionable and its view on the nature and operation of credit markets is simplistic if not outright erroneous". 4.22 The banker's aversion to lending small loans to small farmers is based on three factors: the risks inherent in small scale farming, especially dryland - 27 - farming; the risks that small farmers will not repay, even if the investment is successful; and the high costs of transacting and managing these small accounts. For the bankers to cover these risks and extraordinary costs, they would have to charge interest rates at levels that would not only drive away the customers but invite political reprisal. In that context, the position of the President of LBP seems to be a reasonable compromise between the discipline of the market place and the special circumstances of small farm agriculture. He is prepared to keep interest rates close to market rates, but is not prepared to load these rates with the extra costs of risk and administration. He also will not set the rate loose, because the advantages in reducing misallocation are outweighed by the disadvantages of farmer misunderstanding and mistrust. We may be looking at the sort of compromise that is needed to extend the new credit model down the farm size scale. In this sense, the "exception" to policy (para 4.12) may be no exception at all, but the only way to make the policy viable for the majority of farmers. That includes many of the smaller commercial farmers in the RB network. 4.23 In short, the shift to the new model is not as simple as projected. One must question whether leaving the difficult part of RB reform to the study program was the appropriate strategy for the Philippines Agricultural Credit Project. The appraisal team considered and then dropped the idea of making the studies and their follow-through a condition of disbursement. But some method to ensure rehabilitation and participation by the RB without abandoning the integrity of the program would appear in retrospect to have been needed. If the interest rate differential had not opened in 1988, prompting the KB and their clients to grab onto ALF, the disbursement of project funds would have gone on at a much slower pace, but without any change in the overall share taken by the large farmers and companies. The RB and their clients would have still been outside. 4.24 This is not a finding in favor of the old "supply-led" model of rural credit. That would be a disastrous misreading of the project outcome. Nor does one find in this experience any counter-case against the removal of subsidy, although the argument for practicality put forth by the President of the Land Bank is persuasive. But it does suggest that the removal, as well, of any form of targeting, and refusing special support to finance the extraordinary costs of the small farm sector, is premature, given the nature of the banking network. that presently exist. Special support will be needed; it must be targeted to avoid its abuse; it must be funded from the budget to avoid abusing also the financial intermediaries; and there is no reason for the Bank to decline participation in such small farmer credit programs if properly constructed. 4.25 Thus, the policy reorientation was a necessary condition of credit reform, but was not a sufficient condition to move the system except in the upper strata of farm size. It freed up the supply side. But it did not touch the demand side in the small farm sector - the preferences and requirements of both the bankers and the farmers. For this, substantial progress is also needed in institutional reform, in improving road infrastructure, and in reducing the risks and costs of small farmer agriculture. If these tasks are ignored, future projects will end up just as tilted as the audited project. AICLURAL P~ 1CTI0 LOM GMTED ET IMTIUIffi 191-1989 (Amoun in Miio P~cs) 1985 1986 1987 1988 1989 Amount % Share Amount % Share Amount % Share Amount % Share AManut % Share Government Banks 1,3778 5.0 579.0 23 1,485 SA 3,492.5 9.9 5ß399 18.7 PNB 1,315.7 4.8 5742 23 702.4 2.6 2,631.4 7.5 3,394.6 10.9 DBP 62.1 0.2 4.2w 0.0 131.2 0.5 458.7 1.3 210.6 0.7 LBP n.a. - n.a. 652.20 2.4 402.41 1.1 2,233.7W 7.2 Private Banks 26,122.9 95.0 24,535.4 97.7 25,9742 94.6 31,797.5 90.1 25,367.0 813 PKB 22,479.7 81.7 20,008. 79.7 21,007.1 76.5 23,777.0 67.4 18,5585 59.5 SMBs 12.9 nil 97.1 0.4 90.2 0.3 201.6 0.6 355.9 1.1 PDB 3523 1.3 858.1 3.4 1,013.4 3.7 1,190.3 3.4 1,481.1 4.7 RBs 2,77&9 10.1 2,948.1 11.7 3,456.4 12.6 3ß854 10.9 4,402.7 14.1 SSLAs 499.1 1.8 6233 2.5 407.1 1.5 2,773.8 7.9 568 12 Grand Total 27,500.7 100.0 25,114.4 100.0 27,460.0 100.0 35,290.0 100.0 31,205.9 100.0 PNB - Philippine National Bank SMBs - Savings and Morage Banks DBP - Development Bank of the Philippnes PDBs - Private Development Banks LBP - Land Bank of the Pbilippines RBs - Rural B~ank PKBs - Private CoMMercial Banks SSLAs - Stock Savings and Loan Associations 5 Eutimated based on first semester data. W Estimated based on actual data of preceding eleven monthL. 5 Includes PAB koans wich can not as yet be aggregated. f Source of data from IBP. Sourcer: PNB, DBP, CH-DER, and CB-SRCAD and compiled by ACPC. x001 99lSKzK 190'9919 syf f 1i, g 11l'961 9918 m00r KE1t E1 541f*9999 090*96 200 699 ,00K t," ~*9 K$0°Cet 100o 1 l ot e 9001 90°i90ip Kn* c 9s 20 9Z9°O09 99£*99 zo0* £91 19301 K 5W14t 19161t n 59 Tof S9r°Te 6tvilft IC te rO * 069sv Kl g c l r 0c 1 09-1 K 9 X 0099 Oivt Et 9 IT c59 6to69 l0t %t 'S! y 0 Uo'£ (E6t æ 9 w 0we" 00l0 zo 0 l 1LO*£ 9t6'5 K1 9 go 000 00 30 0 10 "0 0000 K0 # EO 0000 000-0 10 ixl 601019 101 9tO*'s 9999i9 U t IL 3sti018 524"og 91 s 5 L% s t 09t1 E69I9I 19 og T V9iEc 99101 91 IC I £95"51 9'1PI set St T 69"Kut 69"t? 1C tv 1 *osteS 11 16cs9 960,95 n 65 Kl 09W1 991: M KL It z9l', ttK61 1 9 t It6'9 969*5 st i ti SZc*K 9t9 u il 9L*t tt: li i x 9111 ti cre9 999°C t 0 El 0gf"l OgttC 20 1 Kl 66919 691*5 tl 9 S O6L*o tviLo K¥ z 10 000 000 K0 Kl 99t*9 99~'9 19 91 En wolf9 Ko OOtII 061'9 3W Kø 00*It 06*9 to t KO 000*0 00010 20 0 TO 0000 0000 to 0 20 000-0 0000 re 9 Ko 000* 00 to 0 tEiu -tegl8 t9 4Lto t9" 9911ZZ Tv gL K " 1 1*I9i99Ktl i K Z 9e*51 996161 3 61 Te SOIT 9*9 KL 91 at 9'01 KrK KL 1t (99*1 Z99gt It c tuI go9 6i 9ttt C59'9 91 c9 ig eft'69 L9o9 sCc (KL gtt K '9Pst 996*Ki xC9 tIt "E9 69191 tit WC "9Z 1c6190 96919 109 99 Ett 6cC*9 no-Otn tit EL [A ~** w WS°CE 5Z191 t e9 lø CW0I 09t*9 ti 9 1K t 19*9 LS0* %1 9 Kl 919E 9t'1t Kv t Ei 9t91 99Z gr IE f99't 999°8 gø 9E A got9 KOT 590K Mrtog 6I ugt 1 I P9Cze 6161c9 sct K 91 £i silSit 19510c1 Izz 01 i Zzol9 £991 101 l Lt OE'i 9lc 9 91 5 EK 9u0' 966*9 T9 is At Go¥9** ZIC 690-II I9c16011 ZKt 9 £t IL e615 11X*Ug Kig UKl KtK Et'9I i5-99L ug (6 199 9t" r'ett l0-CE iK 0 199 99'occ sz9'9M Et 9 x* c1t iti' 0K 69 u til90*t KL OOISEZ* 5c96 Ez Z9 IC 0006 OSV99 *t I E tg 0C9c"f 519L9 il 9 T 009Lt OOV6 12 5 11 S16IC 09W11 o E0 0*0 OTIO 19 9" 11 -iføl ti 9»l< ItSKEW 19 C£ al l pKt 1s t IC IC St tiK't 09i'9 Ks Ot KZ 009°g øv'9 KE 9 K0 Z[000 s005 to O OO 9SE'* 9g5o "ro oTt g I-l 101 9191*9IES5 ~9919 3001 el 3661 9V096r 999*t9Ei Ks Etz UDI t,"990f 090*96< zool 695 1001 9'l 9 I5O'Tit 00c OL 100! 960*GI9 00efgK 2001 69£ 100 9ool09 99*9£t o0el £9V tet01 Ew 09'it 9969 t i ¥ 0 000*0 00l go 0 UK 09t 996*91 z i ES w000,0 000 to 0 K 000* 00010 to 0 go 0000 00o K 0 utsu KL 36c111 CfE'99 zL i c 1 %a 19 ur99 t, i c0 0000 0000 lo 0 0000 30 0 10 00- 00-0 & 0 tg 0000 000o zo 5 9 190 fp2119138s i (919 60tsi 45*55 1 999 1i tc'9t 19 01 « csg'o 99U9 191 06 il 09'15 Es99 il 95 t 609,9 569*5 tt %t %o 591*1 191,1 2C9 912 -10*9 TWJWT 95E 15"t1CI 690,6101 Et £99 t 6U5K99 596'Z1 19 L6 Kti 099109 1f5191t Z1 9K EK 6t 51 69t*91 t0 O 196 COVM o09tle KIL (9 ti 0' 999O6 e t !Co" 332q13 nL£ 16991t 90ME"C6« 1" O %CL 9t91t tig*9991 19 il "t5 000519 UVl'99 Z9 Z2 199 L06*olc 0c®91 MK SL 16£ 99t9K 9 yct' 39K 06 166 C9C'L6 4CLt' Et Tic Mt t13203 KK9ø ti'9 816*1 toilt tlt £*lK *99ev 1001 991«9Cs 1909919 ~ 1 notK ti 9tt096t 9091t S001 Ltt Kg0K 595'9t 0 090*96 100 69K 001 19ZOO99 Y0hi m0l I1* 1001iselstsO 960l9 e0lg1 00" £91 K0e t19't9 99&09Le 100K £9» =etM wS (999 90'9 KS t 10 999 91 Kl 9 10 95-0 9990 0 I K 091 991"t Ko I to 00t! 09l0 Ki i z 000'0 00,0 IS 0 8a£C **A0 69 9g95 090669 99 rø9 99 ILC9I (r9091 199 KIK 5C s 999t 6£C K*999 T KK E99 990c O*t9" KIt 991 l91 oIL £1909 lo 99 to 099*0 9tt0 to i e cl -I KIK MSttz 1t1691 K 9ur M 1ti£ "t00t W916c9 2l 1tt19 19M99 I'CU 'il 9 K95 (91i 6'E11 19 99 910cs9 tILt6o t69 L 00p z* 5O009 9"slg-) 2$0t 99 Tog Pe0 309e 1 100* 9t9it (959919 Z001 tiOT 1661 9E<'96* 9IieKE m00 119 KOST zc 59501 09006L 00t 695 001 19(099 tsltit lOOK It 2001 960°oe9 00t'it o00£ £9W 100i L t91t9 99L9Lt 100K £~l neso 6K 95-691 99'Ci 359 96 i f#t" 599S01 it itt z09 to9 9 5nt I 90 L 6u9r6 6t1 oK 99 K1og is III 969*Ter K9 9 Kf 1001 9ti*09 9Lzoo] KO0K 991 31UG[« 3p623 o0pn6 Z»9 9911ot M919111 (5991 Z9 Stt Ett f9I'69C cirl9 Mtt «I? 199 9t'iltZ £661t91 Ti OL KOK 619"16 9ff'OL tit IC 10 0990 9Ule KS i 1ev16~ t10 6 LK99 S*Ott 191 E 9-IN'Stt 015*sit ZI: 96 15 66lic (96 19u 99 11 i50l Et9" TI OL Ke 996,I 196 Iu 9 to 0 0 TØ 0 (I%) 1V1T15d3 In208o pe 9i pm~q tt 1t*96"% 19"199 101 9MK &LL 56616 * 191*909 I 06 It 9111tt E1'9L1 1 9X Wlt tsil9 519"rg 191 9 1 OtC9 969'9 19 9 20 0 0 TO 0 M0 932) 993 9 0w0111 1" S a 99 a 9Møfflø 019 z 'Om s 1Mo1"s 018 9 *Oml Z ~oqngS1 1 'Og t ~so 1as t "01 etng j18 •ag v 10 6961 1D 89613 £961 D 96K 10 tot 13 <99996 901111. U. I29p.19 6861 lit jaqU~0 jo sy SIK~VAV NVO' W A0 SaLLSI3lVVH~ Z XRNNV - 31 - ANMX 3a ALF INTEREST RATES 19 - 1990 (in percent) CBP End-ussre Lending Rate Time Deposit Year Ouarter ALF Base Cost Pased-on Cost Cost (Secured Loans) I a te W 1965 3rd 8.82 16.70 n.4. 25.307 17.041 1985 4th 8.82 16.70!V 22.01 17.825 15.456 1968 let 8.50 13.20 20.21 24.487 20.223 1966 2nd 6.50 13.20 19.95 17.074 15.473 1966 3rd 8.23 12.80 17.94 16.238 12.442 196 4th 8.23 11.31 18.67 13.923 9.486 1987 1st 7.92 11.00 17.95 12.335 8.639 1987 2nd 7.92 10.20 16.04 13.140 10.285 1887 3rd 7.76 10.20 17.04 13.031 9.926 1987 4th 7.76 10.00 15.19 16.310 11.148 1988 1st 7.72 10.00 n.a. 15.230 12.137 1966 2nd 7.72 10.00 17.00 16.532 13.893 1988 3rd 7.59 10.0 16.48 16.401 14.104 1988 4th 7.59 10.70 16.62 16.566 13.747 1989 let 7.65 10.70 16.50 16.639 14.626 1989 2nd 7.65 10.70 16.30 18.309 14.109 1989 3rd 7.73 10.80 16.28 20.312 18.020 1989 4th 7.73 12.50 16.28 23.768 18.591 1990 1st 7.75 12.509 n.a. 22.163 18.196 1990 2nd 7.75 13.000 n.a. 22.824 21.192 1990 3rd 7.70 13.00S n.. 25.307 21.669 1990 4th 7.70 14.00W n.. 27.226 23.258 Notes: Until December 8. V Figures from 1985 to 1987 are weighted averages of rates in a quarter. Figures from 198 to 1989 are cumulative weighted averages of rates from 1985. SGV computation: CB rate to PFI + CB-computed weighted average spread of PFl. The LBP adds 2 percentage points to this rate. Weighted average of all maturities. Source: Central Bank of the Philippines. - 32 - NNE 3b CURRENCY DEVALUATION, INFLATION AND INTEREST RATES UNDER ALF I 19 1987 1966 19 1990 5-Year I Average Po vs US Exchange Rde Start Year 19.03 20.53 20.80 21.34 22.45 End Year 20.53 20.80 21.34 22.45 28.00 A. Devaluation Rate of Peace vs US$ 7.9% 1 1.3% 2.6% 5.2% 24.7% 8.0% US Dolar ve SOR Start Year 1.12 1.25 1.40 1.33 1.30 End Year 1.25 1.40 1.33 1.30 1.43 B. International Devaluation Rate 11.6% 12.0% *5.0% -2.3% 10.0% 5.0% C. Tolal Devaluation Efect We) 20.4% 18.6% -2.5% 2.8% 372% 13.4% Philippine inflation Between Year 0.8% 3.8% 8.8% 10.6% 14.2% D. Within Year (Calculated) 2.3% 6.3% 9.7% 12.4% 14.2% 8.9% E. Average Cost of Borrowed $V to CBP jf 8.4% 7.8% 7.7% 7.7% 7.4% 7.8% F. Passed on Rate by CBP 12.6% 10.4% 10.2% 11.4% 13.1% 11.5% G. Estimated End User Cost 17.9% 16.6% 16.5% 17.0% 19.0% 17.4% H. Eallmated CBP Margin (F..Cq -16.1% -11.0% 5.1% 0.9% -1.A% 4.7% 1. Time Deposit Rate 14.4% 10.0% 13.5% 16.3% 21.1% 15.1% J. Bene to PFI (i-F) I 1.8% -0.4% 33% 4.9% 8.0% 38% K. Lending Rate (Secured Loans) 17.9% 15.5% 16.2% 19.8% 24.4% 18.7% L Benlt to Sub-bonowees (1G) f 0.0% -1.1% -0.3% 2.8% 5.4% 1.3% M. Real Iliered Rol at CBP Lending Level (F0) 10.1% 38% 0.4% 0.9% -1.0% 2.4% N. Real Interet Rate at PFI Lending Level (P-0) 15.8% 9.6% 8.2% 4.1% 4.2% 7.8% '/ The Borrowed 'V Is the equivalent currency borrowed by CBP from the World Bank under this Loan. The exchange rate between the borrowed W and the US$ Is that used by the World Bank for capital repaid on 15 January 1991. it Underestimates total benefit compared with time deposits because () there is no reserve requirement for ALF, but there Is for time deposits; and (II) len moblisatlion costs are Incurred. Together, these two items would have averaged about 3%. I True benefit Is probably higher as agricultural Investments are often rated as riskier than average, and so ALF borrowers would not on average have been able to access funds at the reported lending rate for secured loans. Source: U.SAID/SGV Impact Evaluation of the Rural Finance Services Project, March 1991. THE WORLD BNK / IFC / MIGA - 3 OFFICE MEMORANDUM DATE: June 1, 1992 ANNEX 4 TO: Ms. Donna Dowsett-Coirolo, Division Chief (EAlAG) FROM: Rolando Arrivillaga, Principal Operations Officer (EAlAG) EXTENSION: 81351 SUBJECT: PhiliDines: Agricultural Credit Project (Ln. 2570-PH) - PPAR 1. I read with great interest the PPAR for the Agricultural Credit Project. The report is well written and uses extensively the data available, but I am not persuaded that it fairly assesses some important aspects of project performance. This project was innovative and pioneered a new approach to agricultural credit--way ahead of the "Levy" report--in a difficult political and economic environment. We need to learn from this experience, its achievements and its shortcomings, but the PPAR does not do this well. 2. We should urge OED to further review the paper before finalizing it. In particular the report should focus on the primary project objectives, and more is needed to be said to set this project in the proper context--the reader should understand the environment that led to this operation. General Comments - Evaluation Methodology and the Project Context 3. The most disturbing feature of the report is that it evaluates performance against a set of objectives that the project was not designed to meet. For example, the draft PPAR criticizes the project because "participation by smaller commercial farmers was minimal, and the line of refinance came to be dominated by the KB's and the traditional commercial clients." The main purpose of this project was not to deal with the credit problems of the rural poor, or even of what the report calls the strata of smaller commercial farmers, but rather to establish a viable institutional framework for agricultural credit, at a time when the banking system in the Philippines was in complete disarray. InerE alia, this meant transforming the traditional supply-led credit arrangements to a demand-led approach. To assess whether this was a proper objective depends on an understanding of the main differences between the two approaches--their costs and benefits. 4. Proiect Context. Let me first explain the context in which the Bank considered the project, since this is not clearly set out in the PPAR. At the time, we were concerned that directed credit programs and excessive government intervention in the financial market threatened the stability of the Philippine economy. The Central Bank's rediscount operations had destroyed the country's institutional capability to channel financial resources for development purposes and the PNB, DBP, the Republic Planters Bank, the rural Banks and, as it turns out now, even the Central Bank were bankrupt because of the losses incurred in financing of directed credit programs. These programs were used mainly as instruments to gain political support at electon time. The non-performing loans of the two development banks were among the main sources of the external debt problem and the Dowsett-Coirolo - 34 - June 1, 1992 government would have to assume liabilities in excess of 7130 billion (about $7 to $10 billion depending on the exchange rate used). Since then, the government was unable to meet its external payment obligations, it declared a debt moratorium, the economy was in collapse and development prospects were very bleak. The damage was not limited to development finance institutions, as a large number of commercial banks were also destroyed. In this context, this issue was not how to put more credit out to a particular segment of society, but how to rebuild a more resilient and robust financial system. 5. The challenge was to develop a market-based system with only a minimum of government interference, which we felt had previously been too perverse. The government had overemphasized retail operations without paying enough attention to resource mobilization and term transformation. It needed, instead, to restrict its activities to the wholesale level, leaving retail operations primarily to private financial institutions or to public financial intermediaries capable of surviving in a market-based financial sector. It was also essential to stop the hemorrhage of public financial resources used to subsidize ineffective credit programs. The costs, both of the subsidies and the related defaults, meant that the Government lacked the finance for urgent investment requirements in the social and other sectors. We, therefore, decided to establish a system, eligibility criteria, and procedures to limit the government's exposure and risk. The resources were to be transferred to financial intermediaries at about market rates of interest. Banks were allowed to set their own spreads, so that they would not be deterred from lending to agriculture because of the perceived high risks and transaction costs. We understood that the retail network, particularly the institutions that could reach small farmers, had been destroyed and therefore never claimed in the appraisal report that this operation would directly reach many small farmers or even smaller commercial farmers. 6. Given this setting and the objectives set at appraisal, project performance should be assessed in terms of whether the government is consistently and coherently fulfilling its role as a wholesaler of funds, and whether it is mobilizing resources effectively for development purposes. From the fiscal and monetary point of view, a comparison should be made of how the current system compares to its predecessor (i.e. how successful were the reforms supported through the project in reducing fiscal or Central Bank losses). The report should also assess, from both a qualitative and quantitative point of view, the nature of the governments exposure under the new approach. Were the eligibility criteria too soft or too strict- -how much risk did the public sector assume? Finally, the report should assess whether or not an unrestrained spread is sufficient to induce bankers to reach out to smaller farmers. If not, what alternative measures could have been adopted at the design stage to reach this particular clientele more effectively. In sum, the report needs to be more balanced. It dwells too much on the perceived, although not fully demonstrated, shortcomings but it does not give enough credit to its achievements under extremely difficult circumstances. Dowsett-Coirolo - 35 - June 1, 1992 Specific Comments 7. Besides criticizing the project for not reaching any of the strata of smaller farmers, the draft PPAR also faults the project for not engaging the network of rural banks, for supposedly concentrating credit resources among too few large corporations, for adopting a faulty interest rate formula, for alleged substitution of credit sources, and for inadequate resource mobilization. Finally, the report assumes a role for USAID which that agency never exercised, and it exaggerates the impact of the land reform credit program on project objectives. 8. Small Farmers. Although at one point the author acknowledges that the project was not designed to reach small farmers, he devotes so much attention to the subject that most readers would forget the caveat and assume (erroneously) that the project failed in meeting a basic objective. At the Decision Meeting for this project, we all agreed that "the rationale for the project is the policy and institutional reforms which would put agricultural credit on a sound and more viable footing." 9. Not only was the main thrust of the project towards policy reform, but also the circumstances were not propitious for reaching small farmers. First, as noted above, the retail banking network was in collapse. Second, under a demand-led approach, and in contrast to directed credit programs, we did not seek to force financial institutions to direct their resources to any particular clientele. We considered that if the institutions were assuming the financial risk, they had the right to choose their clients. We sought instead to deal with the issue by not putting a cap on interest rates, in the hope that this would encourage participating banks to take greater risks. Because of the problems with the retail network, the high transactiori costs and the higher risks, we also sought to reach small farmers through non- traditional or indirect channels. Hence, we encouraged the idea of channeling resources through dealers of goods and services and agro- industrial enterprises, particularly those promoting contract farming. 10. The main lesson here is that to reach smaller farmers, it is necessary to create the conditions for viable institutions to have an incentive to reach out to this clientele. Under the conditions at the time this was unlikely to occur. The banks almost all were restructuring their portfolios and taking losses; they were not in a position to expand their activities to smaller clients. In fact, we were concerned that the collapse of the financial system would inflict further damage on the economy by financially starving the commercial agricultural sector. In particular, we thought that a further deterioration of the sugar industry, a major source of employment and incomes, would have a devastating effect on the rural economy and the country as a whole. Therefore, I do not find it surprising that many of the activities financed under the loan were related to the rehabilitation/restructuring of the sugar industry. 11. Rural Banks. I do not understand why the author goes to great lengths to fault the project for not reaching rural banks more actively. First, there is a question whether we should use the rural bank necwork as an important vehicle for channeling investment and production credit to the sector, and secondly, whether it was possible under the conditions prevailing at the time. On the first point, I strongly disagree with the Dowsett-Coirolo - 36 - June 1, 1992 notion that rural banks should have been an important vehicle for channelling agricultural credit. Rural banks should have a diversified portfolio and their activities should spread across a broad spectrum of the rural economy-- otherwise they cannot be viable in a country prone to natural calamities and typhoons. This point was even underscored in an earlier OED review of the Bank's Agricultural and Rural Development Program in the Philippines (Report No. 3796, of February 1982). In reference to rural credit operations in the Philippines, the report stated: "However, the impact of these credit projects compounded the arrears and the overall financial condition of the rural banks. Rural Banks had been selected for the four credit projects primarily because they are widely distributed in the rural areas. But they turned out to be unwilling participants in the credit schemes. The projects called for medium- and long-term loans, for which they were not financially geared and on which their profit margin was lower. By the fourth credit project, the increase in rural bank resources was accounted for entirely by Central rediscounts as both their capital and reserves had declined substantially. The rural banks' problems did not arise from Bank supported projects alone; much of the aggravated arrears position resulted from the M-99 program." [Perhaps the most (in)famous directed credit program of the government]. 12. By the time we made the loan, many of the rural banks were in serious financial distress. We could have reached a large number of them only if we had lowered the eligibility criteria used to qualify participating institutions. The appraisal mission did recommend very generous criteria, but at the yellow cover review we decided to be more rigorous. By insisting that no participating banks should have arrears above 35% (instead of 50% as proposed by the appraisal mission), we disqualified the bulk of the rural banks. In retrospect, this was a wise decision because many of the banks that might have qualified under more accommodating criteria have since gone into bankruptcy. We did, however, agree to qualify banks below the cutoff if they had formulated acceptable rehabilitation programs. We considered that this was a sufficient risk, but that there was no justification for channelling resources to institutions that consistently kept their arrears above 35% of their portfolio and showed no willingness to address their financial problems. 13. The report also does not give enough credit to the Bank for its efforts to rehabilitate the retail network. Under other operations, and as a result of the work carried out under this project, the Bank helped rehabilitate DBP and PNB and this was a real success. As to whether the Bank should tackle all problems at the same time, I believe that we have learned over the years to focus our efforts on well defiaed goals. The two major development finance institutions had priority, as did the general policy and institutional framework for agricultural credit, but we did not neglect the rural banks and are still working towards their rehabilitation. 14. Sub-loan Concentration. This operation was designed to finance both agro-industrial and production credit, and it would have been an error to superimpose on the market decisions about how credit should be allocated among competing activities. It is also important to recognize, as we did, Dowsett-Coirolo - 37 - June 1, 1992 that some of the main corporate entities have greater potential to reach farmers more effectively than did the directed credit programs. In an Agricultural Credit Sector Review that provided the background for this operation, we identified an urgent need to find alternative modes to reach farmers, such as through contract farming and suppliers credits (subsequently we also started to promote creditworthiness enhancing schemes like guarantees and crop insurance, bonded warehouses, etc.). We, therefore, encouraged the Central Bank to promote the line of credit among critical agro-industrial ventures. In fact, one of the corporations cited in the PPAR as an example of what should not be done is a major proponent of contract farming schemes that have now proven to be viable alternatives to conventional credit in the Philippines. This particular food corporation has pioneered the development of poultry through contract farming. Its efforts, although motivated by profit, should be lauded. Another corporation cited in the report is one of the main sugar centrals, which is now jointly owned by land reform beneficiaries and land owners. The loans to this corporation helped sustain its viability in the transitional phase when the sugar market seemed to be collapsing. I believe the Bank should not refrain from fostering the development of agro-industrial activities in a country just because the sub- loans may be large. For an economy the size of the Philippines, loans for agro-industrial activities will naturally be quite large compared with production credits. The important point is that the lending decisions were made by the participating financial institutions, which have strict single borrower limits and were under close surveillance by the Central Bank, and that the activities financed had important backward linkages to the rural economy. 15. Interest Rate Formula. The PPAR concludes that the interest rate formula was faulty because the rate lagged when inflation accelerated/decelerated rapidly. I do not want to debate this point, but would like to note the reasons for this formulation. Before this project, interest rates had been highly subsidized, in both financial and economic terms. We sought to induce the financial authorities to define a sound principle to price loans: the rates for participating financial institutions should move along with the cost of money. The system was not yet ready to price loans on the basis of variable rates, as would be desirable in an inflationary environment. We, therefore, accepted a compromise because we thought that the government was committed to price stability and there was no evidence that inflation would become chronic in the Philippines. Historical data indicated that from time to time inflation accelerated, but for the most part this reflected adjustments to external shocks and was not due to fundamental structural imbalances. During project implementation, inflation rose fast at the end of 1990 as a result of the Gulf War (the government devalued the currency sharply and increased oil prices to adjust to the external shock), but then decelerated in 1991 in response to the contraction induced by fiscal and monetary policies. By the end of 1991, inflation had already decelerated to a single digit level, where it stands now. We must not, therefore, conclude that the formula was completely faulty. 16. The PPAR also argues that the Central Bank was refinancing at or close to negative rates, which is not factually or methodologically correct. The author arrived at this conclusion by using point-to-point price deflators instead of annual averages. Point-to-point changes are misleading because of the sharp increases at a particular moment in time--for example the change Dowsett-Coirolo - 38 - June 1, 1992 from November 1989 to November 1990 was high because prices changed dramatically during the last quarter of 1990 due to the Gulf War; the changes are not so dramatic when the averages are used for the same period. Finally, the formula should not necessarily be taken too much to task just because it yielded rates that seem to be slightly below the market under certain conditions. What we consider to be the market rate is severely distorted by the debt overhang (a large volume of Treasury Bills being placed in a thin market), indirect taxes (the gross receipts tax on financial transactions), the high level of reserve requirements, and the large volume of financing transactions that take place outside the formal system (trust funds). The market rate in the Philippines is not necessarily a point, but possibly a scatter of points along a range. 17. Credit Substitution and Incrementality. On a related matter, arguments made in the report to the effect that there was little incrementality as a result of the loan are debatable. During most of the period covered by the loan, the Philippines had in place tightly constrained fiscal and monetary programs agreed with the IMF. In working out the credit ceilings, the Government and the IMF factored in the external resources made available under the project. As a result, more resources were included in the monetary program to accommodate the credit needs of private sector activity, including those of agriculture. Indeed, the data for the period 1986-1990 shows that economic activity expanded rapidly--the agro-industrial sector was particularly dynamic during this period. 18. Resource Mobilization. The report argues that the impact of the project on resource mobilization was not significant. In formulating this operation, we postulated that resource mobilization would only improve under a market determined system of interest rates, although we also understood that this would be a necessary but not a sufficient condition to achieve this end. Other conditions must also exist: e.g. political stability, sound foreign exchange and fiscal policies. But not all of these conditions were met during loan implementation (the country experienced a series of attempted coups; the external debt overhang imposed a serious burden on the economy; and there were several major shocks in the form of earthquakes, typhoons, and a middle-east war). It is unwarranted to call the project to task because all of the conditions for improved resource mobilization were not in place at the same time. Nevertheless, there is evidence that resource mobilization improved significantly compared to the period prior to the project. Real rates of interest rose sharply and remained high, and this certainly fostered higher levels of financial savings. 19. The Role of USAID. The PPAR assumes that USAID played a key role in project design, which is factually incorrect. The truth of the matter is that USAID first indicated (informally) its interest in participating in the operation during the very last days of appraisal, when it had some funds that it was able to commit before the end of its fiscal year. It officially confirmed its interest only after practically all the key decisions had been made by the Bank. We welcomed USAID's participation, and encouraged them to proceed on a parallel basis. At one point (close to negotiations) USAID did query our decision to raise the eligibility criteria and expressed its concern about the possible disqualification of non-performing rural banks. Since there was no cross-conditionality, USAID could have done as it pleased Dowsett-Coirolo - 39 - June 1, 1992 at that time, but presumably they were also convinced of the risks because they adopted the same criteria. 20. USAID played a more prominent role during project implementation, particularly through its support for the technical assistance program. USAID agreed to finance the studies and to monitor closely all the TA, and its field presence meant that it had a comparative advantage in doing so. We consulted periodically on all aspects of the project but, again, this occurred mainly at the implementation phase. It is thus incorrect to argue now that USAID could have made a meaningful contribution prior to Board Presentation; it simply came in too late in the process. 21. Land Reform. The author argues that the land reform credit program dealt a severe blow to the policy reforms advocated under the project, which is quite an exaggeration. Although the Land Bank makes available resources to farmer groups at a predetermined rate, the program reflects many features of the principles which we introduced. First, the rate is not as far apart from the market as it might seem at first glance: the Land Bank has quite rightly sought to reduce transaction costs to beneficiaries by organizing them into groups, which it treats as quasi-participating institutions (so the comparison should be with the ALF's rates to PFIs, not those to the final borrowers). Second, to reduce risks, the Land Bank requires that loans be covered by crop insurance, and borrowers pay for this cost. Third, the farmer groups are free to determine the spreads that they need to cover their costs. And fourth, the credit is financed with the resources of the agrarian reform program and any subsidies are totally transparent- -the Government and the Bank had a clear understanding that if there were to be any subsidies, these had to be transparent and should not affect the health of the financial intermediary. The final interest rate to the end-Borrower, although lower than those under the ALF (or our follow-up loan), was not very far below the market rate when the rate of inflation was not too high (e.g. in the teens), and certainly not when inflation is at a single digit level as it was when the program started, is at present and prevailed during most of the period covered by the loan. More, important, in the aggregate this credit program is a relatively small (perhaps 5%) share total lending to agriculture by the banking system. The lion's share of total lending takes place at market rates, as promoted by the project; and the proliferation of directed credit lines was done away with and has not been re-emerged, so it is incorrect to suggest that the relatively small agrarian reform credit program may have dealt a fundamental blow to project policy reforms. - 40 - B"go Senhatny iPd#nas ANNEX Sa (CENTRAL BANK OP THE PHILIPPINES) MAYNILA, PILIPINAS. OFFICE OF THE DEPUTY GOVERNOR RESEARCH SECTOR. DOMESTIC OPERATIONS SECTOR AND INTERNATIONAL OPERATIONS SECTOR 14 May 1992 Mr. Graham Donaldson Chief Operations Evaluation Department The World Bank 1818 H Street. N. W. Washington, D. C. 20433 U S A Dear Mr. Donaldson: Thank you for your letter dated April 10 , 1992 transmitting a draft of the Proiect Performance Audit Reports (PPAR) on the Agricultural Credit Proiect (Loan No. 2570-PH)., for our comments. The PPAR, appears to provide a comprehensive assessment* and evaluation of operations of the Agricultural Loan Fund Program. Nonetheless, allow us to summarize below our comments/observations which we hope will afford clarity to certain minor items in the report. 1. Paraaraoh 1.18. With the concurrence of the World Bank (WB), the ALF interest rate is reviewed every auarter* instead of semi-annually because of the volatile nature of interest rates, especially during the early period of proiect implementation. The frequency of review has effectively ensured the alignment . of the ALF interest rate with current market rates. 2. Paraoraoh 2.2. There were actually 49*(not 41) PFIs that participated in the ALF program as follows: 13 rural banks (excludino 12 banks involved in retroactive financino in 1985). 21 commercial banks (including LBP), 13 thrift banks, 1 specialized government bank (DBP)? and I non-bank financial intermediary (PDCP). 3. PargaraDh 2.7. In mid-1991, the ALF interest rate formula was changed from the oricinal formula of WAIR of the banking system's demand, savings and time deposits, to Weighted Average Interest Rate (WAIR) of time deposits of 180 days in order to keep the ALF rate -close to market rates. Effective July 1. 1992. a unified formula in calculating the interest for both the ALF and CLF of LBP will be adopted. * Indicates comments that have been reflected in the audit report, either by correction or footnote. "ANG UTANG DAPAT BAYARAN NANG TAYO'Y PAGKATIWALAAN"' 1 - 41 - 4. Paraoragh 2.11. The statement that differences in terms of service between the CBP and LBP "help explain' why few.staff elected to transfer" may not be totally accurate. The imolicit reluctance of the LBP. due to possible employees' union obiection. in taking the ALFU become evident during the negotiation of the ALF transfer. Thusj in the Memorandum of Agreement effecting the ALF transfer from the CS to LBP, the latter saw to it that a provision was incorporated in the' agreement to the effect that the transfer of the ALFU staff from the CB to LBP Aould be optional. 5. Paraorach 2.14. The statement-that Phase (2) was not assigned to and not carried 'out by the ALFU is incorrectA Phase (2) which called for the evaluation of the impact of ALF subloans at the levels'of both the participating farms and the PFI was undertaken by the ALFU on a regular basis up to 1989. This activity was a maior component of the end-use verification survey carried out by the ALFU. The survey covered the verification of the utilization of ALF loan proceeds and the evaluation of the financial and economic performance of proiects financed. Verification compared the planned and the actual results of operations. the term lendino transformation of PFIs! institution building, among others. Later on this task was done occasionally due to the diminished capability of the ALFU brought about by the substantial depletion of its Staff as a result of the imminent transfer of the ALF to the LBP.. 6. Paracrach 3.1. The words "Minister and" in the last sentence should be deleted.* 7. Paraorach 3.2. The acronym CPC at the beginning of the paragraph should be changed to ACPC.* 8. Paraoraoh 3.5. The acronym LDB in line 7 should be changed to LBP.* 9. Paraaraoh 3.10.*' While the extent to which the ALF helped mobilize deposits is difficult to determine. nonetheless there is sufficient indication that decosit liabilities of the banking system grew tremendously by 119% from P164,256 million in 1985 to P359,017 million in 1990 or an average annual rate of increase of 17%. Commercial banks deposits. which constituted 87% of the banking system's deposit liabilities, increased substantially by 117% from P144,059 million in 1985 to P312,979 million in 1990; this*represented an average - annual increase of 19%. It may be underscored that the 27 - 42 - deposit mobilization of commercial banks is carried out on a auid pro quo basis and linked up with their lendino activities. An Account Profitability Analysis (APA) in conducted for every account before any financial assistance is extended. An important feature of the APA relates to deposit mobilization. Thus. for every ALF loan released by a commercial bank, a corresponding amount of deposit is generated. This relationship is not conspicuous in the case of rural banks. 10. Paraorach 3.14.* The fact of additionality.in the ALF may be gleaned from the rise in the loan portfolio of the banking system, especially among commercial banks. The banking system's loan portfolio expanded significantly by 66% from P175,236 million-in 1985 to R290.563 million in 1990, reflecting an average annual arowth rate of 13%. Since ALF loan disbursements to commercial banks amounted tu P2,958.2 million during the oeriod. it may be safe to conclude that the ALF played a significant role in substantially increasing the loan portfolio of commercial banks. 11. Paraoraoh 4.6. The assumption that if the RB rehabilitation program had moved faster, the entry of these banks into the ALF would have made for a more balanced portfolio and a higher incidence of incrementality. seems to us to be flawed by the possible reluctance of the rehabilitated rural banks in availing themselves of the ALF.* considering their quite auspicious performance in the previous CB-IBRD rural credit proiects. Of the outstanding loan portfolio of 9409.69 million of the 294 rural banks that took -part in these credit proiects. 9350.66 million, or a high 86%, remained past due as of February 29, 1992. Again, thank you and regards. Very truly yours, EDDO P. ZIALCITA Deputy Governor - 43 - LANDBANK ANNEX 5b May 15, 1992 MR. GRAHAM DONALDSON Chief, Agriculture & Human Dev*t Division Operations Evaluation Department International Bank For Reconstruction And Development 1818 H. Street, N.W. Washington D.C. U.S.A. 20433 SUBJECT : AGRICULTURAL CREDIT PROJECT (LN. 2570-PH) - PROJECT PERFORMANCE AUDIT REPORT (PPAR) Dear Mr. Donaldson: Thank you for the PPAR on the Agricultural Credit Project. The report gave useful pointers in our continuing administration of the project's credit component, the Agricultural Loan Fund (ALF), and the implementation of the follow-up project, the Rural Finance Project (RFP). As stated in your report, project results indicated that the ALF had undershot some of its objectives, viz: 1. ALF funds were concentrated on large farm clientele. 2. Commercial Banks (KBs) dominated the loan portfolio while Rural Banks' (RBs) participation declined. 3. Additionality in ALF lending is lower than expected. I wish to emphasize the results of LandBank's ALF relending operations. Gross releases since July 1990, date when the ALF was transferred from CB to LandBank, is P1,261,631,471. Availments by small and medium enterprises (SMEs) amounted to P823,898,971 or 65%. Thrift Banks (TBs) and Rural Banks shared 43% or P546,224,746 of the total funds relent. Forty four percent (44%) or P556,229,450 went to finance project expansion with term of more than one (1) year maturities. Loan size of P5.0 million and below aggregated 54% of total availments. Only 8% of an over P20.0 million loan size was released. When CB turned over the ALF loan portfolio to LandBank, only nineteen (19) Rural Financial Institutions (RFIs), composed LAND BANK OF THE PHILIPPINES 319 Sen. Gil J. Puyat Ave. Ext. P.O. Box 1311 Makati, Metro Manila Cable *LANDBANK" Tel. Nos. 818-9411 to 21 - 44 - of RBs and TBs, were recorded. As of December 31, 1991, there were 104 RFIs eligible to participate in the ALF Program. These figures are a strong showing that LandBank's disbursements of the ALF fund reflows are not targeted towards the large farm clientele of the KBs. A serious attempt has been made to come close to the project's objectives. With the new design of the Rural Finance Project, LandBank is prepared to accomplish the objectives of a) RB reform, b) spread of funds to a broader base pf subborrowers and, c) development of countryside economy thru lending to sub- projects with incremental impact and additionality. The eventual merging of the ALF and the Countryside Loan Fund (CLF) is being evaluated cautiously by us and CB particularly the pricing mechanism and the PFI eligibility criteria. To give you an update of the Countryside Financial Institutions Enhancement Program (CFI Enhance), ten RFIs have an approved capital build-up program and 143 applications are presently being processed. The program also aims to bring more RFIs into CLF eligibility. As of April 30, 1992, there are 38 accredited RFIs. On the finding that our ARF rediscounting program has compromised World Bank's credit reform policy position, I would like to inform you that LandBank is moving towards market-oriented pricing by gradually removing the subsidies. From the 12% in,.itial pass-on rate to RFIs, we are now lending at 16% (please take note that CLF pass-on rate for this quarter is 15.47). However, I still believe that the ARF lending program can co-exist with the CLF Program. The former shall target the "real" small farmers, the intended beneficiaries of the country's agrarian reform program, who would otherwise not fit into the CLF Program. I am fully aware of World Bank's project design that eliminates preferential targeting, adheres to market based credit pricing, improves rural credit institutions and stimulates expansion of countryside economy. These are similarly our goals in the Rural Finance Project whose peformance is being closely monitored by our institution. ry tr y yo s, E US F. DIAZ Ac ing P esid t /dmm IBRD 22431 CLASSIFICATION OF PROVINCES BY GEOGRAPHICAL REGIONS l IOCOS VI WESTERN VISAYAS PHIL PPIN ES 2Ila 1 Iocbm Norte 38 Aklan am, 2 ~locos Sur 39 Capic La Union 40 AfiqUO BATANES 4 ~ InU108 41 110110 CORDILLERA ADMINISTRATIVE 42 Ngra OccidntlI REGION (CAR) 43 GulmarPa 5 Abra VII CENTRAL VISAVAS o Province Capitalm 6 KaIsaAayao 44 C 7 iIrvM 0 45 National Capital b I~ 46 e9 gnill 47 r Province Boundarles CAGAYAN VAL,fY IISTRN VIGAYA5 i oC~ 4a NouSemr 6 - Region Boundarles 2 iCabm 49 WNUsn Smflr 12 NlaVsy 50 Satern ~ R T,, international Boundaries 13 NUIRVcyW YC sy ~vi T"u< l 14 Quirlfo 52 Sothern ~Lyt ll CENTRAL LUZON 54 Iliran 1% Nuva fe Ix WESTERN MINDANAO 12 16 TU 54 zimd Nr 12 1 7 Z~nbaure 5 5 ZamboUngI del IAFLME-M 10 300 300 l a1 5 6 sulian † 1 1- 5 awSulu 13 14 MILIS 1 i 1a0 200 TIONAL NCMTAL NORTHERN MINDNAO ROHN TC 59 sur¥o del Not n ISOLJTHERN TAGALOG 60 Comn 17 21 Auab 6A~ndIleN~ ' 16U1 22 Qua=Ol 62 M~am1Ormå L Z 0 23 RI1l 63 Mk~ Occd 7 ar 9 24 Cavita 64 Duk~id0n 25 LWAIL 6s A~um del Sur3 26 6~ XI SOUTNERN MINDANAO C 3 1 P"1-/P^/E SEA 27 Marlnduque 66 Surigbo del Sur 28 Mindoro Or1naI 67 DaVå Orlentav 29 MIndoro OccMdeta 68 Dau del Norte 26 Q 2 C T D 30 Romblon 69 Davao del Sur ATANDUANES 31 lawan 70 South CorAaiog v 6ICOL X11 CENTRAL MINDANAO a 32 Camarinus Norte 71 Lanio del Norte 33 Canarins Sur 72 Lano del Sur 3 34 Cantanlduns ~ 73 North Cotabato 35 A7ay 74 MuIndao 29 36 SonVn 75 Sultan Kudåret 37 M ~ INb0 at748 vl 49 39 50 SVUT C N v 4 S1 51 SEA O 1 PAC.A E a acsaNE 5 114 31 45 M/aaSe 060 Se/e Sea 2 6 54 isx 65 fi3 11 4 55 18 CHIN X -VI . RA4CfFAC 0 b~ 0Q 1w.. Ix x HILIPPINES 75 SEA vn.san piza i SUL.U ie/ebes sea IN 0° 1A e JUNE 11990