Mexico - Ninth Agricultural Credit Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 12616 PROJECT COMPLETION REPORT MEXICO NINTH AGRICULTURAL CREDIT PROJECT (LOAN 2837-ME) DECEMBER 21, 1993 Country Department II Latin America and Caribbean Regional Office 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 Unit - Old Mexican Peso (Mex$) Exchange Rates: At Appraisal: US$1 = Mex$1,214.00 Average During Implementation: US$1 = Mex$2,388.76 On Loan Closing, June 30, 1992: US$1 = Mex$3,122.30 Fiscal Year January I to December 31 ABBREVIATIONS AND ACRONYMS AGROASEMEX Aseguradora de Mexico para la Agricultura (Agricultural Insurance Company of Mexico) ANAGSA Aseguradora Nacional Agricola (National Agricultural Insurance Company) BANXICO Banco de Mexico (Central Bank of Mexico) BANRURAL Banco Nacional de Credito Rural (National Rural Credit Bank) CPI Consumer Price Index CPP Costo Porcentual Promedio (Average Cost of Funds) FEFA Fondo Especial para Financiamientos Agropecuarios (Special Agricultural Trust Fund) FEGA Fondo de Garantia y Asistencia Tecnica (Technical Assistance and Loan Guarantee Trust Fund) FICART Fideicomiso para Credito en Areas de Riego y de Temporal (Trust Fund for Credit in Irrigated and Rainfed Areas) FIRA Fideicomisos Instituidos en Relaci6n con la Agricultural (Trust Funds for Agriculture) FIRCAVEN Fideicomiso para la Restructuraci6n de la Cartera Vencida (Trust Fund to Restructure Overdue Portfolios) FONDO Fondo de Garantia y Fomento para la Agricultura (Trust Fund for Crop, Livestock, and Poultry Credit) FOPESCA Fondo de Pesca (Fisheries Trust Fund) FSAL Financial Sector Adjustment l oan (Loan 3085-ME) GDP Gross Domestic Product GIRA General Interest Rate Agreement GOM Government of Mexico OP Otros Productores (Other Producers) PBI Productor de Bajos Ingresos (Low-Income Producer) PPAR Project Performance Audit Report PRONASOL Programa Nacional de Solidaridad (National Solidarity Program) SARH Secretaria de Agricultura y Recursos Hidra6licos (Ministry of Agriculture) SHCP Secretaria de Hacienda y Credito P6blico (Ministry of Finance) TORs Terms of Reference Note: Figures in 1991 currency were calculated using the average CPI for flow data and the year-end CPI for stocks. Figures in US dollars were calculated using average exchange rates for flows and year-end exchange rates for stocks. Figures in constant 1991 US dollars were calculated by converting local currency to IJS dollars, then adjusting by the US CPI.  FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Ofrice of Director-General Operations Evaluation December 21, 1993 MEMORANDUM TO THE EXECUTTVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on Mexico Ninth Ag-ricultural Credit Project (Loan 2837-ME) Attached is a copy of the report entitled "Project Completion Report - Mexico - Ninth Agricultural Credit Project (Loan 2837-ME)" prepared by the Agriculture and Natural Resources Operations Division, Country Department II, Latin America and Canbbean Regional Office. Part H was prepared by the Borrower. The project outcome is considered satisfactory, because of significant improvements in Government's agricultural sector and rural credit policies during the period and because Government was able also to ensure the financial integrity of FIRA and FICART, the two second tier agricultural lending agencies. The lending period coincided with the decline of inflation and the early effects of the reforms of the new Mexican administration. According to Part I, the impact of the subloans on farm productivity cannot be clearly demonstrated, because the effects of credit cannot be disentangled from those of other inputs. According to Part H, data from FIRA's monitoring and evaluation system nevertheless show that the income and production objectives on participating farms were achieved. Both positions are correct. The PCR "hesitates" to call the project sustainable, reflecting a lapse in the last year in the dialogue over further reform and a follow-on project Nevertheless, the PCR does suggest that institutional as well as policy improvements to date appear well entrenched. All in all, the Ninth Project demonstrates that even a relatively small Bank financial contnrbution can be effective to help implement important policy reform. Further, institutional development would be facilitated by a continuing Bank-Borrower relationship in the sector. The issues of Bank influence, project impact and sustainability warrant further investigation. An audit is planned. Attachment (hI This document has a restricted distribution and may be used by recipients onty In the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.   FOR OFFICIAL USE ONLY PROJECT COMPLETION REPORT MEXICO NINTH AGRICULTURAL CREDIT PROJECT (Loan 2837-ME) Table of Contents Page No. Preface . ................................................................ i Evaluation Summary .................u...................................... iis Part I Project Review frm Bank's Perspective ................................... 1 1. Project Identity .................................. 1 2. Background ................................................... 1 3. Project Objectives and Description ................................ . 2 4. Project Preparation and Design ................................ . 3 A. Project Design .............................................. 3 B. Key Issues Dicussed During Preparation .......................... 4 5. Project Implementation .................................5 A. Key Issues During Implementation .............................. S B. Variances Between Planned and Actual project Implementation .... .... 8 C. Project Risks ............................................... 9 D. Compliance with Loan Covenants ............................... 9 E. Project Costs, Financing, Disbursement and Procurement .... ......... 9 6. Project Results ....................... ......................... 10 7. Project Sustainability ............................................ 15 8. Institutional Performance ................... ..................... 16 9. Project Relationship ......... .......... .. ....................... 16 10. Project Documentation and Data .............. .. .................. 17 11. Lessons Leamed ....................... ........................ 17 12. Conclusions ................................................... 19 Annex A. Operating Performance of the Executing Agencies .................. 20 Annex B: Financial Data ............................................. 23 Part Il Project Review from Borrower's Perspective .............................. 30 Part ITh Statistica Data .................................................... 43 Table 1: Related Bank Loans ........................................ 43 Table 2: Project Timetable/Project Schedule ............................. 44 Table 3: Cumulative Estimated and Actual Disbursements .................. 46 Table 4: Total Project Costs ......................................... 47 Table 5: Project Costs by Source of Financing ........................... 49 Table 6: Distribution of Loan Resources ............................... 51 Table 7: Direct Benefits ............................................ 52 Table 8: Studies .................................................. 54 Table 9: Status of Loan Covenants .................................... 55 Table 10: Use of Bank Resources ..................................... 57 Map IBRD 19728R 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.   MEXICO NINTH AGRICULTURAL CREDIT PROJECT (LOAN 2837-ME) PROJECT COMPLETION REPORT PREFACE I. This is the Project Completion Report (PCR) for the Ninth Agricultural Credit Project for which Loan 2837-ME in the amount equivalent to US$400 million was approved on June 16, 1987. The Guarantee, Loan and Project Agreements were signed on July 31, 1987. The loan was fully disbursed on April 2, 1992, and closed on June 30, 1992, one year behind schedule. 2. The PCR was jointly prepared by the Agriculture Operations Division of Country Department II of the Latin America and Caribbean Regional Office (Preface, Evaluation Summary, Parts I and III) and the Borrower (Part II). 3. Preparation of this PCR was based on project files, data provided by the Borrower, and a PCR preparation mission to Mexico in October 1992. Project papers reviewed include, inter alia, the Staff Appraisal Report for the Ninth Agricultural Credit Project and the President's Report, both dated May 29, 1987, the Guarantee, Loan and Project Agreements, supervision reports, correspondence with the Borrower and internal Bank memoranda on the complete project cycle, the Operations Evaluation Department Report No. 8860, "Project Performance Audit Report," dated June 29, 1990, on the Bank's six previous agricultural credit loans to Mexico, and on the PCR of the Financial Sector Adjustment Loan (Loan 3085-ME).   - iii - MEXICO NINTH AGRICULTURAL CREDIT PROJECT (LOAN 2837-ME) PROJECT COMPLETION REPORT EVALUATION SUMMARY 1. Through the Ninth Agricultural Credit Project, which disbursed from September 17, 1987 through April 2, 1992, IBRD (the Bank) provided US$400 million equivalent to the Government of Mexico (GOM) to support the agricultural credit system. The project was administered by Banco de Mexico's (BANXICO) Trust Funds for Agriculture (FIRA - US$300 million) and BANRURAL's Trust Fund for Credit in Irrigated and Rainfed Areas (FICART - US$100 million)." Objectives 2. The project was approved at a time when the Bank's wider country assistance strategy was focussed on providing large, relatively quick disbursing policy-oriented loans, both to help Mexico's foreign exchange position, and to support crucial structural changes in the Mexican economy. In this context, the project's objectives were to: (a) increase agricultural productivity and thus agricultural exports, real farm income, and rural employment; (b) assist the Government to improve agricultural sector policies through a further reduction of the interest rate subsidy; and (c) assist the participating financial intermediaries to maintain their financial integrity. To achieve these objectives the project was to: (i) finance investment and incremental short-term credit to farmers; (ii) provide training, technical assistance, consultants' services, and demonstration centers, to support FIRA's and FICART'S productive support programs; (iii) support an Agricultural Financial Subsector Study to identify further areas to improve the rural financial system and increase resource mobilization; and (iv) ensure that budgeting provisions were adequate to materially maintain the capital structure of FIRA and FICART. Implementation Experience 3. A critical variance between planned and actual project implementation was the timing of the Agricultural Financial Subsector Study, which was delayed because of lack of agreement on the final TORs and the role of the Bank with respect to the study, the need to carry out field surveys, and later, due to lack of adequate analysis of the data collected during the field surveys. Remedial measures were agreed upon by supervision missions that continued the dialogue on reforming the rural financial system. This dialogue eventually led to the Government's decisions of closing down its agricultural insurance agency (ANAGSA), restructuring the Government agricultural bank (BANRURAL), and piloting a new incentive scheme for commercial bank lending to small farmers who lacked access to the formal credit system. (See para. 34 and paras. 57-59.) 4. A second variance was that the bulk of the Loan, for the credit component, disbursed much quicker than expected. Disbursements were not tied to continuing policy reforms, and, in particular, to the implementation of the recommendations of the Agricultural Financial Subsector Study. Consequently, the unintended fast disbursements may have reduced the influence the project might have had over the pace of continuing reforms. (See paras. 35 and 43.) 5. One factor that positively affected implementation was the sharp decline in the rate of inflation that has occurred since 1988. The project had been appraised and negotiated at a time when inflation was rampant, reaching 159 percent in 1987. In 1988, however, inflation subsided to 52 percent, and continued declining sharply thereafter. (See para. 37.) !' Gos dissolved FICART as of January 1. 1993.  - iv - 6. Compliance with Loan Covenants. GOM generally complied with loan covenants, except for not ensuring that participating banks' financing contributions were fully in line with the two Project Agreements. Beneficiary participation, however, was above the minimum required by the BANXICO/FIRA Project Agreement, and FICART partially assumed BANRURAL's financing responsibility. In addition, there was delayed compliance with the General Interest Rate Agreement (GIRA) and the execution of the Agricultural Financial Subsector Study. GOM's compliance with the GIRA was eventually satisfactory, although initially it was consistently behind schedule. The Bank had to grant two exceptions, and did not disburse against subloans granted during January-March 1988. GOM has been in compliance with GIRA since then. GOM eventually complied with carrying out the Agricultural Financial Subsector Study; however, because of several delays caused by disagreements with the Bank as to the scope and analysis of the data obtained, the study could not be finalized until August 1990, two and a half years behind schedule. (See paras. 40-41). 7. Project Costs, Financing and Disbursements. The total cost of the project at completion has been estimated at US$920.0 million, 8 percent below the appraised cost. The Bank loan financed 44 percent of total project costs, compared to 40 percent estimated at appraisal. Two years after Loan signing, 93 percent of Loan funds had been disbursed. An extension of the closing date was necessary, however, due to delays in the construction of the FIRA headquarters building. The Loan was fully disbursed on April 2, 1992, and closed on June 30, 1992. (See paras. 42-43.) Results 8. Project Objective: Increase agricultural productivity and thus agricultural exports, real farm income, and rural employment. The overall numerical targets for beneficiaries and job creation appear to have been exceeded, and the FIRA productivity support component was carried out. FICART did not carry out the subcomponent with project funding. However, achievement of the project's objectives with respect to agricultural production and agricultural exports cannot be clearly demonstrated and will depend largely on continuing rural credit reforms to stimulate future agricultural investment. (See paras. 45-53.) 9. Project Objective: Assist the Government to improve agricultural sector policies through a further reduction of the interest rate subsidy. The project, through GIRA, was successful in assisting the Government to reduce interest rate subsidies to agriculture, although a major factor contributing to the decrease in these subsidies has been GOM's success in controlling inflation. There were also significant improvements in GOM's agricultural sector and rural credit policies during the project period. At the end of the project, however, the Government remains committed to the selective use of agricultural credit subsidies and controlled interest rates as an element of social policy, the timing and extent of any further rural credit reforms appear uncertain, and BANRURAL remains financially unviable. (See paras. 54-60.) 10. Project Objective: Assist the participating financial intermediaries to maintain their financial integrity through provision for direct and transparent transfers from the Government to cover remaining interest subsidies; improved financial management; and institutional development. The project has accomplished the objective of ensuring the financial integrity of both FIRA and FICART as measured by maintaining equity in real terms at their January 1, 1986 levels. However, in retrospect, these requirements appear somewhat arbitrary, if not excessive, and may have benefited from more ex- ante analysis and discussion with GOM. Generally, GOM transfers have also been made more direct and transparent, although some off-budget, below-cost funding still exists, such as BANXICO's rediscounts of part of FIRA's portfolio, and FIRA is still not financially self-supporting. With respect to financial management and institutional development, the FIRA staff training component appears worthwhile, although it would have better if there had been a formal evaluation. FICART did not carry out the subcomponent with project funding. (See paras. 61-71.)  v Sustainability 11. The rural credit reforms implemented during the project period appear sustainable, and there appears to be relatively little risk of backsliding. However, the Government's rural financial institutions are not currently "sustainable," and there appears to be little prospect for them becoming so in the near future. The momentum for further rural credit reforms appears stalled, and the expected follow-up project is in abeyance, which was to address issues of financial viability of the agricultural credit system and institutions. Consequently, one hesitates to assert that the Ninth Agricultural Credit Project is itself "sustainable." (See para. 72.) Institutional Perfornance 12. Bank. During preparation and appraisal, the main question for the Bank was "how much and how fast" credit subsidies should be reduced to justify proceeding with the project. After considerable preparation delays caused by disagreements over the content and timing of policy reforms to be supported by the project, processing of the Ninth Agricultural Credit Project was finalized when the Bank relaxed somewhat previous positions regarding the extent and timing of the policy reforms to be supported by the project. This was due partly to the wider country assistance strategy then in effect of emphasizing relatively large, quick-disbursing projects to support structural changes in the Mexican economy, as well as helping Mexico's foreign exchange position. In this context, although some of the specific objectives of the project were not fully met, the Ninth Agricultural Credit Project, like other large, quick-disbursing, policy based loans, can be considered part of a quite successful wider country assistance strategy. In addition, the simple existence of the Ninth Agricultural Credit Project provided a good vehicle to discuss with GOM significant reforms that did take place during the project period. Consequently, it is difficult to conclude that the Bank made the wrong decision to proceed with the project in 1987. Project supervision was generally adequate. (See paras. 73-75.) 13. Borrower. The Borrower's performance was good with regard to the reduction of interest rate subsidies to agriculture. Moreover, the Borrower ultimately implemented a series of reforms in the rural credit subsector that went beyond the recommendations of the Agricultural Financial Subsector Study. Possible criticisms, however, may include (a) failure to take prompt action regarding delayed construction of the FIRA headquarters building; (b) inadequate monitoring and evaluation of the productive support component; and (c) unwillingness to provide the Bank the data collected during the second phase of the Agricultural Financial Subsector Study. (See paras. 76-77.) Findings and Lessons Learned (See paras. 83-89.) 14. Measuring achievement of project objectives. In future operations, the Bank and the Borrower should build into project design an effective system for monitoring performance indicators to measure the achievement of project objectives through the linkage between project activities and general objectives such as "increase agricultural productivity," or avoid claiming such broad, difficult to measure project objectives in the first place. 15. Policy reforms. As in other projects, when this project was appraised it was difficult to decide on the degree of policy reform needed to justify proceeding with the project. In the case of this project, the subject was complicated by the wider country assistance strategy at the time, which included substantial resource transfer objectives to support policy reform. Even today, it is difficult to measure whether the project was "successful." The lesson is that the Bank and the Borrower should be more specific about the objectives of a project, which may well include resource transfer objectives, as well as about what would be considered "successful" implementation, especially with respect to policy reforms.  - vi - 16. Agricultural credit policy reforns. A related lesson from this and Bank-assisted agricultural credit projects in other countries is that, if the primary objective of a project is to reform the rural credit system, then simply relying on studies to be carried out during the project may be an insufficient instrument to encourage further reform, especially given the high degree of political sensitivity agricultural credit has in many countries. Ideally, of course, all policy reforms would be made, or at least agreed upon, prior to Loan approval. If this is not possible, then the lesson is to directly link disbursements to the implementation of agreed action plans based on the recommendations of such studies. 17. In addition to the above general point, a number of lessons from the Mexico Ninth Agricultural Credit Project are consistent with the Bank's experience with agricultural credit projects in other countries: (a) financial intermediaries at each link of the chain should pay the full cost of funds, including administrative costs, to avoid distortions and overburdening government resources; (b) agricultural insurance and credit guarantee programs should be actuarially sound; (c) government institutions should be staffed and managed along commercial lines; and (d) to the extent that subsidies are desired to help the poor, they should be transparent and well targeted. 18. Selection of financial performance indicators. In retrospect, the Bank's concern with the possible loss in value of FIRA's and FICART's equity appears overstated. The proportion of equity to risky assets is very high for both institutions, even if measured conservatively. In future operations, the Bank should focus on maintaining an adequate level of equity for a second-tier institution, as required by the Mexican National Banking Commission, by adapting another standard such as the Basle Agreement as it was done for first-tier DFIs in Operational Directive 8.30, or by agreeing on a system in line with the "Convenio" mentioned in para. 24 of Part I, instead of simply requiring maintenance of existing equity in real terms. 19. Civil works construction. Project files do not record that there was a thorough analysis of the proposed FIRA headquarters building during project preparation and appraisal, and Bank supervision missions have concluded that the eventual specifications are excessive. Especially for large buildings such as this one (originally estimated to cost US$11 million), calculations should be made during preparation clearly showing the scope of the proposed investment, and conceptual designs and bidding conditions should be agreed to during appraisal. Without proper information, analysis and justification, such proposed investments should not be accepted. 20. Contractors. Approval of contracts which include disbursement for advances to contractors run the risk of having these disbursements tied up in litigation if the contractor fails to perform. Performance bonds, especially in Mexico, do not pay upon notification of contractor default. They are only effective after legal settlement, which can be delayed for years. In the future, in countries like Mexico, in which performance bonds may not be executed easily, the Bank and the Borrower should put more emphasis on analyzing the financial condition and other qualifications of bidders during bid evaluation.  MEXICO NINTH AGRICULTURAL CREDIT PROJECT (Loan 2837-ME) DRAFT PROJECT COMPLETION REPORT PART I: PROJECT REVIEW FROM THE BANK'S PERSPECTIVE 1. PROJECT IDENTITY Project Name: Ninth Agricultural Credit Project Loan No.: Loan 2837-ME RVP Unit: Latin America and the Caribbean Region Country: Mexico Sector: Agriculture Subsector: Agricultural Credit 2. BACKGROUND 1. Macroeconomic setting. The years prior to the approval of the Ninth Agricultural Credit Project in 1987 were exceptionally difficult for the Mexican economy. The 1980s began with rising world interest rates and falling oil prices, which resulted in severe fiscal and balance of payments deficits and led Mexico to suspend interest payments on its external debt in 1982. This financial and economic crisis prompted the real depreciation of the peso in 1982 and 1983 and initiated an inflationary spiral, as inflation rates rose from 29 percent in 1981 to 159 percent in 1987. The Government responded with strong fiscal and monetary adjustment; the fiscal deficit dropped sharply from 7.6 percent of GDP before the 1982 crisis to an estimated 5.8 percent surplus of GDP in 1988. However, the adjustment program was undermined by the collapse of international oil prices in 1986 and the ensuing currency devaluations. As a result of the restrictive fiscal policy and the unfavorable international environment, real economic growth stagnated between 1982 and 1987 and real wages fell by more than 40 percent. This macroeconomic context was relevant to the preparation of the Ninth Agricultural Credit Project, because the Bank's country assistance strategy at the time was heavily focussed on financing large, relatively quick-disbursing, policy-oriented projects, both to help Mexico's foreign exchange position and to support structural changes in the Mexican economy. 2. The agriculture sector. The agricultural sector accounts for about 9 percent of GDP and contributes about 9 and 13 percent of the country's total and non-oil exports, respectively. During the 1980s, the sector's growth rate was 0.4 percent compared to 1 percent for overall GDP, reflecting the unfavorable macroeconomic environment, low international prices, and poor agricultural policies. Since 1985, the sector is being transformed to have a greater reliance on market forces, although farm production has not yet shown a strong response to the reforms. Bank support for this adjustment process has included the First Agricultural Sector Adjustment Loan (Loan 2918-ME, approved in February 1988) and the Second Agricultural Sector Adjustment Loan (Loan 3357, approved in June 1991). 3. Directed agricultural credit at preferential rates has been key to the Government's strategy to increase agricultural output and the living standards of the rural population. Due to the economic crisis  -2- of the early 1980s and budgetary constraints, outstanding agricultural credit dropped consistently in real terms, from 48 percent of agricultural GDP in 1980 to 22 percent in 1987. 4. Rural finance has been characterized by a distorted incentive structure for both borrowers and lenders. Government institutions have crowded out private lenders. Mobilization of rural savings has been slow. The Governrment's agricultural credit bank (BANRURAL) had little incentive to attract deposits to lend to the rural sector because of the demonstrated security of receiving Government transfers. Moreover, Governrment policy discouraged commercial bank lending to agriculture, particularly to small farmers. BANRURAL's privileged access to funding and relaxed collection practices--a large proportion of its loans went uncollected, with the resultant financial gap being covered with government transfers-- and controls on borrowing and lending rates discouraged commercial banks from mobilizing resources from, and lending to, agriculture and creating a commercially sound rural banking infrastructure. Compulsory insurance of BANRURAL's loans by the government-owned agricultural insurance company (ANAGSA) encouraged BANRURAL's already loose loan recovery policy, and as it did farmers to default on loans. 5. Mexico's rural financial system has never been, and still is not, self-financing. Subsidized interest rates on loans to agricultural producers, sometimes negative in real terms, the lack of borrowers' credit discipline, and the operating inefficiencies of BANRURAL and ANAGSA have had a high cost to the treasury. It has been estimated that during the 1980s, Government transfers to the agricultural credit system have exceeded USS28 billion. 3. PROJECT OBJECTIVES AND DESCRIPTION 6. Through the Ninth Agricultural Credit Project, which disbursed from September 17, 1987 through April 2, 1992, IBRD (the Bank) provided US$400 million equivalent to the Government of Mexico (GOM) to support the agricultural credit system. The project was administered by Banco de Mexico's Trust Funds for Agriculture (FIRA - US$300 million) and BANRURAL's Trust Fund for Credit in Irrigated and Rainfed Areas (FICART - US$100 million)." 7. The objectives of the Ninth Agricultural Credit Project were to: "(a) increase agricultural productivity and thus agricultural exports, real farm income, and rural employment; (b) assist the Government to improve agricultural sector policies through a further reduction of the interest rate subsidy; and (c) assist the participating financial intermediaries to maintain their financial integrity through: provisions for direct and transparent transfers from the Government to cover remaining interest subsidies, improved financial management, and institutional development." (See Staff Appraisal Report (SAR), para. 3.02.) 8. To achieve these objectives the project was to: (a) finance investment and incremental short- term credit to farmers for crops, livestock, agroindustries and other activities (e.g., aquaculture and horticulture); (b) provide training, technical assistance and consultant services, as well as civil works for the construction of demonstration centers, to support FIRA's and FICART's productive support programs; (c) support an Agricultural Financial Subsector Study to identify further areas to improve the rural financial system and increase resource mobilization; and (d) ensure that budgeting provisions were adequate to materially maintain the capital structure of FIRA and FICART. An additional project component not explicitly stated in the SAR, although included in the Loan Agreement, was the construction of a headquarters building for FIRA. pi GOM dissolved FICART as of January 1, 1993.  9. Prior to this Loan, the Bank had made eight similar agricultural credit loans and an "interim"'' agricultural credit loan to Mexico, totalling US$1.4 billion equivalent, all administered by FIRA.3' The project under review was the first Bank Loan in which BANRURAL participated directly through rediscounts with FICART. Until recently, the Bank had been considering a new credit operation to support GOM's policy reforms in the rural financial subsector, but processing is in abeyance due to lack of agreement between GOM and the Bank on the extent and pace of future rural credit reforms. 4. PROJECT PREPARATION AN]) DESIGN A. Project Design 10. Project genesis. Preparation of the Ninth Agricultural Credit Project started early in 1985, as a follow-up to the Eighth Agricultural Credit Project (Loan 2454-ME, approved in June 1984). As noted above, during this period Mexico was suffering from stagnation in both the total economy and the agricultural sector, inadequate support for agricultural credit by the nationalized banks, and severe operating problems in BANRURAL. This project was prepared as a part of a larger package of policy reforms and activities4' designed to assist Mexico's wider economic restructuring, and was thus linked to the 1986-87 Commercial Bank Financing Package for Mexico. 11. Project design. Since the Fifth Agricultural & Livestock Credit Project (FIRA 5, Loan 1217- ME, approved in March 1976), the Bank had been unsuccessfully trying to move towards a "sector approach" for agricultural credit loans to Mexico5". The Ninth Agricultural Credit Project also sought a sectoral approach, including disbursement conditionality in addition to the General Interest Rate Agreement (GIRA)6". During preparation, the Bank made substantial effort to obtain GOM's agreement with respect to the reduction and future elimination of interest rate subsidies and the financial self- sufficiency of the Government agricultural credit entities, especially, FIRA and FICART. However, after lengthy discussions with GOM and as preparation advanced, the Bank relaxed these objectives, partly because of the wider countr-y assistance strategy to rapidly increase resource transfers to Mexico in the wake of the debt crisis. In particular, the Loan Agreement explicitly linked the Ninth Agricultural Credit Project to the 1986-87 Commercial Bank Financing Package for Mexico, dated October 16, 1986. 12. In the end, project design evolved very much along the same lines of previous operations, with no up-front policy conditionality and no significant disbursement conditions, other than the GIRA. The end result was a project which left many important issues to be addressed by an Agricultural Financial Subsector Study that was to be carried out under the project, with its recommendations then implemented during the latter part of the project period. The objectives of the study (see para. 25), moreover, were much broader than the project's specific credit policy objectives, which only contemplated a further 2, During preparation of this project, the Bank also approved an interim loan named "Agricultural Credit Loan" (Loan 2610-ME, also known as FIRA 8A) for US$180 million, processed between March and July 1985. "1 See Project Performance Audit Report No. 8860, IBRD, Operations Evaluation Department, June 29, 1990. '' Such as an agricultural sector review, which led to the first Agricultural Sector Adjustment Loan (Loan 2918- ME, approved in February 1988) and a Financial Sector Adjustment Loan (Loan 3085-ME, approved in June 1989). 1/ See paras. 18, 19, 26 and 33, Project Performance Audit Report No. 8860, IEBRD, Operations Evaluation Department, June 29, 1990. Y The GIRA is an umbrella interest rate agreement, signed in 1984 by the Government of Mexico and the Bank, that covers aU credit operations financed by the Bank in Mexico. Under this agreement, the adjustment of the overall level of controlled interest rates was tied to the Average Cost of Funds (cPP) until 1989, and has been tied to the interest rate on the 28-day short-term Government paper (CETES) since 1990.  - 4 - reduction of the interest rate subsidy, and transparent transfers to the two executing agencies. Disbursements, however, were not linked either to satisfactory completion of the study, or to the implementation of its recommendations. In retrospect, the project could have benefitted from additional preparation, during which the Bank could have ensured that there was undivided agreement between the Government and the Bank as to project objectives and adequate instruments to achieve those objectives. 13. New project features. Compared to previous projects in the subsector, this project included two new features: (a) the direct participation of FICART/BANRURAL, and (b) a study of the rural financial subsector to identify, inter alia, the cost of financial intermediation and subsidies to the agricultural sector. It was expected that FICART would benefit from participating in the project by becoming financially and operationally more efficient, and would positively influence BANRURAL. The project provided for training and technical assistance of FICART's staff, but very little attention was paid to developing a plan of action to ensure the institutional strengthening of the institution, and thereby, its expected positive influence on BANRURAL. 14. Draft terms of reference (TORs) for the Agricultural Financial Subsector Study were agreed during negotiations. The study, then to be financed by GOM, was expected to be finalized and its recommendations implemented by March 1988 (see paras. 25-28 below). 15. Roles and responsibilities. The roles and responsibility of the institutions involved were clear, except for the role that the Bank was supposed to play in the Agricultural Financial Subsector Study. The legal provision for this study (Loan Agreement Section 3.02) did not require the Government to submit a copy of the entire report, but only to "exchange views with the Bank in respect to the conclusions and recommendations of the study." B. Key Issues Discussed During Preparation 16. Credit policies. During project preparation and appraisal, the Bank persistently sought GOM agreement to further reduce and ultimately eliminate interest rate subsidies for agricultural credit. Although the Government recognized that credit subsidies were not a cost effective instrument for either poverty alleviation or increasing agricultural growth, it maintained that they were essential for social policy in Mexico. GOM agreed, however, to revise the GrRA interest rate to further close the gap between interest rates to low-income producers (PBls)" and the reference rate. The GLRA, which had been used in the two previous agricultural credit operations, was considered an effective instrument for overall subsidy reduction. Although the amount of this Loan represented a small proportion of the lending program of FIRA and FICART,8' the Bank, through GIRA, was able to bring agricultural interest rates for PBls closer to the reference rate (cPP or CETES), from 70 percent of the reference rate in 1987 to 95-97 percent in 1989, the level prevailing today. 17. GOM further agreed to examine other alternatives for the delivery of subsidies under the Agricultural Financial Subsector Study supported by the project, and that this study would also identify the true cost of intermediation of financial intermediaries, to adjust existing intermediation margins accordingly. Moreover, the GOM agreed to budget adequate funds to avoid any further decapitalization of F'RA and FICART. 18. BANRURAL's participation. It was not until this project that the Bank agreed to BANRuRAL'S direct participation--through its trust fund, FICART. BANRURAL's earlier participation had been through discounts with FIRA, which the Bank approved during implementation of FIRA 5 because it (a) "wanted 71 Low-income producers (PBms) are those producers with net family income (from all sources) below 1,000 times the minimum regional daily wage. The PBI rate was the only interest rate that remained below the Cpp. F About 6 percent of the appraisal estimate of FIRA/FICART five-year discount program.  -5- to turn BANRURAL into a sound development institution, and (b) "it also needed BANRURAL to get the PBI part of Loan 1217-ME disbursed".9' Ever since, and despite strong evidence of BANRURAL's weak financial position, the Bank has accepted BANRURAL's indirect participation in all subsequent agricultural credit projects. Furthermore, during preparation of the Ninth Agricultural Credit Project, it was decided that BANRURAL could participate directly, through FICART, provided that the Government agreed to undertake a study of the agricultural financial subsector, and BANRURAL's portfolio arrears did not exceed 15 percent. 5. PROJECT IMPLEMENTATION A. Key Issues during Implementation (1) Credit Component 19. Interest rates. At the start of the project, GOM was consistently behind schedule in conforming to the level of agricultural interest rates agreed upon under the GIRA. The Bank had to grant two exceptions in a six-month period. A first exception was made in early 1988 to accommodate subloans made between November 1 and December 31, 1987, in the expectation that the issue would be resolved soon. Subsequently, the Government failed to comply with GIRA requirements which called for an increase of on-lending rates to PBls to at least 85 percent of CPP by January 1, 1988. Disbursements under the Loan were temporarily halted, and the Bank did not disburse against subloans granted during January-March 1988. A second exception was made in August 1988 for subloans granted during April- July 1988, despite non-compliance with the level of interest rates and the use of higher intermediation margins. The Bank agreed to this additional exception provided that GtRA was amended to increase interest rates to PBIs to at least 95 percent of CPP starting in August 1988. In March 1991, the GOM also increased by one point, to CETES + 2, the interest rate applicable to OPs engaged in the production of basic products. Because of the rapid decline in inflation during 1988, the GIRA amendment also provided for monthly adjustments of the interest rates, instead of semiannual. 20. Intermediation margins. During implementation of this project, the Government proposed an increase in the intermediation margins of commercial banks to 12.5 percentage points for PBls and oPs '