Colombia - Rural Development Investment Program Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 16721 IMPLEMENTATION COMPLETION REPORT COLOMBIA RURAL DEVELOPMENT INVESTMENT PROGRAM (LOAN 3250-CO) June 18, 1997 Natural Resources Management and Rural Poverty Division Country Department III Latin America and the Caribbean Region This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  CURRENCY EQUIVALENT Currency Unit = Colombian Peso (Col$) Average Exchange Rate for 1996 US$1.00 =Col$ 1000 WEIGHTS AND MEASURES Metric System FISCAL YEAR OF THE BORROWER January 1 - December 31 ABBREVIATIONS AND ACRONYMS DRI, Fondo DRI - Fondo de Cofinanciacion para la Inversion Rural - DRI (Rural Investment Cofinancing Fund)1/ CMDR - Consejo Municipal de Dessarrollo Rural - (Municipal Council for Rural Development) CONPES - Consejo Nacional de Politicas Econ6micas y Sociales (National Council for Economic and Social Policy) ERR - Economic Rate of Return FY - Fiscal Year IDB - The Inter-American Development Bank IRR - Internal Rate of Return NCS - National Cofinancing System NGO - Non-Governmental Organization PNR - Programa Nacional de Rehabilitaci6n - (National Rehabilitation Plan) RDIP - Rural Development Investment Program (Programa de Desarrollo Integral Campesino - PDIC) SA - Special Account SAR - Staff Appraisal Report UDECO - Unidad Departamental Especializada de Cofinanciacion (Departmental Cofinancimg Unit) UMATA - Unidad Municipal de Asistencia Tecnica Agropecuaria (Municipal Unit for Agricultural-Livestock Extension Services) UNIR - Unidad de Interlocuci6n Regional (Unit for Coordination with the Regions) Vice President: Shahid Javed Burki Director: Paul Isenman Acting Division Chief: Jonathan Parker Staff: Elsie Garfield 1/ Fondo DRI changed its name with Decree 2132 from Fondo de Desarrollo Rural Integrado (Integrated Rural Development Fund) to the Fondo de Cofinanciacion para la Inversion Rural (Rural Investment Cofinancing Fund). Throughout the report it will be referred to interchangeably as Fondo DRI or DRI.  FOR OFFICIAL USE ONLY IMPLEMENTATION COMPLETION REPORT COLOMBtA RURAL DEVELOPMENT INVESTMENT PROGRAM (LOAN 3250-CO) Table of Contents Page No. PR E FA C E .........................................................i EVALUATION SUMMARY ......................................................... ii PART I: PROJECT IMPLEMENTATION ASSESSMENT Introduction ................................................I A. Project Objectives ................................................1 B. Achievement of Project Objectives ..............................................3 C. Implementation Record and Major Factors Affecting the Project ................. 4 D. Project Sustainability ..............................................6 E. Bank Performance ................................................7 F. Borrower Performance ..8 G. Assessment of Outcome ..........................................9 H. Future Operation ....................................... ..9 I. Key Lessons Learned .......................................... 10 PART II: STATISTICAL TABLES 1. Summary of Assessments ................................................... 11 2. Related Bank Loans ................................................... 12 3. Project Timetable ................................................... 12 4. Loan Disbursements: Cumulative Estimated and Actual ......................... 13 5A. Key Indicators for Project Implementation ............................................. 14 5B. Key Indicators for Subproject Accomplishments .................................... 15 6. Studies Included in Project ................................................... 16 7A. Project Costs ................................................... 17 7B. Project Financing ................................................... 18 8. Status of Legal Covenants ..................... .............................. 19 9. Compliance with Operational Manual Statements ................................. 22 10. Bank Resources: Staff Inputs ................................................... 22 11. Bank Resources: Missions ................................................... 23 APPENDIXES: A. Borrower's Contribution to the ICR ................................................... 24 B. Borrower's Comments on the ICR ................................................... 38 C. Cofinancier's Comments on the ICR ................................................... 41 D. Map IBRD 21684RI 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.   IMPLEMENTATION COMPLETION REPORT COLOMBIA RURAL DEVELOPMENT INVESTMENT PROGRAM (LOAN 3250-CO) PREFACE This is the Implementation Completion Report (ICR) for the Rural Development Investment Program in Colombia, for which Loan 3250-CO in the amount of US$75.0 million was approved on July 31, 1990 and made effective on January 31, 1991. The loan was closed on December 31, 1996, one year later than the original closing date. It was fully disbursed, and the last disbursement took place on June 3, 1997. Cofinancing for the project was provided by the Inter-American Development Bank. The ICR was prepared by Elsie Garfield of the Natural Resources Management and Rural Poverty Division, Country Department III of the Latin America Region assisted by Maria Elena Castro, LA3DR, H.J. Nissenbaum (Consultant) and Carmen Nielsen, LA3NR. The report was reviewed by Jonathan Parker, Acting Chief, Natural Resources Management and Rural Poverty Division (LA3NR) and Robert Crown, Project Advisor, Country Department III. The Borrower and Cofinancier have provided comments that are included as Appendixes B and C to the ICR. Preparation of the ICR began during mid-1995 with the Bank working closely with the Implementing Agency (Fondo DRI) to prepare an in-depth evaluation of the program. Due to this continuous involvement over a year and a half, a specific ICR mission was not undertaken and no aide-memoire was prepared. The Borrower has prepared a summary of their lengthy evaluation report which is contained in Appendix A.   IMPLEMENTATION COMPLETION REPORT COLOMBIA RURAL DEVELOPMENT INVESTMENT PROGRAM (LN. 3250-CO) EVALUATION SUMMARY Introduction 1. Twenty years ago the Colombian Government initiated an integrated rural development program (DRI) to increase the productivity, incomes and output of small farmers (campesinos) which the Bank supported with two loans. The third phase of DRI, supported by this loan approved in 1990, represented a significant change in approach which was shaped by Colombia's process of political, administrative and fiscal decentralization. The program's concept and mode of operation changed from one in which Fondo DRI contracted central government agencies to carry out program activities to one which gave municipalities and communities a central role. Project Objectives 2. The loan was to finance a time slice of the Rural Development Investment Program (RDIP) for 1990-94 which covered 602 municipalities located in 21 Departments. DRI would provide matching grants for investment projects to be carried out by municipalities and communities. The project's main objectives were to: (i) increase the incomes of about 280,000 poor farmers and fishermen; (ii) improve living standards of the rural communities through provision of basic infrastructure; (iii) promote grassroots involvement of rural communities in the identification of investment needs and subsequent maintenance; and (iv) strengthen DRI's capabilities to prepare, appraise and supervise projects and help the municipalities administer and cofinance development projects to support decentralization. 3. The program's objectives were not totally clear. There was some tension between the desire to improve the situation of poor farmers and fishermen, and the "productivity" focus which led to working in areas with agricultural promise and with farmers having the potential to become businessmen. This focus was diluted in 1994 when the program was expanded to the entire country. The establishment of quantitative objectives for economic benefits was overly ambitious and unrealistic for an operation which was a pioneer of the decentralization process. The appraisal report recognized the risk that some of the benefits might not materialize during the first five years of the program, but the remedies proposed did not address this issue well. In addition, the program's strategy for rural development left out two critical ingredients--small farmers' access to land and capital-- which also were not being successfully addressed outside the program.  Implementation Experience and Results 4. The major factors affecting the project's implementation were those within the Government's control. The RDIP's design was overly bureaucratic and complex, thus inconsistent with the participatory approach; this was not detected immediately since the program got off to a good start. Although operational limitations soon emerged, DRI steadily improved implementation and by 1993 the new approach, including effective delegation of responsibilities to DRI's regional offices and mastery of procedures and instruments introduced by the RDIP, was taking hold. 5. The Government then issued Decree 2132, establishing the National Cofinancing system (NCS), which mandated the closure of DRI's regional offices by 1994 (functions to be taken over by Departmental govemments), and required it to use a fiduciary to manage cofinancing funds. In addition, DRI was required to expand nation-wide and assume responsibility for new programs which added to its burden. DRI and, to some extent the Government, undertook remedial measures and developed some creative responses to these problems which yielded some positive results in 1996. 6. A major implementation problem throughout was the lack of an adequate monitoring system to track progress in the subprojects being financed. By the time the loan closed, a satisfactory system had been designed, but was still not functioning adequately despite considerable efforts by DRI to assist the departments and municipalities who were the primary providers of the information. 7. Despite the many problems and obstacles, the RDIP achieved two of its four major objectives and partially achieved the others. It cofinanced 12,038 projects in the span of six years of which the bulk of funding was for agricultural extension services and infrastructure. The number of beneficiaries and level of investment far surpassed expectations, with 64 percent of the resources going to the least developed municipalities. Beneficiaries and municipalities have expressed a high level of satisfaction with the program's results, including improved incomes and production (see Table 5A). 8. The main activity aimed at increasing the incomes of small farmers, as part of the first objective, was support for the decentralization of the agricultural extension system. By 1995, municipal technical assistance units (UMATAs) were serving double the number of farmers expected at appraisal and three times the number served by the previous extension system. The RDIP made an important contribution to the second objective by improving infrastructure in rural areas, with roughly 50 percent of program funds allocated to road improvement and construction, water and sanitation systems (benefiting 450,000 rural residents), and rural electrification. The program only partially fulfilled its third objective of promoting community involvement, due to weaknesses in the program's design and the disruptive changes which overwhelmed DRI. However this changed in 1996 when DRI led efforts to make the newly created Municipal Councils for Rural Development (CMDR) effective forums for community participation. Concerning the fourth objective, the strengthening of DRI's capacity occurred in the early years, but the institutional limbo and chaos which resulted from the introduction of the NCS seriously  - iv - undermined it. The related goal of strengthening municipalities was partially achieved. Survey evidence and feedback from the government indicate that DRI and the RDIP are perceived as having made an important contribution to improving municipal and departmental capacity (the latter was totally unforeseen at appraisal). 9. Project sustainability is uncertain. The prospects are good for the UMATAs and water and sanitation projects: (i) half of the water and sanitation projects are being operated and maintained by user committees; and (ii) municipalities have assumed a rising share of the cost of the UMATAs and, in many cases, mayors and farmers are actively supporting them. Uncertainty remains about maintenance of road investments, though as the decentralization process is consolidated, the prospects for this should improve. Finally, concerning the institutional development objectives, Colombia's 1991 Constitution provides a strong basis for sustained involvement of municipalities and communities in rural development. However, there is much uncertainty as to the sustainability of DRI as an institution as Government has announced the eventual elimination of the NCS. 10. The RDIP's outcome is judged satisfactory. As indicated above, despite facing an enormous number of complex problems, the RDIP was a larger program and reached more beneficiaries than expected. The prospects for sustained municipal and community participation are good. The RDIP was a pioneering operation for the Bank and has served as a model for rural development investment funds being tried in other countries. 11. The Bank's overall performance was satisfactory. However, in several respects the Bank's appraisal of the RDIP was deficient due to lack of experience with this innovative approach, particularly the participatory aspects. Supervision was satisfactory. While in hindsight, perhaps the Bank should have realized earlier on the serious impact of Decree 2132, supervision missions continued to find good results on the ground and disbursements were satisfactory. When the negative impact of Decree 2132 became clear, the Bank assisted the Government and DRI to analyze and develop solutions for the RDIP and NCS' problems. The Bank worked closely with the Inter-American Development Bank during all phases of the project. 12. The overall performance of the Government (Borrower) and Fondo DRI (Implementing Agency) is judged deficient, though it is recognized that many problems which arose were not easy to foresee or handle due to the ambitious and politically charged nature of the decentralization process. The objectives of Decree 2132 were laudable, but the mandated changes proved to be impractical in the short-term and adequate transitional arrangements were not provided for. There appeared to be a lack of sufficient recognition that institutional reforms take t.ime to implement and, thus, a tendency to drop a reform without allowing sufficient time for it to take hold. Overall, DRI's performance is also judged deficient, but admirable under the circumstances. Institutional instability and constantly changing rules during the RDIP made it difficult to get things done and eroded institutional capacity. Financial management and monitoring and evaluation remained critical weaknesses throughout.  - v - 13. Conceming future operations, the Government has requested the IDB to support a new phase of DRI's program whose proposed design incorporates many of the lessons of the RDIP. The Bank will continue to support the Government's efforts to reduce poverty through promoting rural development, a priority of the Country Assistance Strategy, including consideration of options for use of the NCS and DRI. Findings and Key Lessons 14. Some key lessons of the RDIP are as follows: * a participatory and decentralized operation which is to finance many small projects in a large number of localities, can not work if it is overburdened by the bureaucratic, centralized rules which characterize normal operations of the Colombian Government and World Bank; likewise, a shift in focus from ex-ante controls to ex-post controls and sanctions is likely to be more conducive to timely implementation and building accountability at the local level; * a flexible, open design is needed for centrally managed programs targeted to local governments due to the enormous diversity of capacity, needs and socio-economic conditions in Colombia; * the process of designing this type of operation should be highly participatory, drawing on the contributions of local communities and municipal governments to define the basic operating rules which are pertinent --.d consistent with local capacity, while ensuring an adequate level of transparency and accountability in use of public funds; * a simple monitoring and control system should be in place from the start in this type of operation in order to rapidly identify and correct problems, and to identify local governments which require special assistance or, in extreme cases, disqualification from the program; * it is advisable to separate monitoring and evaluation functions, and rely on an independent agency such as a university (which was a successful aspect of DRI II) to conduct periodic impact evaluations; * in view of the Colombian Government's rules concerning management of Special Accounts, the experience of this project shows that they do not fulfill their intended purpose and should be dropped; and * major institutional and political reforms, such as those undertaken in Colombia since the conception of the RDIP, are long-term processes and there are bound to be set-backs and unexpected developments; the Bank needs to exercise patience and flexibility to support this process and judge the ultimate results in a time-frame that goes beyond a single lending operation.  IMPLEMENTATION COMPLETION REPORT COLOMBIA RURAL DEVELOPMENT INVESTMENT PROGRAM (LOAN 3250-CO) PART I: PROJECT IMPLEMENTATION ASSESSMENT Introduction 1. Twenty years ago the Colombian Government initiated an integrated rural development program (DRI) to increase the productivity, incomes and output of small farmers (campesinos). The program's Phase I (1976-82) and Phase 11 (1983-90), supported by Bank loans 1352 and 2174, were motivated largely by the Government's pursuit of self-sufficiency in basic foodstuffs (of which small farmers produced some 60 percent) and concern for improving the living standards and productivity of these farmers many of whom were poor. DRI gained a reputation for being a successful program which the Bank repeatedly commended as a model. 2. The third phase, supported by this loan approved in 1990, represented a significant change in approach which was shaped by Colombia's process of political, administrative and fiscal decentralization. National reforms empowered municipal residents to elect mayors for the first time in 1988, assigned municipalities direct responsibility for providing services previously fuirnished by central government agencies, and gave them the right to greater financial resources through revenue-sharing. The program's concept and mode of operation changed from one in which Fondo DRI contracted central government agencies to carry out program activities to one which gave municipalities and communities a central role in deciding investment priorities, financing and implementing projects. An important aspect of the program was support for the transition from centrally financed and organized extension services for small farmers to municipal responsibility for this service. A. Project Objectives 3. The loan, along with one from the Inter-American Development Bank, was to finance a time slice of the Rural Development Investment Program (RDIP) for 1990-94 consisting of small, discrete environmentally sound projects in 9,000 communities (602 municipalities in 21 Departments). The project's objectives, as stated in the Staff Appraisal Report (SAR), were to: (i) increase the incomes of about 280,000 poor farmers and fishermen mainly in the Andean and coastal areas; (ii) improve living standards of the rural communities through provision of potable water and better sanitation systems, roads and marketing facilities; (iii) protect and develop watershed areas; (iv) promote grassroots involvement of rural communities in the identification of investment needs and subsequent maintenance; and (v) strengthen DRI's capabilities to prepare, appraise and supervise projects and help the municipalities administer and cofinance development projects to support decentralization.  4. The program had a participatory planning approach and, therefore, leeway was left for municipalities and communities to decide the investments to be financed as long as they were on the program's "menu" and respected an agreed set of criteria and procedures. The loan financed two types of expenditures: matching grants for specific types of investment projects to be carried out by municipalities and communities, and activities carried out directly by Fondo DRI including institutional strengthening.' The scope of investments to be financed was reduced during design, in line with lessons from the earlier projects, to focus more on improving agricultural productivity, dropping support for education and health; support for credit to farmers was also dropped due to financial sector policy issues and past problems with this component. 5. The program's objectives were not totally clear. There was some tension between the desire to improve the situation of poor farmers and fishermen, and the "productivity" focus which led to working in areas with agricultural promise and with farmers having the potential to become businessmen.2 This was consistent with the Government's strategy at the time which considered that sustaining high economic growth was the best anti-poverty strategy. Subsequent analysis showed that despite high growth rates nationally and for agriculture which resulted in an overall decline in poverty, poverty was becoming increasingly a rural phenomenon with a broadening gap between living conditions in rural and urban areas along with significant differences among regions in Colombia.3 To be fair, the RDIP did have poverty targeting mechanisms: lower income municipalities were offered special help in preparing subprojects and higher rates of cofinancing. However, its overriding objective was to help small farmers (campesinos); with the introduction of the National Cofinancing System (NCS) in 1994, the emphasis on improving productivity was diluted by the expansion of the program to the entire country with resulting political pressures to provide equal access to all municipalities and departments. 6. The establishment of quantitative objectives for economic benefits (i.e. increased yields and production of agriculture and livestock yielding an economic rate of return of 16 percent for the productive projects) was overly ambitious and unrealistic considering the enormous changes which were being introduced.4 The appraisal report recognized the risk that some of the benefits might not materialize during the first five years of the program, because of inexperience and limited capacity of DRI and the municipalities, the possible need for greater time to put the new mechanisms into effective operation, and the possibility that generation of technological packages to increase yields might take longer 1 Fondo DRI differed from a social fund in two respects: it did not have special status which exempted it from normal Government regulations; and municipalities were responsible for implementing and cofinancing the subprojects. 2 The original choice of 602 municipalities gave greater weight to a municipality's agricultural potential than the socioeconomic conditions. This was in part because another Government program, the National Rehabilitation Program (PNR), was targeted to the poorest, most remote and violent areas of the country. So, to some extent there was an agreed division of labor between the two programs. See Colombia Poverty Assessment Report, No. 12773-CO, August 1994 and Colombia Review of Agricultural and Rural Development Strategy, Rpt. No. 13437-CO, November 1995. Infrastructure projects were to have a minimum ERR of 12 percent in order to qualify for financing.  - 3 - than expected. The remedies proposed, however, such as a Program Review at the end of 1992, Annual Operating Plans and technical assistance to municipalities did not seem to address the issue very well. In hindsight, it was unrealistic to expect a major improvement in yields and production for an operation which was a pioneer of the decentralization process, piloting new ways of operating which required major changes in behavior at all levels. In addition, the program's strategy for rural development left out two critical ingredients for achieving the ultimate objective of increasing incomes, employment and productivity: small farmers' access to land and capital. These were not being successfully addressed outside the program either, though a Bank financed project was trying to help the Government improve the credit situation.5 By mid-1990, increased violence had become another factor complicating rural development efforts. B. Achievement of Project Objectives 7. The project partially achieved its major objectives. Concerning the first three objectives which relate to the investments financed by the program, the population benefiting from the program and the level of investments far exceeded expectations (see Table 5A). The RDIP cofinanced 12,038 projects in the span of six years of which the bulk of funding was for agricultural extension services and infrastructure; 64 percent of funding went to the least developed municipalities, while less than 11 percent went to the most developed. No measurement has been done of the impact of the program on incomes or living standards, but survey evidence6 indicates a high level of satisfaction of beneficiaries and municipalities with the program's results. The majority of the beneficiaries (83 percent) expressed their satisfaction with the projects and considered that the program has had a positive impact on their well being mainly because of: (i) the improvement in the quality and coverage of services (63 percent); (ii) the improvement in income levels (50 percent overall, with 53 percent in the case of men and 43 percent in the case of women); (iii) the improvement in agriculture production levels (63 percent); and (iv) the better management of natural resources (54 percent). 8. The main activity aimed at increasing the incomes of small farmers, as part of the first objective, was support for the decentralization of the agricultural extension system. While there was a delay in putting the new system in place, by 1994 municipal units for agricultural-livestock extension services (UMATAs) had been established in nearly all municipalities. In 1995, the number of farmers served was double the SAR target and three times that of the previous extension system. Other productive projects accounted for a small percentage of expenditures and had mixed results. Efforts to help fishermen, recognized as one of the most impoverished groups, appear to have failed. As indicated above, beneficiaries believe that the RDIP has improved incomes and production. 5 ICR Caja Agraria Strengthening Project (Ln. 2909), Report No. 14690. 6 As part of the RDIP's evaluation, DRI undertook a survey in a statistically representative sample of municipalities (67) to get the opinions of municipal authorities (128 interviews), direct beneficiaries of projects cofinanced by RDIP during 1991-95 (550 interviews), and other municipal residents (irrespective of whether they were a beneficiary or not, 827 interviews). The results for the agricultural extension and infrastructure components are representative at the national level. In addition, a social assessment was carried out in three, poor municipalities.  9. Concerning the second objective, the program made an important contribution towards improvement of infrastructure in rural areas, with roughly 50 percent of RDIP funds for road improvement and construction, water and sanitation systems benefiting 450,000 rural residents, and rural electrification projects expected to benefit 61,000 families (see Table 5B). While there were some technical weaknesses in design and implementation, the survey indicated a high level of satisfaction. The third objective, the watershed component, was a relatively minor part of the RDIP and while it was implemented satisfactorily, the impact is difficult to measure. 10. Initially, the program made good progress in promoting the fourth objective of community involvement, but this was undermined by the onerous procedures and excessive time between subproject identification and implementation. The design of the RDIP also underestimated the amount of support required by communities. DRI was overwhelmed by institutional changes and attention to this fell off. However, this changed in 1996 when DRI led efforts to make the newly created Municipal Councils for Rural Development (CMDR) effective forums for community participation. Concerning the fifth objective, the strengthening of DRI's capacity occurred in the early years, but the institutional limbo and chaos which resulted from the introduction of the NCS seriously undermined it. Due to continuous change and uncertainty, the staff suffered from low morale and Fondo DRI's institutional culture of activism and commitment was muted, at least temporarily. The related goal of strengthening municipalities was partially achieved. The expansion of the program to the entire country strained DRI's capacity to attend to this. However, both the survey evidence and feedback from the government indicate that the RDIP and DRI are perceived as having made an important contribution to improving municipal capacity. In addition, through support to the NCS, the RDIP also contributed to building capacity at the departmental level which was totally unforeseen at appraisal. C. Implementation Record and Major Factors Affecting the Project 11. One of the factors affecting implementation was the RDIP's design, for which the two Banks and Government were jointly responsible. The major factors affecting the project were those within the Government's control, though some problems were not necessarily easy to foresee or handle given the ambitious and politically charged nature of the decentralization process. Some factors were within the control of DRI, but the incentives system in the public sector did not encourage attention to problem areas such as monitoring and evaluation. 12. The weaknesses in design were not detected immediately, since the program was launched more quickly than expected. Spurred by substantial demand from the municipalities which had completed participatory planning, Fondo DRI signed agreements for some 639 subprojects totaling US$17.8 million in the first nine months (drawing effectively on the retroactive financing provisions of the Loan Agreement). As early as the first year, however, operational limitations began to crimp DRI's effectiveness. These stemmed from a number of factors within the Government's control: (i) the design of the RDIP was overly bureaucratic and complex, thus inconsistent with the participatory  - 5 - approach;' (ii) tardy Government funding for recurrent expenses aggravated the problem of insufficient personnel to master the new approach and increased volume of work (going from signing 100 contracts a year with public entities to signing up to 2,000 cofinancing contracts per year with 600 municipalities); and (iii) onerous procedures held up disbursements from and replenishment of the Special Account (SA) managed by the Central Bank. Nothing was done about the first. Stop-gap measures were deployed to address the second such as using IICA convenios to hire staff and amending the loan agreement to allow financing of some operating costs related to supervision. Concerning the third, due to problems in the previous DRI operation, special measures were put in place to speed up replenishment of the SA. While these improved somewhat the operation of the SA, DRI found that use of a revolving fund in pesos (constituted from reimbursements received from the loan's retroactive financing) which it managed directly was the most efficient disbursement mechanism. To conclude, an evaluation of the RDIP in 1994 found that implementation had steadily improved and that by 1993 the new approach, including effective delegation of responsibilities to DRI's regional offices and mastery of procedures and instruments introduced by the RDIP, was taking hold. 13. However, by this time the Govemment had decided to change drastically the institutional arrangements for implementation of the RDIP without sufficient consideration of the negative impact on program implementation. The key institutional reform introduced by Decree 2132 which established the NCS was to assign responsibility for key operational decisions and activities of the Government's four cofinancing Funds to each of Colombia's 32 Departments (technical units called UDECOs and Departmental Cofinancing Committees presided by the Governor were to be established). It mandated the closure of DRI's regional offices by 1994 which was when the NCS was supposed to be operational, and required it to use a fiduciary to manage cofinancing funds. In addition, DRI was required to expand nation-wide (from 602 to more than 1000 municipalities, 21 to 32 Departments) and new subprojects were added to the menu of investments to be cofinanced. Around the same time, the Government also assigned Fondo DRI responsibility for new programs including the Plan Pacifico, Plante, and the Red de Solidaridad, which added to its burden. 14. The changes created voids in the project machinery. DRI was now dependent on third parties who, in the short term, lacked the capacity to carry out the program. In mid-1994, only five Departments had created UDECOs; the following year nearly all the cofinancing contracts were signed in December due to poor functioning of the NCS. The programming and budget cycle described in the SAR required 2 years between approval of a project idea and starting implementation; the reality is that 3 to 4 years was normal which undermined the participatory planning process as communities became frustrated. This was one of the key findings of the Evaluaci6n Institucional del Programa de Desarrollo Integral Campesino 1991-1993, prepared in 1994 by a consulting firm. The evaluation was based on analysis, including field visits and interviews, of a sample of 14 percent of the projects cofinanced by DRI between 1990-93 (600 projects in 65 municipalities). DRI's evaluation of the RDIP recognized that its practice (up to 1995) of establishing, at the central level, budget quotas per component/per municipality-department undermined the demand-driven aspect of the program by severely circumscribing local choice.  - 6 - Management of program funds (contracting, disbursing and some reporting) by a fiduciary institution created serious problems; this obligation was rescinded eventually by a Presidential Decree in December 1995. The central Inter-Fund Committee complicated decision-making and created uncertainty as to which decisions DRI's management could take independently. There were serious problems coordinating the four funds which had very different mandates and ways of operating. This improved by 1996 due to a clarification of the Committee's role and progress in establishing a common set of operating procedures for the funds. 15. A major implementation problem throughout was the lack of an adequate monitoring and control system to track progress in the subprojects being financed. Financial management in Colombia's public sector, as in DRI, is characterized by an extremely heavy system of ex-ante controls which make it difficult to spend, but limited follow-up once expenditures are made. The incentives system rewards managers and institutions that fully commit their budget before the end of the year, secondarily it rewards spending the resources. This played itself out in the RDIP by failure to develop and operate adequate monitoring and control systems. DRI was under enormous pressure to commit and spend its budget (especially as the crisis of the NCS ensued in 1995), but under little pressure to put in place an effective monitoring system, though annual audit reports were very critical of this weakness. Under countervailing pressure from the Bank, DRI made steady progress to address this problem, undertaking special measures to monitor problem projects in 1995-1996 and, in 1996, exercising sanctions for municipalities which did not comply with the terms of the cofinancing contracts. However, when the loan closed, a satisfactory system had been designed, but was still not functioning adequately despite considerable efforts by DRI to train, equip and provide technical support to the UDECOs and municipalities who were the primary providers of the information. 16. To its credit, Fondo DRI undertook, with support from the two Banks, remedial measures and developed some creative responses to the problems resulting from Decree 2132. These included: creating the centrally-based staff (UNIR) who acted as liaisons with the local governments (in lieu of staff in the regions which was prohibited), use of the media for disseminating information and contacting mayors to sign contracts, contracting of universities and NGOs to support municipalities and communities in project preparation, and hiring extra personnel to liquidate (audit and legally close) contracts. Based on DRI and the Government's efforts to address the problems affecting the RDIP and the belief that the program would have a satisfactory outcome, the Bank agreed to a one year extension of the Loan's Closing Date. During the last year, the loan supported DRI's efforts to improve the functioning of the NCS, evaluate the RDIP and formulate a new program, strengthen the CMDRs, and cofinanced 580 UMATAs. D. Project Sustainability 17. Project sustainability is uncertain both in terms of the individual investments financed and the institutional reforms supported. Concerning the main physical investments financed, the picture is mixed: (i) the prospects are reasonable for water and sanitation subprojects with survey results showing 50 percent of the facilities being  - 7 - operated and maintained by user committees; and (ii) the prospects are less certain for maintenance of road investments, a municipal responsibility, despite clauses in the cofinancing contracts which committed the municipality to reserve a certain amount for maintenance in its budget. Concerning the decentralized extension service, this is a long-term process of change, but the prospects for sustainability appear good. Municipalities have assumed a rising share of the cost of these services (60 percent in 1995); in many cases, mayors are actively supporting the UMATAs and farmers are becoming more vocal about their needs. However, there is some uncertainty as to whether and when the municipalities will be ready to fully finance these services without support from the central government; payments by farmers are taking place in a few cases. 18. Decentralization and community participation are central aspects of Colombia's 1991 Constitution. In this respect, there is a strong underpinning to sustain the involvement of municipalities and communities in rural development, important objectives of the RDIP. However, there is much uncertainty as to the sustainability of DRI as an institution since the recent announcements about the possible "fusion" of the four cofinancing Funds and eventual elimination of the NCS. E. Bank Performance 19. The Bank's overall performance was satisfactory. It worked closely with the Inter-American Development Bank during all phases of the project. During the project design process, the Bank supported the innovations embodied in the new decentralized and demand-driven approach proposed by the Colombians. However, in several respects the Bank's appraisal of the RDIP was deficient. This type of operation was still relatively new in the Bank; the RDIP's description in the SAR and consequently in the legal agreements tried to fit a new product into an old model, by making the program appear more concrete than it was (e.g. setting targets for yield increases). In hindsight, the project's design also suffered from the Bank's inexperience with participatory approaches. It could have been improved if it had: incorporated the ultimate beneficiary's views into the design of the program's operating rules; actively supported capacity building at the municipal and community level as an end in itself; and established a set of rules and procedures which accommodated the diversity of capacity and needs within Colombia. In addition, the Bank did not adequately assess DRI's capacity to triple its coverage and municipal capacity to implement the subprojects. While there were provisions in the legal agreements for more staff to manage the increased workload and for sanctions to be exercised if municipalities did not respect their contractual obligations, these were not fully respected. 20. Supervision was satisfactory. In hindsight, perhaps the Bank should have realized earlier on the serious impact Decree 2132 (adopted in December 1992 for implementation in 1994) would have on the program's implementation. However, due to a lag between subproject approval and actual implementation, throughout 1993 and 1994 supervision missions found evidence of good results on the ground and disbursements continued to be satisfactory. It was not until late 1994 that the negative impact of Decree 2132 became clear; the two Banks then assisted DRI to organize a workshop which analyzed the situation and how DRI could respond. This together with follow-up letters expressing  - 8 - serious concern got the government to focus attention on the NCS's problems and by mid-1995 it took some positive actions. Subsequently, the Bank sent a mission of staff from sectors affected by the NCS to analyze its problems.' The option of suspending operations was rejected, because DRI and the Government continued to making progress to address the problems affecting the RDIP and the Bank believed that the program's development objectives could be achieved despite the problems. In view of the innovative nature and complex history of this operation, a Performance Audit by OED could be extremely useful. F. Borrower Performance 21. Overall, the Borrower's performance is judged deficient including compliance with the financial covenants of the Loan Agreement (see Table 8). Colombia of the late 1980s and 1990s was a caldron of change. Needless to say, the uncertainties generated by the constantly changing rules of the game--new procurement code, budget law, revenue sharing and assignment of responsibilities among municipalities, departments and central level--and institutional reforms took a toll on the capacity to get things done which affected all Bank operations. The objectives of Decree 2132 were laudable: (i) bring consistency with the new Constitution, by designating a role for Departments as an intermediary between the central government and municipalities; (ii) simplify matters for municipalities so that they do not have to deal with four cofinancing Funds, each with its own rules; and (iii) provide for more flexible management of funds (exoneration from certain budget procedures) through use of a fiduciary. However, in view of widely recognized weak capacity in most Departments9 and inexperience of fiduciaries, the mandated changes were impractical in the short-term and adequate transitional arrangements were not provided for. Ironically, the impact of the decree was to recentralize DRI' s management and make it distant from the municipalities and communities who were the program's raison d'etre. 22. The Government was slow to recognize the problems of the NCS and, when it did, tended to blame the Funds. There appeared to be a lack of sufficient recognition in Government and political circles that institutional reforms take time to implement. There was thus a tendency to drop a reform without allowing sufficient time for it to take hold. Additionally, the NCS had become a political football in the debate about the decentralization process which was being blamed by some as the principal cause of the fiscal deficit. In late 1996, just as some positive accomplishments were coming out of the The mission concluded that while the guiding principles of the NCS were valid, the way it was being operated was not consistent with the principles enunciated in Documento CONPES 2791, Ajustes al Sistema Nacional de Cofinanciaci6n, junio 21, 1995. For example, the NCS was to cofinance investments considered a priority and within the mandate of local governments, yet members of the legislature were earmarking an increasing share of NCS funds for specific projects that were not necessarily endorsed by the local government concerned. Colombia Local Government Capacity: Beyond Technical Assistance (Rpt. No. 14085-CO, July 1995).  -9- experience of the NCS, both the reputable Public Finance Commission and Minister of Finance recommended its eventual abolition. 10 23. Overall, DRI's performance is also judged deficient, but admirable under the circumstances. Like other public sector agencies, DRI responded to the incentive to commit and disburse funds, without paying sufficient attention to financial management and monitoring and evaluation. Institutional instability and constantly changing rules during the RDIP made it difficult to get things done and eroded institutional capacity. The Fund performed well in many key areas and showed resilience in the face of much adversity (para. 16). However, it never overcame the lack of sufficient staff and time to adjust to the changed circumstances. Some analysts argue that within DRI there was resistance to the underlying philosophy of the RDIP which gave a central role to municipalities and communities; directors of the regional offices had enormous power and did much of the work which was to be done by others."1 With the shock of the NCS, some staff left and others "dreamed" of the old days when they were the vanguard of rural development. Just as DRI was settling down to its changed mandate, uncertainty about its future reemerged. G. Assessment of Outcome 24. The RDIP's outcome is judged satisfactory. Despite the many problems and obstacles described in Part C of this report, the RDIP achieved two of its four major objectives and partially achieved the others, contributing to a far-reaching process of change. The number of beneficiaries and level of investment far surpassed expectations, with the bulk of the resources going to the least developed municipalities. Beneficiaries and municipalities have expressed a high level of satisfaction with the program's results, including improved incomes and production (Tables 5A and SB), and the provision of agricultural extension services by municipalities. The prospects of sustainability are good for the UMATAs and water and sanitation projects, whereas they are uncertain for roads and for DRI as an institution. The RDIP was a pioneering operation for the Bank and has been a model for rural development investment funds being tried in other countries; the Bank and DRI have organized several study tours, at the request of other countries, to share the experience with the UMATAs and decentralization. H. Future Operation 25. The Government has requested that the IDB support a new phase of DRI's program, since it is already financing a number of other programs being implemented by DRI. Lessons from the RDIP are being incorporated into the design, including: greater emphasis on support for communities and municipalities, and financing of municipal investment plans instead of individual projects. In the meantime, DRI has agreed with the 10 One reason offered by the Commission was the growing practice of earmarking NCS funds by legislators; earmarked funds accounted for about 30-40 percent of DRI's investment budget for 1996. It seems that some staff of Fondo DRI used their power to allocate resources to further their aspirations for political office. Some analysts say that this was an underlying factor motivating the Decree.  - 10- Bank on a plan of work to complete the liquidation of the cofinancing contracts financed by the loan, and to undertake in 1998 a full impact evaluation of the RDIP (measuring results against the base-line established in 1991). As part of the Bank's Country Assistance Strategy, it has agreed to support the Government's efforts to reduce poverty through promoting rural development including land reform. Proposed Bank sector work should contribute to the Government's thinking about the appropriate role, institutional arrangements, and financial mechanisms for central government support for rural development, particularly in view of uncertainties about the future of the NCS and DRI. I. Key Lessons Learned 26. Some key lessons of the RDIP are as follows: * a participatory and decentralized operation which is to finance many small subprojects in a large number of localities, cannot work if it is overburdened by the bureaucratic, centralized rules which characterize normal operations of the Colombian Government and World Bank; likewise, a shift in focus from ex ante controls to ex-post controls and sanctions is likely to be more conducive to timely implementation and building accountability at the local level; - a flexible, open design is needed for centrally managed programs targeted to local governments due to the enormous diversity of capacity, needs and socio-economic conditions in Colombia; - the process of designing this type of operation should be highly participatory, drawing on the contributions of local communities and municipal governments to define the basic operating rules which are pertinent and consistent with local capacity, while ensuring an adequate level of transparency and accountability in use of public funds; * a simple monitoring and control system should be in place from the start in this type of operation in order to rapidly identify and correct problems, and to identify local govemments which require special assistance or, in extreme cases, disqualification from the program; * it is advisable to separate monitoring and evaluation functions, and rely on an independent agency such as a university (which was a successful aspect of DRI II) to conduct periodic impact evaluations; * in view of the Colombian Government's rules concerning management of Special Accounts, the experience of this project shows that they do not fulfill their intended purpose and should be dropped; and * major institutional and political reforms, such as those undertaken in Colombia since the conception of the RDIP, are long-term processes and there are bound to be set-backs and unexpected developments; the Bank needs to exercise patience and flexibility to support this process and judge the ultimate results in a time-frame that goes beyond a single lending operation.  - 11 - Table 1- Summary of Assessments A. Achievement of Objectives Substantial Partial Negligible Not applicable Macro Policies 0 0 0 El Sector Policies 0 0 0 0 Financial Objectives 0 0 0 0 Institutional Development 0 0 0 0 Physical Objectives 0 0 0 0 Poverty Reduction 0 0 0 0 Gender Issues 0 0 0 0 Other Social Objectives 0 0 0 0 Environmental Objectives 0 0 0 0 Public Sector Management 0 0 0 0 Private Sector Development 0 0 0 0 Other: Decentralization O 0 0 0 B. Project Sustainability Likely Unlikel] Uncertain (/) (v') (/) O 0 0 C. Bank Performance satisfactory Satisfactory Deficient (v') (/) (/) Identification 0 0 0 Preparation Assistance 0 0 0 Appraisal Ol 0 0 Supervision 0 0 0 Highly D. Borrower Performance satisfactory Satisfactory Deficient Preparation 0 0 0 Implementation 0 0 0 Covenant Compliance 0 O 0 Operation 0 0 0 Highly Hijy E. Assessment of Outcome satisfactory Satisfactory Unsatisfactory unsatisfactory 0 0 0 0  - 12 - Table 2: Related Bank Loans Loan/credit title Purpose Year of Status approval Preceding operations 1.2nd Integrated Rural Development Rural development 1982 Closed 2.Small-Scale Irrigation Small scale irrigation 1989 On-going schemes 3.Second Rural Roads Sector Improvement of roads 1990 Closed in rural areas Following operations I.Natural Resource Management Forest & watershed 1994 Ongoing management, support decentralized system to manage environment. 2.Agricultural Technology Reform of system for 1995 Ongoing agricultural technology generation and transfer. Table 3: Project Timetable Steps in Project Cycle Date Planned Date Actual/ Latest Estimate Identification (Executive Project Summary) - September 15, 1988 Preparation Feb-May 1989 Appraisal __November 1989 Negotiations __May 1990 Board Presentation July 10, 1990 July 31, 1990 Signing September 14, 1990 Effectiveness January 31, 1991 Project Completion June 30, 1995 December 31, 1996 Loan Closing December 31, 1995 December 31, 1996  Table 4: Loan Disbursements: Cumulative Estimated and Actual (US$ thousands) FY91 FY92 FY93 FY94 FY95 FY96 FY97 FY98 Appraisal Estimate 6,000 13,500 25,500 36,700 52,500 66,700 74,500 75,000 l Actual - 7,084 22,191 32,766 49,442 63,012 75,000 - Actual as % of Estimate - 52 87 89 94 94 100 - I Date of Final Disbursement: 6/3/97 l  - 14 - Table 5A: Key Indicators for Project Implementation INDICATOR OPERATIONAL RESULTS Stimulation of municipal investment proposals 12,038 cofinancing contracts signed by Fondo DRI Subprojects are implemented on timely basis 45% of projects completed and/or operating by December 1996; only 8% unfinished (survey) Numbers Benefited Program activities benefited (# of persons): Women's groups: 10,000 Water projects: 455,000 Rural electrification: 61,000 families Community development: 84,000 Agricultural extension: 545,000 Fisheries: 20,040 Rural markets: 17,918 50% of water and sanitation projects are operated Subprojects sustained by community over time and maintained by community groups (survey) Involvement of private sector agencies in project 24% of sub-projects executed by private sector activity (survey) Focusing program funds on poorest areas 64% of funding allocated to least developed municipalities (categories 1-3) Decentralization of agriculture extension service 1,022 UMATAs established serving 97% of Colombia's municipalities, with 3,539 professionals (ratio I agent: 154 farmers) Departmental units for handling DRI program CODECOs and UDECOs are functioning in all locally departments with varying degrees of competence Attention to gender issues 1,053 rural women's groups formed, 10,281 training sessions held Transparent, participatory involvement 83% of the beneficiaries expressed satisfaction with extent of community participation under the program (survey) Participation by stage of project cycle 83% of the beneficiaries participated in identification of projects; 60% in definition of priorities; 43% in project execution; 89% said they were informed of monitoring results during execution (survey) Attention to indigenous communities' needs Proposals from indigenous groups awarded highest level of cofinancing; 2% of DRI budget targeted for these projects Note: The SAR did not include key indicators in part due to the nature of the operation. The projects cofinanced and executed were.part of a program of interventions based on a demand-driven mechanism, whose costs, benefits and rates of return could not be determined ex ante.  - 15 - Table 5B. Key Indicators for Subproject Accomplishments Total No. of Project Type Project Output No. of Beneficiaries Agricultural 4,527 1,022 UMATAs established 545,000 Extension Technology Transfer (included above) 1.027 technology trials and Technical 9,320 demonstration plots established in Assistance 1995 999 extensionists' training events Marketing 775 6 regional markets opened 17,918 (direct) 98 other types of market infrastructure built 246 municipalities served Fisheries 65 12 support centers opened, training 34,663 Fish Ponds 327 16,087 reservoirs built 1,628 training sessions 20,040 Watershed 572 Reforestation and better management Management protecting almost 24,429 ha. 1,043 training classes 23,663 groups Irrigation 289 39,463 ha irrigated Roads 1454 2,576 kms newly built; over 4,000 km improved Water and Sanitation Systems 1,917 1,994 facilities built 449,824 Electrification 632 8,497 connections 61,000 families Community 395 10,281 training events involving 1,569 84,055 Development groups Womens Groups 356 1,053 organized and supported 10,000 (estimate) Local Institution 1,018 212 training events l Building 1,143 pre-investment studies UDECOs established in all 32 Departments Note: The SAR did not include key indicators. The Economic Rate of Return for the productive subprojects (extension, marketing, fisheries, fish ponds) was not recalculated as part of the ICR.  - 16 - Table 6: Studies Included in Project Purpose as Defined Study at Appraisal/Redefined Status Impact of Study Institutional Mid-term Evaluation of project Completed Provided important insights Evaluation performance focusing on into weaknesses in the institutional aspects (1991-93). project design and operation, particularly concerning participation aspects, the role of municipalities, and the marketing component. Governmental policy changes taking place at the time of the study diminished its impact, but it did influence thinking on design of DRI phase IV. Assess Policies on To assess implementation of Completed Useful, but creation of NCS Graduation of Municipalities graduation policies for phasing with all municipalities out municipalities from the eligible for financing program. irrespective of level of development made the policy less relevant. Baseline Study to Evaluate To develop a method to measure Completed Some significance. DRI RDIP Impact project impact through plans to use the baseline to systematic surveying in the do the formal evaluation of different regions covered by the impact in 1998, though the program. sample is for the original 602 municipalities.  Table 7A: Project Costs Appraisal Estimate (US$M) Actual/Latest Estimate (US$M) Item Local Foreign Total IBRD IDB Colombian Total Costs Costs Contribution 1. Agricultural Production and 56.1 20.4 76.4 39.1 30.9 197.7 267.7 Environmental Protection 2. Infrastructure 33.9 73.3 107.3 26.2 38.6 253.8 318.6 3. Training and Community 16.7 4.2 20.9 2.6 4.0 4.9 11.5 Organization 4. Strengthening DR] and 7.2 1.5 8.7 7.1 1.5 35.0 43.6 Municipalities 5. Physical Contingencies 5.7 5.0 10.7 6. Price Contingencies 7.8 18.3 26.1 TOTAL 127.4 122.6 250.0 75.0 75.0 297.6 641.4 Due to rounding, totals may not add up exactly.  Table 7B: Project Financing Appraisal Estimate (US$M) Actual/Latest Estimate(US$M) Source Local Costs Foreign Costs Total Total IBRD 13.7 61.3 75.0 75.0 IDB 13.7 61.3 75.0 75.0 Government (DRI) 37.6 37.6 193.8 Municipalities & other 62.5 - 62.5 297.6 X Agencies TOTAL 127.4 122.6 250.0 641.4 Note: The IDB loan will continue disbursing throughout 1997, though only a small amount remains.  - 19 - Table 8: Status of Legal Covenants COLOMBIA Rural Development Investment Program Section Covenant Present Original Revised Description of Comments type status fulfillment fulfillment covenant date date Loan 3.01(a) 12 CP - Borrower shall cause DRI Borrower did not allow Agreement to perform in accord with DRI to adequately Project Agreement (PA), increase its staffing. did will take any action not take action to get the necessary or appropriate to financial and monitoring enable DRI to perform its problems addressed, and obligations under PA, and by adoption of Decree will not take/permit any 2132 took an action with action which will prevent serious negative or interfere with such repercussions on DRI's performance. ability to respect PA. 3.04 5 CD 11/30/92 - Govemment, through Study managed by DRI. DNP, hire consultants to Delayed and with Bank's prepare study of RDIP, agreement combined with submit to Bank for review, mid-term evaluation. agree on action plan to Study carried out in 1994 implement and early 1995. completed recommendations by June 1995. 11/30/92. Ensure DRI implements action plain. 3.05 12 CD 9/30/92 - Govemment to develop a Policy introduced graduation policy for February 1994. phasing out municipalities from program; thereafter implement. 3.06(a) 12 CD 9/30/90 - Govemment cause ICA to Plan submitted November submit action plan and 1992. timetable, satisfactory to the Bank, for establishing additional 19 CRECEDS; and thereafter implement such a plan. 3.06(b) 12 CD 9/30/91 Govemment to submit New system of UMATAs action plan to phase ICA %Nas fully in place by end out of extension. Make of 1994. arrangements to provide extension services to RDIP beneficiaries. 4.01 (a) I CP Cause DRI to maintain Auditors repeatedly records and accounts in criticized lack of adequate accord with sound intemal controls. Though accounting practices. improvements were made. weaknesses remained at end of RDIP 4.01 (b) I CD 6/30 each Cause DRI to submit Reports submitted late. yr annual audit report.  - 20 - Section Covenant Present Original Revised Description of Comments type status fulfillment fulfillment covenant date date 5.01 13 Specific conditions which A number of these can trigger suspension. conditions did arise, but the Bank choose not to exercise its rights under this clause. Project 2.01(c) 5 CP DRI shall not fail to DRI introduced sanctions Agreement enforce the Contratos de and began to enforce the Cofinanciacion or any contratos late in the provision thereof project. Prior to that monitoring information was insufficient to detect problems in a timely fashion. 2.04 9 C 6/30/91 - DRI to complete first Baseline survey 6/30/95 baseline survey, to hire completed in 1991, but 10/31/95 consultants acceptable to Bank accepted DRI's the Bank, and to argument that it was too complete second baseline early to conduct second survey and the impact baseline and impact evaluation. evaluation, in view of major disruptions of NCS. Impact evaluation planned for 1998, meanwhile in-depth evaluation of RDIP done in 1996 for ICR. 2.08(b) 3 C DRI to make suitable Initial disbursements arrangements to ensure lagged under IDB loan that Bank and IDB loans due to delayed start-up. disburse at about the Measures were taken to same rate; to be reviewed accelerate pace of each September. disbursements to achieve pan pasu. 3.01 5 CP DRI shall carry out its Serious problems operations with sound detected related to administrative, monitoring and internal managerial etc. practices controls (though under supervision of corrective actions taken) qualified management and not have adequate assisted by competent number of staff staff in adequate throughout much of numbers. RDIP. 4.01 1 CD - - DRI to submit audit See Loan Agreement, reports by June 30 each para. 4.01 (b). year. Schedule CP DRI municipalities will Proved difficult to 1, para. provide for periodic road monitor; while provisions 3 (b) maintenance; failure to do made in the cofinancing so will preclude their contracts, eligibility eligibility for future apparently not denied. financing of roads as well as ag. extension, rural water supply and san. projects.  - 21 - lmSection Covenant Present Original Revised Description of Comments a lrtype status fulfillment fulfillmect covenant l l ~~~ ~~~~~~~~~~~~date datel l ~~~~Schedule NC DRI should impose Cases of outright misuse | l I~~~~, para. sanctions, deny of funds were not l l ~~~~10 participation to detected during audits,| l ~~~~~~~~~~~~~~~~~~~~municipalities that not field visits by Banks, andl l ~~~~~~~~~~~~~~~~~~~~~respect obligations under special control efforts. | cofinancing contracts. However, DRI failed to detect problems in a timely manner and exercise sanctions until very late in the RDIP. Schedule 2 CD - - DRI to submit Annual Delays accepted by Bank. para. I(a) Operating Plan by AOP proved unwieldy September 30 each year. throughout program; proved to be unsuitable instrument after introduction of NCS. Schedule 2 CD - - DRI to submit annual Delays occurred para. progress reports by frequently, but reports of 1(b) September 30, and satisfactory quality. biannual progress reports by June 30 and December 31, each year. Schedule 2 CP - DRI to retain adequate Not done due to para. staff. restrictions imposed by 4(a) Government. Ad-hoc solutions done. Schedule 2 C DRI to maintain a - para. coordinating unit. 4(b) Covenant types: 1. = Account/Audits 8. = Indigenous people 2. = Financial performance/revenue generation from 9. = Monitoring, review, and reporting beneficiaries 10. = Project implementation not covered by categories 1-9 3. = Flow and utilization of project funds 11. Sectoral or cross-sectoral budgetary or other resource 4. Counterpart finding allocation 5. = Management aspects ofthe project or executing 12. = Sectoral or cross-sectoral policy/ regulatory/institutional agency action 6. = Environmental covenants 13. = Other 7. = Involuntary resettlement Present Status: C = covenant complied with CD = complied with after delay CP = complied with partially NC = not complied with  - 22 - Table 9: Compliance with Operational Manual Statements Statement Number and Title Describe and comment on lack of compliance OP 10. 02 - Financial Management Financial management aspects were not fully complied with. Poor internal controls in DRI's accounting system and insufficient monitoring resulted in delays and inaccurate reporting. Table 10: Bank Resources: Staff Inputs Stage of Planned Revised Actual Project Cycle Week US$ Weeks US$ Weeks US$"' Preparation to Appraisal n/a n/a n/a n/a 66.5 226,100 Appraisal n/a n/a n/a n/a 139.1 472,940 Negotiations through Board Approval n/a n/a n/a n/a 18.8 63,920 Supervision n/a n/a n/a n/a 138.0 469,200 Completion n/a n/a n/a n/a 9.0 30,600 TOTAL 541.4 1,262,760 1/ Assumes US$3,400 per staffweek  - 23 - Table 11: Bank Resources: Missions Perfomiance Rating Number Specialized Implemen- Develop- Stage of Month/ of Days in Staff Skills tation ment Types of Project Cycle Year Persons Field Represented Status Objectives Problems Through Appraisal and 11/88 2 5 A, I Board Approval __ 5/89 6 10 A, F,O _ 12/90 2 5 A,O _ _ Supervision 3/91 7 3 E, A, F, D, 0 1 9/91 5 10 E, F, A, 1 5/92 7 10 E, A,O 1 12/92 4 10 E, EN, O 2 8/93 2 10 E, F 2 5/94 3 10 E, F, A 2 6/94 1 4 F - - 12/94 4 10 E, A, S S S 6/95 6 5 E, A, F, S S S 8/95 1 3 E 12/95 4 10 E, A, F, S S S Completion 10/96 1 10 S S S 1/97 2 6 E, S S S Skills Represented: E = Economist I = Institutional Development Specialist F = Financial Analyst EN = Environmental Specialist A = Agriculturalist 0 = Operations/Disbursement Analyst S = Social Scientist  - 24 - Appendix A Translated from Spanish RURAL INVESTMENT COFINANCING FUND - DRI EXECUTIVE SUMMARY OF EVALUATION OF THE RURAL DEVELOPMENT INVESTMENT PROGRAM - RDIIP, 1991-1996 1. The following is a summary of those aspects that DRI considers most significant from its report Evaluation of the RDIP. This report was the result of a complex exercise including an internal evaluation within DRI, a social assessment with different actors involved in the program, and a quantitative evaluation performed through a survey carried out in a representative sample of municipalities. The findings presented here are based on detailed reports about each of the subjects and program components considered important in the aforementioned evaluation. A. Financial evaluation 2. The amount of funds committed and transferred during the 1991-1996 period (see Tabl_ 1, Annex) shows that the Program exceeded by more than 2.5 times the original financial goals. In fact, during this period, DRI cofinanced subprojects under the RDIP for an amount equivalent to US$641.4 million (US$343.8 million contributed by Fondo DRI including the IDB/IBRD loans of US$75 million each, and US$297.6 million contributed by the municipalities, communities and others). The total amount spent (value of completed projects plus amounts disbursed by DRI for projects under implementation) as of December 1996 was US$356.0 million' for the 12,038 cofinancing contracts which were signed during 1991-96. 3. In 1990, the most optimistic forecasts for the RDIP were that Fondo DRI would sign close to 1,500 cofinancing contracts annually (2 or 3 per municipality per year). Already in 1993, DRI exceeded this estimate by 30%, reaching a level of more than 2,600 contracts per year in 1995 and 1996. In the RDIP's six-year period of implementation, Fondo DRI signed more than 12,000 cofinancing contracts, almost twice the expected number. 4. However, there was no lack of problems, particularly related to the participatory planning process as originally designed. In fact, in order to accelerate the process of establishing the Municipal Projects Banks (Banco de Proyectos Municipales) where projects must be registered prior to DRI's signing a cofinancing contract, the previously defined consultation process with communities became lax (theficha veredal was This refers to the liquidated (audited and legally closed) cofinancing contracts as of December 1996 corresponding to the 1991-94 period (close to 82% of all commitments made have been liquidated, with payment of balance still pending ) and payments (drawn or disbursed ) made by DRI as of December 1996 for cofinancing contracts signed in 1995 and 1996 (projects under execution) with the local or municipal counterpart contribution not yet entered into the books. Therefore, these figures underestimate the real amount financially "executed".  - 25 - dropped, the DRI Committees were reduced to almost nothing, and the fulfillment of the project eligibility requirements contained in the Operating Rules-Reglamento Operativo were not properly enforced). B. Analysis of the targeting of investment 5. The RDIP's targeting mechanisms which, paradoxically, became more appropriate and necessary when the program's coverage was extended to the whole country and new components were introduced, were not used by DRI to their full potential. If there had been a more sincere and transparent use of these targeting mechanisms, the RDIP might have been able to achieve a greater impact in more backward areas, in greater need of this type of assistance. Even so, as explained below, the RDIP achieved an acceptable targeting of investment by component and by category of municipality (as defined by the National Cofinancing System). 6. Targeting by component. Table 2 (see Annex) shows the committed resources of the RDIP disaggregated by component (type of subproject). There are only slight variations in the percentage per component compared to the original plans, due to DRI's practice of setting budget allocations by component rather than relying on the real and effective demand for investment of the municipalities themselves (one of the original principles of the program). In other words, the fact that there are no great differences between the percentages of funds programmed for various components (subprojects) and the amount executed is not by chance or due to good guesses about municipal demand during the initial programming exercise, rather it was predetermined by the annual programming system. In absolute terms, all components received a greater amount of resources in comparison with the amounts originally planned. 7. Targeting by category of municipality. Similarly, an adequate distribution of resources by category of municipality is noted (see Table 3, Annex). The less developed municipalities (Categories 1, 2 and 3), representing a little over 51% of all municipalities, received 64.2% of the resources, while the more developed ones (Categories 6, 7 and 8), representing around 17% of the total, only received 10.3%. These figures reflect the Program's redistributive effect and demonstrate that most of the resources were concentrated on the poorest municipalities. C. Institutional and normative aspects 8. Contribution of the RDIP to the decentralization. The process of decentralizing rural development in Colombia began in 1987 when, through Decree-Law 77/87, new responsibilities for rural development were assigned to municipalities. Also, Fondo DRI was assigned the new function of assisting municipalities in this process through the cofinancing of programs and projects. The planning and design of the RDIP was made on the assumption that the decentralization process, particularly its start-up, was going to be slow and difficult. However, the decentralization advanced far more quickly than expected, even by the most optimistic forecasts.  - 26 - 9. In effect, the enactment of the new Colombian Constitution in 1991 ratified and deepened the principles which had been the origin of the decentralization of rural development starting in 1987. Thus, for instance, in addition to increasing to three years the period in office of mayors, the Constitution provided for the popular election of Governors to head the departments. Consequently, the Government had to assign to the departments, new responsibilities in the direction and orientation of rural development, some of which clearly competed with responsibilities of the municipalities and/or even the central government. 10. As a result of these changes, in December 1992, the Government enacted Decree 2132/92: (i) creating the National Cofinancing System (NCS); (ii) changing the old Fund for Integrated Rural Development - DRI into the Rural Investment Cofinancing Fund - DRI. The Decree required, among other things, the dismantling of DRI's Regional Offices in the departments and cleared the way for departments to participate in the administration of the NCS (for which, later on, the Specialized Departmental Cofinancing Units-UDECOs were created). In addition, also in fulfillment of mandates in the new Constitution, in 1993, Congress passed Law 60 which established regulations concerning the attributions and fiscal transfers to municipalities and departments. So, scarcely two years after starting the RDIP, the norms and institutional framework had undergone a profound transformation that had direct consequences for its implementation. 11. In the context of this decentralization process, the RDIP is recognized as having -made contributions in the following areas: * Successful experiences in participatory planning and democratic management of rural development in municipalities throughout different regions of the country. * Clear understanding among the principal actors of the NCS about the central role and importance of the project as a planning unit. * Greater capacity in a majority of municipalities to formulate and present, in a timely manner, their projects to the NCS. * More direct and effective pressure from municipal residents regarding the management of public resources, forces which require mayors to generate some community backing when defining their actions. * Existence of more than 700 Municipal Councils for Rural Development which, although still in their beginnings, constitute the seed for the new institutional arrangements for promoting and managing the development of rural societies in a more democratic manner. * A basic and explicit concern of communities to guarantee adequate supervision of the projects (this is one of the most important indicators taking into account that Colombia has no experience or interest in the subject). 12. Institutional adaptation of Fondo DRI. Two formal reorganizations of DRI were undertaken during the program: (i) in 1989, for the start-up of the RDIP, and (ii) in 1993, for the start-up of the NCS which required closing DRI's Regional Offices. In addition, in  - 27 - 1995, a functional reorganization of the entity took place through the creation of working groups specialized in critical subjects; this was aimed at reactivating and redimensioning the relations between the center (DRI-Bogota) and the regions. It should be recognized that DRI developed a great capability during the implementation of the RDIP to identify alternative ways of operating which would efficiently respond to the changes imposed from outside and for reasons unrelated to Fondo DRI's mission; these changes were made without allowing time for DRI to consolidate any of the previous institutional changes. Despite this, Fondo DRI succeeded in effectively maintaining its role as a leader in rural development. 13. Efficiency and effectiveness of the institutional framework. DRI ran the RDIP with adequate institutional efficiency, in so far as it succeeded in exceeding the expected results with regard to the number of subprojects contracted and resources transferred. However, from the point of view of institutional effectiveness, despite the fact that the results are not negative according to the perception of the beneficiaries and some intermediary actors, it must be pointed out that the internal difficulties experienced from 1991-96 (particularly, the ambiguity and lack of definition concerning institutional responsibilities for the quality and efficiency of the programs) had negative repercussions on the institutional climate in which DRI had to carry out the RDIP. 14. Similarly, the tension--and in many cases, the conflict--generated between the central and regional levels (first with DRI's Regional Directors, and later on with the departments and UDECOs) as part of the decentralization process (meaning, transfer of real power over planning and allocation of resources) became a great difficulty for the RDIP's implementation that had direct repercussions for the independence of the local and municipal authorities to decide which projects would be supported. D. Community participation in the RDIP 15. The instruments designed to facilitate community participation in decision-making processes, which was one of the fundamental precepts of the RDIP, did not operate in the manner originally anticipated. The ficha veredal (village data sheet) and the DRI Committees, which were put into place at the end of Phase 11 (1988 - 1990), did not hold up adequately to the magnitude and scope of the new dynamic generated by the decentralization process and the countervailing struggle of the old political actors to maintain regional hegemony (in contrast to the new actors who now began participating in the local and regional political game). Thus, none of these instruments was capable of stopping, or even reducing, the effects of this dynamic; although the instruments continued to operate for some time (1991-1993), this was simply to satisfy a requirement, and, after a short while, they were forgotten. Later, in 1994 the Municipal Councils for Rural Development (CMDR) were introduced. Although, they are better structured--to take into account the different local political interests--they have not yet become a central feature of municipal life, with the exception of a few regions where the political culture favors the full and open participation of the people.  - 28 - 16. The decentralization of decisions in order to bring them increasingly closer to the communities did not produce the expected results. Although in the process the central bureaucracy of Fondo DRI yielded and lost much discretionary power, the decision- making power was not effectively transferred to the rural communities; it remained stuck halfway (at the regional, departmental or municipal level) in the center of the political struggle for the control and management of the program's resources and, obviously, of the corresponding electoral "dividends". The clearest demonstration of the struggle to handle and control these resources was the appearance since 1994 in Fondo DRI's budget appropriation of specially assigned or earmarked funds to cofinance projects endorsed by members of Congress. 17. Under these circumstances, community participation in the RDIP was mainly limited to consultation for the identification of the projects and, more recently and to a lesser degree, to prioritizing projects. The decisions about planning, implementation and contracting and monitoring are taken by other actors, mainly local politicians with varying degrees of commitment to the community. The latter depends on several factors, with the regional political culture being particularly important: the more traditional and paternalistic the relationship between politicians and voters, the more passive, dependent and scarce is the community participation. In those regions where the community has achieved greater levels of institutional development, there is less political patronage and community participation is more active and decisive. E. Accomplishments of the Monitoring and Evaluation System 18. Although the general tone of observations made about the monitoring and evaluation system seem to show dissatisfaction, it is also recognized that the efforts made until now by DRI to standardize and systematize these processes are important accomplishments. The development of instruments to identify, plan and evaluate projects as part of the monitoring and evaluation system was a positive step. However, the municipalities are not wholly capable of managing the designed instruments and, therefore, the information they collect is not flowing adequately and does not properly serve the purposes of the monitoring and evaluation system at the central level of Fondo DRI or at the level of the departments/UDECOs. The system is even less effective as a tool to help the municipalities improve their management of cofinanced projects. 19. DRI has the technical and operational design of a monitoring and evaluation system for rural development projects, and great progress has been made in the design of a system for the computerized handling of basic internal management information. However, neither system satisfies DRI's requirements for handling and systematizing management information. DRI is not utilizing the potential of these systems to improve its overall performance, nor is it using them to support efforts of local governments to improve the planning, management and execution of the rural development projects for which they are responsible.  - 29 - F. Accomplishments in environment management 20. Environment management in the RDIP advanced well with regard to the development of basic procedures and technical support for DRI as an entity and for the local governments presenting projects to be cofinanced. Also, the investments made through the Micro Watershed subprogram are recognized as having made a positive contribution to protecting, preserving and/or recovering natural resources. Yet, notwithstanding the important accomplishments and the short time DRI has been involved in environment matters, some difficulties continue to exist in this area--both inside DRI, in the projects, as well as in the "beneficiary" population and its natural setting. G. Main accomplishments of the RDIP's components 21. Institutional strengthening. This component contributed to the successive reorganizations the entity had to tackle, as well as to: the development of training and support strategies in the regions (in particular, the establishment of the "project" culture); and to the dissemination in all municipalities of the principal instruments for regional and local planning needed to get access to NCS resources. However, the accomplishments and results were not totally satisfactory due to the following reasons: (i) continuous change in the rules of the game prevented the development and consolidation of proposals; (ii) actions were unfocused and dispersed; (iii) high discretionary power of DRI's central authorities in the programming of the resources; (iv) lack of leadership to direct the component; (v) excessive emphasis in the planning stage and negligence during the execution stage; and (vi) lack of an adequate methodology to evaluate the component. 22. Rural organization and training. The main contribution of this component was the support given to the formation and consolidation of the CMDRs. The following obstacles affected this component: (i) execution of the component was disconnected from the concept contemplated when it was originally formulated; (ii) the characteristics (more than the weaknesses) of the control, monitoring and evaluation system prevented a feedback process; and (iii) the Association of DRI Users (ANDRI) failed as an organizational option for small farmers; and (iv) the ficha veredal created more expectations than results. 23. Rural woman. According to the survey, women considered the main success factor of this component was the incorporation of activities contributing to their personal improvement; they also specially recognized the role played by the UMATAs in the success or failure of the projects. Despite the progress of Fondo DRI, it was recognized that discriminatory conditions continue to exist in the rural sector. There are not enough explicit and attractive incentives for the municipal authorities and rural communities to guarantee the incorporation of women's needs and potentialities in the municipal plans for rural development.  - 30 - 24. Technological Development. The most remarkable result of this component, and without any doubt one of the RDIP's most important ones, is to have contributed more than any other Entity or Government program to the establishment of the municipal system of free extension services to small rural producers through the UMATA. As a matter of fact, one of the few successes so far obtained--probably the only one--in the process of political and administrative decentralization of the public agricultural sector in Colombia, was to get the municipalities to assume responsibility for this service. The contribution of Fondo DRI in this respect was fundamental with regard to funds as well as to technical orientation through the norms for obtaining financing. 25. Marketing. The performance of this component constituted, again, the great disappointment of RDIP's rural development strategy. The underlying strategy was to support processes to increase supply in coordination with demand in regional markets and terminals in sufficient volumes to achieve efficiency. More than any other component, this one faced interference of many actors and interests that deviated the subprojects from the intended objectives. The participation of the beneficiaries was limited to the initial consultation processes, while planning, approval, execution and operation were concentrated in the hands of RDIP staff (at the local, regional as well as national level) and, to a lesser degree, the leaders of the participating organizations. 26. Micro Watersheds. Reforestation (not integrated) projects prevailed, due to the greater acceptance of this type of project in the communities. As an accomplishment, it is important to emphasize the implementation of projects to mitigate environmental impacts, although they continue to be insufficient in relation to needs. Despite this progress, there are still numerous and important obstacles that must be solved in order for the municipalities to assume fully their responsibility regarding environment management, territorial ordinance and the preservation of natural resources in rural zones. 27. CESPAS - Service Centers for Non-Industrial Fishing. Another of the RDIP's great disappointments was the execution of the CESPA strategy. The problems were similar to though more serious than the agricultural marketing component and less was accomplishment with very unsatisfactory results. The fishermen's cooperatives supported by the RDIP were not capable of assuming the administration of the Centers, much less to operate and maintain them according to economic profitability criteria. The participation of the cooperatives was more formal than real; for that reason, as soon as the support of the executing-promoting entity (NGO) was over, the projects declined rapidly. 28. Fish Ponds. Fulfillment of the financial and physical goals was quite satisfactory. This activity provided an alternative source of supplementary financial and food resources quite attractive for the small rural producers. In addition, it permits a more rational use of the productive resources of the small rural producers, reducing the "dead" periods when manpower of the rural family is available. 29. Rural Roads. According DRI's central staff, this component succeeded in improving the competitive conditions of hundreds of thousands of rural families through  - 31 - better access to input and product markets. In the regional workshops held as part of the evaluation, as well as in the results of the survey of beneficiaries and mayors, it was found that this was the most important and recognized component of the Fondo DRI in its entire twenty year history. According to the survey, 79% of the beneficiaries consider that the roads subprojects are functioning adequately. 30. Water Supply and Basic Sanitation. Historically, this component is considered by beneficiaries and mayors to be one of the greatest contributions to rural development. According to the survey, 77% of the beneficiaries consider that the subprojects are functioning adequately; benefited the neediest population (55%); improved the quality of services (56%); and improved service coverage (54%). 31. Rural electrification. This component was included in the RDIP in the last three years. Since electric power is one of the greatest needs for rural development and given that few entities either public or private are backing rural electrification, Fondo DRI's resumption of financing for this type of project has been very much appreciated. H. Performance of the Fondo DRI Fund in the Implementation of the RDIP 32. The RDIP's implementation was carried out in an agitated international and national context and in the middle of profound changes in the institutional framework. The institutional development of DRI was amply exceeded by changes in the framework, so that the only thing permanent during the RDIP's implementation period was change itself 33. The planning and execution of the RDIP were done within the context of a decentralization process that meant a profound change in the responsibilities and relations between the different levels of government (Nation, Department, Municipality) and between the State and civil society. In effect, the responsibility for rural development was shifted from the central government to the municipalities, and the mechanism for access to resources is now based on demand and free competition between public and private institutions. The central level now performs an advisory and monitoring function for local initiatives. In this way, the RDIP cannot be evaluated based on the original plans, because its relationship with and degree of control over those implementing the program substantially changed. 34. The evaluation of DRI's performance in the implementation of the RDIP by the different actors is not homogeneous. These are some of the conclusions: * Self-evaluation (DRI) As a result of a self-evaluation workshop and a survey of managers (deputy directors and division chiefs), the following were considered to be the principal accomplishments: definition of the mechanisms for access to the cofinancing resources and high capacity to commit resources; introduction of the rural development concept at all levels of government, important increase in the service coverage by the State through DRI,  - 32 - development of instruments for identification, planning and evaluation of rural development projects; institutionalization of the UMATA; and start of environmental work. The following are the views among DRI staff on the difficulties encountered: lack of quality of the projects; lack of ownership of the RDIP's objectives within DRI; intervention of strongly decisive political elements at regional and local levels; resistance of DRI's staff to the practical and daily consequences of decentralization; difficulty of centrally based technicians to prepare simple instruments for the regional technicians who today are performing the work they used to do before; and ignorance of DRI's staff of the cross-cutting components destined to strengthen the management capacity of the RDIP. * Regional Workshops (Beneficiaries, Mayors, Staff of UDECOs etc.) The principal accomplishments and/or strong points (certainly quite numerous) identified by the participants in the 8 regional workshops dealt with the following subjects: 20 years experience and continuity of the DRI's involvement in integrated rural development; credibility in the eyes of the rural community and the local governments; management capability; clear and operational instruments including operating rules, mechanisms of control, monitoring and evaluation, component design; leadership inside the NCS and within the Ministry of Agriculture; power to convene and capability to consult with the community and institutions in the sector; and capability to reach the poor and needy. The following are the weaknesses identified in the workshops, prioritized according to the importance each workshop assigned to the different problems: community participation; political interference; insufficient resources; red tape; high cofinancing percentage or unfair cofinancing matrix; community training; lack of monitoring and evaluation; delayed disbursements; lack of coordination between projects and development plans; and little flexibility in the face of regional diversity. L Evaluation of IBRD's performance in the implementation of RDIP 35. In the RDIP's implementation, the IBRD has fulfilled an important role since the program design process; during the missions, there were discussions originating from differing points of view. However, in general the relations between IBRD and Fondo DRI were always cordial and valuable for the RDIP' s execution. 36. From DRI's point of view, IBRD was quite flexible, taking into account that during the time of Program execution, several institutional changes took place that in one way or another affected the initially defined design as well as the implementationi. In addition, through supervision missions, the IBRD contributed important elements in aspects such as: institutional strengthening of DRI; environmental aspects; control and monitoring system; rural woman; NCS; municipal councils for rural development; extension services; and financial matters.  - 33 - 37. There are some significant aspects to be considered as lessons for future programs, such as: (i) the need to design monitoring instruments that are simpler and more effective in their application, and to assign a prudent amount of time to consolidate any strategy in program implementation, before making changes; (ii) the Bank's support to the extension services component could have been more successful if specialists familiar with the situation in other countries had shared their experiences, thus deriving from international experience the lessons which could be applied in Colombia's programs; (iii) the possibility of using the "cashier account" and the "special account" for the withdrawal of funds from the Loan account, expedited the disbursement of resources from the Loan, since if all disbursements had to be made through the special account, they would have been slower due to the strict rules for handling SA funds; and (iv) the combination of the IBRD and the National Government rules and procedures affect the agility in executing the Program, therefore ways should be found to reconcile them. J. Evaluation of implementing agencies and contractors 38. Cofinancing Agreements. The cofinancing contracts financed by the RDIP were executed by five types of implementing agencies as follows: 37.7% by municipalities; 9. 1% by public enterprises; 23.6% by private enterprises; 23.4% by communities; 6% by others. The performance of the five types of agencies was evaluated taking into account the following aspects: projects whose execution was affected due to problems of technical incompetence of the executor; unfinished projects due to poor execution; service projects with delays at the beginning of the execution due to the executor; and infrastructure projects with delays at the beginning of the operation, also due to the executor. Table 4 (see Annex) presents a summary of the situation for each type of problem. It shows that despite the fact that they are in charge of the bigger volume of projects, the municipalities present a smaller percentage of those whose execution was affected due to their technical incompetence. The communities were responsible for the execution of an important amount of projects: 23.4% of the total, 19.6% of the services and 28.1% of infrastructure; communities presented a very low percentage of unfinished projects attributed to them, and low percentages of projects with delays at the beginning of their execution in the case of services and at the beginning of their operation in the case of infrastructure projects. Public enterprises were not the best executors of infrastructure projects, while on the other hand they were the best in service projects. Private enterprises are very good executors in infrastructure projects. 39. Consultants Contracts. The following conclusions were drawn concerning consultants: (i) in the case of preinvestment studies, this evaluation shows a good performance of consultants (individuals and companies) in all components; (ii) in the case of consultancies on specific subjects and evaluation of the RDIP, the evaluation concludes that consultants in most cases had adequate credentials and did a good job (but in some cases more could have been gained); (iii) with respect to supplemental personnel (the regular staff was insufficient to implement the RDIP both before and after the creation of the UDECOs), they worked directly in the units needing support and their contribution was fundamental for RDIP's development; and (iv) in relation to the Technical  - 34 - Cooperation Agreements with the IICA, the evaluation concluded that they expedited the execution of the institutional strengthening component, but the percentage of resources directed to administrative expenses must be curtailed. K. Principal lessons learned about the RDIIP's strategy: impact and sustainability 40. Relevancy. The RDIP's strategy was relevant: it led and accompanied the decentralization process and, despite the give and take, it has been one of the govemment programs with the greatest impact on strengthening the management capability of the municipalities. Likewise, the RDIP's instruments, components and approach were relevant. It helped establish the "project culture" at municipal and regional levels. The components were also appropriate; with the exception of credit and agrarian reform, the most relevant aspects for rural development were included. Other instruments, such as programming of resources by demand and planning "from the bottom up" with the decisive participation of the community did not fully achieve the expected objectives. 41. Efficiency. The strategy was very efficient. In fact, despite the delays and the constant complaints, in six years (only one more than expected) 2.5 more resources were transferred than originally planned, and the program succeeded in mobilizing greater resources and at the same time tripled the number of projects financed. 42. Effectiveness. Although it is too early yet to evaluate the strategy's effectiveness, which must be measured more appropriately during 1998 (evaluation of the RDIP's impact), according to indirect data derived from this progress evaluation, the strategy was also effective, according to the degree of beneficiary satisfaction with the projects, as well as the results obtained with respect to developing greater management capability of municipalities and departments. At the same time, there are clear indications that the accomplishments were not well oriented to the paradigms or model for rural development. In fact, there is still much to do in order that the projects financed effectively contribute to improving competitiveness, sustainability and equity (both between men and women, as well as youth and adults, campesinos, blacks and indigenous people). 43. Thanks to the combination of credibility and technical capability generated by the DRI programs between 1976 and 1990, and the efforts in institutional development, Fondo DRI succeeded in implementing under the RDIP far more resources than expected, as well as maintaining its leadership in the agricultural sector and in the NC S. This was achieved despite the difficulties, both intemal and extemal, which were the product of Decree 2132/92 and the incorporation of the Presidential Programs and the projects financed with earmarked funds. 44. Sustainability of the strategy: need to incorporate the political variable. The strategy by itself is not sustainable. On the contrary, it was designed to disappear insofar as the municipalities fully and responsibly assumed their functions regarding rural development. Therefore, the support strategy of the National Government, although it must disappear gradually, still has an important task to fulfill: imprinting a greater quality  - 35 - to the process. To this effect, it is essential to incorporate the political variable, in the best sense of the expression. 45. In order that the municipalities really take ownership of the models of rural development policy (competitiveness, sustainability, equity), it is necessary that the new support strategy be structured around the rural communities political participation in municipal life. And, with such a base, to coordinate the municipal plans for rural development with their medium and long term strategic vision, and their short term (three years) investment plan. In other words, the central level must modify the strategy, forget about added value with respect to quality of the projects (justification of Phases I and II and a very important element of the RDIP) and design a new strategy with emphasis on the quality of the rural development processes, honoring and applying the aforementioned models with regard to municipal plans for rural development. Annex, Table I PDIC: Compromisos adquiridos y ejecutados 1991 - 1996 (en US$ millones, 1996) (1 Afho Fondo DRI Municipios y Valor total Liquidado o Nuitmero de (2) otros compromisos girado (3) Convenios 1991 40,0 29,1 69,1 55,9 882 1992 40,5 31,3 71,8 66,5 1.382 1993 58,7 44,0 102,7 85,8 1.899 1994 66,1 64,0 130,1 76,7 2.490 1995 66,2 66,5 132,7 51,7 2.734 1996 72,4 62,5 134,9 19,5 2.651 Totales 343,8 297,6 641,4 356,0 12.038 (%) (53,6%) (46,4%) (100,0%) (55,5%) Fuente: Fondo DRI, Subdirecci6n de Planeaci6n, 1997. Notas: (1) Los pesos de 1996 se convirtieron a US$ d6lares a una tasa de COL$ 1.000 por US$ 1. (2) Incluye aporte del Gobierno y prestamos del BID/BIRF (US$150 millones, entre ambos) (3) Para el periodo 1991-94 se refiere al valor de los convenios liquidados hasta dic/96 (cerca del 82% del total de convenios del periodo); para 1995-96 se contabilizan unicamente los pagos (giros) realizados por el Fondo DRI hasta dic/96 con cargo a los convenios realizados.  - 36 - Annex, Table 2 PDIC: Compromisos adquiridos por Componentes durante el periodo 1991 - 1996 (en US$ millones, 1996) Valor (%) Numero de Valor por COMPONENTES compromisos Convenios Convenio (U ) Organizaci6n y Capacitaci6n 11,5 1,8% 395 29.122 Fortalecimiento Institucional 43,6 6,8% 1.018 42.844 Mujer Rural 8,0 1,2% 356 22.437 Desarrollo Tecnol6gico 176,3 27,5% 4.527 38.948 Comercializaci6n 39,4 6,1% 775 50.867 Centros para la Pesca Artesanal 7,7 1,2% 65 118.256 Acuicultura 14,7 2,3% 327 45.029 Microcuencas 21,5 3,4% 572 37.670 Vias 130,6 20,4% 1.454 89.802 Acueductos y Saneamiento Basico 134,4 21,0% 1.917 70.121 Electrificacion Rural 53,6 8,4% 632 84.789 TOTALES 641,4 100,0% 12.038 53.280 Fuente: Fondo DRI, Subdirecci_ n de Planeaci6n, 1997. Annex, Table 3 PDIC: Compromisos adquiridos por Categorias Municipales del SNC durante el periodo 1991 - 1996 (en US$ millones, 1996) Valor Valor por Numero Categorias municipales compromis (%) Municipio de Os (USS miles) Convenios 1: 1, 2 y 3 (549 mpios.) 411,8 64,2% 750,2 7.547 11: 4 y 5 (332 mpios.) 163,7 25,5% 493,1 3.153 III: 6, 7 y 8 (187 mpios.) 65,9 10,3% 352,1 1.338 TOTALES (1.071 mpios.) 641,4 100,0% 598,9 12.038 Fuente: Fondo DRI, Subdireccion de Planeacion, 1997.  - 37 - Annex, Table 4 PORCENTAJE DE PROYECTOS DEL TOTAL DE LOS REAUZADOS POR CADA TIPO DE EJECUTOR QUE PRESENTARON LOS DIFERENTES PROBLEMAS CAUSADOS POR EL EJECUTOR (PORCENTAJES) TIPO DE PORCENTAJE PORCENTAJE DE PORCENTAJE DE PORCiNTAJE DS PORCENTAJE DE PORCENTAJE DE PORCENTAJE DE LJECtJTOR TOTAL DE PROYECTOS POR PROYECTOS POR PROYECTOS DE PROYEICTOS DE PROYECTOS DZ PROYECTOS DE PROYECTOS TIPO DE TWO DE tzCUIOr SERVICIOS POR TPO SERV'CiOS POR TWO IEFRAESTRUCTURA i4FRARUCTUILA POR POR TIPO DE EJECUTOR INCONCLUSOS Pr DE WEcUrOR DE IJECUTOR CON POR TIPO DE TIPO DE EZJCUTOR CON EJECUTOR AEECTADOS POR MAEA ECUCION DEMORAS EN EL U JECUTOR DIMORAS EN RL WNCIO INCAPACIDAD ?4CMO DR LA Di EA OPIRACI6N TECNICA DEL EJECUION DEBmAS DEDWAS AL EJECUTOR EJECUOR AL LECUrOR MUNICp=IO 3770 5.6S 0.4 5.00 7.95 13.20 0.7 iMPRESA 9.10 57 0. 4.70 0.0 14.40 IS7 PUBUCA PARTICULAR 23.60 11.2 21 *390 6.9 41.40 0. COMUNIDAD 23.40 9.23 0.0 19.60 0.2 2.10 0.7 IOTROS 6.00 3.23 3.81 2.0 15. 2S0 0.0 FUENTE ENCUESTA BENEFICiAOS EVALWACION PDIC.  - 38 - Appendix B Translated from Spanish BORROWER'S COMMENTS ON THE DRAFT ICR A. Comments of the National Planning Department The report presents a general appreciation of Phase III of Fondo DRI (RDIP) with an objective evaluation by the Bank of the quantitative and qualitative aspects of the program's planning and implementation. Among various aspects, the Bank's report emphasizes the negative impact of factors such as Decree 2132 and introduction of the National Cofinancing System (NCS), the lack of an adequate system of monitoring and evaluation, partial achievements in promoting community participation, and uncertainty about the sustainability of Fondo DRI as an institution. In relation to these specific aspects of the report, the Unit of Agricultural Development of the National Planning Department would like to make comments (see below) in order to complement the conclusions presented in the Bank's report. Introduction of the NCS and Decree 2132. To further expand on several observations made in the report about the decentralization process and two of its most important instruments--Decree 2132 and the NCS--it is necessary to remember that the design and implementation of the RDIP was done in the context of a profound change in the political structure of the Colombian State. The development of the new Constitution generated drastic changes in the institutional structure of the State, and in relationships and attributions of the State and Civil Society and between different levels of government. Consequently, this was a period of rapid evolution in the structure of the State which contributed to the deepening of democracy. This was obviously disruptive not only for the RDIP, but for all programs designed to support rural communities, and, more generally disruptive for the central and regional government institutions. As part of this process of change, the responsibility for rural development was for the most part transferred to municipalities on the basis of laws such as Law 101 on agriculture-livestock development, Law 100 which created the Social Security System, and Law 3 which created the Sistema Nacional de Vivienda de Interes Social (National System of Social Interest Housing). As part of the change, the mechanism for access to public resources was based on demand and the free competition of public institutions, NGOs, peasant (campesino) organizations and others for implementation; the use of these funds would be under the supervision of citizens (e.g., las veedurias ciudadanas). It should be emphasized that the role of the central government evolved into one of accompanying and assisting the regional, provincial and local levels of government. Decree 2132 and the NCS enabled a transformation in the traditional dynamic of allocating public resources, by making the decentralization process more agile permitting the appearance of new actors through mechanisms like citizen participation (involvement in identification and implementation of projects, and supervision of the utilization of public  - 39 - resources). Likewise, Decree 2132 and the NCS promoted the strengthening of the institutional structure in the regions by giving more decision-making power to municipalities, departments and by reinstating and strengthening the role of the province and region. These changes supported among other aspects an increase in the effectiveness in the use of the State's resources and an increase in the capture and contribution of these resources by municipalities. This is reflected in the US$ 297.6 million contribution to the RDIP made by municipal governments, communities and other actors for the cofinancing of rural development projects. Sustainability of the RDIP and Fondo DRI. We share the report's view concerning the low sustainability of the RDIP as a project, given that the original strategy and design required a highly flexible program which responded to the different and thus highly variable socio-economic conditions for rural development in Colombia (one factor being the issue of competitiveness). On the other hand, the sustainability of Fondo DRI as an institution should be analyzed in light of the institutional changes which the Colombian State has gone through in the 1990s and the fact that currently no other major institutional changes are foreseen. However, if such changes do arise, the focus of attention on rural development is a State concern which is reflected in the Development Plan and, as such, continuity and strengthening of concern for rural development should be a priority. Community Participation and Institutional Strengthening. Concerning these aspects, other elements not considered in the report which represent important contributions of the RDIP should be emphasized, such as: - successful experiences in participatory planning and democratic management of rural development in municipalities and distinct regions of the country (finding based on the evaluation of Phase III by DRI); * improved capacity in most municipalities to formulate and present (in a timely fashion) projects to the NCS, in this way consolidating the project "culture" and cycle; * preparation of new components as a result of a concept of rural development which takes into account the gender and age perspective, ethnic minorities, and environmental issues; * increase in participation of rural and urban actors in the management of public resources which may take on a more complex form as is the case in situations where the Municipal Councils for Rural Development (CMDR) are involved; * increase in the communities' knowledge of the budgeting system, the procedures for getting access to funds, the instruments for regional and local planning, and the sector policies; * existence of 700 CMDRs and work underway to support their consolidation and strengthening as a central element of building a new institutional framework more in accord with the characteristics of rural development in Colombia;  - 40 - * the CMDRs are contributing to improved inter-institutional coordination through the allocation of public resources (coming from different sources and programs) and the articulation of urban-rural planning; For these reasons, the RDIP has been the leader in the NCS. Through this leadership, it has generated a high impact on the strengthening of municipal capacity and, in general, on rural development in the municipalities under its jurisdiction. Earmarked Budget Resources. The problem of earmarked budget resources undoubtedly affected the NCS. However, it should be noted that this mechanism allows at least the financing of viable, duly prepared projects. Finally, it should be noted that the evaluation of Phase III of DRI was carried out while the RDIP was underway. For this reason, the results should be considered preliminary until the full impact evaluation of the RDIP is carried out either during or after 1998. Likewise, the design of a new phase of DRI should incorporate the results of the completed evaluation and include measures to improve on the problem areas, particularly those which will allow a better response to the needs of urban-rural communities in accord with the principles of sustainability, equity and community participation. B. Comments of Fondo DRI In accord with your request, we are pleased to submit the following comments on the ICR. Ref. paras. 10 and 12 of the summary and paras. 23 and 24 of the report. We believe there is a contradiction between saying that the results of the RDIP are judged satisfactory because it reached more beneficiaries than expected, had a higher level of investment than expected, and achieved most of its objectives, yet rating "the performance of Fondo DRI as deficient, though admirable under the circumstances." From our point of view, while recognizing the problems of institutional instability during the program, the results themselves demonstrate the level of efficacy and efficiency of Fondo DRI in program implementation. For this reason, we consider it too harsh to state that Fondo DRI's performance was deficient. Ref. para. 10 of report. The report states that: "The expansion of the program to the entire country swamped DRI's capacity to attend to this." We do not know what substantiates this statement since the evaluation we undertook shows that Fondo DRI implemented the program in a largely satisfactory manner. The recognition of this by the Government and IDB resulted in the assignment to DRI of other programs such as the Plan Pacifico, Plante and the Program for Generating Employment.  - 41 - Appendix C COFINANCIER'S COMMENTS ON THE DRAFT ICR The Inter-American Development Bank (IDB) finds that the Implementation Completion Report for the Rural Development Investment Program clearly and accurately reflects the challenges and successes of this innovative project. The collaboration of the IDB and the World Bank in this operation was noteworthy and stands as an example of how the two institutions can work together effectively. In examining the performance of this project, we observe several outstanding issues to be addressed during the preparation of the next phase by the IDB. The report correctly points out the significance of this project, both in terms of its approach and in total amounts lent. However, while we do have information regarding the number of municipalities financed under the project, we do not have a clear idea of the socio- economic impact of the program. An effort will have to be made to prepare a more rigorous analysis for evaluating the results of this important project. The report recognizes the problems facing the Fondo DRI in the midst of the changing structure of the National Cofinancing System. However, we would expect that some of the current instability would be diminished in the next year, and subproject approval mechanisms be streamlined so as to more effectively address needs in low- income rural areas in a more timely manner. We concur that ex-post, rather than ex-ante project analysis should be considered toward this end. Finally, decentralized recurrent resources for investment maintenance continue to be an issue in insuring the sustainability of DRI investment. The next phase of the DRI must address weaknesses in rural municipal financial management and project execution.  I  MAP SECTION   7a' 76 74 72 70 CUBA A,,QCACAI C 0 l O M B I A A T-.fBb fANJ J5A ~~fIs* (AR IXUc S DE EAAI RURAL DEVELOPMENT INVESTMENT PROGRAM (RDIP) / J (CEA Dri Municipalities 1 990-94 Incorporation R,ohocho > / 1