Colombia - Second Irrigation Rehabilitation Project

Document of The World Bank FOR OFFICLAL USE ONLY Report No. t6584 IMPLEMENTATION COMPLETION REPORT COLOMBIA IRRIGATION REHABILITATION II PROJECT (LOAN 2667-CO) May 16, 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$1.00 = Col$ 1037 WEIGHTS AND MEASURES Metric System FISCAL YEAR OF THE BORROWER January I - December 31 ABBREVIATIONS AND ACRONYMS CONPES - Consejo Nacional de Politicas Econ6micas y Sociales (National Council for Economic and Social Policy) ERR - Economic Rate of Return FEDERRIEGO - Federacion Colombiana de Distritos de Riego (Federation of Irrigation Districts) FY - Fiscal Year HIMAT - Instituto Colombiano de Hidrologia, Meteorologia y Adecuaci6n de Tierras (Colombian Institute for Hydrology, Meteorology and Land Improvement) INAT - Instituto Nacional de Adecuaci6n de Tierras 1/ (National Institute for Irrigation and Drainage) IRR - Internal Rate of Return ha - hectares O&M - Operation and Maintenance PSR - Public Sector Reform Loan SAR - Staff Appraisal Report SSI - Small Scale Irrigation UNDP - United Nations Development Program WUA(s) - Water User Association(s) Vice President: Shahid Javed Burki Director: Paul Isenman Acting Division Chief: Jonathan Parker Staff: Elsie Garfield 1/ The Institute changed its name and functions from HIMAT (Instituto Colombiano de Hidrologia, Meteorologia y Adecuaci6n de Tierras) to INAT in 1994. Throughout this report, INAT will be used.  FOR OFFICIAL USE ONLY IMPLEMENTATION COMPLETION REPORT COLOMBIA IRRIGATION REHABILITATION 1I PROJECT (Loan 2667-CO) Table of Contents Page No. PREFACE ............................................................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 ..................................................5 E. Bank Performance ..................................................6 F. Borrower Performance ..................................................7 G. Assessment of Outcome ............ .....................................8 H. Future Operation ..................................................8 I. Key Lessons Learned ..................................................9 PART II: STATISTICAL TABLES I . Summary of Assessments .10 2. Related Bank Loans/Credits .1 3. Project Timetable .1 4. Loan/Credit Disbursements: Cumulative Estimated and Actual 12 5. Key Indicators for Project Implementation and Operation .13 6. Key Modified Indicators for Project Implementation and Operation . 15 7. Studies Included in Project .16 8A. Project Costs ....................................... 17 8B. Project Financing ....................................... 18 9A. Investment Costs ....................................... I 9 9B. Irrigated Areas ....................................... 19 9C. Production Impact ....................................... 19 9D. Economic Budget ....................................... 20 10. Status of Legal Covenants ....................................... 21 11. Compliance with Operational Manual Statements ................................. 23 This document hag 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.  Page No. 12. Bank Resources: Staff Inputs ........................... 24 13. Bank Resources: Missions ........................... 25 APPENDIXES: A. Mission's Aide Memoire ..................................... 26 B. Borrower's Contribution to the ICR ..................................... 35 C. Map IBRD 19223 - Project Location  IMPLEMENTATION COMPLETION REPORT COLOMBIA IRRIGATION REHABILITATION II PROJECT (Loan 2667-CO) PREFACE This is the Implementation Completion Report (ICR) for the Irrigation Rehabilitation II Project in Colombia, for which Loan 2667-CO in the amount of US$114.0 million was approved on March 25, 1986 and made effective on February 24, 1987. The loan was closed on June 30, 1996, two years later than the original closing date. Final disbursement took place on November 14, 1996, at which time a balance of US$ 16.9 million was canceled. 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 Juan Morelli, FAO/World Bank Cooperative Program, 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. Preparation of the ICR began during the Bank's completion mission of August- September 1996. The Borrower has reviewed the report and does not wish to make any comments; their contribution to the ICR is contained in Appendix B.  IMPLEMENTATION COMPLETION REPORT COLOMBIA IRRIGATION REHABILITATION H PROJECT (LOAN 2667-CO) EVALUATION SUMMARY Introduction 1. In the early 1980s, the Colombian Government requested Bank assistance to rehabilitate and complete 15 of the 22 publicly owned irrigation and drainage districts, most of which had been constructed by the land reform agency in the 1960s. The Bank agreed to a phased approach: Rehabilitation I (Ln. 1996-CO) covered 8 districts, and Rehabilitation II covered 6 districts. The two operations overlapped from 1986-88. Both projects aimed at expanding agricultural production by improving the contribution of existing irrigated areas which was well below potential. Project Objectives and Description 2. Rehabilitation II's objectives were to: (a) increase crop production in six districts through rehabilitation and improved technical and extension services leading to more intensive land use; (b) improve the incomes of 3,900 poor farmers through small-scale irrigation development; (c) partially restore agricultural production in the volcano- damaged Armero area through irrigation investments; and (d) improve the sustainability of public irrigation investments through full recovery of operation and maintenance (O&M) costs, and the building-up of water user associations (WlUAs) in the districts covered by the project. The project components included: civil works and replacement of maintenance equipment in the 6 districts; design and construction of small-scale irrigation schemes (SSI); civil works and design studies for the Armero region; erosion control; credit for farmers; and technical assistance and training for farmers and WUAs. 3. The project's objectives were fairly clear, but unrealistic in terms of the implementing agency's capacity and country experience. The rehabilitation component was more than triple the size of the first project and new components were added at the last minute in areas where the irrigation agency had little experience. The key project risk, slow intensification of land use, was underestimated and the measures for minimizing it were inadequate. The realism of the objectives was also questionable, because the Government did not fully endorse the strategy giving priority to rehabilitation over new investment and did not recognize the need for supportive sector policies. 4. The project's objectives should be looked at in historical context. Prior to policy discussions under the Bank's Public Sector Reform loan (PSR, Ln. 3278 approved in 1990), focus was on the production objectives for the 6 districts and SSI. Subsequently, both the Government and Bank gave increased emphasis to achieving the sector policy objectives, in particular the transfer of district management to WUAs which was not  iii treated as an explicit project objective at appraisal. The new irrigation law adopted in late 1993 reinforced this. Implementation Experience and Results 5. Project implementation was initially slow and never improved sufficiently to complete the project. Delays in achieving loan effectiveness and completing design studies were compounded by periodic counterpart funding shortages and inadequate sector policies. INAT's ineffectual management caused serious problems throughout, particularly in procurement. The small scale irrigation component suffered from poor quality works and inadequate support to farmers. At the time of the mid-term review in 1991, it was estimated that without a major improvement at least five more years would be required to complete the project. There were improvements, though, when the enabling environment drastically changed. In 1994, new INAT management began to strengthen O&M recovery and speed-up the transfer to WUAs. These events figured importantly in the two extensions of the loan's closing date. Following the second extension, project activity and disbursements accelerated. However, it was too late to make up for all the lost time and the loan closed in mid-1996 after eleven years' disbursements, with 15% of the loan amount canceled. 6. The project partially achieved its objectives. In regard to the first objective, INAT's failure to complete the civil works in four of the six districts combined with farmers' failure to intensify land use in all but one district resulted i 'ittle impact on production. Consequently, the updated economic analysis shows an IRR of 7.8% for this component, far below the anticipated 28%. For the second objective, while the achievements for the area and number of farmers were nearly double the targets, in 1996 a majority of the projects were not working adequately. Therefore, it is unlikely that the ERR of 20% for this component was achieved. In regard to the third objective, there is no estimate of impact; the target for rehabilitation in Armero was met, but the new project for the area was dropped. 7. Finally, relative to the fourth objective, achievements for sector policy are satisfactory. At appraisal, targets were set for O&M recovery based on expectations concerning improvements in irrigation service; transfer of management to WUAs was to take place once full cost recovery was achieved, but no timetable was set until 1990 as part of the PSR. While O&M recovery is 100% in only 2 of the 6 districts, it has substantially improved in all districts which is consistent with the project's original goal that full recovery would be achieved when the civil works were completed. All six districts financed by the loan are now being managed by the WUAs, though some still need substantial assistance to improve capacity. 8. Project sustainability is uncertain. It remains to be seen whether the Government and INAT will respect their commitment to complete the project within a reasonable timeframe so that the benefits of the investments made up to now can be more fully realized. However, even if the works are completed, it is still uncertain whether farmers will make on-farm improvements to intensify land use, particularly in the face of credit problems and violence in rural areas. Yet a positive factor that strengthens the prospects  iv for intensified land use is the changed incentives framework which has put pressure on farmers to improve productivity, reduce costs and increase the efficiency of water use. The sustainability of gains in cost recovery and WtUA management is promising but uncertain. A positive sign is the more active role some WUAs are taking to seek solutions to their problems which could contribute to the long-term sustainability of the investments. 9. The Bank's performance was deficient particularly during project design. It would appear that pressures to lend resulted in the Bank making an overly optimistic assessment of the on-going operation, rushing to appraisal before the civil works had been designed, glossing over some fundamental differences with the Government concerning strategy and policy, and inadequately evaluating and addressing the project's risks. Supervision was inadequate until the later years when it was too late to make up for lost time. The Bank did not monitor closely the key project risk and when the mid-term review recognized that it had materialized, no effective actions were taken (see para. 3). On the other hand, the Bank made repeated efforts to right the project's unsteady course and contributed in its policy dialogue to the Government's thinking which resulted in the pioneering irrigation law. In 1994, the Bank responded energetically to new opportunities offered by the law, new leadership in INAT, and initiatives of farmers in the Maria la Baja district. Intensive supervision accelerated project activities with 22% of total disbursements taking place during the 16 month period following the second extension of the closing date. 10. Both the Government (Guarantor) and INAT (Borrower) performed deficiently, though in the later part of the project the Government adopted policy reforms and the new law which may contribute to a favorable outcome. The Government failed to show a strong commitment to the project, and oversight of INAT was weak and sporadic. The project was badly weakened too by the Government's failure to keep in place a sufficiently competent technical and managerial team in INAT, along with its frequent restructuring of the agency from 1990-94. At critical moments, the Government and INAT have given higher priority to financing new projects, which apparently have higher political returns, than to completing and improving the performance of the existing districts. In this regard, firm commitments to complete the project made to the Bank at the time of extending the loan closing date are now in question. 11. INAT's performance was deficient which reflected in part broader management problems within the institution aggravated by excessive turnover. It failed to maintain a separate project unit with management authority, a key factor in the satisfactory management of the first project. Procurement and contracting were particularly deficient. INAT failed to monitor and improve the availability of critical ingredients for project success: credit, good extension services and appropriate training for WUAs. However, after the second loan extension, INAT's performance did improve substantially and it took a number of positive actions. 12. The project attained several useful ends, but its outcome is rated as deficient. However, if the Government completes the project as promised, an acceptable rate of return of as much as 18.8% could be achieved (para. 14). Furthermore, the changed  v sector and economic policies improve the prospects for the long-term development of the sector. Findings and Key Lessons 13. Some key lessons of the project related to its "quality at entry" are as follows: (a) it is better not to proceed with a lending operation until fundamental agreement on strategy, objectives and key sector policies has been reached; (b) the scope and objectives of a follow-on project should be carefully calibrated to the capacity of the implementing agency as demonstrated in the project under implementation; (c) project risks identified at appraisal should be closely monitored and credible solutions found if the risk does materialize; (d) to be successful, irrigation and drainage projects must pay attention to agronomic, economic, institutional and social issues in addition to engineering; (e) an engineering-minded irrigation agency with a narrow mandate and limited experience in agricultural matters is unlikely to have sufficient interest and capacity to ensure that on-farm development activities are adequately carried-out; (f) the experience in Colombia has shown that WUA management can be an important contributor to improved cost recovery and maintenance; the logic that rehabilitation of irrigation works is a prerequisite to full cost recovery and transfer of management to WUAs was flawed; (g) greater involvement of WUAs and the irrigation agency's regional offices in the design and implementation of the project would have improved implementation; and (h) low recovery of O&M costs is not necessarily a result of farmers unwillingness to pay or the irrigation agency's poor capacity as a bill-collector, but may stem from long-standing socio-economic problems which a good social assessment during preparation would have uncovered. Future Operations 14. The second extension of the closing date was based on Government's commitment to complete the project using its own funds (estimated US$34 million required) by end of 1997; an action plan and set of performance indicators were agreed. However, in response to competing investment priorities, INAT has now programmed completion of the project for the year 2000. The Bank has sought clarification from the Government about when it plans to complete the project. Currently, there are no plans for a new lending operation, but support for strengthening the WUAs through an Institutional Development Grant and an EDI seminar is planned.   IMPLEMENTATION COMPLETION REPORT COLOMBIA IRRIGATION REHABILITATION II PROJECT (LOAN 2667-CO) PART I: PROJECT IMPLEMENTATION ASSESSMENT Introduction I. In the early 1 980s, the Colombian Government requested Bank assistance to rehabilitate and complete 15 of the 22 publicly owned irrigation and drainage districts, most of which had been constructed by the land reform agency in the 1960s. The Bank agreed to a phased approach: Rehabilitation I (Ln. 1996 for US$37 million) covered 8 districts, and Rehabilitation II (US$114 million) covered 6 districts.' Both projects emanated from the goal of expanding agricultural production in the face of a sharp decline in the growth of agricultural output and rising demand due to rapid urbanization; the growth potential of rainfed agriculture was not expected to be sufficient to meet this demand. Since the contribution of existing irrigated areas (accounting for 2% of land used for agriculture) was well below potential, the Government and Bank agreed that investment to improve these returns was a priority. A. Project Objectives 2. Rehabilitation II's objectives, as stated in the Staff Appraisal Report (SAR), were to: (a) increase crop production on 108,000 ha of six districts through rehabilitation and improved technical and extension services leading to more intensive land use; (b) improve the incomes of some 3,900 poor farmers located on about 7,800 ha. through small-scale irrigation development; (c) partially restore agricultural production in the volcano- damaged Armero area through reconstruction and addition of irrigation facilities; and, (d) improve the sustainability of public irrigation investments through full recovery of operation and maintenance (O&M) costs and the building-up of water user associations (WUAs) in the districts covered by the project. The project components included: civil works and replacement of maintenance equipment in the 6 districts; construction of small- scale irrigation schemes (SSI); civil works and design studies for the Armero region; erosion control; credit for farmers; and, technical assistance and training for farmers and WUAs. 3. The project objectives were fairly clear, but unrealistic in terms of the implementing agency's capacity and country experience. The project was presented as a follow-up to Rehabilitation I, which was perceived at the time as being successfully implemented: civil works in over half of the districts were then well advanced and crop output increases were considered promising. However, the rehabilitation component of One district was dropped from Rehabilitation II during appraisal, bringing the total number of districts covered by the two loans to 14.  2 the second project was significantly more demanding: the area to be rehabilitated was 3.5 times and the cost 3.2 times that of the first project. This stemred from a compromise whereby the Bank agreed to finance some large civil works in order to increase the irrigated area instead of focusing solely on rehabilitation. The project scope kept expanding up to the time of Board approval to include new components in areas where INAT had little experience: (i) small scale irrigation; (ii) an erosion control component added after appraisal; and (iii) a new 8,000 ha. irrigation scheme for Armero which was added at the Government's request during negotiations despite the lack of feasibility studies and the fact that INAT had only built one large project in its decade of existence. Thus, during the design of the project, its objectives became increasingly unrealistic, and its risks multiplied. 4. Another critical issue relating to the realism of the objectives was the Government's lack of full agreement, despite the impression given by the SAR and mid- term review, with the strategy giving priority to rehabilitation and intensified use of existing infrastructure over investment in new works.2 Nor did it share the Bank's concern that government policy against financing on-farm development posed serious problems, as a high proportion of land in the districts was under low value crops. Both sides made compromises in order to get the loan approved, but over the course of the project these differences re-emerged and may account for Government's lack of drive to complete the project in the face of competing priorities for use of investment funds. 5. The key project risk, slow intensification of land use, was underestimated. The measures for minimizing it were inadequate: technical assistance, training, access to credit for farmers, and a clause in the loan agreement that committed the Government to take corrective action, if this problem arose. Lack of supportive sector policies was not cited as a risk. Yet, it was not until the Government approved the first framework for irrigation development in mid-1991 (CONPES, Irrigation Development 1990-2000) and then translated the policies into law in late 1993 that the sector policy framework fully supported the project's objectives. 6. Finally, these objectives should be looked at in historical context. Prior to policy discussions under the Bank's Public Sector Reform Loan (PSR, Ln. 3278 approved in 1990), focus was on the production objectives for the 6 districts and SSI. Subsequently, both the Govemment and Bank gave increased emphasis to achieving the sector policy objectives, in particular the transfer of district management to WUAs which was not treated as an explicit project objective at appraisal. The new law which gave farmers a central role in all aspects of irrigation reinforced this. 2 Project preparation nearly broke off when the Government became angry at the Bank's unwillingness to include some new, large scale projects, because of their poor economic justification and the underutilization of existing irrigated areas. While the Government eventually accepted this for purposes of obtaining the loan, evidently it did not accept the strategy.  3 B. Achievement of Project Objectives 7. The project partially achieved its objectives. In regard to the first objective, INAT's failure to complete the civil works in four of the six districts combined with farmers' failure to intensify land use resulted in little impact on production (see Table 5). Consequently, the updated economic analysis shows an IRR of 7.8% for this component, far below the anticipated 28%. Yet, the Prado Sevilla district did achieve satisfactory economic returns (an IRR of 23 percent) and average farm incomes grew 30 percent. Two other districts have positive IRRs (3% for Saldana and 7% for Maria la Baja, assuming that the works contracts in progress at the closing date are completed) and farmer incomes also improved in the La Doctrina district. 8. In regard to the second objective, while the achievements for the area and number of farmers were nearly double the targets (16,500 ha. and 8,000 families) for SSI, in 1996 a majority of the projects were not working adequately. Data on the economic returns are scarce, but it appears that the working schemes did have returns consistent with appraisal estimates, though it is unlikely that the global ERR of 20% was achieved. In regard to the third objective, the target for rehabilitation in Armero was met, but there is no estimate of the impact (the SAR contains no evaluation either); the new project for the area was dropped in part because of concerns about the economic retums. 9. In regard to the fourth objective related to sector policy, achievements have been satisfactory compared to appraisal expectations. The project's underlying logic was that once the infrastructure in the districts was improved, irrigation service would improve, which would lead within 2 years to full recovery of O&M costs, at which time the irrigation agency would transfer district management to the WUAs. Thus, targets for recovery of O&M costs in the SAR were set in function of expectations concerning improvements in irrigation service. No timetable was set for transfer of management to WUAs. Due to substantial delays in the civil works, the targets for O&M recovery were formally revised downward in 1991 and 1995. 10. Since 1993 when the new law was passed, cost recovery substantially improved and the transfer of district management to WUAs accelerated. While O&M recovery is 100% in only two of the six districts supported by the project, it has substantially improved in all reaching 60% for the four WUAs in Prado Sevilla and ranging between 17-50% in the other three districts. INAT needs to phase out the subsidy quickly for the WUAs in Prado Sevilla, whereas the other three districts merit more careful consideration in view of difficult underlying problems such as poor profitability and high debt of farmers. Transfer of management became an explicit objective of the project in 1990 and all six districts under the project are now managed by the WUAs. From 1989-1995, INAT transferred 13 districts to WUAs including the four project districts not yet  4 managed by WUAs and most of the districts financed by the first project.3 While much remains to be done to strengthen the capacity and management of the WUAs, a major break with the past, paternalistic role of the Government in the agricultural sector is underway. C. Implementation Record and Major Factors Affecting the Project 11. Project implementation was initially slow and never improved sufficiently to complete the project. Delays in achieving loan effectiveness (nine months) and completing design studies (two years) were compounded by periodic counterpart funding shortages and, in the early years, inadequate sector policies. Procurement proved to be a continual headache, particularly before a new procurement code came into effect in 1994; delays of 6-12 months between bid opening and contract signature were common. This was aggravated by INAT's insistence on small, inefficient contract lots for civil works, apparently due to political pressures on behalf of local contractors. At the time of the mid-term review in 1991, it was estimated that without a major improvement at least five more years would be required to complete the project. The review also recognized that the agricultural aspects of the project were going poorly: the applied research and extension approach was not having the desired impact, and the project's impact on production and farmers' investment in on-farm development could not be adequately assessed due to unreliable data. 12. The small-scale irrigation component suffered from poor quality works and inadequate support to farmers. INAT failed to provide enough manpower for the SSI program, to coordinate it properly internally and with the communities and other agencies concerned, and to monitor and evaluate its activities. The Bank's 1989 SSI Loan (311 3-CO) financed completion of 30 of the projects and eventually helped address these problems. 13. There were improvements when the enabling environment drastically changed. This began in the early 1990s in the context of the PSR when the Government and Bank agreed to change the underlying logic of the project: WUA management was sufficiently important as a contributor to improved cost recovery that it should take place even before works were completed and full cost recovery achieved (via negotiation of a plan for phasing out the O&M subsidies). While INAT fell short of fully meeting the PSR targets (ICR, Rpt 14882), the amount of attention focused on the sector policy objectives was a major milestone for a project which included them primarily as window-dressing. Meanwhile, other positive changes in the enabling environment were underway: (i) the Government introduced a medium term program of structural reforms; and (ii) the 1991 Constitution supported economic liberalization, popular participation, decentralization, and modernization of the state. Despite this progress, in frustration over INAT' s failure to explain the persistent low levels of O&M recovery, in February 1992 the Bank declared 3 The goal of transferring all the districts (8 are still managed by INAT, accounting for 36% of the total area) by 1997 was adopted as part of the Action Plan for the second extension of the loan closing. However, it is unlikely to be achieved given wavering interest on INAT's part and some difficult social and economic problems which make transfer complex for some of these districts.  5 ineligible for loan financing three of the six districts;4 a subsequent study showed that failure to pay was related to the low profitability of the districts and long-standing problems such as high debt. 14. The improvements in sector policy and replacement of INAT's top management led to a sense of optimism about the project's prospects and the Bank's decisions, in 1993 and 1995, to extend the loan's closing date. Following the second extension, project activity and disbursements accelerated in large part due to improvements in procurement introduced at the Bank's insistence. However, it was too late to make up for all the lost time and the loan closed in mid-1996 after eleven years' disbursements, with 15% of the loan amount canceled. However, based on the Government's plan to finish the project using its own funds, the loan closing did not signal the completion of the project (see para. 27). D. Project Sustainability 15. Project sustainability is uncertain. It remains to be seen whether the Government and INAT will respect their commitment to complete the project within a reasonable timeframe so that the benefits of the investments made up to now can be more filly realized. Another possibility is that the WUAs will organize financing to complete the works, though this seems unlikely for poorer farmers such as those in the Maria la Baja district. However, even if the works are completed, it is still uncertain whether better utilization of the infrastructure and the improved returns will be realized. The cost and availability of credit for on-farm development may continue to be a problem, as well as continuing violence in rural areas which discourages investment. The sustainability of gains in cost recovery and WUA management is promising but uncertain. The Coello and Saldana districts continue to be strongly managed and the prospects are good that the four WUAs in Prado Sevilla will find a way to manage the district properly in view of the high value crops and relatively bigger and wealthier farmers involved. The case is less certain for the other three districts and SSI where the farmers are poorer and in greater need of capacity building and technical support. 16. At the same time, factors which strengthen the prospects that Colombia will make and sustain progress towards the project objectives include: (i) the revised incentives framework which pressures farmers to improve productivity, reduce costs and increase the efficiency of water use; and (ii) the legally recognized importance of farmer participation and WUAs, and signs that they are beginning to more actively seek solutions to their problems. More support for WUAs will probably be provided by the Federation of Irrigation Districts (FEDERRIEGO) which is exploring how it could make itself stronger and more responsive to its members needs. 4 The 3 districts accounted for less than 30% of the area and an even smaller proportion of project funding for rehabilitation. Two districts, La Doctrina and Abrego, were small and rehabilitation works had been completed by the time of ineligibility. In the Maria la Baja district,, rehabilitation was at an early stage and interruption of financing did have a significant impact; the Bank agreed to resume funding the district in June 1995 based on a detailed Action Plan.  6 E. Bank Performance 17. The Bank's overall performance was deficient, although a concerted effort has been made since 1994 to take advantage of the new opportunities. This project is an example of the consequences of "poor quality at entry." While Bank sector work on irrigation addressed a full range of issues, the approach to the project was typical for irrigation projects of this period which focused mainly on the engineering aspects and physical investments, and paid limited attention to the agronomic, economic, institutional, and social issues. Even then, the engineering aspects were not handled well. Contrary to Bank norms, appraisal was done on the basis of recently completed feasibility studies and design studies had not yet been contracted at the time of Board presentation; this resulted in poor specification of the works to be done, cost estimates which turned out to be too low, and a two year delay while the designs were being completed. This situation was exacerbated by lack of technical support by the Bank at critical moments in the project's development. 18. It is striking that nearly all the critical issues and problems which arose during implementation were foreseen by various experts during the project preparation and review process, but credible solutions were not defined. The appraisal inadequately gauged the lessons of the on-going Rehabilitation operation and Colombia's sectoral experience (particularly with WUAs). Insufficient attention was devoted to the efficiency of water use, intensification of production, and the strategy for linking infrastructure rehabilitation with district transfers to the irrigators. There were failures to design remedies for the farmers' credit shortages and other causes of weak incentives for on-farm investment. Questionable judgments were made regarding the adequacy of irrigation technology and the extension methodology. INAT's capacity was not adequately assessed, particularly in procurement and non-engineering aspects of the project. These deficiencies are distressing, given that Rehabilitation II was the Bank's fifth irrigation sector operation in the country. More time should have been spent on additional project preparation and greater effort devoted to shaping measures to cover the numerous risks. 19. One of the weak aspects of project design which was not questioned within the Bank was the fact that the Coello and Saldana districts which had been managed by WUAs for nearly a decade were treated the same as the other districts managed by INAT. If the two WUAs had been involved in project design and implementation, it is highly likely that the project's performance would have been better (more than half of the project funding for civil works and maintenance equipment was for these districts). 20. Supervision was inadequate until the later years when it was too late to make up for lost time. INAT had shown during the first project and preparation of the second that it needed strong technical backstopping, yet only limited technical advice was offered by the engineer supervising the project up to 1991, and no engineer participated in the mid-term review or decision in mid-1993 to extend the closing date. Likewise, despite pervasive problems, there is no record of a senior procurement expert participating in supervision until 1995. The Bank did not monitor closely the key risk and when the mid- term review recognized that the risk had materialized, no effective actions were taken.  7 21. There were positive aspects of the Bank's work however. The diagnosis and advice contained in the Bank's July 1988 irrigation sector report and the work on the PSR project contributed to the Government's thinking about a new sector policy which led to the pioneering irrigation law. Likewise, the PSR created momentum towards fuller cost recovery and transfer of districts. In 1994, the Bank responded energetically to new opportunities offered by the law, new leadership in INAT, and initiatives of farmers in the Maria la Baja district. Intensive supervision with particular attention to procurement matters led to a significant acceleration of project activities. Several innovative activities were also supported, such as a study tour to Chile, development of an integrated Plan of Action for the Maria la Baja district, and presentations on modern design of irrigation infrastructure to INAT and the private sector. F. Borrower Performance 22. Both the Government (Guarantor) and INAT (Borrower) performed deficiently, though in the later part of the project the Govemment adopted policy reforms and the new law which may contribute to a favorable outcome. The Government failed to show a strong commitment to the project, and oversight of INAT was weak and sporadic. The project was badly weakened too by the Government's failure to keep in place a sufficiently competent technical and managerial team in INAT, along with its frequent restructuring of the agency from 1990-94. The project experienced significant delays caused by the complex, inefficient budget process and the cumbersome procurement code. At critical moments, the Government and INAT have given higher priority to financing new projects, which apparently have higher political returns, than to completing and improving the performance of the existing districts. In this regard, firm commitments to complete the project made to the Bank at the time of extending the loan closing date are now in question. In sum, there was a lack of strong political leadership in pursuit of project objectives. 23. INAT's performance was deficient which reflected in part broader management problems within the institution aggravated by excessive turnover. It failed to maintain a separate project unit with management authority, a key factor in the satisfactory management of the first project. When it re-established a unit at the Bank's insistence in 1995, the unit was given a more limited coordinating role. Procurement and contracting were particularly deficient; procurement planning was totally absent and resulted in costly works being inoperable, because a critical link was missing. Audit reports criticized INAT's poor supervision of contracts; problems arose of poor quality construction and inadequate works supervision by contractors. INAT resisted implementing agreements with the Bank to increase the role of WUAs in project implementation. While outside the scope of the project's objectives, INAT also failed to follow through on recommendations to give greater responsibility for project implementation to its Regional Offices, consistent with the decentralized approach embodied in the Colombian Constitution and INAT's revised statutes. 24. On the other hand, INAT did take positive steps to: seek the opinion of the appropriate legal authority on how it might transfer property rights to WUAs for the districts built before 1994; improve its capacity in non-engineering aspects including  8 establishment of a new department; accelerate implementation and increase efforts to achieve sector policy objectives after the second extension of the closing date; obtain UNDP funds to support the WUAs in the Prado Sevilla district; and secure funds to complete the contracts underway when the loan closed. G. Assessment of Outcome 25. The project attained several useful ends but its overall results were disappointing, mainly because of the unfortunate combination of a narrow strategy and unrealistic objectives. The project targets could not be achieved in the time allotted, even with a two year extension of the closing date. However, achievements concerning sector policy, though only partially attributable to the project, have been greater than expected and are important for the long-term development of the sector. While full cost recovery has not been achieved in all the districts, this is partly attributable to underlying problems which have affected farmers capacity to pay. Concerning transfer of management, the outcome is good, but more support is needed to help the WUAs to fully and effectively assume their responsibilities. In sum, based on the project's current status and reasonable estimates about future production which result in an estimated IRR of 7.8%, the project's outcome is presently rated as deficient. 26. However, it is too early to judge the ultimate development impact of the project. If the civil works are satisfactorily completed, the project could in large measure yet fulfill the objective of improving the retums from irrigation for both the state and farmers. Depending on when this goal is accomplished, the IRR could rise from 7.8% to 15.5- 18.8% which is good compared to other irrigation projects financed by the Bank.5 The impact of management by WUAs and effects of the changed economic incentives on farmers will also be decisive factors. H. Future Operation 27. The second extension of the closing date was based on Govermment's commitment to complete the project using its own funds (estimated US$34 million required) by the end of 1997; an action plan and set of performance indicators for the full two and a half year period were agreed between INAT and the Bank. However, in response to competing investment priorities, INAT has now programmed completion of Rehabilitation II for the year 2000. The Bank has sought clarification from the Govemment about when it plans to complete the project. In addition, the Bank continues to work with INAT under Ln. 3113 to correct problems with the SSI projects financed by the loan. Currently, there are no plans for a new lending operation, but support for strengthening the WUAs through an Institutional Development Grant and an EDI seminar to review experience with irrigation management transfer is planned. The lower end of this range would equate the project's outcome to the average economic returns of over two-thirds of the 192 irrigation projects rated "satisfactory" which were examined in OED's 1994 sector review (Report 13676).  9 L. Key Lessons Learned 28. Some key lessons of the project related to its "quality at entry" are as follows: (a) it is better not to proceed with a lending operation until fundamental agreement on strategy, objectives and key sector policies has been reached (para. 4, 6); (b) the scope and objectives of a follow-on project should be carefully calib,ated to the capacity of the implementing agency as demonstrated in the project under implementation (para. 3); (c) project risks identified at appraisal should be closely monitored and credible solutions found if the risk does materialize (para. 11, 20); (d) to be successful, irrigation and drainage projects must pay attention to agronomic, economic, institutional and social issues in addition to engineering (para. 18); (e) an engineering-minded irrigation agency with a narrow mandate and limited experience in agricultural matters is unlikely to have sufficient interest and capacity to ertsure that on-farm development activities are adequately carried-out (para. 11); (f) the experience in Colombia has shown that WUA management can be an important contributor to improved cost recovery and maintenance; the logic that rehabilitation of irrigation works is a prerequisite to full cost recovery and transfer of management to WUAs was flawed (para. 13); (g) greater involvement of WUAs and the irrigation agency's regional offices in the design and implementation of the project would have improved implementation (para. 19); and (h) low recovery of O&M costs is not necessarily a result of fanners unwillingness to pay or the irrigation agency's poor capacity as a bill-collector, but may stem from long-standing socio-economic problems which a good social assessment during preparation would have uncovered (para 13).  10 Table 1: Summary of Assessments A. Achievement of Substantial Partial Negligible Not Applicable objectives Cv') Cv') Cv') C') Macro policies Oi L Li Sector policies ii O Li Financial objectives Oi Li Li Institutional development n [ o [ Physical objectives Li Poverty reduction El O[" ii Gender issues