Madagascar - Irrigation Rehabilitation Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 14671 PROJECT COMPLETION REPORT MADAGASCAR IRRIGATION REHABILITATION PROJECT (CREDIT 1589-MAG) JUNE 26, 1995 Agricultural and Environment Operations Division South-Central and Indian Ocean Department Africa Region This document has a restricted distribution and may be used by recipients only in the performance of their ofricial duties. Its contents may not otherwise be disclosed without World Bank authorization.  EXCHANGE RATES Year Malagasy francs (FMG) Malagasy francs (F`MG) per SDR per US$ 1985 698.4 662.5 1986 941.6 676.3 1987 1,751.0 1,069.2 1988 2,054.1 1,407.1 1989 2,085.4 1,603.4 1990 2,012.0 1,494.1 1991 2,673.8 1,835.4 1992 2,626.5 1,910.2 1993 2,500.0 1,950.0 WEIGHTS AND MEASURES 1 are (a) = 0.0247 acres I hectare (ha) = 2.47 acres 1 kilometer (km) = 0.62 miles I square kilometer (kM2) = 0.39 square mile 1 kilogram (kg) = 2.20 pounds I liter (1) = 0.26 US gallon 0.22 Imperial gallon 1 ton (t) = 2,204 pounds ABBREVIATIONS CCCE Caisse Francaise de Coop6ration Economique (now called Caisse Frangaise de D6veloppement, CFD, French Development Agency) CFD Caisse Francaise de Developpement (formerly called Caisse Centrale de Coop6ration Economique, CCCE, French Development Agency) DAP Directorate for Agricultural Production DRW Directorate for Rural Works FAC Fonds d'Aide et de Cooperation (French Development Agency) FAO Food and Agriculture Organization FED Fonds Europ6en de D6veloppement (European Development Fund) PRWS Provincial Rural Works Service (Service Provincial de Genie Rural) ERR Economic Rate of Return MPARA Ministry of Agriculture and Rural Development (Ministere d'Etat a l'Agriculture et au Developpement Rural) M&E Monitoring and Evaluation NCU National Coordination Unit O&M Operation and Maintenance PCR Project Completion Report WFP World Food Program WUA Water Users' Association FISCAL YEAR January I - December 31  FOR OMCIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-General Operations Evaluation June 26, 1995 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND TIE PRESIDENT SUBJECT: Project Completion Report on Madagascar - Irrigation Rehabilitation Project (Cr. 1589-MAG) Attached is the Project Completion Report (PCR) on the Madagascar Irrigation Rehabilitation project (Cr. 1589-MAG, approved in FY85) prepared by the Africa Regional Office. Part II was prepared by the Borrower. The project was to develop and apply a systematic approach to rehabilitating Madagascar's many medium-sized irrigation schemes. It aimed to rehabilitate 16 schemes with a total irrigated area of 17,000 hectares. The project was consistent with both Government policy and Bank country strategy. After initial difficulties agreeing a common methodology for rehabilitating schemes, the project was effectively implemented with cofinanciers. Physical targets were fully achieved albeit with a 2.5-year delay. Unit costs of rehabilitation works were higher than estimated in the Staff Appraisal Report in local currency, but as a result of the depreciation of the Malagasy franc, total project costs in SDRs were lower than envisaged. The project benefitted from the liberalization of rice pricing and marketing under the concurrent Bank-financed Agricultural Sector Adjustment Credit. These reforms led to higher prices giving farmers a strong incentive to take advantage of the rehabilitation. Farmers responded with expeditious selling at the farmgate. As a result, agricultural yields grew faster than anticipated, and incremental gains to farm incomes proved much higher than estimated at appraisal. The expected economic rate of return is 18 percent. The project outcome is thus rated as satisfactory, sustainability as likely and its impact on institution building as substantial. The PCR is complete and informative. No audit is planned. Attachment l 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.   FOR OFFICIAL USE ONLY PROJECT COMPLETION REPORT MADAGASCAR IRRIGATION REHABILITATION PROJECT (CREDIT 1589-MAG) TABLE OF CONTENTS PREFACE ................................................. EVALUATION SUMMARY ..................................... PART I: PROJECT REVIEW FROM THE BANK'S PERSPECTIVE ...........1 Project Identity ......................................... Background ............................................ 1 Project Objectives and Description ............................. 2 Project Design and Organization ............................... 2 Project Implementation .................................... 3 Project Results ....................... 6 Project Sustainability ....................... 8 IDA Performance ....................... 9 Collaboration with other Donors ....................... 9 Performance of the Borrower ....................... 10 Consulting Services ....................... 10 Project Documentation and Data .......... 11 PART II: PROJECT REVIEW FROM BORROWER PERSPECTIVE ... ........ 12 PART III: STATISTICAL INFORMATION ........... .. ............... 16 Table 1: Related Bank Credits ............................... 16 Table 2: Project Timetable .................................. 17 Table 3: Credit Disbursement ................................ 18 Table 4: Project Implementation .............. ............... 19 Table 5: Project Costs and Financing ........................... 20 Table 6: Project Results ................................... 22 Table 7: Compliance with Credit Covenants ....................... 25 Table 8: Use of Bank Resources .............................. 26 Maps: IBRD 18378 - Mangoky Development Area IBRD 18285 - Location of Medium Irrigation Scheme IBRD 18286R - Project Irrigation Schemes 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.   i PROJECT COMPLETION REPORT MADAGASCAR IRRIGATION REHABILITATION PROJECT (CR 1589-MAG) PREFACE This is the Project Completion Report (PCR) for the Irrigation Rehabilitation Project in Madagascar, for which Credit 1589-MAG in the amount of SDR 12.6 million (US$12.0 million) was approved on April 30, 1985. On November 14, 1988, an amount of SDR 1.5 million was canceled, following Government's decision to exclude the Mangoky component from the project. The Credit was closed on December 31, 1993, two and half years behind schedule. The last disbursement was on June 6, 1994, at which time, the undisbursed balance of SDR 1.094 million was canceled. Parallel financing was received from the Caisse Francaise de D6veloppement (CFD), formerly the Caisse Centrale de la Cooperation Economique (CCCE), equivalent to SDR 6.5 million (FF 90 million), the Fonds Europeen de Developpement (FED), equivalent to SDR 3.9 million (ECU 9.7 million), the United Nations Development Program (UNDP/WFP-PAM/FAO), equivalent to SDR 2.2 million (US$2.9 million) and the Fonds d'Aide et de Cooperation (FAC), equivalent to SDR 0.3 million (FF 2.0 million). The Government contribution is estimated at SDR 4.9 million (FMG 12 billion). The PCR was prepared by the Agriculture and Environment Operations Division of the South-Central and Indian Ocean Department, Africa Region (Preface, Evaluation Summary, Parts I and III). This report is based, inter alia, on the Staff Appraisal Report, the Development Credit Agreement, supervision reports, correspondence between the Bank and the Borrower, internal Bank memoranda, a field mission in December 1993 and an evaluation report prepared by the Borrower in 1993. The Borrower and the other financiers of the project were sent copies of the PCR for comments, but only the Borrower responded; its comments form Part II of the report.  -  iii PROJECT COMPLETION REPORT MADAGASCAR IRRIGATION REHABILITATION PROJECT (CREDIT 1589-MAG) EVALUATION SUMMARY Background 1. Agriculture is an important economic activity in Madagascar, contributing about 31 percent to gross national product and 60 to 65 percent to export earnings. Agricultural performance had suffered in the seventies and early eighties from ill-conceived Government policies that were aimed at protecting the interests of the politically vocal urban consumers; farm prices were set at too low a level to make agricultural investments attractive and as a result, sector performance was well below the population growth rate during the seventies and early eighties. In the early eighties, Government faced a fiscal crisis following its heavy reliance on borrowing to finance its ambitious public investment program. In response to this crisis, Government embarked on a structural reform process which included the liberalization of prices. Prices for most agricultural products were liberalized by 1986, with a strong production response. 2. Prior to this project, IDA financed three other irrigation projects in the country, including two large irrigation schemes in the Lake Alaotra region. Although these two Lake Alaotra projects succeeded in expanding the irrigated areas and in expanding crop production, they failed to strengthen the executing agency and to create a self-sustaining financial mechanism for operations and maintenance. Project Description and Objectives 3. The project's main objective was to develop a systematic approach to the rehabilitation of medium size irrigation schemes and thus progressively redevelo,. the country's 116 such schemes representing over 140,000 hectares of irrigated rice fields. Other objectives were to improve the organization and management of irrigation rehabilitation and thus encourage other external financiers to participate in the overall rehabilitation program. In its emphasis on self- financing and self-management at the scheme level, the project also sought to develop an alternative to reliance on Government management and funding, which had led to widespread deterioration of the schemes. 4. The project's components were: (a) management support to the Ministry of Agriculture  iv and Rural Development (MPARA) for selection of, planning for and supervision of rehabilitation; (b) civil works for rehabilitation of about 16 medium irrigation schemes. Three schemes had been selected earlier during project preparation, while thirteen other schemes were to be selected by MPARA as a first phase of the rehabilitation program; and (c) rehabilitation of the Mangoky irrigation scheme, including technical assistance, rehabilitation works, maintenance equipment and studies. 5. The project was to be financed by the Caisse Francaise de Developpement (formerly called the Caisse Centrale de Cooperation Economique; SDR 6.3 million), the European Development Fund (SDR 4.2 million), the French Fond d'Aide a la Cooperation (SDR 0.1 million), the Government (SDR 7.9 million) and IDA (SDR 12.6 million). Project Implementation 6. During the first three years, there were difficulties in agreeing on a common methodology for rehabilitating schemes. This factor, added to the policy of using a maximum of local engineering and contracting firms, often with insufficient experience and financial capacity, created delays and errors in studies and works. Consequently, some extra studies and works had to be done for some schemes. Because of these delays, farmers were disinclined to pay maintenance fees on time; they preferred to wait until all works had been finished. 7. The coordination of services of the Ministry of Agriculture that had operational responsibilities (bidding, control of studies and works, extension, etc.) was difficult. The Coordination Committee did not work and a reorganization took place. However, the two main Services (Agricultural Engineering and Agricultural Production) remained too centralized and inadequately staffed. 8. Costs of rehabilitation works were higher than estimated: actual costs were US$1,730 per hectare compared to US$1,050 per hectare at appraisal. 1/ This increase was due to: (a) lack of experience and competence during the first years of the project of engineering and contracting firms; (b) delays by Government in paying its share of rehabilitation costs to contractors, who subsequently increased their submissions in follow-up contracts; and (c) political insecurity during the period 1991/92. Average cost of studies manuals, training and supervision increased from US$610 per hectare in 1986 to US$870 per hectare for recent bids. 9. As a result of local political tensions in the Samangoky region, where the Mangoky irrigation scheme is located (a large scheme of 7,000 ha for rice and cotton production) and misprocurement in the selection of a contractor concerning this component of the project, the Borrower and the Bank decided in November 1986 to exclude the rehabilitation of the Mangoky 1/ The actual average rehabilitation cost per ha for the Lake Alaotra Rice Intensification Project was US$650. This cost is, however, not directly comparable because work in the Lake Alaotra area was less extensive and much more experienced staff was available.  V irrigation scheme from the project and to cancel SDR 1.5 million from the Credit proceeds. Since the cancellation happened only 9 months after Credit 1589-MAG was declared effective, this PCR does not include observations about this component of the project. Project Results 10. With a delay of 2.5 years, the project almost achieved the rehabilitation of the planned 17,000 hectares (actual: 16,750 hectares). Total project cost, estimated at appraisal at SDR 31.1 million amounted, in the event, to SDR 27.7 million. 11. After some delays and mistakes, an adequate methodology for rehabilitating schemes was devised and is now implemented. Farmers are now involved early-on in the rehabilitation process. Local consulting firms now have the necessary experience for pre-selecting schemes, undertaking farmers' orientation studies and carrying out technical studies, preparing bidding documents and supervising works. Ministry agents have also gained from that experience and showed a great sense of willingness to draw lessons from past mistakes and difficulties in improving their working methodology. 12. Three schemes representing 3,200 hectares were rehabilitated, that should not have been selected (large scale up-stream erosion, insufficient water, susceptibility to cyclone damage). Rehabilitation works have almost been completed on the remaining 27 schemes. Supplementary works have been carried out on a number of schemes in response to environmental problems, design mistakes and farmers demands. Farmers of all schemes have created Water Users' Associations (WUAs) that are active, but by June 1994, only 45 out of 85 WUAs had been officially recognized by the Ministry, the remaining 40 are still under review. For schemes where all works are finished, WUAs have begun to collect maintenance fees, and for the three schemes that had been fully operational during the 1993/94 agricultural year about 80 percent of participating farmers had paid these fees. 13. Agricultural production results are also successful. Increases of rice yields are higher than estimated at appraisal and the final yield increase of 940 kg per hectare predicted for year 9 after rehabilitation will be attained. The average economic rate of return over 30 years is 18 percent. This rate is the weighted average of 9 schemes for which data exist: 6 "normal" schemes representing 5,870 ha, and 3 "risky" schemes in terms of environmental problems, representing 3,200 ha. Farm incomes are much higher than estimated at appraisal, which is due in part to the effects of price liberalization. The incremental net financial benefit for 1993 is estimated at 310,000 FMG per ha in constant 1986 FMG versus 67,000 FMG per ha at appraisal. Project Sustainability 14. Overall project sustainability is likely. Some individual schemes will require additional investments in order to address environmental issues that had not been adequately analyzed earlier. Sustainability is doubtful only in the case of the three rehabilitated schemes mentioned  vi earlier which experienced serious problems of siltation, climate and lack of water. The feasibility of further remedial works on these schemes is presently being studied. 15. In terms of institution building, WUAs will need to be assisted for some time to come to strengthen their managerial capacity for the collection of maintenance fees and the organization of water usage. Capacity building and decentralization have still to be strengthened in MPARA and the rehabilitation methodology used by Government and donors needs to be further fine-tuned (greater farmer participation in each phase and phasing of rehabilitation conditional on continued farmer participation). Lessons Learned 16. The main lessons learned are: (a) The selection criteria for schemes to be rehabilitated are essential for the success and sustainability of the rehabilitation. In particular, a balanced set of criteria has to be used with respect to: (i) physical characteristics of the scheme: water resources, soil fertility, environmental aspects and climatic hazards (cyclones); (ii) social cohesiveness considering tribal and ethnic aspects as well as the need for cooperation between up and down stream farmers; and (iii) financial potential considering the costs of rehabilitation and the economic costs of accessing the area for works, agricultural inputs and outputs. (b) Water users are ready to pay for maintenance provided that rehabilitation works are adequately designed and formulated with participation of likely beneficiaries, maintenance fees have been agreed with them prior to the rehabilitation and all works have been completed. (c) Regulatory changes should have been initiated at the beginning of the project and made effective before rehabilitation works were finished in order to allow timely management transfer to WUAs. (d) A monitoring and evaluation system of the project needs to be established at the onset of the project.  vii (e) In view of the difficulties experienced by individual agricultural projects such as this project in obtaining the support of national service projects such as extension and research, arrangements for such cooperation, together with the necessary funding, should be formally agreed upon and included in the contracts of the projects requiring such assistance.   1 PROJECT COMPLETION REPORT MADAGASCAR IRRIGATION REHABILITATION PROJECT (CREDIT 1589-MAG) PART I: PROJECT REVIEW FROM THE BANK'S PERSPECTIVE 1. Project Identity Project Name Irrigation Rehabilitation Project Credit Number 1589-MAG RVP Unit : Africa Department South-Central and Indian Ocean Department Country : Madagascar Sector : Agriculture Sub-sector Irrigation 2. Background 2.1 Agriculture is an important economic activity in Madagascar, contributing about 31 percent to gross national product and 60 to 65 percent to export earnings. Agricultural performance had suffered in the seventies and early eighties from ill-conceived Government policies that were aimed at protecting the interests of the politically vocal urban consumers; farm prices were set at too low a level to make agricultural investments attractive and as a result, sector performance was well below the population growth rate during the seventies and early eighties. In the early eighties, Government faced a fiscal crisis following its heavy reliance on borrowing to finance its ambitious public investment program. In response to this crisis, Government embarked on a structural reform process which included the liberalization of prices. Prices for most agricultural products were liberalized by 1986, with a strong production response. 2.2 The Government's goal for the sector is to achieve self-sufficiency, increase the production of export crops, expand agro-industries to process raw materials and increase rural employment. During the sixties, Madagascar was a net exporter of rice, but because of the ill conceived macro-economic policies, the country became a net rice importer (over 200,000 tons in 1984). However, elimination of price and trade controls in the mid-eighties helped check this trend, with only 40,000 tons imported in 1992.  2 2.3 Prior to this project, IDA financed three other irrigation projects in Madagascar (the Morondava Irrigation and Rural Development Project, Cr. 0322-MAG, the Lake Alaotra Irrigation Project, Cr. 0214-MAG, and the Lake Alaotra Rice Intensification Project, Cr. 1337- MAG). The Morondava Project suffered greatly from Government's macro-economic policies. As a result, the project had to be drastically restructured (high local inflation and price controls on inputs and outputs). The PCR rated project performance as unsatisfactory. The two Lake Alaotra projects succeeded in achieving their physical objectives. They both failed, however, to strengthen the executing agency and to create a self sustaining financial mechanism for operating and maintaining the irrigation network. Consequently, project results have not been sustainable. 3. Project Objectives and Description 3.1 The project's main objective was to develop a systematic approach to the rehabilitation of medium-sized irrigation schemes and thus progressively redevelop the country's 116 such schemes representing over 140,000 hectares of irrigated rice fields. Other objectives were to improve the organization and management of irrigation rehabilitation and thus encourage other external financiers to participate in the overall rehabilitation program. In its emphasis on self- financing and self-management at the scheme level, the project also sought to develop an alternative to reliance on Government management and funding, which had led to widespread deterioration of the schemes. 3.2 The project's components were: (a) management support to MPARA for selection of, planning for, and supervision of rehabilitation; (b) civil works for reconstruction of about 16 medium-sized irrigation schemes, totaling about 17,000 ha. Three schemes had been selected earlier during project preparation and their rehabilitation was to be financed by the French Caisse Central de Cooperation Economique (CCCE) and the Fonds Europ6en de Developpement (FED). Thirteen other schemes were to be selected by MPARA, as a first phase of the rehabilitation program; and (c) rehabilitation of the Mangoky irrigation scheme, including technical assistance, rehabilitation works, maintenance equipment and studies. 4. Project Design and Organization 4.1 The project was designed to develop a systematic approach to the rehabilitation of irrigation schemes with the aim of transferring the operation and maintenance (O&M) responsibilities to the farmers at the end of the rehabilitation phase, thus reducing Government's  3 financial involvement to a minimum and increasing the prospects for sustainability of the investment. This conceptual basis for the project was clear to and accepted by the Government and the various co-financiers. 4.2 Project design depended heavily on active collaboration/participation at all stages of the rehabilitation process between farmers and government services (Directorate for Rural Works, DRW, and Directorate for Agricultural Production, DAP) at all levels. The difficulties in achieving this in practice were, however, greatly underestimated. As a result, at the beginning of the project, when the rehabilitation methodology was still in a testing phase, the criteria for selecting schemes for rehabilitation did not take into account some non-technical aspects such as relationships between groups of farmers and environmental hazards (upstream erosion, cyclones). This resulted in rehabilitation delays (an average of about 40 months per scheme between the start of studies and the end of works), the need for additional works after the rehabilitation was finished, and problems in collecting user fees. 4.3 In terms of project organization, the National Coordination Unit (NCU) was responsible for planning, monitoring and evaluation (M&E), and dialogue with the co-financiers. The DRW was responsible for supervising the consultants and contractors for civil works, while the DAP was responsible for the agricultural production components including research, extension and M&E. As the project financed training and only a limited quantity of operational requisites (such as vehicles and equipment) for these two Directorates, it was clear that the project was accorded only low priority. The Coordination Committee, which was to play the role of a facilitator and coordinator for these Directorates, did not assume its intended role despite repeated donor efforts. The NCU was originally located within the Planning Department of MPARA, but in 1992 it was decided to place it directly under the Secretary General of MPARA, which greatly improved its effectiveness. Other services of MPARA were unable to provide the expected level of training and M&E, leaving the NCU overburdened with a number of tasks for which it was not designed. 5. Project Implementation Macro-Economic Environment 5.1 The project became effective at a time when the prices of agricultural inputs and outputs were being gradually liberalized. As a result, real prices of rice inc-'uased more than the cost of inputs, and net incomes of rice growing farmers improved. In general, this project was not negatively affected by conditionality imposed by sectoral adjustment credits. However, the freeze on hiring new staff to replace retiring staff, imposed as part of the conditionality for the Public Sector Adjustment Credit (Cr. 1941-MAG), hurt project implementation because it was impossible for MPARA to hire new engineers and other technical staff to assist in project execution.  4 Project Effectiveness and Closing 5.2 The Credit was declared effective on February 3, 1986, a delay of less than two months. The planned closing date was June 30, 1991; due to delays in rehabilitation studies and civil works execution during the first three years, the closing date was extended three times. The credit closed on December 31, 1993. Project Cost and Financing 5.3 In SDR terms, actual project costs were about 10 percent less than estimated (SDR 27.7 million actual versus SDR 3 1. I million estimated). In US dollars terms, they were higher than estimated at appraisal, because of the depreciation of the US dollar relative to the SDR and the currencies of main trading partners. Actual project costs were US$38.2 million versus an estimate of US$29.6 million. IDA contributed US$12.1 million versus US$12 million estimated. Actual contributions from other external donors (CFD, FED, FAC and United Nations Development Program) were US$17.8 million versus US$10.1 million estimated and Government contributed US$6.8 million versus US$7.1 million estimated. 5.4 Increases in project costs resulted primarily from higher than estimated costs of rehabilitation. Actual rehabilitation costs (excluding studies and supervision) were US$1,730 per hectare compared to US$1,050 per hectare at appraisal. 2/ This increase was due to: i) lack of experience and competence, during the first years of the project, of engineering and contracting firms; ii) difficulties of Government in providing timely counterpart funds to pay contractors who then increased their unit rates in subsequent contracts; and iii) political instability during 1991 and 1992. The average cost of studies, operational manuals, training and supervision increased from US$610 per hectare in 1986 to US$870 per hectare for recent bids. Moreover, the extension of the project by two and a half years necessitated the longer than expected presence of consultants who were providing technical assistance and thereby contributed to the increase in project costs. Disbursements 5.5 Disbursements initially lagged due to the slow start-up of project activities, however, this lag was greatly reduced during the last two years of project implementation. As a result of local political tensions in the Samangoky region and misprocurement in the selection of a contractor for that component of the project, the Borrower and the Bank decided to exclude the rehabilitation of the Mangoky irrigation scheme from the project and to cancel SDR 1.5 million from the Credit proceeds in November 1986. The last disbursement was on June 6, 1994, at which time the undisbursed balance of SDR 1.094 million was canceled. 2/ The average rehabilitation cost per ha for the Lake Alaotra Rice Intensification Project was US$650. This cost is, however, not directly comparable because work in the Lake Alaotra area was less extensive and much more experienced staff was available.  5 5.6 Disbursement arrangements through the Special Account were not satisfactory due to foreign exchange restrictions in effect during most of project Implementation; all payments were made by a New York bank, resulting in substantial delays in payments to local contractors. Also, the cost sharing arrangements in the Development Credit Agreement between Government and IDA for civil works (65 percent IDA and 35 percent Government) proved unrealistic during the later stages of project implementation, when Government was not able to provide its share in a timely manner. Implementation Issues 5.7 The lack, at project start-up, of a well defined rehabilitation methodology, which included farmers' participation in its different phases, led to a number of problems ranging from poor engineering designs to farmers' reluctance to pay O&M fees. It was not until mid-1989 that an acceptable methodology based on stronger farmers' participation and training at all levels, was developed by the NCU and agreed upon by all co-financiers. This methodology included the following steps: (i) pre-selecting schemes; (ii) sensitizing and mobilizing farmers to create WUAs; (iii) carrying out the necessary studies in consultation with the beneficiaries; (iv) selecting contractors; (v) supervising the execution of the contracts; (vi) estimating and collecting annual maintenance fees; and (vii) developing agricultural development plans and annual production plans for the schemes. Following completion of the rehabilitation, regular maintenance was to become the responsibility of the WUAs. 5.8 The Bank recognized the need to develop a well-thought out rehabilitation methodology prior to undertaking any rehabilitation activity. 3/ Other co-financiers, however, went ahead with the physical rehabilitation hoping to solve the participation and training problems along the line. These two different approaches initially sent conflicting messages to water users. Moreover, for some schemes this led to ill-prepared rehabilitation studies, which necessitated supplementary works after completion of the initial rehabilitation. This was the factor which led to the unwillingness of farmers to pay O&M fees before all works had been finished. Nevertheless, with the implementation of the revised strategy in mid-1989, better results were obtained, and the design and execution of rehabilitation works was carried out in closer participation with users. 5.9 Project implementation was based on dynamic coordination between the NCU and the DRW and DAP at the three levels: national, regional and local/scheme. At the national level, the Coordination Committee did not work. At the regional level, the responsibility for coordination was, for most provinces, given to the Provincial Director of Rural Works, which in some instances led to conflicts with the Provincial Director of Agriculture Production. At the local level there were three agents: one Training Advisor (paid by the project) for helping farmers' associations, one extension agent from the DAP and one Head of Irrigation Scheme from the Provincial Rural Works Service (PRWS). Monthly meetings of those three agents were 3/ The Bank's position was also that the Borrower (see para. 2.1, Part II).  6 held under the leadership of the Head of Zone from the Provincial Agricultural Production Service (PADS). This arrangement, involving rehabilitation and shared leadership between PRWS and PADS agents did not facilitate coordination. 5.10 Extension, research and M&E efforts in rehabilitated schemes were to be executed by the DAP, with support from the IDA financed Agricultural Pilot Extension (Cr. 2150-MAG) and the National Agricultural Research (Cr. 2042-MAG) projects and by technical assistance financed by CFD. In practice, however, the project received very limited assistance from these two projects. Schemes rehabilitated with CFD financing received more intensive extension and research support because of closer coordination between the various CFD supported projects. 5.11 The Coordination Committee established at the beginning of the project convened several times, but, due to inter-departmental rivalry, it never took an active role in the implementation of the project. After the first two years, it ceased to function and was slowly dismantled despite frequent reminders from the co-financiers. Eventually, in early 1992, the project was placed directly under the authority of the Secretary General of MPARA, which led to a better coordination between the various directorates dealing with the project. Early weaknesses in staffing and composition of the NCU also contributed to its weak initial performance. However, with the appointment of a new National Coordinator and the replacement of some expatriates, the situation improved dramatically in late 1989/early 1990. New development methodologies were put in place (see also para. 5.7), an M&E system was designed and implemented which was helpful in monitoring results in the field (surfaces, yields, number of farmers, etc), agricultural plans for each scheme were devised, and training programs were implemented. 5.12 Works for the IDA financed part of the project were awarded, through local competitive bidding, to local contractors, even though they did not always have the requisite technical and financial capacity. It was felt that the advantages of this approach, primarily increased local involvement and capacity building, outweighed the ensuing implementation delays. In contrast, some of the large consulting engineering companies, participating in the project, had a tendency to discuss technical problems directly with the donors, bypassing local governments and WUAs, which undermined the much needed farmer participation. However, the limited financial capacity of some of the smaller contractors put them in a difficult situation when Government was unable to make timely payments for its share of the costs of civil works (35 percent) because of the precarious state of Government finances (see also para. 5.4). Most contractors inflated their subsequent bids in order to compensate for this factor, which was a major cause for the higher than estimated rehabilitation costs. 6. Project Results 6.1 The project was successful in achieving its main objective: to develop a systematic approach to the rehabilitation of medium sized irrigation schemes with increased involvement of beneficiaries, as an alternative to reliance on Government for management and maintenance. This rehabilitation methodology is to be applied nationwide in the follow-on project under  7 preparation. About 310 schemes have been identified for possible rehabilitation, and 140 have been pre-selected for inclusion in the follow-on project. 4/ 6.2 The objective of rehabilitating 17,000 ha had almost been achieved by the end of the first quarter of 1994 (12,989 ha had been rehabilitated by the end of 1993, 3,761 ha were to be completed by the end of the first quarter of 1994 and 452 ha scheduled to be completed by mid- 1994, depending on availability of Government funds, for a total of 17,202 ha). Although further repairs and complementary works will have to be carried out on some schemes, mainly due to lack of experience of local engineering and contracting firms, the rehabilitated schemes are generally of sound technical quality. However, three schemes should probably not have been rehabilitated in the first instance, in view of adverse environmental factors that are too expensive to remedy: (i) Behara - insufficient water; (ii) Belamoty - siltation due to upstream erosion; and (iii) Taheza - cyclones. These three schemes represent 3,200 ha or 19 percent of the rehabilitated area. 6.3 Project design called for strengthening existing MPARA services but not for creating an independent organization which would provide the same services outside the Government structure. The NCU coordinated the operations of the DRW and DAP. Three main obstacles remain: i) difficult coordination and collaboration between the two branches at different levels, particularly with the DRW which is too centralized in the capital; b) the weak absorptive capacity of field agents from both directorates due to their poor technical education; and c) the difficulty in improving personnel skills mix by replacing staff going on retirement by younger and better trained new staff, since all recruitment was frozen as part of the effort to rationalize public finances. The impact of technical assistance on staff development was less than planned because of the lack of counterpart staff and poor timing of technical assistance. 6.4 Except for production on the Behara, Belamoty and Taheza schemes, agricultural production has increased following completion of the rehabilitation (including establishment and training of WUAs). On a sample of six schemes representing 5,870 ha (34 percent of the rehabilitated area), rice yield three years after rehabilitation rose to 605 kg/ha compared to 150 kg/ha estimated at appraisal. No scheme has been rehabilitated for long enough to reach a constant yield after 9 years to properly assess project impact, but all data indicate that the appraisal objective will be attained. For the sample of six schemes, the rehabilitation allowed farmers to grow a diversity of off-season crops (wheat, potatoes, vegetables, etc.). About 20 percent of the rehabilitated surface is now used for this purpose (practically none before rehabilitation), adding substantially to farmers' food security. A, schemes have now an Agriculture Development Plan that has been formulated by farmers with the assistance of extension workers. 6.5 During project implementation, about 85 WUAs were created with the assistance of the 4/ This includes medium-sized and small irrigation schemes. The number in para. 3.1 refers to medium-sized schemes only.  8 Training Advisors (temporary employees paid by the project from counterpart funds). Due to administrative tardiness, only 45 WUAs had been officially recognized by MPARA by June 1994. This will delay the transfer of scheme management to the WUAs and thus affect project sustainability. The fact that about 80 percent of the farmers on the three fully operational schemes paid maintenance fees for the first time by the end of the 1993/94 agricultural year demonstrates that farmers are willing to pay for maintenance, but only when all works are completed and of good quality, and maintenance fees have been discussed with WUAs prior to rehabilitation, with follow-up discussions planned to take place every year. It will be crucial for the sustainability of the project to monitor closely the collection of maintenance fees. On some schemes, farmers created associations for obtaining bank credits and for procurement and marketing. 6.6 Economic Results. The economic analysis for the six sample schemes yields an Economic Rate of Return (ERR) of about 20 percent. The same assumptions were used as at appraisal, with the exception that the hypothesis of declining farm income in the without project situation of 7 percent per annum over the first six years of project implementation was dropped because it was found to be an invalid assumption. The ERR for the 3 environmentally "risky" schemes (Behara, Belamoty and Taheza) representing 3,200 hectares is negative. On extrapolating the data for these nine schemes to the entire project, the ERR is found to be 18 percent, as compared to the appraisal estimate of 23 percent. For farmers served by the Soavina scheme (1,200 ha), for which estimated and actual results are known, the incremental net financial benefit for 1993 is estimated at 310,000 FMG per ha in constant 1986 FMG versus 67,000 FMG per ha at appraisal. The main causes for the differences between the actual and estimated financial benefits are higher paddy prices, resulting from the elimination of price and trade controls in 1986, lower production costs, and intensification of land use with off-season crops. For this scheme, the actual ERR is 33.8 percent (21 percent estimated at appraisal). 6.7 An important side effect of project implementation was substantial capacity building at the local level because of the selection of a large number of local firms for participation in engineering studies and works execution. Experience shows that great care has to be taken in selecting and monitoring local firms in order to enable their performance to be sustainable. Local contracting firms were asked to hire a maximum of unskilled labor and hourly paid workers from populations near the site. This policy brought employment and income to local residents, and some of those former workers are now better trained to participate in maintenance works on the schemes. 7. Project Sustainability 7.1 Overall project sustainability is likely. Some individual schemes will require additional investments in order to address environmental issues that had not been adequately analyzed earlier. Sustainability is doubtful only in the case of the three rehabilitated schemes mentioned earlier which experienced serious problems of siltation, climate and lack of water. The feasibility of further remedial works on these schemes is presently being studied.  9 7.2 In terms of institution building, WUAs will need to be assisted for some time to come to strengthen their managerial capacity for the collection of maintenance fees and the organization of water usage. Capacity building and decentralization have still to be strengthened in MPARA and the rehabilitation methodology used by Government and donors needs to be further fine-tuned. 8. IDA Performance 8.1 The project was carefully and regularly supervised by the Bank and the other main co- donors (CFD, FED, FAC). Bank supervision missions, however, consisted usually of only one Bank staff, an irrigation engineer. It would have been helpful to include in some of these missions an agriculturist, in particular, for helping the NCU to make arrangements with the PNVA and the PNRA projects. Supervision missions insisted on the necessity of having a systematic approach to irrigation scheme rehabilitation and involving farmers in the different stages of the process with a view of facilitating the transfer of responsibility for the schemes to users at the end of the rehabilitation process. Several supervision missions raised doubts about the effectiveness of the training program, which was judged to be too theoretical for field agents (extension and water management). 5/ 8.2 The Bank and the Borrower gave insufficient attention to the pre-qualification of local contractors and since all contracts were based on Local Competitive Bidding (LCB), this led to the award of some contracts to contractors with doubtful credentials and insufficient experience. On the other hand, if strict pre-qualification rules had been adopted, very few local contractors could have qualified. Subdividing works into smaller contracts proved to be a good solution. 8.3 Bank staff did not arrange for procurement training of project staff until late in the project, which was partly the reason for several procurement delays. 9. Collaboration with other Donors 9.1 During the first 2-3 years of the project and probably due to the fact that Government and other donors did not share a common approach toward scheme rehabilitation, donors had a tendency to go it alone. Each donor selected separate areas to finance rehabilitation: IDA (Central and North-West), CFD (Central Highlands) and FED (North and South regions). During those first years, MPARA agents (NCU, DRW and DAP) had the impression that the expatriate technical experts were there to assist the donors and not Government. This perception did not help in motivating MPARA personnel. Later on, relations improved as a consequence 5/ The Borrower feels that the resident mission could have played a more active role in facilitating the approval process, the latter, however, lacked the trained manpower to do so (see para. 4.4, Part 11).  10 of some changes in the technical assistance team and changes in NCU management and its relocation within MPARA. 10. Performance of the Borrower 10.1 The NCU performed well in executing the project, especially in the last 4 years. However, the lack of a single rehabilitation strategy and the relatively weak leadership during the initial years did affect to a large extent the course of the project and resulted in substantial delays. The under-staffing and lack of experienced engineers at the DRW, with most staff located in the capital, contributed to the need for additional works to correct errors. 10.2 The Coordination Committee did not function as foreseen and stopped meeting after the first two years of the project. The Borrower did not actively pursue the objective of coordination between donors. The relationship between the NCU and the DRW was not always smooth, in part due to the unclear definition of the role of each agency in contract award, supervision and disbursement decisions. Despite improvements, there is still a lack of active collaboration between agents of the DRW and DAP at all levels (national, regional, local/scheme). 10.3 The lengthy process of awarding contracts was one of the major bottlenecks in the execution of the project (usually taking 6 to 10 months). In addition, lack of counterpart funds invariably led to delays in paying contractors. 11. Consulting Services 11. 1 The project used a large number of consultants for technical assistance. This team did bring in considerable technical expertise, benefiting MPARA staff. However, even though a sizable training effort was undertaken, some expatriates left without having been able to hand over their responsibilities to counterparts, for lack of qualified local staff. 11.2 The project used a large number of expatriate experts in different areas (irrigation, agriculture, extension, institution building, training, finance, and general management) at national and regional levels. Although at the beginning of the project, they had the tendency to run the project as their own with difficult relations with their local counterparts, in general, they performed well throughout project implementation. At one time their number became too high (14 long-term positions in 1991) as some donors preferred to have expatriate technical personnel at both the regional and central levels of the project. This practice led to different implementation approaches for each cluster of schemes, depending on the source of financing. The number of members of the TA team was subsequently reduced to 6, most of whom will  11 leave by late April 1994. 6/ 11.3 The few studies carried out by individual consultants were generally of good quality. Unfortunately, the performance of some local consulting engineering firms was, in some cases, not up to standard, which resulted in weaknesses in design and poor rehabilitation. 11.4 Auditing of project accounts was undertaken over the course of the project by two different auditing firms, both of which proved to be of an acceptable standard. 12. Project Documentation and Data 12.1 The one and five year work programs were available and updated throughout the course of the project. These reports proved to be very useful in supervision and monitoring. A comprehensive M&E report was produced in 1992. This document provided the basis for the preparation of the follow-up project. Orientation studies, pre-feasibility studies, feasibility studies and final designs have been produced all along the project for each scheme. In addition, a large number of documents and reports dealing with the various aspects of the project (agricultural production, rural institutions, financial accounts, etc.) were prepared by project personnel and made available to visiting missions. For each supervision mission, the NCU prepared in advance a set of short reports which proved beneficial in discussions with Government and beneficiaries. 12.2 The SAR was useful in providing a framework for project implementation; it was very often referred to during project implementation although its presentation did not always facilitate the readers' task. Data were, in general, readily available, thanks to an understanding between the NCU and the Bank, on the need for proper databases and M&E. During the first two years of the project, some production data were estimated using ill-suited approaches. The methodology was subsequently adjusted and the reliability of estimates improved. Production data used for the economic analysis of the project were obtained from the M&E section of the project and the DAP. 6/ The Borrower has commented that the timing of bilaterally financed technical assistance did not coincide with its particular needs (see para. 3.1, Part II). The mismatch in timing was a real problem.  12 PROJECT COMPLETION REPORT MADAGASCAR IRRIGATION REHABILITATION PROJECT (CREDIT 1589-MAG) PART II: PROJECT REVIEW FROM BORROWER PERSPECTIVE MINISTRY OF AGRICULTURE AND RURAL DEVELOPMENT (MPARA) NATIONAL COORDINATION UNIT FOR THE IRRIGATION REHABILITATION PROJECT 1. PROJECT COORDINATION 1.1 Project coordination was assigned to the Programming Department, perhaps because it had a rice coordination office whose head was appointed National Coordinator and whose entire staff was incorporated into the National Coordination Unit for the Irrigation Rehabilitation Project. 1.2 In principle, according to project documents, the Coordinator was to ensure the secretariat for the Coordinating Committee. The Coordinating Committee ran smoothly in the early years, but subsequently the project teams within each department gradually severed themselves from their hierarchical attachments and clustered around the National Coordinator. The result of this was a lack of interest on the part of the directors, and the meetings of the Coordinating Committee became merely information briefings on project activities without true involvement of the directors in decision-making. 1.3 The problem of the location of the Coordinator in the hierarchy of the Ministry was resolved by attaching the Project Coordination Unit to the Secretariat, with the rank of Ministry Director for the National Coordinator. 2. PROJECT IMIPLEMENTATION 2.1 The project got off to a premature start; the appraisal document does not appear to have thoroughly analyzed the actual implementation of the various project components. Activities should have started with the inventory and formulation of the methodology for establishing the water user associations, before any studies or rehabilitation works were done. 2.2 Unfortunately, the rehabilitation works at Behara (EDF), Belamoty (EDF) and Soavina  13 (CFD) began immediately, as did the rehabilitation studies for the four central highlands irrigation rehabilitation areas (CFD), when only the major principles of farmer accountability had been mapped out. One of the key objectives of the project was the establishment of autonomous management and maintenance by user associations. The methodology for establishing those associations should have been formulated first, but that document was not finalized until January 1991. The rehabilitation operations should have begun after that date. 2.3 There were three categories of irrigation areas: (a) Category 1: The first three schemes rehabilitated. The water user associations were set up following the rehabilitation works. Because of their ill-conceived structure and organization, the framework for farmer involvement in these areas must be totally revamped. (b) Category 2: The four central highland (Antsirabe) schemes. The farmer organization methodology was based on actual experience in these irrigation areas; the dialogue with users during the studies was not undertaken systematically. Supplementary works were found to be necessary following the resumption of farmer consultations. (c) Category 3: All irrigation schemes after 1991. The methodology was finalized and the results in terms of farmer accountability are promising. 2.4 The schemes to be rehabilitated were scheduled on the basis of a list of classified areas, without their exact status being clear cut. The results of the inventory/survey and pre-selection briefs should have been in before the scheduling of the areas to be rehabilitated was done. This was covered by the project document entitled " 1991-95 action plan" prepared in December 1990. 2.5 Because the technical teams reported to the Regional Coordinator, the technical departments did not consider themselves responsible for the implementation of the pertinent component; for example, the agricultural development component in the rehabilitation studies was monitored only by the central project unit and the Agricultural Department remained totally out of the picture, despite the fact that the implementation of agricultural development programs and cropping plans are the responsibility of the regional agricultural services. 2.6 The inclusion of SAMANGOKY in project financing seems to have been a mistake; in point of fact, the issues at stake were different (large irrigation schemes) and the establishment of water user associations was not planned. Moreover, its implementation was completely outside the overview of the National Coordination Unit. 2.7 One of the project objectives was strengthening management support to the Ministry of Agriculture; vehicles and equipment were acquired under the project. However, the day-to-day running of the project was not expressly covered in the project document. As it turned out, project agents had to make field visits (inventory, surveys, studies supervision and works, etc.)  14 and their pay from the government barely covered the cost of a dish of rice. The cost of the day-to-day running of the project was therefore included in the technical assistance contract, which prevented the Project Coordination Unit from providing management oversight. In the future, the heading of day-to-day project administration should be considered part of the investment and should be mentioned clearly in the Credit Agreement. 2.8 With regard to disbursements, a special account was opened with a bank outside of Madagascar (Irving Trust, New York). The interest on the special account was lost because the payment deadlines were close to those for direct payments. A project account managed by the National Coordinator and opened in a local primary bank made it possible to resolve payment timing difficulties on the one hand and on the other to ensure project cash flow for day-to-day operations. 3. TECHNICAL ASSISTANCE 3.1 Project technical assistance was very badly planned. The timing was poor, given the large-scale arrival of technical assistance at the start of the project (especially regarding rural engineering) and the project activities, which focused on the inventory/survey, pre-selection briefs and farmer organization. Subsequently, the number of technical assistants in rural engineering was reduced precisely when needs swelled with the start-up of the studies and works. 3.2 Better scheduling of technical assistance throughout the duration of the first phase of the project would have had a positive impact on the transfer of know-how to the offices of the Ministry of Agriculture. 4. RELATIONS WITH DONORS 4.1 Although at the outset project appraisal was done jointly by the three main donors (IDA, CFD, EDF), subsequently each conducted project supervision on an individual basis. 4.2 EDF has continued its operations in the south and north of the country, whereas CFD has concentrated on the central highlands in the southeast. 4.3 The CFD and EDF technical representations in Madagascar facilitated contacts and discussions and shortened the decision-making time frame. 4.4 With regard to IDA, the office of the resident representative did not contribute much to the project, other than serving as a mailbox for letters to Washington. Channels were therefore  15 very lengthy owing to mail delivery. 7/ 4.5 Because of the personality of the Task Manager who served until the end of 1990, national coordination for the project was seen by the other donors as the domain of IDA, which was not conducive to good coordination with the donors. After 1991, when a new Task Manager was appointed, the situation improved markedly; discussions with the other donors were effective during supervision missions and the wrap-up meetings always included all parties. 4.6 However, with a view to strengthening coordination among donors, a joint monitoring mission by the donors should be held at least once a year. 7/ The resident mission lacked the trained manpower to play an active role during project implementation (see para. 8. 1, Part I).  16 PROJECT COMPLETION REPORT MADAGASCAR IRRIGATION REHABILITATION PROJECT (CREDIT 1589-MAG) PART III: STATISTICAL INFORMATION Table 1: Related Bank Credits Credit Name Credit Purpose Year | Status Comments Number Approved I PNVA 2150-MAG Strengthen 1990 On-going To be closed agricultural in June 1994 extension PNRA 2042-MAG Strengthen 1989 On-going To be closed agricultural in Dec. 1996 research Lake Alaotra 1337-MAG Large 1983 Closed on irrigation March 31, schemes 1991 Rural Credit 2459-MAG Rural credit 1993 On-going To be closed in June 1997  17 Table 2: Project Timetable Item Date planned Date revised Date actual Identification 1979 1979 Preparation 1982 1982 Pre-appraisal 11/82 11/82 Complementary Studies 01/84 01/84 Appraisal Mission 04/84 04/84 Post-appraisal I 07/84 Post-appraisal II 10/84 Credit Negotiations 03/85 03/85 Board Approval 04/85 04/30/85 Credit Signature 09/85 09/12/85 Credit Effectiveness 01/86 02/03/86 Credit Closing 06/30/91 06/30/93 12/31/93 Project Completion 12/31/91 09/30/93 03/31/94 Comments: This project was viewed as a pilot project to test methods for rehabilitating small and medium size irrigation themes at pre-appraisal, further studies had to be done by the Borrower (scheme surveys, costs, etc.) due to delays in studies and works experienced during the first years, project closing was postponed three times.  18 Table 3: Credit Disbursement (in millions of US$) Fiscal Year Estimated Actual Actual as % of Estimated (Cumulative) Annual Cumulative Annual Cumulative 1985 0.1 0.1 0 0 0 1986 1.3 1.4 1.3 1.3 92 1987 2.5 3.9 0.4 1.7 44 1988 3.2 7.1 2.5 4.1 58 1989 2.8 9.9 1 5.1 52 1990 1.4 11.3 1.7 6.8 60 1991 0.7 12.0 0.8 7.6 63 1992 1.5 9.1 76 1993 1.2 10.3 86 1994 3.2 13.4 112 TOTAL 12.0 13.4  19 Table 4: Project Implementation Item Appraisal Actual Comments Estimate Estimate at Closing Date Technical Assistance IDA Financing 29 man/year 33 man/year Longer than planned Other Donors n.a. 47 man/year Total 80 man/year 6 TA present on Dec. 1993 Local Training man x days man x days For project Coordination n.a. 1,180 For production MINAGRI n.a. 3,400 For Genie Rural MINAGRI n.a. 300 Farmers' Associations n.a. 500 Training outside country man x days man x days For project Coordination n.a. 450 For production MINAGRI n.a. 569 For Genie Rural, MINAGRI n.a. 210 For Program Director n.a. 55 Investment cost per ha US$1,050/ha US$1,730/ha Without technical (studies & works) assistance Note: n.a. = not available  Table 5: Project Costs and Financing A. Project Costs Item A raisal Estimate Revised Estimate Actual (estimate) Mitlion FMG SDR '000 % Million FNG SOR '000 Million FMG SDR '000 % Works 8,900 12,400 40 31,500 15,000 27,500 13,000 47 Equipnent & VehicLes 300 400 1 1,000 500 1,500 700 3 Technical Assistance, 2,900 4,100 13 17,600 8,400 25,000 11,800 43 Studies & Training Operating 2,600 3,700 12 4,200 2,000 3,100 1,500 5 Contingencies 5,800 8,100 26 8,500 4000- - 0 Samangoky 1,700 2,400 8 1,900 900 1,000 700 2 Total 22,100 31,100 100 64,700 30,800 58,100 27,700 100 O Comments: Due to FMG depreciation during the period, project cost in FMG has considerably increased over appraisal while project costs in SDR have been lower; the cost of Samangoky component was planned to be SDR 2.4 million. Actual cost for the component was SDR 0.7 million (component stopped 9 months after Credit effectiveness).  B. Project Financing Item Apraisal Estimate Revised Estimate Actual (estimate) MiLLion FMG SDR '000 | ition F1G | R '000 Million FMG SDR '000 W IDA _ l Works 3,000 4,200 14 3,800 2,550 3,100 2200 8 Equipment & Vehicles 400 500 2 800 510 1,000 700 3 Technical Assistance, 1,800 2,500 8 10,400 6,900 8,900 6,230 22 Sturdies & Training l Operating 600 900 3 - - - - 0 Prefinancing 600 800 3 100 70 100 70 - Contingencies 900 1,300 4 250 170 - - - Samangoky 1700 7 400 1,350 900 1000 700 3 Sub-total IDA 9,000 12,600 41 16,700 11,100 141100 9,900 36 Other Financiers __ UNDP/PAM 0 1,000 700 1,000 700 3 UNDP/FAO - 0 2,800 1,500 2,800 1,500 5 FED 3,000 4,200 14 10,400 4,200 10,200 3,900 14 FAC 100 100 1 900 300 900 300 1 CFD 4_400 62300 20 17,400 6,600 17,100 6,500 23 Government 5, 600 7,900 25 15, 500 6,400 12,000 4, 900 18 Sub-total Others 13,100 18,500 48,000 19,700 44, 000 1 7,800 64 TOTAL Project 22,100 31,100 100 64,700 30,800 58,100 27,700 100 Comments: Due to heavy devaluation of the FMG and due to the obligation of the Government to pay 35 percent of cost of works as counterpart funding, the Government total contribution to the project increased from 25 percent (appraisal) to 34 percent (actual).  Table 6: Project Results A. Direct Benefits Indicators Appraisal Estimate Estimate at Closing Date Estimated at FulL Development Hectares rehabilitated 17,000 12,989 at closing date 17,202 4,213 within 3 months 17,202 total Number of schemes 16 21 completed 30 9 within 3 months Number of WUA's 16 84 84 Number of Agricultural 16 19 completed 30 Development Plans 10 in process Number of beneficiaries 34,000 29,420 29,420 Incremental yield of rice after: year 1 + 50 kg/ha + 286 kg/ha year 2 + 100 kg/ha + 578 kg/ha Year 3 + 150 kg/ha + 605 kg/ha year 4 + 350 kg/ha + 687 kg/ha year 9 + 940 kg/ha + 940 kg/ha + 940 kg/ha Percent of irrigated land with marginal 20% of surface 30% of surface contra-season crops . Annual maintenance cost 15% of revenues 2% of revenues  B. Economic Impact Appraisal Estimate | Actual | Coments A (at Final Development) Economic Rate of Return * of whoLe area, 17,200 ha 23% 18% both rates at constant FMG, vaLue 1986 * of "normal" PPI, 14,000 ha n.a. 20% based on 6 schemes * of "risky" PPI, 3,200 ha n.a. negative 3 schemes with cyclones/siltation or Lack of ____ ___ ___ ___ ___ ___ ___ water Sensitivity to rice revenues for Soavina scheme (1,200 ha) * base IERR 21% 33.8% * rice revenues - 10% n.a. 32.7%; - 3% compared to base * rice revenues - 20% n.a. 31.5%; - 7. compared to base * rice revenues - 30% n.a. 30.1%; - 11% compared to f base Assumptions * analysis done over 30 years at 1986 1986 constant FMG value * paddy price, CIF Taomasina $274/ton $260/ton used for extrapolation over 30 years constant price for years 8 to 30 * shadow exchange rate + 21% average 12% between 1986 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ & 1993 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . constant incremental paddy yieLd 940 kg/ha > year 9 940 kg/ha > year 9 * evolution yields "without" project - 7V//yr, 6 years no deterioration C. Financial Impact Appraisal Estimate | ActuaL Estimate | Comments Net Incremental Financial 67,000 FMG/ha 310,000 FMG/ha In eqLuivalent FMG, value 1986 (example Revenues due to the Project: of Soavina rehabiLitated schemes)  24 D. Studies Name of Study Purpose as Defined at Status Inpact of Studies _ Appraisal t Pre-selection Studies Sort areas to be further done for 140 sites pre-selection of schemes studied l Orientation Appraise potential and done for 64 detected helped in selecting costs estimates schemes proper schemes, except for 3 schemes Preliminary Define basic technical done for 49 detected helped discuss with Engineering Studies choices schemes farmers and sensitize to the need of WUAs Detail Engineering Define detail technical done for 46 detected helped build confidence Studies and economic analysis schemes with farmers for final scheme selection Development After scheme selection, done for 19 schemes in helped discussing with Agricultural Plan define agricultural process for 10 schemes WUAs their production develop 5 year plan potentials__ Annual Agricultural Define each year done each year for 13 helped keep contracts Production Plan agricultural production schemes; done with the with farmers activities ODR method for 5 schemes; can be done only when rehabilitation is finished Operating Manual for Recommendations for completed for all 30 help WUAs to comply with Water Users water management and schemes their responsibilities _ network management Development Actions Methodology to define terminated in Dec. 1990 done by the Coordination in Schemes development actions and Unit their implementation Management Transfer Analyze issues terminated in Sept. base for preparing legal to Farmers concerning the transfer 1992 documients of scheme management to WUJAs Evaluation of WUAs Perception by members of terminated in Aug. 1991 allowed better focus of their WUA and proposals training programs of WUAs for the future members and field agents Creation of WUAs Evaluation of terminated in 1991 same as above methodology concerning consultation of WUAs in ODR l,, Animation Advisors Analysis of jobs, terminated in Sept. useful for better motivation, problems 1992 defining respective roles with Production and Rural G6nie agents WUAs' Capacity Capacity in managing terminated in Feb. 1993 allow to realize that schemes and paying farmers are willing to maintenance fees, in 2 pay maintenance fees if regions works are correctly H ~~~~~~~~~~~~~~~~~~~~finished  25 Table 7: Compliance with Credit Covenants Covenant |Subject |Status DEVELOPMENT CREDIT AGREEMENT Section 2.02 b Opening Special Account Completed. Section 3.01 a Government provides promptly funds as Practically never provided needed when needed. Section 3.01 e Government provides necessary proceeds required by the project, on the basis of Annual Work Programs - to Minagri Completed. - to Samangoky Component cancelled on November 14, 1988. Section 3.03 b Government consults IDA for the Completed. position of Project Coordination for IRP Section 4.01 b i Financial audit Completed. Section 4.02 Government submits annual maintenance Component cancelled on contracts concluded with Samangoky, by November 14, 1988. Oct. 31 Schedule 5. 1 Selection of schemes for rehabilitation Done. Schedule 5.2 Pre-investment studies Done. Schedule 5.3 Detailed studies and rehabilitation works Done. Schedule 5.4 Annual work program Done. Schedule 5.5 Mid-term special report (1988) Done. PROJECT AGREEMENT (SAMANGOKY COMPONENT) Section 2.04 Samangoky to duly perform all its Component cancelled on obligations November 14, 1988. Section 2.05 b Samangoky informs IDA of any Component cancelled on interference with the progress of project November 14, 1988. Section 2.06 Samangoky furnishes an annual work Component cancelled on program November 14, 1988. Article III Management and operation of Component cancelled on Samangoky November 14, 1988. Article IV Financial audit of accounts Done through 1988.  Table 8: Use of Bank Resources A. Staff Inputs Staff weeks 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 TOTAL Preparation 28.3 28.3 Appraisal 30.7 30.7 Negotiation 2.2 2.2 Supervision 2.6 19.6 12.9 12.5 10.8 14.7 9.5 15.7 12.9 5.6 116.8 Project Completion Report 4.0 4.0 TOTAL 28.3 35.5 19.6 12.9 12.5 10.8 14.7 9.5 15.7 12.3 9.6 182.0  B. Mission Data M onth/Year No. of Days in Specialization Performance Type of probtems persons field represented 8/ Appraisal 03-04/84 5 25 M,T n.a. 9/ Supervision 1 10/85 3 10 F,M,T 1 Supervision II 04/86 4 7 F,T 1 Supervision III 11/86 3 10 F ,T 2 Supervision IV 06/87 1 6 T 2 Supervision V 11/87 1 9 T 1 Supervision VI 07/88 1 8 T 2 Supervision VII 11/88 1 11 T 2 Supervision VIll 09/89 1 10 T 3 procurement Supervision IX 06/90 1 13 T 3 procurement, studies Supervision X 09/90 1 6 T 3 procurement, studies Supervision Xi 07/91 1 14 T 2 work quality Supervision XII 02/92 2 15 T 2 work quality Supervision XIII 10/92 2 11 T 2 work quality Supervision XIV 04/93 1 8 T 2 water fee collection Supervision XV 11/93 1 10 T 2 water fee collection 8/ F=Financial; M=Managerial/Economist; T=Technical. 9/ n.a. = not available   IBRD 18378 4 /ZI2 43.5 M D GAC R3' JUMADAGASCAR '3'50 JUNE 19R4 FIRST IRRIGATION REHABILITATION PROJECT PREMIER PROJEr DE REHA 811rAT/ON DES PER/MErRES IRRIGUES 2C20' ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~21'20' MANGOKY DEVELOPMENT AREA PiER/METRE DU MAVGOKY I() a < { v \\ 5 -- Boundary oF AMVR LimlPe de IAMVR I Eli ~~~~~~~~~~~~~~Irrigated Area Zone Ibrig ahle -Main Canal Conal Principol -' w IDike 4,S V X < 1\ Digue River I ~~~~~~~~~~~~~~~~~~~Fleuive et rfviere Main Road I ~~Route Principole -21-30 2-30'- MILES 0 2 4 6 1 I Ts,an,hy KILOMETERS / v X / "oRrfi > O < IAmbh I AN ~ ~ ~ ~ ~ N0A~MNNS /31 [ \ No~~~~ ~ ~ ~ ~ ~ ~ ~~matoo / \ 2 P40 2145-JBe na,r ZAHEE TANZAN/AANDRANIOMANI < W ~ A l o ic aNoI roMonm.be < Tananda,,.^\ |. Z^1RR g ?\JA"NZANIA . < ANG0LAI _ .-- ~MA WI ii / } xO4 > 5 tY I This map has been prepared by/-.. r .IMBABWE S J The World Bank's staff exclusivefy AMIBe IA Ito'Z E \ u AlSO for the convenience of the !MIA > readers and is exclusively for the gegulation BO.SWANA internal use of The World Bank stwctwe SOTOWANA r \ .Jand the Internatonal Finance Berehoi \Desanding / / 4z 3 fCorporation rhe denominations bos,. I \J ! l / used and the boundaries shown }_, fi/AztaNo \~ on this map do not imply, on the INDIAN. 0CfA4 part of The World Bank and the 2P55 International Finance Corporation. n t (Lesr9o 0 * NetbIi~I # 4GN t any judgment on the legal status SOUTHt AFttCA C o 4 w of any territory or any tw- ntattilbd 4ac endorsement or acceptance of d/i/t~ li ' tetw such boundaries A3__0 IToThi5 ToTui., 45   MADAGASCAR FIRST IRRIGATION REHABILITATION PROJECT PREMIER PROJET DE REHABILITATION DES PERIMETRES IRRJGUES / LOCATION OF MEDIUM IRRIGATION SCHEMES -C- LOCAUISATION DES PETITS PERIMETRES ANTSERANANA , WA~~~~~~~~~T \~~ ma^ ...... 1m- -. . L / S @ 1 ~~~~~~~OMASIAJ - AN NA 0 4, J ... r , fs.unns MAI-IAJ GA * ~ C)~~~~ ) ,s . i ~~-A. NW/ "' ' ) i */^ole ) r .~4ev. ,r ar I ., MSrsedofs' /_ arte .. INDIN OCEA j.- . A- a... An anaiv Ci ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~~~~l y ~ ~~~~~~ ~~~~ FianalF -~ ~ ~ IDlQL am. River ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~RsBts |NaoesrmlSo (go ,.Adt7ianade=j IL~~~~~~~CV lImamS sud/az5 \ U \~~~Faeeraaao .{ ,. ' / h .~~~~~~~I t'h..t,I=Ieiol Mat- -Ismoa - AhA ,,AL ,~ ~ ~ ~ . .L. pi. -. - '. S ,- ) } , A& . , .ee.g,.hTWAAv. / \ I :- a . c ,/ , OLE,FOTHO _~~~~~~~~~~uo cpk .~~ ~ ~ ,*Sa ____ LD___________BOU________ i~~~~~~~~~~~~~~~~~~~~c _ APrWI FIAN.A. ZABI -A- TOLIARA ~ ~ ~ ~ ~ ~ ~ ~~~~~~AABA% ZMAW N \M.B5 ZIBA tAAGASCAR BOTSWANA ,/. c ~~~~~~~~~~~~~ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~Lmits d'Etat0   iZ ZAMBIt M COMORoS MADAGASCAR 101286R R e I 2 ' \ZIMBABWE, #Eii1 PREMIER PROJET DE REHABILTATION DES PERIMETRES IRRIGUES PROJECT IRRIGATION SCHEMES -< (SWAZILAND ) t/@ @ *,PERIMETRES IRRIGUES DU PROJET -t souTh .i&,,AFRaCP,rABA PA * Paved Road ..............Dike LSSOThO * ~~~rowRoute Goudronn6e Digue Behara~~~~~~~~~" - ~~~~WW 91T-