e14 - 0£ Ii EXCHANGE RATES YA Malagasy francs (FMG) Malanasv francs (FMGI 22r-SDB wer US$ 1982 405.6 349.7 1983 515.3 430.4 1984 645.0 576.6 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 2,014.0 1,603.4 1990 2,012.0 1,494.0 WEIGHTS AND MEASURES 1 1hectare (ha) = 2.47 acres 1 kilometer (km) = 0.62 miles 1 square kilometer (kM2) = 0.39 square mile 1 kilogram (kg) = 2.20 pounds 1 liter (1) = 0.26 US gallon 0.22 Imperial gallon 1 ton (t) = 2,204 pounds ACRONYMS EEC European Economic Community EDF European Development Fund ERR Economic Rate of Return FAFIFAMA Livestock Development Agency for Western Madagascar FAO/CP FAO/World Bank Cooperative Programme FOFIFA National Institute for Agricultural Research and Rural Development GTZ German Development Aid Agency IDA International Development Association IFAD International Fund for Agricultural Development LCV Central Veterinary Laboratory MPAEF Ministry of Animal Production and Water and Forest Resources OED Operations Evaluation Department PAR Performance Audit Report PCE. Project Completion Report PCV Central Veterinary Pharmacy RMEA Regional Mission in Eastern Africa of the World Bank SAR Staff Appraisal Report SEDES French consulting firm SPA Animal Production Service of the MPAEF SPEL Provincial Livestock Services Agency VLRDP-1 and II Village Livestock and Rural Development Project I and II FISCAL YEAR January 1 - December 31 FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-Geleral Operations Evatuation March 22, 1993 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Performance Audit Report on Madagascar Second Village Livestock and Rural Develooment ty,.!*ct (Credit 1211 -MAG) Attached is a copy of the report entitled "Perfarmance Audit Report on Madagascar: Second Village Livestock and Rural Development Project (Credit 1211-MAG) prepared by the Operations Evaluation Department. The project delivered useful animal health (mainly vaccination) services during the project period, bu'. these were not sustained, the animal production component was ineffective and the impact on producer incomes was minor. The project is rated as unsatisfactory. The earlier Project Completion Report had rated the project as satisfactory because of the positive returns achieved on investments in the animal health component. However, it acknowledged that the support service created by the project was unsustainable for financial and management reasons. Attachment 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 authoriz* t. FOR OFFICIAL USE ONLY PERFORMANCE AUDIT REPORT MADAGASCAR SECOND VILLAGE LIVESTOCK AND RURAL DEVELOPMENT PROJECT (CREDIT 1211 -MAG) Table of Contents Page No. Preface ................................................... i Basic Data Sheet ............................................... iii Evaluation Summary ............................................ v I. BACKGROUND AND DESIGN ................................... 1 Background ............................................... 1 Prior Bank Group Involvement in Livestock .......................... 1 Project Objectives and Design .................................. 2 II. IMPLEMENTATION EXPERIENCE ................................. 4 Project Costs and Credit Disbursements ............................ 4 Animal Health .............................................. 5 Pilot Animal Production Program ................................. 6 Road Construction .......................................... 7 WaterSupply .............................................. 8 FAFIFAMA Support .......................................... 8 Technical Assistance to the Subsector ............................ 1 Imports of Veterinary Products and Equipment ....................... 9 Applied Research ........................................... 10 Ill. OUTCOME ................................................ 10 Mahajanga Component ................................... 10 Animal Health ............................................ 10 Pilot Animal Production Program ............................ 13 Roads and Water Supply .................................... 13 FAFIFAMA Support ........................................ 13 Subsectoral Component .................................... ..13 Technical Assistance and Studies .............................. 13 Imports of Veterinary Products and Equipment ................ 14 Export and income Objectives .................................. 14 Environmental Impact ....................... .. ......... 15 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 Dank authorization. Table of Contents (continued) Paae No. IV. FINDINGS AND LESSONS ............................ 15 Sustainability ................................... 15 Design Issues ................................. 17 Conclusions ............................................... 18 1 Physical Achievements ....................................... 20 2 ProjectCosts .............................................. 21 3 Project Financing ............................ .. 21 4 Numbers of Cattle ('000) Vaccinated Against Blackleg and Anthrax and Percent Coverage per Province (1968-1991) ......... 22 5 Numbers of Treatments ('000) for Liver Fluke and Roundworms in Cattle per Province (1974-1990) .............................. 23 6 Reestimated Economic Rate of Return Calculation ..................... 24 Aonelt A Note on Animal Health Coverage and Cattle Numbers in the Provinces of Madagascar ........................................... 25 Attachment1 Comments from the Ministry of Stockraising and Fish Resources ............ 29 Map: IBRD 15637 .I- PERFORMANCE AUDIT REPORT MADAGASCAR SECOND VILLAGE LIVESTOCK AND RURAL DEVELOPMENT PROJECT (CREDIT 1211-MAG) PREFACE This is the Performance Audit Report (PAR) for the Second Village Livestock and Rural Development Project (VLRDP-Il) for which Credit 121 1-MAG, in the amount of US$15.0 million equivalent, was approved on February 23, 1982 and became effective in March 1983. The Credit was closed on December 31, 1988, one year behind schedule. An amount of US$7.5 million equivalent (51 percent of the original Credit) was cancelled, and the last disbursement was on August 29, 1989. An IFAD loan also supported the project to the value of US$8.0 million equivalent, of which US$3.6 million equivalent (45 percent of the original loan) was cancelled at project closure. The project was the third livestock operation supported by the Bank Group in Madagascar, and has been followed by the Livestock Sector Project.J The PAR is based on the Project Completion Report (PCR)2/ of the Africa Regional Office, the Staff Appraisal Report (SAR) and President's Report, the legal documents, the project files (including a mid-term review of the project) and discussions with Bank staff and Government officials. An OED mission visited Madagascar in June 1992 and the kind cooperation and valuable assistance of Government officials during the mission are gratefully acknowledge:. The PCR provided a summary of the project design and implementation experience and attempted to measure impact of the animal health vaccination program by the implementing parastatal (FAFIFAMA) which formed the major part of the project. The PAR elaborates on particular aspects of the project, puts the project's vaccination program in the context of the national program, revisits the project's impact, and raises questions about project design. Following standard OED procedures, copies o the draft PAR were sent to the Government. Comments from the Ministry of Stockraising and Fish Resources are Included as Attachment 1. 11 Staff Appraisal Report: Democratic Republic of Madagascar, Uvestock Sector Project, dated April 18, 1991 (Report No. 9424-MAG). 21 Project Completion Report: Madagascar: Second Village Uvestock and Rural Development Project (Cr. 1211-MAG) dated December 5, 1991 (Report No. 10128). * III - PERFORMANCE AUDIT REPORT MADAGASCAR SECOND VILLAGE LIVESTOCK AND RURAL DEVELOPMENT PROJECT (CREDIT 1211-MAG) BASIC DATA SHEET A. KEY PROJECT DATA Appraisal Actual or Aetual as % of EstimL Reviged Estimated Actual Aggraisal Estimate Total Project Costs (US$ m equiv.) 25.5 15.2 12.3 48 IDA Credit 15.0 8.1 7.4 49 IFAD Loan 8.0 4.3 4.4 55 Economic Rate of Return 36 la about 25 69 Institutional Performance Mixed 1A For the major component of animal health. ESTIMATED AND ACTUAL DISBURSEMENTS (US$ '000) FY83 FY84 FN 15 FY86 FY87 FY88 FY89 FY90 Appraisal Estimate 700 3,400 7,300 11,000 13.500 15,000 la Actual - 1,338 2,377 3,900 5,457 6,511 7,092 7,374 Actual as % of Estimate - 39 33 35 40 43 47 49 Date of Final Credit Disbursement August 29, 1989 Ia After the mid-term review, SDR 5.85 million was cancelled of the IDA credit on June 1, 1988 (and SDR 3.15 million of the IFAD loan). C. STAFF INPJTS (Staffweeks) Staff Inputs FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 TOTAL Identification 16.3 16.3 Preparation 3.7 3.7 Appraisal 69.8 69.8 Negotiations 3.7 3.7 Supervision 2.8 3.8 13.8 8.8 14.5 11.0 23.4 3.9 82.0 Project Adninistration 0.7 * * * 0.7 TOTAL 96.3 4.5 13.8 8.8 14.5 11.0 23.4 3.9 176.2 - iv.- D. PROJECT DATES Item Original Actual Schedule Date Preparation - October 1979 Appraisal Mission June and July 1980 Post-Appraisal Mission - November 1980 Staff Appraisal Report * January 1982 Negotiations February 1982 IDA Board Approval February 1982 IFAD Board Approval - March 1982 Credit Signature - April 1982 Credit Effectiveness August 1982 March 1983 Project Completion June 1987 December 1988 Credit Closing December 31, 1987 December 31, 1988 MISAION DATA Stages In MonthlYear Number of Days in Field Specializations Performance Trend g/ Types of Project Cycle persons represented a/ Ranking k1 problems gi Appraisat 6/1980 7 * Post-pprafsat 11/1980 2 * Supervision 1 3/1983 1 10 F 2 2 F 2 10/1983 5 10 F,L,O,T 2 2 F,N 3 5/1984 2 20 F,L 2 2 F,M 4 11/1984 1 15 L 2 3 N,F,T 5 6/1985 1 15 L 2 1 H,T,F 6 2/1986 1 17 L 2 7 6/1986 1 20 L 2 8 12/1986 1 15 L 2 9 1/ 10/1987 /f 4 16 E,L 3 Al F: Financial Analyst; T: Training Specialist; L: Livestock Specialist; 0: Other. 1l 1: Little or no problems; 2: Minor problems; 3: Major problems. .Q 1: Improving; 2: Stationary; 3: Worsening. St F: Financial; M: Managerial; T: Technical. Al This was a HO mission; all previous supervision missions were undertaken from RMEA. f/ Major review mission; a 590 prepared in April 1987 also rated the project 3. OTHER PROJECT DATA Borrower (Executing Agency) Government of Madagascar (Livestock Development Agency for Western Madagascar (FAFIFAMA)) Follow-on Project Madagascar Livestock Sector Project .v- PERFORMANCE AUDIT REPORT MADAGASCAR SECOND VILLAGE LIVESTOCK AND RURAL DEVELOPMENT P6OJECT (CREDIT 1211-MAG) EVALUATION SUMMARY Introduction 4. The most prominent component was animal health services and was to cover the 1. The Second Village Livestock and Rural whole of Mahajanga province. Cattle raising Development Project (VLRDP-ll) was the third in this province was in a low-h-put produc- in a series of four projects in which the Bank tion system of extensive grazing on native has assisted the livestock subsector in pastures in a monomodal rainfall regime with Madagascar. In the early 1980s livestock a distinct six-month dry season. Small herds production accounted for about 10 percent of less than 50 head predominated, with of the value of agricultural output, which most families having some subsistence contributed 40 percent of GDP. cultivation (<1.0 ha); draft power was also increasing in the expanding cropping zone. 2. The first project was largely directed at parastatal ranching and failed. The next 5. The project was supported by an IDA project, VLRDP-1, changed the emphasis to Credit and IFAD Loan. The Mahajanga providing animal health services to existing animal health and rural development compon- cattle herders and to rural development ent was to assume 72 percent of project investments and social services in western costs, with a subsectoral support component Madagascar. A parastatal body, FAFIFAMA, consuming the remainder. was created to implement the project. In the PAR on VLRDP-I, it was rated as having a Implementation Exoerience positive, but limited, impact on the national economy; the social impact was significant, 6. Project costs were drastically reduced but concern was expressed for the sustain- in US dollar terms due to a four-fold devalua- ability of the institutional arrangements. tion of the Malagasy franc (FMG) between 1982 and 1988. Disbursements und3r the Obiectives IDA Credit and IFAD Loan were correspond- ingly reduced to 49 and 55 percent re.-pec- 3. VLRDP-1l in many aspects was an tively of the intended amounts. Major reduc- extension and expansion of the previous proj- tions in the road building, pilot animal ect. It had the objectives of increasing meat production and stock watering programs in production (thereby raising incomes of live- Mahajanga also contributed to the cancella- stock owners and increasing meat exports), tion of external funds despite high local improving livestock institutions and inflation rates. supporting subsectoral changes. - vi - 7. The animal health services were carried 11. An extensive training program was out relatively well by FAFFAMA with implemented for FAFIFAMA staff. A increased staffing, training, transport and Monitoring and Evaluation Unit was slow to veterinary infrastructure supplied under the become effective and had a limited impact. project. This resulted in significantly increased vaccination coverage during most 12. In the subsectoral component, a series of the years of the project, but reverted to of studies was uneertaken, but late in the levels at or below those for the preproject implementation period, so that a one-year situation after project closure. Treatments extension ot the Credit was justified on the for internal parasites were also close to critical need for completion of these studins. appraisal expectations but declined substan- Animal health drugs were also imported tially towards the end of and subsequent to under this component by a Government of the project. The sma!!er programs of Madagascar (GOM) parastatal with a vaccination of poultry and pigs exceeded monopoly on imports of veterinary drugs, but targets. The full responsibility for the veter- the management of procurement and distri- inary services reverted to the regular provin- bution was very inefficient. A Bank staffer cial animal production service, SPEL, after was seconded to GOM in the latter years of the project. the project to assist in implementation pro- cedures and in guidance in preparation of a 8. A pilot animal production sub- follow-on project. A research subcomponent component which was to follow a technical was not implemented as it proved to be and socioeconomic survey of village areas inappropriate for a project of this type. was not implemented, and limited commer- cial scale fattening trials gave unfavorable Rlts results. Some of the funds for the produc- tion subcomponent were used for a small 13. There were no herd monitoring dat to dairy plant and milk collection scheme. validate cianges in the herd numbers or structure d-je to the animal health program. 9. The cancellation of most of the live- However, two study/surveys (1978 and stock production plans affected the road 1984-7) in Mahajanga by a French consulting infrastructure subcomponent which was to firm, SEDES, did provide a basis for equating concentrate on the village areas selected in vaccination coverage to herd changes. the socioeconomic survey. The heavy equip- ment (procured under VLRDP-I) was used 14. The cattle saved by the animal health largely to improve access to FAFIFAMA facil- program added to the provincial herd base, ities and for renting to the provincial Public but the program in Mahajanga component of Works for their road program. Project road the project did not have an influence on improvement targets, however, were met. production systems nor on age or weight of turnoff. Using the actual vaccination 10. Only a few investments in the stock coverage, the economic return in reduced water supply program in Mahajanga were herd losses from the investment in the implemented and subsequently not main- Mahajanga component was estimated at tained. The village water supply, on the about 25 percent. The anima production other hand, was successfully implemented, program did not make any contribution to despite a lack of villager participation ir. the this return. The reduced road subcomponent early years of the project. would have generated some (but unquanti - vil - fied) economic response additional to that structure has not been transferred yet and re1lected in access by the vaccination teams. the provision of increased budgets necessary There were also social and economic benefits to maintain the expanded veterinary post from the village water supply programs accommodation built undar the project is through women's labor saved in water doubtful. The transport provided has largely collection. depreciated bayond its useful life and has not been replaced. 15. The primary objectives of the Mahajanga component were to increasq meat Findings and Lessons exports and raise smaltholder incomes through investment and improved services. 19. In short, the incremental investment in The project had no significant impact on animal health services in Mahajanga largely exports, but did have some impact on the realized the intended coverage for the project income of cattle owners by increasing thoir period, and gave a sat;sfactory economic base herds by about 3.3 percent over the return to the GOM. Vaccination against the project period; this is significant but far less major disease challenges in Mahajanga than the estimate at appraisal. Institutional represented a low cost-high return annual performance during the project was mixed production expenditure. Under these but the animal health service performed rela- circumstances, a high payoff is virtually tively well under FAFIFAMA. inevitable on this investment by a herd owner, or by a public service providing 16. The studies in the subsectoral vaccination in accessible areas. The component were useful in contributing to the objective of Bank involvement in this type of follow-on Livestock Sector Project which project, however, must be to establish an became effective in August 1992. effective, sustainable system for the delivery of services to improve herd productivity; this, Sustainabilitythe project did not do. The service institution was dissolved after project 17. The marginal impact of the project on completion. Furthermore, the livestock the herd base in Mahajanga is sustainable; production subcomponent of the project was the herd will be affected by many influences aborted during Implementation. As the in the future, but the additional animals Mahajanga animal health subcomponent of saved have contributed to a new base the project was its major focus, the audit is number of animals. The village water supply unable to rate the project as satisfactory, is likely to be maintained as consumers now despite the apparent acceptable returns on appear to value the benefits and the water the GOM expenditures In Mahajanga and the extraction is manual and, therefore, relatively usefulness of many of the studies. easily maintained. Future road maintenance is doubtful and infrastructure built on 20. Although the appraisal mission's FAFIFAMA farms and facilities is already in acceptance of FAFIFAMA as the imple- disrepair. menting institution for the Mahajanga component was understandable in view of 18. At the end of the project FAFIFAMA the VLRDP-11 being virtually an extension of was dissolved with contract vaccinators the former project, the Bank did not appear being dismissed; few benefits appear to have to seriously address the replicability/ been realized by the regular provincial sustainability issue of FAFIFAMA services. services. Some of the animal health infra- The rural infrastructure components had not - viii - been carried out well under VLRDP-I, so 22. Obvious lessons from VLROP-11 are: there was no overwhelming case to continue with FAFIFAMA for rural development. The (a) Projects providing animal health two components of VLRDP-II which were services are likely to provide attractive implemented relatively well were animal returns on investment, but such projects are health and village water supply; it could be not a suitable development vehicle for Bank argued that some assistance to the provincial involvement unless the support is geared to livestock service in transport and training the establishment of efficient and sustainable would have provided a reasonable and more arrangements to provide these services. sustainable herd coverage, and the village water supply could very likely have been (b) In proposing production system undertaken as well through support to the modifications involving improved technology provincial Public Works Department. and practices, a knowledge of the characteristics of the existing production 21. The Bank was aware of the need for systems and the attitudes of the target group more appropriate subsectoral policies and a should be ascertained before including such rationalization of public sector livestock a component in a project with a limited services. It included the studies and implementation period; that is, the technical assistance as a subsectoral investigation should be made before appraisal component in recognition of this concern. In rather than during implementation, regardless hindsight, it may have been more appropriate of whether the component is of a pilot or for the Bank to finance subsectoral studies wider nature. as a basis for a dialogue with GOM and the development of a project along the lines of (c) In formulating and appraising the current Livestock Sector Project, rather projects, consideration of subsectoral than proceed with the Mahajanga conditions and requirements should take component. The Livestock Sector Project precedence over pressures for continuation includes policy reforms, strengthening the of traditional interventions, especially in the existing subsectoral institutions and use of scarce IDA resources. promotion of private animal health services, epidemiological surveys in extensive cattle areas, and dairy development in the higher potential plateau environment. PERFORMANCE AUDIT REPORT MADAGASCAR SECOND VILLAGE LIVESTOCK AND RURAL DEVELOPMENT PROJECT (CREDIT 1211-MAG) 1. BACKGROUND AND DESIGN lackaround 1.1 In 1980, 85 percent of the population of Madagascar was rural, and half of the population lived below the poverty line. Agriculture contributed about 40 percent of GDP; rice was the predominant foodcrop (accounting for 30 percent of the value of agricultural output) while coffee and cloves exceeded 50 percent of total merchandise exports. Livestock production was important throughout Madagascar and accounted for about 10 percent of agricultural output. 1.2 Smallholder farming on holdings of one to three ha predominated. Cropped land was heritable while pasture land was usually held communally. Government was heavily involved in the marketing and processing of agricultural commodities and had become involved in large scale production, especially since 1975 on nationalized farms. These interventionist policies were part of the socialist political philosophy introduced by the Government of Malagasy (GOM) in 1972. Performance of the agricultural sector had been poor in the 1970s. The stagnation of agricultural output was a major contributor to the economic crisis occurring after 1980 which led to severe budgetary problems and acute foreign exchange shortage. 1.3 Cattle production in Madagascar is overwhelmingly traditional. In the principal cattle regions (northwest, west and south of the country in the provinces of Mahajanga and Toliary - see map) the production system is very ex qnsive with individual herds being up to several hundred but commonly 15 to 70 head. The common age of turnoff is very high at 6 to 9 years, and sometimes is over 10 years of age. There is some degree of intensification in the central plateau area of the country in stall feeding and fattening and there is a substantial and growing use of cattle for draft purposes (cultivation and transport). The herd is free of many of the major cattle diseases of Africa, but is afflicted seriously with the bacterial diseases of anthrax and blackleg and with internal perasites. Sheep and goat production is largely in the south under traditional management. Pigs are also largely in the traditional sector with farmers owning one to three pigs which typically scavenge for food. Although some commercial poultry production is developing near large urban centers, most poultry is still in a scavenging production system with smallholders. Prior Bank Groug Involvement in Livestock 1.4 The first livestock project in Madagascar, the Beef Cattle Development Project (Ln. 585-MAG) called for the introduction of new technology on parastatal ranches and the importation of Brahman cattle to improve the genetic base of the local herd. Like most other parastatal livestock production schemes in Africa, this project was a failure, and the -2- Bank and GOM subsequently targeted subsectoral development through the extensive smallholder livestock production system. 1.5 The First Village Livestock and Rura, Development Project (VLRDP-I - Cr. 506- MAG) was implemented from 1975 to end-1 982. It was to develop an intensive veterinary and livestock production program, as well as rural infrastructure (roads, water supply, health and education centers), in one of the least developed and populated regions of the country - part of the Mahajanga province in the northwest of Madagascar, two livestock-oriented pilot schemes and one including crop extension, credit, educational and health facilities in addition to animal health were to be carried out in the middle-west (see map). The project beneficiaries were expected to be about 2,000 villages and several thousand small cattle owners. The project objective was to increase beef production both for internal consumption and export. The Livestock Development Agency for West Madagascar (FAFIFAMA) was established in 1974 to take major responsibility for implementation of the project. Due to the complexity of the project and the lack of natioial expertise in some areas, expatriate consultants were to be recruited. Other government agencies involved were the ministries and/or provincial services of health, education, public works and water supply for implementing their respective parts of the project under FAFIFAMA's coordination. 1.6 Although no monitoring of the effects of the animal health program was carried out, it was estimated that the animal health program had a substantial impact (through veterinary infrastructure, training and services). A pasture improvement program was hampered by technical and management problems complicated by the communal land tenure and did not have the desired impact. No new husbandry practices were introduced. The rural infrastructure investments had mixed results: the school program was successful; the Agricultural Training Center and Human Health Centers were relatively unsuccessful; the road component eventually was implemented under force account by FAFIFAMA (rather than the intended provincial Public Works) and technical specifications for roads were low, requiring frequent road maintenance; the cattle and human drinking water component fell short of its objectives and end-user responsibility for maintenance was not achieved. Overall, in the PAR the project was assessed as having a positive but limited impact on the national economy.1/ Project Obiectives and Design 1.7 The Second Village Livestock and Rural Development Project (VLRDP-ll) was prepared by FAFIFAMA based on a feasibility study by SEDES, a French consulting firm with a record of assistance to the Malagasy livestock sector. The initial project brief envisaged an expansion of the services under the ongoing project to the whole of the province of Mahajanga. This would include animal health services, a zone of integrated development activities in selected villages, and a zone of intensive cattle production on parastatal farms and surrounding villages. An appraisal mission in June and July 1980 did not accept the complete project package as presented by GOM; the proposals for village 1/ OED Report No. 5403, December 28, 1984. -3- development and the parastatal ranch scheme were substantially reduced, and some assistance was provided to GOM to define an appropriate sectoral strategy. A post- appraisal mission in November 1980 made further reductions in the village intensification scheme and also introduced a national drug import package in response to the rapidly deteriorating economic situation in the second half of 1980. The Bank's Operations Policy Staff noted the poor policies and management of the subsector and recommended that the Bank should concentrate on developing a sound livestock strategy before giving any broad subsectoral support; it also was concerned that any agreement to finance a study of the livestock production and marketing parastatals might give Lhe impression that the Bank accepted and condoned these institutions as desirable for subsectoral development. 1.8 The project eventually was approved with the objectives of increasing meat production to raise incomes of livestock owners and to increase exports, of strengthening livestock institutions and supporting subsectoral policy changes. It had the following components: IN THE MAHAJANGA PROVINCE, IMPLEMENTED BY FAFIFAMA: (a) continuation and expansion of the animal health program in the whole of the province; (b) a pilot animal production program which included (i) intensified infrastructural development and livestock health and extension services for 40 accessible village areas near Mahajanga town; (ii) testing of the economic viability of cattle fattening in both smallholder and large commercial schemes (the former through marketing extension, a market outlet, and credit organized through a state bank, and the latter on a FAFIFAMA farm); (C) all-weather standard road construction (100 km) with concentration in the areas of the villages selected for intensified development, using equipment procured under the former project with a 20 percent contribution by users towards capital and maintenance costs; (d) 35 surface water StL age facilities for livestock (again, concentrated near the pilot scheme villages) and 80 wells in villages for human consumption; and (e) management and material support for FAFIFAMA as the main implementing agency, including training of its staff, studies and consultancies. IN THE LIVESTOCK SUBSECTOR, IMPLEMENTED BY THE NATIONAL ANIMAL PRODUCTION SERVICE (SPA): (f) Technical assistance to improve subsectoral organization and institutions; -4- (g) Import of veterinary products (largely internal parasite treatments) through the Central Veterinary Pharmacy (PCV) - a division of SPA - for a three-year period on a declining finance basis, and some equipment and materials for the Central Veterinary Laboratory (LCV) which produced vaccines for SPA and was operated by the national agricultural research organization (FOFIFA); (h) applied research on cross-breeding trials with Malagasy Zebu and Brown Swiss for dual- and triple-purpose cattle in the plateau environment; and (i) studies on subsectoral issues, documentation of research and scholarships. II. IMPLEMENTATION EXPERIENCE Project Costs and Credit Disbursements 2.1 Bureaucratic procedures caused a delay in effectiveness for one year after Board approval in February 1982. The Credit was eventually extended by one year (closure in December 1988) largely to accommodate the delayed studies on sectoral issues which were considered essential for further subsectoral development and for a new livestock project. Project costs were drastically reduced in US dollar terms from an estimated US$25.83 million equivalent to only US$12.3 million equivalent. Final disbursements on the IDA credit and IFAD loan were correspondingly reduced by 51 percent and 45 percent respectively. The project costs in Malagasy francs (FMG), however, were 52 percent higher than at appraisal. 2.2 The main reasons for the reduction in US dollar costs were (a) the four-fold devaluation of the FMG against the US dollar between 1982 and 1988 which affected the US dollar value of local expenditures;Z/ (b) cancellation of most of the pilot animal production program which also affected the road component; (c) management and conceptual difficulties leading to a substantial reduction in the stock water program, and (d) administrative difficulties which slowed and reduced parts of the subsectoral component. These reductions occurred despite a high level of local inflation which averaged about 23 per.ent annually from 1982 to 1988. The achievements of physical targets under the project is summarized in Table 1, a comparison of appraised and actual costs in FMG in Table 2, and the intended and actual financing in Table 3. The latter table indicates a four percent contribution by GOM to project costs, but this should very likely be reduced as French, Swiss and German bilateral aid contributed to purchase of veterinary medicine. 2_1 This is also a reflection of the less-than-adequate treatment of overvalued exchange rates at Bank appraisal of projects. It is not an issue peculiar to this project, however, and affects a significant percentage of Bank-funded projects; it leads to overcommitment of funds if the exchange rate is adjusted to near real values during the project, especially if this occurs early in the implementation period. -5- Animal Health 2.3 The expansion of the animal health program from the area covered under the VLRDP-1 to the whole of the province of Mahajanga was completed in three years instead of the intended four years. In Table 4, which indicates the number of cattle vaccinated in each year, it can be seen that close to the SAR target of 2.7 million cattle were vaccinated in the peak year of 1985. Although data on cattle numbers in the province are subject to a significant degree of error, this table also shows that the percentage coverage of cattle with blackleg and anthrax vaccination increased from the low 40s to the low 60s under the first project, and peaked at around 80 percent in 1985 under VLRDP-II. Subsequently, the coverage dropped to between 55 and 65 percent, and was even lower at about 52 percent in 1991 after project closure. 2.4 The relatively good performance in implementing the vaccination program was facilitated by the recruitment of many more veterinary assistants and vaccinators than envisaged at appraisal (Table 1), and by substantial accomplishment of the veterinary infrastructure program in veterinary centers and cattle crushes. Expenditure on the infrastructure component greatly exceeded the SAR projections due to upgrading the specifications of vaccinator's offices (of mud construction under VLRDP-I and in the SAR of VLRDP-II), the need for reconstruction of facilities destroyed in the major 1984 cyclone, and the cost of inputs in the relatively remote locations. The improved benefits and transport facilities enjoyed by FAFIFAMA field staff relative to their counterparts in regular SPA service were also conducive to good performance. 2.5 The exceptional vaccination rate in 1985 was apparently partly associated with conditions particularly conducive to blackleg outbreaks, which would also have prompted some repeat vaccinations. The decline in vaccinations after this peak was largely due to delays in deliveries of vaccines, and in later years to a dismissal of one-third of the vaccinators when operating funds were decreased and eventually of all contract vaccinators after project closure. (Both the Credit and the IFAD loan financed the operating costs of the animal health program in a declining proportion for five project years.) 2.6 Treatment for gastro-intestinal parasites peaked at about 150,000 in 1985. Although the intention was to treat calves twice yearly, the extent to which this did occur is unknown, but if half the treated calves received two applications, this would represent about 17 percent of the estimated calf population in 1985, which is a substantial achievement. The coverage subsequently declined, however, and in 1990 only about 16,000 treatments were made. The conditions for liver fluke are not widespread in Mahajanga so that the demand for flukicides (for young and adult cattle) was not as high as in some other provinces; treatment numbers, however, did exceed 50,000 in the peak year of 1986 before progressively declining to very low levels after project completion (Table 5). 2.7 In audition to cattle vaccination and treatment, a pig and poultry health program was also carried out. The achievements far exceeded the targets in vaccination against the main pig disease, Teschen (possibly reaching half the pig population) and treatment of -6- external parasites (lice and mange), which are very obvious inflictions, Increased dramatically (from about 3 percent to 22 percent of the estimated population). Vaccination of poultry for the common fowl pasteurellosis and Newcastle Disease greatly increased as well and exceeded SAR targets. The programs for pig and poultry declined, however, in later years as for the cattle programs. Although all vaccines were provided free-of-charge, the increase in vaccinations of pigs and poultry was also likely associated with the high rate of price increases in these commodities relative to beef. 2.8 After the project, FAFIFAMA was dissc',ed,J/ so that at the time of the audit the animal health services had been transferred back to the provincial animal production service (SPEL). Most of the contract vaccinators had been dismissed; the tertiary (lowest level) veterinary posts had not been transferred to SPEL and, therefore, were not manned or maintained; the livestock assistants and livestock extension staff were distributed over the province to provide the best coverage that resources would allow; material for 35 additional cattle crushes were still in store; veterinary equipment from the FAFIFAMA service had not been officially transferred; and the few remaining functioning vehicles were near the end of their useful life. Pilot Animal Production Program 2.9 This program was poorly implemented and finally abandoned. The socioeconomic and technical survey to select the 40 village areas for intensive support was seriously delayed but was eventually carried out under a consultancy. The delay was caused by the Bank not being willing to accept the firm selected by the GOM until the firm was strengthened with suitable experts. The final study in late 1984, however, was poorly designed and implemented and was unsuitable as a basis for an intensified development program. 2.10 No smallholder fattening trials were carried out,J/ but commercial scale steer fattening trials based largely on delinted, whole cottonseed supplementation were carried out a, FAFIFAMA's Sante Marie field station in 1984. The results of these trials, however, were disappointing. About 3 kg of cottonseed were needed to yield a liveweight advantage of 0.5 kg. This proved highly uneconomic at existing prices. A survey of surrounding villages was made with the intention of leasing animals to villagers for fattening on native and improved pasture. Villagers were generally not interested in the proposal, ostensibly because of the riskiness of such an operation in their security environment where cattle rustling was common. Some improved pasture plots were established at Sante Marie based on legume-grass mixtures, but those suffered from accidental fires, anthracnose infestation on the types of Stvlisanthes spp. grown, and lack I FAFIFAMA was legally abolished on April 25, 1990 and was gazetted as such June 26, 1990. A/ The M&E unit did, however, undertake a feasibility study on smallholder fattening. .- of attention, At the time of the audit the whole station was abandoned (apart from a "caretaker") with buildings and facilities in decay.A/fi 2.11 In view of the failure of the study, the unviability of the fattening enterpris .nd the loss of the EEC export license at FAFIFAMA's abattoir in 1985 (which was to provide an attractive market for fattened cattle), this component was abandoned on the advice of the Bank. Some of the component proceeds were diverted to a study on milk supply in the vicinity of Mahajanga and the subsequent establishment of a small milk pasteurizing plant in 1989. 2.12 The milk plant was leased out to a private operator (ex-FAFIFAMA staff member). At the time of the audit it was collecting from 80 I/day (November) to 400 I/day (March) compared to its capacity of 1,200 I/day. Collection was from six collection points on a 30 km route on six days per week, with the first collection 18 km from the factory. The sales were competing in the city with supplies from farmers closer to the city at similar prices for the less diluted type of farmer-supplied milk. Consumers boil milk. The plant had a mark-up of FMG 350 to FMG 400 per liter, but was having difficulty making a profit. The cattle supplying the milk were local Zebu and the cattle owners were largely a localized immigrant population. 2.13 One of the consequences of the failure of this subcomponent was that the intended monitoring and evaluation (M&E) unit, which was to be heavily involved in the survey and monitoring activities, was late in being established, was not fully developed, and its limited activities declined towards the end of the project. Road Construction 2.14 With the delay in planning for and the eventual cancellation of the intensified animal production program (paras. 2.9-2.11), the road program which was to be concentrated in the selected village areas had to be modified. The plant, which was procured under the previous project, was used to provide or improve access to FAFIFAMA facilities and cut firebreaks on their farms, and was also rented to the provincial Public Works in 1984 and 1985 to rehabilitate major and secondary roads. An attempt was made in 1985 to have villagers commit to and subsequently pay for some of the operating plant costs of access roads, but it was not clear to the audit mission that payments ever eventuated. Despite management difficulties, about 380 km of roads were finally constructed or maintained, compared to a target of 100 km construction in the SAR. f/ It is interesting that the local MADANA Research Center of the national research organization (FOFIFAI had no information on the supplementation trials carried out at Sante Marie, nor did they have data on the improved pasture trials under this or the earlier project. The Director doubted the reliability of the Sante Marie results as he understood that an outbreak of blackleg had marred some of the trials. 6I Some smaller pasture and fattening trials were also carried out at the Milk Farm station (120 ha) owned by FAFIFAMA and located near Mahajanga; this had similar poor results. -8- 2.15 Some of the plant was sold in 1989 and part was leased to a private company which was a subsidiary of the development company (PROCOPS) attached to one of the main political parties in 1990. Components of the plant remaining in 1992 (Unimog, compactor, lorry, 2 Mitsubishi 4-WD and 2 Toyota 4-WD which were beyond repair) were rented out when functional by the remaining unit of FAFIFAMA to pay the wages of its eight staff and 10 guards who persisted despite the dissolution of this parastatal (para. 2.8). Water Supply 2.16 The planned livestock water supply (35 surface dams) were affected by the cancellation of the production component and prolonged argument on whether to build dams or sink boreholes. Eventually one dam was built and five boreholes sunk and equipped with windmills at cattle purchasing points for the FAFIFAMA abattoir. At the time of the PCR no windmills were operational, while at the audit date one had been repaired by the abattoir unit and was operating (the other cattle purchasing points were not functioning). 2.17 The village water supply (wells with manual water extraction) was more successful, with 75 of the planned 80 facilities constructed. No attempt was made until 1986 to have villagers participate in the cost of construction (labor, local materials) or maintenance, but this concept was implemented to some degree after that date. Nearly all human water supply wells were reported to be operational at tne time of the audit. FAFIFAMA Support 2.18 The training component for FAFIFAMA staff was implemented so that virtually all staff underwent a significant training program during project implementation. As already indicated (para. 2.13), the M&E unit was deprived of one of its major functions when the production component collapsed. It served more as an administrative unit and introduced some improvements in accounting. Its contribution did improve, however, in 1985 to 1987 with intensive fostering by supervision missions; a manual system of mapping disease outbreaks was introduced and a program for supervision of the various levels of the animal health service was designed. These activities, however, declined towards the end of the project and the disease mapping has not been continued by SPEL. 2.19 The Commercial Department of FAFIFAMA (managing livestock marketing, meat processing and meat export) was to channel the increased offtake from the production subcomponent to export. The increased offtake, of course, did not occur and the loss - making export activities ceased in 1985 (para. 2.11). Some technical assistance was provided to this Department in processing and cattle procurement. Exports were only recommenced in 1992 after the abattoir was brought up to EEC sanitary standards with European Development Fund (EDF) financing. -9- Technical Assistance to the SUbsector 2.20 This subcomponent in the SAR could have been combined with the Studies and Scholarships component as these items were of a similar nature. The intended study "Meat Supply to the Capital" was changed to "Analysis of Beef and Hides Available for Export"; this was done as the acute problem of meat supply to the capital, Antananarivo, was largely solved by the abolition of the monopoly existing at the time of appraisal. Differences between the Bank and the Ministry of Livestock, Water and Forest Resources (MPAEF) led to a 3-year delay between initiation (1983) and contract award. 2.21 The study on nine parastatals was carried out under a consultancy and this was followed by a USAID-financed study in 1987 on some of the same units (in a package of five) to identify those parastatals from which it would be appropriate for the GOM to disengage. 2.22 On the Bank's recommendation, two study trips were organized to India in 1986 and 1987, largely to look at the cooperative milk collection schemes operating there. Most of the important subsectoral studies were undertaken subsequent to the major review (termed "mid-term" in project files) in late 1987. These included studies of the epidemiology of blackleg and anthrax, legislation concerning industry regulation, livestock statistics, vaccine production, subsector reviews for dairy, pig and poultry, and a review of documentation facilities and services in SPA. 2.23 Institutional changes delayed the implementation of the design of a training program in SPA. Moreover, the scarcity of qualified people limited the extent to which staff could be sent for further training without jeopardizing daily operations, and the level of local qualifications and language proficiency also limited choices. Overseas training that was possible was done through bilateral and multilateral aid (Germany, FAO, USA, Egypt and Japan) rather than the Credit. The significant vehicle and equipment procurement which was assigned largely to provincial and field veterinary staff was completed. 2.24 From June 1986 to September 1989 a Bank staff member from RMEA was seconded to the project to provide technical assistance to the Ministry of Animal Production and Water and Forest Resources (MPAEF) in the livestock subsector. Imoorts of Veterinary Products and.Eauioment 2.25 The Central Veterinary Pharmacy (PCV) was created under MPAEF in the mid- 1970s when foreign exchange shortages were interrupting the importation of veterinary products. In view of the acute shortage of foreign exchange in 1981 the Bank agreed (in a post-appraisal mission - para. 1.7) to include financing of the import of largely antibiotics and internal parasite control remedies over a period of three years. The PCV had a monopoly on drug imports which persisted until 1985. The imports were made using the Credit and IFAD loan as well as bilateral aid funds. However the PCV operation proved grossly inefficient; ignorance of effective demand, administrative problems in determining stock and faulty procurement led to costly mistakes. Consequently, with Bank *10- encouragement, the PCV gradually withdrew from the import business to allow the private sector to take over, and the PCV was finally closed in 1991.Z/8/ 2.26 The planned import of equipment to support the Central Veterinary Laboratory (LCV) in producing vaccines was carried out. At the same time, the LCV was receiving significant assistance from the German aid agency, GTZ. Aoolied Research 2.27 "his component was not implemented with the agreement of the Bank. It was a cattle cross-breeding program, which is very long term by nature, and was inappropriate for a project of this type. III. OUTCOME Mahajanga Component 3.1 Animal Health. If there is a high incidence of disease challenge, an effective vaccination program usually has a high pay-off in reduced morbidity and/or mortality. For the individual animal in which death or severe sickness is prevented, the return is enormous. When a regional herd is concerned, however, the costs of coverage of the herd must be weighed against the production benefits which would be foregone if the vaccine were not applied in specific areas with particular levels of disease incidence. 3.2 In 1986 the French consulting firm, SEDES attempted to analyze the impact of the vaccination program in Mahajanga. They estimated that increasing the vaccination coverage from about 50 percent to 60 percent would decrease the mortality rate due to anthrax and blackleg from 5 percent to 3.5 percent, and an increased coverage from 60 percent to 74 percent would induce a further reduction to 1.5 percent. SEDES also calculated that it would be uneconomic to tty to reach more than about 70 percent of the regional herd because the costs involved would be more than the value of incremental production to be obtained by vaccination. This is undoubtedly caused by the remoteness of many grazing areas and the very low average productivity per animal in the area (about 9 percent extraction rate). ZI The deficiencies in management of the PCV are exemplified by the extent of outdated drugs held in a June 1989 inventory: Centrally held - 83% outdated 1.3% to be outdated 12/89; Provincially held - 87% outdated 4.3% to be outdated 12/89; Most of this stock was fasciolicide (live fluke treatment). I/ FAFIFAMA procured it own supplies so that Mahajanga was not affected by the PCV problems. * 11 - 3.3 Unfortunately, the figures on vaccinations vary according to source. However, using the derived percentages of coverage indicated in Table 4 for Mahajanga, this would suggest that towards the end of VLRDP-1 there was an annual saving of about two percent of the herd compared to 1974, and a further two percent saving in the better years of the second project before declining post-project to levels below that achieved at the end of the earlier project. The SEDES calculation converts to a one percent increase in coverage being equivalent to about 3,000 head of cattle saved. Based on the vaccination record judged most reliable by the mission (Table 4), over the six-year project period the vaccination program can be equated to savings of 9,000, 3,000, 60,00n, 15,000, none and 12,000 head, compared to the pre-project situation which h3d an average 60 percent coverage. 3.4 The original PCR prepared by FAO/CP estimated the saving of cattle by extrapolating the saving calculated in the 1986 SEDES study for the years 1978 and 1986 to the intermediate and subsequent years; this gave a substantially greater saving than that indicated by the vaccination record in Table 4. It then equated each animal saved into 300 kg liveweight which was realized as the economic value of meat with a one-year lag. It assumed that the vaccination service would continue with a coverage of 80 percertt of that obtained with the project, so that the benefits in animals saved would be reduced by half over a period of 8 years (with a reduction of coverage from 70 percent to 56 percent). For pigs, it was assumed that 22 percent of the vaccinations against Teschen and against swine fever would result in a saving of treated pigs at 80 kg liveweight, with 20 percent reduction in coverage after the project. Cumulative annual poultry vaccinations against pasteurellosis, fowl pox and Newcastle Disease were assumed to save 33 percent of this number of poultry at 0.8 kg liveweight, again with a 20 percent reduction in coverage after the project. These assumptions yielded a very high ERR of over 100 percent for the Mahajanga component investment. 3.5 The PCR presented by the AF3 Region made a number of changes which, unfortunately, are not transparent, and finally estimated the ERR at about 50 percent on the project investments. 3.6 As there was no effective monitoring of sample herds to verify impact, these calculations were an attempt to estimate the outcome of the animal health program using the scarce data available. However, a number of factors cast doubt on the validity of some of the assumptions for the cattle vaccination program, which was the main activity: * First, the actual vaccination figures were less than those assumed in the calculations. * Second, it should not be assumed that the saved cattle are marketed as beef one year after being saved, as there was no increased percentage offtake due to the project and it is likely that many of the cattle saved were young animals whereas most cattle are sold between 6 and 9 years of age. - 12- * Third, the actual herd age structure increased (para. 4.3), possibly due at least partly to an increased use of cattle for agricultural and transport draft purposes;9/ this, of course, could be a benefit, even though the contribution of the project would not be to increase the use of draft power but to contribute to the base herd to provide draft animals. In effect, the savings in cattle more appropriately represents a contribution to the herd base in Mahajanga. 3.7 In view of the lack of reliable data, it is not worth attempting to develop complicated models to assess the impact on the herd base. However, a simple approximation would be to assume the incremental herd base would yield the average herd offtake rate, which is about 10 percent (including household slaughter, sales, gifts and thefts in the SEDES survey terminology). This could be arbitrarily and generously increased to 12 percent to allow for the value of cattle for draft purposes, and have the offtake represented by the economic beef price so that one saved animal yields a continuous annual benefit equivalent to 12 percent of 300 kg liveweight, commencing in the year after being saved. 3.8 The outcome from treatment for gastrointestinal parasites in calves and for liver fluke are not included in the benefits as (a) the outcome for the former is very difficult to quantify, especially when the reinfestation conditions after treatment are unknown, (b) liver fluke is not widespread in Mahajanga and, therefore, of limited provincial impact, and (c) most of the animals treated would be also included in the vaccination program on which benefits are assumed. 3.9 In the case of poultry and pigs, the benefits in the PCR were very generous in that the saving was calculated on each vaccination rather than on each animal vaccinated. The benefits, therefore, could be realistically reduced at least by one third. 3.10 On the basis of the above and using PCR assumptions for economic values, the rate of return on the Mahajanga investment would be reduced to about 33 percent (Table 6). However, further analysis using this methodology to examine the impact of recorded vaccinations between 1978 and 1987, which were the years of the actual herd inventory estimates, suggests that the "savings equation" may be too generous (para. 13 of Annex 1). If the benefits due to coverage are reduced by 30 percent (which is suggested by the analysis in Annex 1), the ERR becomes 22 percent. Consequently, the audit concludes that the ERR on the Mahajanga component as reflected in the impact of the animal health program is in the vicinity of 25 percent. This ERR is still relatively high compared to the average for agricultural and livestock, development projects in general, but should be expected for projects of this type, which finance a production input with substantial and immediate effects on the production vehicle - livestock in this case. 9/ The predominant crop, rice, had an average growth rate of 4 percent per annum in the 1980s in Mahajanga province. -13- 3.11 It is worth noting that the successful pig and poultry vaccination programs were directly beneficial to women in the farm households as they are normally responsible for this type of stock. 3.12 Pilot Animal Production Program. As could be expected from the aborted implementation of this subcomporent (para. 2.9), the production improvement aspects of the project were a failure. Even the small milk collection and processina scheme had a limited impact and is running below its processing capacity in competition with raw milk suppliers (para. 2.12). 3.13 Roads and Water Sungly. Not all roads constructed or given some maintenance treatment during the project have continued to be maintained. The improvement in access for the animal health services, however, would have been beneficial (and reflected in the estimated benefits from vaccination) and the improvement in general communication could be expected to have some positive but unquantified economic and social effects on the accessed population. 3.14 The livestock water supply had no significant impact but the human water supply would have social benefits and also some economic benefits including labor saved in walking to distant traditional water sources. This would have benefits particularly for village women normally charged with this task. 3.15 FAFIFAMA Support. The training activities carried out under the project could be expected to have contributed to improvement in the human resource base for subsectoral services. The abolition of FAFIFAMA, however, removed the identifiable effects of this investment, and any influence on SPEL operations were not obvious to the audit mission, (although it could exist in less apparent ways). Subsectoral Component 3.16 Technical Assistance and Studies. Overall, most of the studies carried out appeared to be useful. The study on export potential indicated that national beef production was stagnating due to a high incidence of cattle rustling (which affected the grazing and management of cattle), and continued high mortality and low reproductive performance; export potential was likely to relate to the exchange rate situation. The studies on parastatals emphasized their unprofitability and paved the way for the trend to divestiture by GOM (although only limited divesture had occurred at audit). The visits to India did not result in adoption of the studied model, but may have contributed to the current emphasis given to milk production in the plateau environment. 3. 7 Many of the studies contributed to the formulation of the subsectoral policies which formed the basis for the follow-on Livestock Sector Project.I1/ The seconded staff member provided valuable assistance to the GOM in organizing and interpreting these studies and in preparing the Livestock Sector Project. 1Q Madagascar: Livestock Sector Project, SAR April 18, 1991, Report No. 9424-MAG. - 14- 3.18 Imports of Veterinary Products and Equipment. The provisional funds to support drug imports obviously enhanced their availability in the period of scarce foreign exchange. An earlier removal of the PCV monopoly would have been advantageous in allowing for more efficient private sector intervention, provided importers had access to foreign exchange; this may have been difficult to ensure, however, in the fiscal stabilization being undertaken up to 1985. The equipment supplied to LCV helped in vaccine supply, but this public institution also had annual budgetary and management problems which resulted in serious delays in deliveries and thus inefficiencies in the provincial vaccination programs. Export and Income Objectives 3.19 The stated primary objectives of the project were to increase meat export and increase rural incomes (para. 1.9). 3.20 During the project, cattle exports peaked in 1984 but this only brought the export slaughter to little more than pre-project levels. When the Mahajanga abattoir lost its EEC sanitary clearance in 1985 (as did the other two parastatal export abattoirs in Madagascar), exports ceased completely. After renovation of the plant under an EDF grant (para. 2.19), export processing commenced again in March 1992. The project, therefore, made no significant contribution to meat exports. 3.21 This report has argued that the animal health program did have a positive impact on increasing the base herd in Mahajanga province through decreased mortality, but has not suggested that any improvement occurred in offtake rate due to the project. According to the SEDES studies, the age structure of the herd actually increased during the project period and the reproductive rate did not increase.11 / A very rough estimate of the impact of the animal health program on the income of the target population can be obtained using SEDES study data and the assumptions used earlier to estimate the economic impact. This study indicated the average herd size was 46 cattle, with 66 percent of herds being below 50 head and, therefore, classed as smallherders. If the offtake of 12 percent (which makes a generous allowance for a cattle draft value) is assumed and interpreted as beef offtake, and the cumulative herd size impact of the 11/ Composition of Herd: Mahajanga 1978 1984 mate Female Total mate Female Total 0-1 yr. 9.3 9.6 18.9 8.4 9.2 17.6 1-3 yrs. 15.0 17.7 32.7 12.0 14.0 26.0 4-9 yrs. 15.3 29.5 44.8 15.3 30.0 45.3 10+ yrs. 0.41 3.25 3.06 3.6 7.4 10.3 Sgurce: Derived from SEDES survey data. If fecundity is assumed to be represented by the numbers in the 0-1 yr. age group as a percentage of females over 3 years of age, this indicates a decline from 39% to 32%, while the percentage of both males and females over 10 years has increased. *15- program over the project period is taken as applicable to each herd, then annual impact on each smallherder would be about US$27, assuming additional costs were negligible.j2/ If the impact is allocated to 60 percent of smallholders (coverage of the herd varied from 57 to 80 percent during the project - Table 4), the impact would be USS45 per smallholder family affected. Although this is a cumulative rather than an annual impact, and much less than intimated (but not quantified) in the SA-R.U/ it could be significant for smallholders below or close to the poverty line. 3.22 The small dairy scheme (paras. 2.12 and 3.12) would have substantially helped the cash income of the small number of suppliers (about 30 families) but is insignificant for the total project. This gross income from milk would have typically been about 6 x 8 liters x FMG 600 (or US$16.40) per week per supplier for the wet season and into the early dry season. 3.23 The above estimate of income relates to cattle benefits only. Some benefits were also obtained from the poultry and pig health program, and increased labor availability (or leisure time) was enabled by the village water supply program. Both of these activities had a favorable impact on the welfare of women in the households. Overall, the objective of increasing rural incomes appears to have been partially achieved, albeit to a much less extent that postulated at appraisal. Environmental Impact 3.24 There were no specific environmental objectives in the project. As no changes occurred in the production system due to the project and the herd increase associated with the health program was less than four percent, the project had no impact on the environment in the area. IV. FINDINGS AND LESSONS Sustainability 4.1 It can be assumed that the results of the health program are sustainable, and this is the basis of the economic return estimation for the Mahajanga component (para. 3.10). The size of the provincial herd may go up or down due to seasonal conditions, in- or out- Ji/ Cattle savings annually from 1983 to 1989 being 9000, 3000, 60,000, 15,000, nil, 12,000 and nil respectively compared to the pre-project situation (para. 3.3); this is 99,000 over approximately 3 million head or 3.3%. This percentage applied to the average herd of 46 yields: 46 X 0.033 X 0.12 x 300 kg x US$0.50/kg = $27 per family. ai/ The SAR assumed a 30% increase in the family herd over 12 years. -16- migration, disease incidence, or change in production systems, but the increase in the base herd (which reproduces itself) has been establish ed._A/ 4.2 As the benefits of the village water supply appear to be appreciated by end-users, it can be assumed that these facilities will be maintained in most instances. Continued maintenance of roads constructed or maintained under the project will depend on the funding .and management of the provincial Public Works Department and is an unknown quantity. The abolition of FAFIFAMA implies that the relatively high proportion of road investments in access to its facilities will result in a lack of attention to these roads. Prospects for maintenance of the animal health infrastructure (veterinary posts, cattle crushes) constructed under the project are not good unless improvements are made in the allocation of resources for this purpose and all facilities and promptly and officially transferred to SPEL. The enhanced transport facilities provided to FAFIFAMA animal health services have already depreciated and will make little future contribution. 4.3 The institutional mechanism used to provide the animal health services in Mahajanga in the form of the parastatal, FAFIFAMA was clearly unsustainable. This is evidenced in the dismantling of FAFIFAMA and the return to the regular SPEL to provide provincial animal health services, and in the current emphasis on finding means to further reduce the cost of public sector services. This raises questions about the appropriateness of project design, as discussed below (para. 4.6 and 4.7). To the extent that the SPA statistics can be relied on, an analysis of vaccination coverage in other provinces in the 70s and 80s also raises questions about the justification for the heavy expenditure on FAFIFAMA services in Mahajanga; these data are examined in the note attached as Annex 1 and show how one province, Antsiranana, has obtained good vaccination coverage without intensive project support. 4.4 The lack of effective monitoring of disease outbreaks under the project also inhibits reliable analysis of the value of attempting to have blanket coverage in a blackleg and anthrax vaccination program in all accessible areas. Both of these diseases derive from bacteria which persist as spores in the soil for many years. Improved monitoring of outbreaks would allow consideration of an alternative strategy of restricted public sector coverage in a regular annual program with a capacity to respond to outbreaks in other areas.15/ 4.5 The small dairy project has limited replicability. A consultant study in 1985 estimated that the total milk market in Mahajanga to be about 5,000 I/day. There is an established raw milk supply system from herds near to the city which could very likely be expanded to meet a limited demand increase. The small plant tapping suppliers further 14/ This is a case where annual production input results in a "permanent* asset accumulation. The capital base in the herd has been increased, but not to an extent where the effects of production constraints such as drought will be increased; the native rangeland feed base is not infinite, but certainly has the capacity to absorb the herd increase generated under the project without detrimental effects on the herd or rangeland. J/ In all cases, at least the cost of the vaccine could be recouped from the herders. - 17- away from the cityj_Q/ (para. 2.12) is only working at about 30 percent capacity. The audit was informed that the African Development Bank was examining a project to install large processing plants in major cities, including Mahajanga. This appeared premature to the mission in view of the current supply and demand situation and the difficult production environment in Mahajanga. 2esian Issues 4.6 FAFIFAMA was established to implement the former project, VLRDP-I on the grounds that (a) as a parastatal, it would have the administrative and financial flexibility considered necessary to implement the livestock and rural development activities, and (b) with a cross sectoral Board of Directors, cooperation between the various provincial agencies involved in development could be enhanced. FAFIFAMA experienced many organizational difficulties and underwent a change to the status of a State Farm to improve its operational flexibility; this, however, was at the cost of participation of agencies on the Board with a significant loss of cooperation. Despite the difficulties, the PCR on the VLRDP-I lauded the creation of FAFIFAMA as a major accomplishment of the project. It did raise the question of sustainability after cessation of external support, and suggested FAFIFAMA might eventually concentrate on livestock activities leaving rural development to other agencies. The PAR was sympathetic to this proposed role, noting that the animal health services had been well developed, but stressed the sustainability issue. 4.7 Under the existing circumstances, it was understandable that the appraisal accepted FAFIFAMA as the implementing institution for a VLRDP-II, which was virtually an extension of the previous project. However, the appraisal did not appear to seriously address the replicability/sustainability issue for animal health services through FAFIFAMA. The rural infrastructure components had not begn carried out well under the former project, so there was no overwhelming case to continue with FAFIFAMA for the rural development activities. For the two relatively successful components of VLRDP-II - animal health and village water supply - it could be argued that some assistance to SPEL health services would have been a reasonable option (considering the emphasis traditionally given to livestock vaccination services in Madagascar - Annex 1), and the village water supply could very likely have been undertaken through support to the provincial Public Works Department. 4.8 The legitimacy of the Bank concentrating its livestock subsectoral support efforts in Mahajanga was questioned during the mid-term review. This was on the grounds that subsectoral policies needed to be rationalized, the potential for livestock improvement was limited in Mahajanga, and concentration on milk from cattle and on other livestock species (pigs and poultry) would have been more appropriate. 4.9 The first criticism has some merit. The Bank was obviously aware of the unsatisfactory subsectoral policy environment, as evidenced by the inclusion of a series of studies in the subsectoral component. Most of these studies, however, were only J& The nearest of the six collection points is 18 km from the city. -18- undertaken towards the end of the project (which was the main rationale used by the Bank to justify a one-year extension). In hindsight, it may have been more appropriate to finance studies as the basis for a program along the lines of the current Livestock Sector Project and to reject the emphasis given to the Mahajanga component. Thi Livestock Sector Project includes policy reforms, strengthening of existing subsectoral institutions and promotion of private animal health services, epidemiological surveys in extensive areas, and dairy development in the higher potential plateau environment. 4.10 Regarding the limited provincial potential for livestock improvement, this is correct in that the potential for productivity gains on the Mahajanga rangeland and in the extensive low-input production system was very limited. The difficulty of introducing improved, legume-based pastures under current and prospective herd and pasture management systems was already evident from the firot project. To its credit, however, the appraisal did substantially reduce the production component to a pilot activity in which those villages with potential for improvement in mixed farming systems were to be selected. It would have been preferable, however, to have had the planned technical and socioeconomic survey carried out before the project so as to assess real possibilities; this component failed (para. 3.10). The improved and native pasture commercial schemes were rather naive from the point of view of potential replicability; these also failed. 4.11 The financing of drug imports was obviously included to alleviate the foreign exchange shortage. In retrospect, however, the management and organization of PCV should have been studied more thoroughly as an importing monopoly, and procedures introduced to improve its efficiency as a condition for disbursement. Conclusions 4.12 In short, the incremental investment in animal health service in Mahajanga largely realized the intended coverage for the project period, and gave a satisfactory economic return to the GOM. Vaccination against the major disease challenges in Mahajanga represented a low cost-high return annual production expenditure. Under these circumstances, a high payoff is virtually inevitable on this investment by a herd owner, or by a public service providing vaccination in accessible areas. The objective of Bank involvement in this type of project, however, must be to establish an effective, sustainable system for the delivery of services to improve herd productivity; this, the project did not do. The service institution was dissolved after project completion. Furthermore, the livestock production subcomponent of the project was aborted during implementation. As the Mahajanga animal health subcomponent of the project was its major focus, the audit is unable to rate the project as satisfactory, despite the apparent acceptable return on the GOM expenditures in Mahajanga and the usefulness of many of the studies. 4.13 Obvious lessons from the VLRDP-1I are: (a) Projects providing animal health services are likely to provide attractive returns on investment during implementation, but such projects are not a suitable development medium for the Bank unless the support is geared to the -19- establishment of efficient and sustainable arrangements to provide these services. (b) In proposing production system modifications involving improved technology and practices, a knowledge of the characteristics of the existing production systems and the attitudes of the target group should be ascertained before including such a component in a project with a limited implementation period; that is, the investigations should be made before appraisal rather than during the project regardless of whether the component is of a pilot or wider nature. (c) In formulating and appraising projects, subsectoral conditions and requirements should take precedence over pressures for continuation of traditional interventions, especially in the use of scarce IDA resources. - 20 - Tabte 1: PHYSICAL ACHIEVEMENTS Ind cators Appraisal Objectives Achievement at mid-term ch evement at Credit (January 1982) evaluation (September 19971 Closing (December 1989) Mahalange Component Activities a) Animal health . Area entire province entire province entire province . Vaccination of beef cattle 2,700,000 2,553,000 (1985) 2,162,000 . de-worming of calves (%) 22 17 (1985) 11 . Vaccination of swine 20,000 51.892 (1985) 32,454 . Vaccination of poultry 120,000 700,779 (1985) 459,388 . Construction of veterinary 80 86 86 centers " Construction of crushes 20 16 16 Personnel recruitments veterinarians 6 4 n.a. veterinary assistants 30 53 n.e. village vaccinators 56 106 80 b) Pilot production program . intensification in combined 40 village areas stopped stopped zone . fattening cattle commercial trials stopped stopped farm trials not started not started c) Construction and maintenance of rural roads 100 km 383 km 383 km d) Water supply . livestock 35 ponds 5 boreholes and ponds S boreholes and ponds . village 80 wells 75 wells 75 wells e) Support to FAFIFAMA . Training (internat) for all staff to large extent achieved to large extent achieved . Monitoring and Evaluation partially realized partially realized Subsector Component Activities f) Technical Assistance to improve the sub-sector . Study on Beef supply 1982/1983 1987 1987 for Antananarivo . Studies of key early in project partially achieved partially achieved institutional issues . Training continuous very minor very minor . Material support procurement of completed completed vehicles and other equipment g) importation of products and materials .veterinary products $ 3 M S 2 N S 2 M . Laboratory equipment S 100,000 S 150,000 $ 150,000 h) Applied research S 200,000 cancelled cancelled i) Studies, documentation, S 500.000 partially realized partially realized study grants Source: Largely PCR. -21 - Table 2: PROJECT COSTS SAR Estimates ActuaL Costs FM FMG "ittion %Iion % Regional Program i Animal Health 2,831 40 2.750 25 Pilot Program 812 11 430 4 Rural Roads Program 1,048 15 340 3 Water Supply Program 527 7 510 5 Intensification of 200 3 2.270 21 FAFIFAMA /S Sub-total 5,418 (76) 6,300 (58) Livestock Sub-sector Program Strengthening of SPA /h 550 8 2,430 22 Veterinary Products 1,029 14 1,950 18 Laboratory Equipment 33 1 150 1 Applied Research 77 1 * Sub-total 1,689 (24) 4,530 (42) TOTAL 7,107 100 10,830 100 /g The actual costs include many operational expenditures which could have been altocated to specific subcorponents in a more refined accounting system; hence the high actual figure. /k This category includes alt technical assistance, studies, scholarships and documentation. Funds were reatlocated to cater for an expansion in the study program and for preparation of the Livestock Sector Project which led to the high actual expenditure figure; after the 1987 review, however, project records deteriorated so that the accuracy of the figures is questionable. Source: PCR Teble 3: PROJECT FINANCING SAR Estimate Actual /i Actual as % US mittion USS million S SAA IDA 15.0 58 7.4 60 49 IFAD 8.0 31 4.4 36 55 0N /h 2.8 11 0.5 4 18 TOTAL 25.8 100 12.3 100 48 /S After the mid-term review, SOR 5.85 mittion was cancetted of the IDA credit and SOR 3.15 million of the IFAD loan. /b The Government contribution is only an estimate. Source: PCR data. Table 4: NUMBERS OF CATTE *000) VACCINA7ED AGAINST LACKLEG AND ANnRAX AND PERCENT COVERAGE PER PROViN (968-19911 Yoer Total Ant*frnana Mahajanga Towma~ina Antmanaoro Ranarant~oa Toay No. % /g No. /N % No. % No. % No. % No. % 1968 5.211 51 1971 5.259 1972 5.021 1973 4.443 1974 4.623 47 294 42 1.229 41 184 26 530 43 572 46 1.914 61 1976 5.285 300 1,423 48 260 436 827 2.039 1978 5.311 52 418 57 1.508 51 264 37 607 48 861 68 1.654 52 1980 5.122 306 1.877 62 267 717 755 1.201 1981 4.962 323 1.828 62 382 717 569 1.143 1982 6.229 1,917 62 1983 6.050 1.974 63 1984 5.785 57 399 53 1.914 61 308 48 847 66 666 57 1.598 51 1985 6,662 2.553 80 1986 6.221 707 2,129 65 343 834 618 1.789 1987 5,878 1,899 57 1988 6,547 64 463 60 2.162 64 245 40 858 66 463 41 2,000 64 1989 6.147 1.826 57 1990 6,032 58 459 59 1.904 58 337 47 829 64 581 48 1.923 62 1991 n... n.a. 1.693 52 n.a. n.a. n.a. nl.a, etimated 10.332.200 770,300 3.266.700 709.400 1.285.300 1.200,500 3.100.000 1990 h~td aize Nurnbrs of vaccination. and percentage coverage of estimated hend in a particudar year ae presented for periodic years to indicate relative performnance trenud within and between proucaS; Mahajanga figure am. complete 0 rom 1974. n.a. not avaable Source: D~erved from SEDES repr, interpr~tations of SPA Inomal studiee and SPA record. - 23 - IIIIII 1-illll!------- Ill8l ! - 24- i ___ i I :fat * 25 - ANNEX 1 A NOTE ON ANIMAL HEALTH COVERAGE AND CATTLE NUMBERS IN THE PROVINCES OF MADAGASCAR VACCINATIONi 1. An examination of available SPA Statistics on vaccination of cattle against blackleg and anthrax diseases in each province of Madagascar from 1968 to 1990 (Table 4 in the text of the PAR) reveals the following: (a) The apparent coverage of the national herd as a whole has ranged from 50 percent to 64 percent except for a drop to around 45 percent in the mid 70s. A SEDES study indicated that a coverage in excess of 70 percent would be unlikely to be economic considering the costs involved in coverage of the more inaccessible herds in the main cattle areas. (b) TQamasina, a cattle deficit area in the humid east coast, has the lowest cattle population and has consistently had the lowest vaccination coverage. (c) Antsiranana in the north also has a low cattle population but is a cattle surplus area and has maintained a relatively high vaccination coverage. The likely reasons for this favorable performance in vaccination are (i) the services by (SPEL) have been relatively well-managed, (ii) cattle are an important commodity in the province and are prominent in draft power,1/ (iii) since the mid 70s the province has received some support from Japanese aid in a local animal health training facility. (d) Mahajanga and Igliaty have the highest cattle populations which are largely managed under extensive conditions, and both have had a relatively good vaccination coverage. IliLary was the recipient of an FAO vaccination assistance program in the 70s, and subsequently the SPEL services received special attention, apparently due to the Minister responsible for the subsector being a native of the province and a previous manager of the FAO project. Mahaianna was the recipient of substantial support under the two Village Livestock and Rural Development Projects (1976-82 and 1983-88). The data in Table 4 indicate that the Tollary program was better than Mahajanga (and all other provinces) when it received the very intensive FAO project assistance in the mid-70s; the Mahajanga program was better than Toliary during most of the period of the Bank-supported projects. In recent years, vaccination performance has been similar. .1 The SEDES data for 1987 estimated the percentage of adult males used for draft purposes as follows: Antananarivo 77.7, Toamasina 58.9, Mahajanga 44.6, Antsiranana 40.8. Flanarantsoa 20.1 and Tollary 14.6. - 26 - (e) Antananarivo's vaccination program has been consistently as good as or better than that the other provinces. This program by the provincial SPEL has likely performed welt because of (i) the proximity to the national capital, (ii) generally better accessibility, and (iii) a greater occurrence of more intensive forms of cattle production (fattening and milk production) and many draft cattle, (f) Fianarantsoa's coverage deteriorated in the 80s. This is likely explained by the increased amount of cattle rustling in this zone with negative consequences for services and attention to the industry, and a deterioration in roads for access. 2. The foregoing observations suggest that, while the intensive inputs associated with major projects can increase coverage, it is possible with good management and staff training to have the provincial SPELs undertake vaccination programs with an acceptable performance level (e.g. Antsiranana). Internal Parasite Treatment 3. Table 5 in the text of the PAR provides a record of treatments for liver fluke and roundworms by province from 1974 to 1990. 4. Liver fluke treatments tend to be high in provinces where the environment is favorable for the lymnaeid snail (as an alternate host) and for the encysted stage of the liver fluke life cycle. This coverage is no doubt due to demand as the benefits from treatment of infected animals are obvious. There was a marked drop in treatments in 1990, possibly because of the uncertainties associated with the GOM drug procurement and distribution system changing over to private sector procurement, and with budgetary constraints within SPA and SPELs. 5. The Mahajanga project did appear to increase liver fluke treatments, but this parasite is of limited significance in the province overall. 6. Roundworm treatments have not been as numerous as liver fluke treatments nationally. Adult Zebu cattle tend to develop resistance to these parasites and treatments often do not have obvious results in animal condition or performance. As cattle owners have to pay for these treatments (but not for vaccines), less demand could be expected. 7. The treatment for roundworms was relatively high in Mahajanga under the two Bank- supported projects. This was especially in the period of peak project activity in 1985/6, where the number of treatment reportedly was equivalent to about 20 percent of calves in the province (but many calves could have been treated more than once in a year). This should have reduced morbidity and mortality in this susceptible age group. More recently, however, the coverage has decreased dramatically (see para. 4). 8. The better performance of FAFIFAMA was likely partly associated with its capacity to import its own drug requirements and avoid the procurement and management deficiencies of the PCV. - 27 - Cattle Numbers 9. Livestock population estimates were undertaken by SEDES in 1978 and 1984-87. These provide an indication of the herd population changes in the six provinces: Percentage change in cattle population from 1978 to 1987 Antsiranana + 4.4 Mahajanga + 11.0 Toamasina - 15.7 Antananarivo + 3.0 Fianarantsoa - 11.5 Toliary - 1.6 Madagascar + 1.0 10. The large decline in cattle numbers in Toamasina and Fianarantsoa is consistent witt. the analysis of the vaccination coverage (para. 1(b) and (f)). Insecurity and cattle rustling, a decline in market activity, a decreased purchasing power in much of the east coast market areas, and heavy incidence of liver fluke would contribute to low reproductive performance, less attention to the cattle enterprise, and outmigration in the case of Fianarantsoa. 11. The increase in the herds in Antsiranana is consistent with its relative isolation, the importance given to cattle and the relatively good animal health services (para 1(c)). The increase in Antananarivo is also consistent with the explanation given for good health services in this province (para 1 (e)). 12. Toliary registered a population decline, despite its importance as a cattle region, and despite its reasonable animal health coverage. The SEDES survey data, however, showed a high incidence of insecurity and cattle rustling in many parts of this zone which would explain significant out-migration of herds, especially north to Mahajanga where conditions were better in most of the province and where there was a substantial expansion in cropping. 13. Mahajanaa showed the highest growth rate, which is consistent with its importance as a cattle producing area, the expansion in cropping and use of cattle for draft, and the good animal health coverage achieved under the two Bank-supported projects. If the assumptions used to interpret actual vaccinations into animals saved (para. 3.3) is used, the incremental vaccinations due to the two projects after 1978 to 1987 would imply an increase of 13 percent in the herd due to this single factor, whereas the actual increase was 11 percent. This suggests that the assumptions used in assessing impact of the vaccination program have been generous, as in-migration is almost certain to have been a significant factor in the herd increase. It is more likely that at least 2 points of the 11 percent increase would have been due to in-migration, which would suggest a 30 percent reduction in the vaccination-herd increase equation. - 29- ATTACHMENT 1 State Ministry of Agriculture and Rural Development UNOFFICIAL TRANSLATION Ministry of Stockraising and Fish Resources Office of the Minister Ref. No.: MERH/GAB/GM.4/B Antananarivo, January 26, 1993 Division Chief Agriculture and Human Resources Operations Evaluation Department Washington, D.C. Subject: Second Village Livestock and Rural Develooment Project (Cr. 1211 -MAG) Reference: Project Performance Audit Report sent by your letter of December 22. 1992 Dear Sir: After carefully reading the above-mentioned document, I should like to make the following comments. It would have been instructive to specify the context in which these Seepas 4.6- two projects were begun, and to know if feasibility studies had preceded implemen- 4.11 of PAR tation. In point of fact, the gaps noted in the PPAR raise questions as to the degree of rigor observed. Furthermore, the report highlights positive achievements, in particular as regards the animal health and food components of the project, but questions their sustainability, especially since the dissolution of FAFIFAMA for field operations, and the General Veterinary Pharmacy was not succeeded by a substitute entity or supplementary operational measures. Given the disproportionate means, the re-incorporation of the Mahajanga SPEL will not suffice if the achievements are to be sustained, and privatization of the veterinary profession does not cover a sufficient amount of the country to ensure adequate animal health coverage. One would have hoped that the last studies, having justified extension S pa 2.22 of the project, would be able to shed light on the current Livestock Sector Project, and 3.17 particularly in view of its numerous components. Please accept, etc. Is/ Tsialetra IBRD 15637 MADAGASCAR SECOND VILlAGE UVESTOCK AN1ImANA AND RRAL DEVELOPMENT PROJECT DEU)fEME PROJET D'ELEVAGE VLAGEOIS ET DE ~LflZOPMNT A4t MAAJANGA FARITANY 1 FARITASYEMAA.ANGA Sb0oe OlGNONA pEEflln N INEFIED, VETERINARY BERVICE UNDER CREDIT 503 MAG I V~+ F1VON AN AVAN NEFICIE D'£~DEMNT VERIAWAE IAWF A LEMA CADE oU CRE0/M7om- MAG PILOT DEVELOPMENT PROGR0~S - MIDEWEST UNDER CREDIT El» VAGI l PAGRMES PILO UMEN k UE CAR D CREDIT 6 MAG O OMBY STATE RANCNES / RANHES DE LA FERME YETAT OMSy BITUMINOUS BURFACEO ROADE ROUTES A SURFACE 1LTUMEE Sb - ALL WEATNER ROADS ROUMTES PRATICAOLES TOUTE LANNEE RAILROADS /8EMINSDE FER FARITANY 8OUNDARIES lMITES ES ARITANY PORMER PREFECTURE 80UNDARIES / LIMiTES DES ANCIENNESOUS-PREFECTURES ---PVNRONAN4A BOUNDARIES pHOWN IN M~NJ/GA FARITANY OSLY) 1 -.)\ - mlh - åt TESDES FVONDR~ N ~ASUNQUEMENT E ASITAR YEMAHAANA ALL WEATHER AiRPORTS AeROPORTSOUVERs 1UTE L-ANNEE CAPITAL CAPTALE F FARITANY CAPITALS / CAPITALES DES FARIT ANY OCHIEF TOWN IN PIVONORONANA 1 W RNIAEVLLE DES PI1VONDR~SÅtA7 - X 41F t~ D FORMER OREFECTUR E CU PI~JVILLEDES~NCAES SOUII-PIPECTLXRES 0 Mozombiqu e - ir- Can/Tn dia n Ocean 20V hobo Mo .o.bdndy k- NosyVunkö FIAN'ARANTSOA Se' Amýbo o n C eno 2 2 KM . Ankzoobo~ / Monokoro O 50 160 10 MILES ..be Vohqpno T FarBflngkn 24- -2A AFRICA A nd , Y MADA0~ iA FEa;uJARY i993
Groupe de la Banque mondiale · Project Performance Assessment Report
Madagascar - Second Village Livestock and Rural Development Project
Voir le document original
Le texte intégral est hébergé par l’organisation qui le publie. lawenc.com indexe les métadonnées et renvoie vers la source officielle.
Texte intégral
Informations clés
Organisation
Groupe de la Banque mondiale
Type de document
Project Performance Assessment Report
Pays
Madagascar
Source
Banque mondiale