Documet of The World Bank FOR OmCIL USE orLY Report N. 5775 - ; PROJECT PERFORMANCE AUDIT REPORT HAITI: RURAL DEVELOPMENT PROJECT IP RHE NORTHERL DEPARTMENT (CREDIT 675-HA) June 28, 1985 a Operations Evaluation Department dTs d.cunt hm a nlcu dbirHkM md way be mod by repae uy In the pedfoamm .o * theidr aSahi mudm lb comuku MY mayd oehwlube be dlselese withou WerM Bank mutbe.ulati. WEIGEHTS AN(D MEASURES Metric System I meter Cm) = 39.37 inches I kilometer (kmn) 0.62 miles 1. lectare (ha) = 2.47 acres - 0.78 carreaux I square kilometer (km2) -0.386 square miles I kilogram (kg) = 2.205 pounds I ton Ct) 2,205 pounds ABBREVIATIONS BCA - Bureau de Credit Agricole (Agricultural Credit Bureau) CIDA - Canadian International Development Agency CRUJDEM = Centre Rural de De'veloppement de Milot (Milot Rural Development Center) FAC - Fonds d'Aide et de Coopiration (French Bilateral Aid Agency) FAO - United Nations Food and Agriculture Organization GTZ - West German Bilateral Aid Agency IDA - International Development Association IDB - Inter-American Development Bank MARNDR = Ministry of Agriculture, NLatural Resources and Rural Development ODN - Northern Department Development Agency ROR - Rate of Return TPTC - Miuistry of Public Works WFP = World Food Programm USAID = United States Agency for International Development GOVERNMENT OF HAITI FISCAL YEAR. October 1It September 30 FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT HAITI: RURAL DEVELO?MENT PROJECT IN THE NORTHERN DEPARTMENT (CREDIT 675-HA) TABLE OF CONTENTS Page No. Preface . ........................................................... Basic Data Sheet ................................................... Highlights . .........................................................i PROJECT PERFORMANCE AUDIT MEMORANDUM A. SUMMARY . ................................................. 1 B. MAIN FINDINGS ............................................ 5 PROJECT COMPLETION REPORT I. Introduction .............................................. 13 1I. Background ................................................ 13 III. Project Formulation ....................................... 13 IV. The Project ............................................... 14 V. Implementation ............................................ 15 VI. Agricultural and Economic Impact .......................... 22 VII. Institutional Performance ................................. 25 VIII. IDA Performance ........................................... 27 IX. Lessons Learned ........................................... 28 Annex 1 ....................................................... 30 Map - IBRD No. 12217 This document has a restricted distribution and may be used by recipients only in the performance of their official duties- ts contents may not otherwise be disclosed withoul Wodd Bank authorization. :S:~~~~~~~-i - PROJECT PERFORMANCE AUDIT REPORT HAITI: RURAL DEVELOPMENT PROJECT IN THE NORTHERN DEPARTMENT (CREDIT 675-HA) PREFACE This is a performance audit of the Rural Development Project in the Northern Department of Haiti, for which Credit 675-HA was approved on December 18, 1976, for the sum of US$10 millior.. The final disbursement took place on June 30, 1983. The audit report consists ef an audit memorandum (PPAM) prepared by the Operations Evaluation Department (OED) and a Project Completion Report (PCR) dated May 31, 1984, prepared by the Latin America and the Caribbean Regional Office. The audit memorandum is based on a review of the Appraisal Report (No. 1281-HA of December 13, 1976), the President's Report, the Credit Agreement (dated January 14, 1977) and the PCR; correspondence with the Borrower and internal Bank memoranda on project issues as contained in relevant Bank files were consulted. Bank staff associated with the project have also been interviewed. An OED mission visited Haiti in February 1985. The mission held discussions with the Minister and staff of the Ministry of Agriculture, Natural Resources and Rural Development (MARNDR) and with staff of the Northern Department Development Agency (ODN), the entity responsible for project execution. A field trip to visit the project area and the participating farmers was undertaken. The information obtained during the mission was used to test the validity of the conclusions of the PCR. The audit finds that the PCR covers adequately the project's salient features, and the audit memorandum generally agrees with its conclusions. In addition to summarizing the objectives and outcome of the project, the memorandum devotes special attention to three issues which are seen in numerous rural development projects: insufficient preparation, an error of judgment regarding the time needed to exert a significant impact on agricultural production, and a project organization which, from the outset, did not meet the requirements of integrated regional development. A copy of the draft report was sent to the Borrower for comments on April 8, 1985. On the same date, a copy of the draft report was also sent to the Fonds d'Aide et de Cooperation (France), the Gesellschaft fur Technische Zusammenarbeit (Germany) and the World Food Program (United Nations) which cofinanced the project with the Bank and the Government of Haiti. No comments have been received. The valuable assistance provided by the Government of Haiti, and the project staff met during the preparation of this report, is gratefully acknowledged. - ii- PROEC PEmMCB AUDerr EPOfr AITM: UATAL DELWIIENT PROJECr IN SE NORTHERN DEPARTNT (CREDrrT 675-HA) LASIC DATA SHE KEY PROJECT DATA Actual or Actual am I Appraial Estimted of Appraisal Estimate Actual Estimate Project Cost (USS million) 13.4 18.0 134 Amount of External Financing IDA Credit 10.0 10.0 100 WFPK/ 1.4 1.4 100 CTZBj ) - 2-6 - FAC5J ) Date of hoerd Approval n.&. 12/28/76 - Data of Effectiveness 05/17/77 05/26177 - Date Physical Coaponents Coopleted 06181 06/83 144 Proportion tben Completed U.&. n.e. - Closing Date 12/31/81 06130/83 133 Economie Rate of Return 142 negative - Nuber of Direct llenef iciry Pamiles Agriculture 4.000 U.n. - Weter Supply 3.000 14.700 490 COMlATIYE DISBURSEMENTS ms78 FY7 FM80 81 P62 FY83 Appraisal Estimate (USSVOOO) 2,400 4.800 7,600 L0.000 - - Actual (CS$'000) - 900 2.000 4.200 7.800 10.000 Actual as 2 of Estimte - 19 26 42 78 100 Date o Fin l Disbursement: March 21. 1984 Staff/ Speclall- Perfor- Tpe of No. of days zntone ance Types of tflmston Date Persone in Field rentd dg Eatlng5/ rend tv Frobl *, Identification 1973 - - - - - Preparation 02/05/75 - - - - - -- Appraisal 04105/76 7 - a.b.c,d.e - - - Sb.'ervision 1 03/77 1 10 a n.a. n.a. n.a. 2 06/77 1 9 a Z I n 3 08/77 1 2 a 2 1 L 4 12/77 2 8 a 2 1 M.0 5 05/78 2 8 a.d Z 1 n.0 6 08178 3 9 *.c.d 2 1 N.0 7 01/79 2 11 a.d 3 Z K.F.0 8 05/79 3 6 *.d,- 3 2 Fn.O0 9 10/79 2 S a.d 3 2 FHM.0 10 02/80 2 7 a.d 3 2 F.K.0 11 06/80 1 b a 3 1 n.P.0 12 10/80 2 7 a.d 3 2 r.n.r 13 01181 1 8 a 3 Z n.r.P 14 05/81 3 6 ..c,d 3 1 H.P 15 10/81 3 11 !f a.b.d 3 1 H.P 16 01/82 3 -5 a.b.d 2 I p 17 05/82 3 13 a.b.d 2 2 F.l 18 22/83 2 17 i/ ad 2 2 r 19 07/83 2 121/ a,c 2 Z r Total 103 OTHER PROJECT DATA First mention in files 05/16/73 Negotiation 11/18/76 Credit Agreement Date 01114/77 borrower: Covern ent of Haiti Executing Agency: Agency for the Development of the Nortb (ODN) Fiscal Year: October I - September 30 Name of Currency (Abbreviation) Covrde (C) Appr-esaL Yesr Average SS31.00 - C 5 Intervening Years Average OSS1.00 - C 5 Completion Tear Average 01531.00 - C 5 Follow-on Project: Second Rural Development Project Credit N_ber: 1410-NA Credit Amount: USS19.1 million Date heard Approval: August 23. 1983 */ WFP - World Food Progrm bl GTZ - Cerman Bilateral Aid c/ FAC - French bilateral Aid d/ a-agriculturalist; b-agricultural economist; c-ftuancial analyst; d-gineer; end e-credit specialist. e/ 1-problem free or no problem; 2'eooerste problem; and 3_mjor proble. fI 1-imsprving; 2-atationary; and 3-deteriorating. Lj P-financial; N-managerial; T-technici; P-political; and O-other. bI About 6 days out of 11 devoted to en interis review of the project. Tl About 12 days out of 17 devoted to the preparation of a follow-on project. About 5 days out of 12 devoted to the project. - iii- PROJECT PERFORMANCE AUDIT REPORT HAITI: RURAL DEVELOPMENr PROJECT IN THE NORTHERN DEPARTMENT (CREDIT 675-HA) HIGHLIGHTS The project was the first Bank operation in the agriculture sector in Haiti. It was designed to increase agricultural production and improve the incomes and quality of life of approximately 200,000 inhabitants of the Northern Department. The main components of the project were: (i) rehabilitation and expansion of irrigation systems (3,100 ha); (ii) improvement of support services for both irrigated and rainfed agriculture (approximately 50,000 ha); (iii) construction of 400 wells and five water supply systems; and (iv) construction or rehabilitation of three markets. Execution of the project was to be entrusted to a new, autonomous entity, the Northern Department Development Agency (ODN), whose role it would be to plan, finance and coordinate the actions of existing government departments, which would then be responsible for carrying out their particular components of the project. In view of the complexity of the project and the lack of adequately qualified Haitian personnel, a number of expatriate consultants were to be recruited, their services being financed under West German and French bilateral aid. The project encountered a wide range of difficulties from the outset and was very soon classified as a -problem project. Delays in execution occurred as a result of the slowness of counterpart funding procedures and the late arrival of certain consultants. The project was also affected by the lack of basic data on the socioeconomic and institutional environment and on the topography and hydrogeology of the region. As a result, (i) ODN was obliged to conduct a series of preliminary studies, which concluded that a part of the irrigation component was in fact not feasible; (ii) the agricultural technical packages were virtually non-existent and the project had to engage in research that was more extensive and time-consuming than foreseen; (iii) the farm credit and extension components were cut back significantly for lack of technical packages to recommend to farmers; and (iv) instit'.tional weaknesses in the government technical departments involved obliged ODN to carry out the majority of necessary civil engineering works itself. On completion of the project, two years behind schedule, the rural roads and water supply components exceeded appraisal targets. Only one market was built instead of three. The agricultural component was limited to partial rehabilitation (460 ha) of one irrigation system and a series of research and demonstration programs on irrigated and rainfed crops. Final - iv - project cost exceeded projections by approxim&tely 34Z. Since only a minimum increase in agricultural production was achieved, the project rate of return is negative. A follow-up project, approved by the Bank in August 1983, should make it possible to move into a productive phase thanks to the studies, research and demonstrations carried out during the first project and to the regional development experience ODN now possesses. The economic setback the project suffered can be ascribed mainly to three serious shortcomings: hasty and insufficient preparation for a project of this nature, poor assessment of the time needed to achieve an incremental agricultural production, and poor interagency coordination. These conclusions demonstrate the importance of having a long-term regional development plan (PPAM, paras. 33-34) and the need to plan regional development in several consecutive phases, the first of which should be a pilot project (PPAM, para. 27). The following points may be of special interest: - the absence of a master development plan partly explains ODN's difficulties to coordinate its actions (P?AM, para. 36) and continues to be a handicap to the project (kPAM, 38-41); = pressure from the Bank and the Government for rapid launching of a project had an adverse effect (PPAM, para. 35); weaknesses in government line agencies led ODN to step in and become the executing agency for the majority of project works (PPAM, para. 29); = the transfer of the project water supply facilities from ODN to SNEP is a good example of cooperation both with the benefiting population and government line agencies (PPAM, para. 32); and the Haitian authorities made a concerted and successful effort to simplify procedures and allow ODN adequate financial autonomy (PCR, para. 7.10). PROJECT PERFORMANCE AUDIT MEMORANDUM HAITI: RURAL DEVELOPMENT PROJECT IN THE NORTHERN DEPARTMENT (CREDIT 675-HA) A. SUMMARY Background 1. With a surface area of 27,750 km2, 70X of which is mountainous and arid, and more than 5 million inhabitants, Haiti is heavily populated (400 inhabitants per km2 of arable land) by comparison with its agriculturEl rosources. It is also the poorest country in the Americas. 2. In 1975, agriculture accounted for approximately 45% of GDP and over 50% of exports, providing a livelihood for roughly 802 of the populace. Despite the importance of agriculture to the economy, production was increasing by no more than 1.2% a year, a rate below average population growth, then 1.6%. The Government allocated only 7% of the total budget to agriculture; research, extension and farm credit support services were glaringly inadequate; traditional production methods were employed for the most part, except for sugarcane, rice and tobacco. 3. Because of the constraints encountered in the sector, nationwide f forts had little chance to succeed and the Government was formulating projects and plans for regional development based on on-going projects such as the Artibonite Plain, the Gonaives Plain, Petit Goave and Cayes. These development projects aimed at upgrading basic infrastructure, support services and farming techniques simultaneously. Project Design 4. The project was the first Bank-financed operation in the agricul- ture sector, Bank action in the country up to tt-at point having been limited to the road, education and electrification sectors. Completion of the main road from Port-au-Prince to the north, financed partially by IDA, and the considerable potential offered by the Northern Plain were the main reasons for choosing the Northern Department as the region in which to implement a rural development project in line with the Government's new regionalization policy. 5. The project was identified in June 1974, and prepared in January- February 1975 under the aegis of FAO/CP. After the Bank had sent out two pre-appraisal missions, in October 1975 and January 1976, final appraisal was carried out in April-May 1976. As presented to the Board at the end of 1976, the project focused on increasing agricultural production in the Cap-Ha!tien district (approximately 50,000 ha) and on improving the quality of life for the majority of the region's 220,000 inhabitants, 92% of whom were living below the absolute poverty line. -2- 6. The major components of the project were: (a) rehabilitation of the Quartier Morin (1,200 ha) and Saint Raphail (1,900 ha) irrigation systems and expansion of the irrigated area at Quartier Morin by a further 1,800 ha through canal extension and the construction of deep wells; (b) building or rehabilitation of 66.5 km of feeder roads and 27 km of farm access roads; (c) construction of 400 wells equipped with hand pumps, and five water supply systems served by springs, to provide potable water to local villages; (d) construction or rehabilitation of three markets; (e) institution of a program of agricultural trials and seed multipli- cation; (f) provision of improved agricultural services (extension, credit, animal health) and organization of rural community development; and (g) administration of the project. 7. Project execution was to be the responsibility of a new, regional agency, ODN (Organisme de Developpement du Nord), with headquarters in Cap-Hartien. Its Board of Directors, chaired by the Minister of Agriculture, was to include representatives of the other Ministries involved in the project. ODN was intended to be a relatively small agency, whose main role would be to plan, finance and coordinate the activities of existing govern- ment service departments (the Public Works, Irrigation, Agricultural Extension, Water Supply, and Veterinary Departments, and the Agriculture Credit Bureau), which would be responsible for their own particular compo- nents of the project. Eight expatriate specialists were to be recruited to strengthen ODN's technical capacity. 8. Total project costs were estimated at US$13.4 million, of which US$10 million (75%) would be financed by IDA, approximately US$1.4 million (10%) by the World Food Program (WFP) in the form of food rations for local unskilled labor, and approximately US$2 million (15%) by the Government of Haiti. It had been agreed during project negotiations that the cost of consultants (US$2 million), would be cancelled from the IDA Credit if the funding could be obtained from West German and French bilateral aid. It was estimated the project would take four years to complete. The economic rate of return was calculated at 14%. Project Execution 9. The Credit Agreement became effective in January 1977. Although the French and West German aid agencies agreed to finance the technical -3- assistance component, it was decided to retain the US$2 million to cover possible contingencies. The consultants financed under the bilateral aid arrangements with France arrived on the scene promptly but it took the Wesc German Government two years to assemble its team of specialists. In addi- tion, the directorship of ODN changed frequently, while shortages of Haitian counterpart personnel was a chronic problem during the project's early years. 10. Other factors contributing to project difficulties were: (a) The lack of information on the socioeconomic and institutional environment and on the soils, topography and hydrogeology of the project area, which made it necessary to conduct a series of unscheduled studies. (b) The lack of technical packages necessitating numerous adaptive trials on a range of crops. Although these activities had been planned under the project, finding suitable packages took much longer than expected, leaving both extension and credit idle. (c) A delay in the allocation of counterpart funds and the time- consuming nature of the procedures for authorization of payments meant that until a revolving fund was established in the Central Bank four years later, the project lacked the cash resources necessary for full-scale implementation; 'd) The main government agencies scheduled to carry out project activi- ties had only very limited implementing capability, a situation which forced ODN to assume their responsibilities, principally for irrigation, water supply and road building; and Ce) The sugar mill, a private corporation, was unable to reach the target of 300,000 t of cane p.a.; over 100,000 t (equivalent to US$1 million) were not harvested or were lost each year.l/ 11. As a result of these problems, the Bank and the Government decided to redesign the various parts of the project, particularly its agricultural component. First, the Quartier Morin irrigation component was dropped as studies had found that increased sugarcane production in this area needed improved drainage, not irrigation. Second, in view of the difficulties of organizing an effective irrigation system at Saint Raphael, the program was cut from 1,900 ha to 750 ha, of which only 460 ha were developed in the end. Third, rainfed crops were now to be predominant in the project, so the pro- gram of trials was extended to the whole of the Northern Department and seed production was undertaken. 1/ The sugar mill was subsequently nationalized and its management was entrusted to another private group, considered more efficient and forceful. The latest Ecoromic memorandum on Haiti, is recommending, however, that the sugar mill be closed because of continuing excessive losses. -4- 12. The road building program encountered serious difficulties from the outset owing to faulty distribution of the WFP foodstuffs to be used as remuneration for labor and to slow performance of construction by the Public Works Department. Subsequently, ODN began to carry out a large part of the roads program itself with its own equipment; this resulted in the construc- tion or rehabilitation of 97.2 km of feeder roads and farm access roads, slightly more than originally scheduled. 13. The potable water supply program completed 150 wells and 25 commun- ity systems, which benefitted a population of about 100,000, a much larger number than originally planned. Two notalle aspects of the program were the productive cooperation between ODN and the line agency, SNEP, and the effec- tive participation of the target populationt in constructing the water supply systems (para. 32). 14. The component calling for construction or renovation of markets was limited to rehabilitation of the Cap-Raltien market, the cost of which proved much higher than projected at appraisal. 15. Construction of the ODN office and housing suffered delays and the costs were higher than originally antLcipated. Expenditures for project administration were also higher, since ODN went considerably beyond its initial functions of coordinating, financing and planning to become the major executing agency for almost all components of the project. Moreover, addi- tional consultants were retained to assist with the new needs that became apparent in the course of the project. 16. Delays in implementation, the changes made in the project design and the higher input prices i -ulted in postponement of the project comple- tion date by two years and in a total project cost of US$18 million (includ- ing technical assistance and project administration expenses), representing a cost overrun of 34X. 17. Agricultural prodrction in the project zone has increased only minimally, as a consequence of the major cutback in the irrigation component and the length of the preliminary period taken up in research and demonstra- tion in connection with rainfed farming. Improvements to the road network facilitated the transportation of sugarcane to the mill and opened up certain regions by giving them access to markets. Nevertheless, in the absence of an appreciable increase in agricultural yields and production, the roads have not brought the economic advantages anticipated. The economic rate of return on the project is regarded as negative. On the other hand, the impact of the water supply component on beneficiaries' quality of life has been largely positive. 18. Despite its disappointing-results, the project has laid the neces- sary basis for future agricultural development. Constraints have been identified and appropriate technical packages devised. Efficient input supply systems and farm credit facilities are in place. Seed production has been started on a mucest scale. The project has also enabled ODN to acquire -5- experience of regional development and to train its personnel. A second project (Credit 1410-HA, approved in August 1983), now under implementation, aims at using the experience gained under the first project for increpsing agricultural production in the northern region. B. MAIN FINDINGS 19. The project showed four shortcomings which OED has encountered frequently in rural development projects: insufficient project preparation, an error of judgement regarding the time needed to achieve a significant impact on agricultural production, poor interagency coordination and the lack of a comprehensive regional development plan. These weaknesses, which were only partially corrected in the course of execution, largely explain the economic setback to the project and the difficulties the follow-on project is likely to experience soon. Insufficient Project Preparation 20. The project was identicJed in June 1974, and prepared in January ad February 1975 by two FAO/CP missions. A wide range of components were initially proposed for the project, all of which were included in the preparation report, together with infrastructure for health, education and electrification. 21. The major constraints identified by the preparation mission were: (i) the low capacity and poor management of the Caldos Sugar Eill, which, in the opinion of the mission, was incapable of handling the additional sugarcane that would be grown as a result of the project; and (ii) the appreciable increase in land prices and the possibility the project would favor land speculation. The level of user water charges for the water supply component was also a matter of concern. These problems, necessitating additional surveys and extensive discussions with representative of Caldos Corporation and the Government, delayed appraisal until April 1976, despite the pressure exerted by the Government and tie Bank for early project start-up. 22. The appraisal mission rejected the education, health and electrifi- ,ation components proposed, since they were already the subject of separate financing arrangements, but it adopted the other components endorsed by the Government and the preparation mission. The Issues Paper and the Decision Memorandum of June 1976 proposed that the Government prohibit land transac- tions in order to prevent speculation. Farm credit interest rates and water charges were also discussed at length. The sugar mill problem was regarded as solved, since, in the opinion of the appraisal mission, it would have no difficulty in processing 300,000 t of cane per annum. The question of how 25,000 families living on 52,000 ha would increase their incomes two or three times was hardly discussed, except for a Central Project Staff (CPS) memoran- dum arguing that the constraints on agricultural development, the suitability of a technical package and the strategy as to how these would be alleviated -6- had not been clearly identified. CPS staff offered valid comments on the draft SAR and pointed out (i) the absence of an irrigation land suitability assessment, (ii) the need to clearly establish ODN's role and agree on its authority over, and means of coordinating line agencies, and (iii) the desir- ability of undertaking a separate ERR for irrigated and rainfed interven- tions. These key questions, however, were not satisfactorily answered when the project was presented to the Board. 23. In retrospect, the issues to which the Bank directed all its atten- tion at preparation and appraisal, despite CPS staff warnings, have proved academic and trivial, in contrast to the problems that surfaced once imple- mentation began. Bank supervision reports for the years 1977-1980 highlight many of these difficulties: land tenure was little understood (more absen- teeism than expected), -equiring a special study; maps were not detailed enough and aerial photographing of the project area was necessary (December 1977); the pre-project agricultural environment was poorly comprehended; technical packages were inadequate; seed production was largely irrelevant, since it was not clear what should be produced in the region, extension work was ineffective and agricultural credit used either very little or not at all (May 1978); work on the roads program was delayed (March 1979), as ODN and the Public Works Department were unable to agree on construction standards; the major problem at Quartier Morin was not a lack but an excess of water (March 1980); well drilling was not producing adequate returns; and sugarcane worth US$1 million was being lost owing to breakdown of the sugar mill. 24. The project experience shows the risk of insufficient preparation for a multicomponent rural development project of this nature, particularly when the project is the first Bank operation in the sector. The poor project preparation, mostly due to the Bank's and Government's pressure for quick action, was clearly the first shortcoming of the project and adversely affected both project implementation and impact.2/ Time Needed to Achieve an Impact on Agricultural Production 25. As in many other similar Bank-financed projects, this project in Haiti aimed at increasing agricultural production by two or three fold in the space of four years. Project experience confirms findings from the majority of other regional development projects: the time needed to have a signifi- cant impact on agricultural production is much longer than expected and generally requires one or more follow-up projects.3/ 2/ OPS staff points out that the project shows the need for thorough preparation of agricultural projects, especially for poor countries, and particularly in cases where technical considerations are of prime importance to the success or failure of the Bark's inverventions. 3/ See in particular the following PPARs: Kigoma Rural Development Project Tanzania (No. 4858); Bougouriba Project, Burkina (No. 4547); and PIDER I Mexico (No. 4617). -7 26. OED experience attributes this to three main factors: (i) the time needed to set up a new development institution and place it on a firm opera- tional footing is always longer than scheduled; (ii) beneficiaries, espe- cially those with the lowest income levels are risk averse and thus resist changing traditional agricultural methods; such change usually requires a series of convincing demonstrations, which may have to be extended over several crop seasons; and (iii) providing smallholders with agricultural inputs often presents problems that are difficult to solve. 27. Bank experience demonstrates that regional development projects generaily require several phases, preferably defined at the time of project preparation. An initial pilot phase, during which the techniques and methods of development are tested on modestly-sized areas, is desirable. In this regard, project experience in Haiti is typical: implementation took six years instead of four, and the results achieved at project completion were such as to require a five-year follow-up project, which should itself probably be consolidated by a third project of equal duration. A period of 16 years in which to develop a vast and densely populated region should be the rule rather than the exception. Within this time perspective, careful preparation of the project, its launching through a pilot project and development over several judiciously planned subsequent phases is obviously preferable to a series of projects hastily organized to mitigate the errors of earlier ventures. Project Organization 28. The project organization also followed the typical pattern: a new regional agency, ODN, responsible for the planning, financing and coordina- tion of project works, to be carried out by existing government line agencies; such institutional arrangement has been found in a large number of Bank-assisted rural development projects. 29. In fact, the project was not executed in the manner expected. Weakness in these line agencies led ODN to take on the role of implementing agency for the majority of works. This problem has been encountered in numerous projects and has been mentioned frequently by OED in its PPARs.4/ Experience has shown that organizing a project in this manner is successful only if the government technical departments are themselves strengthened in personnel, equipment, and financial resources to cover their operating expenses, factors which are rarely, or insufficiently, considered or allowed for during project preparation. 30. In this particular project and with the benefit of hindsight, the initial choice of project organization was based on two major misconceptions: 4/ See in particular the reports on the Paraguay Credit and Rural Development Project, (No. 4419) and the Madagascar Village Livestock and Rural Development Project (No. 5403). -8- (a) although the Miaistry of Agriculture is, generally speaking, over- staffed, it lacks qualified field personnel and commands only a limited investment budget; (b) public investment in agriculture is financed, in the main, through external aid and consists of specific, regional projects. In Haiti, it is the project authorities rather than the Ministry of Agriculture which serve as the direct executing agencies for most project works, using their own resources and personnel for the purpose rather than those of the government technical service departments. 31. The problem therefore lies not so much in the execution of projects as in their transfer to the line agencies upon completion. Experience gained under other projects has shown that such transfers are often made under unsatisfactory conditions, and that the investments themselves are frequently badly utilized and poorly maintained for want of resources and adequately qualified staff in the typical service departments.5/ 32. In this regard, it must be pointed out that ODN has performed better than many other agencies of the same type in other Bank-supported projects. The methodology developed for the transfer of responsibility between ODN and the agency responsible for water supply (SNEP) may be cited as an example. From the very beginning, SNEP was associated with the plann- ing and construction of the works, execution of which, however, remained an ODN responsibility. At the same time, efforts were made to have the local populace participate in implementation and be responsible, without remunera- tion, for providing such materials as stone, sand, and cement on site prior to commencement of the works. Before pumps were installed at well heads, the population was required to pay a lump sum, which would be allocated for main- tenance purposes. Soon after construction, SNEP was expected to take over responsibility for the operation and upkeep of water supply structures, this transfer being partly facilitated by the participation of SNEP personnel in the planning and construction phases and by the technical training specifi- cally provided for them. Water mana.gement committees were also organized in the villages. Transfer of responsibility for road maintenance has also been satisfactory. Although road maintenance is still financed by ODN, it has been entrusted to the National Highway System Maintenance Division, which is attending satisfactorily to the task. The Importance of a Regional Development Plan 33. The difficulties encountered by this ambitious and vastly complex project point to a particular lesson which has relevance to other projects of this type-namely the benefits to be obtained from the preparation, prior to 5/ This question is discussed at length in certain OED audit reports, but particularly in those on the Madagascar Village Livestock Project and the Papaloapan Project in Mexico, already mentioned above. -9- project appraisal, of a comprehensive regional/area development plan. Such plans have been put to good use in other projects related to land/area devel- opment6/ and if such a plan had been developed in Haiti, the future short- comings of this project would have been largely avoided. 34. First, the regional development plan could have overcome the poor knowledge of the project zone and its socioeconomic environment and potenti- alities. Such a coherent development plan for the northern region would have provided answers to many of the following questions: why certain courses of action were chosen or why a particular target zone in the Northern Department was chosen rather than another. Was irrigation at Quartier Morin and Saint Raphagl to be given priority over the development of piedmont zones, the anti-erosion campaign or bringing into production unused public land areas (numerous in the Department)? Should such infrastructure components as roads and water supply facilities have been integrated with agricultural develop- ment areas or located randomly? What is the Northern Deiiartment's true agri- cultural vocation and which crops would best serve the national and regional economy? All these questions have been, and still are, answered only partially, all for want of a coherent development plan for the northern region. 35. Second, it would have provided a framework by which the optimal timing and phasing of the project could have been accomplished. Instead, it is an example of the effects of Bank pressure for rapid project launching, the first in the agriculture sector in Haiti, despite scant Bank acquaintance with the country in general and the sector in particular. A few months preparation period was manifestly not long enough to allow the Bank to familiarize itself with the social, ecological and technological environment of the northern region. In addition, project preparation was carried out by FAO/CP and Bank short missions, which paid more attention to issues of minor importance than to constraints to agricultural development and ways to alleviate these constraints soon after project start-up. The project tried to mitigate the consequences of this ignorance as best it could by improvis- ing a series of studies and research efforts which, after six years, brought the project up to the stage at which it should normally have begun. 36. Third, the lack of a master regional development plan was also responsible in part for the failure to coordinate the various actions under- taken by the departments of ODN. Having no guidelines and no overall view of the project to follow, each worked in isolation from the others, concentrat- ing on tasks in its own sphere of specialization that could be carried out in the field effectively and rapidly. The problem was aggravated by both the lack of continuity in the directorship of ODN and its failure, in any case, to provide actual direction. It was not until 1984, eight years after the 6/ See for example, PPAR on Malaysia Jengka Triangle Projects (OED Reports Nos. 2122, 3024, and 5978) and Malaysia Keratong Settlement Project (under preparation). - 10 - first project was launched, that any coherent management strategy was devel- oped. This eventual success can be ascribed to a growing knowledge among ODN personnel of the region itself and new management that is aware of the prob- lem and attempts to address it. 37. The experience of regional development the Bank has now accumulated shows that it ought to have begun in this instance in Haiti with a phase allowing for meticulous preparation by a well-qualified team stationed in the zone for at least two years. This preparatory phase could have been financed by the Bank (Project Preparation Facility or a technical assistance loan), or through bilateral or international aid (UNDP/FAO, for example), and thus at a lower cost to the Government. The preparation work, encompassing the Northern Department as a whole, would have involved: (a) a study of the pre-project socioeconomic and technical environment, and identification of the constraints on development;7/ (b) a study of the potential resources in the northern region, and definition of project objectives; (c) identification of ways and means available for circumventing the constraints, and pinpointing of priority actions and target zones; (d) formulation of a long-term master development plan; se) design of a pilot project to be carried out as a first phase. There is no doubt that such preparatory work would have made it possible to provided more thorough knowledge of the population8/ the region and allowed more account to be taken of local aspirations in planning the development of the Department. 7/ An in-depth study of the pre-project environment would have revealed the lack of technical packages on which to base the project. Applied research could also have been undertaken during this preparatory phase. 8/ OED has made the same observation regarding a number of rural development projects, in particular the Village Livestock and Rural Development Project in Madagascar (OED report no. 5403), an instance .n which it was decided at the beginning of the follow-up project to undertake a socioeconomic study of the country's northwest region, the Bank having realized at the end of the first project that very little was known of the attitudes and motivation of the livestock farmers themselves. In another case, the Papaloapan Rural Development Project in Mexico (OED report in preparation), where a general study of the region was decided on and its cost incorporated in the Bank loan only at the end of the project, when it became apparent that knowledge of the region's potential was meager. - Li - Outlook for the Future 38. The project moved into its second phase in 1984. First-phase results have proved minor in economic terms but fruitful in terms of exper- ience acquired. ODN now knows what it must do in the sphere of rainfed agri- culture and how to do it. It only remains for the agency to move resolutely ahead with the production phase by making the necessary inputs (seed, plants, tools, fertilizer) available to smallholders and to disseminate among them a knowledge of the recommended technical packages tested during the first phase. 39. The success of the follow-on project will depend mainly on the quality of ODN management and on the qualifications, motivation and continu- ity of its staff. In the past, ODN has been severely affected by successive changes in its directorship (five directors in eight years) and by the short- age of suitably qualified Haitian nationals. The situation has now improved appreciably, thanks to a better and more effective management and the continued presence on staff of a group of competent Haitian professionals. 40. The project still suffers, however, from the lack of a master plan (paras. 33-34) and the uncertainty which still pervades certain major issues in the development of the region. The problems of irrigation and drainage of the Northern Plain, for instance, remain as pressing as ever and no clear solution has been found. The tracts of fertile government land on the Northern Plain are still poorly utilized and no action is planned in this regard.9/ Anti-erosion measures, which appear to be essential in the Northern Department, constitute no more than a marginal activity. 41. There is no doubt that ODN's activities are useful. Nevertheless, until they are incorporated into a general framework, one may ask whether they are first-priority activities or not. That, indeed, is the main handi- cap which has weighed on the project throughout, and continues to do so. 9/ The Region comments that the Second Rural Development Project (Cr. 1410-HA) is now supporting better utilization of this area, through the provision of technical assistance and credit. v~~~~~ b~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ..1 - 13 - PROJECT CUMPLkTlUEN REPUKT hlAITI CREDIT b7S-iiA RUStL UEVELUPNENT 1N TIhE NORTHERN UEPARTMENT May 31, 1984 E.rojects Department Latin America and the Caribbean Kegional Uffice . 1. t~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ I, I \ I l I * -15 - I. INTRODUCTION 1.01 Haiti is the poorest country of the Americas, with high population density and limi'ed resources, especially land resources. Agriculture, Haiti's most important sector, has shown little progress over the last two decades. It has been handicapped by major institutional constraints, land tenure insecurity, and, at the time of appraisal, the inadequacy of the road and irrigation network. 1.02 In the mid-1970s the Government's strategy for rural development emphasized the rehabilitation and extension of the existing transport and irrigation infrastructure. It also included an increase in crop production, particularly export crops (coffee and sugar), through increased agricultural credit and improved extension services. However, because of constraints encountered in the sector, the Government realized that nationwide efforts would not be effective. Therefore, it decided to concentrate its IDA- financed rural development efforts in the Northern Department of the country. II. BACKGROUND 2.01 The Rural Development Project in the Northern Department was the first IDA-financed project in the agricultural sector, after IDA and other donors resumed their activities which had been suspended since 1962 for political reasons. An FAO/IBRD/IDA mission reviewed the sector in August-October 1973, and proposed nine projects for priority consideration. Out of these, IDB selected irrigation projects in the Artibonite and Cul-de-Sac plains, and USAID decided to support coffee development. 2.02 This report is based on a review of records kept by the LAC Information Center (i.e., correspondence, internal memoranda, the Appraisal Report of December 15, 1976, the Credit Agreement of January 14, 1977, and supervision reports). A substantial portion of the information was provided by the Interim Review Report issued December 29, 1981. III. PROJECT FORMULATION 3.01 Because of the good agricultural potential of the Northern Plain, and considering that IDA was financing the main road from Port-au-Pri.ce to the North, Government officials proposed the selection of the North for an IDA-supported project. An FAO/CP identification mission visited the country in June 1974. An FAO/CP preparation mission visited the project area in January-February 1975. The mission identified two major issues: (a) the weakness of the regional services that were to implement project actions, and (b) the need to increase the capacity of the Citadelle Sugar Mill (in Caldos) to handle the planned increase in sugarcane production. The adequacy of sugarcane processing facilities was the main topic of discussion with the Government during the pre-appraisal period, and commitment by the Government to increase the technical and financial capacity of the mill to process the increased cane output was a condition of appraisal. Two pre-appraisal missions took place (October 1975 and January 1976) primarily to review sugarcane processing issues. - 16 - 3.02 Project appraisal was carried out in the field in April-May 1976. During the appraisal review process, CPS recommended close monitoring of the Agricultural Credit Bureau (BCA) and the central services of the Ministry of Agriculture during project implementation. 3.03 At Board presentation (December 28, 1976), the Board was concerned that: (a) the US$2 million out of the US$10 million IDA credit, which had been provided for technical assistance, should remain in the credit if bilateral agencies agreed to finance this technical assistance component, considering that additional funds could be needed in view of the country's poor financial situation and possible cost overruns; (b) the project might facilitate the accumulation of land in the hands of few larger farmers; (c) small-scale farmers might not be able to pay user fees for irrigation; and (d) the high level of expatriate technical assistance would raise the foreign exchange cost of the project. IV. THE PROJECT A. Project Objectives and Description 4.01 The project was intended to remove key constraints to development in the Northern Department, especially the Northern Plain, one of the areas with high agricultural potential in the country. In particular, the project objectives were to improve agricultural production and rural incomes, improve the quality of rural life, and help build up regional icstitutions. With some of the constraints tackled under programs financed by other agencies (health, education, soil conservation), the project was designed to finance basic infrastructure (rural roads, water supply and markets) and agricultural services (irrigation, credit, applied research and extension, animal health) for rural development. In order to accomplish the objectives, the project was designed to: (a) rehabilitate the Quartier Morin (1,200 ha) and St. Raphael (1,900 ha) irrigation systems, expand these systems by 800 ha at Quartier Morin, and irrigate up to 1,000 ha based on deep wells north of Quartier Morin; (b) rehabilitate and improve 66.5 km of feeder roads and 27 km of farm access roads; (c) construct up to 400 wells to be operated with hand pumps and up to five small water supply systems served by springs, for the rural potable water supply; (d) rehabilitate and improve the two municipal markets and abattoir in Cap Haitien, and build a market in St. Raphael; Ce) provide short-term and medium-term credit to groups of small-scale farmers and to individual medium-scale farmers; (f) initiate agricultural trials and improved seed multiplication; (g) provide agricultural extension and veterinary services, and organize rural-community development; and - 17 - (h) create the Organisme de Developpement du Nord (ODN) to implement the project under the supervision of a Board of Directors chaired by the Secretary of Agriculture. 4.02 About 4,000 rural families were to benefit from the directly productive agricultural components of the project. In addition, some 26,000 rural families were expected to benefit from better roads, improved agricultural services and availability of improved seeds, and about 3,000 rural households were to receive potable water. B. Project Costs and Financing 4.03 Total project costs were estimated at US$13.4 million, with a foreign exchange component of US$8.2 million. Over one-third of the cost was for irrigation, one-fourth for project administration, and 14% for roads. IDA was to finance US$10 million, or about 75%, of total project costs; the Government was to contribute US$2 million, or almost 15%; and the World Food Program was to provide US$1.4 million, or about 10%. The Government had requested bilateral assistance from the Federal Republic of Germany and France to finance the cost of resident consultants, and if this assistance were forthcoming, IDA would cancel from the credit an amount equivalent to the value of the assistance obtained. V. IMPLEENTATION A. Key Events 5.01 Tho Credit Agreement was signed on January 14, 1977, and the Credit became effective on May 26, 1977. Before Credit effectiveness, the French and German aid agencies agreed to finance fully the technical assistance components. Project implementation started slowly, and soon after effectiveness, the lack of counterpart funds readily available to prefinance project expenditures and slow authorization of expenditures bv Government services emerged as main constraints. In early 1979, project costs were revised, following full assumption of technical assistance costs by France and Germany. IDA and the Government agreed to retain the US$2 million allocated for technical assistance, in view of possible further cost increases and for contingency purposes. 5.02 The French technical assistance consultants in agriculture were fielded promptly in May-June 1977, but the three groups of German-financed engineers did not arrive until May 1978, July 1978 and January 1979. In early 1979, the project was declared a problem project, on account of lack of counterpart funds, delays in the approval of expenditures, lack of counterpart personnel, and the resulting slow disbursement. In order to improve the financial situation, IDA and the Government agreed that the Government would establish a revolving fund in the Central Bank. In October 1980 an advance of US$500,000 was made from the IDA credit to the revolving fund, following streamlining of the expenditure authorization process and allocation of counterpart funds to the fund. 5.03 The supervision mission of January 1981 recommended an extension of the project completion date from July 1981 to September 1982 because of the project's slow start. An amendment to the Credit Agreement was signed - 18 - in August 1981 to this effect; it also provided for the IDA advance to the revolving fund (para. 5.02), reduction in the scope of the irrigation program, expansion of the water supply and agricultural research programs, and a revised financing plan. 5.04 In September 1981 IDA carried out an interim review of project performance including changes up to that time in implementation and their implications. The review concluded that the changes were justified. By early 1982, progress was such that the project was removed from the problem projects list. 5.05 The project began winding down its activities in mid-1982. In April 1983, a reallocation had been proposed for the use of the remaining uncommitted funds. Project activities were then drastically reduced because of the termination of the technical assistance and severe Government budget constraints. However, ODN and the participating agencies continued to function, without disintegration as had been feared. The Credit was closed on September 30, 1983. B. Implementation Problems 5.06 From the outset, the project faced severe implementation problems that can be placed in four main categories: lack of basic information available to the appraisal mission, institutional weaknesses, financial problems, and the collapse of sugar prices. Lack of Basic Information 5.07 The preparation and appraisal missions found limited information on the agricultural environment, and almost no solid information on the institutional and socio-economic environment, because available documents, studies and surveys were incomplete or non-existent. During the design phase of the project, the country was just emerging from a long period of isolation, and the complex socio-economic fabric of the rural areas was not fully understood. Thus, it was not until implementation got underway that a host of unforeseen problems--mostly institutional--surfaced, and project management had to learn by doing. 5.08 To supplement the meager data that were available, the project undertook a number of unscheduled studies with consultant support for a total of US$269,000. These studies included a survey of cropping patterns and population distribution by air-photo interpretation, farm surveys, studies of sugarcane production and milling, climatological surveys, a review of drainage problems, soil and irrigation surveys, topographic surveys, and a hydrogeological study. As a result of tis flexibility during implementation, the project area is now better understood, and ad hoc changes to project design were made as new facts emerged. Institutional Weaknesses 5.09 The main Government agencies scheduled to carry out project activities had only limited or no implementing capability. This wAs especially true of agencies dealing with civil engineering (irrigation and water supply). In the case of roads, construction under the project did not - 19 - start until suitable arrangements had been made between ODN and the Public Works Ministry, which took two years. Agencies also suffered from lack of qualified engineers and technicians (foremen, surveyors, draftsmen, and mechanics), and some- of the consultants who were to reinforce these agencies arrived late. This situation forced ODN to assume implementation responsibilities. Financial Problems 5.10 ODN was created as an autonomous agency, but for nearly four years it did not have the cash resources necessary for project implementation. The problem resulted from several factors: (a) because of budgetary constraints, considerable delays occurred in the allocation of counterpart funds; (b) the amounts allocated were insufficient to prefinance project activities; and (c) the authorization of expenditures required lengthy procedures in Port-au-Prince. As a result, necessary purchases could not be made on time, suppliers refused to deal with the project, tenders had few takers, and disbursements from the IDA credit fell well behind schedule. 5.11 The financial constraints were eventually alleviated: in early 1980 the monthly allocation of counterpart funds increased; in 1981 a revolving fund was established in the Central Bank for the Government's contribution; and the Credit Agreement was amended to provide an IDA advance of US$500,000 to this fund. Collapse of Sugar Prices 5.12 Because an increase of sugarcane output (80,000 t) was expected, a main concern at the preparation stage wva the milling capacity in the area, especially the capacity of the privately owned Citadelle Sugar Mill and the organization of cane collection. The sugarcane situation was reviewed at that time by a consultant. During implementation, the complexity of the situation and persistent problems prompted ODN to engage consultants for further reviews. 5.13 The agronomic review and a review of the climatic environment by an agroclimatologist led to the conclusion that investments in irrigation improvement would not be justified for sugarcane production. Accordingly, it was agreed to give priority to irrigation in St. Raphael, outside the sugarcane area. It was also concluded that no improvement of cane production in the northern plain was possible until variety trials were carried out. After several years of trials, high yielding cane varieties have now been identified and are being propagated for renovation under the follow-up project. 5.14 The combination of depressed sugar prices in the international market, breakdowns and inadequate management at the Citadelle Mill, inadequate roads, insufficient transportation and floods, resulted in a substantial proportion of the cane crop not being harvested, with resulting financial losses for growers. The unharvested sugarcane amounted to over 100,000 tons per year, equivalent to about US$1 million per year. This situation kept alive the issue of cane collection and milling, and a great deal of discussion with the Government focused on this issue. Compliance with a covenant, Section 4.09 of the Credit Agreement), whereby the Government was to ensure adequate milling capacity to process the increased - 20 - production from the project, was alse a main topic of discussion. Compliance with the covenant was made a condition of extension of the credit and processing of a follow-up project. 5.15 This pressure--combined with project actions, purchase of the mill by a Government bank, and its rehabilitation under a management contract--has led to the start of rehabilitation of the sugarcane sector and favorable prospects for increased output. A beginning was made in 1983 with the renovation of cane fields. C. Project Components 5.16 As a result of the above-mentioned problems, certain elements of the project were redesigned. The main change in strategy was to forego irrigation as a priority item. This shift implied that the main increase in agriculture output would come from rainfed crop improvement through research, extension, input supply and credit. Another shift in strategy was the priority given to developing information and technical packages, without which production development could not progress. Such packages began to be available for the main dryland crops in 1980/81. As a result, directly productive actions started on a modest scale. Such actions were based primarily on the promotion of high yielding varieties selected and tested under the field trial program. Specifically, the project was carried out as follows (also see Key Indicators in Annex 1): Irrigation 5.17 ODN and IDA decided to confine the irrigation component under the project to St. Raphael instead of Quartier Morin, as irrigation is better justified at St. Raphael because of the relatively dry climate. The program at Quartier Morin was postponed indefinitely because of: (a) topographic data and design had to be prepared with the help of expatriate staff before work could start on the rehabilitation and expansion of irrigation systems; (b) data on groundwater were needed before starting construction of irrigation wells; with the increase in energy costs, it was estimated that the investment and operating costs of well irrigation would not be financially justified; (c, drainage was found to be as important as irrigation in Quartier Morin, and preliminary requirements for drainage work were studied; (d) it was determined by a meteorology/hydrology consultant in 1977 that investment in sugarcane irrigation was not justified in the Quartier Morin area; and (e) the diversion dam for the Quartier Morin ckmmand area was found to be in need of rehabilitation as the result of iloods and a section of the spillway was carried away in 1981 before repairs could begin; this precluded irrigation in Quartier Morin until the dam was repaired and the economic justification ascertained. - 21 - J.1d Thus, the total irrigation program was reduced from 4,90U to about 75U ha. because or ditticulties in organizinig eftective irrigation at the rarm level, after construction of the secondary system, U'N and IUA agreed in 1YbU to engage an irrigation agronomist. Out of the 75U ha that could be irrigated at bt. Kaphael, UDN was able to reliabilitate only 4oU ha by the end of the aisbursemenc period. The balance ot the area is to be rehabiLitated under the follow-up project. Irrigation Program trom Wells 5.19 This program was deleted trom, the project because of implementation ditficulties, lack ot knowledge of groundwater supply, high costs of pumping and lack of technical staff to easure U&1l. A hydrugeological survey to identity the groundwater resources available (which would be a prerequisite tor drilling for irrigation or other purposes) was, however, rinanced under tne project tor the entire Northern Plain. Water bupply 5.ZU because ot the absence of potable water, low cost per family, good identitication work, and good experience in implementation, this program was expanded. Instead of 4UU wells and tive coimnunity systems to serve a population ot Z,500, the revised program called for 15U wells and Z community systems tor a population of about !UU,JUU. The program was extended beyond the original project area and Ethe tredit Agreement was amended accordingly in 19b1. An expatriate works supervisor was added to the staft in 1961. Koads 5.21 The appraisal target of about 9J km of feeder and farm access roads was increased to lUo km; 97.Z Km were ac;ually built. however, financing remained the same as estimated at appraisal because the Ministry of Transportation dia not keep records which would enaDle it to establish actual costs. Also, because or the slowness ot construction by the Public Works Department, UDN started its owII road program in early 19bl, but unit costs were not determined. Agricultural Trials and Seed Production 5.ZZ tarly project experience pointed to the urgert need tor a trial component going beyond the scope suggested at appraisal (trials by the extension service on tarmers' rields). DUN and LUA agreed thiat a comprehensive program of controlled trials (pcimarily improved varieties) on the major crops was required to develop the technical packages needed for effective extension and credit efforts. The trial program was expanded to cover the entire Northern Department. rhe componenc incLuded the development ot several tield experiment centers, with a main experimental farm of regional scope at Lhe tormer facilities of the Canadian-sponsored Rural Development Center ot iilot (CKUIDk.l). The farm was also developed as a main seed production center. A total of 12U trials were carried out on the major crops of the area. - 22 - Agricultural Credit 5.23 The demand for agricultural credit was limited because farms are small, incomes are low and inputs are scarce. The experience gained with project implementation led to a reduction in the estimates for the volume of credit (short- and medium-term) from the already modest US$800,000 to US$600,000. Actual disbursement for credit was only US$382,000. This downward adjustment was prompted by the realization that the credit demand was even more limited than anticipated because of: (a) lack of proven technical packages; tb) the decision by ODN to concentrate credit on individuals rather than groups; and (c) the limited availability of seed, which is the input in highest demand. The costs of subloan administration were excessively high, mainly because of the small size of short-term subloans (average of about US$25 equivalent). Extension Services 5.24 There was no major change in appraisal objectives, but the impact of this component was limited for lack of credible extension messages. Sugarcane improvement was to be an important objective of the extension program, but no action could be undertaken for lack of a suitable technical package. Markets and Abattoir 5.25 The central market of Cap Haitien was rehabilitated, but work required was much more extensive and costly than anticipated at appraisal. ODN and IDA decided not to schedule rehabilitation of the other two markets and the abattoir, because of the keakness of the municipal institutions at Cap Haitien and the failure of authorities to make land available. Project Administration 5.26 Construction of the ODN office and housing suffered delays and the costs were higher than anticipated. Expenditures for project administration were also higher than anticipated. Permanent consultants were financed by bilateral assistance, but in order to respond to needs identified during implementation, additional consultants were employed (an irrigation agronomrist, a civil works supervisor, and an automechanic). Although ODN was conceived at appraisal as an agency which would coordinate the activities of line agencies and channel funds to them for project actions, it actually operated as an executing agency since some of the line agencies were found co be too weak in the project area (para. 5.09). Some of the regional capability was strengthened under the project (water supply and field trials). D. Changes in Project Financing and Costs 5.27 Along with changes in the project scope and orientation, there were two major revisions of financing and cost plans, with resulting reallocation of the proceeds of the credit (Annex 2). The first was in early 1979, following completion of recruitment of the permanent consultants with bilateral, rather than IDA, financing (para. 5.01). The second major - 23 - revision was in early 1981, after a joint meeting with all donor agencies, following a review of development needs and prospects for implementing the various components. 5.28 A comparison of the actual project costs (US$15 million) with appraisal estimates (US$13.4 million) should take into account that: (a) actual expenditures do not include the salaries of permanent expatriate consultants, estimated at about US$2.5 million between 1977 and 1982, or those of permanent Government staff estimated at US$1.6 million; (b) irrigation expenditures were drastically reduced (from US$1.9 million to US$828,000); (c) credit voltme was reduced by 25%; (d) project administration costs increased substantially, from US$2.2 million to USS 6.8 million. Besides higher staffing and operating costs, an important reason for the increase was consolidation of some project accounts in the administration, compared to the costing at appraisal whereby individual components were accounted for separately; (e) expenditures for water supply and field trials increased as a result of expansion of these programs beyond appraisal targets; and (f) market rehabilitation costs increased. E. Disbursements 5.29 The disbursement rate provei to be considerably slower than anticipated (Annex 2 and Basic Data Sheet). It remained well behind targets during the first four years, but increased markedly in 1981, reflecting progress in implementation and adjustments in design and budgeting. Three major reasons for the initial slow disbursements were: (a) the drastic reduction of the irrigation component, which represented one-third of total costs at appraisal; (b) slippage in the initiation of road works (14% of project costs); and (c) inadequate pre-financing and lengthy procedures for the authorization of expenditures. F. Procurement 5.30 Procurement remained a weak point throughout most of the implementation period, mainly on account of lack of familiarity of ODN staff with acceptable procedures. Thus, procurement matters took a great deal of the supervision time and correspondence between IDA and ODN. Virtually all biddlng documents and evaluation reports proved to be unacceptable and had to be revised or clarified. The main procurement weaknesses were: (a) application of criteria for bidder selection (these were not spelled out in the documents for ICB); (b) the small list of suppliers invited to give quotations for local - 24 - (c) uncritical acceptance of bids for civil works and poor supervision of construction. These weaknesses were substantially corrected by close supervision by IDA. G. Compliance with Credit Covenants 5.31 The Borrower failed to comply with two covenants: (a) A major concern at preparation and appraisal was reflected in the Credit Agreement by a clause whereby the Government would take steps to ensure an adequate milling capacity and an effective system of cane collection, in order to process the incremental production brought about by the project (para. 3.01). In fact there was no increase in production, but milling and collection proved inadequate for the existing crop. A great deal of supervision time was expended on this issue, and several corsultant missions were mounted to analyze the complex situation. By the end of the implementation period, the Government complied with this covenant by nationalizing the mill and signing a management contract with a qualified firm; and (b) A covenant prohibiting land sales in the irrigation areas to be rehabilitated was not complied with, and was, in fact, nearly impossible to enforce. 5.32 Covenants on the reinforcement of the BCA and interest rates for agricultural credit were complied with (with some delay), due to the support of USAID, which provided technical assistance to the credit agency. VI. AGRICULTURAL AND ECONOMIC IMPACT A. Agricultural Impact 6.01 It was anticipated at appraisal that the increased agricultural output would come from 3,870 ha (1,800 farms) in the rainfed area, 3,200 ha (2,200 farms) in the irrigated districts, and another 100 medium-size farms. Yields were expected to increase by 25-100%, mostly as the result of irrigation. Sugarcane was to account for a large part of the increase in output. 6.02 However, because of the drastic reduction in the irrigation program, and the slow pace of implementing the rehabilitation on the remaining 460 ha, there was a limited increase in output from improved or expanded irrigation. A small increase cau be credited to support services (research, extension, credit, input supply) for crops grown both in rainfed areas and in areas already under irrigation. The monitoring and evaluation unit of ODN has been unable to quantify this. At St. Raphael, the improvements in irrigation (on 460 ha) and support services have resulted in a sizable, but unquantified, increase in vegetable output (onions, carrots, and tomatoes); most of the vegetables are shipped to Port-au-Prince. The - 25 - agricultural component of the project reached about 400 farm families, or lOX of appraisal estimates, and irrigation at St. Raphael benefitted some 450 farmers. 6.03 Until 1980-81 there was no proven technical recommendation available to generate any increase in output except for vegetables. The research program has led to the selection of higher yielding varieties for cassava, rice, maize, sugarcane and beans, and seed multiplication has started on a small scale. Yields achieved under the research program suggest that substantial increases are possible at the farm level (e.g., from 2 to 3 t/ha for irrigated rice, 5 to 12 t/ha for cassava, and 35 to 60 t/ha for sugarcane). 6.04 The road network rehabilitated by the project contributed to the rehabilitation of the sugarcane sector which began in 1982/83. Before that time, the lack of passable roads in the area severed by the Citadelle Mill was one of the constraints which had accounted for the deterioration of the sector. rhe improved road network also had a significant impact on traffic increase and movement of agricultural output to Cap Haitian and Port-au-Prince. ODN carried out a traffic survey in December 1983 on two of the roads improved under the project, and compared it with a survey made on those roads in 1980. The average daily traffic (counted during one week) increased as follows: Average No. of Vehicles X Increase 1980 1983 Plaine du Nord-Duty 28.8 101.5 252 Haut du Cap-Lory 30.4 70.3 131 Limonade-Bois de Lance 2.0 62.0 3,100 The increase was mostly in rural buses, which carry both people and produce and assure regularly scheduled services instead of the itinerant bush taxi service. Agricultural output is carried by bus or truck to major centers, such as Cap Haitien, Dondon, Duty and Acul du Nord. Another benefit was that pedestrian and pack animal traffic were made possible even after rains. B. Economic Impact 6.05 The economic rate of return for this project, which is the first stage of an ongoing rural development effort in the North, is considered to be negative. This assessment is based on a qualitative evaluation of the benefits of the individual components as indicated by the physical and financial progress made in each of them, discussions with project staff and farmers and by the surveys carried out during the project (para. 5.08). 6.06 The improvement of living standards of the rural population in the project area is a long-term process, with the project constituting an initial time-slice effort in this endeavor. Because of the complexity of the local environment, the paucity of pertinent information and the lack of effective regional institutional investments, this project focused on institution - 26 - building (administration), training through technical assistance studies, and on applied research and experimentation. The benefits of these investments are by nature medium- and long-term, and will materialize only in conjunction with follow-up actions to directly improve agricultural production and the standard of living of the rural population. Furthermore, the effectiveness of the directly productive investments under the project has been significantly impaired by the absence or inadequacy of suitable regional development institutions and technical, administrative and operational expertise. Irrigation 6.07 Quartier Morin. Compared to appraisal estimates, no significant investments were made under this componeut, except for some administrative and reconnaissance expenditures and cleaning of canals using food for work support. Eventually the Government and IDA agreed to discontinue this component. No benefits accrued from this component for lack of incremental production. The estimate is that these small investments yielded a zero or slightly negative rate of return. Because of the small size of the investments, the impact on the overall rate of return of the project is negligible. 6.08 St. Raphael. Irrigation water constraints limited the area of potential development to 750 ha. Of this, 460 ha have been developed. Farmers have begun to grow two crops a year (rice and vegetables) resulting in substantial increases in cropping intensity and production. Agricultural Credit 6.09 Individual beneficiaries of this component received short-term credit (total of US$137,000) for inputs procured by ODN, as ODN made it a policy to provide credit in kind. Medium-term credit (total of US$115,000) was available for minor farm equipment (pumps, plows) or small processing facilities (sugarcane, cassava). At the farm level, the component can be credited with a marginally positive financial return. On economic grounds, the rate of return may be considered negative because of the high cost of loan processing and administration at ODN (which required annual budgetary transfers from the Treasury). Field Trials, Extension, Animal Health and Markets 6.10 Field trials, agricultural extension, animal health and the rehabilitation of the municipal market at Cap Haitien had the potential fer having a substantial impact. However, their economic benefits were not quantified because of the methodological difficulties of assessing the related benefits and the lack of suitable data. Rural Roads 6.11 This component is estimated to have a rate of return above the opportunity cost of capital, judging from traffic counts and other benefits (para. 6.04). - 27 - Water Supply N.IL lte component is considered to have a high return, as the average investment cost is only some US*25 per beneficiary. Again, a quantitative evaluation has not been possible oecause no suitable data are available to quantify benefits. Administration N.Ui Alchough no direct benefits can be attributed to this component, its pertormance and impact have been assessed. uf total project costs, 4UX were incurred under the heading "administration, excluding the costs of expatriate consuLtants. This means high administrative icosts, which reflect institutional weaknesses. htowever, this has to be viewed in light of the fact that UuN, with strong consultant support, carried out the various project works as a de facto implementing agency, with some af the project costs charged to administration that were actually incurred tor development. UDN is now a viable regional development agency with trained personnel, adequate accounting and financial autonomy. The model is being replicated to promote development in other parts of the country; however, the continuing viability ut UUN depends on sustained expatriate consultant support and Government commitment. VII. INSTITUTIUNAL PYK.rUKMANCk .. UJN and Kegional Agencies 7.uI rhe agricultural sector was institutionally weak, whi-cb led the Government to create UOlN to coordinate project implementation, control funding and monitor relevant regional services. 7.uz The weakness or local services forced U=4 to take over project implementation, except for the road program. The road program was nevertheless delayed while arrangements were worked out between UUN and the ilinistry of Public Works (TPTC) on the definition of their respective responsibilities. In che case of agricultural credit, a separate service was created for the project within UbN, as SCA declined to participate in the project. 7.U3 Project implementation with the support ot consultants has presented an opportunity to organize and/or strengthen regional services (research, extension, seed productioni, irrigacion and water supply maintenance and credit) and to create a viable regional development institution. However, OUN and the line agencies strengthened under the project are fragile institutions, as they remain subject to sudden changes in personnel, budget cuts or reorganization of services. 7.U4 while the DudgeEtor overseas training was not fully used, a research agronomist and a water supply engineer received international training and several visits were also organized to the Dominican Republic to gather agronomic information. ln-service training organized for water supply, research and extension staff resulted in upgrading of professional and specialized labor (masons, plumbers and surveyors). - 28 - Auditiig and Reporting 7.uj The accounts ot UUN were audited each year. The performance of the tirst auditing firm was deemed inadequate and a new firm was engaged for the second year; that audit in turn was deemed inadequate and a new firm-- Peat-tarwick & Mitchell, which has a branch office in Haiti--was selected. besides the regular audit, the firm was requested to provide technical assistance on the improvement of the project accounting system. This resulted in the establishment of accounts for each activity. Further assistance on the reorganization of the accounting system was provided after project completion (with retroactive tinancing under the tollow-up project). Separate accounts were kept tor UDl (administration and components other than roads and agricultural credic), the Ministry of Public Works and tarm credit within UUN. In the case of the credit, the accounts were poorly kept and write-orfs had to be identified to reconcile aggregate portfolio information with updated subsidiary accounts, and then write off long overdues, prior to Lranster by UUN of the portfolio to BCa under the follow-up project. 7.ub Ut" submitted annual and quarterly reports, but the quality of reporting was poor (except for the report for 19dZ/bJ), in spite of comments anid suggestions reiteraced by project officers. The annual reports were usually verbose; they lacked quantitative data and clear analyses of issues. The quarLerly reports were prepared by the manager of each component, but IUIN failed to provide an analysis and a management overview or implementation and issues. The less-than-adequate quality of reporting reflected the turnover of ODN directors and the marginal performance of the deputy director. rMonitoring and Lvaluation 7.U7 Monitoring and evaluation (M&E) or project activities was organized by a consultant tinanced by the French Aid Agency (EAC) for two years (1979 ana 19bu). The unit reported to the uirector or UWN, but received technical support from the 6bh service of the Ministry of Agriculture. based on data collectea, the unit identified proolems and contributed to changes ii: the scope of the project during implemencation (para. 5.27). After the consultant's departure, the unit continued, but at a slower pace and with a decrease in quality. ne unit tocused on monitoring and did little with respect to evaluation of the impact of the project. 7.ub In 19bJ tne monitoring and evaluation team supported by a French-tinanced consultant carried out a baseline survey under the follow-up project. The survey, covering bOU farms, included questions on the impact or the first project. between ZUX and 50X of the respondents, depending on the area, stated that they had been concacted by the extension service; the proportion was higher in villages and along the roads. Road improvement res,lted in a reduction of transportation costs (but this decrease was offset by an increase in petroleum prices), ana in saving transport time. About 2UX of the respondents received better prices for their crops, ano bO7 stated chat they increased quantities sold, especially perishables. With respect to water supply, most respondents indicated that the new systems made it possible to save time and provided better water, resulting In lower incidenice ot disease. - 29 - B. Central Government Institutions 7.u9 Despite a covenant providing tor consultation with 1UA on strengthening the Hinistry, there has been little evidence of improvement, in spite of ettorts by bilateral agencies through tunding and technical assistance. In tact, the Presidenlt deplored the weakness of the Ministry ot Agriculture (UAtilm) in a speech in January 1963. Also, some donor agencies (USAIU, Clk)J have abandonied or scated clOWLI their cooperation with DIAKhDk. The main problems remain lack ot adeqLaL.: budgeting, managerial continuity and well-defined programas. 7.1U Government authorities made a concerted and successful effort to streamline procedures and provide UUN with sutficient autonomy. A marked improvement of the managerial anid tiinancial performance of BCA has taken place at tne national level. however, this improvement is independent of the project actions; it has resulted from technical assistance provided by USA1l. C. Consultants 7.11 Un the whole, long-term consultants (administration, agronomy, credit/marketing, irrigation, roads, water supply, irritation agionomy and automechanics) and short-term consultanEs made a positive contribution. Most outstanding was the contribution ot the water supply consultant, followed by that ot the agronomy consultant. As a result, the water supply and agronomic trial programs were expanded, and viable services were organized. Un the other hana, the performance of the deputy director general (administration consultant) was marginal; however, he was not replaced because he provided conitinuity, considering that OLN saw a succession of four directors during the implementation perioa. The automechanics and irrigation consultants were ineffective because of poor relations with counterparts. Vill. LDA PLKFUKMANch n.Ul T'he hosc ot problems that emerged curing implementation indicate that the institutional, technical and sociological problems were not sufficiently dealt with during the project design stage. Some of the targets coula not be met, and input requirements (staff, investments, operating costs) were higher than anticipated. The project designers were faced with a lack of reliable information on the complex socioeconomic environment and institutional capability. The situation became better known as the result of the studies carried out during implementation. Starting project implementation was perhaps the only way to gain a sufficient knowledge of the area; it took time for Uuh and supervision missions to apprehend the problems and react accordingly. 6.UL Vuring implementation, IUA missions made frequent visits to the project area and worked closely 4ith uLUN on project reformulation as needed to keep the project afloat. Supervision thus involved a great deal of technical assistance to project management, and interventions with Government ministries for steps needed to ensure improved implementation. Supervision missions also tocused heavily on sugarcane and the performance of BCA. A positive point of the supervision process was the continuity ot staffing. - 30 - U.U3 Foor project pertormance has been a characteristic of non-IUA- rinanced projects in the agriculLural sector in haiti. The two projects in the country that have achieved some success (lDb-supported Artibonite Irrigation ProjeCt and Uerman-supportea (onaive irrigation Project) have been carriea out under ravorable conditions not replicable elsewhere in the country. Tlwo regional development projects (supported by USAII) and CiLA) pertormed poorly. 1x. LkbbUNb LkAKrkL 9.Ul Considering the complexity ot the agricultural sector, lack ot sutficient Knowledge on socioeconiomic aspects, and the tact that the project was the tirst IUA involvement in agriculture in haiti, the project proved to be a good learning experience. (a) The major lessun nas been the need for flexibility and active participation in adjustment and redesign by supervision missions. Without substantial changes in organization and scope ot components, the project woula have experienced greater ditficulties. (b) At the same tine, a critical review or issues Dy supervision missioiis waz required before changes were maue, considering that UDh occasionally proposed changes thac could not be justified. (c) Projects of tniis type in haiti require a commitment of sutficient and continuing local manpower to ensure frequent supervision. (d) experience suggests that the main risk in this project (and probably in the follow-up project) was institutional uncertainty. This resultea from a lack ot continuity in staffing, policies and programs, and from uncertainty of funding. (e) Li1A and the Government learned the importance of the availability of viable technical packages and the role ot regional development, as discussed below. Technical PacKages 9.UZ The absence ot technical packages at the start of the project deterred significant impact on agricultural output. The emphasis on agronomic trials during project implementation made it possible to develop technical recommendations, especially higher yielding varieties, which should enable the follow-up project to yield acceptable returns. Kegional Development 9.U3 Considering the pervasive weaknesses of Central Government services, the Government decided at the design stage to entrust the project to a decentralized agency. Two formulas were discussed by the preparation team: a regional development authority, as already existed in the Artibonite Valley, or a small project unit entrusted with programming, budgeting and monitoring actions carried out by line agencies. The Government opted for the second formula, which had not been tried before. - 31 - 9.04 The formula proved difficult to apply in practice, on account of the weakness of the regional line agencies. Those agencies which were active in the area (roads, agricultural credit) resented what they saw as an intrusion of a foreign supported agency into their sphere of influence. Public Works accepted eventually to participate in the project after protracted negotiations, but BCA declined. In cases where the regional services were nonexistent or too weak, ODN assumed responsibility for project actions, including farm credit, and helped build up regional services. Such actions could not have been carried out under the responsibility of the central services in the capital. 9.05 In spite of early growing pains and institutional problems (financial autonomy, high operating costs and over-staffing), ODN has developed into a reasonably viable development institution. It fulfills its role of a coordinating and funding agency and implements activities that cannot be undertaken by the regional services. Under the follow up project, institutional arrangements between ODN and line agencies are governed by signed agreements in which the respective responsibilities are defined. 9.06 The ODN formula has made it possible to avoid or minimize the constraint resulting from the weaknesses of the central services and inefficiencies attached to excessive centralization. In spite of its deficiencies, the ODN formula may be replicated for development of other regions. The Decentralization Law enacted in 1982 attempts to shift responsibility for development to regional agencies. Also, the recently defined USAID strategy opts for the support of regional development agencies of the ODN type. The Organization for the Development of the Northwest was created in 1983. 9.07 Problems encountered during project implementation led to adjustments in organization and scope of the various components. The experience gained with the first project was used in the design of the follow-up project (e.g., establishment of a working fund with an IDA advance, agreements between ODN and line agencies on the definition of responsibilities, definition of project components, and focus on rainfed agriculture). Project experience suggests that the regional development and ODN approaches are reasonable. It al;o suggests that agricultural development in Haiti is a slow, long-term process. .,,~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~p - 33 - AMX I EAm RDRAL DEVELOPMENT PROJECT IN THE NORTHERN DEPARTMENT PROJECT COMPLETION REPORT Mey Indicators U- Appraisal S Estimate Actual Completed (1976) (1983) Irrigation Irrigation canals built (km) U.n. 3.9 Drainage canals built (ckm) n.a. 6_0 - Total Irrigation Rehabilitation (ha) 4,900 660 9 Water Supply No. of systemes comileted 5 Z2 440 Storage capacity D U.n. 930 - Piping (km) u.n. 70 - No. of fountains .n&. 87 - No. of private connections U.&. 38 _ Wells (units) 400 135 34 Population served 20.500 100,000 488 Roads Feeder and farm access Ckm) 93.5 84.2 90 irrigation access (km) n.a. 13.0 - Field Trials No. of trials n.a 120 - Agricultural Credit (USS) Short-term 603.000 138,420 23 Medluu-term 200.000 244,210 122 Expenditures by Component (Total with US$'000 (Base Costs) contingencies) 1/ Irrigation 3,283 1,585 48 Agricultural credit 1,025 385 38 Trials and seed multiplication 83 496 598 Extension 523 411 79 Animal health 89 20 22 Markets and abattoir 161 391 243 Roads 1,247 1,875 150 Potable water 172 1,766 1,026 Project administration 2,227 6,842 2/ 307 Project monitoring 231 8 3 1/ Does not include costs of consultants financed by bilateral agencies and of Haitian civil servants paid from the Government's operating budget. 21 Of which $4.23 million for buildings and equlpments for all project components. iA ITI R{UR(AL 1JJEVELOPNhENT PkIUJIECT IN THEi NURTkiERN DEIPARTMhNT kVKOJECT CUMk'LETIUN k(hPUKT Schedule 1: Withdrawal of the Proceeds of the Credit Amount of the Amou,t Actual Credit Allocated Actually Disbursement/ (Expressed in Disbursed Amount Category US$ Eq.uivalent) (US~ Equivalent) Allocated ().Irrigation I*9UU,UUU bzd,uuu 44 (i)b. Uther Civil Works 1,7UU,UUU 4,UZ4,UUU 437 (2) Vehicles, Equipment 1,300,000 1,615,UUU 14U (3) 1?arm Lnputs 1,UOO,uu0 3u3,OUU 0 (4) Administrationi 2,900,000 2,95U,UUU1 102 (5) Uniallocated 1,2UU,0UUU_____ Total 10,1100,000 10,000,000 IOU 1g''"''''"'~~ i-i '~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ HAITI AI r,- j 't ' -4h RURAL DEVELOPMENT PROJECT * cr A N -, '\< U';t IN THE NORTHERN DEPARTMENT -' t U~~~~~~~~~~e I.:. X f1 Sir r___l bw k.S- A- I --_ [war %.I, D~~~~MW,- A- 40~ *Opua .. S > ' d O _ _ | ) _._0 200 hd.,,4,^ .....h;- -w SW I ', ,.:~~~~~~~~~~~~~~~~~~~2 - , 0 .. -,,...,.-,