Benin - Hinvi Agricultural Development Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 2053 FILE COPY PROJECT PERFORMANCE AUDIT REPORT BENIN - HINVI AGRICULTURAL PROJECT (CREDIT 144-BEN) May 15, 1978 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  FOR OFFICIAL USE ONLY Project Performance Audit Report BENIN - HINVI AGRICULTURAL PROJECT (Credit 144-BEN) Table of Contents Page Preface Basic Data Sheet Highlights PROJECT PERFORMANCE AUDIT MEMORANDUM I. Project Summary Background 1 The Project 2 Implementation 3 Impact of Project 4 Concluding Remarks 5 II. Main Issues Institutional Aspects 6 Insufficient Knowledge of Local Conditions 9 Impact of the Government's Price Policy on the Project 12 Economic Rate of Return 13 III. Conclusions 14 PROJECT COMPLETION REPORT I. Background, Preparation and Appraisal A.1 II. The Project A.2 III. Implementation A.6 IV. Evaluation A.10 V. Economic Results A.18 VI. Conclusion A.19 Annex Map This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  Project Performance Audit Report BENIN - HINVI AGRICULTURAL PROJECT (Credit 144-BEN) PREFACE This report presents the results of an audit of the Hinvi Agricul- tural Project in Benin for which the World Bank granted an initial credit of US$4.6 million (Credit 144-BEN), signed in March 1969, and a supplemental credit of US$600,000 (Credit 144-2-BEN), signed in March 1974. Credit 144-BEN was fully disbursed in June 1976, a few days before the closing date, and undisbursed funds of Credit 144-2-BEN (US$25,000) are fully committed. The report consists of a project completion report (PCR), prepared by the Western Africa Regional Office in March 1977, and a memorandum pre- pared by the Operations Evaluation Department (OED). The PCR contains a frank discussion of major project developments and highlights the difficulties encountered during project implementation. The memorandum comments on certain conclusions of the PCR that warrant more detailed analysis. The audit was based on the PCR, a review of the files available in the Bank, discussions with Bank staff familiar with the project and an OED mission to Benin in August 1977. The points discussed in the memorandum have been selected not only because of their importance in this project, but also because they appear to be relevant to the performance of three other projects in West Africa eval- uated recently by OED, viz. Senegal, Casamance Rice Project; Cameroon, Semry Rice Project; and Sierra Leone, Integrated Agricultural Development Project. The reports'on these projects are in preparation. The writers of the report wish to express their thanks to the Government of Benin and to the Societe Nationale de Developpement Rural (SONADER, now SOBEPALH), which was responsible for execution of the project, for their cooperation.  BASIC DATA SHEET BENTN - HINVI AGRICULTURAL PROJECT Credit 144-BEN A. Amounts As of June 30, 1977 Original Disbursed Repaid Outstanding Credit 144-BEN 2._ D .6 ei 6 Credit 144-2-BEN 0.6 0.575 - 0.575 Total 5.2 5.175 5.175 B. Project Data Original Data Revisions Actual First Mention in Bank Files 2/15/65 Appraisal Mission 1/15/67 Board Approval 2/18/69 Credit Agreement 3/22/74 3/ 5/69 Credit Effectiveness 6/ 5/69 6/20/74 8/ 5/69 Physical Completion 6/30/77 12/31/75 Credit Closing 6/30/76 /1 /1 6/11/76 Total Costs 9.6 10.6- 9.59-- 9.1 Economic Rate of Return 12% Re-estimated 5% C. Mission Data Month, No. of No. of No. of Date of Year days persons Manweeks Report Prepreparation 4/65 Preparation 11-12/66 6/ 1/67 Pre-appraisal 3/67 6 4 3 Appraisal 7/67 28 3 12 /2 Reapnraisal 8/68 12 2 4 2/ 3/69 Supervision I 4/69 1 1 6/26/69 Supervision II 8/69 10 3 5 MissinR Supervision III 3/70 11 3 5 5/18/70 Supervision IV 8/70 14 1 2 9/30/70 Supervision V 2/71 10 2 3 4/20/71 Supervision VI 6/71 12 3 6 6/28/71 Supervision VII 1/72 11 1 2 4/ 7/72 Supervision VIII 10/72 11 2 4 11/21/72 Supervision IX 6/73 3 1 1 6/27/73 Supervision X 9/73 5 2 2 11/28/73 Supervision XI 5/76 7 1 1 6/ 5/74 Supervision XII 2/75 11 2 3 2/27/75 Total 35 No follow-up project. D. Exchange Rates 1S$1 = CFAF 247 (1969) 277 (August 1q6Q) 256 (April 1972) 219 (November 173) /1 Estimates. /I 7eport issued arter reappraisal.  Project Performance Audit Report BENIN - HINVI AGRICULTURAL PROJECT (Credit 144-BEN) HIGHLIGHTS The Agricultural Project provided funds for the establishment of 6,000-ha oil palm plantations, a 6,000-ha expansion of the annual food crop area, and the planting of 1,000-ha of timber, as well as the construction of an oil mill and project infrastructure. The main objec- tive of the project was to develop an efficient and modern system of agricultural production to assur6 participating small farmers improved incomes. The project's accomplishments fell short of appraisal expec- tations. Poor rainfall in two consecutive years caused delays in oil palm development and production, as well as the initial failure of the annual crop development, although during project execution it became apparent that labor constraints were a major factor causing the dis- appointing performance of the food crop subcomponent. To overcome this difficulty, an accelerated promotion of draft animal mechanization was required, for which the project had only insufficient funds available. The project was successful in assisting SONADER the executing agency to continue its good performance and in supporting cooperative developments. The rate of return has been re-estimated at less than 5% as cormared to 12% at appraisal. The following points may be of special interest: - investments in oil palm development despite known climatic constraints (PCR paras. 1.02, 1.04, 4.02-4.05, 4.34, 5.02, 5.08); - shortfall of food crop production due to unexpected labor constraints (PPAR paras. 32-36; PCR paras. 1.03, 1.06, 2.18, 3.01, 4.15, 5.02); - introduction of draft animal mechanization helps break labor bottlenecks (PPAR paras. 18, 37, 38; PCR paras. 4.16-4.19, 5.02); - negative influences of Government pricing policies (PPAR paras. 42-44; PCR paras. 4.11, 4.15, 4.21, 4.36, 6.02); - successful operations of cooperatives and the wholly national project agency (PPAR paras. 24-31; PCR paras. 1.05, 2.10, 2.14, 4.29, 4.31, 4.32).  Project Performance Audit Memorandum BENIN - HINVI AGRICULTURAL PROJECT (Credit 144-BEN) I. PROJECT SUMMARY I/ Background 1. The possibility of Bank involvement in oil palm development in Benin was raised by an FAO/IBRD CP mission in April 1965. European financial aid agencies were also interested but reluctant to carry the full burden of external financing. In 1966 the Bank helped the Societe Nationale de Developpement Rural (SONADER), i a parastatal run com- pletely by Africans, to prepare the proposal. Three Bank missions in 1967 and 1968 reviewed the project and the Credit Agreement became effective in August 1969. 2. The constraints on the agricultural components of the project, located in the Grand Hinvi area about 70 km north of Cotonou, were recognized from the start: (i) marginal climatic conditions for the proposed ten 600-ha oil palm blocks, making it likely that yields would be 50% below those in the most favorable regions of West Africa, and (ii) low yields also likely on individual plots in the proposed ten 600-ha annual crop blocks and competition between cultivation activities on these plots and maintenance of the cooperative plantations. In spite of these drawbacks, the pre-appraisal report concluded that the project was justified because there were no investment opportunities then offering better prospects in the agriculture sector of the country. 1/ Adapted from the PCR. 2/ SONADER is now called SOBEPALH but its functions have not undergone any significant change. - 2 - 3. The appraisal report did not differ substantially from the pre- appraisal report. It described SONADER as having a high level of compe- tence and possessing considerable experience in the organization of cooperatives. It repeated the misgivings regarding the area's suitability for oil palm. But it gave less weight to the constraints on annual crop production; in fact income from these crops was a significant element in the total incremental income forecast in the report and contributed appreciably to the project's favorable rate of economic return (12%). The Project 4. The project consisted mainly of: (i) planting and bringing to maturity 6,000 ha of oil palms; (ii) preparing 6,000 ha for annual crop production; and (iii) constructf g an oil mill with an ultimate annual capacity of 70,000 tons of ffb.- The project also comprised components for forestry, livestock, maize storage, roads and administrative facilities. 5. The participants were to be organized into ten producer coopera- tives - each with 600 ha of oil palm and 600 ha of annual crops (the latter block sub-divided in 1.5 ha individual plots). For the first 25 years SONADER was to promote and direct the activities of the coopera- tives. Subsequently, the cooperatives were to become lessees of the land and owners of the improvements brought by the project. 6. The main objective was to develop an efficient system of agricul- tural production capable of replacing the "shifting cultivation" system and raising the farmers' living standards. The project was to make it possible for Benin to maintain, or even increase, its palm oil exports. The annual value of the incremental production was estimated at US$1.3 million in 1975 and US$2.4 million in 1980 (1969 prices). 7. The cost of the project was estimated at US$9.6 million (or CFAF 2.37 billion), of which IDA was to finance 47.9% (i.e. US$4.6 million), FAC (Fonds d'Aide a la Cooperation, the French bilateral aid fund) 47.9% and the Benin Government 4.2%. 8. The physical implementation of the project was to be executed directly by SONADER. For the 25-year life of the project, SONADER was to manage on behalf of the cooperatives the oil palm plantations, the annual crop blocks and all the productive components of the project except for the oil mill, which would be SONADER property. SONADER was thus to play two distinct roles: (i) that of a development agency, and (ii) that of a managing agent for the cooperatives. 9. Farmer participation in the project was not entirely voluntary, since SONADER was legally empowered to oblige landowners in the project areas to join producer cooperatives and rent their land to these cooperatives. 1/ fresh fruit bunches. - 3 - The farmers could become members of the cooperatives by leasing their land to them ("A" share members), by working on the oil palm plantations ("B" share members), or both. The novel feature of this cooperative system was the formation of producer cooperatives with two categories of members, where it was intended that only the "B" share members were entitled to a voice in the running of the cooperative. In the event of a surplus, they were to receive interest of 3% on their shares in the profits of the cooperative. Implementation 10. Oil Palm Plantations. The planting program got off to a good start and was completed in accordance with the forecasts. Altogether 6,075 ha of oil palms were correctly planted and reasonably well main- tained, despite difficulties in obtaining the labor needed for these operations. 11. Annual Crop Areas. The annual crop program soon ran into diffi- culties. Fewer farmers than expected were prepared to abandon traditional shifting cultivation on plots cleared in the forest and take up 1.5 ha holdings in the project blocks. On these 600 ha blocks, called ZOCA's, input use was minimal and yields far below appraised targets. By 1971, only 450 farmers had taken up plots in the ZOCAs. In 1971-72, a greater flexibility permitted to the farmers by SONADER for choosing the crop rotation, and the introduction of animal traction, stimulated more farmer interest. Thanks to technical assistance financed by FAC, the use of animal traction increased rapidly. But constraints on the supply of animals are believed to limit expansion of ox-drawn cultivation and further development of the ZOCAs remains uncertain. 12. Oil Mill. The original project called for a mill of 24 t/h capacity to be built in two stages. The revised project (April 1971) showed the need for a 20 t/h mill, with provision for expansion to 40 t/h. Construction of the mill was completed in 1974 as planned, but it did not go into operation until mid-1975 owing to water-supply problems. The total investment cost was CFAF 962.4 million (US$3.9 million), as estimated in 1971. The functioning of the mill is satisfactory on the whole, especially since some last adjustments in 1977 which made it possible to raise capacity to 27 t/h. 13. Costs and Financing. The project costs and financing schedule was revised twice. In 1971, with the difficulties encountered on the ZOCAs and a need to reallocate oil palm production among existing and planned mills, project costs were restructured. The Bank then abandoned its support to the ZOCAs, which SONADER pursued thereafter with FAC technical assistance, and the funds thus released helped finance the extra costs of the mill. Total project costs were increased by 26% to CFAF 3.02 billion. 14. In 1973 SONADER faced a financial crisis, caused by successive dollar devaluations and non-payment of the government contribution, and aggravated by the agency's failure to establish financial planning and control systems. To solve these problems, IDA provided a supplemental credit of US$600,000 in January 1974 and FAC contributed an additional US$520,000 to the project. 15. In the end, total project expenditures were lower than estimated (CFAF 2,266.6 million). IDA financing was increased, however, from US$4.6 million to US$5.2 million, thus accounting for just over 54% of the total bill. FAC's contribution amounted to 43% and the Government's to 3%. 16. These costs were determined from the disbursement records kept by SONADER. For certain types of expenditures, the weakness of SONADER's accounting procedures means that actual costs cannot be stated with certainty. However, the accounting system is now being improved. It must be recognized that disbursements and procurement were particularly complicated in this project owing to the sharing of payments between FAC and the Bank. Impact of Project 17. Plantation Yield and Production. Despite the excellent performance in planting, and satisfactory maintenance, production has been very much below estimates. The principal reason is low rainfall. In most of the cooperatives, rainfall for two years (1971-72 and 1972-73) was less than the minima recorded over the previous 30-year period and distributed very unevenly over neighboring cooperatives. Production was also affected by fires, spreading from where farmers were burning their fields, and theft, due to the inadequate organization of ffb collection and the higher prices obtainable on the parallel market. 18. Annual Crops. With the introduction of animal traction, the total number of farmers in the ZOCAs rose from approximately 1,800 in 1973 to 3,000 in 1975, and those using ox-drawn equipment from 352 to 880. Maize yields are double those obtained with hand cultivation; cotton yields, how- ever, remain marginal. 19. Oil Mill. The functioning of the mill is satisfactory, although a certain number of difficulties still persist (peak production higher and more concentrated than anticipated, high proportion of kernel, insufficient oil storage capacity). The free fat acid content is much too high (over 7% as against a target of 4%), due to inadequate organization of fruit collection and the insufficiency of storage capacity. SONADER plans to raise the capacity from 27 t/h to 40 t/h, which will entail an additional cost of CFAF 900 million (US$4 million at 1976 prices). - 5 - 20. SONADER. The assistance from the Bank and FAC helped SONADER pass through its financial crisis in 1f973. Essentially this assistance provided for the injection of more fund and for an administrative reor- ganization designed to bring about tighter control of the plantation operations and more effective guidance for the cooperatives. 21. Cooperatives. The cooperatives had a hard time getting going, mainly because of initial opposition from the landowners. The working members did not perceive themselves as owners of the oil-palm plantations, responsible for their success or failure, and they considered remuneration from the plantations as being too low in relation to work done. At the recoverable price of CFAF 5/kg ffb 1/ the estimated cash flows of the cooperatives are not satisfactory. In these circumstances, the coopera- tives would never be able to pay their debts as provided for in the Credit Agreement. The Government has been asked several times to raise the price, but no decision has yet been taken. 22. Rate of Return. To recalculate the economic rate of return, the wage is set by the PCR to reflect the labor constraints in the area, no deduction was made for the oil and kernel production foregone from wild palms on project land, and annual crop production was expected to stabilize at the 1975 level on the 1,500 ha presently cropped. Under various assump- tions of future oil palm yields, the economic rate of return is estimated at less than 5%. Concluding Remarks 23. The project has been closely supervised since its beginning and IDA demonstrated its flexibility in 1971 and 1973 in coping with the problems encountered. However, the outcome of the project now depends on two emerging problems concerning organization and the financing of the cooperatives. The first was precipitated by Government's decision to entrust all extension work in the country to a newly formed organization; SONADER would henceforth only manage fruit collection and processing. If carried out it would mean abandoning the concept of integration of the different project functions in one single organization and SONADER's supervision of the cooperatives. The second problem results from the small share (37% of net revenue) of the income of the oil-palm plantations received by the cooperatives. The low share jeopardizes their autonomy. Government appears to be aware of the probable consequences of these matters. 1/ This price has been raised since the time of the PCR (para. 44). - 6 - II. MAIN ISSUES Institutional Aspects 24. SONADER. One of the unusual features of the project is that it was prepared and executed by a Beninese corporation with very little technical assistance. SONADER assembled the technical, social, and economic data, and prepared the whole of the Government's application. The agency's management has been sound, both for this project and for activities in other regions covering similarly ambitious programs. Besides incorporating into its operations a number of new activities called for at appraisal, such as stock-raising, forestry plantations, and palm oil processing, SONADER succeeded also in introducing animal traction to improve the situation in the ZOCAs, a component not discussed in the planning period. SONADER also innovated by establishing a women's section in the field service, partly to ensure support from female workers (para. 42). Such weaknesses that drew the attention of supervision, particularly in accounting, were or are being corrected and no longer seem to seriously affect agency operations. To sum up, despite its deficiencies, SONADER is remarkably effective, and this is recognized not only by the Government but also by the commercial banks. 25. The chief reasons for this success are: (i) SONADER has always been able, because of its good reputation, because of a salary structure which remunerates its officers for their good performance, and because of its favorable southern locations (not far from the major urban centers), to attract competent staff and thus to set up an efficient management team; (ii) SONADER has operating autonomy and executes its development programs according to business criteria; (iii) however, this autonomy does not make SONADER into a "foreign body", since it remains first and foremost an executing agency for Government policy, well integrated into the national institutional setting; and (iv) SONADER was able to effec- tively coordinate its different fields of activity, because they were integrated under a single authority. 26. SONADER's ability to analyze its problems and act accordingly is impressive. The introduction of ox-drawn cultivation and the creation of a female worker's section exemplify this ability. It is noteworthy that SONADER has responded to its experience without having instituted a formal monitoring and evaluation function. SONADER delegates authority to its different sections and the cooperatives. 1 1/ As the PCR points out, this efficient network for communication and coordination may be lost if the new agencies established recently (CARDER: Centre d'Appui Regional pour le Developpement Rural; SONACEB: Societe Nationale du Commerce Exterieur du Benin; SODERA: Societe de Developpement des Ressources Animales) start performing the functions officially assigned to them. Government confirmed to the audit mission that it is aware of the dangers of subdividing SONADER's activities among the new entities. - 7 - 27. SONADER's experience raises the question as to whether the Bank in its projects should limit the services of expatriates to advisory functions during the initial stages of execution and institution building and have projects managed by national staff from the onset. On the one hand, the project had characteristics which are usually thought of as requiring expatriate management, numerous components involving a complex task of coordination, and components such as the establishment of estate plantations and the operation of major processing facilities, calling for strict administrative discipline. SONADER has demonstrated that even these difficult tasks are within reach of local agencies. Furthermore, local management seems to facilitate both the integration of the project into national programs and the interaction between the project and the intended beneficiaries. These are important advantages over expatriate management for maintaining project achievements once external financial assistance ceases. On the other hand, Benin is known to be exceptionally well endowed with high-caliber personnel. Furthermore, SONADER had accumulated ten years of experience when the project started and was well established as a trustworthy undertaking to which the government was willing to delegate authority to operate efficiently. It is therefore difficult to determine to what extent SONADER's experience is replicable. Even in Benin SONADER's success seems to be exceptional: OED is currently auditing the Benin Zou-Borgou Cotton Project and the performance of the semi-autonomous executing agency for that project (SONACO) has been pretty poor. 28. The Cooperatives. In contrast to experience in most African countries, and despite valid reservations made in the PCR (PCR paras. 4.33 through 4.36), the project cooperatives seem to be well on the way to becoming efficient farmers organizations, managed to a large degree by the members themselves. This phenomenon can be attributed to the following set of factors: (a) Legislation, which attempts to find original solutions for local conditions. By making a distinction between "A" and "B" members (PCR para. 2.16) for example, this legislation facilitated low-cost land reform by compensating land owners with shares instead of cash while leaving management authority to the workers (only "B" members have voting rights); (b) Good management support from SONADER, which gives the cooperatives efficient back-up in selecting staff, provides technical assistance and supplies inputs. SONADER manage- ment now views its role as a supporting institution to the cooperatives, the reverse of the earlier relationships; (c) A fair system of remuneration (wages, dividends) which increases payment to farmers as the cooperative develops, thereby starting to satisfy the cooperative membet's expectation of increased income in step with project progress; - 8 - (d) An efficient system of supervision. The accounts are kept by SONADER, but are subject to a dual-control system, being supervised by (i) an auditor chosen by the cooperative members, usually a trusted civil servant and (ii) a Govern- ment auditor from the Ministry of Rural Development; (e) The granting of effective powers, from the start, to the Board of Directors, which organizes the cooperative's day-to- day operations and pays and supervises the permanent staff of the cooperative; and (f) A national political context favoring farmer participation in agricultural development organizations. 29. One potential obstacle to the smooth development of the cooperatives has not been mentioned in the PCR;./ it is the risk of a small minority of "B" share members gaining control of the cooperatives' management, at the expense of those who actually contribute most of the work. In one of the cooperatives visited by the audit mission, there were 470 "A" members, 100 "B" members and 300 workers who had not accumulated in four years the 200 days necessary to become "B" members. In another cooperative the composition of the labor force in 1976 was as follows: 4% worked over 200 days per year (semi-permanent workers); 17% worked between 50 and 200 days per year; 37% worked between 10 and 49 days per year; 42% worked less than 10 days per year. Thus, less than a quarter of the manpower reaches the average 50 mandays per year necessary to acquire a "B" share within four years time. For the 10 cooperatives there are 4,731 "A" members and 1,218 "B" members. These figures compare with the estimate of 3,000 farmers having a plot in the ZOCAs (PCR para. 4.18). There is no indication how that group of 3,000 relates to the "A" and "B" categories, but at most only about 40% of farmers in the ZOCAs have "B" status. Since only "B" members are legally entitled to participate in the management of the cooperatives, this minority group could in theory work against the interests of the broadly based participation on which the principle of the cooperatives was based. 30. So far, the workers who are not "B" members are nevertheless permitted to participate in share holder meetings. They also received dividends in pro- portion to their work. Thus, the danger of a take-over by the minority "B" members does not seem to be immediate. Nevertheless, the legislation should be amended to take the actual situation into account and allow formal parti- cipation of all male and female workers (e.g., through the creation of special shares for instance). 1/ Two other problems mentioned in the PCR are the planned take-over of the cooperatives by a newly formed Government agency and the low price paid by Government for the cooperatives oil palm products. - 9 - 31. In sum, project success is related to the strength of its institu- tions. At the outset the Bank decided to take part in the project chiefly on the basis of the proper functioning of SONADER. As indicated in the PCR (para. 1.05) the initial confidence was justified. Today, the fact that SONADER and the cooperatives are functioning satisfactorily augurs well for the future of the project. In three other West African projects visited at the same time as this audit (see the Preface), second phase projects were necessary not only to support additional investments, but also to maintain the achievements of the first phase by providing further management assistance to the project executing agency. This is not so in the Hinvi Project. Insufficient Knowledge of Local Conditions 32. Labor. For the plantations, the Bank assumed that family workers in the traditional system were largely underemployed and had limited opportu- nities for employment outside the project. Thus, the initial scheme envisaged that the cooperatives would benefit from a steady labor force with each member working 50 to 100 days per year once the project was fully under way. This regular supply of labor was expected to be an important factor in ensuring the proper technical functioning of the cooperative and the interest and loyalty of its members. 33. The labor assumptions were overoptimistic. A labor shortage appeared starting in the land-clearing stage, and it was necessary to recruit casual labor, sometimes from outside the area. Even now, although the project as a whole is able to obtain the amount of labor required for maintenance and harvesting, its distribution is extremely uneven (para. 29), 34. The actual labor situation is analyzed in a study made in 1973 in the Mono River Valley, where SONADER implements a similar oil palm venture partly financed by FED, and where the amount of time devoted to agricultural work has traditionally been about 100 days annually per active person. In retrospect it can be said that furnishing to the coops 50 days of additional work was a substantial increment in view of the other economic opportunities and social obligations of the farmers. In fact, it was mainly young people without family responsibilities who worked more than 50 days annually on the plantations. The other farmers worked in the cooperative only when they had no other opportunity for employment. This happened only occasionally, as suggested by the figures in para. 29. The difficulties encountered at the start of the project with the maintenance of the palm plantations arose chiefly from this lack of labor. That it has not become a major problem is mainly due to the cooperatives' acceptance of irregular workers as full members although the original legislation did not provide for this (para. 31). - 10 - 35. The Annual Crop Areas (ZOCAs). The new permanent agricultural sys- tem developed at a local research center, at that time operated by French technicians, was based on the rotation of crops and intensive farming methods. The system had to be followed rigorously to yield results which would make it superior to the old system of shifting cultivation. Previously, the new system had been introduced without much success; in fact, all of SONADER's earlier experiments with it had foundered. Nonetheless, at apprai- sal, all agencies concerned believed that the new system was superior to the traditional one of shifting cultivation and felt that the farmers would adopt it. 36. From the outset the proposed system was not popular with the Hinvi farmers, who, as the PCR indicates (para. 4.15), preferred their traditional practices to the intensive project routine. It would appear that the new system did not meet all the needs of the farmers - the part devoted to food crops was too small, the yields obtained from commercial crops (groundnuts, cotton) were low and the planned rotation created serious labor constraints during the planting and harvesting periods. Since the rotation system was rigid and virtually obligatory, it is not surprising that in the early years the farmers stayed away from or abandoned the ZOCAs, preferring to grow their crops outside the project. The Bank, like FAC and SONADER, seems to have seriously overestimated the benefits that the farmers could derive from the new system. Subject to labor constraints, farmers did not assess the system in terms of its return per hectare - in which it was unquestionably superior to the traditional system - but rather in terms of its return per day worked - in which its superiority remains uncertain and probably small at best. 37. Later, relaxation of the cropping pattern routine and the intro- duction of ox-drawn cultivation brought about a radical change. The use of animal traction makes it possible to almost double the yields of the main crops while reducing the mandays needed per hectare at the critical moments of planting and weeding. 38. The PCR correctly notes that expansion of the use of animal traction faces a number of obstacles: the inability to supply oxen adapted to the conditions of the region, and financial limitations which prevent the farmers from acquiring a pair of these draft animals. The cost of a pair of oxen, together with the risks associated with the purchase (mortality of the animals, difficulty with repayments in years of low rainfall) and the additional work involved in looking after the animals, have persuaded the small and medium farmers to rent oxen. The rental system leaves the risk and the outlay - but also the profit - to the large operators. Without financial assistance for small and medium farmers, the introduction of animal traction is leading toward greater social differentiation. Thus, the Bank may have erred when it decided not to help SONADER with the ZOCAs precisely at the time a solution seemed to be in sight. In 1971, when the project was reviewed, FAC stepped up its technical assistance to ZOCAs for developing animal traction, while the Bank reduced its contribution to this component of the project due to shortage of credit funds (para. 47). - 11 - 39. It is evident that the breakthrough brought about by ox-drawn cultivation has been only partial, and that other factors continue to depress planned activity on the ZOCAs. The audit mission found that: (i) The crop rotation has been virtually abandoned. Maize accounts for about 80% of the ox-drawn cultivated areas and 90% of the hand cultivated areas. This monoculture responds to the greatly improved market for maize, in Benin and Nigeria, but threatens to exhaust the soil; (ii) Fertilizer consumption remains extremely limited (4.5 tons total in 1976 versus 100 tons forecast at appraisal) and has been falling sharply since the rise in fertilizer prices in 1974; (iii) The sum of (i) and (ii) result in a critical threat of soil degradation. This is one reason for the limited use (50%) of available area: farmers prefer to leave fallow land which they believe to be partly exhausted,,in accordance with their traditional shifting system. Thus, ox-drawn cultivation alone is unlikely to bring about a new agricul- tural equilibrium as long as the problem of maintaining the fertility of the soil has not been resolved. 40. Role and Attitude of Women in the Project. During the preparation of the project no specific consideration was given to women, even though they were known to play a key role in food crop production, in the processing of wild oil palm fruits and in the marketing of oil and maize. The women have not yet given full support to the project: It has eliminated a good part of their annual income (processing and marketing), a loss which has been offset only in part by the possibility of obtaining paid work in the cooperative. Their participation in the cooperative is still limited: there are very few women "A" or "B" members. Nonetheless, their role has been important, especially in work requiring meticulous detail (prepa- ration of nurseries and planting of legume cover). In 1973, SONADER recruited female personnel to set up a women's section responsible for a program of home economics and instruction on new cultivation techniques. Some additional measures (participation in cooperative structures, establishment of stores for the supply of staples, etc.) would undoubtedly increase the women's support for the project. That support is essential to its continued success, for the women can undermine the design. They might protest, for instance, by processing more fruit themselves instead of having them delivered to the mill (PCR, para. 4.10). 41. In the early stage of the project, the Bank's lack of attention to the needs of women farmers might be explained by its over-optimism about labor availability, which made the project appear less dependent upon female participation. The emerging labor shortage, however, and the large number of women in the nurseries, should have made the Bank more aware of the special role played by women in a project that had been designed mostly - 12 - with men in mind. The head of SONADER women's section commented that the audit mission was the first World Bank mission that had asked to meet with her. Impact of the Government's Price Policy on the Project 42. Despite a recent increase in the produce price for oil palm fruit from CFAF 4.6-4.9/kg to 5.6-5.9/kg, the parallel market is still more attractive to the growers and the problem of fruit theft continues to be a concern to project authorities. The farmers apparently make more money by processing the fruit themselves than by delivering it to the mill. The mill, already operating below capacity, has again cut back its processing volume in 1977. 43. At the current price, the cooperatives can cover their operating costs but an average yield of at least six tons/ha is necessary to allow them to repay their 25-year loans and provide fair remuneration for manpower. At a yield of 5 tons, as obtained during recent years, the cooperatives will not be able to pay back either long-term loans for project investments or short-term borrowings in the form of cash advances granted by SONADER. This is not a deficiency of the cooperatives, which are operating efficiently and could only with difficulty reduce production costs. 44. The policy to keep producer prices as low as possible has to be seen in the light of the budgetary problems faced by the Benin Government. By becoming the sole beneficiary of the wide margin between low producer and high world market prices, Government is allowed to transfer revenues from this project to other less favored areas. However, the funds that the Government would distribute to Hinvi farmers via a producer price increase would be partly returned to it in the form of loan repayments. Likewise, with mills operating below capacity, the fixed costs per unit processed are excessive. They would be reduced, and Government's average earning per ton increased, by an upward adjustment of the producers' price. The latter would be only a temporary outlay, which would be recovered by the Government over the long run. The increase would also strengthen the cooperatives to play a larger developmental role. For all these reasons, it seems urgent to bring remuneration of the farmers' production into line with market conditions, either by increasing the price itself, or by arranging for the cooperatives to share in the profits of the mill. 1/ The Bank was right in requesting continuously policy changes during project execution and the PCR is correct in its criticism of Government for not giving this matter the consideration it requires. 1/ It was originally planned that the cooperatives would be federated into a cooperative union which was eventually to take over responsibility for the mill. If this idea is abandoned, the question of distribution of the mill's surplus, after normal repayment of loans, will still have to be settled. - 13 - Economic Rate of Return 45. The pre-appraisal mission assessed project risks and concluded that even if the economic return was marginal at best, the project was justified for other reasons, such as a good institutional base and a lack of other investment opportunities in the agriculture sector in Benin. 46. The appraisal report echoed the warnings of the previous mission, but the quantitative analysis provided a more optimistic report. Production values proposed by these two missions, and those originally suggested by SONADER, are as follows: Annual Gross Income SONADER Pre-appraisal Appraisal in Operation Estimate Estimate Estimate (millions CFAF) (millions CFAF) (millions CFAF) Palm oil 486 424 340 Palm kernel 115 65 49 Annual crops 139 46 250 Total 740 535 639 The major role assigned to annual crops in the appraisal design was not found in the earlier proposals. Given the failure of previous SONADER trials with annual crops, appraisal's faith in the successful outcome of this component seems unwarranted. There were, on the contrary, numerous signs that it was overoptimistic and that the economic justification was therefore overstated. 47. The ROR estimate, apparently important at the time of project appraisal, seems to have been ignored during project execution. When, following the 1971 SONADER crisis, the Bank revised project costs, it reduced those attributed to the ZOCAs and increased those to the mill (PCR, para. 3.09). Contrary to the position of FAC, the Bank did not find it appropriate to assist in developing ox-drawn cultivation. At the time, doubts existed about the appropriateness of ox-drawn cultivation as a way to make the ZOCAs more attractive to the farmers and the prospects for livestock and annual crops development remained therefore uncertain. Furthermore, the Bank had limited experience with ox-drawn cultivation. Thus, the Bank was reluctant to support SONADER's initiative. But the main constraint according to Bank staff was financial: the Bank had only US$600,000 available and was committed to finance 100% of the cost of the mill. It therefore withdrew its support to the ZOCAs tomeet the full cost of constructing the mill. The Bank was thereby cutting itself adrift from its own appraisal analysis, since annual crops were intended to provide 40% of project earnings (see table above). - 14 - Abandonment of the ZOCA by the Bank implied that project's ROR would be driven below zero. There is no indication in Bank documents, however, that the decision to reduce efforts in the ZOCAs was taken in association with an analysis of the impact on the ROR. The ROR was no longer empha- sized as a decision-making tool, even when significant changes in project components were being considered. It is fortunate that FAC continued to promote the development of ox-drawn cultivation, because the annual crop component has been essential in maintaining a positive ROR. 48. Unrealistic assumptions in previous ROR calculations have been corrected in the PCR. Yield estimates take into account possible future poor rainfall conditions, areas cultivated in the ZOCAs have been assumed to remain unchanged (PCR, para. 5.01), and labor has been costed at close to market price. For the reasons outlined in the preappraisal report (PCR, para. 1.04), the low rate of return is not an indication that the investment should not have been carried out. III. CONCLUSIONS 49. On the basis of economic returns, the project does not compare favorably with the appraisal forecasts. The new ROR estimate at most reaches 5%, far below the 12% forecast by the appraisal mission. Output has undoubtedly suffered from poor rainfall in the period. Nevertheless, even with palm yields equal to those estimated in the original appraisal, the project would not reach the forecast ROR since the production of annual crops is far below expectations. That result is explained with reasons largely unconnected with climatic conditions and which were fore- seeable at the time of appraisal. A more in-depth analysis then of the project setting, and appraisal figures more in line with the uncertainties expressed in the text of the appraisal report, would have led to a more realistic ROR, and the final result would not be disappointing. 50. The most interesting aspects of this project are the institutional aspects. The institutions set up have proven to be efficient and show a great capacity for self-betterment. SONADER is remarkable in that it is entirely managed by nationals who required only a minimum of foreign technical assistance to complete the development program in the alloted time. The cooperatives evidence interesting characteristics and are a rare example of successful deployment of rural community organization in a Bank project, conspicuously more successful than the institutional evolu- tion in the three other West African projects mentioned in the Preface, which were reviewed by OED at the same time. The ability of SONADER and - 15 - the cooperatives to continue to pursue their development will depend, however, on the support that the Government provides them in the form of continued integration of the activities now embraced by SONADER's single organization, fair prices for cooperative products, and adaptation of the legislation governing the cooperatives to the new needs which have arisen since their establishment. 51. The Bank's review of on-farm conditions was superficial. Benefits that farmers obtained from their traditional cultivation system, and the farmers' perception of those benefits, were underestimated, and potential benefits from the new annual crops program were overestimated. The Bank also seriously overestimated the farmers' availability of labor. The common appraisal assumption in smallholder projects, that the opportunity cost of the small1olders' family labor is zero, holds in none of the four projects visited.T 52. Two lessons for other projects like this one are that farmers' likelihood to participate in the projects should be measured in terms of returns per manday as much or more than returns per hectare; and that detailed studies of seasonal labor bottlenecks, and the role of women and of younger family members who form a major part of the available labor force, should be made at appraisal and during project implementation to ensure that the effects of labor constraints on the project are adequately anticipated. 1/ For the past three years CPS has not accepted assumptions that the opportunity costs of smallholders' family labor are zero in appraisal reports.  BENIN HINVI AGRICULTURAL PROJECT CREDIT 144-BEN COMPLETION REPORT TABLE OF CONTENTS Page I. BACKGROUND, PREPARATION AND APPRAISAL A.1 II. THE PROJECT A.2 A. Project description and objectives B. Project costs and financial arrangements A.3 C. Organisation and management A.4 a) SONADER1 b) The Cooperatives A.5 III. IMPLEMENTATION A.6 A. Chronological Review B. Project Revisions A.7 C. Costs disbursement and procurement A.9 IV. EVALUATION A.10 A. Productive Components A.10 B. Institutions A.16 C. Cooperative and Social Aspects A.17 V. ECONOMIC RESULTS A.17 VI. CONCLUSION A.19 ANNEXES Annex I, Table 1: Cooperative Cash Flow Table 2: Expost Evaluation of the Economic Rate of Return Table 3: Hinvi Oil Mill: Investment Cost Table 4: Rainfall in Grand Hinvi Project Area Table 5: Yields of Fresh Fruit Bunches .i The report refers to SONADER, the agency responsible for project implementation. Since 1976, SONADER has become SOBEPALM (Societe Beninoise de Palmiers a Huile).  - A.1 - I. BACKGROUND, PREPARATION AND APPRAISAL 1.01 The possibility of Bank involvement in oil palm development in Benin was discussed by an FAO/IBRD CP mission in a report dated April 1965. This report concluded that the ongoing program of planting 2,500 hectares of oil-palm annually was likely to be financed by FAC and FED at least until 1968, and there was therefore no room for immediate Bank involvement. The CP report did however point out that production from the ongoing program would not meet projected requirements of palm oil for export. During 1966, because of a growing reluctance of FAC and FED to carry the full burden of external financing for the proposed project, the Government requested assis- tance from the Bank. PMWA then collaborated with SONADER to prepare a project during the latter months of 1966. This was followed by a pre-appraisal mission in March 1967, appraisal in July 1967 and reappraisal in August 1968; the Credit Agreement became effective in August 1969. 1.02 From the beginning it was recognized that the agricultural elements of the project, namely development of oil palm in ten cooperatively owned blocks of about 600 ha each, and development of annual food crops on the holdings of individual cooperator/farmers, would be subject to a number of constraints. For instance, the pre-appraisal report said "conditions for planted (oil) palms are marginal and yields are likely to be 50 percent of those in the most favorable areas of West Africa" (P.A.R. para 2.07 of appraisal report). The incorporation of oil palm plantations with food crop growing and small holdings had also been the subject of a number of negative experiments but it was believed the problem had been overcome (P.A.R. paras 2.10 - 2.12. See footnote). 1.03 As regards the field crops it was feared that the cooperative orga- nisation proposed might mean that the farmer cooperators would neglect either field crops or oil palms, and although it was felt that it had been carefully thought out, "time alone (would) show its effectiveness" (P.A.R. para 3.20). 1.04 In spite of these reservations, the preappraisal report concluded that there were "no other investment opportunities which offer more favorable prospects than this project in the agricultural sector", (P.A.R. para 5.01). If the appraisal mission could satisfy itself that the cooperators could "produce as much from the field crop areas as they could without the project" and that productivity could be increased by the project, the project would probably be judged economically viable (P.A.R. para 5.03). The delay of nineteen nonths between the first appraisal mission and the issue of the appraisal report as "due to the political instability following the change of government in December 1967 and the precarious financial situation in Benin which prevented FAC and IDA from proceeding with the project until the end of summer 1968." (P-670 5/2/69). Note: In this paper, References to documents are as follows: P.A.R. Pre-Appraisal Report dated 1st June 1967, Memorandum from Rowe, Vigie, Bishop and von Czernicki to Evans. P-670 Report and Recommendation of the President, dated 5th February 1969. A.R. Appraisal Report. Report on Hinvi Agricultural Development Project, TO-615b, dated 3rd February 1969. S.M.R. Supervision Mission Report. - A.2 - 1.05 The appraisal report did not differ substantially from the pre- appraisal report in its general assessment of the situation. It confirmed both the misgivings about the suitability of the area for oil palm (A.R. para 3.02) and the judgement about SONADER's capacity to execute the project. "SONADER is completely Africanized and displays a high level of competence, both at its head office and in the field. It is efficieit, and has gained considerable experience in agricultural development and the organization and management of producer cooperatives. SONADER is capable of handling an expanded programme including the proposed project. It is of the utmost importance to the continued success of SONADER's operations that the quality of its senior staff is main- tained" (A.R. para 6.02). This refreshing and rather rare confidence in the ability of an institution to carry out a development project, was, as will be seen, on the whole, well-justified. 1.06 Unfortunately the appraisal report devotes relatively little attention to the constraints on annual crop production. This had been picked out in the reappraisal report as a key issue; income from annual crops was an important element of the total incremental income and in the favorable economic evaluation of the project. In the event the relatively meagre success of annual crop development almost caused the project to founder during implementation. Thus it is unfortunate that this issue was not given more prominence at the time of appraisal. 11. THE PROJECT A. Project Description and Objectives 2.01 The Appraisal Report summarises the project and its principal objectives as follows: - establishing and bringing to maturity 6,000 ha of oil palms; - preparing 6,000 ha for annual crop production; - constructing a palm oil factory with an ultimate annual capacity of 70,000 tons of ffb; - planting 1,000 ha of teak and cassia trees; - purchasing 310 cattle for the development of beef production; - constructing maize storage silo with an ultimate capacity of 3,000 tons; - developing necessary roads and central project facilities. Development, and subsequent production, was to be organized through ten cooperative units - each with 600 ha of oil palms and 600 ha of annual crops. During the first 25 years of project development and operation, the Societe Nationale de Developpement Rural (SONADER) was to have full responsibility for managing the project, and directing cooperative activity. Subsequently, the ten cooperatives were to become lessees of the land and owners of the im- provements described above. SONADER was to train required cooperative staff, and to meet cooperative staffing costs during the development period. 2.02 The main objective of the project was to develop an efficient and modern system of agricultural production, capable of assuring participating farmers of standards of living superior to those obtainable from traditional farming methods. In achieving this, the project was to increase the production, and maintain, - A.3 - while initially increasing, exports of oil palm produce from Benin. 2.03 Incremental production generated by the project was expected to be as follows: (per year) 1975 At Maturity (1980) Palm Oil (Tons) 4,275 10,080 Palm Kernels (Tons) 1,395 2,400 MAize (Tons) 6,535 8,739* Groundnuts (Tons) 1,582 1,920* Seed Cotton (Tons) 1,338 1,600 Livestock (Adult Head) 60 62* Teak (Poles) 70,000 (approx.) The annual value of incremental production was estimated to amount to US$1.3 million in 1975 and to US$2.4 million in 1980 (at 1969 prices). B. Project Costs and Financial Arrangements 2.04 Estimates of Project costs are summarized in the following table: Original Project Cost Estimates Local Currency US$ Local Foreign Total Local Foreign Total Currency Exchange Costs Currency Exchange Costs -------CFAF Million---- -------US$ '000-------- Studies 8 32 40 32 130 162 Oil Palm Development 708 237 945 2,870 956 3,826 Oil Palm Factory 64 492 556 255 1,992 2,247 Annual Crop Development 165 32 197 668 129 797 Livestock 34 3 37 138 12 150 Afforestation 25 4 29 101 16 117 Staff and Training 36 4 40 146 16 162 Villages, Roads, Vehicles,etc. 75 119 194 304 481 785 Maize Silos 4 36 40 16 146 162 Overheads and Maintenance 80 44 124 324 178 502 Contingencies 84 84 168 340 340 680 Total 1,283 1,087 2,370 5,194 4,396 9,590 2.05 Fonds d'Aide at de Cooperation (FAC) of France participated in financing the project as shown in the following agreed financing plan: The annual production of these commodities was expected to increase slightly in subsequent years up to 385,000 poles in the period 1980-84. Financing Plan Source Foreign Exchange Local Currency Total Financing Percentage ---------------------US$ Million----------------- IDA 2.8 1.8 4.6 47.9 FAC 1.6 3.0 4.6 47.9 Government of Benin - 0.4 0.4 4.2 Total 4.4 5.2 9.6 100.0 2.06 As explained in the President's Report and Recommendation, the financing arrangement with FAC meant that "the amount of IDA financing (was) substantially the equivalent of the foreign exchange component of the whole project" (P-670, para 11). The large proportion of external financing was conditioned by the fact that Benin was suffering financial difficulties at that time; was dependent oii France for considerable current budget support; and could not be expected to make a large contribution to development expenditures. 2.07 It was proposed that "IDA and FAC financing would be parallel in part, and the remainder joint. All goods and services financed wholly by IDA, valued at $2.7 million, and all goods and services financed jointly by IDA and FAC, valued at $5.8 million, would be procured through international.competitive bidding, except for contracts of $50,000 equivalent or less which would be awarded on the basis of local competitive bidding in accordance with procedures acceptable to IDA. Goods and services financed wholly by FAC, valued at $0.7 million, would be procured within the Franc zone in accordance with FAC's normal procurement procedures." (P-670, para 12). 2.08 Internal financing arrangements were that Government would (a) on- lend the IDA credit to SONADER at 6% interest for a term of 25 years including 9 years of grace and (b) onlend the proceeds of the FAC grant to SONADER for a term of 31 years with varying interest rates of 0.75% to 2.5%, repayment to be to Fonds Dahomeyen de Renouvellement de lapAijeraie, a fund established, at FAC insistence, for the purpose of financing a continuing program of oil palm development. The Government contribution was to reimburse SONADER for the tax aidduty component of goods purchased by SONADER in Benin. Direct imports were to be free of import duties. C. Organisation and Management 2.09 The organisation of the project,involved two principal elements (a) SONADER, with overall responsibility for management of the project and the affairs of the cooperatives elements; (b) ten producer cooperatives that were expected to comprise 4,000 farm families each working 1.5 ha of oil palms and 1.5 ha of arable crops. (para 2.15). 2.10 SONADER. As its name implied, SONADER had in principle very wide statutory responsibilities for rural development, but in practice, when the project was under discussion, these responsibilities were limited to fostering oil palm development with lesser emphasis on food crop development in the oil palm areas. - A.5 - During project formulation SONADER was also given responsibility for ownership and management of the oil mill which would eventually have to be built; this was done because of the evident weaknesses of SNAHDA, the already existing organisation responsible for managing the four oil mills then in opera- tion. Also during formulation, 'but not discovered by the Bank or FAC until the loan was just about to become effective, SONADER was given responsibility for all rural development activities in the Mono River Valley. The first super- vision mission judged that "These responsibilities cannot be undertaken without adversely affecting the progress of SONADER's existing activities" (SMR 26/6/69 para 7). This, together with doubts about its ability to keep adequate financial records, was the only reservation expressed about SONADER's competence to execute the project. 2.11 Farmer participation in the project was not entirely voluntary, since SONADER was legally entitled to: - oblige land owners in designated development areas to group themselves into producer cooperatives, or alternatively; - oblige land owners in development areas to rent their land to producer cooperatives, established by the agency; and supervise and control, for 25 years, producer cooperatives established in the above manner. 2.12 Physical development was to be directly executed by SONADER with its own staff and funds, and was to include establishment of the oil palm plantations, tree plantings, and cattle herds; road and palm oil factory construction, and the clearing and development of village sites and annual crops areas. 2.13 For the 25-year life of the project SONADER was to manage on behalf of the cooperatives the oil palm plantations, the blocks of annual crops and all productive components of the ptoject, except the oil mill which would be SONADER property. During this period SONADER was to have complete managerial, administrative, and financial control of the cooperatives; all revenues were to accrue to SONADER, and only after all costs, including loan repayments, had been met, would the surplus be paid to the cooperatives. During the development period the ten cooperatives, and an apex organization -- the Cooperative Union, were to be created by SONADER and gradually take over responsibility for manage- ment. SONADER had legal power to take over the direction of the cooperatives or even to dissolve them in the event of unsatisfactory performance. In practice, and since the ten cooperatives would be financially indebted to SONADER, the latter was expected to be able to exert a high level of direct control over them for the first 25 years of their existence. It was recognized that only after this time, and after discharge of their debt to SONADER would the cooperatives become autonomous. SONADER was thus required to play two distinct roles in development and operation of the project. First as a development agency, and second as a managing agent for the cooperatives. 2 14 The Cooperatives. Each of the project's ten cooperatives was expected eventually to manage about 600 ba of oil palms, 600 ha of annual crops, and associated tree plantings, cattle, and buildings and equipment. 2.15 Farmers could become members of the cooperatives by leasing land to - i. 6 - the cooperative on a fifty-year basis, by working on the oil palm plantations or both. In compensation the farmers received shares in the cooperative which entitled them to a fixed return. In addition farmers who contributed labor were entitled to a daily cash renumeration plus a share in the profits of the cooperative. 2.16 Each hectare rented to the cooperative entitled its owner to one "A" share and an interest of CFAF 900 (US$3.60) per year. Farmers working for a minimum of two hundred days per year were entitled to a "B" share. Both types of share were valued at CFAF 30,000 (US$120). Farmers qualifying for "B" shares received CFAF 125 (US$0.50) for each days work on the oil palm plantation. Since the "official" daily wage in the project area was CFAF 275 (US$1.1), the farmers were considered to be contributing CFAF 150 (US$0.6) in value of work per day to the project, or CFAF 30,000 (US$120) total. "B" share farmers also received 1.5 ha of cleared land from the cooperative, whi'ch they were under an obligation to cultivate in the manner prescribed by SONADER. Any surplus after payment of interest on "B" shares was to be used for capital improvements and to increade the daily wage for workers in the plantations. 2.17 The 1.5 (1.4 net) ha holdings of annual crops were to be worked by cooperative members a6d their families as individual enterprises, the produce from these being their sole property. The cooperatives, however, were to require farmers to follow a specific rotation for 1.2 ha of each plot. On the remaining 0.2 ha, farmers were to be free to grow what they please. The cooperatives would provide seeds, fertilizer and other inputs which would be paid for by farmers, and marketing and storage facilities. Farmers would have the choice of whether or not to use the latter. 2.18 By way of comment, the provisions for "B" shares and annual crop holdings, were clearly based on the assumptions (i) that there was a surplus of labor in the project area, so that farmers would be willing to work for low wages on the oil palm plantations, and (ii) that the required rotation for the annual crops was sufficiently profitable to be attractive to farmers. In the event neither assumption was justified, and this was one of the factors which eventually caused the project to be substantially revised. 2.19 The ten project cooperatives were to be grouped into a Cooperative Union, which eventually would own the palm oil factory, maize silos and other central facilities. As in the case of the inditidual cooperatives, the Cooperative Union was to be managed by SONADER for 25 years. III. IMPLEMENTATION A. Chronological Review 3.01 Because of the delays mentioned above, implementation of the project started before the credit became effective. However, while the oil palm plantating programme got off to a good start and was completed in accordance with forecasts, the annual crop element was much less successful. From the beginning, supervision missions expressed doubts about progress. By June, 1971, about 1,500 ha of land for annual crops had been cleared, as compared with fore- casts at appraisal of 1,750 ha, nevertheless only about 700 ha were being cul- tivated, and there was a marked lack of demand for cleared land. 3.02 In addition, by mid-April, detailed planning for the oil mill showed that a 20 ton/hour mill was necessary, with provision for increase to 40 ton/ hour mill, as compared with the 24 ton/hour mill to be constructed in two stages which was provided for in the project. The increased mill size was required because other milling capacity was no longer available to process oil palm bunches produced by the project. 3.03 Finally the increasing responsibilities laid on SONADER by the Government, principally its continuing involvement in the Mono Project and in rapidly increasing oil palm development in addition to the Hinvi project, seemed to be straining its executive ability and endangering its capacity to im- plement the project. This was evidenced by poor financial and administrative control of the cooperatives, by financial control of its own operations, and poor maintenance of the plantations. 3.04 In order to rectify the situation a substantial revision was made to the project in 1971. Essentially this involved abandoning any further annual crop development (though FAC insisted on bringing the remainder of the cleared land under cultivation, financed out of its own contribution); utilising the funds so saved to provide the necessary extra finance for the oil mill; and effecting a reorganisation of SONADER so as to achieve tighter supervision of the cooperatives and field control of the plantations. 3.05 After this substantial revision, the project continued to make progress, but with difficulty. The oil mill was constructed, but could not be commissioned on time, because of problems with the water supply. There were difficulties over procurement for the mill auxiliary installation because of collusion between bidders. On the bright side, the annual crop programme began to show positive results as a result of the introduction of ox-teams for farm operations, and the beef cattle enterprise began to benefit from better management. 3.06 Underlying everything were the continuous financial problems of SONADER. These were compounded of SONADER's inability to establish and maintain an adequate system of financial planning and control, and by its fragile financial situation. These financial difficulties came to a head in 1973, with successive devaluations of the US dollar in relation to the CFAF, and in January 1974, a supplemental credit of $600,000 was agreed upon. 3.07 In the end, the project as revised was completed on time in mid 1976. By then, the 6,000 ha of oil palm were in production; the oil mill began operating in 1974; about 1,500 ha of annual crops were being farmed,part of it cultivated by ox-drawn equipment. Yields from the oil-palms cannot yet be forecast with accuracy, but-have recovered from the adverse effect of the drought. One problem which was never solved was the weakness of the accounting prodedures and financial control of SONADER. B. Project Revisions 3.08 The revisions made to the project in 1971 involved the following: a) change from construction of the mill in two phases (first 16 ton ffb/h and second 24 tons ffb/h) to construction of a single, - pA A - 20 ton/h mill, with room for expansion to 40 ton/h if required later; b) increase in cost of the mill to allow for price escalation and also extra ancillary facilities; c) reduction of annual crop area from 6,000 ha to 1,700 ha; d) reduction of maize silo capacity from 3,000 to 2,000 ton. 3.09 These changes led to revised project cost estimates, as shown below: CFAF Appraisal Revised --------CFAF Million----- Oil Mill .556 1.369 1/ Maize Silos .00 - Oil Palm Development .945 1.076 Other Ag. Development .262 .111 Infrastructure .158 .202 Administration .179 .212 Studies .040 .036 Contingency .189 .035 Total 2.370 3.021 1/ including the second line with total output of 4o t/hour. 3.10 These revisions were agreed to by the Executive Directors on September 13, 1971, (IDA R71-63), and the project description was amended appropriately. 3.11 It should be noted that this revised funding was based on the assumption that FAC would provide CFAF 1,235.6 million (US$4.5 million), IDA CFAF 1,269.9 million (US$4.6 million) and the remaining CFAF 515.8 million (US$2 million) would come from the Government (CFAF 145 million) and SONADER self-generated funds (CFAF 370.8 million). Government contribution has never been paid and this was one of the causes for SONADER's continuing financial problems. The installation of a second processing line has been postponed because of lower production than expected. Subsequently, under the adverse impact of currency realignments (US$/CFAF exchange rate declining from 1:277 to 1:225) Credit 144-IDA was increased in February 1974 by US$600,000 to US$5.2 million. In late 1973, FAC also increased its grant contribution by US$510,000 equivalent. 3.12 The following tables show the finally agreed financing plan, and the planned disposition of the revised IDA credit. - A.9 - - A.9 - Final Planned Disposition of IDA Credit Final Pldft--d'(OMposition of IDA Credit US$ '000 Category I Studies and construction of palm oil Category I milUu:tabb*ft0d_I1midditurcEf qrAm oil vehtdEI4. associated installations and 4,158 vehicles. 4,158 Category II (a) Development of oil palm plantations 830 Category II (b) (Othdlellp lml lahlqialpmqtntations 159830 (b) Other agricultural development 159 Category III Contingencies 53 Category III Contingencies 53 Total 5 200 Total 00 C. Costs, Disbursements and Procurement C. Costs, Disbursements and Procurement 3.13 The following table compares appraisal estimates of project cost with actk.3131isbur&kwmedelowing table compares appraisal estimates of project cost with actual disbursements. Appraisal Actual A raisal CFAF Million CFAF Million US$ 1 = 241 CFAF Million CFAF Million (US$ 1 =24) IDA FAC Govt Total Total IDA FAC Govt Total Total Studies - 40 - 40 40 0ilgathedeveiopment 230.9- 458.5to - - 689.t 945 h0 Oil Qiallnpaft3evelopment 962.930.9 - 458.5 - - 962.89.5 556945 Annill qlp deNOlopment - 962.4 201.0- 4.2 - 205362.4 197556 Livdtaak1 crop development 1.0 - 5.3Z1.0 o.9 4.2 7.A5.2 37 197 AffbatahoWn - 1.0 5.7 5.9.4.7 0.9 20.47.8 29 37 StaNraadiri n d25.9 - 53.75.7 -14.7 78.&0. 4 40 29 VillS m0Wft tnmb fin 7.T5.9 94.F3-7[8.1 - 119. W8.6 194 4o Wbiag usd, roads 7.3 94.0 .18.1 119.4 194 Maize sidlies, etc. - 40.3 - 40.3 40 Ovethed mit&4aintenance 9.0- 66.5140.3.4.9 - 90.10.3 124 40 Contnnade and Maintenanoe 9.0 - 66.9 -14.9 - 90.8 168124 RentContingencies - - - - 10.8 - 10.8- - 168 Rent - - 10.8 10.8 - Total . 1,236.5 966.0 63.6 2,266.2 2,370 Total . 1,236.5 966.0 63.6 2,266.2 2,370 - A.10 - 3.14 These costs have been determined through the analysis of disbursement files kept at SONADER. Costs related to contracts or suppliers invoices are accu- rate; land development cost (oil palm and annual crops at forestation) overheads and infrastructure are based on SONADER's allocations to project cost without re- gard to actual costs which are uni own (even by SONADER) and for which it is impos- sible to obtain further details. IDA supervision missions have recommended many times that SONADER take action to improve the accounting procedures. In 1974 and 1975 FED had financed a consultant, SORGEM, to train staff and establish new pro- cedures but this action has not brought any significant results. Balance sheets and financial statements, although more accurate, are still unaudited and issued late, budgetary control and financial management remain weak. SONADER has generally followed IDA guideline for procurement, in particular for the oil mill. However, an IDA supervision mission recommended in 1974 to not finance the exaggerated cost of a water tower (US$150,000) since the contract was awarded without IDA consultation. This recommendation has not been followed ultimately because SONADER would have had to pay for a Government mistake and this could have aggravated project execution. 3.15 Disbursement and procurement were complicated. Until the 1971 revision FAC paid 100% for some items and FAC and IDA shared in the proportion 54 to k6, for the oil palm plantations; staff and training; village sites, roads etc; livestock; and overheads and maintenance. Thereafter IDA paid 100% for the ancillary facilities; the oil palm plantations and staff and training continued to be shared as before; all the rest were paid for 100% by FAC. Prior to 1969, FAC of course paid everything. IV. EVALUATION k.0l For review purposes the project can be divided into three parts: (a) the productive components; (b) institutions; and (c) cooperative and social aspects. A- Productive Components 4.02 a) Establishing and Bringing to Maturity 6,000 ha of Oil Palms Planned and Actual Plantings Calendar Year 1968 1969 1970 Total ha planted-planned 1,800 1,800 2,oo 6,000 actual 1,835 1,8k2 2,403 6,075 Plantings, as shown by the table, were on schedule and made with high potential seed- lings of good physical quality. Maintenance .was generally good through 1971 and an invasion by the spear grass weed (imperata spp) extending to 1,070 ha had been cleared up by the beginning of the 1972 wet season. In 1972, however, a general fall in main- tenance standards occured and has not been corrected to date. The basic cause is fi- nancial. Cooperators are unwilling to work for the CFAF 125/day paid by SONADER to the cooperatives for maintenance work in the immature plantations, unless no other work is available, as during the dry season, when there is no difficulty in obtaining labour. Thus during the rainy season maintenance is less than would be acceptable an commercial estates elsewhere. To correct the situation, SONADER has begun to use machinery for - A.11 - maintenance specially financed by FAC. However, it is unlikely that yield depression because of low maintenance is significant. Moreover, because of the low yield poten- tial of the area it is doubtful whether a large increase in maintenance costs is eco- nomically justifiable. 4.03 Despite the excellent performance in planting, and reasonable maintenance, it is not clear at present whether or not yields will reach the levels anticipated at appraisal. This is due to the low rainfall experienced in 1971/72 and 1972/73 which occured at a critical time in the early growth of the plantations. Also it may be that the rainfall pattern is more unfavourable than was at first thought. . 4.04 The attached table shows the distribution of rainfall by crop-year and by plantation. The overall average crop-year rainfall for the years 1969/70 to 1974/75 in the project area was about 980 mm as compared with about 1,170 mm during the years 1941-1966 at Niaouli on the boundary of the project area. However, two of the planta- tions had rainfall within 5% of the Niaouli 25-year average. Within the project area there were considerable differences between the plantations, that with the least rain- fall, Dodji-Sehe, having about 1/3 less than that with the most, Agbotagan. These data do not show any apparent geographical pattern when plotted on a map. 4.05 From the point of view of oil palm growth and development, a very significant constraint is the incidence of long dry spells. However, measured by the number of three monthly periods with rainfall of less than 100 mm, the drier plantations do not appear to be much worse in this respect than are the plantations with high rainfall. 4.o6 On the basis of data available, it seems clear that the substantial diffe- rences between plantations was an adverse factor unknown to the appraisal mission. Furthermore the abnormally low rainfall of two years has substantially affected the six-year average. However, all years were below the Niaouli average and only two years were within 10% of it. Thus it seems possible that the long-term average rainfall in the project area, and hence yields, will be somewhat lower than that envisaged at ap- praisal. 4.07 The following table shows original and revised estimates of future yields. Comparative Yields - Estimated and Revised Tons ffb/ha and 12 Years after Planting 5 6 7 8 9 10 11 Subsequent Appraisal Estimates (1969) 3.0 5.0 6.5 7.5 8.0 8.0 8.0 Revised IRHO (1973) 0.5 1.5 3.0 4.o 6.0 8.0 8.0 Revised Bank (1973) 0.3 1.7 2.3 3.6 5.6 6.3 7.0 Last Estimates (1976) 0.5 1.5 3.0 4.0 6.0 7.0 7.0 4.08 Harvesting of ffb started in the first three months of 1973, when 116 tons were produced by the cooperatives of Agbotagon and Goulo, planted in early 1968. There- after the progression of yields and production is as follows: - A.12 - Agricultural Year 1973/4 1974/5 Yield Production Yield Production Tons/ha To-s Tons/ha Tons (6th year after planting) (7th year after planting) Agbotagon 2.939 1.184.890 2.626 1,605.050 Attogon 0.704 429.972 2.490 1,519.360 Goulo 2.318 1,423.473 1.875 1,151.396 (5th year after planting) (6th year after planting) Koundokapoe 0.221 137.954 1.864 1,153.874 Sedje 1.099 661.063 Rodji 0.629 391.600 (5th year after planting) Sehe 68.330 3,176.289 6,550.676 4.09 The above figures show that in the sixth year after planting, average yields over the project area were about 1.4 ton/ha, which is much lower than the appraisal estimate, but about in line with the revised Bank and IRHO estimates of 1973. Furthermore the average yield should be adjusted to allow for the fact that only about 300 ha at Attogon were productive that year, the remainder having been burnt. 4.10 Any estimate of how production is now likely to develop must take into account two further factors, the incidence of fire, and of theft. During the first three months of 1973, 465 ha of trees were burnt as a result of fires spreading from where farmers were burning their fields. However, that was an abnormally dry period, and the problem has not reoccured. 4.11 Theft, accentuated by inadequate organization of ppb collection is however an important problem which could jeopardise the viability of the project. For instance, the Report of SONADER for 2nd Quarter of 1975, para 6.3, refers to "the eternal pro- blem of theft of fruit bunches". Essentially it is due to the low prices paid to the cooperatives for ffb (about 3.8 frs per kilo of ffb) which in turn affects the cash earnings of the workers on the plantations and cooperative cash flow. b) Annual Crop Development 4.12 The object of the annual crop development component of the project was (i) to introduce a system of rotational cropping to replace the "shifting cultivation" generally practised in the project area; and (ii) to promote the use of the seed of improved varieties, fertilisers and other inputs. Through these measures it was ex- pected that substantial gains in productivity would be achieved and that the average participating family, would achieve a cash income from annual crops alone of about CFAF 42,150 (Us$172) in 1980, as compared with about CFAF 12,000 (US$49) cash income annually without the project (constant 1967 CFAF). - A.13 - 4.13 The proposals concerning annual crops were that (i) a standar,d 1.5 ha holding (1.4 ha net) should follow a rotation which, under the two-season rainfall regime of the project area, would permit the cultivation each-year of 0.4 ha ground- nuts, 0.4 ha cotton, 1.2 ha maize, and 0.2 ha of other crops; and (ii) the use of seed of improved varieties, fertilisers and other inputs, would permit average yields to rise by 1980 to 1,260 kg/ha (in shell) for groundnuts, 1,000 kg/ha (seed cotton) for cotton, and 2,300 kg/ha (grai-) for maize. It was assumed that cultivation would continue to be by hand. 4.14 It was expected that the annual crop blocks would be developed quite rapidly as follows: 1968 1969 1970 1971 1972 1973 1974 1975 1976 ha (Cumulative) 317 600 1,150 1,750 2,400 3,300 4,200 5,200 6,ooo The blocks were to be cleared by SONADER which also was to mark out the individual holdings and construct access roads. While 1.5 ha was to be the norm, families with sufficient labor were to be permitted to operate two holdings, but 3 ha was the abso- lute maximum allotable to an individual cooperator. 4.15 The annual-crop development program soon ran into difficulties; fewer than expected farmers were prepared to take up holdings, input use was minimal and yields failed to reach appraisal targets. A Bank supervision mission in early 1971 attribu- ted the failure to (a) farmers' unwillingness to give up traditional systems of cul- tivation and adopt a complex rotation requiring a high degree of discipline; (b) poor results of cotton cultivation, a result of heavy insect infestation; (c) low yields and prices for groundnuts; and (d) the lack of agricultural credit and markets at assured prices. For all these reasons, farmers preferred to clear and cultivate small areas of bush for their own food crop requirements in the traditional manner, rather than take up cooperative plots with the attendant obligations. The mission concluded that by 1976 ( the closing date for the IDA credit) 3,000 ha might be developed if so- lutions were found to the above problems and some changes made to the rotation. Sub- sequently, however, it was agreed to limit further work to the preparation of 1,700 ha for annual cropping. 4.16 Since 1971 an unforeseen development has revitalized the annual crop compo- nent of the project. This is the introduction of draft oxen and plows, harrows, cul- tivators and carts with technical assistance financed by FAC. Original trials with draft oxen started in 1969, and the number of animal traction units (two oxen and equip- ment) has grown quite rapidly,as has the number of farmers using animal traction (farmers who owns units rent these to other growers). Thus in the 1975 crop season a total of 751 ha (442 main season; 309 short season) were cultivated with about 150 pairs of oxen and some 3,300 haulage trips were made with ox-drawn carts. 4.17 In the ten cooperatives 880 farmers used ox-drawn cultivation provided by 155 units. The benefits of ox-drawn cultivation are perceived by farmers to be so substantial that most of the 6,262 ha available for annual cropping have been taken up by farmers who are using hand cultivation until such time as they can obtain oxen and equipment. A requirement to receive credit for oxen and equipment purchase is that the farmer has removed all the stumps and roots from his land, a very substantial under- taking in the so-called ZOCA (zones de culture attelee) where only 1,700 ha were pre- pared by SONADER and where even this involved simply cutting off trees and shrubs at ground level. Despite the work involved arid the further requirement that the farmer must make a downpayment of CFAF 30,000 (US$120) for his oxen and equipment, 746 had been de-stumped by March 1976. - A.14 - 4.18 The benefits of ox-drawn cultivation are the higher yields permitted by timely planting and weeding at times of the year when labor shortage is the prin- cipal constraint. Derived from a sample survey, SONADER estimates of yields for 1973-1975 are as follows: Comparison of Yields from Hand Cultivated and Ox-Drawn Equipment Cultivated Farms rop Year Main Season Short Season Main Season Short Season Farmers Hand Ox-Drawn Hand Ox-Drawn Hand Ox-Drawn Hand Ox-Drawn Hand Ox-Drawn Total Ha Cultivated Yields, kg /ha Numbers Maize 1973 793 111 600 64 1,510 2,585 582 2,505 1,440 352 1,792 .grain) 1974 954 289 624 149 1,000 2,800 900 2,100 2,426 799 3,225 1975 996 380 1,188 304 1,100 2,300 1,000 1,900 2,099 880 2,979 Groundnuts 1973 16 16 7 21 1,580 2,505 1,521 2,550 (in shell) 1974 31 28 28 45 1,000 3,000 1,300 1,900 1975 55 38 13 80 1,800 2,000 800 1,800 Cowpeas 1973 45 5 - 2 N/A 598 571 588 (grain) 1974 20 12 5 8 800 1,100 N/A 800 1975 20 21 7 12 700 1,000 300 400 :tton 1973 - - - 46 - - - 599 Seed 1974 - - - - - - - 500 )tton) 1975 - - - 53 - - - 400 4.19 The table shows clearly the growing number of farmers and area involved in annual cultivation, and the substantial increases in maize and groundnut yields thus achieved. The latter would be significant even if SONADER's recording is faulty, and explains the high demand among farmers to acquire these facilities. The table also shows the generally poor performance of cotton, and the room for improvement in cowpea yields, though the sample is scarcely large enough to draw substantial conclusions. 4.20 It does however, seem fair to suppose that the appraisal estimates of yield foi maize and groundnuts will by and large be reached by 1980, not only on the revised area of 1,700 ha but also on the whole of the originally projected 6,000 ha, provided that there is no check to the introduction of animal-drawn cultivation. The most recent reports suggest that there are some constraints on the supply of animal for this,* which is a matter to which SONADER and Government should devote attention in the future. 4.21 In short, the annual crops program has now reached the point where it should provide a sound basis for further development; however, the above satisfactory results are likely to be eroded: SONADER has not been able to organise the development of ani- mal traction since 1974, because of the difficulty to purchase oxen at low prices fixed by Government. This could rapidly discourage farmers. - A.15 - 4.22 Livestock. The project included the purchase of 310 cattle for the development of beef production. Cattle has been purchased accordingly and allocated to 10 cooperatives, including the four financed by FAC. In 1975, the herd was 1030 heads. Veterinary services are adequate and the herd develop satisfactorily. How- ever, the benefits of that component do not accrue to the cooperatives because animals available for sale are sold at a price fixed by Government about 50% of the market price. The herd consists mainly of Ndama and few of them can be adequately trained for traction because of their small size. 4.23 Afforestation. 1000 ha of cassa and teak have been planted in 1972 and 1973 in line with appraisal targets. The plantations attached to each cooperative have not all been properly maintained and some of them have suffered from bush fire at their early stage of development. However, they have generally developed satis- factorily. No indication are available on the future number of poles made available to the cooperatives. 4.24 Oil Mill. The oil mill contract was awarded to de Wecker (Luxemburg), the lowest bidder in 1971 after an international bidding. Other bidders were SPEICHIM and VOYER (France). The construction of the mill started in 1972 and was completed in 1974 as scheduled. However, processing commenced only in mid 1975 because of delays in installing a tubewell, the inadequacy of water supply (about 10m3/hour instead of 20m3 envisaged) and the procurement of pumping equip- ment. Total investment cost amounts to CFAF 962,4 million (US$3.9 million) as anticipated in 1971 whentheproject was evaluated (detail at Annex 2) 4.25 The functioning of the oil mill is satisfactory but is has had difficulties because of several factors: first, peak productionis higher and more concentrated than anticipated at appraisal,and, as a result, the mill was heavily congested in the peak season of 1976 and 1977; second the mill cannot handle the high proportion of kernel (about 6-7% of ffb, instead of 4-5%) anticipated for which it was not de- signed (possibly because of hybrid seeds); and third, oil storage capacity is ina- dequate and has to be increased. Furthermore, some improvements of the machinery were to be done by WECKER under the guarantee clause of the contract early in 1977. 4.26 The quality of oil is also not satisfactory because of the inadequacy of field organization to collect fruits (see para 4.12). As a result, free fat acid content (ffac) is usually above 7%. While the mill is designed to produce oil with a ffac below 4%. The situation is aggravated by the inadequacy of storage condi- tions either at the mill or at the harbor, on the other hand, this should not affect prices since most of the oil is marketed domestically or in Nigeria. No data are available on the quantities and price at which oil is sold on the local, Nigerian and other export markets. No data are also available concerning processing, main- tenance, transport and storage costs. 4.27 The maintenance of the oil mill is almost adequate although it should be improved and the functionning oil mill is supervised twice a year by an expatriate engineer provided to SONADER by FED under an arrangement related to the AGONVY oil mill. IDA approved the appointment of an oil mill manager as specified in the Credit Agreement. During negotiations in 1971 SONADER agreed to set up a Technical Unit responsible for the maintenance of the three oil mills it was expected to mana- ge. This unit exists but still lacks technical staff and workshop facilities. This becomes a real need now that SONADER also manages the three oil mills formerly under SNAHDA supervision. - A.16 - 4.28 SONADER is now planning to expand the Hinvi oil mill capacity from 20 t per hour to 40 t per hour to cope with expected production in 19(9-80. In the meantime, the FED engineer has been able to increase the output of the presses to about 30 t/hour and has designed plans to immediately increase kernel crushing po- tential (from 1 t to 1.5 t/hour).Altogether, additional investment required at Hinvi is estimated at CFAF 900 million (US$4 million) in 1976 prices and could alone justify the financing of a second project provided that an adequate supply be guaranteed by an improvement of collect organization (para 4.12) B. Institutions 4.29 The main Beninese institutions involved in the project were SONADER and SNAHDA*. The support given by the Bank to SONADER appears to have been well- justified and successful. Essentially SONADER was responsible for promoting pro- duction, while SNAHDA was responsible for processing and sale of palm oil and palm kernel oil. However, because of misgivings about SNAHDA's capacity to construct and manage the mill required to process project output, it was agreed that SONADER should do this also. It may be argued that from the point of view of institution building, it would have been preferable to have retained SNAHDA and attempted to improve its performance. However, given that the project mill was only one of se- veral operated by SNAHDA, it is unlikely that the Bank could have exerted enough leverage to have made an appreciable improvement in SNAHDA's efficiency. The choice taken was probably the correct one. 4.30 A crisis in SONADER's affairs came to a head in 1973. At that time it was trying to carry the burden of a much enlarged responsibility for rural develop- ment, as well as a rapidly accelerating oil palm development program, aggravated problems caused by a very adverse cash-flow and financial resource situation. These factors together meant that the standard of management of the oil palm plantations was slipping (aggravated by the adverse weather), control and guidance of the coope- ratives was lax, and SONADER's own financial policy and control was poor. 4.31 There is little doubt that the intervention of the Bank and FAC at that time succeeded in restoring the situation and enabled SONADER to retrieve itself. Essentially that intervention provided for the injection of more funds into the pro- ject (Us$600.000 from the Bank, and US$520,000 from FAC) and for an administrative reorganisation intented to bring about tighter field control of the plantation opera- tions, and better guidance for the cooperatives. It appears that by and large these actions achieved their purpose, though, as has been mentioned before the problem of ensuring SONADER's financial policy and control does not even now seem to have been fully solved. 4.32 Two more points may be made to demonstrate the basic soundness and viability of SONADER as an institution. One concerns the volume and adequacy of the reports provided by SONADER. These have consistently improved in both quantity and quality once the project started and provide a mass of data concerning all aspects of its ope- ration. The other concerns the introduction of animal-drawn equipment. The fact that when the annual crop programme as originally conceived was obviously failing, SONADER attempted a new approach with FAC assistance, clearly demonstrates its flexi- bility and willingness to learn from experience. * Note: These were changed at various times, and SONADER is now called SOBEPALH, and SNAHDA is called CONICOG. However, their functions have not substantially changed and the same names have been retained here for convenience. - A.17 - C. Cooperative and Social Aspects 4.33 Over this part of the project there still hangs a large question mark. Obviously in the early years of the project, the farmers who were members of a cooperative did not perceive themselves as owners of the oil-palm plantations, responsible for their success ro' failure. Rather the plantations were regarded as places where remuneration was very low in relation to work done, and farmers only sought employment there if no other was available. This still seems to be the situation, so that there is competition between work on the plantations and work on farmers' own fields. This competition is particularly important during the rainy season. 4.34 Thus although an operating framework of cooperatives has been set up, it is difficult to conclude that it has really taken root. Perhaps this will change with the spreading use of animal-drawn implements, in the introduction of which the cooperatives have a real part to play, as also in the distribution of inputs for improved agricultural practices. However, as regards the social aspects, there seems little doubt that the project has had a substantial impact on the farmers concerned. In an agricultural area where natural conditions (climate, soils) are marginal and potential productivity is low, nevertheless a new technology has been introduced and farmers can be seen to be adopting better practices than in other parts of the country. This must be counted as a major achievement. 4.35 Furthermore the provision of roads and village sites by the project will permit farmers to benefit not only from higher incomes but also from a better quality of life. 4.36 Financing of cooperatives. Projected cash flows for cooperation are not satisfactory, even if the expected yield of 7 tons of ffb per ha is reached. First, cooperatives are heavily indebted; and second, SONADER (and government) are not keen to increase the price for ffb paid to cooperatives. Current price of CFAF 5 per kilo of ffb is much below the price paid in Ivory Coast (CFAF 9) (where interests are subsidised). However, in present circumstances, it is unlikely that Government would agree to change its policy because of the com- petition between wild and selected palms, the former being transported and processed at a higher cost. In these circumstances, cooperatives would never be able to repay their debts as provided for in the Credit Agreement. Under present price arrangement, the financial rate of return of the investment is minus 0.45 over 25 years, but it could be raised to 4.7% assuming an ffb price of CFAF 7 Frs per kilo (+40%) that SONADER can reasonably afford to pay given the present and future trend for oil palm products. Government has been asked several times and this was a condition of renegotiations of the credit in 1971 to make proposal satisfactory to IDA to resolve this issue. No answer has been obtained yet. V. ECONOMIC RESULTS 5.01 The rate of return on the project has been re-calculted as follows: In the cost stream: a) labor costs through 1976 for oil palm plantations and palm oil mills are included at cost whether or not provided by cooperative members, and thereafter are costed at a minimum of CFAF 125/day (US$0.5) until 1975 and CFAF 200 (US$0.8) to better reflect the - A.18 - opportunity cost of such labor; b) no further costs are attributed to the annual crop production component of the project post 1975 other than recurring costs for the seed and other inputs employed and the replacement of oxen and equipment, on the grounds that current levels of production can be maintained for at least 10 years without further investment in either infrastructure or extension and other farmer support services. In the benefit stream: c) no deduction is made for palm oil and kernel production for- gone from wild palms on land now occupied by the project on the grounds that the net benefit would be small and difficult to quantify and in the calculation compensated for by attributing a uniformity round-the-year value to field labor of CFAF 125/day (US$0.5) through 1975 and CFAF 200/day (US$0.8) thereafter; and d) assuming annual crop production will stabilize at 1975 levels on the 1500 ha actually cropped under SONADER supervision, and treating as benefits from the annual crop program the difference between actual production and estimated production under traditional hand cultivation practices. Thus, it is assumed that farmers cultivating by hand obtain an increment in yield of 25% through project participation and these using ox-drawn cultivation substantially more. In addition, costs and benefits of the afforestation and livestock components of the project are removed from both streams. The grounds for this are that: (a) experience shows the livestock program must be treated as an experiment and the probably substantial benefits from the afforestation program cannot yet be quantified; and (b) the impact of these components on the overall rate of return is negligible. 5.02 An economic rate of return of about 12% over a life of 25 years was estimated at appraisal. In the calculation the cost of labor provided by cooperative members in both investment and operating phases was estimated at zero. Furthermore a deduc- tion was made from the benefit stream to allow for the estimated value of that produc- tion sold for cash that would have been produced on land occupied by the project were not implemented. The merits of this technique are a matter for conjecture, particular- ly as the success of ox-drawn cultivation demonstrates a labor bottleneck for much of the year, and in practice considerable recourse was made to hired labor from outside the project area for work on the plantation. Consequently, given the major changes that have occured since appraisal, including the delayed entry into production of the palms, the expected low oil palm yields, the sharp rises in the value of outputs, and the failure of the original plans for annual crop development, the calculation of a revised rate of return on the same basis for comparative purposes would not appear appropriate in light of the present circumstances. - A.19 - 5.03 In calculating the economic rate of return, the timing of investment costs was as follows: Up to 1969 1970 971 1972 1973 1974 1975 Total actual: 448.9 181.2 196.6 403.4 707.5 175.8 112.8 2,226.2 constant 1976 terms: 982.3 370.5 378.8 714.0 1056.0 214.4 117.6 5.04 As regards benefits from palm-oil, timing should be based on the assumption that each plantation will begin production five years after planting and yields will follow the progression as follows: Year 5 6 7 8 9 10 11 ffb ton/ha 0.5 1.5 3.0 4.0 6.0 7.0 7.0 5.05 Yieldsof palm-oil are put at 21% and of palm kernels at 5% at full production. 5.06 Price of palm oil and palm kernels are taken to be CIF US$578 and US$ 240 respectively in 1980, equivalent to CFAF 13,400 per ton of ffb, in 1976 prices. These prices are in line with the commodity price forecast. 5.07 There is no reliable data in Benin to calculate the economic benefits of the annual crop component; therefore the calculation has been done on the basis of farmgate prices recently estimated in neighboring countries. 5.08 The rate of return is sentitive only to oil palm yields variation since the benefits of the annual crop components are small. Given the uncertainties concerning the estimates of oil palm production, the rate of return has also be calculated with lower yields and would be 2.5% with an average yield of 6 t/ha and zero % with yield at 5 t/ha. Difference with appraisal estimates results primarily from the high benefits anticipated from the annual crop development, benefits which have not yet fully materialized, although clear indications show that the potential for further development still exists. VI. CONCLUSION 6.01 IDA Performance. The project has been closely supervised since its be- ginning. IDA has demonstrated its flexibility to cope with the problems encountered: a) in 1971, the project has been amended to insure the adequate financing of the Hinvi Oil Mill; b) in 1973, additional financing was also provided to compensate for losses due to exchange rate variations. - A.20 - Supervision missions have been generally more concerned with SONADER finance than with cooperative finance because they considered that SONADER was essential to project success and that no solution to cooperative problems could be found with low yields estimated in 1973-1975 when the effectsof draughts were at their maximum. However, we now have enough evidence that the production would be close to appraisal estimates, and therefore, cooprative financial problems can be resolved through an increase of the price of ffb. 6.02 Future of the project. The project now faces two major issues concerning the organization and cooperative finances. As originally conceived, the project was intended to integrate oil palm and annual crops within a cooperative framework expected to become autonomous with the support of a single agency SONADER. Now Government has decided that the "CARDER de l'Atlantique" would be responsible for the extension work and that SONADER (now called SOBEPALH) would manage the collection of palm fruits, thus making two different agencies intervening in the same project area; thus giving up the integration concept. These new arrangements would certainly aggravate the difficulties already met by SONADER to properly organize the collec- tion of fruits as discussed at para U.02. Furthermore, Government is studying the possibility to take over from cooperative the oil palm plantations which would be maintained and harvested by SOBEPALH with hired labor. The second issue results from the fact'that cooperatives do not receive an adequate share of the value of oil palm plantations. With a price of CFAF 5 per kg of ffb (37% of net revenue), they cannot either pay more than CFAF 200 (US$0.8) per manday to cooperators, nor repay their debts. Thus the important concept of autonomous cooperative is jeopar- dized. Government is aware of those problems and has temporary maintained SOBEPALE responsible for cooperative supervision in the project area. Further discussion with Government on these important issues is still possible because SOBEPALH is looking for external financing to increase the output of the Hinvi oil mill. IDA has, in principle, agreed to examine a new Hinvi project which would finance the Hinvi oil mill extension together with the development of ox-drawn cultivation. This would be discussed again with Government during the forthcoming negotiations of the Technical Assistance project scheduled in March/April 1977. BENIN k I Table 1 HTNVT PROJECT Coonerative Cash Flow (600 ha) 1976 Terms (CFAF'000) PY 1 PY 2 PY 3 PY'4 py 5 PY 6 FY 7 PY 8 7Y 9 PY 10 PY 11 Py 1 1:: ?L O'dS Develoor.ent Loan 102400 12700 9000 9000 Short term advances 3150 990 990 Sales of ffb A! 1500 4500 9000 12000 18000 21000 21000 210, Total Inflow 102400 12700 12150 9990 2490 4500 9000 12000 18000 21000 21000 210( 2lanation Development 79800 9000 9000 990 Adcational land clearing 3150 3150 990 990 Maincenance (tools, f rtilizers) 3420 3420 3420 3420 3420 3420 3420 3420 342 Maintenance (labor) - 3360 2880 2400 2400 2400 2400 2400 2400 24C Farvesting (Labor) 180 540 1080 1440 2160 2520 2520 2520 242 Coop overheads 3060 3060 3060 3060 3060 3060 3060 3060 206 Rent 540 540 540 540 980 980 980 980 98 Annual crops 19700 Livestock 3700 yGrest 2900 Tottl Outflow 102400 12700 12150 9990 11550 10440 10500 10860 11960 12380 12360 1238i (bfore debt services) (deficit) surplus (9060) (5940) (1500) 1140 6040 8620 8620 S62( Xandsy vorked 17700 17100 17400 19200 22800 24600 24600 2460( 1 zximum yield: 7t/ha 2/ CFAF 250 per manday ANNEX I BENIN Table 2 HITVI AGRICULTURAL PROJECT yp5EVALUATION OF THE ECONOMIC RATE OF RETURN (in 1976 constant terms) 68/69 70 71 72 73 74 75 76 77 78 79 80 ross Value of Incremental Production ffb (13 frs) 4.8 56.3 85.2 216.8 332.3 403.9 521.5 552.5 Maize (25) 37.1 37.1 37.1 37.1 37.1 37.1 Groundnut 7.1 7.1 7.1 7.1 7.1 7.1 Cowpeas 0.6 0.6 0.6 0.6 o.6 0.6 Cotton 0.9 0.9 0.9 0.9 0.9 0.9 Tech./Cassia Cattle Subtotal 4.8 56.3 130.9 262.5 .37b.0 44b9.b >3.2 )98.L roject Cost 448.9 181.2 196.6 403.4 705.5 175.8 112.8 quipment Renewal 40 40 I0 lantation Maint-iance 10.8 21.6 36 36 36 36 36 36 nnual Crop Maintenance 17.4 17.4 17.4 17.4 17.4 17.4 abor Cost 10.6 21.0 34.8 38.4 42 45.6 49.2 ,ooperative Management 9.2 18.4 30.6 30.6 30.6 30.6 30.6 Subtotal (current term) 448.9 181.2 196.6 403.4 716.3 217.2 205.6 118.8 122.4 166.0 169.6 173.2 In current 1976 terms 982.3 370.5 378.8 714.0 1069.1 264.9 214.4 118.8 122.4 166.0 169.6 173.2 ANNEX I BENIN Table 3 HINVI PROJECT HINVI OIL MILL; INVESTMIENT COST (CFAF Hillion) 1971 . 1972 1973 1974 1975 Total OIL MILL COMPLEX Oil mill (including civil works) 194.2 479.9 71.5 745.6 Access roads 1.6 1.6 Auxiliary buildings 15.5 15.5 Water and Electricity network 1.3 5.3 6.6 Tools 11.8 11.8 Office Equipment 0.3 0.6 0.9 Telephone Lines 8.5 8.5 Tubewell 21.6 21.6 Pumps 9.8 0.2 10.0 Water Tower 33.5 33.5 Technical Assistance 1.9 1.6 0.8 0.5 3.3 8.1 Subtotal Oil mill 1.9 219.0 539.8 87.1 15.9 863.7 VILLAGE Housing 10.2 h.7 1.2 16.1 Water supply 13.1 1.1 11.2 Electricity line 10.3 12.0 22.3 Subtotal Housing 10.2 28.1 13.1 1.2 52.6 RANSPORTATION EQUIPMENT Fork Cart 3.9 3.9 Vehicles 0.7 0.7 Tip Trucks 11.8 11.P Tank Truck 11.7 11.7 Tractors and Trailers 13.5 5.8 19.3 Subtota. Vehicles 16.4 25.2 5.8 4'.4 TOTAL 1.9 2115.6 567.9 125.4 2?.9 963.7 BENIN Table 4 HINVI PROJECT Table Rainfall in Grand Hinvi Project Area by 3-Monthly Periods, Showing Crop-Year Figures for each Cooperative Plantation. Scdje Koundokpoe Dodji Agbotagon Attogon Zegoulo Adjan Hanafin Dodji- Kpoe- Annual Average Gbeto Dessa Sehe Kpannon 1969.3 225.7 211.5 153.9 258.2 142.6 170.1 222.2 169.1 200.8 156.3 10 Cooperatives .4 234.9 295.2 189.9 173.2 272.2 220.1 289.6 275.3 210.6 207.1 1970.1 10S.1 127.1 127.1 134.5 164.4 94.7 95.3 198.2 104.5 19.1 7C.2 59S.9 503.5 386.9 699.7 495.5 507.1 478.2 6L9.3 442.5 464.2 7C 097.6 11 857.8 992.0 958.4 965.7 1,071.6 ..70.3 98.1 157.3 126.1 151.3 117.5 144.4 153.1 106.0 136.3 139.1 4 325.7 212.4 -320.1 285.3 186.8 302.2 238.2 267.7 265.5 237.9 1771.1 95.9 156.9 179.5 183.2 188.5 168.9 175.4 258.3 193.3 114.6 1971.2 453.6 442.1 308.9 297.3 295.7 417.7 348.7 439.5 200.4 271.3 973.1 968.7 934.6 917.1 788.5 133.2 915.4 1.071.5 795.5 762.9 916.1 1971.3 172.5 246.7 204.9 268.9 149.4 248.0 203.9 246.0 250.9 246.4 .4 40.1 96.6 94.9 87.4 120.7 72.9 46.8 149.1 132.1 39.9 72.1 141.6 166.1 260.0 216.5 223.9 182.6 189.9 116.9 258.7 146.6 2 510.3 541.5 506.7 72S.9 513.3 515.9 509.5 647.8 536.3 552.9 84.5 T.0.0.9 15c-6.5 1 3.07.3 1 950.1 11159.8 1 18.0 955.8 584.0 1972.3 203.8 182.2 143.4 234.8 138.8 234.0 175.6 231.9 128.5 148.6 .4 74.5 151.3 105.6 102.9 53.7 74.9 94.2 108.2 81.4 122.4 73.1 42.1 97.9 80.0 56.1 54.4 71.9 93.1 70.7 103.7 96.9 .2 270. 4 277.4 273.5 400.0 224.4 506.3 217.4 328.1 223.7 284.2 .25.S 73.2 599.5 793.8 471.3 687.1 530.3 738.9 537.3 652.1 656.5 1973.3 "29.8 340.5 292.0 375.7 288.6 518.3 453.4 342.2 340.7 281.5 .4 95.S 235.7 164.4 250.5 133.7 146.1 89.3 209.7 256.9 132.8 74.1 91.7 114.9 143.3 120.1 125.5 90.6 102.4 96.6 145.5 181.3 .2 3F6.9 444.7 L-62.9 537.0 473.9 468.2 515.3 517.4 476.9 442.9 _ )_-11__ _6 2 1 38 1 037 1,023.4 1974.3 290.4 40S.1 325.5 325.6 299.0 335.2 356.3 432.0 325.5 432.0 .4 61.4 163.9 189.0 198.2 165.4 85.2 90.4 115.8 127.5 247.9 75.1 58.5 154.8 93.0 173.2 51.7 152.0 133.3 151.6 84.0 96.6 .2 407.0 56'.4 572.1 643.9 462.6 522.8 525.2 575.8 533.5 529.2 11. 1=89-6078-7 52 l23 1.168.1 10-Year 892.1 1,047.9 951.8 1,167.1- 915.4 1,050.0 952.5 1,116.8 795.0 943.0 983.2 Nate: These figures are taken from SO\ADER 1/4 by reports except for the 3r'* nd 4th 1/4s of 1974, which are taken from Bogaerts tables. 7' ire reasoaably consistent with the anr.ual total given by SONADE. .zcept where marked by asterisk. BENIN ANNEX I Table HINVI PROJECT Table Yields of Fresh Fruit Bunches, Tons/ha 64/5 65/6 66/7 67/8 68/9 69/70 70/1 71/2 72/3 73/4 74/5 75/6 Hinvi Planted 3.0 4.5 4.5 Ovagbo Planted 2.4 4.5 4.2 Agon Planted 2.6 6.1 Assa-Gota Planted 2.3 5.0 Agbotagon Planted 1.9 2.6 Attogon Planted 0.7 2.5 Goulo Planted 2.3. 1.9 Koundokpoe Planted 0.2 1.9 Sedje Planted 1.1 Dodji Planted 0.6 Sehe Planted Hanafin Planted Adjan Planted Kpoe Planted  BE NIN HINVI AGRICULTURAL DEVELOPMENT PROJECT IBR - 2112 A MARCH 1978 UPPER NICER VOLTA- 0 BENIN TOGO NIGERIA TEAK FOREST *ABOMEY • POB AGONAGON ASSAGOTA *POBE AO ALLADAA GR D HINVI " PORTO NOVO COTONOU AGBOTAGON A T L A N T I C 0 C E A N- 1968 OUAGBO 3 HINVI 1970L HINVI E A 1969 11968 NIAOUL PPc B 2 ATTOGON 1968 2 DOODJI-BATA ALLADA 1970 16 -f--$-h-4-- - - - - --A- --A- - ® NURSKRU GRANS HINV PERIMETER ZINVlE PROPOSED PROJECT ROADS O INTERSTATE ROAD 6 TANGB0 PROPOSED BLOCKS EXISTING BLOCKS RAILWAY NURSERY SPOSSIBLE FACTORY SITES 0 5 10KM Note Thi. map is bnsed o IBAD 2112, Dcmber l9, only the nome of the country has ben changsd from Dehom,ey to BnIn