Cameroon - Rural Development Fund Project

Document of The World Bank FIR OFFICIAL USE ONLY Report No. 6282 PROJECT PERFORMANCE AUDIT REPORT CAMEROON RURAL DEVELOPMENT FUND PROJECT (CREDIT 723-CM) June 27, 1986 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. WEIGHTS AND MEASURES 1 hectare (ha) = 2.47 acres 1 kilometer (km) = 0.62 miles 1 square kilometer (km2) = 0.39 square miles 1 kilogram = 2.2 pounds 1 metric ton = 2,205 pounds 1 liter = 1.057 US quarts ABBREVIATIONS FONADER - Fonds National de D&veloppement Rural (National Rural Development Fund) FSAR - Fonds Spfcial d'Actions Rurales (Rural Development Fund) GDP - Gross Domestic Product IRAF - Institut de Recherche Agronomique et ForestiAre (Institute for Agronomic and Forestry Research) OED - Operations Evaluation Department PCR - Project Completion Report PPAM - Project Performance Audit Memorandum PPAR - Project Performance Audit Report RDF - Rural Development Fund (Project) SAR - Staff Appraisal Report SEMRY - Socift& d'Expansion et de Modernisation de la Riziculture de Yagoua (Agency for the Production and Modernization of Rice Cultivation) ZAPI - Zones d'Actions Prioritaires Int&gr6es (Priority Zones for Integrated Action) COUNTRY EXCHANGE RATES Name of Currency (abbreviation) CFAF Currency Exchange Rate: - Appraisal Year Average 1977 US$1.00 = CFAF 245 - Intervening Years Average 1978-83 US$1.00 = CFAF 305 - Completion Year Average 1984 US$1.00 = CFAF 460 FOR OFFICIAL USE ONLY THE WORLD BANK Washington. D.C. 20433 U.S.A. Office of DirectoCeriwal Operstons Evaluatum June 27, 1986 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Performance Audit Report: CAMEROON Rural Development Fund Project (Credit 723-CM) Attached for information is a copy of a report entitled "Project Performance Audit Report - Cameroon Rural Development Fund Project (Credit 723-CM)" prepared by the Operations Evaluation Department. Attachment This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization. FOR OFFICIAL USE ONLY PROJECT ?ERFORMANCE AUDIT REPORT CAMEROON RURAL DEVELOP14ENT FUND PROJECT (CREDIT s23-CM) TABLE OF CONTENTS Page No. Preface ..................................................... i Basic Data Sheet ................................................... ii Evaluation Summary ................................................. iii PROJECT PERFORMANCE AUDIT MEMORANDUM I. PROJECT BACKGROUND ...................................... 1 A. Country and Sector Background ............................ 1 The Country ......................................... 1 The Agricultural Sector ............................ 1 B. Project Context .......................................... 3 C. Project Objectives ....................................0.. 4 II. PROJECT IMPLEMENTATION ...................................... 4 A. The Water Supply Subprojects ............................ 4 B. The Production-Oriented Subprojects ..................... 5 C. Institutional Development ............................... 5 D. Project Impact .......................................... 7 Water Supply Subprojects ............................ 7 Production-Oriented Subprojects ......................... 7 E. Project Follow-Up and Sustainability .................... 9 Follow-Up on Bottomland Development ..................... 9 Follow-Up on Institutional Development ................... 10 F. Human Resources Development .............................. 10 III. FINDINGS AND ISSUES ....................................... 11 Water Supply Subprojects ................................ 11 Production-Oriented Subprojects ............................. 11 Project Management ........................................ 12 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 Word Bak authodiation. TABLE OF CONTENTS (Cont'd) Page No. PROJECT COMPLETION REPORT I. Tntroduction ............................................. 15 II. Project Formulation and Processing ....................... 16 III. Implementation ......................................... 26 IV. Institutional Performance ******************************** 34 V. Operational Performance and Project Impact ............... 39 VI. Financial and Economic Performance ....................... 49 VII. Bank Performance ........................................ 52 ANNEX Tables 1. Project Cost and Financing ............................... 55 2. Partial Farm Budget ..................................... 56 3. Key Indicators at Completion Date ........................ 57 ATTACHMENT I: Comments from the Borrower ........................ 59 PROJECT PERFORMANCE AUDIT REPORT CAMEROON RURAL DEVELOPMENT FUND PROJECT (CREDIT 723-CM) PREFACE This is a performance audit of one of the first three rural development projects financed by the Bank in Cameroon. The other two, identified at the same time as this project, were the Western Highlands Project (Credit 784-CM) and the Priority Zones for Integrated Action Project (Credit 776-CM). Both have also been completed and are expected to be audited in FY87. Credit 723-CM, in the amount of US$7 million, was approved in June 1977 and became effective in January 1978. The closing date of December 31, 1981 was extended by two years but final disbursements were made only in August 1984. An undisbursed balance of about US$ 450,000 was cancelled. The audit report consists of an audit memorandum (PPAM) prepared by the Operations Evaluation Department (OED) and a Project Completion Report (PCR) of June 28, 1985 prepared by the Western Africa Regional Office on the basis of a completion study prepared by a local consultant firm for the Cameroon government. The audit report is based on a visit to the project area in April 1985, w,iich preceded the publication of the PCR, a review of the appraisal repo . (1479-CM) of May 26, 1977, the President's report (P-2052-CM) of May 10, 1977, and the corresponding credit agreement. Correspondence with the Borrower, supervision reports and internal memoranda on project issues contained in relevant Bank files have also been consulted, and Bank staff associated with the project have been interviewed. The PCR provides a good review of project implementation experience. It is lacking, however, a detailed review of how this project's experience has been taken into consideration in the follow-on project and of the project's overall sustainability. Points discussed in the PPAM are similar to experience with many other rural development projects in Africa and elsewhere. The draft report was sent to the Borrower for comments on May 2, 1986. Comments received have been taken into account in the preparation of the final report and are reproduced as Attachment I. The valuable insights, assistance and facilities provided by government staff in Yaound4 and Cameroonian and expatriate staff of the Rural Development Fund (FSAR) in Marouz is gratefully acknowledged. PROJECT PERFORMANCE AUDIT REPORT CAMEROON RURAL DEVELOPMENT FUND PROJECT (CREDIT 723-CM) BASIC DATA SHEET KEY PROJECT DATA Appraisal Actual or Actual as X of Estimate Estimated Actual Appraisal Estiteate Total Project Cost (US0 million) 10.6 12.1k 121.0 a/ Credit Amount (US$ million) 7.0 6.5 94.0 Amount Cancelled (US$ million) - 0.5 - Date of Board Approval 06/30/77 Date of Signing 08/04/77 Date of Effectiveness 01/30/78 Date Physical Components Completed 06/81 12/83 Proportion of Physical Components Completed - 70.0 Closing Date 12/31/81 12/31/83 Economi Rate of Return - - Institutional Performance good Number of beneficiaries 2,400 farm families STAFF INPUT (staff weeks) F76 FY71 FY78 FY79 FY80 FY81 FY82 FY83 FY84 FY85 Identification/Preparation 26.1 12.1 - - - - - - tiegotiation - 6.6 - - - Appraisal - 85.4 - - - - - - Supervision - - 10.1 19.3 16.3 28.3 7.9 6.1 3.4 14.2 TUTAL 26.1 104.1 10.1 19.3 16.3 28.3 7.9 6.1 3.4 14.2 CUMULATIVE DISBURSEMENTS FY78 FY79 FY80 FY81 FY82 FY83 FY84 FY85 Appraisal Estimate (US$ million) 1.2 2.8 4.6 6.2 7.0 - - - Actual (US$ million) - 1.0 1.6 2.7 3.8 4.7 6.3 6.5 Actual as 2 of Estimate 0 16 35 44 54 67 90 93 Date of Final Disbursement August 1984 MISSION DATA Date No. of Specializations Performance Types of Mission (Mo./Yr.) Persons Represented lb Rating /c Trend jd Problems /e Identification 11/75 - - - - - Preparation 04/76 - - - - Pre-Appratal 08/76 - - - - - Appraisal 11/76 5 - - - - Supervision 1 06/79 2 b,e 2 1 -i,F,T Supervision 11 12/79 2 b,d 2 1 M Supervision III 04/80 2 b.e 2 1 M,F,P ';uperaviston IV 11/80 2 b,c 2 1 1,F,t Supervision V 05/81 2 a,b 2 1 F,P Supervision VI 10/81 1 b 2 1 F,P sunervison VII 03/82 1 c 2 2 M,P Supervision VIII 12/82 4 e,dfe 2 2 M,P Supevison 1 0/82 1 c 2d*8, Supervision IX 03/84 1 c 2 1 M,P Supervision X 10/84 1 c 2 1 M'p OTHER PROJECT DATA Borrower Government of Cameroon Executing Agency FONADER Follou-,,n Project: FSAR Itf/ - Credit Number 2567-CM - Credit Amount (US$) 25.5 million - Date Board Approval 06/06/85 a/ Cost overrun due to underestimate of cohLs, high domestic inflation and extended iaplementation period. a - agriculturalist; b - agricultural economist; c - financial analyst; d - economist; e * rural engineer; f - agronomist. c/ I - Problem Free or Minor Problems; 2 - Moderate Problems; 3 - Major Problems. d/ I - liaproving; 2 - Stationary; 3 - Deteriorating. e/ F - Financial; M - Managerial; T - Technical; P - Political; 0 - Other. f/ SAR - Cameroon, Second Rural Development Fund Project, Report No. 4780-CM, - iii - PROJECT PERFORMANCE AUDIT REPORT CAMEROON RURAL DEVELOPMENT FUND PROJECT (CREDIT 723-CM) EVALUATION SUMMARY Introduction The Rural Development Fund Project in northern Cameroon is one of the first three rural development projects identified (in 1976) and approved in Cameroon. As such, it has been designed mainly as an institution building and pilot operation with emphasis on rural water supply. The other two projects were the Western Highlands Rural Development Project and the Priority Zones for Integrated Action (ZAPI) project. The PCRs of both projects are being finalized. Objectives The Rural Development Fund Project was designed to develop, strengthen and suppo..t Cameroonian institutions involved in rural development and thus increase their capabilities to carry out high-priority, small-scale development projects throughout the country. To this end, it has been necessary to create an institution capable of designing, funding and adminis- tering such small-scale infrastructure or production oriented projects. Its main components were the rehabilitation or construction of wells and bore- holes, their equipment with pumps, the construction of small mountain dams and a few small, production-oriented subprojects. Project costs were estimated at appraisal around US$8.5 million equivalent. It ended up costing about 50% more. Implementation Experience In early 1978, six months after appraisal, a Directorate ("Fonds Spfcial d'Actions Rurales" - FSAR) was established within the National Rural Development Fund (FONADER) for the purpose described above. Staffed overall with high quality technicians assisted by a team of expatriate specialists, FSAR's performance was good throughout the project. Lacking, however, the Aecessary support from other government services in the region, it shifted its activities from overall coordination and planning to day-to-day project i mplementation. Through FSAR, the project provided funding for subprojects to be executed by regional development agencies or existing government services. Implementation of these subpr-jects was uneven; those executed by experi- enced, well managed agencies met all their targets while those to be executed by government services did not. For instance, despite the provision of technical assistance and considerable logistical support, the Ministry of - iv - Mines (Groundwater Division) drilled only 232 boreholes (equipped with hand pumps) during the extended project period, against the 430 planned. The shortfall was due to an "over-bureaucratization" of the implementation teams, which should have been more action oriented. Similarly, the heavy admiis- trative set-up of the Ministry of Agriculture's Rural Engineering Service proved to be ill-adapted to implement effectively the teni small dams planned for water catchment in the Mandara mountains. As a result, but also because of inadequate preparation, the funds earmarked for the whole dam program were spent for the construction of only four dams. FSAR's management of project accounts was good throughout the proj- ect and costs were tightly controlled. Nevertheless, a higher than antici- pated domestic tnflation rate, coupled with the underestimation of dam construction costs, led to a cost overrun of 51% at project completion. The project's recurrent cost implication for government is about US$0.5 aillion per annum, of which a large part can be attributed to the maintenance costs for the hand pumps installed on the boreholes. Although the Credit Agreement contained a covenant according to which government was to propose, by the end of 1978, a beneficiaries' participation scheme to recapture part of these costs, none was submitted because of the politically sensitive nature of such a proposal. Given government plans to install about 4,000 pumps throughout the Sahelian zone by the year 2000, the need for a national policy governing this issue had become acute by the time the follow-up project was being prepared. Deliberations between government and the Bank on how best to address this subject led to a substantial delay in processing the follow-up project, the Second Rural Development Fund Project, which was appraised in May 1983 but signed only in Jane 1985. Meanwhile, government (a) enacted in 1984 a decree providing a legal basis for beneficiaries to pay for water; and (b) created a National Water Committee in May 1985 for the purpose of developing and coordinating the policies and activities of the various Ministries involved in water supply. The project achieved substantial results in the rehabilitation of some 900 existing wells (50% more than originally envisaged), drillings of 232 new boreholes (against 430 envisaged) and provision of pumps for these water points. It has thus increased water availability to the villages and reduced the effort needed to fetch the water, a job usually done by women. The project was less successful in the construction of small mountain dams (four against ten planned) in areas where ground water is difficult to tap. The impact of the water supply system on the regional health conditions is not yet known. The production-oriented subprojects, all of a pilot nature and corresponding to less than 15% of total project costs, including their share of overheads, achieved mixed results. The bottomlands development component was a failure, in the short run because of the drought which has beset the region in recent years but also probably in the long run because of the com- ponent's overall technical and economic conception. This component has had a disruptive effect on the local villagers' economy (see PPAM, para. 24). The market gardening component succeeded in expanding onion production but failed in its diversification objaictives because of inadequate research and extension. Marketing of onion proved to be a difficult and yet unresolved - v - challenge. Bullock fattening, although giving limited financial results in the beginning, has a potential for further expansion and improved income generation. Lack of proper monitoring makes it impossible to determine the economic (or even financial) returns on these small components. Institu- tional impact has been limited by the lack of project attention to staff training. Sustainability Tht factors which affect most the sustainability of the project are: (a) a lack of proper cost recovery to meet the maintenance costs of the water supply system; (b) lack of technical packages and the weakness of government agencies whose services are necessary to achieve substantial production increases; and (c) lack of systematic training for government staff, other than the limited on-the-job training of water supply maintenance crews. In the shnrt term, the sustainability of the project depends on the Second Rural Development Fund project, signed in June 1985 after protracted negotiations dealing with the cost recovery issue. The other elements of longer term sustainability are also to some degree being addressed by the second project. Findings and Lessons Maintenance of the water system is putting considerable financial and institutional strains on the responsible agencies. Besides the cost recovery policies sought by the follow-on project, but for which political commitment has proved difficult to generate, villagers should be encouraged to maintain the equipment by relying on the private sector and their own initiatives (PPAM, paras. 27 and 31 and PCR, para. 5.27). The engineering and cost aspects of small mountain dams had not been properly studied (PPAM, para. 32; PCR, para. 7.01). The absence of adequate technical packages has reduced the effec- tivenesa of the maAet gardening and the stall fattening components (PPAR, para. 33). Dual ownership of gardening pumps has hampered loan recovery efforts (PCR, para. 5.07). The rate of loan recoveries depends to a large degree on the competence of credit staff (PCR, para. 5,16). The rate of progress of a subproject is essentially in the hands of the implementing agency, not the coordinating agency. Therefore, the choice of a capable implementing agency becomes a key to those subprojects' success (PCR, para. 4.06). Disparity in remuneration of the staff of the coordinating agency and of implementing agencies creates a poor working environment (PCR, para. 4.10). Bank supervision was intensive but concentrated too much on infra- structure and not enough on institution-building (PCR, paia. 7.03). - 1 - PROJECT PERFORMANCE AUDIT MEMORANDUM (,AMEROON RURAL DEVELOPMENT FUND PROJECT (CREDIT 723-CM) I. PROJECT BACKGROUND A. Country and Sector Background The Cuntry 1. The United Republic of Cameroon is one of the rare African coun- tries enjoying political and social stability, a considerable economic poten- tial and virtual self-sufficiency in food. At the same time, it faces accelerating population growth (presently about 9 million and expected to reach 15 million by the year 2,000, by which time 57% of the population is expected to live in urban areas, compared to 33% in 1980). A limited range of exports (coffee and cocoa alone accounted for close to 50% prior to the development of oil production), vulnerability to dr.ught, limited physical infrastructure, institutional weaknesses and a lack of indigenous entrepre- neurs resulting in part from the social relationships prevalent in most African countries are other features of the country's economy. 2. The government has been successful in avoiding economic disruptions and a collapse of agriculture which could have resulted from the oil boom. Cameroon'a fiscal and monetary policies have remained prudent. Government expenditures have been financed mainly from its own revenues. Some of the oil revenues generated since 1978 have been placed abroad, taking into account the limits of domestic investment absorption capacity. There has also been less recourse to external financing. On the other hand, interest rates have become negative in real terms in recent years, resulting in implicit subsidies and credit rationing. 3. The last two decades have seen an era of sustained economic growth (6.3% per annum in the seventies, up to 8% in thz early eighties). Per capita income rose at an average rate of about 3.1% per annum from 1966 to 1980. It was estimated at around US$880 in 1982. The Agricultural Sector 4. The economy remains essentially based on agriculture, especially considering that oil production is expected to have a limited lifespan. Not- withstanding its decreasing relative importance, agricultural production has grown faster than population, and the rural sector still accounts for about a quarter of total exports and a quarter of GDP. It employs about 70% of the labor force. One-third of the country's land area is classified as agricul- tural, but less than 20% of cultivable land (2 million ha) is actually being tilled. Yields are low, but agricultural production is diversified, food - 2 - production has been growing, and eA efficient network has been established for food distribution to the main cities. 5. Traditional smallholder farming (about one million farm families) provides 93% of agricultural production. The remaining 7% come from industrial plantations. The productivity of small farms has been slow to increase mainly because of lack of extension services, bad road infrastruc- ture and poor input supply and marketing arrangements. There is also growing concern about the exodus of young males from rural areas, resulting rural population imbalances and growing urban unemployment. To meet growing urban demand, agricultural productivity will have to increase substantially. The ongoing fifth development plan emphasizes investment in agricultural parastatals, notwithstanding their past lackluster performance. 6. Some of the key issues in agriculture relate to (a) efficient export crop production; (b) improving the quality of cash crops; (c) raising the productivity of small farmers; (d) government action on agricultural prizing and other policies; (e) how to improve and whether to expand agricul- tural parastatal companies; (f) how far to pursue food self-sufficiency; (g) the balance to be achieved between economic efficiency and socio-political objectives; and (h) the need to generate agricultural employment fast enough to absorb more than one fou-th of the future labor force increase. 7. The implementation of agricultural development is generally carried out by parastatal and government agencies, while the central services of the Ministry of Agriculture focus on policy-making, planning, budgeting, monitor- ing and evaluation. Among the development agencies, some are crop or area specific, such as the Cocoa Development Company (SODECAO), the Agency for the Production and Modernization of Rice Cultivation (SEMRY)1/ or the various tree crop parastatals.2/ Other agencies are specialized along functional lines such as seed multiplication or agricultural machinery. Entrusting the implementation of a development project to a union of former cooperatives, as was the case in the Western Highlanda project (Credit 784-CM), is a very exceptional occurrence. Rural development operations are financed through the National Fund for Rural Development (FONADER), iscussed in the following paragraph. 8. The broadest mandate was given to FONADER, established in 1973 to organize and supervise both rural development programs and agricultural credit operations. FONADER is legally a separate financial entity, under the tutelage of the Ministry of Agr±culture, but with administrative autonomy. It operates from its Yaound6 headquarters but has been decentralizing its activities since 1978. However, it still relies on other government field services for supervision and execution of activities for which it provides finhacing. 1/ See PPAR - Cameroon Second SEMRY Rice Project (Loan 1512-CM/Credit 763-CM), OED Report No.5156, dated June 25, 1984. 2/ See PPAR - Cameroon Coastal Treecrop Plantations, OED Report 6019, dated December 31, 1985. - 3 - 9. By contrast with these development agencies, which have financial and operational autonomy, the services of the Ministry of Agriculture at the provincial and departmental levels are hampered by insufficient and inade- quately trained personnel, low salaries and a lack of coherent work programs and the means to implement them. A program for financing urgent, small and diversified rural needs which are not being met under existing projects and programs has been considered essential to complement the activities of the rural development agencies. B. Project Context 10. The Rural Development Fund project, focusing on the Mandara Mountains and their immediate surroundings in Northern Cameroon, was identi- fied by the Bank's Regional Mission in West Africa in 1975/76 along with the Western Highlands Rural Development project and the Priority Zones for Integrated Action project (ZAPI) in the East. The latter was prepared and initially supervised by the Rural Development Division,3/ The three projects, initiated in the early years of the Bank's poverty orientation in agricultural lending, represented a first generation of Bank lending for poverty oriented rural development in Cameroon. Agricultural projects previously financed by the Bank in that country dealt with tree crop estates, livestock and large scale irrigated resettlement (SEMRY) or mechanized rice cum soja (Plaine des M'bo) operations.4/ 11. The project reviewed in this audit was located in a densely populated area of the country, where very small farming units predominate; this choice was in line with government priorities set out in the Fourth Development Plan (1977-81) in favor of the traditional smallholder sector. For social and political reasons, the project aims at improving agricultural and living conditions in its area, as an alternative to resettlement in the undeveloped regions of the country, notwithstanding the substantial pressure on land resources in the project area. The Rural Development Fund project agglomerated a large water supply component with a number of small, somewhat pilot, subprojects to be implemented each by a different government agency under the coordination of a single planning entity. 12. The Rural Development Fund Project was formulated with a view to create the institutional capacity to plan and execute small rural projects of high priority, with initial emphasis on the Northern province. Planning and coordination of the project was entrusted to FONADER. Within FONADER, the "Fonds Special d'Actions Rurales" (FSAR) was created soon after loan signature as a directorate responsible for this project and for future opera- tions, countrywide, which would eventually stem from this project. Both 3/ The ZAPI project has recently been completed. Its PCR is expected to be available by mid-1986. The PCR for the Western Highlands project is being redraf ted. Both projects are expected to be audited by OED in FY87. 4/ PPAR, Cameroon - Plaine des M'Bo Rural Development Project (Credit 672-CM), OED Report No.4362, dated March 28, 1983. -4- FONADER and FSAR enjoy operational autonomy although they fall under the supervision of the Ministry of Agriculture. This autonomy allowed them to perform much more satisfactorily than the other government agencies responsi- ble for the implementation of the various subprojects (see PPAM, para. 17). C. Project Objectives 13. The project was mainly oriented at providing a water supply infra- structure. Two of its five subprojects, corresponding to about 60% of all project costs (both at appraisal and in actual achievements) dealt with the construction, rehabilitation and equipment of wells and boreholes and with the construction of small dams, higher in the mountains, where underground water was not readily available. The three production-oriented subprojects plus funds for activities to be identified during implementation were expected to absorb 20% of project costs. At completion, they had utilized only 10% of total costs, whereas project administration, including technical assistance, absorbed 30% of project costs instead of about 20% initially envisaged. Thus, little was planned by the project in terms of incremental agricultural production, and little was achieved. II. PROJECT IMPLEMENTATION A. The Water Supply Subprojects 14. The project was designed to rehabilitate some 600 existing open wells and cap 60 of them; drill about 430 boreholes; and equip about 490 water points with hand or pedal operated pumps. Ten small dams were also to be built in the Mandara mountains for human and animal consumption. In physical terms, 900 wells were rehabilitated, compared to 600 planned at appraisal, a substantial achievement even if in the drought year of 1983 one third of them were found to be dry or have insufficient water. The boreholes component reached only about 50% of appraisal targets, or 232 out of 430. More than 75% of the drillings were successful. The pumps component reached slightly more than 60% of appraisal targets. All three components have suffered from sand infiltration, which clogs up the water points, reverting rehabilitated wells to their original condition within a year or two, and leads to pump breakdowns. The maintenance problems of the water points are discussed in the section on institutional development (see PPAM, para. 21). 15. Only four of the ten projected dams were built. However, because of massive cost overruns with respect to the unduly low appraisal estimates, which were based on nothing more than preliminary identification data (see PCR, paras. 3.10, 3.11 and 5.20 through 5.23), the funds earmarked for 10 dams were totally spent on the four dams. All four dams were located within the densely populated mountain areas inside the Mora-Meri-Mokolo triangle, on sites of very difficult access. Their choice was based, in principle, on a preliminary inventory of 151 potential sites classified in terms of their distance from population centers, the technical advantages of the sites and a rational geographical spread of the dams. A first identification made by a consultant firm reduced the choices to 57, six of which were classified as - 5 - "very a4vantageous" in terms of cost, three as "advantageous" and another six as average. Only one of the four dams finally built was in the first cate- gory, one in the second and the remaining two in the third. None of the four or five sites with a capacity to reach the highest number of people were selected, indicating that technical and social considerations may, in some cases, have weighed less than political considerations. All four dams are quite small structures with a reservoir capacity going from 22,000 m3 for the smallest to 144,000 m3 for the largest. The second largest (89,000 m3), not fed by a permanent stream, was not yet completely filled at completion of the project. The third largest (47,000 m3) lost water when filled and required additional works. All these difficulties appear to be the result of the absence of good engineering studies, which should have normally followed the preliminary inventory and were intended to be implemented during the first year of the project, but for which no funds were earmarked as it was felt at the same time that such studies would have increased the cost of what was intended to be inexpensive works. The burden of designing the dams as they were built fell on the expatriates who were successively given the job of building the dams. No wonder that that position witnessed a turnover of three expatriates over the life of the project. B. The Production-Oriented Subprojects 16. The three production oriented subprojects comprised the development of 300 ha of bottomlands for irrigated rice production; small scale market gardening through the provision of loans for the purchase of 60 motor pumps, each to be owned jointly by two farmers; and the establishment of a line of credit to provide loans for some 2000 stall fattening operations. Although in terms of physical implementation all three subprojects claim to have substantially reached their appraisal objectives, with 300 ha equipped for rice cultivation, 60 loans extended for water pumps and more than 1300 loans distributed for stall fattening, only the latter subproject has achieved any immediately replicable results. The dyke built for bottomland irrigation has never held enough water, especially in the recent years of drought. Market gardening, for lack of proper extension, was limited to onion production, which rapidly exceeded the marketing (if not the market) capacity of the region. Stall fattening resulted in less animal weight gain than projected, attributable indirectly and in part to an insufficient number of extension- ists and possibly to the insufficient size of individual loans (although higher than envisaged at appraisal), but succeeded to increase the income of mountain farmers in areas with few options for diversification and to gradually orient this type of livestock operation from exclusively ceremonial use purposes towards more commercial objectives. C. Institutional Development 17. Since most of the project's production oriented activities were conceived as pilot, learning, operations, a key though implicit objective of the project was to achieve institutional development. The concept of basing institutional development on pilot operations and socially indispensable activities like water supply is a laudable one. However, overall, it appears to have been implemented in an amateurish fashion. The project did not initially provide for staff training. A small training component was later - 6 - introduced at the instigation of a supervision mission. The project imple- mentation agency, i.e., the FSAR office in Maroua (DOFSAR), lacked proper financial control mechanisms until its expatriate chief was replaced by a more capable Cameroonian in 1980. 18. DOFSAR lacked proper means of communication with the central FSAR office in Yaound6, relying on a rarely functioning telephone, which, in the absence of a telex, could have usefully been complemented by radio communica- tions. Possibly becaus; of higher pay, FSAR and DOFSAR enjoyed exceptional staff stability. At the same time, the difference in pay levels created friction between FSAR and the other agencies responsible for the implementa- tion of various project components. Those other agencies, namely the Rural Engineering Service of the Ministry of Agriculture (Genie Rural), the Hydro- ge.ological Service of the Ministry of Mines and Energy (PES), the Institute for Agronomic and Forestry Research (IRAF) and the "SocietG d'Expansion et de Modernisation de la Risiculture de Yagoua" (SEMRY) performed below expecta- tions. 19. Both the Genie Rural and PES had little implementation experience, were poorly organized and suffered from low salaries. SEMRY, which has otherwise been developing rice cultivation in Cameroon very successfully, apparently did not feel committed to the bottomland irrigation subproject entrusted to it, because of its other priorities and because of the distance of SEMRY headquarters from the bottomlands to be developed. Results of research into market gardening, undertaken by IRAF, became available only four years after pumps had been distributed to their potential users. 20. Extension services were almost non-existent both in the market gardening and stall fattening subprojects. Although the weakness of exten- sion and other agriculture services was well documented, the project made no attempt at improving them. As a result, whatever success was achieved in the physical implementation of the project was achieved mainly by FSAR staff, who got excessively involved in day-to-day operations at the expense of the broader objective of developing FSAR into an institution capable of identify- ing and preparing rural investment projects and later coordinating their implementation (see PCR, para. 4.06). 21. The most successful components of the project, such as the mainte- nance of project financed vehicles and pumps or the management of stall fattening credits and their recovery are to be attributed to the capability and dedication of a few individuals or groups of individuals rather than to the establishment of a successful overall management system. With respect to the maintenance of wells, there appears to have been confusion as to the respective responsibilities of the Genie Rural and FSAR, with the latter end- ing up doing the job. Institutional weaknesses were also accompanied by an absence of systematic project monitoring, which explains the lack of reliable data on which to judge the project's impact to date. -7- D. Project Impact Water Supply Subprojects 22. There is little doubt that the water supply subprojects, which account for more than 75% of project costs, and particularly the wells and boreholes components have had a positive effect on the region. It appears that, even if the appraisal target of one water point for 560 people has not been fully met, more than a million people have benefitted from improved water supplies. This either by virtue of improved wells, boreholes and wells fitted with pumps, reduced distance of the source of water (1 km after the project, compared to 5 km before), amount of water consumed (25 liters per person per day, against 15), lesser effort in drawing water, or water supply for animals. The weak points of the wells and boreholes components related to sand infiltration, the costly and precarious organization of pump and well maintenance, absence of cost recovery and limited beneficiary involvement, all of which augured poorly for the longer term sustainability of the subproject (see PCR, para. 5.24). 23. Results achieved by the mountain dams are more ambiguous. No com- prehensive survey has been made of the number of people using the water, especially towards the end of the dry season, when supply is most precari- ous. The 1976 identification study estimated that the four dams which have since been built could, together, reach about 1400 families of fiv-- (each also owning a few sheep and goats) within a radius of 2.5 km, or 2500 families within a radius of 5 km-implying a walk of up to 10 km on very difficult terrain each time water is fetched. Insofar as this has been achieved, it still represents an improvement with respect to the before project situation. The dams are further used by nomad herders who temporarily settle around the dams. This entails serious risks of overgrazing, accelerated erosion and conflicts with the local population. The most serious risk, however, is that of disease. As observed by the audit mission, animals drink directly from the pond and defecate in it. This is officially prohibited but there is no one to enforce the prohibition. Women fetching water often dig holes some distance from the pond in the hope that water seeping into those holes will have been partially filtered through the earth. Filtering systems are being installed at some of the dams, but there is no clear indication as to how they would be managed and how the villagers would be educated to use them. A side benefit expected from the dams, is the production of fish. The fish population of the dams was too young to be harvested at the time of project completion, and fishing remained prohibited. The benefit from fish population, in terms of water oxygenation but more particularly of human diet, would be subject to the limitations of the reservoir size and particularly of the amount of water remaining behind the dams towards the end of the dry season. The size of the dams would also limit their potential impact on regional rainfall. Production-Oriented Subprojects 24. Production-oriented subprojects were all small, somewhat specula- tive pilot operations. Without enough water for irrigated rice production, the bottomland development project has lead to a disruption of activities and - 8 - serious losses for the farmers involved. Only about half the area managed tc be irrigated in 1983, when most of the infrastructure had been completed. Yields were far below those achieved by SEMRY in other schemes. No irriga- tion was possible in 1984, and, again in 1985, no activity could be observed at the time of the audit mission. Research into irrigated rice, a component of that subproject, was interrupted after two years, when the seed beds were destroyed in a flood. The November 1984 supervision mission concluded that "the investment into the development of 300 ha of bottomland in the Moulv iday region has most likely to be written off; it is almost certain that, -jue to an error in the topographical lay-out, the earth dam will never retain water." The land, the Bank wrote to FONADER, should be returned to the farmers for (traditional) food cultivation.5/ 25. The market gardening subproject achieved ambiguous results. For lack of extension and research, it failed to produce anything but onion, already traditionally grown in the region, or to improve the storage of onions. Since the beneficiaries of the sixty loans had been growing crops of lesser value before the project, their income increased considerably. How- ever, their yields after losses did not substantially exceed those of onion growers using traditional wells, although the owners of motor operated pumps saved on labor and were thus able to plant a larger area into onion. Market- ing limitations and inadequate storage practices may have resulted in more than 40% post-harvest losses (compared to 11% losses by non-project farmers). The PCR estimates the economic rate of return of the market gardening subproject at about 20% (see PCR, para. 6.02) but provides little evidence in support of that conclusion.6! It isn't clear, either, if the calculation takes into consideration that one-third of the pumps inatalled by the project were no longer functioning at project completion because of improper maintenance resulting from the dual ownership of the pumps. In the absence of monitoring, the impact of the heavy water tapping generated by this subproject on the local water table also remains unknown. Onion growers have been beset by enormous price fluctuations resulting from the absence of (a) prcper storage facilities; (b) better organized transport; and (c) inter- mediaries capable of larger purchases. Thus, the subproject may have had a negative impact on the traditional onion growers, but if this was the case, it remains undocumented. At the time of project completion, recovery on the amounts due by beneficiaries of the market gardening loans stood at slightly under 70%. 26. Out of the 2000 loans envisaged for stall fattening, the project disbursed 1330, i.e., about 65% of appraisal objectives. Again for lack of monitoring, there is no precise evaluation of the subproject's achievements 5/ Cameroon - Rural Development Fund Project Supervision Report, November 14, 1984. Following good rains in 1985, Bank staff report that the reservoir now contains water. However, there is no information as to how this water is being utilized and with what results. 6/ The PCR estimate is almost identical to the appraisal estimate, although actual net yields were only about one-quarter those expected at appraisal. The PCR also states (para 6.07) that the economic performance of credit recipients has not been evaluated. - 9 - or impact. It appears that the fattened animals gained less weight than anticipated, and this over a longer time period than initially envisaged. The shortfall is attributed to the unavailability of supplementary feed com- bined with difficulties in water supply; the use of lower grade animals than foreseen; and insufficient extension and credit funds. However, credit recovery, which had initially been poor, improved in the later years of the project. At completion, 78% of all credits had been repaid. Credit recipi- ents are believed to have achieved modest financial returns, but there appears to be substantial interest in stall fattening, and lessons from this project could help achieve better results in follow-up activities. The eco- nomic returns of the subproject have not been computed. E. Project Follow-Up and Sustainability 27. The RDF project is now being followed by a Second Rural Development Fund project appraised in May 1983 and approved by the Board, after protract- ed negotiations, in June 1985. Without such a follow-up, the sustainability of the first project would have been, at best, dubious. The long negotia- tions were due to a disagreement between the Bank and the Cameroon Government on cost recovery policies and beneficiary contribution to the maintenance of the wells, boreholes and pumps. The second project was appraised and approved prior to the completion of the first project and drafting of its PCR. Essentially, the second project does more of the same things financed in the first project and expands these activities to a wider geographical area. Some lessons of the first project appear to have been given proper consideration in the second project. Thus the Bank has insisted on proper cost recovery procedures, making the elaboration of a satisfactory policy a condition of disbursement for future well drillings. The government enacted a decree on water resources at the end of 1984, which provides the legal basis to make beneficiaries pay for water. It also created a National Water Committee to develop and coordinate the policies and activities of the ministries involved in water supply. The Committee was created just before President Clausen's visit to Cameroon in June 1985, when the second project was signed. The impact of the various government measures remains to be seen. Market gardening loans have been held in abeyance until a marketing study (financed by the new project) is completed. The second project does not finance any new mountain dams, considered to be too expensive. However, the government has made clear its intention of building more dams, since the population of the Mandara montains has no other source of water supply, and displacing these populations would be socially, politically and probably economically unfeasible. While dams are expensive, there is as yet no clearly proven cheaper alternative. The second project will attempt to drill boreholes in the mountains. If results of such attempts are not favorable, the Bank should again consider assisting the government in dam construction. However, there is no specific provision for that in the second project. Other lessons have been either improperly evaluated or even ignored. They are discussed in the following paragraphs. Follow-Up on Bottomland Development 28. On the other hand, the Bank should have been more circumspect about financing further bottomland development in the second FSAR project. - 10 - Notwithstanding the very poor performance of the bottomlands subproject of the first project (see para. 24), which had became clear by the time the second project was presented to the Board, the appraisal report of the second project distributed in May 1985 but relying on a 1983 mission's findings, gives an uncautiously optimistic evaluation, mentioning that 300 ha of bottomland were "developed" (i.e., provided with physical infrastructure), but remaining totally silent about the Bank's own conclusion that there was little hope of ever growing any rice on those 300 ha.7/ The SAR of the second project proposed extending bottomland development to another 1,000 ha in the same region without proper indication of why this new component was expected to be more successful than its predecessor. The skimpy financial information available in the SAR of the second project shows, furthermore, that net returus per day of labor would be lower with the subproject than without - a consideration to which African farmers are very sensitive.8/ Overall returns would less than double while the area planted by each farmer and their laber input would both double, implying that more profitable uses of both land and labor could have been considered. All these aspects should have been taken into consideration even though the new bottomlands subproject represents only a small fraction of the second project. At the time of this audit, however, the region has already substantially modified its technical proposals with a view to a simpler and less costly development of bottom- lands. Follow-Up on Institutional Development 29. The institutional difficulties of the first project steamed mainly from the weaknesses of the various agencies of the Ministries of Agriculture and of Mines and Energy. The second project relies much less on such agencies than did the first, entrusting many project components to more reliable though somewhat peripheral institutions which have long proven records of capable management. The main project component, water supply, would, however, remain under the Ministry of Agriculture, with coordination with other ministries to be provided by a newly created National Water Committee, about which little can be known at this stage. Furthermore, extension for bullock fattening and the health aspects of water supply would have required an institutional strengthening which the new project does not provide. F. Human Resources Development 30. The project has provided no training for Cameroonian nationals. Through its provision of more ample and more readily available water resources to a large population, it is believed to have had a positive impact on the health of those populations. However, there are no monitoring data or 7/ SAR - Cameroon - Second Rural Development Fund Project, Report No. 4780-CM, dated May 10, 1985; para 1.14. 8/ SAR - Cameroon - Second Rural Development Fund Project, Report No, 4780-CM, dated May 10, 1985; Annex 6-2, p.57. - 11 - studies on that subject. Women, traditionally responsible for fetching water, have particularly benefitted from the closer availability of water and may have put to better use the time they save as a result of the project. Here again, there is no study of the project's impact on the women's daily occupational patterns. III. FINDINGS AND ISSUES Water Supply Subprojects 31. The project has been quite successful in the implementation of its well rehabilitation and borehole construction program, although the mainte- nance of all the wells, boreholes and pumps is putting considerable strain on the limited institutional capacity available in the region, and the issues related to sand infiltration are not yet resolved. It appears that in some places, villagers have taken the initiative to have defective pumps repaired at their own expense in private shops. Such initiatives should be regularly encouraged rather than viewed simply as proof of failure of government agencies in maintaining that equipment. Properly designed maintenance cost recovery procedures could further enhance villagers' initiatives along that line. 32. The mountain dams subproject has met only a small part of its objectives, operating as it was in a "mission impossible" situation. Taking into consideration the social and health implications of safe water supply in the Mandara mountains, the Bank should not have totally ignored this line of activity in its follow-on project. Dams which have already been built will require improvements both in the quantity and quality of the water deliver- ed. Local populations need to be educated to avoid a dangerous pollution of the stagnant waters accumulating behind the dams. And, if the government continues building new dams (PPAM, para. 27) and alternative solutions fail to deliver, it is hoped that, Bank assistance will remain available to government in its efforts. The major error of the Bank in the mountain dams subproject was its decision to proceed with a complex engineering operation on the basis of nothing more than a preliminary identification study (PPAM, para. 15). Doing without further studies and adequate designs on the argu- ment that they would have cost too much shows to what extent economic reason- ing can be distorted by bureaucratic and political pressures (see PCR, para. 3.11). Production-Oriented Subprojects 33. The bottomlands subproject apparently suffered from the same type of bureaucratic optimism, which may have persisted into the second RDF project even after the first bottomland development resulted in almost certain total failure. The main weakness of the market gardening and stall fatcening subprojects was the absence of proper technical assistance for lack of sufficient and qualified personnel and, in the case of the gardening operation, of proven technical messages to transmit. The market gardening subproject should also have paid more attention to the downstream (marketing, - 12 - storage and transport) requirements of large scale onion cultivation and its potential negative impact on neighboring traditional onion growers. Project Management 34. The project identified, and rightly so, the lack of investment planning and financing mechanisms, as a key hurdle to the development of the Mandara mountains and surrounding lowlands. Having established a project planning and coordination unit, FSAR, within a financial inst!.tution, FONADER, the project needed to collaborate with government agencies which were both ill-equipped to provide the services indispensable to the good implementation of the project and reluctant to cooperate with the better paid staff of a "newcomer" institution. In that situation, project authorities, rightly sensing that they would be judged on the short term achievements of the project, channeled their efforts into day-to-day infrastructure construc- tion activities and away from coordination and institution-building objee- tives. As a result, the sustainability of the project components which did not fail remains dependent on follow-on financing. The Maroua office has a few young and competent Cameroonian technical staff. But, its general administration, lacking means of communication with Yaound6 headquarters, has possibly remained too conservative and bureaucratized. 35. Institutional development efforts of the second RDF project aim at strengthening FSAR's project management capability, including the supervision of a few special studies and the always difficult-to-handle monitoring and evaluation activities. Agricultural planning, however, would be developed within the regional structures of the Ministry of Agriculture. Implementa- tion of the identified subprojects would be entrusted to successful parastatal enterprises. The second project continues, however, to rely on weak government agencies for agricultural extension and health and sanitation activities. Some progress appears to have been made in the issue of cost recovery in the rural water supply components. The adoption of a satisfac- tory policy for the recovery of operation and routine maintenance costs has been made a condition of disbursement for the well-drilling component. 36. The RDF project was a relatively small and somewhat pilot opera- tion. As such, it has avoided many of the mistakes of more "intregrated" rural development projects with an excessive number of components. The experience it has generated should help achieve better results in follow-on operations. The test of whether its immediate achievements in terms of infrastructure will be sustained by proper maintenance and further expansion still lies ahead. - 13 - CAMEROON RURAL DEVELOPMENT FUND PRCJECT PROJECT COMPLETION REPORT June 28, 1985 Projects Department West Africa Region ・一／ゾー－, が、戸タグ・ご冷hダ - 15 CAMEROON RURAL DEVELOPMENT FUND PROJECT (CREDIT 723-CM) PROJECT COMPLETION REPORT 1. INTRODUCTION 1.01 The Rural Development Fund Project, prepared by the Bank in April and August 1976 at Government's request, had been designed co develop, strengthen and support Cameroonian institutions involved in rural development and thus increase their canabilities to carry out high priority* Fmall-scale development projects throu#out the country. More than half of the project's development involved a water supply investment program. It also assisted farmers through small-scale, production-oriented sub-projects, that were designed as tests to be replicated elsewhere in Cameroon, if proven successful. Finally, it provided for a US$500,000 equivalent line of credit for sub-projects to be identified, prepared and appraised during project execution. Nationally, the project focused on a much needed structure for channelling funds for rural infrastructure development and was tailored to the Government's strategy as outlined in the Fourth Development Plan (1977-1981). 1.02 The Bank identified and prepared the project in October/November 1975 and April/August 1976, respectively. It was appraised in October/November 1976, the Credit was approved In Zune 1977, and declared effective in January 1978. The original closing date, December 31, 1981, was postponed twice: the first time because lmplementation took longer than expected and the second time to avoid a hiatus between the first and second phases; it was finally closed in December 1983. Total costs of the project amounted to US$12.8 million equivalent, US$6.6 million of which was fin&nced by the IDA Credit and the rest from Government's investment budget. The credit was closed on December 31, 1983 with an undisbursed balance of approximately US$448,000 which was cancelled. 1.03 The findings in this report are based on: (a) a completion study prepared in January 1984 on behalf of Government and the project unit by SEDA (consultants) of Yaoundil; (b) a completion mission carried out in February 1985; W a detailed review of project files at the Bank's headquarters; and (d) interviews with people in the country as well as Bank staff having been involved with the project. The main thrust of the completion mission's findings was discussed in a meeting with Government representatives; there exist no fundamental - 16 - differences of opinion regarding the conclusions of the mission between Government and the Bank. II. PROJECT FORMULATION AND PROCESSING A. Background and Project Origin 2.01 Two separate ideas were at the core of the initial formulation of the project. The first was to support rural development, particularly small projects by creating an institutional capacity to plan rural development actions and channel funds for importRnt activities, and to provide a management framework to accelerate execution - i.e., a rural development fund. This was in line with Government's priorities set out in the Fourth Development Plan (1977-1981) which, among other things, stressed more support to the traditional smallholder sector to help raise rural incomes and improve the quality of life. The second idea, which was to support regional development, is not specifically discussed in the official project 6ocumentation except in indirect reference to the importance of the sub-projects and to the low level of income in the Northern Province. In 1974/75, there was a considerable emphasis by Government to develop the North and extreme North of the country; this was reflected by the large number of donors formulating projects in that area at the time. Although there is little discussion on how or why five sub-projects under FSAR I were selected all in the northern region a RMWA back-to-office report (1975) mentions the importance of finding projects in this area. Having small sub-projects scattered throughout the country would have made the project impossible to administer and would have transferred control from the region(s) to the capital -- thus reinforcing the general tendency to overcentralize. B. Identification and Preparation 2.02 The R1WA identification mission of 1975, which was essentially a reconnaissance mission, identified three project possibilities: (a) a rural development project in the Western Highlands, (b) an extension of the ZAPI (Priority Zones for Integrated Development) project in the East and (c) the creation of a rural development fund. The third proposal, which was seen as a means to group several micro-projects given priority in the five year Plan, was received favorably by Government. The main problems identified were of an organizational nature, namely how to strengthen the government departments in charge of the micro-projects and how to assess FONADER's suitability to manage a rural developt'pt fund. 2.03 The preparation report that resulted from a visit to the country by a RMWA team in April 1976 reflects in essence the implementation of the project. The organizational structure, with FONADER as the lead agency, was established then; the five sub-projects remained and their concept and scope changed little from the preparation report (June 1976). The - 17 - following points made in the preparation report happened to have particular relevance later. - The small dams sub-project was limited to 10 dams because of the "physical difficulties expected in implementation," especially difficult access to dam sites and problems in organizing a large labor force. In spite of this caution, the requirements, especially of labor for the construction of 10 dams, to be considerably underestimated. - * The initial idea for bottomland development was for a simple project involving contour bunding and drains for better water control and, thus, for the cultivation of traditional rice varieties. This concept was later changed to a more sophisticated one, providing for a reservoir and enough water to cultivate irrigated rice. - The market gardening sub-project proposed a credit for both small motor pumps (40) and hand pumps (20), but indicated that the motor pumps (5HP were three times more expensive than the hand pumps (CFAF 330,000 versus CFAF 100,000). It was also noted that "no hand pumps are, however, at present operating in the area and the cost estimates are therefore mainly based on motor pumps". The pumps eventually provided by the project (3HP motor pumps) cost as much as the hand pumps. - The line of credit for stall fattening was initially intended for 900 farmers; 834 credits were actually distributed in the first four years of the project, out of the line of credit which had been increased during appraisal to 2,000 credits. - Under "Organization and Management" it was stated that "despite the weakness of their field services, ministries would be charged with executing the five identified sub-projects rather than having them executed by a special projects authority," an important principle which caused difficulties during implemention but which would probably have been sound, if 'xecuting agencies with more capacity had been selected. - The issue whether Government and the population would agree to contribute either unpaid labor or funds to the water supply and bottomland development sub-projects. Little or no progress was made in this area during implementation. 2.04 In August 1976, a RMWA pre-appraisal mission (one economist) visited the proposed project and its report, Preparation Report Addendum, made a number of observations and recommendations for changes. It proposed that the market gardening and bottomland development sub-projects be transferred from Agricultural Extension to G6nie Rural -- both departments of MINAGRI. He suggested that the bottomlands "...might not be suitable for the modest improvements proposed in the preparation report..." and that - 18 - maybe improvements should be made to allow the cultivation of irrigated rice. It also pointed out the weakness of information on market demand and market capacity for onions and other market garden crops, and proposed that a component be included in the market gardening sub-project for a program to find better means of storage. Finally, its rec)mmendation for financing the re-equipment of the vehicle and equipment workshops of G6nie Rural (only technical assistance was previously included) proved to be correct; however, the estimate (CFAF 3 million), was low compared with the actual expenditure (CFAF 16 million for equipment and CFAF 21 million for civil works). C. Apprqisal 2.05 No substantive issues were raised in the pre-mission project brief, which had not been flagged in the preparation report or the Preparation Report Addendum. The appraisal mission departed for Cameroon as planned on 22 October 1976. Probably due to the continuous dialogue maintained with Government during the preparation process and Government's strong interest in this type of action, the appraisal process was relatively free of problems. 2.06 Based on information gathered for the preparation mission, minor adjustments were made to some sub-projects. In July 1976, G6nie Rural carried out an inventory of the wells in the Northern Province as requested. Of the 1,366 wells surveyed, over 40% required deepening or major repairs; many others required minor repairs or maintenance. As a result the appraisal mission increased the number of wells from 360, as originally proposed, to 600 to be rehabilitated and 800 to receive maintenance; of these 60 were to be capped and equipped with hand pumps. The number of new boreholes was increased from 210 to 430 to help Government meet its target of one water point for every 500 people. These increases, which aimed to meet the total demand for well rehabilitation in the area and Government's ambitious target for water points, did not take proper account of the limited implementation capacity of G6nie Rural. G6nie Rural was weak in terms of administration and management, especially to handle such a large program -- 165 wells to rehabilitate each year compared with 68 usually achieved. Similarly, the borehole program of 130 boreholes was to be executed by the "Programme des Eaux Souterraines" (PES) of the Ministry of Mines and Energy which had had limited experience in this type of activity. 2.07 The bottomland development sub-project was discussed in depth during appraisal. The complementary studies on topographic and hydrological aspects that were to be completed by IRAF (Institute for Agronomic and Forestry Research) before the end of February 1977 were delayed and therefore not available before appraisal. However, it was decided to go ahead with the sub-project on the condition that these studies did not bring to light new information to invalidate the concept. As mentioned in the appraisal report, the area was reduced from 500 ha to' 300 ha and a more intensive approach adopted, following the recommendation - 19 - of the Preparation Report Addendum. It was decided to build a barrage and canals to provide water control for cultivation of irrigated rice. 2.08 The only other change was to the stall fattening sub-project which increased from 900 credit recipients at preparation to 1,000 in the Issues Paper and to 2,000 in the final appraisal report. Again questions might have been raised regarding the assessment of FONADER/FSAR's capacity to disburse and administer the increased number of credits. 2.09 From preparation (April 1976) to appraisal (October 1976) the project*investment costs rose by 50%, from CFAF 1,318 million (US$5.6 million) to CFAF 2,043 million (US$8.3million). The cost analysis highlights several reasons. All the components increased in cost except, surprisingly, the small dams sub-project which was reduced by 13% in cost and the unidentified sub-projects component which was reduced by 50%. The 44% increase in the cost of the water supply sub-project brought it up to 60% of total costs. The cost of the support services increased by 50% mainly because of the doubling of the number of vehicles and operating costs and of adding a provision for civil works; the allocation for expatriates changed, but costs remained about the same. Most interestingly the cost of the stall fattening sub-project increased only 27% while the number of credits went from 900 to 2,000 at a time when the cost per credit increased from CFAF 30,000 to CFAF 38,700. A closer examination has revealed that although the preparation report and related documents referred to 900 credits the budget provision -- possibly due to a mistake in cumulative totals -- provided for 1,900 credits, almost the same number as in the appraisal report (2,000). 2.10 The Yellow Cover appraisal report was issued on February 22, 1977, the Loan Committee cleared the project on April 23, 1977. D. Negotiations and Board Presentation 2.11 Negotiations took place from 3 to 6 May 1977. There were no substantial changes in project design vis-A-vis the draft Credit Agreement. However, three modifications were made, one of which focused on a sensitive area of contention between Government and the Bank, an area which was to pose problems later and which was one of the main elements that delayed the approval of FSAR II, that is, the implementation of a cost recovery system. The Cameroonian Delegation readily agreed to the need for recovering costs from beneficiaries, but it had more difficulty in accepting the formulation proposed in the draft Credit Agreement, section 3.15(a) (cost recovery). It was particularly adamant regarding the levying of local taxes which was considered as politically unacceptable. The Delegation stated that the beneficiaries of the village water supply sub-project were among the poorer people of Cameroon and that the same services were being provided free of charge in more affluent parts of the country. After considerable discussion a new formulation was agreed upon, which (a) caused the "Socigtg d'Expansion et de Modernisation de la Riziculture de Yagoua" (SEMRY) to enforce the provisions of the Cabiers des Charges (equivalent to a system of cost recovery) for those benefiting from the investment in bottomland - 20 - development; (b) regarding the dams sub-project, required the beneficiaries to make contributions in kind (or the equivalent in cash) to maintenance and Government to make adequate budget appropriations to the local communities to cover monetary expenditures; and (c) caused Government to continue the program of establishing water points with proceeds from a cost recovery scheme that would cover "at least a portion of investment costs" and required the beneficiaries, through their contributions, to meet the maintenance costs of these facilities. The effect of the modifications was essentially to remove all references to "taxes levies or charges" and to generally soften the wording while maintaining the principle of cost recovery. 2.12 The other two modifications were less important in policy terms, but both were key points in facilitating project implementation. Government requested first that SEMRY be given the responsibility for implementing the bottomland development sub-project instead of Gfnie Rural, and second that the provision for technical assistance and key financial staff be increased to compensate for the weakness of those institutions that would be implementing the planned project activities. Both were agreed upon and incorporated into the Credit Agreement. 2.13 The Grey Cover Appraisal Report was issued on may 26, 1977, Board Presentation was on 30 June 1977, and the project was approved with little discussion and no major issues raised. E. Project Descri;tion and Objectives 2.14 According to the appraisal report, an important objective of the project was to strengthen Cameroonian institutions and train staff involved in rural development so as to create a local capability for physical planning and execution of high-priority small-scale agricultural development projects. The principal beneficiaries of this institution building were to be (a) FSAR (FONADER), (b) Gfnie Rural (the Rural Engineering Service of the Ministry of Agriculture), and (c) the Programme des Eaux Souterraines (Hydrogeological Service of the Ministry of Mines and Energy -.PES). As a first step, these institutions were to be involved in the execution of a series of specific sub-projects financed by the project. The project also provided financing for additional sub-projects, which at the time of appraisal remained to be identified and for project monitoring and evaluation to incorporate and develop the local ideas and capability for effecting rural development. Lastly, the project provided for the testing of cost recovery procedures which were intended to supply information to judge the degree of replicability, from the standpoint of future fiscal burden, of the type of sub-projects financed by FSAR. 2.15 While the FSAR was ultimately intended to operate nationwide, the project concentrated in the first phase on the four Northern Province departments of Diamare, Margui-Wandala, Mayo-Danay and Logone et Chari, covering some 30,000 km2 and a population of about 1.1 mill-on. The project, which provided financing for five small-scale rural development sub-projects, was to be implemented over a four-year period by FONADER with - 21 - FSAR as the main fund-channelling institution, and by several government technical agencies and units responsible for executing the particular investments. The five sub-projects were: % of ProJect Costs At Appraisal Actual (a) improvment of about 300 ha of bottomland near Moulvouday for rice cultivation; 7 4 (b) development of Irrigated market 2 2 gardening (onions and other vege- tables) in a 250-ha area near Maroua by providing loans and technical assis- tance to about 60 vegetable growers; (c) development of commercial stall fattening 5 3 of steers by providing credit facili- ties and technical assistance to about 2,000 farmers in the Mandara mountains area; (d) construction of 10 small dams in the 20 19 Mandara mountains for human and animal water supply; (e) increase and improvment of village water supplies mainly on the Diamare Plain by (i) rehabilitating about 6-90 existing 20 20 open wells, and capping 60 of them; (ii) drilling about 430 boreholes; 14 14 (iii) equipping about 490 water points 7 6 with hand- or pedal-operated pumps. The project also provided funds for unidentified sub-projects; 6 1 and, through FSAR, for supporting services 19 31 including technical assistance, on-the-job training for local personnel, project monitoring, preparation of other sub-projects and assis- tance in preparing a follow-up project. 100 100 - 22 - F. Approach and Rationale 2.16 The Project. The project would "help establish and operate a mechanism for channelling and monitoring the effective use of funds for high-priority small-scale rural development schemes." I/ This statement reflects very succinctly the essence of the project. While it might be unfair to say that the project lost sight of the importance of creating this mechanism, it is clear that the individual sub-projects and their problems became the focus of the project and absorbed most of the time and attention of the FSAR team -- very naturally so since they were being judged according to their performance. Furthermore, the emphasis placed by supervision missions on sub-project performance could not help but reinforce this tendency. Building up institutions and systems for planning, monitoring and evaluation is a much more difficult task, but in the long term it is the only justification for setting up a rural development fund. Developing means for identifying sub-projects, criteria for selecting, and systems for monitoring them should have been the major activity of FSAR, not the concern of the sub-projects. As it was, FSAR was created for implementing the five sub-projects, but it took an incieasing responsibility for day-to-day execution -- primarily because of the weakness of the executing agencies. The approach and rationale for each sub-project are discussed below. 2.17 Bottomland Development. The principle of having a pilot activity that could eventually lead to the exploitation of some 4,000 ha of underutilized bottomlands (yaeres) in the North would appear to have considerable merit. Nevertheless, there is an open question as to whether the specific project approach of using one depression to provide water for an adjoining one could or should be replicated in other yaeres. Unfortunately, by the close of the project there had been no concrete proof to show whether the concept was correct or whether it could be replicated. The dyke constructed with project financing has never held enough water to properly irrigate the adjoining area, and still does not. Although there is no clear understanding of why this has happened, it appears that a number of factors have come into play, two of which would put into question the basic viability of developing the Moulvoudaye yaere as a pilot area. First, doubt has been expressed regarding the accuracy of rainfall statistics for the depression. If the rainfall proves less than predicted the dyke might never hold enough water to fully irrigate the 275 ha proposed. Second, although the situation has not been fully investigated, it appears that water collected behind the dyke during the rainy season has been escaping either by infiltration or by reverse flow. Both of these questions will remain open until the current drought, which has afflicted the area sinci the completion of the dyke in 1982, is over and the validity of the ccncept can be measured. 1/ Covering note for project submission to the Loan Committee, April 21, 1977, para 2. - 23 - 2.18 Market Gardening. The sub-project as presented in the appraisal report is attractive -- the adoption of an intermediate technology approach aimed at the poorer farmers, which would increase gradually the production of vegetables, especially onions, in order to give the market time to absorb the increased production and not to produce more than it can absorb. This concept derived from the experience gained in a UNDP-financed project in Chad and focused on hand or pedal-operated pumps instead of motor pumps. Furthermore, it was decided that boreholes rather than shallow wells were the most effective way of tapping the water table. The appraisal report provided for 40 boreholes and 20 shallow wells. 2.19 The approach changed from the beginning of project implementation. The FSAR team first requested that the hand pumps be substituted by motor pumps because these were fouAd to be marginally cheaper than the manual pumps and they could provide enough water to irrigate a much larger area (between 0.5 and 1.0 ha versus 0.2 and 0.5 ha). Second, boreholes were found to be too expensive and unnecessary -- the shallow wells provided sufficient water. Although the revised package was more attractive to farmers and readily adopted, the fear expressed in the appraisal report of possible overproduction in relation to market demand and capacity, did materialize (para 5.13). 2.20 It might still be questioned whether the modified concept, in spite of its popularity, was sound. The fact that large-scale onion production is a high risk/high profit venture only suited to entrepreneurial farmers leads one to ask (a) whether they needed the credit at all and (b) whether the provision of an organized credit program to encourage onion production does not adversely affect a straight commercial judgement. The effect of this approach is that in a short time many additional producers are brought into the system before it is possible to assess if the market can absorb the additional production. 2.21 Stall Fattening. The basic concept of the sub-project was well founded. By expanding and modifying an existing practice, stall fattening of cattle for traditional slaughter, it has been possible to increase the disposable incomes of families in the Mandara Mountains -- a densely populated area with extremely scarce arable land and limited agricultural potential. However, during implementation some flaws appeared in the detailed design: the uptake of the credit was slower than anticipated primarily due to the few extension staff with limited mobility who were unable to adequately cover such a large and difficult region; and the animal weight gain was less than projected (230 gm/day versus 300 gm/day) due primarily to the inability of the farmers to buy an animal of sufficient high quality (the credit provision was insufficient) and to the difficulty in getting the proposed supplementary feeds (cotton seed cake and rice flour). While it might be said that these factors somewhat invalidated the concept, as a pilot activity it should be seen as a success. Not only did the credit recipients increase their incomes, they modified their traditional perception of keeping animals (for customary slaughter), which resulted in a better utilization of limited resources. - 24 - 2.22 Small Dams. Because of the high cost of the dams in relation to the original estimate (2.7 times more on average per dam) there has been considerable debate regarding the validity of using dams as a means of providing drinking water to the farmers and their cattle. The main argument against the approach is that water from a reservoir is much more costly per ms consumed than water provided by boreholes or wells. The main arguments in favor of the dams have been that (a) there is no reason to sink boreholes or wells as there is no reliable underground water in the area and (b) there are supplementary benefits attributable to dams, which are not normally considered in a straight cost analysis of such an option. 2.23 It is important to look into these arguments in more depth. Even taking into account the supplementary benefits, it is clear that boreholes or wells costing.approximately CFAF 1.8 million to serve 350 people are less expensive than dams costing CFAP 155 million and serving 2,000 to 3,000 people and their animals. This assumption is based on a high rate of successful drillings (or diggings). However,*at a drilling rate of say 10 boreholes per productive well, then the "construction" costs per person/animal of the two options become similar. The overriding factor is the argument that, boreholes and wells are not physically feasible in many parts of the Mandara Mountains, 2/ especially at higher altitudes where the four FSAR I dams have been built and where most of the others are proposed. Following the logic that boreholes and wells are not feasible and dams are too expensive, a study was carried out to investigate the feasibility - both physical and economic -- of other means of providing water to the people and livestock living in these mountainous areas. The study 3/ concluded that the tapping of underground water was not feasible, but it presented the following arguments for and against barrageL versus boreholes. - Ecology - by establishing a body of water (with a barrage), a micro-climate can be created that would increase rainfall in the area. - Population and Agriculture - boreholes provide a source of water close to the existing concentrations of populations whereas reservoirs tend to attract population to them. Barrages provide readily accessible water for livestock and can also permit irrigated agriculture even if it is only minor vegetable growing. 2/ This was the conclusion of the Italian engineering firm La Recci after a study of the underground resources of the region in 1969; it was confirmed by SOGREAH in 1976. 3/ Construction des barrages, Projet de ressources hydrologigues des Monts Mandara, FSAR, la section Barrages, June 1983. - 25 - - Acquaculture - barrages can be easily stocked with fish. - Health and Sanitation - boreholes will almost always give cleaner water while barrages can be polluted by animals and can pose risks of schistosomiasis, malaria and gastrointestinal diseases. In both cases, education of the population in proper use of water is important; in the case of the dams, it is essential. - Maintenance - in contrast to dams which are almost maintenance free, boreholes require continuous maintenance which in turn requires training and payment of technicians to ensure backup service and availability of spare parts. - Investment - the study does not draw a conclusion on investment, except to say that the cost per m3 of water from the reservoirs could vary between CFAF 137 and CFAF 9 depending on whether one writes off the investment over 20 years or 300 years. Furthermore, it states that investment in a borehole (no figures quoted) can give spectacular results for the first five years, but what happens thereafter is another story. 2.24 The above observations appear well founded but it should be stressed, regarding the maintenance of boreholes, that it is not only the training and payment of technicians and availability of spare parts that constrain borehole development, but also and more importantly it is the organization required to ensure regular maintenance and the timely provision of spare parts.. Even if the communities were trained to carry out most of the maintenance themselves, certain repairs would still require the services of a specialized technician. 2.25 Following this evaluation, the study analyses six options for providing surface water, either on an individual or on a collective basis. These options and their viability are summarized below. (a) Family catchment in masonry - same basic approach as the dams but on a small scale. The constraints are the requirement of a skilled mason, disease problems and the high cost estimated at CFAF 240,000 per person served. (b) Family water tanks - prefabricated containers in plastic or metal (10m each). The constraints are the difficulty in acquisition and transport and the high cost at CFAF 200,000 per person served. (c) Family reservoir/roof catchment - as the style of housing in the area is almost exclusively a rondal with a thatched roof, to have a successful roof catchment system a new house would need a flat tin roof. The estimated cost is CFAF 300,000 per person served. (d) Collective dug-out catchment - in the few areas where one finds sufficient depth of earth (normally intensively planted with - 26 - millet and other food crops), bulldozers could scrape out a reservoir of,say 25,000 ma that could serve 1,200 persons and their livestock. The constraints are few suitable sites, difficulty in ensuring impermeability, and health risks. When no lining is required (clay base), this solution is the cheapest at CFAF 43,000 per person served; when a plastic lining is needed the cost rises to CFAF 94,000. (e) Dam and reservoir - either in grouted rubble masonry (as constructed by the sub-project) or in earth reinforced with cement anchoring. For both, the constraints are the high level of organization required to execute the works, problems of health and the risk that the selected site will not hold enough water to guarantee supply until the end of the dry season. The costs per person served are relatively low, especially compared to the individual solutions, at CFAF 64,000 for the former and CFAF 79,000 for the latter. (f) Reinforced concrete dam and reservoir - this is a lightweight and precise structure requiring limited but high quality materials such as high grade cement which is often not available in Cameroon. The major constraint is the high level of precision required; however, this option is quoted as being the least expensive one (along with (d) - unlined) at CFAF 43,000 per person served. 2.26 While the analysis was not exhaustive, the conclusions drawn are convincing and point clearly toward the system being implemented. The only conceptional question remaining is whether a high cost investment should be sunk into such mountainous areas -- the alternative being to encourage the people to move where they are easier and cheaper to serve. There is no obvious answer to this question but since it is primarily a social and political issue and not an economic one, it should be the responsibility of Government to determine the appropriate policy. 2.27 Village Wells, Boreholes and Pumps. The basic concept of this sub-project -- to deepen or rehabilitate existing wells, to sink new boreholes and to equip them and some of the wells with hand or pedal pumps proved sound. Its performance is discussed in Section V. III. IMPLEMENTATION A. Effectiveness and Start-up 3.01 The project started up relatively soon after Board approval due mainly to Government and FONADER taking an active part in meeting the prerequisites. FSAR was officially created on August 12, 1977, the bank account was opened shortly thereafter with the required deposit of funds and the Chief Operations Officer was appointed, all prior to.the first supervision mission in October 1977. The final conditions for - 27 - effectiveness were met by the beginning of 1978 and the project was officially declared effective on January 30, 1978. B. Progress and Revisions 3.02 The implementation of the project as a whole followed relatively closely the concept described in the appraisal report; there were no major changes or revisions in any of the main sub-projects or components. However, while the project itself started on schedule, it took some time for it to become truly operational and to get the individual sub-projects on track. No project funds were disbursed during the first 18 months. 3.03 The project, originally scheduled to be implemented over four years, was extended to the end of 1983, and thus became a six-year project. The need for this extension was due mainly to the late takeoff of some sub-projects and the slow progress of others. However, it was more important to reach a sufficiently advanced level of implementation and to make as much progress toward the projected goals as possible than to end the projeci at four years when much of the experience would have still been forthcoming. Furthermore, as the project was intended not only to execute the technical aspects of the five sub-projects, but also and more Importantly to develop a new institutional and management structure, it was essential that this structure be given time to evolve and mature. In retrospect, one might question whether the original four-year implementation period was not too short to build up a new institutional capability, particularly if it were expected to assume an expanded role by the end of the project. The building up of management capability derived from the implemention of the five sub-projects and the lessons learned from them regarding technical, logistical and financial management. Progress of each sub-project is discussed below. 3.04 Bottomland Development. Execution of the major works for this sub-project did not commence until 1982, four years after the projected starting date and one year after the expected completion date of the project. The cause for the delay can be attributed to three main factors. - The difficulty in finding the technical solution to what was initially thought to be a simple water control scheme. The underestimation, both by the appraisal team and FSAR management, is reflected in the original plan which expected that the works would be studied and designed in six months, using the resources of one expatriate engineer and the existing staff of Ggnie Rural in Maroua "as required". This responsibility was switched to SEMRY during negotiations but no provisions had been made for SEMRY to handle this additional work. Questions of partial or total water control, of how to manage a heavy inflow of water over the short rainy season combined with a gradual and controlled provision of irrigation water in the dry season, and of how to ensure that the flat poorly drained depression would hold the necessary amount of water, were all difficult to answer. - 28 - - The second factor, which is linked to the first, is the insufficient-detailing of the organizational responsibilities and staff requirements for carrying out the design studies and executing the works. This is reflected in the several attempts to find the correct technical solution and later in the apparent reticence of SEMRY to commit itself to its execution. The fact that SEMRY had previously (under SEMRY I and at the time under SEMRY II) carried out much larger projects in the same area might have led the implementation team into underestimating the difficulty of implementing what appeared to be, by comparison, a relatively small and simple project. Furthermore, the fact that the resources of SEMRY were to be heavily committed to the execution of SENRY II was not sufficiently taken into account. - Finally, management of the sub-project lacked a strong guiding hand. No one took full responsibility for designing and executing the sub-project. The expatriate engineer (6 months provision) was not hired and the responsibility fell on the expatriate project director. As he had many other responsibilities, he was not able to devote sufficient time to this sub-project. 3.05 These factors resulted in the following sequence of events: (a) the first study prepared by the project manager in 1979 with the topographic data provided by PES was rejected as being too sophisticated, too costly and too risky because of drought and, as such, did not follow the spirit of the sub-project as presented in the appraisal report; (b) subsequently, in 1980, a consulting firm, Bfture, was asked to carry out a study but never completed the work; and (c) finally, SEDA was requested to prepare the study and final design, which were completed in April 1981. 3.06 SEMRY started executing the sub-project in April 1983 and completed it in June 1983. The farmers eager to cultivate land under the scheme contributed to the construction of the bunds. About 250 ha of bottomlands were distributed to 500 farmers. 3.07 Market Gardening. This sub-project started as envisaged early in 1978. However, from the beginning the project team requested the World Bank to provide motor pumps instead of the hand and pedal pumps provided for in the appraisal report (para 5.08). In July 1979, the Bank authorized the purchase of motor pumps for the first 20 credits. After reviewing the satisfactory progress of credits for the first 20 motor pumps, the Bank authorized in May 1980 the purchase of motor pumps for the remaining 40 credits, in substitution for pedal or hand pumps, on the condition that they first serve the less advanced farmers in the area instead of the well established ones, as had been the case with the first 20 credits. On one hand, one could comment that, at appraisal, a more detailed and thorough investigation of the technical solution best adapted to the specific circumstances in the area would have facilitated the use of a better adapted line of credit and would have helped to cr'anize the appropriate support.services; on the other hand, as these mic. ,-proje' ts were pilot in - 29 - nature, some experimentation in project implementation can be justified as the experience gained could help to execute future lines of credit. 3.08 Stall Fattening. The approach proposed for this line of credit was properly followed. Only minor adjustments were made. As the price of calves for fattening was about CFAF 37,000 in 1978 (at the start of the project) i.e., 25% higher than estimated at appraisal, the amount of each credit was raised from CFAF 38,700 to CFAF 45,000, and again in 1982, to CFAF 52,000. The other adjustment during implementation was the stopping of all credit disbursement in 1981/82 due to poor repayment rates (para 5.21). ,The credits were reinstated in 1982/83 after the recovery rate ad improved. Of the 2,000 credits foreseen, 1,330 were distributed, but over a 6-year period instead of the original 4-year period. 3.09 One addition to the sub-proj-ct activities, not considered at appraisal, was the associated research program. As little information existed on the best feed rations and balanced diet, on the needs for maintaining good animal health and on the means to achieve the best economic return, FONADER signed in May 1981, in behalf of FSAR, a protocol with the "Centre National Zootechnique et Vttrinaire" (CNZV) to carry out research on the above issues. In line with the agreement, CNZV purchased 18 animals and studied over a 2-year period how to achieve optimum animal performance under conditions similar to those of thd villages. The results are not yet available. 3.10 Small Dams. Although the sub-project officially started within a month of project effectiveness with the arrival 1f the expatriate engineer, work for the first construction season -- November 1978 to June 1979 - was hampered by the late arrival of equipment and by the inadequacy of the preparation studies and surveys. It was the latter that, over the life of the project, caused the major delays. The precise location of the dams, the type and depth of foundations, the topographic characteristics and the provision for access roads to the dam sites required detailed study and interpretation in order to produce comprehensive work plans and budgets. General data were provided in the SOGREAR study -- which served as the basic document for the sub-project - but as this document was mainly an inventory of 151 dam sites of which 57 were judged acceptable, considerable supplementary information was required to facilitate the implementation of the specific dams proposed under the project. 3.11 The main changes during the implementation of the sub-project were the method and rate of work. It was decidel in the first year that it was impracticable to bring materials in by head load and consequently roads/tracks were constructed to reach the dam sites by truck. The amount of work required to prepare the site and build the dam was considerably underestimated; this is somewhat understandable since, first, little comprehensive experience existed on small dams in such terrain and, second, at appraisal the engineer had only the SOGREAR study on which to base his estimates. More detailed studies of the 10 proposed sites prior to appraisal would have made a more realistic evaluation possible; on the other hand, they would have further raised the costs of providing the - 30 - planned low cost small dams. If these design studies had been properly carried out in the first year of the project, as proposed in the appraisal report but for which no proper cost provision was made available, the potential problems and increased magnitude of the works would have been flagged early. Nevertheless, the original proposal to construct 3 dams a year (4 in PY4) in the 6-month dry season in difficult terrain was an extremely optimistic assumption, particularly for a pilot activity and given the limited resources provided -- two construction teams, each consisting of 47 laborers and a foreman. Over the extended life of the project (6 years) only four of the proposed ten dams were completed. 3.12 Village Wells, Boreholes and Pumps. The implementation of this sub-project followed closely the approach proposed in the appraisal report and no major revisions were made; howevet, there were considerable delays in execution. Since the sub-project was implemented as three relatively independent components, the progress will be discussed separately under each heading. 3.13 The wells component, divided into reconstruction (generally deepening) and maintenance (often work on the superstructure), had implementation rates of 95% for reconstruction and 57% for maintenance by the end of PY4; they rose to 150% (904 actual wells versus 576 at appraisal) and 72% (568 actual versus 790 appraisal) respectively by the end of PY6 (December 1983). In addition, 60 wells were capped as foreseen at appraisal and 45 new wells, not originally included in the sub-project, were dug. Table 1 in Annex I of the SEDA completion study provides the annual progress in each category. The only major modification was to increase the number of wells to be maintained from 104 as proposed in the appraisal report to 790, based on the reevaluation of the maintenance capacity of G6nie Rural at the beginning of the project, which proved overly optimistic. 3.14 The implementation of boreholes fell considerably short of the rate achieved for wells. Over the four-year period (1977/78 to 1980/81) only 137 successful boreholes were drilled compared with the appraisal estimate of 430. Their number increased to only 232 by the end of the project. The major reason for this poor performance was the unavailability of drilling rigs. Three drilling rigs were forecast in the appraisal report. The "Programme des Eaux Souterraines" (PES) owned two rigs, one suitable for crystalline rock areas (Stenwick 66D), the other for sedimentary areas (Failing 250); the third rig (Percussion) was to be purchased by the project. The Stenwick rig was to continue to be used by PES for exploratory drilling, the other two were to be for project work. The first problem was that the Failing 250 was 6 years old (at the start of the project), in poor condition and operated for 3 months only during the first two years of the project. The second problem was the long delay (over two years) in purchasing the new drilling rig. This was caused, among other things, by a difference of opinion between the Bank and Government over which rig to buy. Third, even though requested and agreed in mid 1979, PES did not make the Stenwick rig temporarily available to the project until the 1980/81 dry season to compensate for the Failing. Thus, - 31 - the borehole component did not become fully operational until PY4. Only the new rig (Ingersoll Rand) operated for the full season in PY3. 3.15 The root of the problem can be traced back to the evaluation of the PES Failing 250 drilling rig. If at appraisal it had been possible to properly assess the remaining working life of the rig, it would have been either written off or possibly provision would have been made for a complete overhaul. A proper assessment might also have prevented the difficulties in selecting a new drilling rig. It was required that this new rig be able to operate in both crystalline and sedimentary areas (to compensate for the Failing) and it is this requirement which, to a large extent, caused the confusion and delays. Furthermore, to achieve the targets set in the appraisal report, both machines would have had to operate at a high level of efficiency (about 3 boreholes per week) throughout the dry season. 3.16 The puMEs component experienced the worst delays. By mid-project (:979) only 3% of the appraisal total (490) had !een installed; by the end of PY4 this had only increased to 18%. The last two years saw a rapid increase in pumps installed and by the end of 1983 the percentage had increased to 64%. The main delays were due to the slow progress in sinking boreholes and capping wells and to the late delivery of pumps. 3.17. Unidentified Sub-projects. The US$500,000 equivalent line of credit has only been used to fund preparatory activities for the follow-up project. The principal reason was the non-availability of qualified staff and the lack of a mechanism to identify, prepare and appraise small-scale rural development interventions. FSAR staff designated to perform these functions, were totally occupied with managing the project. In retrospect, it was probably unwise to assume that these time-consuming tasks could have been performed along with day-to-day management by FSAR senior staff, and provision should have been saade accordingly at appraisal -- possibly by setting up a separate unit within FSAR. C. Procurement 3.18 The August 4, 1977 Credit Agreement stipulates ICB for all contracts worth more than US$100,000 equivalert. They were to include civil works related to the dam construction, FSAR office buildings and bottomland construction as well as some equipment. The total package was estimated at about US$3.6 million equivalent. As it happened, none of the civil work could be done by contractors, as the office building of FSAR was rented and the bottomland and small dam construction work was too specialized to interest any international or local construction firm. Consequently the only ICB was for vehicles (including the drilling rig), equipment and pumps, worth approximately US$1.9 million equivalent. 3.19 Procurement for these items was not without problems. While no major problems in the awar,4 of contracts for rolling stock, the selection process for pumps and the , illing rig presented some difficulties. As previous experience with hand pumps in Cameroon was limited, the project - 32 - procured 4 different types of pumps (Vergnet 54, Biau 117, Abi 55 and Robbins 11). At the end of the project, the project pump section-chief project still maintained that no single make can be said to be ideally suited for the Northern Province conditions. The selection of a drilling rig for boreholes was long and protracted. Instead of emphasizing maintenance, Government applied criteria which clearly favored the most powerful and efficient type of rigs. The selection of the Ingersoll Rand at a cost of nearly US$500,000 proved a costly mistake, precisely because spares had to be ordered overseas since the manufacturer is not represented in Cameroon. 3.20 The recruitment of consultants was also difficult as FONADER lacked experience in selecting such experts; the subsequent replacement of ill-suited personnel and the periodic absence of key personnel ultimately contributed to the slower than anticipated execution of the project. D. Costs 3.21 Actual cost at completion was CFAF 3,905 million, including some investments undertaken to prepare the follow-up project, based oh the average exchange rate of CFAF 304.9/US$1 prevailing during project execution (January 1978 - December 1983). This translates into US$12.8 million as opposed to US$8.5 million or CFAF 2,595 million, at appraisal. i.e., a 51% cost overrun in US$ terms. If, however, a technical completion rate of 70% were applied over the project period, costs more than doubled. Contributing factors, in order of importance, are: underestimation of the costs of dams, borehole drilling and well rehabilitation costs at appraisal; higher than anticipated domestic inflation (14% as opposed to 8% at appraisal); and an extended project implementation period. E. Financing 3.22 While at appraisal the financing was set at 28% and 72% for Government and IDA, respectively, the corresponding percentages at completion in December 1983 were 49% and 51%. Government's contribution to project costs represents local staff salaries and local operating costs. The shift in financing percentages is entirely due to the higher than expected domestic inflation, in particular the 15% estimated annual salary increase for civil servants. Given the substantial appreciation of the U.S. dollar during project implementation, from CFAF 218 per US$1 in February 1978 to CFAF 488 in December 1983, the savings under the IDA Credit (US$7.0 million) were sufficient to cover all expenditures planned at appraisal despite the two year extension of the project implementation period. 3.23 Government's project funding was exemplary, with funds always flowing to FSAR's account with little or no delay; moreover, instead of the budgeted CFAF 640 million at appraisal, Government contributed CFAF 1,900 million to total project costs. Due to FSAR's consistent overbudgeting of its annual expenditures -- the budget consumption rate was on average - 33 - 65%--the project never suffered from cash flow problems even though IDA disbursements often lagged one year (para 3.24). It is worthwhile to mention that although FSAR never really mastered the technique of budgeting and it soon became apparent that there would be delays in implementation, FSAR always complied with the PY costs in the appraisal report. F. Disbursement 3.24 Of the US$7.0 million IDA credit, US$6.6 million was disbursed and US$0.4 million was cancelled. There were 217 disbursements averaging CFAF 9.2 million (US$30,200). As a great share of project expenditures occurred in the North of Cameroon and as communications between Maroua and Yaoundf (FSAR headquarters), are erratic, the preparation by the project of disbursement applications lagged actual expenditures by an average of 6 mon:hs throughout project execution. Contributing to the delay was the external debt office's requirement for painstaking substantiation of each franc spent by the project and claimed for reimbursement from the IDA credit. This problem has not yet been satisfactorily resolved and Bank-financed projects in Cameroon lead to a substantial, additional amount in administrative costs. G. Reporting 3.25 Up to the end of 1980, the project produced only annual reports with statements of expenditures, although the legal documents requested that FSAR submit half-yearly reports about physical progress and budget consumption. It was only.in FY 80/81 that FSAR began to regularly produce the necessary reports and submitted them to the Bank. Budgets, on the other hand, were regularly prepared and submitted for approval to both Government and the Bank. 3.26 Section 4.01 (b) of the Credit Agreement stipulates that FSAR accounts be audited annually and the report, prepared by an independent auditor acceptable to the Bank, be submitted within four months of the end of each FY. While the accounts were closed to meet this deadline, the actual audits never took place earlier than one year after each FY. Since FSAR is not a legal independent p&oject executing agency and exists only on paper as a Department of FONADER, its accounts are audited jointly with those of FONADER. However, given FONADER's weak accounting department and rather complex banking and credit operations, its accounts were, and still are, not ready until at least one year after each FY. Therefore Section 4.01 (b) of the Credit Agreement was never complied with. It is noteworthy to mention that all Credit and Loan Agreements for agricultural projects in Cameroon include a legal covenant providing for FONADER audit reports within four to six months of completion of its FY. This covenant has not been complied with since the Bank Group's started financing operations through ?ONADER. - 34 - H. Project Closing 3.27 The project closed on December 31, 1983. The FSAR and its operational division (DOFSAR) in Maroua continued to operate the sub-projects previously supported by the project. New stall fattening credits are being disbursed and payments are still made by the former credit recipients; work has started on a fifth dam with financing being provided directly from Government budgets; boreholes continue to be sunk and pumps installed. IV. INSTITUTIONAL PERFORMANCE A. Institutional Design 4.01 Institutionally, the project focused on FONADER. This was a logical choice as FONADER's mandate broadly, to fund and supervise rural development projects and agricultural credit operations--includes specifically the following responsibilities: (a) to provide financial assistance to farmers, individually or in groups; (b) to carry out or help to carry out rural development projects; and (c) to guarantee loans made by other credit institutions to individuals or groups in agriculture and livestock. FONADER falls under the supervision of MINAGRI but is legally a separate entity with administrative autonomy, FONADER's broad ability to intervene in the rural sector, its operating autonomy and, most importantly, the operating autcnomy given to FSAR within FONADER, were key aspects in the project design. The above arrangement, after some initial difficulties, worked out satisfactorily and FONADER provided the necessary financial and administrative backup required to support the project. The success in following this approach reinforced the idea to utilize and reinforce existing institutions to the maximum. 4.02 The responsibilities of FSAR are clearly defined in the appraisal report: "FSAR would be responsible for overall supervision and monitoring, including the establishing of annual programs, budgets, and progress reports, and for procurement. In addition, it would identify and prepare as-yet-unidentified sub-projects." The report does not mention execution or implementation. Although FSAR had to follow its terms of reference, it spent too much time in the day-to-day management of the sub-projects and not enough in planning, coordinating and monitoring. 4.03 The activities of FSAR were to be split between a head office in Yaounde (DFSAR) for overall management and accounting and a regional office in Maroua (DOFSAR) for administration of the sub-projects. Because of poor communication - no telephone and inadequate telex linkage -- this arrangement that on paper appeared logical became in practice difficult, with the passing of information and the decision making process being cumbersome and time consuming. - 35 - B. Management and Organizational Effectiveness 4.04 FSAR went through a considerable learning process. Poor management control was reported early but little systematic monitoring of the sub-projects was established. Furthermore, there was lack of budgeting or financial control at DOFSAR, which resulted in low morale. Much of the blame was placed on the head of the DOFSAR office, an expatriate, who was ill-qualified for the job and who at the Bank's recommendation was replaced by a Cameroonian in early 1980. Thereafter, management and morale improved noticeably and the performance of the project picked up. The distance between'the DFSAR and DOFSAR offices, however, continued to be a major constraint to management effectiveness, particularly in the processing of "bons de commande," procurement of materials and equipment, and in signature delegation at DOFSAR. One of the recommendations of the SEDA completion study was to bring together all the services of FSAR in Maroua. This would have been a mistake as the aim of the project was to create a national capacity to support the implementation of small priority projects in the rural sector. Too much attention was already focused on resolving problems specific to one region and to the five sub-projects in that region. 4.05 The operation and management of the garage and associated support services deserve special mention. It was and still is one of the best run sections of FSAR/Genie Rural and is a good example of the establishment of a proper physical and financial control system to ensure good maintenance and operation of the project's machinery and equipment. 4.06 One of the major problems for the project was its relationship with the agencies respoasible for implementing the sub-projects -- Ggnie Rural, PES and SEMRY. It was found early in the project that Genie Rural and PES had little experience and capacity to implement investment activities of this nature. Neither had an organized system of administration or accounting to keep track of FSAR operations and expenses. Records of wells rehabilitated or maintained were in the beginning almost non-existant and, later, often incomplete. Largely because of the lack of capacity in these two agencies, DOFSAR became more and more involved in the day-to-day administration of these sub-projects, a task which was beyond the capacity on hand and, in a technical sense, outside the scope of most DOFSAR staff. However, the wells, boreholes and pumps sub-project achieved an acceptable rate of implementation, but at the expense of the broader institution building required to create in FSAR a development planning/coordinating capacity. SEMRY had a different problem. As a well administered and efficient organization with established cost control procedures, it was well suited to implement the bottomland development sub-project. However, because of its commitment to execute SEMRY II, it paid little attention to the FSAR project until PY4. Two facts emerged from these experiences: first, the rate of progress in implementing a sub-project is essentially in the hands of the executing agency, not FSAR; and second, the only way for FSAR to operate within its mandate (i.e., no direct participation in execution) is for it to select executing agencies already in place and, with minor reinforcement, can bring up to the - 36 - required standard administrative, financial and technical capabilities to Implement and monitor a project. 4.07 For FSAR to have met the objectives in para 4.02, it would have been necessary to create a system for (a) identifying new projects; (b) evaluating their technical, economic and social validity and importance; (c) assessing the capacity of the proposed executing agencies; (d) determining how to best set up the interface between FSAR and these agencies, including financial and physical budgeting, monitoring during execution, and impact evaluation after execution; and (e) establishing working'protocols with these agencies. This cycle would have obliged FSAR to establish close relations and regular working sessions with people and bodies at the national, provincial and even district levels if the rural development fund were to fully realize its mandate. The neglect of this aspect was perhaps the major shortfall in the first phase of FSAR. 4.08 FSAR operations were divided between two offices; DFSAR (Direction du FSAR) at Yaounde and DOFSAR at Maroua. The staffing at the end of 1983 was as follows: DFSAR (Yaoundf) - Director Total: 17 - Advisor - Administrative Officer - Accounting Service - 8 persons headed by an accountant - Secretaries (3),' administrative assistant, drivers (2) and office boy. DOFSAR (Maroua) - Field Manager Total: 16 - Assistant Manager - Accounting Service (3) - Administrative personnel (6) - Credit Unit Chief - Extension Agents (4) The appraisal report did not provide for the advisor (expatriate) or the Administrative Officer, the accounting service of DFSAR expanded from one to eight, and the service staff increased by 50%. Most staff were seconded from FONADER and are still employed under FONADER terms and conditions. The increase in accounting staff may reflect an overemphasis on this aspect, but it is at the same time an indication of an underestimation at appraisal. 4.09 To staff FSAR, FONADER drew heavily on its existing staff as well as existing staff within its parent Ministry, MINAGRI. Of the total permanent personnel .of 33 (17 in DFSAR and 16 in DOFSAR), only 2 outsiders were recruited; the DOFSAR accountant and the Head of the Credit unit. During the 6-year project implementation period there were only 6 changes in personnel, i.e., staff continuity was good. The project developed several good local staff, the most notable being the Chief of Operations - 37 - DOFSAR, the Head of the Credit Unit and the two accountants at Marous and Yaound6. All benefited from project-financed training courses abroad (not foreseen at appraisal but approved during supervision). 4.10 One continuous cause of friction, which could have been avoided with better organizational planning, was the discrepancy in salaries and terms of employment between technicians working together on the same sub-project. The FONADER staff were well paid and benefited from good social insurance and generous living allowances. Comparable staff of Genie Rural, PES and other regular government technical staff often received salaries lower than those of FONADER drivers, and their social security benefits and housing allowances were considerably less attractive. This disparity in remuneration not only created a poor working environment but also made it difficult to attract good technical staff within the regular government services. 4.11 The man-month budget for technical assistance increased by only 10% over apuraisal estimates even though the project was extended by two years. The distribution changed as shown below. Appraisal Actual (man months) Field Manager 48 30 Credit Specialist 24 47 Maintenance Engineer 48 61 Dam Const. Engineer 48 59 Borehole Unit Manager 48 48 Consultant Engineer 6 - 222 245 To a large extent the changes were a result of the competence and suitability of the individual experts in their positions. The inability of the Field Manager to execute his job was noted in 1978 and in June 1979 a supervision mission recommended to FONADER that a replacement be found. Even though the Directeur General of FONADER acknowledged the situation and stated he would take up the matter with SATEC, the consulting firm that provided the technical assistance team, it was one year before the expert left and was replaced by a competent Cameroonian manager. The other experts performed well, particularly the maintenance engineer who, in addition to being responsible for the garage, also coordinated the installation of the hand and pedal pumps. The three dam construction engineers who kept their positions over the six years of the project tended to be practical and experienced field engineers and were reported to have done a good job under difficult conditions without any comprehensive site plans or final engineering designs to work with. 4.12 Although some funds for training were approved later on (para 4.09), the project definitely neglected training; the appraisal report did not mention training funds. The example of the Head of the Credit Unit, a - 38 - former departmental delegate of MINAGRI, who was recruited by the project and sent for a 4-month training course abroad and, as a result, performs his duties as well as any technical assistant did before him, demonstrates the potential impact of well-focused training. C. Covenants and Sectoral Developments 4.13 FSAR has been reasonably efficient in complying with legal covenants under its own control, but two other major covenants outside its control have never been met. Both covenants are financial, one dealing with the annual audit of FSAR's accounts and the other with Government's rural water supply sector policy. 4.14 Although FSAR is a self-contained department within FONADER and its staff are solely working for FSAR, in the legal sense it is not independent and, therefore, its assets belong to FONADER. Thus FSAR's annual accounts are consolidated with those of FONADER and subsequently audited together from Balance Sheet and Profit/Loss Statement. Due to considerable delays in the closing of FONADER's annual accounts, Mainly because of permanent enquiries about its financial operations by state audit commissions, FSAR's accounts automatically suffer from the same delays. Thus, the deadline of four months after the close of each fiscal year for submission of audited accounts was never met. This constraint has been taken into consideration when preparing the second phase of the project and provisions have been made for a separate audit of FSAR accounts. To that extent, FONADER's status has to be amended to provide for a separate audit and subsequent consolidation. 4.15 Official cost recovery policy, or rather the lack thereof, has been a growing concern to Bank staff ever since project identification. It is estimated that at any given time about 50% of all water points in the two northernmost provinces of Cameroon are inoperative for lack of maintenance. Therefore, the immediate question to be addressed is the sensitization of beneficiaries. As long as the village people do not consider themselves owners of the installations, they will never maintain them. The FSAR project addressed this issue by asking Government, within a given time-frame, to submit to IDA for discussion a cost recovery policy involving the participation of the population. 4.16 However, given the political implication of implementing such a policy in one region only, Government evaded the issue by maintaining that it could not address a macro-policy issue within a regionally confined project, thus missing the deadline for submission of a proposal. Just how thorny a problem this really is, is best demonstrated by the fact that negotiations for the second phase project, which foresees the construction of a large number of additional water supply installations, lasted from July 1984 to April 1985 to define a beneficiaries' participation scheme. On the other hand, FSAR frankly admits that its major problem is to keep up with the maintenance of the pumps installed under the first project. Indeed, given the wide dispersion of the installations of over some 80,000 km2 it takes a full time repair team (a mechanic, an hydraulic engineer and - 39 - a driver) to ensure simple maintenance of some 300 pumps installed under the first project. The estimated cost of this service is approximately US$300,000/year (including technical assistance). V. OPERATIONAL PERFORMANCE AND PROJECT IMPACT 5.01 The operational performance of FSAR is most easily seen through the sub-projects, the performance and impact of which are given below. A. Bottomland Development 5.02 Physical Performance and Production. After numerous attempts at an &cceptable design to serve as the basis for implementation of the project (para 3.05), the dyke was constructed, the canals were dug and 230 ha of the originally proposed 275 ha were prepared for cultivation. This work started by SEMRY in 1982, was completed in time for the 1983 rainy season and the 230 ha were sown with rice. 5.03 Prior to the completion of the dyke and canals, 230 ha were redistributed in plots averaging about 0.5 ha to approximately 500 farmers, according to traditional land rights. 5.04 The first crop production was dismal and does not provide useful information on the performance of the scheme. Rainfall was very low (estimates vary from 300 am to 500 mm) and the level of water behind the dyke only reached 0.68 m against the projected retention of 1.20 m. Taking into account evaporation, infiltration and other losses, 163 ha received insufficient water to be productive the remaining 67 ha were reported by SEMRY to have produced 104 mt (1.6 mt/ha), or 0.45 mt/ha if one takes the full 230 ha into account, which is probably more reasonable since it is difficult to draw the line between those areas that received enough water and those that did not. This compares to an average yield from SEMRY schemes of over 4 mt/ha. The repercussion of this bad harvest is greater than is !'u_ediately apparent. In years with poor rain, the farmers do not attempt to grow rice but grow their traditional staple food crops (millet and sorghum) which require less water and are more resistant to drought. The rainfall did not improve in 1984 and the dyke did not retain enough water to provide irrigation during the dry season. 5.05 A secondary activity of the sub-project, research, extension and seed multiplication), which was to be carried by the Institute for Agronomic Research (IRA) over a period of three years, was terminated in 1980/81 after less than two years. Crop variety trials were initiated in 1980/81 on 10 ha of the undeveloped yaere but an unexpected flood destroyed the seed beds. Since the extension activities (primarily to assist in seed distribution and encourage fertilizer usage) and the seed multiplication activities depended on the successful implementation of the water catchment and distribution system, neither could be implemented during the project period. - 40 - 5.06 Impact. The project justification stated in the appraisal report was to raise faru family cash incomes from CFAF 17,000 (without project) to CFAF 61,000 (with project), based solely on rice. As food crops are traditionally cultivated in the yaeres, a scheme based solely on rice is somewhat unrealistic. However, as the delay in implementation and the drought negated most of the sub-project, no information exists on which to base a new scheme. One can only say that, the impact of the sub-project has been negative so far in terms of disruption of the population and labor loss: only a year of "normal" rains will tell if the forecasted benefits can be achieved. B. Market Gardening 5.07 Physical Performance. The 60 credits proposed were distributed as planned in two tranches the first 20 in 1978/1979 and the second 40 in 1979/80. In each case a motor pump (3.5 HP) and accessories were provided to two recipients who took responsibility for the credit. Each dug a new well or rehabilitated an existing well, and the pump was used to serve both farmers fields. As might be expected the dual responsibility caused some problems in ownership and repayment. The table below summarizes credit disbursements and recovery rate for 1978 to 1983. - 41 - CREDIT DISBURSEMENTS AND RECOVERY RATE No. of Actual Amounts Annual Payments Annual Credits Credits Disbursed Amoun" Received Recovery (SAR) Distri- (CFAF Due - Millions Rate Year buted 000) (CFAF (CFAF (%) - -_ - 000) 000) 1978/1979 20 20 8,944 - - - 1979/1980 40 40 16,989 3,130 2,013 64 1980/1981 9,970 2,074 21 1981/1982 16,101 7,845 49 1982/1983 15,812 10,662 67 1983/1984 9,822 1,899 b/ 19 b/ Total 25,933 54,835 24,493 45 a Principal, interest plus accummulated outstanding debts. b/ Only applies to the first half of 1983/84 (to 31 December). Source: Rapport d'Ach4vement du Projet FSAR I, SEDA January 1984. 5.08 Even though the distribution of motor pumps was successful two separate studies 4/ showed that the farmers had trouble with breakdowns primarily due to the unavailability of spare parts. The two studies estimate that 19 and 22 pumps respectively, out of the 60 supplied, were not operational. 4/ Ministry of Agriculture, "Rapport d'Enqu'te Socio-Economique du Projet FSAR I", February 1983. SEDA Completion Study, "Rapport d'Ach&vement du Projet FSAR I", January 1984. - 42 - 5.09 The extension service staffed with only one extension agent had little impact in assisting the farmers to improve their cultivation practices, since his main function was that of a credit agent. The applied research program carried out by IRAF in Maroua has performed somewhat better. The two objectives of the program were to: (i) find improved varieties of onions and test improved methods for their cultivation, and (ii) explore better methods of drying and storing onions. Since the program started two years late, the results of the research were not available until 1984, four years after the distribution of the last motor pumps provided under the line of credit. The research program thus made no positive contribution to the credit recipients at the time, but now it has available for distribution improved varieties and viable means of drying and storing onions have been successfully developed. 5.10 1 Yields and Production - Onions are the main crop cultivated under the sub-project and in the area in general; only small areas are planted with other vegetables. The farm model presented in the appraisal report assumed a "with project" yield 5/ for "onions and other vegetables" of 30 mt/ha; this estimate was over optimistic as the average yields in the area, as recorded in 1983 by the MINAGRI socio-economic survey, were 7.9 mt/ha for non-credit recipients and 13.6 mt/ha for farmers who received FSAR credits. The farmers reported having difficulty getting improved varieties of onions -- the new varieties were not available from the research program as foreseen. They had little or no technical support from the project (inadequate extension) and consequently cultivation practices and irrigation techniques were far from ideal. 5.11 It is difficult to arrive at an accurate figure for the incremental production because of a large discrepancy in the estimate of the average holding size. The MINAGRI socio-economic survey mentions an average size of 0.432 ha; in the SEDA completion study it is 1 ha. There is probably an oversight in the MINAGRI report regarding the number of farmers for each motor pump (two, not one). In this case, the hectarage for two farmers (0.864 ha) approximates the 1 ha used by SEDA. Assuming 1 ha per pump (i.e. per credit) and 13.6 mt/ha, the gross production from the project would be over 800 mt per annum. If one adopts the appraisal report assumption that no onions and other vegetables were cultivated by these farmers prior to the project, this figure would also represent the incremental production. The production marketed was much lower, approximately 480 mt due primarily to storage losses (40%). This 5/ No "without project" yield for onions-and vegetables was provided, as it was assumed that credit recipients would cultivate only millet, sorghum and groundnuts prior to receiving the credit. - 43 - loss is high compared to an estimated 11% for non-project farmers who each cultivate a smaller hectarage (0.2 ha), and achieve a yield averaging 7.9 mt/ha. 5.12 Impact. The impact of the sub-project can be looked at in terms of farm income, income distribution, national supply of, and demand for, onions, and the environment. The farm budget, presented in the Annex, (Table 2) shows that at present levels of production and sales, the average credit recipient has an income about the same as one cultivating onions by the traditional method (handoperated water lever/(chadouf). Howeve-, his disposable income has risen considerably if one assumes, as does the appraisal report, that prior to the project the farmers cultivated only sorghum, millet and groundnuts; in that case the income per family would be CFAF 90,000 versus CFAF 52,000. If the storage and marketing losses are reduced, as should be possible with the new research findings, if a large enough and accessible demand is developed, and if the yields increase, which might be possible with the improved varieties and improved cultivation techniques, the farmers could considerably increase their profits. Furthermore, the return per man-day is much higher with a motor pump than with a chadouf, which requires heavy labor. 5.13 Even with the current relatively low production -- compared to the potential of the area -- the farmers had difficulty selling their crop. Furthermore, the price fluctuations are enormous ranging from CFAF 40,000 at peak production in March to over CFAF 200,000/mt in September. The difficulty in selling the additional production can be attributed primarily to the inexperience of the farmers in marketing such large volumes - i.e., the need of storage, better organized transport, and larger buyers-not to the lack of demand. A study carried out by SCETAGRI for FONADER in 1984 6/ showed the total Cameroon consumption at 27,000 mt, of which 15,700 mt was marketed; the rest represents on-farm consumption and marketing losses. About 6,000 mt were estimated to be imported: 3,000 mt from Holland through Douala and 3,000 mt unofficially from Nigeria to the north of Cameroon. Furthermore, the report estimates that about 3,000 mt of onions are produced for market in the north, a large part of which comes from around Maroua. The 800 mt produced by the project credit recipients is thus a large increment to local Maroua production (27%) - if it is all considered as incremental - but represents only a small portion of the national market (5%). The conclusion is that provided the private sector succeeds in working out market and storage logistics, there is considerable scope for local production of onions. 5.14 Although it was feared that the switch from manual pumps to motor pumps would limit the credit line to the richer farmers in the community, the MINAGRI survey showed that the credit recipients average holding size 6/ SCET-AGRI, "Production et Commercialisation des Cultures Maratchares dans le Nord Cameriun, SCET-AGRI, Mars 1984, p. 115. - 44 - was 0.2 ha compared to 0.3 ha for the non-participating farmers. The credit allowed them to increase their holding to over 0.4 ha (i.e., about 1 ha for the two owners of a pump). 5.15 The main environmental question is the impact of the heavy water tapping on the water table in the area. No comprehensive data exists and no lowering of water levels in the project wells was reported during the recent drought year. As over-irrigation appears to be common practice, the water table should be monitored and better water management encouraged. C. Stall Fattening 5.16 Physical Performance. The table below shows the performance of the line of credit: 1330 loans were distributed out of 2,000 planned, i.e., 67% of the original estimate. However, as the size of the loans increased (para 3.08), the total amount of credit disbursed represented 82% of the original provision. The recovery rate was initially poor: 47% in the first year and 352 in the second year; however after freezing the credit program for one year the rate improved noticeably. By the close of the project (end of 1983), 782 of the outstanding credits had been repaid; this was equivalent to 49% of the total outstanding balance. By the end of 1984, with some final payments still not due, of approximately CFAF 5-10 million, the overall repayment rate had risen from 47% to 78%. This relatively high rate is due in large part to the competence of the Cameroonian Credit Unit Chief and in general to the success of the stall fattening operation. No. of Actual Disburs- Amounts Annual Annual Annual Year Credits Credits ment Disbursed Amounts Due Payments Recovery (SAR) Disbursed Rate (%) (CFAF '000) (CFAF '000) Received Rate (%) (CFAF '000) 1978/79 200 123 62 5,535 - - - 1979/80 500 434 87 19,530 1,353 637 47 1980/81 900 277 31 12,465 9,983 3,503 35 1981/82 - - - - 21,427 20,076 94 1982/83 400 300 75 15,600 10,665 11,400 107 1983/84 - 200 - 10,600 3,300 887 27 Total 2,000 1,334 67 63,530 46,728 36,503 78 a/ Only applies to the first half of 1983/84 (to 31 December). b/ Distribution by year made by FSAR. Source: SEDA Completion Study, Op. cit., and mission revisions (1984). - 45 - 5.17 The weight gain of the credit recipients' animals was somewhat below the appraisal estimate (300 kg); furthermore, it took normally 2.5 years instead of the 2-year period anticipated. This estimate could not, however, be verified as the animals' performance was not regularly monitored and the weight at the time of purchase and at the time of sale was not systematically recorded. The expected shortfall in performance has been attributed to three factors. First, the supplementary feeds envisaged in the appraisal report--cotton seed cake and rice flour--were not readily available. However, in some areas the difficulty in providing the animal with enough water reduced its ability to consume as much feed as desired. Second,*the tendency to purchase a somewhat lower grade animal than originally foreseen might have inhibited the ability of the animal to take on sufficient weight. Third, in general the four extension/credit agents were not sufficient to provide the required technical support to such a dispersed population. In this context, an interest charge of 10% was notably not sufficient to cover the cost of maintaining a greater number of agents. 5.18 Impact. Although the sub-project went a long way toward meeting its objectives, the lack of proper monitoring by FSAR and the inability of the socio-economic study carried out by MINAGRI 7/ to provide hard information on weight gains and changes in farm family incomes means that no complete evaluation of the impact of the sub-project can be made. However, from the limited information available it can be concluded that project performance exceeded the appraisal estimate (para 6.08). D. Small Dams 5.19 Physical Performance. Of the ten dams proposed in the appraisal report four were built by the end of the project in (December 1983). The first one, No. 16 (8.15 m by 81 m; capacity 22,000 is) was completed in May 1980 after two construction seasons. The second dam No. 17 (9.15 m by 54 m; capacity 47,000 as) was also completed in May 1980 but as it lost water when filled additional work had to be done which was only completed in August 1982. The third dam, No. 9 (9.0 m by 49.5 m; capacity 144,000 ms) which started in January 1980 and completed in September 1982 was the largest dam built during the project (2,431 as of masonry); it has double the volume of the others. The fourth and final dam No. 32 (7.0 by 53 m; capacity 89,000 as) started at the end of 1980 and completed in August 1982. To date it has not been possible to completely fill its reservoir. The rainfall is lower at this site which has been exacerbated by the drought over the past two years and it is not fed by a permanent stream. 5'.20 Paragraphs 3.10 and 3.11 discuss the main reasons for the delays and shortfalls in dam construction. Annex 3 proviees details on the scope 7/ MINAGRI."Rapport de 1enquite socio-4conomique du projet FSAR I," Ffvrier 1983. - 46 - of the work completed and shows clearly the underestimation of the work required in terracing (actual quantities were four times the original estimates) and in masonry (60% higher than the original estimates). No detailed figures were available on the number of man-days of labor used to construct the four dams, but the total labor cost is quoted 8/ as CFAF 240 million for the project period, compared with CFAF 88 million (for ten dams) at the time of appraisal. The additional volume of rocks and earth to move (terracing), the increased volume of the dams (masonry), the inexperience of the supervision and labor teams and the necessity to construct 7 km of access roads by hand instead of by bulldozer as proposed in the Appraisal report (no bulldozer was included in the project's investment costs) contributed to this massive increase in labor costs. 5.21 Impact. The dams have had an important impact in the area. Prior to their construction few reliable sources of water existed. The population had to walk long distances on steep terrain to find water. The number of livestock, mostly sheep and goats, that could be maintained by the average family was also limited by the scarcity of water. The four dams were estimated to serve 2,510 saris 9/ or 1,2,550 persons. No comprehensive survey of the number of sards using the water has been made and no regular monitoring is carried out. However, in 1983 the project did make one survey of the number oi visits to the four sites by'people and livestock between March and June. These figures cannot be directly translated into households served, but, the survey indicates that the number of people (and livestock) increases continually from March to June as the dry season progresses to the point where water usage has doubled by June. The number of individual visits to the four reservoirs (i.e., trips to fetch water) increased from 50,000 in March (30 days) to over 100,000 in June. Two related factors have been observed: first, after an initial reluctance by the families in the area to use water from the reservoirs, their number quickly increased each year; and second, there has been an increased use of the reservoirs by nomadic people who temporarily settle around them to provide water for their herds. While this action increases the utilization of the dams, it could also, if not properly controlled result in overgrazing of the surrounding areas, accelerated erosion and even conflicts between different groups. A comprehensive study is needed to look into these factors and to recommend measures for better management of resources with the communities taking the main responsibility themselves. 5.22 One of the main areas of concern in developing the dams was the potential risk of disease. Although there has been considerable discussion 8/ FSAR. Comte-rendu d'activitAs, Exercice 1982-1983," Tableau 5, Page 17. 9/ Group of 5 people with 7 sheep or goats, 0.25 head of cattle who consume 180 1/day. SOGREA Stud, op cit. - 47 - and most people agreed to the health hazard posed by the reservoirs, little was done during the project to either find what the true nature and magnitude of the problems was or to propose workable solutions. The animals drink directly from them and defecate in the water. This leaves little doubt that there are serious health risks for the people using the same water. The solution put forward during appraisal was to install a filtering system below.the dam to provide clean water to people who use it. This system is now almost ready for operation. However, it is not certain that people can familiarize themselves with using filtered water. The cleaning, maintenance replacement of charcoal (filter) and, in general, the management of this filtering system is another question. All this leads to the conclusion that there is a lack of usable duta on health, no system of monitoring and, as yet and most importantly no education and organization of the communities living around the reservoirs on how to use, manage and treat water. 5.23 One positive side benefit of the dams has been the introduction of fish, mostly tilapia and carp, to the reservoirs. This action was initiated by the project and FONADER signed a protocol with the "Ministare de 1'Elevage, des P9ches et des Industries Animales" to stock the four dams in 1981. The fish population is still too young to exploit and fishing has been prohibited on the reservoirs to date. However, the stocking of the dams will have two major benefits (a) supplement local diets and sale for cash, and (b) oxygenation of the water. E. Village Wells, Boreholes and Pumps 5.24 Physical Performance. This sub-section will look briefly at how successfully the works were completed, in quality terms, and the difficulties experienced. The wells component, which deepened and rehabilitated over 900 wells during the project had, according to two surveys 10/ problems with about one third of the wells in terms of either being dry or having insufficient water. These surveys indicate that 37% and 33% respectively of the wells were dry when surveyed. However, 1983 was a drought year and it was thus impossible to know what percentage would be dry in a normal year. The first survey found that the wells had on average the depth stipulated. However, due to the poor and incomplete state of the records it was not possible to make a more complete analysis. This record keeping problem which was continually raised during supervision missions was the result of a number of factors including: a confusion in responsibility between G&nie Rural and DOFSAR, lack of a proper recording system, and little or no training of the site crews in how to fill out the necessary forms. From an engineering view point, the major difficulty was infiltration of sand. This was, and is, a constant problem, and in some 10/ First survey by FSAR and Gnie Rural of 77 reconstructed wells and 50 maintained wells in February 1983. Second survey by "Le Laboratoire Central d'Hydraulique de France" (172 villages) in May/June 1983. - 48 - areas it is reported that the deepened wells have reverted back to their original condition one to two years after their completion. Although different screens and filters were tested, no truly successful solution has yet been found. 5.25 While the boreholes component was less successful in reaching its target -- about 50% of the appraisal estimates -- the rate of successful drillings, i.e. boreholes that became reliable water points was high (over 75%). As was the case with the well, the boreholes in the sedimentary areas had and still have major problems with the fine sand which wears out the pumps and leads to their breakdown. The main solution proposed is to use a double filter and to use in those areas pumps that are less sensitive to sand. 5.26 In the pumps component, four different makes of pumps were installed. Even though this gave the opportunity to test the different pumps under the varying local conditions and depths to determine the most suitable ones (no study summarizing the results was produced), this benefit was outweighed by the difficulty in maintaining adequate stocks of spare parts for all four types of pumps. It is recommended that the number be limited to two or possibly three makes.. This will become increasingly important as more pumps are installed (3,000 to 4,000 pumps planned over the next five years. 5.27 A related and equally, if not more important, problem is maintenance. The appraisal report mentions but not in detail, that a pump maintenance unit would be.established by Genie Rural in Marous and that villagers would be trained in routine pump maintenance. The G4nie Rural maintenance unit was set up and is still operating relatively efficiently. Nevertheless, it was and is too centralized and consequently the operating costs are too high. Whenever a pump needs repair a mechanic must go from Maroua to a village that can be 500 km away. However, little progress was made in getting the communities to take responsibility for part of the maintenance themselves. The high cost of a government-run maintenance service at about US$950 per pump/year and the difficulty in keeping an effective service emphasizes the importance of community participation in simple maintenance and repairs. The constraints were: lack of manpower to train and "sensitize" the communities, insufficient mobility, no organized institutional structure to take responsibility (the Department of Community Development of MINAGRI had almost no staff and no budget) and no political commitment. 5.28 Impact. In a zone which regularly suffers from drought and has been experiencing a drought over the past three years, reliable and clean water supply is crucial. A socio-economic survey carried out by MINAGRI/DEP (February 1983), studied the impact of the village water supply program and indicates that - about 816,000 people benefit from the improved wells (without pumps) and 226,000 from boreholes and wells fitted with pumps. Of these 85%.to 90% owned animals which also benefited -49 - from the water--a crucial aspect as can be seen from the abandonment of certain villages when the water point was still functioning but not providing sufficient water for the . . community's livestock; - the increase in operational water points has improved from the ratio of 1 to 1,200 people prior to the project to 1 to 792 by the end of 1983 a marked improvement but well short of the 1:560 target set in the appraisal report. - *the walking distance for people around the project water points surveyed decreased markedly (in the dry season) from 5 km on average prior to the installation of the borehole or improvement of the well to less than 1 km (525 m) after; - the quantity of water consumed increased from 15 to about 25 litres per person per day; - because of the installation of pumps, the time required to draw water decreased by a factor o: two or three; and - as yet it has not been possible to see the impact of cleaner water on health in the communities. VI. FINANCIAL AND ECONOMIC PERFORMANCE A. Financial Results 6.01 (a) Bottomlands. As the Moulvoudaye bottomlands did not retain sufficient water, and at this stage it is not known whether this is due to poor rainfall or design error (para 5.04), no meaningful result can be calculated at this juncture. Yet the likelihood that the estimated financial return to farmers of 22% as per the appraisal report can be achieved is doubtful. (b) Market Gardening and Stall Fattening Credit. The net annual financial return per family (i.e., gross value of production minus costs thereof, debt-service and losses) increased ranging from a modest 12% the stall-fattening to well over 100% for market gardening. Results for the stall-fattening are slightly weaker but in line for the market gardening when compared with the appraisal report estimates. It could be argued that the stall-fattening credit yields marginal returns to farmers, even more so if FSAR were to charge "real interest rates" which are closer to 15% than . the 10% charged now. However, given the limited options in the Mandara Mountains and the long tradition of this type of credit, it merits continued support. (c) Government. The price for not designing and implementing a cost recovery system for the borehole and well-digging is estimated at CFAF 145 million (US$300,000) per annum (1984/85 FSAR budget for these - 50 - components). This is based on a maintenance cost of about US$950 per pump/year and 315 installed under the project. It is estimated that about two thirds of this amount represent fixed costs to ensure the necessary infrastructure for maintaining pumps. Thus, annual operating costs per pump installed under the project can be estimated at US$315/year or about 33% of the purchase value. This cost of course will decrease as more pumps are being installed in the North and, assuming the Government envisaged program of 4,000 pumps as of the year 2000 will be achieved, it is estimated to average out to about CFAF 50,000 (US$110). As a result, recurrent cost implications per year are about CFAF 200 million plus replacement costs of CFAF 1.1 billion (US$2.5 million) per year and Government's costs (net of amortization of investments for borehole construction) for ensuring clean drinking water for the entire North of Cameroon (US$3.0 million) per year. B. Economic Results 6.02 (a) Market Gardening. With 1/3 of the FSAR's credit unit administration costs charged, a farmgate price of CFAF 65/kg of produce, an average useful life of 5 years for the pump and 15 years for the dugout, the rate of return was 20.9% at project completion. Although this looks good at face value, there are instances where farmer's post harvest losses reached beyond the 40% limit built in to the above analyses. This is mostly due to inadequate shortage practices. (b) Stall Fattening. Contrary to the optimistic economic rate of return estimate of 35% at appraisal, reality has clearly demonstrated that, with 21%, the rate is rather marginal. The same assumptions have been used as at appraisal, with the addition that benefits from a slightly increased production of groundnuts and sorghum, due to the application of manure from the steers, have also been included. The appraisal team underestimated farm labor input to fetch drinking water for the steers; the 20 days per operation originally estimated proved far too little and double that amount comes closer to reality. In some instances it is known that farmers' wives often spend as much as half a work day every day of the year for this activity. Moreover, the analysis suggests great sensitivity to cost fluctuations with a 10% increase in cost already yielding an ERR of 0%. Yet, given the severe lack of alternative economic activities in the northern region stall fattening must be considered a worthwhile undertaking. 6.03 Cost Recovery. One of the major issues during nepotiations and one which was never adequately tackled during project implementation was cost recovery and farmers participation. Each supervision mission discussed the issue with Government and various promises were elicited; the closest the issue came to being confronted was a proposition put forward regarding community taxation by MINAGRI in 1980. This was rejected by the local authorities. The issue was again discussed and, finally in 1983, was referred to the FSAR II project under which it was proposed that a "pilot recovery component" would be included, 6.04 Employment Created. The project had an important impact in the north. About CFAF 7i3 million was paid in local salaries during the - 51 - project period, the majority in the north (82%). The payment of labor over the project period was equally important, with the following figures recorded: Component CFAF millions Wells 278 Small dams 161 Support services of DFSAR and DOFSAR 140 579 At about CFAF 500 per man-day this would be equivalent to 1.2 million man-days of temporary employment. C. Farm-level Impact 6.05 Three of the sub-projects could be classified as productive in the agricultural sense. However, only the Market Gardening and Stall Fattening schemes could be analysed, as the Bottomland Development produced only a partial and very low harvest in one year. Thus the results are neither statistically meaningful nor representative. 6.06 Market Gardening. No detailed production nor cost of production statistics was collected during the project for the 60 motor pump credits provided under the project. However, two independent surveys were made of the sub-project in 1983 -.one a completion study by SEDA team which was responsible for the Rapport d'AchAvement and the other a socio-economic survey by MINAGRI. The tables detailing SEDA and MINAGRI analyses are in Annex 1, Tables 4 to 10 of the SErA study; the assumptions and results are summarized below: - the analysis compared farmers with a motor pump with farmers without a motor pump. MINAGRI refers to them as "before" and "after." Neither makes it clear what the without project case is but it can be assumed to be onion cultivation by chadouf. - those "without" were assumed to have 0.5 (SEDA) and 0.2 ha (MINAGRI) under onion cultivation (the latter is more realistic) and those with a motor pump 1 ha (SEDA) and 0.4 ha (MINAGRI). As it is known that a motor pump can irrigate 1.0 to 1.5 ha, the MINAGRI case is probably only for one of the two farmers sharing the pump. - the assumptions regarding input costs are similar for both and total about CFAF 180,000/ha with a motor pump and CFAF 125,000/ha without, but MINAGRI assumes no land preparation costs for the "before" case but CFAF 60,000/ha for the "after" case, exactly the same figure as SEDA study. - 52 - - after deducting debt service and family consumption the SEDA study shows a "Revenu Agricole" of CFAF 264,705 "with motor pump" compared with CPAP 59,225 "without". - to get a clearer picture, the table below compares the results on a per hectare basis. before/ after/ without with -- (CFAF/ha) '---- SEDA 113,450 264,705 MINAGRI 425,490 231,366 Neither of the analyses is satisfactory and the results are not reliable as indicators; furthermore, neither compares the economic performance of the credit recipients with the performance anticipated in the appraisal report. 6.08 Stall Fattening. No evaluation of the economic performance of bullocks fattened under the credit scheme was included in SEDA consultantion study. However, using figures generated by the project's research component, a farm budget (Annex 1, Table II of the SEDA study) compares the performance as forecast in the appraisal report with that actually achieved. The total net benefits for the two years were as follows: TR YR2 Total . -----(CFAF '000)---- Appraisal 92.4 200.8 293.2 Actual 119.7 200.6 320.3 In other words, the actual performance was better than the forecast in aggregate. However, a closer examination of the figures shows that there was a major difference in the estimation of the other farm production; this was balanced by an over estimate of the weight gain - 315 grams/day forecast versus 195 grams/day actual. Therefore, the performance in terms of weight gain was considerably below expectations due to poor weight gain in the rainy.season and difficulty in acquiring the cotton seed cake. It has been recommended that an 18 month fattening period with only one rainy season would give a better performance than 24 months as was used during the project. VII. BANK PERFORMANCE 7.01 The Bank correctly identified the need for a rural development fund in Cameroon. No mechanism previously existed for identifying, preparing, evaluating and channelling funds to small but priority projects in the rural sector. The project concept as developed by the Bank was thus sound and the preparation/appraisal process succeeded in narrowing down the - 53 - scope of the project for this first phase - which although not specifically referred to as such, was a pilot phase - to a limited number of projects in a geographically manageable area. However, looking at the final formulation of the project, particularly in the.light of implementation experience, two weaknesses appear. First, the state of preparation and to an extent knowledge of certain sub-projects was such that they were not ready for implementation. This was particularly true for the bottomland development sub-project and for the small dams sub-project. On the other hand, since the intention was to select small projects that could be replicated, a higher level of risk should have been accepted as, in many cases, such projects explore new grounds. If, however, one is to accept projects which have not been fully prepared, then proper provision must be made to do this, both in terms of timing and resources, in the early stages of project execution. Also, while the concept of using existing agencies to implement the sub-projects was sound, insufficient stress was placed on the selection of agencies with strong management and accounting structure, to handle the execution of the projects and associated monitoring and reporting activities, without intensive help from FSAR personnel. 7.02 In retrospect, there should have been greater emphasis on institutional aspects during project formulation. The selection of FONADER was a wise decision, but in structuring the management of FSAR within FONADER, it was necessary to make provision not only for facilitating the implementation of the five sub-projects already identified but more importantly to provide for the identification, preparation, evaluation and monitoring of new projects, including the establishment of a dialogue with potential executing agencies. The "unidentified projects" component was no more than a financial provision to initiate this process, but no staff was allocated nor organizational structure specifically designed for this task. Possibly, a special unit concentrating solely on this aspect should have been provided for in the Yaound6 DFSAR office -- a unit removed from the day-to-day implementation problems that took up so much of DOFSAR's staff time. 7.03 The Bank supervision of the project was intensive but concentrated too much on the performance of individual sub-projects and not enough on the build up of an institution capable of continuing after the project, to identify and facilitate the implementation of new micro-projects, including institutional arrangements with the executing bodies and the monitoring of micro-projects. The agencies selected to implement the sub-projects were weak; FSAR officials took an increasing role in their execution and the Bank supervision missions, which were generally oriented toward achieving numerical targets (number of boreholes sunk, number of credits disbursed, quality of onions produced and marketed, etc.), while knowing the increased FSAR involvement in execution, were not in a position to change things. If FSAR had restricted its direct involvement in sub-project execution, in all likelihood the sub-projects would have performed badly or failed. FSAR itself would probably have also been considered as a failure. Thus, the emphasis on institution building would have been wasted. In spite of this dilemma, more attention could have been paid by supervision missions to building up FSAR as a Rural - 54 - Development Fund instead of as an executing agency -- even to the extent of providing short-term technical assistance and for developing a system, for identifying, evaluating and funding the small priority projects in the rural sector. WAPAD June 1985 CM D M!MM CiT 723-CM Project Cst and FInng (in W '000) Appra1 Eat Atual >2·t£~1 at % Total Us$ Total I$ of Appral _ _e Ical Forg ~ ! alent oc re ~i ent Est te Botto~land 303 251 554 231 191 422 76 Maret Gardenrng 63 85 148 38 52 90 61 Stal1 Fattenrig 357 34 391 153 15 168 43 Unidentified 249 249 498 189 189 378 76 Small Dæs 748 876 1,624 1,096 1,287 2,383 142 Open Weils 730 950 1,680 1,125 1,465 2,590 154 lemp Infta11atonn 192 431 623 232 522 754 121 Boreholes Unit 280 976 1,256 363 1,266 1,629 130 Project Supporting Servics 688 878 1,566 1,927 2,459 4,386 280 Sub-Total 3,610 4,730 8,340 - - - Physical Cntingencies 370 470 840 - - - Price Contingencies 610. 800 1,420 - - - Total 4.600 6&000 10.600 5254 7,446 12800 121 Financing: Bank/IDA - - 7,000 - - 6,600 94 Borrawer - - 3,600 - - 6,200 172 1/ US$ = CFAF 245 at Appr~ial. 2/ US$ = FCA 305 (exchange rate prevailing during project execution). 3Ii - 56 - Annex Table 2 CAMEROON RURAL DEVELOPMENT PROJECT CREDIT 723-CM Partial Farm Budget Without Project With Project No PuMp* Chadoof Hand Pump* Motor Pump Area (ha): Sorghum/millet 1.3 1.3 1.3 1.3 Groundnuts 0.2 0.2 - - Onions (and other Veg.) - 0.2 0.2 0.5 Yields (kg/ha): Sorghum/millet 800 800 800 800 Groundnuts 750 750 750 750 Onions, etc. - 7,900 30,000 13,600 Production (kg): Sorghum/millet 1,040 1,040 1,040 1,040 Groundnuts 150 150 - - Cnions, etc. - 1,580 6,000 6,800 Losses - Onions () - 11% - 40% Value of Productip (CFAF '000) Sorghum/mil t - 52 52 52 52 Groundnuts 9 9 - - Onions, etc. - 79 375 255 Operating Costs-c/ (CFAF '000) Seeds - 5.4 5.4 13.5 Fertilizer - 4.8 4.8 12.0 Pesticides - 4.7 4.7 12.0 Hired Labor - 12.0 12.0 20.0 Fuel for Pump - - - 20.0 Maintenance - - 7.5 7.5 26.9 34.4 95.0 Net Value of Prod.(CFAF '000) 52 113.1 392.6 212.0 Home Consumption (onions) - 13.2 24.0 15.0 Net Income 52 99.9 368.6 197.0 Amortization d/ ~ - 17.0 17.0 Net Income after amortization 52 99.9 351.6 180.0 e/ *Appraisal Estimates a! Assumes 1 motor pump shared by two farmers and E.!rves 1.0 ha in total b/ Prices taken as average price in 1983: sorghum/millet CFAF 50/kg; groundnuts - CFAF 60/kg; Onions = CFAF 62.5/kg. c/ 0-tons only, other crops used an insignificant amount of paid inputs. d/ Five years. e/ This total must be divided by 2 to give the return to one farm family and thus be comparable with the other 3 totals. - 57 - Annex Table 3 CAMEROON RURAL DEVELOPMENT FUND PROJECT CREDIT 723-CM Key Indicators at Completion Date Item Unit Appraisal Actual Estimates Performance Wells No. reconstructed 1/ 600 904 No. maintained 790 568 No. capped 60 65 Boreholes Boreholes drilled - 285 Positive boreholes 2/ 430 232 Pumps Installed Pumps installed 490 315 Pumps serviced - 253 Small Dams Sites under construction - 2 3/ Sites completed 10 4 Bottomlands Cultivated area (ha) 300 300 Stall Fattening Loans approved 2,000 1,734 4/ Loans disbursed 2,000 1,334 Market Gardening Loans approved 60 60 Pumps installed 60 60 1/ Figures for actual performance may include some double counting because of reclassification from maintenance to reconstruction. 2/ Developed, tested and handed over for pump installation. 3/ On Government's own financing. 4/ Includes 400 loans already approved and to be disbursed under FSAR II. ．一．文了－ 州涌才p叮才右＇·抓 - 59 - Attachment I COMENTS FROM THE BORROWER (Page 1 of 5) E-126/87 August 11, 1986 French (Cameroon) OEDI ENMcM:bas Fonds National de D6veloppement Rural - FONADER (National Rural Development Fund) Yaound6, Cameroon July 22, 1986 The Director General Directorate of Programs No. 5820/DG/DPR/CT The Resident Representative of the World Bank Yaound6 Subject: FSAR I (Credit 723-CM): Project Performance Audit Report I have the honor to request you to forward the attached memorandum regarding the Project Performance Audit Report for the Rural Development Fund Project (Credit 723-CM) to the World Bank's Operations Evaluation Department. Thank you for your assistance in this matter. Yours$ etc. Isl Albert Ngome Kome Director General Encl. - 60 - Attachment I (Page 2 of 5) FONDS NATIONAL DE DEVELOPPEMENT RURAL Yaounde July 21, 1986 No. 5791/DPR/CT COMMENTS ON THE PROJECT PERFORMANCE AUDIT REPORT FOR THE RURAL DEVELOPMENT FUND PROJECT (Credit 723-CM) 1. Introduction This memorandum records the comments and observations of FONADER on the Project Performance Audit Report for the first phase of the FSAR project (FSAR I), the Credit Agreement for which was closed on December 31, 1983. The draft report was prepared by the World Bank's Operations Evaluation Department in May 1986. 2. Comments and Observations by FONADER 2.1 General In general, FONADER's views coincide with the conclusions set forth in the referenced report, particularly as regards the following points: - The involvement of beneficiaries in maintenance of wells and pumping facilities was inadequate. This deficiency could jeopardize continuity of FSAR's village water supply program. - The Dams subproject was insufficiently studied, from both the technical and the financial point of view (volume of work and final cost both underestimated). - Insufficient attention was paid to the training of Cameroonian nationals. - The success or failure of the subprojects depended essentially on the abilities and interest of the executing agency, which explains the variability of the results obtained during FSAR I. - The management of FSAR failed to develop adequately its role in planning, supervising, identifying and appraising new subprojects. - The Bank supervised the Project intensively, but it focused too much on the physical infrastructure and not enough on institution-building. - The bottomlands in the North required less sophisticated and less costly development than those in the Moulvoudaye region. However, the report does call for a number of comments from FONADER. - 61 - Attachment I (Page 3 of 5) 2.2 Comments on the Project Performance Audit Memorandum (PPAM) 2.2.1 There are a couple of typing errors on page iii: - The date of signing of Credit Agreement 723-CM should be "04/08/77" in the French text, instead of "08/04/77." - The number of the Credit for the second FSAR project should be 2567-CM, instead of 723-CM, which is the number of the Credit for FSAR I. 2.2.2 Page v of the French text (page iv of the English text). The total cost of the FSAR, including technical and financial contingencies, was appraised at US$10.6 million -- not US$8.5 million, which reflects only the base costs (cf. Appraisal Report No. 1479-CM dated May 26, 1977, page 18 of the French text and page 15 of the English text). Thus the actual cost, including physical and financial contingencies, of US$12.8 million represents an overrun of 21% compared with appraisal, not 50% as stated in the PPAM. 2.2.3 Page vi of the French text (page v of the English text). The Dams subproject was carried out by the Service de Genie Rural (Rural Engineering Service) of MINAGRI, and not by the Service des Travaux Publics (Civil Works Service). 2.2.4 Page vii of the French text (page vi of the English text). According to FSAR's records, at least 250 (and not 232) boreholes were drilled during FSAR I. Moreover, more than 1,400 (not 900) wells received maintenance (568) or were re-excavated (908). This was in addition to 45 new wells dug (cf. PCR, para. 2.13, page 19 of the French text and page 16 of the English text). 2.2.5 Page x of the French text (page ix of the English text). With regard to the disparity in remuneration between the staff of FONADER and the staff of the implementing agencies, it should be borne in mind that the objective of the FSAR is to strengthen existing services, not to replace them or absorb their staff. To this end, FSAR provided the implementing agencies with equipment, supplies, etc. and extended substantial benefits to staff working in the subprojects (efficiency bonuses, lump-sum relocation allowances, etc.). If the entire staff were to have the same terms of employment as the staff of FONADER and be supported by FSAR, the cost would be exorbitant. This is yet another reason for choosing "development companies" as executing agencies, since their staff have more or less the same terms of employment as that of FONADER. 2.2.6 Para. 15, page 9 of the French text (page 8 of the English text). Three expatriates, not four, successively headed the Dams subproject (cf. PCR, page 27 of the French text and page 23 of the English text). 2.2.7 Para. 18, page 9 (sic - should be page 11) of the French text (page 10 of the English text). Until early 1984 there was a radio link between FONADER and the North of Cameroon (Ngaound6r', Garoua and Marou4). Communication between SDFSAR and DOFSAR in Maroua has improved since the installation of a telex link between headquarters and the FONADER agency in Garoua, as the latter's telephone line to Maroua functions more or less - 62 - Attachment I (Page 4 of 5) normally. It is planned to install telex equipment in the Maroua agency. A weekly mail system has been set up linking FONADER's headquarters with all its agencies. 2.2.8 Para. 24, page 15 of the French text (page 13 of the English text). Regardless of the drought, it is normal for the yields from the Moulvoudaye bottomlands to be lower than those in other SEMRY schemes, given the better growing conditions to oe found in the latter (full control over irrigation use, land leveling, transplanting of seedlings, and intensive use of fertilizer). In 1985, when rainfall amounted to 720 mm (average for the region = 750 mm), the Moulvoudaye reservoir was filled to the brim (1.2-1.5 meters of water). The decision to cultivate the bottomland area was taken late (in July), and as a result only 30 hectares were planted to rice. The yield is estimated to have been 2.8 tons/ha. The volume of paddy rice brought to market was 48.5 tons, giving a marketed yield of 1.6 t/ha. The experience of 1985 shows that the Moulvoudaye scheme can function when rainfall is close to average. 2.2.9 Para. 25, page 17 of the French text (page 15 of the English text). At the time of project completion, loan recoveries stood at 75%, not 42% (see comments under 2.3.6 below). 2.3 Comments on the Project Completion Report (PCR) prepared by the Western Africa Regional Office 2.3.1 Para. 3.09, page 18 of the French text (page 15 of the English text). The final report on research on livestock fattening was published in November 1983 by the "Centre National de Formation Zootechnique et V6t6rinaire" (CNFZV). It was sent to the World Bank under cover of letter No. 1918/DG/DPR of March 15, 1984. It was also discussed at length in the FSAR I completion study prepared by SEDA. 2.3.2 Para. 3.13, page 19 of the French text (page 16 of the English text). The reason why the Wells component involved much more reconstruction (deepening) and less maintenance than expected was that the 1976 inventory prepared by G4nie Rural (which served as the basis for the appraisal) showed many wells as requiring only simple maintenance whereas 2-3 years later they had sanded up and become unusable, thus necessitating their re-excavation. This fact argues strongly in favor of continuing well maintenance operations in FSAR II if they are to be kept in working order. 2.3.4 Para. 3.14, page 18 of the French text (page 16 of the English text). According to FSAR's records, at least 250 (not 232) tubewells were drilled. 2.3.5 Para. 3.21, page 21 of the French text (page 18 of the English text). When comparing the actual cost of the project with the estimates made at the time of appraisal, it would be preferable to take the estimated total cost (including technical and financial contingencies) rather than the base cost. With a total appraisal cost of US$10.6 million and a final cost of US$12.8 million, the overrun was 21% rather than the 51% shown in the PPAR. - 63 - Attachment I (Page 5 of 5) 2.3.6 Para. 5.07, page 31 of the French text (page 27 of the English text). We feel that the recovery rate on the Market Gardening subloans should be calculated in terms of total recoveries as a percentage of total amounts payable (principal + interest) rather than of total annual amounts due (principal + interest + arrears) as yearly arrears are not cumulative. Against total amounts payable of CFAF 32.4 million (cf. SEDA completion study, p. 83, Table 16), payments received amounted to CFAF 24.5 million, giving an overall recovery rate of 75.6% at December 30, 1983 (and not 45% as computed by the Bank).