Burkina Faso - Third Rural Development Fund Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 13225 PERFORMANCE AUDIT REPORT BURKINA FASO THIRD RURAL DEVELOPMENT FUND PROJECT (CREDIT 1218-BUR) JUNE 28, 1994 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. CURRENCY EQUIVALENTS Currency Unit: SDR US$1.00 = SDR ABBREVIATIONS AND ACRONYMS BOAD - Banque Ouest Africaine de D6veloppement West African Development Bank CNCA - Caisse Nationale de Cr6dit Agricole National Agricultural Credit Bank CRPA - Centre R6gional pour la Promotion Agropastorale Regional Center for Agropastoral Promotion FAER - Rural Equipment Assistance Fund FDR - Fonds de D6veloppement Rural Rural Development Fund FEER - Fonds de l'Eau et de I'Equipement Rural Water and Rural Equipment Fund HER - Direction de I'Hydraulique et de l'Equipement Rural Rural Water Supply and Engineering Department ICRISAT - International Center for Research in the Semi-Arid Tropics MAE - Ministbre de l'Agriculture et de l'Elevage Ministry of Agriculture and Livestock ME - Ministbre de l'Eau Ministry of Water MQP - Ministbre de la Question Paysanne Ministry for Peasants' Affairs MRD - Ministre du Ddveloppement Rural Ministry of Rural Development NLM - National Land Management ORD - Organisation Rdgionale de D6veloppement Regional Development Organizations PAR - Performance Audit Report PCR - Project Completion Report PNGTV - Village Land Management Program PPD - Popular Development Program PY 1 - project year 1 PY 4 - project year 4 RAF - Loi sur la R6organization Agraire et Foncire Agrarian and Land Tenure Reorganization Law RDF - Rural Development Fund Project FISCAL YEAR January I - December 31 FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-General Operations Evaluation June 28, 1994 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Performance Audit Report on Burkina Faso Third Rural Development Fund Project (Credit 1218-BUR) Attached is the Performance Audit Report on Burkina Faso - Third Rural Development Fund Project (Credit 1218-BUR) prepared by the Operations Evaluation Department. The project was the third of a series designed to provide financial support through the Rural Development Fund (RDF) for small scale rural operations with an emphasis on the promotion of productive activities, particularly land development. By project closing, most of the physical targets had been achieved, but there remains concern over the sustainability of some of the investments. The audit agrees with the conclusion of the PCR that the project was marginally satisfactory and that its sustainability is uncertain. The project was implemented during a period of great upheaval in Burkina Faso, characterized by sharp changes in policies and institutional structures. Most specifically for the project, institutional changes gravely weakened local level capacity to select and coordinate the implementation of small projects, financed through an agency such as RDF and its successor (FEER). In response FEER attempted to offset this by increasing its own capacity to support implementation. However, this created a centralized implementing agency, whose size and top down operating style became a matter of concern to the Bank and the cofinanciers, who pressed FEER to narrow its role and streamline its operations. For this reason among others, the Bank did not support further FEER operations. Changes have been made by FEER in response to these pressures, but it is the judgement of the audit that their impact may be limited, because of more fundamental issues related to the nature of an intermediary institution such as FEER. These relate to the inadequacy of the local institutional capacity in Burkina Faso to coordinate the development of local infrastructure; the pressures on FEER as an intermediary institution to broaden its range of activities; and the very different operational demands which many of these activities create. The principal lesson of this experience is that the structure and modus operandi of an institution such as FEER, designed to channel funds to local projects, cannot be considered independent of the institutional structure in which it operates. Robert Picciotto by H. Eberhard K6pp 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.  PERFORMANCE AUDIT REPORT FOR OFFICIAL USE ONLY BURKINA FASO THIRD RURAL DEVELOPMENT FUND PROJECT (CREDIT 1218-BUR) TABLE OF CONTENTS Page No. PREFACE .........................................i BASIC DATA SHEET .................................iii EVALUATION SUMMARY .............................. v 1. INTRODUCTION ........................... 1 Background .................................... 1 2. PROJECT DESIGN .......................... 2 Project Objectives and Description ...................... 3 Negotiations .................................... 5 3. IMPLEMENTATION ......................... 5 Background to Implementation. ........................ 5 Project Results. .................................. 7 Physical Achievements............................7 Economic Returns ............................... . Compliance with covenants. .......................... 9 Mid-Term Review. ................................ 9 The Termination of Bank Involvement........................ 11 4. ISSUES AND LESSONS............................. 13 Was the Bank Criticism Justified?.......................... 13 The Role of the Intermediary Institution....................... 13 The Different Operational Demands......................... 13 Local Infrastructure.................................. 14 Agricultural Investments............................... 14 Commercial Investments............................... 14 Institutional Structures.................................. 14 What Should Happen Now?.............................. 16 Postscript. ........................ ......8.... 17 This report was prepared by John English (ask Manager), with administrative support provided by Constance Frye. This document has a restricted distribution and may be used by recipients only in the performance of their oficaldte Ifferonent ma oeria Demadlse .ihu .ol Bnauhrzto.1  PERFORMANCE AUDIT REPORT BURKINA FASO THIRD RURAL DEVELOPMENT FUND PROJECT (CREDIT 1218-BUR) PREFACE This is the Performance Audit Report (PAR) on the Third Rural Development Fund Project for which Credit 1218-BUR in the amount of SDR13.6 million was approved 30 March, 1982. The original Closing Date, 30 June 1987, was extended by three years to 30 April 1990. The Borrower obtained project co-financing for this project from the Netherlands, US$7.8 million, and from the West African Development Bank (BOAD), US$3.7 million. The PAR is based on the Project Completion Report prepared by the Africa Regional Office (Parts I & III) and by the Borrower (Part II), on the Staff Appraisal Report, the President's Report and the legal documents, on study of the project files, on the External Evaluation Report prepared in 1988, discussions with Bank staff, and on the findings of an OED mission which visited Burkina Faso in September/October 1993. The mission discussed the effectiveness of the project with Government officials from the project authority, FEER (Water and Rural Equipment Fund), the Ministries of Agriculture, Water and Finance, as well as private citizens and representatives of international organizations. The mission met with representatives of DANIDA and Swiss Aid in the field and has held discussions with Netherlands Aid. The PCR provides a generally frank and comprehensive account of project experience, including performance of the Bank and the Borrower, including executing agencies and other participants in the project. It correctly concludes that the project was only partially successful in supporting viable small investments in rural areas of limited agricultural potential in Burkina Faso. The audit addresses the appropriateness of the measures pressed by the Bank to improve this performance. The draft report was sent to the Borrower and cofinancier for comments but none were received.  PERFORMANCE AUDIT REPORT BURKINA FASO THIRD RURAL DEVELOPMENT FUND PROJECT (CREDIT 1218-BUR) BASIC DATA SHEET Key Project Data Item Appraisal Actual Actual as % of Estimate Appraisal Estimate Total Project Costs (US$ million) 28.9 26.1 90 Credit Amount (US$ million) 16.0 16.67 104 Cofinancing - BOAD (US$ million) 3.7 n.a. n.a. - Dutch (US$ million) 7.8 5.0 64 Economic Rate of Return Cumulative Estimated and Actual Disbursements (in US$ m) FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 Appraisal Estimate 2.1 5.1 8.7 12.7 15.5 - - - Actual 0.69 2.0 3.5 3.9 7.4 11.6 15.6 16.7 Actual as % of Appraisal 33 39 40 31 48 73 97 104 Date Final Disbursement April 25, 1990 Project Dates Original Revised Actual Appraisal 03/11/81 - 03/16/81 Negotiations 12/16/81 - 12/16/81 Board Approval 03/30/82 - 03/30/82 Credit Signing 10/28/82 - 10/28/82 Credit Effectiveness 05/04/83 - 05/04/83 Credit Completion 12/30/86 06/30/88 12/31/89 Credit Closing 06/30/87 12/30/88 04/30/90 a! Not available. b/ Estimates made for directly productive sub-components only. - iv - Staff Inputs (Staff weeks)&/ FY77 FY80 FY81 FY82 FY83 FY84 FY85 FY86 FY87 FY88 FY89 FY90 FY91 FY92 Total Preparation 0.4 1.4 5.3 24.9 - - - - - - - - - - 32.0 Appraisal - - 37.0 23.0 - - - - - - - - - - 60.0 Negotiation - - - 30.1 - - - - - - - - - - 30.1 Supervision - - - 6.3 16.9 10.3 21.3 8.6 13.3 14.7 7.1 3.7 - 1.7 103.9 PCR - - - - - - - - - - - 0.1 0.7 - 0.8 Total 0.4 1.4 5.3 61.9 16.9 10.3 21.3 8.6 13.3 14.7 7.1 3.8 0.7 1.7 226.8 Mission Data Date No. of Staff Days Performance Month/Year Persons in Field Specializationji/ Ratingg/ Problemsf/ Identification Preparation Appraisal Supervision 1 04/82 2 12 a,b 2 M Supervision 2 01/83 2 7 a,b 2 M Supervision 3 06/83 1 3 a 2 T Supervision 4 10/83 2 10 a,b 2 F,M,T Supervision 5 05/84 2g/ 6 a 2 M,T Supervision 6 12/84 4g/ 17 b,c,d 2 P,T Supervision 7 06/85 1 8 d 2 F Supervision 8 10/85 2 3 d hi Supervision 9 04/86 1 5 a Supervision 10 10/86 1 11 a Supervision 11 03/87 1 10 a Supervision 12 11/87 Supervision 13 02/89 C/ No information is available from the MIS on FY78-79. A/ a = agriculturalist; b = financial analyst; c = economist; d = rural engineer. C/ 1 = no or minor problems; 2 = moderate problems. V/ F = financial; M = managerial; T = technical; P - political. g/ IDA staff and consultants and cofinancier staff or consultants. h/ With introduction of new 590 format, problem identification no longer took this form. PERFORMANCE AUDIT REPORT BURKINA FASO THIRD RURAL DEVELOPMENT FUND PROJECT (CREDIT 1218-BUR) EVALUATION SUMMARY Background 4. The ORDs coordinated RDF activities at the local level. RDF I and II were considered to 1. The Bank has been engaged in lending have been successful in creating an efficient operations in Burkina Faso for over 20 years. organizational tool for appraising small, rural The primary thrust of agricultural lending has investments, providing sound financial control evolved from explicitly commodity-oriented during their implementation, and building up the operations, through area development projects, capability for monitoring such projects. to countrywide service projects. Project Design 2. Parallel to these projects was a series of three supporting the Rural Development Fund 5. Steps to prepare a third project began in (RDF). The objective of this fund was to fi- 1978, half way through the life of RDF I. By nance a variety of small investments which, the late 70s concerns were arising about the size because of their size and scattered locations, and cost of the ORDs and their ability to finance were otherwise not suited to conventional Bank themselves. It was suggested that they be scaled or other external lending. The first of the series back and focus more on agricultural extension was approved in 1972. This operation was and related activities. The potential role of RDF considered to have been successful, and a if such change occurred was raised early in second project (RDF II, which was essentially preparation but not pursued. The more specific similar to RDF I) was approved in 1976. issues in the appraisal process related to: the geographical location of operations to be sup- 3. These two projects were heavily depen- ported; the range of activities; RDF's role in dent upon the Regional Development Organiza- implementation and the use of contractors; and tions (ORDs) for their implementation. ORDs the extent of recurrent costs to be covered by were created in 1966 as means of decentralizing external financing. agricultural development activities. They were empowered to undertake activities in agricultural 6. Several specific sub-project types were marketing and input supply and to generate included in the project: land development funds to support agricultural activities. In the (including anti-erosion bunds, bottomland devel- 1970s they became the locus for integrated rural opment, and small scale irrigation); credit to development activities and were expected to farmers for the purchase of oxen and animal execute multi-sectoral functions and coordinate drawn equipment, and to villages for purchase of the activities of departments and ministries in grain mills; village level storage and cereal their areas. banks; village tree plantations and water sup- plies. These were designed primarily to increase - vi - crop production, and also to improve the avail- implementation, namely the performance of ability of safe drinking water for rural communi- some of the components, and the role of FEER. ties, increase fuelwood production, and increase the value of agricultural produce through pro- 9. Component performance. In August cessing and better storage. The organizational 1984 the Sankara regime, anxious to speed up arrangements, as before, relied heavily on the the provision of social and economic infrastruc- ORDs. The choice of villages and communities ture in the rural areas, pressed the Bank to agree for support, and the type and size of invest- to an expanded rate of implementation under the ments, were to be decided at the local level RDF Project. However, Bank staff had already during the course of implementation. The become concerned about both a lagging rate of project was approved in March 1982 with cofi- implementation under the project and the short- nancing from The Netherlands and the West fall in the Government's contribution to this African Development Bank. program and, therefore, had doubts as to whether an expanded program was possible. Implementation Experience 10. A joint Bank/Netherlands Aid mission 7. The audit concurs with the conclusion of raised concerns about the effectiveness of several the PCR that the project was marginally satisfac- of the components and concluded that an expan- tory, and its sustainability is uncertain. It was sion on the scale envisaged by Government was implemented during a period of great upheaval. not feasible or desirable. A limited expansion of The principal turning points were the revolution the cereal bank and borehole programs was of 1983, which brought the Sankara Regime to finally agreed. The outcome of the mission power, and the events of 1987, which ended it. generated considerable tension between the Bank During the whole project period there were and Government, but it also resulted in FEER frequent government reshuffles and changes in focussing more directly on the performance of the structure of ministries and other government the major components in its 'portfolio'. The entities. The most significant changes were the results of this review did substantiate some splitting of the Ministry of Rural Development questions raised by the 1984 mission and, for into the Ministry of Agriculture and Livestock example, a different approach was adopted to and the Ministry of Water (ME), and the trans- erosion control work. formation of RDF into the Fund for Water and Rural Equipment (FEER), which reported to 11. By project closing, most of the physical ME. In 1987 it was decided to scale back the targets had been achieved but there remains ORDs to Regional Centers for Agropastoral concern over the sustainability of the invest- Promotion (CRPAs) with a primarily extension ments, in particular poor maintenance or function. abandonment of erosion control measures and improved land. In the audit's judgment these 8. These upheavals had two main results. problems may have been exacerbated by the First, they created uncertainty which affected project selection process (see below). civil servants in particular. These conditions inhibited innovation and led to a decline in 12. Role of FEER. The latter half of the quality of work. Second, RDF-FEER attempted extended project period coincided with the to maintain the momentum of its activities by development of what has become known as the increasing its involvement in technical matters "Village Land Management" (Gestion de and in project execution. The problems outlined Terroir) approach to dealing with land issues. in para. 7 were also the major factors in the two From 1985 this issue began to dominate the principal issues which arose during project interaction between the Bank and the - vn - Government. The supervision of RDF I Was the Bank Criticsm Justified? became increasingly linked to the discussions of this issue, and the preparation of a project to 15. While changes have been made by attempt to tackle it. Discussions about the FEER, it is the judgement of the audit that their potential role of FEER in this new initiative impact may be limited. This is because of the became focussed on two perceived issues related existence of more fundamental issues related to to the modus operandi of FEER; the need for the nature of an institution like FEER. These FEER, first, to avoid becoming directly involved relate to: the range of activities which FEER in implementation of projects it was financing finances and the very different operational and, second, to attempt to increase the degree of demands they create; FEER's position as an local participation in the process of the selection intermediary institution; and to the institutional of projects to be financed. vacuum for development of local infrastructure in which FEER currently operates. 13. In addition, Netherlands Aid, the other major financier, became uneasy about the effec- 16. Operational demands. The range of tiveness of FEER operations at the local level. activities now supported by FEER can be divid- In 1988 an external review was undertaken ed into three broad categories: which expressed concern, first, over the high share of project expenditure which was going to (a) local infrastructure, such as schools, FEER's administrative costs; second, that opera- health and maternity clinics, and water tions were target driven and village groups supplies. (which often were weak and lacked cohesion) felt little obligation to repay loans and; third, (b) agricultural investments, such erosion that this was exacerbated by the fact that the control measures, as the development of institutional changes which had occurred had bottomlands, and small scale irrigation. weakened the relationships between FEER and its collaborating agencies, so that technical back (c) small commercial type enterprises, such up to village groups was often inadequate. In as village stores, pharmacies, and grain particular the report stressed that, if a FEER mills. type program was to continue, much clearer contractual arrangements between the collaborat- This range of activities inevitably places quite ing agencies should be required as a condition different demands for back up support on a for the provision of funds. financing agency such as FEER. While the ORDs existed they were involved in most of 14. The external review and the Bank's these activities and could, in principle, provide views led to FEER carrying out a review of its field support to implementation. However, approach and operations in 1989. That review trying to cope with such a range of activities was resulted in changes being instituted which did a major factor in the problems which led to their respond to some of the issues raised by the Bank dissolution. and the Dutch. However, despite these measures, both donors decided not to provide 17. The role of the intermediary Institu- furthier support to FEER. ton. FEER is primarily funded from external sources, the main vehicle currently being the FEER I Project (which followed RDF IM). While FEER is now trying to operate through contractual agreements with its collaborating agencies, many of them are weak and lack - viii - resources. However, FEER will be judged by 20. At present there is no structure in the extent to which it has used the resources Burkina to resolve these dilemmas. The ORDs made available to it and will be expected to have had this function but have been eliminated. achieved measurable results on the ground. If There is now no formal requirement for review its contractual partners fail to utilize the funds and approval at the local level, for example, by made available, then FEER inevitably will have the Prefect or High Commissioner. Agencies difficulty obtaining further funding. Thus, it is such as FEER meet on a non-formal basis to almost inevitably going to be drawn in to fill the keep abreast of one another's plans, but such gaps in the performance of the collaborating meetings are periodic and carry no weight. agencies, as it was during the earlier Bank Even if FEER was designated by government as funded projects. the agency responsible for coordinating such activities, it is doubtful that it could do so 18. Institutional Structures. The current effectively. arrangements put FEER in much more of an arms-length relationship with the villages being Conclusions and Lessons assisted. The primary points of contact between FEER and the villages are now the CRPAs. 21. The 1988 external review and Nether- Their function is to promote agricultural produc- lands Aid were critical of the share of project tion, and their best contacts are likely to be with expenditure which went for FEER's own operat- those villages which either have greater agricul- ing expenses. They stressed the need for a tural potential or have individuals (especially in smaller, leaner agency to serve as the source of leadership positions) who at least are willing to finance for local level projects. However, for try out new agricultural ideas. These are, the reasons noted above, this appears to be therefore, most likely to be linked up to the unrealistic. There is not the credit, commercial FEER program. The effect may be that the or institutional infrastructure in place for such a choice of villages for support under programs lean agency to function. Thus, the efforts made such as FEER reflects the priorities of the by the donors in the latter stages of RDF m, intermediary agency rather than those of the area and which played a significant role in the Bank's as a whole. deciding not to back FEER 1, were not able to effectively resolve the basic causes of the prob- 19. Given the current conditions in most lems identified. villages in Burkina Faso, improvements in basic social infrastructure are a high priority for the 22. There are two main lessons from this villagers and the FEER I program has increased project: the emphasis on them. The villages selected by the CRPA will press to obtain some improve- (a) While changes to improve the effective- ments in social infrastructure from the FEER ness of FEER operations were required, program, and are likely to undertake agricultural changes were also necessary in the local measures promoted by CRPAs to obtain them, institutional structure, whose capacity to setting the stage for subsequent poor agricultural select and coordinate the implementation performance. It is also quite likely that, for of small local projects had been gravely example, the priorities for improved health weakened by the turmoil and change facilities from the point of view of improving during the 1980s. the overall access to care, might be quite differ- ent from those thrown up by the response to (b) In regions such as that covered by this agricultural priorities. project, with low agricultural produc- tivity and/or limited market potential for - ix - increased production, making agricul- 24. Although the Bank is no longer financing tural production activities the key to FEER this issue continues to be a live one for support of local projects is likely to lead the Bank in the context of the ongoing Environ- to distortion in the pattern of investment mental Management Project. This project is and to poor performance of agricultural attempting to help rural communities to design activities. and implement sustainable management plans for their lands and natural resources. However, in many cases effective land management is ham- Next Steps pered by the lack of access to markets for poten- tial products which would justify investment in 23. What appears to be desirable is a recre- resource improvement. Unfortunately, the ation of the local coordinating function held, in project only has funds to finance natural principle, by the ORDs, before the changes of resource management measures, such as erosion the past decade. This might best be managed control efforts. Attempts are being made to from the offices of the Prefect and High Com- access other funds to support linked local infra- missioner, and would identify more general structure development but, in the absence of any development and investment priorities, as identi- general local planning and implementation fied by the villages. A structure of this type is coordination structure, everything has to be done being used on a limited scale in the ongoing on an ad ho basis, greatly increasing the proj- Environmental Management Project (Cr. 2229- ect's management load. BUR). The identified priorities could then be addressed, on the basis of available resources and personnel, by the agency most suited to do so. At the same time FEER might undertake a narrower role. For example, it might concen- trate either on the construction of rural social infrastructure, or on agricultural investment.  -1- PERFORMANCE AUDIT REPORT BURKINA FASO THIRD RURAL DEVELOPMENT FUND PROJECT (CREDIT 1218-BUR) 1. INTRODUCTION Background 1.1 Burkina Faso is a landlocked country in the semi-arid (Sahelian) zone of West Africa. Rainfall ranges from about 500 mm per year to over 1000 mm, but is highly variable. Vegetation is largely scrub or tree savanna, and soils are poor. Water supplies are erratic and population density is relatively low. In most of the country, subsistence cropping, based on shifting cultivation and an extensive fallow period, is typical. In a few, more arid, but relatively tsetse free, areas of the country a limited share of the population is dependent on livestock herds for subsistence. Because of highly variable rainfall, the risk of drought and crop or livestock loss is high. Farmers and herders employ risk limitation strategies, including transhumance for livestock. 1.2 The population of Burkina Faso is almost entirely rural. Currently known and economically exploitable natural resources, other than land, are minimal. Seasonal migration to seek employment or some other source of income is common. Considerable migration to the coastal countries, especially Cote d'Ivoire, has taken place. The low level of natural resources, including the difficult conditions for agricultural activity, and the relatively low population, have meant that much of the Sahelian region, including Burkina Faso, is economically marginal, thus compounding the problem of fostering development. 1.3 The Bank has been engaged in lending operations in the country for over 20 years. Its operations in the agricultural sector have had two main elements. The primary thrust of agricultural lending can be categorized as having three phases. The first was explicitly commodity oriented (primarily cotton). The second broadened the focus to agricultural development as a whole, especially to grain production, and to the desirability of integrating farming and livestock rearing in order to maintain soil fertility. This was reflected in a series of area development projects. The third phase, currently ongoing, stresses country wide service projects. 1.4 Parallel to these, largely production oriented, projects were a series of three supporting the Rural Development Fund (RDF). The objective of this fund, which was established in the National Development Bank, was to finance a variety of small productive investments which, because of their size and scattered locations, were otherwise not suited to conventional Bank or other external lending. The first of the series was approved in June, 1972 and included support for wells, small village storehouses, the development of valley bottom lands and small irrigation schemes, erosion control works, and improvement of feeder roads. An essential point in the selection of specific investments for funding was the interest of the people concerned, expressed through their willingness to provide labor for the initial investment, and for subsequent operations and maintenance. -2- 1.5 This operation was considered to have been successful. In fact in 1973-74, when efforts were being made to undertake activities to ameliorate the effects of the drought of that period, the RDF was to prove an invaluable channel for funds and was considered to have been a major reason for the more effective use of funds in Burkina Faso than elsewhere. A second project (RDF II) was approved in May 1976. At that time it was considered that the major problems experienced under the first project were (a) the need to strengthen staff in the technical and executing agencies (principally the Regional Development Organizations (ORDs)); (b) the problem of ensuring the maintenance of assets created with RDF financing; and (c) the importance of securing credit to enable farmers to purchase complementary inputs and equipment, in the absence of which the full value of the investments undertaken was unlikely to be realized. 1.6 The ORDs were the principal collaborators with RDF. They were created in the 1960s to permit a devolution of authority for agricultural development activities to the regional level, and were supervised by the Ministry of Rural Development (MRD). In the 1970s they evolved into implementing agencies for multisectoral activities such as development of infrastructure, agricultural development (including marketing and input and credit delivery), and literacy training. Some of these activities were potentially revenue generating. In addition the ORDs were expected to coordinate the activities of national departments and ministries within their jurisdiction. In this they collaborated with the Prefect who was a member of their administrative council. To enhance collaboration between RDF and the ORDs, RDF was transferred from the Development Bank to MRD. 1.7 For RDF operations the project cycle for the investment subprojects at local level was conceived in order to involve the beneficiaries in the identification and design of these subprojects with the backup of technical agencies at the regional (ORD) and the central (RDF) levels. Pre-appraisal was to be done by the ORDs and the appraisal by RDF. The latter was to decide on the financing, supervise the implementation and carry out ex-post evaluations. The RDF Board would approve the annual program, to be executed by the competent agencies, with the beneficiary communities providing labor, and occasionally other inputs, in kind. 1.8 RDF II, concentrating in particular in the Central Plateau, continued with the subprojects considered successful under RDF I, but excluded feeder roads, for which a national agency had been created. It also included elements to provide credit and for the strengthening of the staff of the technical and implementing agencies. Results were generally considered satisfactory. The major success of the project was in creating an efficient organizational tool for appraising small, productive rural investments, providing sound financial control during their implementation, and building up the capability for monitoring such projects. 2. PROJECT DESIGN 2.1 Steps to prepare the Third project began in 1978 half way through the life of the second project. By the end of the 1970s, some concerns were being raised over the effectiveness of the ORDs. While in principle they had been encouraged to engage in commercial type activities to stimulate market oriented agricultural production and as a source of funds, the extent to which they could do so varied widely across the country. For reasons typical of parastatals they also often generated losses rather than profits. They were also becoming unwieldy and very difficult to manage and concerns were expressed that they were not providing effective, basic extension support to farmers. A commission was created to study the role and functions of the ORDs. This commission recommended that the ORDs be scaled -3- back and become the regional operating arm of the MRD and that an Economic and Social Coordinating Committee be set up, headed by the Prefect, to coordinate the activities of the various agencies in the regions. However, a Bank sector report noted that "if the Coordinating Committee is not created, it is unlikely that the ORD alone can coordinate development activities over which it has no authority." Unfortunately these proposals were overtaken by events and the problems never effectively resolved. 2.2 Such changes would also have had implications for the role and operations of RDF. The question of the Fund's future role was referred to, somewhat obliquely in some of the materials related to the preparation of RDF I, in particular how all of the different externally funded rural development activities were to be coordinated. However, this fundamental question was never really resolved and to this day the issue remains a live one. 2.3 Preparation and appraisal focussed primarily upon nuts and bolts issues of the project. The more specific issues addressed during the process related to: the geographical location of operations to be supported; the range of activities; RDF's role in implementation and the use of contractors; and the extent of recurrent costs to be covered by external financing. The appraisal mission was accompanied by representatives of Dutch Aid (who cofinanced the project). In their supplementary report they raised a number of issues related to the nature of village groups (which were the primary interlocutor with RDF at the village level and managed many subprojects), and the nature of their interaction. These appear not to have been addressed in any depth and, in fact, the Dutch report indicated a belief that not enough time was allowed in the mission for discussion and resolution of strategic issues of this type. Subsequent events suggest that this was a relevant concern. Project Objectives and Description 2.4 The project aimed at completing the process pursued under the first two projects, that is the development of a reliable and cost-effective capability for managing small-scale rural investments. The specific sub-project types identified were designed primarily to increase crop production, and also to improve the availability of safe drinking water for rural communities, increase fuelwood production, and increase the value of agricultural produce through processing and better storage. 2.5 The project, as approved, included the following components: (a) land development comprising: engineering design services, implementation and supervision of a program of (i) anti-erosion bunds to protect about 20,000 ha of dryland used for cropping, (ii) simple bunding on 640 ha of 'bottomland' (areas which are liable to flood during at least part of the wet season during normal years) so as to provide at least partial control of water, (iii) construction of earth dams, intake structures, spillways and flood protection dykes to provide for complete control of flooding and supplementary irrigation on 240 ha of existing bottomland developments, and (iv) small scale irrigation works to provide full water control irrigation on 160 ha downstream of existing reservoirs or from ponds and shallow wells; (b) ggdij (i) to farmers for the purchase of ox drawn implements and oxen to be administered by the National Agricultural Credit Bank (CNCA); (ii) to groups of farmers, to finance small rice dehullers, small grain mills and initial stocks of cereals and general merchandise for group-owned and operated village cereal banks or stores, to be -4- administered by RDF; and (iii) to the Rural Equipment Assistance Fund (Fonds d'Assistance a l'Equipment Rural - FAER) to finance the purchase of inputs required for the manufacture of animal drawn implements; (c) village storage which would be used either to store inputs and implements (in which case they would be operated by ORD staff), or as community-owned and operated stores for traded cereals (cereals bank) or general merchandise; (d) tree plantations consisting of investments in six nurseries and about 400 ha of trees to provide wood, fruits, and fodder, and the equivalent of a further 160 ha in trees to fix bunds on soil erosion sites; (e) village water supplies consisting of the establishment of 60 new wells and the deepening of 200 existing wells; the drilling of 150 boreholes; a provision for the equipment and operating funds for well and borehole maintenance brigades; and a stock of spare parts for pumps, which would be sold for cash to villagers; (f) support to ORDs consisting of training, and the provision of initial funds for the departments responsible for training of extension staff and farmers; credit administration and input supply; and research and demonstration; and (g) project management costs consisting of a headquarters building for RDF and other costs associated with managing the investment program outlined above. 2.6 The project was also to provide funds for the financing of other small-scale rural projects, which would be identified in the course of project execution. Such small projects would be selected using substantially the same criteria of technical and economic feasibility as for the investments which had already been identified. 2.7 Thus, the project was designed to be flexible in its implementation, and the basic operational arrangements were as before. The choice of villages and communities, and the type and size of investments, were to be decided at the local level during the course of implementation. Emphasis was put at appraisal on the selection criteria for the investment which were defined as follows: "(a) need, as evidenced by a request from potential beneficiaries; (b) effective demand, as evidenced by a willingness on the part of such beneficiaries to contribute in kind or in cash to the investment; (c) technical feasibility, implying a design simple enough to be realized by local farmers under the technical supervision of staff of the project executing agencies; (d) economic efficiency, which is deemed to be satisfied by a minimum economic rate of return of 12 percent on productive projects and, for infrastructure projects, by a design judged to produce the lowest cost per beneficiary; and -5- (e) social acceptability, which requires RDF to examine proposed subprojects, taking into account the larger social context to ensure their acceptance by target populations." While on their face these were reasonable criteria the project experience suggests that they were not sufficient to ensure a satisfactory outcome (see below). Negotiations 2.8 The principal proposal of the Loan Committee was that IDA finance extension agents on a declining basis over the life of the project and the Government was asked to finance about 80 percent of this cost (increasing from 68 percent in PYI to 95 percent in PY4). This was the major issue in the negotiations since Government took the view that, given the state of their finances, the most they could cover would be an average of 35 percent. 2.9 The Dutch delegation strongly supported the Government and expressed the view that, given the public finance position, there was a real risk that, under the proposed financing plan, the necessary extension services would not be provided. It was also argued that the extension services should in this case be seen as assistance during the start up phase of the investments being supported, and not as a simple recurrent cost. These arguments were accepted and the finalized agreement called for a reduced, but increasing, local share of financing for this component. 2.10 The project was approved in March, 1982 and became effective in May, 1983. It was cofinanced with the Netherlands and the West African Development Bank (BOAD). BOAD's support was limited to financing the village water supply component. During implementation this component was handled entirely separately from the rest of the project. There was little communication between BOAD and the Bank and no joint supervision missions were arranged. No evaluation of that component has yet been carried out. Relations with Dutch aid were, however, close throughout. In fact, the latter provided an initial grant to cover a financing gap between the completion of FDR II and the effectiveness of FDR III, and several joint supervision missions were mounted. 3. IMPLEMENTATION Background to Implementation 3.1 The project was dogged by political upheaval throughout its life. The principal turning points were the revolution of 1983, which brought the Sankara Regime to power, and the events of 1987, which ended it and brought the present President to power. In addition to fundamental changes in institutions and policies, these events created conditions of great uncertainty for civil servants in particular. They also did little to help the society cope with growing economic problems. 3.2 During the whole project period there were frequent government reshuffles and changes in the structure of ministries and other government service structures. These included: - The splitting of the Ministry of Rural Development (MRD) into two ministries, the Ministry of Agriculture and Livestock (MAE), and the Ministry of Water (ME). -6- - Dissolution of one of the existing central agencies, the Rural Water Supply and Engineering Department (HER). - Transfer of part of its activities to the RDF, which was then merged with the Rural Equipment Assistance Fund (FAER) to create the Water and Rural Equipment Fund (FEER). Responsibility for FEER was given to the Ministry for Water (ME). - Redefinition of the status of FEER, which became an Administrative Public Corporation with much less administrative and financial autonomy than RDF had had as a Specific Public Corporation. - Creation of a Ministry for Peasants' Affairs (MQP) in 1986. This was renamed the Ministry of Peasants' Cooperative Action in 1987, but never really got off the ground. - Redefinition of the mandate and statutes of the Regional Development Organizations (ORDs), responsible for the execution of some of the project sub-components and for providing technical support. 3.3 At the beginning of the project MRD was the ministry responsible for the supervision of the project implementing agency, the Rural Development Fund (FDR). In the reorganization that split MRD into MAE and ME, the ORDs were placed under the MAE, while the new FEER was placed under ME. Therefore, the ORDs and the new FEER came under different ministries, with FEER in the ME controlling the funds and, as a result, the MAE and the ORDs became less committed to the project. 3.4 After the dissolution of HER, the key technical role which it was expected to play in project implementation (to assure quality control for all interventions related to water erosion control and water collection, storage and distribution) could not be effectively assumed by FEER since it did not have the necessary expertise. There was a general perception that the quality of the work declined as a result. 3.5 The ORDs were formally abolished in 1987, but for about one year the decision was not implemented and ORD staff remained employed but uncertain as to their future. In 1988, the ORDs were changed into Regional Agro-pastoral Promotion Centers (CRPAs) with purely administrative status. The function of these new agencies was much more focussed on extension activities and they lost the more general developmental role which had been the function of the ORDs. The CRPAs did not have the autonomy and commercial status of the ORDs and were handicapped in undertaking direct responsibility for activities implemented under the project. They also did not have any responsibility for coordinating rural development activities and no formal committee structure was set up to replace the role of the ORDs. 3.6 "Popular Tribunals of the Revolution" were established to investigate those suspected (or accused) of working against the interests of 'the people'. The effect of these tribunals was to paralyze civil service employees who became reluctant to do anything without clearance from the highest level. They created a climate of fear among civil servants, especially those linked to externally funded projects and to agencies such as RDF which were 'foreign creations', dependent upon external support. In addition, allowances, including travel allowances, were suspended. This hampered the work of field staff, whose regular interaction with the communities was essential for project success. There were at -7- the same time budget cuts on related items such as vehicles and office supplies, which further compounded the problem. 3.7 The results of all this upheaval were two fold. First, it created great pressure and uncertainty on civil servants in particular, which inhibited innovation. In addition, all of the institutional changes disrupted existing patterns of support and collaboration which, in the best of circumstances, would have taken time to rebuild. There was a general consensus that, for these reasons, the quality of the work declined. Secondly, RDF-FEER attempted to maintain the momentum of its activities by increasing its involvement in technical matters and in project execution, a major shift in its mandate. These problems were the origin of the two major issues which arose during project implementation, and which are discussed further below, first, the performance of some of the components (particularly the erosion control bunds) and, second, the role of FEER, (especially in the context of the evolving approach to agriculture and resource management in Burkina during the project period). Project Results 3.8 Physical Achievements. In purely physical terms the basic objectives of the project were achieved. These have been discussed at some length in the PCR and the audit has little to add to its conclusions. Anti-erosion measures were undertaken on over 58,000 ha, nearly 300 percent of the area targeted. Simple management of bottomlands was supported on 677 ha (107 percent of target). However, the more ambitious use of irrigation did not materialize and, for the improvement of existing bottomland developments and the small-scale irrigation components, only 19 percent and 14 percent respectively of the targets were met, mainly because of the disbanding of HER and its technical capability. 3.9 Similarly, for the tree planting elements of the project it was reported that between 103 percent and 423 percent of targeted numbers of seedlings had been distributed and planted. However, the results in terms of survival and growth were believed to be very low, because of adverse soil and climatic conditions. 3.10 Support services, such as training and extension, were provided at a rate in person-days, above the targeted objectives: training of extension agents (107 percent), extension (105 percent), workshops (181 percent), and the training of tractor drivers (128 percent). On-farm trials and demonstration plots were established at over 200 sites. In addition, basic literacy courses (which had not been planned) were organized for village groups. 3.11 A numerically poorer outcome was achieved in the credit program, where the amounts utilized for the acquisition of grain mills, animal drawn implements, and rice dehullers were respectively 63 percent, 18 percent and 0 percent of the targets. The demand for grain mills was high, but poor management and maintenance and low credit repayment rates forced the project to close down the operation. Their was no interest in borrowing for rice dehullers, and that for animal drawn equipment was confined to the cotton areas. The separate line of credit set aside to promote the establishment of animal drawn production units was not utilized as the agency designated to administer the funds (FAER) could not meet the preconditions laid down. 3.12 The target for the construction of village supply stores and stores for cereal banks was fully met. Credit was made available for the village groups responsible for management of these facilities - 8- but, perhaps not surprisingly, problems were encountered in financial and inventory management (theft, in simpler terms). There were also some problems of deterioration of cereal stocks and a reluctance in times of adequate supply (when stock values were below those envisaged by the village management group) to dispose of stocks at these low values and accept loss. Credit repayments were consequently low. 3.13 The rural water supply component of the project, consisting of digging new wells and boreholes, and deepening existing ones, was successfully completed. This was entirely financed by the West African Development Bank (BOAD) and there is no reporting on it in Bank files. The free supply of pumps, envisaged at appraisal, was to be dependent on the communities' commitment to finance the maintenance and operating costs. It became clear that this could not be guaranteed and pumps were not provided. 3.14 Economic Returns. Neither the appraisal nor completion reports attempted to calculate an overall economic return for the project. They did make estimates for the returns to individual 'productive' sub-components: Component Appraisal Completion (%) (%) Erosion Control 13 n.a. Simple Bottomland Management 60 40 Improved Bottomland Management 12 11 Small-scale Irrigation 12 -7 Village Tree Plantations 15 n.a. 3.15 The PCR did not recalculate a rate of return for the erosion control component because the mid-term review of the project (see below) indicated that significant numbers of farmers were not effectively maintaining the bunds, and that field trials by ICRISAT indicated that a significant increase in yields could not be demonstrated from use of bunds. It implies, but does not say, that it expects the return to be very low. This may be a bit harsh. The ICRISAT data does indicate a higher average yield of grain where bunds were used, but the difference was not statistically significant.' The stone bunds, used in the second half of the project, appear to have been more satisfactory, but yield comparison data is not available. 3.16 The returns for bottomland development were in line with appraisal estimates. The low returns for small-scale irrigation probably partly reflect the institutional hiatus and lack of technical back- up which followed (para. 3.6), and the high capital cost per hectare of this form of development. The analysis of irrigation was based on rice production, assuming a substantial increase in yields. However, 1. One problem in undertaking a statistical analysis of crop yields under these types of conditions is that crop failure is not uncommon. Thus, the standard deviation of yields is very high and, therefore, the improvement in yields has to be large in order to be statistically significant. Even if the percentage of failures using the conservation technology is low, the variation in yields is still high. A further factor that may attract the farmer is that in years when crop failure is common, crop prices are likely to be higher than in good years, so that the improvement in yields may greatly underestimate the increase in crop value. -9- this depends on effective water control and requires an effective group operation to manage it. Villagers may have little experience in this type of operation. Another requirement is use of improved technologies and therefore, technical support and availability of inputs (which may be linked to the technical support) are crucial. Irrigation is also likely to be most attractive to farmers where vegetables and other high value crops can be grown. If markets for surplus production are not available it may be difficult to maintain the level of commitment to the irrigated area which is necessary for adequate returns. 3.17 The PCR raises the issue of the sustainability of the investments undertaken under RDF III. It points out that, although evidence is incomplete, significant 'abandonment' of the investments has taken place. This has taken the form of the failure to maintain erosion control bunds and bottomlands, failure to pay the agreed share of operation and maintenance costs for wells, and closure of village stores. The audit confirms that this has been a significant problem. Given the nature of many of these undertakings, where the full implications of the investment could have not been apparent to the beneficiaries when they agreed to the action, some drop-out rate should be expected. At appraisal the base case did not make any allowance for such drop out. The analysis did consider as a variant the effect of a reduction of the life of the productive assets by one third, and the importance of ensuring beneficiary support from the outset was noted. The audit argues below that this problem may have been aggravated by the approach to project selection and the changing relationships between RDF, the collaborating agencies, and the beneficiaries. Compliance with covenants 3.18 Compliance with covenants was spotty. The most serious breach was with respect to the commitment to ensure enforcement of cost recovery policy for irrigation and water supply components. This had first been agreed with the pre-Sankara administration for the Niena-Dionkele project (Cr. 1013). The major concern was over ensuring equity with existing irrigation schemes where limited charges were levied. This issue had not been resolved by the time of the 1983 revolution and the Sankara regime saw access to water as a right and made an increase in availability a policy goal in its PPD program, under which it wanted to drastically increase the FEER program (para. 3.21). Subsequently the issue was never resolved. 3.19 Other covenants were also overtaken by events. The requirement for the transfer of titles to land to participating farmers was rendered moot by the 1984 land law (para. 3.8). Similarly the requirement that HER maintain qualified staff was also rendered moot by the agency's abolition in 1986. Basic reporting and financial covenants were observed, although fitfully. Mid-Term Review 3.20 In August 1984 the Sankara regime announced a new "Popular Development Program" (PPD) with the objective of greatly increasing the provision of social and economic infrastructure in the rural areas, particularly for water supply. The Government approached the Bank about the possibility of Bank and cofinancier participation in this program through an expanded rate of implementation under RDF III. Bank staff had already become concerned about a lagging rate of implementation under the project and there was already a shortfall in the Government's contribution to this program, even though the latter was only five percent. As a result, there was skepticism about the feasibility of the proposed expansion. A senior, four person joint Bank/Netherlands Aid mission was sent in November to determine - 10 - whether (and if so, how) they should participate in the PPD, and to assess the real constraints to more rapid development of RDF-type activities.' 3.21 The mission concluded that "although overall physical progress is satisfactory, and the project is unusually well managed, the impact and cost effectiveness of several components are open to severe questions. These arise in part from the effect of poor rainfall in the last few years and in part from doubts about appraisal assumptions." Government had requested that the project be expanded to include an additional 35 earth dams, 100 cereal banks, and 50 boreholes. Based on progress to date and the apparent effectiveness of some of these efforts, the mission concluded that an expansion of this scale was not feasible or desirable. It was finally agreed that the project would be extended to cover 16 further cereal banks, and 50 boreholes (but these should be equipped with hand rather than motorized pumps). The mission, however, concluded that the proposed dams were not cost effective as a source of water for general purposes and that utilization of the existing sites where water was available suggested that the real demand was not as great as might be anticipated. It was proposed that FEER and the ORDs concentrate on assisting villagers to make better use of existing facilities. 3.22 The mission was particularly concerned about the Erosion Control program, which had been based primarily on mechanically constructed bunding. They noted that very little was left of the extensive (120,000 ha) program undertaken between 1962 and 1965, maintenance of bunds implemented under the RDF projects was often poor, and that more than 90 percent of bunds constructed under more recent programs (including the RDF) appeared to be constructed on sites with slopes of less than 1 percent. Evaluations by ICRISAT of a number of project sites indicated that no significant yield increases had been observed under these, limited slope, conditions (see para. 3.17). 3.23 The mission also noted that, although it was not part of RDF's policy, farmers had been receiving food aid after construction of the bunds. This increased motivation and participation during the construction phase, but was considered detrimental for maintenance, as no aid was provided for that purpose. They also noted the simpler systems being introduced by other aid programs, such as bunds constructed with stones, branches and crop residues. These seem to have achieved results as least as good as the more expensive earth bunds. As a result of this mission FEER made major changes to its anti-erosion program and began to assist farmers in construction of the stone bunds in particular. These have often been supported by the planting of appropriate grasses (eg Vetiver) or trees and shrubs to make more permanent, contoured, plot boundaries. 3.24 The outcome of the mission generated considerable tension between the Bank and Government, but in retrospect the active intervention by the Bank may have protected FEER from Governmental pressure to some degree. As noted (para. 3.2), at that time failure to respond to such requests was considered as sabotage. FEER was also relieved to agree to undertake an extensive program of evaluation of the major components in its 'portfolio'. The results of this review came forward over the period to the end of the project and did substantiate questions raised by the 1984 mission as to the benefits of some of the components of the project. The effort also substantially increased FEER's capacity to carry out studies of this type. 2. BOAD was unable to participate in the mission. - 11 - The Termination of Bank Involvement 3.25 The combination of the drought periods of the 1970s and early 1980s, and the increased pressure of population in the north-center of the country (the Mossi Plateau), led to increased movement of population into the less populated and generally higher rainfall areas of the south and west of the country. This movement was further stimulated by the reduction of the scourge of river blindness in these regions. It also led to increasing tensions in the areas of settlement between the indigenous groups and the Mossi immigrants.2 The Agrarian Reform Law of 1984 (RAF) was in part a response to these pressures. Paradoxically, as noted earlier (para. 2.3), the law itself may have acted to encourage further movement and in some respects added to, rather than solved, the land problem. Following the promulgation of the law, Government became increasingly interested in seeking ways to build upon it to provide some security for local groups, which would validate the locally recognized rights. 3.26 From this concern the Village Land Management (Gestion de Terroir) approach to the management of land began to evolve. From 1985 this issue began to dominate the interaction between the Bank and the Government and the supervision of the RDF project became increasingly linked to the discussions of this issue and the preparation of a project to attempt to tackle it. RDF III was scheduled to close at the end of 1986, but because of initial implementation delays, and the devaluation of the CFA franc relative to the SDR in the early 1980s, disbursement lagged. The project was twice extended, both to achieve its development targets and to continue to support the FEER program until a follow-on project could be appraised. 3.27 It was initially assumed by all sides that FEER could be the vehicle for channeling the finance for a National Land Management (NLM) project. However, as discussions proceeded differences of opinion began to emerge. The Bank considered that effective involvement in the program as then perceived would require a heavy emphasis on community participation. However, it was felt that FEER's modus operandi continued to be heavily top-down in orientation. At the same time the view was expressed that FEER was too heavily involved in sub-project implementation and that it should sharpen its focus as a financing agency, tighten its financial controls, and arrange with other agencies (eg. the CRPAs) to take responsibility for implementation. 3.28 At the same time the Netherlands Government, the major co-financier, was also becoming concerned with the performance of FEER, although more from the perspective of the overall impact and effectiveness of the effort. An external review of the project was, therefore, carried out in 1988. This review was critical of FEER in three main areas; the strength of the village groups with whom they worked, the quality of the technical backup given to them, and the effectiveness of FEER's financial controls. 3.29 The Review Mission was concerned about the feasibility of the FEER approach at the village level. They felt that the approach overestimated the cohesiveness of the village groups and their ability to plan and execute the projects. Doing so involved undertaking work with which they may have had little experience and often without adequate technical back-up from the relevant ministries. Also several of the projects (eg. grain mills, cereal banks, village shops and pharmacies) required new management skills with which they were not familiar such as book-keeping, administering loans, and 3. This was further complicated in some areas by increased use of the area by transhumant pastoralists. - 12 - managing inventory. Most of the villagers were not able to read and write and it was, therefore, not surprising that many of the village groups ended defaulting on their loans from FEER. 3.30 As a result, the mission questioned the real impact of the program on the living conditions of the rural population, and the return on the investments made. The results in both respects were disappointing, partly because of poor design, partly because of inadequate monitoring, supervision and feedback. (This reflected the involvement of FEER in execution, also of concern to the Bank.) The mission was also concerned that FEER, pressed by the targets set by donors under the projects, and by Governments desire to speed up rural development efforts, in turn pressed quantitative targets on the ORDs who in turn pressed the villages to collaborate with the authorities' wishes. The result, not infrequently was that 'the way the villagers were approached by the technical ministries, responsible for the execution of the sub-projects, did not contribute to their confidence in government services'. The effect was to reduce the real effort made by villagers to sustain activities which were not obviously immediately successful. 3.31 The mission also believed that the various institutional changes, including the change in legal status of FEER and the ORDs (the latter superseded by CRPAs with much reduced powers), had weakened the relationships between FEER and its collaborating agencies. It strongly recommended that, under any future project, FEER develop a contractual system with collaborating agencies to clearly define the responsibilities of the various parties and the financing arrangements. This could involve the use of multi-year protocols which would specify the quantity and quality of the tasks to be carried out and their implementation schedule, establishing a clear basis for monitoring progress and evaluating results. 3.32 In order to respond to these criticisms, FEER undertook its own review of operations in 1989 and during the following year took steps to institute many of these recommendations. However, the Netherlands felt that the proposals for revision of procedures, and the development of contractual relationships with collaborating agencies, would not address their basic concerns and decided not to support FEER I, as the follow on project was then called. They indicated to the audit that, while there is a role for an institution such as FEER to support small scale productive and social infrastructure activities in the rural areas of Burkina Faso, they still had reservations about FEER's ability to fill this role. At bottom this is a question of judgement but, if FEER is not to carry out this role, who is? The audit is left with the impression that the basic unease was with the near impossibility of the task confronting FEER in the environment in which it has to operate, and that this is the basic cause of the unease and the reluctance of external agencies such as the Netherlands to continue support. 3.33 The Bank also decided not to support the FEER I proposal. Bank staff believe that this decision was clearly conveyed to FEER in 1989. Certainly supervision reports and other documentation indicate that, at the time, staff believed they had done so. However, FEER staff maintain they were caught by surprise when the Bank did not agree to any further extension of RDF III in early 1990 and stated again its decision to cease funding. Whatever the truth of the matter, the perceived abruptness of this termination left a nasty taste with FEER, which still continues. - 13 - 4. ISSUES AND LESSONS 4.1 The audit broadly concurs with the judgement of the PCR on the physical and related economic performance of the project. It addresses two related issues not addressed by the PCR; were the criticisms of FEER which led to the termination of Bank support justified and; what should happen now? Was the Bank Criticism Justified? 4.2 In 1989 and 1990 FEER and the Government of Burkina Faso instituted a number of changes to FEER's modus operandi. These changes did respond to some of the issues raised by the Bank, and other donors, at various times during the long period of Bank support, particularly in the latter part of the life of RDF III. These criticisms were that there was a need, first, to avoid becoming directly involved in operations (i.e. to focus on financing projects implemented by others), second, to clarify the responsibilities of the various collaborating parties to each sub-project and, third, to increase the degree of local participation in the process of the selection of projects to be financed. 4.3 However, while changes have been made by FEER, it is the judgement of the audit that their impact may be limited because of more fundamental problems, which appear to have been most specifically raised by the external review and the Netherlands' comments. These problems relate, first, to the nature of an intermediary institution such as FEER in a context such as Burkina Faso, second, to the range of activities which it finances and their very different operational demands and, third, to the institutional vacuum for local infrastructure in which it currently operates, particularly since the demise of the ORDs and their implementation coordination capacity. The Role of the Intermediary Institution 4.4 FEER is primarily funded from external sources, the main vehicle currently being the FEER I Project.' While it is now attempting to operate through contractual agreements with its collaborating agencies, many of these agencies are weak and lack resources and have different priorities than FEER. FEER itself will be judged by the extent to which it has used the resources made available to it to achieve measurable results on the ground. If its contractual partners fail to utilize the funds made available, then FEER inevitably will have greater difficulty next time arguing that it needs additional funds to meet the substantial demand. Thus, it is almost inevitably going to be drawn in to fill the gaps in the performance of the collaborating agencies, as it was during the earlier Bank funded projects. The Different Operational Demands 4.5 The range of activities supported by FEER can be divided into three broad categories: (a) local infrastructure, such as schools, health and maternity clinics, and water supplies; (b) agricultural investments, such as erosion control measures, the development of bottomlands, and small scale irrigation; 4. This is primarily funded by Swiss Aid and DANIDA. - 14 - (c) small commercial type enterprises, such as village stores, pharmacies, grain mills and grain storage. 4.6 These different types of activities inevitably place quite different demands for back up support on a financing agency such as FEER. In a more developed economy these functions are likely to be performed by quite different types of organizations. While the ORDs existed, they were involved in most of these activities and could provide field support to implementation. However, trying to cope with such a range of activities was a major factor in the problems which led to their dissolution. 4.7 Local Infrastructure. This is the simplest from the point of view of FEER and its contractors. The structures have just to be constructed and are then turned over to another agency for them to operate.' FEER operates purely as a channel of funds with the responsibility of insuring that the structures are properly constructed. In principle, the operation of a health clinic is well known to the health workers who will be engaged to operate it. 4.8 Agricultural Investments. The types of investments being supported do not involve just the application of well known technologies. In most cases the local community will have limited experience with their management. Furthermore, some of them, such as the development of bottomlands, are not straightforward. The amounts of water which may be available at a given small site, and the best layout for utilizing them, will not be known until development is started. The operation of these sites introduces a whole new element into the users' farming system and, therefore, throws up a series of management as well as technical problems. This, therefore, will require experienced support for an extended period of time. This was recognized at the outset of RDF III when the Bank agreed that project funds could be used to support the salaries of extension agents for up to eight years, on the grounds that this was part of the investment. With the present two year contractual agreement for each sub-project it is not clear that support will be available in the amounts necessary, because of other demands on the extension workers time, unless the CRPAs develop a special corps of staff. 4.9 Commercial Investments. These also involve the local community in an activity of an entrepreneurial nature and require skills or technologies which are unlikely to be well known to villagers eg. bookkeeping. These activities are also financed at least partly through loans. In a situation such as Burkina, FEER cannot just bundle a set of these repayment obligations and sell them in a secondary market in order to recover its funds. The obligation to recover cannot really be effectively contracted to any other agency in this situation. In the end, FEER will have to do it itself. Institutional Structures 4.10 The primary points of contact between FEER and the villages are now the CRPAs. The Bank and others have long criticized FEER (and RDF before it) of operating in too top-down a manner. The current arrangements put FEER in much more of an arms length relationship with the villages being assisted. However, in the present institutional structure in Burkina Faso, this may be creating some problems, which are also arising in the Environmental Management Project (Cr. 2229-BUR). 5. This, of course, ignores the issue of the capacity of the community or others to operate the facilities once they are constructed. - 15 - 4.11 Stress was placed in the donor criticism of FEER and in the PCR on the desirability of greater beneficiary input in sub-project selection and design. Given the current conditions in most villages in Burkina Faso when discussions turn to steps to improve 'welfare' at the village level, basic social infrastructure ranks high for the villagers. For example, a simple health facility where the sick and women in child birth can be attended to by other women as much as by a medical attendant, is a significant asset. Such a center does at least give some degree of privacy and some access to basic health services. Similarly, additional classrooms do at least reduce the level of overcrowding in the village primary school. 4.12 However, every agency has its own agenda. The CRPAs' priority is to promote agricultural production, and their best contacts are likely to be with those villages which either have greater agricultural potential or have individuals (especially in leadership positions) who at least are willing to try out new agricultural ideas. These are, therefore, most likely to be linked up to the FEER program. In all probability, the villages selected by the CRPA will press to obtain some improvements in social infrastructure from the FEER program. However, it is also quite likely that, for example, the priorities for improved health facilities from the point of view of improvement of general access to care, may be quite different than those thrown up by the response to agricultural development priorities. Several studies by the Bank and others have shown that extension agencies tend to concentrate their activities in more accessible areas. For a start, the economic incentives to the farmers to intensify and invest are greater close to a highway than five miles away from it. On the other hand, from the point of view of improvement in basic welfare, the priority for maternity centers may be greater in the village five miles from the highway. 4.13 Over much of Burkina Faso climatic conditions are highly variable and market demand for agricultural products is weak. Thus, the incentive for farmers to invest in resource enhancing measures and intensify production is limited. The result is for some pressure to be exerted on villages to undertake some agricultural measures, which would not otherwise interest them, in order to obtain social infrastructure. The probability that these agricultural activities will subsequently be abandoned is high. In some of the villages visited by the audit mission there was strong evidence that this had happened. Thus, it is not clear that the change in approach will necessarily result in a real 'bottom up' determination of investment priorities for FEER funds. To some degree the priorities of the intermediary agency will be a major factor in determining the content and location of the activities funded by an agency such as FEER. This problem undoubtedly existed before with the ORDs. The latter did, however, have a wider rural development mandate and were in a better position to collaborate with other ministries etc. at the local level. 4.14 At present there is no structure in Burkina to resolve this dilemma. There is no formal requirement for review and approval at the local level, for example, by the Prefect or High Commissioner, although they can informally review programs and attempt to coordinate. Agencies such as FEER meet on a non-formal basis to keep abreast of one another's plans, but such meetings are periodic and carry no weight. Even if FEER was designated by government as the agency responsible for coordinating such activities, it is doubtful that it could do so effectively. Other agencies would, in all probability, continue to follow their own ministerial priorities and, in the case of major conflicts, raise issues to a higher level for resolution. 4.15 The 1988 external review and Netherlands Aid were critical of the share of project expenditure which went for FEER's own operating expenses. They stressed the need for a smaller, leaner - 16 - agency to serve as the source of finance for local level projects. However, for the reasons noted above, this appears to be unrealistic. There is just not the market or institutional infrastructure in place for such a lean agency to function. If there were a well developed credit market, as in the industrial countries, then FEER could act as a channel for external finance, refinancing loans made by intermediaries.' If there were a well developed institutional structure for selecting and coordinating projects at the local level, then FEER could, for example, finance a share of all, or particular types of projects, approved in an annual national program. Also, under these conditions, there would be an adequate capacity at the local level to develop and implement projects and an agency such as FEER would not have to get involved in project implementation. However, care must be taken to avoid a vicious circle where the lack of supporting structures justifies bloated institutions. 4.16 In practice, neither of these preconditions apply. Thus, the efforts made by the Bank and its cofinanciers in the latter stages of RDF III, and which played a significant role in the Bank's deciding not to back FEER I, were not able to effectively resolve the basic causes of the problems identified in the type of program which had been supported under the RDF and FEER projects, with the objective of fostering local level investment. What Should Happen Now? 4.17 There are two main lessons from this project: (a) While changes to improve the effectiveness of FEER operations were required, changes were also necessary in the local institutional structure, whose capacity to select and coordinate the implementation of small local projects had been gravely weakened by the turmoil and change during the 1980s. (b) In regions such as that covered by this project, with low agricultural productivity and/or limited market potential for increased production, making agricultural production activities the key to support of local projects is likely to lead to distortion in the pattern of investment and to poor performance of agricultural activities. 4.18 It is the conclusion of the audit that, while the changes made by FEER may have improved its operation, they have not resolved the basic dilemma facing it. Moreover, it is not in a position to tackle this problem. What appears to be necessary is a broader structure, perhaps managed from the offices of the Prefect and High Commissioner, which would identify more general development and investment priorities, as identified by the villages and which could act as the basic agent for implementation of the types of small projects funded by FEER and RDF before it, and have some power to coordinate. A structure of this type is being used on a limited scale in the ongoing Environmental Management Project (Cr. 2229-BUR). These projects could then be implemented, on the basis of available resources and personnel, by the agency most suited to do so. 4.19 This might argue either for a narrower role for FEER, for example it might concentrate either on the construction of rural social infrastructure, or on agricultural investment, or at least make a clearer distinction in its operations between the two. Depending on this choice it might most appropriately report to a ministry other than the Ministry of Water. Increased emphasis on support of 6. However, the problems of small borrowers in the inner cities with their preponderance of low-income minority borrowers are well documented and show that similar problems exist in developed economies. - 17 - agricultural production investments would require for success an extension of activities into areas of higher production and market potential than FEER's traditional areas of operation. 4.20 Although the Bank is no longer financing FEER this issue continues to be a live one for the Bank in the context of the ongoing Environmental Management Project. This project is attempting to help rural communities to design and implement sustainable management plans for their lands and natural resources. However, in many cases effective land management is hampered by the lack of access to markets for potential products which would justify investment in resource improvement. Unfortunately, the project only has funds to finance natural resource management measures, such as erosion control efforts. Attempts are being made to access other funds to support linked local infrastructure development but, in the absence of any general local planning and implementation coordination structure, everything has to be done on an adLho basis, greatly increasing the project's management task. Postscript 4.21 While the withdrawal of the Bank from support of FEER may have been perceived as abrupt, the audit believes that the decision to terminate Bank support for FEER appears on balance to have been beneficial for the agency. The imminent threat of termination did result in it taking a hard look at its operational policies. The resulting changes, while subject to the caveats noted above, have resulted in a sharper focus to operations. The Bank had supported FEER and its predecessor FDR for almost 20 years and it is reported that FEER was seen, at least in some quarters, as a creature of the Bank. The need to develop contacts and support from a wider range of external financial institutions will also have been beneficial in opening it to different ideas and opportunities.