The Burundi Coffee Sector project, supported by Credit 2123-BU for SDR 21.3 million (US$28 million equivalent) was approved in FY90. The credit was closed on schedule on July 31, 1996. A total of US$13.3 million was canceled, because of a reduction in scope of the works undertaken by the project. Cofinancing was provided by three French agencies, CFD for US$6.1 million, FAC for US$1.5 million, and CIRAD for US$0.5 million, and by Belgian Aid for US$1.0 million. The Implementation Completion Report (ICR) was prepared by the FAO/World Bank Cooperative Programme (FAO/CP) for the Africa Regional Office. Because of the security situation in the country it was not possible for FAO/CP to mount a mission to Burundi and the ICR is, therefore, based on a review of documentation available in the Bank. The Borrower undertook its own evaluation and the executive summary which is appended to the ICR. The draft ICR was not sent to the cofinanciers for comment.
The objective of the project was to increase the quality and hence the export and farmgate prices of Burundian coffee. The project had two components: (i) a package of policy and institutional reforms to improve the efficiency of the industry including reorganizing existing institutions and opening up marketing, processing and exporting to private capital; and (ii) investments in 100 coffee washing stations (to increase the share of washed coffee, which receives a higher price), and a program of coffee research and training.
Prior to the project, marketing and export of Burundian coffee had been undertaken by state enterprises. The institutional reform program aimed to (i) achieve a more rational and efficient organization of coffee marketing by separating it from government and dismantling the monopolies; (ii) introduce competition in export marketing by establishing an auction system; and (iii) replace fixed prices with an incentive based payment system and the levy on the marketing margin with a direct export tax. There was strong opposition to several of these changes within the Government, slowing implementation or causing modifications. As planned, existing institutions were reorganized into four different types of entity, most of which were to be independent and open to private capital. So far, however, there has been little private sector interest. These entities inherited many of their staff from the former parastatals and operational practices have changed little. Government insisted on linking the incentive based payment system to a floor price. This proved cumbersome, reduced the incentive for timely and higher quality production, and may have led to some smuggling. The introduction of an auction system for export sales has had a more dramatic impact and 90 percent of exports are now handled by private traders and prices to farmers have risen as a share of world price.
In the first two years of the project 39 washing stations were constructed. However, since coffee prices had fallen sharply during that time, production increases were lower than projected. But, washing station unit costs were much higher than planned, and the stations were able to handle up to 50 percent more beans than expected. Thus, far fewer stations than planned were needed. Therefore, at the 1993 Mid-Term Review, it was agreed to suspend the program, and to investigate smaller prototype stations, which might be more attractive to private investors, more widely dispersed and therefore more accessible to growers. However, due to a rapid deterioration in the security situation in October 1993, this program was not implemented, and no further stations were constructed. The research program had mixed results. Farm level work was hampered by the security situation, but useful results were obtained on ways to overcome an “off-flavor” problem that caused Burundian coffee to be downgraded. Training activities also had mixed results as there were problems in trainee selection. Technical assistance, especially for the development of the auction system and management of washing stations, had good results. Because of a lack of data, the ERR is not reestimated in the ICR.
The Borrower’s evaluation is in broad agreement with the ICR. It does note, however, that the research program was handicapped by the slow arrival of several bilaterally-funded experts and by their precipitate departure. It also notes that a factor in overcoming resistance to the liberalization measures was that neighboring countries were also making similar changes, and this was having an adverse impact upon marketing and exporting entities in Burundi.
The Operations Evaluation Department (OED) concurs with the ICR in rating project outcome as marginally satisfactory, institutional development as modest, and Bank performance as satisfactory. OED rates sustainability as uncertain, rather than unlikely as in the ICR, since the problems currently faced are related more to the current economic and social situation in the country than the viability of the project actions and investments. The ICR presents no evidence that the institutional changes and policies are likely to be reversed.
In the circumstances, the ICR provides a clear and concise account of project implementation and is rated satisfactory.
The principal lessons from the project are first, that attempting to introduce flexibility into an administered commodity pricing system will increase the complexity of the system but do little to provide more effective incentives to producers. Second, that it may be useful in projects of this type to have some second stage issues in mind for the mid-term review, which could not be included in the original project. For example, in this case, it might have been desirable, based on the experience of the first three years, and given that reasonable momentum had been built up, to have pressed at the MTR for greater and faster movement toward a fully liberalized pricing system.
No audit is planned.