Mali - Industrial Sector Development Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 8401 PROJECT COMPLETION REPORT MALI INDUSTRIAL DEVELOPMENT PROJECT (CREDIT 98611) FEBRUARY 23,.1990 Africa Regional Office Sahelian Department Industry and Energy Operations Division ThIb document has a restrict dibid and may be used by nIde only In the pedornune of "hir officld dute ltb contens my not otherwise be dicosed wwoS Wodd Bank auoration.  FOR OFFICIAL ONLY TMH WOILD BANK Washington. D.C. 20433 U.S.A. OtS* of oIttVCdwil February 23, 1990 MEMORANDUM TO TEE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECTs Project Completion Report on Mali - Industrial Development Pro3ect (Credit 986-MLI) Attached, for information, is a copy of a report entitled 'Project completion Report on Mali - Industrial Development Project (Credit 986-NLI)) prepared by the Africa Regional Office. No audit of this project has been made by the Operations Evaluation Department at this time. Yves Xwvani by Ram K. Chopra Attachment This document has a restricted distribution and may be used by recipients only in the performance of their offcial duties. Its contents may not otherwise be disclosed without World Bank authorition.  FOR OMCILL USE ONLY MALI PROJl`T COMPLETION REPORT INDUSTRIAL DEVELOPMENT PROJECT CREDIT 986-MLI TABLE OF CONTENTS Page No. Preface ...........................................................i Evaluation Summary ...................................................... i PART Is PROJECT REVIEW FROM BANK PERSPECTIVE ......... ........ 1 A. ProJect Identity ..................................... ...... 1 B. Background ......... . . .......... 1 C. Project Objectives and Description .. ................. 2 D. Project Design and Organization ...................... 3 E. Project Implementation ...... ........................... 4 F. Project Results ... ......................... 5 G. The Bank's Performance ............. ............. 8 B. The Borrower-Is Performance ...* ....... ........... 9 I. Consultant Services . .............................. ... 9 J. Conclusions and Lessons Learned ..... ................. 9 PART IIs PROJECT REVIEW FROM BORROWER'S PERSPECTVE ........ 11 Borrower's Perspective of the design and Implementation and Its Development Impact on the Project PART III: STATISTICAL INFORMATION .......... *................ 46 Related Bank Credits . ................................... .. 47 Schedule of Actual vs Estimated Cumulative Disbursements .. 48 Subprojects Financed Under the Credit ..................... 49 Mission Data ....................................................... 50 Staff Input ....................................................... SO This document has a resrcted distrbution and may be used by reipients only in the performance of their officil duties Its contents may not othewise be disclosed wWtout World Bank auhoization.  MALI PROJECT COMPLETION REPORT INDUSTRIAL DEVELOPMENT PROJECT CREDIT 986-MLI PREFACE This is the Project Completion Report (PCR) for the Industrial Development Project in Maii for which Credit 986-MLI in the amount of US$8 million was approved on February 26, 1980. The credit was closed on June 30, 1988, three years behind schedule, and the last disbursement was in December 1988. The PCR was prepared jointly by the Industry and Energy Operations Division of the Sahelian Department of the Africa Regional Office (Preface, Evaluation Summary, Parts I and III) and by the Borrower (Part II). Preparation of this PCR was started during the Bank's final supervision mission of the project in January 1988, and is based, inter alia, on the Staff Appraisal Report, supervision reports and other documents relevant to the project.  - ii - MALI ?ROJECT COMPLETION REPORT INDUSTRIAL DEVELOPMENT PROJECT Credit 986-d6 EVALUATION SUMMARY The project as originally conceived had two main objectives: (a) to encourage private productive investment by improving the incentive 3ystem and by providing technical and financial assistance to private entrepreneurs and artisansl and (b) to demonstrate that financial and economic viability can be achieved for certain state enterprises when efficient management is allowed to operate under commercial conditions. (Para. 7) With the benefit of hindsight, it seems that the degree of prepa- redness and ability of the project institutions and of the Government to adequately implement the project had been overestimated during project pre- paration, particularly with regards to the line of credit. (Para. 10). The lack of understanding of, and /or commitment to the project objectives on the part of the project institutions and the opposition of the Ministry of Plan to give assistance to the Banking system contributed to impede the project implementation. Finally, the Government commitment to the project was also insufficient to ensure appropriate project coordination. (Paras. 11-13) on account of these shortcomings the project got off to a slow start. Declared effective in September 1980, only the rehabilitation of ULBI's equipment started immediately. Technical assistance to SERVULART became ope- rational only in June 1982, the first subprojects were approved in September 1983, training of amall entrepreneurs started in Ncvember of 1985, and ULB initiated a study of the expansion of its production capacity at the end of that same year. The credit closing date was postponed three times and imple- mentation ol the project lasted almost 8 years instead of the 4.5 years ori- ginally planned. No changes were made to the project objectives, but another component was added in 1987 so that funds could be used to fight a locust invasion. Project proceeds were reallocated four times. (Para. 14) Two unforeseen events contributed to hamper the implementation of the two main components of the line of credit and the rehabilitation of ULB rehabilitation. First, no measure had been included in the project to either remedy BDM's financial situation which had been found critical during project preparation, or prevent it from using the line of credit, as a result, BDM became the only bank to finance subprojects with the proceeds of the line of credit and its worsening financial situation and weak organization affected the line of credit quality. A third of these subprojects do not perform satisfactorily. Secondly, ULB began to face marketing problems as soon as its production capacity increased following the renovation of its equipment, and its expansion envisaged In the project became questionable. Moreover, con- trarily to the assumption underlying this component, the Government did not ensure that ULB would operate under strict commercial conditions, it started to post huge losses and this contributed further to make its expansion unwar- ranted. (Para. 16-17)  - iii - This project was conceived more than 10 years ago. The lessons to be drawn from its Implementation have since been learned from similar bank projects in Africa. The scope of this particular project, designed as an IDF-type project, was broadened to include training of entrepreneurs, assis- tance to artisans and support to the rehablltation of a public enterprise. Therefore, it was destined to have the merits and weaknesses attached to its own design as well as those attached to a classical IDF project. (Para. 35) For a first project its scope was far too broad to be carried out efficiently by institutions and/or the Governument whose degree of prepared- ness, strength and commitment to the project objectives was overestimated. Furthermore, it is clear that the conceptual basis of the IPGP and the ULS components was wrong. It is now well established that the sesistance to the emergence of entrepreneurs involves much more than providing training in management skills. With regard to the ULB component it is also clear now that the rehabilitation of individual public enterprises cannot succeed if the rationalization of the public enterprise sector as a whole is not undertaken. These two components should have been excluded from the project as soon as it appeared that they had little chance of succeeding. With regards to the SERVULART experiment, it showed that it is possible to improve the effi- ciency of artisans ani to create new jobs in the artisan sector as long as they receive assistance to improve their production techniques and product design without being brought into competition with the modern sector. It was the most successful component of the project. (Paras. 36-38) On the whole, the performance of the project was average as were generally the IDF-type projects in the region.  tw1 INDUSTRIAL DEVELOPMENT PROJECT PROJECT COMNPLETION REPORT PART ONE A. PROJECT IDENTITY Project name a Industrial Development Project. Credit number X 986-MLI RV Unit i Africa Regional Office Sector t Industrial and Aitisan Sectors B. 8ACKGROUND 1. Mali is a vast, landlocked, and resource-poor country located in the Sahelian zone of Western Africa. Only about a quarter of its 1.2 million square kilometers is arable. Nea'rly 902 of the population estimated at 7.9 mllion in mid-1987, is dependent on the agricultural sector which accounts for roughly 50S of GDP. Per capita income was estimated in 1986 at US$ 170 and social indicators, such as life expectancy and school enrollment rates, were among the lowest in the world. 2. The Industrial Sector. In the early 1980s, the Malian industrial sector was characterized by the dichotomy, typical to West Africa, of a small modern sector made up of some 28 large public enterprises and numerous medium- to small-scale private enterprises along side an active traditional sector of more than 200,000 artisans. The large state-owned enterprises, which accounted for two thirds of modern sector sales revenues, employed some 12,500 persons or almost 901 of the labor force of the sector. Their performance had, however, started to deterlorate and continued to decline, and by 1985 most of them were virtually bankrupt. A significant number of them ceased operations and, since 1988, the Government has, with IDA's assistance, undertaken a project (the Public Sector Adjustment Program, Cr. 1937-MLI) to rationalize and liberalize the sector, reduce state participation trough divestiture from the industrial and commercial enterprises and liquidate those that are not viable. 3. In 1979, project preparation had identified about 20 privately owned enterprises operating in subsectors ranging from textiles to foundries or bakeries. Most of them were medium- or small-scale enterprises and, notwithstanding the dynamism and entrepreneurship of their owners, their performance w,s clouded by the lack of appropriate financing and assistance to improve their technical and management skills. 4. The Artisan Sector. The activity of the artisan sector was based on the communal nature of the Malian society and was providing considerable benefits to the population. In this sector, which had shown vitality and resilience, the utilitarian crafts (metal and woodworking, textile weaving and spinning) were growing rapidly. A highly successful model, implemented by the Malian cotton company, had set sn example in helping blacksmiths adapt  2 modern technical methods to the production of small agricultural implements. Overall, however, its potential development was hampered by unavailability of small credit, weak work organization and inefficient * and unadapted production techniques. 5. Sector Development Objectives. At appraisal, several areas of the industrial sector h.,d been identified as having potential for development among ihich the processing of agricultural and livestock products, the production of agricultural and mechanical tools and spare parts, metal products and light chemicals. It seemed that the development of productive capacities in these areas could increase linkages between the different sectors of the Malian economy. In the artisan sector, textiles and decorative crafts vere thought to have good prospects for expanding productivity and employment. However, the emphasis put on the demand for financing and institutional support of the public sector had deprived the non public sector investors of the ingredients required for the development of the private sector. Therefore, the purpose of the Industrial Development Project was to help Mali take full advantage of its private sector by Improving the technical and financial assistance needed to develop it. 6. Policy Context. The inadequacy of the investment code was identified during project preparation as a policy issue. With the help of a PPF-financed consultant, the Government had substantially modified the code's provisions to make them more conducive to the establishment of productive projects. As In most industrial sector development projects of that period, no other major policy issue was addresseds the analysis of the domineneo of the public esector was l4i tad to the d.privatozo of revource,w that the private sector suffered and the banks' inability to finance private investments was only attributed to their lack of funds to make term financing. In retrospect, it appears that the problem of the public enterprises (including BDM, the government development bank) were so serious that they should have started to be addressed at the time of project preparation and appraisal (1979-1980) rather than in 1985 when the whole sector had collapsed. C. PROJECT OBJECTIVES AND DESCRIPTION 7. Pro1ect Obiectives. The primary objectives of the project were to s (a) encourage private productive investment by making a start towards improving the investment environment and incentive system, and by providing technical and financial assistance to private entrepreneurs and artisans; and (b) demonstrate that finmancial and economic viability may be achieved for certain state enterprises, when efficient management is allowed to operate under commercial conditions.  3 8. Protect ComnoneAts. To achieve these objectiv4s" the US$8 million It project consisted of the six following components (a) an *apex* line of credit (US$3.5 million) and technical assistance to the banklug system to develop comon acceptable standards for subproject appraisal and supervision (US$100,000); (b) financial and technical assistance to strengthen C$PI's role as a promotion and technical acsistance agency (US$350,000); and, (c) technical assistance to help IPGP carry out a training program for small entrepreneurs (US$150,000); (d) the setting up of SERVULART as an extension service agency providing training and assistance to artisans (US$850,000) and the establishment, with UNDP's assistance, of a fund to promote and develop Malian craft exports (US$50,000); and, (e) technical end financial assistance to ULB, the state milk- processing factory, for Its rehabilitation and for the expansion of its production capacity; The project also earmsrked US$65,000 to finance the operating cost of a project coordination unit in the Ministry of Plan. 9. Proiect Institutions. To achieve the project's objectives, six fovernment institutions were going to be involved in, ana benerit from the project. These were s (a) le Centre d' etude et de promotion industrielle (CEPI)s (b) le Service de vulgarisation artisanale (SERVULART), to be set up in the course of the project as an appendage of CEPI; (c) l'Institut de productivit6 et de gestion pr6visionnelle (IPGP), a vocational training Institution; (d) la Banque centrale du Mali (BCM); (e) l'Union laitiere de Bamako (ULD), a government milk factory; and (f) la cellule de coordination, the coordinating unit. D. PROJECT DESIGN AND ORGANIZATION 10. The scope of this first project in the sector was very wide. It was conceived as an IDP-type project of the late 706 that included institution-building components and technical and financial assistance to private entrepreneurs. Two self-contained components (the assistance to artisans and the rehabilitation of ULB) were, however, added to the project which became, as the SAR admitted, comulex. The 'apes"-type line of credit (also the first in Western Africa) was another innovative feature Intended to Increase the number of banks involved in lending in the sector contributing, thus, to further complicate the project design. In retrospect it seems that the degree of orenaredness and ability of the lDroiect institutions and of the Government to adequately $mplement the project had been overestimated during project preparation. The project institutions were weak and their role, In particular in the implementation of the line of j/ US$360,000 refinanced a PPF and US$200,000 were allocated to contingencies.  credit, were not clearly perceived. Furthermore, the lack of understanding of, and/or commitment to the project objectives, on their part as well as on the part of the Government, contributed to impede the project implementation. 11. This affected especially the implementation pace of the line of credit which required extended technical assistance to upgrade the subproject appraisal capability of CEPI, which was expected to play a pivotal role, to an acceptable level. Extensive supervision work was also needed in the first years to finally clarify the role of each institution concerned (CEPI, BCH, the banks and the Ministry of Plan) in the processing of subprojects to be financed by the project. The supervision missions were not, however, successful in convincing the Ministry of Plan that the assistance to the benking system, envisaged in the project, was critical to induce the participation of all coimercial banks in the line of credit and this assistance was consequently not provided. 12. Similarly the training of small entrepreneurs was impeded by IPGP's 'nexDerience. This weak institution was already receiving assistance from LO to help train micro-entrepreneurs. Under ;he project, it was envisaged that another ILO assistance, financed by IDA, wuuld provide, in coordination with CEPI and the banks, training for the :repreneurs benefiting from the line of credit. Instead, IPGP's management, unable to carry out both tasks separately, integrated the two training sessions into a single program and kept organizing short term seminars for micro-entrepreneurs that did not benefit at all those entrepreneurs concerned by the project. i. Finaily, the government commitment to the project was insufficient to ansure appropriate project coordination. It did not also take the decisions needed to ensure the implementation of the ULB component (para. 18) or to facilitate that of the artisan component (para. 27). E. PROJECT IMPLEMENTATION 14. On account of the shortcomings mentioned above, the project got off to a very slow start. It was declared effective in September 1980 but, with the exception of the rehabilitation of ULB's equipment which started imediately, it remained inactive for years. The technical assistance to SERVULART became operational only in June 1982, the first subprojects to be financed out of the line of credit were approved in September 1983 and the training of small entrepreneurs did not start until November 1985 and the decision by ULB to initiate, as envisaged, a study of the expansion of its production capacity from 30,000 to 50,000 liters of milk per day, was taken at the end of 1985. The June 30, 1985 credit closing date was therefore postponed three times: first to June 30 1986, then to June 30, 1987 and finally to June 30 1988. Thus, the implementation of the project lasted almost 8 years instead of the 4.5 years originally planned, which in retrospect was very optimistic 21 15. No changes were made to the project objectives. In 1987, however, another component was added so that the project funds could be used to help, 2/ According to the region's profiles, the disbursement period of this type is 9 to 10 years.  5 along with other donors, fight a locust Invasion. The project proceeds have been reallocated four tinens once to establish a revolving fund and speed up disbursements; another time to reallocate the funds initially earmarked for W' s expansion to other components; and twice to increase the amount allocated in US$ to subprojects under construction, when the CFAF appreciated significantly against this currency. 16. Two events, unforeseen at project appraisal, contributed also to hamper the implementation of the line of credit and of the ULB rehabilitation component. As regard the former, -the SAR had rightly indicated that one of the reasons for adopting an *apex'-type mechanism was BDMI's fragile sltuation. However, no measure was included in the project to prevent this bank from using the line of credit as long as its finanv:ial situation had not been restored. Moreover, the willingne3s of the other commercial banks to participate in the project had not been correctly assessed. As a result, BIN became the only bank to finance subprojects with the project's proceeds and its worsening financial situation and weak organization have affected the quality and operations of the investments financed and explain why about a third of them (para. 21) did not perform satisfactorily. More importantly, in retrospect, if the project had addressed the critical situation of BDM that project preparation had identified, the collapse of this institution which took place in 1986/1987 might have been avoided. 17. With regard to the ULB component, while the SAR had projected sales increasing to 50,000 liters per day by 1985, the factory started facing marketing problems when it reached, after having renovated its equipments, a production oi 30,000 liters per day. This unexpeuted development made uBI's expansion questionable. In addition, following a government policy consisting of paying high producer prices to encourage local production, ULB started in 1986 to process huge quantities of local fresh milk purchased at a price well above the selling price of its finished product. ULB which was until then profitable started to post losses because, contrarily to the assumption underlying this component, the Government did not ensure that this factory would operate under strict 'commercial conditions". P. PROJECT RESULTS 18. On the whole, notwithstanding the impediments to its implementa- tion mentioned earlier, the project achieved some of its expected results. The project's overall objective and expected benefits of creating a nucleus of essential technical and financial support for Mal! private entrepreneurs and artisans were obviously not reached. However, three of its five components, the line of credit, the strengthening of CEPI and the training of artisans, were fully and quite reasonably well implemented. The following paragraphs analyze the results of the project in more detail. 19. The Line of Credit. Despite a slow implementation pace, this line of credit, as compared to those granted in the region, has been quite successful in terms of number of subprojects financed as well as in terms of proportion of problem-free subprojects. It has finally used more funds and financed more subprojects than expected: the US$3.5 million initially allocated to this component was increased to US$4.4 million to finance 27 subprojects. About 30 Z of the line of credit (instead of the 35? originally planed) was used to finance 15 small-scale subprojects, i.e., those,  6 subprojects. About 30 1 of the line of credit (instead of the 35X originally planed) was used to finance 15 emall-scale subprojects, i.e., those, according to the Project Agreement, whose investment cost was lover than PCAF 50 million (about US$180,000) or investment cost per job created vas under CIA? 4 million (about US$14,000). 2C. Of the 27 subprojects financed, 19 are now in operation and 8 still under construction. Five of those now operating are reported as having difficulties because of marketing problems, lack of working capital or inexperienced management. Similarly, 3 of the subprojects under construction might not be completed because of underestimated investment cost and lack of complementary financing. On the whole, therefore, 21 ot the 27 subprojects established with IDA financing can be considered as problem-free. Problem subprojects where concentrated in two subsectoro where either the demand had been overestimated (bakeries) or the managerial capability of subproject sponsors had been overestimated (jewelry stores). 21. As anticipated at appraisal, the bulk of these subprojects operate In the food and beverages subsector (11 subprojects or 41Z) and in agro- industries (5 subprojects or 202). The number of subprojects in construction materials, light chemicals, textiles and services, has been modest. On the whole, the subprojects financed by the line of credit have created about 450 new permanent jobs. The average cost per job created of US$19,000, although high, is in line with the cost per job created by similar projects in the region. Annex 3 gives detailed information on these subprojects. Given their short history, however, data on their financial and economic performance are not yet available. 22. CEPI. This institution received technical assistance to improve its capability in subproject appraisal %,ad in assisting medium- and small- scale enterprises. The project also provided budgetary assistance to support its operations and to modernize itg organization and iwternal procedures. This assistance, combined with that of other donor institutions (UNIDO and MNDP), has considerably improved its performance. Today, CEPI can be considered, in relatives terms, as the best promotion institution of the region. 23. CEPI has a staff of about 50 employees of which 30 are professionals. It is organized in two main departments: the project appraisal department and the department of technical assistance to industrial enterprises. With IDA's assistance, the former has become more and more active over the years. The number of projects that it has appraised, which includes all those financed by the line of credit, increased from about 10 only in 1983 to more than 20 by 1986. Although less efficieint than the appraisal department, the technical department has also becore very active in the latter years of the project Implementation, when it received adequate from UNDP. The number of enterprises that it has assisted also increased to 7 in 1984 to 15 by 1986. It should also be noticed that CEPI played a major role in setting up SERVULART, the body which successfully carried out the training of artisans (paras. 26 - 29). 24. The sustainability of CEPI's operations in the future is, however, questionable. This government institution did not receive, because of budgetary constraints, the financial support expected from the Government,  7 other than payment of salaries to its employee3. It has, so far, been able to develop its operations thanks to the technical and financial assistance of external donors such as IDA, UNDP etc.. In the final years of the project, it has started to follow the recosmendations of the supervision missions to charge a fee for the services that it provides. Fees are now being paid for subproject appraisals though not for technical assistance. However, in 1987, these fees hava represented less than 302 of CEPI's budget requirements. It is very doubtful that they will ever cover its operating expenses or that CEPI could, therefore, survive without continued external assistance. 25. The Trainian of ortisans. This component was innovative and its design was original. The training was provided by foreign master artisans and was based on the cooperative concept, a social feature firmly established in the rural areas where it was implemented. It demonstrated that training can markedly improve the production skills of artisans whenever there is room for improvement of technology and product design without brinfins the artisans in direct comt*tition with the modern sector. This explains why the training of weavers was more successful than that of the metalworkers, while the training of mecianics did not lead to substantial improvement in their perforrvnce. 26. To carry out this component, SERVULART was established, under the umbrella of CEPI, as independent extension services in Segou and Sikasso. These extension services were very lightly staffed with foreign master artisans and Malian counterparts. The main problem was to identify foreign Ma3ter artisans who would be competent and motivated trainers willing and able to leave in the difficult environment of small Malian citiess some of the trainers initially recruited had to be replaced after a short time in the field. Notwithstanding a slow start mainly due to these unforeseen difficulties (18 months for getting technical assistance contract signed, adeqr'te assistance and operating workshops in the field) this component was implemented with a relative success. About 110 artisans (weavers and mechanics in Segou, metalworkers and blacksmiths in Sikasso) received adequate training. They were taught to improve their techniques by ameliorating their own equipment or purchasing new but unsophisticated tools. They were also trained in basic production management. After having completed their training, about 80 artisans undertook to pool their skills and establish cooperatives. This attempt faced two major difficulties: the long and intricate administrative and legal procedures and the lack of financing to establish appropriate workshops. After years of efforts, the weavers, the metalworkers and the blacksmiths each set up a cooperative with the support of SERVULART and with French (CCCE) financial assistance. The artisans in mechanics did not create cooperatives, partly because they could not secure the substantial financing required, but more importantly because they would have become uncompetitive with either the well established garages or with their fellow artisans. 27. The three cooperatives employ about 270 workers in addition to the 80 trained artisans themselves. The SERVULART component has, therefore, heloed create or consolidate about 350 iobs. The weavers cooperative operates very successfully and its products are highly appreciated and in high demand. Although potentially successful, the two other cooperatives still face difficulties in getting credit to purchase the equipment required to complete  8 the settiag up of their workshops. The three cooperatives are still benefitting from SERVULART's support for administration and management. 28. Encouraged by the results of the SERVULART component the Government is presently seeking donors assistance to expand this experience to other professions and to cities other than Segou, Sikasso and the capital, Bamako. This initiative should be supported. However to carry out this project, the creation, as envisaged by the Government, of a central institution should be avoided. Instead the setting up of several light and decentralised extension services similar to SERVULART should be encouraged. 29. TPGP. The training provided by this institution under the project was not at all consistent with the project's intent to execute a Obasic management training program for local entrepreneurs ...refer_ed to the program by CEPI and the banks.* Instead, IPGP developed, with ILo assistance financed by the project, general seminars of one or two days, attended by people as diverse as small entrepreneurs, civil servants and labor union staff and students. This has had virtually no impact on the objective assigned to this component. 30. The supervision reports suggest that although this institution has received, in parallel to this project, YDA's assistance under an Education Project, it has remained very weak. Its management, although replaced twice during the last years has been largely incompetent and the objectives and priorities of the institution, in relation to its means, have remained unclear. 31. ULD. The first objective of this component consisting of Increasing its production capacity to 30,000 liter of milk and dairy products per day has been achieved. IDA decided not to flnance, as envisaged originally, a further expansion to 50,000 liters per day because the market was uncertain and, most importantly because ULB's profitability was affected as mentioned earlier. In retrospect, it seems that it was a mistake to provide assistance to this government enterprise without attempting first to reform the overall institutional framework regulating the relations of the Government with its public enterprises to ensure that they can really operate under strict ccomercial conditions. G. THE BANR' S PERFORMANCE 32. Project preparation was, understandably for a first operation, a long process (about 107 staff weeks until negotiations). The ULB and SERVULART compor-'its were thoroughly prepared by PPF-financed studies. Although the projuct design was complex, it does not seem that the ability of the Government and of the project institutions to implement it were ever questioned. Thus, in retrospect, the main weakness of the preparation stage seems to have been an underestimation of the difficulties involved in the implementation of the project. For this reason, IDA supervision had to be very intensive at the early stage of the project (about 55 staff weeks for the first 4 years or 581 of total supervision time spent over the 9 years). The review of the project documentation suggests that it might have been more advisable to drop the IPGP and the ULB components as soon as it appeared, in 198311984, that they had little chance for success.  9 R. THE DORROVERS' PERPOMANcE 33. As mentioned earlier, lack of commitment and understanding of the project objectives were the main impediments to the implementation of some components. This was aggravated by frequent changes in the local staff in charge of the project implementations the managers of IPGP, CEPI and of the project coordination unit were replaced two to three times over the implementation period. Although financial assistance had been provided to this unit, its inefficiency appears to have been the main weakness on the borrower's side. Complex project design and difficulty, common in Mali, to coordinate institutions belonging to different ministries certainly affected its performance. Its usefulness is, on the whole, questionable. I. CONSULTANT SERVICES 34. Consultants had been provided to all project institutions. Their overall performance was acceptable. Two lessons can, nevertheless, be drawn from this project. Firstly, very specialized consultants such as those needed by SERVULART should have been identified before project implementation because it was difficult and lengthy to find competent candidates willing to work under the difficult conditions of small Malian cities. Secondly, it can be counterproductive to use consultants from UN sister institutions, mostly when they are already involved with a body receiving our assistance such as in the case of IPOP, because conflict of objectives might arise. J. CONCLUSIONS AND LESSONS LEARNED 35. This !project was conceived more than 10 years ago and the lessons that can be drawn from its implementation have since been learned from similar Bank projects in Africa. The scope of this particular project, designed as an IDF-type project of that time was, however, broadened to include training of entrepreneurs, assistance to artisans and support to the rehabilitation of a public enterprise. The project was destined, therefore, to have the merits and weaknesses attached to its own design as well as those attached to a classical IDF project. 36. The merit of the project design was its broad scope designed to tackle the institutional and financial obstacles to the development of the private sector while attempting to set up an example for the rationalization of the operations of public enterprises. This proved to be also its main weakness. For a first Drolect, its scope was far too broad to be carried out efficiently by institutions andlor the Government whose degree of prepared- ness, strength and commitment to the project objectives was overestimated. Furthermore, with the benefit of hindsight, it is clear that the conceltual basis of the IP