Tunisia - Agricultural Research and Extension Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 17224 IMPLEMENTATION COMPLETION REPORT REPUBLIC OF TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT (LOAN 3217-TUN) December 15, 1997 Rural Development, Water, and Environment Department Middle East and North Africa Region 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: Tunisian Dinar (DT) Appraisal average (1990) US$ 1 = DT 0.949 Completion average (st term 1997) US$ 1 = DT 1.020 WEIGHTS AND MEASURES Metric system FISCAL YEAR OF BORROWER January 01 - December 31 ABBREVIATIONS AVFA Agence de la Vulgarisation et de la Formation Agricoles (Agricultural Extension and Training Agency) CIV Campagne Intensive de Vulgarisation (Intensive Extension Campaign) CRA Centre de Rayonnement Agricole (Local Extension Unit) CRDA Commissariat Regional de Developpement Agricole (Regional Agricultural Development Commission) CTV Cellule Territoriale de Vulgarisation (Sub-Regional Extension Unit) FAO Food and Agriculture Organization INRGREF Institut National de Recherche en Genie Rural, Eaux et Forets National Research Institute in Rural Engineering, Water, and Forests INRAT Institut National de Recherche Agronomique de Tunisie (Tunisian National Institute of Agricultural Research) 10 Institut de I 'Olivier (Olive Institute) IPM Integrated Pest Management IRA Institut des Regions Arides (Arid Zones Institute) IRESA Institut de Recherche et d 'Enseignement Superieur Agricoles (Institute of Agricultural Research and Higher Education) ISNAR International Service for National Agricultural Research PBO Programming and Budgeting by Objectives PNM Projets Nationaux Mobilisateurs (National Priority Projects) PRR P6le Regional de Recherche (Regional Pole of Research) SERST Secretariat d'Etat a la Recherche Scientifique et Technique (State Departnent for Scientific and Technical Research) SMS Subject Matter Specialist TA Technical Assistance T&V Training and Visit Extension Methodology UNDP United Nations Development Program I Vice President: Kemal Dervi I Sector Leader: Mark D. Wilson Acting Director: Salah Darghouth Task Teamn Leader: Charles Ameur  FOR OFFICIAL USE ONLY IMPLEMENTATION COMPLETION REPORT TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT (Loan 3217-TUN) TABLE OF CONTENTS PREFACE .............................................. i EVALUATION SUMMARY ...............................................ii PART I: PROJECT IMPLEMENTATION ASSESSMENT . ..............................1 A. STATEMENT AND EVALUATION OF OBJECTIVES ...............................................1 B. ACHIEVEMENT OF PROJECT OBJECTIVES ...............................................2 C. MAJOR FACTORS AFFECTING THE PROJECT .......... .................................... 10 D. PROJECT SUSTAINABILITY .............................................. 10 E. THE BANK'S PERFORMANCE ..............................................1 l F. THE BORROWER'S PERFORMANCE .............................................. 12 G. ASSESSMENT OF OUTCOME .............................................. 12 H. FUTURE OPERATIONS .............................................. 12 I. KEY LESSONS LEARNT .............................................. 13 PART II: STATISTICAL ANNEXES Table 1: Summary of Assessments Table 2: Related Bank Loans Table 3: Project Timetable Table 4: Loan Disbursements: Cumulative Estimated and Actual Table 5: Key Indicators for Project Implementation Table 6: Key Indicators for Project Operation Table 7: Studies Included in Project Table 8A: Project Costs Table 8B: Project Financing Table 9: Economic Costs and Benefits Table 10: Status of Legal Covenants Table 11: Compliance with Operational Manual Statements Table 12: Bank Resources: Staff Inputs Table 13: Bank Resources: Missions 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.  APPENDICES A. Aide-Memoire B. Re-allocation of Loan Tranches C. Detail of Loan Disbursement and Achievement of Objectives D. IRESA's Contribution (Summary) E. AVFA's Contribution F. Map  IMPLEMENTATION COMPLETION REPORT REPUBLIC OF TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT (Loan No. 3217-TUN) PREFACE This is the Implementation Completion Report (ICR) for the Agricultural Research and Extension Project in Tunisia, for which loan 3217-TUN in the amount of US$17 million was approved on June 5, 1990, signed on September 24, 1990, and made effective on May 21, 1991. The loan was closed on June 30, 1997, one year behind schedule. Final disbursements took place on October 31, 1997, at which time a balance of US$1.7 million was canceled. The United Nations Development Program (UNDP) provided cofinancing for the project, used for the extension component, in the form of a US$1 million grant. The Food and Agriculture Organization of the United Nations (FAO) and the World Bank Cooperative Program prepared the draft ICR for the Middle East and North Africa Rural Development, Water, and Environment Group (MNSRE) of the World Bank.' This ICR was based, among other things, on discussions in Washington with Mr. Charles Ameur, task team leader, on material gathered from the project files, and on the findings of the Cooperative Program mission that visited Tunisia from May 12-23, 1997. The Borrower contributed to the preparation of the ICR by drafting its own Completion Report (May/June 1997), by arranging field visits and meetings, and by making substantive comments for the mission's aide-memoire (Appendix A). Mr. Charles Ameur finalized the report and Mr. Mark D. Wilson, sector leader, MNSRE, reviewed it. Messrs. F. Dauphin (agronomist, mission leader) and E. Hidier (economist). i  IMPLEMENTATION COMPLETION REPORT TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT (Loan No. 3217-TUN) EVALUATION SUMMARY Introduction The Agricultural Research and Extension Project was appraised in October 1988 and approved by the Board on June 5, 1990. At appraisal, project costs were estimated at US$34 million, toward which the Bank provided a loan of US$17 million. The Borrower was the Republic of Tunisia. Created soon after loan signature, the Institute of Agricultural Research and Higher Education (IRESA) was entrusted with the research component of project implementation and the Agricultural Extension and Training Agency (AVFA) with the extension component. The original closing date was June 30, 1996, which was extended to June 30, 1997. Final disbursements took place before October 31, 1997. The total investment costs amount to US$21 million of which US$15.3 million was disbursed against the Bank loan. This represents 90 percent of the appraisal target. During implementation, part of the loan was re-allocated in favor of civil works and equipment, whereas 10 percent of the loan instead of the planned 26 percent was allocated for technical assistance (TA) and extemnal training. The United Nations Development Program (UNDP) provided an additional US$1 million grant to cofinance the TA and training expenses of the extension component. Project Objectives The project's objectives were to improve the overall efficiency and relevance of Tunisian agricultural research and to enhance the dissemination of research results that were adapted to the needs of the farming community through a unified and revamped extension system. Implicitly, the ultimate objectives were to promote greater efficiency, increase productivity in the agricultural sector, and improve farmers' incomes. The description of the weaknesses of the research system in the Staff Appraisal Report (SAR) suggested more specific objectives, including, for the research component, (a) coordinating research institutes by defining priority objectives and improving linkages with extension; (b) improving overall research management by adopting appropriate tools, methodology, and procedures; (c) making research more relevant to local needs by developing regional research centers and strengthening socioeconomic surveys; (d) upgrading research quality by appointing staff and providing training, evaluation, and better facilities; and (e) providing direct support to ten priority programs. The extension component was designed to (a) improve the overall management of extension services by unifying them under a single agency and introducing programming, monitoring, and evaluation; (b) upgrade the quality of extension messages by training staff and ii  ensuring close linkages with research; and (c) expand the audience via mass media communication and increased staff mobility. Implementation Experience and Results Although the project was slow in starting up, mainly because of the establishment of IRESA and AVFA on the onset of the project, most of the objectives have been achieved and the outcome is satisfactory. Research Component. IRESA has played a critical role in coordinating the activities of the various institutes involved in agricultural research. It has expended considerable effort on the programming exercise, which is one of the project's major benefits. IRESA set up national commissions with a wide spectrum of participants (academics, extension officials, representatives of the farming community), defined and ranked priority research programs, and assessed available human and budgetary resources. In parallel, based on project recommendations, IRESA also recruited additional staff and strengthened facilities, although with some delays because of a commendable willingness to assess needs carefully prior to disbursement. It is only towards the end of the Project that a monitoring and evaluation system was implemented due to the lengthy period of time spent on programming. Also, there was a need to devise a system that was specific to agricultural research. The expected improvement of research efficiency, especially at the laboratory and station levels, was hampered by external factors to the Ministry of Agriculture. These included among others, the rigidity of research management procedures and the lack of financial authority delegation. In addition, partly because of IRESA's limited leverage over research budgets, the programming exercise fulfilled only part of its objectives, which ultirnately consisted of re-orienting research budgets. Along the same lines, the absence of a revamped system of career development which was not part of the project, remains a major constraint to the promotion of research teams that would focus on farmers' needs. The understanding of these needs was to be improved by recruiting agro-economists at the regional level. IRESA recruited only three out of the six agro-economists envisaged due to the limited number of positions opened by the Govenmment. Also, such specialists are few on the labor market and in high demand, which lead to financial conditions that IRESA could not match. Two major objectives were partially met, narnely: (a) the restructuring of research stations, and (b) the regionalization of research activities due to delays in preparing the restructuring plan for the former, and in approving the necessary Decrees for the later. The initial proposal to strengthen three regional centers was modified to create seven regional research poles whose mandate was to coordinate research at the regional level. Three such poles were established during the project but they lack a clear definition of their attributes. It is worth noting that those poles created can tap critical masses of researchers both from regional centers and from higher education entities. Extension Component. Throughout project implementation, AVFA has played a leading role in energizing the extension system nationwide, developing new programming and monitoring methodologies, focusing on staff training, and expanding the audience of extension messages through mass media communication. The TA cofinanced by the UNDP was generally helpful, though lacking a clear strategy. The disbursement of the Bank loan to finance construction and rehabilitation of facilities and to supply additional equipment to local extension teams (such as vehicles and photocopiers) went according to plan. Originally envisaged for three pilot regions, the iii  women extension program, which focused mainly on home-based income generating activities and energy conservation, was successfully extended to eight zones. Among various obstacles, the ambitious status of the extension coordinators at the level of the Ministry of Agriculture's regional offices (CRDAs) was generally considered a weakness in that the coordinators are not in a position to prevent CRDA division chiefs from using extension teams for other tasks. In addition, the lack of a formal status for extension staff hindered the retention of motivated personnel. As for the research component, evaluation exercises will not produce sufficient quantitative data on the impact of extension services. In any case, the monitoring and evaluation units at both the center and the CRDAs were markedly understaffed. Overall, a clearer strategy for the future of extension still needs to be defined taking into account the diversity of farming systems, the sustainability of the extension services, and ways to involve further the farming community and the private sector. Research-Extension Linkages. Mechanisms have been developed throughout project implementation to promote research-extension linkages, including having extension workers and CRDA personnel participate in national and regional research commissions, increasing the number of open house days at research stations, and inviting researchers to make greater contributions to the development of extension material. However, the process is not yet truly interactive, and the feedback from the farming community to define national, or even regional, priority programs is still weak. Meanwhile, the Ministry of Agriculture created technical centers outside the IRESA-AVFA system to further involve the farming community. These centers need however a clear mandate and significant input from farmers and professional associations in defining their programs. Major Factors Affecting the Project and Sustainability The evaluation programs, the re-orientation of research budgets, the decentralization of authority at the station and laboratory levels, the appointment of staff outside Tunis, the drastic restructuring of the research stations, the promotion of research staff involved in development work, and the better understanding of local farming systems have not yet attained the levels originally envisaged in the project objectives. However, all these reforms are under way and should generate tangible results. The future of the extension system developed by the project will depend greatly on the capacity to adapt to farmers' needs and to the emergence of new extension channels. The improved managerial and technical capacities gained by extension during project implementation constitute a definite asset. A remaining issue is the lack of consensus in defining a national strategy for extension. Bank and Borrower Performance The Bank's satisfactory project identification and preparation, carried out with the assistance of the Food and Agriculture Organization of the United Nations/World Bank Cooperative Program, ensured a high degree of project ownership. Staff continuity, regular supervision missions, and a thorough mid-term review in April 1994 provided the Project with appropriate guidance during implementation. In the case of extension, the single training and visit (T&V) methodology was substituted for more creative approaches. As early as 1993, the Bank strongly recommended that Tunisia define and embark on a clear extension strategy that would iv  shift more responsibility to farmers and would allow the public sector gradually to divest itself of extension activities while inviting the private sector to participate as a partner. Meanwhile, supervision missions could have pushed some important aspects of the project more forcefully, especially regionalizing research activities and improving research management. The issue is that the TA component to support restructuring and regionalization as weli as to assist the various commissions was never used. The Bank was somewhat late in reacting to the creation of the regional poles which was seen as afait accompli especially as the latter had neither a clear mandate nor proper funding. These concerns were subsequently tackled by IRESA. The Borrower has made sound use of the resources the Bank provided and has always kept in mind the main objectives of the project. Project Outcome Both research and extension, and thus the agricultural sector indirectly, have gained from the Project, albeit the late devising and implementing of the evaluation mechanisms made quantifying the project's benefits difficult. Multiparty commissions listed, discussed, and ranked research programs, available resources were assessed, and a number of dynamic programs benefited from the recruitment of new staff and the strengthening of facilities. Budget re- orientation according to priorities and program evaluation are being introduced. Extension is much better organized. Beneficiary participation in defining priorities was initiated. Staff are better trained and more mobile. Also, activities are programmed, monitored, and are beginning to be evaluated. In sum, progress toward project objectives as defined in the Staff Appraisal Report has undoubtedly been achieved. However, in the case of research, the objectives might have been overambitious for a five-year implementation period. To reap the full benefits of the Project, further reforms will have to be initiated. The extension component has strengthened the agricultural sector's capacity to cope with future restructuring made necessary by the diversity of Tunisia's farming community. Meanwhile, a clear strategy for extension still needs to be defined and implemented. Future Operation While the Tunisian economy is rapidly opening to foreign markets, and while the division of responsibilities between the private and public sectors in delivering farming services is being discussed, both research and extension need renewed vision and reform. In the extension sector, agricultural organizations are urged to play a greater role in providing and financing extension services. The training of young farmers has also been identified as a new priority. For research, regionalization should be firmly favored, while IRESA's role in effectively evaluating and guiding research needs to be confirmed. These challenges should be addressed through a proposed Agricultural Support Services Project, for the preparation of which a Technical Assistance grant has been obtained. Key Lessons Learnt The key lessons learnt from the project include the following: v  * Programming and budgeting by objectives is a complex approach that required some three years for IRESA to adopt properly. * Because this approach was a critical factor in the further implementation of the research component, this delay had a negative impact on project implementation. * When agricultural research is to undergo complex restructuring, expecting rapid disbursement from the start of the project is unrealistic. * Given the previous point, when a project design includes a new and complex approach, the local context, capacity, and constraints should be carefully analyzed and assessed. * To achieve regionalization of research, the Project should have placed more emphasis on the necessary prerequisites. In particular, a consensus with other Ministries involved should have been sought to solve such issues as the delegation of financial authority, the appropriate posting of staff, and the development of critical masses of research activities. * Considering the time and energy needed to develop comprehensive systems for programming and budgeting by objectives, simpler procedures, like projets nationaux mobilisateurs, to redirect research budgets effectively during project implementation constitute a good compromise. * Implementation was clearly obstructed by the need for regulatory changes that were not envisaged at appraisal. * The obvious need to up-grade the level of the arrondissement vulgarization was a case in point; never tackled it resulted in extension remaining under the authority of all the other directorates at both the central and the regional levels, thus having the lowest priority. * The need for unanticipated regulatory changes suggests that to facilitate reforms such as those promoted by the Project, more thorough preparation to understand better and to tackle decisionmaking processes and bureaucratic procedures might be envisaged. * The introduction of a national extension system based solely on the T&V system proved inadequate in the Tunisian context. * Modifying the approach to extension, especially at the national level, requires having a clear vision and defining a strategy for extension at an early stage, together with intense and effective participation by farmers. * When elaborating a strategy, decisionmakers must decide whether extension should remain the sole responsibility of the public sector or whether it should rely on a mix of public and private interventions, the level of producers' contributions to the costs of extension, and the amount of funds the government is ready to commit to extension on a regular and long-term basis. * To improve the motivation of field staff, project management should give serious consideration to preparing proposals for legislation that would upgrade the status of extension staff to be submitted to decisionmakers for their consideration. vi  IMPLEMENTATION COMPLETION REPORT TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT (Loan 3217-TUN) PART I: PROJECT IMPLEMENTATION ASSESSMENT A. STATEMENT AND EVALUATION OF OBJECTIVES 1. The Bank's assistance strategy for Tunisia's agricultural sector emphasized the need for increasing productivity and efficiency, given that horizontal expansion of rainfed agriculture had exceeded the limits of sustainable resource use and the scope for further expansion of irrigation areas was becoming limited. For this strategy to work, research needed to be more efficient and better focused and to be served by an effective dissemination system. The project's general objectives were simply stated in terms of improving the institutional and organizational framework to increase the efficiency of both research and extension and rectifying a number of weaknesses highlighted during the preparation process. Thus for research, the main objectives were to * Guide and coordinate research through a new institution responsible for setting priorities, programming activities, and creating linkages with extension * Improve research management by introducing programming and budgeting by objectives and systematic monitoring and evaluation and by restructuring the research facility network 3 Adapt the research programs to local conditions by establishing three regional centers, developing on-farm trials, and strengthening socioeconomic studies * Raise the quality of research by hiring staff, providing training and evaluation, upgrading research facilities, and managing research facilities better * Support directly priority programs corresponding to ten themes identified during preparation. For extension, the main objectives were to * Improve overall management by uniting extension systems under one agency; introducing programming, monitoring, and evaluation of extension workers' activities; establishing coordination and supervision units at the regional and subregional levels; and introducing the training and visit (T&V) approach * Improve the quality of extension by providing initial and continuous training of staff and by establishing close linkages with research for regular updating and adaptation of messages * Expand the audience for extension by developing mass media communications and achieving better geographic distribution and mobility of staff.  2. To achieve these objectives, the project consisted of a program of investments associated with a series of organizational reforms, staff appointments, and targeted training and technical assistance (TA) and the provision of adequate budgets to cover related operational costs. The research component envisaged (a) establishing three regional research centers at Beja, Mornag, and Sahline and strengthening eight existing research stations; (b) providing scientific, farm, and office equipment and improving laboratories and staff housing at the five central research centers and at the regional centers and research stations; (c) setting up gradually a system of programniing and budgeting by objectives; (d) appointing forty-two additional research staff and providing TA and short-term training with an emphasis on research organization and programming; and (e) improving linkages between research and extension by strengthening the liaison unit with extension, appointing six agro-economists, and establishing working groups at each of the three regional centers. 3. The extension component included (a) strengthening about 550 extension centers (CRAs) by providing vehicles, office and demonstration equipment, and staff training and constructing 54 new CRAs; (b) providing TA and training to build up methodological tools for extension; (c) introducing a pilot extension program for rural women; (d) strengthening 15 regional extension units and 104 supervisory subregional units (CTVs) by assigning staff and equipment; and (e) strengthening a central extension support unit, the Agricultural Extension and Training Agency (AVFA), by assigning and training staff and providing equipment. 4. Proposals for research under the project were based on a comprehensive analysis of the existing system and of national research needs conducted by a national team from 1985 to 1987 with the help of the International Service for National Agricultural Research (ISNAR) and the World Bank Cooperative Program. For extension, a review had highlighted existing deficiencies in organization and management, but an analysis of farming systems and related extension needs was totally lacking. For both subsectors, a masterplan had been prepared (with support from the TA project, loan 2197-TUN). The government endorsed the recommendations and started implementing reforms during project preparation, which formed the basis for the project. In general, the objectives and the means chosen to achieve them were sound, and could be expected to have an impact on agricultural production. The difficulties of maintaining the pace of reforms and adequate levels of recurrent expenditures in a context of budgetary constraints had been correctly identified as serious risks, but perhaps not fully taken into account in the design of the project. B. ACHIEVEMENT OF PROJECT OBJECTIVES Research 5. Institutions. IRESA, officially established in 1990 to coordinate research and higher education in agriculture, becane operational in April 1991. Despite delays in filling senior staff positions, and since that time has played a growing role in coordinating research, IRESA has brought together researchers from the various research institutes and educational institutions by obtaining a similar status for both and by involving them in the preparation of national priority programs. It was instrumental in setting up programming commissions and guiding the programming process (paragraph 9), in setting up a number of regional research poles (PRRs), in synthesizing available research results, and in fostering closer linkages with extension. Its role in managing agricultural research still remains modest, as IRESA has little control over the budgets of the research institutes under it (paragraph 10). While the creation in 1991 of the State 2  Secretariat for Scientific and Technical Research (SERST) initially raised doubts about IRESA's survival, excellent collaboration between the two organizations has since been established. 6. The SAR envisaged the creation of a national research council to provide policy guidance and evaluate research. With the creation of IRESA, the view taken was that IRESA's board would play the same role, thereby satisfying the related covenant. However, it has become apparent that the board is ineffectual in its policy and evaluation roles for reasons that should be clarified. 7. TA support (six man-months) and training of financial and administrative staff (twelve man-months) was intended to strengthen research organization and management. However, only 17 percent of the training aimed at institutional development was achieved (that is, two man- months). Perhaps more important, the loan agreement stipulated that suitable procedures would be developed to improve the management of research activities, in particular, that financial authorities would be established at the regional centers. This has not happened, partly because the PRRs, which have replaced the concept of regional centers, were only conferred a coordinating role. However, the lack of delegation of authority in general was and remains one of the most serious impediments to doing research in the regions and to enhancing efficiency in research management. The Government has just passed legislation that recognizes laboratories and research stations as "Entites Fonctionnelles', which represents a first step in that direction. It should allow institutes to designate laboratories and research stations as "Ordonnateurs Secondaires"and entrust them with minimal financial authority. 8. One of the necessary start-up activities presented in the SAR was the development of new methods of financial allocation and analytical accounting. However, the appropriate tools started to be created only in 1994, when the Tunisian National Institute of Agricultural Research launched a new analytical accounting and budgeting computer system based on ISNAR's information management system with the assistance of a consulting firm. To date, the system still needs to be tested and installed at the other research institutes under IRESA. 9. Programming and Budgeting. In 1992 IRESA created national commissions for each of the ten priority sectors that were identified during project preparation. The methodology followed ISNAR's recommendations, and was piloted throughout 1993 with TA support.' Systematic programming was new and bound to be painful initially. The main difficulty stemmed from the limited and irregular participation from extension services, the public sector as well as from farmers organizations. However, preliminazy programs were ready for most commissions by 1994, and by 1995 for others. Although of unequal quality, they provided a common framework for discussing research activities and funding requirements, particularly for activities that were to be supported by the loan. Moreover, their preparation drew together scientists who had never cooperated before and made them listen to potential users and justify their proposals. By 1996, forty-five of the seventy programs the commissions had identified were being funded and implemented. A second programming cycle started in 1996 for inclusion in the ninth five-year plan. Through a consultant from Morocco, the second exercise benefited from that country's experience in developing a programming methodology with ISNAR. Priorities are now ranked and take into account staff resources that exist or that can realistically be made available. The methodology envisaged seven steps, from sector analysis to an action plan and budget for individual projects through analysis of sector constraints and past research results, priority setting and project identification, and assessment of existing and supplementary resources needed. 3  The process encompasses both existing and new projects (a total of 291) and has improved markedly. The programming exercise represents one of the most significant benefits of the project. However, room for improvement still exists, and efforts should be pursued to overcome a number of weaknesses, for example, the scarcity of multidisciplinary research projects and the lack of monitorable objectives. Programs also need to be published. 10. Budgeting by objectives, the last step in the programming exercise, was fonnally achieved in 1995 or 1996, depending on the commission. However, while research projects are now documented with their human and financial requirements, the influence of programnming over resource allocation-a process that remains essentially within the purview of individual research institutes-is unclear. In practice, the main resource over which IRESA has direct leverage is the budget SERST allocates for projets nationaux mobilisateurs, an annual program of national scientific priority projects. This resource represents less than 10 percent of total budget resources allocated to the four institutes under IRESA, but most of that is available to finance new projects. As to the institutes' main budgets, they have hardly been affected by the programming exercise. IRESA, which is responsible for ensuring the relevance of on-going programs in relation to priorities set by the national commissions, formally reviews their budgets. 11. Monitoring and evaluation of programs was slow to be implemented until the new programs were prepared. Since then, the emphasis has been on the design and implementation of a computer system that would monitor inputs (research staff, program budgets, equipment, and so on), delaying the assessment of scientific outputs. So far, however, only projets nationaux mobilisateurs have been submitted to a consistent monitoring and evaluation process using standard procedures that are built into the programming exercise. No established mechanism to monitor other programs is available, although IRESA plans to initiate this in conjunction with the programming commissions, and in 1997 hired a consultant for advice. The officer of the Sous-Direction du Suivi et de I 'Evaluation, who was finally recruited in 1996, received some training overseas and is now developing a database of researchers as well as supervision and evaluation forms for testing. Because of the researchers' misconceptions about evaluation, the Tunisian National Institute of Agricultural Research had to sensitize its research staff and launch preliminary stocktaking exercises. While the SAR had envisaged the appointment of ad hoc commissions to carry out regular external evaluations of research progress, this did not happen. 12. Facilities. Procurement for construction and equipment was initially slow, but gradually picked up after the mid-term review in 1994, and was essentially finished before project completion in June 1997. The proceeds of the loan allocated to civil works was increased by US$600,000 and that allocated to equipment was increased by US$2.75 million from the original amounts specified in the loan agreement. This was made possible by the re-allocation of US$1.3 million initially budgeted for both TA and external training (paragraph 32) and the allotment of US$2.05 million, which represented two-thirds of the unallocated portion of the loan. While delaying the deployment of regional research, IRESA took the commendable decision to check facilities and equipment specifications against the precise identification of research priority programs and to make an accurate inventory of existing equipment. 13. Research Network Restructuring. The excessive number of research stations, many of them ill-equipped and barely maintained, was to be drastically reduced from forty-six to about twenty-five by December 1992, following a plan to be agreed on in 1991. A restructuring plan 4  prepared by consultants following a thorough review of existing facilities was finalized in May 1996. Under the maximum closure scenario, the plan envisaged the closure of thirty-two stations and the retention of sixteen, including seven PRRs and nine substations, which would have required significant investment (DT 11.5 million) and additional recurrent costs of DT 0.5 million per year, for which no provision had been made.2 IRESA's decision is to maintain seven main stations and twelve program stations to ensure that on-going programs are retained; however, it has taken no action yet. While new construction could be phased in over time, an immediate difficulty is the loss of fanm income from stations to be ceded, for which institutes expect budgetary compensation. 14. For most stations their allocated budgets do not cover basic maintenance, resulting in obsolete, increasingly out of repair facilities, and in subverted commercial production progressively replacing scientific work. In the meantime, significant amounts of funds have been earmarked in the 9th Plan to cover the restructuring plan. Also, IRESA is consulting the institutes concerned by the restructuring plan in order to come up with a final proposal for submission to the decisionmakers. 15. Staff Development. One of the project's initial achievements was to obtain similar status for research staff and professors. IRESA recruited a total of forty-seven additional researchers during project implementation, exceeding the target of forty-two proposed in the SAR. However, while two programs (forestry and arid zones) benefited from significantly more human resources than originally planned, only three of the recommended six agro-economists were recruited. This is unfortunate, because basing agro-economists in the regional centers would have added to the overall understanding of farming systems and to the multidisciplinary approach that is generally lacking in the programming process. The geographical distribution of staff recruited also illustrates the difficulties of the regionalization process. Apart from the nine research staff that joined the histitute for Arid Region's main center in Medenine, only four new researchers were posted in the regions. 16. The current promotion system still favors academic results, for example, article publication, and does not encourage research work directly targeted toward the needs of the farming community. This both hampers the process of regionalization and contradicts one of the main objectives of the project, which was to make research programs more relevant to the needs of the agricultural sector. Revamped, dynamic promotion procedures are certainly among the lacking, but necessary, tools of the overall evaluation system. 17. The project also aimed at achieving staff development by financing external training. Only 57 man-months of training were eventually used, that is, 23 percent of the 243 man-months proposed in the SAR (paragraph 32). In fact, IRESA took advantage of bilateral and international fundings to supplement its overall training program. 18. Regional research. The decision to create seven PRRs instead of four regional centers (the three envisaged at appraisal plus the Institute for Arid Regions at Medenine, previously established) was made in 1992. The new policy was to emphasize local cooperation between research and teaching institutions in implementing regional programs and to increase agro- 2 The minimum closure scenario retains eight main stations, seventeen substations, and thirteen experimental sites and requires the same total investment, while incremental operating costs increase to DT 0.72 million. 5  ecological coverage. The PRRs were designed to have a coordination rather than an implementation role. Meanwhile, the research capacity remains uneven among the various regions. They remain weak in the center West and South West while being stronger elsewhere (ESA Kef and ESA Mateur for example. 19. Delays in issuing the related decree (signed in June 1995) and the lack of resources earmarked for PRRs have contributed to a difficult start. Only two PRRs have been formally created so far (at Sousse and Le Kef). These are currently developing regional research programs, for which funding from regional governments and internationally funded projects is expected, and are promoting modest on-farm research activities. Two additional PRRs at Mornag and Beja should be established by the end of 1997. 20. PRRs should play a major role in mobilizing researchers to serve local priorities. However, their impact will be limited until a real capacity for research exists locally, that is, the regional centers, now main stations, are effectively strengthened. While construction has been completed and equipment is in place in the three original sites (Mornag, Sousse, and Beja), little progress has been made in posting staff in the regions, nor has station management been improved, although both were seen as critical before the project. 21. With the exception of the Institute for Arid Regions at Medenine, which received nine new staff, practically all new recruitment under the project (thirty-four of thirty-eight recruited, excluding the institute, were for Tunis, whereas the SAR envisaged that three-quarters of such staff would be posted outside the capital. The reluctance of staff to work in the regions is due not only to the lack of facilities and financial incentives, but to the fact that all decisions remain centralized. Apart from the staff motivation aspects, improved station management, particularly the possibility of managing sufficient funds locally to cover daily expenditures, for instance, machinery repairs, small equipment purchases, is crucial to achieve good quality research. 22. Research Programs. Programs supported under the project have benefited to some extent from improved facilities and new equipment, as well as from staff recruitment, as identified during preparation. This has permitted research that was not possible before. A good example is research into mechanization, which thanks to new equipment was able to prototype and test cereal threshers and grain cleaners for small farms and to test combined equipment for land preparation and sowing for large-scale rainfed farming, which has met with considerable interest from farmers. 23. While in most cases the programs that received such support only started in 1995 and remain to be evaluated, the support provided has probably been put to good use. This cornclusion is justified because of the care with which the new programs were selected and designed during programming, and because the resources were given to institutes and teams that despite all the shortcomings the project was addressing, have continuously demonstrated their vitality. This is supported by the number of new varieties and techniques the various institutes have passed on to extension services during project implementation. Among the most recent releases are nevv varieties of durum, bread wheat, barley, triticale, and oats, most of which show higher resistance to prevalent diseases. New varieties of apricot, cucumber, and potatoes have also been created and successfully introduced. Other accomplishments include the production of various feed rations using local produce, the biological control of olive tree pests, the introduction of integrated pest management techniques for citrus crops, and the screening of drip irrigation equipment. 6  24. Scope exists, however, for further improving national programs. In the absence of systematic evaluation of on-going or new research programs from proposal to impact assessment, research priorities likely do not receive adequate resources and unsound programs could, although in rare cases, continue to be funded. SAR recommendations to promote multidisciplinary teams, to develop economic and social inputs, and to study production systems have had a limited impact. Extension 25. Institutions. AVFA was established in 1991 to coordinate all extension in the country. It had a leading role in program and methodology development, staff training, monitoring and evaluation, and dissemination via the mass media. AVFA rapidly succeeded in establishing an effective capacity for training staff, setting up regional programming and monitoring commissions, establishing an effective programming and monitoring mechanism, procuring project equipment, and putting in place the conditions for meeting project objectives. By means of a dynamic core team, and despite limited resources, AVFA has energized the entire extension system. However, the agency has encountered difficulties in confronting issues that had not been well identified at appraisal-for example, diversity of demand, cost efficiency of delivery, financial sustainability- and generally in charting the future of extension. As concerns research, AVFA's board was expected to provide policy guidance, which has not materialized. 26. Regional Level. Fifteen of the twenty-three regional offices of the Ministry of Agriculture (CRDAs) were designated at appraisal to receive project support for extension over the whole or a part of their region. In each CRDA, a coordinator for extension was appointed to take responsibility for all extension activities in the region. However, the coordinators' low status gives them little authority over extension staff, who often have to perform tasks at the direction of the CRDAs' division chiefs. This issue, repeatedly raised since 1992, has not yet been solved, possibly because it is linked to other, more fundamental, issues.3 27. CRA and CTV Strengthening. At the local level, 550 CRAs were to be strengthened and about 54 new centers were to be added. A total of 616 CRAs benefited from training and methodology support, and 40 new centers were built, while 60 were improved. Less new construction took place than envisaged, as many unused buildings could be made available for extension. For transport, 347 vehicles were procured, including 151 motorcycles and 196 cars, which is approximately according to plan. The procurement of office and demonstration equipment was only 53 percent of SAR forecasts; however, computers and photocopy machines were procured for all twenty-three regions, whereas the provision of only five copiers had originally been foreseen. At the subregional level, 163 CTVs have been supported, which seems excessive compared to the number of CRAs (one supervisor for eight to ten field staff is usual). 28. Staff Development. Training received much emphasis from the beginning of the project. Initially envisaged at one site, Sidi Thabet, for all 1,100 staff, the approach was subsequently changed to a two-stage system, whereby trainers trained at Sidi Thabet were then responsible for training other staff at five regional centers. Eleven modules were developed with the assistance of United Nations Development Program (UNDP) and Food and Agriculture Organization (FAO) The difficulty arises not so much from the status of the coordinators as from the numerous national programs that other divisions have to implement with limited or no field staff. These tasks, such as planting fruit trees on private farms, for which quantitative targets have to be achieved, take precedence over extension programs, even if these respond to farmers' demands. 7  consultants, totaling fifty-five days of instruction, and covering diagnostic surveys, an introduction to faniing systems, communication techniques, extension methodology, and monitoring and evaluation of extension activities. The mid-term review recognized that extension workers who had received the training were clearly better equipped to perform their duties than those who had not. Virtually all 1,100 staff were trained during project implementation, including CTV and CRDA supervisors. While subject matter specialists provide some form of on-the-job training at the regional level, the need to pursue formal training with refresher courses and with new modules is fully recognized. Regarding the training center at Sidi-Thabet, the Bank suggested that the training personnel include at least one female trainer, and this has been done. 29. The Bank regularly raised the need for better incentives and recognition of extension workers during missions from as early as 1992. Estimates indicate that about 10 percent of the individuals trained by the project have since left the extension service. Although this figure is not high, the chances are that those who have left include some of the more dynamic staff, as extension staff have no career prospects. Without better status, recruiting and retaining the high caliber personnel that extension increasingly needs will be difficult. 30. Programming, Monitoring, and Evaluation. AVFA has also emphasized programming and monitoring. Programmning is a complex exercise that combines national priorities with a participatory process to identify local needs. Researchers and subject matter specialists are invited to take part in the process. Programs are thus defined annually at the CRA, CTV, and CRDA levels. Combined with rigorous monitoring, programming has imposed discipline and focused activities on agreed priorities. Regular statistics have been produced on the numbers of meetings, demonstrations, and so on. Although purely quantitative, these are nevertheless informative.4 AVFA is now attempting to introduce extension objectives in the programs; however, as now planned, this runs the risk of making the system even more top-down, whereas it would gain by becoming more decentralized and participatory. 31. By contrast, evaluation received less attention and resources. Indeed, the monitoring and evaluation unit remained understaffed and underequipped for most of the project's life span. In addition, AVFA lost some valuable personnel to other agencies where work conditions and promotion opportunities were more attractive. AVFA produced guidelines on monitoring and evaluation in 1993 with TA support. Based on these, an survey of the adoption of ten technical themes was carried out in 1994, and an evaluation of intensive extension campaigns was undertaken in 1996, whose results are now being analyzed. Information is being obtained that can, to some extent, be used to re-orient programs, but because of methodological difficulties (such as not using a control sample), not to measure the impact of extension. These surveys are costly. Decentralized, informal surveys are preferable for providing guidance on themes and methods. Qualified external teams should be used to measure impact, and a few cases (regions and themes) selected. The evaluation by consultants of mass communication programs was rated as satisfactory. A wide audience, including up to the highest level in the country, followed those programs found to be effective. 32. Women Extension. The SAR (Annex 1) envisaged a small, pilot extension program for women in two areas to start in year three, once the general extension program was well established. 4 For instance, when it showed that the numbers of visits to demonstration plots was inferior to the number of demonstrations, this raised obvious doubts about the quality of the demonstrations and their usefulness. Indeed, as a result their number has been reduced to emphasize quality 8  Given the importance of women in Tunisian agriculture, such a cautious approach is difficult to understand. AVFA was bolder. Starting in 1993, extension aimed at women initially covered three zones, and was progressively expanded to cover eight zones by 1996. Female staff received the same basic training modules as men, to which was added training in audiovisual communication, in home-based income generating activities (food processing, cottage industries, and so on), and in energy conservation. Unlike for male farmers, programs for women only respond to local priorities determined by the beneficiaries. They tend to focus on fruits, vegetables, livestock, and home industries. Despite the scarce resources allocated for these activities, participation rates are high and the program is considered a success. No formal evaluation has been undertaken, however, nor have any studies been carried out on which to base future strategies. As recruiting female extension workers to work in the field is difficult (of thirty-eight female extension workers only eleven are working in the field), one issue is whether and how their male colleagues could reach women farmers. 33. Extension Strategy. Encouraged by Bank supervision missions since 1992, during the last two project years AVFA has prepared a strategy for extension and training with a team of national consultants. While the strategy amply demonstrates the need for well-trained farmers and correctly underscores the role of training as a complement to extension, it provides insufficient information about how to accomplish this at realistic costs. Such issues as sustainability, cost recovery, and responsiveness to the diversity of production systems and farmers' needs remain to be tackled. 34. Encouraged both by the Bank and by FAO TA, a number of pilot activities were carried out to involve farmers' organizations in the delivery of extension services, particularly through agricultural chambers, cooperatives, and the private sector. These valuable experiments are worth pursuing, but need to be complemented by a well-organized drive to strengthen farmers' organizations, which remain notoriously weak in Tunisia. Research-Extension Linkages 35. Linkages between research and extension were developed during implementation through a series of new mechanisms, including (a) the participation of representatives from CRDAs and extension services in the identification of national priorities, and now also in regional research priorities where PRRs exist; (b) the attendance at research station information days by both farmers and extension workers; (c) the involvement of researchers in programming extension activities at the CRDA level and their contribution in solving problems brought to their attention by extension staff, and (d) the contribution of scientists to the preparation of brochures and other extension materials and to intensive extension campaigns. In addition, some limited on-farm experimentation is being initiated at the PRRs. This was to be developed vigorously at the envisaged regional centers. 36. Despite the above mechanisms, linkages remain loose. On the side of research, adaptive work has received insufficient priority, mainly because the present system of staff evaluation favors academic work. In addition, many scientists are based in Tunis and lack mobility. As concerns extension, the weight of national strategic objectives and insufficient knowledge of local production systems may lead to simply pushing those new technologies that seem to offer less resistance. On-farm testing of new technologies would offer scope for a truly interactive process through which scientists would come to understand farmers' objectives and constraints and could orient their research accordingly. Unless researchers and extension workers are encouraged to work together in the field, linkages may remain theoretical or bureaucratic. The Government's decision 9  to create conmnodity technical centers, starting with one for cereals and one for potatoes, illustrates the need for more concrete collaboration between research, extension, and farmers. Technical Assistance 37. TA for extension was cofinanced by the UNDP and provided by the FAO. It covered extension methodology, monitoring and evaluation, women extension, computer systems, training evaluation, communication techniques and the mass media, and gender analysis for a total of sixty- two man-months. TA was generally found to be satisfactory and useful, with individual variations in quality. It certainly helped to organize staff training and establish methods and procedures for programming and monitoring extension activities. For research, mobilization of TA was slow until 1994 and was essentially limited to supporting the work of the two programming commissions for mechanization and irrigation. In 1994 IRESA and the FAO entered into a contract whereby the FAO would provide TA. However, the national procurement committee rejected the contract in 1995, by which time it was too late to prepare alternative arrangements for the comprehensive TA program envisaged at appraisal (fifty-two man-months of technical TA plus eighteen man-months for management TA). Instead, in 1996/97 limited TA was obtained on programnming and regionalization. Most of the funds originally intended for external training and TA, US$]1.3 million, or 72 percent of the total, were re-allocated toward other categories of expenditures during project implementation. More liberal use of TA for research could have contributed to improving project implementation, for instance supporting restructuring and regionalization of research and assisting the various research commissions. C. MAJOR FACTORS AFFECTING THE PROJECT 38. Budgetary constraints unforeseen at the time of appraisal affected both the research and extension components to varying degrees, especially with regard to recurrent expenditures. Together with the lack of flexibility in administrative procedures, this was a major factor in the slow recruitment of new staff, particularly researchers, and in the slow strengthening of the main research stations. 39. As is now clear, the decision to create PRRs made necessary to establish a consensus in MOA, though not entirely external to the project, affected the regionalization objective, which was expanded rather than changed. In addition to diluting efforts over more sites, the regionalization focused on coordinating existing research rather than on building a credible research capacity at regional levels, and set the whole process onto a different path dependent on lengthy political and administrative decisions. As to AVFA, it struggled during most of the Project with the arrondissement vulgarization problem, the lack of a statute for extension personnel and a controversial extension strategy that has still not been agreed upon and finalized. D. PROJECT SUSTAINABILITY 40. Research investments financed under the project are likely to be sustainable to the extent that programming has now been adopted across the board, that equipment acquired under the project corresponds well to agreed priorities and programs, and that the status of researchers is sufficiently attractive to retain qualified staff. Financially, the setting of clear research priorities, the wide dissemination of results, and the greater involvement of researchers in development 10  activities have all contributed to improve the visibility and credibility of research, which is crucial when budget decisions are made. The benefits are seen in recent increases in investment budget allocations. Between 1988 and 1996, investment budgets allocated to agricultural research institutes increased from DT 580,000 to DT 1.17 million, representing a 28 percent increase in real terms. After correcting for the overall increase in research staff, the trend in real terms is still a 10 percent increase from DT 4,000 per staff member in 1988 to DT 6,900 per staff member in 1996.5 Meanwhile, budgets allocated to recurrent expenditures should have been given more emphasis. In real ternms, they increased by 10 percent between 1991 and 1996; however, per researcher they actually decreased by 9 percent.6 The overall situation of the research system also remains a source of concern insofar as the long awaited restructuring of facilities has not been carried out, because a proper evaluation system has yet to be established. 41. Without a shared vision for the future, the sustainability of the extension system as developed under the project is less clear, because agriculture is becoming ever more global. Modern agriculture, which has benefited most from extension, may be increasingly looking for private, specialized services to become more competitive. The more traditional farmers may expect little from a system that has concerned itself primarily with production increases. However, the managerial and technical capacities developed by the project can only be an asset when extension reforms itself to tackle future challenges. When this is done, the issue of the status of extension staff should not be forgotten. E. THE BANK'S PERFORMANCE 42. In agreement with the government, the FAO/World Bank Cooperative Program helped with project preparation. The Bank's satisfactory performance at identification, preparation, and appraisal ensured a high degree of project ownership, so that reforms in the research and extension sector had already started during project preparation. However, Bank staff did not fully appreciate the difficulties involved in implementing some project components, especially in the case of research, where complex changes were envisaged. A clearer statement of project objectives and their articulation with project components in the SAR would have made it easier to stay the course during implementation. Nine supervision missions were carried out, and a mid-term review took place in April 1994. Continuity was optimum, with the same task manager leading all nine supervision missions. The support these missions provided at every stage of implementation was useful and appreciated. For example, following the creation of SERST, prompt intervention by the Bank in late 1991 averted the dismantling of the newly established IRESA. As concerns extension in particular, continuity and firm, clear-sighted guidance prevented the introduction of pure T&V, which was not appropriate for Tunisian circumstances. Instead the Bank encouraged taking initiatives in new approaches to extension and seeking to involve the private sector and professional organizations. Based on International Monetary Fund and Central Bank of Tunisia data, compounded inflation between 1988 and 1996 was estimated to be 57.5 percent. 6 Based on Central Bank of Tunisia data, compounded inflation between 1991 and 1996 was estimated at 27 percent. 11  F. THE BORROWER'S PERFORMANCE 43. The Borrower's performance was particularly satisfactory during project formulation through appraisal, as a firm will to confront the institutional issues in research and extension was evident. During implementation, the commitment to carry out the activities defined in the SAR never weakened. For research, the decision to defer procurements to ensure consistency with decisions reached by the programming commissions certainly postponed implementation, but was sound and was supported by the Bank. Progress reports were prepared regularly, and audits, which were submitted annually, although with delays, confirmed the confonnity of disbursemnents. Yet despite this commitment, a number of delays have affected implementation, particularly in the area of research decentralization, where the formulation of a new research decentralization policy and related administrative decisions took several years. Regionalization of research was also hampered by parallel delays in the restructuring of the research network. These delays, as well as those introduced by changes in the leadership of both IRESA and AVFA at mid-term, may also have weakened the initial commitment to serious reform. G. ASSESSMENT OF OUTCOME 44. Much has been accomplished in the course of project implementation to modernize and energize research and extension. Research programs are better targeted and more consistent, monitoring and evaluation of programs and staff is being introduced, regional priorities are being identified, higher education institutions contribute actively, and better facilities are available and being used. Extension is well organized, beneficiaries participate more actively in the definition of priorities, staff are much better trained and more mobile and their activities are programmed and monitored, and mass media communications are used effectively. Because of the weakness of the evaluation system, measuring the benefits of the project is not possible. However, examples abound of new technologies that have been or are being successfully transferred to farmers thanks to better organized extension and improved linkages with research, and these technologies are in domains as diverse as cereals, livestock feeding, plant protection, and agricultural mechanization. 45. In sum, progress has been made toward achieving each of the project's original objectives, and such progress is sufficient to justify further assistance for completing the reforms undertaken. As concerns research, the SAR's objectives are now generally seen as too ambitious for a five- year implementation period. Clearly, appraisers had not fully appreciated the complexity of programming and budgeting by objectives, given the lack of experience with introducing these procedures in other countries. As concerns extension, the initial objectives were less ambitious and have been essentially achieved. The reforms needed now were not envisaged during project formulation, because the then prevalent approach to extension in the Bank, which favored a uniform government funded and operated service geared toward technology transfer, coincided largely with the government's own aspirations for extension. H. FUTURE OPERATIONS 46. Tunisia's economy is rapidly opening up, and the Government is also redefining public and private sector responsibilities in the delivery of services to agriculture, thus both extension and research need renewed vision and further reform. The authorities increasingly understand the need for a new strategy for extension, and have directed that farmers' organizations should assume a 12  greater role in providing and financing extension services. A strategy for training farmers, particularly young fdrmers, which the present project did not cover, should be developed. In the area of research, regionalization needs to be vigorously pursued. While IRESA's role in effectively guiding research by allocating resources and evaluating research should be firmly established, regional and international cooperation should be strengthened and network restructuring boldly undertaken. These various challenges would be tackled through a proposed Agricultural Support Services Project, which preparation will be supported by a Technical Assistance grant. I. KEY LESSONS LEARNT 47. The key lessons learnt from project implementation include the following: * Programming and budgeting by objectives is a complex approach that required some three years for IRESA to adopt properly. * Because this approach was a critical factor in the further implementation of the research component, this delay had a negative impact on project implementation. * When agricultural research is to undergo complex restructuring, expecting rapid disbursement from the start of the project is unrealistic. * Given the previous point, when a project design includes a new and complex approach, the local context, capacity, and constraints should be carefully analyzed and assessed. - To achieve regionalization of research, the Project should have placed more emphasis on the necessary prerequisites. In particular, a consensus with other Ministries involved should have been sought to solve such issues as the delegation of financial authority, the appropriate posting of staff, and the development of critical masses of research activities. - Considering the time and energy needed to develop comprehensive systems for programming and budgeting by objectives, simpler procedures, like projets nationaux mobilisateurs, to redirect research budgets effectively during project implementation constitute a good compromise. l Implementation was clearly obstructed by the need for regulatory changes that were not envisaged at appraisal. * The obvious need to up-grade the level of the arrondissement vulgarization was a case in point; never tackled it resulted in extension remaining under the authority of all the other directorates at both the central and the regional levels, thus having the lowest priority. * The need for unanticipated regulatory changes suggests that to facilitate reforms such as those promoted by the Project, more thorough preparation to understand better and to tackle decisionmaking processes and bureaucratic procedures might be envisaged. * The introduction of a national extension system, based solely on the T&V system proved inadequate in the Tunisian context. * Modifying the approach to extension, especially at the national level, requires having a clear vision and defining a strategy for extension at an early stage, together with intense and effective participation by farmers. 13  * When elaborating a strategy, decisionmakers must decide whether extension should remain the sole responsibility of the public sector or whether it should rely on a mix of public and private interventions, the level of producers' contributions to the costs of extension, and the amount of funds the government is ready to commit to extension on a regular and long-term basis. * To improve the motivation of field staff, project management should give serious consideration to preparing proposals for legislation that would upgrade the status of extension staff to be submitted to decisionmakers for their consideration. 14  PART II: STATISTICAL ANNEXES Table 1: Summary of Assessments A. Achievement of objectives Substantial Partial Negligible Not Applicable Macro policies Sector policies Financial objectives [I Institutional development Physical objectives Poverty reduction Gender issues ] L Ir Other social objectives l I Environmental objectives _ _ _ Public sector management L / I I I Private sector development Il I Other (specify) [ / B. Project sustainability Likelv Unlikelv Uncertain (9') (s/) (/) 9' W Lcell  H~i&hli C. Bank performance satisfactory Satisfactorv Deficient (/) (/) (il) Identification l l Preparation assistance [ill] Appraisal [Il] Supervision [l D. Borrower performance satisfactorv Satisfactory Deficient (1) (/) (/) Preparation _ Implementation Covenant compliance Operation (if applicable) l l l___r llighlY Hi E. Assessment of outcome satisfactorv Satisfactory Unsatisfactorv unsatisfactory w()() / /  Table 2: Related Bank Loans Loan/credit title Purpose Year of Status l_____________________ approval 2. Second Forestry Sustainable development and 1993 Ongoing Development Loan management of the Tunisian forests (Loan 3601-TUN) with an active participation of forestry populations, and to improve the cost effectiveness of a forest resource l___________________ development program. l 2. Agricultural Sector Promotion of sectoral growth through 1993 Ongoing Investment Loan improved management of public (Loan 3661-TUN) resources, sub-sector policy reforms and support of the public investment budget. 3. Northwest Mountains Reduce poverty, diminish natural 1993 Ongoing Areas Development resources degradation and improve the Project population's well-being by providing (Loan 3691-TUN) for some basic needs in NW Tunisia. 4. Rural Roads Project Reduce disparities between urban and 1995 Ongoing (Loan 3840-TUN) rural areas, reduce rural migration, and improve the delivery of social services to poorer population, distribution of manufactured goods at affordable prices, facilitate construction of other rural infrastructure and permit further social and economic development. Table 3: Project Timetable Steps in project cycle Date planned Date actual/Latest estimate Identification 17/7/87 17/7/87 Preparation 23/8/88 23/8/88 Appraisal 23/9/88 23/9/88 Negotiations 2/4/90 2/4/90 Board presentation 5/6/90 5/6/90 Signing 24/9/90 24/9/90 Effectiveness 21/5/91 21/5/91 Mid-term review Project completion 1/4/96 1/4/96 Loan closing 30/6/96 30/6/97  Table 4: Loan Disbursements: Cumulative Estimated and Actual (US$ million) FY90 FY91 FY92 FY93 FY94 FY95 FY96(a) FY97(a) Appraisal estimate 0.8 3.1 6.5 10.0 13.5 16.2 17.0 Actual (a) 0.0 0.8 2.9 5.4 8.2 10.3 16.5 Actual as % of estimate (a) 0% 25% 44% 54% 61% 64% 97% 97% Actual (b) 0.0 1.3 3.1 6.0 8.2 10.4 14.0 14.0 Actual as of % estimate (b) 0% 42% 48% 60% 61% 64% 82% 82% Date of final disbursement: 31 October 1997 Last cancellation: Extended to 30 June 1997 (from 30 June 1996) (a) Based on aggregate data provided by the two PMUs. All disbursements pending as per May 15, 1997, including these related to bids in process, were included in the first semester of 1997 (FY 96). The exact situation will need to be assessed after June 30, 1997. (b) Based on Bank figures, as per June 1997, which do not include bids in process.  Table 5: Key indicators for Project Implementation: Research Component Indicators Unit Estimated Actual Actual as% l l l | ~~~~~~~~~of Estimate A. Research Component Civil Works - Laboratories m2 1,235} 3,523 99% - Housing / Offices m2 2,310} - Others m2 3,030 1,927 64% - Rehabilitation m2 2,060 4,681 227% Vehicles - Light cars unit 26 26 100% - Long distance vehicles unit 8 0 0% - Pick-ups unit 19 30 158% - Minibus unit 1 1 100% - Trucks unit 2 0 0% - Four Wheel-Drives unit 11 0 0% Computers - PCs and printers unit 21 25 119% - Mainframe for IRESA unit - 1 Technical Assistance - Research Training man-months 22.5 n.a - Farm Machinery man-months 30 n.a - Research Budgeting man-months 12 n.a - Financial Management man-months 6 n.a Training man-months 243 57 23% Research Staff Recruitment Nb. 42 47 112% Research Programming - Programmes and Draft Budgets Nb. 9 10 111% Prepared Establishment of Regional Research Nb. 3 3 100% Centres (a) (a) The three Regional Research Centers were replaced by seven PRRs, of which three are operating.  Table 5: Key indicators for Project Implementation: Extension Component Indicators Unit Estimated Actual Actual as% l l l 0 of Estimate B. Extension Component Civil Works - Constructions (CTV and CRA) Nb. 54 40 74% - Rehabilitation (CTV and CRA) Nb. 30 60 200% Vehicles - Long distance (A19) unit 10 15 150% -Light vehicles unit 103 103 100% - Pick-ups unit 45 59 131% - Four Wheel-Drives unit 17 8 47% - Minibus unit 9 11 122% - Motorcycles unit 141 151 107% Mobile Equipment unit 175 (a) Audio-Visual Equipment Nb. 1323 (a) Demonstration Equipment unit 408 (a) Computer Equipment - PCs unit 20 29 145% Photocopying machines unit 5 23 460% Technical Assistance - Operations co-financed with UNDP (b) Nb. 5 5 100% (a) The acquisition of audio-visual, movable and demonstration equipment was originally planned in sets. However, during project implementation, needs became clearer and were expressed in units, following the requests of the CRDAs. Obviously, the number of units cannot be compared to the original number of sets. In terms of value, the actual amount spent on these three categories is 111% compared to the original proposal. (b) These operations correspond to activities that came under the US$ 1 million UNDP TA grant, co- financed by the Bank.  Table 6: Key Indicators for Project Operation No key indicators for project activities were defined in the SAR. The future development of evaluation systems for both the research and the extension components should allow to derive a set of significative indicators.  Table 7: Studies Included in Project Study Purpose as defined at appraisal! Status Impact of Study redefined l A. Research Component 1. Accounting System (SIMAC) Setting-up of a new analytical Started in April 94, to be Not yet implemented accounting and budgeting system completed in June 1997 at INRAT. 2. Restructuring of the Network Take stock of existing situation Started in November 1994, Restructuring not yet of Research Stations (SCET) and propose streamlining and completed in June 1996 implemented, results reorganizing of research stations will be used for the IXth Plan 3. Assessment of Scientific Cf. Title Started in April 1995, Help r ationalize the Equipments owned by the completed in May 1995 purchase of equipment Research Institutes under IRESA for the last phase of the (ORSTOM) project 4. Assistance to the Help commissions to prioritize Started in November 1996, Objective met Programming of Agricultural research objectives completed in May 1997 Research (ISNAR) 5. Regionalisation of Propose options for the Started May 1997 Not yet completed Agricultural Research (INRA- regionalisation of agricultural France) research activities, in the context of the new regional "poles" 6. Evaluation of Research Provide methodology and tools to Started in January 1997, Results just being Programmes (PROMOVIENDO) set-up a system of evaluation of completed in March 1997 exploited, supervision research programmes, and evaluation forms laboratories and individual developed by IRESA's researchers evaluation Unit B. Extension Component (a) I. Credit and Saving Schemes Define activities for the pilot Completed in 1995 Orientation of the pilot for Rural Women extension program for women extension program for women 2. Training Needs of Rural Id. Id. Id. Women 3. Impact of Mechanization and Id. Id. Id. Other Technologies on Rural Women 4. Practical Guides for the Strengthen the methodology and Studies carried out from Objective met Planning/Supervision/Evaluation organization of the 1992 to 1994 of Extension Activities Planning/Supervisionl/Evaluation Unit of AVFA 5. Strategy for the Agricultural Serve as basis for national Started in 1995 Not yet implemented Extension Sector strategy of agricultural extension for the IXth plan (a) For the extension component, studies were mostly financed by UNDP  (page 1) Table 8A: Project Costs L Appraisal Estimate (US $'000) (a) | Actual/Latest Estimate (US$'000s) (b) Item Local Costs Foreign Total Local Foreign Total l Costs l Costs (c) Costs (c) A. Research Investment Costs Civil works 1,025 872 1,897 n.a 2,490 Equipment and Vehicles 2,953 4,617 7,569 n.a. 8,060 Training and Technical Assistance 468 2,412 2,880 n.a. 650 Subtotal Investment Costs 4,446 7,901 12,346 n.a. 11,201 Subtotal Operating Costs (d) 3,734 2,371 6,105 n.a. n.a. Subtotal Research 8,179 10,272 18,451 n.a. n.a. B. Extension Investment Costs Civil Works 1,472 1,253 2,725 n.a. 2,265 Equipment and Vehicles 2038.7 2299.9 4,339 n.a. 5,727 Training and Technical Assistance (e) 510.4 1189.3 1,700 n.a. 1,742 Subtotal Investmnent Costs 4,021 4,742 8,764 n. a. 9,734 Subtotal Operating Costs (d) 3973.5 2783.3 6756.8 n.a. n.a. Subtotal Extension 7,995 7,525 15,520 n.a. n.a C. Project as a Whole Investment Costs Civil works 2,497 2,125 4,622 n.a. 4,755 Equipment 4,991 6,916 11,908 n.a. 13,787 Training 979 3,601 4,580 n.a. 2,393 Subtotal Investnent Costs 8,467 12,643 21,110 n.a. 20,935 Subtotal Operating Costs (d) 7,707 5,155 12,862 n.a. n.a Total 16 174 17,798 33,972 n.a. n.a. (a) Appraisal estimates include contingencies, as indicated in Implementation Report dated December 1990. (b) Latest estimate is bassed on data provided by AVFA and IRESA on 8/12/97 further to the audit of project accounts. Tunisian dinars were converted into dollars at a rate of D 1 for US$ 1. (c) The distinction between local and foreign costs was not monitored. (d) Project operating costs were not monitored. (e) UNDP's contribution to the Training and Technical Assistance expenses of the extension component was estimated at US$ I million.  (page 2) Table 8B: Project financing Appraisal Estimate (IJS $'000) (a) Actual/Latest Estimate (US$'000s) (b) Item Local Costs Foreign Total Local Foreign Total Costs Costs (c) Costs (c) A. Research World Bank 0 7,600 7,600 n.a. 9,317 Government (c) 3,800 0 3,800 n.a. 1,884 Taxes and Duties 4,400 0 4,400 n.a. Subtotal Research 8,200 7,600 15,800 11,201 B. Extension World Bank 0 6,400 6,400 n.a. 6,007 Government (c) 3,500 0 3,500 n.a. 2,728 Taxes and Duties 4,500 0 4,500 n.a. Cofinancing (UNDP) (d) 0 0 0 n.a. 1,000 Subtotal Extension 8,000 6,400 14,400 9,734 C. Unallocated World Bank 0 3,000 3,000 n.a. 0 Governmuent 0 800 800 n.a. 0 Subtotal Unallocated 0 3,800 3,800 0 D. Project as a Whole World Bank 0 17,000 17,000 n.a. 15,323 Government (c) 7,300 800 8,100 n.a. 4,612 Taxes and Duties 8,900 0 8,900 n.a. cofinancing (UNDP) (d) 0 0 0 n.a. 1,000 Total 16,200 17,800 34,000 n.a. 20,935 (a) The breakdown between the two components (research and extension) was based on (i) the loan agreement for the World Bank financing, and (ii) the costs tables of the implementation volume for taxes and duties and local costs financed by the Government. Foreign costs to be financed by the Government (US$ 800,000) were considered unallocated. (b) The distinction between local and foreign costs was not monitored. Latest estimate is base on data provided by AVFA and IRESA on 8/12/97 further to the audit of project accounts. (c) For actual financing, project operating costs are not included in what is referred to as Government, as these were not monitored. Hence a major discrepancy between total estimated financing at appraisal and the latest estimate of actual financing shown in this table. (d) An additional US$ 1 million grant, financed by UNDP, was identified after project appraisal. Its proceeds were used to finance part of the technical assistance and training expenses of the extension component. The reallocation of World Bank funds for the extension component further to the identification of the UNDP grant is shown in Table 1, Annex B.  Table 9: Economic Costs and Benefits The Economic Rate of Return (ERR) was not estimated at appraisal, nor were the benefits of the object quantified in monetary terms. To date, the evaluation systems of both the research and the extension components are not yet operational, making any tentative to quantify project benefits quite difficult.  Table 10: Status of Legal Covenants Agreement Loan/ Text Reference Covenant Original Revised Description of Covenant Comments Credit Classes States (b) Fulfillment Date Fulfilbment Date Number (a) Loan 3217 TUN 4.01 1 C Borrowerto have accounts for each fiscal year audited With delays Sch 2 Part A a (i) 5 C 30/7/91 Borrower to establish IRESA Sch 2 Part A a (ii) 5 CP . Set-up three multidisciplinary research centres Poles established instead of centres Sch 2 Part A a (iii) 5 NC Set-up a national research and extension council Council of IRESA instead Sch 2 Part A b (i) 5 CP . Develop a national system for agriculture research prograrning and budgeting by On-going objectives Sch 2 Part A b(ii) 5 NC - Establish procedures for improving the fnancial management ofresearch activities Reluctance to through the transfer of authorities to regional entities transfer authority Sch2Part B 1 5 C 31/12/90 31/7/91 Borrowerto establish AVFA Sch5 A2 5 C 31/12/90 Borrowerto establish and maintainaProject Coordination Conmuittee Sch5A3a 5 CD 31/12/90 BorrowertoestablishProgramnmeCommissions Sch 5 A3 b 5 CD, CP 31/12/93 Borrower to establish working group to each ofthe three multi-disciplinary regional On-going (one is research centres operational) Sch 5 A 4 5 C every 31/8 Borrower to prepare and submit each year proposed work programme and corresponding budget Sch 5 A 5 9 C 31/12/93 4/94 Borrower and Bank jointly cany out Mid-Term Review Sch5 A 2 5 CD 30/6/91 Borrower to cause programmes and budgets to be prepared for: animal husbandry and farm mechanization 30/6/92 for vegetables, citrus, other fuit, and olives 30/6/93 for irrigation, forestuy, soil conservation, and arid zone farming. Sch 5 B 3 5 NC 1/1/92 Borrower to ensure that policies and procedures to Improve the financial Few progress due management of research activities are applied to lack of harmonization with MOA and MOF Sch 5 B 4 5 SOON 31/12/91 Borrower to submit plan of actions for review ofthe number of agricutural research No decision made stations yet Sch5 C 1 5 C 31/12/90 31/7/91 Borrowertoestablishacentralextensionunit Sch 5 C 3 9 C Borrower to establish programming and monitoring units in each ofthe 15 CRDAs Sch 5 C 4 5 C 31/12/90 Key stafffor progmming and traming to be remuited at Ctral Unit (a) Covenant classes are: 1 - Accounts/audit, 2- Financial performance/generate revenues from beneficiaries, 3- Flow and utilization of proJect funds, 4- Counterpart fsndi-Ing 5- Management aspects of the pfoject or of its executing agency, 6- Environmental covenants, 7- Involuntary resettlements, 8- Indigenous people, 9- Monitoring, review and reporting, 10- Implementation, 11- Sectoral or cross- sectoral budgetary or other resource allocation, 12- Sectoral or cross- sectoral regulatory/institutional action, 13- Others. (b) Status: C- Complied with, CD- Compliance after Delay, NC- Not Complied with, SOON- Compliance Expected in Reasonable Short Time, CP- Complied with Partially, NYD - Not Yet Due.  Table 11: Compliance with Operational Manual Statements Statement Number and Title I Describe and Comment on Lack of Compliances There was no major deviance from Bank Operational Manual Statements. Table 12: Bank Resources: Staff Inputs Stage of Project Cycle Planned Revised | Actual Weeks TUS$ 'OOOs Weeks I US$ '000s Weeks | US$'000s Through appraisal 25.9 66.7 25.9 66.7 25.9 66.7 Appraisal-Board 60.0 145.4 60.0 145.4 60.0 145.4 Board-Effectiveness - - - - - - Supervision 122.1 408.9 122.1 408.9 118.7 390.6 Completion 7.0 28.5 7.0 28.5 4 15.6 Total 215.0 649.5 215.0 649.5 208.6 618.3  Table 13: Bank Resources: Mission Stage of Project Month/Year Number of Days Specialized Performance Rating Cycle Persons in Staff Skills Types of Field Represented Problems Implementation Development Status O;bjectives Identification Preparation 7/7/87-23/8/88 Appraisal Supervision 1 Apr-May-91 1 13 A S S 2 Oct-91 2 14 A, R HS S 3 Jul-92 3 13 A, R, F HS S 4 Mar-Apr-93 3 10 A, R, F S S 5 Sep-93 3 10 A, R, F n.a. n.a. 6 Apr-94 4 18 A, R, F, M&E HS S 7 Oct-94 2 10 A, Ec S S 8 Jun-95 2 14 A, Ec S S 9 Oct-96 3 12 A, Ec, F S S ICR (a) May-97 2 12 A, Ec Skills of specialized staff: A= Agronomics, Ec=Economics, F=Financial Analyst, R=Research, M&E= Monitoring Types of problems: Rating: HS= Highly Satisfactozy, S=Satisfactory (a) Carried out by FAO/CP  TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT APPENDIX A AIDE-MEMOIRE  REPUBLIQUE TUNISIENNE PROJET DE RECHERCHE ET DE VULGARISATION AGRICOLE (PRV) PRET 3217 - TUN Mission du Programme de Cooperation FAO/Banque mondiale Pr6paration du rapport d'ex6cution du projet Aide-m6moire I. INTRODUCTION 1. A la demande du Gouvernement Tunisien et de la Banque mondiale, une mission du PC- FAO/BM composee de MM. F. Dauphin (agronome, chef de mission) et E. Hidier (economiste) a sejourn6 en Tunisie du 12 au 23 Mai 1997, afin d'initier la preparation du rapport d'achevement du projet. La mission, dont le travail a ete efficacement organise par l'Institut de la Recherche et de l'Enseignement Sup6rieur Agricoles (IRESA) et par I'Agence pour la Vulgarisation et la Formation Agricole (AVFA), a travaille avec des responsables du Ministere de I'Agriculture, du Secr6tariat a la Recherche Scientifique et Technique, et le personnel des principaux departements de l'IRESA et de 1'AVFA, et a rencontre les Directeurs de l'lNAT, de l'INRAT, de l'INRGREF, ainsi que certains des chefs de programme et coordinateurs de poles. Des visites effectu6es dans les regions d'Ariana et de Sousse ont pernis de rencontrer des responsables de la vulgarisation aux niveaux CRDA, CTV et CRA et des agriculteurs. La mission a pu egalement visiter les stations de Oued Beja, d'El Afareg, et du Kef, ainsi que l'Ecole Superieure d'Agriculture du Kef. Le representant de la FAO en Tunisie a ete tenu inform6 des travaux de la mission et I'a conseille efficacement. La mission tient A remercier les Directions de L'IRESA et, de I'AVFA pour la fructueuse collaboration dont elle a beneficie durant son s6jour. 1I. LES OBJECTIFS DU PROJET 2. Le projet se proposait essentiellement d'ameliorer l'organisation et le fonctionnement de la recherche et de la vulgarisation agricoles afin de fournir les bases techniques et scientifiques permettant au secteur agricole de remplir les objectifs du 8'me Plan, essentiellement en matiere de production agricole. Les efforts prevus devaient porter en priorite sur les principales zones de production en pluvial et en irrigue ; en particulier les investissements en matiere de vulgarisation au niveau local ne devaient porter que sur 15 des 23 Gouvemeras. 3. En matiere de recherche, il s'agissait: (1) d'orienter les travaux vers les priorites du developpement par la mise en place de structures chargees de definir les priorites, d'etablir et de suivre les programmes, et par le renforcement de la liaison avec la vulgarisation ; (2) d'ameliorer la coherence des recherches et la gestion des ressources humaines, physiques et financieres par la mise en place d'une programmation et budgetisation par objectifs, la mise en place d'un systeme de suivi et d'evaluation, et par la restructuration du reseau des stations; (3) d'ameliorer l'adaptation des recherches aux conditions locales par la regionalisation, le developpement de recherches en milieu reel et le renforcement des etudes socio-economiques; et (4) d'ameliorer la qualite des travaux par le developpement et la meilleure utilisation des ressources humaines, 1'etablissement d'un systeme d'evaluation des chercheurs, l'amelioration de l'infrastructure et des .6quipements, et une meilleure gestion des stations. En outre, le projet prevoyait d'appuyer l'ex6cution de  programmes de recherche prioritaires autour de 10 themes definis lors de la preparation, et portant sur le developpement de productions strategiques (cer6ales et l6gunineuses alimentaires, lait et viande, maraichage et arboriculture. 4. En matiere de vulgarisation, il s'agissait de facon similaire d'ameliorer la gestion des ressources et la qualite de la vulgarisation. Pour eviter la dispersion des efforts, il fallait en particulier r6unir l'ensemble des systemes de vulgarisation sous une seule agence charg6e de coordonner les actions, de mettre en place la programmation, le suivi et I'evaluation des activites, d'organiser la mise a niveau et la formation continue des agents, et de mettre en place wue methode de vulgarisation utilisant les principes de I'approche 'formation - visites'. Le projet pr6voyait en outre d'augmenter la portee de la vulgarisation par la fourniture de logements et de vehicules et par le developpement de la vulgarisation de masse. L'amelioration de la qualite devait etre obtenue par un important effort de formation, compte tenu d'un mveau initial generalement faible des agents de vulgarisation, ainsi que le renforcement des liaisons avec la recherche au niveau national et regional. 5. Les propositions du projet en matiere de recherche etaient fondees sur une analyse detaillde de l'organisation et du fonctionnement du systeme, ainsi que des besoins nationaux de recherche, menee de 1985 a 1987 par une equipe nationale appuyee par l'ISNAR. L'analyse du systeme de vulgarisation fut beaucoup plus sommaire. Dans les deux cas cependant, les objectifs institutionnels et les investissements prevus etaient de nature a ameliorer de facon tres sensible le fonctionnement et la productivite des deux systemes, et il etait l6gitime d'esperer im impact significatif en matiere de production agricole. III. DEROULEMENT ET RESULTATS DE L'EXECUTION A. Generalit&s 6. La mise en couvre du projet, si elle a connu certains retards, s'est deroulee de facon relativement r6guli6re et satisfaisante. Les retards ont porte essentiellement sur les recrutements, en particulier dans les nouveaux organismes, IRESA et AVFA, ce qui a ralenti l'ensemble de l'execution. Cependant, des la mi-parcours, des progres importants etaient notes en matiere de renforcement des deux nouvelles structures, de programmation de la recherche, de synthese des resultats acquis, de mise en place des p8les r6gionaux; et pour la vulgarisation, de formnation du personnel, de reorientation et de dynamisation du dispositif regional et local, de vulgarisation f6minine et de liaison avec la recherche. Depuis, les retards ont e combles tant en matiere de recrutements que de construction et d'equipements et, moyennant une extension d'un an de la durde d'execution, les objectifs physiques du projet sont pour l'essentiel atteints. En outre les institutions prevues ont ete mises en place et ont deja fait la preuve de leur utifite, meme si de nombreux problemes institutionnels, dont celui du statut du vulgarisateur, restent a regler. 7. Parmi les facteurs imprevus qui ont handicap6 la r6alisation du projet, il faut noter les restrictions budgetaires, et surtout les difficultes de recrutement de personnel nouveau, parfois meme de remplacement de personnel existant. D'autres facteurs, non totalement ext6rieurs au projet, ont cependant joue, comme l'elargissement du projet de regionalisation qui est passe, de trois implantations a sept.  B. Recherche Creation de l'IRESA 8. L'IRESA, cre des le demarrage du projet, a beneficie largement de du soutien de ce demier et de l'appui des missions de concertation, qui ont en particulier permis l'affectation du personnel requis. Le r6le de l'IRESA a ete d6cisif dans la mise en route du processus de programmation et dans la creation des poles regionaux de recherche. L'IRESA a aussi permis le rapprochement statutaire et fonctionnel de la recherche et de l'enseignement sup6rieur, permettant ainsi la mobilisation d'un important potentiel de recherche et de connaissances. Si son existence a ete un moment remise en cause par la creation du Secretariat a la recherche scientifique et technique (SERST), la collaboration entre les deux institutions est maintenant exemplaire. Son efficacite est cependant linitee par l'absence d'une ligne budgetaire specifique pour la recherche. Programmation 9. Les commissions, formees en 1992, ont realise une synthese des acquis qui, confrontee a un inventaire des besoins, a abouti a l'elaboration de programmes couvrant dix secteurs prioritaires, recouvrant grosso modo les themes retenus au moment de l'valuation, les themes horizontaux ayant e integres dans les differents secteurs. Les ressources additionnelles ayant pu etre degagees pour financer les actions nouvelles proviennent surtout du SERST. Leur attribution a donne lieu a une seconde programmation, beaucoup plus pr6cise, et favorisant des projets multidisciplinaires et pluri-institutionnels. I1 est a noter que seuls ces projets font l'objet d'un suivi- evaluation systematique. Pour le deuxieme cycle de programnmation, engage pour le 9efe Plan, et qui a b6neficie d'un appui de l'ISNAR, les priorites sont hierarchisees et tiennent compte des ressources humaines disponibles ou envisageables de faon r6aliste. Le processus s'est donc deja bien ameliore, et cette programnmation constitue peut-ere le benefice le plus significatif du projet. Budg6tisation par objectif 10. La budgetisation par objectifs avait pour but de permettre non seulement de quantifier les besoins budg6taires relatifs aux priorit6s identifiees mais aussi de reorienter les moyens financiers mis a la disposition de la recherche en fonction de ces priorites. Le travail considerable des commissions a pennis d'identifier et de hierarchiser un certain nombre d'objectifs, en concertation avec un nombre important d'acteurs interesses a differents titres par la recherche agricole. Par ailleurs, les moyens humains et financiers necessaires A la poursuite de ses objectifs ont ete largement clarifies. Par contre, il semble que la reorientation effective des budgets en fonction des priorites 6tablies par les Commissions soit pour l'instant assez limit6e. De fait, le ballon d'oxygene constitue par les fonds supplementaires alloues par le SERST comme contrepartie locale du projet (i.e. les PNM "BIRD", environ 300,000 TD par an) a permis de privilegier certaine priorites. 11. Il est difficile d'estimer a quel point les budgets des etablissements de recherche, hors PNM, ont integre les nouvelles priorites. Il semble que la plupart des programmes preexistants ont e repris dans les nouveaux objectifs et ont continue de beneficier de budgets similaires. Ceci resulte sans doute d'une volonte initiale de ne pas brusquer les acteurs en place lors d'une phase qui s'est voulu la plus consensuelle possible. Mais en l'absence d'un systeme de responsabilisation et d'evaluation individuelle ou collective des acteurs en presence (chercheurs, chefs de labo, chefs de station, etc.) par rapport aux objectifs qui leur sont proposes, une reorientation dynamique de la recherche en fonction de criteres exterieurs restera difficile. De ce point de vue, les objectifs du  projet ne sont aujourd'hui que tres partiellement atteints. Cependant de nombreuses actions sont en cours, en particulier la reconnaissance des laboratoires et des stations en tant qu'entites fonctionnelles pouvant b6neficier d'une certaine autonomie de gestion, la mise en place d'un systeme de suivi et d'evaluation generalise des programmes de recherche au sein de l'IRESA. D'autres devront etre initiees, telie la prise en compte des activites de d6veloppernent dans la gestion des carrieres des chercheurs. Regionalisation 12. Le choix initial de concentrer les efforts sur trois centres regionaux qui seraient dotes de moyens significatifs de recherche, a e reniis en cause par l'adoption en 1992 d'une politique plus ambitieuse portant sur la cr6ation de sept P61es r6gionaux de recherche, structures tres legeres et s'appuyant sur les differentes instituts, stations et ecoles existant dans la region. II est encore beaucoup trop tot pour faire le bilan de ces p6les, puisque deux seulement sont op6rationnels. Leur merite essentiel sera de concretiser sur le terrain la collaboration inter-institutions et le rapprochement recherche-vulgarisation-utilisateurs, ce qui devrait deboucher sur la definition de programmes regionaux de recherche. Pour autant, ces rapprochements ne seront fructueux que dans la mesure ofi la recherche regionale dispose de moyens suffisants. Or it faut noter que l'objectif de placer plus de chercheurs sur le terrain n'a pas e atteint, et que les stations de recherche ne disposent souvent ni des moyens ni de la souplesse de gestion necessaires pour fonctionner de facon satisfaisante. Si des ressources financieres supplementaires peuvent etre trouv6es localement aupres de differents partenaires (CRDA, offices, projets), la paucite de chercheurs affect6s dans les regions est un probleme majeur. Restructuration 13. Le rapport d'6valuation envisageait la fermeture avant fin 1992 d'environ la moitie des quelque 50 stations gerees par les differents instituts. Une etude, conduite en 1996, confirmait la n6cessit6 et l'urgence de cette restructuration, puisque les ressources des instituts ne leur permettent pas de les entretenir. Certaines des stations qui ne sont pas utilisees pour la recherche g6n6rent des revenus importants pour les instituts, et l'on comprend que le besoin d'assurer des ressources de remplacement retarde le processus. Pour autant, cet objectif ne semble pas avoir ete suivi avec toute la determination necessaire. C. Vulgarisation Mise en place de I'AVFA 14. L'agence, egalement mise en place avant le demarrage du projet, est maintenant bien fonctionnelle. Elle a permis la fusion reussie des differents systemes de vulgarisation, la formation des vulgarisateurs, le rapprochement d'avec les institutions de formation, l'instauration d'une veritable progranmmation et du suivi des actions sur le terrain, la production d'un support audiovisuel non n6gligeable, et l'amorce d'une grande reflexion sur I'avenir de la vulgarisation et de la formation.  Formation 15. Avec l'appui du programme PNUD-FAO, un effort considerable de formation a e consenti, qui a permis la mise a niveau de l'ensemble des vulgarisateurs tant sur le plan des methodes que sur le plan technique. Les evaluations faites montrent qu'ils sont a present beaucoup mieux armes. La mise a niveau etant terminee, des recyclages periodiques devraient etre prevus, pour completer la formation continue donnee au niveau des CRDA par les specialistes-matiere. Programmation, suivi, evaluation 16. La programmation mise en place par l'agence combine un processus descendant, refletant les priorites nationales, avec un processus participatif, par lequel les groupes-cibles definissent leurs priorites. L'association a ce mecanisme complexe des specialistes-matiere et de chercheurs aboutit a definir un programme a chaque niveau (CRA, CTV, CRDA) qui concilie des exigences pourtant trks diverses. Cette programmation, d'abord orientee vers les activit6s, tente maintenant de prendre en compte des objectifs de vulgarisation. Combinee a un suivi rigoureux des activites, elle a introduit discipline et concentration des efforts de vulgarisation vers les activites jugees prioritaires. Le systeme, bien qu'efficace, parait pouvoir etre allege et decentralise. Des evaluations ont e realisees sur la vulgarisation de masse (radio, television), et sur un certain nombre de themes techniques. Elles ne permettent pas encore de se faire une idWe, meme tres partielle, de l'impact de la vulgarisation. De nouvelles enquetes sont en cours sur les campagnes intensives de vulgarisation (CM), qui devraient mieux cemer les conditions de l'adoption. Les ressources disponibles pour le suivi-evaluation sont tres insuffisantes, des intervenants exterieurs et de nouvelles techniques seront n6cessaires pour mesurer l'impact et assurer un feed-back exploitable par la vulgarisation et la recherche. Liaison recherche vulgarisation 17. La liaison recherche vulgarisation s'est am6lioree de facon tres sensible par une serie de mecanismes; il faut en particulier citer: la participation de representants des CRDA et de la vulgarisation a la definition des programmes de recherche au niveau national et, avec la mise en place des p6les, au niveau regional; la tenue de journees d'information sur les stations de recherche, 6galement favorisee par les p6les; la participation de chercheurs a la progranumation de la vulgarisation au niveau des CRDA, et leur appui ponctuel pour resoudre les problemes qui leur sont pos6s par la vulgarisation, leur r8le enfin dans la preparation de brochures et autres supports de vulgarisation. La preparation des CIV et leur suivi sont egalement l'occasion d'une collaboration fiuctueuse. Enfin, Ia ou existent des p8les, comme la oii des offices l'ont demande, se developpe une activite, encore tres modeste mais cruciale, d'experimentation en milieu reel. II est sans doute dommage que ces operations n'aient pas ete d6veloppees plus vigoureusement, comme prevu a l'origine. 18. Malgre ces mecanismes varies, force est de constater que les liaisons sont encore laches. Du c6t6 de la recherche, la recherche adaptative n'a pas 6te favorisee autant que souhaitable (le systeme actuel d'evaluation tire la recherche vers le laboratoire) et l'exp6rimentation en milieu reel demande des ressources supplementaires; du c6te de la vulgarisation, la prise en compte prioritaire des objectifs nationaux strategiques et l'absence d'une strat6gie fondee sur une connaissance fine des systemes de production a conduit a se limiter a choisir, parmi les themes proposes par la recherche, ceux qui sont le plus facilement acceptes par les groupes cibles. La decision de cr6er des  Centres Techniques, en particulier pour la pomme de terre et les cereales, traduit bien la necessite de liens plus concrets entre recherche, vulgarisation et utilisateurs. IV. OPERATIONS FUTURES 19. Au total, le projet a permis des avancees tres significatives en matiere de recherche et de vulgarisation. Les evolutions envisagees ont ete toutefois souvent plus longues a metlre en ceuvre que prevu, et ne sont encore que partiellement realisees. C'est vrai notamment de la regionalisation de la recherche, qui devra s'accompagner de la constitution d'un petit nombre de centres d'experimentation performants, donc dotes et de moyens, et d'une autonomie de gestion. La programmation a bien progresse, mais doit etre poursuivie par la mise en place de la budgetisation par objectif, le suivi et l'evaluation des programmes et des projets. 20. Par ailleurs, les circonstances ont change depuis le debut du projet, avec la liberalisation de l'6conomie et la volont6 de redefinir le role de l'Etat et de responsabiliser la pro:fession. Les objectifs assign6s initialement a la vulgarisation doivent etre revus en fonction d'une strat6gie qui reste a pr6ciser. Ceci passe par la reconnaissance (1) de la multiplicite des types d'exploitation, donc de la n6cessite d'une meilleure connaissance de leurs objectifs et strategies; (2) de la multiplicit6 des canaux d'information, dont la vulgarisation gen6rale et rapprochee represente sans doute la forme la plus cofiteuse, qu'il faut n'utiliser qu'a bon escient. Des experiences de prise en charge d'activites de vulgarisation par la profession sont prevues. 11 faudra trouver les moyens pour les elargir et les multiplier. V. SUITE A DONNER 21. Dans ces differents domaines, des investissements significatifs seront necessaires, qui peuvent justifier une nouvelle aide exterieure. Cependant, la reflexion doit etre au prealable approfondie pour s'assurer que les difficultes qui n'ont pu etre surmontees pendant 1'execution du pr6sent projet, bien que parfois identifiees lors de la preparation, soient resolument abordees. 22. De retour a Rome, la mission elaborera une version provisoire du rapport d'achevement du projet qui pourrait etre envoye a Washington vers la fin Juin. Ce rapport suivra le format habituel, et comprendra notamment un chapitre sur les lecons a tirer du projet. De leur c6te, AFVA et IRESA devront egalement envoyer leurs rapports d'achevement.  TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT APPENDIX B RE-ALLOCATION OF LOAN TRANCHES  Table 1: Re-Programming of Loan Tranches and Allotment of Unallocated Tranche Planned after Re- Planned in Loan Agreement programming and Allotment Difference Main Categories of Unallocated Tranche % of Total excl. % of Total % of (in US$ 000's) Unallocated (in US$ OO's) (in US$ 000's' Total A. Research Civil Works 1,000 1,600 600 Equipment 4,800 7,550 2,750 TA&Training 1,800 500 -1,300 Total Research 7,600 54% 9,650 57% 2,050 68% B. Extension Civil Works 1,000 1,800 800 Equipment 3,500 4,650 1,150 TA&Training 1,900 900 -1,000 Total Extension 6,400 46% 7,350 43% 950 32% C. Unallocated 3,000 0 Total 17,000 17,000 Planned after Re- Planned in Loan Agreement programming and Difference Main Categories Allotment of Unallocated Tranche % of Total excl. % of Total % of (in US$ 000's) %Unallocated (in US$ 000's) (in US$ 000's) Total I- Civil Works Research 1,000 1,600 600 Extension 1,000 1,800 800 Total Civil Works 2,000 14% 3,400 24% 1,400 I- Equipment Reserch 4,800 7,550 2,750 Extension 3,500 4,650 1,150 Total Equipment 8,300 59% 12,200 87% 3,900 m- TA&Training Research 1,800 500 -1,300 Extension 1,900 900 -1,200 Total TA&Training 3,700 26% 1,400 10% -2,300 IV- Unallocated 3,000 0 Total 17,000 17,000  TUNISIA AGRICULTURAL RESEARCH AND EXTENSION PROJECT APPENDIX C DETAIL OF LOAN DISBURSEMENT AND ACHIEVEMENT OF OBJECTIVES  Table 1: Research Component - Training (in man-days, unless specified) Program of Research 92 93 94 95 96 97 Total Total Planned at % (man-months) Appraisal ActuaV (a) (man-months) Planned Arid zones - - 26 90 31 42 189 8.6 32 27% Vegetables - 7 - - 23 32 62 2.8 6 47% Livestock production - - - - 30 44 74 3.4 21 16% Forestry - - - 60 30 80 170 7.7 24 32% Major crops - 15 - - 73 - 88 4.0 32 13% Olives - 7 68 21 70 - 166 7.5 24 31% Potatoes - 17 30 60 93 17 217 9.9 20 49% Fruit Trees and - - - - - 14 14 0.6 24 3% Citrus Irrigation - - - - 52 - 52 2.4 27 9% Farm machinery - - 136 - - - 136 6.2 21 29% Institutional - - - 15 21 8 44 2.0 12 17% development Budget Control - 30 - - - - 30 1.4 0 Total 0 76 260 246 423 237 1,242 56.5 243 23% (a) Man-days were converted into man-months using a 22:1 ratio  Table 2: Research Component - Disbursement by Categories of Expenditures (in US$ '000's) Categories of Expenditures 92 93 94 95 96 97 (a) Total Planned % Actual (b) Planned 1. Civil Works 0 75 128 309 630 457 1,600 Subtotal Civil Works 0 75 128 309 630 457 1,600 1,600 100% II. Equipment - Equipment for Laboratories 61 592 613 1,214 755 1,493 4,727 - Prnting Equipment, Computers, etc. 7 24 0 47 103 442 624 - Vehicles 69 219 0 56 228 143 715 - Farm Equipment 125 260 19 136 643 111 1,296 98