Tunisia - Gabes Irrigation Project

Document of The World Bank FOR OFFICIAL USE ONLY Report No. 14444 PROJECT COMPLETION REPORT TUNISIA GABES IRRIGATION PROJECT (LOAN 2605-TUN) MAY 4, 1995 Agricultural Operations Division Maghreb and Iran Department Middle East and North Africa Regional Office 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.  PROJECT COMPLETION REPORT TUNISIA GABES IRRIGATION PROJECT (Loan 2605-TUN) CURRENCY EQUIVALENTS Currency Unit = Tunisian Dinar (D) Appraisal Year US$1 = DO.77 Implementation Average US$1 = DO. 89 Completion Year Average US$1 = Dl.002 ABBREVIATIONS AIC: Association of Common Interest (Association d'Interet Collectif) AGSI: Irrigation Management Improvement Project (Am6lioration de la Gestion du Secteur Irrigu6) BNDA: National Bank for Agricultural Development (Banque Nationale pour le D6veloppement Agricole) BNA: National Bank for Agriculture (Banque Nationale Agricole - aleas Banque Nationale de Tunisie, BNT) CNEA: National Centre of Agricultural Studies (Centre Nationale d'Etudes Agricoles) CRA: Agricultural Extension Service (Centre de Rayonnement Agricole) CRGR: Rural Engineering Research Centre) (Centre de Recherche du G6nie Rural)  CRDA: Regional Commissariat for Agricultural Development (Commissariat R6gional de D6veloppement Agricole) CTV: Local Extension Unit (Cellule Territoriale de Vulgarisation) DGR: Department of Rural Engineering (Direction du Genie Rural) DGPC: Highway Department of the Ministry of Public Works (Direction Generale des Ponts et Chaussees) DPSAE: Department of Planning, Statistics and Economic Analysis (Direction de la Planification, des Statistiques et des Analyses Economiques) DRE: Department of Water Resources (Direction des Ressources en Eau) DRS: Department of Soil Resources (Direction des Ressources en Sols) FOSDA: Special Funds for Agricultural Development (Fonds Speciaux Pour le D6veloppement Agricole) GIH: Group of Hydraulic Interest (Groupement d'Interet Hydraulique) GIAF: Association of Fruit and Citrus Growers and Traders (Groupement Interprofessionnel pour les Agrumes et les Fruits) GIL: Association of Vegetable Producers and Traders (Groupement Interprofessionnel pour les Legumes) IPB: Initiating Project Brief INRAT: Tunisian National Institute of Agricultural Research (Institut National de Recherche Agronomique Tunisien) IRA: Research Institute for Arid Zones (Institut de Recherche pour les zones Arides) M&E: Monitoring & Evaluation MOA: Ministry of Agriculture (Ministere de l'Agriculture)  MOF: Ministry of Finance (Ministere des Finances) MOP: Ministry of Planning (Ministere du Plan) OC: National Cereal Board (Office des Cereales) OEP: National Livestock and Range Agency (Olfice de l'Elevage et des Paturages) OH: National Olive Oil Board (Office National de l'Huile) O&M: Operation & Maintenance OMV: Office de Mise en Valeur OMVGM: Gabes and Meddnine Irrigation Development Office (Office de Mise en Valeur des Perimetres Irrigues de Gabes et Medenine) PPI: Public Irrigated Perimeter (Perimetre Public Irrigue) SAR: Staff Appraisal Report SONEDE: National Company for Water Supply (Societ6 Nationale d'Exploitation et de Distribution des Eaux) STEG: Tunisian Electricity and Gas Company (Societ6 Tunisienne d'Electricite et du Gaz) STIL: Tunisian Milk Processing Company (Societ6 Tunisienne d'Industrie Laitiere) SNPSA: National Strategy for Promoting AICs (Strategie Nationale de Promotion et Suivi des AICs) WUA: Water User Associations  FOR OFFICIAL USE ONLY THE WORLD BANK Washington, D.C. 20433 U.S.A. Office of Director-General Operations Evaluation May 4, 1995 MEMORANDUM TO THE EXECUTIVE DIRECTORS AND THE PRESIDENT SUBJECT: Project Completion Report on Tunisia Gabes Irrigation Project (Loan 2605-TUN) Attached is the Project Completion Report on Tunisia-Gabes Irrigation Project (Loan 2605- TUN). Parts I and III were prepared by FAO/CP for the Middle East and North Africa Regional Office and Part II by the Borrower. This successor to the Southern Irrigation Project (Ln. 1796) improved existing irrigation and was supposed to improve fanning practices on 43 oases around Gabes. The 7,000 farms on 5,000 ha suffered from aging date palm groves, declining artesian pressure, ill-maintained irrigation systems, salt build up, and tenure problems. Some land was abandoned. Civil works-watercourses, and deep wells and pumps to compensate for declining artesian pressure-were well designed and built on schedule and at less than projected cost, thanks to efficient bidding and supervision. Institutional and agricultural development suffered however. Government froze hiring and did not provide its share of project finance in the early years; the project "office" was abruptly abolished in mid-implementation. Consequently, extension to help farmers adjust to new irrigation was minimal. Some farmers resisted drain installation. The program to replace aging date palms with higher-return fruit trees and vegetables moved slowly even after Government reverted to high subsidies, which had initially been reduced under the project. Projected involvement of the "office" and irrigators' associations in marketing was wisely dropped. Irrigators' associations strengthened themselves and now pay for 80 percent of operation and maintenance. The irrigation facilities are now working well. Agricultural intensification is proceeding, thanks to strong mark-ets, and strong farmer market orientation and initiative. In the absence of useful monitoring and evaluation, evidence is fragmentary. The PCR estimates that, due especially to delayed or non-implementation of agricultural components such as extension, benefits will be delayed by 7 years, but that appraisal full-development targets will then be surpassed. The project outcome is rated as satisfactory. There are still problems, especially with major maintenance, drainage and salinization. Farmers' interest in and complete dependance on irrigation suggests that they will solve these problems. Sustainability is rated as likely. While large-scale institutions have not evolved as expected at appraisal, small-scale irrigators' institutions are flourishing. Institutional development is therefore rated as substantial. The PCR is factual, concise, provides a balanced assessment of the project's achievements, and some perceptive lessons and insights. An audit of this and other Tunisian irrigation projects is planned. Attachment , This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank authorization.  II ac  FOR OFFICIAL USE ONLY PROJECT COMPLETION REPORT TUNISIA GABES RRIGATION PROJECT (Loan 2605-TUN) TABLE OF CONTENTS Page No. Currency Equivalents and Abbreviations PREFACE ....................... ......................... i EVALUATION SUNIM ARY .................u.................... i PARTI ........................1.......................... 1. Project Identity ........................................... 1 2. Background ............................................. 1 3. Project Objectives and Description ................................ 2 4. Project Design and Organisation ................................. 4 5. Project Implementation ...................................... 6 6. Project Results ............................................ 9 7. Project Sustainability ....................................... 10 8. Bank Performance ......................................... 11 9. Borrower Performance ...................................... 12 10. Project Relationship ........................................ 13 11. Consulting Services ........................................ 13 12. Project Documentation and Data ................................ 14 PART II - PROJECT REVIEW FROM THE BORROWER'S PERSPECTIVE ... .... 15 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. I  PART m - TABLES .......................................... 17 1. Project Timetable .................................. 17 2. Cumulative Disbursements ..................................18 3. Project Implementation .................................. 19 4. Project Financing in US$ .................................. 20 5. Financial Impact ................................... 21 6. Economic Impact ................................... 22 7. Status of Covenants ................................... 23 8. Missions Carried Out .................................. 24 9. Staff Inputs ........................................... 26 ANNEXES i . Agricultural Development .................................. 27 2. Financial and Economic Analysis . ............................. 37 Table 1. Financial Prices .................................. 39 Table 2. Economic Benefits and Costs .......... ................ 40  PROJECT COMPLETION REPORT TUNISIA GABES IRRIGATION PROJECT (Loan 2605-TUN) PREFACE This is the Project Completion Report (PCR) for the Gabes Irrigation Project, for which Loan 2605-TUN in the amount of US$ 27.7 million was approved in July 1985. The loan closed in September 1993, nine months behind schedule. A sum of about US$ 9.3 million was cancelled between December 1988 and January 1994, date of the last disbursement. Parts I and m of this PCR were prepared by a mission from the FAO/World Bank Cooperative Program which visited Tunisia in September 1993. The PCR is based on a review of the Staff Appraisal Report (No. 551 1-TUN) dated May 30, 1985, WB supervision reports and project files, field visits to the Gabes Governorate, and discussions with relevant Government officials and project beneficiaries.  II  - ii - PROJECT COMPLETION REPORT TUNISIA GABES IRRIGATION PROJECT (Loan 2605-TUN) EVALUATION SUMMARY Objectives: The main project objectives were to: a) provide more and better distributed water to 43 oases covering 5,000 ha in the Gabes Governorate, b) increase and diversify agricultural production, c) strengthen the public and private organizations directly involved in the production and marketing of agricultural products, and d) set up sustainable arrangements for the operation & maintenance (O&M) of the irrigation infrastructures. Total costs were estimated at US$ 58.3 million (D 43.8 miUion) over six years and the Bank loan was equivalent to US$ 27.7 million. Through increased water availability, decreased salinity, and orchard reconversion, the project was meant to enable around 7,700 farming families to intensify the production of fruits and vegetables, dates and forage, on a profitable and sustainable basis. Implementation Experience: Loan effectiveness started in Year 1986 and two loan cancellations totaling US$ 8.5 million were effected following substantial under-disbursements in equipment/materials and agricultural development. An additional $876,000 was cancelled in January 1994, date of the last disbursement. The physical targets for the civil works have been exceeded, mainly because of efficient bidding procedures and the tight supervision of the work. A total of 37 deep wells were sunk (although 8 were dry - 15% in terms of overall depth), compared to 26 foreseen at appraisal. Some 4,225 ha are now irrigable - less than the 5,021 ha planned - but an additional 885 ha will also become irrigable when all the wells are finally connected/operating, the conduits damaged by flood repaired, and the remaining land tenure problems resolved (para. 5.6). Agricultural development has not proceeded at the pace envisaged at appraisal. It was planned that 5,021 ha. were to be fully reconverted - ie. aging date palm trees uprooted and new fruit trees and horticultural crops planted - at the end of 1991. However, of the 4,225 ha now irrigable, it is estimated that at least 25 % are still not being exploited on a yearly basis. Only 23,000 trees have been uprooted so far, out of a target of 50,000 palm trees, and some 250,000 fruit trees have been planted or 56% of a target of 450,000. The 22 additional extension agents envisaged at appraisal were never hired and, as a result, most farmers have never been properly assisted in improving their water use practices and cropping systems. On the other hand, there was the unexpected establishment of around 25 ha of geothermal green houses which allow some off-season sales to Europe. Moreover, the domestic market for horticultural products has revealed to be more buoyant than predicted. The market prices have been strong and the private marketing channels dynamic. Partly as a result of this, some few  - iii - investments planned at appraisal (mainly cold storage and grading/packing facility) turned out to be unnecessary and were cancelled. The main reasons for the slower than expected pace of agricultural development were the beginning of national economic austerity at project start (shortage of Government's counterpart funds and freeze on hiring national/incremental staff), the lack of balanced project coordination (dominance of irrigation engineers with little input by agronomists and socio- economists) and of a proper monitoring & evaluation (M&E) system, and the absence of an analysis (and therefore knowledge) of the potentials, constraints, and priorities of the beneficiaries. At different periods during project execution, various farmers resisted the installation of irrigation infrastructures (including drainage networks) in their fields as well as tree uprooting, mainly for reasons related to land tenure (land fragmentation - too many owners with very small plots or on the same lands - and absenteeism). Results: The lack of a proper M&E system means that the information available on the project results is insufficient and very fragmentary. Nevertheless, the mission feels that - in general terms - it is a matter of time before most of the expected benefits at appraisal materialize. The problems indicated previously will mainly result in delaying these benefits rather than reducing them. Basically, the irrigation infrastructures put into place are operating satisfactorily, the domestic and international (off-season) demand for horticultural products is strong, and the farmers in the Gabes Governorate have a long/recognized experience with growing/marketing these products. It is estimated that full development will now come in Project Year 23, compared to Year 16 (2002) at appraisal. The incremental production at full development would be at slightly higher levels than expected at appraisal; some 24,000 tons in the case of vegetables of which 8,600 out of season (including 4,500 tons produced in geothermal greenhouses). The mission has calculated that the project's internal economic rate of return will be around 16%, compared to 21% at appraisal. Sustainability: The sustainability of the project is promising. There is now a water users association for each deep well in operation and around 80 % of the O&M needed on all the irrigation infrastructures are being financed regularly by these associations. There is also an upcoming nationwide institution building programme, financed through American and German aid, which will reinforce the institutional support already provided through the project to these associations. Moreover, an ongoing Bank-financed National Research and Extension Project is expected to pursue most of the project's previous goals for applied research and extension in the Gabes area. The main risk to project sustainability is increasing land fragmentation which might hamper farmers' rational use of the lands made irrigable. Lessons Learned: The main lessons to be drawn from the experience of this project are not new. The first is that irrigation projects should always be formulated taking into account the often limiting institutional/financial absorptive capacity of the Government and beneficiaries. Second, and related to the first, is the necessity to analyze the socio-economic conditions of beneficiaries before trying to help them. In the present project, this could have resulted in: i) starting the project with oases where the problems were found to be least important in terms of  - iv - land tenure, and ii) in parallel (still at project start), experiment with (pilot) solutions aimed at solving these problems. This approach - combined with splitting the project into two phases over a longer period, to reflect the constraining absorptive capacity - would probably have yielded a higher overall economic rate of return than now expected. Third, there is always the need to set up (and ensure the efficient functioning of) an overall project coordination system as well as a proper M&E. In the case of Gabes, this perhaps could have avoided - or at least minimized - the loss of synchronisation and balance between the civil works and agricultural development.   PROJECT COMPLETION REPORT TUNISIA GABES IRRIGATION PROJECT (Loan 2605-TUN) PART I 1. Project Identity Project Name Gabes Irrigation Project Loan No. 2605-TUN RVP Unit Middle East and North Africa Country Tunisia Sector Agriculture Sub-sectors Irrigation 2. Background 2.1 Despite a more rapid growth in Tunisia's petroleum and tourism activities over the 1980s, the agricultural sector still represents about 13% of GDP and provides work to around one-third of the labor force. Repeated droughts had, however, slowed the pace of agricultural production during 1978-82. At the time of the formulation of the Gabes Irrigation Project - under the Sixth Development Plan (1982-86) - the Government's main objectives were to decrease national unemployment and regional economic disparities. The need to maximize investment returns in agriculture grew in importance due to the necessity of adjusting the economy to a post-hydrocarbon situation. The overall approach adopted was to promote small/labor intensive investments, including the rehabilitation of existing infrastructures, and to create an economic environment more favorable to agriculture (mainly more realistic prices).  2.2 The southern part of Tunisia (60% of total land area) is a predesert zone with less than 200 mm of rainfall per year. Given the relatively low level and high variability of these rainfalls, important efforts are being made to develop irrigated agriculture. A Master Plan for Water Resources Development in Southern Tunisia was elaborated in 1974 and a Southern Irrigation Project was prepared by the World Bank in 1979, as a first implementation phase. The project essentially provided for the rehabilitation and new development of 16 oases in the South and the Project Completion Report produced in January 1988 indicated that: "In general, rehabilitated oases and new plantations are almost fully in operation; farmers are dynamic, fully participating and satisfied." The main lessons drawn were the need to "bulk" the contracts as much as possible at project implementation, for avoiding long administrative procedures, and to involve the beneficiaries in project preparation. The Gabes Irrigation Project constitutes the second phase of the 1974 Water Master Plan. 3. Project Objectives and Description 3.1 The overall objective of the project was to reverse the decline of agriculture in the southeast of the country. This decline was mainly due to: a) a shortage of water; lowering of the artesian head of wells, deteriorating irrigation networks, aging of existing wells, b) an increase in salinity; poor drainage and insufficient leaching, and c) the aging of trees (mainly date palm and olive) and subsequent partial abandonment by farmers. To reverse this situation, the project was aimed at: a) providing more/better distributed water and improving drainage, b) increasing and diversifying agricultural production, c) strengthening public and private organizations directly involved in the production and marketing of agricultural products, and d) setting up sustainable arrangements for the Operation & Maintenance (O&M) of the irrigation infrastructures. 3.2 The project area consisted of 43 oases covering around 5,000 ha on the coastal strip of the Gabes Governorate. The oases are of four different types: i) the oases of the El Hamma region mainly with varieties of date palm/olive trees and forage, ii) the oases of Gabes where vegetable production predominates, iii) the oases of Gabes-South where fruit trees (pomegranate) predominate, and iv) the oases of Ghannouche M'Ziraa where there is only vegetable production. A characteristic of several oases is the large degree of farm fragmentation, especially close to Gabes where each ha is typically fragmented into 3-4 plots owned by different people. A total of 7,700 families are estimated to exploit the 43 oases.  - 3 - 3.3 The project components were essentially": (a) Development of Infrastructure: (i) rehabilitation of 42 oases totalling some 4,800 ha; digging 26 wells and installing 45 pumps where artesianism is insufficient, setting up distribution systems, improving drainage networks, uprooting aging trees (mainly date palm and olive) and planting high-yielding ones (ex: fruit trees) and horticultural crops. (ii) expansion of irrigation over new areas totalling 200 ha; one new oasis and expansion of three existing ones). (b) Institutional Development: (i) strengthening of the central and regional administration of the Ofrice de Mise en Valeur des perimktres irrigues de Gabes et Medenine (OMVGM); additional technical/administrative staff, the construction of offices and houses, and the provision of cars. (ii) strengthening of the extension services of the OMVGM; construction of extension centers, additional extension agents and their training, the provision of technical assistance'. (iii) strengthening applied research; the construction of nurseries and the provision of equipment for a small soil and water laboratory. (iv) support to marketing; the construction of a cold storage room, the provision of one grading/packing facility and three trucks. (v) strengthening the financial/technical capacity of the Associations d'Int6rAt Collectif" (AICs); through study tours/seminars and workshops with the OMVGM. Staff Appraisal Report, May 1985, page 15. 21 For the OMVGM in general; 22 new extension agents, four research technicians, three mechanics, 10 buildings and 26 vehicles. 31 The AICs are water users associations. The issues of their reorganization and improvement of their legal status was already being dealt with through a national project, Ameiioration de Ia Gestion du Secteur Irrigu6 - AGSI, (2573- TUN).  - 4- (c) Support Action: (i) promotion of credit and input supplies; for the uprooting of aging trees and the planting of 450,000 new trees, small farm machinery, reservoirs, greenhouses (20 ha), and plastic tunnels (80 ha). (ii) strengthening of services to supervise the execution of the works and to monitor and evaluate the project impact; the hiring of one agro-economist with assistant. (iii) provision of technical assistance and training; mainly for organization, management, extension, hydromechanics, and marketing. 3.4 Total costs were estimated at US$ 58.3 million (D 43.8 million) over six years, of which 51 % in foreign exchange and 25 % in taxes/duties. Some 77% of the costs are allocated to the rehabilitation of irrigation infrastructures; ie. civil and on-farm works, equipment and operating costs. The World Bank loan was equivalent to US$ 27.7 million and the Government's counterpart funds to US$ 8.4 million. The Banque Nationale Agricole (BNA)"was expected to provide US$ 2.9 million in credit and the beneficiaries would contribute the remaining US$ 5 million. 3.5 Through increased water availability, decreased salinity, and orchard reconversion (essentially the uprooting of aging date palms and other fruit trees and their replacement by more productive/profitable fruit trees and vegetable crops), the project was expected to enable the 7,700 farming families to intensify the production of fruits, vegetables, dates and forage, on a profitable/sustainable basis. The incremental production - with the project at full development - was expected to be of the order of 18,000 tons of vegetables (55% in winter), 6,000 tons of dates, 5,000 tons of pomegranates, and 28,000 tons of forage. The ERR was calculated at 21 % over 20 years. The main risks were said to be related to cost overruns and delays in project implementation. Other risks were the possibility of price shortfalls, marketing bottlenecks, and slow orchard reconversion by farmers. 4. Project Design and Organization 4.1 The project was identified in 1983 by the Bank as a follow-up to the Southern Irrigation Project. The Center National d'Etudes Agricoles (CNEA) carried out a feasibility study which was followed by two Bank missions to supervise project preparation by the Govemment. Two other missions were subsequently sent to appraise the project. This gradual approach, in close collaboration with the Govemment, had also been used for the formulation I Called the Banque Nationale de Tunisie at appraisal.  -5- of the previbus project. The concept of the project is classical but sound, and seemed to have been relatively well understood by all the Government institutions involved". Numerous contacts were taken at project formulation with the future beneficiaries but it is reported that the discussions were essentially limited to agronothic matters2". 4.2 One innovative aspect of the project design was the uprooting of aging trees, which was to be carried out by private contractors and reimbursed at 50% by the farmers through credit. This was in contrast with the Govermnent's approach to pay the farmers full subsidies for carrying,out the uprooting themselves. The project approach worked well for the first years, mainly - according to the Government - because the program was started in the Mareth oases where uprooting is easier (lower tree density and more accessible for machines). The cost of carrying out uprooting mechanically was said to be much higher in other oases and several farmers refused to participate. At that point, tlhe Government - reversing its initial acceptance of the project's approach - chose to offer farmers an increased subsidy of DlI per tree for executing the work themselves. But this did not help because, as will be seen in the next chapter, the difficulty remained that several faimers could not remove trees for reasons related mainly to land tenure. 4.3 The organizational arrangements of the project were as follows. The Ministry of Agriculture (MOA) was the executing agency through its OMVGM in Gabes. A project management unit (Bureau de gestion) was to be set up for the purpose of coordinating the various activities. All the major civil works were to be managed and supervised by the Direction du Genie Rural (DGR). The AICs - which constitute the main formn of water users' associations in the region - were going to be strengthened and made directly responsible for the Operation & Maintenance (O&M) of the irrigation infrastructures put at their disposal. Major maintenance works and special repairs would be carried out by the OMVGM at cost, until the AICs could take care of these services before project termination. The OMVGM was to establish a contract with the Institut de Recherche pour les Zones Arides (IRA) to carry out applied research on water use and fruit/vegetable production. The OMVGM was also in charge of operating the cold storage, giading/packing facility, and trucks until some AICs (or existing cooperatives) could take over the managemeuit of these assets. Credit was to be granted by the BNA. 4.4 The roles/responsibilities of the institutions involved were relatively clear at project start. The OMVGM was abolished in 1989, three years after project start. The already existing Commissatiat regional de developpement agricole (CRDA) became the project executing agency. This created disturbances in the operations which, along with the excessive turn-over of the top management during project execution3", can still be felt today. Furthermore, the " However, there were fundamental institutional changes during project implementation, see next chapter. 2/ A communications specialist was to be hired at project start and help organize a campaign to inform/sensibilize the beneficiaries. This, however, did not materialize. 3' Three President-Directeur-Generals for the OMVGM and three Directeur-Gknrals for the CRDA.  - 6 - situation was exacerbated by the fact that the Bureau de gestion was never given enough recognition and power by The Borrower. This bureau, as well the agro-economist heading it"', do not even figure officially in the CRDA's organigramme. Although there is validity in the MOA's argument that the bureau is only temporary, the practical consequence is that it has no direct authority to coordinate the project activities, particularly over the powerful DGR. As a result, these activities quickly grew out of balance and agricultural development ended up lagging far behind the civil works (see next chapter). 4.5 In retrospect, the Mission feels that project coordination was not given sufficient attention at formulation. Given the overwhelming importance of the civil works in the project costs - and consequently of the engineers in project implementation - there was a concrete danger that agricultural development would be "neglected" as well as the land tenure problems. A coordinating committee, set up at the Directeur-G6n6ral level in the MOA at Tunis, could perhaps have assisted. This committee could have included officials directly concerned with agricultural development (as well as irrigation) who would have ensured that matters such as the organization of tree uprooting and applied research/extension were sufficiently dealt with during project execution. Representatives from the AICs could also have been included, for helping to understand better the constraints/priorities of beneficiaries. 5. Project Implementation 5.1 Loan effectiveness started in June 1986, only four months later than planned. Two cancellations totalling US$ 8.5 million were effected following substantial under-disbursements in the expense category "equipment/materials (mainly irrigation pumps)" due mainly to efficient bidding procedures and favorable exchange rates, and in the expense category "agricultural development". The total loan amount was consequently decreased to US$ 19.2 million and the allocated sums per expenditure categories were modified, upward for the irrigation infrastructures and drastically downward for equipment/materials and agricultural development. A final cancellation of US$ 876,000 was effected on January 26, 1994, date of the last disbursement. 5.2 The physical targets for civil works have been exceeded, mostly because of efficient bidding procedures and the tight supervision of the work2'. A total of 37 deep wells were sunk (although 8 were dry - 15% in terms of overall depth), compared to 26 foreseen in the Staff Appraisal Report (SAR). A total of 61 irrigation pumps were procured (16 more than in SAR), of which nine are still to be installed or used as back-up. The equivalent of 381,700 meters of conduits and accessories were installed, slightly more than envisaged in the SAR. Some 4,225 ha are now irrigable - less than the 5,021 ha planned - but an additional 885 ha will also become " His only assistant, along with the computer, were taken away from the bureau two years ago. 2/ The slowing down of oil drilling in other areas also increased competition and resulted in lower unit costs than expected.  - 7 - irrigable when all the wells are finally connected/operating, the conduits damaged by flood repaired, and the remaining land tenure problems in limited areas resolved (see para. 5.6). Eight buildings (offices/houses) were constructed under the project and 24 vehicles provided. 5.3 Regrettably, agricultural development has not proceeded at the pace envisaged at appraisal. The SAR envisaged that the 5,021 ha. planned were to be fully reconverted - ie. unproductive trees uprooted and new/more valuable trees and horticultural crops planted - at the end of 1991. But of the 4,225 ha now irrigable, it is estimated that at least 25 % are still not being exploited on a yearly basis. Out of a CRDA's revised uprooting program of 50,000 trees (as opposed to an SAR target of 270,000 trees"), only 23,000 palm trees have been uprooted so far under both the project's approach (3,000) and with Government's subsidies (the remaining 20,000, see para. 4.2). In addition, some 250,000 fruit trees have been planted or 56% of the SAR target of 450,000. Concerning applied research and extension, the contract with the IRA was cancelled due to the institute's strong orientation towards livestock/pastures and its consequent inability (mainly in terms of competent staff) to carry out the program envisaged. The 22 incremental extension agents planned in the SAR were never hired and the limited number of existing field agents have been far too busy with statistical surveys and data collection/compilation. As a result, most farmers have never ben properly assisted in improving their water use practices and cropping systems. 5.4 On the other hand, some unexpected and positive developments have taken place during project implementation. First, there has been the establishment of around 25 ha of geothermal green houses2'which allow some off-season sales to Europe (total of 707 tons of tomatoes and melons in 1992). Second, the domestic market for horticultural products has revealed to be more buoyant than predicted. Contrarily to what was feared at appraisal (see para. 3.5), the market prices have been strong3'and the private marketing channels proved to be dynamic. Partly because of that - and the proximity of main markets such as Sfax and even Tunis - the project investments in marketing (see para. 3.3) have not been critical to the outflow of the project's incremental production4'. These investments were eventually cancelled since, in any event, the Gabes Wholesale Market - to which the cold storage was offered - was not prepared to finance their O&M costs. Finally, although the experience with the IRA was disappointing, the OMVGM's research station in Zerkine5'carried out useful trials mainly on the use of geothermically heated water for greenhouses and on vegetable production6". The This target was deemed to be grossly overestimated by the CRDA and reduced following new field inventories during project implementation - see Annex 1. 2/ Using hot water from some deep wells to increase the in-house temperature, enabling earlier production of out-of- season vegetables. 3/ The prices have generally been at higher levels in constant terms (after inflation) than expected at appraisal. 4' Although it must be pointed out that the bulk of this incremental production has not yet materialized. 5/ This station was not targeted for research work by the project because the Government's policy was to rid the OMVs of their separate stations and consolidate them under one authority. 6/ Most of these trials were conducted (unexpectedly) by an expert financed through the UNDP. The trials on greenhouses were instrumental in the establishment of geothermal greenhouses.  -8- station has also produced disease-free seedlings and rooted cuttings for farmers. Under the project, it was envisaged that the nursery activities would be taken over by private entrepreneurs before project completion, but no interested party could be found. 5.5 One of the reasons for the slower than expected pace of agricultural development was that project start coincided with the beginning of an austerity period in Tunisia. As a consequence, there was a shortage of Government's counterpart funds in the first years and a freeze on hiring national (incremental) staff during most of the investment period. The mission could not establish how i-nuch of these problems could have been foreseen before implementation". Certainly, the fact that the 22 (incremental) agents planned in the SAR were never hired, was also a major problem. In addition, as indicated in para. 4.5, the lack of an efficient overall coordination was an other constraining factor. Better coordination could perhaps have avoided - or at least minimized - the loss of synchronization/balance between the civil works and agricultural development. The situation was exacerbated by the absence of a proper Monitoring & Evaluation (M&E) system/unit in the OMVGM (and later in the CRDA). Again, the main difficulty was the lack of provision of resources to the PMU, notwithstanding the clear official recognition of the "bureau de gestion" in the legal documents (Section 3.02 and Schedule 5) with the resulting scarcity of manpower and annual/operating funds during project implementation2". 5.6 The slow pace of agricultural development was also due to the lack of analysis (and therefore knowledge) of the potentials/constraints/priorities of the beneficiaries. Several farmers, at various periods during project implementation, resisted the installation of irrigation infrastructures in their fields (including drainag,. networks) as well as tree uprooting - ex: in the Kettana and El Hamma oases. Some did not want to pay the O&M costs of irrigation/drainage while there was still artesi4nism or when salinity was not yet a problem, others were not exploiting their land (absenteeism) and therefore were not interested, and many could not take any decisions without the consent of all the other owners of the same land (land fragmentation). A better knowledge of the beneficiaries could have allowed to: i) start the project with oases where the problems were found to be least important in terms of land tenure and capacity/willingness to borrow (see para. 9.3), and ii) in parallel - still at project start - experiment with (pilot) solutions aimed at solving these problems; ex: land consolidation activities to rationalize land use- . The danger of slow adoption by farmers of orchard reconversion was indicated in the SAR, although last in the list of possible risks. The fact that the loan amount was reduced by 30% would suggest that, in any event, the project size and its financing plan were somewhat over-optimistic. 21 At the MOA's level, the DPSAE had initially the overall responsibility For M&E but was later replaced by an other direction in MOA which limited itself to monitoring the physical/finarcial realisations. " The Government has been trying recently to convince the absentees to exploit their lands or to rent them to others.  9 (5. Project Results 6.1 The lack of a proper M&E system means that the information available on the project results is insufficient and very fragmentary. Although the physical/financial realizations were reasonably well recorded, the project impact - ex: in terms of the farmers' incremental production/incomes - was never analyzed, apart from the preparation of indicative farm models by the CRDA's staff in 1991 and 1992. Nevertheless, the missidn feels that - in general terms - it is only a matter of time before most of the SAR's expected benefits materialize. The problems indicated previously will mainly result in delaying these benefits rather than reducing them. Basically, the irrigation infrastructures put into place are operating satisfactorily, the domestic and international (off-season) demand for horticultural products is strong, and the farmers in the Gabes Governorate have a long/recognized experience with growing/marketing these products. 6.2 For the time being, although the con; eyance systems installed under the project have resulted in greatly reduced water losses, there, are still technical problems - ex: some farmers receive water at too long intervals for optimaN production of summer vegetables, some others practicing different crop intensities have the same irrigation schedule. Salinity is also becoming a problem for the equivalent of around 10% of the areas irrigated, often because no drainage networks have been installed due to farmers' objections (see para. 5.6) or because of insufficient maintenance. The reduced scale of the applied research/extension work carried out"so far has seriously limited the possibilities for the farmers to maximize the use of the water available. The mission has visited fields where the owners are obviously still trying to find out by themselves how to adapt best their cropping patterns to the new situation. I 6.3 The mission is of the view that the six farm models produced in the SAR are still relevant (with modifications, see Annex I and Table 5 of Part III). One model representing the emergence of geothermal green houses has been added. The financial profitability of all the revised models is quite high (FRRs above 50% over 20 years), except for dates intensification (20%) which was a type of cropping meant to be phased out in the SAR, precisely because of its low potential value and longer gestation period. Farmers will gradually adopt the cropping patterns implicit in the most profitable models as artesianism runs out and as they have to pay the full cost of pumping water2'. Although the FRR for the geothermal greenhouses is relatively low at 29 %, the net incremental income per ha/year at project full development is of the order of $ 53,000. Of course, the investments necessary are also in proportion and out of reach for ffost small farmers. I A special one-man supervision mission carried out in 1990 indicated that the ratio of farmers per field agent varies between 250 and 600 ahd that extension is only a small part of the agents' work anyway. 2l Roughly half the project wells still have artesianism. In the models, the O&M costs for the irrigation infrastructures was calculated on the basis of the mid-point between with artesianism and with fully pumped water. With pumped water, the FRR for dates intensification is 13%.  - 10 - 6.4 It is estimated that project full development will now come in Project Year 23, compared to Year 16 (2002) in the SAR. The incremental production would then be of the order of 24,000 tons of vegetables of which 8,600 of out-of-season vegetables (including 4,500 tons produced in geothermal greenhouses), 7,000 tons of dates, 11,000 tons of fruit, and 28,000 tons of forage. These production levels are slightly higher than those forecast at appraisal. There is a larger production of out-of-season vegetables and still little of the expected new crops, such as table olives and pistachios. On the basis of the (implicit) average farm holding size used in the SAR (0.65 ha per family), the 4,225 ha now irrigable under the project would be exploited - although not regularly, see para. 5.3, - by around 6,500 families. The mission has calculated that the project's internal economic rate of return will be of the order of 16% over 30 years, compared to 21 % over 20 years at appraisal". 6.5 Other tangible results could also be credited to the project. Before project start, a number of farners - reportedly around 1,200 families exploiting some 800 ha - had abandoned their lands, following the problems of water shortages, salinity and aging of their trees. It is now said that about 25 % of them have returned with more families expected to follow in the years to come. The return of these families will no doubt contribute (although modestly) to the overall economic growth of the Gabes region. This is in line with the overall objective of the project which was to stop the decline of agriculture in the southeast of Tunisia. The project also served to establish the principle and practice of the water users paying the O&M of the irrigation infrastructures put at their disposal, thereby helping guarantee the sustainability of irrigation. 7. Project Sustainability 7.1 The sustainability of the project is promising. There is now an AIC for each deep well in operation and around 80% of the needed O&M for all the irrigation infrastructures are being financed regularly by these associations2'. Consequently, the farmers are becoming increasingly aware of the cost of water and the need to economize it. In addition to the support provided through the project (although mainly technical), these AICs will now be further strengthened (mainly in managerial and financial training) by a nationwide institution building program (Strat6gie Nationale de Promotion et Suivi des AICs - SNPSA), financed through American and German aid. 7.2 The major repairs and replacement of the main parts on the irrigation infrastructures continue to be the responsibility of the CRDA. Up to the end of 1992, the financial matters of the AICs in the Gabes Governorate - basically book keeping and disbursements - were indirectly taken care of by the CRDA (and the Governorate's officials) through an established Groupement d'interet hydraulique (GIH). The GIH is now meant to be tnerely an adviser " Part of the negative impact of the delays in benefits is compensated by lower project costs and higher market prices for products. 2/ The remaining 20% of the needed O&M is either "neglected' or is carried out by the CRDA.  - 11 - but the situation of dependence continues de facto since most of the AICs would not be capable of maintaining proper financial records"'. It is to be hoped that the benefits of the SNPSA will reach the AICs in the project area before major repairs become necessary on the irrigation infrastructures. There is the danger that the CRDA will not be able to finance all these repairs and that the AICs would then lose interest in the O&M of the infrastructures. 7.3 There is an ongoing Bank-financed National Research/Extension Project which is determining priority objectives and strengthening the regional structures. This project is expected to pursue most of the project's previous goals for applied research and extension in the Gabes Governorate. The monitoring of the water situation in the southeast of the country by the Water Resources Department of the Ministry of Agriculture (DRE) is based on mathematical simulation models developed with the help of UNESCO in 1968-72 and systematically updated and refined as more detailed data have become available since then. The water balance in the project area is closely monitored and no serious problem is expected for the foreseeable future. The two main risks to project sustainability are increasing land fragmentation, which might hamper farners' rational use of the lands made irrigable, and recurrent salinity due to insufficient maintenance of the drainage networks. 8. Bank Performance 8.1 The project was the object of an Initiating Project Brief (IPB) in January 1984. The main components and their content have not been changed substantially throughout formulation. The financing by the Govermment of the counterpart funds had already been singled out in the IPB as one of the main issues2'. The matter was raised, first, by the Task Manager in an internal Bank memo and, later officially, with the Tunisian Ministry of Planning (MOP) and MOA in July 1984. It was agreed that the financing of the project's operating costs would be made/shown separately in the OMVGM's annual budget. In retrospect, this measure was not sufficient to counter the negative impact of the austerity period coinciding with project start. Furthermore, the transfer of overall responsibility to the CRDA in 1989 has resulted in the abandonment of this practice. Conceming the potential problems with M&E, the issue was raised at the Bank's review of the "yellow cover" SAR. It was suggested that particular attention should be given to this aspect at appraisal. But, in any event, the lack of Government's counterpart funds prevented the setting up of a proper system. 8.2 The project was systematically supervised each year and the reports often diagnosed the problems ahead of time and proposed practical solutions. For crucial issues, such as the shortage of staff for extension3'and M&E - as well as the controversies over the tree uprooting " Including a correct depreciation of their assets and the setting aside of money for replacements. 21 This was already an issue with other Bank projects in Tunisia. 31 The Bank also sent two one-man special missions to analyze the problems with extension. But the shortages of manpower and financing proved to be intractable.  - 12 - program (see para. 4.2) - specific telexes stressing the problems, and the Bank's concerns, were sent to the relevant Government officers in Tunisia. But in most cases, the problems were not or could not be solved. Certainly, the fact that 30% of the initial loan amount was cancelled lessened the temptation on the Bank's side to consider more serious steps". There is also the point that the economic viability of the project was diminished as a consequence of the problems but never seriously threatened. The project performed reasonably well, when compared to other irrigation projects in the country (and elsewhere in the developing world). 8.3 However, the mission feels that the SAR tended to exaggerate the potential benefits of the project's extension component, which was expected to solve most of the problems related to land tenure and orchard reconversion. Consequently, the possible impact of these problems was underestimated at appraisal and the uprooting program (for instance) was certainly over- optimistic. 8.4 There are a number of lessons to be drawn for the Bank, all of them not new. First, irrigation projects should always be formulated taking into account the institutional/financial absorptive capacity of the Government and of the beneficiaries, which therefore need to be analyzed. Nornally, this absorptive capacity constitutes a real limitingfactor to the project size. In this case, the adoption of the approach suggested in para. 5.6 - combined with splitting the project into two phases over a longer period, to reflect the constraining absorptive capacity - would probably have yielded a higher overall economic rate of return than now Pxpect(ax1. In addition, a critical analysis of the outlook for the Government's public finance would perhaps have reinforced the case for implementing the project over a longer period. 8.5 Second, as indicated in para. 4.5, project coordination should be given sufficient attention in the design if the right balance/synchronization between the components is to be kept. In the present project, the lack of balance between the civil works and agricultural development was a result of the fact that the officers responsible for irrigation in the MOA were allowed to dominate the situation. Third, the setting up of a proper M&E system - meaningful but not excessively complex - is more difficult than often thought. It requires speciali1ed technical assistance which is seldom available in the borrowing country and of an intensity which might need to be more than for other types of assistance2". 9. Borrower Performance 9.1 It is very difficult to judge to what extent the Tunisian MOP and MOA could have foreseen the upcoming austerity measures at the time of project formulation. II so, they (in particular MOP) should have insisted on a more modest project (in a first phase) or a different financing plan. It is reported that the abolition of the OMVGM in 1989 was a rather sudden Such as a Bank intervention with the Government at higher levels of management. 21 In this respect, the CNEA's contribution was deemed by the Bank to be unsatisfactory, see para. 11.1.  - 13 - decision on the part of the MOA. Certainly, the Bank was not advised officially in advance and the coordination of the project still suffers from this institutional clhange. The transfer of authority to the CRDA should have been carried out with more consideration for the project, in which - after all - the Government has a high stake. 9.2 It could also be said that the MOA should have seen the need for a better coordination/balance between the civil works and agricultural development. The main lesson for the Government is to ensure that irrigation projects are being coordinated and managed by people having the capacity to take an overall view of the matter, in particular of the ultimate objectives pursued". On the other hand, it is to the Government's credit that the two loan cancellations requested and effected were fully justified, on the basis of the developments in project implementation (taking advantage of possible savings) as well as in terms of public finance (minimizing commitment charges and the public debt). 9.3 Although the BNA was part of the project's financing plan, the BNA's regional office in Gabes was not sufficiently familiar with the project. This office is basically managing credit lines for other externally-financed projects such as the Fonds Speciaux pour le Developpement Agricole - FOSDA. Very little of these funds is actually going to small farmers - who constitute the bulk of the population in the project area - mainly because of the existing land tenure problems (with the resulting lack of collateral) and some farmers's preceding bad debt situation. The main lesson is that credit is an essential element of development projects targeted to small farmers and that the terms and conditions of it must be analyzed and agreed as an integral part of the projects. 10. Project Relationship 10.1 The relations between the Bank and the Government were satisfactory throughout the project, except that in the case of the uprooting program the Government went back on understandings reached with the Bank. The situation concerning uprooting is briefly described in paragraph 4.2. 11. Consulting Services 11.1 The contracting and supervision of the civil works during project implementation were efficient - mainly because of the Bank's experience with previous/similar projects - and the procurement packages were constituted in larger sizes than foreseen at appraisal due to greater homogeneity/better synchronization in the contracts. As mentioned in paragraph 4. 1, the CNEA prepared the feasibility study for the project. This study was meant to be preliminary only but it drew criticisms which were partly unjustified. Some Government officials felt that the work i The comments previously made about the M&E are also valid for the Government.  - 14 - carried out was not detailed and specific enough for implementation purposes. It seems that the study tended to be considered by all parties involved as a preparation report which was not the intention. This explains partly the absence of socio-economic studies. Technical assistance contracts have also been signed with the CNEA to assist in the fields of monitoring and marketing. The report on M&E did not correspond to the project needs and was eventually cancelled. The study on marketing (Etablissement de la Situation de Reference et du Plan d'Action) was rather vague and suggested essentially to transfer the activities in input supply and marketing from the Government structures to the private sector, ie. the existing services cooperatives and the AICs. Although the project has not taken any concrete action in this field, the situation has naturally evolved towards a greater privatization of these activities. 11.2 The mission feels that more competition should have been introduced in the selection of the contracts, in order to avoid the risk inherent with having only one outside consultant responsible for most of the technical assistance to be provided. Of the 25 man/months allocated in the project for technical assistance, only four were eventually used. In contrast, 46 of the 47 man/weeks allocated for overall training were used. 12. Project documentation and Data 12.1 The SAR (in two volumes) provided a useful framework for implementation. There are some inconsistencies in calculations (ex: buildup of incremental production to full development) but these have no practical impact on project implementation. The lack of an efficient M&E system meant that the information necessary for the preparation of a more detailed PCR was not available.  - 15 - PART II: PROJECT REVIEW FROM THE BORROWER'S PERSPECTIVE (a) Data in Part m There are no objections to the data given in Part mI of the project completion report. (b) Comments on the analysis in Part I The analysis in Part I refers to the major problem relating to the low rate of oasis development, despite the satisfactory construction of infrastructure. The factors responsible for this situation were in general were analyzed. Several comments that could be of use in the future should nevertheless be raised in the light of project experience. In point of fact, the objective to carry out simultaneously the construction of the irrigation infrastructure and agricultural development was not realistic. After a long period during which the oases were abandoned owing to water shortages, awareness-raising and adaptation of farmers to the new oasis development conditions clearly require considerable time. A project specifically focused on oasis development should normally follow the infrastructure project. (c) Evaluation of the Bank's performance Relations between the Bank (task manager) and the project managers were characterized by clarity and confidence. The nine monitoring missions that were made during project implementation did not run into any difficulties. With regard to the financial aspects, the payment of suppliers and enterprises was made in a timely manner using the funds available through the Central Bank or IBRD. (d) Evaluation of the Borrower's performance It is true that the Project Management Unit is the weak point in project implementation. Because its structure was not formalized, the PMU was at times not given the necessary importance and power, and its role was hindered by the multiple changes in the regional officials and offices of the Ministry of Agriculture (Office/CRDC). Because it was in "limbo", the Project Management Unit lost all motivation in carrying out its functions, which had an adverse impact on the monitoring of the "development" component of the project.  - 16 - As a result, the lesson to be drawn for the future would be to give the PMU legal status and the necessary powers to act vis-a-vis the various project partners (Rural Extension, Agricultural Production, Land Tenure, Rural Engineering, AICs, BNA, farmers). Current regulations in fact make it possible to establish project units that are sufficiently flexible from the hierarchial standpoint. The experience of the Gabes project also indicates a need to include technical assistance to the Project Management Unit with a view to training its personnel in monitoring and evaluation techniques and identifying the main indicators and the procedures for their application. (e) Relationship between the Bank and Borrower The relations between IBRD and the Borrower were very satisfactory.  - 17 - PART m - TABLES Table 1. Project Timetable Stage Date Planned Date Actual Identification Oct. 1983 Oct. 1983 Preparation Sept. 1984 July 1984 Appraisal Oct. 1984 Nov. 1984 Loan Negotiations May 1985 Board Approval March 1985 July 1985 Loan Signature Oct. 1985 Loan Effectiveness Feb. 1986 June 1986 Loan Closing Dec. 1992 Sept. 1993 Loan Completion April 1993 Jan. 1994  - 18 - Table 2. Cumulative Disbursempents (US$ Millions) Bank FY 1986 | 1987 | 1988 | 1989 | 1990 | 1991 1992 |1993 1994 (1) Appraisal 1.0 3.0 6.0 12.7 21.2 25.7 27.7 27.7 0 Estimate (2) Revised 1.0 3.0 6.0 6.7 15.2 19.7 21.6 19.2 0 Estimates' (3) Actuals 0 1.7 6.8 9.1 13.3 15.9 17.1 18.2 18.3 (3) as a %of (1) 0% 56% 113% 72% 63% 62% 62% 66% - (3) as a % of (2) 0% 56% 113% 135% 87% 81% 79% 95% - a/ The loan was closed on 30/09/93. An amount of $6 million was cancelled in December 1988 and a further $2.5 million as of December 1992.  - 19 - Table 3. Project Implementation Key Indicators Unit Appraisal Actual Tar-get Deep Wells No. 26 37 of which 8 negative (15% in terms of meters of depth) Conduits and Accessories % 100 103% in terms of meters (or 381,700 m.) Pumps No. 45 61 of which 9 as back-up or hot-installed yet Works Ha 5,021 5110 of which 200 on new lands as foreseen. The remaining 4,910 consist of tehabilitation Tree Uprooting No. a/50,000 o 3,000 of which 3,000 through the Bank loan Buildings - ctv/cra No. 10 8 - offices No. 3 4 - houses No. 4 3 Vehicles No. 26 24 Demonstration Plots No. 750 597 Area irrigable Ha 5,021 5110 of which 4225 now operating (see Part 1) a/ The SAR initial target was 270,000 trees (see Annex 1). It was subsequently revised following field inventories during project implementation.  - 20 - Table 4. Project Financing in US$ Expenditure Categories As in Loan Amended Actual Agreement By World Bank Civil Works - for Irrigation Infrastructure 8,000,000 9,700,000 10,617.031.88 - for Agricultural Development 2,700,000 40,000 24,943.50 - for Extension/Institutions 700,000 700,000 382.814.65 - equipment and materials 11,400,000 8,670,000 7,202,798.60 - Consultant's Services and Training 200,000 90.000 96,167.63 - Unallocated 4,700,000 0 0 - Total 27,700,000 19,200,000 18,323,756.26 By Government - regular budget 8,400,000 - b'1,886,339 - BNA 2,900,000 - 3,011,490 By Beneficiaries 5,000,000 - 1,387,195 Total (before taxes and duties) 44,000,000 -35,500,000 24,608,780.26 a/ The financing targets for the Government, BNA and beneficiairies were not ajusted. b/ Mission estimates in the field, using an average exchange rate of 0.89D per $1 over the project's investment period (1986-93).  - 21 - Table 5. Financial Impact Models No. of Ha No. of FRR in Net at Appraisal Ha in PCR a' Incremental PCR Income per Ha/Year dl A. Dates intensification 1,023 860 20% 1,306 B. Fruit/vegetables intensification 2,784 3,030 133% 4,198 C. Olives intensification 422 500 71% 1,137 D. Olives/crops intensification 311 230 91% 5,975 E. New irrigation 200 230 130% 5,152 F. Fruit tree development in vegetable area 231 260 b 3,540 G. Geothermal greenhouses 25 29% 53,460 a/ The SAR simply indicates that the FRRs for all the models were above 50%. b/ This model does not have any negative cash flow. c/ This model was not envisaged at appraisal. d/ At project full development Comments: The FRRs were calculated using an O&M cost for the irrigation infrastructures which is the mid-point between with artesianism (DO. 15/m3) and fully pumped water (D0.40/m3). Other things being equal, the FRR for each model will decrease as artesianism runs out and pumping becomes a necessity. Rougly half the deep wells in the project area are now operated with pumps.  - 22 - Table 6. Economic Impact Economic Rate of Return Appraisal Estimate PCR Estimate 21 % over 20 years 16% over 30 years Conversion Factors jUsed I /1 Foreign exchange 1.00 Unskilled labour 0.60 Domestic fertilizers 0.90 a/ Ref: FAO/CP Preparation Report of Oroject d'allegement de la pauvrete, March 1992, for the World Bank.  - 23 - Table 7. Status of Covenants As in Loan Agreernenta' Status Comments - Borrower to establish within OMVGM a project done But this unit never had any management unit legal standing and this limited seriously its efficiency. - OMVGM to provide to the Bank a plan for organization done Only since the creation of the and operation of extension services CRDA. But the incremental staff necessary was never provided - OMVGM to enter into management agreement witlh each Model agreements being used AIC by CRDA. - Borrower to prepare annual budgets done - Borrower to transfer OMVGM's vegetable and fruit Still under preparation seedlings and research activities - Borrowet to have special research and salinity done The IRA program was monitoring programs included in IRA and DRS cancelled for lack of progress - OMVGM to provide an action plan for marketing done Completed in 1988. But was services satisfactory to the Bank never followed-up - Borrower to maintain records and accounts of the project done For the loan's portion almost exclusively - Borrower to provide audit reports of annual project Since 1989 accounts - Borrower to maintain separate records of SOE to be Included in audit reports audited separately - Borrower to prepare schedule of repayment for the tree done only when uprooting was uprooting and incorporate resulting obligations in carried out by contract (see agreements with landlords para. 4.2 of Part 1) - OMVGM to provide audit reports of selected AIC done in 1992 accounts - DRE to monitor water balances, salinity in the area done And published yearly - Borrower to provide semi-annual progress reports done from 1987 to first part of 1992 - Borrower to prepare a completion report a number of elements were prepared, including farm surveys and models a/ The OMVGM (see Abbreviations) was amalgamated with the CRDA in 1990.  - 24 - Table 8. Missions Carried Out Stage of Month/ No. of Days in Specialization Performance Types of Project Cycle Year Persons Field Rating Problems Identification 10/83 2 2 TM (IE),DC - - Identiricationa' 12/83 3 1 TM (IE),AG, ECN - - Preparation 2/84 4 11 TM (IE) AG, DC, - - EC Preparation 7/84 4 17 TM (IE), AG, - - ECN, Irrigation Drainage Specialist Preappraisal 11/84 3 15 TM (IE) AG, ECN - - Appraisal 3/85 2 14 TM (IE) AG - - Supervision 12/85 1 3 TM (1E) 2 Financial Supervision 6/86 I 4 TM (E) I Supervision 12/86 1 5 TM (1E) I Project Management Supervision 5/87 2 5 TM (IE), AG 2 Project Management Supervision 12/87 1 5 TM (lE) I Supervision 4/88 2 5 TM (GE), AG - Supervision 11/88 1 5 TM (G E) I Development Impact Supervision 9/89 2 4 TM (IE), EC I Devel.impact, Technical Assistance Progress KEYS: TM=Team Leader, DC=Division Chief, IE=lrrigation Engineer, Ag=Agriculturalist, Ec=Economist, FA=Financial Analyst, SS=Soil Specialist, I=Problem free or minor, 2=Moderate problems, 3 =Major problems. a/ A feasibility study was conducted by "Centre national d'etudes agricoles". N.B.: Two special one-man missions (with extension specialists) were conducted in 1988 and 1990 for the purpose of looking at the situation and making recommandations in the field of agricultural extension.  - 25 - Table 8. continued: Missions Carried Out Stage of Month/ No. of Days in Specialization | Performance | Types of Problens Project Cycle Year Persons Field Rating Supervision 9/90 1 6 TM (IE) 2 Project Development Objectives, Availability of Funds Supervision 11/90 1 4 AG (ext/research - specialist) Supervision 12/91 1 4 TM (IE) 1 Availability of Funds Supervision 10/92 1 5 TM (IE) I Completion 8/93 2 13 AG, ECN KEYS: TM=Tearn Leader, DC=Division Chief, IE=Irrigation Engineer, AG=Agriculturalist, EC=Economist, I= Problem free or minor, 2=Moderate problems, 3=Major problems. N.B.: The agronomists in 4/88 and 11/90 supervision missions looked at the situation and made recommendations in the field of agricultural extension.  - 26 - Table 9. Staff Inputs Preparation Appraisal Negotiations ARIS Supervision PCR Total F'Y 84 14.9 14.9 FY 85 31.1 27.1 4.6 62.8 FY 86 0.4 7.0 7.4 FY 87 5.2 5.2 FY 88 0.1 7.3 7.4 FY 89 4.7 4.7 FY 90 4.2 4.2 FY 91 6.4 6.4 FY 92 4.8 4.8 FY 93 3.2 3.2 FY 94 9.4 9.4 Total 46.0 27.1 5.0 0.1 42.8 9.4 130.4  - 27 - ANNEX 1 AGRICULTURAL DEVELOPMENT Introduction 1. Agricultural development in some 5,000 ha rehabilitated for irrigation by the Project was to be achieved through: a) increased availability of irrigation water; b) improved drainage to reduce salinity; c) thinning of too dense or unproductive stands of date palms and other fruit trees; d) their replacement by more productive and profitable fruit crops and vegetables; e) the adoption of improved cultural practices suitable for the area; f) resumption of cultivation of abandoned fields, either by returned absentee owners or through alternative forms of management; and g) on-farm development through credit for such items as greenhouses, farm equipment. In addition, some 200 ha of rainfed land would be provided with an irrigation and drainage infrastructure enabling more intensive forms of land use. Irrigation 2. Despite a number of technical problems during Project execution, the Project's objective of providing the 42 scattered oases with sufficient water resources to enable more intensive forms of production were largely achieved. Moreover, the distribution network installed results in greatly reduced water losses in the conveyance system. However, in some areas, problems of water distribution still need to be resolved. Some farmers receive water at too long intervals for optimal production of summer vegetables. In other cases water demand can be met by extending the hours of pumping into the night, but farmers are unable or unwilling to practice night irrigation. To alleviate this problem, on-farm reservoirs are considered, with partial financing from credit. In another case, areas with different crop intensities were receiving the same irrigation schedule, which caused complaints from farmers having only fruit trees and no vegetables. An important achievement of the Project is that water charges are now to be paid by all members of the AICs, to cover at least the 0 & M costs. Farmers are therefore more aware of the need to economize water use. At field level, improved irrigation methods are needed to ensure optimal efficiency. These methods should be adapted to the individual farmer's needs, depending on soil type, tree and crop intensity and other factors. To meet this need, one oasis has recently been selected in the Project area to test and demonstrate methods to increase water use efficiency. This current work is done in the framework of a FAO/UNDP assisted Project for training on economic water use (TUN/91/002). In the early stages of the Project, work in this field had been initiated with the help of a consultant recruited by the Project. This resulted, interalia, in the adoption of improved irrigation practices for out-of-season vegetables, and lining of permanent field ditches in some farns. Applied on-farm trials in this field are needed and should include the comparison of  - 28 - various systems of intercropping and tree densities, under the main soil conditions in the Project area. Another shortcoming is that no initial winter leaching is applied in oases with high initial salinity (El Hamma and Zrig), although this is partly compensated by the growing of forage as a first intercrop. Drainage and Salinity 3. The Project achieved a large part of its objectives to provide drainage, where needed, to reduce salinity and to prevent its recurrence. However, in an estimated 10-15% of the Project area, salinity is becoming a problem after additional irrigation became available, for one of the following reasons: a) no drainage network was established due to farmer's initial objections; b) due to the low groundwater table at that time and on cost considerations, drainage was not considered essential at Project design; or c) insufficient or no maintenance caused drains to be clogged up, thereby allowing the water table to rise. Some farmers who initially resisted the construction of drains are now asking the authorities to rectify the situation, and funds have been made available at the Governorate level for this purpose. Similarly, funds are expected to be made available to construct drains where this was not considered necessary initially. Furthermore, Government funds are made available to carry out at least part of the required maintenance cleaning of drains. AICs are expected to bear full responsibility for this task in future. 4. During field visits, the Mission was made aware several times of the need for proper functioning drains. In some areas, residual salinity prevents cultivation of the more profitable fruit crops, such as apricots and apples. In this respect, the salinity monitoring program, carried out by the Soils Directorate in the Gabes CRDA as a result of an agreement under the Project, should be given adequate means to continue and intensify salinity monitoring in the field. It should also be provided with the means to analyze and interpret the collected information without undue delay. This work had started during project execution, and two reports have been issued thus far. A seminar or workshop is planned to be held in December 1993 to discuss the findings, to contemplate corrective measures where necessary, and to receive feedback on the need for expanded or intensified monitoring. It would be of interest for regional and local extension staff to be informed as well, and to discuss the potential of various crops under different salinity conditions in their respective areas. At regional and local level, AIC representatives should also participate in discussions on the findings of the monitoring program. Orchard Renovation 5. This has perhaps been the most controversial component of the Project. The older oases were originally fed by springs which led to the development of dense stands of date palms with only a relatively small number of productive trees. Ownership of the date palms was often shared among various heirs. As springs were gradually replaced by wells and artesian water made way for pumped water, the parts near the dried up springs were partly abandoned because of water shortage and salinity. Excessive land fragmentation was another factor leading to the  - 29 - neglect of parts of these old oases. More recently developed oases, meanwhile, usually have a wider spacing of date palms, with better maintenance and productivity. On the basis of a detailed analysis of date palm and fruit tree densities in all the oases covered by the Project, CNEA's Project preparation report recommended thinning of unproductive and low-productive date palms and other fruit trees and their replacement by more profitable or more productive species at regular spacing, to allow the maximum potential for vegetable and forage intercropping. Prior to the CNEA study, a detailed study of one of the oases in the Project area had been made (Bouchemma, 1978). It arrived at 46 palm trees/ha as optimal density to retain their effect on the micro-climate (slightly higher and lower extreme temperature values, reduced wind velocity, up to 10% lower evapotranspiration). 6. In total, some 270,000 date palms and other fruit trees were to be removed, of which some 50% in the El Hamma cluster of oases. To ensure their continued role as windbreak and microclimate regulator, minimum densities of 60-100 palm trees were to be retained, depending on their original density (some 350/ha in parts of El Hamma oasis). The proposed lower densities of date palms would minimize their competition for light, water and nutrients with the proposed more profitable intercrops. Similarly, in certain oases where olives were the dominant crop prior to irrigation intensification, their density would be decreased in favor of higher value intercrops. The financial benefits of lower palm tree densities and corresponding higher intercrop intensities are demonstrated in the farm models. 7. Although aware of many farmers' reluctance to date palm and olive tree removal in favor of other crops, the Government retained this basic approach to date palm thinning, which was in turn adopted at appraisal. To enable the earliest possible benefits stream resulting from Project investments, orchard renovation was to be carried out more or less simultaneously with irrigation rehabilitation. In order to ascertain the highest possible participation rate by farmers and orchard owners, an intensive public relations campaign was to be carried out from the start of Project execution. Mechanized uprooting through contract was adopted as the most efficient and speedy method wheirever tree density and other accessibility factors would allow it. The World Bank loan would finance part of the cost to be repaid by the farmers. In 1987 and 1988, some 3,000 date palms were uprooted in the Mareth area as part of this program. Extension agents were assigned to the area to assist in the reconversion program on a farm to farm basis, giving advice to farmers on farm development plans. Subsequent to the amalgamation of OMVGM with the CRDAs of the concerned Governorates, the orchard reconversion program was reconsidered and an updated survey was carried out to reduce the number of trees recommended for thinning. The new criterion applied was to retain 100 date palms in the interior oases and 80 per hectare in the coastal oases. Moreover, based on the new criteria uprooting would be limited to date palms, whereas the approach described in the SAR also included thinning of olive trees and removal of unproductive or other old fmit tree species. The outcome of the updated inventory and the application of the revised criteria resulted in a reduction of the estimated number of trees to be uprooted to about 1/3 of the original estimate. However, no detailed breakdown of the new inventory could be obtained by the Mission. To date, some 23,000 date palms have been uprooted, out of a total of some 50,000 according to  - 30 - the new estimate (uprooting of some 20,000 date palms was financed through Government subsidy). 8. The SAR projection that the total program of tree thinning could be executed during the six years of Project execution, in order to enable part financing from the loan, seems to have been highly optimistic". Farmers' resistance to uprooting had been experienced in the Southern Irrigation Rehabilitation Project, notably in the old oasis near Tozeur. The Govermment had also experienced considerable delays in olive renovation programs in the Sousse area due to farmers' often psychological rather than financial attachment to olive trees. Moreover, when the date palms still provide a certain yield, even at a low level, farmers often try to retain this low level of income until income from the newly established intercrops has reached a certain level. 9. Apart from farmers' reluctance due to psychological or financial reasons, a major factor which slowed down the renovation program is the high degree of joint and absentee ownership in some of the oases. The Mission was informed that in some oases a number of previously absentee owners have returned after irrigation water became available, and that thinning of dense stands of trees on such properties is expected to take place as farmers are establishing increasing amounts of intercrops. Moreover, there is an increasing interest among farmers to plant young date palm offshoots of higher market value where space is available. 10. It is understood that the discrepancy between the CNEA estimates and the updated inventory is largely due to the inclusion of young offshoots and "spontaneous" date palms (often in the form of clusters of seedlings rather than the planted offshoots used for productive trees) as well as other unproductive fruit trees in the CNEA survey, whereas the updated survey is based mainly on adult single date palms. In addition, since the CNEA survey of 1983, a certain amount of spontaneous uprooting of trees has been going on without Government assistance or without having been otherwise recorded. 11. At appraisal, three alternative approaches could have been used to avoid tying up credit funds from the loan for orchard renewal, which were ultimately cancelled when it became clear that they were not needed for this purpose. One would have been to exclude from the Project all old, partly abandoned oases with a high salinity level, high date palm densities and complex land tenure problems (notably El Hamma, and parts of other oases, grouped in farm model A), and to limit the Project to areas with lower date palm densities. This approach presumably would have been difficult to accept by the borrower. Another possible approach would have been to include all the oases in a long or medium term rehabilitation program, of a duration of e.g. 25 years, whereby financing would be made available in successive 5-year slices. National staff would then have been able to make a more detailed socio-economic scrutiny of each slice, enabling more accurate cost calculations and assuring a larger degree of " While the disbursement projections (COSTAB) in the SAR include the full cost of uprooting during the life of the Project, the phasing of Farm Model A (SAR implementation volume, Annex 7) imply that for part of the farms uprooting would continue, beyond the 6 years of the Project.  - 31 - farmer participation as the benefits were becoming more tangible. Farmers could then have played a larger role in the planning and execution of the program. This approach, however, would have had the following two disadvantages: a) a more rapid and accelerated abandonment of the oases to be rehabilitated in the later stages of the program, due to the steady decline of the Jefara aquifer which provides about 50% of the Project area's water supply; and b) the loss of economies of scale in bidding and contracting for drilling, irrigation and drainage network construction, other works and equipment. Moreover, several knowledgeable spokesmen doubted that a more detailed socio-economic scrutiny would have made much difference in the achievement of the goals of orchard renovation. 12. A third possible approach could have consisted of the exclusion of the financing of tree removal through the loan altogether and to seek alternative sources of financing. This is what is happening now, whereby the farmers are receiving Government subsidies for tree removal. The subsidy level of D. 11 per date palm covers roughly 50% of the cost of tree removal (including removal of the stump). 13. While mainly in the older oases, delays in date palm and other fruit tree thinning has contributed to a slower than projected intensification process, in areas with relatively low tree densities, farmers embarked on tree planting and intercropping at a relatively fast pace. In 1992, the total number of fruit trees exceeded the number prior to project-generated irrigation improvements by about 50%, thereby reaching about 60% of the SAR estimate at full development. Similarly, intercrops increased by almost 1,000 ha or 70% of the SAR's estimated increase at full development. Research and Extension 14. The Project aimed at strengthening regional research and extension to provide the farmers with technical and farm managerial information adapted for local conditions. Moreover, channels would be strengthened to inform the Research Institutes of the major problems confronted by the farmers, to enable these Institutes to consider including such topics in their programs. 15. Prior to the Project, OMVGM had already established a network of extension staff for the irrigated areas under its responsibility. The Project would strengthen this network in the Project area through the appointment of 22 additional extension workers, 4 subject-matter specialists, the purchase of vehicles and light motorcycles, extension equipment, and the construction of additional extension centers. Detailed work programs would be prepared and the work would be organized using management elements of the "T & V" system. A total of 7 man-months of consultant services were foreseen to assist with the organization of the program and with training of local staff. Special attention would be given to periodic evaluation and monitoring of the effectiveness of the program.  - 32 - 16. OMVGM would enter into an agreement with the Research Institute responsible for applied research in the region (Institute for Research in Arid Zones=IRA) to establish an experimental program aimed at resolving priority problems confronted in agricultural development during project execution. Agreements were also to be established with the specialized Departments concerned on salinity monitoring and the monitoring of water resource availability and quality. 17. A consultant was recruited to assist with staff training, mainly in technical aspects such as irrigation water management. The duration of this mission was shorter than foreseen, and no communications specialist was recruited. Moreover, OMVGM was unable to recruit additional extension staff, and no systematic work program using T&V management elements was established. Moreover, existing staff continued to be in charge of various support activities, such as orchard renovation, input distribution (e.g. seed and seedlings, fertilizers) and plant protection. OMVGM provided seasonal credit in kind, and credit recovery was among the tasks of the field staff. After the amalgamation of OMVGM with CRDA in 1989, input distribution ceased to be among its responsibilities. Staff re-assignments took place and extension staff became responsible for both the irrigated and adjacent rainfed areas. 18. The Gabes Governorate is affected by the extension component of the World Bank assisted national Research and Extension Project for training, and guidelines on methodology and organization, including the use of audiovisual and other communications means, and monitoring and evaluation. Extension is no longer the responsibility of the Plant Production section within the CRDA, and in some cases field extension units have both Plant Production and Livestock Production oriented staff. The Director of Agricultural Production has overall responsibility for extension. His directorate includes eight specialized regional bureaux, including extension, plant protection, vegetable and fruit crop production. The former chief of agricultural development in OMVGM is entrusted with the task of extension coordination. The program has a strong emphasis on plant protection for the main fruit tree crops (pomegranates, apricots, apples) and vegetables. Demonstration plots are established at contact farmers' fields, with free provision of pesticides, and the provision of spraying equipment on loan. Treatment of epidemic infestations is often provided free of charge, in the expectation that farmers in the future will carry out such treatments on their own account. The Project area is served by the Gabes Governorate's 5 regional extension centers (CTV) at Gabes, Mareth, Methioua, El Hamma and Matmata. Each CTV covers a number of local extension units (CRA, "Centre de Rayonnement Agricole"). Of the 34 CRAs in the Gabes Governorate, 12 are at present vacant, their area being temporarily served by adjacent CRAs. The national extension project has embarked on a program of staff selection and training at the national level. Each selected extension agent would get specialized training in extension methodology as well as in technical subjects, for several weeks a year over a 5-year period. Extension in the Governorate is benefiting from the physical support (extension centers, vehicles) which was provided under the Gabes Irrigation Project, and from the national Project for training, organization and management. 19. At the field level the main weakness of extension in the Project area is the large amount of time spent by field agents on administering various assistance programs, such as the  - 33 - tree planting program (PAAF) and various financial encouragement programs available at the governorate level, in addition to the collection of statistical data and surveys (including surveys in the framework of the Project). However, starting from early 1993 ten field staff have been exclusively assigned to extension duties. In some cases, extension staff are restricted by the shortage of transportation, which in principle is made available 2-3 times a week. In other cases, extension staff have been able to have regular contacts with over 50 % of the farmers of their area, largely as a result of the proximity of the CRAs to the farmers, and to commercial and social centers. These regular contacts, as well as field days and regular meetings with input distribution salesmen, have led to an increased demand for fertilizers and plant protection support. 20. Initially, OMVGM had assigned a special role in technology transfer to the Zerkine Agricultural Development Support Station ("Station d'appui"). Almost all OMVs in Tunisia had established such centers, to develop adapted cultural practices and test varieties of vegetables and fruit trees under irrigation. The Zerkine Station carried out a number of trials, mostly on under cover vegetables. It also carried out the first trials on the use of geothermically heated water for heating of greenhouses in winter and early spring. This technique was subsequently further developed for the use of other geothermic well water in the El Hamma area and elsewhere in Southern and Central Tunisia with bilateral and multilateral cooperation. The Zerkine Station also produced large numbers of disease free seedlings and rooted cuttings for farmers in the area. Such tree propagation material initially was not readily available from other sources. 21. During the early stages of project execution large numbers of farmers visited the Zerkine Station. Observers agree that the field days held at this station were in particular instrumental in promoting out-of-season covered vegetable production. In the late 1980s early vegetables from the project area were exported to France under the trade name "Oasis de Gabes". 22. Upon the amalgamation of OMVGM with CRDA, and in partial compliance with the relevant WB loan covenants, trials and demonstrations at the Station were discontinued, and nursery propagation was scaled down. The Government has hired a consultant to determine the future use of all the OMV support stations in the country. Some may be fully or partly incorporated in the regional research networks, others may be retained for training or transferred to the private sector as part of the young technician settlement program. 23. On Research, the Project aimed at engaging the responsible research institute in establishing practical objective-oriented applied research. Two agreements were established; one with IRA and the other with the Soils Department for salinity monitoring. Equipment was also provided to the Water Resource department to expand its program of water resource monitoring. The salinity monitoring program faced some initial delays, and needs to be strengthened to allow timely analysis and interpretation of data. The agreement with IRA resulted in a number of trials on winter forage crops, but no more elaborate applied research program was established for the project area due to shortage of research staff and the orientation towards pastures and livestock production of most existing research staff. With bilateral cooperation, IRA has  - 34 - established a research centre at Kebilie, an adjacent Governorate, where research and testing is done, interalia, on plant protection of crops grown in the region. Due to the above constraints, the agreement with IRA was subsequently cancelled. No new attempts to involve other research institutes were made because a selection process of priority research at national and regional level had meanwhile been initiated through the WB-supported Research and Extension Project. 24. The importance of water resource monitoring by DRE should be emphasized in view of the need for data on well head output and its relation to maintenance of sustainable aquifer levels. DRE also monitors the artesian pressure in existing wells and assists in projecting the need for additional pumping or lowering of pumps. It is also the agency entrusted with the task to advise the Government on the risk of aquifer pollution, which is particular relevant in the Gabes area due to the existence of chemical industries. For example, it advised against the deposit of chemical residues at locations where there would be a risk of seepage into the Jefara aquifer. 25. In conclusion, it can be stated that the Project was unable to accomplish the extension and research tasks in the manner envisaged at appraisal. However, the Project was able to provide important support to agricultural development through demonstrations on covered vegetables, plant protection, water and salinity management. The subsidized distribution of fertilizers and other inputs, previously considered an essential component of extension by local staff, has been successfully taken over by private and cooperative sales outlets. The quality of fruit tree propagation material from private sources, however, remains a source of concern. Farrm Models 26. Farm models were developed which to a large extent are similar to the models in the SAR. One new model was included to demonstrate the viability of out-of-season vegetable production in greenhouses heated with warm water from deep wells in the El Hamma area ("geothermic" greenhouses). Areas under crops and yields were adjusted in the light of the farm surveys carried out by the Project monitoring and evaluation unit in 1991 and 1992, and field observations. Except for the geothermic greenhouses, full development is projected in year 12 in each model, while the phasing of the benefit stream is projected to have a delay of 2-3 years as compared to appraisal projections. A summary description of each model is given below and more detailed information has been put in the project file. 27. Farm Model A (1 ha). Without Project, this model represents neglected and partially abandoned stands of date palms with soils of increasing salinity, low and decreasing water availability and decreasing production. The assumed date palm density without Project (250) has been adjusted downward to 183 and the number of trees retained was adjusted upward to 125 from the SAR estimate in view of CRDA's updated survey. Similarly, the total area represented by this model was adjusted downward to 860 ha. With Project, improved irrigation and drainage would enable intercropping between wider spaced date palms which would include some 15 newly planted date palms of higher than average market quality. Fifty other fruit trees  - 35 - would be planted, and intercropped summer and winter vegetables, forage and industrial crops would cover about 0.5 ha. 28. Farm Model B (0.5 ha) represents 3,030 ha or 60% of the Project area consisting of oases, mostly in the coastal area, where date palm density was already lower (124 trees/ha) and intercropping already takes place without Project. However, a slight decrease in production would take place due to deteriorating irrigation and drainage. With Project, adequate water would be available for a higher crop intensity: date palms, olives and other fruit trees (100, 20 and 120 trees per ha respectively, winter and summer vegetables, forage and industrial crops covering 0.26 ha). 29. Farm Model C (1 ha, total 500 ha or 10% of project area) represents oases which consist of olive plantations with little or no supplemental irrigation without Project. With Project, the density of olives (for oil production) would be decreased from 100 to 60 per ha and intercropping would take place with other fruit crops and vegetables (0.5 ha). 30. Farm Model D (1 ha, total 230 ha or 4% of project area) represents those oases where old olive trees (30/ha) are intercropped with vegetables. The development with Project would consist of a further intensification, including the planting of 80 fruit trees per ha and the production of out-of-season vegetables including some greenhouses: 0.07 ha. 31. Farm Model E (1 ha, total 230 ha or 4% of project area) represents areas newly equipped for irrigation by the Project. Contrary to SAR expectations, no newly introduced crops would be grown (such as autumn potatoes, due to market constraints and non-availability of seed potatoes for the local farrners). Higher-value date palms would be planted on the borders of the farms. 32. Farm Model F (1 ha, total 260 ha or 5 % of project area) represents a few areas where vegetables are grown without Project, partly irrigated from shallow wells. Lack of drainage and insufficient water thus far prevented establishment of fruit crops, which would be planted at a density of about 90 per ha. Winter and summer vegetable cultivation would also be intensified. 33. Farm Model G: Geothermic greenhouses (0.05 ha), total 25 ha. Deep wells rehabilitated in the El Hamma area produce water of a high temperature. Following experiments carried out over several years, private investors have established greenhouses adjacent to these wells and adjacent to the oases irrigated by these wells. The warm water circulates through the greenhouses and is subsequently used for irrigation in the oases concerned. In the cases of high artesian pressure, certain amounts of water in excess of the winter irrigation requirements is lost. The out-of-season vegetables (cucumbers, tomatoes, watermelons, peppers and others) are to a large extent exported. The cropping sequence used as demonstration in this model consists of autumn tomatoes followed by early melons. Full development is expected to be reached in year 5, when some initial problems (e.g. seed or seedling use for autumn tomatoes, nematode problems) are expected to have been resolved through research efforts.  - 36 - Agricultural Production 34. Based on the estimates in the farm models and the areas they represent, the total incremental production of the Project at full development (year 23) would amount to some 7,000 t of dates (including 6,000 t of better quality dates), 11,000 t of fruit, 24,000 t of vegetables and 8,600 t of out-of-season vegetables, including 4,500 t, produced in geothermal warm-water greenhouses. The main difference in the estimates of the SAR consists of a considerably higher incremental production of out-of-season vegetables, no decrease in olive production, and an incremental production of better quality dates. 35. Due to a slower than anticipated completion of irrigation and drainage works (2- 3 years), slower orchard reconversion and the persistent land tenure constraint, the phasing of benefits is retarded with full production being projected for year 23 after the start of the Project. Unless efforts towards land consolidation are made, it is likely that even at full development some 10-15% of the Project area, especially in the old oases (model A), will not be intensively cultivated as a result of land fragmentation, absentee ownership or joint family land ownership.  - 37 - ANNEX 2 FINANCIAL AND ECONOMIC ANALYSIS Financial Analysis 0.1 The seven farm models prepared in the PCR are described in Annex 1 and more details are available in the documentation put in the project file. Internal Financial Rates of Return (FRR) have been calculated for all the models. The farmgate prices used are given in Table 1 of this annex. All the investments (mainly uprooting/planting) and production costs have been included at their full value, to reflect the real financial viability over the long-term, ie. when the Government does not subsidize uprooting and planting any more. The FRR for each model, as well as the net incremental income per ha and per year at full development, are shown in Table 5 of Part m. 0.2 The SAR simply indicated that the FRRs for all models were above 50%. The FRRs for all the PCR models are also quite high (above 50% over 20 years), except for dates intensification (20%) which was a cropping pattern meant to be phased out in the SAR, precisely because of its low potential value and its longer gestation period. The FRRs were calculated using an 0&M cost for the irrigation infrastructures which is the mid-point between with artesianism (D 0. 15/m3) and fully pumped water (D 0.40/m3). Roughly half the deep wells in the project area are now operated with pumps and this proportion is naturally increasing. Other things being equal, the FRR for each model will decrease as artesianism runs out and pumping becomes a necessity. Farners will gradually adopt the cropping patterns implicit in the most profitable models. Although the FRR for the geothermal greenhouses is relatively low at 29 %, the net incremental income per ha/year at full development is of the order of $53,000. But the investments necessary are of the order of $190,000 per ha which is out of reach for most small farmers. Economic Analysis 0.3 In the SAR, the project's internal Economic Rate of Return (ERR) was calculated at 21% over 20 years. The ERR has now been recalculated at 16% over 30 years, mainly because of the seven-year delay in the full development of the benefits compared to the SAR (see Table 2 and documentation in project file). However, the negative impact of the delay is partly compensated by the fact that the total project costs are lower by 40% (estimated to be US$ 34.9 million compared to US$ 58.3 million in the SAR) and that the market prices for the products are higher in constant terms (after inflation) than expected at appraisal. For the off-season exports, the revenues in dinars have also increased due to the depreciation of the currency vis-a- vis the clollar.  - 38 - 0.4 The exchange rate of the Dinar has been allowed to depreciate from US$ 1.33 in May 1985 (date of SAR) to US$ 1.01 in September 1993. Furthermore, most of the protective customs duties and taxes in force in 1985 have now been substantially reduced. In these conditions, much of the economic distortions existing at appraisal time are no more present. In line with recent SARs" produced by the Bank, international prices have therefore been converted in dinars at the going commercial exchange rate. Exported goods have been valued at their CIF import prices. The project's total incremental benefits are the summation of all the incremental benefits for all the seven models when phased over the first twelve years of project life. These benefits are net of the farmers' on-farm investment and production costs - whether in cash or credit through the BNA - and shadow-priced on the basis of the conversion factors shown in Table 6 of Part III. 0.5 The project's total incremental costs include the reduced loan amount ($19.2 million), a rough estimate of the Government's counterpart funds (mainly the financing of few additional staff plus the O&M of the additional buildings and vehicles), and all the recurrent/replacement costs after the investment period. No residual value is assumed at project end. 0.6 The calculated ERR does not include other tangible benefits which are also important. Before project start, a number of farmners - reportedly around 1,200 families exploiting some 800 ha - had abandoned their lands, following the problems of water shortages, salinity and aging of their trees. It is now said that about 25 % of them have returned with more families expected to follow in the years to come. The return of these families will no doubt contribute (although modestly) to the economic growth of the Gabes region. This is in line with the overall objective of the project which was to stop the decline of agriculture in the southeast of Tunisia. Although the "without project" situation on the lands which were half-abandoned and abandoned is reflected in the farm models, the families' overall economic impact on the project area, through an income/job multiplier, is not. 0.7 The project also served to establish the principle and practice of the water users (AICs) paying the O&M of the irrigation infrastructures put at their disposal, thereby helping guarantee the sustainability of irrigation. It is estimated that about 80% of the needed O&M is now being financed by the AICs. The remaining 20% is either not carried out or is taken care of by the CRDA. In retrospect, the SAR cost-recovery target was too ambitious, ie. 100% of the O&M costs plus 5 % for the constitution of a reserve fund to finance the replacement costs. On the other hand, only 3,000 date palms were uprooted under the project approach calling for the farmers' payment of 50% of the cost (either cash or by credit). The cost of the other 20,000 trees uprooted was entirely subsidized by the Government but, in any event, farmers' resistance to uprooting (see main report) has been the bottleneck rather than the financing of it. The benefits derived from (and the cost of uprooting) these 20,000 trees are included in the calculation of the ERR since this would not have taken place without the project. " For example; Second Forestry Development Project, April 1993.  - 39 - Table 1. Financial Prices (Farmgate) Dinars/Ton Dates, Low Quality 400 Dates, High Quality 600 Pomegranates 278 Olives 220 Apricots 450 Apples 660 Henne 2,500 Peppers 500 Carrots 120 Green Onions 266 Marrons, early (tunnel) 700 Tomatoes (early, greenhouse) 330 Tomatoes 190 Melons 250 Alfalfa 80 Autumn Tomatoes 520 Spring Melons 950 NB. Given the absence of noticeable trade distorsions (including in the exchange rate), the financial prices of the outputs have been used as approximations of the economic prices, for the purpose of this PCR. The economic value of the main inputs has been calculated on the basis of the conversion factors given in Table 6 of Part RI.  - 40 - Table 2. Economical Benefits and Costs (in thousands of Dinars) Year Incremental Incremental Net Benefits" Costs Benefits I 0 37 -37 2 -31 3,005 -3,036 3 -74 4,907 -4,981 4 -123 3,518 -3,641 5 -83 2,418 -2,501 6 54 5,786 -5,732 7 372 1,642 -1,270 8 1,015 613 402 9 1,853 330 1,523 10 3,115 4,548 -1,433 11 4,586 330 4,256 12 6,045 330 5,715 13 7,354 330 7,024 14 8,431 330 8,101 15 9,354 330 9,024 16 10,128 330 9,798 17 10,833 330 10,503 18 11,456 330 11,126 19 11,915 330 11,585 20 12,244 330 11,914 21 12,446 330 12,116 22 12,563 330 12,233 23-30 12,610 330 12,280 ERR = 16% The incremental benefits are net of the famers' investment and production costs.   IMAGING Report No; 14444 Type: PCR