Mauritania - Livestock Development Project

Document of The World Bank FILE COPY FOR OFFICIAL USE ONLY Report No. 3208 PROJECT PERFORMANCE AUDIT REPORT MAURITANIA LIVESTOCK DEVELOPMENT PROJECT (Credit 273-MAU) November 19, 1980 Operations Evaluation Department This document has a restricted distribution and may be used by recipients only in the performance of their official duties. Its contents may not otherwise be disclosed without World Bank autherlation.  FOR OFFICIAL USE ONLY PROJECT PERFORMANCE AUDIT REPORT MAURITANIA LIVESTOCK DEVELOPMENT PROJECT (Credit 273-MAU) TABLE OF CONTENTS Page No. Preface I Basic Data Sheet ii Highlights PROJECT PERFORMANCE AUDIT MEMORANDUM I. Project Sumary .......................................... 1 1. Project Preparation and Appraisal ................... 1 2. Project Implementation ................. ......... 2 3. Project Impact ............................ 3 4. Bank Performance .................................. 4 II. Main Issues .......................................... 4 1. Project Design and the Pastoralist System .......... 4 2. Lack of Post-Project Financing for Maintenance ..... 9 3. Economic Rate of Return ............................ 9 PROJECT COMPLETION REPORT I. Background ............................................... 13 II. Project Formulation and Appraisal ........................ 14 III. Project Implementation ................................... 20 IV. Project Costs and Financing .............................. 29 V. Project Impact .......................... ................ 33 VI. Financial and Economic Results ........................... 34 VII. Institutional Performance ................................ 35 VIII. Bank Performance ......................................... 38 IX. Conclusions .............................................. 39 Map 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 PERFORMANCE AUDIT REPORT MAURITANIA LIVESTOCK DEVELOPMENT PROJECT (Credit 273-MAU) PREFACE This is a performance audit of the Mauritania Livestock Development Project for which Credit 273-MAU was approved in November 1971 in the sum of US$4.15 million and closed, fully disbursed, on March 31, 1979. The audit report consists of an audit memorandum prepared by the Operations Evaluation Department, and a Project Completion Report (PCR), dated November 1979. The PCR was prepared by the Western Africa Regional Office on the basis of a country visit in December 1978. The audit memorandum is based on a review of the President's Report (No. P-989) of October 20, 1971, the Appraisal Report No. PA-102a of October 15, 1971, the Credit and Project Agreements dated December 17, 1971 and the PCR; correspondence with the Borrower and internal Bank memoranda on the project issues as contained in the Bank files have also been consulted and Bank staff associated with the project have been interviewed. Two previous OED reports were also consulted.! The draft report was sent to the Borrower on August 25 for comments, but none have been received. The audit finds the PCR a thorough and critical document, properly covering the project-s principal achievements and shortcomings. The points discussed by the audit have been selected because of their relevance to this and other livestock projects in pastoral regions in Africa. 1/ OED, PPAR on Mauritania Drought Relief Fund Project (Credit 444-MAU), Report No. 2808, January 2, 1980, and OED, Delays in Project Implementd- tion, Report No. 2946, April 11, 1980, Case Study on the Mauritania Livestock Project.  - ii - PROJECT PERFORMANCE AUDIT REPORT BASIC DATA SHEET MAURITANIA LIVESTOCK DEVELOPMENT PROJECT (CREDIT 273-MAU) KEY PROJECT DATA Appraisal Actual or Item Expectation Current Estimate Total Project Cost (US$ million) 5.95 5. Overrun (%) - 9'- Credit Amount (US$ million) 4.15 4.15 Disbursed 4.15 4.15 Date for Completion of Physical Components 06/30/76 12/31/78 Proportion Completed by Above Date (Z) 80 Proportion of Time Overrun (%) 60 Incremental Economic Rate of Return (%) 19 21/b Cumulative Estimated and Actual Disbursements (US$ thousands) FY73 FY74 FY75 FY76 FY77 FY78 FY79 Estimated 1233 2352 3287 4100 (sic) - - - Actual 90 570 900 1360 3020 3460 4150 Actual/Estimated (%) 7 24 27 33 73 83 100 OTHER PROJECT DATA Original Actual or Item Plan Revisions Est. Actual First Mention in Files - - 11/06/68 Negotiations 09/23/71 - 09/23/71 Board Approval 11/02/71 - 11/02/71 Credit Agreement Date 12/17/71 - 12/17/71 Effectiveness Date 04/01/72 - 05/02/72 Closing Date 12/31/76 05/31/76 12/31/78 Borrower Islamic Republic of Mauritania Executing Agency Ministry of Rural Development Fiscal Year of Borrower January 1 - December 31 Follow-on Projects none MISSION DATA Month/ No. of No. of Man- Date of Item Year Weeks Persons Weeks Report Preparation I 02/69 2 3 6 06/05/69 Preparation II 11-12/69 3 4 12 05/15/69 Preparation III 09/70 0.3 2 0.6 11/27/70 Preparation IV 02-03/71 4 1 4 05/26/71 Appraisal 03-04/71 3 2 6 10/15/71 Total 12.3 28.6 Supervision I 05/72 1.4 1 1.4 07/05/72 Supervision II 11/72 1 1 1 01/25/73 Supervision III 04/73 2 1 2 05/25/73 Supervision IV 09-10/73 1.4 2 2 01/25/74 Supervision V 06/74 1.4 2 2.8 08/13/74 Review Mission 11/74 1.4 2 2.4 03/14/75 Supervision VI 03-04/75 1.3 2 2.6 04/28/75 Review Mission 06/75 0.3 1 0.3 07/08/75 Supervision VII 12/75 0.6 1 0.6 02/02/76 Supervision VIII 03/76 0.9 1 0.9 05/05/76 Supervision IX 09/76 0.9 1 0.9 11/03/76 Supervision X 01-02/77 1.3 1 1.3 02/22/77 Supervision XI 08/77 0.9 1 0.9 09/28/77 Supervision XII 03/78 1.4 1 1.4 04/05/78 Completion 12/78 0.7 1 0.7 02/17/79 Total 16.9 21.2 COUNTRY EXCHANGE RATES Name of Currency (Abbreviation) Franc de la Communaute Financiere Africaine (FCFA) Ouguiya Mauritanien (UM)- Year: Appraisal Year Average Exchange Rate: US$1 = 275 CFAF Intervening Years Average US$1 = 45 UM Completion Year Average US$1 = 45 UM /a On basis of costs to date. On basis of costs to complete construction of all wells originally planned, the overrun would have been 30%. /b Very little data are available on which to base this calculation; see paras. 2.17-2.25 for an explanation of the revised calculation. /c Since May 1973.  - iii - PROJECT PERFORMANCE AUDIT REPORT MAURITANIA LIVESTOCK DEVELOPMENT PROJECT (Credit 273-MAU) HIGHLIGHTS The project was to restore infrastructure essential for livestock activities, which had badly deteriorated during the previous 10 years. Major project components were: the reconstruction and/or repair of approximately 300 pastoral wells; the establishment of three well maintenance teams; explor- atory drilling for water within a sector then water scarce (and construction of 15 new wells contingent on the results of the exploratory drilling); strengthening of the Livestock Service to enable regular animal vaccination and treatment in the project area; and provision of equipment for maintenance of approximately 6,500 km of firebreaks. The original project goals required substantial modification because of the 1972 drought and because of major implementation problems and delays. The facilities actually provided have proved useful; however, their re- estimated return is low. Project start-up suffered considerable delays because: i) a severe drought in 1972 required all surviving animals to be temporarily removed from the project area, ii) contracts for a preliminary well survey and tender documents were delayed because of misunderstandings on the Bank's procedures for competitive bidding, and because the Bank had seriously underestimated the prices which contractors would propose, and iii) establishment of the adminis- trative and financial arrangements proved difficult. Only after the second- ment by IDA, in 1976, of a Project Coordinator did project implementation accelerate. Achievements fell short of the original 1971 targets, but close to the revised 1975 targets. It is estimated that 60-70% of livestock were lost in the 1972 drought; the major impact of the project was therefore to help herds recover. The vaccination program reduced animal mortality, but its precise impact is unclear. The benefits of the firebreak program were reduced by the drought, which diminished the forage which could burn. Only 137 wells were reconstructed or repaired; a larger number of wells required reconstruc- tion instead of repair than was originally expected. The distribution of wells actually constructed or repaired under the project is much better, however, than those initially planned. At project closing, the Hydraulics Department had 12 well-equipped teams to carry out well maintenance in the three administrative regions covered by the project. The following points may be of special interest: - appraisal design proposed continued government development and maintenance of livestock infrastructure, without alloca- tion of grazing rights and maintenance responsibilities to individual transhumant tribal groups (PPAM, paras. 1.11; 2.01- 2.11); - iv - - project assumed that infrastructure established would be main- tained by government through resources obtained via increased livestock taxes, which were already resisted by herders and were subsequently eliminated by Government as a result of the 1972 drought damage (PPAM, paras. 2.15-2.16); - rate of return is hard to re-estimate because of very limited data on project impact. Evidence suggests that the original assumptions have since become much less optimistic, particu- larly as a result of drought effects, although the project has doubtlessly contributed to the recovery of livestock in the area (PPAM, paras. 1.10, 2.17-2.25); - the necessity of checking carefully whether the works to be carried out under contract are in fact suited to contracting (PPAM, para. 1.12; PCR paras. 9-11, 28, 29, 36-38); - the necessity of planning for a longer start-up period in view of local administrative, financial, physical and procurement conditions (PCR paras. 25-28, 36, 45, 65-67, 79); - the need to reduce the number of ministries and departments involved in decisions regarding project implementation (PCR paras. 28, 29, 39, 65, 77); - the importance of the frequent and regular supervision which probably saved the project (PPAM, para. 1.13; PCR paras. 28, 29, 71, 77, 79); and - impact of 1972 drought was catastrophic and required a reorien- tation of project, which was not fully achieved (PPAM, paras. 2.18; PCR paras. 54-60). PROJECT PERFORMANCE AUDIT MEMORANDUM MAURITANIA LIVESTOCK DEVELOPMENT PROJECT (Credit 273-MAU) I. PROJECT SUMMARY 1. Project Preparation and Appraisal 1.01 Livestock herding has traditionally accounted for most of agricul- tural output in Mauritania. In early 1967, an IBRD technical mission identi- fied several key limiting factors to increased livestock production. In 1968, PMWA suggested a possible livestock project emphasizing disease control. Project preparation was entrusted to the FAO/IBRD Cooperative Program (CP), which carried out two preparation missions. The first mission proposed a project which would reconstruct and/or repair 250 wells in the southwestern part of the country, and vaccinate all cattle in this area against rinderpest and CBPP..1/ The second mission proposed a similar, but slightly different project which would construct 60 new wells and repair 500 wells on force account, establish six teams to maintain the wells in the project area, vaccinate all cattle in Mauritania against rinderpest and CBPP, build and equip a small diagnostic laboratory near Nouakchott, and establish and equip teams to rehabilitate and maintain the existing national firebreak network. The Bank sent a mission in September 1970 to discuss whether: 1) well con- struction should be carried out by private contractors instead of on force account, 2) the specific coverage of the animal health subcomponent, and 3) the inclusion of the firebreaks component. 1.02 The project was appraised in March/April 1971. The appraisal mis- sion decided that, under the current nomadic system, it was not feasible to introduce improved methods of herd management and animal husbandry. The principal project goals were instead to avoid the sharp decline in livestock production, which would otherwise occur as existing livestock infrastructure deteriorated in the absence of effective maintenance, and to permit an increase in herd size. The mission thus recommended the reconstruction and/or repair of 300 existing pastoral wells; the establishment, equipment and train- ing of two well maintenance brigades, exploratory drilling for underground water within a small sector not previously tested; vaccination against rinder- pest, CBPP, anthrax, blackleg, and botulism, and provision of equipment and maintenance of a network of firebreaks within the project area. The estimated project costs amounted to nearly US$5.0 million. The Credit proposed amounted to US$3.4 million. 1.03 The Loan Committee was concerned whether to require that additional taxes be levied on livestock to cover the services provided to livestock 1/ Contagious Bovine Pleuropneumonia (CBPP). - 2 - owners under the project, but concluded that the Credit Agreement should include a covenant under which Government would undertake to provide the funds necessary for the maintenance of the wells and firebreaks and the continuation of the animal health program. During negotiations, the Credit was increased from US$3.4 ti IS$4.15 million to provide for an expanded wells component, for an increase in operating costs for the health component, and for an increase in contingencies. It was also agreed that the Ministry of Equipment might be able to carry out by force account, in lieu of contract, up to 15% of well repair work. The Credit was approved in December 1971 and became effective in May 1972. 2. Project Implementation 1.04 Project start-up suffered considerable delays. Because of the severe drought, nc aiimals remained in the project area in 1972; both the animal health and firebreak maintenance components were delayed one year. Contracts for the preliminary wells survey and the preparation of tender documents were delayed because of misunderstandings on the Bank's procedures for inviting competitive bidding, and because the prices proposed by contrac- tors were higher than expected. Long negotiations and a remodeling of the contract followed, so that surveys could not start until early 1973. Estab- lishment of the administrative and financial arrangements for project imple- mentation also proved difficult. 1.05 Because of start-up difficulties, several revisions in project design took place from 1972 to 1975. These included agreement that the repair and reconstruction of about 200 wells (as opposed to 300) would be undertaken on force account, and that 15 new wells would be constructed by contractors. The nature and extent of well repair, as opposed to reconstruction, had proven unforseeable because every well was in a different state of disrepair and it was impractical to pre-determine the exact nature of the work by inspecting each one. A consulting firm was invited to provide two well diggers and two mechanics for the well repair and reconstruction program for two years, and to renegotiate the existing supervision contract of this program. 1.06 The project completion date was postponed to June 30, 1977 for both the animal health and firebreak components, and to June 30, 1978 for the hydraulic component; the closing date of the credit was postponed until December 31, 1978. 1.07 By June 30, 1978, 137 pastoral wells had been completed (68 repaired, 44 reconstructed, and 25 newly constructed). The smaller number of wells requiring repairs was due to the unexpected geographical dispersion of the wells, as identified by the consultants. Most of the wells which required repair were found in the western part of the project area, where water is plentiful but the danger of overgrazing very real. Thus, only a limited number were repaired, these being judged sufficient for this area. The high - 3 - number of reconstructions, in other areas, occurred because degradation was unexpectedly great on a number of wells initially chosen for repair. Imple- mentation was also hindered by difficulties in obtaining funds through the local financial administration and reimbursement procedures, and because of poor management of teams, material and equipment within the Department of Hydraulics. Implementation was further impeded by considerations related to the location of wells to be rehabilitated or constructed, with various offi- cials in the Government favoring different areas. Only the secondment by IDA in 1976 of an expatriate project coordinator made it possible to enforce a rational choice on all parties and accelerate the well rehabilitation program. Work was also carried out by relatively inexperienced teams, and the technical assistance provided by the consultant was of uneven quality. The execution of physical works was frequently disrupted by difficulties in obtaining adequate financing, and management of project financed teams proved inadequate in various areas. 1.08 The animal health component began in 1973/74 and proceeded to 1976/77. The campaigns were conducted regularly, but not always with the expected thoroughness. Since the percentage of cattle vaccinated against rinderpest (55-70%) never achieved the 75-80% level required for elimination of the disease, the veterinary service continued to vaccinate all classes of animals after the second campaign instead of merely those less than one year old. Treatments against parasites were implemented sporadically so that their effect on herds was too dispersed to offer significant results. 1.09 The firebreak component also began in 1973/74. Contrary to the appraisal forecast, it was never necessary to maintain the entire existing firebreak network in the project area. Because of drought, pastures were insufficiently high and dense in two-thirds of the area to pose a risk of bushf ire propagation. Moreover, the existing network design was found un- satisfactory because the layout of many firebreaks did not take into account the direction of the prevailing winds. 3. Project Impact 1.10 Approximately 800,000-850,000 cattle were in the project area at the time of appraisal; the project was expected to permit a herd increase to slightly more than 1 million animals. However, the 1972 drought caused losses estimated at 70% of the herd, reducing it to 250,000. The drought caused considerable shifts in population, with large numbers of people fleeing rural areas to the surrounding towns and places near food distribution centers. Numerous herders abandoned livestock activities and sought other occupations. The project, therefore, contributed mainly to a recovery of livestock activi- ties. Due to a combination of heifer and cow purchases from the southeastern part of the country, and the reduction in mortality and off-take, the cattle herd had recovered to about 450,000 by 1979. The sheep and goat population, which suffered similar losses, appeared to have regained its previous level. The fact that the region, seven years after the drought, has only one half the animals previous to the drought, suggests that it is animals and not (water) pasture, which has been the limiting productive resource in recent years. Nonetheless, many pastures are still well below their previous conditions and it is not clear when, or if, they will fully recover. It is therefore doubt- ful that the number of cattle will soon reach its past level. 4. Bank Performance 1.11 The project as designed at appraisal provided for the rehabilitation of previously existing livestock infrastructure which had deteriorated sharply during the 10 years following independence. The project sought to both fore- stall a decline in livestock activity, which would otherwise result, and to permit an increase in livestock production. No efforts were made, however, to improve herd management or pasture utilization. Similarly, the approach used, by supporting existing government construction and maintenance of wells which substituted for traditional tribal responsibilitie-, aay contribute to a long run deterioration of the traditional pastoral system. In the short term, however, individual tribes had neither the equipment nor the expertise to repair and maintain modern wells. 1.12 The appraisal overestimated the Government's implementation capac- ity, leading to many implementation problems. The project suffered from the absence of a strong project unit free of other administrative duties, and the need for coordination among several ministries and operating agencies, each of which was involved in decision making, unnecessarily complicating matters. Finally, the costs of implementing the wells component through bidding pro- cedures were significantly underestimated. The work was poorly suited to contracting, and the need to redesign this component of the project was a source of considerable delay. 1.13 Supervision of the project was good. Particularly noteworthy was the project review mission of November 1974. With very strong management support, this mission resulted in a number of remedial measures, including the appointment of a project coordinator, which permitted the project to reach a much more satisfactory conclusion. II. MAIN ISSUES 1. Project Design and the Pastoralist System 2.01 This was the first livestock project the Bank supported in West Africa. Its design was carried out over a substantial period, and involved two FAO/IBRD Cooperative Program missions, a Bank preparation mission, a Bank appraisal mission, and substantial involvement from the Bank's Loan Committee. Such time and effort involved reflect the lack of a clear strategy regarding -5- how best to improve traditional pastoral systems and how to transform effec- tively a traditional into a modern society. This is still the case today; this type of project remains among the most difficult to design and imple- ment, despite 10 years of experience and research. Despite this, however, it must be concluded that the final project design was not fully satisfactory. The main project deficiencies regarding the strategy adopted to develop pastoralist systems are discussed in this section. Further, Central Proj- ects staff have taken the view that, although additional studies might have improved project design, the principal deficiency was failure to take suffi- cient account of the uncertainties in behavior of the institutional, physical and cultural systems within which the project had to be implemented and to reflect these uncertainties in a more flexible project design. Since this was the first Bank involvement in livestock development in West Africa, it may be argued that many of the difficulties could not have been foreseen. Explicit recognition of this might have improved project design since it is clear, with the benefit of hindsight, that the project, like its antecessor, has poten- tially harmful as well as beneficial effects and that the harmful effects received much less attention. 2.02 The project was not very innovative. It was conceived to follow the initiatives of the French colonial government which, between 1950 and 1960, had established a well system to open new areas to regular pasture use (or permit more intensive use of areas already used), had implemented a widespread system of firebreaks, and had established an animal health service for live- stock vaccination. This infrastructure had deteriorated significantly since independence (11 years), due principally to reduced financial resources on the part of government and lower administrative capabilities. The project was expected to return these services to their previous level, and somewhat more, and it was hoped that a mechanism could be found to ensure the con- tinued maintenance of these services at their higher level after the project ceased (see paras. 2.15-2.16 for further discussion of this issue). However, although the project was expected to resolve the financial and administrative problems in the short run, it did not adequately deal with the financial issue in the longer run nor, and probably more importantly, did it question the role which the infrastructure being established was expected to play. The latter issue will be addressed first. 2.03 Nearly two-thirds of Mauritania-s population is dependent on live- stock production. The overwhelming majority of these producers are trans- humant herders who move their cattle, sheep and goats over the natural grasslands in regular response to seasonal rainfall patterns. Each year, after the rainy season in the more arid northern area has passed and pastures there begin to decline, migration to the richer pastures in the south begins. Migration continues, particularly in the drier years, to the banks of the Senegal River (or beyond), where cattle graze on crop stubbles of the seden- tary farmers. These farmers benefit from the manure left behind and the reduction of crop residues; the herders obtain forage and access to the river. The return of rains in the north permits a return to those pastures later in the year. - 6 - 2.04 Although the pastoral system appears primitive at first glance, it has evolved over many generations of experience and reached an equilibrium which permits rather efficient use of the available forage, whose seasonal variation within each area is very great. There are no economic means by which supplemental forage can be produced. The establishment of wells in different areas, however, permits the use of grasslands which otherwise, for lack of natural watering holes, would have had much less grazing value. They there- fore permit an increase in the herd. The major problem was to ensure that the wells dug, and the associated grasslands grazed, were adequate to support the herd along the whole migratory path, even though some variation in the forage available could be accommodated by the herd, or by varying stock numbers. The balance between water, forage, and animals was rather complex.1/ 2.05 Although the grasslands were operated on a communal basis, indi- vidual tribal groups had a claim on lands in different regions. This claim was largely the result of tradition, arrived at through complex means, but it also depended on the control of pastoral wells. Most wells were in areas where water was available at fairly shallow depth, usually in the southern area. These wells could be dug and maintained with fairly rudimentary tech- nology, although the work involved was nonetheless substantial. Each tribal group was responsible for the wells in the areas which they traditionally used, and each had an incentive to undertake this effort because the pasture surrounding was recognized to be its own. Thus, the need to dig and maintain the wells introduced a significant degree of private control into the system. Each individual tribe could control its own members, who grazed their animals in semi-joint fashion, and thereby limit the overstocking which normally results from communal grazing systems. It was accepted, however, that herders could cross the lands of others on their annual migrations and could enter and use the pasture areas serviced by the wells of other tribes when drought had made their own pasture areas uninhabitable. The frequent occurrence of drought in one area or another meant that each tribal group would, under specific conditions, cooperate with others. Such a system was their only insurance against frequent disaster. 2.06 The introduction of central authority under the colonial government improved this traditional system in several dimensions. For example, animal vaccination campaigns, carried out among all animals, could be carried out at lower cost and offered the potential for control of major epidemic diseases which limited private action could not have done. Similarly, the system of firebreaks, spreading across the entire area, could only have been planned and implemented by a central authority../ 1/ CPS points to the fact that the above-mentioned equilibrium had been seriously disturbed long before this project due to a number of programs, inter alia the ones mentioned in para. 2.06 below. 2/ It is still unclear, however, whether the costs of the endeavor were justified by the benefits. - 7 - 2.07 However, the introduction of central authority weakened the tradi- tional pastoralist system in some fundamental aspects. Government construc- tion of new wells introduced a fundamental reordering of the land resources available, and changed the grazing rights potentially available to different tribal groups. For instance, the government could expand the well system using modern machinery, thus opening many areas to more intensive grazing. Wells could be constructed in areas where water was encountered at relatively deep levels; wells of larger capacity and improved design and construction were also feasible. But these wells had to be maintained once constructed; if not, they deteriorated and were gradually lost to service. Individual tribes had no equipment to maintain modern wells, nor did they have the expertise. Although the government provided materials to herders for maintenance, after independence the wells system deteriorated badly. During project prepara- tion 70% of the wells examined needed repair. The Bank estimated that 38 wells were going out of production each year, or roughly 4% of the total. The actual situation encountered during project implementation proved to be even worse. 2.08 The change in maintenance responsibility also converted the wells into communal resources. Construction and maintenance of wells by the central authority began to reduce tribal responsibility and to weaken tribal authority over grazing areas, and through this, to induce an increase in herd size which was not in balance with the natural resources available. This was probably the issue of greatest potential danger. Although some expansion was doubt- lessly feasible and desirable, no alternative system for restraining animal numbers in the whole region was developed../ The new system thus made it likely that significant overgrazing would occur with time, and this over- grazing would make the impact of drought, whenever it occurred on a widespread basis, much worse.2/ 2.09 The Bank project accepted the desirability of an expanded public wells system and further herd growth. While the Bank believed that, for sociological, climatic, and economic reasons, a communal range system was the only feasible mechanism for use of such pasture resources, it also believed that livestock management could not be improved within the context of trans- humant conditions. Thus, it sought to maintain and expand the area within which traditional practices were possible. The Government had long since undertaken responsibility for providing pastoral wells, and the Bank saw no alternative to maintain an expanded area except through continued Government effort. While all of these conclusions seem sensible, the complexity of the traditional system, and particularly the problems imposed by use of communal ranges, are not easily resolved. 1/ CPS states that at the time of appraisal no proposals could be made because it was then considered that such measures could not be enforced successfully under the circumstances obtaining in Mauritania. 2/ The Region states that the wells are not equipped with motorized pumps and consequently have a low discharge rate which per se does not permit the build up of excessive herds in their vicinity. -8- 2.10 Insufficient information is available to determine conclusively the degree to which the expansion of livestock wells through Government initia- tive has reduced the efficiency of the traditional pastoralist system, but a reduction appears a likely possibility. The 1950-60 well expansion program promoted excessive herd growth, and overgrazing, leaving the livestock system more vulnerable. The impact of the 1972 drought was probably more severe as a result. 2.11 The 1972 drought intervened before the Bank-financed wells could be implemented. Once project implementation began, special care was exercised by the Project Coordinator to ensure that the wells constructed or rehabilitated provided water resources which were well balanced with regard to pasture availability. Since the number of livestock which can be maintained at each pastoral well is limited by the amount of water that can be drawn manually or by animals, the danger of overstocking (because of the common range) is reduced from what it would be if the wells were equipped with windmills or engine-driven pumps. Nonetheless, without a means to allocate the areas which have been opened, overgrazing in the long term remains a concern. 2.12 Given that this was the first Bank livestock project in West Africa, and that there was considerable uncertainty with respect to the institutional, physical, and cultural systems within which the project had to be implemented, This audit agrees with CPS that the inclusion of a formal monitoring component would have been valuable. 2.13 After a decade of experience, there is a general feeling in the Bank that it has been more difficult to make significant contributions to tradi- tional pastoralist livestock systems in Africa than was originally expected. Perhaps the most important lesson is that it is very difficult to make piece- meal additions to any traditional system which has achieved equilibrium over a long time span without creating some important imbalances. Thus, extreme caution is probably advisable before large scale changes in such systems, such as the very large well program was designed to achieve, are proposed.!. 1/ In March 1975, a supervision mission recommended the establishment of a development planning team which would study the means for improved use of livestock, range and water in Mauritania. It suggested that some form of long term grazing controls were needed, in addition to improved range management practices, and proposed that funds be reallocated to this purpose from the wells component. Although it appears that this might have been a sensible alteration in project design, the Government declined, indicating that a study similar to that suggested was already being financed by USAID. A Bank technical assistance project subse- quently provided a livestock specialist as part of a team seconded to the Ministry of Planning. Although his performance proved to be poor, with the result that his contract was terminated, a new livestock specialist is to be hired as part of the second technical assistance project team. -9- 2.14 The Bank has also considered shifting its emphasis toward semi- intensive and intensive livestock development along the Senegal river. Although such a proposal might be economically attractive, the riverine area is important - and probably crucial - to the transhumant herders who, during the dry season, need access to crop stubble and river water. If the sedentary farmers there are transformed into agricultural/livestock producers, the transhumant herders may find that their own equally important system will be rendered inviable.!/ 2. Lack of Post-Project Financing for Maintenance 2.15 The financial and administrative difficulties of the system adopted under the project were not sufficiently resolved. The livestock sector con- tributed a major proportion of the GNP. The services which it was to receive through the project, and continued maintenance of the same in future years, were expensive. The Government had to raise funds to finance the project. Yet there were little means by which the livestock sector could be made to pay for the benefits which it would receive. A livestock tax had already been imposed, but it was strongly resisted by herders and significant efforts were made to evade the tax.1 Animals were exported illegally, and fear of government tax collectors had made the vaccination programs difficult to implement. Nonetheless, the Bank pushed hard at appraisal for a commitment from the Government for an increase in livestock taxation which would provide funds with which to maintain the works which were to be built or repaired under the project. 2.16 During negotiations this approach was altered. A commitment from the Government that funds would be obtained for maintenance of livestock infrastructure was obtained, but the commitment did not specify the origin of these funds. The Bank recognized that it could not force the Government to levy new taxes, or ensure collection of the taxes if levied, and thus the post-project financial situation was left unresolved. The Government clearly wanted to continue the measures financed by the Bank loan, but it also had insufficient funds. The final project design did nothing to alter this situation. After the project ceased, the Government would, regardless of its intentions, be in the same situation as before the project. Thus, the impact of the project was simply to postpone the need to secure dependable domestic financing. It does not appear that systematic efforts were made during project implementation to arrive at any viable long run solution. 1/ A similar issue has been raised in two other recent OED reports: PPAR Ethiopia Amibara Irrigation, Report No. 3035, June 18, 1980, paras. 10-11 and, particularly, in PIER Sudan Roseires Dam, Report No. 3051, June 30, 1980, paras. 2.12-2.16. 2/ The livestock tax was eliminated after the 1972 drought, to assist the badly damaged livestock sector. - 10 - 3. Economic Rate of Return 2.17 The appraisal mission, which found little data on which to calculate the project's expected rate of return, estimated that the latter should be 19.6% on the assumption that Mauritania would continue full maintenance of the investments established under the project for 30 years. If such maintenance ceased at the end of the project, the estimated rate of return was 14.6%. The appraisal report also reported that the wells component was estimated, by itself, to provide a 15% return, while the vaccination and firebreaks com- ponents were to jointly provide a 28% economic rate of return. 2.18 The PCR provides a very useful discussion of project impact, and of the financial and economic results achieved, but stops short of re-estimating the project's rate of return. The PCR argues that the impact of the 1972 drought, which impc.sed great changes on project implementation and design, also made it impossible to calculate the rate of return because essential data regarding livestock numbers, production, and marketing are simply not avail- able. Accordingly, the impact of project activities on these key paramaters is also unknown. 2.19 The PCR-s point is well taken. Nonetheless, there is considerable evidence indicating that the assumptions on which the appraisal report rate of return estimate were based, have since become much less optimistic. Many factors have intervened to this effect. First, although project investments were maintained through June 1978, the fiscal situation has since worsened progressively due to the war; maintenance has suffered accordingly. No follow-up project has been implemented. Maintenance will likely suffer for several years, at least. 2.20 Second, the appraisal calculation for the firebreak component assumed that 40% of the forage would be lost during the dry season due to bushfires and that the reduction in lost forage by the firebreaks would permit an increase of 20% in the herd carrying capacity. In fact, it appears that the estimate of forage lost to fires in the pre-project situation was exag- gerated. Moreover, the decline in the height and density of pasture following the 1972 drought meant that bushfire propagation would have occurred much more rarely during the last eight years even without the firebreaks. Further, the establishment of fire fighting teams, equipped with light vehicles for move- ment of personnel to directly provide fire control, appears to be an essential supplement to the firebreaks. This element of human-mechanical control, and the heightened sensitivity of the rural population to fire safety precautions, appear to be more important factors than the firebreaks. Finally, the decline in the herd, due to the drought, meant that forage was not a constraint on herd production during most of the recent period. Firebreaks have served frequently as transport roads, thus providing a benefit which was not con- sidered at appraisal, but it is reported that their use as roads has also led to an increased number of fires. These factors would all suggest that the return to the firebreak component was very low. It appears that it would have been wise to reduce or even abandon the firebreak component following the - 11 - drought, recognizing that the danger of fire was very low. The funds so saved could have been used better to other purposes. Although about two thirds of the existing firebreak network did not require maintenance, and did not provide any benefit, total firework expenditures were 15% above those origi- nally planned. 2.21 Third, although the vaccination program was reasonably implemented, fewer animals were vaccinated and the program's coverage was systematically less complete than originally expected. Because of the incomplete coverage, a much larger proportion of the herd had to be repeatedly vaccinated against rinderpest than had been planned - at higher cost. Similarly, the irregular treatment of animals against parasitic diseases, on a limited basis, suggests that the program's impact in this area was also small. The vaccination program, which would otherwise surely have been completely stopped because of a lack of funds, did indeed prevent any major outbreak of epidemic dis- eases,' but it would have had to prevent the deaths of about 10,000 animals each year simply to recover its costs. This audit doubts that its impact was so strong.1 2.22 Fourth, although the wells component has resulted in a much better balance between water supply and forage availability in the area where the wells were constructed, it has provided for fewer fully operational wells (137 vs 315) than was envisaged at appraisal. Part of the shortfall is due to higher costs resulting from inflation, which should be reflected in propor- tionately higher future project returns. But most of the shortfall was due to the need for more substantial well reconstruction than was expected, and to organizational and operational inefficiencies. Thus, on an equal cost basis, the wells will have provided water to about 67% of the area originally expected, with consequently lower productive impact. 2.23 The 1972 drought, however, reduced the herd in the project area from about 850,,000 to 250,000 head. Pasture quality suffered badly, but it recovered more rapidly than did the herd. The wells investments were expected to relieve a pasture shortage, and permit higher production than would other- wise occur, but pastures (water) proved not to be the limiting resource during the first seven years following project initiation (see para. 1.10). 2.24 Herd size has now recovered to about 450,000, which is 45% the herd projected in the appraisal report for the seventh year. Rebuilding of the herd, however, is not, in itself, evidence of the profitability of project investments. Some of the herd increase is due to the retention of animals 1/ The PCR mentions that several thousand animals were indeed lost in both 1975 and 1978 in Southeast Mauritania, a region not covered by the project. 2/ The PCR differs on this point, stating that the vaccination campaigns have had a high value. Insufficient information exists to resolve this disagreement. - 12 - produced from within the region, but this value is already inclulded in the estimate of production from the existing herd. The rest is due to purchases of animals from other regions. Even if real livestock prices are 50% above appraisal assumptions, low past and current herd levels indicate that the project's contribution during the first seven years will be well below that predicted. The first seven years have a strong impact on the estimated rate of return. 2.25 The carrying capacity of some ranges appear as if they will remain below their pre-drought levels. If so, each of the wells rehabilitated or constructed by the project will permit fewer animals to graze within its area than was originally estimated, reducing long run benefits. 2.26 The combination of these factors - drought, overoptimistic appraisal estimates, reduced post-project maintenance, imperfect ve cination coverage, and the problem of commons - suggests that a conventional re-estimate of the project's actual economic rate of return will be low. This audit estimates that the rate of return on project investments, as implemented, but without the impact of the drought, would have been about 11% had maintenance been continued after the first four years, and about 8% had such maintenance ceased. The rate of return, given actual project implementation and the effects of the drought, would be about 4% with continued maintenance, and about 2% without maintenance. Currently, the last estimate would appear to be the most representative of actual events. Given the scarce information available, all of the previous estimates are based on crude assumptions.!1 2.27 The estimated rates of return may not fully reflect the value of the livestock project to Mauritania. Following the 1972 drought, Mauritania's pastoral economy was badly disrupted and destitute. The assistance given by the project, and the demonstration of Government concern, were important to herder morale and to the rebuilding of livestock numbers. 1/ These estimates assume that, as a result of the drought, real livestock prices have been about 100% above those predicted at appraisal. Such high prices are assumed to gradually return to the levels predicted in the appraisal report as the range becomes fully stocked. In the Region's view real prices are about 120-150% above appraisal projections and also that there will be no reduction of this price level in the foreseeable future. It is also pointed out that due to the lack of pertinent data the ERR calculation appears too uncertain. - 13 - MAURITANIA LIVESTOCK DEVELOPMENT PROJECT (Credit 273-MAU) PROJECT COMPLETION REPORT I. BACKGROUND 1. The Livestock Development Project was the Bank Group's fifth opera- tion and first in the rural development sector in Mauritania. It was moreover the first livestock development project in West Africa. Financed by Credit 273-MAU (US$4.15 million), the project was designed to promote livestock development in the fourth, fifth and sixth administrative regions of Mauritania in the southwestern part of the country. 2. In the long term, the development of livestock production was seen to be of obvious primary importance to Mauritania because most of the country's usable land was suitable for this purpose only. Further, two thirds of the population were engaged in the livestock sector. As the other main developing sectors, i.e. mining and fishing, could be expected to employ but a small fraction of the country's work force, the standard of living of the majority of the population could be improved only through increased livestock production, at least in the short or medium term. 3. The growth of local livestock production was hampered by several factors. There was a lack of watering points, and maintenance of existing wells was poor, making it difficult if not impossible to use much of the available grazing land and thus limiting the number of cattle that could be raised in the country. The grazing potential was further reduced by periodic bushfires. Finally, livestock development was handicapped by the complete lack of extension services and highly insufficient sanitary practices which had sharply deteriorated in the preceding years because of a shortage of funds and qualified staff. 4. Credit 273-MAU was negotiated from September 23 - 30, 1971 and approved by the Board on November 2, 1971. It was signed in December 1971 and declared effective in May 1972. 5. This report on the Mauritania Livestock Development Project is based on the findings of Bank supervision missions; a report from OED review- ing the causes of delays in project implementation; a review of files; and the project coordinator's completion report forwarded to the Bank by the Mauritanian Government. - 14- II. PROJECT FORMULATION AND APPRAISAL A. Origin 6. In early 1967 an IBRD technical mission visited Mauritania at the request of Government. Its report, "Guidelines for a Four-Year Development Program for Mauritania", was probably the true origin of the project, because it emphasized the fact that livestock herding accounted for the greater part of agricultural income in Mauritania and specified that the principal factors limiting livestock production were: (i) the quantity and quality of pastures; (ii) the availability of water; and (iii) the pathology of the animals. Soon after, in 1968, a memorandum was issued by PMWA on a possible livestock project in Mauritania. It stated the need for and importance of a project in the livestock sector with emphasis on disease control which could be broadened into the fields of production and marketing. B. Project Preparation 7. Following a suggestion from the Bank's Area Department, project preparation was entrusted to the FAO/IBRD Cooperative Program. Two prepara- tion missions were then carried out, one in February 1969 and one in November/ December 1969. (a) The first mission proposed a project incorporating the following components: (i) reconstruction of 52 old wells and repair of 200 wells for stock watering in a project area covering administrative regions 4, 5 and 6 in the southwestern part of the country; (ii) vaccination of all cattle (estimated at 800,000) in this project area against rinderpest and contagious bovine pleuropneamonia (CBPP), and of some 200,000 cattle in the same area against botulism, anthrax and blackleg over a four-year period; and (iii) establishment of services to carry out the program during the following six years. (b) The second mission proposed a somewhat different project, incor- porating the following components: - 15 - (i) construction of 60 new wells and repair of 500 wells in the same project area on force account; (ii) training and equipping six maintenance teams to maintain the wells in the project area; (iii) an animal subproject involving annual vaccination of cattle through- out the country against rinderpest and CBPP and building and equip- ping a small diagnostic laboratory near Nouakchott; and (iv) establishing and equipping teams to rehabilitate and maintain the existing national firebreak network. 8. The first mission estimated the project's cost at US$1.53 million, and the second at US$7.1 million. Estimates for components were as follows: First Mission Second Mission CFAF 000 US$000 CFAF 000 US$000 Well scheme 188,000 752 1,259,000 4,578 Animal Health 194,100 776 630,000 2,291 Firebreaks - - 68,000 247 Total 382,100 1,528 1,957,000 7,116 Although the project proposed by the second mission was larger, the difference in cost estimates was mostly due to different estimates of the physical inputs required to perform certain functions, and of unit costs. 9. The main issues raised by the second mission were (i) the need for a well survey to provide technical justification for well replacement and an inventory of wells to be repaired; (ii) disease control; (iii) the need for a range survey to indicate the carrying capacity of pastures around the various wells; (iv) the organization and management set-up; and (v) the feasibility of establishing a livestock development unit. 10. The Bank's view, based partly on the findings of an economic mission to Mauritania in March/April 1970, focussed on a number of issues related to the proposals of the second mission. These were: (a) Under the FAO/IBRD program, Government would have had to agree to continue well maintenance, the vaccination program and maintenance of fire- breaks, in a manner satisfactory to the Bank, after full disbursement of the then proposed credit in project year four. The Bank did not believe that Government would be able to do this, since the estimated annual cost was US$851,000, far above the existing budgetary allocation for such activities; (b) The Bank considered that it would be more efficient to carry out well construction and repairs through private contractors than on force account; - 16 - (c) The Bank deemed that more detailed information would be needed to evaluate the work to be carried out under the well component; and (d) The Bank had serious reservations about the propos!d country-wide vaccination program. There was also some doubt as to whether a campaign to vaccinate against rinderpest and CBPP should have highest priority. 11. The next step was a visit to Mauritania by a two-member Bank mission in September 1970 which discussed the issues with Government officials and came to the following conclusions with regard to (b), (c) and (d): (a) The wells subproject should be modified along the following lines: (i) provision should be made for training and equipping two or three well maintenance teams in the southwestern region of the country, this part of the project to be coordinated with assistance from FAC; (ii) construction of the 60 wells as proposed in the FAO/IBRD report should be undertaken by contract rather than by force account; (iii) the greater part of the program of repairs on 500 wells should also be carried out by contract rather than by force account; and (iv) a rough estimate of the costs of the wells subproject as modified was: Construction of 60 wells at US$14,000 each US$ 840,000 Repair of 500 wells at US$3,500 each 1,750,000 2 or 3 well maintenance teams (for five years) 500,000 Contingencies (10%) 300,000 Total: US$3,390,000 (b) The animal health subproject should be confined to the same area as that of the wells subproject and would have to provide an adequate level of technical assistance to local services; and (c) The firebreaks subproject was suitable for incorporation in the project. 12. Because of the delays already incurred, the Bank requested Govern- ment to react expeditiously to the main issues still to be resolved so that agreement could be reached on the following points: (a) hiring of a well specialist to carry out a more detailed survey in the field and prepare, with officials of the Department of Hydraulics, general specifications for well construction and repair and preliminary cost estimates for these works; (b) modification of the animal health subproject; - 17 - (c) specification of the amounts Government was prepared to provide to cover the recurrent costs of the vaccination program and well maintenance after full disbursement of the Credit. 13. Government immediately reacted positively, and discussions were then held with FAC and the Bank to coordinate plans. FAC was to finance two well maintenance teams and provide technical assistance to the Department of Hydraulics. Meanwhile, FAC had also agreed to finance, together with Yugo- slavia, the Nouakchott laboratory and three French veterinarians who were to be in Mauritania at the time the project began. C. Appraisal 14. An appraisal mission visited Mauritania in March/April 1971. It was joined in the field by the hydrologist hired by the FAO/IBRD Cooperative Program, at the request of the Bank, to carry out a survey of the pastoral wells in the project area. The mission estimated that under the current nomadic system of herd management, it was not feasible to introduce improved methods of range management and animal husbandry. The mission considered, however, that to maintain production at the existing level, it was essential that (a) a satisfactory network of pastoral wells be maintained, (b) adequate measures be taken to protect livestock against the major epidemic diseases, and (c) provision be made for the protection of pastures from destruction by fire. Compared to the most recent project design, the only further major modification then proposed was that the number of wells to be repaired or reconstructed under the project should be 300 instead of 500. 15. The project as proposed by the appraisal mission was comprised of the following components, all within the country's fourth, fifth and sixth regions: (a) reconstruction and/or repair of 300 existing pastoral wells; (b) establishment, equipment and training of two well maintenance brigades; (c) exploratory drilling for underground water within a small sector of the project area not previously tested; (d) an animal health component involving provision of vehicles, equipment, vaccination yards and buildings to enable the Livestock Service of the Ministry of Planning and Rural Development to carry out a program of vaccination against rinderpest, contagious bovine pleuropneumonia, anthrax, blackleg and botulism; and (e) provision of equipment for establishment and maintenance of a network of firebreaks within the project area. Estimated total costs of the project amounted to nearly US$5.0 million. The Credit would then have amounted to about US$3.4 million, representing the foreign exchange cost of the project and a substantial amount of local costs. - 18 - D. Loan Committee Nexotiations and Board Approval 16. Issues raised by the Loan Committee concerned: whether to require that additional taxes be levied and collected on livestock within the project area to cover the costs of services provided to livestock owners under the project, as proposed in the appraisal report, or whether to seek from Govern- menran understanding that its yearly budget should provide for adequate resources to cover the cost of these services. The Loan Committee concluded that the Credit Agreement should include a provision of funds covenant under which Government would undertake to provide the funds necessary for the maintenance of the wells and firebreaks included in the project.and for continuation of the animal health program. 17. Negotiations took place from September 23-30, 1971. They led to an increase in the amount of the Credit from US$3.4 million to US$4.15 million because of (i) a provision for construction of 15 new wells conditional on the results of exploratory drilling for water; (ii) a provision for a third well maintenance team and an increase in investment costs per team to provide for the purchase of a stock of spare parts; (iii) a provision for operating costs of the well maintenance teams - three in lieu of two; (iv) an increase in operating costs for the animal health component to provide for the treatment of sheep, goats and calves against internal and external parasites; and (v) an overall increase in contingencies for both investment costs and operating costs from 10 to 20%. 18. The only other substantive change agreed to was that the Ministry of Equipment might be able to carry out by force account, in lieu of con- tract, up to 15% of the well repair work. 19. The Board approved the Credit on November 2, 1971. E. Targets and Goals 20. In retrospect, this type of project certainly cannot be regarded as innovative. The rehabilitation of disease control services, the reconstruc- tion or repair of a number of wells and the maintenance of firebreaks could merely restore the essential infrastructure needed for maintenance of live- stock activities and did not really constitute a new investment program for increased production or even for inciting herdowners to better management of herds and pastures. However, given the weakness of livestock services at that time, it was probably right not to have aimed at more sophisticated objec- tives. As it was assumed that without the project meat production would decline or even disappear, the project was aimed at at least maintaining livestock production and exports or possibly increasing them, thereby maintaining Or improving the standard of living of livestock owners. F. Project DescriDtion 22. The project approved by the Board was comprised of the following components to be implemented in the southwestern part of Mauritania: - 19 - (a) Wells. It was specified that most construction and repairs should be carried out by contract, and that consultants should be hired to make a detailed survey of the wells to be repaired or reconstructed, to prepare bidding documents, to make recommendations on the bids received and to supervise the execution of works. FAC was to provide three specialists who would be responsible for training the well maintenance teams and for supervis- ing their operations. (i) reconstruction and/or repair of approximately 300 pastoral wells; (ii) establishment of three well maintenance teams and imple- mentation of a program for regular maintenance of pastoral wells; (iii) exploratory drilling for water at 10 sites within a sector of the project area where there was inadequate information on groundwater supply; and (iv) construction of 15 new wells contingent on the results of the exploratory drilling. (b) Animal Health. This involved strengt4ening of the Livestock Service of the Ministry of Rural Development to enable it to carry out a campaign to vaccinate animals in the project area against the major epidemic diseases and provide regular treatment of sheep, goats and cattle against parasitic diseases. The following treatments were to be carried out during the campaigns: (i) vaccination of all cattle against rinderpest in each of the first two years of the project; (ii) annual vaccination of all cattle less than a year old against rinderpest after Year 2; (iii) annual vaccination of all cattle against contagious bovine pleuropneumonia; (iv) vaccination against botulism, anthrax and blackleg of all cattle in zones where outbreaks occurred; and (v) regular treatment of sheep, goats and calves against parasitic infestations. A qualified and experienced veterinary officer was to be stationed in the project area to supervise the animal health component of the project. (c) Firebreaks. This involved provision of equipment for regular maintenance of approximately 6,500 1m of firebreaks within the project area. Some parts of the existing network were to be reopened by bulldozer while maintenance was to be carried out with graders. - 20 - 23. Total project costs were estimated at approximately US$6.0 million with a foreign exchange component of US$4.1 million, or 69% of total costs. The breakdown of project costs by major categories was as follows: US$ '000 % foreign Equivalent Exchange Investments Wells 2,553 83 Animal Health 270 78 Firebreaks 215 100 Contingencies - physical 358 84 - price 250 84 Subtotal 3,646 83 Operating Costs Wells Maintenance Teams 520 9 Animal Health 1,029 55 Firebreaks 367 49 Contingencies - physical 195 59 - price 195 59 Subtotal 2,306 44 Total Project Costs 5,950 69 The IDA Credit was designed to finance the foreign exchange component of the investment costs and a declining contribution to the operating costs from 80% in Year I to 20% in Year 4. III. PROJECT IXPLEMENTATION A. Effectiveness and Start-up 24. Conditions of effectiveness were the establishment of a project coordinating commission and the hiring of an expatriate veterinarian to act as coordinator of the animal health component. These conditions were ful- filled in early 1972 and the Credit was declared effective on May 2, 1972. 25. Project start-up suffered considerable delays for various reasons: (i) because of the severe drought in 1972, no animals remained in the project area, so that both the animal health and firebreak maintenance components were delayed one year; (ii) contracts for the preliminary wells survey and the preparation of tender documents were delayed because of misunderstandings on the part of the Mauritanian authorities with respect to the Bank's procedures for inviting competitive bidding, and because the prices proposed by the contractors were higher than expected at appraisal. This situation led to long negotiations and remodeling of the contract with the result that surveys - 21 - could not start until early 1973; and (iii) establishment of the administra- tive and financial arrangements for project implementation proved difficult. 26. Retrospectively, the factor that appears to have had the most impact on project start-up and the project's overall performance, is that the appraisal report emphasized the establishment of a project coordinating commission but did not emphasize the necessity for a strong and independent project management unit. The result was that the director of the livestock department was appointed as project manager. Although he was one of the ablest of the local professionals, he proved unable to handle his project assignment for the following reasons: (i) he was frequently away from head- quarters on official missions; (ii) he had no alternate and his staff was not large enough; (iii) in addition to his regular workload, he had to implement the FAC and FED development projects. Obviously, with 'hc help of only one veterinary advisor and not even a qualified bookkeeper, the project manager did not dispose of the necessary means to make the project run properly. 27. Equipment contracts for well maintenance, the firebreaks program and the animal health campaign were awarded in December 1972, but the material was not delivered until late 1973. B. Revisions 28. Because of the difficulties encountered in starting up the project and making it run, several revisions of varying importance took place from 1972 to 1975. They were: (a) in mid 1972, the Mauritanian authorities, namely the Minister of Rural Development, requested substantial changes which would provide for the firebreak rehabilitation equipment to be operated by the Water and Forest Service of the Ministry of Rural Development instead of the Civil Works Service of the Ministry of Equipment. This was agreed to by the Bank in early 1974; (b) in May 1973, a supervision mission discussed the drilling sites with Government and the consulting firm which was carrying out the preliminary surveys on wells. An agreement was reached that the geophysical survey could be satisfactorily checked by test borings on only four sites instead of ten as foreseen at appraisal. In spite of the reduced number of test borings, costs were still to exceed appraisal estimates; (c) tenders for the construction and repair of vaccination yards were not called for until January 1974. The tender proved unsuccessful, as no qualified firm submitted bids. Consequently, IDA agreed to construction of the yards by the Livestock Department on force account, with the Department purchasing the required materials by negotiated contract from a local supplier; (d) the bids received in late 1974 for part A(1) of the project, recon- struction and repair of 236 pastoral wells, and for part A(4) of the project, construction of 15 pastoral wells, proved unacceptable to IDA. Bids were - 22 - received from only two Mauritanian firms, both of dubious competence, and the bids were about twice the amount originally budgeted. The tendering process failed for three main reasons. First, publicity was somewhat restricted. Second, international contractors were reluctant to operate in Mauritania because of customs and fiscal problems. Finally, the nature and extent of well repair and reconstruction work, as compared with well construction, were unforeseeable because every well was in a different state of disrepair, and it was impractical to determine the exact nature of the work by inspecting each one. Hence, bidders, if interested at all, could reasonably be expected to allow for a considerable price contingency. Thus, the supervision mission recommended that the project's management and technical services be appropri- ately strengthened and that parts A(1) and (4) of the project be carried out totally on force account by the Hydraulic Service of the Ministry of Equip- ment. 29. By late 1974, it had become clear to IDA that the project should be modified to achieve effective and expeditious implementation. Indeed, project implementation was satisfactory only for the animal health com- ponent. The firebreaks component was behind schedule. Particularly serious was the delay in water development -- the most important component of the project. The project was revised in November 1974 in order to accelerate its implementation. The proposed changes were discussed with Government in March 1975 and the following agreements were reached: (a) secondment by IDA of a project coordinator and a project accountant, each for 1-1/2 to 2 years; (b) repair and reconstruction of about 200 wells on force account instead of 64 and 236 by contractors; and construction of 15 new wells by contractors; (c) wells north of latitude 170 37' would be repaired or reconstructed only with IDA's approval; (d) Government would invite the consulting firm to: (i) negotiate a contract to provide two well diggers and two mechanics for the well repair and reconstruction program for two years, and (ii) to renegotiate the existing supervision contract of this program; (e) representatives of the Hydrauli, Livestock and Environmental Departments and the project coordinator would visit the wells to be repaired or reconstructed; and (f) IDA would finance 80% of the equipment costs of five new well repair and reconstruction teams and 80% of the operating costs of 11 teams during a two-year period. 30. By July 1975, the Board had approved the following amendments to the Development Credit Agreement: - 23 - A. The table in paragraph I of Schedule 1 is amended to read as follows: Amount of the Credit Percentage Allocated (expressed of Expenditures in dollar equivalent) to be financed Category Repair, Reconstruction and Maintenance 1,200,000 80% of total of Pastoral Wells on Force Account expenditures Exploratory Drilling for Underground 160,000 80% of total Water expenditures Engineering Consultants' Services 220,000 100% of foreign expenditures Equipment, Structures, Vehicles, 1,000,000 Graders and Bulldozers: (a) Imported Items 100% of foreign expenditures (b) Structures 70% of total expenditures Operating Costs of Maintaining Fire- 720,000 80% of total breaks, and Carrying Out Vaccination expenditures of Livestock Salaries of Project Coordinator, two 450,000 100% of foreign Well Diggers and two Mechanics expenditures Unallocated 400,000 Total 4,150,000 B. Schedule 2 is amended as follows: (1) Part A is modified to read "Part A: A program to improve pastoral wells, including: (1) reconstruction, repair, and maintenance of approxi- mately 200 pastoral wells on force account; and (2) exploratory drilling for underground water at four sites. (2) Parts B and C remain unchanged. - 24 - (3) A new Part D is added, as follows: "Part D: Employment of a Project Coordinator and an accountant, each for a period of 18 months." (4) The last sentence of Schedule 2 is deleted and replaced by the following: "The project is expected to be completed for the reconstruction and repair of pastoral wells by March 31, 1977, and for the animal health and firebreaks components, by March 31, 1976." 31. The project completion date was later postponed to June 30, 1977 for both the animal health and firebreak components, and to June 30, 1978 for the hydraulics component. Simultaneously, the closing date of the project was postponed until December 31, 1978. C. Physical Progress 32. All three components of the project were substantially delayed. Their implementation having proceeded rather independently, it is appropriate to assess the status of each one separately. The project was (technically) closed on June 30, 1978, with disbursements amounting at that date to 83% of the Credit. The Credit was finally fully disbursed on March 7, 1979, after the last supervision mission brought back withdrawal applications that exhausted the remaining Credit funds. On the basis of disbursement figures, the project was thus, at that date, almost three years behind schedule. 33. Wells. By June 30, 1978, 137 pastoral wells had been completed (68 repaired, 44 reconstructed and 25 newly constructed). If one assumes that the phasing of expenditures forecast in the appraisal report is proportional to the expected schedule of well repair and construction and, on the other hand, that construction or reconstruction of one well is almost equivalent in time and cost to repairs on four wells, the appraisal and revised work schedule estimates can be compared as follows with the actual rate of completion: Well Program FY 1973/74 FY 1975 FY 1976 FY 1977 FY 1978 Total App. App* App. Est. Act. Est. Act. Est. Act. Act. Act. Est. Rev. Act. Repair 90 7 90 10 90 16 29 6 270 170 68 Reconstruction 10 17 10 16 10 4 6 1 30 30 44 New construction 5 - 5 - 5 4 1 20 15 15 25 Total repair equivalent 150 75 150 74 150 48 57 90 450 350 344 - 25 - 34. Compared to appraisal expectations, the overall program actually completed represents slightly more than two years' work, suggesting an esti- mated total delay of almost three years. On the other hand, the amount of work actually carried out differs very little from the revised schedule established by the 1975 review mission. 35. The smaller number of wells requiring repairs was related to the geographical dispersion of the wells as identified by consultants. Most of those necessitating repair work were found to be in the western part of the project area where water is plentiful but the danger of overgrazing very real. On the other hand, the high number of reconstructions completed was due to unexpected degradation found on a number of wells initially chosen for repair. 36. A number of causes of delays were specific to the wells program; these relate to the appointment of consultants, the late decision to carry out well construction work by force account and the choice of wells to be rehabilitated. Another set of problems was of a more general nature and involved in particular (i) difficulties in obtaining funds through the local financial administration and reimbursement procedures; and (ii) poor management of teams, material and equipment within the Department of Hydraulics itself. 37. A team of consultants was appointed in January 1973 to survey existing wells and ascertain underground water availability in the project area, prepare tender documents for exploratory drilling as well as for the repair, rehabilitation and reconstruction of 315 wells to be designated. Signature of the contract with the consultant was delayed for a few months by the need to incorporate IDA's comments on an earlier draft, in particular to provide sufficient guarantee concerning the consul- tant's performance. Following a survey of 20 wells, the consultant prepared bidding documents for repair and reconstruction of 236 wells and for construction of 15 wells which were found unacceptable by project staff; revised bidding documents were approved by IDA in February 1974. Calls for tenders were advertised locally and - on a limited scale - internationally. Bids received (by July 1974) were totally unacceptable to IDA and were about twice the amount originally budgeted (see para. 28). 38. It took a number of months for the Association to accept the princi- ple of having the well program carried out by force account by the Department of Hydraulics of the Ministry of Equipment. By that time the project was already almost two years behind schedule; construction costs had risen and it proved necessary to reduce the well program from 300 to 200 wells so as to remain within planned budget allocations (para. 29). 39. Implementation of the well program was further impeded by con- siderations related to the location of wells to be rehabilitated or con- structed, with various officials in the Government favoring different areas. In this regard, only the appointment in 1975 of an expatriate project coordi- nator made it possible to enforce a rational choice on all parties and accel- erate the well rehabilitation program, although the pace of construction works never reached the level expected at appraisal. This had to be expected as - 26 - work was carried out by relatively inexperienced teams, and the permanent technical assistance provided to them by the consultant (two well-diggers and two mechanics) proved of uneven quality. 40. The e:ecution of physical works was frequently disrupted by dif- ficulties in obtaining adequate financing. Well construction and maintenance teams sometimes went on strike for not having been paid for several months, and the work of teams in the field was occasionally stopped because suppliers refused to deliver goods to 'the hydraulics service while waiting for payment of a previous bill. This resulted from (i) the acute shortage of local funds of recent years, and (ii) (indirectly) the shortage of adequate staff for management of the project. 41. The managem-Ant of project-financed teams and of supplies and equip- ment within the Depar%ment of Hydraulics (Ministry of Equipment) proved inadequate in various areas, mainly: (i) insufficient staff monitoring and recording of information, leading to misknowledge of hiring and leave dates, miscalculation of staff rights, frequent transfers of assignments, salary differences and so forth; (ii) inadequate planning for providing teams at work with the supplies or equipment required; (iii) poor control and monitoring of teams; and (iv) misadministration of supplies and equipment, leading to considerable losses. The result was chronic irritation of the staff and frequent technical interruptions which significantly reduced the efficiency of field works. 42. Animal Health. The animal health component started one year later than expected because of the severe 1972 drought which drove most domestic animals out of the project area. In subsequent years the campaigns were conducted regularly, although not always with the thoroughness expected initially. Number of Animals Vaccinated or Treated 1973/74 1974/75 1975/76 1976/77 Total campaign campaign campaign camvaign Vaccination against: Rinderpest 165,475 198,042 262,170 255,123 880,710 CBPP 114,350 104,694 186,487 255,076 660,607 Blackleg 3,170 21,447 54,902 75,229 154,748 Anthrax 2,026 76,652 4,533 3,912 87,123 Pasteurellosis 800 3,875 4,531 15,655 24,861 Botulism 79,759 49,076 144,389 191,305 464,529 Treatments against: Try-panosomiasis 11,565 17,676 12,510 8,592 50,343 Internal parasites 32,905 75,780 77,426 75,731 261,842 External parasites 41,943 33,552 28,746 22,189 126,430 - 27 - These figures show a regular increase in the animal health project component activities from the 1973/74 campaign to the 1976/77 campaign. However, the implementation of vaccinations against the main cattle diseases was not carried out exactly as designed at appraisal. Since the percentage of cattle reached was never more than 75%-80%, the veterinary service continued to vaccinate all classes of animals after the second campaign instead of merely those less than one year old. Vaccination against CBPP was not carried out systematically, except during the 1976/77 campaign. On the other hand, serum inoculations against botulism were systematic because of the strong demand on the part of cattle owners, although for economic reasons these are recommended in case of outbreak only. Finally, treatments against parasites were implemented spora- dically without preliminary planning so that their effect on herds was too dispersed to offer significant results. Despite these deviations, however, the impact of this project component was considerable as it maintained a rather sound health situation in the project area, thus allowing a rapid restocking after the 1972 drought. 43. Firebreaks. After a difficult start-up with a delay of one year in this component also, firebreak implementation proved satisfactory. Works completed during the four campaigns financed under the project are as follows: Firebreak Maintenance Firebreak Openings Actual Covered by Actual Covered by Length Machinery Length Machinery (km) (km) (km) (km) 1973/74 1,058 2,848 299 897 1974/75 2,199 5,007 162 486 1975/76 2,406 6,021 - - 1976/77 1,713 4,280 - - Total 461 1,383 Contrary to what was forecast at appraisal, it was never necessary to proceed with maintenance of the entire existing firebreak network in the project area. The need for firebreaks depends on the quantity of rain received during the rainy season and the consequential growth of pastures. It was found during project implementation that more than two thirds of the 6,500 km firebreak network in the project area did not require maintenance because pastures were not high or dense enough to pose a risk of bushfire propagation. Moreover, the existing network design was found unsatisfactory because the lay-out of a substantial number of the firebreaks (east-west or even southwest - northeast) did not take into account the direction of the prevailing winds which generally blow the same direction. - 28 - D. Reporting 44. Although the project director was made aware of the Bank's reporting requirements, he was unable to comply with them because of h:.s numerous other duties. Thus, not one implementation report was sent to the Bank until the project coordinator took up his assignment in early 1976. The latter provided the following reports: June 1976 - current status of project implementation Sept. 30, 1976 - quarterly report Dec. 31, 1976 - quarterly and annual report June 30, 1977 - semi-annual report Dec. 31, 1977 - semi-annual and annual rep'rt June 30, 1978 - completion report These reports, providing information on all main aspects of project imple- mentation and financing, were found satisfactory to the Bank. E. Procurement 45. Procurement was the source of a number of delays for several reasons: (a) Unavailability of professional staff able to deal efficiently with Bank guidelines on the preparation of tender documents and Bank requirements with respect to procurement procedures; (b) Government's favoring of local suppliers and enterprises; (c) Difficulties in finding contractors willing to undertake certain works; (d) Complexity of meeting local procedural requirements through the "Commission Nationale des Marches" and of numerous visas from both the financial administration and a number of ministries. 46. These problems led to a number of reviews of tender and bidding documents and further negotiations and delays in waiting for administrative approval. Changes in procurement procedures became necessary on two main occasions: in 1972, to prevent any further delay in implementation of the wells components, the Bank approved an alteration of procedure in tender documents aimed at hiring a consultant to carry out the preliminary well and hydrological survey. Later, after the project reappraisal, the Bank accepted that well repair and reconstruction works be entrusted to the Hydraulics Department instead of private contractors because of the failure of the tendering process. - 29 - IV. PROJECT COSTS AND FINANCING A. Cost and Financing 47. As the project was considerably delayed and went through major changes, actual costs were somewhat different from those expected at appraisal. A breakdown of actual and estimated project costs by major equipment, works and services is shown below: Actual and Estimated Project Costs Cost Estimates Actual Costs % of Expected Cateor7 (UM '000) (US$ '000) (UM '000) (US$ '000) Costs 12/ Wells Repair, reconstruction and construction of new wells 109,900 1,979 - - 0 2/ Exploratory drillings 16,660 120 6,903 153.4 127.8 Equipping of maintenance teams 11,340 240 51,793 1,151 564.2 2/ Consultant services 13,880 250 33,759 750.2 300.1 2/ Subtotal 141,780 2,553 92,455 2,054.6 80.5 Animal Health Equipment 15,000 270 13,648 303.3 3/ 112.3 Firebreaks Equipment 11,940 215 11,757 261.3 121.5 Consultant Services - - 13,108 291.3 - 2/ Contingencies Physical 19,880 358 - - Price 13,900 250 - - Subtotal Investments 202,500 3,646 130,968 2,910.5 79.8 Operating Costs Well Maintenance Teams 28,880 520 86,140 1,914.2 368.1 2/ Animal Health 57,140 1,029 30,029 667.3 64.8 Firebreaks 20,380 367 19,053 423.4 115.4 Contingencies Physical 10,820 195 - - Price 10,820 195 - - Subtotal Operating Costs 128,080 2 306 135,221 3,004.9 130.3 Total Project Costs 330,580 5.952 266,189 5,915.4 99.4 1/ Exchange rate assumed at appraisal: US$1 - 275 CFAF = 55 U.M. Average exchange rate was about: US$1 = 45 U.M. 2/ Large differences due to the reappraisal decision to build the wells on force account instead of by a private company. 3/ Including US$20,000 in building facilities. - 30 - 48. Actual costs can be considered reliable in the aggregate. They are the result of the project coordinator's reconstruction of accounting informa- tion based oi Cisbursement records of the National Treasury for local expenses and the Bank's data on foreign expenditures and direct payments. However, the figures shown above represent the results of "a posteriori" bill control and recording. The Bank could have had better statements of expenses for each project component and much more sound accounting by providing for a strong and independent management unit at appraisal with a professional accountant being able to control "a priori" foreseen expenses of the various departments involved in the project. 49. Comparisons4 of actual and appraisal estimated costs for each project component show tha: (a) Total actual costs of the wells component finally amounted to slightly more than appraisal estimates - US$3.9 million compared to US$3.7 million 1/ for a completion rate of almost 75% of appraisal expectations, i.e. an underestimation of some 30%. Still, it is necessary to specify that cost overruns can be explained by considerable delays, insufficient management and unforeseen changes; (b) The animal health component was apparently less costly than expected (US$0.97 million versus US$1.56 million 1/),.almost all of the difference lying in operating costs. In fact, the actual structure of animal health operating costs appears somewhat different than expected. There was a definite overrun in some categories (vehicle operation and maintenance: 172%; medecine: 118%; miscellaneous supplies: 155%), and a considerable underrun in others (salaries: 13%). Moreover, some categories that were expected to be financed under the project were financed under Government budget (vaccines and most salary expenses). 2/ (c) Firebreak component costs were slightly underestimated with regard to equipment (84% of actual costs), and slightly overestimated regarding operating costs (105%). The latter is fully justified by the fact that the average amount of work actually required and completed by year was carried out on approximately 40% of the total existing firebreak network, thus considerably reducing fuel needs. B. Disbursements and Financial Sources 50. The share of Bank and Government financing was reviewed in 1975 and the Credit was comprehensively reallocated based on the various changes. A comparison of appraisal estimates, the 1975 reallocation and final disburse- ments is shown below: I/ Including physical and price contingencies. 2/ Expenses related to the project are not included in actual project costs. Final Initial Allocation __Revised Allocation Disbursements Category Amount of Percentage of ex- Category Amount of Percentage of Final the Credit penditures the Credit expenditures to disbursements allocated financed allocated be financed (JS$ equiv) (US$ equiv) 1. Repair and recon- 1,400,000 80% of total ex- I. Repair, recon- 1,200,000 80% of total 779,314.84 struction of pas- penditures struction and expenditures toral wells maintenance of pastoral wells on force ac- count II. Exploratory 100,000 80% of total ex- II. Exploratory 160,000 80% of total 93,004.63 drilling penditures drilling expenditures III. Construction of 180,000 80% of total ex- III. Engineering con- 220,000 100% of foreign 244,773.00 new wells penditures sultans' ser- expenditures vices IV. Engineering 225,000 100% of foreign IV. Equipment, struc- 1,000,000 1,416,886.97 consultants' expenditures tures, vehicles, services graders and bull- dozers; (a) imported items 100% of foreign expenditures (b) structures 70% of total expenditures V. Equipmentsatruc- 615,000 V. Operating costs for 720,000 80% of total ex- 938,400.76 tures, vehicles, animal health and penditures graders and bull- firebreaks dozers: (a) imported items 100% of foreign expenditures (b) structures 70% of total expenditures VI. Operating costs 930,000 80% of total VI. Salaries of Project 450,000 100% of foreign 667,619.80 of maintaining expenditures Coordinator, two expenditures pastoral wells, in PYI, 60% well-diggers and two firebreaks and in PY2, 402 in mechanics animal health PY3 and 20% in PYl, VII. Unallocated 700,000 VII. Unallocated 400,000 Total 4,150,000 4,150,000 4,150,000 - 32 - 51. Compared to appraisal expectations, disbursements were very slow, representing only 45% of the Credit as of the initial closing date (see graph attached). The main reason for slow disbursement, as a number of supervision reports emphasized, was inadequate accounting practices in the three depart- ments charged with project implementation as well as in the p.oJect's central accounting unit attached to the Department of Livestock. Until 1976, sub- stantial expenses for the purchase of materials and for salaries were not recorded and were therefore not submitted to IDA for reimbursement. This explains the discrepancy between the rate of disbursement under the Credit and that of physical implementation. Withdrawal applications began to flow regularly in the controller's office only after the project coordinator assumed respon- sibility for the project's accounts and began to reconstruct these accounts from the beginning of the project. 52. Obtaining counterpart funds for the project proved to be difficult. To begin with, Government's fiscal position progressively worsened during project implementation as Mauritania's war effort intensified, draining most of the country's resources. Then, in 1976, 1977 and 1978, following a drop in foreign assistance, the budget deficit was such that severe cuts had to be made in the administrative budget. The creation, in 1972, of a special account for the project was not sufficient to ensure that funds would flow freely to their ultimate destination. Obtaining funds was made even more difficult by the numerous administrative channels involved. Approval from a series of Government offices was necessary to obtain the release of either local or foreign funds. In this context, a close follow-up of reimbursement requests as they filtered through the administrative network should have been a priority. However, as the position of project accountant was, for a number of years, either left vacant or occupied by a person of insufficient compe- tence, Government approval of bills was often an extremely lengthy process. This appears to have been particularly detrimental to project implementation as the same routine was required for even the smallest invoices. 53. Sources of funds were the following (U.M. '000): Government IDA Other Total 1972 24,636 - 38 3/ 24,674 1973 26,500 848 - 27,348 1974 35 3,725 - 3,760 1975 18,490 14,303 - 32,793 1976 - 65,722 - 65,722 1977 3,500 1/ 54,900 - 58,400 1978 25,000 2/ 41,490 - 66,490 1979 - 5,760 38 5,760 Total 98,161 186,748 38 284,947 USS Equivalent 2,181 4,150 1 6,332 ('000) 1/ Received from Government for continuation of firebreaks and animal health components after their completion. 2/ Received from Government for continuation of wells component after its completion. 3/ Received from sale of bidding documents. - 33 - V. PROJECT IMPACT 54. Basic data concerning the status of livestock in the project area were completely overturned by the severe drought of 1972 before project implementation started. From various sources, it is assumed that losses in the herds of the fourth, fifth and sixth administrative regions covered by the project involved some 70% of the cattle and 60% of the small ruminants. The drought also caused considerable shifts in population, with large numbers of people fleeing the rural areas to the surrounding towns and places near main roads and food distribution centers. In the meantime, numerous herders were forced to abandon livestock activities and seek other occupations. Thus the major impact of the project was not so much to develop livestock activities in the southwestern part of the country, as to help herds recover from the consequences of the drought by improving livestock living conditions in the project area. Also, by reducing the risks of animal mortality, the project contributed to restoring confidence in livestock activities on the part of herders. 55. With regard to the hydraulics component, it is assumed that the program actually completed, although not as large as expected at appraisal, is more closely adapted to the needs of the project area. As the selection of works to be carried out was made in strict accordance with livestock require- ments, the quality and expected duration of pastures as well as hydrological conditions, the project area not only has at least 137 more properly con- structed wells than before the project, but the overall distribution of existing wells is much better balanced with regard to pasture usage. Also, new pastures were opened by creating a dozen new wells inside or around the Aleg-Mali-Monguel-Kaedi zone where almost no water points existed before. Moreover, subject to Government compliance with operating cost financ- ing requirements, the Hydraulics Department now disposes of 12 well-equipped teams which are in a position to ensure well maintenance in the three adminis- trative regions previously covered by the project. Most vehicles and the majority of the heavy hydraulics equipment financed under the project were still fully operational at the time of project completion. 56. The project's impact was particularly significant in animal health. Without the project, the four campaigns carried out under this component from 1973 to 1977 would simply not have been possible because of lack of funds and inadequate equipment. In 1972, the livestock service only had some old, worn out vehicles bought in 1967 for the joint project campaign against rinderpest, and almost no residual freezing equipment for the conservation of vaccines. The existing equipment was generally inadequate and in very poor condition. Moreover, Government had no means to either re-equip the service or to ensure its operating costs. Thus, it was only with the assistance of the project that the vaccination campaigns were realized without major outbreaks. Furthermore, the expatriate coordinator of the animal health component had a large impact in reorganizing the Livestock Department staff and replanning the vaccination and treatment campaign at the beginning of the project. His plans were followed quite closely by his successors. - 34 - 57. The continuation of vaccination campaigns was essential, mainly to prevent the recurrence of rinderpest which can cause high mortality in individual h'rds. In terms of prevention of cattle mortality, the campaigns carried out under the project have certainly had very high returns. On two occasions, in early 1975 and in 1978, major rinderpest outbreaks occurred in the southeast of the country. Although numerous contaminated animals were found in the project area within commercial herds going to Nouakchott or individual herds fleeing the sites of outbreaks, very few animals native to the project area were actually contaminated. Consequently, mortality in the project area was limited to a few young animals which had not been vaccinated. In the southeast, on the other hand, where the animal health component of the FED-financed livestock project was not yet well implemented, several thousand animals were lost. 58. Although the benefits of the firebreak component were somewhat overestimated at appraisal, the rehabilitation and maintenance carried out during the project as required by pasture conditions certainly reduced sub- stantially the destruction of pastures by grass fire. Only a few small grass fires were reported from 1973 to 1978. 59. Other returns from the firebreaks component resulted from their use in communication and as transport roads between key wells and main towns. Before the Nouakchott-Kiffa concrete highway was completed, firebreaks were in fact the only means of communication and transport between Nouakchott and the southeastern part of the country. 60. It is roughly estimated that the number of cattle increased from 250,000 to almost 450,000 due to a combination of heifer and cow purchases from the southeastern part of the country and the reduction in mortality and off-take. The sheep and goat population appears to have regained its pre- drought level. Average pasture growth, however, has not reattained its previous level. As an example, the pastures around the towns of Boutilimit or Moghta-Lahjar, which were suitable for cattle before the drought, are now grazed almost exclusively by camels and goats, while cattle are found from 50 to 100 km further south. It is thus assumed that, if the present situation continues, the number of cattle will not reach its past level, except if a better integration between livestock and crops is developed along the Senegal valley. VI. FINANCIAL AND ECONOMIC RESULTS 61. Because of the overturn of all basic data as a result of the drought, the project's economic justification as assessed at appraisal is no longer valid as a basis for analysis. In particular, it was assumed at appraisal that the benefits of the well component would arise from the fact that the program of well reconstruction, repair and maintenance would pre- vent the virtual collapse of the existing network, leading to a progressive decrease in livestock of about 60%. As the drought itself drastically reduced herds by 60 to 70%, this assumption is no longer applicable for the medium term. - 35 - 62. Likewise, it was expected at appraisal that the firebreak program would increase the pasture available to livestock during the latter half of the dry season, thus enabling livestock owners to increase their herds in the project area by about 20%. Although overall pasture production was dramati- cally reduced because of the drought and will take years to recover, there is still, and there probably would have been even without the project, consider- able potential for increasing the size of the remaining herds. Thus, this appraisal expectation cannot be considered for the medium term either. 63. Because of the large losses resulting from the drought, many live- stock owners completely changed their activities and moved to the towns to seek jobs or Government assistance. Large herds of cattle or sheep and goats had become rare, and animal husbandry practices were closely adapted to the necessity of herd reconstruction. As a result, extraction for marketing was reduced to a minimum and marketing channels were totally disrupted while local prices of meat and animal products increased threefold. Under these condi- tions, appraisal estimates regarding the production of carcass meat and animal production values cannot be taken into account. 64. Actual project benefits are to be found in the fact that restocking of the project area was probably much more rapid than it would have been without the project. This is due to the fact that the project considerably improved and secured the living conditions of livestock, primarily by making available a more rational network of watering points and ensuring protection against main diseases. These benefits can hardly be quantified. Official Government data concerning livestock numbers, production and marketing are not reliable at all; and as the project did not include monitoring activities, we do not dispose of data concerning either the evaluation of herds or animal production and its value, mortality, fertility, extraction rates, etc. VII. INSTITUTIONAL PERFORMANCE AND DEVELOPMENT A. Government Agencies and Departments 65. Apart from technical and financial factors, the pace of project implementation suffered mainly from the lack of a central project authority. Although the project was officially under the jurisdiction of the Ministry of Rural Development - the Director of the Department of Livestock being the Project Director - the largest part of the project (namely, the wells program) was carried out by the Department of Hydraulics in the Ministry of Equipment. The two other project components were carried out by separate departments in the Ministry of Rural Development. Difficulties of coordination between these different agencies were particularly acute during the first few years of implementation (1973-75) as the Director of Livestock, charged with several other duties which were given precedence over the project, was unable to devote sufficient time to coordinating project execution. The Project Coordinating Commission, established in 1972 as a condition for Credit effec- tiveness to advise management on various matters, performed very poorly. Personal rivalries between senior officials and furthermore, between the - 36 - different ministries involved in the project (Planning, Rural Development and Equipment) contributed to the ineffectiveness of the Commission. After the project coordinator took up his assignment in early 1976, he organized, approximately once a month, meetings of an unofficial technical commission comprising the department heads of Hydraulics, Livestock, Water and Forests, and sometimes representatives of the Budget Department and the Ministry of Planning. This unofficial technical commission played an essential role in resolving a substantial number of administrative and financial problems. Composed of persons directly concerned with project implementation, it was, in the final analysis, much better suited to its purpose, and therefore more useful, than the official Coordinating Commission comprised of ministers and high officials who met only two or three times and were totally inappropriate for coordinating such a project. 66. The Treasury Department kept good records of disbursements from the project's special account opened at the beginning of the project. This complete recording, although not classified by category, was the basis for the coordinator's reconstruction of the project's accounts and for the filling out of withdrawal applications for Bank reimbursement. 67. The performance of the various Government departments was generally hampered by the cumbersome administrative procedures in Mauritania and by the lack of adequate accounting practices in the three departments charged with project implementation. The different departments also suffered from poor organization due to the lack of high level staff assigned to the project. This was particularly evident in the Department of Hydraulics, from which only one technician with an undergraduate diploma was assigned to the project. Although this technician was active, he was unable to handle by himself the overall organization of 12 teams divided into three "brigades", their sup- plies, staff management, equipment and recording of expenses and salaries, and all the more so because he often had to devote time to the programs implemented under FAC or UN financing. The same problem arose when the young veterinarian assigned to Boghe as coordinator of the animal health component also became Deputy Director of the Livestock Department in Nouakchott. Having to deal with administrative problems in the capital and help the Livestock Director in various areas, he did not have enough time to devote to the project's activities or close monitoring of the teams working in the field. B. Accounts and Audit 68. From the beginning of the project to June 1975, no qualified accoun- tant was assigned to record the project's accounts. Accounts were, in princi- ple, kept within the Livestock Department by an unqualified Mauritanian bookkeeper who was in fact the Livestock Director's secretary. He proved to be unable to maintain a meaningful project accounting/reporting system or to fill out correctly withdrawal applications requesting disbursement from the Bank. This situation was pointed out by all supervision missions since early 1974, when it became evident that slow disbursement was hampering project implementation. In 1974, the French Assistance Fund (FAC) promised to second an expatriate accountant to the Director of the Livestock Department, but he did not take up his assignment until June 1975. From June 1975 to September - 37 - 1976, when he left the country, he kept the project's accounts in good order with respect to ongoing expenses, but failed to update accounts dating from before his arrival. Because it was feared that the accounts would fall behind again, project supervision missions asked Government to entrust to the project coordinator the task of keeping the accounts with the help of an accountant/ secretary. By the end of 1977, they had succeeded in straightening out the accounts for the four previous years, and by June 1978, all accounting infor- mation about the project was available, including details of all expenses by category, equipment purchased, costs of technical assistance, salaries and supplies. 69. As of June 1977, the accounts were ready to be audited from the beginning of the project to 1976. The Treasurer General was appointed to this task in late 1976, but was fired from his post in early 1977. In October 1977, due to insistence from a supervision mission, Government nominated a financial inspector from the Ministry of Finance as auditor of the project. The latter met with the project coordinator twice to discuss the audit proce- dure and check the information available, but failed to come again to audit the accounts properly. It is unlikely that any external audit of the project's accounts would or could be performed at this time given the political condi- tions in Mauritania and the fact that there is no one who knows about the project's accounts or the information left by the coordinator. C. Expatriates 70. A veterinary officer was seconded by IEMVT under FAC financing in March 1972 to act as the coordinator of the animal health component. Although he was expected to be located in Boghe, in the center of the project area, he was obliged to stay in Nouakchott for nearly a year because of the numerous administrative problems that hampered the start-up of the project and delayed the delivery of vehicles and equipment as well as repairs on his place of residence in Boghe. Despite these particularly difficult conditions, however, he performed very well in planning the animal health campaign, building up 12 teams to run the yearly campaign and organizing the delivery and dispatch of all equipment and provisions required for this purpose. He finally resigned as of the end of November 1973 because of the numerous administrative difficulties encountered during his assignment in Mauritania of almost two years, and was later replaced by a young Mauritanian veterinarian. 71. The project was comprehensively reviewed in November 1975 in order to accelerate its implementation. One of the chief decisions of this missionl/ 1/ Besides cutting down the size of the wells program, reallocating IDA funds among the three components of the project and undertaking the construction of wells by force account. - 38 - was to appoint a full-time project coordinator to strengthen project manage- ment, as well as two mechanics and two well-diggers to help the Hydraulics Department carry out the wells program by force account. 72. The Project Coordinator (seconded by SEDES but under IDA financing) took up his assignment in February 1976. Although it was doubted whether he would be entrusted with sufficient power to obtain agreement on decisions to be made with respect to various matters (such as the work program, procurement and payments to suppliers or contractors), his impact on project implementa- tion was significant. In particular, he was able to deal with local realities with effectiveness, gain the confidence of the relevant Mauritanian administrative services, and thereby acquire sufficient authority to assure the effective organization and implementation of the three components of the project within the l.mics imparted by the capacity of technical services in charge of the execution of works. 73. On the other hand, the performance of the mechanics and well-diggers seconded by the consultant under IDA financing was uneven. The incompetence of some of the mechanics and the unwillingness of some of the well-diggers to travel in the field led to a rapid turnover of this technical assistance staff and consequently poor monitoring of the well teams in the field and inadequate maintenance of equipment. VIII. BANK PERFORMANCE 74. As previously mentioned, the project as designed at appraisal did not really constitute a new investment program for increased production or for inducing herdowners to adopt better management and improved pasture usage. It has been suggested too that it was not feasible to prepare a better project, as the drought of 1972 would have probably led to the elimination of any other component not directly related to the basic and emergency needs of the subsector. However, given the situation of the project area which borders the Senegal River over a considerable distance, the project could have included trials and extension of semi-intensive or intensive animal husbandry techniques among the farmers/herdowners living along the Senegal Valley to test the necessary integration between crops and livestock. But, in any case, the drought would have precluded the proper conduct of these trials. 75. The key issue obviously was the assessment of Government's implemen- tation capacity. The basic problems faced by the project were of an administra- tive type. To start with, the project suffered from the absence of a strong project unit free from other administrative duties; in particular, the normal functions of Department Director continued to be carried out by the Project Director. The changes in project accountants and, for a number of years, the absence of a project coordinator, were factors which prevented a tight -control over project implementation. This absence of a central authority directly - 39 - resulted in a lack of coordination among executing agencies. Also, the lack of high level staff directly responsible for implementation of each project component to a unique project authority contributed to a considerable diffu- sion of efforts and means. 76. Scheduling for implementation of the wells component works and procurement for the three components proved to be unrealistic. A major source of delay was the time required to decide on the way the wells construction program would be executed. Supervision reports repeatedly cite the tendering procedures utilized by the project office and the restricted publicity as factors contributing to the total failure of ICB procedures when seeking contractors for the wells program. This, however, does not alter the fact that the type of work involved was in itself poorly suited to contracting. Procurement of equipment also proved to take much longer than expected and prevented the start of animal health and firebreak implementation foreseen for 1972. 77. Supervision was adequate. It always tried to promote steps to avoid additional delays, even though internal rivalries between ministries or departments meant that recommendations made by supervision missions often went unheeded. As already mentioned, a number of remedial measures were taken on various occasions, the most significant being on the occasion of a major project review conducted in November 1974 (see para. 29). Although it is regrettable that these measures, or similar ones, were not included in the original project design, their introduction after 2-1/2 years of Credit effectiveness was instrumental in having the project reach a satisfactory conclusion. The application of these measures was somewhat facilitated following a strong cable sent to Government by the Regional Vice Presi- dent in early 1976. This cable should probably have been sent earlier, since it made the subsequent supervision missions more efficient and more useful in assisting the staff in charge of project implementation. LX. CONCLUSIONS 78. The project achieved its immediate aim of improving nomadic live- stock production through the provision of watering facilities and their maintenance, animal health control measures and fire protection works. Its impact was all the more significant in that the project area was affected during the implementation period by the most severe drought in several decades. Apart from this exceptional situation, which modified considerably the benefits to be expected from the project and delayed some components, most of the difficulties faced by the project might have been foreseen at appraisal with a better evaluation of local realities. In this sense, the overall schedule was overoptimistic. 79. Several lessons from the project should be retained: (a) the necessity of forecasting a longer start-up period in view of local conditions regarding administrative and financial organiza- tion, local procurement conditions and physical conditions of the country; - 40 - (b) the primary importance of setting up and enforcing, each time it is needed, a strong management unit independent of other administra- tive agencies and duties, comprising sufficient and nualified staff to plan and control activities, make funds available ii time and keep reliable project accounts; (c) the need to prevent the situation of several ministries and depart- ments being involved as decision makers in project implementation and, when required, the necessity of setting up inside the manage- ment unit an efficient coordinating staff; (d) the difficulty of setting up a project coordinating commission as an advisory committee comprising high local parsinalities not directly involved with execution, in particul. Ahen there is a risk of rivalry between various ministries, and thus of disagreement on the implementation approach. Experience with this project has shown that in such cases it may be better to have technocrats from the relevant departments involved in project execution, with the minis- tries not directly involved being informed from time to time of the project's progress and of their responsibilities; (e) the necessity of checking carefully whether or not the type of works to be carried out under contract are in fact suited to con- tracting to reduce the risk that bidders, if interested at all, will include in their proposals a considerable price contingency; (f) the difficulty for staff operating in the field to understand and comply with Bank procedures regarding procurement, disbursements, covenant compliance, etc.; a preliminary mission from Controllers' could train executive staff on these matters; (g) the need for frequent and regular supervision missions. In this respect, frequent IDA supervision and the resulting decisions prob- ably saved the project; (h) the necessity of a strong and capable accounting unit with the means to obtain necessary funds, and the need to enforce covenants relating to the auditing of project accounts as well as the need for regular accounting reports. 80. As the project has definitely been successful given the local condi- tions and the major events which occurred, it is likely that livestock activi- ties will continue to develop assuming that: (i) Government is able in the future to comply with project covenants related to financing/ the continua- tion of animal health campaigns as well as firebreak and well maintenance; (ii) no major catastrophes occur again; and (iii) a suitable livestock policy is designed, financed and applied, emphasizing a real integration between livestock and crops wherever feasible in order to maximize livestock production. 1/ Government allocated UM 3.5 million (US$78,000) in 1977 for the 1977/78 animal health and firebreak campaigns, and UM 25 million (US$555,000) in 1978 for continuation of the three project components. - 41 - 81. Given the major importance of livestock in the Mauritanian rural economy, the uncertainty of natural resources with regard to pasture avail- ability, the ongoing development of modern agricultural projects along the Senegal Valley (in which the Bank is involved), and the small number of high level staff available in the country, the Bank Group should take advantage of its experience in the most promising part of the country to remain involved in livestock development in Mauritania. Its participation could focus on an advisory role in helping Government to elaborate an overall livestock develop- ment policy and in identifying and financing a follow-up project partly oriented to the introduction of semi-intensive and intensive livestock prac- tices among the modern and traditional farmers of the Senegal Valley.  긔