Over the last twenty years, Mauritius has undergone major structural changes from an agricultural mono-crop economy with a rapidly growing population, high unemployment and low per capita incomes to a situation characterised by fairly stable population growth, near-full employment and an economy which is undergoing rapid diversification with the emergence of new sources of growth in export manufacturing and tourism. Rapid economic growth, besides creating employment, also altered the occupational structure of the country. In 1972, the number of people employed in the agricultural sector was 2.5 times the number in manufacturing. In 1983, this ratio had fallen to 1.2 and by 1990, this situation had reversed. Mauritius is heavily dependent for its food supplies on imports. It imports the total amount of its two basic staples, rice and flour (both of which are subsidized), and over 80% of its consumption needs in milk, beef and lamb.

There has been a definite improvement in the education levels of the labour force over the last twenty years largely as a result of the introduction of free education in 1976. The proportion of workers with no eduction has been reduced to a negligible level and the proportion of those with secondary education has doubled over the same period. The country has a comprehensive social security assistance system with a wide coverage which may be under threat due to the high recurrent expense.

The current Government's rural development policies emphasise improving the agricultural productivity of smallholders and encouraging agricultural diversification whilst paying due attention to natural resource conservation and rural employment creation.

The completion evaluation mission

In 1992, IFAD fielded a General Identification Mission to Mauritius to explore opportunities for future IFAD investment. In this connection it was decided that a completion evaluation of the project could contribute to the process of defining a conceptual and strategic framework for future IFAD assistance in Mauritius.

The Completion Evaluation mission was fielded in June 1993 and findings and recommendations are based on field visits, comprising interviews with beneficiaries and on-site observations, interviews with bank staff, project and other government officials, and the review of project documentation. Although the mission was unable to access sufficient data in the banks on the performance of the individual small-scale loans, the conclusions reached by the mission on this subject generally coincided with those of the government officials involved in the project and were confirmed by field observations.

Project rationale and objectives

At the time of appraisal, the economy of Mauritius relied heavily on the mono-crop cultivation of sugarcane and was therefore vulnerable to international market fluctuations of this crop in addition to being subject to sporadic incidence of cyclones. In rural areas, unemployment had risen to over 10% of the rural workforce. Some 12% of the rural population lived below the absolute poverty line (USD 190 per caput in 1981) and Government policies were geared towards eliminating unemployment and improving living conditions in villages which often lacked basic services and amenities. Crop diversification on land marginal for sugar-cane cultivation was identified as a main opportunity to both raise domestic food production and to increase the resilience of the rural economy.

The Mauritian Small-Scale Agricultural Development Project aimed at improving the incomes and living conditions of small-scale farmers other than those engaged in sugarcane growing, to promote crop cultivation and livestock raising, promote productive use of available land and support a number of discreet activities for the development of the rural sector. In addition, the project was to create local capacity to carry out and monitor effectively small-scale agricultural development projects. The project had a total cost of USD 8.70 million. This was made up of the IFAD loan (078-MT) of SDR 5.25 million (equivalent to USD 6.3 million), GOM contribution of USD 1.7 million and farmers contribution totalling USD 0.7 million. The project was to be implemented over a five year period following loan agreement signing, with implementation to be coordinated by the Rural Development Unit (RDU) of the Ministry of Economic Planning and Development (MEPD). The cooperation institution was the World Bank.

The components of the project were:

• support for an agricultural credit programme to be implemented by the Mauritius Cooperative Central Bank (MCCB). This was to finance the following: three small irrigation schemes for production of mixed vegetables (95 ha); a small-scale irrigated litchi production scheme (25 ha); drainage of a swamp for rainfed production of rice and vegetables (105 ha); planting of mulberries for silkworm production (42 ha); improving goat production by providing selected breeding stock; funds for unspecified small-scale agricultural schemes; and technical assistance to strengthen MCCB capability in provision of effective lines of small-scale credit;
• support for rural development activities. This was to finance the following: equipment for research into food crops; six small rural health centres; self-help schemes in 100 villages for the provision of kitchen gardens; improving the goat breeding stock and strengthening the veterinary outreach of MOA; a study on marketing of fruit and vegetables and construction of storage facilities; and, training and transportation for RDU field officers;
• support for M&E activities of MEPD. This component was to provide technical assistance, local consulting services, training, vehicles and equipment.

Due to slow disbursements in the initial years a reformulation of the project took place that introduced a general line of credit administered by the MCCB and the Development Bank of Mauritius (DBM), to provide credit for agricultural and non-agricultural enterprises to individuals rather than to cooperatives.

General observations

General outcome of the project: Among the more successful outcomes of the project were: the strengthening of the small-scale lending capabilities of MCCB and DBM; the improvement of the goat breeding stock of MOA; the introduction of credit for small-scale enterprises in rural areas; the provision of decentralised health centres; provision of three veterinary centres; and a stimulus for successful kitchen gardens for more than 900 families. The following were the main failures of the project: the low productivity and profitability of the small-scale agricultural projects; the abandoned irrigation projects because of non-availability of land, the initial failures to improve the goat breeding stock by the MAO; the marketing study did not lead to any improvements in storage or marketing of fruits and vegetables; the MEPD failed to implement M&E functions to any useful degree; the project failed to disburse at the anticipated rate and took almost twice as long to complete as was envisaged at appraisal; and no system was designed to monitor the impact of the project on beneficiaries.

Targeting: Beneficiaries targeted under the project were to have an annual per capita income not exceeding the absolute poverty level of USD 190. In 1987, this was raised to USD 300. The attempt was to target the unemployed and/or marginally employed. Initially, Regional Development Officers (RDOs) from the RDU were involved in the selection of beneficiaries. Even though income criteria were adjusted to take account of Mauritius' economic boom, targeting criteria appear to have been less rigidly applied later as it became increasingly difficult to find beneficiaries. Many of them, originally targeted for cooperative loans during the first phase of the project, left altogether or farmed only on a part time basis due to better, more secure employment opportunities elsewhere. For the banks implementing the project credit lines, the viability of projects was considered to be more important than income targeting. Furthermore, given the nature of the extended family system in Mauritius, targeting and genuinely assessing income is a complicated exercise accentuated by the lack of a strong grassroots appraisal and monitoring system.

Beneficiary participation: Beneficiary participation was not part of the original project design or of the reformulated GLC. The farmers' community was hardly involved in any of the stages of the project cycle. Project identification, preparation, implementation, and monitoring and evaluation were carried out with little prior participation of the farmers. This was consequently reflected in poor performance of several subprojects. Several of the technical activities undertaken by the project failed due to their imposition on farmers rather than testing them under their local conditions and discussing their proven technique and economic viability with the beneficiaries. However, where the RDOs were actively involved in project implementation, there were opportunities for interchanges between beneficiaries and project implementors. But when the RDU was moved to the Prime Minister's Office (PMO) in 1988, and the functions of National Development Officers (NDOs), the ex-RDOs, were redesigned, formal channels for beneficiary grassroots participation were severed.

Impact on beneficiaries: The loan agreement provided for a monitoring system which in theory would have enabled an estimate of the economic impact of the different project components. However, apart from the irrigation sub-projects, no data on yields, production and cash flows of the small-scale enterprises were available to the mission. Activities carried out under the Agricultural Credit Programme have suffered from absenteeism of farmers which has effected productivity and the efficient use of the land. Farmers remaining in operation (as part-timers) employ seasonal labour, particularly for crop maintenance and harvest activities. Very few employ permanent workers. Skilled labour is hired locally and is commonly associated with soil preparation. Low cropping intensities have diminished the significance of income generated from farming. Although the mission found that farm income could reach or even surpass income obtained from manufacturing if land is efficiently used and well managed, nevertheless most farmers sought other employment also for more security and income stability. Both the full-time farmers (20%) and part-time farmers (80%) seem to have better economic conditions (than at project design) with significant improvements in income levels.

Without a tracer study it is extremely difficult to disaggregate the general impact of the economic boom from the impact of the project. At the very minimum, the project created credit facilities for a section of the population who would not normally have been eligible for loans and provided for training of entrepreneurs and their employees. An unquantifiable success of the project is that it created a special institutionalized window of loans for poorer people in Mauritius. The down side of this is that unsecured loans had a higher rate of arrears than those which were secured. The project financed 16 Health Centres, 10 more than originally planned. The overall impact of these centres has been highly positive (especially for women) both in improving the quality and the coverage of the health services. IFAD financed health centres reach a total population of 264 000, which is 36 percent of the population served by all centres on Mauritius. The majority of the centres are fully utilized, are performing well and patients interviewed were very satisfied with the services provided. The overall impact of the services provided by the veterinary services sub-project is satisfactory. The centres provide fully-integrated veterinary services to the community of livestock breeders. The impact of the animal health laboratory has also been positive as the rate of parasitism is approaching the target of 75% and the struggle against stomoxys nigra is improving.

Women: The project does not include a specific component for the promotion of rural women. They were thought to benefit indirectly from land development and intensification of cultivation. As they contribute more than 70% of total seasonal labour and an estimated 18.5 percent of the total households in Mauritius are headed by a woman, this has proved to be a valid assumption. It can be said that sub-projects still in operation contribute significantly to the access of rural women to productive work and to their participation in the development process. However, there are no formal women extension programmes or training facilities, though the issue of the role of women in extension and agricultural development is economically important. Women are traditionally in charge of livestock, as well as fruits and vegetables, which gives them a key role in production and also in shifts towards high-value products. The rural health component had a very positive impact on women as to family planning. Women have been visited at their homes and benefited from medical personnel providing them with mother and child health care, nutrition information for children and family planning service. The birth rate has been decreasing significantly due to the services and women were reported to be in a better health condition.

Institutional impact: Both the MCCB and the DBM had dealt with small enterprise credit before IFAD's project. Nevertheless, the experience of each institution under the GLC increased their experience immensely. Unfortunately, the experience gained by the Project Unit at the MCCB, which was supported by the project was lost when first the advisor left and subsequently the unit was altogether disbanded and its functions placed under the general Credit Department. As a result of the shift of the RDU in the MEPD to the PMO where it was renamed the National Development Unit (NDU), the project lost considerable grassroots support from the RDOs whose functions changed from socio-economic to infrastructure advisors. The NDU assumed the role of the principal implementing agency; however, it lacked the executive authority necessary to obtain data from the banks or to verify the status of the revolving funds at either. At the same time, the reorganization effectively sidelined the MEPD whose function was reduced to macro-economic monitoring and to producing evaluation reports.

Sustainability: Few of the plots cultivated under the agricultural credit sub-projects seem to be sustainable, except where full-time farming is practised. As for the rural development activities, only the health centres and the veterinary activities represent a sustainable activity as their operations have been integrated into the national services. Although the GLC was instrumental in terms of delivering credit to persons who would otherwise not have had access to it, the high cost of lending and the problematic position of arrears on these loans, brings the sustainability of the credit schemes into question. The benefit from the technical assistance is not sustainable as the trained counterparts left the project to take better paid jobs.

Extensions: The project was to close in 1986, but low disbursements in the initial years allowed it eventually to be closed in December 1991, having received 5 extensions. The main concern in granting the extensions seems to have been the low level of disbursements and less attention was given to verify if the reformulated project was serving the IFAD target group.

Implementation

Financial performance: Until the changes in project scope introduced in 1986, disbursement stood at only 6.7% of the loan amount, but thereafter it quickly increased to 31.9% in 1987 and to 76.0% in 1989 to reach 99.4% as at the loan expiry date. Due to the extended period during which the project was implemented, exchange rate fluctuations played an important role. In total, a devaluation of 76% occurred causing a substantial increase in the loan amount expressed in the local currency. Part of these funds have been used to finance additional health centres and the institutional strengthening of the NDU. Originally planned at 20%, the Government's contribution to the project eventually amounted to some MUR 32 million or 25 % of project costs. It is not possible to estimate the contribution of the farmers toward the project costs. Reimbursements of the loan have always been regular and timely.

Project coordination and management: Coordination of project implementation was the responsibility of an inter-ministerial committee, the Project Coordination Committee (PCC). The PCC did not perform optimally and in November 1984 a Project Implementation Committee (PIC) composed of senior working-level staff of the project's implementing agencies was set up to closely monitor progress under all sub-projects. Policy issues would be referred to the PCC. It met frequently and served as a useful forum for thrashing out many implementation problems at the field level. The implementation of the various project activities was to be carried out by Government agencies, parastatal bodies and cooperatives under the overall direction, supervision and coordination of the RDU. In 1988, when the RDU was transferred to the PMO it was restructured to become the NDU. The main effects of the shift were a greater visibility of the Unit as part of the PMO resulting in an increased authority conferred by this allegiance and at the same time a more political role through the attachment of Parliamentary Secretaries to the PMO. These are elected politicians and are responsible for the identification and follow-up of development projects in their respective Village Council Areas. Whereas the point of reference of the RDOs in the village had initially been the informal village or council leaders, it was now an elected politician. One of the results has been a shift toward infrastructural projects, away from small-scale productive development initiatives. It would therefore appear that the rational for the project to be under the aegis of the RDU is no longer valid.

Cooperating institution and supervision: Regular supervision was carried out and it appears that sufficient attention was given to emerging issues regarding project progress, although this was not always followed up by corrective actions. Agricultural irrigation and livestock did not receive adequate coverage (as reflected in the choice of specialisation of consultants) and this was reflected in the poor performance of these sub-components. For the needed reorientation of the project towards the creation of a general line of credit, the CI takes full credit, but there has been a lack of emphasise from both the CI and IFAD on the importance of data on project impact and individual loan performance.

Agricultural credit programme: Sixty-five percent of the loan was earmarked for on-lending by MCCB to small-scale agricultural borrowers through producers' cooperative societies. Three small irrigation scheme (total 295 ha) earmarked at design were replaced by a single scheme at Riche Terre due to unsuitability of land. This scheme is only yielding sub-optimal results, as most farmers have opted for part-time farming, or have even abandoned their plot in favour of a more secure income generating activities outside the agricultural sector. Cultivation on the one arpent (0.422 ha) plots was further constrained by inadequate and uncertain irrigation, lack of credit (due to unsettled land deeds, insufficient extension services, and problems O&M of irrigation equipment. The same is true for the Litchi Production scheme where yields are not promising, many plots are heavily infested with weeds and intercropping is only sporadically practised. As for the rice-subproject, the farmers decided to change to other crops because of low-returns on paddy rice and preference of farmers for traditional crops such as sugarcane and vegetables. The Sericulture sub-project was also abandoned, because traditional crops regained their attractiveness, as their prices went up relative to that of silk and the free family labour that was assumed to be available did not materialise because of the changing structure of the Mauritian economy and the labour market. The Goat Upgrading Sub-project was initially hampered by the poor management of the newly introduced breed. There was a lack of control and monitoring of the animals' health and adaptability to the new environment and in addition the breed was more suitable for milking than for meat which does not concur with the preference of Mauritians. In the second phase, three rounds of breeding took place at the Veterinary Service Division, which assured that the goats were able to adapt to their new environment. However, results were similarly unpromising with less than a quarter of the offspring distributed to farmers, and half of the population remaining unsold at the breeding station. No criteria were used in the selection of farmers and no monitoring of the breeders was being done.

The general line of credit general line of credit: Following a major project reformulation and amendments to the loan agreement in 1986, two schemes, the Small Agricultural and Rural Industries Development Scheme (SARIDS) at MCCB and the Small Rural Enterprises Scheme (SRES) at DBM, eventually became the central plank of the whole project. The system of loan administration was designed with an intention to rely on the grassroots involvement of RDOs in preparing and monitoring loans. Once their responsibilities changed, this role was lost and an important step in the process of loan administration was severed. The MOA, SIDO and MOC were all to be involved in the appraisal of loans, however, this system was never implemented, and presently the loans are appraised entirely by the bank's loan officers. A subsidiary credit agreement was made between the GOM and the banks, but the institutional developments described above generated a junction in the project, by effectively separating the project management from the credit line which was not foreseen in the subsidiary agreement.

Publicity of the availability of the credit had been successful and beneficiaries were reached through the media, the network of RDOs and by word of mouth. Collateral requirements changed from general floating charges to personal guarantor, in view of the high arrears and non-performance of outstanding loans. Although cases are reported to have been filed for legal action the results of these steps are not yet clear. The institutional support provided to the banks consisted of an internationally recruited TA Financial Analyst for MCCB and staff of RDU/NDU seconded to both MCCB and DBM. Although the TA provided some staff training and assisted in designing a loan administration system, the subsequent departure of the individuals trained and the dismantling of the Project Unit effectively minimised the impact on the operations of the Bank.

The inadequate provision of data by both banks but in particular the MCCB, has impaired the efforts of the evaluation mission in evaluating the financial or economic performance of the schemes. At MCCB 65% of the loanees have loans in arrears, and total arrears amount to 32% of the outstanding balance. Although MCCB management claims that some funds are being recycled, both NDU and the mission were unable to verify that these funds are being kept in a special account and outlent at the agreed upon interest rate. The situation for DBM is similar, with 65% of the small industry loans in arrears representing 51.6% of the value that is due. Of those, almost 60% have arrears greater than 12 months, while 91% of the amount due has been in arrears for more than a year. For the agricultural portfolio the situation seems somewhat better with 47% of the borrowers in arrears for 35% of the amount outstanding. More disturbing is the fact that 95% of the amount in arrears have been outstanding for 12 month or more. It seems that the situation has rapidly deteriorated over the last two years, with arrears doubling, suggesting perhaps high accumulated interest charges on loans overdue by more than 12 months. According to the banks, the following factors contributed to the high level of arrears: (i) technical capability of beneficiary, (ii) marketing capacity of loanee, (iii) equipment failures, (iv) insufficient returns on investment, and (v) some mala fide borrowers. The mission was of the opinion that apart from these factors, poor managerial and business organisation of the enterprises also contributed to the situation of high arrears. The low level of support these credit windows received from the banks also created an environment were loan appraisal and monitoring was extremely difficult.

Rural development: The self-help scheme for the cultivation of kitchen gardens, although successful in terms of the number of households reached (930), could not be continued for its high cost and heavy monitoring requirements. Both the marketing study and the sub-component for foodcrop research equipment found alternative funding. This may have been foreseen at appraisal with a more careful screening of funding options of the GOM. The rural health component proved successful and 26 health centres were eventually built and equipped, whereas only six had been planned. These centres reach a population of 264 000. The accessibility to health care services increased and pressure on hospitals was relieved. Three veterinary centres were fully equipped and provide around-the-clock services to the farming community. Livestock extension services and feed supplies are also provided by these centres.

Monitoring and development: Responsibility for M&E was assigned to the Programme Implementation and Monitoring Section (PIMS) of MEPD with RDU only involved in providing field support and ex-post evaluation to be undertaken in cooperation with the University of Mauritius. Eventually a division of tasks between the Monitoring and Evaluation Division (MED) of MEPD (formerly PIMS), RDU, MCCB and DBM evolved, with MED largely responsible for baseline surveys and evaluation studies and the other three agencies for the monitoring of their respective activities. The University was never involved in the project, apparently because of a lack of interest and absence of technical capacity. The former RDU became the NDU in 1988 and was moved from MEPD to the Prime Minister`s Office. The MEPD became responsible for macro-programme evaluation, the NDU for specific monitoring and evaluation and the banks for the monitoring of their credits. An internationally recruited consultant was temporarily stationed in PIMS. He established a M&E system, and trained one economist and two sociologists. Part of the funds from cost savings on a few other IFAD sub-projects were allocated for the purpose of strengthening the Monitoring Section of the NDU by way of computerization, procurement of vehicles and oversees staff training. Under this arrangement, two vehicles and four personal computers have been acquired. However, the training of staff has not been achieved. Although these arrangements seemed conducive for an effective system, in practice the division of responsibilities resulted in inadequate monitoring and evaluation by all parties and the major problem encountered in making the M&E system operational was the difficulties in obtaining the required information from all the executing agencies involved in the project. The actual monitoring is presently done by one RDO and the whole operation is centralised at the MEU (of NDU) which is responsible for the monitoring of all NDU projects.

In retrospect, the project did not develop an effective M&E capacity. The project managers did not have access to the data required to ensure efficient management. There is no method of objectively determining impact on beneficiaries. Management seemed more concerned with the progress of implementation than with analysing what the various projects are producing in terms of socio-economic results. IFAD support for the design of the project did not bring the expected results, mainly because of institutional changes introduced throughout the project.

Lessons learned

Design elements: The objectives set for the project appear to have been over-ambitious. The risk analysis made in the appraisal report overlooked some of the main constraints to the project. With the economic development of the country, labour availability turned out to be a constraint. The assumption that land for irrigation projects would be available proved to be misplaced in a country with an intensive land use pattern. The appropriate authorities concerned with land utilization were not effectively consulted.

Only limited consultation of targeted beneficiaries took place during the design stage. This led to delays as well as failure of some activities. A thorough and effective consultative process with the beneficiaries and implementors of the project from the outset of project preparation could have avoided a number of problems encountered by the project.

The appraisal report failed to address the issue of extension services and the need to improve or expand those to promote the input packages needed for the project.

Institutional Capabilities: The project was not based on a realistic assessment of the capacity of the institutions involved in implementation. Whilst RDU/NDU had an adequate planning capability, it did not have the technical skill, nor the political and administrative support required to implement the project effectively. It would therefore appear that the rational to place a small-scale agricultural development project under the aegis of the NDU was not valid, especially after its transfer to the PMO.

The experience with cooperatives as vehicles for credit and other support services for farmers in this project has not been positive.

The implementation of the small-scale enterprise credit schemes suffered from an absence of staff at the grassroots for appraisal and follow-up either from the banks themselves or from the Ministry implementing the project. Qualifications and suitability of international TA staff should be carefully screened by IFAD and fully endorsed by the recipient country.

Imposition of technical solutions: Several of the technical activities carried out under the project failed without being tested at pilot scale or being based on proved techniques. In particular the Goat Breeding, the Rice Production and sericulture proved not economically viable. This contrasted with the more successful results of the GLC. The lesson is that local entrepreneurs and farmers have a better idea of the techniques that will work than outsiders. Where new technical solutions are needed, they should be tested under local conditions prior to distribution to beneficiaries for production.

Monitoring and Evaluation: The project did not generate an effective M&E system. Monitoring should pay more attention to assessment of changes in socio-economic indicators at the level of the beneficiaries. Experience from the General Line of Credit taught that data on performance of credit lines should at all times be available. Project managers should have access to data to ensure efficient management, flexibility of approach or identification of issues affecting implementation. An effective M&E system must be installed from the outset. This includes identification of criteria to be monitored in assessing project impact. M&E operations are highly dependent on the institutional set-up of a project and they should be fully re-appraised after any reformulation or change in the institutional arrangement of a project takes place.

Reformulation: The project was subject to a fast changing social and economic situation of Mauritius. This required rapid responses from project management that was not built into the project from the start. The issue of "part-time" farming became a factor of some importance which should have been dealt with in a more decisive manner. The project reformulation that took place, relied too much on the capacity and willingness of the banking community and the Government authorities to cooperate and subsequent restructering of the government services was not followed by an adjustment in the appraisal and monitoring arrangements. However, improved disbursement records after the reformulation led the CI and IFAD to belief that the project was performing satisfactory and little was done to verify that project funds were targetted towards the poor and were supporting economically viable projects.

Recommendations and issues for follow-up

Future interventions should be designed in a flexible manner so as to enable the beneficiaries to engage in profitable activities, taking into account the open character of the Mauritian economy and its propensity for fast change. A thorough and effective consultative process should be entered into with potential beneficiaries from the outset of project preparation.

Extension of closing dates should considered less on the possibility it will give to the project to disburs all the loan funds and more the ability of the project to channel those funds to activities that will clearly serve the intended beneficiaries.

For agricultural activities, due attention should be given to the capacity of the extension services and the importance of the farmer systems approach. Land requirements should take into account present and future alternative uses and should be based on realistic assumption regarding farming needs.

The choice of implementing agency for any future IFAD project in Mauritius should take into account the ability of the agency to follow up on project implementation at the grassroots level. The rational for placing a future small-scale agricultural development project with the NDU is no longer obvious.

Credit programmes: When a line of credit is administered by a commercial bank, clear guidelines should be laid down from the beginning on the access that project authorities should have data relating to the individual loan performance. It is recommended that data on credit should be made available by the banks in a standard format which is cleared by IFAD or the CI.

Unless the banks, implementing the small scale credit schemes have separate project units from which thorough screening of applications and active follow-up of the loans is guaranteed, the credit schemes and their expected results in terms of increased incomes of beneficiaries will not materialise to an extent that would justify IFAD's intervention. To establish a viable small enterprise in the context of Mauritius' booming, relatively open economy is extremely difficult. Sound appraisal of loan applications, careful monitoring and extension support services is required.

Consideration should be given to the possibility to make some training (e.g. in business administration) compulsory for credit takers. The availability of support services should also be assured. The redeployment or creation of additional countrywide RDOs to support appraisal and monitoring of loans is to be considered.

The Evaluation Mission was not able to obtain satisfactory information on the revolving funds. IFAD should receive assurances that the revolving funds are properly managed and used for the purpose for which they were created.

Irrigation authority: Breakdown of irrigation equipment occurred frequently but repairs were not carried out promptly, leading to insufficient irrigation of the two irrigation schemes financed under the project. The Irrigation Authority should enhance the efficiency of its maintenance services to irrigation schemes.

Extension services: The extension service should endeavour to make periodical visits to farmers, prepare an inventory of resources available to the farmers and of their constraints. Therefore, the extension worker should start with the concept of the farm-household system, analyze constraints and potentials and list possible solutions before preparation of an extension program. Efforts should be made to pay more attention to train farmers and extension officers in management, farming systems approaches and farmer's participation. Sufficient technical training for women (extensionists workers and farmers) should also be given, and a rational strategy for extension including programs for women and small farmers should be worked out. In addition to basic agricultural extension, the extension service needs to concentrate on water savings and other environmental aspects of agricultural production, thus playing a leading role in the development of sustainable agriculture.

Marketing: The need for storage of vegetables should be considered. There is at the moment a low capacity for storing vegetables in the country whereas there are frequent occurrences of oversupply. The construction of refrigerated cold stores would provide such storage, but it is unlikely that the high cost involved could be offset with higher market price. It would be better to use present storage capacity more efficiently and to develop additional market outlets coupled with improved market price intelligence. To reduce the margin of the middlemen, MEPD could play a role in supervising the marketing process.

Cooperatives: The Ministry of Cooperatives should focus on helping cooperatives to reestablish financial viability, to advise on basic management reforms, and to undertake, or commission feasibility studies for medium and long term investment by farmers under cooperative societies. The validity of cooperatives as vehicles for credit and other support services for farmers should be reviewed in the light of the negative experience of the closed IFAD project.