Completion evaluation report summary

The project area comprises two distinct swamp areas located 7 km apart on the southern banks of the river Gambia. The climate, soils and topography are favourable for intensified rice cultivation. At the time of project formulation, most of the Jahaly swamp covering some 1.230 ha was already cleared and levelled as a result of previous development interventions. The Pacharr swamps, covering about 1.050 ha was more or less in its natural state. Smallholder farmers in the JP area produce a variety of crops during the short rainy season, mainly groundnuts (the main cash crop) and cereals (millet, sorghum, maize) in the rainfed upland areas and rice in the lowland swamps. The upland cropping productions are based on traditional shifting slash and burn cultivation with fallow. Opportunities for production in the longer dry season are scarce and this made the cultivation of dry season irrigated rice attractive.

Project objectives and design

Target group

The target group comprised 1.000 compounds in some 40 villages with an estimated population of 15.000 "having user rights" in the two swamps. Women, regarded by the appraisal as principal beneficiaries under traditional systems of swamp rice production, were expected to "receive major benefits from the project".

Objectives and components

The JPSP set out to increase Gambian national self-sufficiency in rice production, improve food security and raise farm families incomes.

The components were:

• land development: 1.510 ha of swamplands were to be developped. Pump irrigation would comprise 560 ha and "improved rice" 950 ha. In an average year, 380 ha could be tidally irrigated in the wet season and 240 ha of these could be tidally irrigated in the dry season. Land would be reallocated on the basis of submissions by villages of their traditional land rights, existing use rights and existing areas cultivated by compounds.
• training and institutional support: the development of a local institutional capability was considered important far irrigation-based agriculture and related extension.
• credit input supply and crop marketing: the credit scheme was to be administered by the Gambia Cooperative Union (GCU). The cooperatives were also to handle actual input supply and crop marketing.

Then, the project was to fund the construction of five cereal banks and two day-care centres to take care of children of working mothers. A monitoring and evaluation unit was to provide the project management unit and the coordinating commitee with regular information concerning the performance of the project.

Expected effects and assumptions

Yields would increase up to 3,3 t/ha on improved swamp land (1,2 t/ha without project), to 4 t/ha on pump irrigated land for the wet season, 4,5 t/ha on pump irrigated land for the dry season. Production was expected to rise to 7.874 t (PY 6). Incomes were expected to rise from 247 USD without project to 1.079 USD with project.

The project was based on strong assumptions as to the permanence of the economic environment prevailing at the outset. Asumption was also made that other aspects of compound production and income would not affect rice production and income and that the main interest of farmers was in rice production even during the rainy season.

Evaluation

Implementation context

The 1985 structural adjustment programm led to significant policy changes in the Gambia's tariff and tax structure. Relative agricultural output and input prices turned against input intensive, high yield rice production. The reduction of price subsidies had doubled fertilizer cost in real term.

Natural conditions for rainfed upland agriculture improved which contributed to reduced food shortages and less demand for water-controlled rice production in the lowlands.

Project achievements

Land development: a total of 1.392 was developed, 545 ha for pump irrigation, 703 ha for tidal and 144 ha of improved rainfed land. The total area that may be double cropped under tidal irrigation is 364 ha.

Irrigation and drainage: they were both pump-based. The command areas were hermetically sealed off through raised ciment levees. A total of 13 pumpsets were procured, 11 installed and 2 spares. Water delivery and drainage inefficiency have been seriously compounded by little maintenance and sabotage of the pump-irrigated areas' infrastructure.

Input supply, credit, marketing and loan recovery: the GCU was responsable for input supply,credit and onward marketing. Initial performance was quite satisfactory. Between 1984 and 1987, loan recovery rates for seasonal credit averaged not less than 97% at Jahaly and 83% at Pacharr. In 1897, paddy production and loan recovery rates dropped (loan recovery rate: 30%). In 1988-89, the GCU withdrew from the rice sub sub-sector. The project became responsable for input supply, credit administration, loan recovery and marketing. There were a break down in input supply and credit delivery.

Revolving funds: the GCU deposited a proportion of recovered water and land charges and loans into two revolving funds. Lack of clarity about the accounts and the amounts involved have persisted since 1985. GCU retained control over funds accumulated in the revolving fund accounts, despite its withdrawal from the project.

Extension and agronomy: a modified T&V system of extension has been used. Meetings are held weekly between the extension staff and chosen "contact" farmers. One progressive farmer is chosen for each 10 ha block of the scheme. In addition, there is a demonstration plot at each swamp. Communications between farmers and extension staff were not easy because of the rigid water scheduling and became less than cordial when the staff in recent years had to undertake additional responsabilities over and above their primary extension activities (collecting loan repayments, organising water scheduling and sharing and maintenance of primary canals).

Cereal banks have not been built. JPSP day care centres never operated properly and are now defunct.

M&E: the technical and socio-economic M&E has been fragmentary. The data collected were of limited use in improving project performance, water scheduling and extension recommendations.

Effects assessment and sustainability

Beneficiaries: land developed under the project is presently available to 2.200 registered leaseholders from 1.253 compounds, dispersed over 67 villages. Current direct and indirect beneficiaries of the project are estimated at between 19.000 and 22.000 members of farm households. Poorer farmers were successfully targeted. In 1991, smaller households cultivated a larger per capita area in the project than larger ones. The number of participants in upland villages which used to be marginally involved, or not at all, in swamp or irrigated rice cultivation is considerably higher than expected because of the precarious food situation between 1983 and 1984.

In the early years 1984 to 1986, pump irrigated rice production performed very well. Average rice hields per hectare were significantly higher than anticipated at appraisal: 6.5 t/ha during the dry season and 5.4 t/ha during the wet season. But, between 1987 and 1992, average yields reached only 4.65 t/ha for the dry season and 2.28 t/ha for the wet season. The recent IFAD completion evaluation survey (CES) found even lower yields for 1990 and 1991: for the dry season, interviewed farmers reported yields between 3.3 (1991) and 3.7 tons per hectare (1990). Average wet season yields were between 1.7 and 1.8 tons/ha. In the end, low production caused low loan repayment rates of project charges by farmers. Eventually, this led to a cancellation of all ploughing, puddling and irrigation activities in pump-irrigated plots for wet season 1992. In wet season 1991, on average only 74 percent of all irrigated plots were cultivated. But for the dry season, a remarkably high cropping intensity was recorded of 95 percent.

On improved swamp rice plots (tidal-irrigated and improved rainfed swamp), yields of 2.5 to 3 tons/hectare can be achieved during the wet season. Yet, there are many plots in this production system where yields are significantly lower; many fields present yields hovering around 1 ton per hectare, particularly those which are purely rainfed, not transplanted, or where water-control through tidal irrigation is insufficient.

Effects on beneficiaries incomes: in 1985-86, the average household with a plot in the project area recorded a net increase in household income of 13%, compared to those without such land in the project area. But, in 1991, farm average gross margins per hectare and personns per day did not exceed those achieved in traditional swamp rice production.

One of the major accomplishments of the J.P. project is the beneficial impact of increased food security during the traditional lean or "hungry" season between June and September. By providing an additional dry season crop, the project increased level, and stability, of food consumption.

Specifics effects on women: the nutritional status of children and women improved, as measured by body weight and growth. The project diminished mothers' seasonal stress that is caused by increased workload, low food intake, and increased disease prevalence during the rainy season. A deliberate attempt to put women in charge of pump-irrigated plots by formally registering them as lease-holders was not successful. The net impact of the J.P. project on women's access to resources is difficult to assess. Overall, women certainly gained from the increased availability of food in the household and the significant reduction of back-breaking labour through mechanical land preparation.

Effects on the environment: the SAR referred to the threat to cultivation along the River Gambia because of an upstream movement of the salt water mixing zone associated with pump irrigation. But, when pumping was carried out on JPSP, the position of the salt water mixing zone was not monitored. This means, at least in theory, that the environmental impact over the project period could be negative.

Sustainability: the technical assistance had not transferred skills to national staff. Farmers' interest in rice cultivation decrease because of the declining profitability of paddy cultivation and lead to the break down of input and credit delivery. Farmers have limited interest in paying user charges and allocated their time on canal maintenance. The scheme's physical infrastructure and technical support system are deteriorating rapidly. There is no institutional structure for assuring stable and adequate credit, agro-input supplies and marketing facilities necessary for viable scheme operation. There is no beneficiary organisation in place for overall scheme management and field-level water management. This project does not seem to be more sustainable than the three previous interventions to develop rice production in the area.

Main issues

The project design is not adapted to the situation.

Project design must tie in with farmers' strategies and respect their comparative advantage in production by season, crop and intensity of effort. Farmers long term comparative advantage in cultivation, by crop and season had not been understood, and, as well, the probability of an imminent reduction of rice production support subsidies had not been taken into account. Pump irrigation in the wet season was an essential feature of project design, in spite of the uncertain nature of farmers' demand for such cultivation.

The design of the irrigation component permitted little or no flexibility in response to varying preferences in production and water demand.

The design of th irrigation component was not a "least cost" alternative. Design engineers opted for a total water control approach. Irrigation and drainage were pump-based. The local land form was radically and irreversibly transformed. The possibility was blocked of a reversal to tidal irrigation or to swamp cultivation. Bay must be completely flat to prevent the creation of ponds; to secure even-wetting for cultivation and seeding; and to obtain optimal drainage. Farmers have difficulty in fitting levelling into a tight cultivation schedule at a point in time when soils are workable. The flat bay system heavily restricts the type of crops which can be grown, since it is suitable only for flood irrigation. For tidal irrigation, the effectiveness of the watering system rests on the degree of skills in controlling water supply, and timeliness of opening and closing flood gates. Although the technology is relatively simple and cheap, tidal dependence and the variation in water levels places a premium on rigid scheduling and farmer cooperation.

The high initial yields in pump irrigated rice were not sustained.

The project was not well prepared for the sequence of unfavourable agronomic and economic conditions that started around 1986-87. Despite water controlled conditions, production in the scheme became more risky. Early rains and crop pests cut rice yields in dry and wet season 1987 almost by half, starting a cycle of loan default and delayed planting. Economic conditions turned against rice production (see 8). Input supply became more erratic after the withdrawal of GCU from the project in 1988. Improved climatic conditions for alternative, rainfed crop production of groundnuts and upland cereals further eroded the attractiveness of water controlled rice production for many producers in the area.

Recommendations and lessons learned

Under present technical, market and socio-economic conditions, in the Gambian agricultural sector, the most realistic scenario for future project operations is to: (i) support continued dry season rice production; and (ii) to seek to improve the technology of low-cost tidal irrigation.

In the short-to medium-term, the project should focus on the preferred production technologies, which means less emphasis on pump irrigation in the wet season, a scaling down of such irrigation, or its elimination. The situation could change were a repeated and prolonged drought to reoccur, negatively affecting upland crops and food availability. Hence, the longer-term scenario would be to maintain at minimal level a pumping capacity also for wet season irrigation.

Dry season production should focus on enhancing household food security for a maximum number of farm households, in low and upland villages. Wet season production should be targeted for women in the tidal/rainfed command area, who have the lowest opportunity costs and the highest interest in rice production during this period.

The schemes require considerable investment to rehabilitate pumping, water conveyancing and plot water control systems to improve their performance. If these physical investments are not carried out, the scheme will deteriorate at an accelerated rate over the next few years. These investments are necessary before any of the functions or responsibilities are transferred to the farmers, i.e. the scheme should function to design specifications if farmers are expected to pay for services.

A reorganization of the institutions concerned with operation of the JPSP is necessary to meet the post-project situation. With the premise that: (i) beneficiary farmers themselves would be responsible for at least partial management of water services and preferably land preparation through draught animal traction; and (ii) support services would be taken over by line ministries, the future role of the PMU should be redefined. With the premise that management functions would be assumed by beneficiary farmers themselves, the setting for developing farmers' organisations needs to be explored.

Lessons learned:

When an investment project includes heavy infrastructure development like important irrigation schemes and expensive equipment (pumps and tractors), project design must be preceeded by a far more careful analysis including sensitivity test for major changes in economic and institutional environment. Stated project risks should be reviewed carefully against lessons learned. Risks considered major and which would impact negatively implementation should be clearly spelt out. Proposals for alternative design(s) need to be considered, before a final decision is taken. In this kind of project, remedial actions have limited or negligible effectiveness once the project has been already designed.

Farmers' decision criteria need to be derived through a farming systems analysis. Gains to production and productivity must reflect the behaviour of the economic agents to optimise returns on scarce assets. Yield maximization may fail as long as land as a productive asset is not yet scarce. Farmers' portfolio of activities must be fully ascertained, and their economic returns on the different assets indicates their opportunity costs. If empirical data have not been systematically collected, or cannot be referred to, for use of on-farm labour in alternative activities (including non agricultural activities), a targeted study for such a purpose should be undertaken as part of formulation.

The rules for appointing consultancy firms need to be far stricter, even if bilateral co-financing on a grant basis is involved. Because of vested interests and potentially conflicting interests, a single consulting firm should never be engaged by the project to carry out feasibility studies, engineering design, agricultural design, supervision of contractors during construction, as well as be in charge of agricultural management of the project. Expatriate technical assistance should never be permitted to assume managerial roles at the outset of a project. There is then little scope for the nationals to gain experience in decision-making, and hand-over is always delayed.

Mechanisms must be developed within IFAD, to ensure that resources are made available to follow-up at the field level recommendations from evaluation missions to modify project design. Otherwise, such exercises with recommendations become purely academic exercises.

Beneficiaries must be involved in the design and eventual re-design through a structured process for participation. Self-reliance and self-management of production systems can only come about by putting the proposed grass roots institutions firmly in place from the outset, as it takes a long time for them to reach maturity. Participation in managing project assets, or cultivating the land, is strongly linked to the economic environment. When the economic environment shifted from rice towards upland crops; the economic surplus in rice production fell, interest for rice cultivation in the wet season subsided. The interests of participants in sharing, or assuming the responsibility for service provision, let alone cost recovery, diminished.

Projects which seek to promote the socio-economic development of the relatively underprivileged and powerless will meet with conflicts. They have to be designed with particular care. There is little or no guarantee for that the target group can successfully compete with more influential members of the local community and assimilate and consolidate gains arising from project implementation. For this reason, a detailed design and full transparency is required with regard to the activities related to: (i) distributing the most sought after asset, the irrigated land; and (ii) collecting and using core indicators with which the management can monitor, or ensure, that those less privileged continue to draw lasting benefits in the use of their assets. Administrative regulation of land tenure at the level of a project can not easily dominate the practices of customary land tenure. This is definitely the case with regard to changes to support the rights of women as against customs or controls exercised by household heads, or men. The preferences by gender for type of crop, and land use, generally reflect relative labour inputs and associated cash returns coupled with preferences for food security. Clear commitment to promote equity is required from the Government and other involved agencies. If such commitment is not obtained, project management is bound to be caught in the role of mediating local vested interests without adequate political support.