Completion evaluation

The geographic area covered by the project coincides with the twelve districts of the Governorate of Sidi Bouzid, totalling 755 500 hectares. In 1994 the total population in the Governorate was 377 700 inhabitants, of whom 78% living in rural areas. The cultivated area is 412 000 hectares, with 33 000 farms (381 000 ha) plus the Government's "organised sector" (31 000 ha). It consists of plateaux intercepted by rocky chains. The climate is semi-arid to arid in the south (300-200 mm). Agriculture primarily based on rainfed tree crops, localised but dynamic irrigation and sheep production. Rainfed cereal crops cover large areas but production is erratic. Until some forty years ago the region was dominated by cereal crops and extensive sheep farming.

The development of tree crops, sustained by a series of programmes or projects since 1960, has been the major factor in the evolution of the pastoral system towards sedentarized agriculture. Irrigation is practised locally. Tree crops covered some 195 000 ha (physical area) in 1996, mainly olives (4 million trees) and almonds (3 million trees). Mixed cropping (intercropped olives and almonds) is practised on approximately 43% of the physical area. Although sedentarized, extensive sheep and goat production is still important (545 000 head, equivalent to 16 head per farm). Most farms using irrigation also grow rainfed crops. Tree crops are important throughout most of the Governorate except in the north-west (unsuitable soil) and the south-east. Small holdings (less than 10 ha) account for 61% of all privately owned farms (25% of the area); medium-sized holdings (10 to 50 ha), 36% of the total number and 57% of the area; large holdings (more than 50 ha) account for 3% of the total number and 18% of the area.

Project design and objectives

Target group

The rainfed agriculture development project is derived from a more far-reaching project which was intended to develop "small and medium-sized holdings" in the Governorate of Sidi Bouzid, using rainfed and irrigated farming. During the appraisal phase, the irrigation component was detached from the rest to become a separate irrigation development project. Since the Government and IFAD also wished to develop rainfed agriculture, a second project was set up. It was implemented by OMVPI, the agency which already had responsibility for the irrigation project. FADES was chosen as cooperating agency, while the World Bank had that role in the irrigation project. In both projects the objective was to have an impact on "small and medium-sized farms" through actions aimed at attaining a sustainable improvement of productivity in both agriculture and employment. Approximately 6 600 farms were involved, equivalent to 20% of the total number of farms in the Governorate. The target group was thus composed of 70% small holdings (less than 20 ha with an income of around USD 130 ) and 30% medium-sized holdings (20 to 50 ha with an income of USD 400).

Objectives and components

The development of rainfed agriculture is considerably limited by agroclimatic constraints. In this context, fruit trees, which had developed considerably since the 1970s, appeared to be the only activity capable of producing sustainable and relatively regular incomes. In the early 1980s, it was estimated that a further 50 to 60 000 ha were suitable for this type of crops – compared to the then (physical) planted area of 140 000 hectares. The development of tree crops was therefore chosen as the principal strategy for rainfed agriculture. Clearly less emphasis was given to livestock production which was nonetheless a major activity in the Governorate . Project support to farmers was to be the provision of works together with credit for planting. The project envisaged the planting of 13 400 ha of which 6 300 ha in almond trees, 4 300 ha of pistachio trees and 2 800 ha shrub fodder crops. The project intended to finance up to 3 ha of plantations for each beneficiary.

The project components were:

Agricultural development

• Provision of agricultural credit to cover the investments required for the planting of fruit trees and fodder crops and for the purchase of 1 000  animal-drawn tanks for watering the plantations.
• Provision of agricultural equipment to OMVPI for support activities
• Rental of tractors and bulldozers for 210 000 hours for soil preparation.
• Construction and fitting out of 12 shallow wells and 3 boreholes to irrigate the plantations.
• Construction of 15 drinking troughs and of 12 disinfestation dips.

Institutional strengthening

• Strengthening OMVPI's intervention capacity; maintenance equipment, setting up input distribution centres.
• Strengthening OMVPI extension services.

The Ministry of Agriculture was to entrust OMVPI, already responsible for the irrigation project, with the execution of the project. A single project management unit and a single monitoring and evaluation unit would serve both projects.

Expected effects and assumptions

Estimated yields at full production (about 20 years after project start-up) were 2 814 t of almonds and 944 t of pistachios. The planting of fodder crops (14 000 t) was expected to generate an additional production of 186 t of meat and 7 t of wool. The rate of return was estimated at 13% over a period of 30 years.

Evaluation

The mission had at its disposal the information contained in the annual reports, the mid-term evaluation report (1987) and the completion report drawn up by CRDA in January 1996. The quality of the information was affected by the fact that numerous unclassified documents were inaccessible following the dissolution of OMVPI and by insufficient monitoring and evaluation data. The evaluation information was integrated by field work and special farmer surveys during the final mission as well as during the preparatory missions.

Implementation context

The principal change during the project was that of the institutional framework. Indeed, the dissolution of OMVPI occurred in 1989 and its responsibilities were transferred to the regional Commissariat (CRDA) representing the Ministry of Agriculture. Said transfer was accompanied by the liberalisation of the economy and the Government's disengagement from some of its functions. Several direct Government service components such as the provision of inputs were thus no longer justified. But the mechanical work to support the investment was maintained. The mid-term evaluation, which was performed after these changes, gave rise to a number of recommendations which were taken into account.

The project, originally envisaged for a period of five years, used its funds over a period of nine years (1986-1994). This extended closing was possible thanks to remainders resulting from certain prices that were lower than expected and from local funds derived from exchange rate differences. During the course of the project, a number of changes were made to adapt the components to the farmers' requirements as well as to changes in the political and economic context.

Project Achievements

The Tunisian dinar comparison of the costs envisaged (TND 11 million) and the effective costs (TND 10.5 million) shows only a slight difference. In US dollars, however, the difference is notable, the effective cost in USD (11.7 million) being decidedly less than that originally estimated (USD 13.3 million). This discrepancy can mainly be explained by exchange rate variations. The accounting structure of the costs also reflects important changes: the purchasing cost of equipment originally envisaged at 30% of the total project cost was actually 10%; farm credit dropped from the originally estimated 40% to 18%; equipment rental – entered under other items in 1984 – represented 49% of the effective costs, whereas the cost of wells and boreholes rose from the envisaged 4.6% to 16%. The cost of the institutional development component dropped from 24.7% to 9%.

Over the period 1986-89, total project expenditure and the IFAD loan accounted respectively for 38 and 19% of the OMVPI cumulative investment budget. During the CRDA period (1990-94), these figures were 20 and 13%. Compared to the OMVPI and CRDA cumulative investment budget, from 1986 to 1994, these figures were 22 et 16%. IFAD financed 65% of the costs instead of the originally planned 45%. Subsidies accounted for TND 2.3 million (TND 402 per beneficiary). The participation of beneficiaries, not quantified by the project, was estimated by the mission at TND 2.25 million. The cost of the project, including this self-financing element, is therefore TND 12.8 million instead of TND 10.54 million, according to the project accounts. Self-financing would therefore account for 21.4% of the actual total cost.

Fruit trees planted. The project contributed to the planting of 13 720 ha of fruit trees - 2 705 ha of almonds, 4 375 ha of pistachios and 6 638 ha of olives. The latter species, excluded at the time of the evaluation due to unfavourable price projections, was then included again following farmer demand and better market prices for olives. The planted area is approximately the same as that originally planned but with a different structure. All the districts in the Governorate were included in the fruit tree project but 66% of the total planted area and 83% of the olive trees are concentrated in four of them. Project activities involved the selection of the sites and the beneficiaries, soil preparation work financed by the project (3 957 ha of clearing et 10 715 ha of deep tilling ), the provision of saplings, monitoring and maintenance of the plantations. The average cost of the plantations was 742 TND/ha for almond trees (originally TND 534), 832 TND/ha for pistachio trees (initially TND 630 ), and 482 TND/ha for olive trees, corresponding to an average of 645 TND/ha. The average cost per beneficiary was TND 1 311 .

The 5 720 beneficiaries of the fruit tree plantations received an average maintenance subsidy of 209 TND/ha (originally TND 200). The subsidy was very unevenly distributed due to differences in access to credit. Some 1 155 beneficiaries obtained credit : taking into account the various project contributions (loans, subsidies and works), they received 862 TND/ha, which is 217 TND more than the effective investment cost (645 TND/ha). The other 4 565  beneficiaries, without access to credit, only received from the project 410 TND/ha (in works and subsidies); they therefore self-financed 235 TND/ha, equivalent to one third of the investment.

Creation of a watering capacity. Watering the saplings during the first years is the critical element for this type of investment. By abandoning the customary practice (whereby watering is the responsibility of the local authority), the project intended to build up in the beneficiaries a self-sustaining watering capacity. It therefore financed the purchase of 735 animal-drawn tanks, thus covering 13% of the beneficiaries involved in the fruit plantations. It also implemented 12 boreholes (three originally planned) and 12 shallow wells (initially 12). These various funds also covered the supply of drinking water. Several boreholes are already or can be used for irrigation. Drinking troughs (11 out of the 12 initially envisaged) were provided for livestock but, into account the existing equipment, the disinfestation dips were considered of no use.

Soil and water conservation measures (CES). These measures were not originally envisaged in the project but proved necessary to counteract the risk of erosion in areas just cleared or tilled with mechanical means. They involved 2 000 ha of banquettes on privately-owned farms.

Planting of perennial fodder crops. The project had planned to plant perennial species on privately-owned farms (by providing loans). The farmers preferred fruit trees, so the perennial fodder crops were planted on land reserved for collective use. The project thus funded, as a public investment, 1 150 ha of acacias and spineless cactus (2 000 ha planned). The cost was  400 TND/ha for cactus and 850 TND/ha for acacia (the average cost of fruit plantations being 645 TND/ha).

Provision of credit. The planned amount of credit was TND 3.6 million, of which 70% (SDR 1.46 million) to be financed by IFAD. The total credit distributed amounted to TND 2.0 million, entirely financed by IFAD. The plantation loan covers 15 to 24 years with a grace period of 8 to 12 years. It is guaranteed by the title-deed. The medium-term loans for equipment (water tanks) have already matured, and 64% of the repayments are outstanding. The long-term credit only concerned 20% of the farms receiving support to plant fruit trees and approximately 30% of the areas planted. It only accounted for 18% of the project costs compared to the forecast 40%.

Institutional support. The measures involved constructions (OMVPI headquarters, which were subsequently passed on to CRDA, and two territorial extension units) and extension (external training courses, local extension activities, purchase of equipment). They were also to cover monitoring and evaluation activities. The Planning Unit did not have time to operate effectively before the dissolution of OMVPI. The CRDA took over some of OMVPI's activities but within the framework of a service that performed more general functions.

Project effects

Effects on the target group. The beneficiaries were, in theory, chosen following the logic of farmer "demand". But in fact, a somewhat "spatial" logic was used, since the experts had to determine which areas were suitable for the plantations. The distribution of the 5 720 beneficiaries involved in the tree crop plantations shows that 43% of the areas were planted on land belonging to 18% of the beneficiaries who planted more than the established limit of 3 ha (30% of the area belongs to 10% of the beneficiaries, the latter having planted more than 5 ha). Small-scale planters (with a planted area of less than 2 ha) are the most numerous (64% of the beneficiaries) but they only own 35% of the areas planted with support from the project. The structure of the plantations shows that the larger holdings have a proportionately larger area under pistachio trees and the smallest have a proportionately larger area under olive trees - almond tree plantations remaining approximately stable (20% of the planted area for all farm sizes). A comparison of these data with those of beneficiaries by farm size shows that 22% of the beneficiaries (farming 56% the total land area owned by the beneficiaries) have areas of more than 20 hectares. The beneficiaries with less than 10 ha account for 47% of the total (17% of the farmed areas).

The comparison of this breakdown by farm size with the total farms in the Governorate shows that the larger categories (20 to 50 ha and more than 50 ha) were favoured because they account respectively for 26 and 30% of the beneficiary holdings, whereas farms of less than 5 ha and 5 to 10 ha only account for 10 and 17%. On average, 18% of the total farms benefited from the project's fruit-tree plantation measures. The overall objective achieved is of the order of that forecast (20% of the farms) but it is distributed differently: 57% of the areas planted should have been on farms of less than 10 ha (compared to an actual 36%), and farms larger than 50 ha were not intended to be included (13% of the planted areas did benefit however).

Medium- or long-term loans were obtained by 1 890 BNA clients. The distribution of these loans shows that the most widely represented category (less than 10 ha, 61% of the farms) received the least credit (18% of the clients). The holdings with more than 20 ha (15% of the farms) provided 45% of BNA's clients. The amount of the loan ranges from TND 364 per farm for the category of less than 10 ha to TND 1 986  per farm of more than 50 hectares. It is observed that 40% of the farms of more than 20 ha and which benefited from the tree crop plantation component, did receive some credit. The proportion is reduced to 20% for farms of less than 10 hectares. The subsidies having been distributed according to the areas planted, it is the larger growers who received the most: 43% of the subsidies thus went to planters who exceeded the established threshold of 3 hectares.

The beneficiaries of the other measures of the rainfed project include some 300 livestock producers who use the collective fodder crop plantations. They include the beneficiaries obtaining drinking water from the boreholes and wells (14 000 persons served). Approximately 90 farmers irrigate their crops with water from the boreholes.

Effects on production. The forecast fruit tree yields are generally lower than those estimated: almonds 300 kg/ha instead of 350 kg; pistachios 210 kg/ha instead of 250. Olive yields are 700 kg/ha. Bearing in mind that the planting of the trees is staggered in time, full production will be reached in 2012 (25 years after the first plantations). Additional production will be 4 976 t of olives, 811 t of almonds and 919 t of pistachios. The net value of the additional production (including cereals and by-products) will be TND 8.4 million, equivalent to TND 616  per hectare planted. The net value of fodder crop production achieved with the project is estimated at 49 TND/ha, for an investment cost roughly equal to the average cost of the trees planted. It is observed that the area under almond trees is 60% less than that forecast, while that of pistachio trees is approximately the same as that envisaged. The olive tree, however, which is well-known and reliable, is less profitable than the other species : it accounts for 50% of the areas planted but only 20% of the net value of the additional production.

Impact on income and employment. The gross income obtained on one hectare planted is TND 668  for almonds, TND 1 794  for pistachios and TND 222  for olives. The remuneration of a day's work, for each species, is valued at TND 33, 69 and 15. The gross income before the project (cereal crops and extensive livestock production) was of the order of 100 TND/ha. The annual income projections with respect to the debt schedule show that the repayments will not affect the farmer's income in the case of almonds or pistachios. But the problem exists for olives which show a negative balance between the eighth and the twentieth year.

Economic rate of return on investments with the project. The rate of return of the project over 30 years is 16.4%, compared with the appraisal's estimated 13%.

Impact on the environment. The project has had two opposite impacts on the environment: on the one hand it has exposed the planted area tilled with mechanical means to water and wind erosion, which is partly compensated by the soil and water conservation works (CES); on the other hand it has increased the biomass which adds to that of the plantations set up in the last thirty years or so, thus modifying what was formerly an arid and steppe-like ecosystem.

Recommendations

The overall outcome of the project is positive. It could be improved in the future if all efforts are made to increase yields where possible, notably for pistachios. The basic lesson learned is that it is possible to invest in an economically profitable manner in areas with poor potential in Central Tunisia under two conditions:

• the interventions must be part of a sound regional development policy;

• the Government's direct financial aid to private investment must integrate the self-financing of farmers.

That being said, it is also to be stressed that, contrary to the intentions of the project, smallholders have in proportion benefited less from the project than farmers belonging to the more prosperous categories. The lack of integration and coordination among the various technical services involved, the inefficiencies of the financing systems, and the geographic concentration of the project's actions have greatly contributed to this imbalance.

Action targeting

• The beneficiaries must be better identified in order to more effectively direct the financial aid towards the target group (primarily holdings of less than 20 ha). The larger farms must be oriented towards financing mechanisms different from those envisaged for smallholdings.

• The project accounting plan system should be revised so that the data on actions implemented are readily retrievable – other than in the form of budgetary items as defined in the loan agreements.

Action integration and technical approach

• The range of actions should be extended in order to better respond to the diversity of the farms. The rainfed agriculture development project was primarily for rainfed tree crops, and operated according to poorly differentiated procedures. The development of rainfed farming calls for more diversified solutions adapted to the different types of holdings. Livestock production, which is essential for rainfed farms, should thus be the object of targeted approaches.

• The technical solutions should be more effectively diversified according to the areas and farms concerned. The planting techniques were, overall, well assimilated. However, there are weak points as far as the pistachio, objectively the more interesting crop, is concerned. The planting actions were generally well accepted, particularly thanks to project support for preparatory operations.

Financing

• The credit needs and conditions should be more efficiently evaluated. The credit regulations, which comply with existing regulations, did not take into account the contributions of the project, so that the credit required was overestimated. Loan reimbursement over a period of 18 to 20 years is unrealistic and leads to forgetfulness and outstanding payments. The financing plan should take into account other sources of farm income - notably irrigation and livestock production. It should be possible to make a specific case-by-case analysis.

• The duration of Government support with regard to initial maintenance should be extended. As in the preceding programmes, this type of support, in the form of works or subsidies, proved to be indispensable. It is indeed very difficult for most smallholders to go beyond the phase of plantation start-up. The experience acquired in other programmes show that these farmers should be monitored over a period of three to five years. The subsidies should, however, be reduced or suppressed for large- and medium-sized holdings for which more important sources of financing are readily available.

Institutional support and monitoring

• Consultation mechanisms should be set up so that the technical services adopt an "integrated" and multidisciplinary approach in their work. Systems should be introduced to disseminate information on the actions and results obtained in each sector.

• The national extension strategy should be better adapted to the local needs and conditions and programmes should be draw up with the users and adapted to farm diversity. Approaches originating from central bodies and based on "national orientations" should be more flexible.

• A monitoring and evaluation system should be devised to make better use of the capacity to adapt already shown by the CRDA agents.