Interim evaluation report

Despite an estimated real GDP growth of 10% an outcome of expansion on mining, 50% of the population live at below subsistence level. Agriculture remains the dominant sector, though by 1980 its share of GDP dropped from 40% to 20%. At appraisal time over 80% of the population still depend on agriculture for employment and subsistence. Yet due to semi-arid climate and erratic rainfall, arable land is limited to about 5% of the total area of the country; hence, the country relies heavily on food grain imports. The Arable Lands Development project (ALDEP) is designed within an overall programme in which other multilateral donors participate e.g., AfDB. The programme aimed at raising the production of food grains by small farmers and make the economy less dependent on imported food.

Project design and objectives

The project was based on the Arable Lands Development Programme originally conceived by the Government of Botswana as early as 1977. Within the context of this Programme a pilot phase was initiated in 1979. IFAD programming Mission (August 1979), ADB/IFAD/Identification and IFAD Preparation missions drew up the design for ALDEP from the experiences of this pilot scheme as well as from other research projects in Botswana.

Target group

Within the project area all farm families, except those with more than 40 head of cattle, constituted the target group of ALDEP. Targeting was articulated around cattle ownership because the number of cattle was taken as an indicator for access to draught power (and obviously cattle would be a proxy for wealth in such communities). Draught power is the single greatest non-physical or non-climatic constraint to arable farming in Botswana because of the bearing it has both on the total area a farmer can plough/plant in a season and on the timeliness of those operations. "Access" to draught power was therefore used to categorize the traditional arable farmers.

On the basis of cattle ownership and access to draught power the farmers of the target group were ranked into:

• Model I - "No draught power" farmers - defined as those owning no cattle; 3 000 farmers (27%), each with an average total land area of 7 ha, cropping 5 ha annually.
• Model II - "Inadequate draught power farmers" - defined as those owning 1-20 head of cattle; 3 000 farmers (27%) with an average holding of 9 ha, cropping 6 ha annually.
• Model III - "Adequate draught powers farmers" - defined as those owning 21-40 head of cattle; 4 000 farmers (36%) with an average holding of 10 ha, cropping 7 ha annually.
• Model IV - "Molapo" - dryland crop farmers;1 000 farmers (molapo, 9%) with an average holding of 4 ha cropping 3 ha annually.

Objectives and Components

Objectives. The principal objective of the ALDEP was to assist small subsistence farmers to increase the production of basic food grains (sorghum and maize) and legumes and sunflower in order to achieve self-sufficiency at household and national levels and raise rural revenues and improve income distribution.

Components. The project would comprise the following components: (a) on-farm investment; (b) seasonal inputs; (c) strengthening of the extension service; (d) strengthening of the credit service; (e) strengthening of the marketing input supply and distribution system; (f) project management and coordination; and (g) monitoring and evaluation.

Expected effects and assumptions

The incremental food grain production (maize, millet, sorghum) and pulses by these farmers would be about 18 500 tons annually a full development. Additionally, there would be incremental production of about 3 000 tons of cash crops (mainly sunflower) that would be directly marketable. At full development farmers would not only be self-sufficient in basic food grains but would also realize substantial marketable surplus to improve their income. The project would make a significant contribution to the GOB's goal of redistributing wealth from the mining sector to the poor segments of the rural population. Furthermore, the project would contribute to an improvement in national food self-sufficiency. The project's contribution through external and on-the-job training to institutional building would be considerable and would strengthen the extension service and agricultural credit institutions beyond the project life.

The benefits expected from the project are likely to be adversely affected by several conditions. These conditions relate to the level of subsidy to farmers and to measures designed to provide relief to farmers and credit institutions in the event of severe drought resulting in crop failure.

Evaluation

A team consisting of IFAD Senior Evaluation Officer and five IFAD consultants in the fields of extension, agronomy, credit, sociology, livestock, and monitoring and evaluation visited the project during January/February 1992 for a period of three weeks approximately. The objectives of the interim evaluation mission were to review project implementation, evaluate the extent to which project objectives were met and to review design and implementation of research policy and output of relevant agricultural technology. The mission reviewed extensively the project activities in the field, had extensive discussions with project and government officials, analyzed data on project impacts and beneficiaries participation. It consulted vast literature relevant to Botswana and the project as well as project reports.

Implementation context

The period during which the project was implemented had six years of consecutive droughts and eight years of drought out of ten. As a result, on-farm investment component was seriously impeded and the extension activities were mainly limited to input distribution and drought relief programme. In response to drought the Government of Botswana (GOB) initiated the Drought Relief Programme (DRP) and the Accelerated Rainfed Arable Programme (ARAP) through which GOB distributed subsidized inputs and grants to the project target groups jeopardizing project activities, particularly the input supply distributed through credit.

The GOB had windfall income from mining and has continued disbursement on project activities without requesting reimbursement from the financiers. For example, at interim evaluation the project disbursement from IFAD components have been estimated at 30%, but only 8% were requested for reimbursements. For AfDB loan disbursements were negligible prompting a cancellation of the loan, and the redesign of the project and its scaling down to about 50%. However, in terms of local counterpart fund the project experienced no problem.

Project achievements

Production effects. The project in many ways has performed well in institution building but expected production effects were not achieved, largely because of external factors. The immediate goal of ALDEP was to increase production among the low-income households, but firm data with which to determine positive effects are not available.

Raising potential food security on-farm. A very substantial physical participation of farmers in the adoption of implements and a building up of a reservoir of equipment among Botswana's poorest farmers. Some 43 000 packages were distributed to 32 000 farmers compared to planned 23 470 packages, i.e., an over-achievement of more than 80%

Institutional Strengthening

(a) A strengthened extension system with the position of the extension work (AD) in the village being reinforced.

(b) The conduct of national tillage trials whose results show promise for the future, but also confirm the difficulty in disseminating standardized recommendations for farmers in situations with pronounced location-specific variations in climate, moisture and socio-economic conditions.

(c) A management system that has grown in strength during implementation and has established, not without difficulty, satisfactory relations with other parts of the MOA.

Adoption of Draught Power. Draught power available immediately at the onset of rains for the typical range of 10-15 days of ploughing/planting is a primary requisite in arable farming. But a foremost concern has remained the low adoption of draught power; only 14% of the ALDEP beneficiaries adopted draught power compared with the appraisal objective of 54%.

Limited Adoption of Equipment. Less than 20% of households without cattle received packages compared with 27% envisaged at appraisal, and the 40% set out in the Reformulated Programme (there is now some evidence that adoption is rising among these farmers).

Credit and Stores. The credit component and the construction of the nine lock-up stores did not perform. This was not a fault in implementation, rather the capability of the National and Cooperative Development Banks was assessed too optimistically at appraisal.

Monitoring and Evaluation. The establishment of an M&E system which has been extremely valuable in monitoring the project in spite of limited computer support. If strengthened, it would be able to conduct useful in-depth evaluations.

Effects assessment and sustainability

Without the necessary statistical data, the mission itself compiled and processed data from some 400 available questionnaires. Analysis of these data for the years 1989, 1990 and 1991 (including a separate 1991 sample of female-headed households) does not attest clearly to difference in production achieved by ALDEP farmers compared with those who did not participate in the project. Nor could production increases be universally attributed to particular types or sequences of equipment.

The majority of rural households persist with their traditional methods - broadcasting the seed and ploughing it in - rather than adopting the ALDEP-recommended double ploughing and row-planting.

Beneficiaries. Model II farmers were the largest group to take up ALDEP packages. Model I participation remained low especially in the southern region where there are fewer such farmers. The Francistown region recorded slightly higher than average participation by Model I farmers.

Area cultivated was not a primary targeting variable and for all the three years of the impact surveys, more than half (52%) of farmers had land areas in excess of 10 ha. Less than one-third (28%) had fields below 6 ha. In 1991, 37% of all farmers fell within the appraisal target ceiling of up to 7 ha for Models I, II and III.

Food Self-Sufficiency. According to Impact Survey (1991), half of the sample (48%) produced no cereals in 1991. Of the cereal-producing households, only 16% produce more than 200 kg of cereal grain per resident member. Most of the remainder produce about half of the resident members' needs per season but 40% produce roughly a mere 15% of annual cereal needs. The food security situation at the household level is precarious. Farmers have to depend on other sources of food - either purchased or obtained from other sources. There is a dire need to look carefully at production systems and to develop sustainable income sources such as goats or small stock.

Specific Effects on Women. About 20% of the participating households are headed by females. This percentage is consistent across years and regions except for the Southern (7%) and Maun (9%) regions where a lower than average female participation was observed in the 1989/90 survey. Most female households are in the Model I category whose participation was low. An increase in Model I participation in the project would mean an increase in female household participation.

Credit. The credit component of the project failed and was abandoned unilaterally by the Government in October 1983. The high cost of servicing a large number of small loans. The National Development Bank (NDB) and the Botswana Cooperative Bank (BCB) were not staffed or structured to deal with small-scale rural credit; they had a poor network of branches and long distances were involved in reaching farmers, creditworthiness, and hence the rate of repayment by recipients was very poor.

Nevertheless, the NDB did in fact make olans for some 2 464 packages under ALDEP indicating that there was a considerable interest in credit amongst the farmers at the beginning of the project. It is also notable that the uptake of packages obtained on credit was better balanced (more draught power and planters than ploughs) than under the down-payment scheme. Subsequently, many of these loans were converted into grants less down-payment.

Extension. The project has been partially successful in implementing its extension component. The construction, early in the implementation period, of 130 houses with adjoining offices for extension workers has made a positive contribution to their morale and sense of job satisfaction. Many people under-estimate the difficulty facing extension workers in the ALDEP project. They are not dealing with progressive farmers keen to adopt new technologies and risk changes in their farming practices.

It may not be feasible for farmers to respond to the main extension messages. For example, since the mid-1970s, recommended practice has been to plough in spring and row-plant - sorghum at 50-70 000 plants per ha and maize at 35-40 000 plants per ha. Recommended fertilizer application is 250 kg of single super phosphate per ha. Recommendations on planting and fertilizer application have not been widely adopted by farmers, largely because they require more draught power and labour than the traditional plough/broadcast method and yields are not much higher. Farmers perceive little net benefit in changing from a well-understood traditional method.

Environmental Effects. The project has been gravely affected by the adverse climatic conditions, rather than the opposite. Consistent with the rainfall pattern, the number of farmers planting was on the decline. While 88% of farmers were able to establish a crop in any given year. However, only about 68% harvested their crop in 1991, because crops failed to reach their maturity. In 1989, because of good rains the harvest was particularly good, about 80%.

Sustainability. Given the competition from other government programmes, which have large grant elements, the sustainability of project activities is questioned. Indeed, the whole strategic orientation of the project would be seriously impaired, unless GOB reverses the grant-based rural interventions. Credit institution sustainability has been further undermined by this policy, in addition to the common perils which rural financial institutions are familiar: with lack of appraisal of rural enterprises; collaterals and creditworthiness; low repayments etc.

Some project activities were given significant assistance in physical terms, such as office buildings and housing to extension staff. Given that GOB does not have serious problems with funding at the time of the evaluation, extension service sustainability hinges on the availability of the technical messages, and on the mobility which the project and GOB put at their disposal. In the previous years, drought conditions did not allow extension to demonstrate its potential.

Main issues and recommendations

Target Group classification. The present target group classification based on number of cattle is simple, reasonable and equitable and should be maintained. Since draught power is the limiting factor in poverty alleviation, the case for shifting to livestock units including small stock is not supported.

Food security. It could be argued that there is little or no rationale for further assistance to the ALDEP target group through production support if food security is already guaranteed, but there is insufficient evidence to confirm this supposition. Indeed, the data suggest that household cereal production is not sufficient to obtain food security. The vast majority of farmers require outside sources of income to meet cereal consumption requirements. Although farmers have access to additional resources, there is no statistical base for the assertion that these resources are sufficient for a minimum standard of household food security. Furthermore, remittances from RSA have likely fallen over the last two years.

Development and dissemination of appropriate technical packages. The farm budget analysis confirms that at present technology levels donkey draft is superior to oxen draft for cash- and labour-constrained households, i.e., especially for the FHHs. Hence, a strong case can be made for credit provision for such households to purchase donkey draft. At the same time, efforts should be made to furnish donkeys to FHHs outside of the Maun area, where donkeys are in ample supply.

The National Tillage trials have yielded valuable information and should continue and be extended: (a) technology generation and refinement need continued support and should be viewed as an ongoing process; (b) with the technology tested to date, smallholders are more likely to see greater productivity gains during periods when rainfall is below rather than above average; and (c) tillage trials should also explore the possibilities inter alia of: reducing production cost through the use of a smaller number of better trained draught animals; and using alternatives to the mould board plough.

The demonstration farms represent worthwhile interventions but need to be supplemented by the setting up of a network of farmer-managed on-farm trials. In such trials, a representative selection of farmers is essential, and farmers, not researchers, should be in command. The reward structure should be neutral to the cultivation method used and farmers should not receive free or subsidized ploughing services. They should be encourage to adapt and improve upon available extension recommendations.

Diagnostic assessments are needed in order to better target project interventions. The project should support the setting up of diagnostic teams with the extension system to provide the in-depth feedback to the project management that is all too often missing. The diagnostic teams would also provide the information base for more effective research-extension linkage.

The evaluation report recommends a review of present extension policy. In part, such a review is predicated by the mismatch between the old and illiterate farmers and the mostly young and inexperienced extension workers (ADs).

Rural credit. There has always been a demand for credit amongst ALDEP target farmers; but the two main rural credit institutions (NDB and BCB) are not in a position to administer small farmer credit schemes. Local NGOs have particular advantages in the implementation of credit schemes. They are: (a) community based; (b) have local knowledge (for example, of the creditworthiness of groups and individuals); (c) have a self-interest in promoting the development of their areas; (d) often have management structures already involved in provision of small-scale credit; and hence, (e) have an even stronger self-interest in ensuring that any credit arrangements they operate are properly administered. Credit under a pilot scheme should therefore be administered through the same outlets that currently provide inputs to the farmers, i.e., the ALDEP/BCU-supported NGOs. The role of the AD and DAO would be to determine the eligibility of applicants and to make recommendations to the NGO. It is important that the extension staff not become debt collectors.

Credit should also be considered for the acquisition of small ruminants. The ownership of goats represents a security- and risk-reduction strategy, especially for poorer households. The feasibility of providing credit for the acquisition of small ruminants, not only for high-risk arable production, needs to be pursued.

Access to land and water resources. Meanwhile, as complementary activities, the project should target its interventions on the basis of availability of local micro-catchments and boreholes so as to ensure a minimum of drinking water for humans and animals. Government policy on groundwater use and boreholes should be reviewed to determine whether it can be amended to stem primary causes of rural poverty and better support GOB and IFAD objectives. There may even be a case for a temporary subsidy for borehole development that favours Model I and II farmers to offset climatic risk (instead of subsidy on-farm investment).

The broader conclusion emerges that government policy needs to be examined to see how it may be shaped to improve entitlements of small farmers-livestock owners and their access to common grazing and water resources; what are the trends with regard to their access to drinking water (boreholes) and to pastures for draught animals. Government policy on groundwater use and boreholes should be reviewed by a consultant to determine whether it can be amended to stem primary causes of rural poverty, and better support GOB and IFAD objectives.

Complementarily of project. The continued involvement of IFAD has assisted in promoting technology generation with a better focus on the risks involved in arable farming. The competing programmes, ARAP and DRP, have been closed. The GOB is intent on reducing and better targeting subsidies to alleviate poverty and wants to shift to a credit programme to achieve this objective. But at present productivity, the scope for reaching this objective is limited.

Lessons learned

IFAD's experience with ALDEP demonstrates the need for continuing involvement in order to obtain a lasting effect from technology generation and diffusion. A project that aims at poverty alleviation in fragile environments needs to have at least a 15- to 20-year perspective.

The Government's intention to reduce subsidies and shift towards the provision of implements and inputs on credit requires a finely orchestrated programme design based on data which are not yet available. The risk in such programme design is high. Pilot and phased projects are essential. The rationale for further IFAD support would have to rest in support for a shift in policy direction that serves to reduce costs associated with poverty alleviation and obtaining improved food security at the farm level. Hence, the rationale or economic justification should be clearly presented in terms of "cost savings", or "cost minimization" to achieve a given policy objective over a specified period of time.

Need of diagnostic studies. IFAD needs to pursue more in-depth the mechanisms and processes of rural poverty and regional and social mobility. It is especially important to understand the implications of present Government land tenure policy affecting access to drinking water, boreholes and the commons IFAD's mandate would be well served by such an analysis, in order to review with the GOB the primary causes of poverty and how to address them.

Household demand for on-farm investment is related to several factors. They comprise the need to acquire a minimum of basic food staples on-farm for household consumption, employment opportunities in non-agriculture in general and in the RSA in particular, risk in production, the availability and affordability of draft power, the availability of improved technology, and evolving GOB policy on subsidies. It is impossible to weigh in all these variables and to foresee production risk and location-specific variations so as to design in detail necessary technology and project interventions in advance.

For this reason, prior to passing from a small to a larger scale project, the technology proposed and generated needs to be based on a number of farmer-managed on-farm trials in representative sites to verify the improved technology generated. Such verification is essential since so many factors influence household demand for on-farm investment. The farmer managed trials become a natural point for extension meetings and demonstrations.

Implements as well as extension messages should only be disseminated if confirmed by financial and economic analysis. This has not been the case. The costs associated with the additional labour required for double ploughing and row planting should have come to the forefront at an earlier stage. A series of indicative farm budgets should have been prepared on an annual basis and been discussed with the extension service; they should have been specific to the different types of IFAD target groups and to regional variations, and then they should have been updated on an annual basis.