Interim Evaluation

Background

Objectives, components and project area

The Project aims at improving the incomes and food security of Swaziland's disadvantaged smallholder families, particularly women on Swazi Nation Land (SNL). It sets out to improve incomes of smallholders and standards of living of disadvantaged smallholders on SNL. The main objectives of the project are to: (i) assist the Government in its overall efforts to achieve food self-sufficiency; (ii) strengthen the institutions already supported under the Swaziland Credit and Marketing Project (SCMP); and (iii) address constraints to raised productivity in rainfed crop production and the livestock sub-sector, particularly the need to introduce new systems of production that will reduce the number of cattle in the range. The project intended to develop grassroots institutions to achieve the project objectives.

The four major project components of the project are: (i) crop production including irrigation; (ii) livestock production; (iii) credit; and (iv) institutional strengthening. About 5 000 households would receive the full complement of inputs and services. Other families on SNL would benefit from general extension and adaptive research services.

The project area consists of the SNL areas, which comprise up 60% total arable land 70% of its population. It was intended to include all the four agro-ecological zones. Support for agricultural development particularly in terms of adaptive research and small scale irrigation development, would be directed to the Lowveld and those parts of the Middle and Highveld, where the irregular rainfall pattern is a major constraint to agricultural production.

Projects costs, financing and implementation arrangements

The total project costs for the seven-year disbursement period were estimated at E 31 million (USD 8.4 million, including physical and price contingencies). The foreign exchange costs of the project comprised 63% of the total base costs, equal to USD 5.2 million. Physical and price contingencies represented 12% of project base costs.

IFAD had approved a loan of USD 6.9 million to cover 83% of total project costs. Government of Swaziland /SDSB/NAMBOARD had agreed to finance USD 1.5 million, of which USD 0.8 million would be taxes and duties. Beneficiaries would contribute about USD 0.1 million reflecting their down payments for loans received under the line of credit and labour input for irrigation development and construction of grazing land management areas (GLMDA.)

Project organisation and management

Responsibility for the implementation of each component was assigned to the relevant technical department or agency. The project would seek to direct existing services to meet specific target group needs. It would provide any necessary incremental support for this purpose. The Department of Agriculture would be responsible for the dryland and irrigated crop production programmes, for extension and adaptive research services. For small-scale irrigation development, responsibility would be reorganised in Small-scale Irrigation Section within the Department of Agriculture. The Departments of Veterinary Services and Co-operative Development, respectively, would be responsible for the livestock production and co-operative development component. NGOs or other entities under the auspices of the PCC would undertake group development. SDSB would be responsible for credit and the NAMBOARD / Nokwane market for input supply and produce marketing. The African Development Bank was assigned to be the Cooperating Institution (CI).

Financial performance and procurement

Financial performance

IFAD appraised the project early in 1993. Total project costs are US$ 8.56 million, of which the IFAD loan represents SDR 5.1 million (USD 6.9 million). The Project Accountant could not provide details of the balance available for disbursement during the remainder of the project life. The TA has been ineffective in assisting the Project to design the systems and procedures. The situation was compounded by a high rate of staff turnover. The budgeting process in the project was top-down, providing little or no space for participatory processes, reflecting the government's traditional budgeting system. The community organisations had no role in budget planning exercise. Beneficiaries, project implementers - and the mission – fail to ascertain the quantum of investments and cost effectiveness of investments made by the project at each community/group

IFAD's cumulative disbursement as of 2 May 2001 was SDR 2 694 499, or 47% of the total allocation of SDR 5 100 000. The actual rate of disbursement has been much lower than the projected rate. The AfDB operates with a different set of data, which is not reconcilable with IFAD data. The AfDB data indicates a lower disbursement of merely SDR 2.02 million, representing 40% of the IFAD loan amount.

Procurement

The project design indicated that the Government procurement procedures were satisfactory to both IFAD and the CI and would be followed under SADP. But procedures are excessively complicated for smaller procurements. For any procurement exceeding E 5 000 up to USD 50 000, the PCU has to obtain three quotations and seek authority to spend from the tender board in the Ministry of Finance. These procedures are unpractical to follow in a project setting and easily cause non-justified delays.

The CI has remained too passive regarding involvement of NGOs for micro-schemes and engagement of private contractors for small-scale schemes. The project had initially engaged NGOs from developing micro-schemes. The project has made partial payments towards these contracts. The NGOs due to non-payment on their own had to switch to other donors to obtain funds for completing the works. The non-payment of NGOs and the absence of a well-designed procedure to engage the NGOs to implement project activities contributed to the decision by NGOs to disengage themselves from the project.

Project management capability

The Project established a Project Co-ordination Unit (PCU) with a PC, Counterpart Project Co-ordinator (TA), Project Financial Advisor/Financial Controller (TA), Project Accountant, Economist and Secretary. A Project Co-ordinating Committee (PCC) was established. The project subsequently established a system of monthly meetings at the PCU level chaired by the Under Secretary.

But this system for coordination did not meet expectations. The different complementary project supported services did not materialize as expected at the site level. Instead, each sub-component is being implemented in isolation, independent of other activities. Project activities have become fragmented. The co-ordination of the various project-implementing departments has been inadequate. Two issues are evident. First, the project work plan and budget preparation was component driven contrary to the design expectations. Second, the regional level officials were not involved in preparing the project work plan and budget to ensure co-ordinated services to the community / farmer organisations.

The capacity of the project management has been inadequate. The project was established as a small unit with only three staff. The project technical assistance became ineffective due to the turnover of trained staff and inability of the TA to conceptualise community institution development as the core activity and as the basis to fund other interventions. The project staff are drawn from the MOAC cadre; they consider that their compensation levels do not match the project workload. The component heads were required to undertake the project activities in addition to their regular work with little difference between them. As a result, project activities became an appendage of the regular department work.

The project design had provided an initial framework to develop sustainable farmer organisations by using NGOs. The framework, though not well articulated, could have been adjusted based on the experience in participatory development in other countries. The project initially contracted NGOs but with the unresolved dispute over contract payments, NGOs withdrew. The project was neither assisted in settling the payment dispute with NGOs, nor to come up with an alternative framework for developing farmer organisations. The project's intention of developing farmer organisations in the irrigation schemes was thwarted.

The supervision by the CI was inadequate. The Aide Memoires of the CI indicate simple recording of performance as against the appraisal targets. Efforts of the supervision missions to set out the issues and recommendations to address the issues with clearly articulated implementation steps are not evident. In addition, strategy development and adjusting the project activities based on the international experience gained in participatory development and lessons learned from implementation experience have been neglected. The issues related to NGO involvement in micro-irrigation and group development, non-payment of contract amount to NGOs, constraints to deliver credit to the target group through SDSB, financial management, the flow of funds and use of private contractors for irrigation works remain unresolved.

Project design weaknesses

Overall, design features have been inadequate. First, the processes to develop community institutions/farmer organisations as the core of development architecture have not been articulated. The sectoral interventions were not based on the assessed needs and priorities of these institutions/organisations. Second, small stock development (pigs and poultry), as preferred by the IFAD target group, was largely ignored. Third, processes for engaging the NGOs, including selection, contracting and payment procedures were not spelt out. Fourth, the risks of involving only a single financial institution – the SDSB – to deliver credit to resource poor households (given the experience in the earlier SCMP) were not explored. Fifth, the capacity and the need for direct execution of irrigation works by the Irrigation Section were not analysed: the limited capacity and the residual option of contracting irrigation works to private contractors were not made clear.

Targeting

The project design identifies four categories of homesteads on the basis of cash or resource endowment, homestead labour, and associated farming practices. They are: (i) cash/resource rich - labour rich; (ii) cash /resource rich - labour poor; (iii) cash/resource poor - labour rich; and (iv) cash/resource poor - labour poor. It was suggested that the target group for the project could be found in the latter two categories.

The project design did not set out a plan to operationalise the targeting criteria to select those households that should benefit from the project. It did not even specify an area with concentration of target group households, or establish a methodology to identify such areas with concentration of target group households. The project allocated 24% of the funds for cattle development catering to farmers, who do not fall within the target group criteria. The situation is worsened by inadequate emphasis on small stock development (poultry and pigs), which are the activities preferred by the cash/resource poor and labour poor households.

Implementation progress by component

Irrigation development

The project had intended to: (i) expand two existing SCMP schemes by 5 ha each, to make use of water delivery infrastructure already provided; (ii) develop or rehabilitate about 75 ha of small-scale irrigation schemes; (iii) develop/rehabilitate one hundred hectares of new, intensive micro-smallholder irrigation schemes at an average size 1 ha; and (iv) consolidate and reorganise another 257 ha of existing schemes to promote farmer-management. The irrigation component was to benefit 2 215 farmers

Twelve small-scale irrigation schemes have received attention from the project to date, including one of those intended for expansion (Ntamakuphila). The total irrigable area of these schemes at full development will be some 214 ha. However, even though substantial work has been done on some of these schemes, none of them is yet fully operational, even though seven years have elapsed since project effectiveness. Physical works have not yet been started on four of the schemes Compared with the ‘without project' situation, the incremental area developed/rehabilitated and brought into operation/production at the time of the IE is less than 10 ha, compared with the target of 85 ha. Three micro schemes, amounting to a total of about 4.5 ha, have been completed by SFDF and are operational. A further three micro schemes totalling 3-4 ha are reported by the Irrigation Section to have been completed (possibly by Women Resource Centre).

Thus, the sum total achieved to date is less than 10 ha of small-scale schemes and about 4.5 ha of micro schemes, compared with the targets of 85 ha and 100 ha, respectively. Nothing has been achieved on the intended consolidation and reorganisation of the 257 ha of ‘old IFAD schemes'. Moreover, regrettably, the quality of design and construction work leaves much to be desired. The approach to development has also sometimes been top-down

The main reasons for these disappointing results are: (i) inadequate manpower within the Irrigation Section as it has only one irrigation engineer to cover the whole country and the intended overseas training for two irrigation staff apparently never took place; (ii) difficulties with procurement of construction materials and the flow of funds, mainly as a result of the lack of a special account, combined with cumbersome procurement procedures; (iii) difficulties in arranging for force-account work through the Land Development Service which apparently is poorly equipped; (iv) apparent reluctance to contract out for plant hire or construction services even when the Land Development Service is unable to assist; (v) lack of awareness of the state of project finances and consequent inadequate budgeting; and (iv) failure on the part of the PCU to successfully contract NGOs.

However, there are two encouraging points: (i) the schemes have generally been initiated by the farmer groups, who have unstintingly provided their labour (and a small amount of cash) towards scheme construction; and (ii) the groups generally appear to be committed to meeting operation and maintenance costs. The completed micro-scale schemes appear to be functioning well.

With one notable exception (Lavumisa), targeting has generally been satisfactory. Again with the exception of Lavumisa and in the case of Mhlangeni, women are in the majority in the memberships of the scheme. They far outnumber men in the memberships of the micro-schemes. The case of Lavumisa is cause for concern. This 51 ha sprinkler scheme has not been commissioned since it is not yet fully equipped. It was developed together with a 120 ha drip irrigation scheme for sugar cane and a 73 ha drip irrigation scheme for vegetables (which has now apparently been abandoned). This does not appear to be in keeping with project guidelines for targeting.

With the exception of Lavumisa, allocated plot sizes appear to be reasonably in keeping with expectations, although the process may be a cause for concern, since the project has not documented ‘ownership identity' of the schemes. There is little in the way of documentation on file to show that the Chief has given his permission for the use of the land on any of the schemes and that the users (particularly women) have security of tenure.

There is no doubt that the few developments completed to date are having a significant impact on household incomes and nutrition and that these benefits in most cases will far outweigh the costs. However, this situation will continue only if the present limited market access is maintained or improved and the schemes are sustainable. Neither is at present guaranteed – and prospects have not been improved by the failure to contract NGOs for group development.

The quality of supervision is a cause of concern. Only one supervision mission included an engineer. All irrigation sub-components require an engineer to participate in supervision missions, for example to endorse subproject investment decisions. If there are budgetary constraints to including an engineer in supervision missions, future project designs should concentrate on even lower technology than employed on this project – perhaps restricted to micro-schemes.

Livestock development

The services to be provided under the livestock component, as per project design, were not to specifically directed to assist the core poor homesteads. Instead, they were to reach a wider spectrum of SNL farmers. The emphasis was on disease control and prevention, destocking and education and training of farmers and staff.

The interventions proposed were: (i) completion of 15 km outer border Foot and Mouth Disease (FMD) fence on the border to Mozambique and vaccination of animals in the buffer zone against FMD; and (ii) reduction of the dipping intervals by researching the partial substitution of some of the dippings with vaccinations against tick-borne diseases; (iii) stall feeding experiments; (iv) support to the Livestock Marketing Unit to develop a marketing information system; (v) training of lead farmers and professionals on alternative production systems through one-week residential training courses at the VFTC in Mpisi.; and (vi) grassland management activity with the aim of involving the community to rehabilitate nine existing GLMDAs and to establish three new ones.

Activities to prevent the Foot and Mouth Disease (FMD) are based on well-established Government practices and procedures. Until an outbreak of FMD in November and December 2000 in the regions of Manzini, Lubombo and Hhohho, Swaziland had been free of the disease for more than twenty years. It had, in May 2000, obtained status, granted by the World Organisation for Animal Health, as a ‘FMD Free Country Without Vaccination'. The fencing work has been efficiently done.

Results are analysed from five dipping tanks to be compared with data from vaccination against tick-borne diseases. Seventy animals were vaccinated in 1996. Only one died, although none of the surviving 69 had ever been dipped since 1996. This result suggests that vaccination can be used to reduce the number of dippings. The current system of dipping with the government subsidising the entire cost is not sustainable.

The pilot feeding trials have failed. The trials were conducted on government farms and efforts were not directed towards finding cheap local feed materials. The dissemination of price data by the Livestock Marketing Unit is so far limited to cattle prices. The Unit has also conducted courses for farmers in grading of cattle. The Unit has completed building three sales yards and expects to complete the remaining three within the life of the project.

The Veterinary and Farmer Training Centre (VFTC) at Mpisi has played a useful role in training 1 511 farmers of whom 76% were women. The VFTC has established units for fattening of cattle, dairy production, broiler, egg and pig production with project funds. The GLMDAs are technically well established and demonstrate that rotational grazing leads to higher grass production. Pasture sites visited show very good stands of grass under fenced conditions.

The project has rendered partial support to the Philani Poultry Co-operative Society Abattoir, which has a daily slaughter and sale of 1 000 broilers. This is a co-operative with 53 women poultry producers that have built an abattoir, initially by members contributing E 12 000. Subsequently, funding was sourced from an NGO based in the UK to expand the abattoir. The present IFAD project has assisted with training programmes in management, preparation of business plans, training in marketing and record keeping.

The project design had put undue emphasis on cattle. This happened even though the SAR had reported that no less than 42% of the SNL homesteads did not own cattle. The baseline study conducted for the Mid-Term Review found more than 60% of the homesteads without cattle. Yet, for all practical purposes the appraisal report equated cattle with livestock. No attention was given to other animals – even though non-ownership of cattle might be taken as a proxy for poverty. The project has put most of its resources into cattle – with an inherent strong gender bias towards men – with no uptake of stall-feeding among SNL farmers. Women have taken to poultry and pig production in spite of very limited support from the project.

Crop development

Adaptive Research, Extension and Farm Mechanisation Support are the three sub-components of the crop development component. The adaptive research sub-component includes development of: (i) low risk packages for maize production with appropriate inter-cropping; (ii) suitable cotton package; (iii) fodder production; (iv) varietal testing on irrigated vegetable crops; and (v) production techniques to improve marketable quality and priority would be given to tomatoes.

The approach to improving rainfed agriculture in Swaziland has been that of trying to find substitutes for maize wherever rainfall is close to or below 800 mm/year (representing around 40% of the arable land). Thus, the adaptive research work in the project has taken two basic directions: (i) develop crops designed for irrigation schemes; and (ii) develop sorghum, sweet potatoes, cassava, groundnuts, cotton, etc., as somewhat drought-resistant alternatives to maize. The extension program has dedicated some of its efforts to the management and training of farmers in irrigation schemes. The farm mechanisation work has been much more focused on rainfed farming.

Some 16 varieties of some 10 species of field crops (maize, sorghum, cowpeas, groundnuts, sweet potatoes, and cassava), vegetables (onions and tomatoes), and fruits (avocados and mangos) are now ready to be multiplied and tried more widely by farmers; none of the seeds of these species have been multiplied. Thus, none of these investigative efforts have benefited more than a handful of farmers. The project has not recognised the strength of cultural habits, of the desire to grow maize and the need to improve maize cultivation practices. Seed producer groups for spreading drought resistant crops are part of the regular menu of donor supported activities in other countries such as Zambia and Zimbabwe: but the SADP design did not provide such support.

Extension work has reached virtually all of its recent objectives in terms of classroom-based trainings and field days conducted, but the impact on farmers' practices is once again fairly minimal, being restricted almost entirely to the spread of certain varieties of seed material. The farm mechanisation programme could compete with commercial tractors, only because the service is subsidised.

Credit

The project would establish a revolving fund with the SDSB to provide credit to the beneficiaries under the project. In this context, the project design comprised two important features. First, attempts would be made to improve the viability of the Agricultural Development Fund through group lending to reduce transaction costs and increase lending. Second, it was intended to use NGOs to train group leaders in leadership, group dynamics and record keeping and savings mobilisation.

The project had made an allocation of SDR 510 000 towards incremental credit. A subsidiary loan agreement was signed between the MOAC and SDSB as early as 1993/94. The subsidiary loan agreement is weak. It is not clear as to who would bear the credit risk and also about the subsidiary loan repayment modalities. Owing to its preoccupation with restructuring, SDSB has not participated in the project until recently. Only during the year 2001 did it submit a work plan and the MOAC submitted a withdrawal application during July 2000. Surprisingly, the entire amount allocated under the project was released to SDSB without taking into account the SAR stipulations on release of incremental credit funds and also without testing the ability of SDSB to fund project beneficiaries.

A statement provided by SDSB indicates that disbursements have been made to the tune of E 1 634 986 to farmers. However, SDSB is yet to deliver credit to the farmer groups mobilised under the project. The responsibility for the inability of SDSB to deliver to project farmers partly lies also with MOAC. The latter has not engaged NGOs to mobilise groups and link them to SDSB. SDSB on the other hand had received 24 person months of TA from the project. In spite of this TA, no SDSB strategy for group lending materialised. The bank even did not request the MOAC to engage NGOs to promote groups while submitting the work plan. SDSB is not geared to providing loans to the farmer organisations promoted under the project, as the transaction costs to reach such farmers are higher. Farmers who are interested in small working capital loans find it difficult to meet the requirements of 25% contribution and to prove their credit worthiness. The current situation of SDSB does not favour experimentation in service delivery to the resource poor households on SNL.

Marketing

The IFAD project has intended to address the constraints encountered in SCMP namely: (i) failure of Encabeni (formerly Nokwane) to attract smallholder produce; (ii) absence of Encabeni market/domestic producer linkages and (iii) lack of well planned and phased production programmes for market oriented production. Moreover, NAMBOARD's statutory and commercial functions were to be separated.

NAMBOARD is mandated by the parliament to regulate imports as well as to determine levies on scheduled produce (vegetables, maize wheat and wheat products and poultry and poultry products) in conjunction with MOAC. It is required to facilitate procurement of scheduled agricultural products to meet any shortfall in production for which import permits are issued and the levy at the border is collected. But, it does not yet have a mechanism to collect data on a scientific basis on local production and determine shortfalls to determine the quantum of import.

NAMBOARD has established a system of procurement of vegetables from the farm gate. It manages two depots (Lomahasha and Piggs Peak) in the country. NAMBOARD staff directly sell procured vegetables to major consumers and retailers through these two depots. NAMBOARD did separate the statutory and commercial functions for a short time. But, as the commercial functions are not financially viable, the separation of these two activities was rolled back. The statutory function continues to subsidise the commercial functions. It competes with private sector by offering better prices, using its purchasing power derived from import levy collection. The net result is that the levies collected from the imports of scheduled produce do not reach the national exchequer but are spent on maintaining the loss making commercial activity. At the same time, development of the private sector marketing sector is stifled to the detriment of the interests of the farmers and the consumers in the longer term.

The NAMBOARD under instructions from MOAC is now in the process of working out joint venture arrangements with a private trader for the Encabeni market, which is the hub of vegetable marketing. Encabeni is currently negotiating with Pick and Pay to export baby vegetables to South Africa. In addition, Encabeni also has a marketing agent in South Africa to supply vegetables to the EU market; the current export volume is very small.

Other institutional strengthening efforts

Drought early warning and contingency planning

The project intended to support the National Early Warning Support Unit (NEWU), which became operational in 1987. The Unit is located within the Economic Planning and Analysis Section. It was intended at the outset that the unit would have a Nutritionist, but that post is yet to be created. The Unit produces monthly food security reports. It acts as a secretariat for the national early warning technical advisory committee, which meets on a quarterly basis. All equipment and training have been provided. But the long-term training as agreed in the Mid-Term Review is yet to be provided.

CCU

The Project proposed a study to prepare a strategic plan for strengthening CCU to ensure that it emerges as a financially viable institution. FINTEC consultants Cairo was selected to undertake this study. The study amongst others proposed internal reorganisation of the CCU, increasing its trading activity, broadening its membership and implementation of several projects. This report was submitted in 1996. The CCU has now embarked on a restructuring process, assisted by the Enterprise Development Fund. Discussions with the officials of the CCU indicate that the CCU would be even interested in moving from trading towards producing fertiliser and animal feed. Yet, there is a need to carefully review benefits and costs of such aspirations since trading – but not production - remains the core strength of CCU.

MOAC Reorganisation

Zimken Management Consultants have been engaged to conduct a study on the reorganisation of MOAC. They have submitted a report in February 2001. This report has been discussed and approved by the MOAC and forwarded to the Ministry of Public Services and Information. The Ministry of Public Services Affairs and Information is now in the process of implementing the Public Sector Reform Programme. The MOAC reorganisation would be considered as a part of this programme. A weakness of this study is that it does not analyse the service delivery approach, diffusion and outreach methods and the organisational requirements to assist the resource poor SNL farmers.

Monitoring and evaluation

In October 1994, a two-day Monitoring and Evaluation Seminar was organised with the help of Mananga Management Centre. This included a session introducing the Logical Framework Approach. The log frame concept is being used in a rudimentary way to report project progress. The Monitoring Officer made some attempts to draw formats to report component-wise activities, outputs and objectives. But these were not adopted and used by the Component Managers. Subsequently, a three-month TA was provided to develop the project monitoring and evaluation system. Even now, the formats designed by the TA are not being fully used by the Component Managers to report progress in project implementation.

The project engaged a consultancy company to undertake base line survey. The baseline survey was conducted in 1996. The baseline survey provides very good details on overall poverty status and farming systems. But it fails to identify the impact points at the community and household level that need to be monitored during the subsequent evaluation/impact surveys. The evaluation surveys proposed to occur once in every two years have not been undertaken. The project was required to undertake an impact evaluation study as an input to the Project Completion Report. Actions to initiate the process are yet to be taken.

51. The project document did not set out the outcomes that could be measured to assess the project impact. It indicates calculating net worth of the project beneficiaries over a period of time as a means to assess the impact of the project. Collection of data on assets and liabilities to assess the net worth is subject to a large number of assumptions. Such an attempt would not provide accurate or meaningful data with which to assess changes in IFAD target group well-being.

Compliance with loan agreement and covenants

The loan agreement between IFAD and the government comprised two covenants. The first one pertains to periodic revision of the interest rates. It would be applied to credits to be made out of the proceeds of the loan so as to take appropriate measures consistent with the policies of the borrower in order to harmonise the interest rates on credits with the Fund's policy on relending rates. The second covenant requires SDSB to minimise its costs. Compliance with these covenants was of little consequence as the SDSB received project funding only in 2001.

Impact and sustainability

Poverty is not being reduced. The project was designed to improve the incomes and food security of Swaziland's disadvantaged smallholder families, particularly women on SNL. However, during the course of project implementation the food and nutrition security situation in Swaziland has deteriorated. According to the national nutritional survey conducted in the year 2000, 30% of children under age five are stunted while those wasted were less than 2%. Under weight, as measured by weight-for-age was found in 10% of the children. This means that a significant number of children start schooling when they are nutritionally compromised. The deterioration of nutritional status was the highest in Shiselweni region. Chronic malnutrition measured by children's stunting rate almost doubled, from 21% in 1995 to 38% in 2000. Nutritional status of children improved marginally from 1995 to 2000 in respect of Hhoho and Manzini whilst the situation was static in case of Lubombo

The IE mission commissioned a small survey to collect data on trends. A total of 244 households were interviewed. The majority of the heads of household were married (61 %), with about 27% widowed. About half of them were not employed, 27 % were in skilled employment and 22 % in unskilled employment. Close to 50% of the households received remittances. Half of the households had no cattle, while 18% had more than ten. Only 5% of the sample had no farming land. Nationally, the homesteads that never produce enough or do not produce maize represent more than 50%; in Shiselweni and Lubombo the proportion is higher than the national figure. Most of Lubombo and Shiselweni consist of the drought-prone Lowveld agro-ecological zone.

From this survey, it becomes apparent that the poor do not perceive their lives to be improving. First of all, very few of them identified sale of agricultural produce as an income source. Secondly, even fewer perceive that there has been an increase in income or consumption from the various agricultural produce. Due to the scattered nature of the IFAD projects, it will always be next to impossible to determine their impact on the socio-economic situation of households. For definite charting of the impact of the project, there need to be reliable indicators that can detect improvements in quality of life over time.

The undertaking of this survey further demonstrates first, the feasibility even by a rapid rural assessment to obtain impact data. Second, it demonstrates the high cost of missed opportunities to IFAD and the MOAC of not ensuring that at least minimal data sets are provided to gauge progress of a USD 8 million project, evidence of emerging impact or its absence, and seek to take corrective action within the project's lifetime.

The irrigation component did much less than expected. Incrementally, it added only 11.3 ha, of which about 4.0 ha were in small-scale schemes and about 7.3 ha are in micro schemes – compared with the targets of 85 ha and 100 ha respectively. Nothing has been achieved on the intended consolidation and reorganisation of the 257 ha of ‘old IFAD schemes'. However, the micro schemes that have been developed to date are producing a range of vegetables for home consumption and local marketing and appear to have a significant impact on household incomes and nutrition. From a consideration of the costs mentioned above and the indicative crop budgets and representative farm model these benefits far outweigh the costs. The project did not envisage a system of cost recovery for the investments made in the irrigation infrastructure. The farmer organisations were not developed to take over the responsibility of operation and maintenance. As a result, the sustainability of irrigation investments is in doubt.

Most of the planned adaptive research has been undertaken on maize, cowpeas, groundnuts, sorghum, tomatoes, cabbage, amaranthus, potatoes, onions, corchora and okra. However, there has been virtually no positive impact yet of all this work on the smallholder farmers' productivity. Of course, in some cases, the reason for the lack of impact was that the trials were unsuccessful, as in the case of the sorghum trials, in which bird damage was severe. Little attention was given at time of design to the step-wise processes that are required to move from the stage of adaptive research to widespread dissemination and uptake. Such steps were neither envisioned nor implemented.

The livestock component has been relatively successful when compared to other components. Until an outbreak of FMD in the regions of Manzini, Lubombo and Hhohho in November and December 2000, Swaziland had been free of FMD for more than twenty years. A partial cost-recovery practice was evolving in the dips was evolving owing to budgetary constraints. However, this was discontinued when the project funds were made available for this activity. This is contrary to what is required for a sustainable arrangement, where farmer involvement is necessary in management, investment and cost recovery of the dips. But an important spin-off effect has been the increased capability of the staff of the Epidemiology Unit to conduct on-farm research.

The training programmes conducted by the Livestock Training Centre at Mpisi have helped in upgrading the capabilities of farmers. Even though the SAR did not place any special emphasis on training in small stock development, the training centre developed a few modules in pig rearing and poultry farming. This effort has had a positive catalytic effect on uptake of small-stock related activities by the women. These activities at present depend on commercial inputs. Efforts to develop local feed resources are vital for long-term sustainability of these interventions. The income levels of the women undertaking pig rearing and poultry production have increased substantially. The inability of the women to access credit has been the main constraint; inadequate working capital has led to closure of a number of poultry units.

Credit and marketing components were the ones that have had least impact on the project target group. Despite the fact that the project disbursed the entire credit allocation to SDSB, no disbursements were made to the project beneficiaries. The activities under the marketing component never were geared to develop linkages between the private sector and the project beneficiaries. Produce has been purchased at relatively high prices with revenues generated from import levies: development of the small-scale private sector marketing has been discouraged.

Lessons learned

Targeting: SADP rightly targeted the SNL where most resource poor households live. The IE survey clearly demonstrates two issues related to identifying resource poor households. First, the project design did not take into account the regional disparities. This resulted in a scattered project implementation, increasing logistic and management problems. In turn, this has contributed to the difficulty in identifying and quantifying impact of the project. Second, the project provided merely broad or vague guidelines for identifying resource poor households. As a result, the project funded irrigation interventions appear to have been targeted to the resource poor households, but in reality reached the better off households. The project design did not pursue or reflect household survey data that show that some 50% of the households possess no cattle and about 5% have no farmland. The IE is of the view that these categories of households form the focus target group of IFAD.

A three-pronged approach is necessary to identify IFAD target group and to assess impact of the project. First, prior to project design, using relevant, valid and low cost data, the most vulnerable areas requiring project interventions will have to be identified. Chronic malnutrition data is one such indicator that represents overall wellbeing of the household/community. Subsequently targets need to be fixed as a part of the objectives of project to reduce chronic malnutrition. Second, with detailed survey data for these households, projects will have to design interventions that reach the poorest groups. These interventions are generally low cost / low capital in nature: they increase the risk taking ability of the households by focussing on their core strengths, and solutions that are found at the level of the community. Third, the welfare of the community over a period of time can be tracked by conducting low cost surveys to collect data on trends across sites in prevalence of chronic malnutrition. Reduction in these rates indicates gains in the welfare of the community.

Participatory processes: This project design had realised the importance of participatory processes as a vital instrument to permit resource poor households to articulate their preferences. It had also visualised the need for engaging NGOs to develop farmer organisations. But unfortunately, the project design did not have community mobilisation as a separate supporting ‘integrating' component that would form the basis for designing the interventions in other activities, for irrigation, livestock development and adaptive research.

Participation is generally not well understood. Some even regard mere consultation with the beneficiaries as participation. Four steps are essential to empower resource poor households to achieve participatory development. First, ascertaining the level of interest of the target group in participatory development is vital. This includes formulation of a social charter outlining the vision of the community with regard to overcoming issues that may relate to health, education, environment, natural resources, women and resource poor. The most important is the social action phase by the community to demonstrate their ability to work together by undertaking activities using their own resources. Second, capacity building of the project staff, NGO staff and the selected community members is the next important step covering community development planning and financial management. The third step involves community institution development and implementation of the interventions; these involve community level planning, facilitation of the community to identify low cost options and sustainable solutions, provision of necessary financial resources and implementation of the intervention by the community themselves. The fourth step represents monitoring the progress including financial discipline, evaluating the impact of the project interventions and adjusting the project priorities and interventions based on the lessons learned.

Supervision: Most if not all of the persistency of the problems encountered can be traced back to poor supervision. Supervising process-oriented projects is not as simple as sector specific projects. The CI needs to use for referral and periodic assistance a core group of consultants with hands-on experience in participatory development and technical interventions such as in the case of irrigation. Even if supervision costs were a constraint, the savings made by the CI by not including an engineer in the supervision are not comparable to the cost of unproductive investments made in irrigation infrastructure. There is reason both for IFAD PMD and the CI to consider the opportunity costs, the wider dimension and missed opportunities of an excessive economy in curtailing supervision budgets.
Project Co-ordination and Management: Too much reliance was placed for too long on a Project Co-ordinator drawn from within the ranks of MOAC, with little in the way of incentives to good performance – or disincentives to poor performance. Experience with other projects in the region suggests that a national Project Co-ordinator recruited from the private sector might have been more effective. Knowing that his/her job depended on it, he or she would have seen that a PCC was essential for effective project management and would have insisted that the supervision took the issue up at the right time and place.

Even though the irrigation section has no capacity to undertake direct implementation, oddly all irrigation works were assigned to it. As a result, the project had to rely heavily on government procedures for procurement and funds flow. This has resulted in piece-meal approach to irrigation infrastructure development. This issue could have been remedied if the project were to engage contractor for implementation and the irrigation section for supervision. It is essential that the government take up the role of the facilitator rather than that of the implementer.

Technical Assistance versus NGO and Private Sector: This project is yet another example of one that makes a feeble attempt to engage the NGO and private sector. More than three years have gone by: substantial funds were invested in technical assistance for the Irrigation Section without any sign of increased capacity. There was and is a dynamic private sector within Swaziland and across the border in South Africa that could have undertaken the design and implementation of most, if not all, of the schemes targeted. The Appraisal Report and Loan Agreement should be much firmer on the issues related to involvement of NGOs and private sector.

Future

The project is due for closure on 31 March 2001. The AfDB vide letter dated 8 November 2000 has informed MOAC that the loan account will close on 30 September 2001; thereafter only eligible expenditure properly entered into before this date and winding down activities will be settled from project funds. The last date of withdrawal from the loan account will be 31 December 2002. This means that the project can continue to implement its activities beyond another 12 months.

Project Co-ordination: It is important that the PCC system is revived headed by the PS and attended by all the Department Heads, with the PC acting as secretary. Such a forum would pass on the ownership of the project to the heads of the department. These meetings could be held once each three months. In addition, the Ministry has a system of fortnightly meeting of department heads. It is recommended that the PC also be invited to attend these monthly meetings to discuss the project issues on a regular basis.

Project management capability: This project intended to use line departments. Hence the PCU had a small staff compliment. However, there is a need to strengthen its capacity by employing a community development officer from the NGO/private sector with experience in working with the NGOs. It is also recommended that the PCU officially request IFAD to provide an Irrigation Specialist and an Engineer to join future supervision missions for the project. The project accounting system will have to be improved by employing a local short-term Financial Management Specialist. Two project vehicles be allowed special registration so as to allow mobility for the staff to work during weekends/holidays and after office hours. The PS of MOAC may write to Ministry of Public Service and Information seeking approval for paying the government approved incentives to the project staff.

Shift from direct implementation to facilitation: The government should move away from direct implementation responsibility and focus on facilitation and supervision. Implementation should be undertaken by private contractors/NGOs.

Support to field level staff: training and motorcycles: It is recommended that the field level staff be provided with training in leadership, communication, use of PRA tools to identify farmers' priorities and to focus on existing groups instead of building new groups. The PCU be permitted to process pending long-term training applications and motorcycles for use by the field level staff of Agriculture and Livestock Department.

Irrigation development and support of NGOs The outstanding payments for the NGOs will have to be settled by forming a committee consisting of the PC, the Project Accountant, the Irrigation Engineer, and a representative from CANGO. It is recommended that the project should now focus on completing and consolidating the small-scale schemes in Nkwene, Mhlangeni, Lesibovu, Kholwane, Mgeza and Ekuthuleni: this would result in an achievement of approximately 90% of the Appraisal Report target for small-scale schemes. The TAG for implementation support should be used to employ a national or regional technical assistant to assist the Head of Section in completing all outstanding groundwork (design, procurement of materials, planning/organising construction or alternatively procuring contractors) over a period of 5-6 months. In the case of Ekuthuleni, the scheme design should be obtained from a reputable irrigation equipment supplier on a turnkey basis including supply and installation, following local competitive bidding; in view of the topography and proposed use, consideration should be given to drip irrigation, rather than a sprinkler system.

The CI use TAG funds to recruit a regional irrigation consultant to participate in the next two supervision missions and provide approvals for the procurements at the time of supervision mission. NGOs be contracted as soon as possible to initiate group development on the above seven schemes. Smallholder Poultry Development: The evidence across IFAD projects is very consistent in showing that first, women headed households tend to be poorer than male headed households and second, their priorities are centred around small animals, particularly scavenging poultry. It is recommended that a team consisting of members from the Livestock Department and the Home Economics Section and the SDSB should visit all small poultry farmers to establish linkages with extension and credit. On the marketing side, the Philani Poultry Co-operative Society Abattoir is a model to follow and the project should attempt to replicate this in as many sites as possible. The small animals can help poor families in starting an asset creation process as clearly demonstrated in the IFAD/Danida sponsored Smallholder Livestock project in Bangladesh. Swaziland should build up human capacity in this area. The first step would be to contact the Network for Poultry Production and Health in Developing Countries (NPPADC) to request its co-operation in planning and conducting a study tour to Bangladesh for three women poultry extension workers and one regional level Home Economics Officer.

Concurrently, the Livestock Department will have to initiate a Poultry Sector Study to understand the strengths and weakness of all the actors involved in poultry production in Swaziland, comparative advantages of smallholder poultry based on commercial inputs vis-à-vis large commercial farms and possibilities to reduce production costs by developing semi-intensive systems. It is recommended that a TA with background in commercial as well as semi-intensive smallholder system be recruited to undertake this study.

Pig Rearing: This sector also suffers from high dependence on commercial feed. It is recommended that Livestock Department of MOAC engage the services of the Animal Production and Health Department, University of Swaziland to conduct a literature review and identify cheap feed resources within the country. Trials in the SNL households in collaboration with farmers and piggery extension officers would have to be set up.

Analysis of the tick-borne disease data, development of a strategy and study tours: The field-based data be analysed in order to come up with a new, probably reduced dipping strategy. It is recommended that TA for this activity be recruited immediately. It is also recommended that the technical staff and some of participating farmers of the sites for the five dipping tanks installed visit neighbouring countries to update themselves on current tick control strategies.

Livestock marketing: It is recommended the Livestock Marketing Unit will have to gather and publish prices of not only cattle, but on all commonly marketed animals. In addition, the sales yards built under the project should be handed over to the farmer organisations.

Information technology: The training courses on the use of computers for the FMD and Livestock Marketing should be initiated without delay. The VFTC at Mpisi should get the four telephone lines it has applied for and be provided with Internet access and the required training to use the equipment.

Crop Development: Intercropping of maize with green manure/cover crops (gm/ccs) will largely control weeds while providing a good increase in the productivity of subsequent crops. It decreases the cost of applying chemical fertilisers and weeding the maize plots. The introduced varieties of cowpeas, groundnuts and sweet potatoes seem to be especially popular among Swazi farmers, and should be spread as far and wide as possible. It is recommended that the Adaptive Research and Extension establish linkages with the existing Zenzele groups and develop a system of a seed revolving fund for quicker distribution of drought resistant crop varieties. It is also recommended that the Adaptive Research conduct field trials involving the farmers in the Zenzele groups to test the gm/ccs.

Credit: A study tour of officers from the Ministry of Finance, the Central Bank, the SDSB and the Co-operative Department may be arranged to visit Palli Karma Sahyak Foundation (PKSF) in Bangladesh and Small Industries Development Bank of India (National Micro Finance Support Programme).

It is also recommended that the PCU establish an Expert Committee consisting of members from MOF, Central Bank, SDSB, MOAC and CANGO to: (i) assess the performance of existing micro-finance institutions, evolve accreditation criteria and develop systems and procedures to start a refinance window for these institutions; and (ii) review the subsidiary loan agreement to address its weaknesses and institute measures to use part of the funds released to SDSB to finance micro-finance institutions.

There has been no effort to develop community institutions in the rain-fed areas. It is recommended that the Home Economics Section identify women groups that have been trained under the project and empower these women to form about 10 savings and credit groups. In addition, a contract for mobilising 10 savings and credit groups may be given to an NGO (ACAT is suggested) so as to ensure that the project at least develops about 20 groups. These groups need to be the focal points for seed multiplication, small stock development and testing green manure cover crops. For this purpose the Home Economics Section and the selected NGO would have to develop the required linkages with research and extension.

Marketing of Horticultural Produce: In the medium and long-term, NAMBOARD will have to completely privatise its vegetable marketing activity. NAMBOARD, using its revenues from the statutory functions, will have to proactively facilitate growth of the private sector in marketing of agricultural produce.

Drought Early Warning and Contingency Planning: The current approach to developing forecast on drought focuses on assessing supply of food and not on the capability of the household to acquire food. In order to fine tune the data collection and to develop required drought mitigation strategies, it is suggested that a post of nutritionist be created within the unit and the Swaziland National Nutritional Council be included as a member of the Early Warning Technical Advisory Committee.

CCU: The CCU is on a path of restructuring with the help of Enterprise Development Fund. The CCU is interested in moving from trading to production of fertiliser and animal feed. It needs to make a full assessment of: (i) its own institutional strengths and weaknesses to undertake such businesses; (ii) its comparative advantage vis-à-vis private sector; and (iii) techno-economic feasibility of the project, before venturing into diverse commercial activities that are not its core strengths.

MOAC Reorganisation: It is recommended that the MOAC revisit the proposals made in the reorganisation study to look into the service delivery strategy needed to reach, and provide facilitation and empower the resource poor households on the SNL.

Monitoring and Evaluation: The project should immediately initiate actions to locate a consultancy company to undertake an impact assessment study. The PC and the Monitoring Officer should make frequent field visits to ascertain the progress in implementation of the project activities and get feedback from the primary stakeholders for taking corrective actions in implementation.

Framework to assist resource poor households: Four factors need attention while developing future rural development strategies for Swaziland. First, a study conducted by the International Food Policy Research Institute (IFPRI) indicates that any meaningful poverty reduction strategy needs to address women's education, health, status and food availability. Experiences in Swaziland indicate that community institutions play an important role in building education facilities. There is increasing control of resources by women in case of small stock development and vegetable production interventions. Savings and credit group development have taken off with the involvement of NGOs, Co-operative Department and Home Economics Section. The interest of farmers is shifting towards low cost gravity fed systems that are easy to maintain.

Second, the land tenure system favours those with privileged access to human and financial resources. At the same time, it restricts investment in land by those with limited resources. Resource poor farmers may fear loss of land and also investment made owing to inability to cultivate the land as a result of labour shortage. Or they may be subject to land reallocation by the Traditional Chiefs. These in-built fears drive the rural communities on SNL to invest in cattle, which is a low risk and mobile investment but with low economic return. Literature published as early as 1991 indicates that there are strains in the social framework in SNL areas. These will become more evident as the number of landless increases along with the pressures to privatise communal lands to satisfy the demand by commercial companies and entrepreneurial farmers. Resolution of issues related to security of tenure becomes central to any meaningful investment strategy in SNL.

Third, the AIDS epidemic has its debilitating human and social impact; it also affects the scope for productive investment and production in the rural areas. The average life expectancy in Swaziland is now as low as 38 years. These factors do not bode well for any investment that requires a high labour input. The strategy for any future rural development needs to focus on investments, where the priorities are determined by the communities and not investment strategies designed and delivered from outside.

Fourth, the projects so far implemented in Swaziland suffer from top-down implementation mode with little participation of the communities. Inadequate participation of the communities is masked when governments provide high subsidies. To achieve sustainability, participation of the communities needs to be enhanced.

Interventions required to improve livelihood systems of the resource poor households vary from community to community. Projects with a set of activities that are inflexible will not be able to respond to the requirements and priorities of the community. A system of flexibility will have to be introduced in the project design. Impact and sustainability are enhanced when the ability of the communities to identify constraints, and to prioritise and implement the interventions are strengthened. This will have to be coupled with adequate control by the community over the financial resources for implementing activities at the community level. Such devolution of financial powers would empower the communities. It increases their ownership of interventions.

Rehabilitation of Old Irrigation Schemes: The data available suggest that some 300 ha are cultivated by existing small-scale irrigation schemes in Swaziland. If the sample visited by the mission is representative, most of this hectarage is under-performing. It would appear logical to concentrate first on improving the performance of these schemes – through rehabilitation and group development – before embarking on a specific project for expansion.