Integrated Rural Development Project (1991) - IOE
Integrated Rural Development Project (1991)
Mid-term evaluation
This report distils and synthesizes the principal findings and recommendations of partial evaluations by IFAD consultants of the Integrated Rural Development Project (IRDP), co-funded by IFAD and Government of the Commonwealth of Dominica (GOCD). It also incorporates a mid-term partial evaluation and baseline survey conducted by OAS in 1990.
Project background and objectives
The IRDP was formulated in concert with the GOCD's overall objectives of agricultural diversification, income growth, reduced unemployment, increased exports, and overall economic growth. The project aims to achieve its objectives by supporting the settlement of former estate workers with the provision of credit, inputs, technical services, and infrastructural improvements.
IFAD funded the IRDP as a follow up to its previous support to the agricultural sector through the Agricultural Credit for Food Production and Related Services Project (hereafter IFAD-1).
The IFAD loan to the Commonwealth of Dominica for the IRDP has a 20 year term, a grace period of 5 years with an interest rate of 4% per annum. The total planned project cost is US$ 2,996 million; of which IFAD is financing US$ 1,502 million, or 50%, with the remaining 50% financed by the GOCD.
The Caribbean Development Bank (CDB) is the cooperating institution, and the Ministry of Agriculture (MOA) is the main executing agency. The Dominica Agricultural and Industrial Development Bank (DAIDB) is responsible for administering the credit fund for small farmer lending. The project commenced officially in November 1986, and has a completion date of December 1991, although it is likely to be extended.
The project is targeted to approximately 640 settlers/farm households on the former Geneva and Castle Bruce estates; most of them were landless or near landless, and were plantation labourers. Their per capita income was estimated at US$ 180 per year, as compared to the gross national per capita income of US$ 970 (1983), and below the absolute rural poverty per capita income level of US$ 185. 90% of the target group cultivated less than 2 acres of land. Credit was targeted to roughly 600 project participants outside the estates, mostly artisanal fishermen, small agro-processors, small farmers with less than 10 acres of cultivable land, and rural people with gross family incomes of less than US$ 2,000 per year. 28% of the proposed settlers were women and heads of households, in a country where women make up 35% of the labour force. Other participants included fish vendors, hucksters, landless livestock rearers and cottage industry workers.
IRDP components and implementation
Development of the estate settlement areas, through developing agricultural stations, with input supplies, training centres, farm settlement access roads, marketing facilities, and slaughter houses Physical improvements have essentially been completed on both estates. Farm lots were assigned and approved by 1986 in Geneva, and by 1988 in Castle Bruce. Approximately 371 farmers with lots ranging in size from 0.4 to 9.7 acres are settled in Geneva, and 235 farmers with lots from 0.1 to 14.6 acres are settled in Castle Bruce. Land was allocated on the basis of existing holdings, and on the recommendations of a committee made up of small farmers and government officials. 26% of the lots in Geneva and 20% in Castle Bruce were allocated to women.
The agricultural research stations have been upgraded on both settlements, and the plant propagation unit has been upgraded to introduce new crops such as hot peppers, passion fruit, papayas, mangos and avocados. Farm settlement roads have been completed on Geneva, providing all farms with ready access to roads. The access roads in Castle Bruce are almost completed, leaving all farms except for 23 across the river within 100 meter access to the roads. Implementation has been delayed by the conflicting demands for equipment and labour from the institution responsible for road construction and maintenance. Road maintenance, envisioned to be conducted by project beneficiaries, is deficient. Marketing facilities are operating in Geneva, and are under construction in Castle Bruce, with expected completion by the end of 1991. The proposed slaughterhouses have been given low priority as farmers have much greater interest in crop production.
The credit component consists of approximately US$ 1.2 million, representing 73% of IFAD's contribution to the Project, and 37% of total project costs. Credit was expected to reach approximately 1,100
borrowers; 500 settlers, and 600 non-settlers. 43% of these funds were targeted to estate settlers. Additionally, US$ 218,518 (20% of credit funds) are to be lent from DAIDB to the Dominica Cooperative Credit Union League (DCGUL) for on-lending through its 23 credit union affiliates. This agreement has only recently been effectuated ~May ~, 1991), after four years of negotiations over the interest rate to be charged the final borrower by the credit union. The Appraisal report included the participation of credit unions in credit delivery under IRDP, as in IFAD-1. However, the credit unions were requested to lend to the final borrower at 12% annual interest rates, which they felt would put them in unfair competition with DAIDB who lends IRDP funds at an annual interest rate of 8.5%. Under the current agreement, the credit unions will be lending at 9.5 - 10% annual interest rates. The credit unions' refusal to participate in IFAD-1 was due to the lack of an insufficient spread to cover costs.
DAIDB currently administers 8 lines of credit for agriculture, each with its own specific target groups, lending criteria, and interest rates. IFAD funds accounted for 49% of the bank's agricultural credit portfolio in 1983, and for 25% in 1990. IFAD accounted for 68% of the number and 28% of the value of agricultural loans from July 1989 to June 1990. The average loan size was US$ 1,185. The main loan activity was for bananas (80% of the number and value), root crops (6% of the number and 5.5% of the value), and livestock (3.5% of the number and value).
Lending rates charged by commercial banks and DAIDB have been positive in real terms over the IRDP life to date.
DAIDB measures up adequately to standard financial indicators. Asset and capital formation have both been increasing. DAIDB has generated surplus income in all years from 1984 to 1990, except for 1985 and 1987, although a large proportion is from non-lending activities.
The credit component encouraged DAIDB to lend to small farmers. However, although DAIDB has been implementing agricultural credit programs for 10 years, it has not acquired the characteristics necessary for lending to small farmers; rural presence either through branches, vans, or extension staff; character and feasibility based lending; group lending; etc.
IRDP loan ceilings are viewed as a constraint by DAIDB which claims that borrowers usually seek larger loans. At times, therefore, a borrower may receive 2 loans from 2 different credit schemes to make up the amount sought.
Farmers with larger land holdings received a greater proportion of loans. Whereas 62% of the farmers have under 2 acres of land, the remaining 38% of farmers with 4 or more acres of land received 60% of the total number of loans. This concentration of credit is due to the fact that those with larger holdings can meet DAIDB's collateral requirements more so than small farmers, and are in an overall better position to qualify for credit from the bank.
Women headed households which account for 28% of the farms in the settlement areas received ll% of total loans disbursed. Women's access to credit is constrained by a lack of awareness of the IRDP credit program, and lack of assistance in accessing it.
It is generally felt that the lack of credit union participation from project inception has excluded women and smaller farmers from being benefited, particularly since 60% of the credit unions' membership base is made up of women. In addition, the credit unions have several comparative advantages in terms of reaching the target group; flexibility, more convenient operating hours, island-wide penetration, a membership base from over 50% of the population, less stringent collateral requirements, diversified range of services, and a community orientation. However, they are also over-leveraged, and lack significant experience in agricultural or targeted lending. Given the credit union's recent agreement to participate in the project, which had been pending on the resolution of lowering the interest rate charged to borrowers, it is expected that larger numbers of the target group will be reached.
Arrears as a percent of principal outstanding has increased from 11% in 1985 and 1986 to 35% in 1989 and 45% in 1990. Again, in the period 1985-1990, in 4 of the 6 years over 70% of the delinquent loans were in arrears of over 12 months, indicating the lesser probability of recovering them.
The increase in arrears can be directly attributed to changes in staffing patterns and arrangements from IFAD-1 to present. Under IFAD-1, PIU staff were housed in DAIDB, and directly appraised, approved and disbursed loans, as well as provided continuous and close supervision, resulting in rapid approvals and disbursements, and high repayments. The PIU has since been moved to the MOA. Although DAIDB has replaced IRDP staff, these replacements were not individuals with the same sensitivity and knowledge of small farmer credit needs. The departure of the credit RDO in 1988 has left a void unfilled by either IRDP or DAIDB.
The increased borrower transaction costs, the lack of supervision, and communication, and the separation of responsibilities for loan disbursement and loan collection in 2 divisions in DAIDB exacerbates the arrears problem.
DAIDB assumes a higher transactions cost than its other lines in processing the small loans stipulated by IRDP. It is not compensated by a higher interest rate it can charge on these loans, which may be an automatic disincentive for promoting the IRDP credit line.
DAIDB's low profitability of lending operations can be attributed to several explanations: high operating costs and inadequate spreads to cover them, high arrears, low lending volumes, and under-utilization of staff resources.
The subsidy per borrower, given average loan sizes of US$ 1,185, is approximately US$ 148, which is equal to 9% of per capita income in 1989. This transfer of resources has affected 575, or 8% of the farmers in Dominica.
Improvement of fish landing sites
At project appraisal, this component involved the improvement of 18 fish landing sites, including eight sites where work would complement improvements already introduced under IFAD-1.
This component has since been absorbed by a US$ 2.9 million WFP and GOCD funded fisheries development project, administered by the Fisheries Division of MOA, with no formal links with IRDP.
Assistance to promote an agro-processing sector
This component has been relatively inactive, in large part due to personnel shortages within the agro-processing laboratory. It is managed separately from the IRDP. This component has not led to the development of new products for local processing or for export.
Project organization and management
IRDP management has benefitted from the continuity of competent staff from IFAD-1. At project initiation, the project manager, two rural development officers (RDOs) for Geneva, and the project secretary were appointed. This group made up the same group that managed IFAD-1.
However, IRDP has been under-staffed over its course. At inception, staffing plans included a project manager, and 4 RDOs: 2 marketing specialists and 2 credit/extension specialists, 1 of each per estate, and a project secretary. For instance, the Geneva credit/extension RDO has not been replaced since his departure in 1988. A marketing assistant, not an IRDP core staff member, has been assisting in the development and marketing of ginger at Geneva. The full time marketing officer is on a training program in the U.K. One RDO with livestock expertise was assigned to Castle Bruce in late 1989. It is planned that the marketing RDO for Castle Bruce will be hired once the marketing facilities are completed. It should be noted that the project manager is also responsible for the Melville Hall Estate Project, funded by the CDB which is modelled on the IRDP.
This component is financed by the Organization of American States (OAS), and is the responsibility of the Economic Development Unit (EDU) within the Ministry of Planning. To date, a baseline survey and a mid-term survey of the Geneva settlement area has been conducted. The M&E officer from EDU has been sent on several training programs.
The M&E component has not served its role as an important management tool. While the M&E system has complied with project plans, the design of this function in the project itself was inadequate. The M&E function has not fully assisted project management. It has acted instead as a recipient of consolidated reports prepared by the project manager, instead of the collector and provider of this information to the project manager.
Key issues and recommendations
Targeting issues: appropriate criteria and credit policies: Through applying a low loan ceiling, IFAD appears to have 'democratized' the credit operations of DAIDB, where loans were made with IFAD funds to small borrowers who would have been excluded by the criteria of the other lines. However, although loan ceiling have directed credit to smaller farmers, low loan ceilings in of themselves do not limit loans to the smallest borrower. Borrowers often receive 2 or more loans from different credit schemes to make up the amount requested. Limiting all loans to those with under 7 acres of land would be a more effective targeting tool than loan ceilings, provided that actual land holdings could be estimated, and differences in land quality be taken into consideration.
DAIDB should consider establishing positive market interest rates. It is more expensive to process small loans than large loans, and being required to charge a lower interest rate for costlier lending operations acts as a disincentive for the bank to promote the IRDP credit line, given its relatively low ceilings on loan size. DAIDB interest rates under IRDP could be revised upwards to 10%, comparable to the rates on the other DAIDB lines, and to the credit unions' rate to be charged under IRDP, to make the IRDP line more attractive for DAIDB to promote. IFAD funds are competing with higher interest earning credit lines within DAIDB. The issue of credit for the rural poor is often a question of timely access to rather than the cost of credit.
Appropriate implementing institution: Given that DAIDB has failed to adopt the characteristics necessary for small farmer lending, the question arises as to whether DAIDB is the appropriate primary lending channel to reach IFAD's target population. Perhaps the total remaining undisbursed loan funds could be channelled through the credit unions, BASED on their performance in managing and monitoring the first tranche of funds under the project. In addition, alternative sub-contracting arrangements discussed under New Directions could be explored.
Lack of savings mobilization: DAIDB currently does not mobilize savings, in large part since it believes not to be ready to manage this function well. DAIDB must learn to operate on the basis of mobilized
savings and earned income, to progressively reduce dependency on external donor capital infusions.
Increase gender equity: IRDP has not been very successful in targeting women directly, although un-single women benefit indirectly through improved household income. Although 26% (96) of the lots were allocated to women on the Geneva settlement, and 2% (8) were jointly held, only 4 women have received credit. Although there is no overt discrimination on the part of project implementors, there has been no overt steps taken to increase women's direct participation. There is a need to include a 'compensatory bias' so that women are encouraged to access project resources.
Improve borrowers' credit access: An important factor determining farmer access to credit is the current lack of credit supervision to farmers. No or minimal extension services are provided to farmers by
DAIDB, which, according to persons interviewed, has lost its orientation and mandate to service the small farmer. There are no credit experts on the current IRDP team to provide this supervision. The IRDP project needs to address this critical 'missing link' to ensure project impact. This can be addressed by hiring staff with credit expertise, and/or training current staff in credit management and loan feasibility analysis. Ideally, DAIDB should be the institution providing this expertise, given its mandate.
Project contribution to agricultural diversification: IRDP is contributing to agricultural diversification, particularly through introducing ginger, as well as through assistance in passion fruit and hot pepper production. Although agricultural diversification from bananas was stated as an objective at the time of appraisal, it was not specific about what crops to diversify to. IRDP has been successful in identifying and promoting crops suitable to Dominica with export potential.
The diversification process has been slower than expected due to the rise in banana prices over this period, and farmers' obvious response to market forces. Consequently, almost 80% of all loans approved from IRDP sources in 1988-1990 were for bananas. The introduction of new agricultural crops is constrained by the efficient infrastructure provided farmers for bananas by the Dominica Banana Marketing Corporation (DBMC).
Project management: Staff continuity of the project manager, and of the credit RDO initially, and the competence of the project manager have enhanced project management. However, the project has been understaffed over its course. Staff levels were not maintained as originally planned, and key staff departures were not replaced by persons of similar technical capabilities.
Participatory approach: IRDP has benefitted from incorporating a participatory approach in the preliminary stages. The project manager was brought on board during the design stage, and was key in contributing to the project formulation, drawing on hi~ experience of having managed IFAD-1.
The participatory means through which land distribution and settlement was decided upon gave project beneficiaries a sense of ownership of the project.
Institutional development and sustainability: The IRDP has made some important strides towards institutional development and sustainability. IFAD funds have contributed to sustaining DAIDB. The bank measures up adequately to standard financial indicators. Asset and capital formation have both been increasing. DAIDB has generated surplus income in all years from 1984 to 1990, except for 1985, and 1987.
IRDP has assisted in the formation of Farmer Advisory Groups, resulting in the development of a strong group at Castle Bruce. At Geneva estate, ginger producers have organized a Ginger Producers' Group, which currently has 23 members, and is seeking cooperative status. It is expected that this group will take over all ginger marketing activities after project end. Increased production has also encouraged a private exporter of fruits and vegetables. Farmer ownership of the project is already occurring, through the progressive phase-out of earlier subsidies provided for inputs, transport, and marketing.
IRDP: replicability and demonstration effects: IRDP has served as the pilot test, and is currently the model upon which other rural development projects are being designed and implemented in Dominica.
Impacts: The IRDP is making a significant contribution to national agricultural output. The Geneva settlement area alone contributes a significant proportion of the national production for ginger (27%), dasheen (5%) and avocados. Ginger production increased from 15.5 tons in 1988/89 to 57.9 tons in 1989/1990.
Fifty one percent of farmers surveyed felt that their standard of living was higher since joining the project, and 39% felt that it remained the same. 45% of women respondents felt their living standards had improved, compared with 53% of men. 38% felt that they enjoy a better standard of living than non-project farmers. In fact, farmers doubled their stocks of consumer goods since the baseline survey conducted in 1987. Roughly 70% of the settlement population have doubled their income since appraisal, from approximately US$ 93-185 per month to US$ 220-370 per month. 98% of fishermen surveyed felt that their standard of living improved since project initiation.
The land settlement was a significant channel for a more equitable land distribution pattern. Prior to settlement only 30% of the population on the estates owned land, while the IRDP land divesture made all settlers on the estates land owners.
The project has stimulated both backward and forward linkages. Increased production has led to increased demand for inputs and tools, provided by other farmers and private traders. In terms of forward linkages, increased output has increased the activity of the hucksters in marketing produce. Given the increased number of farm household members working off the farm since the 1987 baseline survey, increased income, or access to land has permitted farm members to engage in other income generating activities, and in turn hire wage labour to work the farms. Given that labour shortages are repeatedly cited as a major constraint by farmers, it can be assumed that the project has had a positive effect in decreasing unemployment in and around the settlement areas.
Appropriateness of technical packages: The technical packages are considered appropriate by those farmers who received them; primarily men. One hundred ninety seven farmers, 29 of which were women, had adopted them as of mid-1990. Womens' adoption of these packages is constrained by labour shortages, high labour costs, and the length of time (two months) for land preparation for the packages. Proper land preparation and larger plots of land are the critical elements in adopting these packages for new crops. In addition, women's smaller plots make them more risk-averse to allocating scarse resources for uncertain returns. These gender-specific constraints call for devoting specific efforts to women, including the design of a technical package tailored to their resource availabilities.
IRDP has made use of existing technologies in training and extension. It has applied training videos developed by FAO's 'Videos for Rural Development' project in its extension services, and often worked directly with the FAO project staff in identifying needs and providing technical expertise for the production of videos for extension. This has reportedly been highly successful, leading to noticeable changes in farming systems based on the videos, particularly in the case of ginger.
Physical contributions: The project has successfully met planned physical outputs, making adaptations of the original plan as necessary and practical during project implementation. Progress has been slower than planned, due to limited availability of equipment from the Ministry of Communications and Public Works (MCPW). Although most of the planned physical infrastructure is either completed or underway, facilities and roads maintenance is a problem. Although envisioned that project beneficiaries would be responsible for maintenance, in reality this has not yet occurred.
Marketing of agricultural produce: IRDP has been responsive to the marketing constraints faced by farmers. The project has encouraged the development of a private exporter of fruits and vegetables, as well as the development of the Ginger Producers' group, which is expected to take over ginger marketing activities. IRDP has also been successful in developing an agreement with CATCO to export ginger to the U.K., thus identifying a marketing channel for sustained production.
The marketing problems have implications for the crop diversification program. Unless there is a marketing channel like the DBMC or that created for ginger through CATCO, or greater links are developed between the local market and other regional islands for market expansion, the goals of agricultural diversification can not be reached.
Increase support in agricultural marketing: The marketing of increased agricultural output from IRDP interventions has not been dealt with sufficiently. Discussions with the Dominica Export Import Agency
(DEXIA) highlighted the difficulty in responding to demand for produce, given the basic lack of infrastructure and links set-up from the producer to the distribution systems in other regional islands for Dominican produce. There is a dearth of cold storage equipped transport and storage facilities. A project revision or follow-on should incorporate a component which would assist DEXIA in setting up the links to actually deliver the products it promotes. Possibilities for assisting the National Huckster's Association should be explored, given their role in exporting approximately 70% of the non-banana produce to other islands. These would include adequate grading and packaging centres, as well as cold storage and storage facilities. The possibility of DBMC extending its current services for the banana sector to other crops is being explored by DBMC and the MOA. This appears to be a viable option. The DBMC is well-perceived by farmers, and has an established and efficient infrastructure island-wide.
Include microentrepreneurs in target group: Given evidence of increased off-farm employment, IRDP should seek to reach other participants of the informal or micro and small enterprise (MSE) sector. This group was identified as part of the target group at the time of appraisal. While agricultural diversification has been given priority in the overall development strategy, assisting the MSE sector, to promote off-farm employment and income growth will lead to a more diversified economic base. Given increased information on the structure of rural poverty, an extended approach that caters to the multiple income sources of the poor should be better incorporated and addressed in IRDP.
The expanded target group can be reached through: i) Expanding the credit eligibility criteria currently used by DAIDB to include loans for small enterprise activities; ii) Providing support through local institutions such as the National Development Foundation of Dominica (NDFD), the credit unions, and others already assisting MSEs; iii) Providing training in food-processing to expand the agro-processing sector; iv). Not targeting loan use. Loan use should be left to the discretion of borrowers, with project emphasis on loan repayment rates, to serve as a proxy of productive loan use; v). Publicizing the availability of credit for an expanded range of activities.
Sub-contract service arrangements for project implementation: Given the NDFD's success in lending to MSEs, as well as their proven methodology of linked credit and technical assistance, a possible approach would be to 'sub-contract' the NDFD to manage DAIDB's IRDP portfolio, and thus provide the assistance in promoting services, qualifying applicants, and loan supervision and follow-up. The NDFD itself is too small to manage the entire loan portfolio, but they could provide the expertise necessary for successful implementation of the IRDP credit component. They would in effect serve as the bridge between the target group and DAIDB; a capability currently not available within IRDP ~r DAIDB.
According to the NDFD credit officer (the former IRDP credit RDO), farmers from the Geneva estate are contacting him to receive loans from the NDFD, since they feel they have no access to DAIDB anymore, given the current void in the services initially provided under the project. The explanation offered by DAIDB for low IRDP credit disbursements is that credit demand reached an early saturation point, particularly in the Geneva settlement, and had been over-estimated. This contradiction could be due to DAIDB's concentration on the not-so-poor farmers.
Gender concerns: Increasing Women's Participation: Although no overt discrimination was cited as a factor limiting women's participation, no overt steps were taken to increase it. What is being proposed is preferential treatment of women's constraints. Certain mechanisms for better integrating women in IRDP are: i). Support off-farm and microenterprise activities that up to 40% of women reported to be engaged in; ii). Provide training for project personnel to sensitize them to the productive roles of women, as soon as project execution begins. A lack of gender awareness, knowledge or understanding of women-specific constraints on the part of the PIU limited women's participation; iii). Train extensionists to work with women producers, or incorporate female staff in extension teams. Women 'barefoot extension' workers should be trained among rural women themselves, as a step in a long-term strategy to develop local capacity; iv). Use gender-disaggregated data, in order to track the differential performance of women; v). Promote the project to women through women's organizations, and extension workers, maternal clinics, etc. Lack of information about the existence of the DAIDB credit source was one of the major reasons why so few women approached the bank for loans. Bank intimidation and reticence by women should be overcome by several 'outreach' activities; such as organizing special meetings in the settlement areas at hours convenient to women, such as late afternoon; vi). Minimize collateral based lending, and rely instead on character, prior repayment records, income earning potential, use of an internal guarantee fund, and incentives of future access to credit; vii). Target training programs appropriately to user needs. Schedule sessions at - times and places convenient to women.
Project management: The project should acquire credit expertise through hiring at least one credit expert. The Project Technical Advisory Committee should include representation from the agro-processing laboratory and from DEXIA as originally conceived, given their importance to overall project integration.
Credit targeting and administration: Reflow funds should be added to the IRDP revolving loan fund and lent at prevailing IRDP interest rates.
An appropriate management information system need~ to be established in DAIDB, to help improve the types and quality of information flows.
Loan approval, and collection should lie within the functions of the same department, to ensure accountability. An incentive system which rewards credit officers to make sound lends could be used. Such a system has been used with positive results by an NGO in the Dominican Republic, which makes loans to microentrepreneurs.
Arrears control mechanisms need to be installed. These may be through: hiring/training staff within DAIDB in small farmer lending; loan feasibility analysis-based lending; appropriate loan supervision; making group loans; making more frequent shorter-term loans; and tying current loan repayment to the possibility of more credit. It needs to be assessed whether the increased costs of these measures will exceed the cost of arrears.
DAIDB needs to have a larger rural presence in order to service its client base of the IRDP target group. This does not necessarily imply establishing branches, but could be done through the use of rural credit officers, mobile banks, part-time location at the banana boxing plants on the settlements during banana collections and payment disbursement time.
For most small farmers, reliable access to small and short-term loans is more valuable than having large and long-term loans. These should be tested in the case of Dominica, where current loan terms average two years.
The credit unions should be closely monitored in their administration of the credit funds, given their very recent participation in the project, and their relatively limited experience in agriculture or targeted lending.
DAIDB and the credit unions should be allowed greater discretion in establishing positive market rates of interest, after assuring certain efficiency levels.
If DAIDB does not assume the characteristics of a lending vehicle for small farmers, based on the recommendations of this report, it must be decided whether DAIDB should continue in the project as the primary lending institution. Perhaps the DCCUL could assume this role, based on the quality of its management of the initial tranche of funds.
Loan use should not be targeted. It should be left to the discretion of borrower~, with project emphasis on loan repayment rates, to serve as a proxy of productive loan use. The multiple income sources and fungibility of funds within a farm household often diffuses the impact of targeted loan use.
The use of credit for ginger, and the development of ginger as an export crop, an unplanned yet very positive outcome under IRDP, demonstrates the benefits of unrestricting the actual application of credit. On the other hand, the highest arrears were for loans for food crops for consumption, which were targeted in the Appraisal report. It is a merit of the project and IFAD that the project's flexible approach incorporated activities of great benefit to the target group, which were not initially contemplated. In addition, the credit unions are not experienced in targeted lending, and requiring them to assume the burden of tracking actual loan use would deviate them from more important lending and collecting operations.
Technical assistance to increase farmers', and in particular womens access to credit should be provided both to DAIDB and to the IRDP project management team. The appropriateness of the current criteria used for targeted lending, such as the loan ceilings, need to be tested and possibly revised if necessary.
Further assistance to the sector should promote savings mobilization, and should assist the financial implementing institutions with improving savings opportunities for the rural poor. Further injections of external donor credit prolong the delay in savings mobilization.
Measure project impact: A future project design should include the design and provision for the systematic collection of data that will permit an assessment of project impact. It is costly and near impossible to retroactively reconstruct the dynamic flow of funds and their use and impact on the farm household. Impact with respect to assets, income, education, savings, nutrition, and other variables should be collected and compared with the pre-project status, even though a project can only partially be credited with any positive or negative differential.
IFAD funds should be provided for M&E. IFAD has earned a 'pioneer' reputation in this function, and should draw on its comparative advantage in directly supervising and supporting this function. Again, donor agencies are often concerned with certain impact issues, related to the appropriate allocation of resources for development programs. There is a cost associated with this function, and IFAD should fund this component, given its interest in the results and outcome of its projects for further programming lessons.
Improve project information flows: CDB supervision reports have not been submitted to IRDP project management, and there were extensive delays in the project manager receiving the various evaluation reports and WID case study from IFAD. Delays in information sharing defeat the very purpose for which these reviews are often conducted: improving project performance.
It is expected that incorporating the recommendations outlined above will further improve the IRDP's performance in delivering services and generating benefits for the rural poor of Dominica.