The Government of the United Republic of Tanzania has requested IFAD to provide continued support for rural finance and agricultural marketing in a second phase programme that builds on and merges the successful support interventions of two IFAD-supported programmes. One of the programmes closed in 2010, the other one is closing in 2011: (i) the Rural Financial Services Programme (RFSP); and (ii) the Agricultural Marketing Systems Development Programme (AMSDP). IFAD procedures at the time of this evaluation required that an independent evaluation be undertaken for consideration of a second phase of a programme or project and the Independent Office of Evaluation of IFAD (IOE) has the responsibility of implementing such interim evaluations. While interim evaluations are normally limited to one programme or project, IOE responded to the request for a merger and, in this case, covered two programmes in one interim evaluation.

The main evaluation objectives were: (i) to assess the performance and impact of the two programmes; and (ii) to generate findings and recommendations useful for ongoing and future programmes, and specifically for the design of the follow-on programme which has the working title: "Marketing Infrastructure, Value Addition and Rural Finance Support Programme" (MIVARF). With respect to the latter, the evaluation attempted to identify gaps and challenges in the programmes but most importantly, what worked well and how it may be scaled up.

After preparatory work, a field mission of four contracted consultants was implemented during 29th April – 14th May 2010. The mission visited selected sites in all programme zones in the vast geographical area covered by the two programmes and presented and discussed an Aide Memoire with IFAD and the Prime Minister's Office (PMO) which, on behalf of the Government, is responsible for the coordination and implementation of the two programmes. 

Description of the programmes. Both programmes were designed and approved under the 1998 country strategic opportunities paper (COSOP). At the time of design, Tanzania was struggling to fill the vacuum left after the collapse in the late 1980s of the state controlled/managed cooperative and banking sectors. The collapse had reduced farmers' access to markets and financial services. During the liberalisation/privatisation in the 1990s of the financial sector and agricultural markets, the private sector had not responded as quickly and forcefully in the rural areas as one could have hoped for. In order to fill the gaps government and development partners launched a series of initiatives including the IFAD-supported RFSP and AMSDP. Improving farmers' access to financial services and markets has over the last decades been a top government priority and the Government of Tanzania continues to consider these two areas as being crucial in its strategy to boost the relatively modest agricultural sector growth (only marginally above the increase in population).

Before RFSP and AMSDP, IFAD had provided support (at micro level), with limited success, for rural finance and agricultural marketing in the form of components and sub-components in more diverse programmes. For example, rural finance or agricultural credit had been supported in six previous programmes with disappointing results. The design of RFSP and AMSDP represented IFAD's first attempt in Tanzania to apply a more holistic programmatic approach (and in the case of RFSP a financial-sector approach) addressing issues and constraints at the macro, meso and micro levels. At micro level, both programmes had very ambitious targets for geographical coverage, each targeting areas the size of Italy - in the case of RFSP covering the Northern, Central and South Western parts of the country while AMSDP was targeting the Southern and Northern marketing zones.

RFSP. In December 2000, the Executive Board approved a loan of US$16.3 million for RFSP which was designed with a nine-year implementation period using the Flexible Lending Mechanism, providing for three phases, where the transition from one phase to the next was conditioned upon fulfilment of defined "triggers". Swiss cofinancing of US$2.2 million was obtained while scheduled cofinancing from the OPEC Fund did not materialize. The completion date was December 2010 and the closing date was June 2011.

RFSP was designed to improve access to financial services of the rural poor and to develop a rural financial architecture with roots at the village and ward level in the form of semi-formal Savings and Credit Cooperative Societies, Ltd. (SACCOS), ¹ and more informal village banks, Savings and Credit Associations (SACAs) and other informal groups, referred to as grass roots microfinance institutions (gMFIs). The design had five components of which one was for programme coordination and management while another separate component was for monitoring and evaluation.  The other three components - though not systematically structured around the macro, meso and micro levels – included: (i) at macro level, support for improving the institutional, policy and legal framework; (ii) at meso level support for strengthening of apex organizations and linking gMFIs to banks and other formal sector financial institutions including capacity development support for banks and NGOs; and (iii) at micro level, support for developing the human and physical capacity of gMFIs and informal groups as well as for micro entrepreneur development, including vocational training.

In addition, the RFSP design included a microfinance facility which was to provide: (i) credit lines to financial intermediaries for on-lending to the more mature gMFIs; (ii) lease funds to existing leasing companies and NGOs to allow them to extend their operations in the programme area; (iii) equity funds for community banks and the Tanzania Association of Microfinance Institutions for expanding their lending to mature gMFIs; and (iv) an insurance fund (credit guarantee) to cover part of the lending risks of the microfinance institutions (MFIs).

AMSDP. In December 2001, the Executive Board approved a loan of US$16.3 million for the Agricultural Marketing Systems Development Programme. The major financing was provided by the African Development Fund (ADF) with about US$14.5 million for marketing infrastructure (feeder roads, market places and storage). In addition, Irish Aid provided US$1.1 million while another scheduled cofinancing of US$4.5 million was not obtained. The programme was completed in December 2009 and was closed in June 2010.

AMSDP was designed to improve the structure and performance of the agricultural marketing and pricing systems in order to improve food security of the rural poor, raise their incomes and diversify their production. IFAD's funding was allocated for programme management and coordination and three components: (i) policy development aimed at improving the efficiency of the marketing system; (ii) capacity development of producer groups, grass roots organizations and small and medium scale traders and processors with the aim of developing their efficient participation in markets and enhance their bargaining power in relation to larger organized market participants; and (iii) financial market services which included a guarantee fund for a commercial bank, introduction of a warehouse receipt system, lending for processing activities and technical assistance to support the financial activities.

Results. Both programmes achieved their quantitative outreach targets. RFSP had by December 2009 assisted 276 gMFIs against a design target of 275 but total membership of the assisted gMFIs was only 116,000 against a target of 190,000. AMSDP had by programme closure assisted 1,202 groups with 46,500 members against a target of 1,000 groups with 25,000 members. However, while surpassing the targets for producer groups, targets for groups of traders and processors, in particular medium-sized agribusinesses were not achieved. In the ADF-funded market infrastructure component AMSDP achieved most of its quantitative design targets except for market places for which the design had significantly underestimated the costs.

RFSP. While achieving the quantitative outreach targets, it was more difficult to bring the gMFIs to the targeted level of institutional development, partly because the gMFIs were significantly weaker than assumed at design. The targets at meso level were only partly achieved; while most gMFIs were linked to formal financial institutions, the support for the apex organizations did not achieve the expected results. The support at macro level was relatively modest and achieved some minor results. Finally, the Microfinance Facility was not implemented.

Partly due to the large geographical coverage, the achieved outreach had very high transaction costs. With a programme area the size of Italy, there is a long distance in-between the 276 gMFIs. While the budget at design allocated substantial amounts for programme coordination (17 per cent of the total) and monitoring and evaluation (8 per cent), these budgets were significantly surpassed. By December 2009, 29 per cent of total expenditure had been spent on programme coordination while 12 per cent on monitoring and evaluation (M&E), thus in total 41 per cent of programme expenditure. This over-expenditure was made possible by exchange rate gains and by a significant under-expenditure (78 per cent) on component C: Empowerment of the Rural Poor.   

AMSDP faced problems in bringing the assisted producer groups to the targeted levels of institutional development, ensuring their viability and sustainability. In its financial market services component, AMSDP achieved to introduce the warehouse receipt system where 12 SACCOS have been assisted to access loans from banks for on-lending to depositors. However, the guarantee fund and the scheme for lending to value addition and processing activities were not implemented.

AMSDP also had high transaction costs, though relatively less than RFSP partly due to the heavy weight of the infrastructure investments in the overall budget. At programme completion, programme coordination accounted for 17 per cent of total expenditure against 10 per cent in the original budget. This over-expenditure was financed by exchange rate gains and under-expenditure on the budget for Producer Empowerment (21 per cent against 31 per cent in the original budget).

Performance (the combined assessment of relevance, effectiveness and efficiency) is for both programmes assessed as moderately satisfactory with the following combination of ratings: relevance is rated only moderately satisfactory, effectiveness is rated satisfactory while efficiency is rated moderately unsatisfactory for RFSP and moderately satisfactory for AMSDP.

Relevance. Considering the issues of the rural poor in Tanzania and government and IFAD policies and priorities, it was highly relevant for IFAD to support rural finance and agricultural marketing. The partners should also be commended for applying a holistic programmatic approach. However, the design of both programmes had a number of deficiencies so that relevance overall is assessed as being only moderately satisfactory.

Both programme designs included budgets for which the full financing was not secured. In relation to available budgets, the designs had highly ambitious area coverage, implying dilution of support on selected "dots" within a large area. This resulted in high transaction costs which the designs underestimated. Considering that the programmes covered many of the same areas and were implemented by one and the same partner, the designs did not adequately exploit opportunities for savings in delivery costs and for synergies (for example why include a financial services component in the AMSDP). However, such opportunities were pursued during implementation.

The designs underestimated the time and resources required to build viable and sustainable rural grassroots organizations. Furthermore, the designs applied a demand-driven approach instead of pro-actively "picking the winners", i.e. working only with those groups and institutions that have good prospects of viability. Related to this, the designs had the common dilemma that outreach targets had to be phased with a part of the beneficiary groups entering towards the end of implementation, thus receiving support only for a limited period. In the case of RFSP, the design was based on too optimistic assumptions about the level of institutional development of SACCOS and SACAs in spite of past lessons.

The RFSP implementation period was structured in three phases where initiation of the second and third phases was subject to satisfaction of detailed targets (triggers), which to some extent were defined based on uncertain assumptions, e.g. about the level of institutional development of the SACCOS. The Loan Agreement included these details which implied a risk of loan cancellation. While the Flexible Lending Mechanism was intended to introduce flexibility and provide a framework for testing and experimentation, definition of detailed targets/triggers and inclusion of these in the Loan Agreement obviously reduced the space for flexibility and experimentation. Throughout implementation, management felt forced to gear its efforts towards achieving numbers (triggers) and broad coverage in order to move to the next phase, however at the expense of piloting different models and approaches for quality development of rural intermediation institutions.

Effectiveness. With respect to achievement of objectives, both programmes have generally had a satisfactory performance, however with some exceptions. It should also be highlighted that in most cases the appropriate judgement would be to say that the programmes made a contribution to the objective rather than saying that the objective has been achieved and that the achievement can be attributed entirely to the programmes.

RFSP has developed the capacity of 276 SACCOS and some 571 informal self-help groups, in some cases linking them to SACCOS, introducing the concept of "group membership". In many remote places, the support has contributed to giving the rural poor access to semi-formal financial services for the first time while support for SACCOS located in district towns has mainly provided people with an extra choice or service in addition to the services offered by banks and microfinance NGOs. The support has helped many SACCOS to develop their capacity but there is still for many some distance to go before they become self-sustainable; they need to increase their membership, improve their management skills and reduce their portfolio at risk. RFSP has also provided support to improve the entrepreneurial and business skills of the members/borrowers, but inability to properly assess investments and make business plans remains a major challenge. RFSP also contributed to enhanced rural access by facilitating the urban-focused MFI/NGO, Pride, to move into rural microfinance.

The Microfinance Facility included in the design various innovative features, but was not implemented. RFSP only partly achieved its objective of developing new products though there is demand for longer term loans for investments, insurance products, and money transfer services. 

RFSP contributed to improving the capacity of district cooperative officers who support and supervise the SACCOS. However, with the increase in the number of SACCOS, there is need for increasing the number of officers. Furthermore, the cooperative officers, who cover all types of cooperatives, do not generally have the skills to assist SACCOS with more specialized support related to financial services

At meso level, linkages between SACCOS and banks were facilitated and some of these relationships play an important role in the institutional development of the SACCOS. The support for two apex organizations, one for SACCOS [Savings and Credit Cooperative Union League of Tanzania (SCCULT)] and one for MFIs [Tanzania Association of Microfinance Institutions (TAMFI)], did not facilitate these organizations to play an important supportive role for the gMFIs. At macro level, RFSP has worked with the Ministry of Finance and Economic Affairs, Bank of Tanzania and the Ministry of Agriculture, Food Security and Cooperatives to improve the regulatory framework for micro and rural finance, an area that has been receiving significant support from other development partners, notably the World Bank and the Financial Sector Deepening Trust (a major joint initiative of development partners and government).  

AMSDP contributed to the development of an Agricultural Marketing Policy but there is still a long way to go to before the policy paper can have impact and influence on the ground. During the implementation period, Tanzania experienced a decline in its ranking with respect to "trading across borders", partly because government temporarily introduced restrictions on food exports due to local food shortages. Because of these restrictions, some AMSDP beneficiaries in surplus regions with food exports to neighbouring countries suffered a reduction in their prices and income. In some of the AMSDP districts, support was provided to rationalize local taxes and levies that impacted negatively on agricultural trade while at the same time assisting district councils with raising revenue through other means.

In 36 districts, AMSDP contributed directly to formation of producer groups or further development of existing producer/trader groups, in total 1,202 groups with 46,500 members. Per district this comes to an average of 33 groups and some 1,300 beneficiaries, a small fraction (<1 per cent) of the population of most districts. However, in a number of cases people have spontaneously copied the initiative without receiving programme support. Though many groups are at an early stage of institutional development, the evaluation also met many groups that had been able to increase sales and obtain better prices and in a few cases also engage in simple processing, adding value to the produce. A special and innovative feature in developing the capacity of the groups was that the capacity development services were outsourced to contracted partner agencies, i.e. NGOs and private service providers. With a few exceptions, this was a positive experience.

The better prices obtained by many of the groups is not only due to group bargaining power. The support for market information systems, the warehouse receipt system and the construction of feeder roads and market places also contributed. The warehouse receipt system was implemented in 11 districts and is being copied. The warehouse receipt system provides the depositors with liquidity so that they do not have to sell their produce immediately after harvest when prices are lowest. The upgrading of feeder roads has significantly reduced transport costs and in some cases provided communities with access to markets.

The warehouse receipt system was the only intervention of AMSDP's financial services component that was implemented and it was done in close cooperation with the RFSP and the SACCOS, while commendable efforts were made to separate warehouse management from the business management of the SACCOS.

Efficiency. Both programmes faced problems of high delivery/transaction costs partly due to the large geographical coverage. However, AMSDP's expenditure on programme management and coordination had less relative weight in overall expenditure thanks to ADF's US$25 million infrastructure investment which was implemented within standard unit costs, using local and cheaper contractors. 

Both programmes have had very high investments per grassroots institution supported. In the case of AMSDP, an average of US$8,000 was invested in each producer group (excluding programme management costs) ² while many of the groups have annual turnover of less than US$2,000 and a much smaller net income. Investment per direct beneficiary is also very high for the warehouse receipt system but an increase in the number of depositors in supported facilities as well as spontaneous copying of the system would contribute to justifying the high investment.

RFSP focused on SACCOS as the means of expanding financial inclusion. Investment per SACCOS member was high (>US$200). Observers consider it feasible to financially include a person at a cost of US$20 – US$50 which would avoid a continued rise in the total number of financially excluded Tanzanians. However, the SACCOS model has potential for efficiency improvements e.g. by raising membership per SACCOS through mergers or growth from 200-400 to 2,000–3,000. This would reduce unit costs of supervision and capacity development, including training of management and board and committee members. Larger SACCOS would also make it feasible to engage more professional management, which in turn may improve efficiency of operations and improve portfolio quality.  

Impact. The evaluation finds that both programmes have had a satisfactory impact on rural poverty. Available information combined with interviews and observations in the field clearly suggest that AMSDP and RFSP have to a satisfactory degree assisted groups of rural poor in their institutional and business development and individual poor farmers with improving their income and quality of life. While part of this improvement is attributable to other factors, unrelated to the two programmes, most beneficiaries indicated a linkage to programme outputs and activities, clearly perceiving AMSDP and RFSP as crucial to the improvements. Interviewed beneficiaries expressed a high degree of satisfaction with the interventions.

In addition to improving income, assets and food security, the significant investments in developing the capacity of SACCOS, producer groups or other grassroots organizations have made a very positive contribution to empowering members to take better control of their lives and operations. In the area of gender equality, the achievements are significant. Women account for 44 per cent of the total membership in the RFSP-supported SACCOS and groups and do often occupy key managerial positions. In several cases, this experience has induced the women to take part in local politics. In the AMSDP-supported producer groups, women account for 42 per cent of total membership and some producer groups are only constituted of women.

RFSP has contributed to giving some 100,000 households access to financial services. There are some 85,000 borrowers in the RFSP-supported SACCOS and in many cases the credit has contributed to higher agricultural productivity and income as well as establishment of non-farm income generating activities. The increase in income has often been re-invested and/or used for purchasing household assets and financing the children's education. However, in cases of loan default some members are worse off as they lost their collateral. Lack of business and entrepreneurial skills is key reason for defaults. Savings services have allowed households to build up a capital to better withstand a food crisis or other emergency. 

AMSDP has impacted on the target group through various interventions. The support for producer/marketing groups has enabled producers to practice collective marketing, determine profitable prices and thereby enhance their bargaining power. Some groups have also entered into simple processing and some groups have been facilitated to engage in cultivation of "new" crops. Part of the target group benefited from the warehouse receipt system and some depositors improved their prices by more than 100 per cent. The infrastructure investments, in particular the feeder roads, had a major impact on prices and income. The increases in prices and incomes have in turn contributed to more investments in farm inputs, thus raising yields, and/or expansion of the cultivated area.

In the domain of "institutions and policies", the two programmes had a less remarkable impact. Their contribution to developing national regulatory frameworks and the capacity of central institutions was relatively limited. More noticeable impact was achieved at district level where the programmes facilitated the development of public-private partnerships and consultations. RFSP did facilitate an enhanced capacity of district cooperative officers and in some cases district councils now assign priority to rural finance more than they did before the cooperation with RFSP. AMSDP introduced the concept of outsourcing service provision to private and civil society organizations and the programme-supported districts are now familiar with handling the contractual processes. Some districts also benefited from support to rationalize local taxes and levies.

It should be highlighted that over the implementation period there has been an impressive improvement in the capacity of most district administrations. The education and skill levels of district staff are much superior today as compared to the situation when the programmes were initiated. Though resource constraints remain a challenge, the evaluation finds that most of the visited districts are capable implementers of support for rural finance and agricultural marketing.    

Issues of sustainability. While the two programmes have achieved positive results and impact, there are substantial and critical issues of sustainability, involving two sets of questions: (i) will the districts be able to continue the support that is required to make SACCOS, producer groups and other beneficiary groups commercially viable and self-sustainable and, in the case of AMSDP, will the districts have resources to maintain and sustain the investments in marketing infrastructure? And (ii) are the supported beneficiary groups (SACCOS, SACAs, producer/trader groups) commercially viable and self-sustainable, what is the likely mortality rate, and how should one deal with those that have no prospects of commercial viability?

As to the first set of questions, the evaluation found that many of the districts are committed to continuing support for rural finance and agricultural marketing using resources allocated for the District Agricultural Development Plans as well as other resources. However, it was also obvious that they would not have the resources to continue the same level of support provided while the two programmes were on-going. In the case of AMSDP, it is unlikely that districts will have the resources to contract private and civil society service providers for development of producer/trader groups. Development of grassroots institutions for rural finance and agricultural marketing has generally lower priority than education, health, agricultural extension, and social and economic infrastructure. The prospects of maintaining the infrastructure investments, in particular feeder roads, are generally better as this is within what the districts consider as their core mandate. 

As to the second set of questions, the situation is mixed. Several informal self-help groups as well as some of the RFSP-supported SACCOS may collapse or become dormant as they face problems of low repayment and management. However, some may be "saved" if there is a push and support for mergers and a departure from the perception that each ward should have its own SACCOS. Though RFSP has brought the SACCOS a long way in their institutional development, it is still only the minority which are completely self-sustainable. Many of the SACCOS need to increase their scale, and improve on management and repayment performance. Only few of the SACCOS meet the commonly used parameter of 95 per cent loan repayment.

However, given the importance of SACCOS in the local community, it is a possibility that there will be a political push to attempt to save them, in particular those in areas that are not covered by other financial services. Such political priority is not likely to be given to the much smaller producer/trader groups supported under AMSDP. According to self-assessment data, only 16 per cent of the groups have reached what is termed "the performing stage", i.e. they are commercially viable and completely self-sustaining. The remaining groups are at various stages of development and the mortality rate, in particular among those in the early stages, is likely to be high.

It is relevant to ask if major efforts should be invested in SACCOS and groups which are in trouble and have limited prospects of viability in the near future. SACCOS and producer groups are private businesses, and it is normal that governments and donors accept a certain mortality rate when facilitating private business development, in particular start-ups. The approach of the two programmes, which is inline with IFAD's targeting policy, has been to try to accommodate demand and local political priorities for support rather than a more business-like approach where support is limited to those with the best prospects of achieving commercial viability.

Given the above considerations, the sustainability of RFSP is assessed as moderately unsatisfactory. While for the AMSDP supported producer/trader groups which may have even more severe problems, sustainability is rated moderately satisfactory because by far the largest part of expenditure has been invested in marketing infrastructure for which the sustainability prospects are reasonable.

Innovative features and scaling up. In the context of local experiences and the IFAD programme in Tanzania, the two programmes do have certain innovative features. However, rather than their innovativeness per se, both programmes have provided a wealth of experiences and lessons that are useful for scaling up. Indeed, both programmes are often referred to as pilots and the Government's request for a follow-on programme is motivated by the desire to scale up the successful experiences. If speed and cost-efficiency shall be achieved in the scaling up phase, one would need practical what-to-do manuals, developed based on the experiences and lessons learnt. Academic discussion papers on what may have been learnt would not serve as an efficient tool for future implementers.

RFSP was IFAD's first attempt in Tanzania to apply a programmatic and holistic approach comprising support for macro, meso and micro levels. Another design innovation, in the context of IFAD's Tanzania programme, was that RFSP was designed using the Flexible Lending Mechanism. Compared to other development partners and programmes, RFSP has gone to very remote areas. This is emerging as an apparent challenge and strength of IFAD-funded programmes in Tanzania, effectively fulfilling its mandate of reaching out to the rural poor who are not served by others but with considerable challenges of achieving financial viability.

Strategies aimed at getting women to join SACCOS demonstrated innovativeness. For example, very poor women first joined a SACCOS through informal self-help groups such as a SACA. As their economic wellbeing and incomes improved, some of the women were able to join individually, either as owners buying shares or as members.

RFSP introduced subtle mechanisms to subject supported gMFIs into a sense of competition. Monthly reports on key performance indicators of the gMFIs were compiled, summarized into easily comprehendible tables and redistributed to the gMFIs. Both leaders and members of the gMFIs assessed and compared the position of their organization against the others in the report. Visited SACCOS members acknowledge that this was instrumental in pushing them to set higher standards and targets, e.g. for loan repayment and savings mobilisation.

AMSDP had two innovative elements: i) utilisation of private sector partners as the outreach mechanism for developing the capacity of beneficiary groups; and ii) involvement of beneficiary group representatives in marketing research and market information. The partner agencies were mostly locally-based business development service providers, NGOs, or private consultants and they were generally well received by the producer groups. The interventions delivered by partner agencies were assessed as valuable and many beneficiaries give credit to the partner agencies for the fact that they have reached a certain stage of development and empowerment. The best performing partner agencies effectively served as mobilizers, motivators, trainers, and mentors, and they were instrumental in getting the produce marketable. The evaluation visited a number of groups where one or two members were part of a marketing research team headed by AMSDP's marketing specialist, usually a local marketing expert contracted to assist and mentor the groups. This was an efficient and innovative way of engaging beneficiaries in identifying new potential rural markets and buyers and accessing the marketing information available.

A number of AMSDP interventions are being copied and replicated spontaneously without AMSDP support, including: (i) producer group formation and collective marketing (at times facilitated by AMSDP producer groups); and (ii) the Warehouse Receipt System for which communities have requested support through their members of parliament and/or the District Councils.

Performance of partners. For both programmes IFAD and government performance is assessed as moderately satisfactory. While the partners are responsible for a number of deficiencies in the programme designs, their performance during implementation was overall satisfactory. IFAD's support for implementation improved after establishment of country presence in 2004. An innovative and productive supervision modality was introduced in 2008, where programme coordinators and M&E staff participated in supervision of other IFAD-supported programmes in order to facilitate learning and peer review. RFSP and AMSDP were also supervised jointly by IFAD in 2009 to strengthen complementarity between the two. Until introduction of direct supervision, the United Nations Office for Project Services (UNOPS) performed satisfactorily as cooperating institution supervising implementation.

While the minor bilateral cofinanciers were relatively passive but delivered on their financial commitments, the cooperation in AMSDP with the African Development Bank as cofinancier had various challenges. The Bank had its own systems for monitoring, evaluation and reporting and has recently submitted its own Project Completion Report (in addition to the one prepared by the Programme Coordination Unit/Government of Tanzania). Though attempts were made at the end of the programme to introduce common and joint reporting and supervision, the Government was during most of the implementation period unduly burdened with having to comply with two different organizational reporting and supervision systems. In the spirit of the Paris Declaration, both organizations need to adapt in terms of coordination and harmonization.

Conclusions. Support for rural finance and agricultural marketing was and continues to be highly relevant to the context and in line with government and IFAD priorities. The current effort to merge support for the two areas into one programme is commendable given the opportunities for synergies and cost savings. The design of the new programme should avoid the deficiencies in the design of RFSP and AMSDP where certain parts were based on uncertain or too optimistic assumptions. Furthermore, the area coverage was too ambitious in relation to the available resource envelope, resulting in dilution of support and high transaction and delivery costs.

Both programmes delivered on most of their quantitative targets, made satisfactory contributions to most of their objectives and had a satisfactory impact on individual beneficiaries and beneficiary groups, in terms of improving incomes and food security and empowering beneficiaries, in particular women. However, many groups are still struggling to achieve commercial viability and become self-sustainable, and investment per beneficiary group was high and local governments do not have the resources to continue the programme levels of support.

Thus, the follow-on programme needs to give priority to consolidating the successful developments that were initiated by the two programmes. However, when doing so it needs to be recognized that the support is for facilitation of private business development hence a business-oriented approach, emphasising commercial viability, should be adopted. Furthermore, in an up-scaling phase, careful consideration needs to be given to reducing the delivery costs and investment per beneficiary.    

Recommendations. On the above background and considering preliminary proposals on the design of the follow-on programme (MIVARF), the evaluation offers four recommendations for consideration:

Recommendation 1

Scale up the positive experiences of RFSP and AMSDP and avoid support to organizations and activities with little potential for success

As highlighted in this report, there are many positive experiences from RFSP and AMSDP that MIVARF can build upon, scale up and consolidate. Many of these positive experiences may need to be refined and further developed but they do represent a valuable foundation for designing the future support. There are, however, also cases where the support did not produce the expected results and where it is unlikely that continued or different forms of support will produce acceptable results. Therefore, it is important not to continue the support "in order to save the organization". Certain levels of mortality are to be expected for the supported gMFIs, producer groups and agribusinesses. The design of MIVARF should attempt to define acceptable mortality rates (or targets/success indicators for survival) for the different target groups/enterprises.

Recommendation 2

In design, maintain the focus on improving access to financial and marketing services of the rural economically active poor but emphasize financial and commercial viability and sustainability in the support to the beneficiaries.

While focus should continue to be on the rural active poor, it needs to be emphasized in selection and support strategies that grassroots financial institutions as well as producer/trader/processing groups are market-based and that a good social rationale only is insufficient to ensure their survival. Financial and commercial viability must be at the forefront when deciding whom to support and how to do it.

Recommendation 3

Reduce delivery costs and investment per beneficiary, and improve cost-effectiveness

The high delivery costs and investment per beneficiary would not be acceptable in the scaling up phase. In RFSP and AMSDP, a major factor was the combination of operating largely in project mode and the ambitious area coverage, distributing relatively limited budgets over vast geographical areas and many districts. The consideration for the design of MIVARF is to scale up to national level coverage. This will multiply the challenges faced in RFSP and AMSDP unless MIVARF activities are mainstreamed into the local government planning and financial management system. Potentially, this could reduce delivery costs. However, while the district administrations remain the appropriate place for anchoring the field implementation, they will require substantial external hand-holding and support (from a PCU) in order to handle the business and value chain development activities. Such hand-holding is costly and will result in very high delivery costs if the investment per district is modest (because the MIVARF budget may be spread over many districts). 

The lessons and experiences of RFSP and AMSDP make it possible to avoid costly searches and experiments on how to do things. IFAD should facilitate the production of practical what-to-do manuals, developed based on the experiences and lessons learnt. Such manuals should inform district staff and other implementers about what to do and how to do it for different activity areas of the programme. In developing such manuals, it should be an overriding priority to reduce delivery and unit costs.    

Recommendation 4

Emphasize development of partnerships and linkages at all levels

There are many organizations and programmes that support the development of rural finance, agricultural marketing and value chains. Some address levels (e.g. investment finance for large agribusinesses) that are beyond the scope of IFAD's support but still important for development of the value chains, some may fill other gaps in the MIVARF support while some have a time horizon that goes beyond the completion date for MIVARF. For these and other reasons, it is important that the MIVARF develops operational partnerships with such programmes and organizations and also facilitates the MIVARF target districts to access such cooperation.

It is also important to eliminate the perception of "MIVARF groups" and instead promote an approach whereby producer groups and gMFIs that are supported by MIVARF are facilitated to access support and services from other organizations and programmes. From the start of MIVARF's support for a group, attention should be given to avoiding the development of "MIVARF-dependency" and to promoting self-reliance.

---