Completion evaluation ¹

Introduction

The economy. The ADP became effective in 1997 with the country still suffering from the abrupt withdrawal of Soviet Union capital at independence in 1991 and the collapse of the centralised economy. At the time of project design in 1995/6, industrial output had fallen to around one-sixth of its 1989 level and inflation was rampant. The United Nations Development Project (UNDP) National Human Development Report of 2001/2 estimated that 45 per cent of Georgian households were living below the poverty line, at which time the average per capita income stood at around US$680. It was not until the removal of the Shevardnadze government in 2003, that measures were put in place to address corruption and liberalise the economy.

Agriculture. Hit by the lack of investment and markets and the emigration of labour, the agricultural sector effectively reverted to subsistence farming during the 1990s. The share of agriculture in Gross Domestic Product (GDP) fell from around 50 per cent in 1990 to around 16 per cent in 2004. The unofficial land privatisation following the break-up of collective farms was regularized, with priority given to existing farmers and other rural residents. The IFAD Country Strategic and Opportunities Paper (COSOP) of 2004 saw the major challenge in rural areas as ‘the transformation of the new recipients of farmland from farm workers to decision-making farmers'.  

Rural finance. The rural finance landscape of Georgia is represented by commercial banks, credit unions and MFIs, with the National Bank of Georgia (NBG) acting as supervision and regulating agency. Commercial banks have started to move into rural microfinance, but credit unions are effectively the only institution with the potential to reach vulnerable groups.

The project. The ADP constituted IFAD's first involvement in Georgia². It was designed by the World Bank/International Development Agency (WB/IDA), with IFAD agreeing during the negotiation phase to co-finance the components for credit unions and land registration. Total project costs were calculated at US$26.3 million financed largely by an IDA loan of US$15m and an IFAD loan of US$6m. The IFAD loan became effective in August 1997. Project completion was scheduled for April 2002 and progressively extended until 30 June 2005. The project financed five components: Credit to Enterprises; Credit Union Development; Land Registration; Development of Agricultural Services; Project Management.

Objectives. The overall objective of the IFAD-funded components was the reduction of the poverty level and food vulnerability of rural households. Specific objectives included the establishment of a network of village-based rural financial intermediaries and the emergence of an active land market. The project area of the credit union component covered the entire country, while the Land Registration component concentrated on two districts only. The target group comprised rural households with low incomes, typically with less than a hectare of land and an annual income of US$200-400.

Components. The Credit Union component aimed to support the development of up to 120 sustainable credit unions (CUs), with funding provided for loans to members and for initial costs, and to provide the basis for the legislative regulation of CUs. The aim of the Land Registration component was the establishment of land registration offices in two districts and the registration of 130 000 land ownership titles, with a view to countrywide replication of the scheme. Under the Credit to Enterprises component, a line of credit was provided to commercial banks to finance selected private enterprises, the majority of which were expected to be among the agro-processing companies that make up three-quarters of all enterprises in the country. Under the small Agricultural Services component studies were to be made to assist in restructuring agricultural services and identify appropriate investment programmes. The responsibility for implementation rested with the Ministry of Agriculture, Food and Industry.³ A Project Coordination Unit (PCU) was established to manage planning, budgeting and procurement as well as project monitoring and evaluation (M&E).

The evaluation. The evaluation aimed to assess the relevance of project objectives to the rural poor and to IFAD strategies, the extent to which these objectives were achieved, the sustainable impact of the project and the performance of the partners involved.  An evaluation mission was conducted from May 15 to 8 June 2006.  The mission held discussions with the relevant government ministries, the National Bank of Georgia, WB/IDA and other donors, microfinance institutions, project officials, staff and clients of the Land Registry Offices, and the management and beneficiaries of ten credit unions. Field visits and mini-surveys of beneficiaries were carried out for both IFAD-funded components, which constitute the chief emphasis of the evaluation.

Project performance

Project rationale. The ADP placed appropriate emphasis on re-building the rural economy after the cessation of Soviet capital investment. The establishment of a viable rural finance system and the setting-up of an efficient land registration system addressed basic needs in rural areas where such systems did not previously exist. The priorities of ADP were confirmed in the IFAD's subsequent sub-regional and country strategy papers.⁴

CU component design. Targeting mechanisms were set out to ensure that the CU component would reach the very poor, including the stipulation of 30 per cent CU membership for 'vulnerable' households, although the only logframe indicator relating to poverty impact was 'access to credit'. No survey of regional credit unions was carried out and the CU model was to be developed through 'learning by experience', placing a heavy responsibility on the Credit Union Development Centre, set up under the PCU to manage the component. Sustainability was defined at appraisal, but no relevant statistical targets were put in place and the exit strategy was not adequately developed. The start-up grant of US$3 000 was large enough to cause the formation of CUs solely for the sake of obtaining the grant; in some cases, the credit unions existed on paper only.

Changes during implementation. The Mid-Term Review (MTR) recommended a reduction of targeted numbers of CUs, from 120 to 55, with an attendant re-allocation of funds. The share of the CU component in the total budget thus fell from 31 per cent to 25 per cent while that of the Land Registration component rose from 20 to 25 per cent. The MTR revision reflected the unsatisfactory implementation of the CU component and allowed for the concentration of resources on sustainable CUs. The project completion date was extended three times for further institutional development in the Land Registration component and the restructuring of project-supported credit unions.

Disbursement. Around 93.0 per cent of the project funding of US$21.5 million was disbursed, including the same proportion of the IFAD loan. The principal and interest repaid by the credit unions was deposited in a revolving fund which, together with the Insurance Fund, contained nearly US$2 m in June 2006. The sum has been earmarked for credit union support under the RDP.

Credit Union component. A total of 164 credit unions were created in the first two years of the project. This over-rapid expansion was caused by politicisation of the process and resulted in widespread fraudulence and mismanagement. The financial situation of CUs was revealed in 2001 when the majority of CUDC loans were scheduled for repayment. By 2002, loan repayment performance had reached a low of 30 per cent. Auditing was not carried out, the monitoring of loans made by CUDC was weak and performance standards were neglected. Managers of credit unions were poorly trained and many credit unions were essentially structures created by village elites to harvest donor money. There was little emphasis on savings mobilisation or sustainability.

These developments took place against a challenging background which included the Russian financial collapse of 1998, severe droughts and floods in 2000 and 2001, frequent staffing changes in the relevant ministries and allegations of corruption in high places. The result was a crisis of confidence within the Government and the country at large, the withdrawal of the Finance Ministry's representative from the credit committee and a consequent two-year hiatus in implementation. Of the 164 CUs, 35 were liquidated or merged and 71 were taken to court. The underlying causes of the collapse of the network were addressed by a fresh management team in early 2003 through the restructuring of loans and the re-commencement of lending. At project completion, there were 58 functioning CUs, three more than the MTR target, but one year later the number had fallen to 39. 

Although the demand for CU membership has remained strong, actual membership fell from 12 000 at mid-term to around a quarter of that currently, with three CUs accounting for nearly 20 per cent of the total. The pro-poor quotas were abandoned after resulting in high delinquency rates. Mobilisation of savings increased up to 2001 but fell sharply thereafter, leaving the CUs dependent on project credits. At completion, only around 40-45 per cent of the total loan portfolio of the 15 top-performing ADP-supported credit unions was financed by savings, less than half of the internationally recommended benchmark. In 1998-2000, 80 per cent of loans were agricultural, but this led to repayment problems and the proportion has fallen to around 60 per cent, with successful CUs turning to commercial loans. Interest rates paid by borrowers ranges from 20 per cent to 42 per cent annually.

Land registration component. The component established modernised land registration offices in two districts close to the capital which have registered around 170 000 land parcels and issued 155 000 land titles, 85 per cent and 119 per cent of appraisal targets. Aerial photography was carried out on 2 000 sq. km of the selected districts and 10 000 sq. km elsewhere. The ground survey of land parcels was conducted by private survey companies supported by the project in the form of technical assistance. Savings were made by means of efficient surveying methods, with the average cost per parcel reduced by almost half. To these savings were added the budget increase at mid-term, enabling the component to refurbish and computerise 11 regional and 37 district registries countrywide. Over 200 staff were trained in legislative, managerial and technical aspects of land registration resulting in the emergence of an extensive cadre of officials with geodetic and cartographic skills. Project staff contributed to the first manual of registration, based squarely on the experience of the two supported district offices.

The component was implemented originally through the State Department Land Management, set up in 1997 with technical assistance provided by the project. It since been replaced by the National Agency for Public Registry (NAPR) in the Ministry of Justice, a transition physically and financially supported by ADP.  NAPR currently acts as an independent and self-financing land registration and cadastre agency coordinating 67 offices countrywide. Around 4 million land parcels have been registered in different parts of the country.

Credit to enterprises component.  A total of 48 loans were made for investment and working capital in agribusiness enterprises, to a total value of US$8.56 million. Two-thirds of supported enterprises were made up of canneries, wineries, livestock farms and hazelnut processing plants. Demand for medium and long-term credit was high and the repayment rate was satisfactory. Repayments were placed in a revolving fund from which additional loans were made until its transfer to the state budget. Despite the bankruptcy of two participating banks, the government eventually received a sum greater than the value of the initial credit line, with interest payments more than compensating for the deficit on principal. Over 90 per cent of sub-projects were calculated to have financial rates of return of around 25 per cent at mid-term, well in excess of the projected rate of 15 per cent.

Agricultural services component. Studies were conducted in the areas of irrigation development and the agricultural research and extension system. These were used in the design of two WB-financed projects, the Irrigation and Drainage Community Development Project (IDCDP) and the Agricultural Research, Extension and Training Project (ARET).

No viable CU network. An effective CU network was not achieved under the project, with the actual figure of viable CUs now at around 15-20, 13-17 per cent of the original target or 27-36 per cent of the revised target. What has been achieved, however, is official recognition that sustainable village-owned CUs are feasible and desirable, given the appropriate management approach, training programmes and level of commitment. The existence of two or three successful CUs and a further 10-12 with reasonable prospects of sustainability, the emergence of an embryonic CU association and the corpus of lessons learned mean that the component can by no means be written off as a failure. The economic and political contexts posed formidable challenges to implementation in the early years of the project.

Increased liquidity in land markets. Previous land registration procedures were lengthy and expensive, taking up to two months and involving extra-legal charges. This has now been reduced to six steps, which can usually be concluded in a single day on payment of the official fee only. There has been a marked increase in land transactions in the two pilot districts and land values have increased sharply, although limited collaboration among the eight different donors involved in land registration schemes hindered the development of a uniform system.

Agricultural credit system. The Credit to Enterprises component fulfilled its main objective by increasing credit flows to rural areas through the development of a rural commercial credit system using commercial banks. The loans made under the component constituted a significant increase in medium-term lending to the agro-processing sector, which essentially did not exist at the start of the project. The individual loans have been successful in providing rural employment, increasing agricultural output for export and achieving a satisfactory rate of return.

Poverty focus. CUs agreed that the bulk of their membership was drawn from 'average' and ‘poor' socio-economic categories, which accounted for 65-90 percent of village populations and 80-90 per cent of CU membership. Loans to better-off families were rare and to the poorest almost non-existent after the first two years. The consensus among the donor community, the microfinance sector and former project staff is that membership of credit unions will be dominated by the 'economically active' population, a reality apparently unforeseen at the time of project design.

Limited integration of components. In essence, the ADP was made up of four distinct components having only a management structure in common. The lack of zonal planning meant there was no integration between the Credit for Enterprises component and the CU component. Nor did the two components co-funded by IFAD have any close connection, with no credit unions established in the two districts selected for Land Registration activities. The small Agricultural Services component has the appearance of being tacked-on to a project with other priorities.

Assessment: relevance. ADP addressed significant constraints to rural development, in particular the lack of access to rural credit and of functioning agro-processing enterprises, but the design of the CU component was flawed by insufficient attention to capacity building, sustainability factors and exit strategy. There was also no clear rationale in terms of IFAD's comparative advantages for the Fund's involvement in the Land Registration scheme. Security of tenure for smallholders and the use of their land as collateral were included among the component objectives but were secondary to the facilitation of land transactions..

Assessment: effectiveness. The Credit for Enterprise component achieved its objectives in terms of the emergence of profitable enterprises and influence on the banking sector. The Land Registration scheme greatly accelerated the registration process and substantially assisted in the emergence of an active market in land. However, the Credit Union component failed in its stated objective, even while setting important precedents and, eventually, influencing government attitudes.

Assessment: efficiency. Direct beneficiaries of the ADP include nearly 50 enterprises, over 12 000 CU members who gained access to retail financial services for the first time, and over 100 000 landowners who received land titles or mortgage registrations. The Land Registration component was efficiently managed, allowing for expansion of the project to new areas, but donor harmonisation was weak. Efficiency ratios calculated for CUs reveal acceptable expenses/portfolio ratios, improving (but still high) expenses/assets ratios and very high staff/client ratios. The CUDC management team was overstaffed and underskilled and no component resulting in an implementation hiatus of nearly two years and 70 court cases can be considered efficient.

Performance of IFAD. The goal of building a CU network from scratch was courageous and may yet prove a sound development option. A more realistic approach to targeting, a clearer insistence on the elements of sustainability and a consistent IFAD presence on supervision missions would have been desirable. No major problems with the disbursement of funds, procurement or contractual arrangements were noted.

Performance of Government. A supportive legislative framework was established by the enactment of a Land Registration Law in 1996 and a Law on Credit Unions in 2002. Official support for the project has been fairly consistent, although the crisis in the CU component provoked the MOF to cause the suspension of the loan facility. CUDC was obliged to step in as an on-lending institution when the arrangement with MOF and the commercial banks failed to materialise, which it did without proper training or appropriate legal basis.

Performance of National Bank of Georgia. The partnership between ADP and NBG worked well in the preparation of the enabling legislation, and NBG's assistance to credit unions in meeting prudential regulations is appreciated by credit union managers. NBG regulations generally correspond to international best practices. The 2002 Law is acknowledged to have certain defects, among them the prohibitive expense of external audit.

Performance of the World Bank. The collaboration between WB and IFAD continues under RDP. IFAD is thereby involved in larger programmes with greater resources, while WB's influence in terms of policy dialogue is reflected in the close relation between ADP activities and the developing legal and institutional framework. Supervision of the project was regularly carried out by the WB, latterly with an IFAD representative on the team. WB supervision missions were slow to identify the causes of the crisis in CUs and to take preventive measures.⁵

Impact ⁶

Physical assets. The main impact under this heading has been the emergence of an active land market, the first beneficiaries of which were small landholders able to get a fair return on the sale of unutilised land. Increases in the price of newly-registered land were intensified close to the district capital, with some plots now selling at up to five times as much as three years ago.

Agricultural productivity.  Around 60 per cent of loans, with a total amount of US$1.36 m, were made by CUs for agricultural purposes. This represents a direct investment of US$7 000-10 000 in 150 villages. From 50-100 ha of new or revived cultivation per village may have been financed in this way, but no figures for increased agricultural production are available at village level and no impact indicator was included in the logframe. Some agricultural diversification was noted, for example the commercial cultivation of strawberries and greenhouse vegetables in new areas.

Food security. Some CUs have made emergency loans to very poor households, but the impression is that families affected by food insufficiency rely on relatives in the lean months. Among CU villages visited by the mission, estimates of food-insecure households varied between 5 per cent and 25 per cent, for the most part substantially below the national average of 29 per cent. Some CUs visited described a proportion of the communities they served (up to 10 per cent) as 'hungry', but these households did not constitute the target group of CUs after the first two years of the project.

Human assets. At least 2,000 CU loans were used for consumption purposes, with medical and educational expenses among the first priorities. At one CU, 40 per cent of all loans were used for health and education purposes. Of the households interviewed in 2003, nearly a third utilised their loans for medical treatment, educational expenses, house renovation, household goods and wedding expenses. Training was provided to CU staff in financial management, loan appraisal and book-keeping, and the emergence of well-trained staff in the land registry offices has been a notable achievement.

Social capital. Villagers stated that the CUs supported them during difficult times and gave them a sense of hope. This was borne out by the 2003 survey, in which views of overall CU operations were 99 per cent positive. CUs are often found at the heart of village life and are gradually becoming stronger self-help institutions. Women account for 50-55 per cent of CU membership and the same percentage of loans.  In many CUs, women dominate the committees, and the managers are often women.

Institutions and services. The 2002 Law on Credit Unions was drafted with the support of CUDC. New reporting forms and prudential standards were introduced and training was provided for credit unions in applying them. Under the land registration component, land registration offices using modern techniques and streamlined procedures revolutionised an inefficient system of doubtful legality. The establishment of the NAPR was a direct result of ADP initiatives.

Financial assets. More than 12 000 rural people gained access to retail financial services for the first time since independence. Over US$1.5 m were mobilized in savings and around UDS 7 m in rural lending was generated by CUs. All surveys indicate that the CUs constituted the only available source for the great majority of these loans, which enabled borrowers to initiate new income generation activities or to re-start productive enterprises relinquished through lack of finance. The household survey of  2003 showed that 28 out of 40 household used their loans primarily for business purposes. The range of income increase was wide (17 per cent to 200 per cent) but never insignificant. 

Sustainability and ownership. Project achievements in the spheres of land registration and credit to enterprises seem assured, on the one hand through the institutionalisation under NAPR, and on the other through the boost to the banking system. Of the 15 CUs for which sufficient data exists, two have achieved a financial sufficiency ratio of 85 per cent and a further six 65 per cent, but the sustainability of the remainder is problematic. The picture is more favourable in terms of 'ownership': there is everywhere strong evidence of commitment among members and managers, including community initiatives funded by members and managers working on reduced or unpaid salaries.

Innovation, replicability and scaling-up.  The major innovation of ADP was in its attempt to set up a village-based network of financial institutions in a country where there was virtually no access to formal credit in rural areas. This was a bold initiative in a country with a generally negative attitude towards cooperatives in the wake of the Soviet experience. With the benefit of the hindsight, the limitations of the project were to (i) underestimate the risks of developing financial institutional innovations under a volatile policy environment and (ii) provide limited technical assistance to support the development of CUs (the CU component was based on WB experiences in Albania, where the 'learning-by-experience' method had given positive results). Under the Land Registration scheme, the ADP registration system has been chosen as the model of countrywide replication.  This case provides an example of replicating an innovation by forging an alliance with a partner (EU) with practical experience and relevant technical expertise (through its former pilot project).

Conclusions and recommendations

Conclusions

Overall assessment.  ADP was a project implemented in difficult conditions.  The establishment of a replicable system of registration and titling was a major achievement. The lack of integration between components, however, meant that the project was no more than the sum of its parts. If the credit unions had served to finance the production of the raw material for supported enterprises, the overall impact of the project might have been more impressive.

Disturbing financial trends in CU development were overlooked by supervision missions and the MOF decided to withdraw its support rather than seek remedial measures. The result was a setback to the reputation of the CU movement in Georgia from which it is only now recovering. Yet the obstacles to sustainability are not insuperable.

The road to CU sustainability.  In order to achieve long-term sustainability, credit unions need to concentrate on savings mobilisation, loan diversification, closer attention to loan assessment and control of delinquency, and the minimum of loan refinancing and rescheduling. The best performing credit unions seek their clientele over areas wider than a single village, have de-emphasised agricultural loans, accept only members with strong repayment capacity, and are considering moving their headquarters to trading centres.

Vulnerable groups. The risk is that CUs will focus on the more attractive urban loan market and lose touch with the poor in rural areas. Particularly affected would be groups lacking in entrepreneurial initiative and unable to meet collateral requirements. The Government considers that farming households which the lack the labour or the ability to exploit their land in a businesslike way can only be reached by welfare programmes, but such programmes are presently scarce.

Alternative approaches. It is important for IFAD to explore other approaches.  An immediate objective must be to ensure that all cultivable land is productively utilised. Movements in this direction have been taken under RDPHMA, notably the setting up of cooperative ventures to provide inputs and machinery, but the success of this approach is not yet clear, also due to implementation management problems the project is facing. 
 
Markets. Alternatives to the banned Russian market are being pursued by the Government, but solutions will take time. Private initiatives are being undertaken to set up processing plants in rural areas but these are still isolated examples. Meanwhile, the issue of access to urban markets within Georgia needs to be urgently addressed. In remote and mountainous regions, roads and transport are one aspect of this problem, but there are also indications that large urban markets, particularly Tbilisi itself, are controlled through unofficial monopolies.

Partnership with the World Bank. This report stresses the advantages of partnership with WB in Georgia, while drawing attention to its problematic aspects, which include differences over the importance of targeting and a neglect of poverty impact in supervision. IFAD has also moved out from its area of expertise by co-financing the Land Registration component, probably as a result of the Fund's involvement after the main features of the project had been determined.

Recommendations

IFAD should clarify its priorities and conditions for involvement in co-financed projects. The Fund should ensure the inclusion of impact indicators connected with its concern for marginalized groups and should maintain a regular presence on supervision missions.

Measures will be required to support households that are unable to "automatically" take dvantage of interventions aiming at agricultural commercialisation. The Government needs also to intensify the development of an effective extension system and take measure to liberalise access to urban markets.

Credit unions are part of the solutions to rural poverty but not a panacea.  Measures are necessary to strengthen the economic tissue of rural areas and increase the number of potential borrowers.  These measures may include: (i) more direct solutions to the problem of uncultivated arable land, based on access to mechanised equipment for example through customised financial services; (ii) improved marketing methods for agricultural products, including attention to post-production phases; (iii) the liberalisation of urban markets within Georgia.

Future microfinance projects in Georgia and the region should learn from the principal error made under ADP, which was to allow the numerical target of CUs to become the predominant indicator. Project designs should recognise that the sustainability of CUs depends on high repayment rates and seek appropriate ways of reaching the poorest groups. CUs that have reached a certain level of assets could be offered incentives to undertake pro-poor initiatives.

Regarding land registration, the mission recommends measures to protect rural households from the effects of land speculation in areas where the demand for residential property is likely to catapult land prices beyond the means of local farmers. The priority given to village residents in the original distribution of land was aimed at guaranteeing the basic livelihoods of those depending on the land for their survival and this aim should not be jeopardised.

---

1/ The evaluation team was led by Mr Roger Norman, Agricultural Economist; Ms Margarita Lalayan, Rural Finance Specialist, and Mr Giorgi Badrishvili, Rural Sociologist.  The main evaluation mission took place from 15 May to 8 June 2006.  The wrap up meeting (8 June) was attended by Mr Pietro Turilli, Country Programme Manager/IFAD-PN.  Mr Roger Norman and Ms Thuy Thu Le, Evaluation Research Analyst/IFAD-OE conducted a preparatory mission in April 2006.  Mr Fabrizio Felloni, Lead Evaluator/IFAD-OE supervised the evaluation process.

2/ ADP was followed by the Rural Development Programme for Mountainous and Highland Areas (RDPMHA) and the Rural Development Project (RDP), co-funded with WB-IDA and other partners. The three IFAD loans amount to an investment of around US$25 million over a period of 15 years.

3/ Now the Ministry of Agriculture (MOA).

4/ Sub-regional Strategy Paper 1999, Country Opportunities and Strategy Paper 2004.

5/ Duly acknowledged by the WB in their December 2005 completion report.

6/ Two impact indicators only were selected for the IFAD-funded components. There was no monitoring of food security or income levels among beneficiaries, or of changes in the utilisation of cultivable land, and no baseline survey for these indicators. Partially offsetting these omissions were a 1998 Rural Poverty study and a 2003 OE Thematic Evaluation, which both contained useful household level data. The present mission conducted 34 interviews with CU members and non-members in ten villages and with a further 32 beneficiaries of the Land Registration scheme.