FINAL DRAFT Promoting Good Local Governance through Social Funds and Decentralization Andrew Parker Rodrigo Serrano May 2000 A study jointly funded by the World Bank's thematic groups on Decentralization, Municipal Finance and Social Funds; the United National Capital Development Fund, New York; and trust funds of the Government of Switzerland Supporting good local governance through social funds and decentralization ii ABBREVIATIONS AMHON = Asociación de Municipios de Honduras CPC = Community Project Committee FDC = Fondo de Desarrollo Campesino FHIS = Fondo Hondureño de Inversión Social (Honduras) FIS = Fondo de Inversión Social (Bolivia) FISDL = Fondo de Inversión Social y Desarrollo Local (El Salvador) FISE = Fondo de Inversión Social y Económica (Nicaragua) FONCODES = Fondo de Compensación y Desarrollo Económico y Social (Peru) LDF = Local Development Fund MASAF = Malawi Social Action Fund MDF = Municipal Development Fund MP = Member of Parliament MPU = Microprojects Unit (Zambia) NGO = Non-governmental organization O&M = Operation and Maintenance PROFINES = Proyecto para la Focalizacion de la Inversion Economica y Social SDF = Social Development Fund (Zimbabwe) SFKC = Social Fund of the Kingdom of Cambodia UNCDF = United Nations Capital Development Fund UNDP = United Nations Development Program ZAMSIF = Zambia Social Investment Fund Supporting good local governance through social funds and decentralization iii CONTENTS Abbreviations .................................................................................................................... ii Contents ............................................................................................................................ iii Acknowledgments ............................................................................................................ iv Executive Summary .......................................................................................................... v Introduction....................................................................................................................... 1 Establishing Participatory Planning Processes.............................................................. 6 Creating Effective Local Financing Mechanisms ........................................................ 14 Building Local Capacity ................................................................................................. 20 Increasing Accountability in Local Institutions ........................................................... 25 Ensuring Project Sustainability..................................................................................... 29 Conclusions ...................................................................................................................... 33 References........................................................................................................................ 36 Supporting good local governance through social funds and decentralization iv ACKNOWLEDGMENTS This study would not have been possible without the support and assistance of many different people and organizations. The authors would first like to express gratitude to the social fund organizations visited: Parker (Learning Specialist, PRMVP) carried out the field work in Malawi, Zambia and Zimbabwe; Serrano (doctoral candidate, MIT) in Bolivia, Honduras and Peru; and Pagaran (doctoral candidate, MIT) prepared a specially commissioned country study on Cambodia (Pagaran 1999). Many thanks also to the staff of United Nations agency missions who also provided valuable time for discussion and field visits. We would also particularly like to thank Bank staff working in country offices who went out of their way to assist us. Of course the biggest vote of thanks goes to those many people in communities who took the time out of their busy days to participate in discussions during the field work for this study. The authors would like to thank the following for comments on earlier drafts: Anush Bezhanyan, Julie van Domelen, Laura Frigenti, Maurizio Guadagni, Jim Hicks, Steen Jorgensen, Jennie Litvack, Dinah Mcleod, Laura Rawlings, Lynne Sherburne-Benz, David Warren and Dana Weist of the World Bank; Doug Porter, Leonardo Romeo and Roger Shotton, of UNCDF; and Lourdes Pagaran, Paul Smoke and Judith Tendler from MIT. Any errors and omissions remain those of the authors. Supporting good local governance through social funds and decentralization v EXECUTIVE SUMMARY Introduction 1. In recent years, the Bank and other donors have increased their focus on local governance and are now supporting initiatives that allow local institutions to play a much greater role in the planning, design, implementation, and operation and maintenance of smallscale investment projects. Two important aspects of Bank support for local governance are social funds, which seek to target and empower poor communities to improve participation and local service delivery, and support to governments committed to decentralization of responsibility and power for local development to local governments and other local institutions. 2. There are concerns that these two approaches, which address different elements of governance, may sometimes work at cross-purposes. For example, if social funds establish parallel channels for local expenditures and community participation without building proper channels for local accountability or financial sustainability, they can weaken nascent local governments and impede decentralization efforts. Conversely, if decentralization programs build financially healthy local governments without provisions for orienting spending to the poor, they will be at odds with social fund efforts to increase access of the poor to local services and infrastructure. 3. This study was therefore commissioned to examine the interaction between social funds and decentralization in seven countries at different stages of decentralization: Bolivia and Honduras (advanced decentralization), and Peru and Zimbabwe (some decentralization) and Cambodia, Malawi and Zambia (little or no decentralization). It explores the relationships between social funds, decentralization and local governance. Approach 4. Field work comprised of two elements: (i) a series of interviews with key actors working on social funds, central and local government officials, elected representatives, NGOs as well as Bank Task Managers and other staff working on local development programs; and (ii) site visits to social fund projects in the seven countries visited to meet with community members to discuss social fund operations. The field work was supported by a detailed desk review of relevant information and secondary data, including internal government documents, external evaluations, and academic research. 5. For comparative purposes, some alternative approaches to infrastructure and service delivery that focus on support through local government bodies are also examined, including municipal development funds, which are designed to remedy some of the typical constraints of local governments (lack of capacity, limited resource, and limited access to financial markets), and the UNCDF's local development fund (LDF) strategy for supporting local governments. Using a combination of capacity building programs together with a fiscal instrument designed to simulate an unconditional block grant from central government, the LDF approach aims to test innovations in the "how- to" of decentralization policy. Supporting good local governance through social funds and decentralization vi Establishing Participatory Processes 6. Social funds encourage communities to participate at many key junctures in the project cycle, being one of the main instances in the project selection process. Based on the case-study countries two main approaches were distinguished: (i) individual communities select projects from a social fund's menu through a community assembly meeting and send these projects to the social fund which generally approves them on a first-come, first-served basis; or (ii) local governments coordinate a systematic local planning process in which each community first selects its priority, and those priorities are then discussed and ranked at the municipal level by local government authorities and community representatives. The social fund finances the top priorities that fall within the fund's menu of eligible project-types. 7. Allocative Efficiency. Through decentralization of project selection to the community level, social funds aim to improve allocative efficiency by delivering a mix of public goods and services that fits local preferences better than centrally-controlled and implemented programs. An important finding from the field work suggests that social fund investments will have better chances of achieving allocative efficiency if sub- projects come through a comprehensive local planning process rather than from isolated community proposals. 8. A local planning system increases allocative efficiency by promoting a process by which all communities express their preferences rather than only a few, and it creates an instance at the local government level where these preferences are discussed and investment choices are made, rather than leaving this decision to the social fund headquarters. However, the viability of having an effective local planning system is partly contingent on the country's decentralization framework. Countries with decentralization processes in place, such as Bolivia and Honduras, find it easier to support local planning processes than those without them. In countries where there is no functioning representative local government and administrative structures are likely to be weak, as in the three African cases and Cambodia, the absence of a local coordinating authority makes it difficult to introduce local planning processes. In these circumstances social funds have appropriately started by using the individual-communities model. 9. Where implementing a local planning system is feasible, two conditions were identified as important for achieving greater allocative efficiency: (i) project prioritization should be based on an open menu of choices though it is appropriate for social funds to limit the project-types they financially support, and (ii) safeguards are needed to prevent central and local actors from dominating or unfairly distorting the expression of local preferences. These distortions may be related to the "sponsorship" of local planning, opportunistic behavior of private firms, and/or benefit-capture by dominant local political elites. 10. Incorporating Strategic Planning into Social Fund Operations. Two elements facilitate making good local decisions: first, the possibility of discussing tradeoffs between alternative investments to arrive at the option with the highest benefits; and second, the possibility of integrating investments across sectors and levels of government Supporting good local governance through social funds and decentralization vii to achieve synergistic development gains. These two elements as are collectively referred to as strategic planning. 11. Social funds that do not work through a comprehensive local planning process miss the opportunity to discuss trade-offs among competing investments within a local government jurisdiction. Projects are approved on a case-by-case basis with little or no consideration of broader development plans and priorities. Discussion about investment trade-offs only occurs within a local government jurisdiction at the community level. 12. Besides the opportunity for discussing investment trade-offs, social funds relying on local planning have realized that to improve their development impact it is sometimes advisable to fund packages of integrated sectoral investments rather thanindividual projects (e.g., in the education sector in Bolivia). These integrated investments are selected through consultations with local communities, ensuring that local preferences are taken into account. The bundling of investments generates the economies of scale necessary to justify the cost of technical assistance needed to design strategic interventions. Creating Effective Local Financing Mechanisms 13. Social funds are central government programs that are national in scope and provide local funding through a matching-grant mechanism that earmarks funds for smallscale investments. To understand social funds' contribution to good local governance, the analysis of social-fund financing must be placed in the broader context of an overall system of intergovernmental and local development financing. 14. Balancing National and Local Preferences. National and local preferences will not always coincide. These various preferences can be accommodated within the framework of an overall fiscal system, which can be designed to allocate fiscal resources to different levels of government in line with their duties and responsibilities. But in practice things are often different and there is evidence that the design of fiscal systems is often flawed (Shah 1994). Therefore social funds often have to operate in the context of unbalanced fiscal systems dominated by centrally-controlled earmarked funding and a very low level of untied funding. In all the seven countries studied, social funds represent one of the few central government programs that is intended to provide a consistent source of funding for local investment. However, in doing so they often end up dominating the aggregate pool of resources available for local development. 15. Social Funds as Matching Grants. One type of transitionary mechanism for integrating social funds into a broader fiscal system is for central governments to use social fund resources as matching grants to give an incentive to local governments to invest in central governments' priorities. In Bolivia for example social funds operate in a decentralization context where local governments have considerable own-resources from automatic, revenue-sharing agreements. Through the use of a matching-grant approach, the social fund has successfully redirected local government spending toward the social sectors. Supporting good local governance through social funds and decentralization viii 16. Using social funds and other funds as part of the government's cofinancing policy requires careful design. Two lessons emerge: first, if the supply of matching grants is only for a few sectors local governments with scarce own-revenues will have an incentive to use all their own untied funds as counterpart funding for investments in these sectors, reducing their capacity to respond to other local priorities. And second, when the central government has several sectoral funds providing matching grants to local governments, it seems appropriate that the counterpart rates of these funds should be coordinated to reflect central government's preferences. Otherwise if each fund defines them independently, there is no guarantee that the group of cofinancing rates will be consistent with national government priorities. 17. Matching-grant principles are incorporated into basic social fund design through a "community contribution" which is intended to promote a sense of ownership and help establish a culture of joint cost-sharing. Even in countries with weaker fiscal systems such community contributions have provided good incentives for communities to mobilize additional resources. Where local governments play a more active role in social fund operations local governments are also making contributions. 18. Releasing Funds and Handling Procurement. The practical success or failure of a government program is partly determined by how well implementation is carried out. In the seven countries studied social funds had adopted four basic approaches for disbursing funds and procuring equipment and services (community contracting, centralized contracting and disbursement, joint local government-social fund contracting, and local government contracting) 19. It appears that decentralized contracting can provide greater benefits in terms of effective supervision, accountability, local capacity building, and production efficiency. Decentralized arrangements for contracting, however, should be careful to match project size and complexity to existing local capacities. 20. Once communities and local governments demonstrate their ability to manage financial resources effectively, there is no reason why social-fund resources should not be decentralized to them. In countries with functioning local governments (such as the three Latin American cases), many local governments are already contracting projects of a similar or greater complexity than the ones funded by the social funds. These governments are easy candidates to be responsible for contracting out the services that fall in their mandate. Building Local Capacity 21. Having enough capacity to get the job done is one of the biggest challenges confronting local institutions as well as the managers designing and implementing programs of local development. An important element of all social funds is the emphasis placed on capacity building. Social funds have addressed capacity building in different ways--for some funds it has beenan explicit, central goal of their activities; for others an implicit, secondary goal. Nevertheless, the size of the job of building capacity in social funds pales in comparison to the massive effort needed to build essential capacities in local governments as part of broader decentralization initiatives. Supporting good local governance through social funds and decentralization ix 22. Social funds and decentralization utilize different strategies and approaches for building capacity. On the one hand, social funds focus on quick program start-up with capacity building focused on the specific skills needed to implement social fund activities. While community groups are always at the center of social fund operations, the extent to which local governments participate in social fund operations varies from fund to fund. On the other hand, processes of decentralization tend to concentrate on the gradual assimilation of a comprehensive set of the different skills needed to manage local development. Local government institutions are the focus for learning and training programs associated with decentralization. 23. Helping Local Institutions Build Capacity through Social Funds. Social funds have helped build capacities of many types of local institutions. Community groups have benefited from the training provided by social funds in financial management, chairing and managing project committees, and taking and distributing minutes from committee meetings. Local governments have learned the basics of participatory planning (Honduras) and project supervision (Bolivia). Government staff have also benefited from participating in the systematic project cycle process that is part of every social fund. (Africa) Social funds have injected much needed resources and have helped reinvigorate the operations of local planning committees (Malawi and Zambia). 24. Transferring Social Fund Responsibilities to Local Governments. Successful decentralization of powers and responsibilities for delivering local goods and services to local governments holds out the potential for scaling up the types of goods and services that social funds are currently financing. Social funds are becoming an important instrument of these broader efforts to support decentralized local government. 25. Based on the field work three particular aspects of transferring responsibility appear to have importance. First, social funds have adopted a steady pace and are using gradual approaches to transferring functions, first piloting and learning from experience before expanding to a larger number. In terms of the scope of the functions transferred, funds are using a differentiated approach, assigning different responsibilities according to the capacities of particular local governments. A set of objective indicators of capability or performance can help inform social funds and local governments of their own abilities. 26. Second, social funds are exploring three support mechanisms to enhance the capacities of local governments: (i) providing direct technical assistance to local governments through the social fund's regional offices, NGOs, private sector firms, universities, or other sources; (ii) supporting the formation of associations of municipalities that can collaborate for specific purposes, and (iii) financing a municipal representative whose salary can be progressively taken over by the local government. 27. Third, social funds that transfer responsibilities to local governments will face the challenge of reengineering their organizations from agencies specializing in managing the direct implementation of projects to agencies specializing in supporting the creation of local capacity for local institutions that will take over project implementation responsibilities, as well as monitoring and evaluating implementation performance. Supporting good local governance through social funds and decentralization x Increasing Accountability to Local Institutions 28. Improving accountability, or the "answerability for performance" is an essential part of decentralization. Social funds indicate two different approaches to accountability: one that assigns important responsibilities to local governments as agents for managing social fund resources, and the other in which the role and accountability of local governments is limited with social funds focusing on their own internal system of accountability. A variation on the latter is when social funds deconcentrate some authority to regional offices. 29. In important ways, though not often stated explicitly as an objective, social funds have demonstrated some success in improving local accountability. For instance: · by requiring local governments to prepare development plans using participatory procedures, social funds aim to reorient mayors to pay less attention to their political "clients" and more to broader community needs. (Honduras) · by teaching mayors that if the agreements between levels of governments were not followed, sanctions would be applied (Bolivia). · by openly including communities in the project cycle, they have helped rebuild some community faith in public institutions to provide essential services, and communities appreciate their own role and responsibilities in these activities (African cases). · by using transparent and objective criteria to fund projects, social funds foster accountability by reducing discretion and establishing clear rules of the game that can be enforced. (e.g., the objective, ex-ante allocation formula used in Honduras to allocate funds to municipalities.) · by demonstrating the benefits of the social funds methodology, social funds have encouraged greater responsiveness in some local governments, even in countries where the social fund have minimized their role (Peru). 30. Despite the contributions that social funds make to strengthen accountability, in other ways social funds may act to constrain broader, more institutionalized improvements in accountability. · By excluding elected local governments fromthe planning of investments, some local governments perceived their legitimacy as coordinating agents of local investments eroded. If a local government is perceived as making little difference to communities, they will pay less attention to ensuring local government accountability, giving local governments less incentive to be accountable, and communities more reason to distance themselves from local governments (Peru, Zambia). Ensuring Project Sustainability 31. Real development impact comes only from the sustained provision of basic services over many years. But factors outside the immediate control of social funds or local governments often determine whether or not social fund investments will have a lasting impact. If the components of operation and maintenance (O&M) are Supporting good local governance through social funds and decentralization xi disaggregated, three conditions are necessary to achieve sustainability (rebuilding and upgrading of existing infrastructure; preventive maintenance; and staffing and supplies). 32. The specific division of O&M responsibilities between communities, local governments and line ministries varies across countries, partly depending on the legal and operational circumstances in each sector. Social funds have found it necessary to work with all these actors, at different levels of government and civil society. The degree of sustainability hinges on forging effective collaborative arrangements between these different groups and agencies. The experience of having line ministries take on responsibility for O&M has been mixed in the countries studied: generally more positive in operations than in maintenance. Community involvement and contributions seem to have worked better in Africa than in Latin America. 33. Partly in recognition of local government's contribution to O&M, social funds in Latin America have increasingly been involving local governments in the planning of investment maintenance. Yet local governments are often constrained by limited capacity and resources. So social funds in both Bolivia and Honduras have been implementing an approach to sustainability that includes helping local governments produce maintenance plans in which local governments and communities specify their contribution for the maintenance of all social infrastructure. One innovative option being planned in Honduras is for the social fund to implement a performance monitoring system that rewards local governments that show sustainable finances and good maintenance performance. 34. Two necessary conditions associated with good performance is (i) to involve local governments early in the planning process and in the provision of the investment (though not necessarily in its production); and (ii) to ensure there is an institutional strengthening strategy to help local governments increase their revenue-raising capacity and thus fulfill their maintenance commitments. Although it is not necessarily social funds' responsibility to provide this institutional strengthening themselves, it is in their interest to encourage it to happen. 35. In Africa, few local governments are in a position to take on project management responsibilities for social fund projects within a short space of time (given few or no autonomous sources of own-revenue, poorly functioning councils, etc.) Community groups (those whose members have a common self-interest) stand a far better chance of managing specific individual investment projects. Conclusions 36. Both social funds and decentralization policies aim to improve local governance. Yet operational staff working on related projects have often viewed each other as if from opposite sides of the fence, when in fact they are on opposites sides of the same coin! Both approaches are clearly complementary parts of an over-arching governance agenda that may nevertheless take decades to implement. 37. The core message of this report is that a social fund will maximize its impact on local governance when key decentralization policy reforms are in place and the social Supporting good local governance through social funds and decentralization xii fund is aligned with them. When designing a social fund policymakers should identify social-fund program goals within the context of the constraints imposed by the decentralization framework, and decide how to deal with them. If the social fund does not take into account the decentralization framework it is likely to have negative impacts on local governance. In the absence of a decentralization framework, social funds are unlikely to reach their greatest potential, though ironically they can make an important contribution to jump-starting a broader decentralization process. 38. From an intergovernmental perspective, social funds are earmarked grants intended to increase sustained access of the poor to smallscale infrastructure and services. They operate within the context of a legal, fiscal and political intergovernmental framework. Where that framework includes some form of decentralization, social funds can contribute to local governance if their procedures are consistent and reinforce those of decentralized structure. 39. The difficulty arises because social funds often operate in countries where the intergovernmental framework is undeveloped and unclear at best, and contradictory and ill-defined at worst. Key decentralization policy instruments are also absent. Therefore, social fund designers have to make a decision about how to deal with the deficiencies of the framework and what are the trade-offs that the program will confront. 40. In the absence of a formal decentralization policy, social funds can still play a positive role in enhancing local governance and at the same time be supportive of decentralization initiatives. For example, in countries with little experience of local government and limited exposure to community-based approaches, social funds can act as an important catalyst for change by demonstrating the potential for local institutions to play a larger role in local development. Those social funds that, though being in countries with no decentralization framework, want to support the local governance agenda will face the following challenge: how to gradually build an integrated local government/ administration system that clearly links accountability, capacity, and performance. ------------------------------------- Supporting good local governance through social funds and decentralization 1 INTRODUCTION 1. In recent years, the Bank and other donors have increased their focus on local governance and are now supporting initiatives that allow local institutions to play a much greater role in the planning, design, implementation, and operation and maintenance of smallscale investment projects. This trend is occurring at the same time as many countries are showing renewed interest in processes of decentralization with local governments around the world being assigned roles and responsibilities once controlled by the center. 2. Two important aspects of Bank support for local governance are social funds, which seek to target and empower poor communities to improve participation and local service delivery, and support to governments committed to decentralization of responsibilityand power for local development to local governments and other local institutions. There are concerns that these two approaches, which address different elements of governance, may sometimes work at cross-purposes. For example, if social funds establish parallel channels for local expenditures and community participation without building proper channels for local accountability or financial sustainability, they can weaken nascent local governments and impede decentralization efforts. Conversely, if decentralization programs build financially healthy local governments without provisions for orienting spending to the poor, they will be at odds with social fund efforts of increasing access of the poor to local services and infrastructure. 3. This study was therefore commissioned to examine the interaction between social funds and decentralization in seven countries at different stages of decentralization: Bolivia and Honduras (advanced decentralization), and Peru and Zimbabwe (some decentralization) and Cambodia, Malawi and Zambia (little or no decentralization). It explores the relationships between social funds, decentralization and local governance. Specifically, it seeks: (i) to identify a range of issues for improving local development that are relevant for social funds and decentralization; (ii) to discuss and analyze research findings based on field work in the seven country cases; and (iii) to present conclusions that inform the debate on social funds and decentralization, and provide some guidance to policymakers and program managers on important issues and different policy options. Social Funds 4. Social funds usually have two goals: increasing sustained access of the poor to local services and infrastructure, and empowering communities through participation in the selection, implementation and on-going operation and maintenance of development projects. They target resources by providing direct financing for community projects designed to have a quick impact on improving basic services and reducing poverty. Since 1987, the Bank has approved about 100 social fund-type projects in more than 60 countries with a total value of about $3.4 billion. Though a variety of institutional models have been adopted, social funds have usually been established as specialized units set apart from traditional government line ministries. To promote timely support to communities, funds typically utilize procedures that provide a waiver from standard government procurement and other regulations. Supporting good local governance through social funds and decentralization 2 5. Although the specific objectives of each social fund varies depending on the economic, social and cultural context prevailing in a country at a given time, most social funds share common design elements, including: · A focus on community empowerment · Poverty-targeting mechanisms designed to ensure that poor communities benefit substantially · Direct financing of community projects · Community participation in project selection, design, implementation, and operation and maintenance · A cash or an in-kind community contribution toward project costs · Quick-disbursement procedures that typically bypass government procurement regulations and permit social funds to handle contracting on behalf of communities and local governments · An operational manual that codifies processes and procedures for social fund activities and prescribes roles for the different institutions involved 6. Operating in institutional environments where government agencies had singularly failed to improve the living standards of the poor, social funds have developed mechanisms for channeling resources directly to community groups; have empowered these groups to take on responsibilities for such activities as organizing community labor contributions, procuring goods and services and other project management tasks; and have successfully mobilized both public and private sector institutions in an effort to improve basic services. 7. Resources allocated by social funds to communities have financed thousands of smallscale community-based products, and they have often done so more rapidly than publicly-financed programs. Whether social fund investments have a longer-run effect on development remains unclear, however.1 There remain many challenges for the social funds: how to scale up operations to support the many communities not yet reached? how to ensure sustainability of the services provided by social fund investments? how to make sure a participatory planning process is followed that gives voice to all in a community? how best to integrate social funds into systems of intergovernmental fiscal relations? 8. Social funds were not initially conceived of as having any special relevance for decentralization initiatives largely because few countries in which funds were established had proceeded very far in their efforts to decentralize. Developments during the 1990s, however, especially in Latin America, led to a rapid advancement of decentralization arrangements. As central governments began to reallocate responsibilities for providing basic services to subnational governments, the potential for the community-oriented social fund approach to conflict with constitutionally-mandated changes in local government responsibilities increased. 1 A three-year, multi-country impact evaluation study currently being carried out by the Bank, Social Funds 2000, aims to address this issue. Supporting good local governance through social funds and decentralization 3 Decentralization 9. In contrast to social fund programs, processes of decentralization focus on the formal institutions of government, both political and administrative. Decentralization has come to be widely regarded as an important instrument for supporting the factors needed to create good local governance and promote local development. There is a growing recognition that central governments cannot do it all, and that the active involvement of both communities and representative local governments is required for successful local development. 10. Decentralization is complex and encompasses a wide range of elements (Parker 1995) but evidence from many countries suggests three important conditions are necessary if decentralization is to lead to improved governance (Manor 1999): · Significant powers and responsibilities for local service delivery should be devolved to representative local authorities in line with their capacities (political decentralization); · Sufficient resources, through a combination of local taxes and grants from higher- level governments, must be provided to enable local governments to fulfill their responsibilities (fiscal decentralization); and · Proper channels for accountability are needed to encourage strong accountability between bureaucrats and elected representatives, and between elected representatives and their electorate (institutional decentralization). 11. The political reasons why countries begin to decentralize are complex. Donors usually consider constitutional, political and legal changes in favor of greater decentralization as mainly an internal, domestic issue and not amenable to policy intervention.2 This is not the case with the fiscal and institutional aspects of decentralization, both of which have substantial technical and capacity-related elements that represent an important part of donor support for decentralization. AnalyticalFramework 12. As the persistently large numbers of people living in poverty demonstrate, it has proved difficult for development agencies and governments to implement programs that have a lasting impact on reducing poverty. While there have been notable successes, poor countries still often find it hard to deliver basic services. Many factors determine the ability of local institutions to deliver services. For the purposes of this study we have selected five issues: planning, financing, capacity, accountability and sustainability (Figure 1). Poor communities face constraints and challenges with all of these issues: · local planning is often not very participatory resulting in the rich and powerful often dominating decisionmaking and marginalizing the poor; 2 In Bank's Articles of Agreement, Artic le IV states: "The Bank and its officers shall not interfere in the political affairs of any member; nor shall they be influenced in their decisions by the political character of the member or members concerned. Only economic considerations shall be relevant to their decisions, and these considerations shall be weighed impartially in order to achieve the [Bank's] purposes..." (www.worldbank.org/html/extdr/backgrd/ibrd/arttoc.htm) Supporting good local governance through social funds and decentralization 4 · availability of local financial resources is often limited making it hard to provide a sufficiently high standard of service delivery; · local institutions, especially those located in remoter rural areas, often lack sufficient capacity to be able to plan, implement and operate services effectively; · accountability mechanisms between elected representatives and their electorate and local administrative staff are often weak; and · ensuring the sustainability of local investments is difficult given the shortage of resources and limited capacity. Figure 1: Framework for Analyzing Elements of Local Governance Policy and program interventions designed to improve local governance Effective Participatory financing planning Local Accountability Sustainability mechanisms systems capacity Improved delivery of infrastructure and services Sustainable local development Method and Approach 13. As the purpose of the study is to provide quick feedback on a range of issues facing social funds and decentralization, it was decided to adopt a case study approach in order to compare and contrast different country experiences. Seven countries were selected for study: Bolivia, Honduras, and Peru in Latin America; Malawi, Zambia, and Zimbabwe in Africa; and a commissioned report on Cambodia (Pagaran 1999) (Table 1). 14. The countries were selected to provide a range of experiences in terms of decentralization and local government involvement with social funds. In Latin America, although the three countries have political decentralization,3 functional and fiscal decentralization as well as political support for decentralization policy vary 3 All three have direct elections of mayors and local councilors--Peru since 1985, Bolivia since 1994, and Honduras since 1993 Supporting good local governance through social funds and decentralization 5 considerably,4 with Peru being at the lowest end of decentralization, Honduras in the middle and Bolivia at the highest. In the three Africa countries studied government functions have remained largely centralized, though this is beginning to change. In Malawi long-delayed local elections are now planned for the end of 2000, which will reconstitute district level authorities as elected bodies to whom substantial powers and functions are to be devolved. Zambia is also moving toward empowering their long- standing district councils, which have been sidelined by local line ministries who have most of the budget and authority. While in Zimbabwe the government has announced a recent major initiative to improve arrangements for decentralization that is focused on empowering rural district councils and improving local coordination. In Cambodia, certain responsibilities for local service provision and local administration and some tax and non-tax revenues have been devolved to the provinces under the 1998 Provincial Budget Management Law. Table 1: List of Social Funds Studied Country Name of Social Fund Studied Bolivia Fondo de Inversion Social (FIS)--Social investment fund (World Bank projects 1987, 1991, 1994) Honduras Fondo Hondureño de Inversion Social (FHIS)--Honduran social investment fund (World Bank projects: 1991, 1992, 1995, 1999) Peru Fondo Nacional de Compensacion y Desarrollo Social (FONCODES) (World Bank projects: 1993, 1996) Malawi Malawi Social Action Fund (MASAF) (World Bank projects: 1996, 1998) Zambia Microprojects Unit (MPU)/Zambia Social Investment Fund (ZAMSIF) (World Bank projects: 1991, 1995, 2000) Zimbabwe Social Development Fund (SDF) (World Bank project: 1998) Cambodia Social Fund of the Kingdom of Cambodia (SFKC) (World Bank projects: 1995, 1999) 15. The field work comprised two elements: (i) a series of interviews with key actors working on social funds, central and local government officials, elected representatives, NGOs as well as Bank Task Managers and other staff working on local development programs; and (ii) site visits to social fund projects in the seven countries visited to meet with community members to discuss social fund operations. The field work was supported by a detailed desk review of relevant information and secondary data, including internal government documents, external evaluations, and academic research. 16. For comparative purposes, some alternative approaches to infrastructure and service delivery that focus on support through local government bodies are also examined. The Bank and other donors have a number of program-types that provide direct support for decentralization. These programs have focused on building the capacity 4 Shares of total public spending are approximately 6% in Peru (World Bank 1994), 12.3% in Honduras (IDB 1998), and 23% in Bolivia (WB 1999). The functions assigned to local governments also vary: building and maintenance of mainly education, health, and micro-irrigation in Bolivia; water and sanitation in Honduras; and urban maintenance and trash collection in Peru. Central government commitment to decentralization is relatively high in Bolivia, moderate in Honduras, and quite low in Peru. Supporting good local governance through social funds and decentralization 6 and performance of local governments. In some cases they have also promoted systemic reform of intergovernmental fiscal systems. Local government support programs include municipal development funds, most of which operate in urban areas; sector-specific funds, such as road funds; and capacity building programs, which may include a policy component designed to bring about reform of intergovernmental fiscal systems 17. Municipal Development Funds (MDFs) act as intermediaries to provide credit to local governments and to other institutions investing in local infrastructure. Around the world, more than 50 developing countries have established intermediaries of this kind. The Bank has financed projects in 26 countries since the early 1980s. MDFs have dual objectives. In the short run, they are charged with disbursing publicly provided credit to local government and seeing that local investment is efficiently implemented. In the medium and longer run, they are intended to pave the way for a self-sustaining credit system that can participate in private capital markets. MDFs are designed to remedy some of the typical constraints of local governments (lack of institutional capacity, limited local resource mobilization, and limited access to long-term financing for investment programs) while preparing them to borrow in capital markets.5 18. UNCDF's local development fund (LDF) strategy for supporting (mainly rural) local governments comprises both capacity building and investment financing of local authorities, and through these of communities, using funding mechanisms designed to simulate an unconditional block grant from central government. LDFs aim both for direct impact through smallscale investments, but also serve as pilots to test innovations in the "how-to" of decentralization policy for wider replication. 19. LDFs are designed to work within prevailing statutes, regulations and institutions as much as possible, rather than establishing autonomous Project Units or creating new institutions. Within a pilot area, usually a cluster of Districts, or an entire Region or Province, the funds are designed to test decentralization reforms or innovations. These are usually aimed at: (i) improving mechanisms and rules for central-to-local financial transfers; (ii) establishing procedures for local participatory planning, programming, budgeting, and for implementing local public sector spending; and (iii) instituting structures and procedures through which the state supports and supervises local authorities and encourages their performance. Since the mid-1990s the agency has supported 17 such programs in 14 developing countries in East and West Africa, South and Southeast Asia, and the Middle East. About $150 million has been committed to these funds, which affect some 27 million people. ESTABLISHING PARTICIPATORY PLANNING PROCESSES 20. Historically only a few government programs have tried systematically to incorporate community decisionmaking; most decisions were implemented top-down with national political structures rather than consultation with communities usually determining the type and location of local projects. In contrast, social funds encourage 5 For an evaluation of four Bank projects in Brazil and the Philippines see Lee and Gilbert (1998). For general reviews of MDFs see Peterson and Hammam (1997) and Ferguson (1993). Supporting good local governance through social funds and decentralization 7 communities to participate at many key junctures in the project cycle--i.e., project planning, implementation, and operation and maintenance. One of the main instances of community participation used by social funds has been for project selection. Based on the case-study countries two main approaches were distinguished: · individual communities6 select projects from the social fund's menu through a community assembly meeting, and send these projects to the social fund which generally approves them on a first-come, first-served basis (Cambodia, Malawi, Peru, Zambia and Zimbabwe); · local governments coordinate a systematic local planning process in which each community first selects its priority, and those priorities are then discussed and ranked at the municipal level by local government authorities and community representatives (Bolivia, Honduras).7 The social fund finances the top priorities that fall within the fund's menu of eligible project-types. 21. In this section the two approaches are compared and contrasted and an assessment is made of the factors that are likely to lead to efficiency gains at the same time as promoting enhanced opportunities for community involvement. Allocative Efficiency 22. Through decentralization of project selection to the community level, social funds aim to improve allocative efficiency by delivering a mix of public goods and services that fits local preferences better than centrally-controlled programs.8 An important finding from the field work suggests that social fund investments will have better chances of achieving allocative efficiency if sub-projects come through a comprehensive local planning process rather than from isolated community proposals. 23. From a normative perspective, there are two reasons why having a local planning system increases allocative efficiency: first, it promotes a process by whichall communities express their preferences rather than only a few; and, second, it creates an instance at the local government level where these preferences are discussed and investment options are made, rather than leaving this decision to the social fund headquarters. Since local actors have better information about local needs than central actors, they are better positioned to make choices about local investments. In addition, making public the decision-making process increases accountability as informed local citizens have a basis for assessing local government performance. 6 By "community" we understand the lowest level unit of planning. In urban areas, this could be for example a neighborhood or a self-defined group; in rural areas, this could be a village, a group of hamlets, etc. 7 A variation of this approach is to create municipal participatory committees, formed by community representatives, NGOs, church, local government officials, and usually presided by the mayor. Peru's social fund did a pilot in one region between 1993-4. Other examples are FUMAC and FUMAC-P (northeast Brazil), and Mexico's Municipal Fund (see Wiens and Guadagni 1998; Fox and Aranda 1994). 8 "Allocative efficiency" is an economic concept that refers to the degree of convergence/divergence between the preferences of individual community members and the tax and service packages reflecting aggregate community preferences. The political parallel of allocative efficiency is responsiveness and accountability. In either case, the core concern is with the kind of arrangements that increase the congruence between public preferences and public policy. (Wolman 1990). Supporting good local governance through social funds and decentralization 8 24. From an operational perspective, however, the viability of having an effective local planning system is partly contingent on the country's decentralization framework. Countries with decentralization processes in place, such as Bolivia and Honduras, find it easier to support local planning processes than those without them. In countries where there is no functioning representative local government and administrative structures are likely to be weak, as in the three African cases and Cambodia, the absence of a local coordinating authority makes it difficult to introduce local planning processes. In these circumstances social funds have appropriately started by using the individual- communities model. 25. Where implementing a local planning system is feasible, two important conditions were identified in order to achieve greater allocative efficiency: (i) project prioritization should be based on an open menu of choices rather than on a proscriptive, limited menu; and (ii) safeguards are needed to prevent central and local actors from dominating or unfairly distorting the expression of local preferences. 26. Communities need to express their preferences without being constrained to choose only from the social fund's menu, since that does not allow them to reveal their real preferences. The social fund itself does not need to have an open menu; but citizens need to be able to express their preferences freely (guided by technical assistance). Those projects of the local plan that fall within the social fund's menu can be financed through the social fund. Having an open menu, however, does not mean much without availability of some untied funds to finance those choices. We provide further elaboration and evidence on these points in the section on financing mechanisms below. 27. A second condition is ensuring that neither the central agency promoting the local planning process nor local actors systematically distort the process of expressing preferences and analyzing investment options. On this issue, our fieldwork produced three findings that relate to the "sponsorship" of local planning, opportunistic behavior of private firms, and dealing with benefit-capture by dominant local political elites. 28. First, if the local planning process is sponsored by an entity that has a specific sectoral focus, there is a greater risk of distorting local preferences than if it is sponsored by a more neutral actor. Social funds do have sectoral interests--they have a closed menu and tend to have priorities within that menu (e.g., education or water projects). This sectoral interest may distort local preferences since communities to increase their chance of getting funding may tailor their demands to the priorities of the social fund. In Honduras, for instance, when FHIS organized the Participatory Municipal Development Plans in 1998, the fund had education as one of its priorities. Although in theory communities were not constrained in their options, interviews with social fund staff and other local actors suggested that several communities chose education to increase their chances of getting funding. By contrast, in Bolivia, where the role of sponsoring local planning was played by a sector-neutral agency (the Viceministry of Popular Participation, in charge of supporting decentralization in the country), there appeared to have been no systematic interference in the planning process. Thus, whenever possible, support to local planning processes should be carried out by the agency responsible for strengthening local governments. Supporting good local governance through social funds and decentralization 9 29. Local governments themselves may also have their own sectoral interests, especially if their mandate implies exclusive or shared responsibility in specific sectors. Thus, they may also have an incentive to bias the planning process in favor of its sectoral interests. As with social funds, the challenge is not to negate these interests but to introduce them at the appropriate stage--i.e., not while communities are making their preferences but while local government officials and community representatives elaborate the investment plan. 30. A second source of distortion of local preferences may be motivated by the opportunistic behavior of a broader range of actors (e.g., engineers, architects, and small contracting firms, particularly in Latin America) who by exploiting their control over knowledge and power are able to distort project selection for their own benefit. This is the case for local contractors.. Even though in many cases these actors have played a key role in informing rural communities about social funds, structuring their demand and turning it into projects, in some cases they have distorted local preferences by convincing communities to choose projects that serve their own, rather than the communities' interests. This has been termed "engineer-driven" instead of community-driven funds. For example in both Peru and Honduras, professionals prepare standardized designs for a few project types, which they offer to communities. This allows contractors to cut down on designing costs, but it does not always serve communities' best interests. It is difficult to know the extent of this phenomenon but it was reported as happening in both Peru and Honduras by social fund officials, mayors, and donors' staff involved in supervision missions.9 31. Engineer-driven influences seem more pronounced in social funds where individual communities present projects10 than where there is a local planning process. This was the case in Honduras. In Honduras, there were no reports of engineer-driven cases during the local planning processes of 1998, in contrast to the previous situation where communities could present them individually. The reason was that as part of the local planning process the social fund hired facilitators to assist communities in project selection, neutralizing the influence that private firms had. Since not all local planning processes involve this 32. Finally, a third possible source of distortion of local preferences comes from local elites capturing project benefits. To avoid this problem some social funds (e.g., FONCODES) have excluded local governments from project selection, resisting the idea of using a systematic local planning process. The development literature (Mohan and Stokke 2000) and our fieldwork, confirms that although this is a valid concern, there are alternative ways to deal with benefit-capture rather than seeking to exclude local governments. 33. In the three Latin American country cases complaints were voiced about some mayors manipulating the relationships between communities and the social fund. And in Africa some national MPs appeared to have had undue influence over project selection. 9 It has also been reported in other countries in Latin America. (e.g., Rodriguez (1999) for Central America, and Serrano (1996) for northeast Brazil). 10 This situation is especially pronounced in social funds where the same firm that designs the project is allowed to build it, since it becomes more rewarding for the firm to influence the process. Supporting good local governance through social funds and decentralization 10 Local governments cannot therefore always be relied upon to effectively support and implement participatory planning processes. However at the same time as some politicians in the Latin American cases sought to manipulate decisions, other mayors were facilitating access of all communities to the local planning process and making decisions more transparent. Similarly in Africa some MPs work hard to facilitate social fund operations without seeking to circumvent community choice. The operational task is to understand the conditions under which local governments do and do not act as facilitators of local governance. There is a growing literature on this issue that should help inform project design.11 34. Moreover, attempting to avoid local capture by excluding local governments from the planning process is not effective. Local governments who want to influence community decisions will seek to do so even if they are not part of the formal process of project selection.12 In the case of FONCODES, for instance, many local governments were actively involved in how communities chose investments even though this was supposed to be done by communities on their own. Sometimes, mayors intervened to replace community choice in authoritarian ways. Other times, however, they intervened to enhance it. For instance, some mayors followed a systematic process of organizing meetings in every community to select proposals, and hiring professionals to design them. 35. Rather than excluding local governments, the best approach appears to be to design a local planning system in a way that reduces the opportunities for local capture, e.g., through allocation rules can target poor communities. In Mexico, capping the fraction of resources allowed for town centers (where elites are usually concentrated) is a measure that has been used in a social fund-type program to favor poor areas (see Box 1). Box 1: Giving Incentives for Pro-Poor Investments in Local Planning Processes Mexico's Municipal Fund program was launched with an official discourse that mandated broad community participation in project selection and implementation; government-community co- responsibility in project management; and civil society oversight of the use of project funds. But this was not enough to encourage autonomous, informed participation and power-sharing. To improve the accountability and targeting of resources, additional measures were put in place: · encouraging better anti-poverty targeting within municipalities by capping the fraction allowed for town centers; and · setting a sliding scale for varying levels of community counterpart contributions intended to encourage investments with higher social impacts--participants had to contribute more towards the cost of projects that were either less directly related to fighting poverty or were not public goods These measures not only promoted better anti-poverty targeting, they also encouraged broader participation. For example, if most of the funds had to be spent outside the town center, that measure potentially magnified the voices of the poorer communities that lived outside the town centers. The targeting measure created an entitlement that had often not existed before. This 11 For a review of the literature on local capture see Das Gupta, Grandvoinnet and Romani (2000); Bardhan and Mookherjee (1999). There is also a literature on decentralization and corruption--e.g., Wade (1997), Brueckner (1997), Fisman and Gatti (1999). 12 For a similar case in northeast Brazil see Serrano (1996). Supporting good local governance through social funds and decentralization 11 entitlement in turn created an accountability benchmark that served to encourage participation. Source: Fox (1999). 36. Rather than promoting ad-hoc mechanisms for participatory planning, social funds can encourage the establishment of statutory local planning processes that have transparent mechanisms built in that require local governments to consult local population preferences, and accountability mechanisms by which both communities and central governments can sanction misbehavior. Whether there is an appropriate role for social funds to intervene or not in the statutory processes will depend on the country situation. For example, the social fund in Honduras not only sponsored the local planning process, but it also introduced changes in the statutory planning process to ensure a greater level of participation by community organizations. By contrast in Bolivia the social fund found existing statutory processes satisfactory and did not intervene. In Malawi the social fund has played an instrumental role in introducing an specific procedures for community participation (Box 2). Box 2: Ensuring that local planning processes are participatory Bolivia's Decentralization Law includes mechanisms to reveal local preferences, to incorporate them in the municipal plan, and to hold local governments accountable if they do not. Local governments should have a five-year development plan, based on consultations with all the communities. An oversight committee, made up of community representatives, participates in drafting the annual budget and is empowered to freeze central government transfers to the local government if the annual plan does not incorporate the community's demands. A central government agency (the Viceministry of Popular Participation) assists both local governments and communities in this process. In Honduras, in contrast, although there is a technology to reveal local preferences, there are no mechanisms to hold local governments accountable. Local governments must hold at least five cabildos abiertos (townhall meetings) a year to determine community preferences, but local governments are not bound to include community demands in their investment plans. Between 1994 and 1997 FHIS relied on local governments to organize the cabildos abiertos as the law required. Compliance with the law was poor. Mayors often failed to hold cabildos at all. Where cabildos were held, they were often not representative or had little impact on local government decisions. The Local Government Ministry (Ministerio de Gobernacion) proved ineffective at improving the situation. Frustrated by this situation, FHIS proposed a new mechanism that called for town meetings and systematic consultations with community members and the approval of investment plans by all actors. In Malawi, the social fund begins with the formation of a community project committee (CPC) that is responsible for preparing, managing and supervising project activities and serves as an intermediary between the community as a whole and MASAF. The process beings with an open community meeting to discuss priority needs and problems and to identify a priority project. The meeting is convened by a traditional authority, usually a village headman, who are targeted in MASAF publicity campaigns to raise awareness of its activities and application procedures. Once the community chooses a project, it elects a CPC. An application form is completed that contains details of the project-types and funding requested. It also includes a description of the participatory process followed during preparation. Following project appraisal and approval, MASAF provides four days of training in the district headquarters to members of the CPC. Training covers MASAF procedures, leadership skills, procurement, bookkeeping, accounting and other project management skills (Ayres 1999). Supporting good local governance through social funds and decentralization 12 37. In this section we have focused on the potential efficiency gains that a comprehensive local planning process has over the single-project approach, and on some of the conditions needed to realize these benefits, rather that on the factors that affect community choice13. Good local planning, however, is not only about responding to local preferences. It also entails making technically sound decisions. We turn next to consider this issue. Incorporating Strategic Planning into Social Fund Operations 38. Two elements facilitate making good local decisions: first, the possibility of discussing tradeoffs between alternative investments to arrive at the option with the highest benefits; and second, the possibility of integrating investments across sectors and levels of government to achieve synergistic development gains. We refer to these two elements as strategic planning. 39. Social funds that do not work through a comprehensive local planning process miss the opportunity to discuss trade-offs among competing investments within a local government jurisdiction. Projects are approved on a case-by-case basis with little or no consideration of broader development plans and priorities. Discussion about investment trade-offs only occurs within a local government jurisdiction at the community level. 40. For social funds that rely on local planning processes, discussing trade-offs is easier as there is an opportunity to discuss the comparative merits of different proposals. In practice, however, the political pressures to respond to all communities has often made it difficult to have strategic discussions. When local governments consult communities about investment decisions resources end up being distributed in the most politically neutral manner, which usually means one project per community. Investments that may have a larger developmental impact but directly benefit only some parts of an area tend to be neglected. This has been reported as a tension present in the local planning processes of both Honduras and Bolivia.14 41. In Bolivia, local governments have addressed this tension by introducing a multi- year planning cycle in which one group of communities agrees not to receive any resources for the first year, with a guarantee that they will receive funding in the second or third year. As Fox (1999) points out for the case of Mexico, this generally requires the existence of trust among communities and with the local government, since some communities have to trust that they will be served in the future . 42. The townhall meeting approach that FHIS adopted in Honduras is a useful mechanism for determining the needs of the local population, but it does not facilitate 13 Other studies have focused specifically on how communities make decisions and have showed the complex set of factors that intervene. Zambia's Beneficiary Assessment, for instance, shows that the main reason why 70 percent of the projects were for schools resulted not only from a strong community preferences for education but also from the extended presence of PTAs at the local level, asymmetrical information (most communities heard of SF by word of mouth), influence of counterpart requirements (easier for communities to make bricks than come up with cash), and the technological simplicity of school repair (Owen and van Domelen 1999). 14It has also been the experience of UNDCF's local development funds (UNCDF 1999) and of the Municipal Development Fund in Mexico (Fox 1999). Supporting good local governance through social funds and decentralization 13 discussion of trade-offs and project prioritization. In a meeting with 100 people, it is difficult to have a meaningful debate about the tradeoffs between alternatives. For this reason, a two-stage process, such as that used in Bolivia, seems more effective in which members of the community first express their priorities; then a smaller group of community representatives and local government officials discuss the priorities and drafts an annual plan. This plan can be presented to the local population in a townhall meeting. 43. Finally, a key condition to enable strategic discussions is that local governments operate with some kind of hard budget constraint that forces local actors to face the real revenue situations and plan spending commitments accordingly. This is also important to avoid creating false expectations among the local population and to encourage local revenue mobilization. The fund that has worked more consistently in this direction is FHIS in Honduras, which has an ex-ante allocation mechanism that informs the local population the amount of resources it will receive in a three-year period. In Zimbabwe following an intensive analytical effort to rank all of the country's 57 districts according to their poverty status, it was agreed to target social fund resources at the 26 most disadvantaged rural districts with resources being earmarked for individual districts in differing proportions based on a coefficient that combines the district's level of deprivation and population size. Allocations are announced in advance enabling local councils and social fund staff better opportunity to plan with communities. 44. Besides the opportunity for discussing investment trade-offs, social funds relying on local planning have realized that to improve their development impact, it is sometimes advisable to fund packages of integrated sectoral investments rather than individual projects (Box3). These integrated investments are selected through consultations with local communities, ensuring that local preferences are taken into account. The bundling of investments generates the economies of scale necessary to justify the cost of technical assistance needed to design strategic interventions. Box 3: Funding education in Bolivia In 1999, FIS started to allocate funds in the education sector through a "Funds Contest Modality" that integrates pedagogical reforms with infrastructure investments. It allocates resources to municipal governments and educational nucleos integrated into the central government's Curriculum Transformation Program. The goal is to generate a new logic for educational management that grants a higher priority to longer-term planning, management autonomy and investment sustainability, thus avoiding the traditional case-by-case response to scattered demands often adopted by social funds. Groups of municipalities will contest every year for education investment resources (4 contests in total--one per year). The rules and criteria for the bidding and allocation system are made explicit. Local governments receive technical assistance to prepare their bids. Besides taking into consideration positive discrimination criteria for the poorer municipalities and ceilings for each Municipality, the process established priorities for integral and long term planning that are reflected in the Municipal Educational Development Program and in the Educational Projects for nucleos. This allows municipalities to have greater or lesser access to resources depending on the technical quality of their bids. Supporting good local governance through social funds and decentralization 14 45. Another approach seeks to integrate planning across different levels of government. To prepare a strategic local development plan, local actors need to access information on programs and resources managed by sectoral and other central agencies. This implies the creation of a supra-local forum where local governments and sectoral agencies can exchange information and agree on funding for local investments. In Peru, the central government piloted such a forum through a program called PROFINES (Proyecto para la Focalizacion de la Inversion Economica y Social). The program aimed to match local demand--identified through participatory municipal development plans-- with a range of funding sources, including FONCODES. Under the PROFINES program, central government agencies cofinanced projects with provincial and district municipalities. The program has allowed long-desired projects that exceeded the funding ceilings of individual programs to be funded. It has also reduced duplication between programs funding similar investments (PROPAL-PASA 1999). Unfortunately, despite its positive results the program is being scaled-down due to the lack of central government support for decentralization. CREATING EFFECTIVE LOCAL FINANCING MECHANISMS 46. Social funds are central government programs that are national in scope and provide local funding through a matching-grant mechanism that earmarks funds for smallscale investments (most commonly social sector infrastructure). To understand its contribution to good local governance, the analysis of social-fund financing must be placed in the broader context of the overall system of intergovernmental and local development financing. The effectiveness of this system lies upon the achievement of three main goals: (i) achieving a balance between the funding of local and national preferences; (ii) mobilizing local resources to strengthen local accountability and increase financial sustainability; and (iii) designing disbursement and contracting channels that establish clear accountability mechanisms and encourage productive efficiency. Balancing National and Local Preferences 47. National and local preferences will not always coincide. There maybe situations where the central government has no interest in satisfying what it deems to be unreasonable requests from the local level, for example the repeated reconstruction of local infrastructure. On the other hand there may be important initiatives that central government feels strongly about but which local groups do not prioritize because they are unaware of the potential benefits (environmental projects), or because of cultural sensitivities (HIV/AIDS programs). These various preferences can be accommodated within the framework of an overall fiscal system, which can be designed to allocate fiscal resources (in the form of grants or taxing authority) to different levels of government in line with their duties and responsibilities. The aim of the system is to give sufficient resources for local governments to have autonomy over local decisions (through untied block grants) while providing incentives for local governments to access or utilize resources that are earmarked for specific purposes by central governments. This system provides for a more balanced set of local investment programs that reflect both local and national preferences. Supporting good local governance through social funds and decentralization 15 48. In practice things are often different and there is evidence that the design of fiscal systems is often flawed (Shah 1994). Therefore social funds often have to operate in the context of unbalanced fiscal systems dominated by centrally-controlled earmarked funding and a very low level of untied funding. In all the seven countries studied, social funds represent one of the few central government programs that is intended to provide a consistent source of funding for local investment. However, in doing so they often end up dominating the aggregate pool of resources available for local development. This is particularly the case in local governments with scarce own-revenues. In the poorest municipalities in Honduras, for example, FHIS funding represents about 80 percent of total local government investment, leaving local government limited autonomy to respond to local priorities not covered by FHIS (Nufio de Figueroa 1996). 49. Because of the restrictions on which projects are eligible for social fund support important community needs (i.e., those not covered in the social fund's menu) can go unmet. In Malawi and Zambia, for example, more than three-quarters of the social fund projects have been for primary school construction and rehabilitation. While the capital stock of primary schools in undoubtedly run down in both countries, it seems unlikely that schools represented the most pressing development need in every community. 50. A recent report from Nicaragua provides illustrative evidence of how the balance of funding affects local population preferences (Romeo 1999). In the same municipality, the same NGO organized two local planning exercises using an identical approach. One, in 1998, was sponsored by the social fund (FISE) whose funding was limited to social infrastructure. Another one, in 1999, was sponsored by a demand-driven rural investment fund (PROTIERRA), whose funding was mostly untied. Despite their differences in funding, both funds used the same participatory approachwhich stressed that the local population should express freely their needs and preferences. However, results differed markedly. In the initiative promoted by the social fund, 70 percent of the value of investments prioritized were in the social infrastructure sector. In contrast, in the one promoted by the rural fund, social infrastructure constituted only 26 percent.15 This confirms that it is not sufficient to tell the local population to openly indicate their preferences if there is not at least some untied resources available to satisfy them. (Romeo 1999). 51. Just as having funding dominated by earmarked grants may bias local preferences, it is also not undesirable when all local funding is untied. UNCDF's local development funds (LDFs) in Cambodia, Malawi and Zambia are experimenting with the provision of untied block grants to local governments. Their experience suggests that given a budget constraint and a consensual local needs identification and prioritization process local governments tend to give priority to proposals with broad local benefits, with a focus on basic social facilities and services. Other types of investment will therefore not tend to emerge from a funding model based entirely on local discretion either. To address other concerns it is necessary to introduce selective earmarked financing mechanisms to address these concerns (UNCDF 1999, Shotton 1999). 15 The rest was for natural resource and environmental protection (20%), local economic development (18%), municipal services and infrastructure (15%), economic infrastructure (14%), organization and training (7%). Supporting good local governance through social funds and decentralization 16 52. It is interesting to note that identical projects can be requested by communities regardless of the participation mechanism used. Communities in LDF pilot areas in Cambodia, Malawi and Zambia tended to prioritize the same type of smallscale infrastructure projects as were selected through the social fund process of participation. With two possible sources of funding (the social fund and the LDF block grant) available to finance the same type of project, there is potential for the two sources to overlap leading to a duplication of effort that wastes time and dissipates scarce localcapacity. Social Funds as Matching Grants 53. One type of transitionary mechanism for integrating social funds into a broader fiscal system is for central governments to use social fund resources as matching grants to give an incentive to local governments to invest in central governments' priorities. In Bolivia for example social funds operate in a decentralization context where local governments have considerable own-resources from automatic, revenue-sharing agreements. Through the use of a matching-grant approach, the social fund in Bolivia for example, has successfully redirected local government spending toward the social sectors (Box 4). Box 4: Meeting central government objectives with matching grants in Bolivia After Bolivia adopted its decentralization policy, the Fondo de Inversion Social (FIS) made municipal counterpart financing mandatory for all FIS projects. The measure aimed: (i) to give local governments incentives to invest in central government priorities by providing matching funds; (ii) to increase local government's sense of ownership over the infrastructure financed, thereby enhancing the prospects for adequate operation and maintenance; and (iii) to provide local governments with incentives to administer transfers more efficiently and generate new sources of revenue. Cofinancing varied by sector and, in the case of formal education, it also varied by size of municipality. FIS used counterpart funds to cover administrative expenses, new investments and to provide counterpart funds for very poor municipalities. Through this matching grant program, FIS was able to redirect local government investment toward the social sectors. For each dollar of FIS transfers, local governments invested an average of $0.55 per capita in social sectors (Gray Molina and Molina Ballon 1998). 54. Using social funds, and other funds, as part of the government's cofinancing policy requires careful design. Two lessons come from the case of Bolivia. First, if the supply of matching grants is only for a few sectors (e.g., education and health infrastructure), local governments with scarce own-revenues will have an incentive to use all their own untied funds as counterpart funding for investments in these sectors (even if they are not a priority), reducing their capacity to respond to other local priorities. For these governments, the possibility of leveraging two or three times their own untied sources through a matching grant represents a very strong incentive to trade-off autonomy for greater resources. In Bolivia, for example, despite strong community Supporting good local governance through social funds and decentralization 17 preferences for productive projects, social sector projects received twice as much funding as productive projects.16 This mismatch was even greater for poor local governments. 55. Second, when the central government has several sectoral funds providing matching grants to local governments (one fund for social infrastructure, another for rural development, and so on), it seems appropriate that the counterpart rates of these funds should be coordinated to reflect central government's preferences. For local governments counterpart rates are the implicit prices they face to execute different types of investments. If each fund defines them independently, there is no guarantee that the group of cofinancing rates will be consistent with national government priorities. Also, this can establish an unproductive competition between funds, each trying to lower the cofinancing rate to attract more municipal demand. In Bolivia, for instance, a local government had to contribute only 10 percent to access a matching grant for a rural development project (provided by the Small-Farmers Rural Development Program), and up to 50 percent to access a matching grant for a school (provided by the social fund). Negotiations between the different development funds are taking place to address this issue. 56. Matching-grant principles are incorporated into basic social fund design through a "community contribution" which is intended to promote a sense of ownership and help establish a culture of joint cost-sharing. Even in countries with weaker fiscal systems such community contributions have provided good incentives for communities to mobilize additional resources. Although the minimum level of contribution requested varies across countries, from 25 percent in Zambia and Malawi to 10 percent in Peru, it is not uncommon for communities to contribute more than the minimum. Villagers have been willing to contribute to the financing of a project as long as they see the need for the project and they are kept informed about the terms and applications of their financial contributions. Where local governments play a more active role in social fund operations local governments are also making contributions (in Bolivia and on a pilot basis in Honduras). For example, since 1996 local governments in Bolivia have been required to contribute at least 35 percent of project costs. Releasing Funds and Handling Procurement 57. The practical success or failure of a government program is partly determined by how well implementation is carried out. Channels for disbursing funds and procuring equipment and services are needed to encourage the appropriate use of funds and to promote openness and accountability. In the seven countries studied social funds had adopted three basic approaches. 58. The first model, used in the three African countries and in Peru, is community contracting. Community Project Committees (CPCs) are responsible for coordinating and managing procurement activities with the social fund officer responsible for monitoring the contracts entered into by communities and providing assistance when requested. The 16 In Bolivia, there was a rural development fund (Fondo de Desarrollo Campesino) that could have complemented FIS by providing grants for rural development; however, it did not function effectively for several years due to administrative problems and thus was not attractive to local governments. Since 1999, an administrative reorganization has helped the fund gain credibility as a co-financier. Supporting good local governance through social funds and decentralization 18 social fund channels resources into community bank accounts set up for the project being financed. The committee is granted temporary legal status that allows it to receive and disburse public funds. Access to these accounts is usually controlled by requiring that checks be jointly signed by a social fund project committee member and the local social fund officer. Community contracting often has a project ceiling--in Peru, for instance, it is limited to projects costing less than $60,000. 59. A second model, represented by the social fund of Honduras, is centralized contracting and disbursement. Even though local governments are allowed some participation--for projects under $50,000 mayors can propose a contractor from a pool certified by FHIS--the key feature is that the contractor is hired by the social fund who is also responsible for supervising and disbursing funds directly to private contractors. Contractors tend to see themselves as primarily accountable to the social fund rather than to the communities they are intended to serve. 60. A third model in Bolivia, grants additional autonomy to local governments, with private contractors selected through a bidding process done by the social fund (with the local government present at the opening of bids). The fund disburses resources to a municipal bank account into which the local government must deposit its cofinancing share. Supervision, and thus payment authorization, is done by a social fund supervisor. 61. An emerging fourth model is being piloted in Bolivia and Honduras where full devolution of contracting responsibilities is being granted to a group of local governments. The next phase of the Zambia social fund is also planning a dual disbursement system in which resources for projects that benefit more than one or two communities would go to district councils, while resources that benefit individual communities would continue to go directly to communities. The option of allocating community resources through local governments is also being considered through a system of graduated responsibilities based on demonstrated performance. Innovations in other types of local funds are also being piloted that have relevance for social fund design (Box 5). Box 5: Decentralizing project management in Cambodia In the Local Development Fund (LDF) in Cambodia, the management of the LDF has been decentralized to provincial and commune levels. Funds, whether they originate from the national budget or external sources, are released directly to a provincial account and managed by a provincial rural development committee (PRDC) through its Executive Committee. The PRDC, which is chaired by the provincial governor and whose members include heads of deconcentrated line departments and agencies, releases funds to commune development committees (CDCs) after review and approval of CDC project proposals or if requested by the CDCs it pays contractors and suppliers directly. Under this arrangement, CDCs are responsible for procurement with assistance from the district/provincial technical support team. Source: Pagaran (1999) 62. In comparing the models in our country cases it appears that decentralized contracting can provide greater benefits in terms of effective supervision, accountability, local capacity building, and production efficiency. Supervising thousands of projects dispersed across the country is less effective when done only by the fund's supervisors Supporting good local governance through social funds and decentralization 19 (centralized contracting), than when done both by the fund's supervisors together with local actors (decentralized contracting). Funds with centralized contracting usually find it difficult to find supervisors willing to go to remote areas, where projects are often located, and the few located in those areas are overburdened by projects. In Honduras, for instance, although FHIS had the appropriate number of supervisors for its portfolio at the national level, the distribution of projects was very uneven across supervisors, with some having over 25 projects and others with only one,17 suggesting that certain areas were not well supervised (Walker et al 1999). 63. The advantage of decentralized contracting is that over time it enables local actors to become effective supervisors of contractors, complementing social fund's supervision. Without local contracting, even if social funds promote community oversight of firms work, communities lack leverage over the firm, and thus contractors have no incentive to be accountable to them. This was largely the case in Honduras with the use of community overseers (contralores sociales). 18 64. One of the concerns with community contracting is the possibility of mismanagement. In Peru, however, not only did few problems occur in this regard,19 but on the contrary there was an empowering effect on communities, many of whom gained new capacities for funds and project management. When procurement was mishandled, it tended to occur in regions where socio-cultural control mechanisms were weak20 or where contractors had tried to influence the local committee. To minimize these problems, social fund supervisors oversee expenditures and have pressed charges against people caught mishandling funds. The same was true in the Cambodian social fund where there have been few cases of mismanaged projects. This was primarily due to the regular presence of project monitors, including members from the provincial/ district technical team and village technical subcommittee, and villagers themselves. 65. In terms of production efficiency, social funds that decentralize subproject contracting to local governments and/or communities appear able to mobilize more local resources than those with centralized contracting. FHIS found, for example, that a local government in Honduras could build three schools with the money that FHIS used to build one. Similar anecdotal examples were cited by social fund staff and local government officials in other countries. Local governments and communities appear able to implement projects more cost effectively than centralized social funds because they can use local materials and mobilize voluntary nonqualified labor. Decentralized 17 Part of the explanation for this huge differentials has to do with computing projects that are finished but have not been turned over to the owners. 18 Evidence from the Honduras impact evaluation suggests that it is not just involving local actors what can increase supervision effectiveness but also involving the future owners of the facility in supervision and project reception. Of the five sectors where FHIS has invested in, the best supervision by far was achieved in sewerage, the only case where the sectoral central government agency was responsible for supervision (Walker et al 1999). 19 A similar outcome was found in evaluations of other funds relying on community contracting-- e.g., Argentina's FOPAR (WB 1998). 20 For example, mismanagement of procurement has been more common in the Costa region, where communities are heterogeneous and social bonds are weak, than in Cuzco, where communities are cohesive and tied by ancestral family bonds. Supporting good local governance through social funds and decentralization 20 arrangements for contracting, however, should be careful to match project size and complexity to existing local capacities. 66. Once communities and local governments demonstrate their ability to manage financial resources effectively, there is no reason why social-fund resources should not be decentralized to them. In countries with functioning local governments (such as the three Latin American cases), many local governments are already contracting projects of a similar or greater complexity than the ones funded by the social funds. These governments are easy candidates to be responsible for contracting out the services that fall in their mandate. 67. Even though local governments have a preeminent role in terms of planning and coordinating local investments, the same is not true for producing those investments. There is a variety of local actors that can handle this: community organizations, NGOs, local governments, private contractors. Which arrangement is chosen will depend on the type of good involved, as well as the performance that these actors have shown in specific countries. In certain sectors, such as water supplies for dispersed rural communities, there is a strong rationale for an active community role in producing the service. In other sectors, economies of scale or externalities may call for local governments , central line ministry, other actors to produce them. BUILDING LOCAL CAPACITY 68. From an operational perspective, having enough capacity to get the job done is one of the biggest challenges confronting local institutions as well as the managers designing and implementing programs of local development. Technical, managerial, financial and organizational skills are all needed to ensure adequate performance and sustainability over the longer term. Yet much of this knowledge is concentrated in urban areas leaving poorer more remote regions with an even greater obstacle to overcome. 69. Many individual donor programs and projects have sought to build capacity in a wide range of institutions. However, earlier donor-supported interventions tended to assume that communities themselves did not posses the requisite skills to manage and implement local development programs. Most of the "hard" aspects of project implementation (design, contracting, payments, record-keeping, etc.) were assigned to program management units of one type or other, leaving only a marginal role for community groups. Increasingly, however, development practitioners have come to appreciate the critical importance of involving communities directly in decisions about programs that operate at the community level. At the same time it is recognized that poor communities will typically lack many of the formal professional and technical skills necessary to manage and operate local services, and an explicit effort is needed to overcome these shortcomings. 70. An important element of all social funds is the emphasis placed on capacity building. This not an easy challenge to address as social funds rely on a plurality of local institutions (not only local governments and communities but also local contractors and NGOs) covering a variety of activities (identifying projects, co-financing, contracting, Supporting good local governance through social funds and decentralization 21 etc.). Social funds have addressed capacity building in different ways--for some funds it has been an explicit, central goal of their activities; for others an implicit, secondary goal. For example, one of the main goals of Peru's fund was "building local capacity in project planning, execution, operations and maintenance of small scale infrastructure and public services" (World Bank 1996). In Honduras, in contrast, this does not appear among the development objectives (see World Bank 1998). 71. Nevertheless, the size of the job of building capacity in social funds pales in comparison to the massive effort needed to build essential capacities in local governments as part of broader decentralization initiatives. One reason UNCDF's local development funds support pilot activities is the need to test different local modalities before scaling up to the national level. Of course this experimentation takes time to design and implement, such that decentralization programs can often take many years to implement, e.g., the latest Bank project for local government support in Zimbabwe envisages a 12-year time-span of support through an Adjustable Lending Program (APL)21. 72. This section explores these two different approaches to building local capacity. On the one hand social funds focus on quick program start-up with capacity building focused on the specific skills needed to implement social fund activities. While community groups are always at the center of social fund operations, the extent to which local governments (both the political and administrative structures) participate in social fund operations varies from fund to fund. We investigate these issues. 73. On the other hand processes of decentralization tend to concentrate on the gradual assimilation of a comprehensive set of the different skills needed to manage local development. Local government institutions (again both political and administrative) are the focus for learning and training programs associated with decentralization. Where appropriate, part of the responsibility that local governments take on has included playing a greater role in social funds activities. We investigate some of the issues that are emerging from such an approach. Helping Local Institutions Build Capacity through Social Funds 74. Field studies in the seven country cases reveal that social funds have helped build capacities of many types of local institutions. Of course community groups are the primary focus of social funds. Many community committees visited during the field work said that they had benefited from the training provided by social funds, and some indicated their intention to apply skills learnt in financial management, chairing and managing project committees, and taking and distributing minutes from committee meetings to other local development programs. The need to persist continually with capacity building programs is indicated by the case of Peru, where community groups 21 The notion underlying an APL is that some development processes cannot be addressed in the context of a single investment operation and may take many years to resolve. The APL provides for funding of a long-term development program starting with a first set of activities designated according to agreed milestones and benchmarks for realizing the program's objectives. Supporting good local governance through social funds and decentralization 22 risk limiting their possibilities of learning new skills when contractors usurp project functions the community is intended to handle. (Box 6) Box 6: The need for persistence: empowering communities in Pe ru According to the FONCODES operations manual Community Project Committees (CPCs) in Peru are supposed to handle the management and disbursement of project funds. However, CPCs may not have understood the procedures correctly; they may feel intimidated by an "expert" project engineer; or they may lack the necessary skills to do their part, e.g., for accounting or procurement. The outcome is that contractors have often taken on the responsibility for managing funds that was not envisaged. One CPC treasurer indicated that: "The engineer didn't allow me to handle the money. For me, it's fine because he is the engineer...he knows about this. I wasn't afraid he would steal because the supervisor was checking on him." Although the engineer may not steal, the potential for strengthening community capacity for procurement and funds management is hampered. A more systematic training of the CPC (Community Project Committee) on project funds management is needed to give the necessary tools and help build confidence for communities to be more demanding of contractors and to help balance the potential asymmetry of the contractor-CPC relationship. FONCODES does hold a two-hour meeting with the CPC to go over the basics of the implementation process, but this is probably insufficient to address this problem. 75. Local governments in many of the countries studied have also benefited from training provided by social funds. Local governments in Honduras, for instance, learned the basics of participatory planning through the mass consultations that FHIS sponsored in 1998, and in Bolivia FIS requests local governments to appoint a municipal supervisor to accompany the FIS supervisor to learn on the job22. 76. Government staff have also benefited from participating in the systematic project cycle process that is part of every social fund. This has been especially true in Africa where district-level planning structures had witnessed their modus vivendi downgraded from having decisionmaking authority to the more limited role of coordination for implementation. Social funds have injected much needed resources and have helped reinvigorate the operations of local planning committees, as in Malawi and Zambia. Now, instead of meeting only to coordinate field trips and the implementation of higher-level decisions, sub-committees of the District Executive Committee in Malawi and the District Development Coordinating Committee in Zambia both have been given important project appraisal functions that result in the recommendation of projects for funding to the social fund head office. These committees have provided a forum for joint learning that has benefited both local government and social fund staff. 77. Some political structures also received training and information from the social funds studied though with differing approaches (Box 7). 22 The effectiveness of this arrangement might be improved if FIS supervisors were required to train the municipal staff as at present the successful transfer of knowledge depends on the good will of the FIS supervisor, who is not formally obligated to train the municipal supervisor. Supporting good local governance through social funds and decentralization 23 Box 7: Contrasting roles for national politicians in Malawi and Zambia In Malawi local political representatives are not democratically -elected and often lack credibility and legitimacy. In order to launch the social fund on a nationwide basis it was judged necessary for the fund to utilize national Members of Parliament as facilitators in mobilizing communities to participate in social fund operations. In many constituencies members confined their role to facilitation, though there were cases reported of where members had overstepped their responsibilities and either influenced or possibly determined the outcome of local participatory processes. By contrast in Zambia, the social fund has adopted a lower-key approach and generally used its own officers or consultants to conduct most of the capacity building activities sometimes together with district line ministry staff. It was judged that the direct involvement of central politicians at an early stage might undermine the program from the start. Transferring Social Fund Responsibilities to Local Governments 78. Successful decentralization of powers and responsibilities for delivering local goods and services to local governments holds out the potential for scaling up the types of goods and services that social funds are currently financing. Social funds are becoming an important instrument of these broader efforts to support decentralized local government. The social fund approach to building capacity is expanding in recognition of the need to target a broader range of local institutions. For example, a new social fund project under preparation in Zambia aims to give full responsibility for supporting community projects to district councils that meet certain minimum performance criteria. And in Zimbabwe district councils already have the power to vote on social fund projects as part of the budgets they approve. In Latin America some social funds are deepening the shift towards including local governments. The funds from Bolivia and Honduras, which have made progress by involving local governments in planning investments, have recently started to address the more difficult issue of transferring responsibility for project design, implementation and supervision. 79. As more countries move to initiate decentralization programs (Cambodia, Malawi and Zambia are all reviewing decentralization arrangements), social funds will need to be prepared for the changing local institutional environment. Where it is deemed appropriate--and this will definitely not always be the case given the constraints discussed above--responsibility for many aspects of social fund operations can be transferred to local governments. Based on the field work three particular aspects of transferring responsibility appear to have importance: (i) the pace and scope of the transfer; (ii) institutional strengthening of local governments; and (iii) organizational reforms needed in the social fund to support this process. 80. The social funds studied are using gradual approaches to transferring functions, first piloting new arrangements in a limited number of local governments, as a learning experience before expanding to a larger number. In terms of the scope of the functions transferred, funds are using a differentiated approach, assigning different responsibilities according to the capacities of particular local governments. However, the problem of this Supporting good local governance through social funds and decentralization 24 gradual and differentiated strategy is that social funds wishing to retain control may deem local governments as never "ready" to assume responsibilities. 81. A set of objective indicators of capability or performance can help inform social funds and local governments of their own abilities. This is being tried in Honduras, for example, where FHIS has launched a pilot that will support a longer-term strategy of transfer to all local governments throughout the country. After classifying local governments into four categories--based on their level of fiscal revenues, administrative capacities, and experience with participatory management--FHIS will test different modalities of transfer and support with a group of 30 local governments representative of the four categories. It is hoped that the pilot will allow FHIS to identify which functions can be transferred to which type of governments and how to help them acquire the capacities to move from one category to the next. The most recent social fund project in Zambia defines five different levels of access to resources based on a specified set of minimum access conditions and performance criteria. 82. A second strategic element involves how to enhance the capacities of local governments to assume their new responsibilities. Social funds are exploring three support mechanisms: (i) providing direct technical assistance to local governments through the social fund's regional offices, NGOs, private sector firms, universities, or other sources; (ii) supporting the formation of associations of municipalities, such as those springing up in Bolivia and gaining importance in Honduras, that can collaborate for specific purposes, such as the collective purchase of equipment for road maintenance or hiring a technical team for pre-investment designs; and (iii) financing a municipal representative whose salary can be progressively taken over by the local government. In Zimbabwe, for example, the social fund officer, known as the poverty alleviation program officer, is hired as a member of a district council's technical staff. 83. Social funds are taking an instrumental approach to capacitybuilding, teaching procedures that are mainly applicable only to social fund operations--e.g., the design specifications for different type of projects. This limited focus seems appropriate since most funds are not agencies specialized in local government strengthening. Nor should they become one. Though social funds may take a more instrumental approach to capacity building, decentralization policies take a more systematic and comprehensive approach to building the capacity of local governments to be more responsive to communities; improve revenue generation and tax collection capacity; create or improve cadasters; improve administrative and budgeting procedures; manage and sustain projects after they are implemented; and so forth. Even if these functions are beyond the responsibility of social funds, it is in their best interest that they take place since they affect the attainment of social fund's goals--e.g., the financial sustainability for social funds projects depends partly on the local government's financial capacity. 84. Both LDFs and MDFs have adopted strategies to support decentralization and help strengthen local governments' capacities. The main approach has been to use access to grants and loans as incentives for local governments to adopt better practices or achieve improved financial performance. Interestingly, this is the same approach being pursued by those social funds (e.g., in El Salvador) that have redefined their core mission Supporting good local governance through social funds and decentralization 25 as the promotion of local development. This suggests that some funds may be "mutating" from a social fund into an MDF. 85. The third and final strategic element involves the organizational changes social funds need to make to be prepared to work with local governments. Social funds that transfer responsibilities to local governments will face the challenge of reengineering their organizations from agencies specializing in managing the direct implementation of projects to agencies specializing in supporting the creation of local capacity for local institutions that will take over project implementation responsibilities, as well as monitoring and evaluating implementation performance. INCREASING ACCOUNTABILITY IN LOCAL INSTITUTIONS 86. Improving accountability, or the "answerability for performance" (Romzek 2000: 22), is an essential part of decentralization. Successful decentralization requires mechanisms for ensuring that local governments are held accountable to their electorate, and also that local bureaucrats, be they local officers of central line ministries or local government technical staff, are accountable to elected local governments (Manor 1999). In practice local accountability mechanisms are often ineffectual in poorer countries, a product of overly-centralized systems that force accountability to a weak center that usually cannot cope. Accountability mechanisms can also be overridden by strong political forces or be subject to subversion by other powerful elites (Mohan and Stokke 2000). 87. Establishing a proper system of accountability for decentralized institutions to each of their different constituents is an important part of decentralization, together with some system of sanctions that penalizes institutions if they fail to carry out their functions appropriately. As Fox and Aranda observe from experience in Mexico: "...authoritarian and/or bureaucratic concentrations of power at local and state levels can prevent decentralization from leading to increased public-sector efficiency and accountability" (Fox and Aranda 1996). Where a clear chain of command is absent, as defined by a concrete set of accountability rules and procedures, the legitimacy of a decentralization initiative can be quickly lost.23 Social Fund Approaches to Accountability 88. Institutional arrangements for social funds indicate two different approaches to accountability: one that assigns important responsibilities to local governments as agents for managing social fund resources, and the other in which the role and accountability of local governments is limited with social funds focusing on their own internal system of 23 Crook and Manor (1994) cite examples from Ghana, where District Administrators were able to retain effective power because of their ties to the center, and local scandals involving embezzlement of public funds went unpunished. Also in Bangladesh, because sub-district council chairpersons derived most of their power from the central regime in Dhaka, they were permitted to do more or less as they wished and were able to make considerable private gain. Supporting good local governance through social funds and decentralization 26 accountability. A variation on the latter is when social funds deconcentrate some authority to regional offices. In practice most funds lie somewhere between these two contrasting approaches. 89. Each approach has its own challenges. The approach that gives responsibilities to local governments relies on being able to reward good or sanction poor performance. Without clear agreements and appropriate incentives local governments have little reason to comply with social fund guidelines and may instead look for ways to divert resources or avoid their responsibilities, for example for operation and maintenance. This approach also relies on local governments by having the capacity to implement social fund operations, which can be difficult for newly-established local governments trying to meet many simultaneous obligations. 90. In the second approach, social funds operate their own internal system of accountability. This usually ties community groups to social funds whose staff act as facilitators between communities and those responsible for taking decisions about social fund projects. In Malawi, Zambia and Zimbabwe local social fund officers participate in the various local planning committees and encourage line-ministry staff to take an active interest in supervising social fund investments. This has certainly helped to give communities a sense of empowerment but there is a potential future danger that the direct relationship between local bureaucrats and community groups acts to circumvent the authority of local, democratically-elected governments. Also, if there is an increase in the scale of operations, local offices are likely to be unable to cope with the demand. Hence the importance of local councilors and local technical staff getting together with communities to plan an annual set of activities in an integrated manner, including resources earmarked for social fund projects. 91. A variation on this approach is when social funds deconcentrate authority to regional offices, as in Peru. Although the purpose is to increase accountability by being closer to the local population, the main difference with devolution is that it does not establish institutional checks on government action (Haggard 1995). That is why in a devolved system citizens can sanction elected officials through their vote, but in a deconcentrated system communities lack mechanisms to sanction social fund and line ministry staff. Deconcentration also tends to encourage growth in the social fund bureaucracy and the proliferation of local offices established to deal with growing demand, and can also permit more discretional treatment from the center than devolution. 92. As Romzek points out accountability relationships are built up over time--some formal, some informal--but the "...result is the weaving of a thick web of multiple, overlapping accountability relationships within which public administrators work" (Romzek 2000:23). Establishing the net effect of social funds on accountability is therefore difficult to ascertain. In the following sub-sections some of the factors that create favorable or unfavorable conditions for improving accountability are discussed. Favorable Conditions for Improving Accountability 93. In important ways, though not often stated explicitly as an objective, social funds have demonstrated some success in improving local accountability. There are many Supporting good local governance through social funds and decentralization 27 examples of where social funds have helped strengthen accountability by devolving decisionmaking and resources to both communities and local governments. In Honduras for example, FHIS has required local governments to prepare development plans using participatory procedures. These plans, discussed in town hall debates, specify how social fund resources will be spent and identify the criteria for allocating funds. It is hoped that with the encouragement of FHIS it may succeed in reorienting mayors so they pay less attention to their political "clients" and more concern to broader community needs. 94. In Bolivia FIS taught mayors that if the agreements between levels of governments were not followed, sanctions would be applied. Political instability was so great in Bolivia that municipalities were often headed by three different mayors within a single year. Social fund agreements reached with one mayor were not honored by a successor, who feared that funding previously approved projects would benefit his predecessor. To solve the problem, social fund representatives made it clear that they would cut off future financing unless agreements were kept. 95. In the African cases, communities make their own choices over many elements of the project cycle including project selection, procurement, contracting and so on. This openness of operations appears to have helped rebuild some community faith in public institutions to provide essential services, and has helped communities appreciate their own role and responsibilities in these activities. Because social funds sometimes play a coordinating role at the local level they have, within the limits of their defined sphere of operation, successfully brought together community groups, local government staff and line ministry staff in ways that were not common before. The selection of a social fund project provides a tangible point of intervention around which many local institutions have been mobilized. 96. Donor-supported programs have historically often been implemented directly by central line ministries, with the location of investments being predetermined. This is particularly applies to the district-level planning structures in Africa (discussed in the section on capacity). By providing a concrete program of investments that involves district line ministry staff and assigns specific roles to district staff, social funds have realized improvements in accountability. For example in Zambia many proposals are often initially rejected by the district Planning Sub-committee; however, rather than just returning submissions to communities, social fund and line ministry staff use the opportunity as a learning experience to teach and work with community committees on how to develop a proper, well-justified proposal, helping to rebuild trust between communities and government staff. 97. If correctly implemented decentralization can also help improve local accountability by bringing all government technical staff working in a local government jurisdiction--that is, local staff and staff that report to line ministries--under the control of a single local executive who reports to the local council. Doing so is difficult, however, because of the differences between the two groups. Line ministry staff usually have university degrees; local technical staff are more likely to be graduates of technical colleges. District councilors may be illiterate and lack knowledge of the processes of government. Supporting good local governance through social funds and decentralization 28 98. Just as the central government should be able to hold local governments and communities accountable, local governments and communities should have mechanisms for holding social funds accountable. Using transparent criteria to fund projects fosters accountability by reducing discretion and establishing clear rules of the game that can be enforced. An example of this is the objective, ex-ante formula that FHIS uses to allocate funds to municipalities in Honduras. In Zimbabwe social fund resources are allocated only to the poorest group of districts based on various indicators of poverty. And the next phase of the Zambia social fund will allocate resources to individual districts by providing an indicative planning figure based on an agreed formula. 99. We found that even if a social fund does not work through local governments, as in Peru, it may have positive if unintended effects on the accountability of particular local governments by encouraging greater responsiveness to community needs (Box 8). Box 8: Creating a culture of responsiveness to community needs in Peru By demonstrating that communities are capable of selecting and implementing investments, FONCODES was able to promote a culture of responsiveness and accountability to communities. Encouraged by the approach and performance of FONCODES some more progressive mayors sought to find ways of adopting the social fund's methodology by allocating their own revenues through community selection of projects and implementing them through community procurement. One mayor justified community procurement by saying, "If [the community] is handling the money, then they can't tell me I'm stealing, as they did with the previous mayor. Now the local government will control that communities don't steal." Some municipalities have even established small FONCODES-type municipal funds with their investment resources (Guerrero 1998). 100. Finally, the relationship between social funds and local governments is not one- way, and local governments have found that they can be in a stronger position to work with social funds when they work collectively, as they have in Honduras through an association of local governments, the Asociacion de Municipalidades de Honduras (AMHON). This association has played an important role in bringing problems in project selection or implementation to the attention of FHIS. It has notified FHIS, for example, when mayors were incorrectly prohibited from selecting contractors or were not consulted in the selection of the project, though FHIS rules stipulate that they should. AMHON has been successful partly because it has a significant and sustainable source of funding through membership dues of 1 percent of each municipality's current income. Unfavorable Conditions for Improving Accountability 101. Despite the contributions that social funds make to strengthen accountability, in other ways social funds may act to constrain broader, more institutionalized improvements in accountability. Social funds can sometimes perpetuate dependency on the center by centralizing power and excluding or giving insufficient voice to elected local institutions. For instance, the social fund in Peru minimized the role of local governments in its operations, despite the fact that they were the elected local authority. As a result, some local governments perceived their legitimacy as coordinating agents of local investments eroded, especially in municipalities that were not able to find ways of influencing the allocation of FONCODES resources, or where the value of social fund Supporting good local governance through social funds and decentralization 29 investments was significantly greater than that of local governments. If a local government is perceived as making little difference to communities, they will pay less attention to ensuring local government accountability, giving local governments less incentive to be accountable, and communities more reason to distance themselves from local governments. 102. Maintaining a direct center-community relationship for social funds enables central governments to retain some discretion over the use of resources that undermines accountability. In Peru for example, social fund expenditures rose significantly before two recent presidential elections and projects were funded in provinces in which the political impact was likely to be greatest (Schady 1998), though similar effects of political influence were not found in Bolivia (Graham 1992). Close scrutiny of accountability mechanisms is therefore important to assess the feasibility of having local governments manage the relatively large pool of social fund and other potential locally- available resources. 103. Similarly to Peru, in Zambia the elected district councils have been largely excluded from decisionmaking process concerning the allocation of social fund resources, and their technical staff have not been closely involved in project implementation. In this case, the social fund approach runs the risk of establishing a process in which real decisionmaking authority lies with technical staff of central line ministries, rather than supporting the established democratic (albeit weak) institutions of the district councils. The next phase of the Zambia social fund (ZAMSIF) plans to address this problem by establishing a graduated system based on performance that will grant greater autonomy to better performing districts. ENSURING PROJECT SUSTAINABILITY 104. Real development impact comes only from the sustained provision of basic services over many years. But many factors outside the immediate control of social funds or local governments can determine whether or not social fund investments will have a lasting impact. If the components of operation and maintenance (O&M) are disaggregated, three conditions are necessary to achieve sustainability. First, existing infrastructure needs rebuilding and upgrading--social funds focus on the financing and processes related to this condition. Second, infrastructure financed by social funds requires preventive maintenance (painting, minor structural repairs, etc.)--communities and sometimes local governments are expected to mobilize sufficient resources to cover these costs. And third, newly completed facilities require adequate staffing and supplies--this is usually the responsibility of the relevant line ministry or other regional authority. 105. The specific division of O&M responsibilities between communities, local governments and line ministries varies across countries, partly depending on the legal and operational circumstances in each sector. Social funds have found it necessary to work with all these actors, at different levels of government and civil society. The degree of sustainability achieved in different social funds therefore hinges on forging effective Supporting good local governance through social funds and decentralization 30 collaborative arrangements between these different groups and agencies.24 Based on the range of different options available and significant variations in local environmental and institutional conditions it is not surprising to find that different solutions for addressing the challenge of sustainability have been tried in the seven countries studied. Tackling Sustainability in Social Funds 106. Social funds in Latin America have progressed significantly in their treatment of sustainability. The early social funds were not so concerned with the sustainability issue as there was a substantial emphasis place on creating of short-term, emergency employment opportunities for the poor. However as the funds' mission evolved to include longer-term development goals, sustainability gained a more prominent place in their agenda. 107. In Africa, for the two social funds (Malawi and Zambia25) that have been in existence long enough to have some experience of the sustainability issue, the picture is broadly similar. In both countries community groups have overwhelmingly selected primary school projects as their preferred choice of project-type, and both have adopted similar arrangements for O&M responsibilities once classrooms have been completed. As part of their institutional arrangements all schools have a parent-teachers association (PTA), which has a range of responsibilities. Of course the extent to which these PTAs function effectively varies greatly from community to community. Nevertheless, one important responsibility they have is to collect schools fees. Minutes from PTA meetings in Zambia indicate that fee mobilization can become a contentious issue between those who pay and those who do not. With a high rate of poverty there are many households for whom even the lowest fee imposes a heavy burden and who may therefore not be able to pay. But there are those who try to avoid payment and free-ride on others' contributions. Bearing these constraints in mind the finding that over 60 percent of parents pay fees for school maintenance on social fund projects in Zambia (Nkhata et al. 1999), would appear to represent quite a success and indicative of community willingness to financially support important local infrastructure. This relationship between perceived benefits and sustainability is also borne out by the social fund in Cambodia (Box 9). Box 9: Improving the chances for sustainability in Cambodia Several factors were identified as affecting project sustainability in the Social Fund of the Kingdom of Cambodia (SFKC), including participation, benefits and mechanisms for maintenance. Projects funded under the first phase of the social fund were selected and implemented with limited community participation in decision-making and in local contributions. Control over these projects by local communities was compromised, e.g., the marginalization and weakening of local institutions, such as the pagoda committees or Village Development Committees which generally organize and mobilize community support for maintenance, contributed to lack of sustainability of some SFKC projects. Lack of established mechanisms to follow-up after the handover, of training maintenance committees, and of maintenance plan guidelines also contribute to difficulties in sustaining some SFKC projects. 24 There is contradictory evidence as to the sustainability of social fund investments. While data from OED finds sustainability low, many individual beneficiary assessments carried out for social funds have indicated high levels of community satisfaction with social fund projects. 25 The social fund in Zimbabwe only started operations in 1999. Supporting good local governance through social funds and decentralization 31 Despite these problems, some projects have been maintained by communities because they have produced benefits for them. The direct relation between benefits and sustainability has been particularly evident for wells and school building projects. Impressionistic evidence suggests that communities do maintain projects if they respond to their needs and produce benefits, despite a lack of participation by beneficiaries in decision-making. Contrary evidence indicates that participation is not a sufficient for maintaining community support. Even when communities have participated in decisionmaking but projects do not provide expected benefits or satisfactory outputs (such as badly built schools or non-functioning wells), villagers subsequently withdraw from further contributions or involvement. SFKC is addressing the issue of sustainability on several fronts, at both organizational and policy levels. At the organizational level, SFKC has established a Sustainability Unit under the Office of the General Director. It develops strategies and work plans on participatory methods for project selection and implementation, and coordinates implementation among various SFKC staff and institutional partners. There is also a plan to establish three regional offices to decentralize and implement activities of the Sustainability Unit. At the policy level, the project cycle is being modified to incorporate several new steps including pre-appraisal, ground-breaking, project launch, and hand-over to ensure greater community awareness and participation by beneficiaries as well as feedback and interaction from SFKC representatives. Training and capacity building of SFKC staff and counterparts are envisioned under the second phase of the fund on various aspects of participatory methods including project selection, appraisal and monitoring. Source: Pagaran (1999) 108. The experience of having line ministries take on responsibility for O&M has been mixed: generally more positive in operations than in maintenance. In all the cases reviewed, social funds have had formal mechanisms requiring line ministries to "sign off" on O&M commitments. Staffing and supplies (operations) is a central government responsibility in all of the cases studied. Preliminary results of recent impact evaluations show that social funds-financed facilities are staffed and functioning, have higher utilization rates, and have better equipment. (World Bank 2000). Line ministries have not been so effective in ensuring that maintenance is provided, however, and ministries have often failed to fulfill their commitments. 109. Community contributions for maintenance helped but did not fully address the problem. Only 20 percent of project beneficiaries in Peru were estimated as paying the monthly fees intended to cover services for potable water, irrigation, sewerage and waste disposal, and secondary rural electrification (Instituto Apoyo 1999). And even when beneficiaries do contribute, the amounts paid often do not cover all recurrent costs, e.g., 83 percent of beneficiaries in Honduras said that user fees did not cover O&M costs of water projects (Walker et al. 1999).26 Supporting Decentralized Operation and Maintenance 110. Partly in recognition of local government's contribution to O&M, social funds have increasingly been involving local governments in the planning of investment maintenance. In Honduras, for instance, social fund's studies showed that "in many cases, municipalities have already been finding ways to pay for routine maintenance expenditures [of FHIS investments] and even operating costs such as staff salaries, using 26A similar situation, though, was also found in comparable programs in the country. Supporting good local governance through social funds and decentralization 32 resource transfers and other local revenues" (World Bank 1998).27 Encouraged by this finding, FHIS has been implementing a comprehensive approach to sustainability that includes helping local governments produce maintenance plans in which local governments and communities specify their contribution for the maintenance of all social infrastructure. FHIS has also worked with the Comptroller's Office so that the municipal budgeting process includes O&M expenses associated with FHIS-supported projects. Bolivia's social fund also requires municipalities to include the O&M expenses of all their projects in their annual plans and budgets. FIS has warned communities that it may not fund future projects if they do not comply. 111. One innovative option being planned in Honduras is for the social fund to implement a performance monitoring system that rewards local governments that show sustainable finances and good maintenance performance. The incentive system will reward local governments that fulfill a set of good maintenance practices with access to a Preventive Maintenance Incentives Fund. These rewards are needed, at least in the medium term, since local governments have strong political incentives to spend their resources in new investments rather than O&M. 112. Social funds that have been working with local governments (such as FHIS) have learnt that if they expect local governments to be committed to maintenance they should involve them early in the planning process and in the provision of the investment, but not necessarily in its production. However, involving them from the start is not enough. A second condition is to ensure there is an institutional strengthening strategy to help local governments increase their revenue-raising capacity and thus fulfill their maintenance commitments. This is an area whose importance social funds are beginning to recognize and deal with. As noted in the capacity section although it is not necessarily social funds' responsibility to provide this institutional strengthening themselves, it is in their interest to encourage the implementation of a program or instrument within a country's decentralization policy that is assigned this responsibility. 113. Few local governments in Africa have the same degree of constitutional autonomy as compared to local governments in Latin America. Local district councils have few (Zimbabwe) or no (Malawi and Zambia) autonomous sources of own-revenue. While all major towns have some sort of elected district or municipal council taking political decisions, many remoter rural areas have poorly functioning councils. Anecdotal evidence from Zambia suggests that many councils do not even have sufficient resources to pay council members' sitting allowances. Under these circumstances it is hard to think of these types of local institutions as being in a position to take on project management responsibilities for social fund projects within a short space of time. However, community groups (those whose members have a common self-interest) stand a far better chance of managing specific individual investment projects. Hence the generally better O&M performance of social fund projects over other centrally-controlled programs in both Malawi and Zambia, provided projects are perceived as meeting a felt need. 27Similar studies in Nicaragua produced the same findings (WB 1999). Supporting good local governance through social funds and decentralization 33 CONCLUSIONS 114. Both social funds and decentralization policies aim to improve local governance. Yet operational staff working on related projects have often viewed each other as if from opposite sides of the fence, when in fact they are on opposites sides of the same coin! Both approaches are clearly complementary parts of an over-arching governance agenda that may nevertheless take many decades to implement. This study has sought to address some of the critical questions concerning: (i) the appropriate division of responsibilities between social fund mechanisms and support for decentralized institutions and over what timeframe; (ii) how to realize important synergies of poverty-targeted, community-based approaches with the efficiency gains generated through decentralized implementation; and (iii) the main contradictions to avoid. There is no clear consensus or unambiguous evidence to answer these questions. However, a collaborative, team-based approach that seeks to build bridges between would appear to hold out the best prospects for success. 115. The core message of this report is that: a social fund will maximize its impact on local governance when key decentralization policy reforms are in place and the social fund is aligned with them. When designing a social fund policymakers should identify social-fund program goals within the context of the constraints imposed by the decentralization framework, and decide how to deal with them. If the social fund does not take into account the decentralization framework it is likely to have negative impacts on local governance. In the absence of a decentralization framework, social funds are unlikely to reach their greatest potential, though ironically they can make an important contribution to jump-starting a broader decentralization process. 116. From an intergovernmental perspective, social funds are earmarked grants intended to increase sustained access of the poor to smallscale infrastructure and services. They operate within the context of a legal, fiscal and political intergovernmental framework. Where that framework includes some form of decentralization, social funds can contribute to local governance if their procedures are consistent and reinforce those of decentralized structure . For instance, if there are locally-elected authorities responsible for coordinating local investment planning, social funds can finance projects that emerge from local planning exercises. Not doing so will risk undermining the legitimacy of local governments and, thus, the possibilities for strengthening accountability between a local government and its population. 117. The difficulty arises because social funds often operate in countries where the intergovernmental framework is undeveloped and unclear at best, and contradictory and ill-defined at worst. Key decentralization policy instruments are also absent. Therefore, social fund designers have to make a decision about how to deal with the deficiencies of the framework and what are the trade-offs that the program will confront. 118. For instance, in order to maximize the allocative efficiency of their investments social funds require a country context where there is a balance between untied and earmarked financing for local investments. Often, however, this is not the case. A social fund may therefore be tempted to open up its menu and allow financing in sectors that it may have no established comparative advantage in supporting. Or it may decide to stick with a closed menu of selected national priorities losing some allocative efficiency in the Supporting good local governance through social funds and decentralization 34 short-run but ensuring that central government addresses deficiencies in the fiscal framework over the medium term. 119. In the absence of a formal decentralization policy, social funds can still play a positive role in enhancing local governance and at the same time be supportive of decentralization initiatives. They may even be successful in stimulating decentralization. For example, in countries with little experience of local government and limited exposure to community-based approaches, social funds can act as an important catalyst for change by demonstrating the potential for local institutions to play a larger role in local development. In Bolivia, FIS influenced the emergence of the 1994 decentralization law by demonstrating to the central government that a demand-based policy was feasible. In a country with a history of strong centralization and few channels of communication between central and local decisionmakers, policymakers were very skeptical about the transfer of decisionmaking power that the new law involved. The social funds' experience helped allay policymakers' fears by showing the success of the participatory approach in delivering basic services. 120. Those social funds that, though being in countries with no decentralization framework, want to support the local governance agenda will face the following challenge: how to gradually build an integrated local government/administration system that clearly links accountability, capacity, and performance. What is the appropriate strategy to pursue? Where to start? Should the focus be on project delivery or building governance reforms? Under what circumstances can they be built together, and when might tradeoffs need to be made? How can the strategy build on pilot, start-up activities, and when and how to scale up and introduce more sophisticated reforms? 121. Unfortunately strategies for implementation has been a neglected area of research.28 UNCDF has been piloting options for implementing decentralization for the ground-up through a variety of Local Development Funds it has supported in low-income countries (Box 10). Box 10: A strategic approach to implementation: the Local Development Fund in Cambodia The LDF in Cambodia had as its first priority to build political credibility at the village level. Thus, even though it was recognized that the "one community-one project" approach was problematic (since the smaller village projects might be less well implemented and have more modest development impact than commune-wide projects), it was selected as the starting-point to build the needed credibility. Also, early experiments were menu-driven (rather than with an open menu, which would be more consistent with a local development focus) to make sure that villages would choose and benefit from basic-needs projects that required only simple skills and technology to implement. The menu-driven approach also facilitated the implementation process- -local units were given clearly defined pre-disbursement requirements--in terms of adoption of new procedures, technical and cost standards, local consultation, etc. If they met these requirements in preparing the project, the project was "automatically" approved, facilitating the appraisal process and placing well-defined limits on the extent to which higher-level institutions could intervene ex-ante. 28 For two recent exceptions see Bahl (1999) and Smoke (1999). Supporting good local governance through social funds and decentralization 35 There were of course problems with this system, but in retrospect it appears to have been an appropriate way to start. Over time, however, as credibility and capacity gradually developed and people came to think beyond their immediate local needs, communes have begun selecting fewer village projects, concentrating on larger, more complex projects that benefit development in several villages or the commune as a whole. Some village level projects are still important (both for their inherent benefits and out of political necessity), and they continue to be funded. This has been made easier by the strategic introduction of a multi-year budget allocation, so that villages know that a project will benefit them sometime in the next few years. They are willing to wait as long as there is a reasonable timeframe that is made explicit. At the same time, the limited menu used in the initial process has been relaxed: the local planning process has been broadened to include planning for all funding sources, including requests to provincial sectoral ministries; and the system has been moving from ex-ante approval of specific projects to ex-post technical and financial audits with penalties against future allocations for non- compliance. Most of these evolutionary steps were consciously designed into the development of the decentralized system. Source: Smoke (personal communication). Supporting good local governance through social funds and decentralization 36 REFERENCES (1996). Staff Appraisal Report. Peru. Second Social Development and Compensation Fund Project. Washington DC: The World Bank. (1998b). Project Appraisal Document. Argentina: Fourth Social Protection Project. Washington DC: The World Bank (2000). "Report on Social Funds 2000 Research Project," mimeo. Ayres, Wendy (1999) Participation Process Review Case Study: Malawi--Social Action Fund Project. Operations Evaluation Department, The World Bank. Bahl, Roy (1999). Implementation Rules for Fiscal Decentralization. International Studies Program. Working Paper #10. Atlanta: Georgia State University: Bardhan and Mookherjee (1999). "Capture and Governance at Local and National Levels," Draft. Brueckner (1999). "Fiscal Decentralization in LDCs: The Effects of Local Corruption and Tax Evasion," Department of Economics, University of Illinois at Urbana- Champaign, mimeo. Crook, Richard and James Manor (1994). Enhancing Participation and Institutional performance: democratic decentralization in South Asia and West Africa, London: Overseas Development Administration. Das Gupta, Monica, Grandvoinnet and Romani (2000). "State-Community Synergies in Development: laying the basis for collective action," Washington DC: The World Bank, mimeo. De Jager, Patricia (1998) "Social Investment Funds, Local Governments, and Communities in Central America" in Bigio, Anthony (ed.). Social funds and Reaching the Poor. Experiences and Future Directions. Washington DC: The World Bank. Ferguson, Bruce (1993). The Design of Municipal Development Funds. In Review of Urban and Regional Development Studies No. 5. Fisman and Gatti (1999). "Decentralization and Corruption: Evidence Across Countries," Washington DC: The World Bank, mimeo. Fizbein, Ariel (1997). "The Emergence of Local Capacity: Lessons from Colombia," World Development vol. 25, no. 7. Fox, Jonathan (1999). "The Inter-Dependence Between Citizen Participation and Institutional Accountability: Lessons from Mexico's Rural Municipal Funds," in Katherine Bain, ed., Thinking Out Loud: Innovative Case Studies on Participation Instruments Washington: World Bank. Fox, Jonhatan and Josefina Aranda (1996). Decentralization and Rural Development in Mexico: Community Participation in Oaxaca's Municipal Funds Program (La Jolla: University of California, San Diego, Center for US-Mexican Studies. Graham, Carol (1992) "Politics of protecting the poor during adjustment: Bolivia's Emergency Social Fund" World Development 20:1233-1251. Supporting good local governance through social funds and decentralization 37 Gray Molina, George y Molina Ballon, Gerardo (1998). "La Programacion de la Inversion Publica Municipal: Adonde Fueron los Recursos?" in Gray Molina, G. (ed.) Participacion Popular: Diseno y Gestion. Construyendo Politicas Publicas Locales en Bolivia. La Paz: UDAPSO-HIID. Guerrero, Luis (1998) "Exposicion Inaugural. Seminario de Cajamarca" in Garcia de Chu, I. and M. del C. Piazza (eds.) Sociedad y Gobierno Local. Espacios de Concertacion y Democracia. Lima: DESCO. Haggard, Stephan (1995). "State Reform in Latin America" in Burki, Shahid et al (eds) Annual World Bank Conference on Development in Latin America and the Caribbean-1995. Washington DC: The World Bank. Husain (1998) in Bigio, Anthony (ed.). Social funds and Reaching the Poor. Experiences and Future Directions. Washington DC: The World Bank. Instituto Apoyo (1999). Quinta Evaluacion Ex-post de FONCODES. Lima: Instituto Apoyo. Inter American Development Bank (1998). IDBAmerica. January-February Issue. Lee, Kyu Sik and Roy Gilbert (1998). Building Institutions and Financing Local Development: Lessons from Brazil and the Philippines. Impact Evaluation Report. Report No. 18727. Washington DC: The World Bank. Manor (1999). The Political Economy of Democratic Decentralization. Washington DC: The World Bank. Mohan, Giles and Kristian Stokke (2000). "Participatory development and empowerment: the dangers of localism" Third World Quarterly 21 (2): 247-268. Nufio de Figueroa, Cristiana (1996). "Participacion Futura de las Municipalidades en los Proyectos Financiados por el FHIS. Perspectiva de las Municipalidades," Paper presented at the Regional Seminar: Rol Futuro de los Fondos de Inversion Social. Tegucigalpa. Owen and Julia Van Domelen (1998). Getting an Earful: A Review of Beneficiary Assessments of Social Funds. Washington DC: The World Bank. Pagaran, Lourdes (1999). "The Social Fund, the Local Development Fund, and Decentralization: The Case of Cambodia," mimeo. Parker, Andrew (1995). Decentralization: The Way Forward for Rural Development? Policy Research Working Paper No. 1475. Washington DC: The World Bank. Parker, Andrew (1995). Decentralization: The Way Forward for Rural Development? Policy Research Working Paper No. 1475. Washington: The World Bank. Peterson, George with Sonia Hammam (1997). Building Local Credit Systems. Urban Management Program Policy Paper. Washington DC. August. PROPAL-PASA (1999). Estrategia Focalizada de Lucha contra la Pobreza Extrema. Sistematizacion. Lima: PROPAL-GTZ, PASA-Union Europea. PROPAL-PASA (1999). Estrategia Focalizada de Lucha contra la Pobreza Extrema. Sistematizacion. Lima: PROPAL-GTZ, PASA-Union Europea. Supporting good local governance through social funds and decentralization 38 Rodriguez, Rene (1999). Presentation at the VI Annual Conference of the Soc\ial Network of Latin America and the Caribbean. Tlaxcala. Romeo, Leonardo (1998) Decentralized Development Planning. Issues and early lessons from UNCDF-supported LDF Programmes. New York: UNCDF. Romeo, Leonardo (1999). Informe de Evaluación de la Microplanificación Participativa. FISE-PROTIERRA-INIFOM. Draft, mimeo. Romzek, Barbara (2000). "Dynamics of public sector accountability in an era of reform" Internation Review of Administrative Sciences 66: 21-44. Schady, Norbert (1998). Seeking Votes: The Political Economy of Expenditures by the Peruvian Social Fund (FONCODES), 1991-1995. Washington DC: Princeton University and the World Bank. Serrano, Rodrigo (1996). Who Knows What's Best for the Poor? Demand-Driven Policies and Rural Poverty in Northeast Brazil. Master's Thesis. Dept. of Urban Studies and Planning, Cambridge, MIT. Shah, A (1994). The Reform of Intergovernmental Fiscal Relations in Developing and Emering Market Economies Policy Research Series No. 23. Washington DC: World Bank Smoke, Paul (1999). "Rebuilding Local Government in Kenya," Department of Urban Studies and Planning, Massachusetts Institute of Technology. Thorp, Rosemary (1995). Challenges for Peace. Towards Sustainable Social Development in Peru. Report of the Pilot Mission on Socio-Economic Reform of the Inter-American Development Bank. Washington DC: IDB. UNCDF (1999). Taking Risks. New York: UNCDF. Wade (1997). "How Infrastructure Agencies Motivate Staff: Canal Irrigation in India and the Republic of Korea," in Ashoka Mody, ed. Infrastructure Strategies in East Asia. Washington DC: World Bank. Walker, Ian et al (1999). Evaluacion Ex-Post del Fondo Hondureno de Inversion Social (FHIS 2). Report prepared for the World Bank. Tegucigalpa : ESA consultores. Weins, Thomas and Maurizio Guadagni (1998). Designing Rules for Demand-Driven rural Investment Funds. The Latin American Experience. Washington DC: The World Bank. Wolman (1990) World Bank (1994). Peru. Public Expenditure Review. Washington DC: World Bank. World Bank (1998). PAD- Participatory Rural Investment Project. Washington DC: World Bank World Bank (1998). FOPAR World Bank (1998). PAD Education Quality and Equity Strengthening Project. Washington DC: The World Bank World Bank (1998). Project Appraisal Document. Republic of Honduras. Fourth Social Investment Fund Project. Washington DC: The World Bank.