COMMUNITY DRIVEN DEVELOPMENT: A VISION OF POVERTY REDUCTION THROUGH EMPOWERMENT COMMUNITY DRIVEN DEVELOPMENT IN AFRICA: A Vision of Poverty Reduction through Empowerment Foreword want to share with you an exciting and, in my view, important journey that we in the I Africa Region of the World Bank are initiating, but which we can not even begin but together with you. Quite simply, the vision is of Community Driven Development (CDD) - a key means of help- ing to achieve the development objectives that are becoming apparent through the Poverty Reduction Strategy Paper (PRSP) process. Across many countries, the resilience and determination of communities continues to inspire. The intent of CDD is to help empower local communities in Africa to shape their future by giving them more resources and the authority to use these resources to improve their standards of living. Empowering communities is a smart and dignified way to go, and is an integral part of effective poverty reduction strategies. Indeed, many projects we and other donors finance are already moving in this direction. But still too few resources filter down to the communities. CDD is potentially an important part of the emerging architecture for development sup- port. As PRS Papers highlight, poverty reduction is the leitmotif for prioritizing develop- ment assistance. Further, recognizing that participation is a crucial buttress of sustainability, PRS Papers are intended as the outcome of a country-driven process of extended in-country discussion on priority challenges and actions. PRS Papers will open the door to more coher- ent support for poverty reduction by the World Bank and other donors - some as debt relief for heavily indebted countries, some via Poverty Reduction Support Credits and some as more direct resource transfers to communities and local governments. The attached vision statement focuses on this last mechanism. It details the ways in which resource trans- fers to communities - and support for Community Driven Development (CDD) more broadly - can be a powerful contributor to the objective of participatory poverty reduction. To date, the World Bank has made a variety of contributions to local empowerment, but not in an integrated manner, not as a central focus of its strategy. Yet the Bank's compara- tive advantage is precisely that it can systematically link CDD to overall fiscal and state 3 reform issues. It can broaden the debate on decentralization. And it can assist in scaling up successful approaches. The long-term agenda is huge; no one (including the Bank) can make much progress alone. We are eager to work with others, many of whom have a longer track record in this area. It will not always be easy to reconcile th e need for speed with the need for new Jaws, institutions and processes. Initiating a dialog with countries on CDD and assisting participa- tion and institutional development can be a long, gradual process. The level and speed will differ widely from country to country, depending on its history and initial conditions. Our aim in issuing this vision statement is to deepen the dialog on CDD among all our development partners-national governments, other donors, NGOs and community groups. We are eager to find common ground, and are eager for the opportunity to further discuss the potential and challenges of CDD with you and your staff and constituencies in the near future. Callisto Madavo, Vice President, Africa Region, The World Bank WHY COMMUNITY EMPOWERMENT? The Vision Poverty has remained stubbornly high in Africa for decades. Top-down plans and donor- driven investment programs have been less than successful. We need not just a new approach but a new vision: a vision of prosperity through the empowerment of local communities. The starting point must be to empower communities by giving them more resources and authority to use these fl exibly. But this process cannot depend forever on emergency funds or short-lived donor programs: it must at some point be embed- ded in a permanent institutional structure. This can be provided by local governments. Past experience suggests that decentralization will not work without vibrant, participa- tory communities. And enhanced participation will at some point need a local govern- ment structure for sustainability. The two can evolve together dynamically, strengthening one another. The new vision seeks to put local governments and rural and urban communities in the driver's seat, and give them a new set of powers, rights and obligations. These include: • The right to be treated as people with capabilities, not objects of pity. • The power to plan, implement and maintain projects to serve their felt needs. • The right to hold politicians and officials accountable. • The power to command local bureaucrats instead of being supplicants. • The power to hire, pay and discipline all who provide them with frontline local services like education, health , municipal and agricultural services. • The right to a share of central government revenue. • The power to levy user charges and local taxes. 4 • The obligation to enable women, ethnic minorities, the poorest, and other long excluded groups to participate fully in economic development. • The obligation to be accountable to local people, not just central governments or donors. Every country will also need good macroeconomic and sectoral policies to acceler- ate economic growth, which is vital for reducing poverty. Many sectoral programs will continue to be carried out by national governments, especially those involving economies of scale and externalities. Community-driven development (CDD) will complement these other efforts. Communities can be geographical entities (urban neighborhoods, villages) or groups with common interests ( water users associations, parent-teacher associations, herders, members of a microcredit society, women's groups). CDD can harness the considerable social capital in these communities. Donors can support a country's efforts to decentralize but cannot impose it from outside. This is an attempt to move political, administrative and fiscal power from the top to the bottom. Such change will take time and effort - 10 to 15 years or more. What is Poverty? Many people think of poverty as simply a lack of income. Others extend the concept to lack of education and health facili ties. However, as highligh ted in the 2000 World Development Report, Attacking Poverty, economists like Nobel prize-winner Amartya Sen now emphasize a much broader approach. Poverty is also • lack of voice: people need avenues to express their n eeds or obtain redress. • lack of empowerment; people need the resources and auth ority to take charge of programs meant for their benefit. • lack of good governance: people are worse off when officials are corrupt, unre- sponsive to local demands, and unaccountable. Seen in this ligh t, local empowerment is a form of poverty reduction in its own right, quite independent of its income effects. Box 1. Local Capacity Exists, But can Only be Harnessed by Empowerment This is a major conclusion of a technical consultation organized in Rome in 1997 joint- ly by the World Bank, FAO, /FAD, and others. Some extracts: Considerable institutional capacity already exists in local governments or communities. This capacity has been cloaked by a lack of local empowerment to use it. Any definition of capacity that focuses only on technical capacity will miss the huge potential that exists, Existing capacity is best defined as the ability to solve problems. People who have survived by trying to solve problems in difficult economic and political conditions have considerable capacity to put their experience and skills to work, once they are empowered. What is perceived as a lack of local capacity is often a reflection of the fact that what local people want is different from what central planners want. Once it becomes clear that local people have the power to solve problems, they will at last have the incentive to organize, assess current ills, and work out solutions. The very act of wrestling with prob- lems develops new skills. So the process of capacity creation can be described as learn- 5 ing by doing, learning by use of power, learning by solving problems, and learning by making mistakes.This dos not mean that skill development is not required . It means that skill development should be demand-driven, not imposed from above. General purpose capacity-bu ilding funds and tra ining programs that support initiatives from below may be more effective than train ing funds or prog rams for specific sector. Once local govern- ments are in place, they will soon identify which skills are most needed and in what sequence. The best solution is to empower local governments and local communities, and then offer col- laborative support as skills are upgraded. Agencies should not attempt to create skills in a vacuum, or as a pre-condition for empowering local governments. Local capacity cannot be created unless local govern- ments are given resources that enable local people to experiment. If resources are pro- vided first, capacity creation is likely to follow. The gaps in skill can be plugged as they appear.Loans should not be burdened with too many conditions. Flexibility of use, wh ich allows local people to choose what they want to do, should be emphasized. Local capaci- t y is necessarily multisectoral because local commun ities have to deal with a wide range of issues. Block grants may be more effective than schemes following the preconceived priorities of donors and/or central governments. Sen lists five dimensions of poverty: political space, economic space, social space, transparency, and protective security. He views poverty as deficits along these five . dimensions, which limit the ability of people to develop their capabilities and func- tion as empowered persons. Poverty reduction, broadly defined, requires processes that help people improve their capabilities and functioning, that enable people to take charge of local affairs instead of being supplicants before higher authorities. CDD aims to create such processes, with safeguards to provide voice to groups tradi- tionally excluded from the decision-making process such as women and ethnic or religious minorities. Economists now accept that communities have considerable capacity to plan and implement programs, which has often been cloaked by a lack of empowerment (see Boxl.) Vibrant community structures constitute social capital, a much-neglected asset that can yield high economic dividends. CDD aims to build on social capital by harnessing community participation, and also to improve social capital by strengthen- ing incentives for participatory development. Detailed studies the world over show that well-designed participation and decentral- ization can reduce poverty's many dimensions (see Annex). However, the current reali ty in Africa is very different. The Reality Excluding South Africa, Sub-Saharan Africa has the greatest centralization of devel- opment adm inistration in the world, the lowest per capita income , and high inequal- ity. In a typical country the poorest 20 % of households get only 4% of GDP, which is less than half the net aid received on average by African countries. Fiscal policies and public services do little to narrow the income gap: in both health and education , for every $1 that goes to the poorest quintile, $2.50 goes to the richest quintile. Extreme centralization and poor governance mean that little money gets through for intended purposes: prior to Uganda's recent reforms, only one-third of non-salary allocations 6 to schools reached their intended destinations. Accountability is poor, and what little exists is upward to donors and central governments, not downward to local people. Besides, different donors use different channels for aiding community-based develop- ment. In many countries, this balkanizes foreign aid into multiple high-cost donor boutiques, each with limited geographical and sector coverage. Community development and integrated rural development are ideas that have been tried in the past but with disappointing results. Recent research suggests that the problem lay in the inability of central government agencies and donors to respond to local priorities or take advantage of local skills. Funds were donor-driven, failed to generate local ownership or enthuse locals to maintain assets, and simply created a dependency syndrome. Instead of strengthening national institutions, donors raided their best staff to create parallel institutions, which could not be sus- tained when the donor programs ended. The result: poverty remained deep although countries of the region got some of the largest flows of net aid per capita in the world. After falling in the 1980s, African incomes started growing again in the 1990s. Political change in many countries brought a new interest in decentralization and community participation , though the process has a long way to go. Social funds and other similar funds were created to channel emergency money to needy communities. The immediate success rate of these schemes was high. They showed that participation by beneficiaries in projects meant for them improved project design, implementation and outcomes. However, such projects remain almost totally dependent on outside financial support, and this reduces local owner- ship as well as sustainability. A recent book on Latin American social funds (Tendler and Serrano, 1999) suggests that these have significant li mitations. In a recent internal analysis of World Bank projects in Africa, 75 percent of projects with some level of community participation were rated satisfactory against 60 percent for all African projects in 1994-97. So, Africa has a comparative advantage in commu- nity-based projects . But while 75 percent of such projects were rated satisfactory, only one-fifth were rated sustainable. Why? Because they were almost totally donor-financed; they were Box 1A. From Participation to Decentralization: North-East Brazil Brazil's North-eastern Region and Africa have some things in common: for decades, poverty and social indicators stubbornly defied remedial attempts using large-scale exter- nal assistance. But in the 1990s Brazil has shifted from its traditional top-down approach to a bottom-up approach, and registered remarkable successes in the north-east. Three differ- ent approaches were experimented with. In all three the local community had to provide 10% of subproject costs, and obtained matching grants. Communities had to devise cost- recovery strategies to ensure sustainable maintenance and operation. The first approach was PAC (Programa de Apoio Comunitario). Rural communities sub- mitted subproject proposals to the state government, which screened and approved them. This was the least decentralized of the three approaches. The second, more decentralized approach was FU MAC (Fundo Municial de Apoio Comunitario). Here a municipal council (which included community and civil society groups) chose a bundle of subprojects, and obtained financing from a social fund run by municipalities. The third and most decentral- ized approach has recently been introduced in FUMAC II. Untied resources are made avail- able directly to selected municipalities against an annual operating plan. Since the resources are untied to specific end-uses, the local communities have much grater flexibili- ty in using them, and are more fully empowered. Local ownership, beneficiary participation and enthusiasm all have increased. Some major outcomes: • Benefit-cost ratios were high, 3.0 or more. Social internal rates of return were as high as 50 % • 95% of funds are reaching targeted beneficiaries, who are mostly landless. ·The cost of subprojects, executed by local contractors or communities, was 20-30 % less than hose executed by municipalities or government contractors. The cost of PAC and FU MAC subprojects was similar, but FU MAC had on average 40 % more beneficiaries. So the greater decentralization of FU MAC meant a greater reach and lower cost per beneficiary. FUMAC is a more transparent setting than PAC for community debates and decision-making, and helps dampen political interference. FU MAC II now takes decentralization further, by providing untied funds directly to municipalities. This will further promote ownership, participation and sustainability. temporary programs, not embedded in permanent institutions; their scale and geo- graphical spread was very limited. While they provided for some beneficiary participa- tion, they were typically earmarked for specific sectors and did not empower commu- nities to set priorities or manage projects. This was a significant shortcoming: it deprived communities of the ch ance to exercise real choice and build skills through learning by doing. Communities and local governments had no guaranteed sources of revenue, and little or no power to raise local resources. The lesson: much can be achieved through social and other funds, but the process needs to be taken further. Social funds-as designed in much of the 1990s- represent communi ty development without institutionally sustainable empowerment, and this limits its impact and sustainability. CDD aims to remove these limitations, and so improve on the su ccess registered by social funds. First, the impact of CDD can be enhanced by providing untied block grants to communities across entire countries. This will give resources as well as voice to peo- ple who h ave rarely enjoyed eith er. Second, complementing the participatory resource transfer, CDD aims to encourage accountable local governments which can support communities on a permanent basis. Creating strong communities in the ini- tial stage should h elp create accountabili ty norms and leadership patterns which in due course will h elp improve the quality of governance at the local level. And good local governments in turn will widen and deepen participatory processes. CDD aims to help foster a mutually reinforcing process between community development and local government development. Accountable local governments, embedded in a coherent framework of inter-gov- ernmental relations, cannot be a precondition for community empowerment, but are an ultimate institutional. Recent experience in north-east Brazil shows how enhanced participation can evolve logically into greater decentralization. (see Box lA). 8 EMPOWERMENT: HOW TO DO IT To embark on local empowerment, we need first to enunciate its key principles. We can then consider the main elements of a set of interventions to enhance participa- tion and decentralization, tailored to the stage of development in each country. Key Principles of Community Empowerment 1. CDD empowers communities by giving them untied funds which allow them to choose their own priorities and implement their own programs . This will enable local skills to be created and improved through learning by doing. A negative list can be d evised to avoid undesirable spending. 2. Where n o decentralization or local funds exist, we need to start small and grow gradually, so experiments that fail will not be costly. Grants as small as $ 5,000- 10,000 p er community can kick-start the process. 3. The aim should be to cover communities across entire countries within a short time. Experience in Africa shows that this is feasible (see box 2). 4. Development must be participatory and demand-driven. To provide voice to the voiceless, it should have safeguards against social exclusion and elite capture. 5. Devolution to communities and local gove rnm ents implies that th e direction of accountability should be downward to local people, not only upward to central governments and donors. 6 . Basic skills garne red through learning by doing must be upgraded steadily with outside support. Local gove rnm ents and communiti es must be able to ge t techni- cal and managerial support on demand. 7. Developmen t has too often been viewed as mainly about physical inves tm ent o r hardware. CDD emphasizes institutional change and the software of development- empowering and sensiti zing people, enhan cing tran spare ncy, chan ging attitudes. 8. Decentralization should be based on the principle of subsidiarity. That is, responsibility fo r tasks should be devolved to the lowest level of government that can deal effectively with them. Each level should perform tasks according to its comparative advantage. 9. Eve n after decentralizatio n , many tasks involving scale eco no mies and externalities will re main with higher levels of governmen t. Decentralizatio n sh ould n ot pit local governm en ts agains t ce ntral governments or sectoral p rograms. It sh ould be a joint venture of different levels of government. 10. Decentralization must give local governments a predictable, transparent share of revenue (including fo reign aid ) .T his wi ll make th em fin ancially viable. 11. To promo te local "own ership", communities and local governments must con- tribute to project costs and operation and maintenance costs, apart from helping with design , impleme ntatio n , m ain ten an ce and m o nitoring. 12. Targeted schemes sh o uld be d esign ed by don o rs and central gove rnm en ts for obj ectives whi ch may no t be given p ri ority by co mmuniti es (reaching the p oor and th e mino riti es, avoiding e nvironm en tal damage, combating AIDS). To tran slate the key principles of empowermen t in to an ac tio n pla n, it is useful to consider th e main d imensio ns of CDD. Box 2. Entire Countries Can Be Covered Quickly by COD 9 Two years ago, Benin covered a little over half of the 500 villages in the Borgou region within six months with participatory appraisal, using resources already available in the region . This implied that entire countries could be covered with additional resources from outside. More than 6,000 villages have now been covered in other African countries such as Burkina Faso, Cameroon, Cote d'Ivoire, Guinea, Madagascar, Malawi, Mali, and Uganda. Tanzania, Togo and Zambia are about to follow suit. Lessons learned from these experiments are: 1. Start with a national exercise for harmonizing participatory appraisal of projects. In the past, the balkanization of donor efforts has been a major hurdle to national coverage. This can be overcome if donors and central government pull together. 2. The next step is to train the staff of the various agencies. Successful countries have used small but strong teams of national trainers at the regional (or district) level. The faster local agencies pool their resources, the faster national coverage can be achieved. 3. Start simultaneously in as many regions as possible: Madagascar has started pilot proj- ects in all its 29 regions. Each region can then establish its own action plan to reach regional coverage, tailored to available resources. - - - -0 Dimensions of COD The five main dimensions of CDD are: empowering communities, empowering local govern- ments, re-aligning the center, improving accountability, and building capacity. 1. Empowering communities Experience in Africa shows that communities can be organized quickly and produc- tively to diagnose local problems, come up with solutions, lay down priorities, elabo- rate acti on plans, and strengthen community organizations and accountability. However, participatory processes will be discredited and atrophy unless communities are empowered with resources and authority. Social/ other funds already provide matching grants to communities in many countries, but these are typically balkanized funds with a narrow focus tied to donor priorities. Communities will be truly empow- ered only if th ey get untied grants which enable them to decide their own priorities and h one their decision-making skills. Experience sh ows that rules for participation can be designed to give voice to socially excluded groups like women and minority tribes (see Box 3) Fiscal rewards and penalties can spur competition between communities and accel- erate the learning process. Communities wh ich make the best use of funds deserve increased grants. T h ose fou n d guilty of embezzlement should suffer a suspension of gran ts till a n ew management team takes over. Box 3.Social Inclusion of Women and Minorities If not handled well, women will refra in from mentioning AIDS because of fear, shame and denial. TANESA, a Tanzania based NGO, has developed many tools which encourage communities to discuss AIDS. One of these tools is "risk mapping": people make a map of their village and indicate where they are more at risk of contracting AIDS. Because the "risk mapping" takes place in different gender groups, women and men can express themselves more freely. Many women in Tanzania indicated water wells or woodlots as high-risk areas: they had experienced sexual aggression and rape while being away from the village to collect water or fuel-wood. In the light of this, priority was given in some village plans to improving stoves to diminish fuel -wood consumption (and hence the time women spent in collecting wood), and to digging village wells which made it unnecessary for women to wander afar for water. This is just one example of how the participatory methodology, if implemented well, can empower vulnerable groups. In many West African countries, latent tension between semi-nomadic pastoralists and farmers creates conflicts that can turn violent. Farmers see their crops being destroyed by cattle, and herders feel that they cannot make use of the land for grazing.In countries such as Guinea, Benin, Mali or Burkina Faso, the participatory process has provided a plat- form for the farmers and herders to discuss these issues in a peaceful way, and to find common ground. Using mapping techniques, they demarcate grazing corridors and reach agreements on the rights and regulations for the use of the land. This is an exam- ple of how both majority and minority interests ca n be accommodated in a community framework. 2. Empowering Local Governments Community empowerment is unsustainable if based on donor-driven funds. It needs to be embedded in a new institutional framework of local governments. Earlier attempts at decentralization the world over were hamstrung by a failure to harmonize the political, administrative and fiscal elements involved. So, harmonization is a top priority. This typically requires a high-powered task force backed by the head of gov- ernment to work out new legal and constitutional arrangements. The political leader- ship in each country can decide how many tiers of government are appropriate and what powers each will enjoy. This is best done on the principle of subsidiarity. Arrangements for local elections can consider providing quotas for women and other socially excluded groups. Administrative decentralization must keep pace with political decentralization. Central government staff for frontline services may need to be transferred to lower levels of government. Ultimately, each level of government should be able to hire and discipline staff in its jurisdiction. It should also have the administrative machinery to collect local taxes and user charges. Fiscal decentralization has three elements: revenue generation, revenue sharing and matching grants. Local governments will need powers to levy taxes and user charges, once again on the basis of subsidiarity. Studies show that the greater the share of local revenue , the more productive local spending is likely to be. Municipal / local finances currently represent only 2 to 3 percent of national revenue in most countries, and so are not paid attention to by Ministries of Finance or macro- economists. Sustainable decentralization requires that local governments get an assured share of central revenue. This is typically based on a formula worked out by all stakeholders. Over and above this , the central government may give earmarked grants for areas which may be neglected by local governments such as combating AIDS, environmental damage and social exclusion. Once local governments are up and running, donors can switch untied aid from 11 social / other funds to budgetary support for central governments, conditional on binding arrangements to share central revenue (including foreign aid) with local governments. At that stage, local governments can take over from donors the task of providing matching grants to communities. Local governments and municipalities often have limited capacity, and need to har- ness the local private sector to implement projects. 3. Realigning the Center Decentralization implies a far-reaching change in the role of the center, as many responsibilities and resources will shift from the center to local governments. Management and control processes geared to supporting a monolithic hierarchy will need refashioning to support arms-length relationships between multiple centers of power. Traditionally, central governments in Africa have fo llowed the "blue collar" approach of operating all services. After decentralization, they will need to shift to a "white collar" approach. Instead of running services directly, they should focus on facilitating local government activities, setting standards, monitoring outcomes, pro- Box 4. The Vision That Conquered Blindness The conquest of riverblindness in West Africa is one of the most successful multi-donor programs in the history of aid. This once-rampant disease causes severe skin problems before escalating into blindness. It has virtually been eliminated in subregions with a pop- ulation of 34 million in 11 countries. It has saved the sight of 600,000 people and spared 15 million children the risk of infection. By removing the disease from 25 million hectares of arable land, it has opened up huge areas for spontaneous settlement and agriculture. One reason for the program 's success was an appropriate division of labor between central governments and local communities. Transmission of the disease by blackflies was controlled by governments through the aerial spraying of environmentally-safe insec- ticides. And local communities managed the supply and dosage of invermectin, a drug that has cured over four million people. In effect it was a joint venture between the cen- ter and peripheries, based on the principle of subsidiarity. Communities did what they could best handle (distribution of medicines) and the center did what it alone could (aeri- al spraying plus arranging the import of medicines). viding training to lower levels, and providing rewards and penalties to improve local government performance. Reformers will come up against fears and resistance from civil servants and other powerful groups who believe they will lose from the change. The reform program needs to send the message that d ecentralization does not mean the withering away of the center; it implies a j oint venture between different levels of government, each contributing on the basis of its comparative advantage. The conquest of riverblind- ness is an excellent example of a successful joint venture of this kind. (see Box 4) Even after decentralization, large projects with scale economies will remain with the center. Strong sectoral policies and institutions wi ll be needed even after decen- tralization to strengthen capacity building at local levels, set minimum standards for sch ools or health centers, devise procedures to ensure accountability, social inclusion and environmental safety. If decentralization succeeds in accelerating economic growth, this could ultimately mean increased resources for the center and all other levels of government. So CDD aims at a win-win situation for central and local gov- ernments. 4. Improving Accountability Existing mechan isms have failed to provide adequate accountability. Almost all accountability is upward to donors and central governments, not downward to users of services. Even upward accountability is balkanized because different donors have d ifferent requirements and create parallel systems of accountability, which sap limit- ed local resources. A plethora of documents and reports are often required in the language of donors, which may not be locally understood. This reduces transparency and participation, and helps elite capture of programs (only elites understand for- eign languages) . T h e success of community driven projects in Africa shows that accountability exists as social capital in user groups. Social pressure and peer pressure create accountability within a community. The empowerment of communities and local governments will Q enable this social capital to be harnessed, and provide downward accountability to users of frontline services. Greater participation in all projects and programs will improve the voice of local people. Decentralization must go to the grassroots, as social capital works best in small groups. Rural local governments should cover no more than a few tens of thou- sands of people, increasing transparency and responsivity. Fiscal rewards and penalties for communities and local governments can induce competition between them. Losers will be answerable to their constituents for being left behind in the race for matching grants. Such competition will improve accounta- bility and performance. Traditional forms of upward accountability also need to be strengthened. One way of doing so is to train communities to conduct monitoring and evaluation. Users of services are likely to provide more relevant, reliable and timely data than traditional government channels. Donors need to reduce or end the balkanization of accountability. Ideally, they should help create a single line of accountability with all documents in the local language . 5. Building Capacity Untied matching grants to communities will help develop their inherent capacity for problem-solving through learning by doing. As they take on more responsibilities, they will find they need to upgrade their skills. This can be facilitated by technical assistance from central governments and NGOs (see box 1). Local governments will also develop skills initially through learning by doing, and later through technical assistance. The central government needs to upgrade its skills for carrying out large projects, for training local governments and communities, and more broadly for taking on its "white collar" planning, facilitation and regulatory roles. CDD can play an important role in private sector development. Existing procure- 13 ment procedures are often cumbersome and keep out local private sector suppliers on technical grounds. Programs of need to devise procurement procedures that enable local governments/ communities to procure local materials and manpower without hassles. This will help develop the local private sector. AGETIPS in African countries started as independent agencies to fund and implement emergency works in crisis-hit countries. Today AGETIPS are confined to what they do best-delegated contract management and works execution-while the planning, cofinancing and maintenance of projects is left to local governments. Based on these five dimensions, each country can consider how to design a CDD program adapted to its particular circumstances and stage of development. Designing COD to Fit a Country's Circumstances A program of local empowerment can start by consulting all sections of society on how to proceed in the light of the country's history and traditions, and by drawing on the most successful experiences in the country and in neighboring ones. Countries vary widely, so the approach to CDD will also vary widely. Yet most will need a two-pronged approach. First, they will need an immediate action plan to empower communities with resources and authority. Second, they will need a long-term plan to enhance participation and decentralization. The immediate action plan can be a building block of the long-term plan. Each country will have different ideas on what programs and services should be devolved to communities and local governments on the basis of subsidiarity. Each will have a different long-term vision for the share of total fiscal resources to ultimately be at the disposal of local governments and communities( 20-40 per cent would be area- sonable range for many African countries). A guiding framework for CDD can spell out how to get from the current reality to that target in phases. Conditions vary greatly across countries: some have no community participation at all, others may already have functioning local governments. Each country needs to diagnose its particular problems and tailor solutions. In doing so, countries may find it conceptually useful to consider three stages of decentralization-initiation, scaling up and consolidation. Initiation Many African countries today have little or no local participation, decentralization or local funds. Reformers in some of these countries are keen to initiate local empower- ment. They could act simultaneously on three fronts: enhancing participation, tar- geting specific groups (notably people affected by AIDS/ HIV) and starting a dialog with stakeholders on decentralization. An important starting point is to harmonize and strengthen ongoing efforts in community participation. In many countries, a plethora of uncoordinated donor-driv- en funds already exist. Ideally, these should give way to a single local investment fund supported by all donors. The aim must be to get resources and the authority to spend these to as many communities as possible, as quickly as possible. By starting small and 14 growing gradually, entire countries can be covered quickly. Many projects will contin- ue to be executed by central agencies, and new procedures can make it mandatory, wherever feasible, to use potential beneficiaries to appraise, implement and monitor such projects. Reformers in this stage will typically give priority to empowering com- munities over targeted schemes to reach the socially excluded. But AIDS needs to be targeted from the very start. It is a major development challenge which is unlikely to be met by community action: it will need targeted central schemes from the outset. Countries with weak or no local governments need to engage in a dialog with stake- holders and donors on the merits and feasibility of decentralization. Some countries may be convinced of the need to proceed quickly, and donors can assist them in drawing up a framework that harmonizes the political, administrative and fiscal aspects of decentralization. Not all countries are convinced of the need for decentralization in the immediate future. In such cases reformers should keep the dialog on this topic going, and attempt to mobilize opinion in civil society groups too. While that dialog proceeds, the process can start of empowering communities with resources through social/ other funds. Enhanced participation is in any case the first building block of CDD, and the foundation for this must be laid quickly even if the debate on the mer- its of local government goes on for a long time. Q Scaling Up Many countries have crossed the initial stage and achieved some progress in partici- pation and decentralization. The challenge they now face is to scale up operations. Experience shows that community empowerment can be scaled up rapidly. Entire regions can be covered within months, and entire countries soon afterwards. Madagascar has embarked on 80 pilot projects in all 29 regions of the country simul- taneously, and other countries have covered more than 6,000 villages quickly (see Box 2). For late starters in decentralization, the dialog on decentralisation in the initiation stage can give way to concrete action in this stage. The legal and constitutional frame- work can be drawn up and implemented. For early starters, which have already put local governments in place, the challenge will be to strengthen their capacity and accountability mechanisms. Successful com- munity empowerment will typically create a good environment for effective local gov- ernments, and the two can evolve in a mutually reinforcing manner. In the scaling up stage, local governments can gradually assume full responsibility for frontline service delivery. Communities in the initial stage typically rely on temporary funds from social / other funds financed by donors. Local governments could gradually take over this task in the case of early starters. This is being attempted in Zambia. Responsibilities are being transferred in stages from a social fund to local govern- ments, and local governments which perform well are able to advance faster along the path to additional resources and responsibilities. This has proved is a good way to promote competition in good governance, accountability and capacity creation. It enhances community empowerment as well as local government empowerment, and helps both develop a mutually profitable interaction. Consolidation 1 Few African countries have gone beyond the scaling-up stage to consolidation, but that is the logical third stage. In this stage decentralization will be fully operational, but many glitches and problems will remain. This is the stage for fine-tuning political, administrative and fiscal arrangements, and for further upgrading local skills. Where local governments are already functioning well , donors should consider a switch in strategy. They could switch progressively from funding communities directly to providing budget support for central governments, conditional on arrangements to channel an assured share of central government revenue to local governments, which in turn will support communities. This stage may witness much greater emphasis on targeted schemes to fill gaps in social inclusion left by local governments. The local revenue base should increase substantially in this stage, and dependence on donors should decrease. Sequencing: Combine Immediate Action with the Long Term Agenda These three stages represent an idealized conceptual framework: the reality will differ widely from country to country. The framework for will vary accordingly: great diver- sity should be expected. Some countries could choose to begin by focusing on strengthening community participation through social/ other funds. Others could early on move aggressively to initiate or deepen decentralization. Learning by doing is important for all actors; communities, local governments, the private sector, civil society, and donors. By empowering communities quickly with resources, a start can be made and momentum gained even in difficult environments, like post-conflict situations. Since country ownership is essential, donors must pay due regard to the pace of change which countries think is suitable. A dialog is desirable on accelerating action where it seems feasible, but donors must resist the temptation to impose accelerated time-tables on reluctant governments. Throughout the process, the emphasis should be on creating and strengthening the potential synergy between community empow- erment and gains in the capacities of local governments. Annex: Using Empowerment to Reduce Poverty: the Evidence .. Community development and decentralization are old ideas which in the past have yielded very mixed results. Many studies in the past have cast doubt on the ability of decentralization to improve service delivery or reach the poorest of the poor. However, more recent studies have greatly strengthened the case for CDD. One strand of the literature now focuses on a broader approach to poverty that goes well beyond income effects. Another shows that participation and decentralization tend to succeed if well designed and fail if poorly designed. Broader Approach to Poverty The new approach views poverty as not just a lack of income, education and health facilities, but also as a lack of voice, empowerment, good governance and security against shocks. Seen in this light, local empowerment is a form of poverty reduction in its own right, quite independent of its income effects. WDR 2000/ 01 builds on the ideas of Amartya Sen (1999),who lists five dimensions 16 of poverty: political, economic, social, lack of transparency, and lack of protective security. He views poverty as deficits in peoples' capabilities and functioning along these five dimensions. Poverty reduction, broadly defined, requires processes that improve the capabilities and functioning of people in these five ways. Ostrom (1999) argues that intrinsic economic motivation is diminished by low lev- els of self-esteem and self-determination, which arise from lack of voice and empow- erment. The literature on social capital (Putnam et al 1993, Fukuyama 1995, Narayan and Prichett 1997) show that vibrant community structures have major economic exter- nalities. The importance of local interaction to build capable, responsive and responsible institutions has been stressed by Douglas North(l990), Sola and Ebbe ( 1995), Burnside and Dollar ( 1997). Sujai Shivkumar(2000), Li (1999) and McGinnis (1999a, 1999b) argue that devel- opment and institutional governance evolve from associational patterns, which are determined in communities. Well Designed Decentralization and Participation Works A comparative study of decentralization in four countries ( Crook and Manor 1998) shows that it yields good results only if there is strong government ownership; appro- priate legal, administrative and fiscal arrangements; actual empowerment of locals with local elections; sufficient and reliable funding; and substantial freedom for com- munities to choose projects. Not surprisingly in this regard, the evidence on the asso- ciation between levels of decentralization and levels of corruption is mixed ( Gatti and Fisman 2000, Crook and Manor 1995, Huther and Shah 1998, Treisman 1998, Wei 2000). WDR 2000/ l sees great promise in decentralization, but only if it is tailored to reach the poor and voiceless, receives adequate support as well as sufficient autono- my from the center, and if institutionalized mechanisms of wide and regular partici- pation are put in place. Integrated rural development failed in north-east Brazil when imposed from above, but has registered remarkable success in a new, well-designed framework of decentralization (Van Zyl et al 1995 ) . Decentralization can benefit poorer regions if a revenue-sharing formula between regions gives due weight to the needs of backward areas. Sola and Ebbe (1994) find that in many cases of decentralization, communities have responsibilities but little authority or resources, and so achieve little. Social funds have positive outcomes but also shortcomings. Tendler (2000) shows that they have an important supply-driven component that undermines community empower- ment, that they do not represent real decentralization, and have a limited impact on poverty reduction. Wiens and Gudagni (1998) shows that demand-driven investment funds in several Latin American countries did not reach the very poorest. Manor (1999), Bardhan and Mookerjee (1999) and Kerr et al. (1998) find that elite capture and social exclu- sion may prevent the benefits of CDD from reaching the most vulnerable (minority groups, women, the poorest of the poor), so central interventions may be needed to 1 target them . The importance of geographical clustering in promoting skills and productivity has been emphasized by Michael Porter (1998) , Paul Krugman (1991 ,1999) and David Audretsch (1998). Porter( 1995) argues that externalities from clustering can be used to increase incomes and reduce poverty in US inner cities. This logic can be extend- ed to reducing poverty in communities in poor countries. Leaming by doing is an important way of creating capacity in communities and local governments (Fiszbein 1997, Foster and Rosenzweig 1995). Technical assistance schemes in the past often failed because they did not correspond to community needs and priorities. In fact communities have immense latent capabilities, shown by the way they cope in extraordinarily difficult situations including post-conflict ones. Once communities and local governments are given the power and resources to choose and implement projects, the very act of wrestling with problems builds capaci- ty, which can later be supplemented by training (FAQ 1999). Beneficiary participation has greatly improved economic outcomes and helped reduce poverty in Bolivia (Faguet 1997), and in several African countries such as Benin, Zambia, Guinea, Malawi and Uganda. Community management of medicines - - - -Q has played a key role in the conquest of river-blindness in West Africa. Water Users' associations have been a major success in Andhra Pradesh (India), Mexico and Turkey (Oblitas and Peter 1999). Community participation helped curb absenteeism among school teachers in El Salvador and India. (King and Orazem 1999, PROBE team 1999). Local participation has succeeded where centralized schemes failed in forestry, biodiversity and wildlife management (Caldecott and Lutz 1996). In South Africa's poverty reduction program , the cost of transferring one rand to a poor per- son more than halved when it was done solely by communities (Adato and others 1999). Collaboration between communities and donors also increased returns to aid (WDR 2000/ 01). References: Adato, Michelle, Timothy Besley, Lawrence Haddad andjohn Hoddinott. 1999. Participation and Poverty Reduction: Issues. Theory and New Evidence from South Africa. Mimeo. Audretsch, David, 1998, Agglomeration and the Location of Innovative Activity, Oxford Review of Economic Policy (U .K.); 14: 18-29, Summer. Bardhan, Pranab. and Dilip Mookherjee. 1999. Capture and Governance at Local and National Levels. Mimeo. Burnside, Craig and David Dollar, Aid. Policies and Growth, Policy Research working Paper 1777, The World Bank, Washington, D.C. Caldecott,Julian and Ernst Lutz (eds), 1996, Decentralization and Biodiversity Conservation, The World Bank, Washington, D.C. 18 Crook Richard C and James Manor, 1995, Democratic Decentralization and Institutional Performance: Four African and Asian Experiences Compared, Journal of Commonwealth and Comparative Politics, November. FAO, 1999, Technical Consultations on Decentralization. Fiszbein, Ariel, 1997, The emergence of local capacity: Lessons from Colombia. World Development (U.K.); 25:1029-43,July. Foster, Andrew and Mark R. Rosensweig, 1995, Learning by Doing and Learning From Others: Human Capital and Technical Change in Agriculture, Journal of Political Economy (U.S.); 103:1176-1209, December 1995. Fukuyama, Francis. Social Capital and Civil Society, IMF Working Paper WP / 00/ 74 Gatti, Roberta and Raymond Fisman, 2000, Decentralization and Corruption : Evidence Across Countries, Policy Research Working Paper 2290, The World Bank, Washington, D.C. Huther Jeff and Anwar Shah, 1998, Applying a Simple Measure of Good Governance to the Debate on Fiscal Decentralization, Policy Research working Paper 1498, The World Bank, Washington, D.C. Kerr Graham, Keith McLean, Suzanne Piriou-Sall and Melissa Williams, 1998, Decentralization and Rural Development: Characterizing Efforts in 19 Countries, The World Bank. King, Elizabeth and P. Orazem. 1999. Evaluating Education Reforms: Four cases in Developing Countries. World Bank, Washington, D.C. Krugman, Paul R. , 1999, On the Evolution of Hierarchical Urban Systems, European Economic Review (Netherlands); 43:209-51, February Krugman, Paul R., 1991, Geography and Trade, Leuven University Press. McGinnis, Michael D. 1999a. Polycentric Governance and Development: Readings from the Workshop in Political Theory and Policy Analysis. Ann Arbor, MI: University of Michigan Press. McGinnis, Michael D., 1999b. Polycentricity and Local Public Economies: Readings from the Workshop in Political Theory and Policy Analysis. Ann Arbor, MI: University of Michigan Press. Manor,James, 1999, The Political Economy of Democratic Decentralization, The 19 World Bank, Washington, D.C. Narayan, Deepa and Lant Pritchett, Cents and Sociability : Household Income and Social Capital in Rural Tanzania, Policy Research working Paper 1796, The World Bank, Washington, D.C. North, Douglas. 1991. Institutions.Journal of Economic Perspectives 5 (Winter): 7- 112. Oblitas Keith, and]. Raymond Peter, 1999, Transferring Irrigation Management to Farmers in Andhra Pradesh, India. The World Bank, Washington D.C. Ostrom, Elinor. 1999. Crowding Out Citizenship. Scandinavian Political Studies. Putnam, Robert D. with Robert Leonardi and Raffaella Y Nanetti, 1993, Making Democracy Work: Civic Traditions in Modern Italy, Princeton University Press Porter, Michael, 1998, The Adam Smith Address: Location, Clusters, and th e "new" Microeconomics of Competition, Business Economics (U.S.); 33:7-1 3, J anuary. Porter, Michael, 1995, Toward a new Conception of the Environment- Competitiveness Relationship, Journal of Economic Perspectives (U.S.); 9:97-118, Fall. PROBE Team. 1999. Public Report on Basic Education in India, New Delhi: Oxford University Press. Li, Shuhe. 1999. The Benefits and Costs of Relation-based Governance: An Explanation of the East Asian Miracle and Crisis. October 20. Working Paper. Sen, Amartya Kumar. 1999. Development as Freedom, New York, Knopf. Shivakumar, Sujai. 2000. The Constitution of Development. Workshop in Political Theory and Policy Analysis. Unpublished manuscript. Tendler,Judith, 2000, Why are Social Funds so Popular?. mimeo. Treisman, Daniel, 1998 , Russia's Taxing Problem, Foreign Policy Van Zyl,Johan, et al, 1995, Decentralized Rural Development and Enhanced Community Participation : a Case Study From Northeast Brazil, Policy Research work- ing Paper 1498, The World Bank, Washington, D.C. Wei, Shang:Jin, 2000, Natural Openness and Good Government, NBER Working Paper 7765, June. Wiens, Thomas B. and Maurizio Gudagni, 1998, Designing Rules for Demand-driven Rural Investment Funds : the Latin American Experience. World Bank Technical Paper 407 World Bank, 2000, Attacking poverty. World Development Report 2000-1 , The World Bank, Washington, D.C. COMMUNITY DRIVEN DEVELOPMENT