Many African countries are embarking on donor-supported processes of political and administrative decentralisation. In theory, decentralisation holds great potential for development. Decentralised government can provide space for people to participate in local development. It can ensure a more efficient allocation of resources (including development aid), enhance local resource mobilisation and improve local governance. This, in turn, may pave the way for more effective poverty reduction strategies. In development practice however, the ways in which decentralisation can make a positive contribution to poverty reduction have not been widely explored.

This paper reports on a study of the linkages between decentralisation and poverty reduction. Drawing on three cases from Ethiopia, Guinea and Mozambique, the study explored ways to make poverty reduction strategies more effective, focusing on three major questions:

• Can decentralisation contribute to poverty reduction?
• What is the added value of local governments in poverty reduction strategies?
• How can donor agencies support stronger linkages between decentralisation and poverty reduction?

It should also be stressed that decentralisation is not seen as a panacea for development. While the study assumed that greater participation by grassroots communities through established statutory structures is a prerequisite to democratic governance, it also recognised that national security, stability and social order require solid central government authority. In exploring the role of decentralisation in reducing poverty, this study does not advocate dismantling the central State. It examines the lessons that real-life experiences in three countries can provide for improving the balance between central guidance and local initiative.

The Nature of Decentralisation Processes

Why are governments pursuing decentralisation? While some sort of decentralisation is underway in each of the countries studied the term can embody very different political realities. Is decentralisation taking place ‘by design’ (because central government wants to improve its overall development performance) or ‘by default’ (because central government lacks the fiscal capacity to deliver basic services to its citizens)?

Key Concepts Decentralisation is the devolution of resources, tasks and decision-making power to democratically elected lower-level authorities that are largely or wholly independent of central government. Poverty is a multi-dimensional concept, encompassing economic, social, political and cultural aspects. In the framework of this study, no attempt was made to agree a common definition, as perceptions of different actors and stakeholders varied considerably on the constituent elements of this concept and their relative importance.

While professing a commitment to political and administrative reform, most African governments place poverty reduction high on their political agendas. But is poverty reduction an explicit objective of decentralisation policies? In attempting to assess this, it is helpful to imagine the kinds of linkages one might expect to find between decentralisation and poverty reduction policies:

• Political devolution can aim to reduce poverty through an empowerment strategy that creates ‘space’ for people to effectively participate in decision-making processes (including the setting of priorities on the allocation of aid budgets). An attempt to increase the access and influence of local citizens vis-à-vis government may be driven by the belief that institutions of the State will thus become more responsive to the needs of the poor.
• Poverty reduction and decentralisation might also be linked through a resource mobilisation strategy. When people are given greater control over local statutory structures, they may be motivated to commit more assets to the common good.
• Decentralised government may also be seen as a more effective means to deliver basic social services, thus alleviating many of the common causes of poverty such as illness, decrepit economic infrastructure and illiteracy.
  

Each of the three countries studied (Ethiopia, Guinea and Mozambique) lived through a long period of centralised, one-party-State systems. These regimes embodied post-independence (Marxist) ideological choices. They did not exclude the existence of administrative structures at the local level (Ethiopia), nor the setting up of new administrative and political decision-making structures at the local level (Guinea). However, these local structures lacked political autonomy and credibility among local people. They were generally conceived with a view to control, steer and bring the population into line with the political doctrine rather than to increase political, economic and social choices for the poor.

Deconcentration versus Devolution An important distinction concerns the administrative deconcentration of resources as against their political devolution. In cases of deconcentration, resources such as civil service positions or budgetary funds are reallocated from central government to a lower administrative unit (e.g., a district or prefecture). Decision-making authority remains with the central government and local staff answer to their upstream superiors. When authority is devolved down to a lower level, decisions about the use of the resources are made locally, and administrative staff tends to be accountable to local political leadership.

In the 1980’s, the political hegemony of the highly centralised developmental State began to unravel. This often led to political crisis. It also catalysed political and administrative reforms, some initiated from below, but usually orchestrated by the national political leadership. In Guinea, a process of liberalisation was initiated in 1985, which also included a programme of political and administrative decentralisation. Mozambique introduced multi-party democracy in 1990, a general peace treaty in 1992, and a programme of local government reform. This resulted in two statutory instruments on decentralisation, in 1994 and 1997. In Ethiopia, the transitional Government of 1991 saw the removal of the centralised governance system as a key to preserving political stability and to ensuring its own legitimacy. A federal State, based on a far-reaching policy of regionalisation was introduced along ethnical-cultural lines, granting regions the right to self-government (including the right to secede).

In assessing and comparing these decentralisation processes with respect to their potential contribution to poverty reduction, a number of common features emerge.

Decentralisation is a highly political process

Decentralisation is largely about power and access to resources. In Mozambique, for instance, views on the purpose, pace and sequencing of the decentralisation process varied according to political background and position occupied in the national administrative hierarchy. One categorisation of views might distinguish between actors pursuing (or adapting to) decentralisation as part of a patrimonial agenda (aimed at preserving the monopoly of power and ensuring control over resources), and those with reformist objectives (seeking to create more transparency, accountability and efficiency in the management of local affairs). These diametrically opposed positions induce quite different approaches towards poverty reduction.

These two strands can also be found in Guinea, where they tend to be seen as two sides of the same coin. In a country where party politics are strongly influenced by the balance between the four main ethnic groups and where a civil war raged just across the border, decentralisation was promoted with a view to granting greater political independence to local communities, while preserving national unity.

The Mozambique case illustrated how the fear of releasing power and control over resources to the opposition led the Frelimo government to fundamentally redirect the decentralisation process after the introduction of multiparty politics. While the decentralisation legislation of 1994 envisaged a far-reaching devolution of power to new local bodies, a law passed in 1997 reversed this trend by limiting the number of areas where they could be established and subordinating them to higher levels of government. This, in turn, led to temporary erosion of donor interest in the process. Since then, however, it appears that the elections and the establishment of local authorities has revived interest in decentralisation.

A similar trend is evident in Ethiopia where the federal government’s initial preparedness to foster regionalisation appears to be moving towards re-centralisation. Thus, the central government’s policy to coordinate NGO action (starting with registration requirements) and to take the leading role in sectoral development planning is helping it to (re-) gain control of the regions.

Against the background of structural adjustment and a shrinking public sector, resistance to devolving authority is widespread in central ministries in each of the countries studied. As a result, central governments appear reluctant to clearly delineate the respective roles, rights and responsibilities of the different levels of government. This may be partly related to the technical complexity of moving towards more decentralised forms of governance, but also reflects political power struggles between different layers of government. There is also a broad consensus that local governments cannot become effective agents in poverty reduction and local development without a substantial increase in their financial and human resources. Direct external support to local governments is thus often seen as a solution that allows central ministries to postpone the devolution of power, while local governments and the population call for stricter political conditionalities and for the monitoring of the use and transfer of resources.

Decentralisation has poor links with poverty reduction objectives

In each of the three countries, administrative and fiscal considerations seem to have been the driving force behind decentralisation. Political considerations tend to enter the process largely as constraints on the radical downward devolution of authority. The idea that decentralisation could bring positive political outcomes via greater participation and improved livelihoods is not widely embraced. Thus, a constructive linkage of decentralisation to poverty reduction via strategies of empowerment and resource mobilisation does not emerge from government policy statements. None of the countries’ national decentralisation policies have provided local government institutions with an explicit mandate to effectively combat poverty. Only in Guinea were poverty reduction objectives, i.e. the desire to improve the dire social and economic situation inherited from the former dictatorial regime, referred to as a motive for decentralisation from the beginning.

In other cases, e.g., within the education sector programmes in Ethiopia and Mozambique, decentralisation was explicitly linked to enhanced service delivery. These programmes have generally promoted sectoral deconcentration. This can increase the level of resources available for local development, but doesn’t necessary strengthen local democratic institutions. Sectoral deconcentration tends to empower line ministries and central authorities, leaving local populations without the means to guide the deployment of deconcentrated resources via statutory local government structures. Hence, the decentralisation policies studied did not make strong linkages between empowerment, resource mobilisation and poverty reduction, nor did poverty reduction strategies systematically elaborate on the potential of decentralisation to improve the living conditions of poor and marginalised people.

Decentralisation is a top-down process

It appears, then, that the most typical strategy underlying decentralisation processes is one of incorporation. Inasmuch as there is an explicit design, decentralisation tends to be masterminded and implemented by ruling elites. This can leave citizens and organisations at the grassroots level feeling alienated and excluded. Rent-seeking behaviour and the continued search for control over their people and resources on the part of central and local government officials spawns widespread mistrust.

An abstention rate of 86% in the recent municipal elections in Mozambique is a clear indicator of ordinary people’s indifference (or outright scepticism) towards local government structures. Such a low rate of participation reflects the top-down culture of Mozambican politics, which is hard to reconcile with community empowerment.

In Ethiopia, the regionalisation process in the early nineties brought about a major shift of decision-making powers and responsibilities to the regional states. Yet the developmental impact of this depends on the capacity of the regions to take advantage of the process. The benefits of regionalisation for ordinary citizens is contingent on the preparedness of regional state governments to further devolve power to lower administrative tiers (to the so-called woredas), and on the prevailing political culture — it is difficult to break away from decades of centralised governance. Often, local people do not believe that their interests will be represented in a decentralised political system and are afraid of being manipulated by party-related organisations. Linkages between central and local-level politics remain organised in a hierarchical, control-oriented way, thus reducing the scope of local arenas for collective decision-making. A case is the Gemgema-system, whereby the performance of local civil servants is measured more by their allegiance to central paymasters than by their achievements on the ground.

Similar problems can be observed in Guinea. Interviews suggested that the legitimacy of the elected representatives in the local development communities and urban communes is greater than that of the local structures of the former revolutionary regime. Nevertheless, there remains a great deal of distrust of the elected communal representatives. This appears to be particularly true for the population of poorer, isolated rural areas and for the poorer urban quarters. Here, the population has often seen the same politicians re-emerging at the local level and claiming new political functions. Thus, the new local governments tend to be associated with the control structures of the old regime. In spite of efforts to provide adequate information on the nature, mandate and procedures of local governments, citizen’s opportunities to participate in local planning and decision-making are still insufficient.

Donors lack coherent strategies towards decentralisation

It is also hard to find a clear articulation of the link between support for decentralisation and for poverty reduction in donor policies and practices. In general, the backing that donors give to decentralisation seems to combine a mix of economic, technocratic and political objectives. In Mozambique and Ethiopia, poverty reduction only features in this agenda via the enhancement of social service delivery via sectoral programming. In Guinea, a more consistent reference to this objective has been made in donor strategies and project documents since the mid-1990’s.

Still, interesting experiments abound. In all three countries, donor agencies are funding poverty-oriented projects of decentralised public authorities and are supporting the capacity of local communities and civil society organisations to collaborate with the local governments. However, these interventions do not amount to a comprehensive, multilevel strategy that could fully tap the potential of decentralisation for more effective poverty reduction. For instance, in Guinea it is difficult to get an overview of the different components of support by the donor community. This is partly because some donors consider support to decentralisation as a specific area of cooperation while others see it as a cross-sectoral objective, to be addressed within different projects and programmes. Another case is the tendency for donors to concentrate their support on building the planning and financial control capacities of local representatives of the central administration. This trend has been reinforced within the framework of sectoral programmes. The technical and financial capacities of local governments receive much less attention. Yet empowering the local representatives of line ministries can undermine the accountability and democratic legitimacy of the elected bodies of local governments. Another example is the apparent lack of coherence between donor support to decentralisation processes (and improved service delivery through local government structures) and their support for NGO projects, often targeting the same sectors (e.g. health, education) and local beneficiaries.

While the link between decentralisation and poverty reduction policies is weak, the case studies suggest a number of factors that may lead to positive change. It should be stressed that in each of the countries studied, public debate is emerging around decentralisation and local governance issues. At the same time, poverty concerns have become increasingly central in donor and domestic strategies alike. At the moment, donor agencies see the potential to harness local government to the task of poverty reduction more clearly. Inasmuch as democratic decentralisation opens up space for the participation of the poor and the marginalised in local political debates, one can expect these linkages to become stronger and more explicit.

It will take time before the benefits of decentralisation become tangible for the poor. However modest these benefits may currently be, decentralisation improves chances for poverty reduction by:

• Creating local avenues to strengthen political democratisation and power sharing (Mozambique).
• Helping people identify with the nation State; broadening space for local populations to speak their own language, to elect their own political leaders and to take responsibility for their own development plans (Ethiopia).
• Stimulating institutional innovations, such as decentralised authorities and civil society organisations exploring the scope for joint decision-making and implementation of local development priorities. This may gradually lead local governments to be more responsive to the needs expressed by local communities.
• Enlarging space for donor agencies to directly cooperate with provinces (Mozambique), regions and decentralised local structures (Ethiopia) or rural and urban local governments (Guinea) in the delivery of development goods (food security) and services (education, health) that can improve the living conditions of the poor.

What role should local governments play in promoting local development? What does it mean for local government to act as a ‘facilitator’ rather than as a ‘controller’ of processes? What capacities need to be developed at the local government level to enable these new roles? What are the implications of local capacity building for central governments?

Decentralisation policies in the three countries studied formally entrust statutory regional or local bodies with varying degrees of responsibility for economic and social development, self-government and self-administration. In each of the countries, the ‘empowerment’ of local governments includes a responsibility to raise a significant part of their own revenues. As a new forum for bottom-up political decision-making, local governments are considered to be well-positioned to mobilise scarce public and private resources to priority needs of local communities.

Can local governments deliver on these expectations? The cases reveal several reasons why local governments experience difficulties in doing so:

• An unclear task division between central and local government structures;
• Weak human, material and financial capacities. This produces a vicious circle of poor performance and mistrust. Central governments ‘postpone’ the devolution of resources and authority to local government structures, arguing that they lack capacity for accountable management. Local governments are nonetheless expected to meet local demands for service provision and must attempt to mobilise local resources to deliver development goods and services while lacking the human resources and management systems to raise revenue, control private sector operators and attract investment. When citizens fail to see concrete outputs, their faith in local government remains weak, and they are often reluctant to pay rates and taxes. This perpetuates local dependency on (often erratic) resource flows from central governments and donor agencies.
• The political dominance of centrally nominated officials or civil servants over local decision-making can negate the efforts of democratic bodies to influence local development and thus undermine the legitimacy of statutory structures;
• The tendency of local and international NGO’s to occupy the gaps left by an absent State in the provision of basic services can have a similar outcome. For various reasons, constructive partnerships between NGO’s and local governments remain the exception rather than the rule;
• The poor involvement of non-state actors such as community groups, traditional authorities and the private sector in the design of the decentralisation process and local government decision making. Consequently, there is a danger that these actors adopt a ‘wait-and see’ attitude. Most commonly, they cooperate with local officials via patrimonial networks, or simply avoid any cooperation with formal authority.

• Providing vertical and horizontal information and insights to central government and other development agencies and on the specific needs of urban neighbourhoods and rural communities;
• Coordinating the mapping and mobilisation of local capacities and resources. This holds especially true for the informal private sector. If properly supported, it could play a key role in promoting local economic development, in employment creation and in the generation of an economic surplus in the form of taxes that can help local governments assume their new roles as catalysts of local development;
• Providing a domestic framework to promote the participatory formulation, conceptualisation and operationalisation of local development plans. A particular challenge is to ensure the adequate participation of women. During the consultations in Guinea, for instance, it became clear that cultural factors prevent women from participating in local development processes in a way that is commensurate with their economic role. This points to a need for more thorough gender-differentiated consultations and participation modalities.
• Ensuring the fair and equitable targeting of poverty reduction programmes at the local level;
• Facilitating the development of economic, social and physical infrastructure;
• Generating greater trust and accountability between the State and its citizens by involving local leaders, entrepreneurs and civic organisations in democratic dialogue and in the workings of local government.

In Guinea, for instance, one cannot expect relatively better-off communes in the coastal zones to operate in the same manner as communes in the poor and sparsely-populated communes of the Northeast, or those on the Liberian border that absorb a sizeable influx of refugees. The need for locally adapted structures and modalities also holds true in Ethiopia, where the decentralisation process takes different shapes and velocity in different regions. This is only logical given the huge diversity between regions in terms of financial, human and organisational capacity, as well as in terms of size and geographical scope. Regions like Tigray and Amhara have already internalised their constitutional right to self-government and, as a result of their strong organisational capacities, can harness the potentials of decentralisation development. Regions like Afar and Gambella have not arrived at this stage and are essentially run by the central administration.

The Role of Donor Agencies

How can donors support democratic decentralisation? In highly dependent aid economies, donor agencies are powerful players in the process of decentralisation. They influence the agenda for political, administrative and economic reforms. They fund local elections, and provide support to the decentralisation process. Yet donors are not accountable to local actors, but only to their domestic and international constituencies. The efficacy of donor support to democratisation is thus constrained by a classical Catch-22 situation: A high-profile intervention in a domestic political arena by a powerful, unaccountable actor can undermine the legitimacy of all actors in that arena.

To what extent have donor agencies consistently supported decentralisation with a view to enhancing poverty reduction? The three case studies indicate that donor agencies do not systematically link these two objectives in their policy statements. Yet at the same time, most donors agree that local governments can make a significant contribution to social-economic development and poverty reduction provided that 1) a genuine devolution of power takes place, and 2) that this is accompanied with clear mandates, capacities and resources. In recognition of this, donors in each of the case countries channel resources directly to decentralised authorities on an experimental basis. These measures aim, among other things to:

• Improve the local capacity for participatory planning;
• Empower local communities to articulate their interests and needs vis-à-vis local governments (policy advocacy);
• Facilitate joint action between development players in specific sectors such as health, education and the environment.

Still, the cases show that donor involvement in decentralisation is fraught with risks. For instance, the reversal of the original blueprint for democratic decentralisation in Mozambique led some donor agencies to re-orient or diminish their support to the process. Instead, donors are supporting service delivery via sectoral programmes in education and agriculture. These programmes deconcentrate resource deployment, but do so in a way that bypasses the democratic control mechanisms of statutory bodies at the local government level. Such programmes can lead to the proliferation of ‘participatory’ structures for sectoral development activities that have no formal link with the democratisation process in the country. This move may be understandable in the light of donor pressures to achieve value for money and to demonstrate short-term results, yet it can also undermine the task of building viable local governance systems.

In Ethiopia, the government insists that bilateral funds be channelled directly through the treasury and not be used to support NGO activities at the local level. In a similar vein, it invites donor agencies to buy into centrally coordinated sectoral development plans. While this may improve the overall coherence of donor interventions, it may reduce the scope for effective support to local authorities.

In Guinea, a national framework programme to promote decentralisation and the capacities of civil society was launched in October 1998. The elaboration of this programme was assisted by UNDP and preceded by an analysis of constraints to decentralisation with stakeholders at different levels. The programme aims to revitalise and strengthen the poverty focus of decentralisation, but mainly with the help of external resources. In view of the recent proliferation of framework programmes many bilateral donors fear that this initiative may remain an empty shell.

In sum, there are many questions regarding the capacity of donor agencies to support complex, highly political processes such as decentralisation. Involvement in such delicate negotiations of power requires a variety of skills, which one cannot assume to be readily available in donor agencies. Such skills include the capacity to:

• Analyse and make sense of the quickly moving politics of decentralisation;
• Develop long-term institutional development strategies in support of local governments;
• Broker new partnerships between local governments, non-state actors and communities;
• Prevent sectoral planning from simply institutionalising the deconcentration of tasks within a line ministry structure without a genuine devolution of authority to statutory bodies;
• Facilitate national dialogue processes involving the different actors and stakeholders on the content and operational modalities of the decentralisation process, avoiding the imposition of foreign models of local governance.
• Review the role and the funding modalities of NGO’s involved in basic service delivery, ensuring a linkage with local government action;
• Reconcile the risk-taking nature of decentralisation policies with the requirements and procedures of bureaucratic accountability;
• Monitor the impact or at least the medium-term effect of aid to the decentralisation process from a poverty perspective.

1. Link support to decentralisation with support for poverty reduction

It is in the interests of both donor agencies and their partners to coordinate actions targeted at the twin objectives of democratisation and poverty reduction. The case studies suggest that there are good reasons for such an explicit linkage. First, it is now widely agreed that poverty is closely linked to political factors such as access to power and resources and the accountable and transparent mana-gement of local affairs. A genuine devolution of resources and authority can create openings for local communities, traditional leaders, private sector operators and NGO’s to become more fully involved in local development processes. In other words, democratically-controlled local governance systems are a precondition for poverty reduction.

Second, much development potential remains dormant at the local level. Most stakeholders concur that efficient local governments can play a useful role as a catalyst and coordinator of bottom-up development initiatives. A process of decentralisation that best serves poverty reduction is one that combines the strategies of political empowerment, resource mobilisation and enhanced service delivery in a coherent and balanced mix.

2. Adopt an institutional approach that privileges statutory structures

If the development of viable local governance systems is seen as a priority task, it logically follows that the decentralisation policies of donor agencies need a long-term institutional vision. This implies investing in local governments as an essential element in the institutional framework for development. Because the process of decentralisation is highly political, fragile and risky, this long-term institutional perspective is crucial. Hence, donors should resist the tendency to move away from local government support simply because central government shows little commitment to decentralisation, or because of short-term efficiency concerns. Poverty is a complex phenomenon. The institutional expertise and capacities of local governments to effectively address this issue should not be expected to develop within a few years. Providing and monitoring support on a long-term basis and taking time to develop locally adapted approaches may sometimes be more important than the volume of aid provided.

It is also vital that the institutional focus of external support is compatible with the generally accepted principles of good governance. When donor programmes bypass statutory structures at the local government level (often in name of short-term efficiency), this taxes the credibility of the legal framework of democratic governance. Democracy requires participation, but not all channels of participation are equally democratic. Donors should be rigorous in their respect for statutory bodies when promoting decentralisation, but also in any other institutional arrangements at the local level.

3. Invest in both local governments and civil society

While strengthening the institutional capacity of local governments is vital, the risk of re-creating top-down bureaucracies at the local level is also real. Donor agencies are therefore advised to follow a dual track and to invest in the capacity of civil society actors to exact greater ownership of the decentralisation process — and its potential benefits in terms of poverty reduction. This means improving the access of civic actors to information on the roles and responsibilities of local governments in local languages, promoting communication between officials and local communities, as well as strengthening the capacity of local communities to articulate their demands, and to hold elected officials accountable.

It is worth noting that channelling donor support through NGO’s is not necessarily equivalent to strengthening civil society. Funding a service-delivery NGO in isolation from local government may not promote dialogue, but rather competition and even confrontation between local authorities and civic groups. The challenge is to support civic actions at the local level while ensuring that their autonomous activities are linked, to the extent possible, with broader development processes and institutional dynamics in a given territory. This has major implications for the future of NGO co-financing systems of the European Commission or the EU Member States. The rationale for supporting initiatives stemming directly from civic organisations remains valid (especially in countries that do not pursue a serious decentralisation policy), yet ways need to be found to ensure that these activities are properly integrated in sectoral or local development plans.

4. Promote joint action

Rather than funding isolated projects for distinct categories of actors, donor agencies and NGO’s could help to engineer new relations between local governments and people. This is best achieved by involving both parties in the planning and implementation of local development plans and programmes. This offers guarantees for the consolidation of local governments (as a key player in local development), as well as for bottom-up participation and control by the population. In promoting joint action, donor agencies are well advised to call on the expertise of (local or Northern) institutions that can play a facilitating role (e.g. associations of local governments, service NGO’s, etc.)

5. Support central agencies in charge of the decentralisation process

Development from below requires development from above. In the absence of strong leadership at the central government level, little sustainable progress is likely to be achieved in the implementation of a decentralisation process conducive to poverty reduction. Political leadership and commitment is needed to introduce a coherent legal framework, to agree on a functional task division between different layers of government, to ensure the transition from a ‘controlling’ State to a ‘facilitating’ State, and to promote and facilitate the access of citizens to an effective judicial system. Donor support to central government level agencies is an essential pre-requisite for an effective linkage between decentralisation and poverty reduction.

6. Ensure coherence and complementarity

Promoting stronger linkages between decentralisation and poverty reduction is a complex task. No donor agency is in a position to provide comprehensive and coherent support on its own. Their challenge is to share expertise and to strive towards a functional coordination and division of labour among themselves and with NGO’s, based on the principle of comparative advantage.

The current situation leaves much to be desired. Donor agencies tend to come up with their own frameworks, programmes and projects, without much consultation with other players. Often these are based on the donor’s own experiences and traditions with decentralisation and governance systems. Weak consultation and a poor flow of information easily produce a multitude of approaches at the expense of a locally owned and coherent national strategy. These risks appear particularly prominent in highly aid-dependent countries. If donor agencies truly intend to achieve their central objective — poverty reduction — it is essential that policy and operational coordination is brokered to the benefit of all parties involved.

To reiterate, it is important to stress that decentralisation is not a universal panacea for all social or development problems. Territorial cohesion and a minimum level of political stability are basic preconditions for any policy of political decentralisation. Indeed, the over-zealous promotion of decentralisation can tip a dangerous balance between micro-regionalism and national cohesion in favour of civil conflict. African leaders are all too aware of this danger and this is one explanation for the snail’s pace at which decentralisation policies are put into effect.

Donors should recognise that ‘decentralisation by default’ is one of the major threats to social and political stability in Africa today. The identification of local populations with the State is largely contingent on the capacity of the government to provide access to basic social amenities like health and education. For the foreseeable future it appears that donors will play a role in ensuring service delivery to marginalised groups in a great many countries. NGO’s and local governments offer alternative, even competing, options for doing this. Each channel has its benefits and shortcomings. Given the emphasis here on promoting accountability and sustainability, it should be clear that despite the various risks involved, local governments are attractive partners.

Policy Management Briefs are intended for policy-makers in public and private organisations concerned with aspects ofinternational development cooperation. Drawing from work of ECDPM, its partners and its associates, our aim is to provide succinct and readable summaries of current issues in the management of development policy. This brief was adapted from the synthesis chapter of a joint IDS/ECDPM policy paper. [Jong, K. de, C. Loquai, and I. Soiri. 1999. Decentralisation and poverty reduction: exploring the linkages. (IDS Policy Papers, 1/99). Helsinki: Institute of Development Studies / Maastricht: ECDPM]. Copies of the original study are available from IDS, Box 59, 00014, University of Helsinki, Finland ids-finland@helsinki.fi. Financial support by the Finnish Ministry of Foreign Affairs and the European Commission (DG Development) is gratefully acknowledged.