This chapter introduces an examination of the recent in the literature on fiscal federalism and empirical assessments of decentralisation processes in Bolivia, Uganda, Poland and Spain. It argues that decentralised approaches to development are least likely to succeed where they are most needed – where levels of inequality are high. Particular attention needs to be paid to: 1) the risk of local capture; 2) partial decentralisation; and 3) fiscal institutions (especially the assignment of functions, the structure of intergovernmental transfers and limits on local debt).
Most countries have experimented with decentralisation reforms in the last three decades. Alongside these are other developments, such as contracting between different levels of government and the proliferation of intergovernmental coordinating bodies. In some countries, the newer institutional forms have been introduced without the preconditions for their effective use being in place. These preconditions include responsive legal systems and modern budget frameworks that can identify outcomes.
Traditional fiscal federalism theory (on which many empirical approaches to decentralisation have been based) has proven to be a poor source of policy advice because it does not consider political dimensions. This has been particularly problematic for developing countries, where the burden on local government is increased by the task of implementing poverty alleviation policies, often with very unclear boundaries or accountability. The advantages typically associated with decentralised approaches to development are in fact unclear. Multilateral donors’ concerns with decentralisation can lead to community preferences being overwritten, and elites may be able to divert funds. Further, the success of participatory projects is dependent on information flows among stakeholders and on social stratification in communities (affecting how power is distributed).
Partial decentralisation can occur if local governments lack control over tax instruments at the margin. They then remain dependent on central government grants and are less accountable to their electorates. Partial decentralisation also occurs when weak constraints are introduced on borrowing, increasing the chances that the local government will have to be bailed out by the central government. Other findings are that:
- Differing incentives can make problematic overlapping responsibilities between different levels of government and the design of effective special purpose programmes (financed by central and implemented by local government).
- In many decentralisation processes, village-level governments have limited access to tax bases and often no ownership of a clear policy space. This lowers incentives to pursue effectiveness and facilitates clientelism.
- The benefits of decentralisation seem to be less evident in more advanced developing countries.
It is important to check that a decentralised system of government is able to provide services effectively and that these are less expensive, of better quality, or closer to citizens’ preferences than those provided by central government. The effectiveness of decentralised poverty alleviation can be enhanced by better design of intergovernmental financial relations and of specific grant programmes. Also recommended are:
- Actionable contracts between central and local government, based on specific outcomes.
- Competitive political processes, the free flow of information, a network of NGOs and the reduction of inequality. (These can reduce the risk of elite capture.)
- A consistent decentralisation strategy. If, for example, devolution of tax responsibilities does not follow devolution of expenditure functions, local politicians will lack incentives to ensure accountability.
- Making central-local government transfers conditional on policy reforms or past performance. Designing a multi-year budget framework, including multi-year appropriations, would allow competition across lower-level jurisdictions, involving both rewards and sanctions.
