Fiscal decentralisation, institutional quality and ethnic conflict: A panel data analysis, 1985–2001

Abstract

Fiscal decentralisation is increasing throughout the world, especially in developing countries where it is argued to foster good governance and delivery of public goods. Fiscal decentralisation is also widely promoted as an institutional device to manage ethnic conflict. Proponents of fiscal decentralisation claim that it helps accommodate ethnic minorities by granting them considerable policy-making authority. However, the empirical literature on ethnic conflict has mainly focused on federalism and political decentralisation while the few studies that have included fiscal decentralisation have produced mixed results. In this paper, I test the effect of fiscal decentralisation on ethnic conflict while emphasising state capacities as a crucial mediating variable. I assume that fiscal decentralisation is unlikely to produce any effect in countries characterised by low state capacities and weak institutions. The rationale is threefold. (i) State capacities are usually lower at the local level than at the central level; yet implementing fiscal decentralisation requires that subunits are endowed with sufficient bureaucratic and technical competences. (ii) Devolution of policy-making authority to lower tiers of governments is usually assumed more genuine in countries characterised by good governance. (iii) When state capacities are weak, ethnic groups may be tempted to claim more than fiscal decentralisation and seek independence. I assume also that minorities that are the most ethnically distant from the rest of the population are those that should benefit most from fiscal decentralisation. The system GMM estimations confirm that ethnically distinctive minorities benefit more from fiscal decentralisation. Regarding state capacities, findings are radically different with respect to the indicators that are used. Fiscal decentralisation is found to reduce the likelihood of conflict if GDP per capita is considered as a proxy for state capacity, while opposite results emerge when governance indices are used.