Centralisation versus Decentralisation of Public Policies: Does the Heterogeneity of Individual Preferences Matter?

This paper explores the role of the heterogeneity of fiscal preferences in the assignment of policy tasks to different levels of government (decentralization vs. centralization). With reference to a sample of European countries, a median-voter mechanism of collective decision is assumed to work at both a national and a supranational level. Using data from a large international survey (ISSP), a series of econometric models has been estimated in order to make individual attitudes representative of different categories of public expenditures and of different countries. The dominance of decentralization over centralization or vice versa is determined on the basis of the utility loss that each individual suffers in connection with the distance between his own most preferred level of public expenditure and that one chosen by the national/supranational median voter. The main finding is that, differently from the predictions of Oates\u2019 decentralization theorem, the assignment of responsibilities at the supranational level (centralization) for a number of public expenditure programs (healthcare, education, employment benefits) dominates (or is close to dominating) decentralization, even in the absence of economies of scale and interregional spillovers. However when the possibility of inter-jurisdictional mobility is explicitly considered, in line with the predictions of Tiebout\u2019s model, decentralization dominance becomes more and more substantial and also prevails in the sectors where, under the non-mobility assumption, the assignment of responsibilities at the supranational level is efficient