Electoral Control under Decentralization: Decentralization as unbundling of public goods provision.

This paper addresses the question of whether a decentralized government is subject to a stronger level of electoral control than a centralized government. When electoral control is strong an incumbent investing a low level of effort in providing public goods will face a serious threat of being voted out of office. This threat should provide the incentives to the incumbent to exert effort in order to be re-elected as shown by Barro (1973) and Ferejohn (1986). According to the literature decentralization should increase electoral control due to the fact that under centralization the incumbent only needs to please the half plus one of the electorate in order to be re-elected. This paper presents analytically two new sources of differences in electoral control: assuming that public goods can be classified in lower tier public goods (e.g. sub-national or local level) and upper tier public goods (e.g. national public goods), then under centralization there are potential advantages derived from bundling the provision of both types of public goods, whereas under decentralization there are potential advantages derived from a clear delimitation of the responsibilities of the provider of each type of public good. We show that the trade-off depends on the probability distribution of the shocks and on the size of these shocks.