We consider a country whose government provides a bundle of goods and services through a multi-tier administrative organization. We compare the optimal architectures of public governance (i.e. the division of the state into several tiers, the distribution of services among them, their number of jurisdictions and the performance ability of their administrations) of two governments, one centralized and the other decentralized. Under a decentralized government, national and subnational decision-makers only consider the impact of their decisions on the welfare of their constituents, neglecting other tiers' policy. The resulting architecture is generally different from the (first-best) centralized one, and depends on how citizens weight the performance ability of the administrations and the range of goods they provide. If the relative weight on the performance ability is large, the decentralized architecture entails more tiers, less jurisdictions per tier with reduced scope of services than the centralized one, and the reverse if this relative weight is small. We use our results to estimate this weight on U.S. data. We find that the country exhibits two zones (“Northeast & West” and “Midwest & South”) where the estimated values are statistically different.
decentralization fiscal federalism public governance