رانتخواری ، سرریزها و مزایای تمرکز زدایی

In the presence of spillovers, decentralized provision of local public goods may lead to a higher surplus than centralized provision even though localities have identical preferences. Indeed, free-riding costs associated to decentralization can be lower than the costs of rent-seeking and influence activities under centralization. Actually, centralization yields a higher level of regional surplus only if both the spillover effect from local public spending is sufficiently large and the elasticity of the influence function is sufficiently small.

The existing literature on fiscal decentralization can be divided into three branches. The first branch, pioneered by Tiebout [22], emphasizes the benefits of competition among local jurisdictions. The basic idea is that, under a head tax-regime, competition for mobile citizens should match bundles of public goods to citizens’ preferences more accurately. Tiebout further argued that in a system with many jurisdictions, competition among them would ensure an efficient outcome both in the production of local public goods and in the distribution of total population overcommunities.1 The Tiebout sorting equilibrium, however, crucially depends on a number of unrealistic assumptions one of them being the absence of spillovers among localities. The second branch of the literature on fiscal decentralization focuses precisely on the spillovers among jurisdictions resulting from the mobility of the tax base. The standard analysis assumes immobile citizens but mobile capital that is taxed by local governments to finance public good provision. Mobility of capital across borders raises the marginal cost of public funds perceived by each jurisdiction since it reduces the available tax base. As a result, suboptimal levels of tax rates and of local public goods emerge in a decentralized equilibrium.2 In general, however, this literature abstracts from any heterogeneity between consumers. Brueckner [6,7] was the first to attempt to reconcile these two branches of the literature by analyzing a tax-competition model where consumers are heterogeneous and can move from one jurisdiction to the other. One of his main conclusions is that the dispersion of preferences is a critical factor to evaluate the relative performance of fiscal or political decentralization.

of local public goods. We show that, in the presence of cross-boundary externalities stemming from locally provided public goods, centralization may not be desirable due to the conflict for political influence it may generate. Indeed, centralized provision of local public goods implies cost sharing and then creates a conflict of interest between citizens in different regions. This distributive conflict exists even though all regions have the same preferences for public spending and is reflected in rent-seeking or influence activities. The cost of the rent-seeking conflict must then be balanced with the benefits of internalizing spillovers to determine which regime performs better. When the influence function is sufficiently elastic, it is shown that decentralization always dominates centralization. This is because the prisoner’s dilemma of rent-seeking, under a centralized system, is too costly relative to the costs of the free-riding incentives under a decentralized system. If, however, the marginal influence of rent-seeking activities is sufficiently small, the performance of centralization relative to decentralization depends upon the size of the spillover effect. In this situation, we obtain a standard conclusion but from different reasons than emphasized in the literature. Throughout, we have assumed that, under centralization the costs of providing public goods are shared equally among regions. This is justified on empirical grounds since most centralized systems of government operate (roughly) according to such rules.14 This is the case, as in France for example, even though the central government does not redistribute equally tax revenues among jurisdictions or groups through the provision of differentiated levels of public spending to account for the heterogeneity in the economy. But, in this paper, the two localities have identical preferences and value public good consumption to the same extent. Hence, our assumption of equal cost sharing can also reflect the outcome of a constitutional stage in which the two regions must agree on the budgeting process. In turn, sharing the costs of local public spending creates a budgetary externality and a distributive conflict that does not exist under decentralization, when each region must pay for its own public good. The decentralized structure, however, has also a drawback since interregional spillovers are not internalized. This involves lower levels of public good provision than the optimal levels obtained under centralization. In other words, decentralization is represented as a market failure whereas centralization is represented as a political failure. Consequently, a potential trade-off emerges between the two systems even though regions are perfectly identical.