بازبینی "قضیه ی تمرکز زدایی" ــ درباره نقش اثرات جانبی

The “Decentralization Theorem” [Oates, W.E., 1972. Fiscal Federalism. Harcourt Brace Jovanovich, New York] is central to the discussion of fiscal federalism. We revisit the role of consumption spillovers in evaluating the merits of (de)centralization. Unlike the general prediction, a higher degree of spillovers may reduce the difference in utility of centralization and decentralization. The non-monotonicity result relates to the difference in expenditures on public consumption. Provided decentralized choices yield higher levels of public expenditure, a rise in the amount of spillovers allows residents to enjoy larger gains in public consumption (and thereby utility) under decentralization relative to centralization.

The question of whether fiscal responsibilities should be assigned to a (de)centralized authority has long been debated in public economics. The discussion refers to Oates’ Decentralization Theorem (Oates, 1972) stating that in the absence of cost savings from centralization and interjurisdictional externalities, fiscal responsibilities should be decentralized. This argument implicitly assumes that the center is unresponsive to preference heterogeneity and thereby is only able to implement uniform policies. More specifically, “[. . .] individual local governments are presumably much closer to the people [. . .], they posses knowledge of both local preferencesand cost conditions that a central agency is unlikely to have” (Oates, 1999, p. 1123).1 If the geographical scope of a jurisdiction falls short of the spatial pattern of spending benefits, the optimal assignment of policy tasks is deduced by trading off the welfare costs of policy uniformity against the welfare gains from internalizing spillovers in policy-making.2 Consider a country consisting of two regions which differ in their preferences for local public goods, whichexhibit regional spillovers. In this setting, fiscal decentralization allows for a better matching of public good provision to local tastes, whereas under centralization uniform provision ignores local taste heterogeneity, but internalizes spillovers. The central question to be examined in this paper is how the difference in the utility of centralization and decentralization changes with respect to the level of consumption spillovers.3 Using quasi-linear, iso-elastic preferences, the welfare difference turns out to be non-monotone in the strength of spillovers. A larger amount of spillovers may reduce the welfare differential between centralization and decentralization. The rationale for this result is that decentralization may yield higher expenditures on public goods than centralization.

The paper provides a formal treatment of how relative welfare with (de)centralized policy relates to the strength of spillovers in public consumption. Most of the discussion on the costs and benefits of fiscal federalism rests on a welfare trade-off which is taken to be monotone in the primitive of the economy. In contrast to the presumption, the analysis points to a non-monotone trade-off. A marginally higher degree of spillovers may promote the well-being of constituents under decentralization compared with centralization. The analysis reveals that a non-monotonicity of W will only arise when decentralization is welfareenhancing. The finding may not extend to models of fiscal federalism which differ from the specification adopted in the paper. Suggestively, a non-monotone sign of W may arise in models in which decentralization generates distortions beyond the failure to internalize spillovers or in which centralization exhibits allocative advantages in addition to the internalization of spillovers. In these cases, theW-curve potentially shifts upward and multiple crossing points with the 0-line may exist. We leave a rigorous analysis of the reasoning to future research.