The aim of the paper is to analyze a simple model of local public good provision with positive interjurisdictional spillover effects - as the case of environment protection spending - comparing decentralized and centralized system, when spending and taxation decisions are made by nonbenevolent politicians. As in the recent so-called Second Generation Theory (SGT) of fiscal federalism (Seabright, 1996; Besley and Coate, 2002; Lockwood, 2002; Oates, 2005, Weingast, 2009), we adopt a political economy approach to look at the trade-off between centralized and decentralized provision of local public goods. The main differences between our paper and theirs are that we model the public good taking into account two important aspects: the size - in terms of population - of local jurisdictions providing it, which is relevant for the scale effect in the financing mechanism of non-rival public goods; the detail of political opportunistic behaviour introducing a "rent equation" directly into the model to represent the additional gain of "non-benevolent" politicians, who levy higher taxes than the costs of the public goods. Considering these two elements, our results appear to be partially different from the SGT. In particular, the convenience of having decentralization versus centralization changes with the degree of spillovers and the size of regions. Three elements have to be considered: i) the implicit transfers ("cross subsidiation") from high scale economy regions to low scale ones; ii) the free-riding gains in receiving positive externalities; iii) the gain of internalization of externalities. When spillovers linked to public goods provision are low, only the first item is relevant. Thus, smaller regions prefer the centralized solution, since through it they can charge bigger regions for some costs of production. On the contrary, bigger local jurisdictions would like decentralization.When beneficial spillover effects increase (and many regions producing them), the other two factors start to play a crucial role, and the opposite situation takes place. The basic trade-off is between the internalization process and the free-riding tendency, whose efficiency gains are different for large and small local jurisdictions. Hence, from a positive viewpoint, decentralization should not be necessarily pursued only in the absence of externalities, but it depends on the relative size of the local jurisdictions.
