Available online at International Tax & Public Finance is “Switch toward tax centralization in Italy: a wake-up for the local political budget cycle” by Massimiliano Ferraresi, Umberto Galmarini, Leonzio Rizzo, and Alberto Zanardi. Here is the abstract:

Are the incentives to expand expenditure before local elections affected by the composition of local governments’ revenues? We explore this issue by exploiting the Italian government bill that in 2008 replaced the municipal tax on main residence with a vertical transfer. Relying on staggered dates of municipal elections to identify the effect of the reform, we find evidence of a political budget cycle, but only for municipalities that in 2008 were in their pre-electoral year. The result suggests that a lower degree of municipal tax autonomy strengthens the incentives to expand the size of the budget before the elections.

You get a sense of the identification strategy from the abstract, but it seems worth noting that the paper finds support for a political budget cycle in the broadest terms — that the revenues and expenditures are correlated with election cycles — than any specific political budget cycle theory. In particular, they find that after centralization both current expenditures and revenues from charges/fees increased by equivalent amounts. Under some theories of political budget cycles, governments become more deficit prone as they cut taxes and increase spending just prior to the election cycle to protect incumbents, but that isn’t obviously occurring here. The authors seem to think it is a relative change in the political price for own source revenues that translates into spending:

Our interpretation of the result is that the primary channel through which the reform prompted incentives for expanding the budget in pre-electoral years was by lowering the costs for financing expenditure increases with own revenue sources. Before the 2008 reform, the three main sources of own revenues of Italian municipalities were the property tax on main residence, whose burden is entirely borne by residents who are also voters at the local level, users’ fees and charges, whose burden is mostly borne by residents directly benefiting from municipal services, and the property tax on additional residence, whose burden is in some cases mostly borne by non-residents, as in municipalities located in touristic areas. Since taxing the main residence is highly unpopular, before 2008 most municipalities applied tax rates on main residence just above the minimum level imposed by the central government, while they exerted more effort in the taxation of additional residences and in the determination of fees and charges, though the room to rely heavily on the latter was narrow, since residents were already taxed on main residence. Under these conditions, local administrators had little capacity of making leverage on own revenues to finance expenditure expansions for electoral purposes. The 2008 reform changed the setting. By relieving local administrators from having to impose a burden on residents on their main property, and by substituting the revenue loss with a compensating transfer, the reform considerably reduced the political costs incurred by local administrators to make leverage on user fees and charges to finance pre-electoral expenditure hikes.

I’ve previously blogged recent research on PBCT in the post Research on Overlapping Political Budget Cycles.