Do fiscal decentralization and government fragmentation affect corruption in different ways? Evidence from a panel data analysis

government fragmentation on perceived corruption. We innovate on the previous literature by considering a panel, rather than a simple cross-section of 23 developed and developing countries in the 1995-2007 time interval. Furthermore we have examined the impact on corruption of alternative ways (own revenues vs. common pool situations) in which decentralized governments finance their expenditures. Our results show that fiscal decentralization, more than government fragmentation, has the largest impact on corruption. More precisely, the reduction of common pool type of situations, by increasing the ability of sub-central government levels to selffinance their expenditures proves to be the most effective restraining force on corrupt practices. These results turn out to be robust in all specifications of the model. We also find evidence that the virtuous role of fiscal decentralization in reducing corruption is strengthened if it is accompanied by a higher degree of government fragmentation at the lowest possible tier, i.e. that of municipalities. On the other hand, reducing common pool, while at the same time increasing the number of government levels, does not show any impact on corruption; this supports the hypothesis that more complex vertical organizations of the state worsen the citizens’ ability to control public officials. These findings are interesting also in terms of policy implications, as they allow to identify the facets of decentralization that seem to be more effective in reducing corruption and improving the quality of governance.