Document Type

Working Paper

Publication Date



This paper contributes to the existing literature on the determinants of fiscal decentralization by motivating theoretically and exploring in depth the empirical relevance that geography has as determinant of fiscal decentralization. The relationship between decentralization and geography is based on the logic that more geographically diverse countries show greater heterogeneity among their citizens, including their preferences and needs for public goods and services provision. Communications and physical distance are also a very important issue and play a key role on the effect of geography over time. The theoretical model builds on the work by Arzaghi and Henderson (2002) and Panizza (1999). For the empirical estimation we use a panel data set for approximately 91 countries for the period 1960-2005. Physical geography is measured along several dimensions including elevation, land area, and climate. We construct a geographical fragmentation index and test its effect on fiscal decentralization. In addition, we interact the geographical fragmentation index with time variant infrastructure variables, in order to test the effect that infrastructure and communications have on the relationship between geography and fiscal decentralization. For robustness, we construct Gini coefficients for in-country elevation and climate. We find a positive and strong correlation between geographical factors and fiscal decentralization. We also find that while the development of infrastructure (in transportation, communications, etc.) tends to reduce the effect of geography on decentralization, this effect is rather small and mostly statistically insignificant, meaning that the impact of geography survives over time. The additional value added of this strategy is that geography and its interaction with infrastructure development may be used as an instrument for decentralization in future econometric estimations, where decentralization is used as an explanatory variable but it may be suspected to be endogenous to the economic process being studied (economic growth, political instability, macroeconomic stability, income distribution, etc.)


International Center for Public Policy Working Paper Series #1211, Andrew Young School of Policy Studies, Georgia State University.

Included in

Economics Commons