Inequality, Clientelism and Political Budget Cycles: The Case of Italy

This paper argues that inequality, by increasing the electoral rewards of clientelistic practices for incumbents, affects the size of Political Budget Cycles (PBCs). Specifically, clientelistic practices are especially successful in swaying voters at the lower rungs of the distribution ladder, and therefore incumbents will resort to them more in political systems characterised by relatively high inequality. Moreover, as incumbents want to be re-elected, clientelistic spending will concentrate around elections, thus increasing the size of PBCs. By testing this argument on an original dataset of the Italian regions, we shed light on the causal links both between clientelism and inequality and between clientelism and fiscal policy. We assess the impact of elections on the size of health spending, the largest component of regional expenditure in Italy, over 20 years. As a proxy for the distribution of life chances across voters we use educational attainment. We focus on the inequality in educational attainment rather than on income inequality because, due to the redistributive practices of the central government, the latter varies across regions less than education levels. Moreover, and relatedly, using education mitigates the endogeneity problems that arise from the impact of government spending on income. In order to increase the causal leverage of the empirical evidence, we assess the impact of different levels of education across the regions on two measures of health spending: general spending and personnel spending. Based on the connection between inequality and clientelistic practices, we expect education to be negatively correlated with personnel spending, a typical area of clientelistic spending, but not with general spending. We indeed find significant increases in personnel spending, but not in general spending, close to elections in regions where voters have lower levels of education.