Association for Public Policy Analysis & Management

Decentralization of education to cities is theorized to have positive effects on educational performance. However, little is known about the effects of decentralization on educational inequalities or segregation within cities. Furthermore, the literature has not explored potential mechanisms that moderate the relationship between decentralization and educational inequalities and segregation. This article aims to establish the effect of decentralization of education on a set of measures of educational inequalities and segregation within cities. Moreover, it tests the moderation effects of cities’ local state capacity and political competition. The analysis uses a regression discontinuity design that takes advantage of a population threshold to grant autonomy with respect to the education system to Colombian cities in order to identify the effects of autonomy on inequalities. The study finds that autonomy increases educational segregation, as measured by the intraclass correlation of math and language scores in two periods of analysis (2005-2013 and 2015-2016). Autonomy does not have a substantial effect on other measures of inequality (i.e. Gini index and standard deviation of scores, and scores’ gaps by mothers’ education). Furthermore, the study does not find any moderating effects of state capacity and political competition. Results suggest that decentralization reforms should address the potential for political capture of local authorities and establish mechanism to safeguard for that potential.