Electing Educated Leaders during Democratization: Evidence from Indonesia

Using manufacturing plant-level census data from Indonesia, we show that the effect of democratization on manufacturing performance crucially depends on the education level of the newly elected local leaders. In districts that elect a mayor without college education, employment drops by five percent in the first few years after democratization, while employment stays constant under college-educated mayors. We also identify mechanisms: manufacturing plants in districts with non-college educated mayors face a much larger increase in local taxes, but also worse provision of local infrastructure and no extra spending on other public goods. A novel hand-collected dataset on corruption cases further suggests that democratic mayors without a college degree are more corrupt. Our estimates are plausibly causal since the year of local democratization varies exogenously across districts, and districts with different mayor education levels exhibit parallel trends in manufacturing prior to democratization.