• A novel dataset to measure local corruption in China that controls for officials’ rank, tenure, investigation reason and corruption latency. • Corruption has a significant negative effect on manufacturing firm TFP, conditional on firm and city heterogeneities. • Publicly owned, export-oriented, profitable, faster growing firms and firms in concentrated industries are less sensitive to corruption. • Firms in cities with higher levels of human capital and public spending on education and scientific research are less sensitive to corruption. • Corruption also hurts private ownership, investment rate, export intensity, innovation, leverage, employment growth, and profitability. This paper examines the effects of local corruption on total factor productivity (TFP) of manufacturing firms in China. The empirical analysis is based on a novel corruption dataset we developed on local corruption in China at various disaggregation levels. The empirical results using fixed effects and instrumental variable estimation methods suggest that corruption has an economically and statistically significant negative effect on firm productivity. The estimated economic cost of corruption is found to be high; a one standard deviation increase in corruption reduces firm TFP by around 3.8%. We also find that firm heterogeneity shapes business reactions to corruption in a given geographical location. Increasing corruption hurts firms less when they are publicly owned, export-oriented, more profitable, have faster growth, or operate in industries with lower levels of competition. We also show that firms in cities with higher levels of human capital and higher levels of public spending on education and scientific research are less sensitive to corruption. As for transmission channels, we find that corruption is likely to hurt TFP through its negative effects on private ownership, investment rate, export intensity, innovation, leverage, employment growth, and profit.
Read full abstract