Conclusion
The influences of political connections on firm performance have been examined intensively in the literature. However, previous studies find mixed evidence concerning the relationship between political connection and firm performance. Moreover, defining political connectedness is difficult. This study uses government officials' corporate site visits as a measurement of political connections, which enables us to identify the exact timing and strength of the political connections and capture the dynamic nature of political connections. We find that government officials' site visits are not random and the number of visits by officials increases over time. Firms with better past performance, more assets, higher financial leverage, younger age, and low largest shareholder's holding are more likely to receive government officials' visits. Also, firms located in provinces with a higher GDP growth rate are more likely to receive official visits. We examine the effect of officials' site visits on firm performance and find that firms' return on assets, return on equity, and total factor productivity increase following officials' site visits, suggesting that the political connection formed through officials' site visits are beneficial to firms. Moreover, firms gain more access to investment projects and bank loans. Firms' corporate governance improves, and information asymmetry decreases following officials' visits. The effect of government officials' visits is stronger for nonSOEs than SOEs, suggesting that corporate on-site visits by government officials is more valuable for firms that lack political connections. Additionally, the stock market reacts positively towards government officials' corporate site visits, suggesting that investors interpret official visits as government endorsement and support. We also provide evidence that the effect of political connection is stronger in the absence of political corruption, indicating that political corruption is not the precondition for political connections to be valuable.
