Public Displays of Alignment - Jaya Y. Wen

Speaker: Jaya Y. Wen, Harvard Business School

Location: online or at Kiel Institute for the World Economy, Chausseestraße 111, 10115 Berlin

Organizers: Kiel Institute for the World Economy, CEPR

Registration: Please sign up to receive the newsletter and zoom links for the seminar. If you plan to attend in person, please let us know here beforehand.

Abstract: Why do firms use rhetoric to publicly align themselves with governments? We study this question in autocracies, where firms frequently echo regime language in public communications. We develop a repeated-game model in which public alignment can arise as noise, government consumption, or screening, whereby governments identify which firms are best able to support them with costly political tasks. The model shows that each mechanism can be sustained in equilibrium and generates distinct empirical predictions for firms' exposure to political shocks, their behavior during periods of political stress, and how alignment responds to regulatory investigations. To test these predictions, we construct a text-based measure of public alignment using annual reports of listed Chinese firms. We extract phrases that distinguish Chinese Communist Party language from standard Mandarin and aggregate their use to the firm-year level. Empirically, firms with higher public alignment experience larger stock price declines following major political scandals and expand employment more, or contract it less, in response to local labor unrest. Public alignment also increases after the onset of regulatory investigations. These patterns are robust to controlling for state ownership, political connections, Party cells, and other correlates of state proximity. We also show that our measure captures firm behavior distinct from existing measures of political risk. Taken together, the evidence points to screening role for public alignment, in which it helps sustain cooperation between firms and an autocratic regime under limited commitment and private information.