Blockholder Exit Threats in the Presence of Private Benefits of Control: A Quasi-Experiment from Chinaas Split-Share Structure Reform

Abstract

Using the Split-Share Structure Reform (SSSR) in China as a quasi-experiment, we examine whether and how non-managerial blockholder exit threats discipline managers when private benefits of control (PBC) are prevalent. SSSR removed the selling restrictions on previously-dominant non-tradable shares, thereby providing a large, exogenous, and permanent shock to the cost for non-managerial blockholders to exit. However, because such exit potentially reduces non-managerial blockholders’ threats to control rights and managerial blockholders could benefit from the exit, it is ex ante uncertain whether exit threats would be effective in this setting. Using a difference-in-differences design, we find that the shock to exit threats, on average, improves firms’ operating performance. Specifically, we find that firms whose non-managerial blockholders experience a larger exit-threat increase have a greater improvement in performance than those whose non-managerial blockholders experience no or a smaller increase. Within the treatment sample, better performance is observed for firms whose managers’ wealth is more closely tied to stock prices. Our results are robust to numerous controls, different specifications, and alternative measures of operating performance. The findings suggest that exit threats discipline managers even when PBC are prevalent. Further cross-sectional analyses show that the governance effect of exit threats is weaker when PBC are a stronger concern. In particular, the governance effect for firms headquartered in coastal provinces and with less concentrated ownership (indicators for lower PBC) is significantly stronger than that for firms headquartered in inland provinces and with more concentrated ownership. Finally, we identify improvement in investment efficiency and reduction in agency costs as two potential channels through which the shock affects operating performance, and we show that the changes in performance occur around the time when the expectation of a future increase in blockholders’ exit threats is formed.