A government policy aimed at the reduction of state shares in state-owned enterprises (SOE) triggered a crash in Chinas stock market. The sustained depression and spillover even after the policy adjustments were over constitute a puzzle the so-called state-share paradox. The empirical study finds evidence in two dimensions. First, a regime switching model with an absorbing state suggests that government policy switches the regime to liquidity black holes. Second, there is no evidence of light-to-liquidity during the crash, suggesting to model the crash as an aggregate phenomenon of the whole market. To carefully match the evidence, a theoretical model is set up within the framework of market microstructure. The state-share paradox is not a simply instance of news-driven crash. The model shows that Chinas stock market has distinctive features of liquidity production and price discovery. The irregularities of a representative liquidity supporter generate an inverted-S demand curve and give rise to potential liquidity black holes. Multiple equilibria and the resulting large drop in prices arise from supply dynamics of short-run investors, who buy the stock from the primary market liquidate their long positions in the secondary market. This study contributes a rational panics hypothesis to the literature. The rational panics hypothesis is neither an rational model with noise traders, nor a standard rational expectation model under the asymmetric information framework. It is based on homogeneous agents with incomplete information, and is consistent with the evidence of absorbing regime switching and the recent literature on state-dependent preference. Our findings have larger implications for ine¢ ciency of Chinas stock market.

A government policy aimed at the reduction of state shares in state-owned enterprises (SOE) triggered a crash in the Chinese stock market. The sustained depression and spillover even after the policy adjustments were over constitute a puzzle---the so called "state-share paradox". The empirical study finds evidence in two dimensions. First, a regime switching model with an absorbing state suggests that government policy switches the regime to liquidity black holes. Second, there is no evidence of flight-to-liquidity during the crash, suggesting to model the crash as an aggregate phenomenon of the whole market. To carefully match the evidence, a theoretical model is set up within the framework of market microstructure. The model shows that the Chinese stock market has distinctive features of liquidity production and price discovery. The irregularities generate an inverted-S demand curve, gives rise to potential liquidity black holes, and are key features to explain the state-share paradox. This study contributes a rational panics hypothesis to the literature. The rational panics hypothesis is neither a herding model with or without behavioral assumptions, nor a standard rational expectation model under the asymmetric information framework. It is based on homogeneous agents with incomplete information, and is consistent with the evidence of absorbing regime switching and the recent literature on state-dependent preference. Our findings have larger implications for theoretical modeling and policy design.