This paper dissects the lottery-like anomaly in Chinese A-share stocks by decomposing total stock returns into overnight and intraday returns. Our findings indicate that the negative overnight returns are concentrated among lottery-like stocks, and the lottery-like anomaly is mainly driven by the overnight returns component. Considering the unique Chinese institutional features, our mechanism analysis reveals that the overnight returns induced lottery-like anomaly is more pronounced in stocks with high retail investors' gambling preference and high limits of arbitrage. Overall, our results suggest that investors optimism and trading constraints have a substantial impact on market efficiency in China.

Using data on stock trading and accounting information from 2000 to 2018, we construct 426 anomalies and propose the multiple hurdle of 2.85 in the Chinese A-share stock market. With single sort portfolio analysis on value-weighted returns, we find that 98 (27) anomalies have significant raw returns at the 5% level with absolute t-value larger than 1.96 (2.85). After risk adjustment using the Liu, Stambaugh and Yuan (2019) three-factor model, 16 (2) anomalies have significant alphas for single (multiple) tests, about half of which are based on liquidity information, while alphas for accounting anomalies are less significant. After regressing on the four-factor model with turnover, the liquidity anomalies become insignificant. We construct the composite anomalies, and find that the majority can pass the multiple test hurdle.