A Study on the Level of Market Efficiency Based on CSI 300 and 300 Constituent Stocks

This paper analyzes CSI 300 index and its 300 constituent stocks with seven market efficiency measures: autocorrelation of daily returns, autocorrelation of absolute daily returns, runs test, forecast ability of other historical data on daily return (the predictive ability of yesterday fs change of trading volume on today fs return in this paper), the return of specific trading strategy, variance ratio and pricing errors contained in daily return. We do a Principal component analysis to convert these indicators to a single indicator representing the market efficiency. Then we try to find the co-movement among different measures through correlation coefficient and among different stocks through OLS regression: market efficiency values of individual stocks are regressed on market efficiency values of CSI 300 for seven measures respectively. We found that different market efficiency measures are indeed consistent to each other to some extent and the individual stocks are somewhat consistent with the whole market indicating there is a systematic market efficiency in stock market in China. Our finding also support the idea that the market efficiency in Chinese stock market is changing all time without showing a clear upward trend from 2005 to 2020. In the end, we set three hypotheses to explain relatively high level of market efficiency in 2005, 2012, 2017 and 2019: the ability of market detecting and reacting to pricing errors, public information or private information is becoming quickly and accurately. We found that when the market is in a bad condition, the market contains more pricing-errors in daily returns and the ability of market detecting and reacting to private information is also bad.