The Research Of Momentum And Contrarian Effects In China's A-share Stock Markets

The efficiency of financial market is a long-debated issue in finance field. The efficient market hypothesis, price of stocks will react to new information sufficiently and instantaneously, which implies investors cannot get abnormal return using strategies based on previous prices. However, during the last two decades, through extensive empirical studies on stock markets throughout the world, researchers identified the momentum effect and contrarian effect readily existing in these markets, which means that investors can make abnormal returns from following momentum strategies and contrarian strategies. These findings contradict with the efficient market hypothesis.Many researchers also analyzed the momentum and contrarian effect in China's stock market. But they got different even conflicting results when using different data sets, different time periods or different empirical methods. And they seldom gave theoretical models to justify their empirical findings. Therefore, a more in-depth analyses on momentum and contrarian effect in China's stock market can provide an insight on the distinct characteristics of China's stock market, together with advice for investors.After surveying the relevant literature, we constructed a model to reflect the distinct feature of Chinese investors. The model predicts an ultra-short-horizon momentum effect, a short-horizon contrarian effect, and an intermediate-horizon momentum effect. Using the daily transaction data of all"A"share stocks in Shanghai and Shenzhen stocks markets between Jan. 1st, 2000 and May 30th, 2008, we find the prediction of the model is supported by the data. Further analyses confirm the result is robust to bid-ask spread. Another check by lagging one year between ranking period and holding period doesn't support the hypothesis that momentum/contrarian effect is due to differences across stocks, but support the hypothesis, as in our theoretical model, that the momentum/contrarian effect is due to the autocorrelation of return time-series because of under-reaction and over-reaction to stock-specific shocks.