Co—movement And Return Predictability On Style Investing

Due to the limitation of the individual cognitive competence, it is unreality to construct the portfolio from a huge amount of the stocks. To simplify the investment process, investors have a tendency to category the stocks according to different characteristics, and then construct portfolios according to different categories. This investing process is known as the style-investing. In the practice, the style-investing is easy for investors to follow, and is also useful to evaluate the fund managers’ performance. Therefore, it gains the comprehensive attention in the academia and operation circle.Barberis and Shleifer (2003) have conducted a theory on style-investing, and two conclusions were put forward based on their research. Firstly, style-investing behavior generates both style and asset-level momentum. Investors can gain abnormal returns with the style momentum strategy. Secondly, style-investing behavior generates excess co-movement within styles. Based on these two conclusions, Wahal and Yavuz (2013) have finished the empirical analysis. They researched the role of the style-investing on the asset-level return predictability and the co-movement. According to their paper, it is proved that the style-investing behavior shows the predictability, and the momentum strategy adjusted by the co-movement can gain the abnormal return. High co-movement momentum portfolios have significantly higher future returns than low co-movement momentum portfolios.In China, the style-investing behavior has caused the dispute. Based on the monthly stock trading data in China A-share stock market from 1995 to 2004, Xiao et.al. (2006) agreed that there were statistically significant profits with the style momentum strategy. After the amendment on the Xiao’s empirical method, Wang et.al. (2007) insisted that the style momentum effect did not exist in China A-share market,-which was consistent with the volatility characteristics in China stock market.This essay focuses on two effects of the style-investing, the style momentum effect and the co-movement effect, based on the weekly stock trading data from 1997 to 2014 in China A-share market. Firstly, with the Fama-Macbeth regression, the predictability of the style-investing will be tested. Secondly, adjusted by co-movement, Jegadeesh and Titman (1993) momentum model will be applied to see if the style-investing can generate the co-movement effect.From the empirical test on the predictability and the co-movement effect of the style-investing, this essay reached the following conclusions. Firstly, the style-investing behavior shows a certain degree of the predictability in the sample period of the 1 to 3 weeks. Secondly, the small-cap and large cap stocks show a different predictability. As for the small-cap stocks, its predictability of the style-investing is statistically significant within one week, while the large-cap stocks’predictability of the style-investing is statistically significant in 1 to 3 weeks. Fourthly, the predictability of the style-investing before and after the non-tradable shares reform are different. Fifthly, the style-investment behavior is one reason of the momentum effect or reverse effect in China. Sixthly, the co-movement of the style-investment in China is different from that in America. In China, High co-movement momentum portfolios have significantly lower returns than low co-movement momentum portfolios. That is, the higher the co-movement of the stocks is, the lower the future momentum return is.The innovations of this essay are:Firstly, it is the first paper to study the style-investing behavior in China A-share stock market based on the weekly stock trading data. Secondly, this essay focuses on two effects generated by style-investing behavior, the style momentum effect and the co-movement effect. Thirdly, this essay studies the style-investing behavior before and after the non-tradable shares reform. Fourthly, this essay researches the return of the co-movement momentum portfolio controlled by the different stock cross-section characteristics.