Study On Cross-sectional Returns In Chinese Stock Market

Effective Market Hypothesis (EMH) suggests that stock markets should always reflect all available information sufficiently and correctly. However, the real systems of finance are not always in that case. We call those phenomena inconsistent with Capital Asset Pricing Model (CAPM) abnormalities, which illustrate some imperfection on the model specification and limitation of estimation method with those traditional theories. Thus, the research on capital pricing model suitable to real financial markets must be an important subject for the academic and practice fields.In this paper, we focus on the central reasons for the abnormalities deviating from classical theories, such as EMH and CAPM, including problems of model specifications and the shortcomings of the traditional conditional average regression method. The main characteristics and innovations are in the following four aspects:Firstly, we improve the Fama-French Three-Factor Model (FFM, Fama and French, 1993) , and put the volume variable into the model ( we call it FF-Volume Model), according to the close relations between trading volume and average returns. After that we do empirical analysis for the cross-sectional relationship between average returns and relative factors for all the A shares from stock market of China by combination of the FF-Volume Model. The results indicate the volume variable is positively related with cross-sectional average returns, and much more significant than other explanatory variables.Secondly, research on conditional cross-sectional relationship of returns with relative factors in two different situations, upward and downward markets, provides valuable empirical results, and overcomes the detrimental effects of putting realized returns into the FF-Volume Model instead of expected returns required by CAPM.Thirdly, behavioral finance is more and more popular in recent research on complex financial markets. In this paper, we introduce Overreaction Hypothesis (Bondt and Thaler, 1985) into analysis of cross-sectional returns, and test whether the explaining abilities of volume and book-to-market value ratio to cross-sectional returns are due to the overreaction effects of investors. The results about the reversal effects of value and growth stocks confirm the hypothesis.At last, the quantile regression method, suggested by Koenker and Bassett (1978), is introduced into our research on cross-sectional returns. Furthermore, we suggest a systematic and integrated model including firm-specific variables, market economic variables and time controlling variables. Quantile regression, compared with other explaining variables, is more suitable for the real financial markets, and provides more integrated information on the different quantiles for the cross-sectional relations.