A Study On The Characteristics And Intrinsic Formation Mechanism Of Cross-Sectional Returns On Chinese Stock Market

The cross-section of stock returns are anomalies in financial market which have greatly been challenging modern finance theory system. It has been becoming a hot issue in finance field for financial economists how to explain the financial anomalies. The article systematically investigated the basic characteristics,variational rule and intrinsic formation mechanism of cross-sectional returns on Chinese stock market based on the combination of modern qualitative methods and quantitative methods from the theoretic and empirical aspect. Concretely the main research contents and the innovations are listed as follows:1. An empirical analysis of the basic characteristics,robustness and variational rule of cross-section of Chinese stock returns based on sample data from June 1995 to December 2005 found which Chinese cross-sectional stock returns was not robust and emerged from some variational rule; An empirical conclusion was different from the foreign conclusion that size effect existed but was not significant and book-to-market ratio effect existed and was significant in China.2.The analysis methods of portfolios grouping and Fama-MacBeth(FM) regressions of the cross-section were used to analyse empirically the relations between different cross-sectional returns on Chinese stock market. The empirical results showed that BEME(book-to-market equity) and LNME(log market equity) are the most important two variables that affect the cross-section of Chinese stock returns. However the relation between BEME and LNME was faint in China, though the relation was stronger in America.3. The unconditional pricing models which have currently been utilized to explain the cross-sectional returns, including CAPM,Fama French three-factor model and characteristics model and so on, were used to explain the cross-section of Chinese stock returns, and the empirical result of being different from America showed these models could not explain perfectly Chinese cross-sectional returns though these models could explain Chinese cross-sectional returns on certain extent.4. Based on the empirical relations between different Chinese cross-sectional returns, unconditional Fama French three-factor model was improved by constructing new unconditional three factor model and new unconditional four factor model. 5. Based on beta conditional varying informations, unconditional Fama French three-factor model was changed to conditional factor pricing model which was empirically found to increase explaining power of three-factor model on Chinese cross-sectional returns, however the conditional pricing model of beta conditional varying could not explain perfectly Chinese cross-sectional returns.6. Based on mean-variance time-varying informations, unconditional Fama French three-factor model was changed to conditional factor pricing model which was empirically found to be able to explain perfectly Chinese cross-sectional returns, however a stable model shape did not exist in the conditional pricing model of mean-variance time-varying.7. Based on state switch informations, unconditional Fama French three-factor model was changed to conditional factor pricing model which gained tremendous success, Markov regime switch three-factor model was empirically found not only to be able to explain perfectly Chinese cross-sectional returns, but also the explain power was robust.