Analysis Of The Influence Factors Of The Shanghai A-share Stock Returns

It is well known that Markowitz Mean-Variance Model has established the foundation for the modern portfolio theory by index of quantified risk and return, while Capital Asset Pricing Model (CAPM) proposed by William Sharpe et al. has also demonstrated the factors which affect the individual Return on Assets theoretically. In1970s, many empirical studies have manifested that strict-hypothetical CAPM Model lacks general explanatory power on stock yield. From then on, the Scholars have turned the studies into the multifactor model which may affect the stock yield, among which the representative study is Fama-French Three-factor Model. In addition to stressing the influence of market premium factor over the individual stock yield, Fama-French Three-factor Model has also introduced size factor, book to market ratio factor which can exert an important influence as well.Although Fama-French Three-factor Model has been proved applicable in major stock markets abroad, the viewpoint was still not prevailing during its domestic practice of early stage. This article is to find out whether Fama-French Three-factor Model would apply to Shanghai A-share market by reference to its monthly data over the last decade, the result of which turns out to be positive as the Model could well explain the stock yield, especially market premium factor and size factor, though comparing to which the explanatory power of book to market ratio factor is relatively weak. The article later studies the stability of the Model’s structure and regression coefficient by demonstration, showing that the stability of three-factor model is not good enough and its predictive ability is limited.Notwithstanding the well explanation over the stock yield change to a large extent, Fama-French Three-factor Model could still be improved. Since there exists heavy speculation in Chinese stock market which also makes turnover rate as a considerable factor, the article attempts to improve the Model by joining liquidity index-turnover rate factor. Considering certain linear dependence between turnover rate factor and size factor, we make the regression analysis separately by firstly replacing size factor by turnover rate factor and then joining turnover rate as the fourth factor, the results of which show that the fitting degree of the improved model is distinctly raised meanwhile the significance of the regression coefficient is improved as well.