Study on Returns of Shanghai A Stock Exchange
|Keywords||three-factor model stock liquidity quantile regression|
Asset pricing theory is one of the core issues of finance, which plays an important role in theasset allocation and risk control. Traditional asset pricing model (CAPM) received extensiveattention since it was proposed, but the empirical research on the performance is not satisfactory. Inrational pricing framework, Fama和French proposed three-factor model including the scale factor,the value factor and the market risk factor. this model has a better explanation of the force to thestock market. Tradition CAPM model assumes that investors trading securities do not have animpact on asset price. This assumption ignores the reality of the market liquidity and risk. It isimportant to put liquidity into asset pricing model. Finance time series data exist leptokurticcharacteristics. Quantile regression method is able to describe the data of local information. It hasbeen widely used in the study of the recent financial problems.In this paper, we use the shanghai A shares data from January2006to December2010. Theportfolios are constructed by the size of the company, the market value ratio and liquidity. Usingthe CAPM, FF three-factor model and the four-factor model joined the liquidity to make empiricaltests, we conclude that: The shanghai A-share market exists size premium, value premium andliquidity premium; Market risk on explanatory power of stock returns has been strengthened; thefour-factor model adding liquidity fits higher than the CAPM model and FF three-factor model; theimproved Amihud indicator and turnover can characterize the mobility well, there is littledifference in the explanatory power of the four-factor model based on both liquidity.On this basis, the paper uses quantile regression method, select the Beta coefficient, the size ofcompany, the value ratio, the improved amihud indicator and turnover as risk factor, researchcross-sectional returns in shanghai Astock market. The conclusions are as follows: risk factorscoefficients and significant are difference in the different quantile regression, which shows stockreturns have different forms of dependence in different quantiles; it is significance the risk factorscoeffiences are difference among different quantile.