Based On Common Factors Extraction Method To Research The Effectiveness Of Stock Pricing

Capital asset pricing model (CAPM) which is proposed by Sharpe in1964is the milepost on financial market pricing theory. CAPM theory argues that the system risk, namely market premium factor plays an irreplaceable role financial market, while the non system risk can be eliminated by diversified investment. However, affected by the incomplete information, transaction cost and other factors, investors failed to hold diversified portfolios, so the non system risk has been existing, consequently, the pricing function of the single factor of CAPM gradually weakened. Then multivariate asset pricing model become a hot research topic in the modern financial world.However, with the development of research on the pricing model, more and more variables were considered into the category of securities pricing to explain a large number of financial vision. Regressing excess returns on so many factors may exist multicollinearity, the result is incorrect. This paper attempts to employ the asymptotic principle component analysis, in order to transfer high-dimensional time series into low time series, namely common factors or latent factors. But a few common factor sequences reflects lots of information from the panel data.After obtaining the number of latent factor and the factor process itself in the process of estimation, this paper will regress the observed variables on the common factors to determine the effectiveness of factor through a series of statistics. With common factor extraction method introduced, providing effective idea for many empirical factors test.This article selects A-share as the research object (a total of796individual stocks, the sample period from February2009to December2013) to clarify the effectiveness of observed variables. The main conclusion as follows:Firstly,in latent estimation phase, the estimation of the factor loading matrix and the factor process itself is carried out via an eigenanalysis of a N*N non-negative definite matrix, as a result, high-dimensional time series were converted into three-dimensional time series. Secondly, we found that, in the sample period, there is no scale effect in Chinese A-share market, to the contrary, the average monthly returns of the listing corporation to rise along with the company’s market value. Research on momentum factor found in Chinese A-share market exists short and medium term momentum effct, as well as long-tern contrarian effect. Thirdly, regressing the observed variables on latent factors, the market premium variable and book to value variable can be regarded as a good proxy for latent factors. Research on the relationship between macroeconomic variables and latent factors, found that those macroeconomic variables have a weaker influence than microeconomic variables, relatively speaking, consumer satisfaction index and inflation rate have a strong correlation with the latent factors. At last, research in stock returns in different industries, found the market premium factor has significant effect on the stock returns, and the inflation rate of macroeconomic variables can not be neglected.