Quantile Regression Method Research On Influencing Factors Of Stock Return For China A-share Market

The theories and models of influencing factors on the stock return have been aproblem for scholars. From the original the Capital Asset Pricing Model（CAPM）andthe Arbitrage Pricing Theory（APT）to the three-factor model, Scholars have found thatsingle Beta isn’t enough to explain the stock return. At the same time, they also havefound some phenomena including the Small Firm Effect and BM Effect. Althoughscholars have been changing the influencing factors and improving the models, all thesetesting methods’ results were the effect of the mean value of stock return. Differentmodels and different data samples would often come to different empirical results,especially in Chinese market.In this paper, we based on the Chinese A-share market, combined with the resultsof the scholars’ researches and selected26factors that might affect the stock return.Then we used the Cluster Analysis method to select12factors from all factors. At last,we tested how the factors influence the stock through the least square method and theQuantile Regression（QR）which could test the effect at different points of distribution.Comparing the results of these two methods, we found that:（1）For the same factor,the regression result showed the unique relativity and the significance in the leastsquares method, but different quantiles had different relativity and the significance inthe quantile regression. In addition, the relationship between the same factor and stockreturn was different at different quantiles in different industries. Therefore, it’s necessaryto analyze the relationship by the QR.（2）For the overall data, the goodness of fit forquantile regression was better than the least squares method. But for the industry data,the result was opposite.（3）Financial indicators have not strong explanatory power forstock returns. In China, investors paid more attention to the influence of specialindicators that reflected the company’s market performance than the financial factorswhen they did the value analysis for companies.The results showed that compared with the least squares method, QR could be acomprehensive analysis method between stock return and various influencing factors.