An Empirical Study Of Multi-factor Asset Pricing Model

The maturity and growth of capital market promote the continuous development of multi-factor asset pricing model.The research on multi-factor asset pricing model is of great significance both in academia and industry.However,after the Q-factor model and Fama-French five-factor model came out,it seems that there is no big breakthrough.In recent years,foreign scholars' research literature mainly focuses on improving the factor construction method of the five-factor model or introducing new risk factors on the basis of the five-factor model to construct a more explanatory asset pricing model.In the domestic literature,the research on factor model is still stagnated in the three-factor model.Research on five-factor model and Q-factor model is rarely mentioned,and the conclusions are different.In order to test the applicability of these asset pricing models in China's stock market and explore a more suitable asset pricing model for the domestic market,this paper studies the applicability of each factor model based on the sorting and testing of the anomalies existing in China's stock market.At the same time,this paper extends the best performance model and draws the following conclusions:(1)Among the 1269 anomaly factors constructed in this paper,with the extension of the portfolio holding period,the yield of anomaly factor increases significantly,which indicates that long-term investment strategy is more likely to bring benefits than short-term investment strategy in China's stock market.(2)There are only 9 significant conservative anomalies in each holding period,which indicates that it is difficult to sustain various style effects in different holding periods in China's stock market.The number of significant anomaly factors in the past five years is significantly higher than that in the past ten years,and the effectiveness of the stock market does not seem to improve with the passage of time.(3)Most scholars believe that there is no significant value factor in China's market.This paper constructs the value factor by quarterly updating the portfolio,and finds that it is still not significant.However,when using the portfolio with longer holding period to construct the risk factor innovatively,it is found that the value factor constructed is significantly positive at the confidence level of 1%,which is consistent with the general conclusion reached by foreign scholars.(4)The three-factor model of China version is better than Fama-French threefactor model in explaining market anomalies,which is consistent with the conclusion of Liu-Stambaugh-Yuan(2018).The value factor VMG based on profitto-market ratio(E/P)can produce an average monthly return of 0.5%(t=1.83),which solves the problem that the value factor is not significant in the Chinese market.(5)Compared with Fama-French three-factor model,Carhart four-factor model and Fama-French five-factor model,Q-factor model has obvious advantages in explaining anomalies in China's A-share market.(6)Expansion research based on Q-factor model finds that after eliminating 30% samples of small market value companies,the explanatory ability of Q-factor model for anomalies has been improved,with smaller error of regression,which shows that "shell pollution" exists in China's market.In addition,after introducing two kinds of operational capacity factors that Q-factor model can not explain,the explanatory ability of the model to market anomalies has been significantly improved.