Analysis On The Influencing Factors Of SSE Constituent Index Stock Return Rate

This paper attempts to identify which of the hundreds of stock yield impact factors reported in previous studies are statistically significant cross-sectional predicting factors of the average yield of SSE 180 stocks.By improving the pioneering research method of Fama and French,Fama-Macbeth regression estimation is made byclassifying and combining samples and research methods,i.e.the value weighted least squares(VWLS)for all stocks and the ordinary least squares(OLS)for non-micro-cap stocks(SSE180).The results of the two regression methods are brought together as a generalized method for inferring the cross-sectional determinants of non-micro-cap stock yield.In order to prove the impact of Chinas micro-cap stocks,to facilitate comparison with previous studies,to avoid excessive weight on micro-cap stocks,and to adjust the data mining deviation,this paper also provides the regression results of OLS for all stocks.The mimber of independent determinants identified by the multi-factor model is the same as the number of significant factors identified by the one-factor models,indicating that a few independent factors cannot absorb the information contained in a large number of individual significant factors.On the contrary,in essence,there are very few factors that can independently predict the average yield of SSE 180 stocks.The characteristics and properties of independent determinants identified by the multi-factor model are quite different from those of the significant factors identified by the one-factor models.In contrast,of the 12 independent factors identified by the multi-factor models,only one is the fundamental factor,i.e.number of quarters of year-on-year earnings growth,and seven are transaction behavioral factors,i.e.earnings announcement return,one-month momentum change rate,six-month momentum change rate,yield volatility,stock turnover rate,volatility of stock turnover rate,zero trading days.These differences indicate that in the model design for researching the determinants of the average stock yield,better effect can be achieved by using independent factors determined by multi-factor models.This paper proves that the independent determinants in the SSE 180 stock group are often the independent determinants in the micro-cap stock group,but not vice versa.Of the 12 independent determinants of the SSE 180 stock group,10 are independent determinants in the OLS regression of all stocks,while among the 23 independent determinants(OLS for all stocks)in micro-cap stock group,13 are not independent determinants in the SSE 180 group.In addition,11 of the 12 independent determinants are different from the factors in the benchmark factor pricing models,sueh as the Carhart momentum factor pricing model,the Fama-French five-factor model and the q-factor pricing model,while the book-to-market ratio is an exceptional factor.The above benchmark factor pricing models were adopted in previous studies to control the cross-sectional changes of the average stock yield except for specific independent variables.However,the results of this study suggest that these models perhaps fail to achieve the expected control degree in the study of SSE 180.According to the method of this paper,it is determined that 12 factors provide significant independent decision information about the monthly average yield of SSE 180 stocks from 2002 to 2018 while the remaining 82 factors fail to provide.This paper demonstrates that the reason why few factors provide independent decision information is that the number of independent determinants of stock yield is essentially small,rather than that these few factors can absorb information from many other factors with univariate significance.In summary,through determining the independent determinants of the stock,s monthly average yield and relaxing the constraints of the independent determinants on the same size and time of the companies,this paper provides new evidence to prove that the inferences drawn from the regression studies of hundreds of factors are doubtful,thus explaining that many of the factors seen in the research reports of the seller's researchers in the securities companies only appear to be effective in the reports.