The Cross Section Of Stock Returns Based On The Time-Varying Coefficient Factor Models

Based on the Fama-French three-factor and five-factor model,this paper interprets the time-varying ? and ? coefficients with the growth rate of retail and food service scale as the new information variable,along with the commonly used information variables.We provide an explanation for the difference in the performance of the factor model in different samples from the angle of time-varying coefficient.In terms of time series,the empirical results strongly reject the constant coefficient factor model,and the new information variable has better explanatory power than the commonly used macroeconomic information variables in the past.At the same time,we find that the new information variable can not only explain the risk transfer between portfolio returns of F-F factors and stock returns,but also directly explain the change of stock returns.When explaining the cross-sectional variation of stock returns,we find that the F-F three-factor and five-factor ? lose their significance if the fitting value of multiple information variables to stock returns is the control variable.After the contra controlling? of F-F factors,the fitting values still have strong explanatory power for excess earnings.In further analysis,we find that the time-varying coefficients based on the new information variable can explain the change of 38%to 40%of the excess return of stock.This paper not only rejects the common factor model with constant coefficients,but also provides a new idea for researchers in the field of asset pricing.It has important theoretical and practical significance in asset pricing,risk assessment,asset allocation and other related fields.