The Research And Empirical Analysis Of The Effectiveness Of Factor Model In Our Stock Market

The problem of asset pricing can be traced back to year 1738,when Bernoulli submitted his paper to St Peter's Royal Academy of Sciences.The general theory about selection of asset allocation and asset pricing established until 20 th century.Most research supported CAPM model as the early term until Roll(1977,1978)pointed out some queries about that.That was called “Roll's Criticism”.From that time on,there are more and more criticisms on CAPM model,and scholars were going to find out other factors to explain stock price.Based on the No-arbitrage equilibrium principle,Ross put forward APT model at 1976.But APT model didn't refer to specific factors which can affect stock price.The three-factor model of Fama-French is a great extension of APT MODEL.With time goes on,the effectiveness of three-factor model vanished gradually.While more anomalies have been found with researches go deeper.The model has been extended,and Carhart came up with four-factor model,while Fama-French referred to five-factor model.At first,I tested that if there is the inversion-momentum effect in China's stocks market,And find out that there are obvious momentum effect.Then,I tested market size,B/M,investment,profit effect,and the results show that they are all conspicuous in our stock market.However,the correlation between inv and return are positive,which is on the contrary to the result of Fama-French five-factor model.After testing for effectiveness of all factors,I use the portfolio spreads method to construct the others variables,test and adjust them depend on the result of stability test.Quarterly data I used is from Jan.2001 to Dec.2016.Variables include Rm-Rf?SMB?HML?RMW?CMA and WML.RMW has a high correlation with other factors.Then I choose Rm-Rf?SMB and one to four variables from HML?RMW?CMA?WML to construct different 16 models.And then have GRS test for all 16 models respectively.The results show that Fama-French three-factor model and Six-factor model are more stable and better in explanation relatively.Then I did some adjustment on variables which have high correlation with other factors.And test for the effectiveness of the two models in different sets and industries.In the group based on size-B/M 5×5,variables of three-factor model have higher t value,and the model has higher adjusted R-squire for each group.In different industries,Six-factor model has higher ability of explanation of return except transportation industry.Both factor models have better effect in food tourism,black metal,information equipment industries.While,factor models did worst job in financial industry.In general,there are anomalies in Chinese stock market significantly,and the market is ineffective.Multi-factor models have strong applicability in Chinese stock market relatively.Fama-French three-factor model and Six-factor model are more stable and better in explanation process.In a set with less stocks,three-factor model does a better job,while in sets basis on industries,six-factor model performs better.