Study On An Improved Model Based On Fama-French Factor Model And Its Quantitative Investment Management Strategies In The Chinese GEM

Asset pricing theory has always been a hot research topic in the theoretical and practical circles.In its half a century's development,scholars with different research backgrounds have enriched its theoretical foundations from various aspects,and have constantly tested it in the ever-evolving market.Although the Capital Asset Pricing Model(CAPM)is the most widely known asset pricing model,it only describes the relationship between market risk and excess return.As the market continues to evolve,its shortcomings emerge,The single relationship between market and income cannot explain the various anomalies of market returns,Therefore,in order to improve asset pricing theory,many scholars expand the pricing model based on CAPM.The most completed system of Asset pricing theory should be Fama-French factor model,The Fama-French factor model was continuously tested between different countries and varies of markets from it being proposed in 1993.Finally,Fama and French extended the model to five-factor in 2015.This paper will conduct an empirical study in the Chinese GEM based on the theory of factor model.Firstly,this paper tests five factors(including liquidity factors)that affect the excess return of cross-sectional stock by Fama-MacBeth two-step regression method.Secondly,we construct four simulation factors named HMB,SML,RMW and CMA,for the another factors we called situation factor—building MOM simulation factors to measure investor's herd behavior and LMS simulation factors to measure profit-seeking behavior;again,we conducted empirical tests and comparative studies on the three-factor,five-factor,and six-factor models in the GEM;further,The six-factor model is optimized for the actual situation of the GEM,and the optimized six-factor model is also tested empirically.Finally,we use the optimized six-factor model as the theoretical basis to build a Value investment strategy for the GEM.In addition,in the selection of data,this paper selects the stock monthly income data,the company's financial statement data and market transaction data from the GEM in 2011.07-2018.06 in total of 8 years to construct the simulation factor data.The study found that five factors(theoretical factors)can well explain the cross-sectional returns of GEM stocks.At the same time,by observing the average income of the 5X5 portfolio,it is found that there is no obvious book-to-market ratio effect in Chiese GEM market.The time series regression of the three-factor,five-factor and six-factor models in each portfolio shows that the six-factor model has the strongest interpretation ability for grouping income,and the five-factor model has less explanatory ability in the Size-B/M group.The factor model,further,the results of the GRS test prove that the six-factor model has the strongest ability to interpret the average return of the group.Finally,with reference to the above research results,a liquidity pricing factor(state factor in this paper)is constructed to optimize the six-factor model for the real situation of the GEM,still,the empirical research finds that it has the strongest ability to interpret the average return of the portfolio.Then the factor model is applied as a multi-factor stock selection model,and the value-based investment strategy is used to backtest the sample and obtain a return exceeding the benchmark.The research results of this thesis provide reference for the operation rules,pricing mechanism,pricing efficiency and existing problems of the GEM,and also enrich the empirical literature of asset pricing model applied in the Chinese GEM.