Research On CSI 500 Index Enhancement Strategy Based On Barra Multi-factors Model

The quantitative investment has been developed for more than 40 years in the mature market.Since the first quantified hedge fund was issued by Guotai Junan Securities in 2011,the management scale of the fund for quantitative investment has expanded and developed rapidly.Among them,index enhancement fund is an important type of large public offering funds in China.Such active funds have been gradually favored by a large number of private offering funds because of their success in winning the index and controlling risks.Multi-factors stock selection model is an important way to study index enhancement strategy,and Barra model is a common multi-factors stock selection model in the industry,which is widely used in real investment.Based on the latest version of Barra model,Barra USE4,this paper conducts a index enhancement strategy based on the pool of CSI 500.This paper first combs the theoretical background of the multi-factors model that is popular in the industry.Starting from the classic capital asset pricing model(CAPM),Fama-French three factor and five factor model,this paper reviews the success of behavioral finance in explaining the "anomalies" in the market.In this paper,the multi-factors stock selection model is classified into the characteristic model based on behavioral finance,and the theoretical origin is found for the multi-factors model.This paper discusses the logic of the multi factor stock selection model that can be profitable in practice,and introduces the sorting method,Barra USE3 and Barra USE4,and compares Barra USE3 with Barra USE4,and proves the inherent consistency of the two versions of Barra model.Then,this paper introduces the technology of the Barra model,including the calculation of the pure factors portfolio and the multi-factors portfolio.The quality of factor is the key to decide the success or failure of the multi-factors stock selection model.Therefore,this paper specially discusses the factors,including the introduction and acquisition of style factors and industry factors,and the factor evaluation system of Barra.It is inferred that the multi-factors model which is classified into the characteristic model should have the instability itself,so the failure of the factor is a normal phenomenon.This is in line with practical experience.In practice,the switching of factor performance is called factor wheeling,which is a major problem in the multi-factors stock selection model.Based on the in-depth consideration,this paper proposes a dynamic Sharpe ratio threshold method to solve the factor wheeling problem based on the consideration of whether there is a long-term persistence rather than a single direction based on the characteristic factors.It is different from the static Barra factor evaluation system.Based on the framework of Barra USE4 and the Sharpe ratio threshold method,this paper carries out an empirical study of the index enhancement strategy.The data selected in this paper range from January 2012 to September 2017.A total of 12 styling factors were selected,and 28 SWS level-1 industry factors were merged into 7 categories.Based on these factors,after validating the collinearity of the style factor,we get the result of factor investment in each window period,and then calculate the stock holding weight in each window period.In this paper,a rolling stock transfer method of 25%per week is adopted to make the actual holding length for a month,so as to calculate the capital curve of the strategy.The index enhancement strategy in this paper has reached over 20%of the annual excess rate of return(simple interest),with an information ratio of more than 3.Subsequently,sensitivity analysis is carried out on the three parameters,namely factor wheel threshold,factor wheel observation period length and individual stock upper limit,which verify that the performance of strategy return test is stable.This shows that the index enhancement strategy in this paper is successful.The investors in China's small and medium market can overcome the market through the model of this article.