Quantitative Model Of Multifactor Stock Selection

In recent years,the size of China's asset management industry has grown rapidly,reach-ing nearly RMB 20 trillion by the end of 2014 to RMB 54 trillion by the end of 2017,which has nearly tripled in three years.Quantitative investment,as a relatively mature overseas development and relatively late domestic investment tool,is increasingly fa-vored by investors and institutions due to its high yield and low volatility.This paper presents a series of introductions to the most popular multi-factor stock se-lection model in the quantitative investment field,and finally obtains an investment strategy with a high return rate and high Sharpe rate.Based on the characteristics of quantitative investment,the model is divided into three stages:pre-investment,invest-ment-in and post-investment:In the pre-investment phase,the main content is to make all risk factors a valid test.First,historical data of each period of all stocks are obtained and converted into risk factor exposures by statistical methods.Second,use the next period's rate of return and the factor exposure of this period to perform the regression and perform statistical tests.Finally,select the style factor and industry factor applicable to the domestic A-share market as the model's candidate factor.In the investment stage,the main content is to construct a pure factor combination and a portfolio.Firstly,the CSI 300 index was chosen as the benchmark for comparison.The weights were optimized to find a risk exposure on only one factor,and the remain-ing factors had the same combination of exposure and benchmark as a pure factor com-bination.Second,observe the historical performance of pure factor combinations for each style factor,and understand the benefits of different factors in different stages.Finally,based on the degree of exposure of all stocks in the current period,the 50 stocks with the highest expected return are selected as portfolios and backtested.In the post-investment phase,the main content is the analysis of earnings attribution to the previous investment portfolio.Use the factor style attribution and Brinson industry attribution to analyze the investment portfolio and find out the source of portfolio re-turns,investment style preferences,and good investment methods.Investment portfolios constructed through the multi-factor stock picking model main-tained annualized returns of over 50%between 2012 and 2017,and the Sharpe rate exceeded 2.1,which met the standard of an excellent investment strategy.