Research On Multi-factor Quantification Strategies Based On Behavioral Financial Factors

Behavioral finance is a very popular branch of finance in recent years.At the same time,the research of multi-factor quantification strategies has always been an important content of quantitative financial research.The core work of this paper is combining behavioral finance with multi factor quantification strategies to research multi-factor quantitative strategies based on behavioral financial factors.There are three main research links in this paper:factor construction and computation,model building and algorithm design,and empirical analysis.In the factor construction and computation link,this paper takes two articles published in the top financial journals JFQA and JFE as the main reference to construct two behavioral financial factors which are Continuing Overreaction Factor and Trend Factor,and calculates two factors with the data of Chinese stock market.In the model building and algorithm design link,this paper studies the single factor validity test and the multi-factor stock selection model respectively.In this paper,a complete single factor validity test system is established based on the stratified backtest and regression analysis.Through the statistical analysis of the result of factor grouping,the result of factor return,the result of factor information coefficient and so on,the validity of the stock selection is studied.For the multi-factor stock selection model,this paper classifies the model from different angles.In the selection of factors,this paper constructs two models of static multi-factor model and dynamic multi factor model.In the stock selection method,this paper studies two algorithms of combination stocks after single factor stock selection and stock selection after multi-factor combined scoring.In weight allocation,this paper puts forward a variety of algorithms,such as information coefficient weighting,information ratio weighting,factor expected return weighting,and so on.The cross combination of the above multiple angles and multiple algorithms has formed a number of multi-factor quantized stock selection models.The empirical link of this paper will be selected from the model for back test and analysis.In the empirical analysis link,this paper takes the constituent stocks of Shanghai and Shenzhen 300 index as the main research object,and successively carries out the test of the effectiveness of behavioral financial factors and the multi-factor quantitative stock selection model based on behavioral financial factors.The test results show that,the continuing overreaction factor and trend factor constructed in this paper have obvious stock selection effectiveness in the Chinese market.Buying stocks in the biggest continuing overreaction decile and selling stocks in the smallest continuing overreaction decile generates an average return of 0.9%per month and the average annual yield was close to 11%,and Buying stocks in the smallest trend factor decile and selling stocks in the biggest trend factor decile generates an average return of 1.1%per month and the average annual yield was about 13%.The historical performance of factor return is better than the market benchmark,and the factor regression analysis also shows that the two factors have significant stock selection effectiveness.In the multi-factor quantitative stock selection strategy backtest,this paper builds a multi factor model based on the two behavioral financial factors and the size,the book to market ratio,the momentum and so on,which is proposed and verified by the authoritative scholars.Considering the different algorithms in the process of constructing the multi-factor model,this paper deals with four deferent models:Multi-stock selection after multi-factor combined scoring and Rank-IC moving average weighting model,stock selection after multi-factor combined scoring and IC IR weighting model,combination stocks after single factor stock selection and Rank-IC moving average weighted model and combination stocks after single factor stock selection and IC IR value weighting model.The results show that the multi-factor model based on the behavioral financial factor using the multi-factor comprehensive score selection method has shown good performance in the Chinese Shanghai and Shenzhen 300 market.The annual rate of return of the strategy is over 28%and the annual rate of excess return is about 20%.Excluding the abnormal fluctuation period of China's stock market in the second half of 2015,the overall performance of the strategy is stable.