Investigation On Large Portfolio Selection Approaches Based On China Stock Market Data

In the past 27 years,china's stock market has experienced 10 times of bear market and 10 times of bull market.The high fluctuating rate stirs great attention among investors.The great stock market disaster in 2015 and series limit down in July make investors re-consider the efficiency of current portfolio models.Using two different formation approaches,this paper selects 6 portfolio models to study the efficiency of Chinese stock market from April 1991 to December 2017.The data used in this paper is all monthly data from Chinese stock market.Specially,the first part of this paper is the literature review introducing related domestic and foreign studies of portfolio models.Secondly,giving a brief introduction of the 6 selected models and several special rules in Chinese stock market,meanwhile transforming these rules into particular formula to modify the initial portfolio models.Thirdly,the Monte Carlo simulation evaluating the efficiency in one factor economic environment.Fourthly,the portfolio efficiency empirical study on two formation approaches and 6 models using monthly data from Chinese stock market during April 1991 to December 2017.Fifthly,robust test on three sub-periods and different estimation window length.The main conclusions are as follows:(1)if estimation error can be moved away,optimal portfolio strategies are better than naive diversification strategy;(2)among the five optimal strategies,the three fund model performs best;(3)using single stock is better than using initial portfolio to form optimal portfolio;(4)a moderate data dimensionality is better than high dimensionality or low dimensionality.