Study On Portfolio Optimization Of Stock Based On Copula-CVaR-EVT

In view of the increasing importance of stock risk management,it is very important to carry out research on stock portfolio optimization.At present,Va R or ES are widely used to measure the risk of portfolio assets.However,Va R and ES have exposed obvious deficiencies in the risk measure,which makes the research on portfolio optimization of stocks more urgent.Based on the above analysis,this paper adopts the CVaR framework to overcome the defects of Va R and ES,to measure the risk of the sample of various stock samples in the portfolio.Copula method is used to characterize the nonlinear relations of various stocks.EVT is introduced to measure the tail risk to build a portfolio optimization model based on Copula-CVaR-EVT.Through empirical analysis,the following conclusions are drawn:(1)The return series of each stock sample presents typical factual characteristics such as peak,thick tail,biased,non-normal,heteroscedasticity,volatility clustering,etc.(2)The GPD distribution in the EVT method can fit the empirical distribution of the tail of all samples well,which shows that the GPD distribution is very accurate in describing the tail of all samples,and further proves the necessity of the EVT method in describing extreme risk.(3)Compared with other types of Copula,the hybrid Copula model can fit the data most accurately and has the best performance in describing the non-linear dependence among assets.(4)Comparing the upper-tail dependency coefficient with the lower-tail dependency coefficient calculated by the mixed Copula model,it can be found that the lower-tail dependency coefficient is significantly larger than the upper-tail dependency coefficient,which indicates that the spillover effect of risk among assets is significantly stronger than that of income,thus indicating that the risk among assets is more contagious.(5)By comparing the risk dependence of individual stock assets with other stock assets and the dynamic risk intensity of individual stock assets,it can be found that among all stock assets,Suning Easy-to-buy,Donghua Science and Technology,Rui Maotong not only have the least risk dependence with other stock assets,but also have the lowest risk intensity,which is suitable for investment portfolio.(6)Based on the risk minimization strategy,the portfolio under the optimal weight can obtain higher investment returns than the portfolio under the equal weight.At the same time,based on the income maximization strategy,the portfolio under the optimal weight has lower risk value than the portfolio under the equal weight,which shows the investment proposed in this paper.The combinatorial optimization model has excellent portfolio functions.Based on the above analysis,the Copula-CVaR-EVT portfolio optimization model proposed in this paper can provide scientific application tools and methods for market participants to establish an efficient,scientific and accurate portfolio risk prevention and control mechanism.Market participants can use the optimization model proposed in this paper to optimize the investment portfolio of stock,which can guarantee investment returns,and ultimately promote the healthy development of the national economy.