Risk Measurement Of Stock Portfolio Of Shanghai And Shenzhen 300A Fund Based On High Dimensional Copula Function

The core content of asset portfolio risk management is the identification and estimation of portfolio risk.This article chooses different models to quantify the risk,so as to effectively estimate the risk of the stock portfolio.Based on the t-GARCH model and the D-Vine Copula framework theory,a portfolio risk estimation model is constructed for the four typical characteristics of "high-dimensional","stock","portfolio" and "risk" to make decisions in risk management activities for reference.This paper introduces the Vine Copula structure to describe the joint distribution of multiple assets.Based on the Vine Copula of Family Pair-Copula,the Monte Carlo simulation method is used to calculate the VaR of the multi-asset portfolio.The Kupiec return test method was used to test the VaR prediction effect of the Vine Copula model,and compared with the traditional historical simulation method and the variancecovariance risk management method to discuss the estimation method of highdimensional stock portfolio risk value.First,this article collects five stocks invested by China Ping An,Guizhou Moutai,Wuliangye,China Merchants Bank,and Gree Electric Appliances from the ShanghaiShenzhen 300 A Fund from January 4,2017 to December 30,2019 to form highdimensional portfolio research data and describe The logarithmic rate of return characteristics of them,make a time chart and volatility chart.Second,the time series model is constructed to fit the edge distribution.Performs logarithmic first-order difference processing on stock data to obtain yield series,and then performs stationarity test and ARCH effect test.The results show that the sequence is stable and has conditional heteroscedasticity,which is suitable for establishing GARCH model.Assuming that the residuals follow normal distribution,t distribution,biased t analysis,and generalized error distribution,a GARCH model comparison analysis is established for the sequence,and finally the optimal t-GARCH model is selected as the most suitable edge distribution for fitting.Third,the VaR of the portfolio is predicted by the data rolling method.Through the fitting of the edge distribution,the rattan Copula structure was constructed,and the corresponding rattan Copula parameters were estimated,so that it could be used as the dependent structure among the assets of the portfolio.The inverse function of the corresponding edge distribution was used to estimate the corresponding stock return rate series,and the estimated value of VaR was obtained according to the weight ratio.And compared with the traditional historical simulation method,variance-covariance method to predict the estimated value of VaR,it is concluded that the Vine Copula model can be effectively used to predict the VaR of multi-asset portfolio.The empirical analysis shows that the traditional historical simulation method and the variance-covariance risk prediction method have not passed the VaR prediction return test of the multi-asset portfolio,while the Vine Copula model based on family Pair-copula can be effectively used to predict the VaR of the multi-asset portfolio.