Research On The Risk Of Precious Metal Investment Portfolios Based On Time-Varying Vine-Copul

In recent years,the international situation has been turbulent,wars have occurred from time to time,and the trend of the global economy is not very clear.At this time,the hedging function of precious metals such as gold and silver and their low dependence on other financial assets make investors gradually keen on precious metal investment.In this context,based on the daily closing price data of six precious metals:Gold,Silver,Copper,Nickel,Lead and Palladium,this paper studies how to more scientifically and reasonably describe the nonlinear dynamic correlation structure between high-dimensional financial variables and the Risk Value of precious metal portfolio.As the relevant structure of the precious metal market will change over time,based on Patton’s time-varying Copula theory,this paper makes the time-varying parameters of binary normal Copula function,binary tCopula function,binary Clayton Copula function and binary rotated Clayton Copula function obey ARMA(1,10)model,and constructs a timevarying Vine Copula model based on Patton’s thought to analyze the correlation between precious metal combinations.In the demonstration,using the two-stage modeling method,the ARMA-GARCH model is used to fit the edge distribution of the log return series of six precious metals.It is obtained that the edge distribution of gold is ARMA-GARCH-Norm model,and the edge distribution of the other five precious metals is ARMA-GARCH-GED model.Then different Vine structures are selected,and eight static and time-varying Pair Copula Functions are selected according to AIC criteria to ensure a better description of each edge of each tree,so as to establish the joint distribution function of precious metal portfolio.The empirical results show that the time-varying binary copula function can better describe each edge under different Vine structures,and the time-varying binary normal Copula function and time-varying binary t-Copula function have the best depiction ability;Comparing the fitting effects of models under different vine structures,the fitting effect of timevarying C-Vine Copula model on financial return series is slightly stronger than that of time-varying D-Vine Copula model,but the AIC value of the model is not different.This paper makes an empirical study on the risk of precious metal portfolio return through Monte Carlo simulation method.According to the Kupiec failure rate test results,under four confidence levels(90%,95%,99%,99.5%),the time-varying C-Vine Copula and D-Vine Copula models based on Patton’s thought have the robustness and reliability of sample measurement,and can better fit the fluctuation of precious metal portfolio risk.Under different asset ratios(1/6,1/6,1/6,1/6,1/6,1/6,1/6)and(Gold:25%,Silver:25%,Copper:20%,Lead:15%,Palladium:10%,Nickel:5%),the VaR2 calculated by the time-varying C-Vine Copula function based on Patton theory is the best at the 99%confidence level;The VaRl calculated by the time-varying D-Vine Copula function based on Patton theory at 99%confidence level and the VaR2 calculated at 99.5%confidence level can well measure the risk of precious metal portfolio.