Based On Multivariate Copulas Connect - Sv - Var Model Of Open-end Fund Portfolio Risk Measure

Open-end fund of our country requires higher risk measurement techniques in risk management, and now VaR is the main indicator in financial risk measurement, there are many calculation methods for it, as a result of forming many models. Considering the accuracy and simplicity of calculation, this paper proposed a multivariate Copula-SV-VaR model. Therein, Copula was a joint function, which was used to describe the correlation between financial assets in a portfolio, and the correlation was nonlinear; SV was a Stochastic Volatility model, which was used to describe the marginal distributions of financial assets in a portfolio, and describe leptokurtosis and fat-tail of a single asset better compared with GARCH model;VaR was the indicator of risk measurement, which was used to describe the size of volatility risk of a portfolio, calculated by Monte Carlo Simulation method.This paper first introduced Copula theory, putting emphasis on introducing characteristics and application scopes of common Copula factions, and parameter estimation methods and their application scopes of Copula models, and pointing out the significance of probability integral transform in Copula theory, then introduced SV models, putting emphasis on introducing classification and parameter estimation methods of SV models, finally introduced VaR, putting emphasis on introducing two calculation methods, which were Variance-Covariance method and Monte Carlo Simulation method, based on summarizing calculation methods for VaR detailedly, so as to understand and construct the model. In empirical study, the author selected GuangFa Jufeng Fund as the object of study, and measured the risk of the portfolio composed of its top ten holding stocks. The author compared the size of Copula-SV-VaR with the size of Variance-Covariance-VaR, in order to test the accuracy of the model, while the author compared the size of Copula-SV-VaR with the size of Copula-GARCH-VaR, in order to test the precision of the model. To introduce the multivariate Copula-SV-VaR model detailly in theory and in empirical study, the author hoped that it can provide a reference for fund management companies or fund supervision department of our country to measure the risk of portfolio of open-end fund.