基于Copula的股票市场VaR和最优投资组合分析

在股票市场风险分析中,对不同股票相关结构的讨论有着重要意义,从而引出对如何选取好的相关结构模型来捕捉股票间的相关变化规律的讨论.本文选取Gauss Copula、t Copula、Gumbel Copula和Mixed Gumbel Copula,运用两步半参数估计法对股票市场相关结构建模,并依据建立的模型进行VaR分析,采用Wald型检验方法来判断VaR估计效果.从效用最大化角度出发,确定最优投资组合.

Research on the VaR and optimal portfolio of stock markets based on Copula

It is of particular relevance in stock markets if we are required to study the dependence between the stock data.In this situation it is essential to find a good dependent model for reflecting the dependence variety.This paper is concerned with the modeling of the dependence structure of stock markets using the concept of copulas.We select four common copulas and use twostage semiparametric way to estimate the copulas.VaR is analyzed particularly from the fitted models,we compare the effect of the copulas using Wald type test.Utility maximization is used to optimize the allocation of the portfolio.