中文摘要：

针对近期中国股票市场的剧烈波动给投资者带来的影响,本文选取中国沪深股市和香港股市中具有代表性的3种股票指数的日收益率数据进行建模分析。选用ARMA-GARCH（1,1）-t模型拟合其边缘收益率序列,利用Gumbel、Clayton和t-Copula的线性组合函数刻画其相关结构,再运用蒙特卡罗模拟方法计算不同置信水平下各股指的风险价值、预期损失和中位数损失,其中中位数损失是最近才提出的一种新的风险度量指标。本文的研究结果可以为有着不同风险偏好的投资者和管理者监管系统性风险提供借鉴。

英文摘要：

In view of the influence of recent Chinese stock market volatility on investors,three representative stock index＇s daily returns of Shanghai,Shenzhen and Hong Kong stock markets are selected for modeling and analysis.First,ARMA-GARCH（1,1）-t models are chosen to fit the marginal returns.Then,a linear combination function of Gumbel,Clayton and t-Copula is applied to describe the correlation structure of returns.Finally,Monte Carlo simulation method is employed to compute Value at Risk,expected shortfall and median shortfall of the three stock index under different confidence levels.Here median shortfall is a new risk measurement index proposed recently.The research provides references for investors and managers with different risk preferences to regulate the system risk.