中文摘要：

在给定我国主要储备货币为美元的基础上，使用英镑、日元、欧元兑美元每日汇率，应用时变条件t—copula函数描述主要储备币种对美元汇率收益序列之间的时变相依结构，建立了用于描述包含时变自由度在内的所有时变相依模型参数的演化方程。采用蒙特卡洛仿真方法计算了各种币种组合的VaR，并对结果进行后验测试，时变条件t-copula函数仿真估计VaR可以覆盖最大损失风险，并分析了主要储备货币组合兑美元汇率风险演化趋势，作为调整外汇储备结构的参考依据。

英文摘要：

Given the major currency of foreign reserve of China is US dollar, we use the daily exchange rates of Pound Sterling, Japanese Yen and Euro to US dollar to study the risk of foreign reserves. A time-varying t-copula model is used to investigate the dependence between these exchange rates and the evolution equations are developed to describe the time-varying parameters including related correlation coefficient and degree of freedom. Simulated protfolio return series is generated by Monte Carlo method to get VaR of different portfolios. The VaR results are back tested, and the result of these tests shows that VaR series calculated by time-varying T-copula model have a good coverage rate to factual lost. An analysis on the risk trend of these major reserves currency portfolios is also presented. It could be a reference to adjustment of the structure of foreign reserves of China.