基于EGARCH-ε_t-GPD模型的VaR计算

The Calculation of VaR Based on EGARCH-εt-GPD Model

在传统的利用极值理论来计算VaR的过程中,一般先是对时间序列建立GARCH模型,再对残差序列运用极值理论建模,从而估计得到VaR.但建模时在GARCH模型的条件方差方程中,人们只考虑了以前时刻的随机误差项对方差的影响,而忽视了当前时刻的随机误差项对方差所作出的贡献.故作者在对时间序列建立EGARCH模型时,在方差方程中引进了当前时刻的随机误差项,然后再对残差建立GPD模型来研究风险价值,并进行了相应的实证分析,结果表明加入当前时刻的随机误差项后估计得到的VaR准确性更高.

In the process of calculating the VaR by the traditional extremum theory,generally in the first the GARCH model is set up on the time sequence,and then,the extremum theory is adopted to model the residual sequence so as to estimate the VaR.But when modeling in the conditional variance equation of the GARCH model,people only consider the influence of the random error term of the past on the variance while ignoring the contribution made by the random error term of the current to the variance.Therefore,when the EGARCH model is set up on the time series by the author,the random error term of the current is brought in the variance equation.And,the GPD model is built on the residual error to study its value at risk and to analyze it empirically.The result shows that the accuracy of the VaR is higher by the way that the VaR is incorporated with the random error term of the current.