偏态t分布下FIGARCH模型的动态VaR计算

Calculation of Dynamic VaR Based on the Skew Student-t Distribution of FIGARCH Model

针对金融时间序列多是有偏分布和"长记忆性"的特征,讨论偏态t分布下分数维GARCH模型的动态VaR测算问题。在分析正态分布、学生t分布、广义误差分布下和偏态t分布的基础上估计模型参数,得出了动态VaR,并进行了失败率检验。实证结果表明:基于偏态t分布下的FIGARCH模型测算的动态VaR值克服了其他分布假设上的不足,能够较好地反映金融收益率的实际风险,并在该分布下的Pearson吻合度检验也证实了模型分布选择的正确性。

Because the financial time series have the characteristics of multi - distribution and ＂long memory＂, this article discussed the problem of dynamic VaR calculation of the FIGARCH model based on the situation of the skew student-t distribution. After we got the estimated parameters of the model through the analysis of the normal distribution, the student - t distribution, the generalized error distribution and the skew student - t distribution, we calculated the dynamic Var and the test failure rate. The empirical results show that the FIGARCH model based on the skew student - t distribution is better than other distribution on the assumption in estimating the value of the dynamic Var, because it is better to reflect the risk of actual rate of return and it is proved to be better to choose the skew t - distribution model through the Pearson test.