极值理论高频VaR区间估计模型的构建

Interval Model Estimation of VaR of High Frequency Extreme Value

为了对在险值的估计精度进行度量,更为精确和有效地衡量极值VaR(value at risk)的估计风险,基于广义极值理论构建了极值VaR的区间估计模型,并进一步利用高频数据重点考察了不同置信水平和不同样本容量分块下的极值VaR区间估计结果的精度和模型的有效性。结果表明,极值VaR的动态区间估计模型与参数法和非参数法区间估计模型相比,不仅能够更为有效地捕获极端条件下收益率时间序列的动态特征,而且具有很好的估计精度,VaR估计风险的精确度更高。

In order to capture the character of return series in extreme condition and improve VaR（value at risk） precision,a model of extreme value VaR is established.Aolopting high frequency data,the result precision of confidence interval of extreme value VaR and the validity of model are mainly studied under different confidence levels and blocks.The empirical results show that comparing our model with parametric method and non-parametric method in estimation of the confidence interval of VaR,our model can not only captare the risk character of Chinese stock markets,but also achieve better estimation accuracy and describe the estimation risk of the VaR more accurately.