基于经验分布下GARCH模型对VaR的金融测度  

The financial measurement of VaR under the GARCH model based on empirical distribution

作　　者：李翠霞 陈媛媛 LI Cuixia;CHEN Yuanyuan(School of Mathematics and Statistics,Xuzhou University of Technology,Xuzhou 221111,China;School of Mathematics and Statistics,Lanzhou University,Lanzhou 730000,China)

机构地区：[1]徐州工程学院数学与统计学院,江苏徐州221111 [2]兰州大学数学与统计学院,甘肃兰州730000

出　　处：《中山大学学报（自然科学版）（中英文）》2021年第4期177-182,共6页Acta Scientiarum Naturalium Universitatis Sunyatseni

基　　金：国家自然科学基金(11301236)。

摘　　要：VaR (Value at Risk)在风险管理领域一直深受银行业和金融机构的重视,GARCH模型对VaR的测量是一个重要研究领域。然而在实际应用中,利用传统参数GARCH模型建模时需要指定条件分布形式,一旦分布指定错误将会导致模型失效。因此,我们在标准的GARCH(1,1)模型下,结合累积经验分布函数对残差进行修正,避免了传统参数分布由于事先指定错误带来的模型风险。经过实证研究发现,我们采用的方法比指定参数分布下的标准GARCH(1,1)模型在测量VaR方面有了很大改进,其失败频率和相对误差都显著降低。因此,文中采用这种创新的尾部分布形式在估计VaR值方面具有一定的实际应用价值。In the field of risk management,VaR has been highly valued and the measurement of VaR by GARCH model is an important research. In practical application,it is necessary to specify the conditional distribution form when using traditional parameter GARCH model for modeling. However,the model will become invalid if the distribution is specified incorrectly. Therefore,under the standard GARCH(1,1) model,the residual error is corrected in combination with the cumulative empirical distribution function,so as to avoid the model risk caused by the pre-specified error of the traditional parameter distribution. In the empirical research,it is found that the proposed method is greatly improved compared with the standard GARCH(1,1) model based on specified parameter distribution in the measurement of VaR,and its failure frequency and relative error are significantly reduced.

关 键 词：VAR 标准GARCH(1 1)模型 累积经验分布函数 尾部分布 

分 类 号：O212[理学—概率论与数理统计]