VaR准确性检验的T检验法

The T-test of VaR Accuracy Test

根据VaR失败率检验法的定义,通过利用组合的方式获取检验样本和构造服从T分布的统计量,得到检验VaR准确性的新方法——T检验法。T检验法所接受的模型不仅在似然比检验法下能被接受,且在同一置信度下T检验法所接受的模型比似然比检验法所接受的模型的准确性更高。最后通过实证分析进一步说明T检验法比似然比检验法能更好地检验VaR的准确性。

According to the definition of VaR failure test,a new method to test the accuracy of VaR is obtained,which using combination method to get testing samples and structure statistics that obey T distribution,that is T-test. The model accepted by the T-test can also accepted by likelihood ratio test,and at the same confidence,it is more accurate. by the end,through the empirical analysis,it is further explained that,compared to the likelihood ratio test,T-test can test the accuracy of the VaR preferably.