商业银行操作风险的度量——基于非参数方法

在损失分布方法的基础上,本文基于非参数方法对商业银行操作风险的度量进行了研究。非参数方法对损失额的分布不作过多的设定,避免了由于分布误设可能出现的偏差。古典的核密度估计对损失额拟合的效果不太好,特别是尾部的拟合效果更差。变换后的核密度估计的拟合效果比古典的核密度估计改善很多.基于变换后的核密度估计对商业银行操作风险损失度量可以得到不同置信水平的VaR与ES,并且不同置信水平的差距比较大。基于非参数与基于参数方法得到的各个置信水平的VaR与ES有一定差距。

The Measurement of the Operational Risk of the Commercial Bank Based on the Nonparametric Methods

On the basis of the loss distribution approach,this paper has discussed the measurement of the commercial bank operational risk based on the nonparametric method.Nonparametric method can avoid the possible deviation due to distribution misspecification,for the nonparametric method has fewer specification.The fitting effect to losses is not very good of the classic kernel density estimation,and especially at the tail,the fitting effect is poor.The fitting effect is better of the transformation kernel density estimation than that of the classical kernel density estimation.Different confidence level of VaR and ES can be obtained based on the transform of kernel density estimation,and it has remarkable difference between different confidence level.There has some difference of VaR and ES at the confidence level based on nonparametric and parametric method.