现在有不同的模型来度量银行操作风险,然而选取哪种模型作为自己的方法成为银行面临的问题。本文通过不同方法对结果影响的比较分析来研究操作风险度量中模型选择的问题,即选取不同的模型对操作风险度量会产生怎样的影响。本文采用极值理论(EVT)和损失分布法(LDA)分别度量操作风险。对我国商业银行操作风险损失数据的研究分析结果表明,采用两种度量方法的结果具有较好的一致性,而LDA方法两种分布产生的结果差异性大于两种方法的差异性。因此从政策角度讲,对于EVT和LDA方法模型风险来自于模型的应用过程.而非模型选择:重要的是银行如何应用所选的模型。
There are different approaches to measure operational risk. For a bank, the question is which approach should be used to quantify its operational risk. In this paper, model selection for operational risk measurement is researched by comparative analysis of influence of different approaches on measurement results. In detail, Extreme value theory (EVT) and loss distribution approach (LDA) are used to measure operational risk. The empirical results reveal the difference of the two results of EVT and LDA is smaller than the difference of the two results of GED and SGED in LDA by using the data of operational risk related loss suffered by Chinese banks. Therefore, in terms of bank policies, how to use a chosen model is more important than which model to choose.