信用风险内部模型评价

Evaluating Internal Credit Risk Models

随着金融风险数量化研究的深入与发展,研究者给出了形式各异的信用风险度量模型,这些模型依赖不同的假设和信用数据基础。如何在诸多模型中选择适用于自身风险度量的工作就显得十分重要。本文首先介绍Z-计分和EDF模型,然后利用AR方法对这些信用风险度量模型的预测精度进行比较分析,发现对于所选定样本AR值能较好的反映两个模型的预测精度。

Various credit risk modeling approaches have been developed in the past few years. But these models are based on different assumptions and credit data. So, it is very important to choose the suitable one for risk measures among the models. This paper introduces the Z- score model and EDF model, and then compares the forecast accuracy of these models by the Accuracy Ratios（AR）. From the selected samples, we find that the AR can preferably reflect the accuracy of both models.