坏帐率测定和应用中存在的误差

对两种典型的坏帐率统计方法进行比较，发现传统的统计方法存在很大的误差，从而影响信用风险的测定精度。由于信用机构迟缓更新转移概率和个别较低信用质量企业的影响，坏帐率区间存在很大的重叠，且平均坏帐率显著高于中值坏帐率。解决问题的方法是基于期望坏帐频率（EDF）的非重叠区间来划分企业，指定这些区间对应债务等级的区间。

Some Error in Measurement and Application of Default Rate

By comparing two typical statistics methods of default rate, the article found a serious error in the statistics method that will impact on the precision of credit risk measurement. Because credit institutions delay to update transition matrix and the impact of a few corporations with lower credit quality, there is a substantial overlap in the default rate ranges and the mean default rate significantly exceeds the median default rates. The way to resolve the problem is to classify firms into groups based upon non overlapping ranges of expected default frequencies and choosing these ranges to correspond to the rating grade ranges.