The small sample credit risk rating and its empirical study

Abstract

Credit risk rating is used to judge the risk degree of different enterprises and provides some references for measuring the default rates of different enterprises. The sufficient sample number can guarantee the accuracy and rationality of credit rating. It is very difficult to find the ideal large samples in reality, especially in the banking. This study finds out the distribution rule of rating scores by use of testing the score distribution of small sample commercial bank credit evaluation. According to the distribution rule, the sample data are expanded and the bank credit risk ranks are divided. Firstly, this study reveals the logarithmic distribution rule of commercial bank credit scores in China by testing the distribution rule of rating scores. In the face of limited sample date of commercial banks, the problem how the data are expanded in credit rating is solved. Secondly, the credit ranks of 41 Chinese banks who publish the annals can be fixed based on the credit scores of credit ranks of 1000 simulated banks, which solves the problem that nine credit ranks can't be divided reasonably with small samples of 41 commercial banks. So this study is of importance for those countries which have few and concentrated banks to do some ratings.