Study of Credit Risk for Listed Companies under Logistic Model Based on Case-control Data

In risk management, we often assess the credit risk for companies; especially study the default risk for companies with a high liability-to-asset rate. In this paper, logistic regression model is used to assess the listed companies' credit risk. Selecting listed companies' 63 financial indicators reflecting companies' credit characteristics as initial independent variables, we establish logistic regression model based on case-control data. We also determine financial indicators measuring listed companies' credit risk by hypothesis testing and factor analysis, and establish logistic regression model based on case-control data by changing different sample proportion of ST companies and non-ST companies. Finally, we can get the conclusion that predicted accuracy for logistic regression model based on case-control data with sample proportion 1:1; 1:2 is high. For risk assessment, it is an effective method to establish logistic regression model based on case-control data. It is more practical and the probability of committing Type I error or Type II error is also low.