A Study On Performance-Based Enterprises' Credit Risk Models

With the further opening of Chinese financial market, especially in the period between China's entry into WTO and the expiry of the term of banking industry protection in 2007, the issue on how to take precaution measures against credit risk and improve banks' risk resistance ability has become more and more important.In foreign countries, studies on credit risk models got started long time ago. However, Chinese scholars' research in this aspect still remains in an early stage and lags far behind compared with overseas study in both study method and study angle. In addition, current credit risk models still have many defects. Therefore, based on a review of material and documents in this field home and abroad and a summarization and evaluation of past research results, this thesis makes a comparative analysis on the effect of some established credit risk models through empirical analysis. By taking the default rate as dependant variable and the index of cash flow as independent variable, this thesis analyzes andscreens 24 indexes of cash flow via stepwise regression and cluster analysis; then 6 indexes are selected as discriminating variables owing to their outstanding qualities to forecast enterprises' credit risk. Furthermore, the Multivariate Discriminant Analysis and Logistic model are employed in this thesis to establish Chinese enterprises' credit risk models. A comparison between these two models' effects, applications and limitations are made upon the completion of model establishment.The arrangement of this thesis is as follows: chapter 1 introduces relevant theories about enterprise-performance-based credit risk models; chapter 2 focuses on literature review and chapter 3 presents an empirical study on information coverage level of each cash flow index; in chapter 4 two enterprise-performance-based credit risk models are established; and in the last chapter research summary, application conditions and limitations are highlighted and discussed.