Research Of The Credit Risk Measuring Models For China's Listed Companies

Stock market has developed for 17 years in China, the way of securities investment instead of saving money has been accepted by people. However, the phenomenon of disobeying the rules is serious, that listed companies are "special treated" because of financial abonomity occurred usually, which makes all interest-relative people of listed companies, include shareholders, creditors, managers, government and workers, pay attention to the credit status of these companies. So developing a credit measuring model adapt for China's stock market becomes a hot issue in the academical and financial area. This thesis will do a deep research to this issue based on summing up former researchful productions, the concrete contents are as follows:Firstly, the thesis developed the factor analysis model and BP neural network model of credit risk measuring based on the financial ratios.Secondly, the credit risk measuring KMV model was amended according to the characteristics of China's stock market, the adaptability of this model was validated, so the dynamic credit risk measuring model based on the data of market (price of stocks) was obtained.Lastly, a frame of Credit Decision Support System was developed.These conclusions were obtained:The whole accuracy of BP neural network model is better than the factor analysise model. The amended KMV model was adapt to China's stock market in a certain extent, but if you want to develope the corresponding relation between Distance to Default and Expected Default Frequency, the default database of listed companies must be constructed. Distance to Default computed from KMV model can be used to either monitor the credit risk status, or divide listed companies to three kinds, and the BP neural network model was suitable for divide listed companies to two kinds. If you want to put the researchful productions into industrialization, you must develop a Credit Decision Support System.