Comparison between Static and Dynamic Models of Credit Risk Measurement: Evidence from Chinese Listed Companies

This study attempts to choose an applicable credit risk measurement model to forecast the future status of Chinese listed firms among three traditional credit risk models: Z-score model, logit model and hazard model by using the sample of A-shares listed firms in Shanghai and Shenzhen Stock Exchange during the period from 2008 to 2014. Z-score model and logit model are static models that only use one period data to forecast, while hazard model as a dynamic model uses multi-period data. Based on existing western and Chinese studies, seven financial ratios from seven classifications of profitability, liquidity, leverage, asset management, capitalisation, cash flow and firm size are chosen as independent variables. Results show that profitability, capitalisation, the efficiency of asset management, and firm size play more important role than others in predicting the bankruptcy of Chinese firms after financial crisis. By comparing the number of correct classification in both in-sample and out-of-sample forecasts, the performance of Z-score model appears to be better than the other two models, which is inconsistent from previous studies. Key Words: Special Treatment, Z-score Model, Logit Model, Hazard Model, Financial Ratio