Early warning model of crisis in Chinese commercial banks based on gray relational analysis with double standard

China lacks examples of bank crises and reference standards for classifying such crises. This study addresses these problem by designing an ideal bank crisis based on four types of early warning indices of endogenous risk, capital and asset safety, profitability, and liquidity. An early warning model of commercial bank crises is created based on a novel modified Gray relational analysis. This model is capable of classifying banks into three levels of crisis and of evaluating the final crisis value. This study also performs an empirical analysis. Results demonstrate that an early warning model of commercial bank crises classifies and orders the sample data based on forecasting criticality. The model was tested during the risky events in China's commercial banks in mid-2013. The 62 sample banks were all preferably safe in 2012. The risk control of the city commercial banks was better than that of large and joint-stock commercial banks. However, the possibility of crisis growth in China's commercial banks remains.