Abstract: According to the shortcoming of the current literatures that stress more on the diffusion process, variable structure Copula models are constructed for credit risk correlation evaluation under the condition of jump-diffusion process. By applying the Chinese Listed Companies data from 1991 to 2010, the industry indexes of credit risk are calculated, and the empirical results show that the common factor and the industry idiosyncrasy factors lead the jump several times during the sample period. On the basis of identifying the variable structure points, the variable structure copula models are constructed. The results show that the models can describe the credit risk correlation accurately, and the dynamic correlation coefficients are above 0.5 for all industry match-up. Meanwhile, it can also found that the credit risk of different companies rise together, but it may not fall together. The models and conclusions are helpful to understand the credit risk correlation or contagion, thus they can provide more methods and instructions for the credit portfolio management and risk management.

Key words: jump-diffusion process, double exponential index distribution, variable structure Copula models, credit risk correlation