标题：The quantile regression - Mixture copula model applied in the financial tail risk contagion
作者：Liu, Ning ;Wang, Peizhi ;Dong, Jieyu ;Liu, Jing
作者机构：[Liu, Ning ;Wang, Peizhi ;Dong, Jieyu ;Liu, Jing ] School of International Trade and Economics, Shandong University of Finance and Economics, Jinan, C 更多
来源：IPPTA: Quarterly Journal of Indian Pulp and Paper Technical Association
关键词：Mixture copula function; Quantile regression; Risks contagion; Tail dependence
摘要：The joint distribution of financial market data has the characteristics of non-normality, heteroscedasticity, obvious peak and fat tail. Therefore, we choose the quantile regression model to filter the return of financial market and obtain an innovation sequence which obeys independent identical distribution. Then we fit the marginal distribution of innovation sequence. Considering that there are asymmetric dependent relations between the tail structure of different financial market, we build a mixture Copula function model based on Gumbel Copula, Frank Copula and Clayton Copula to analyze the tail dependence and risk contagion between the Chinese and US stock markets, especially the tail risk contagion effect of the two markets drastic rising together or falling together. The results show that, when we analyze the tail dependent risks, we use this method to overcome the estimation error which happens because the non-independent identical distribution of sequence can’t meet the requirements of building mixture Copula model. For another, mixture Copula model based on quantile regression has more advantages in depicting financial markets’ tail structure, especially in the aspect of lower tail dependence. © 2018 Indian Pulp and Paper Technical Association. All Rights Reserved.