跳跃—扩散条件下信用风险相关性度量的变结构Copula模型

发布时间：2018-12-16 01:52

【摘要】：针对现有研究大多只考虑扩散条件的不足,构建了跳跃—扩散条件下信用风险相关性度量的变结构Copula模型。运用1991~2010年中国上市公司的数据构建了行业信用风险指数,运用双指数跳跃扩散模型来识别行业信用风险的跳跃扩散点,发现在样本期,共同因素与行业特质因素引发了行业信用风险的多次跳跃。在识别跳跃点的基础上,构建了变结构Copula模型,该模型能较准确地描述信用风险相关性的变化,各行业之间的信用风险相关系数在0.5以上,并且上市公司信用风险的变化呈现出"一损俱损"的特征,而"一荣俱荣"的特征并不明显。构建的模型及实证结论将有助于理解信用风险相关或传染,从而为信贷组合管理和风险管理提供更多的方法与经验。
[Abstract]:In view of the fact that most of the existing researches only consider the deficiency of diffusion condition, a variable structure Copula model for credit risk correlation measurement under jump-diffusion condition is constructed. Based on the data of China's listed companies from 1991 to 2010, this paper constructs the industry credit risk index, uses the double index jump diffusion model to identify the jump diffusion point of the industry credit risk, and finds out that in the sample period, The common factor and the industry characteristic factor caused the industry credit risk to jump many times. On the basis of identifying jump points, a variable structure Copula model is constructed. The model can accurately describe the change of credit risk correlation, and the correlation coefficient of credit risk among different industries is more than 0. 5. Moreover, the change of credit risk of listed companies is characterized by "all losses", but the characteristics of "prosperity and prosperity" are not obvious. The model and empirical conclusions will be helpful to understand credit risk related or contagion, so as to provide more methods and experience for credit portfolio management and risk management.
【作者单位】：湖南大学工商管理学院;湖南商学院财政金融学院;
【基金】：国家自然科学基金创新研究群体项目(71221001) 国家社科基金资助项目(11BTJ011) 教育部人文哲学社会科学青年基金项目(13YJCZH123) 中国博士后科研基金项目(2013M542111)
【分类号】：F830.91;F224

【参考文献】