系统性信用风险的网络传染联动效应研究  

Research on the Network Linkage Effects of Systemic Credit Risk's Infection

作　　者：茆训诚[1] 王周伟[1] 

机构地区：[1]上海师范大学商学院,上海200234

出　　处：《北京理工大学学报（社会科学版）》2014年第4期57-63,130,共8页Journal of Beijing Institute of Technology：Social Sciences Edition

基　　金：教育部人文社科基金资助项目(11YJA790107)

摘　　要：利用信用资产网络模型,演绎了信用资产网络传染引发的股价联动机理,估计了一阶自回归的不对称广义异方差模型描述的边缘分布参数。选用多元t分布的连接函数,静态拟合计算了全局相关系数;选用时变过程为动态条件自相关的多元连接函数,动态计算了C藤分解结构下的4种时变条件相关系数,描述分析了其时变特征。研究结果表明:信用资产关联企业的股价之间具有较强的网络传染联动效应;多元t分布的连接函数的静态与动态相关度量方法,可以较好地度量网络传染联动效应;无论无条件相关系数还是条件相关系数,都表现出不同程度的显著时变性。网络核心企业应当审慎管理信用资产关联上的系统性风险。By using credit asset network model, we describe the price linkage mechanism of infection caused by a network of credit assets, and for more representative of Baosteel supply chain, estimate the distribution 's parameters of the asymmetric generalized heteroscedasticity model. And then we choose multivariate distribution function, to fit correlation coefficients, make use of multiple connections function of the dynamic conditional autocorrelation to calculate four correlation coefficients for C-vine structure decomposition dynamically, describe and analyze the time variation characteristics. Research shows that, there are strong network transmission linkage effects on the stock price of enterprises that have associated credit assets. The static and dynamic correlation measurement method based on the connection function of a multivariate distribution can be used to measure the network transmission linkage effect, both unconditional and conditional correlation coefficients are showing varying degrees of significant time-varying.

关 键 词：信用网络传染联动 非线性股价联动 平方欧式距离 

分 类 号：F224[经济管理—国民经济] F832.51