系统性信用风险的网络传染联动效应研究

发布时间：2018-03-30 20:24

本文选题：信用网络传染联动　切入点：非线性股价联动　出处：《北京理工大学学报(社会科学版)》2014年04期

【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。
[Abstract]:By using the credit asset network model, this paper deduces the stock price linkage mechanism caused by network contagion of credit assets, estimates the marginal distribution parameters described by the first order autoregressive asymmetric generalized heteroscedasticity model, and selects the connection function of multivariate t distribution. The global correlation coefficient is calculated by static fitting, the time-varying process is chosen as the multivariate connection function with dynamic condition autocorrelation, and the correlation coefficients of four time-varying conditions under C-vine decomposition structure are calculated dynamically. The paper describes and analyzes its time-varying characteristics. The results show that there is a strong network contagion linkage effect between the stock prices of credit assets associated enterprises, and the static and dynamic correlation measures of the connection function with multiple t distribution are presented. The linkage effect of network contagion can be well measured, both unconditional correlation coefficient and conditional correlation coefficient show significant variability in varying degrees. Network core enterprises should prudently manage the systemic risk of credit asset association.
【作者单位】：上海师范大学商学院;
【基金】：教育部人文社科基金资助项目(11YJA790107)
【分类号】：F224;F832.51

【参考文献】