Abstract: The tail risk network of financial institutions including its correlation mode and structures plays an important role in the formation and evolution of financial systemic risk. This paper measures the tail risk contagion among financial institutions by using CoVaR(conditional value-at-risk)and quantile regression method, and then constructs dynamic tail risk networks of financial institutions. We analyze the topology structures and their evolution rules of fully connected and filtered tail risk networks. At last, we investigate the effects of network topology structures on the financial institutions' systemic risk contribution. The empirical study show that the node strength can well demonstrate the tail risk contagion and tolerance strength and their dynamic evolution of different institutions. The ranks of tail risk contagion and tolerance strength are different from each other for different institutions. As time evolves, the average tail risk contagion paths shorten and the systemic risk tend to be accumulated faster. On one hand, the lag one node out degree, lag one node in degree and lag one node clustering coefficient have significantly negative effects on the systemic risk contribution of institutions. On the other hand, the lag one node betweenness and node closeness centrality have significantly positive effects on the systemic risk contribution of institutions. The results provide a completely new perspective of tail risk networks for the macroprudential regulation of financial institutions and systemic risk management.

Key words: tail risk network, conditional value-at-risk(CoVaR), systemic risk contribution, risk contagion, network topology structure

摘要： 金融机构的尾部风险关联模式及结构在金融系统性风险的形成演化中发挥重要作用。利用CoVaR指标及分位数回归方法,衡量金融机构之间的尾部风险传染强度,进而建立金融机构尾部风险动态网络。分析全连接网络及阈值法下过滤网络的全局和局部拓扑结构特征及其动态演化规律。建立面板数据回归模型,研究网络拓扑结构特征对金融机构系统性风险贡献的影响。实证研究发现,全连接网络的节点强度,能有效地衡量金融机构尾部风险传染强度及承受强度,并揭示其动态变化规律;各机构的尾部风险传染强度及承受强度排序匹配性存在差异;随着时间推进,各金融机构间的平均尾部风险传染路径缩短、系统性风险更易迅速累积和爆发;滞后一期的节点出度、节点入度及聚集系数越大,相应金融机构的系统性风险贡献越小;滞后一期的节点介数和节点接近中心度越大,相应金融机构的系统性风险贡献越大。研究结果对于金融机构的宏观审慎监管及系统性风险管理,提供了一个全新的基于金融机构尾部风险网络的视角。

关键词: 尾部风险网络, 条件在险价值, 系统性风险贡献, 风险传染, 网络拓扑结构