基于时变状态网络的银行风险传导研究  

Risk transmission between banks based on time-varying state network

作　　者：邱路 黄国妍[1] Qiu Lu;Huang Guo-Yan(School of Finance and Business,Shanghai Normal University,Shanghai 200234,China;School of Business,East China University of Science and Technology,Shanghai 200237,China)

机构地区：[1]上海师范大学商学院,上海200234 [2]华东理工大学商学院,上海200237

出　　处：《物理学报》2020年第13期310-321,共12页Acta Physica Sinica

基　　金：教育部人文社会科学研究青年基金(批准号:18YJC910010);教育部规划基金项目(批准号:18YJA790036)资助的课题;上海师范大学校级人文社会科学研究项目(A-7031-18-004023)。

摘　　要：银行风险传导研究是系统性金融风险测度与防范的重点.文献中主要研究静态银行网络拓扑结构和银行系统性风险量化,较少考虑银行网络之间的状态转变.针对上述问题,提出时变状态网络模型.根据模型,首先用kmeans方法对各个时间段的银行网络分类,然后通过有向最小生成树(DMST,directed minimum spanning tree)分析每一类银行网络的拓扑结构,最后联合利用平面最大过滤图(PMFG,planar maximally filtered graph)方法构建时变银行状态网络,该网络可用作银行风险源头的寻找和传导时变性分析.利用时变状态网络模型研究我国15家上市商业银行2007年第四季度到2019第一季度同业拆借数据.结果表明,银行状态网络之间的短期连续跳跃性可有效描述金融危机的发生,比如2008年全球金融危机发生前后出现了两个状态间的短期跳跃,从2013年“钱荒”到2015年的股灾阶段,先后出现了四个状态间的短期跳跃.同时,各个有向银行状态网络的出度和传染效应成正比,入度和银行面临风险的稳健程度成反比.时序的银行状态网络具有记忆性特征,这可以为央行防范系统性风险提供决策依据.Aiming at the state transition between bank networks,we propose a time-varying state network model.In this model,we classify the bank networks in each time period by the kmeans method,and use directed minimum spanning tree(DMST)method to describe the topological structure of each kind of bank network.We also construct a time-varying bank state network by the planar maximally filtered graph(PMFG)method.The state network can be used to find the source of bank risk and conduct the time-varying analysis.We put into the model the inter-bank lending data of 15 listed Chinese commercial banks from the fourth quarter of 2007 to the first quarter of 2019.The results show that the short-term continuity jump between the bank states can effectively describe the occurrence of financial crisis.For example,before and after the global financial crisis in 2008,there was a short-term jump between two states.From the“money shortage”in 2013 to the stock market crash in 2015,there were four short-term jumps between states.At the same time,the outgoing degree of each directed bank state network is directly proportional to the contagion effect,and the incoming degree is inversely proportional to the steady degree of the risk faced by the bank.The sequential bank state network has the memory characteristic,which can provide the central bank for decision basis to prevent the systematic risk.

关 键 词：银行网络 有向最小生成树 风险传染强度 平面最大过滤图 

分 类 号：F832.33[经济管理—金融学]