基于DebtRank算法的银行系统性风险仿真研究  

Research on Bank Systemic Risk Simulation Based on DebtRank algorithm

作　　者：范宏[1] 庞琮远 FAN Hong;PANG Cong-yuan(Glorious Sun School of Business and Management,Donghua University,Shanghai 200051,China)

机构地区：[1]东华大学旭日工商管理学院,上海200051

出　　处：《计算机仿真》2023年第9期256-261,共6页Computer Simulation

基　　金：国家自然科学基金面上项目(71971054);上海市自然科学基金项目(19ZR1402100);东华大学研究生创新基金资助,项目编号:GSIF-DH-M-2022013。

摘　　要：银行系统性风险是指一个或几个重要银行机构的违约通过银行网络引起的大范围的银行机构违约风险。目前,大部分的学者通过单一渠道来仿真研究银行系统性风险,而且以银行倒闭的数量来判定银行系统性风险,但是,现实世界中,发生银行倒闭的事件很少,很难用银行倒闭的数量来判定银行系统性风险。而债务等级法的判定,不需要有银行倒闭,就可以用来衡量整个银行系统的系统性风险,但是目前,采用债务等级法(DebtRank)来判定银行系统性风险的研究还缺乏。另外针对双渠道传染的银行系统性风险的研究也少见。为解决上述问题,首先构建双渠道传染模型,然后基于DebtRank算法构建银行系统的债务等级,利用银行系统的债务等级评判银行系统的系统性风险,进一步研究杠杆、平均连接度对银行系统性风险的影响。研究结果表明:杠杆和平均连接度对银行系统的系统性风险有较大影响;杠杆对银行系统性风险具有单调增加的作用,且直接传染渠道占优下作用更大;平均连接度对间接传染渠道占优下的银行系统性风险具有单调增加的作用,而直接传染渠道占优下只有当其较小时才具有单调增加的作用,当平均连接度较大时,反而是单调减少的作用。进一步研究发现,不同传染渠道之间存在一个(杠杆和连接度的)阈值,随着平均连接度的增大,阈值会在更大的杠杆下产生;同样的,随着杠杆的增大,阈值需要在更大的平均连接度下产生。上述研究可以为中央银行的政策制定提供一定的决策依据。Banking systemic risk refers to the large-scale default risk of banking institutions caused by the default of one or several important banking institutions through the banking network.At present,most scholars use a single channel to simulate and study the systemic risks of banks and use the number of bank failures to determine the systemic risks of banks.However,in the real world,bank failures are rare and it is difficult to use bank failures.Quantity to determine the bank's systemic risk.The determination of the debt ranking method can be used to measure the systemic risk of the entire banking system without bank failures.However,at present,there is still a lack of research on the use of the debt ranking method(DebtRank)to determine the systemic risk of banks.In addition,research on the systemic risks of dual-channel infections in banks is rare.Therefore,this article first builds a dual-channel contagion model,and then builds the debt rating of the banking system based on the DebtRank algorithm,uses the debt rating of the banking system to evaluate the systemic risk of the banking system,and further studies the impact of leverage and average connectivity on the systemic risk of the bank.The research results show that:leverage and average connectivity have a greater impact on the systemic risk of the banking system;leverage has a monotonous increase in the systemic risk of banks,and direct infection channels dominate and the effect is greater;the average degree of connectivity has a monotonically increasing effect on the systemic risk of banks under the dominant indirect infection channel,while the direct injection channel has the advantage of monotonous increase only when it is smaller.When the average degree of connectivity is large,it will increase monotonically.The effect of monotonous reduction.Further research has found that there is a threshold(leverage and connectivity)between different channels of infection.As the average connectivity increases,the threshold will be generated under greater leverage;simil

关 键 词：债务等级 杠杆 平均连接度 同业拆借 共同持有资产 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]