基于EVT-Copula的操作风险度量  

Operational Risk Measurement Based on EVT-Copula

作　　者：宋加山[1] 张鹏飞[1] 王利宏[1] 王彪[2] 

机构地区：[1]西南科技大学经济管理学院,四川绵阳621010 [2]中国科学技术大学管理学院,安徽合肥230022

出　　处：《预测》2015年第3期70-73,共4页Forecasting

基　　金：四川省科技厅软科学研究计划资助项目(2013ZR0097);西南科技大学科研基金资助项目(13sxt012)

摘　　要：新巴塞尔协议把操作风险纳入风险量化和监管领域,要求国际活跃商业银行开发的操作风险计量模型能够处理操作风险损失概率分布厚尾特征。并明确建议通过损失分布法等高级方法来度量操作风险。而使用损失分布法的计量模型没有考虑业务线/事件类型之间的相关性,这与实际情况是不相符合的。为此本文运用极值理论模拟损失分布,建立计算操作风险总Va R值的EVT-Copula模型,并在此基础上运用Copula函数度量银行各类业务操作风险之间的相依性,得到整体Va R值的模拟值。The New Basel Capital Accord takes operational risk into the region of risk quantification and supervision, demands operational risk measurement models developed by active commercial banks can be capable of handling the situation of thick tailed distribution on probability distribution losses in the operational risk, definitely suggests measuring operational risk by high-level approaches such as losses distribution. And yet the measurement model of losses distribution leaves out of consideration of the relevance between business line and event type, which is not conform to the actual situation. For this reason, this paper builds a EVT-Copula Model of total value VaR to calculating operational risk, using the extreme value theory to simulate losses distribution, to get an analogue value of the whole VaR value.

关 键 词：COPULA函数 极值理论 在险值 

分 类 号：F832.2[经济管理—金融学]