数据整合视角下商业银行操作风险度量研究  

Research on Operational Risk Measurement of Commercial Banks from the Perspective of Data Integration

作　　者：谢俊明 胡炳惠 Xie Junming;Hu Binghui(School of Economics and Management,Xiangnan University,Chenzhou 423001,Hunan,China;Shenzhen Branch of China Construction Bank,Shenzhen 518060,Guangdong,China)

机构地区：[1]湘南学院经济与管理学院,湖南郴州423001 [2]中国建设银行深圳市分行,广东深圳518060

出　　处：《征信》2023年第4期72-77,86,共7页Credit Reference

基　　金：国家社会科学基金一般项目(22BJL041)。

摘　　要：随着商业银行改革不断深化,操作风险呈高发态势。整合数据是商业银行准确度量操作风险的基础保证。选取1994—2020年间我国四大国有商业银行1846例操作风险事件,充分利用彼此的损失信息互相补充,结合损失分布法、POT模型以及信度模型对四大商业银行的操作风险进行度量。结果表明:对于一般操作风险损失,对数正态分布的拟合效果优于其他分布;对于极端操作风险损失,POT模型能够较好地对操作风险的尾部进行拟合;分段拟合更能准确描述操作风险的损失特征;运用信度理论可以更准确地对操作风险水平作出合理的估计,弥补数据缺失造成的不足。With the deepening of commercial bank reform,operational risks are at a high level.Integrating data is the basic guarantee for accurate measurement of operational risk of commercial banks.This paper selects 1846 operational risk events of the four major state-owned commercial banks in China from 1994 to 2020,makes full use of each other's loss information to complement each other,and measures the operational risk of the four major commercial banks by combining loss distribution method,POT model and credibility model.The results show that for general operational risk loss,the fitting effect of log-normal distribution is better than other distributions;for extreme operational risk loss,the POT model can fit the tails of operational risk well;segmental fitting can more accurately describe the loss characteristics of operational risk;using the credibility theory can more accurately make reasonable estimates of operational risk levels and make up for the deficiencies caused by missing data.

关 键 词：数据整合 操作风险度量 损失分布法 POT模型 信度模型 风险资本金 

分 类 号：F832.3[经济管理—金融学] F224