基于极值理论和多元Copula函数的商业银行操作风险计量研究

中国管理科学 ›› 2013, Vol. 21 ›› Issue (3): 11-19.

基于极值理论和多元Copula函数的商业银行操作风险计量研究

陆静, 张佳

重庆大学经济与工商管理学院, 重庆 400030

收稿日期:2011-10-02 修回日期:2012-03-29 出版日期:2013-06-30 发布日期:2013-06-20
基金资助:
国家自然科学基金资助项目(71232004,71272085);国家社会科学基金资助项目(09BJL024);教育部人文社会科学基金资助项目(12YJA630135);重庆市自然科学基金资助(2009BB2042)

Measurement of Commercial Bank’s Operational Risk Based on Extreme Value Theory and Multivariate Copula Functions

LU Jing, ZHANG Jia

School of Economics and BusinessAdministration, Chongqing University, Chongqing 400030, China

Received:2011-10-02 Revised:2012-03-29 Online:2013-06-30 Published:2013-06-20

摘要/Abstract

摘要： 基于操作风险呈厚尾分布的特征,本文按照巴塞尔协议的要求,采用POT极值模型分别估计了多个操作风险单元的边缘分布,然后用多元Copula函数来刻画这些操作风险单元之间的关联性并计算在险价值。通过对中国商业银行1990-2010年操作风险数据的实证分析表明,Clayton Copula能更好地反映各操作风险单元之间的相关性结构,且采用Copula考虑操作风险相关性下的VaR值要比简单加总下的VaR值减少约32.3%。因此,应用Copula函数计量操作风险相关性,不仅可以提高估计的准确性,还能够达到资产组合的风险分散化效应,减少操作风险资本要求,为商业银行提升盈利能力创造条件。

关键词: 操作风险, 极值理论, 多元Copula函数, 商业银行

Abstract: Owing to the fat tail of operation risk and based on the requirement of Basel Accord, marginal distributions of several operational risk cells are measured with the Peaks-Over-Threshold model of extreme value theory, dependency of operational risks is analyzed with multivariate copula functions and the accumulated VaR is calculated. By using the data of Chinese commercial banking from 1990 to 2010, it is shown that Clayton Copula can better reflect the dependent frame of operation risk and the VaR calculated with Clayton Copula is about 32.3% less than the one calculated by directly summing VaRs of all operation risk cells. Therefore, consideration on dependency of operation risk cells with Copula functions can meet the risk dispersion effect of asset portfolio, greatly reduce required operation risk capital and provide commercial banks with a better way to get more profits.

Key words: operational risk, extreme value theory, multivariate copula, commercial banks

中图分类号:

F832.33

陆静, 张佳. 基于极值理论和多元Copula函数的商业银行操作风险计量研究[J]. 中国管理科学, 2013, 21(3): 11-19.

LU Jing, ZHANG Jia. Measurement of Commercial Bank’s Operational Risk Based on Extreme Value Theory and Multivariate Copula Functions[J]. Chinese Journal of Management Science, 2013, 21(3): 11-19.

参考文献

[1] Fontnouvelle, P, Rosengren E S, John S J. Implications of alternative operational risk modeling techniques[J]. The Risks of Financial Institutions, University of Chicago Press, 2007.
[2] 中国银行业监督管理委员会. 商业银行资本管理办法[Z]. 2012.
[3] 周艳菊,彭俊,王宗润. 基于Bayesian-Copula方法的商业银行操作风险度量[J]. 中国管理科学, 2011, 4: 17-25.
[4] 丰吉闯,李建平,高丽君. 商业银行操作风险度量模型选择分析[J]. 国际金融研究, 2011, 8: 88-96.
[5] 樊欣,杨晓光. 我国商业银行操作风险的蒙特卡罗模拟估计[J]. 系统工程理论与实践, 2005, 5: 12-19.
[6] 张宏毅,陆静. 运用损失分布法的计量商业银行操作风险[J]. 系统工程学报, 2008, 4: 411-416.
[7] Frachot A, Moudoulaud O. Roncalli T. Loss distribution approach in practice. Working Paper, Credit Lyonnais, 2003.
[8] Turk A B. Quantitative operational risk management[M]//Ginancarlo N Advances in risk management Sciyo, 2010.
[9] 吴恒煜,赵平. 我国商业银行操作风险的度量——基于极值理论的研究[J]. 山西财经大学学报, 2009, 8: 109-115.
[10] 张文, 张屹山. 应用极值理论度量商业银行操作风险的实证研究[J]. 南方金融, 2007, 2: 12-15.
[11] 高丽君,李建平,徐伟宣,等. 基于POT方法的商业银行操作风险极端值估计[J]. 运筹与管理, 2007, 2: 112-117.
[12] 高丽君,李建平. 我国商业银行操作风险模拟估计[J]. 山东财政学院学报, 2009, 5: 55-58.
[13] 谭德俊,邹敏烨. 操作风险损失的广义帕累托分布参数估计及其应用[J]. 财经理论与实践, 2010, 6: 22-25.
[14] 李宝宝,王言峰. 基于CVaR-POT模型的我国银行业操作风险度量研究[J]. 华东经济管理究, 2011, 7: 76-79.
[15] 厉吉斌. 商业银行操作风险管理[M]. 上海:上海财经大学出版社, 2008: 215-221.
[16] 丰吉闯,李建平,陈建明. 基于左截尾数据的损失分布法度量操作风险:以中国商业银行为例[J]. 管理评论,2011, 7: 171-175.
[17] Embrechts P, Hoing A, Juri A. Using Copulae to bound the Value at Risk for functions of dependent risks[J]. Finance and stochastics, 2003, 7(2): 145-167.
[18] Embrechts P, Kaufmann R, Patie P. Strategic long-term financial risks single risk factors[J]. Computational Optimization and Applications, 2005, (32): 61-90.
[19] Nelsen R. An introduction to Copulas [M]. New York: Springer, 1999.
[20] 高丽君. 我国商业银行系统操作风险资本金的度量[J]. 山东财政学院学报, 2007, 6: 42-45.