基于POT方法的商业银行操作风险极端值估计  

Estimating Extreme Values of Operational Risk for Commercial Banks via Peak over Threshold Method

作　　者：高丽君[1] 李建平[2] 徐伟宣[2] 王书平[3] 

机构地区：[1]山东财政学院工商管理学院,济南250014 [2]中国科学院科技政策与管理科学研究所,北京100080 [3]北方工业大学经济管理学院,北京100041

出　　处：《运筹与管理》2007年第1期112-117,共6页Operations Research and Management Science

基　　金：国家自然科学基金资助项目(70531040);中国科学院基金资助项目(yjjz946);科技政策与管理科学研究所所长基金(0600281501)

摘　　要：对于商业银行而言,操作风险已经成为与市场风险和信用风险同样重要的风险。本文利用极值理论超越样本的估计能力,采用极值理论中对数据要求量较少,可以进行单步预测的超阈值(POT)方法对我国商业银行操作损失极端值分布进行估计,以均值超额函数图和拟合直线的交点确定阈值,估计出给定置信水平之下操作风险损失的分位数,从而使得国内商业银行操作风险监管资本的计算成为可能。During the recent years, the operational risk has become an important research field in banks risk management. It has been recognized as the major risks for commercial banks together with the market risk and the credit risk. By using peak over threshold method （POT）, which requires less data than other methods, and based on the extreme value theory method, this article estimates the extreme loss distribution and the p-percentile operational loss under certain belief range for Chinese commercial banks, then calculates the operational risk supervise capital. The results are encouraging and the proposed method may be a useful tool for estimating the operational risk.

关 键 词：金融学 POT方法 极值理论 操作风险 均值超额函数图 

分 类 号：F830.9[经济管理—金融学]