基于分块极大值模型的商业银行操作风险计量研究

尽管高级计量法由于具有计算精确和节约监管资本等优点而被多数商业银行所青睐,但对于采用哪一种方法来刻画低频高危的操作风险尾部数据却没有一致认识。本文根据巴塞尔委员会关于操作风险计量的原则,采用分块极大值方法和概率加权矩参数估计法,对中国商业银行1990—2009年间的操作风险数据进行了实证。从图形检验和数值检验结果来看,该模型估计的参数具有较高的拟合优度,能够较好地拟合操作风险极端值的尾部分布,为商业银行计量操作风险资本提供了较高的参考价值。

Measuring Operational Risk of Commercial Banks Based on the Block Maxima Model

Basel II theory asserts that operational risk market risk and credit risk are three major risks for commercial banks and commercial banks to measure operational risks and allocate regulatory capital.However,in practice commercial banks have been challenging with operational risks that have high loss and low frequency.These events have a significant fat-tail nature of modeling..The traditional econometric models generally require a considerable amount of loss data be used.Therefore,the construction of operational risk measurement model to measure credit and market risks is more difficult.This paper uses the publicly disclosed data of operational risks from 1990 to 2009 about China′s commercial banks,and adopts the block maxima method and probability weighted moments to estimates the operational risk of Chinese commercial banks. Although the Basel Committee does not provide any specific methods and statistical distribution assumptions,banks must demonstrate that their risk assessment approaches take into account the probability distribution of serious tail loss events.The bank must also show that the operational risk measurement methods comply with IRB method and fairly robust standards when assessing credit risk.The Basel II is more concerned about the size of operational risk of banks in a given time period.Although traditional POT model can better describe the tail of the distribution of loss events,but it neglects the time factor and cannot fully meet the requirements of Basel II.Therefore,we use the block maxima method(BMM) to estimate the tail of the operational risk data.BMM selects the extreme according to the time period of loss data,which is also consistent with the principle of stability Basel II. The study finds that China′s domestic operational risk cases concentrate in the commercial banking business and payment and settlement business.These two lines of businesses account for 50.34% and 42.14% of operational risks,respectively.All of these risks can be categorized into internal fraud and external fraud,accounting for 71.53% and 28.02% respectively.Classified according to size and scope,the sample data is divided into four groups: all banks,four state-owned banks,joint-stock banks and local commercial banks.In accordance with Basel II 99.9% confidence level,the four group banks need to prepare the following operational risk capitals: 1.195 billion yuan,1.09 billion yuan,2.33 billion yuan and 266 million yuan respectively.This finding indicates that the operational risk capital of state-owned banks is 4.68 times and 4.1 times more than that of joint-stock banks and local commercial banks.However,if considering the size of banks,the results are different.Taking the total assets of 2009 as a standard and comparing the ratio of operational risk capital to total assets,the ratio of state-owned banks is 0.0136%,the ratio of joint-stock banks is 0.0233%,and the ratio of local commercial banks is 0.39%.These findings indicate that the state-owned bank′s operational risk is relatively small,while the local commercial banks need to allocate more operational risk capital and the operational risk of local commercial banks is too high.