基于Copula的商业银行操作风险度量模型及其应用

发布时间：2018-04-13 13:27

本文选题：操作风险 + LDA　；参考：《江西师范大学》2012年硕士论文

【摘要】：巴塞尔《新资本协议》明确地提出,把商业银行操作风险纳入风险量化和监管领域.继信用风险和市场风险之后,操作风险成为了商业银行面临的主要风险之一. 本文首先应用极值理论中的POT模型对损失分布法LDA进行改进,并计算得出我国商业银行的两类主要操作风险（内部欺诈与外部欺诈）的损失分布以及用于度量两者风险大小的VaR值,在有效捕捉损失厚尾性的同时遵循了操作风险准备资本计提的规律;然后,为便于比较分析,论文运用三种Copula函数来分别构造内外部欺诈损失间的相关性结构,并在此基础上构建用于计算内外部欺诈联合损失分布VaR值的Copula模型.最后,在实证研究中,本文分别利用三种Copula模型与传统计量模型相比较后发现,基于Gumbel Copula的模型能够最有效地降低VaR,且降幅为11%以上,这让银行有效地降低了计提的监管资本,增加了资金的流动性.
[Abstract]:Basel New Capital Accord clearly proposes that the operational risk of commercial banks should be brought into the field of risk quantification and supervision.After credit risk and market risk, operational risk has become one of the main risks faced by commercial banks.In this paper, the POT model of extreme value theory is used to improve the loss distribution method (LDA).The loss distribution of the two main operational risks (internal fraud and external fraud) and the VaR value used to measure the risk of the two types of commercial banks are also calculated.In order to compare and analyze, three kinds of Copula functions are used to construct the correlation structure between internal and external fraud losses.On this basis, a Copula model is constructed to calculate the VaR value of internal and external fraud joint loss distribution.Finally, in the empirical research, this paper compares the three Copula models with the traditional measurement model, and finds that the model based on Gumbel Copula can reduce the VaRs most effectively, and the decrease is more than 11%, which makes the banks effectively reduce the regulatory capital.Increased liquidity.
【学位授予单位】：江西师范大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：F224;F830.33

【参考文献】