摘要
商业银行操作风险的计量存在两个重要的问题,一个问题是损失数据缺乏,另一个问题是各部分之间的风险相关问题.结合Bayes估计和Copula函数解决了上述两个问题并基于中国商业银行操作风险的损失数据对对操作风险的计量进行了实证分析.实证分析的结果表明无论考虑风险相关与否,基于极大似然估计的VaR与基于Bayes估计的VaR具有一定的差距.
To the business bank, two important problems are the lack of the loss data and the correlation of the operational risks. Combining the bayesian estimation method with the Copula function the paper has resolved the two problems and has demonstrated based on the data of the business bank of China. The demonstration has showed that considering the correlation of the risk or not, the VaR has some difference based on the maximum likelihood estimation and the bayesian estimation.
出处
《数学的实践与认识》
CSCD
北大核心
2014年第15期154-163,共10页
Mathematics in Practice and Theory
基金
教育部人文社科项目"非参数方法及其在商业银行操作风险计量中的应用研究"(10YJC910001)
国家人文社科基金项目"商业银行操作风险度量中的贝叶斯推断方法及其应用研究"(12CTJ020)