基于BMM模型的金融极端风险研究及实证分析  

Research on Financial Extreme Risk Based on BMM Model and Empirical Analysis

作　　者：李侦 吕文元[1] 

机构地区：[1]上海理工大学管理学院,上海200093

出　　处：《科技和产业》2017年第11期148-152,共5页Science Technology and Industry

摘　　要：介绍极值理论,包括广义极值分布、区组最大值模型的拟合方法、检验方法等,重点考虑极值相依机构,引入极值指标,考虑不同频率下收益率的极值分布,并计算相应的VaR。之后对股票的对数收益率进行分析,首先从偏度峰度等基本统计量可知服从左偏尖峰分布,然后拟合边缘分布参数,判断都属于Frechet分布,最后用BMM模型和正态分布进行连接,得出结论正态分布模拟得到的极端损失和极端收益远远低于BMM方法。This paper introduces the Extreme Value Theory,including the Generalized Extreme Value Distribution,the Group Maximum Model of the fitting method,test methods,focusing on extreme value dependent institutions,the introduction of extreme value indicators,consider the different frequency of the extreme rate of return,And calculate the corresponding VaR.Then,we analyze the logarithmic yield of the stock.First,we can get the distribution of the left spike from the basic statistic of the skewness kurtosis,and then fit the edge distribution parameter.The judgment belongs to the Frechet distribution,and finally the BMM model and the normal distribution The conclusion is that the extreme loss and extreme benefits of the normal distribution simulation are much lower than the BMM method.

关 键 词：区组最大值模型 极值指标 VAR 

分 类 号：F830.9[经济管理—金融学]