基于C藤Copula模型的混业经营下聚合风险的度量  

C-Vine Copula Based Aggregated Risk Measurement Under Mixed Operation

作　　者：周全[1] 陈振龙[2] Zhou Quan;Chen Zhenlong(School of Information and Mathematics,Yangtze University,Jingzhou 434023,China;School of Statistics and Mathematics,Zhejiang Gongshang University,Hangzhou 310018,China)

机构地区：[1]长江大学信息与数学学院,湖北荆州434023 [2]浙江工商大学统计与数学学院,浙江杭州310018

出　　处：《湖南科技大学学报（自然科学版）》2018年第4期113-119,共7页Journal of Hunan University of Science And Technology：Natural Science Edition

基　　金：国家自然科学基金资助项目(11371321);浙江省高校人文社科重点研究基地资助项目(1020XJ3314004)

摘　　要：本文首次提出了对混业经营下聚合风险度量问题的实证研究,以VaR作为风险度量的工具,根据混业经营下金融机构所拥有的资产组合的特点,首先利用ARMA(1,1)-GARCH(1,1)模型提取了各单个资产的标准残差,再选取C藤Copula对各单个资产之间的相依结构进行了拟合,最后根据拟合的参数借助于蒙特卡罗模拟法进行了逆向模拟,得到了C藤Copula模型下VaR的回溯测试结果.同时,采用类似的方法,分别对R藤和D藤Copula模型下的VaR进行了回溯测试,并将3种模型下的回测结果进行了对比,得出了C藤Copula模型下得到的结果误差最小的结论.此外,文中还给出了用C藤Copula对多维数据之间的相依结构进行拟合的算法步骤以及利用蒙特卡罗模拟法求解多维资产组合VaR的步骤和方法.The empirical study about the aggregated risk measurement under mixed operation is first proposed.For VaR calculation of the portfolio that the financial body owns,the standard residual of every financial variable was extracted by using the ARMA(1,1)-GARCH(1,1) model,and then all the eight financial variables were fit by using the C-Vine Copula to describe the dependence structure,according to the characteristic of the portfolio under mixed operation.Finally,the back-tracking test results of VaR under C-Vine Copula model are obtained by means of Monte Carlo simulation based on the fitted parameters..At the same time,using similar methods,VaR under R and D Vine Copula models were retrospectively tested,and the results under the three models were compared,and the conclusion that the results under C-Vine Copula model had the least error was obtained.The algorithm steps of fitting the interdependent structure of multi-dimensional data with C-Vine Copula and the steps and methods of solving multi-dimensional portfolio VaR by Monte Carlo simulation were given.

关 键 词：混业经营 C藤Copula VAR 蒙特卡罗模拟 

分 类 号：F830.91[经济管理—金融学] N945.12[自然科学总论—系统科学]