基于Pair-copula方法的高维相关结构模型的构建

文章在Bedford、Cooke、Joe的工作基础上,利用条件独立的简单构造模块(pair-copula)对多元数据进行建模,模型方案是将多元密度函数分解成一系列pair-copula,为copula理论在高维下的应用提供理论基础,并给出在该模型下进行参数估计的算法.     

基于pair-copula的全国社保基金委托投资组合La_VaR测度研究

社保基金是社会保障事业健康发展的物质基础,安全性和流动性是其投资的首要原则。全国社保基金作为一类特殊的、可以进入资本市场投资的社保基金,其风险管理显得尤为重要。针对全国社保基金投资组合风险测度研究不足的现状,提出了全国社保基金投资组合经流动性调整的市场风险（La-VaR）测度的pair-copula-GARCH-EVT模型。与传统的copula模型相比,pair-copula方法不仅考虑了维数的影响,而且还能灵活地选择 copula 的类型。实证研究表明,pair-copula 对社保基金经流动性调整的市场风险建模的效果优于传统的多维copula模型。     

基于Pair-Copula的社保基金投资组合风险测度研究

社保基金是社会保障事业健康发展的基石,风险管理是社保基金保值增值的关键问题之一。提出pair—copula—GARCH—EVT模型以测度社保基金投资组合风险,与传统的n维。copula—GARCH—EVT模型相比,该模型不仅考虑了维数的影响,而且还能灵活地选择copula的类型。实证研究发现,基于pair—cOpula—GARCH--EVT模型测度社保基金投资组合风险的准确性要高于传统的copuIa--GARCH--EVT模型。     

A method for constructing asymmetric pair-copula and its application

Pair-copula has become a hot spot in the research of both theory and application of statistics. Therefore, proper construction of pair-copula is an area worthy of study. Considering the asymmetry of variate dependence in practical applications and the disadvantages of the widely used asymmetric copulas, we propose a method to construct asymmetric pair-copula. In our method, we treat the asymmetric bivariate copula constructed by Liebscher's method as a generator, using this generator to construct asymmetric pair-copula. Also, on the basis of our method, we propose and prove a reference for selecting copula family in the construction. To verify the method, we construct asymmetric copulas and asymmetric pair-copulas using the daily runoff data collected at Yichang hydrological station to describe the extreme drought events of Yangtze River. After comparing the models in some aspects, we accept a model we construct, and the result displays the feasibility and practicality of the method we propose.