尾部相关系数的渐进变化特征及其应用  

Asymptotic property of tail dependence coefficient for Copula function and its application

作　　者：秦学志[1] 王玥[2] 

机构地区：[1]大连理工大学管理学院,大连116024 [2]沈阳大学工商管理学院,沈阳110044

出　　处：《系统工程理论与实践》2011年第2期193-204,共12页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(70771018);高等学校博士学科点专项科研基金(20090041110009);大连理工大学人文社会科学研究基金(DUTHS2008203)

摘　　要：在Copula函数的尾部相关性研究的基础上,针对其不足进行了两个方面的推广:1)一个变量趋于某个非尾部值与另一变量趋于尾部的相关系数;2)两个变量都趋于非尾部值之间的相关系数.传统的尾部相关系数可用于考察证券价值或企业价值之间在尾部的正相关关系或负相关关系,上述推广可进一步用于考察一个企业处于违约边界或某个状态对其他企业违约可能性的影响,及两个证券或企业的价值均趋于某些特定值之间的相关关系.进一步给出了上述推广的相关系数在证券或企业价值处于某些区间条件下的期望值及其计算方法.这些工作从理论和实证两个方面丰富与刻画了相关系数的渐进变化特征.Based on the studies about the tail dependence with Copula functions,this paper provides two kinds of generalizations as follows：1） the dependence between one variable which approaches to some non-tail value and the other variable which approaches to upper tail or lower tail;2） the dependence between two variables which are not upper tail or lower tail.The classical tail dependence mainly expresses the positive or negative correlation in tails,and these generalized in this paper can be used to express the influences of one security or firm on the default possibility of others when it defaulted,and the correlation between two firms and securities when their values are some non tail values.Furthermore,this paper provides the formulae and calculation methods for the expectations of the generalized dependence in the case that the values of two variables are in some intervals.All the above enriches and characterizes the asymptotic properties of tail dependence in theory and in practice.

关 键 词：COPULA函数 尾部相关系数 误差分析 联合分布 

分 类 号：F830.9[经济管理—金融学] F224.7