债券组合信用风险模型及实证研究  

Modeling Credit Risk of Bond Portfolio and Its Empirical Analysis

作　　者：马勇[1] 张卫国[1] 许文坤[1] 姜茜娅[1] 

机构地区：[1]华南理工大学工商管理学院,广东广州510641

出　　处：《系统工程》2012年第3期25-30,共6页Systems Engineering

基　　金：国家杰出青年科学基金资助项目(70825005);广东省高等学校珠江学者岗位计划项目(2010);广东省自然科学基金资助项目(S2011040005723)

摘　　要：隶属于不同行业的各债券发行公司,不仅受系统风险影响,而且还需承受其行业特有的风险。将债券按发行公司所属行业进行分类,然后利用分组t关联函数研究债券组合的信用风险。实证结果表明:相比t关联函数而言,利用分组t关联函数刻画债券组合信用风险相关性能得到更大的VaR和ES值。这说明当债券的异质性程度大时,分组t关联函数比t关联函数能更好地刻画债券组合的极端风险。Because the companies issuing bonds belong to different industries,which are exposed to different risks besides system risks,we classify the bonds issuers according to the industries they belong to.Then we take advantage of the grouped t-copula to model the correlation between credit risks of the bonds.Empirical analysis shows that the VaR and ES of credit risks of the bonds portfolio are larger when the correlation is characterized by grouped t-copula than by t-copula.Therefore grouped t-copula can capture the extreme risk more efficiently than t-copula when the bonds are heterogeneous.

关 键 词：信用风险 债券组合 分组t关联函数 t关联函数 

分 类 号：F830[经济管理—金融学]