基于行业特性的多元系统风险因子CreditRisk+模型

本文提出了基于行业特性的多元系统风险因子CreditRisk+模型。假设行业风险因子之间相互独立是原CreditRisk+模型存在的不足,随后试图对其进行修正的单因子模型、复合Gamma CreditRisk+模型和两阶段CreditRisk+模型仍存在问题。本文在引入多元系统风险因子的基础上,将行业风险因子的形参数表示为系统风险因子的线性组合与一反映该行业风险因子内在特性的参数之积,对原CreditRisk+模型行业风险因子相关性进行了拓展,使得拓展后的基于行业特性的多元系统风险因子CreditRisk+模型解决了两阶段CreditRisk+模型忽视行业风险因子自身特性这一缺陷,将系统和行业两重风险因子有机地结合起来;新模型能够将一般情形的行业风险因子协方差矩阵纳入该模型框架内,从而克服了复合Gamma CreditRisk+模型要求行业风险因子之间的协方差必须相等的缺陷。本文证明了原CreditRisk+模型、复合Gamma CreditRisk+模型和两阶段CreditRisk+模型都只是新模型的极端情形,这些情形难以将行业风险因子协方差矩阵很好地纳入模型框架内,从而影响贷款组合非预期损失计算的精度。

Multi-Systematic Risk Factors CreditRisk+ Model Based on Sector Character

We proposed the multi-systematic risk factors CreditRisk+model based on sector character.The assumption of independence between different sector risk factors is a drawback of the original Credit-Risk+model.The amended model such as the single factor model,the compound gamma CreditRisk+model and the two stage CreditRisk+model still have their own problems. In this paper,we introduced multi-systematic risk factors,and denoted the shape parameter of sector risk factor. by a product of linear combination of systematic risk factors and a parameter that reflects and inner character of sector risk factor,that is a expansion to the CreditRisk+model. The developed model overcame the problem that inner character of sector risk factor is neglected in the two stage CreditRisk+mo del,and derived compatible combination of two-fold systematic and sector risk factors. The new CreditRisk+model can adopt general covariance matrix of sector risk factors into the framework of the model,thus overcame the drawback that covariances between different sector risk factors are required to be equal in the compound gamma CreditRisk+model.We proved in this paper that all those original Credit Risk+model,compound gamma CreditRisk+model and two stage CreditRisk+model are just extreme situations of the new model,and in those extreme situations the covariance matrix of sector risk factors can not be adopted into those models properly,thus will influence the accuracy for calculating unexpected loss.