基于Copula函数的信用风险管理模型及实证分析

An Analysis on Credit Risk Management Model Based on Copula Function

传统的信用模型,均是在假设信用评级及违约机率的共同变化服从多变量常态分布的前提下,进行投资组合信用风险的评估.但在实际的应用领域中很少有真正是服从多变量常态分布的数据.尤其当投资组合标的的数量庞大时,要准确估算出联合概率分布几乎是不可能的.本文所介绍的Copula方法的综合运用,可将上述问题简化以匹配出更合乎实际的联合概率分布,从而更加准确地测量出银行可能面临的风险.

The credit model is on a presupposition that credit evaluation default odds change subject to multi-variables distribution.But there is seldom data subjecting to multi-variables distribution.Especially when portfolios are huge,any one that wants to estimate association odds distribution is impossible.The way I introduces here could deduces the questions above and give the right evaluation of the risk when banks is about to face.