运用条件随机变量方法处理信用资产组合分布

Conditional Random Variable Approaches for Distribution of Credit Portfolio Value

信用资产组合模型的分布计算是信用计量模型中一个非常重要的主题,对于这个问题一般常用的方法是进行蒙特卡罗随机模拟.本文的基本模型是将信用组合的价值分解为市场共同因素和个体因素两个部分,然后引入三种全新的基于条件随机变量的方法来处理信用资产组合分布问题,并通过实例对三种方法以及标准随机模拟方法加以比较分析.

The calculation of the distribution of credit portfolio value is very important in the CreditMetrics model. Traditionally Monte Carlo simulation is used to calculate the full distribution of portfolio value. The model discussed decomposes the credit portfolio value into two parts：market factor and idiosyncratic factor, then three new conditional approaches are introduced to handle the problem. Finally the three methods are compared with the standard Monte Carlo simulation in concrete cases.