回收率随机的信用组合损失的极限分布  

The Limit Distribution of Credit Portfolios' Loss When Recovery Rate Is Random

作　　者：王国栋[1] 詹原瑞[1] 

机构地区：[1]天津大学管理学院,天津300072

出　　处：《系统工程》2009年第7期28-33,共6页Systems Engineering

基　　金：国家自然科学基金资助项目(70573076);高等学校博士点基金资助项目(20050056057)

摘　　要：估计组合损失常用且有效的方法是蒙特卡洛模拟,但是这种方法需要耗费大量时间。文章假设回收率是随机变量,且与违约率是相关的,得到了组合损失的极限分布函数,拓展了V asicek关于组合损失极限分布的模型。根据模型还求得了组合损失的期望、方差、受险价值和预期短缺。对比发现将回收率看做常数而忽略其波动性会低估组合损失的V aR。另外还与用蒙特卡洛模拟具体组合的结果进行对比,发现得到的模型可以很好地近似包含资产个数较多组合的损失分布,可方便地用来估计大型信用组合的损失。Monte Carlo simulation is a commonly-used and valid technique to estimate portfolios＇ loss, but it consumes a great deal of time. In this paper,the recovery rate is presumed to be a random variable,correlated with the default rate. The limit distribution function of the portfolios＇ loss is achieved, which extends Vasicek＇s model. The expectation, variance, VaR and ES of the portfolios＇ loss are further obtained. Comparing the results, we find out that the VaR of portfolios＇ loss will be underestimated if recovery rate is considered as a constant and its volatility is omited. At the last, all the results obtained above are compared with the results obtained by Monte Carlo simulation. We find out that the model achieved in this paper can approximates the true distribution of the portfolios＇ loss quite well, and the time it consumes is short, so it can conveniently estimate the loss of large portfolios.

关 键 词：组合损失 蒙特卡洛模拟 回收率 受险价值 

分 类 号：F830[经济管理—金融学]