基于CVaR和GARCH(1,1)的扩展KMV模型  

The Extended KMV Model Based on CVaR and GARCH(1,1)

作　　者：王秀国[1] 谢幽篁[1] 

机构地区：[1]中央财经大学应用数学学院,北京100081

出　　处：《系统工程》2012年第12期26-32,共7页Systems Engineering

基　　金：国家自然科学基金资助项目(70901079);中央财经大学科研创新团队支持计划项目

摘　　要：根据我国还没有大量的违约历史数据库,违约距离(default distance,DD)和预期违约概率(expecteddefault frequency,EDF)的映射关系还没有建立的情况下,提出了基于条件在险值(conditional value at risk,CVaR)和GARCH(1,1)的扩展KMV模型,其中CVaR值和其极端波动下的违约距离,作为新的信用风险度量指标。最后,选取了中国沪市A股的两组共14个公司进行了实证分析。结果表明用扩展的KMV模型对样本公司的信用风险评估具有良好的效果,并且很好地对市场的信用风险状况做出预警。Since the database including vast amount of default historical data is absent in China and the mapping relationship between the default distance （DD） and the expected default frequency （EDF） has not been established, we propose an extended KMV model based on CVaR and GARCH （1,1）, in which two parameters are used to measure the credit risk： CVaR and the extreme-DD which concerns extreme volatilities. Finally, empirical analysis is made for 14 companies selected from Shanghai stock market. The results show that the new model can effectively access the credit risk of different public companies and make a better prediction of the market crediting risk copared with the traditional KMV model.

关 键 词：KMV模型 极端违约距离CVaR GARCH(1 1) 

分 类 号：F224[经济管理—国民经济]