登入
選單
返回
Google圖書搜尋
Default Prediction with a Multiple-Spell Discrete-Time Hazard Model
Matej Jovan
出版
SSRN
, 2019
URL
http://books.google.com.hk/books?id=a3X8zgEACAAJ&hl=&source=gbs_api
註釋
We argue that the true transition-to-default dynamic in banks' credit portfolios can only be fully described with a multiple-spell discrete-time hazard model. This paper develops such a model for default prediction. The model permits the use of all data available to the bank or to the bank regulator, which entails recurrent defaults and other recurrent events. The estimated PDs from such model are consistent and more efficient. The results show that the inclusion of historic performance improves predictive power over models lacking such inclusion. This reduces bias in the capital requirement and impairment for credit risk.