Credit Rating Models for Central Banks
Najla Al Barrak, Hessa Al Sanousi, Irene Moulitsas, Salvatore Filippone

Abstract
Credit default is the hottest topic in the banking sector. Central banks are in pressure to conduct studies in order to have a suitable and robust regulations. Recent research on Kuwaiti banks provide evidence that internal rating models are more efficient than the standard approach for predicting the rate of credit default. Having a comprehensive model for the banking system, for Kuwait as an example, will help central banks to test and evaluate the internal models available for the banks. Having two type of banking, conventional and Islamic, will make the system accommodatable in all jurisdictions. Through machine learning solutions and long skewed data, our research is comparing different Ensemble models to reach the appropriate model structure. We have covered two type of loans, instalment and consumer loans to overcome the difference of the use of such loans.

Full Text: PDF