The aim of the validation tools is to monitor the performance of models in the following areas:

Discrimination tests

The analysis of discriminatory power is aimed at ensuring that the ranking of customers that results from the rating process appropriately separates riskier and less risky customers.

Calibration

The analysis of predictive ability (or calibration) is aimed at ensuring that the PD parameter adequately predicts the occurrence of defaults – i.e. that PD estimates constitute reliable forecasts of default rates.

Stability

The analyses provide insight with regard to the stability of rating model outputs over the observation period. The stability of risk estimates is assessed using customer migrations, stability of the migration matrix and concentration in rating grades.