In our previous publication, we explored the inclusion of Machine Learning (ML) in the updated ECB Guide to Internal Models and the regulator’s expectations for its use within Pillar 1 models. Building on that, this article shifts focus to credit risk, where the revised Guide introduces more granular expectations for institutions using IRB models, reflecting a firmer supervisory stance, the integration of EBA Handbook elements, and alignment with CRR3 requirements.

This publication breaks down the major changes in credit risk requirements - from stricter PD and LGD estimation rules to clarified definitions of default and enhanced internal validation standards. These revisions signal a move toward more conservative, harmonised, and auditable IRB practices across the euro area.

As Banks prepare for model reviews and permission applications, understanding the revised Guide is essential to ensure compliance and avoid supervisory roadblocks. Our summary offers a practical entry point for risk teams, model owners, and senior management navigating this evolving regulatory landscape.

Access the full publication: ECB’s Updated Guide to Internal Models: Key Credit Risk Changes [ 586 kb ]

 

Authors:

Andreas Spyrides, Quantitative Risk Services Leader

Kyveli Kyriacou, Consultant, Quantitative Risk