The aim of this paper is to present model risk situations and a methodology to measure and quantify the associated risk at model level, with different types of assumptions. Then, considering that in practice, a model risk management at model level is hardly feasible, this paper also outlines a method to measure and quantify model risk at risk category level (ex: Credit Risk).

In fact, one of the overarching drivers of this paper is to provide a model risk “value” which will enable you to analyse if the model risk is sufficiently covered. Indeed, although banks already allocate funds regarding this risk (portion of RWA attributed to conservative margins for credit risk, portion of Op risk Value at Risk, etc.), assessing the appropriateness of those funds remain complicated.

Huge congratulations to all the GRA team for putting so much effort in this new insightful analysis, demonstrating once again the high quality of our research department!

You can download the white paper below.