Abstract:
EU legislators mandated the European Banking Authority to propose a stress scenario methodology for capitalising non-modellable risk factors (NMRF) as foreseen under the Basel Fundamental Review of the Trading Book (FRTB) rules for market risk. In this paper, we present the foundations of such a methodology. By design, it is universally applicable to all kinds of risk factors to which a bank may be exposed, and it caters for a wide range of data availability by adjusting the stress scenario for the number of returns observed in the calibration period. It captures non-linearities in the portfolio loss profile against changes in the NMRF, while reducing the computational effort and being simple. To motivate the values set for some parameters in the methodology, we use a set of skewed generalised 't' (SGT) distributions as a generic tool for describing a wide universe of real historical returns from all asset classes. Finally, we extend the methodology from single risk factors to segments of curves or surfaces as envisaged in the FRTB.