The distinct risks from ‘non-models’

Regulators now expect financial institutions (FIs) to identify and manage risks not only from traditional models but also from quantitative methods that fall beyond the conventional definition of models.

These ‘non-models’ pose distinct risk management challenges.

Historically, FIs have tiered models and applied risk management frameworks commensurate with the risk. However, this approach overlooks potential risks from quantitative tools and techniques not meeting traditional model criteria.

A robust framework to manage risks from non-model methods is required, particularly with the adoption of artificial intelligence (AI) and machine learning (ML) techniques, which may not be classified as models.