Good model risk governance is essential for all financial institutions. For example, banks exposed to accounting credit impairment reporting (IFRS 9/CECL), evolving stress testing programs that affect capital planning and new regulatory requirements such as Solvency II and Basel III and IV all require comprehensive model validation. Effective model risk management is also essential to gaining shareholder confidence and complying with regulatory specifications of the European Banking Authority, the US Federal Reserve and the UK’s Prudential Regulation Authority’s (PRA) 2022, per its consultation paper on model risk management principles (CP6/22).
Governance is particularly vital for machine learning (ML) models, which can predict more accurately in some use cases but also develop unfair biases that can impact the institution or its clients. ML models need more MRM care – from frequent performance monitoring, constant data review and benchmarking, and better contextual model inventory understanding to wellthought-out, action-ready contingency plans.
- Model inventory & documentation
- Model risk reporting
- Model performance monitoring
- Address & manage regulatory reviews