Complying with the UK’s New Model Risk Regulation

Yields recently hosted a webinar featuring Carl Ireland, the Head of Regulatory Risk at Jaywing, to discuss the implications of the newly introduced SS1/23 guidance on model risk management from the Bank of England. With SS1/23 now in effect, Carl, with nearly 20 years of experience in the field, shared valuable insights on how UK banks are responding to this regulation. This article explores the key elements of SS1/23, its impact on the industry, and the evolving practices in managing model risk within financial institutions.

Key Changes in SS1/23: What to Know

SS1/23 introduces significant enhancements to model risk management frameworks, particularly by expanding the definition of what constitutes a “model” within organizations. One of the most notable aspects is the inclusion of “deterministic quantitative methods,” which, though not traditionally classified as models, now require a certain level of oversight as well.

Another important shift is the heightened focus on governance. The regulation emphasizes greater involvement from senior leadership, calling for independent validation, comprehensive documentation, and robust controls to ensure model risk is managed as a distinct part of the institution’s overall risk management strategy.

Navigating Implementation Challenges

Implementing SS1/23 requires a cultural shift within organizations. Historically, model risk was managed by specialized teams, often within risk or analytics functions. However, SS1/23 promotes a more integrated approach, encouraging businesses to embed model risk management practices across all functions. For example, HR departments using models for compensation decisions will now need to adopt governance practices around model risk.

The growing complexity of model inventories, especially with the increasing use of AI and machine learning, presents additional challenges. Many institutions rely on vendor models, adding further complexity. SS1/23 requires third-party models to meet the same rigorous validation standards as in-house models, which can be challenging when proprietary data or algorithms are involved.

Using Technology to Address Model Complexity

As model inventories expand, technology becomes crucial in managing model risk efficiently. Automation tools and AI can streamline processes, ensuring models are properly documented, validated, and governed before being deployed.

Technology also helps manage third-party risk by enabling financial institutions to assess vendor models more efficiently. Automated testing and access to APIs, for instance, can help to evaluate model performance without needing full access to proprietary algorithms, ensuring compliance while protecting intellectual property and data security.

The Future of Model Risk Management Under SS1/23

Over the coming months, the Prudential Regulation Authority (PRA) will publish guidance for smaller institutions advising on the proportional application of the guideline. Simultaneously, they are reviewing how financial institutions are implementing SS1/23. These two aspects together will drive the field forward and help institutions to further refine model risk management frameworks, expand their model inventories, and ensure model risk is embedded into key decision-making processes. tions.

Conclusion

SS1/23 marks a significant evolution in how UK financial institutions approach model risk management. By broadening the definition of models and increasing governance expectations, the regulation reinforces the importance of treating model risk as a key organizational priority. While implementation may pose challenges—especially as firms adapt to the regulation’s broader scope and the complexity of modern models—the regulation also offers opportunities for organizations to enhance their risk management frameworks and leverage technology to improve oversight.

Firms that successfully implement SS1/23 will be well-equipped to manage the growing complexities of model risk and meet the evolving expectations of regulators with confidence.