ÀÖÓ㣨Leyu£©ÌåÓý¹ÙÍø

The model risk management landscape has evolved significantly, driven by regulatory changes and the increasing complexity of quantitative methods. In the UK, the PRA’s Supervisory Statement SS1/23 has brought Deterministic Quantitative Methods (DQMs) into sharper focus.

DQMs, which include decision-based rules and algorithms, play a crucial role in business decisions and risk management. Advances in technology and data processing power have enabled them to become more complex, requiring robust governance and validation frameworks.

The DQM framework continues to develop, with firms making significant strides in identifying and validating DQMs. The integration of advanced tooling can support these efforts and accelerate the validation process. The adoption of AI enhancements and the extension of tooling capabilities to End User Computing (EUCs) will further empower firms to manage their DQM and EUC populations with confidence and precision.

Appropriate focus by firms on DQMs can help to ensure that they are accurate, reliable and compliant with regulatory standards, ultimately supporting better decision-making and risk management.

As noted above, firms have made substantial progress in identifying DQMs, and their ability to recognise and document DQMs within their systems has matured. However, there is more to do as they refine their approaches to help ensure comprehensive identification and governance of these methods.

Key challenges for firms include:

• Diagnosing the complexity of DQMs â€� firms are adopting various approaches to define and assess the complexity of DQMs, ranging from broad definitions with tiered governance parameters to narrow definitions focusing on the most material and complex DQMs. Ongoing efforts aim to strike a balance between comprehensive governance and practical implementation. Firms are also leveraging existing processes to support DQM discovery and ensure completeness of DQM inventory.
• Proportionate validation â€� firms need to ensure that DQMs are not only identified but also validated to meet regulatory and internal standards, applying documentation and testing requirements proportionate to the DQM’s materiality and complexity. Current hurdles include identifying the correct skill level and team to carry out validation testing, managing the additional time burden on existing teams, and drafting DQM documentation.

Advanced tooling can be a valuable asset in supporting firms' DQM tiering efforts. Such tools can analyse individual file characteristics, apply tailored risk factors and provide a comprehensive complexity assessment. 

The integration of AI can accelerate the validation process further. AI-driven features, such as automated decision matrices and code similarity analysis, can enhance tools� ability to identify and validate DQMs with greater precision and efficiency. 

ÀÖÓ㣨Leyu£©ÌåÓý¹ÙÍø professionals' extensive experience in model risk management and validation, coupled with expertise in DQM and EUC development, enables us to support firms in their DQM journey.

We can provide in-depth test and review services, including qualitative reviews of DQM documentation, assessment of overall DQM methodology, and verification of DQM calculation logic.

For more information on any of the above, please reach out.