Model risk management: questions for decision-makers
Model inventories, ownership, validation and limitations: practical questions for financial institutions using models to support decisions.
Facultas Advisory · Originally published
Revised edition
Why model risk matters
Models help translate data and assumptions into decisions. Errors, unsuitable assumptions or use outside the intended purpose can produce misleading results. A technically sophisticated model still needs accountable owners and informed users.
Five areas to examine
The PRA’s SS1/23 organizes its expectations around model identification and classification, governance, development and use, independent validation, and risk mitigants. It is UK supervisory guidance for institutions within its stated scope—not a universal Canadian requirement. The current version should be checked when applying it.
Turn the framework into questions
Which models influence material decisions? Who owns each model and approves its use? What independent challenge has occurred? Which data gaps and assumptions matter most? What happens when performance deteriorates or a model is used in a new context?
Make limitations visible
Decision materials should explain what the model supports, what it cannot establish and when escalation is needed. Possible responses include additional analysis, restrictions on use and remediation with named owners. Human judgment should be documented and challenged too; it does not automatically correct a weak model.
Reference
PRA — SS1/23 (opens in a new tab)
This website provides general information about Facultas Advisory and its services. Its content is not professional advice for any particular circumstances; advice requires an agreed engagement scope.