Definition:Model change policy
📐 Model change policy is a governance document required of insurers that use internal models for calculating their solvency capital requirements, setting out the criteria, processes, and approval thresholds for modifying those models after they have received supervisory approval. Under Solvency II, any insurer operating a full or partial internal model must maintain a written policy that classifies potential changes — distinguishing between minor adjustments (such as routine data updates or parameter recalibrations) and major changes that could materially alter the model's risk profile or capital output. Comparable expectations exist in other risk-based regulatory regimes, including the Swiss Solvency Test and elements of the ORSA framework globally, though the specificity of the policy requirement is most codified in the Solvency II delegated regulations.
⚙️ A well-constructed policy defines a taxonomy of changes — often categorized as major, minor, or administrative — with explicit quantitative and qualitative thresholds for each tier. Major changes typically require prior approval from the supervisory authority before implementation, while minor changes may proceed with notification and subsequent review. The policy assigns clear ownership: risk management, actuarial, and model validation teams each play defined roles in proposing, assessing, and signing off on changes. An internal change log records every modification, supporting audit trail requirements and enabling regulators to trace the evolution of the model over time. Robust validation procedures — including back-testing, sensitivity analysis, and independent review — are triggered according to the severity classification, ensuring that the model's outputs remain reliable and that the insurer's solvency ratio accurately reflects its risk exposure.
💡 Without a disciplined approach to model changes, an insurer risks either operating with an outdated model that fails to capture emerging risks or introducing modifications that inadvertently compromise the model's integrity and regulatory standing. Supervisors scrutinize model change policies closely during supervisory review processes and on-site inspections, and weaknesses in this area have led to remediation requirements and, in some cases, restrictions on internal model use. For insurers, maintaining a clear, well-documented policy also has practical benefits: it streamlines the approval process with regulators, reduces the burden of ad hoc negotiations over individual changes, and provides the board with confidence that changes to a complex quantitative engine follow a controlled, transparent pathway. In an industry increasingly reliant on sophisticated catastrophe models, machine learning techniques, and stochastic simulations, the governance infrastructure around model evolution is as important as the models themselves.
Related concepts: