Summary Catastrophe risk sits at the intersection of exposure management, capital and reinsurance strategy. It is one of the areas where model governance most directly affects underwriting and capital decisions.
Vendor models Multi-vendor model use is now standard. Blending decisions, view-of-risk adjustments and vendor-model versioning should be governed and documented.
Non-modelled perils Non-modelled perils — some secondary perils, some geographies, some emerging exposures — are a recurring source of adverse experience. Explicit loads and portfolio limits are the correct response.
Data quality Exposure data quality (location, occupancy, construction) drives model output more than model choice. Data quality metrics and remediation programmes are essential.
Accumulation control Real-time accumulation control at bind is the operational core. Aggregation across portfolio and inuring reinsurance treats a single event coherently.
Reinsurance Cat reinsurance structure — per-risk, per-event, aggregate — should be selected against a defined appetite and stress-tested against realistic disaster scenarios.
Governance Committee oversight covers view of risk, PMLs, exceedance probabilities, cat budget and remediation of data-quality issues.
Limitations Cat models are structural approximations. Recent event experience should update the view of risk rather than validate it.
Related expertise See [Insurance Risk & Solvency](/expertise/insurance-risk).
Frequently asked questions
What should risk leaders know about vendor models?
Multi-vendor model use is now standard. Blending decisions, view-of-risk adjustments and vendor-model versioning should be governed and documented.
What should risk leaders know about non-modelled perils?
Non-modelled perils — some secondary perils, some geographies, some emerging exposures — are a recurring source of adverse experience. Explicit loads and portfolio limits are the correct response.
What should risk leaders know about data quality?
Exposure data quality (location, occupancy, construction) drives model output more than model choice. Data quality metrics and remediation programmes are essential.
What should risk leaders know about accumulation control?
Real-time accumulation control at bind is the operational core. Aggregation across portfolio and inuring reinsurance treats a single event coherently.
What should risk leaders know about reinsurance?
Cat reinsurance structure — per-risk, per-event, aggregate — should be selected against a defined appetite and stress-tested against realistic disaster scenarios.