Begin with a decision
Climate scenario analysis can become a large data exercise with little influence on underwriting or capital. A better starting point is the decision: portfolio steering, pricing, reinsurance, accumulation control, investment strategy or operational resilience.
The scenario scope should then follow the material exposure and decision horizon. Short-horizon physical risk may require event and claims analysis, while longer strategic questions may require transition paths and changes in insurability.
Separate the analytical components
Physical risk analysis combines hazard, exposure and vulnerability. These components should remain conceptually distinct so that users understand whether a result is driven by the climate signal, the location and value of assets, or assumptions about damage and protection.
This separation also improves model governance. Data gaps and uncertainty can be attributed to the relevant layer, enabling targeted improvement rather than a generic confidence score.
Design proportionate scenarios
Scenarios should cover plausible adverse conditions and reveal non-linearities, concentrations and management constraints. They need not imply that a precise climate path will occur. The institution should explain time horizon, baseline, peril coverage, adaptation assumptions and exclusions.
EIOPA's work on open-source climate tools illustrates the value of transparent exposure, hazard and aggregation choices. Institutions can use such tools as benchmarks or starting points while retaining responsibility for suitability and limitations.
Connect results to the ORSA
The ORSA should show how climate findings affect the risk profile, strategy, solvency assessment and management actions. If the result never changes an underwriting rule, reinsurance decision, investment limit, data plan or strategic discussion, its decision value should be questioned.
Metrics should be monitored over time, with triggers for deeper review. That creates a practical link between periodic scenario analysis and continuous risk management.