Summary Climate scenario analysis is the discipline of testing balance sheets against defined physical and transition pathways. Its value depends on whether the outputs are actionable.
Scenario selection Reference scenarios (NGFS, IPCC) are a starting point. Firm-specific scenarios that stress the exposures the firm actually holds are more informative than universal scenarios.
Physical and transition Physical risk analyses exposure to acute and chronic climate hazards. Transition risk analyses exposure to policy, technology and market shifts on the path to net zero. Both should be modelled.
Time horizons Long horizons — 20, 30, 50 years — challenge conventional stress-testing techniques. Horizon-appropriate methodologies and clearly stated assumptions matter.
Data Emissions data, physical hazard data and counterparty-level exposure data are typically incomplete. Data-quality documentation is part of the analysis, not an appendix.
Actionability Outputs should inform credit policies, underwriting, portfolio strategy and capital planning. Analyses that produce heat maps without decisions are incomplete.
Governance Board oversight covers scenario design, methodology, results, limitations and management actions. Climate should not sit in a silo separate from other risk types.
Limitations Climate scenarios are uncertain by construction. They are used to test decisions under uncertainty, not to predict.
Related expertise See [Insurance Risk & Solvency](/expertise/insurance-risk) and [Enterprise Risk Management](/expertise/enterprise-risk).
Frequently asked questions
What should risk leaders know about scenario selection?
Reference scenarios (NGFS, IPCC) are a starting point. Firm-specific scenarios that stress the exposures the firm actually holds are more informative than universal scenarios.
What should risk leaders know about physical and transition?
Physical risk analyses exposure to acute and chronic climate hazards. Transition risk analyses exposure to policy, technology and market shifts on the path to net zero. Both should be modelled.
What should risk leaders know about time horizons?
Long horizons — 20, 30, 50 years — challenge conventional stress-testing techniques. Horizon-appropriate methodologies and clearly stated assumptions matter.
What should risk leaders know about data?
Emissions data, physical hazard data and counterparty-level exposure data are typically incomplete. Data-quality documentation is part of the analysis, not an appendix.
What should risk leaders know about actionability?
Outputs should inform credit policies, underwriting, portfolio strategy and capital planning. Analyses that produce heat maps without decisions are incomplete.