The issue

Climate risk models can shape lending, insurance pricing and capital allocation. But their outputs depend on uncertain scenarios, incomplete data and assumptions linking climate hazards or transition pressures to financial losses. Validation must establish whether a model is fit for its intended use.

Supervisory expectations differ across jurisdictions:

  • UK: PRA SS5/25 expects banks and insurers to challenge and periodically review climate scenario tools. SS1/23 adds independent validation expectations for banks within its scope.12
  • Canada: OSFI B-15 expects institutions to understand climate models, including third-party assumptions and limitations. Revised E-23, effective 1 May 2027, extends enterprise-wide model-risk expectations across federally regulated financial institutions, including insurers.34
  • EU: EBA environmental scenario-analysis guidelines apply from 1 January 2027, strengthening expectations for banks’ assessment of financial and business-model resilience.5
  • US: SR 26-2 replaced SR 11-7 in April 2026. It provides risk-based model-management guidance, principally relevant to banks above US$30 billion in assets; it is not a climate-specific mandate.67

Why businesses should care

For banks and insurers, weak models can misprice risk, distort loss estimates and misdirect capital. A technically sophisticated model can still mislead if used beyond its design limits.7

The practical implication extends to corporate borrowers and policyholders: modelled exposures can influence credit terms, insurance availability and investment decisions. Incomplete asset data or unrecognised resilience measures may leave a company assessed through broad assumptions that poorly reflect its actual risk. These supervisory expectations generally reach companies through financial counterparties, rather than imposing equivalent obligations directly.

What businesses should do

Financial institutions should bring climate models into existing model-risk governance, with controls proportionate to materiality and use:

  • Assign owners and record models, vendor tools, approved uses and limitations.
  • Trace data and assumptions through to financial outputs, including exposure, vulnerability and loss estimates.
  • Apply independent challenge, sensitivity testing and benchmarking.
  • Where historical back-testing is limited, test conceptual soundness and explain remaining uncertainty.
  • Document approvals, use restrictions, material changes and remediation; revisit validation when data, methodology or use changes.147
  • Companies should prepare verifiable asset, emissions and resilience-investment data, and engage lenders and insurers on how those inputs affect assessments.

The objective is defensible decisions: understanding where model outputs support financing and capital allocation, and where further evidence or judgement is needed.

Sources

  1. Prudential Regulation Authority, SS5/25 — Enhancing banks’ and insurers’ approaches to managing climate-related risks, December 2025, particularly paragraphs 4.67–4.79.
  2. Prudential Regulation Authority, SS1/23 — Model risk management principles for banks, current version effective 23 April 2026.
  3. Office of the Superintendent of Financial Institutions, Guideline B-15 — Climate Risk Management, particularly Chapter 1, paragraphs 7–12.
  4. Office of the Superintendent of Financial Institutions, Guideline E-23 — Model Risk Management (2027), effective 1 May 2027.
  5. European Banking Authority, The EBA publishes its final Guidelines on environmental scenario analysis, 5 November 2025; guidelines apply from 1 January 2027.
  6. Federal Reserve, SR 26-2 — Revised Guidance on Model Risk Management, 17 April 2026.
  7. Federal Reserve, OCC and FDIC, Supervisory Guidance on Model Risk Management, April 2026.