PRA SS5/25:
climate-risk supervisory expectations

A practitioner guide for UK insurers — what the PRA expects to see in 2026, the four pillars, and where insurers most often fall short of capability-over-compliance.

What SS5/25 actually requires

Supervisory Statement 3/19 — published April 2019, refreshed through subsequent Dear-CEO letters and the climate biennial exploratory exercise commentary — sets the PRA's expectations for managing the financial risks from climate change in the banking and insurance sectors. It rests on four pillars:

1. Governance — clear board ownership, named senior-management responsibility, climate built into the firm's risk-management framework.
2. Risk management — climate risk identified, measured, monitored, and managed across the firm's existing risk taxonomy (insurance, market, credit, operational).
3. Scenario analysis — forward-looking quantitative scenarios used in business planning and capital adequacy assessment, with both physical and transition framings.
4. Disclosure — public disclosure consistent with TCFD recommendations (now ISSB IFRS S1/S2 in the UK adoption pathway).

From "we comply" to "we can demonstrate"

The first cycle of SS5/25 evidence was largely narrative. The current cycle is not. Supervisors expect to see the modelling itself — the data, the calibration, the stress, the back-testing, the governance and the use in decisions. The ORSA section on climate-risk integration should now contain a worked example of a specific climate stress traced through the capital model, not just a statement that climate has been considered.

The shift is from compliance (we have a policy) to capability (we can run the question in real time, including questions the supervisor invents on the day).

Where insurers most often fall short

• Generic scenarios. Off-the-shelf NGFS scenarios reproduced verbatim. Firm-specific scenarios that combine the firm's actual exposures are increasingly the standard.
• Bolt-on climate. Climate risk lives in a separate workstream that doesn't connect to capital, underwriting, reinsurance or operational decisions.
• Missing the regulatory channel. Climate risk is treated as an insurance loss. The supervisory consequences (additional disclosure burden, capital add-ons, prudential restrictions) aren't modelled at all.
• Cat-model opacity. Vendor cat models are run as black boxes without SS1/23-style validation, with no explanation of how climate-conditioned versions differ from base versions.
• Disclosure–prudential mismatch. The TCFD/ISSB disclosure asserts climate is material; the ORSA quietly treats it as immaterial. The supervisor reads both.

The Abgalis approach

The Abgalis Engine treats climate as one of seven domains in a cross-domain transmission framework. SS5/25 evidence isn't a separate workstream — it's a view onto the same channel architecture used for capital, underwriting and disclosure. That makes the worked example the supervisor wants to see a configuration choice, not a build-from-scratch exercise.

Specifically: a firm-specific climate scenario, traced with explicit channel latency through the market, credit, liquidity, insurance, operational, and strategic & emerging domains, with management actions tested in time-indexed steps. Output is consumed by the SCR, the reverse-stress test, the operational-resilience evidence under DORA Article 25, and the disclosure narrative — all from the same model.