Both Sides of the Balance Sheet: Climate Risk and Opportunity Across Insurance Lines

Climate change does not hit every insurer the same way, and most climate scores miss that. At Pico Analytics, we built a model that measures climate risk and opportunity across four insurance lines, on both underwriting and investments. Tested on a real Lloyd's specialty book, it shows a 9 to 13% hit to capital by 2050, driven mainly by what the insurer chooses to underwrite. Now any insurer can run it on its own figures in minutes.

One climate lens does not fit every insurance line

Property and casualty insurers feel climate change as storms, floods and wildfires. Life insurers feel it through heat-related mortality and long-dated investments. Re-insurers carry the largest catastrophe losses alongside large investment books, and specialty insurers meet it mainly through litigation and clients in shrinking industries. Generic climate scores flatten these differences, and rarely measure opportunity at all.

What we built

We started with a scoring model for four lines, property and casualty, life, specialty and reinsurance, built from the public disclosures of a major European insurer in each. We then rebuilt it to measure impact in money terms on both sides of the balance sheet: extra claims and business lost or gained on the underwriting side, and repricing of bonds and equities on the investment side.

It runs three NGFS scenarios to 2050, the framework behind the Bank of England's 2021 climate stress test, with physical hazard from IPCC AR6. Where client data has gaps, each one is filled with a sourced, confidence-graded assumption, so every result shows how much of it rests on observed data.

What we found

Scoring model, 2025 annual reports of four major European insurers

Property and casualty is the most exposed line and life the least. Refining each score with the insurer's own disclosures mattered: the re-insurer's score fell from 3.6 to 2.0 once its transition exposure was clear.

Tested on a real Lloyd's specialty book, the full model shows climate risk outweighing opportunity in every scenario, with a net hit of 9 to 13% of capital by 2050.

Underwriting data of a Lloyd's specialty book, 2021-2025 inceptions

The biggest driver is the underwriting book. Fossil-fuel clients now outweigh renewable-energy clients by around twenty to one in premium. Moving 20% of the insurer's investments out of fossil fuels improved the result by only about 0.1% of capital.

Try it on your own book

We turned the model into an interactive simulator. An insurer selects its line, enters premium by line and industry and investments by sector, and picks a scenario. In seconds it sees the impact on underwriting and investments, its net impact as a share of capital, and a 1 to 5 risk score for each line and sector.

Insurers can try the simulator on the Pico Analytics website, and we welcome conversations about running the full model on your own data.

What this means for insurers

1. Know which channel drives your line. Property and casualty insurers should start with physical catastrophe, but re-insurers and life insurers may find transition exposure larger than they assume, and specialty insurers should look first at their underwriting book. In the specialty book we tested, tracking the share of premium written for fossil-fuel and renewable-energy clients was a simple, early indicator of where the book is heading.

2. Measure opportunity alongside risk, and report both. Growth in renewable energy and other transition sectors is real, but in this book it offsets only a small part of the risk. Reporting gross risk and gross opportunity separately, before the net, keeps a large exposure from hiding behind a small gain.

3. Imperfect data is no reason to wait. No insurer has complete climate data. Filling gaps with sourced, graded and tested assumptions produces a result that is honest about its uncertainty, and it shows which data would improve the answer most.

Limitations and next steps

The results are only as strong as their inputs. The investment side uses a proxy portfolio re-scaled to the size and asset mix of a typical Lloyd's insurer, and sectors for some holdings are classified from security names. Derivatives are excluded, and the four-line scores rely on public disclosures, which vary in detail between insurers. The two assumptions that move the result most are how far a rise in extreme weather feeds through to liability claims, and how fast climate litigation grows; both are reported as ranges. Physical risk differs little between scenarios by 2050, because most of the warming gap between them opens up after mid-century.

Next, we plan to run the model on insurers' own matched underwriting and investment data, add sector and country classifications for every holding, and extend the horizon beyond 2050. We would welcome conversations with insurers who want to see where their own book stands.

Sources

Next
Next

New Workshop Announced: Nature-Related Risks for Banks: A Practical Guide for Banking Professionals