When Insurance Stops Covering Risk, Who’s Holding It?
In February 2025, Federal Reserve Chair Jerome Powell told the Senate Banking Committee that in 10 or 15 years, there would be regions of the United States where you simply cannot get a mortgage. Banks and insurers, he said, are already pulling out of coastal areas and fire-prone regions.
What made that statement remarkable was not the prediction. People in insurance and real estate already know the market is contracting. What made it remarkable was the person saying it. The head of the Federal Reserve was acknowledging that climate risk is migrating from one part of the financial system to another, and that the receiving end may not be ready.
That migration is the subject of this piece. Not the insurance crisis itself, which has been widely covered, but the less apparent question of what happens to the institutions on the other side when the wall starts to move.
Two Industries: One Risk, Two Clocks
Insurance and banking are both in the business of pricing risk, but they operate on fundamentally different timescales.
An insurer can reprice or withdraw coverage every year. A property insurable in January can be non-renewed by December. A reinsurer can reprice an entire portfolio at the next renewal. The insurance industry’s clock runs fast because its exposure is short-dated by design.
A bank holding a 30-year mortgage does not have that option. The loan was underwritten based on conditions at origination, including the assumption that the property would remain insurable for the life of the loan. If that assumption breaks five or ten years in, the bank holds an asset whose collateral has fundamentally changed character. The property hasn’t moved. The hazard hasn’t changed overnight. But the financial architecture that made the loan viable has quietly shifted underneath it.
This is not a hypothetical:
Within the span of three years, Florida lost over 30 insurance providers as these providers sought to limit their risk by nonrenewing policies (Bankrate, 2023).
Turns out, since 2018 over 1.9 million home insurance policies have been non-renewed across the United States. (WWF, January 2026)
And theshare of completely uninsured U.S. homes has more than doubled, from 5% in 2019 to 12% in 2025. (WWF, January 2026)
California’s state-funded FAIR Plan saw its exposure more than double in under two years, reaching $650 billion due to the LA wildfires event and private insurers previously withdrawing coverage. (Resources for the Future, December 2025)
American homeowners saw insurance premiums increase by an average of 24% between 2021 and 2024, rising at twice the rate of inflation, and amounting to a $21 billion total increase nationally. (Consumer Federation of America, April 2025)
Rising insurance costs, alongside higher home prices and elevated mortgage rates, are pushing borrower debt-to-income ratios to new highs, meaning homebuyers must take on significantly more debt relative to their income to purchase the same home. (US Department of Housing and Urban Development, 2025)
The banks holding mortgages secured by these properties are watching from across a wall that is, in regulatory and operational terms, largely opaque. Insurance and banking have historically been regulated in silos. In Europe, Solvency II governs insurers while CRR/CRD governs banks. In the United States, insurance is regulated state by state while banking operates under federal oversight. The granular hazard and claims data that insurers use to make withdrawal decisions is commercially sensitive and rarely shared with lenders.
The result is a time-horizon mismatch that is becoming one of the more consequential structural vulnerabilities in the financial system. Once again: An insurer can reprice every year. A 30-year mortgage cannot.
Where It Shows Up in Hard Numbers
The pricing evidence is materializing.
In China, research published in 2025 found that a one-standard-deviation increase in climate risk exposure raises loan spreads by 40 basis points on average. Notably, banks in that study responded to transition risks but not yet to physical risks, suggesting that physical climate risk remains under-priced in significant parts of the global banking system.
A confidential Citigroup analysis, prepared for the Federal Reserve’s first climate scenario exercise and reported by Reuters, estimated the bank could face $10.3 billion in loan losses over a decade under an accelerated climate transition scenario. On the physical risk side, Citi found that a severe hurricane in the Southeast could trigger $142 million in losses on a $15 billion portfolio in one year assuming no insurance, rising to $571 million when chronic flooding was factored in. The qualifier “assuming no insurance” is the important part. As coverage retreats, that qualifier starts to describe reality.
The BIS published a working paper in July 2025 specifically on incorporating physical climate risks into bank credit risk models. Its conclusion is that physical risks have shifted from a niche concern of reinsurers into a systemic risk factor affecting banks primarily through their loan and trading books. The paper noted that a major obstacle for banks is the absence of generally accepted industry models for quantifying these exposures.
The Protection Gap Is a Credit Gap
In Europe, roughly 75% of climate catastrophe losses are uninsured. Between 1981 and 2023, natural catastrophes caused approximately €900 billion in losses across the EU, with more than a fifth of that total concentrated in the past three years alone.
When insurance doesn’t cover the loss, somebody still pays. That somebody is increasingly the borrower (whose operating costs rise and creditworthiness declines), the bank (whose collateral is impaired and whose default probabilities increase), and the sovereign (which absorbs disaster relief and faces fiscal strain). The ECB’s Financial Stability Review has described the resulting dynamic as self-reinforcing. That insurance retreat leads to asset repricing, fiscal strain, sovereign risk, and higher bank funding costs.
The ECB and EIOPA recognized this in their December 2024 joint paper, proposing a two-pillar EU-level scheme: a public-private reinsurance mechanism to pool catastrophe risk, and a public disaster fund for infrastructure reconstruction. The proposal exists because the current architecture has no mechanism for managing the risk that falls between the two sectors.
In the United States, the response has been more fragmented but is accelerating. At least 18 states introduced insurance reform legislation in 2026, many building on Colorado’s 2025 model, which requires insurers to disclose their risk models, identify risk-reduction actions for consumers, and factor mitigation into pricing. These are important steps, but they address the insurance side. The banking side, where long-dated loan books sit exposed to the consequences of insurance repricing, has received comparatively little attention.
Regulators Are Starting to Connect the Dots
Three regulatory developments in the past year indicate that the institutional response is beginning to catch up.
First, the UK’s PRA published Supervisory Statement 5/25 in December 2025, replacing its 2019 framework with climate risk expectations for banks and insurers under a single supervisory framework for the first time. The fact that this was a novelty tells you how siloed the approach has been. The PRA was measured but blunt: progress was uneven, meaning the majority of banks and insurers were not doing enough. Firms had until June 2026 to complete an internal review and produce a board-approved action plan.
Second, the ECB’s November 2025 system-wide stress test (MaSTER) broke ground by quantifying contagion channels between banks, insurers, and investment funds under stress. Previous stress tests assessed each sector in isolation. MaSTER modeled what happens when shocks propagate across sectors through forced asset sales, cascading price effects, and cross-holdings. The finding: sector-by-sector stress testing understates risk because it misses the amplification that occurs when institutions are interconnected.
Third, the NGFS published its first short-term climate scenarios in May 2025, the first publicly available framework for near-term (three-to-five-year) climate financial risk assessment. Under a severe physical risk scenario, European GDP losses could reach 4.8%. Previous NGFS scenarios looked out to 2050 or beyond. Short-term scenarios close a critical gap for banks that need to make credit risk assessments on a three-to-five-year horizon.
Together, these three developments mark a shift from treating climate risk as a disclosure exercise to treating it as a prudential one. The question is whether the pace of institutional response matches the pace of the underlying risk.
A Deeper Pattern
For sustainability professionals who don’t work in financial services, the insurance-banking story might seem sector-specific. But it isn’t.
The underlying dynamic shows up across industries. It turns out that the institutions we rely on to absorb risk are starting to hand it back, and the institutions receiving it don’t yet know they’re holding it.
This happens in supply chains when a supplier quickly exits a commitment, and the risk migrates upstream without anyone flagging it. It happens in regulatory environments when one jurisdiction acts, creating spillover obligations for another that hasn’t prepared. It happens in corporate risk management broadly, where the assumption that “someone else is covering that” turns out to be wrong by the time the exposure is discovered.
Insurance and banking happen to be the clearest current example because the numbers are enormous, the evidence is central-bank-sourced, and the consequences are materializing every day. People cannot get the mortgages they used to. Property values are falling in exposed areas. State budgets are absorbing disaster relief that would have been covered by private insurance a decade ago.
But the operating lesson generalizes. Risk does not disappear when an institution stops managing it. It just migrates. And the organizations that get hurt are not the ones that saw the risk and repriced. They are the ones that assumed the old arrangement was still in place.
For Sustainability Leaders
The insurance-banking disconnect is what happens when two systems measure the same underlying risk on different timescales, with different tools, and no shared language.
Sustainability professionals are, almost by definition, the people trained to see across these boundaries. Systems thinking, cross-functional risk assessment, the ability to translate between scientific evidence and business decision-making. This is precisely the skill needed when risk migrates between institutions that don’t share data, timelines, or until recently, regulators.
That’s the skill set that matters here. For financial institutions, and really any organization whose risk assumptions were built for a world that is already changing underneath them.
Ready for the Future, Ready Now
Tackling the challenges of today and creating the opportunities for tomorrow requires an end-to-end approach that turns challenge into opportunity. Partner with us to see how sustainable innovation augmented by AI can make meaningful business impact.
