Financial stability is the condition in which the financial system can take a shock without breaking the flow of credit, savings, and cover to the real economy. Central banks and supervisors hold that job. Many of the shocks that will test it begin somewhere physical: a floodplain that no longer spreads the water, a wetland drained upstream of an insured town, a forest stand that stopped holding the slope.
Supervisors can now describe that cascade in more detail than ever. They can stress-test it, set expectations for it, and ask banks and insurers to hold capital against it. That is serious work. It is also not the same thing as keeping the wetland wet.
what financial stability actually means
The European Central Bank's working definition is a good place to start. Financial stability is a condition in which the financial system — intermediaries, markets, and market infrastructures — can withstand shocks and the unravelling of imbalances without disruptions severe enough to impair the flow of savings into productive investment.
Two things follow from that definition.
First, stability is a property of the system, not of any one institution. A bank can be solvent while the system is fragile. That is why supervision splits into microprudential work (is this bank or insurer safe?) and macroprudential work (can the system as a whole absorb the hit?).
Second, stability is about absorbing shocks and leaning against the financial imbalances that feed them — not preventing the physical event. The toolkit — capital, liquidity, reserves, supervisory expectations, stress tests — exists so that when the loss arrives, the system bends instead of breaks. That design is correct for a central bank. It also marks the edge of what a central bank can do about a floodplain.
how climate and nature reach the balance sheet
Nature does not hit a balance sheet directly. It arrives through the usual channels: a borrower's revenue falls, collateral loses value, a claim is paid, a market reprices.
The Network for Greening the Financial System (NGFS), a network of 151 central banks and supervisors, describes the physical side of nature-related risk as losses that result from the degradation of nature and the loss of the ecosystem services that flow from it. Water supply, flood and storm buffering, pollination, soil. When those stop arriving, the damage shows up later as a credit, market, or underwriting problem.
The exposure is not hypothetical. National assessments collected in the NGFS's 2026 note on supervising nature-related risks found:
| study | finding |
|---|---|
| De Nederlandsche Bank (2020) | €510 billion of Dutch financial-sector exposure — 36% of the exposures examined — to companies highly or very highly dependent on at least one ecosystem service |
| Banque de France (2021) | 42% of the value of securities held by French financial institutions came from issuers highly or very highly dependent on at least one ecosystem service |
| EIOPA, with the ECB (2023) | Almost 30% of European insurers' corporate bond and equity investments sat in activities with a high direct dependency on at least one ecosystem service; the largest vulnerabilities were surface and ground water, and flood and storm protection |
Read the last row twice. The biggest dependencies in insurers' investment books are water and flood protection — the work of watersheds, wetlands, and floodplains. The same insurers often underwrite the property those systems protect. When the living system fails, both sides of the balance sheet move together. That is what a supervisor means by a systemic channel.
Which specific forest, basin, or wetland a given exposure depends on — the pathway, not just the location — is its own question. The missing object is the graph covers nature-related financial risk at that level.
what the ngfs guide for supervisors asks for
In September 2026 the NGFS published an updated Guide for Supervisors on climate and nature-related financial risks, building on the original 2020 guide. It draws on a survey of supervisory authorities representing 70% of NGFS member jurisdictions, and it sets out five recommendations:
- Understand transmission — how physical and transition risks move into the economy and the financial system.
- Fit it to the mandate — consider how these risks relate to supervisory mandates, with strategies and organisation to match.
- Assess exposures — measure institutions' exposures and potential losses.
- Set expectations — establish and monitor what supervisors expect of banks and insurers.
- Act where appropriate — use micro- and macroprudential frameworks and tools to mitigate the risk.
The update treats nature in greater depth, including how climate change and nature loss compound each other. It gives more room to transition plans, to emerging efforts to bring adaptation into supervisory assessments, and to scenarios and stress tests, including early applications to nature. It is explicitly progressive and proportionate: each supervisor adapts it to its mandate, its resources, and the materiality of the risk.
It is the right document for its job. Every recommendation acts on the institution — its understanding, its exposure, its expectations, its capital. None is designed to fund the wetland. That is not a flaw. It is the mandate.
what a stress test can and cannot do
A stress test asks one question: if this scenario happens, how much does the balance sheet lose, and is there enough capital to survive it? For nature, the scenario might be a basin losing its water supply or a coast losing its buffer.
Stress tests are good at three things. They show where exposure concentrates. They force institutions to build the data. They tell a supervisor whether the system can absorb the hit.
They are also hard to get right for nature. When De Nederlandsche Bank explored nature scenarios in 2023 — mostly transition shocks, plus one on pollination decline — it found a limited financial impact, and the authors noted that most of the study's limitations pointed to a likely underestimation of the risk. The physical side, where the living system itself fails, is harder still to put on a balance sheet.
And even a perfect stress test changes the book, not the place. Capital held against a floodplain failing does not keep the floodplain working. A buffer absorbs the loss when it arrives. The wetland is what keeps it from arriving.
| job | what it produces | who acts | is the living system funded? |
|---|---|---|---|
| describe the cascade (scenarios, stress tests) | Loss estimates and capital needs | Supervisors and risk teams | No |
| disclose the cascade (TNFD, climate reporting) | A public record of dependencies and exposure | Companies and financial institutions | No |
| fund the living function | A funded, priced condition on a named place | Whoever depends on that place | Yes |
The first two rows are necessary. The third is the one that shrinks the loss. A disclosure is not a transaction makes the same point about the second row.
more cover after the loss is still after
The same timing gap sits on the insurance side. In her September 2026 State of the Union address, European Commission President Ursula von der Leyen said that only around 25% of catastrophe losses in Europe are covered by private insurance, which too often leaves national budgets as the insurer of last resort. The Commission will set up a Climate Insurance Alliance to close that gap; its terms are not yet defined. Wider cover matters, because an uninsured loss lands on households and treasuries. But more cover after the loss is still after. It changes who pays for the flood, not whether the floodplain holds. The book that walks away still pays covers the retreat side of that story.
the leftover: fund the living function
At the root, financial stability sits on something that is not financial. A wet floodplain spreads a flood before it reaches the town. An intact wetland holds water through a drought. A mangrove fringe breaks a storm surge before it reaches the coast. A living stand holds the slope above a road. When those systems work, the shock is smaller, and the balance sheet has less to absorb.
The watershed, the wetland, and the stand exist whether or not anyone books them as a systemic-risk sleeve. Ensurance funds that living function. It is not a stress-test score.
In practice, that means turning the function into something a beneficiary can hold. The condition of a named place is priced with ecosystem-service accounting and funded now, and the proceeds go to that place. The position is a certificate tied to one named natural asset, or a coin that funds across many. Each place has an agent, an onchain account that holds the funds and routes them where they are meant to go. The insurer exposed to a basin, the treasury that ends up as insurer of last resort, and the long-horizon owner whose portfolio is the whole economy all depend on the same wetland. Funding it is an investment in the thing that keeps the loss small. It is not a donation, and it is not a cost booked after the flood.
Three honest limits. This is not a supervisory tool, and it does not earn capital relief. It is not a hedge with a documented beta. And the price is a bridge, not the worth: a wetland is not worth its stress-test haircut. The haircut is only the part a balance sheet can see. Ensurance is live today with agents, coins, and certificates at small volumes, and the stocks and flows behind the pricing are public. None of this is investment or insurance advice.
what this means for you
- If you supervise, or advise supervisors: your mandate is the book, and the 2026 guide is a good map of it. As adaptation enters assessments, ask which living systems an exposure depends on and whether anyone is funding them.
- If you underwrite or invest for an insurer: the same watershed can sit on both sides of your balance sheet. Funding it now is a different ticket from reserving against its failure.
- If you hold a broadly diversified portfolio: you cannot diversify away a basin that a large share of the index depends on. You can fund it.
frequently asked questions
what is financial stability?
Financial stability is a condition in which the financial system can absorb shocks without disruptions severe enough to harm the real economy. Central banks and supervisors protect it with capital and liquidity rules, supervisory expectations, and stress tests.
how do climate and nature threaten financial stability?
Through the ordinary channels of credit, market, and underwriting risk. When a watershed, wetland, or forest stops supplying water, flood protection, or stable soil, borrowers lose revenue, collateral loses value, and insurers pay claims — often in the same place at the same time.
does a stress test reduce systemic risk?
Not directly. A stress test estimates how much a balance sheet would lose and whether capital can absorb it. That makes the system sturdier when the loss arrives, but it does not shrink the physical event. Funding the living system that interrupts the loss does.
what is the ngfs guide for supervisors?
It is the NGFS reference for integrating climate and nature-related financial risks into the prudential supervision of banks and insurers. The 2026 update sets out five recommendations — transmission, mandate, exposures, expectations, and prudential tools — and treats nature, adaptation, and stress testing in more depth than the 2020 original.
keep reading
- what systemic risk actually is — the pillar for this series
- a disclosure is not a transaction — what TNFD does, and what it does not fund
- the missing object is the graph — nature-related financial risk as dependency pathways
- exposure is not risk — what a corporate biodiversity stress test reaches and misses
- resilience is not a donation — next in this series