A catastrophe bond is one of the cleanest instruments in finance. Cash goes into a trust. A named peril goes into the contract. If the peril stays quiet, the investor collects a coupon and gets the principal back at maturity. If it fires, the principal is released to the sponsor to pay claims.
Every part of that works. No part of it touches the storm.
If you are reading about cat bonds because you want a return driver that is not the same rate cycle wearing a different hat, you have landed somewhere real. This page defines the instrument in the market's own vocabulary — sponsor, SPV, collateral trust, trigger, attachment point, spread, principal at risk — and then names the one job the structure was never built to do. We apply the same test to ourselves at the end.
the structure, in the order it happens
1. A sponsor needs cover. An insurer, a reinsurer, a state wind pool, or a sovereign wants protection against a specific peril in a specific place — Florida hurricane, California earthquake, European windstorm, Japanese typhoon. Rather than buy all of that cover from other reinsurers, the sponsor goes to the capital markets.
2. A special purpose vehicle is formed. The SPV is a standalone issuer, commonly domiciled in Bermuda, the Cayman Islands, or Ireland. It does exactly two things: it sells notes to investors, and it writes a reinsurance or retrocession contract back to the sponsor. It has no other business, no other liabilities, and no balance sheet to be surprised by.
3. The cash is collateralized. Note proceeds go into a collateral trust, typically held in short-dated government money market funds. This is the structural feature that makes a cat bond different from a reinsurance promise: the limit is already funded, in cash, before anything happens. The reinsurer's credit risk is gone; what remains is the collateral itself, which is why the market moved to Treasury funds after 2008. That is precisely why sponsors pay for it.
4. A trigger is defined. The contract specifies what counts as the loss for payment purposes. The four common forms are indemnity (the sponsor's own actual claims), industry loss index (an estimate of insured losses across the whole market), parametric (a measured physical value such as wind speed, ground acceleration, or storm-track location), and modeled loss (the sponsor's exposure run through an agreed catastrophe model). The trigger has an attachment point, where losses start hitting the notes, and an exhaustion point, where the notes are wiped out.
5. The coupon runs while the trigger stays quiet. The investor's coupon has two components: whatever the collateral earns in the money market, plus a risk spread paid by the sponsor for standing behind the peril. The spread is the price of the peril. That is the whole economic transaction.
6. If the trigger fires, principal pays the claim. Collateral is released to the sponsor up to the limit. Investors lose principal — partially at attachment, entirely at exhaustion. No one has to chase a reinsurer. For indemnity deals, that release runs through a loss-development period — the notes can be extended while claims settle. Parametric and index triggers are the fast ones.
Terms are commonly multi-year with an annual risk period. Most deals are placed privately and traded under Rule 144A to qualified institutional buyers. Direct notes go to QIBs (or offshore buyers under Reg S); most other exposure runs through cat bond funds. That gate is its own post, and it is not legal advice from us.
A catastrophe bond is pre-funded, collateralized reinsurance in note form. The loss gets paid faster and more certainly than a promise can pay it — and the loss still happens in full.
the one job the structure was never built to do
Read those six steps again and notice what every one of them is about: who eats the loss. The sponsor moves it. The SPV holds it. The collateral funds it. The trigger measures it. The spread prices it. Six well-engineered steps, all of them downstream of the event.
You might be reading that as criticism. It isn't. Rearranging who eats a loss is genuinely valuable work — it is what risk transfer is for, and the cat bond does it better than most instruments in the market, because the cash is already there. A homeowner whose insurer stays solvent after a category four storm is better off because this market exists.
But it has a ceiling, and the ceiling is physical. The wetland that would have taken the surge is still gone. The watershed that would have moderated the flood peak is still degraded. The forest whose condition sets fire behavior is still overloaded. Every dollar in the trust is arranged around the moment of loss, and the moment of loss is the one thing the money cannot move.
the timing table
Three instruments, one axis. Where the capital sits relative to the event determines almost everything else about it.
| catastrophe bond | insurance policy | ensurance | |
|---|---|---|---|
| what the capital buys | a spread for standing behind a named peril | compensation after a defined loss | the work that keeps a named living system in its present condition |
| when the money moves | after the trigger fires | after the claim is adjudicated | now, while the system is functioning |
| who receives it | the sponsor's claims budget | the named insured | the steward of the named natural asset |
| what has to be true | the loss or index crosses the attachment point | the peril is covered and the limit is intact | the condition is observable today |
| effect on the loss itself | none directly — it reassigns who pays | none directly — it reassigns who pays | it aims to make the loss smaller |
| how it fails | the peril fires and principal is gone | the peril is excluded, or the limit is exhausted | the funded work underperforms and the loss lands anyway |
That last row matters. The honest failure mode of funding protection is that you paid for the work and the loss arrived regardless. Anyone who tells you otherwise is selling you a floor that does not exist.
the market is real, and it is large
This is not a fringe corner of finance. Artemis tracked more than $11.3 billion of new catastrophe bond risk capital in the second quarter of 2026, across a record 48 transactions — the largest single quarter in the market's history. First-half issuance came in near $18 billion, and the outstanding market stood at roughly $65.6 billion at the end of June.
Read that number for what it actually tells you. It is not a measure of how protected anything is. It is a measure of how much capital has already accepted physical-world perils as a return driver. That is the audience this series is addressed to. If your committee has already accepted physical-world perils as a return driver, you have already crossed the conceptual line most books still treat as exotic. What we cannot hand you is the modeled expected loss and priced multiple that made that acceptance easy. What we can show is the timing.
the trigger is the definition of the loss, not the place
One clarification that trips up newcomers: the trigger does not describe what happened to a place. It describes what the contract will treat as having happened.
An industry loss index pays on a market-wide estimate, not on the sponsor's actual book. A parametric trigger pays on a measurement — a wind speed at a station, a storm track through a defined box — not on damage. Speed is the point of those designs, and speed is worth real money when a community needs cash in weeks rather than quarters. The trade is that the payout and the damage are two different numbers, and the gap between them is a feature of the design rather than a bug in it.
That gap has a name, basis risk, and it gets its own treatment in the payout that misses the place. The takeaway here is narrower: a trigger is a financial definition, never a description of a living system's condition.
"but cat bond money does fund resilience"
It is the strongest objection to this page, and it deserves a straight answer rather than a dodge.
Payouts do fund real work. Sovereign and public-sector sponsors use cat bond proceeds for emergency response, temporary housing, debris clearance, and reconstruction. That work is necessary and the speed of pre-funded capital genuinely improves it. Nothing here argues against it. But it is response — it happens after, because that is when the trigger releases the money.
Sponsors do have a mitigation incentive. A sponsor that lowers its expected loss lowers what it has to pay for cover. That incentive is real, structural, and underrated — the most promising channel by which insurance capital could end up funding physical work in advance. It is also indirect, uneven, and entirely at the sponsor's discretion. The note does not fund it. The note prices the peril the sponsor still has.
And some sponsors are conservation vehicles. Nature-linked covers exist where a trust or a response fund receives the payout and pays for post-event ecological work. That is a thoughtful use of the structure — and it is still pay-after, which is the whole argument in nature-based insurance is still insurance.
None of these make the structure worse. They make it what it is: excellent capital arranged around the event.
same hunt, different job
The reason an allocator ends up in cat bonds is rarely the acronym. It is a conviction that physical risk in the real world is a legitimate, differently-driven place to put capital. That conviction is correct. We think it is early rather than exhausted.
What we do is aim the same appetite at a different point on the timeline. Two instruments, and both need a one-line gloss before they mean anything to an ILS reader:
- A certificate is a holdable claim tied to one named natural asset. Buying it routes funds to the stewardship of that asset now, while the system is functioning — not after an event, and not against a projected counterfactual.
- A coin is the fungible, protocol-wide version. Trading activity generates fees that fund protection indirectly, which makes it the tradable, small-ticket way to take the position.
Neither one is a note. Neither one has an attachment point, a limit, a term, or a sponsor. What they have is a named place, a steward, and money that moves while the system is still intact. The timing argument in full sits behind that.
what we are not selling
Stated in one place so it cannot be misread later.
We are not selling a catastrophe bond. There is no SPV, no collateral trust, no trigger, no attachment point, no rated note, no 144A offering, and no coupon. Nothing here is a substitute for an ILS allocation, and if your mandate needs collateralized event cover with a defined limit, the cat bond market is the right market and it is very good at its job.
We also make no claim to uncorrelated returns. We have no track record to point at, no documented beta, and no return series to hand your risk team. Our instruments are live and our volumes are small. The honest description of what we offer is a position in the funding of a named place, held in the present tense — with capital at risk, an ecological outcome that is not guaranteed, and a horizon measured in years.
Same hunt. Different job. Funding protection before loss is not new — land trusts and public conservation budgets already do it. Investment capital has done it too: forest resilience bonds and environmental impact bonds fund the work before the event and repay from the beneficiaries. What we add is holding that funding at the level of one named place, as a position you can take. We would rather you hold us to that than to something grander.
frequently asked questions
what is a catastrophe bond?
A catastrophe bond is a note that transfers a specific natural-disaster risk from an insurer, government, or corporation to capital markets investors. Investors post cash into a collateral trust and earn a spread while the covered peril stays below its trigger. If the trigger fires, the collateral pays the sponsor's losses and investors lose principal.
how do catastrophe bonds work?
A sponsor forms a special purpose vehicle, which issues notes to investors and writes a reinsurance contract back to the sponsor. Proceeds sit in a collateral trust invested in short-dated government funds. A trigger — indemnity, industry loss index, parametric, or modeled loss — defines the qualifying event. While the trigger stays quiet, investors receive the collateral return plus a risk spread. When it fires, collateral is released to the sponsor up to the limit, and principal is reduced or lost. Terms are typically multi-year.
are catastrophe bonds uncorrelated?
The industry pitch is low correlation with financial markets, and the reasoning behind it is sound: returns are driven by peril incidence and reinsurance pricing rather than by earnings or rates. Treat it as a claim about the return driver, not as immunity. Spreads widen, liquidity thins in stressed windows, marks move, and a large enough event is a capital event. Low correlation with markets also says nothing about correlation with the physical world — which is the exposure that is actually growing. And to be explicit about our own position: ensurance instruments have no documented correlation profile at all, because there is no track record to measure. We treat "uncorrelated" as a claim requiring evidence, including when it is made about us. The three ways those claims fail applies to everyone, us included.
where to go next
If you are still mapping the category, the next two posts finish the definition work:
- insurance-linked securities are still insurance — the wider ILS family (collateralized reinsurance, sidecars, industry loss warranties) and why the link is the capital, not the job.
- the payout that misses the place — basis risk in plain language, and why a parametric miss is the product working as designed.
If you are weighing what role any of this plays in a book, go to a hedge against what? — it separates hedge, diversifier, third leg, and protection into four jobs with four tests, and applies them to us.
And if you want to work through where present-tense funding of physical risk could sit in a real mandate, the investor page is where that conversation starts.
the series
catastrophe bonds pay after. ensurance pays now.