You get paid a fat coupon to stand in front of the storm. The whole question is whether the coupon is big enough for the storm you are actually holding.
Key takeaways
- Catastrophe bond: a security that pays a high coupon in exchange for covering an insurer’s losses from a specified disaster. If the disaster hits the defined trigger, you lose principal.
- The market is a record $61.3bn outstanding after a record $25.6bn of 2025 issuance.
- Returns are driven by disasters, not the economy, at roughly 0.24 correlation to equities over two decades.
- Three record years running: +19.69%, +17.29% and +11.40%.
- The market is structurally global. Bonds are largely issued and listed out of Bermuda, structured through London and Zurich, and reach European investors through UCITS funds. Access is now real: a US ETF (NYSE: ILS) and $19.12bn of UCITS funds. But the spread over risk is the tightest in six years.
The 60-second version
Most investors have never heard of this market, and it just had the three best years of its life. Catastrophe bonds are securities that pay you a fat coupon for agreeing to cover an insurer’s losses if a named hurricane, earthquake or wildfire strikes. They returned +19.69% in 2023, the highest annual return in the index’s history, then +17.29% in 2024 and +11.40% in 2025. New issuance hit a record $25.6bn in 2025, up 45 per cent on the year, taking the outstanding market to $61.3bn. What makes the market worth a look now is not the headline yield. It is that the return is driven by hurricanes rather than interest rates, which is why it kept paying while stocks and bonds fell together in 2022.
The investment case rests on one property that almost nothing else in a portfolio has. The return comes from whether a catastrophe happens, not from the economy. Correlation to equities has run at roughly 0.24 over two decades, and near zero to core bonds. You get paid a spread over a cash-like Treasury yield for taking a modelled, disclosed risk. Late in 2025 the market yield sat near 8.70 per cent against a modelled expected loss of 2.33 per cent. The people who run this market, from John Seo at Fermat Capital Management to the portfolio teams at Twelve Securis, Leadenhall and Schroders, have built it from a post-Hurricane-Andrew experiment into a $60bn-plus institutional asset class over three decades. The wrappers have finally reached ordinary investors too: a US-listed ETF launched in 2025, and UCITS funds crossed $19.12bn in assets.
Then there is what you are actually holding. You are selling insurance on the worst days imaginable, and one bad landfall can wipe a bond’s principal to zero. When the World Bank’s pandemic bonds triggered in 2020, Class B investors lost 100 per cent of their principal. Climate change is pushing insured catastrophe losses structurally higher. Secondary perils like wildfire and convective storms hit a record 92 per cent of insured losses in 2025, and the spread you are paid over that risk recently compressed to its tightest since November 2019. Buy at the wrong point in the cycle and you are underpaid for a fatter tail. The opportunity, the risks, and the most intelligent ways to get exposure all turn on where you buy in that cycle.
Catastrophe bonds belong alongside private credit. They are a form of privately negotiated, collateralised lending where the “default event” is a natural disaster rather than a company failing.
I. What it is: insurance you can buy on the open market
A catastrophe bond, or cat bond, is a security that turns an insurance policy into something you can buy and sell. An insurer or reinsurer, the firm that insures the insurers, wants to offload the risk of a rare, enormous loss: a Category 5 hurricane hitting Miami, a major earthquake under Tokyo, a wildfire season that torches Los Angeles. Rather than buying that cover from another reinsurer, it raises money from capital markets instead. You, the investor, hand over your principal. It sits in a collateral account. In return you receive a coupon every quarter. If the specified disaster happens and hits the defined trigger, your principal is used to pay the insurer’s claims, and you may lose some or all of it. If the disaster does not happen, you get your principal back at maturity, having clipped the coupon the whole way.
In plain terms, you are the house in a very particular casino. You collect the premium every day the storm does not come. On the rare day it does, you pay out, and the payout can be everything you staked on that hand. The bet is priced by catastrophe models, the same wind and quake models the insurance industry runs, so the odds are disclosed up front rather than hidden.
These are insurance-linked securities (ILS), the broad family of instruments whose value is linked to insurance risk rather than to companies or governments. Cat bonds are the most standardised, tradable corner of that family. Structurally they are floating-rate, principal-at-risk notes. The coupon is the yield on the Treasury money-market collateral plus an insurance risk spread, and the principal is genuinely at risk, per Morningstar’s structural breakdown. That two-part coupon is why cat bonds behaved so differently from ordinary bonds when rates moved.
Over roughly 27 years from 1997 to 2023, the cumulative loss rate, meaning money paid out to sponsors as a share of all the notional ever issued, was just 2.69 per cent. Well under three cents of every dollar issued was ever lost to a trigger. That low base rate of loss is what separates a diversified cat bond portfolio from a straightforward gamble on disaster insurance.
II. The market: history and growth trajectory
Cat bonds were born out of a single event. Hurricane Andrew flattened south Florida in 1992, bankrupted several insurers, and exposed how thin the global reinsurance balance sheet was against a truly large loss. The industry needed a bigger pool of capital than reinsurers alone could provide, and the answer was to tap the capital markets. The first deals followed in the mid-to-late 1990s, and the Swiss Re Global Cat Bond Index, the benchmark most investors track, begins in 2002.
For two decades the market grew steadily but quietly, the preserve of specialist funds and pension allocators. Then came the recent surge. Full-year 2025 issuance reached a record $25.6bn across both the standard 144A public format and private deals, up 45 per cent year-on-year and the first year annual issuance ever cleared $20bn. That pushed the outstanding market to $61.3bn at year-end, a net increase of $11.9bn, or 24 per cent, in a single year. Outstanding 144A property cat bonds alone reached $57bn, with the combined 144A market at $60.7bn, past $60bn for the first time.
The 2025 numbers broke records at every level. The market crossed 122 transactions, the first year it exceeded 100 deals. Fifteen new sponsors issued for the first time, an annual record for fresh entrants, which matters because it means the demand for this cover is broadening, not just deepening among the usual names. The fourth quarter alone saw over $7bn from 27 transactions across 45 tranches, at an average deal size of $259.6m.
A market that grew 24 per cent in a single year has outgrown niche status. It is an institutional asset class that individual investors have only recently been able to access.
| Year | Milestone | Significance |
|---|---|---|
| 1992 | Hurricane Andrew | Exposed the reinsurance capital gap; the reason cat bonds exist |
| Mid-late 1990s | First cat bonds issued | Insurance risk becomes a tradable security |
| 2002 | Swiss Re Global Cat Bond Index begins | The benchmark investors now track |
| 2023 | +19.69% total return | Record annual return in index history |
| 2025 | $25.6bn issued, $61.3bn outstanding | First $20bn+ issuance year; first 100+ deal year |
III. The demand drivers
Three forces explain why this market exists and why it is growing this fast.
Disasters are getting more expensive, structurally. Global insured natural-catastrophe losses reached roughly $107bn in 2025, the sixth straight year above $100bn, even though that figure was down 24 per cent from 2024’s $141bn. Economic losses, most of them uninsured, ran to $220bn. Every one of those dollars is a dollar an insurer would rather not carry alone, which is exactly the demand that feeds cat bond issuance. Swiss Re estimates the long-run trend growth in insured losses at 5-7 per cent real per year, a path that could take annual insured losses to around $186bn by 2030.
Secondary perils have become primary. The old mental model was that cat risk means the big Florida hurricane or the California quake. That is out of date. In 2025, secondary perils such as wildfire, severe convective storms and floods made up a record 92 per cent of insured losses. Severe convective storms alone caused around $50bn. The Los Angeles wildfires in the first quarter of 2025 caused around $40bn in insured losses, the most costly insured wildfire event on record. Balz Grollimund, Head of Catastrophe Perils at Swiss Re, put the trend plainly:
“We are observing a steady rise in losses from severe convective storms. Urbanisation in hazard-prone areas, rising asset values, higher construction costs and factors such as ageing roofs have made these storms a key peril for insurers.”
Balz Grollimund, Head of Catastrophe Perils, Swiss Re, via Artemis
The tail is genuinely fat. This is the driver that sets the price. Swiss Re estimates a 1-in-10 chance of an insured-loss year around $300bn from a major hurricane or earthquake. That peak-year scenario is precisely the risk cat bond investors are paid to hold. The bigger and more probable that tail becomes, the more cover insurers want, and the more capital the market must attract to provide it, which, all else equal, should push the coupons investors are paid higher over time.
IV. The players
A small number of named specialists run this market, and the expertise is concentrated among them.
Dedicated ILS asset managers are the core. The largest is Fermat Capital Management, co-founded and run by John Seo, which manages around $11bn, up from $10.2bn. Fermat’s offshore cat bond fund NAV reached $2.3bn, up 170 per cent in a year, and its UCITS version $2.6bn. The other heavyweights sit largely in Zurich and London. Twelve Securis, formed by the merger of Zurich’s Twelve Capital with Securis, runs offices in Zurich, London, Munich, Bermuda and Tokyo, per its own account. Leadenhall Capital Partners is London-based, Schroders runs its cat bond strategy through the Schroder GAIA range, and GAM built one of the earliest UCITS cat bond funds in the GAM Star Cat Bond Fund, before Schroders overtook it as the largest. Boutiques such as Icosa, Plenum and Franklin K2 fill out the field. These are the people who read the trigger documentation, run the models, and decide which bonds are underpriced.
Sponsors are the other side of the trade, the insurers and reinsurers offloading risk, plus a growing set of first-time issuers. The record 15 new sponsors in 2025 tells you the buyer base for protection is widening.
The modellers and reinsurers set the terms of the whole market. Swiss Re publishes the benchmark index and the loss estimates the entire industry references, and its catastrophe-perils and cat-bond specialists are among the most widely followed voices on where losses are heading.
| Player type | Role | Key names | What to know |
|---|---|---|---|
| ILS asset managers | Buy and manage cat bonds for investors | Fermat (John Seo), Twelve Securis, Leadenhall, Schroders, GAM | The largest, Fermat, runs ~$11bn; expertise is concentrated, largely in Zurich and London |
| Sponsors | Insurers/reinsurers offloading risk | 15 new entrants in 2025 | Record 15 first-time sponsors shows widening demand |
| Modellers / reinsurers | Set loss estimates and pricing | Swiss Re | Publishes the benchmark index and industry loss data |
John Seo is worth listening to on the pace of it. Asked about the 2025 surge, he called it “pretty breathtaking” and said “the issuance surge we’re seeing is far from over,” according to SWI swissinfo.
V. Geography
Cat risk is global, but two different maps matter here. One is where the perils sit, which decides what can go wrong with a bond. The other is where the market itself sits, which decides how you buy it. They are not the same place, and that gap is the whole reason a European or Asian investor can own US hurricane risk through a fund based nowhere near a hurricane.
Start with the perils. North America dominates the risk, because that is where the most expensive perils live. US hurricane, Florida and the Gulf Coast in particular, is the peak peril the market is built around, joined now by California and wider US wildfire. The $40bn Los Angeles wildfire loss in early 2025 and the roughly $50bn of US severe convective storm losses show why. Most cat bond notional references US perils, so a US-heavy fund is where the fattest coupons and the fattest tails both sit. Japan and the Asia-Pacific are the second concentration. Japanese typhoon and earthquake are long-standing cat bond perils, priced into the same $300bn peak-year tail that hurricanes drive. Quake risk is a way to hold cat exposure that is genuinely uncorrelated with the Atlantic hurricane season. Europe contributes windstorm and, increasingly, flood and convective-storm cover, part of the 92 per cent secondary-peril share of 2025 losses. Emerging markets and multilateral risk are the frontier: sovereign and development-bank issuance, of which the World Bank’s pandemic bonds (Section X) are the most famous and cautionary example.
Now the money map, which almost nobody outside the industry knows. The perils are American and Japanese, but the market is booked in Bermuda, structured through London and Zurich, and sold to European savers through Dublin and Luxembourg. Bermuda is the centre of gravity. It is the leading ILS domicile, regulated by the Bermuda Monetary Authority, which registered 36 ILS-related entities in 2025, the fourth-highest since 2010. Roughly 92 per cent of outstanding cat bonds are listed on the Bermuda Stock Exchange, whose ILS listings grew to $65.2bn at end-2025. When you buy a cat bond fund, you are almost always buying a claim on paper that lives on a Bermudian exchange.
London and Zurich are where much of the risk is structured and managed. Since January 2021, Lloyd’s has run London Bridge Risk PCC, the first UK ILS vehicle approved by the Prudential Regulation Authority and the FCA under the UK’s post-2017 ILS regime, giving investors a regulated route into Lloyd’s risk. Zurich is home to Twelve Securis and to GAM’s cat bond desk. Dublin and Luxembourg are where the retail wrapper lives: the UCITS funds that give individual investors entry are domiciled there, which is why a German or British saver can hold Florida hurricane risk in a daily-priced fund denominated in euros, sterling or dollars.
For an individual investor, the practical point is that geography is chosen for you by the fund you buy. A global UCITS fund spreads across US wind, Japanese quake and European storm; a US-tilted vehicle concentrates the peak peril. You access it through the wrappers, not by buying single-country bonds direct, and the wrapper you buy will almost certainly route through Bermuda, London or Zurich to get there.
| Region | Dominant perils | Why it matters | Access for individuals |
|---|---|---|---|
| North America | US hurricane, wildfire, convective storm | Peak peril; fattest coupons and tails | Via global/US-tilted funds and the ETF |
| Japan / Asia-Pacific | Typhoon, earthquake | Diversifies away from Atlantic hurricanes | Via global UCITS funds |
| Europe | Windstorm, flood, convective storm | Home of the UCITS wrapper (Dublin/Luxembourg) | UCITS funds in EUR, GBP or USD share classes |
| Bermuda / London / Zurich | The money, not the peril | Issuance, listing and management hubs | Where your fund’s bonds are booked and run |
| Emerging / multilateral | Sovereign, pandemic, development risk | The frontier; higher trigger risk | Rare in retail funds |
VI. How to actually invest
For most of this market’s history individual investors could not get in at all. Cat bonds were sold in large clips to institutions. That changed in 2025, and there are now three real routes in.
The ETF route is the newest and the simplest. The Brookmont Catastrophic Bond ETF, ticker NYSE: ILS, launched in April 2025 as the first US-listed cat bond ETF. It invests at least 80 per cent of net assets in cat bonds and carries a total expense ratio of 1.58 per cent. As an exchange-traded fund, the minimum is one share, with no institutional minimum to clear. The trade-off is cost. John Seo, who runs the largest ILS manager, has cautioned that cat bond ETFs “are perhaps subject to too much in frictional costs to run efficiently.”
The UCITS fund route is the main global access path, and it is where the money is going. UCITS cat bond funds are daily-priced, regulated European funds domiciled in Ireland or Luxembourg and available to individual and wholesale investors in most jurisdictions, usually in euro, sterling and dollar share classes. They reached $19.12bn in assets at end-2025, up $5.3bn or roughly 39 per cent in the year. The largest are the Twelve Cat Bond Fund at $4.55bn and Schroder GAIA Cat Bond Fund at $4.05bn, with Leadenhall’s UCITS ILS fund near $2.0bn and Fermat’s UCITS fund having added around $1.79bn, growing 238 per cent. GAM’s GAM Star Cat Bond Fund is another long-running option in the same wrapper. Minimums vary by share class, often $100,000 to $1m-plus for the institutional classes, with retail classes lower, and the exact figure lives in each fund’s key information document.
The specialist offshore fund route is for larger allocators: the Bermuda and Cayman-domiciled funds run by the same managers, typically with higher minimums and less frequent dealing than the UCITS wrappers.
| Vehicle | Liquidity | Minimum | Key risk | Best for |
|---|---|---|---|---|
| ETF (NYSE: ILS) | Daily, exchange-traded | One share | 1.58% fee; frictional costs | Smallest allocations, simplicity |
| UCITS cat bond funds (Dublin/Luxembourg) | Usually daily/weekly | Share-class dependent (often $100k+) | Cycle timing; gating risk in stress | Most individual investors globally, in EUR/GBP/USD |
| Offshore specialist funds | Monthly/quarterly | High | Illiquidity; complexity | Larger, longer-horizon allocators |
VII. Unit economics
A cat bond’s coupon has two clearly separated parts, and separating them shows exactly what you are being paid and for what.
Take a representative bond priced against late-2025 market conditions. The collateral, your principal, sits in a Treasury money-market fund earning the risk-free yield, roughly 4 per cent in late 2025. On top of that, you are paid an insurance risk spread for taking the disaster risk. The market’s modelled expected loss, the annual probability-weighted loss the models assign, sat at 2.33 per cent, and the spread over that expected loss at 2.56 per cent, taking the market yield to around 8.70 per cent as at 28 November 2025.
The worked example on a $100,000 position:
| Component | Rate | Cash on $100k |
|---|---|---|
| Treasury collateral yield | ~4.0% | ~$4,000 |
| Insurance risk spread (expected loss plus spread over) | ~4.7% | ~$4,700 |
| All-in coupon | ~8.70% | ~$8,700 |
| Modelled expected annual loss | 2.33% | ~$2,330 (probability-weighted) |
The number the professionals watch is the multiple, the risk spread divided by the expected loss. At a 2.56 per cent spread over a 2.33 per cent expected loss, the multiple is roughly 2.1x. You are paid about 2.1 times the modelled loss as premium. In a no-loss year, you clip the full ~8.7 per cent. In a year the bond’s trigger is hit, the principal covering that layer is wiped for that bond, and the coupon does not save you. So the question a buyer weighs is not yield against duration, as with an ordinary bond, but whether the premium adequately pays for how much the model might be understating the risk.
The multiple is the single most useful number for judging whether a cat bond is well priced. At ~2.1x, you are paid roughly twice the modelled loss to hold the risk. When that multiple compresses, you are being paid less for the same exposure.
That spread has been narrowing. The 2.56 per cent spread over expected loss recorded in late 2025 was the tightest since November 2019, a direct consequence of the record capital that flooded in chasing three years of double-digit returns. More buyers, same risk, lower premium. The unit economics still work. They just work less well than they did in 2023.
VIII. Macroeconomic sensitivity
Cat bonds earn a place in a portfolio because they respond to weather, not to the macro regime. The floating-rate structure means the collateral yield rises and falls with short rates, so duration risk, the losses ordinary bonds suffer when rates rise, is largely absent.
| Regime | Impact | Rationale |
|---|---|---|
| High inflation / rising rates | Positive | Floating-rate collateral yield rises with short rates; near-zero duration means little price loss, the opposite of long bonds |
| Low inflation / falling rates | Mixed | Collateral yield falls, trimming the all-in coupon, but the insurance spread is unaffected |
| Recession | Resilient | Returns come from disasters, not the economy; ~0.24 correlation to equities held the class up when stocks and bonds fell together |
| Stagflation | Resilient | No corporate-credit or growth linkage; the risk is a hurricane, not a downturn |
The counterintuitive point is that the one thing that reliably hurts cat bonds is precisely the thing that has nothing to do with the economy. In 2022, as rates spiked and both stocks and bonds fell, the cat bond market fell only about 2 per cent, and that drawdown was caused by Hurricane Ian, not by the Federal Reserve. Post-storm pricing sheets fell 5-10 per cent, a loss driven by wind speed, not interest rates. That episode is the diversification case made concrete. Because the drawdown comes from peril losses, it does not line up with anyone else’s bad year.
IX. Tax considerations: a global overview
This is general information, not personal tax advice. Treatment depends entirely on where you are resident and which wrapper you hold. Model your own after-tax return, and check with an adviser in your jurisdiction before acting.
The core fact to understand is the character of the income. A cat bond’s return is the collateral yield plus the insurance risk spread, and in most jurisdictions that is taxed as ordinary interest or income, not as qualified dividends and not as capital gains. In practice that makes cat bond exposure more efficient held inside a tax-sheltered or retirement wrapper than in a taxable account, because the headline ~8-11 per cent is a gross figure taxed at your marginal income rate.
The structure adds a second layer. The special purpose vehicles that issue cat bonds are usually domiciled in tax-neutral hubs, Bermuda or the Cayman Islands, per Morningstar’s structural notes, and the retail wrappers in Ireland or Luxembourg, as with the UCITS funds. Those domiciles are chosen so the vehicle itself is not a taxable drag; the tax event lands at the investor’s own residence instead.
| Consideration | What to ask your adviser |
|---|---|
| Income character | Is the coupon taxed as interest/income in my jurisdiction? |
| Wrapper choice | Is it more efficient inside a retirement/tax-sheltered account? |
| Fund domicile | How is an Irish/Luxembourg UCITS or offshore fund treated where I live? |
| Withholding | Does any withholding apply at the fund or SPV level? |
| After-tax yield | What is the ~8-11% gross worth to me net, at my marginal rate? |
The headline yield is the number that sells the asset. For a cat bond investor what matters is the after-tax, after-fee yield at your own marginal income rate, and that gap is wider here than for an asset paying capital gains.
X. Case studies
Three real episodes each carry a different lesson.
The cautionary tale: the World Bank pandemic bonds, 2020. In 2017 the World Bank issued cat bonds through its Pandemic Emergency Financing Facility, designed to pay developing countries fast when a pandemic struck. COVID-19 triggered them, and on 17 April 2020 they paid out $195.84m, 100 per cent of Class B principal and 16.67 per cent of Class A. Class B investors lost their entire principal. That is the base case a cat bond investor must sit with: a trigger can wipe a tranche completely. Worse, critics argued the parametric trigger fired too slowly to help. The Center for Global Development called it “a good idea executed badly” and urged the Bank not to renew it. The lesson cuts both ways. The structure did what it said it would, and yet the trigger design was flawed enough that both the payer and the payee came away unhappy. The trigger document is the thing to read closely before buying.
The near-miss: Hurricane Milton, October 2024. When Milton bore down on Tampa, early estimates put potential cat bond losses at anywhere from 2 to 15 per cent. A more southerly track and wind shear spared the city, and losses came in at a maximum of around 4 per cent, likely single digits, with the insured-loss range revised down to $20bn to $60bn. Tanja Wrosch, Head of Cat Bond Portfolio Management at Twelve Capital, captured the knife-edge:
“Cautiously, we’d say it’s likely less. It played out in the end a little bit better.”
Tanja Wrosch, Head of Cat Bond Portfolio Management, Twelve Capital AG, via Insurance Journal
The lesson is that outcomes hinge on landfall geography measured in tens of miles. Where the eye crosses the coast can decide whether a bond pays in full or takes a loss.
The drawdown and recovery: Hurricane Ian, September 2022. Ian was a top-five US insured-loss event, and the cat bond market absorbed it with a full-year decline of only about 2 per cent. Pricing sheets fell 5-10 per cent post-storm, implying $1.6bn to $3.6bn of value loss, but much of that was mark-to-market rather than realised claims, and a portion recovered as loss estimates firmed. The more encouraging lesson is that even a genuinely huge storm produced a modest, recoverable drawdown rather than a wipeout. Whether an investor meets resilience or fragility depends on the single bond and the single storm.
XI. The core constraint
The bottleneck that defines the future of this asset is the quality of the catastrophe model. The entire market prices off modelled expected loss, the 2.33 per cent figure that sets what you are paid. If the models are right, you are being fairly compensated at a ~2.1x multiple. If the models understate the risk, you are being underpaid for a fatter tail than you think you hold.
The constraint is getting harder, not easier, because the perils are changing faster than the historical record the models are trained on. The rise of secondary perils to a record 92 per cent of insured losses, meaning wildfire, convective storms and flood, is exactly the category where models have the least history and the most uncertainty. A hurricane model has a century of data. A model for how climate change reshapes wildfire behaviour in a warming Los Angeles has far less.
The industry is solving this by reinvesting relentlessly in the models, widening the data, and pricing conservatively where uncertainty is high. For investors, the practical implication is that manager skill in this market is largely model skill, the ability to judge which bonds are priced off honest models and which are priced off wishful ones. That is why the expertise, and the assets, concentrate in a handful of specialists like Fermat’s $11bn rather than spreading thinly across the market.
XII. Inside the asset
What are you actually holding? A cat bond is, structurally, a short-dated floating-rate note wrapped around a special purpose vehicle. Typical deal size runs $50m to $500m, occasionally $1bn to $2bn for the largest programmes, with maturities of one to five years, three to four the norm. Your principal does not go to the insurer. It goes into a collateral account, held in Treasury money-market funds by the SPV, which is why the collateral earns the risk-free rate and why credit risk on the sponsor is largely stripped out.
The trigger is the beating heart of the instrument, and there are different kinds. An indemnity trigger pays out based on the sponsor’s actual losses. A parametric trigger pays based on a measured physical parameter, a wind speed at a location or an earthquake magnitude, regardless of actual losses. The World Bank pandemic bonds were parametric, which is why they could pay 100 per cent of a tranche on a formula while critics argued the formula fired at the wrong time. Parametric triggers are fast and transparent but can miss reality; indemnity triggers track reality but settle slowly. The document that defines the trigger is the single most important thing to understand about any cat bond. It is the difference between the ~4 per cent maximum loss of Milton and the total loss of the World Bank Class B.
Hold a cat bond and you own a slice of a very specific bet, written in legal and meteorological language. This much money says this named storm, of this severity, in this place, will or will not happen before this date.
XIII. The central dilemma
You are paid for correlation to nothing, which means you are also exposed to something the rest of your portfolio cannot warn you about.
The strength is genuine. A ~0.24 correlation to equities and near-zero to core bonds means cat bonds keep paying when everything else is falling. The 2022 episode, where the market fell ~2 per cent on a hurricane while stocks and bonds cratered on rates, is the proof. Nothing in your macro dashboard, no yield curve, no earnings season, tells you a Category 5 is forming.
But that same independence is the trap. Because the risk is uncorrelated, it is also unhedgeable within a conventional portfolio and invisible until it arrives. And the tail is real: a 1-in-10 chance of a $300bn insured-loss year is a scenario in which a concentrated cat bond position takes serious, simultaneous losses across bonds that all reference the same peril. The diversification that protects you against a recession does nothing for you in a bad hurricane season, which is the one risk your macro dashboard cannot flag.
Most serious allocators resolve this by treating cat bonds as a question of how much to hold rather than whether to believe in them. The uncorrelated return earns the asset a place; the fact that a single bad season can inflict severe damage is why that place stays small.
XIV. The next frontier
The mainstream cat bond thesis is US hurricane and Japanese quake. The more interesting frontier is the widening of the perils and the wrappers.
On perils, the growth is in secondary-peril and parametric structures, meaning bonds referencing wildfire, severe convective storms and flood, the category that reached 92 per cent of 2025 insured losses. The record 15 new sponsors in 2025 are bringing new perils and new geographies to market, which is where the least-modelled, highest-uncertainty risk sits, and therefore the potentially best- or worst-priced risk too.
On access, the frontier is the democratisation of the wrapper. The NYSE: ILS ETF launching in 2025 and UCITS assets jumping to $19.12bn mean the capital base is broadening from pension funds to individual investors for the first time. In Europe, the reforms strengthening the UK’s ILS regime and the deepening Dublin and Luxembourg fund market point the same way, widening the routes a retail saver can take into risk that used to be institutional only. That is a genuinely different thesis from the mainstream one. Not “should I own hurricane risk,” but “what happens to the pricing when a $60bn market suddenly has millions of new potential buyers.” John Seo’s warning that ETFs may carry too much frictional cost is the frontier’s open question: whether the cheap wrapper can deliver the asset’s returns cleanly.
XV. Lessons from history
Hurricane Andrew, 1992, the founding lesson. Andrew showed the world that the reinsurance system was undercapitalised against a truly large loss, and it created cat bonds as the answer. The lesson: this asset class exists precisely because catastrophe risk is bigger than the traditional insurance balance sheet can hold. That is a structural reason the market keeps growing, not a cyclical one.
Hurricane Ian, 2022, the resilience lesson. A top-five US insured-loss event produced a full-year cat bond decline of only ~2 per cent, much of it mark-to-market that later recovered. The lesson: the asset absorbs even huge storms far better than intuition suggests, because losses attach only above defined trigger points and are spread across many diversified bonds.
The World Bank pandemic bonds, 2020, the humility lesson. A well-intentioned, sophisticated structure delivered a total loss to Class B investors and drew accusations of being “executed badly”. The lesson: the trigger design, not the headline peril, decides your outcome, and even the smartest issuers get triggers wrong.
The common thread across all three is that cat bonds reward investors who read the specific risk and punish those who buy the headline yield. Andrew explains why the market exists. Ian shows that it is more robust than it looks. And the pandemic bonds are the reminder that more robust than it looks is not the same as safe. A strong run also has a cost that shows up later. The 2023-24 stretch delivered just over 40 per cent combined, and returns like that draw in capital, compress the spread, and leave the next wave of buyers underpaid for the risk.
XVI. The case for it
Genuine diversification, proven under stress. The ~0.24 correlation to equities and near-zero correlation to bonds is not a backtest artefact. It held live in 2022, when the market fell only ~2 per cent on a hurricane while a 60/40 portfolio had one of its worst years in a century. Very few assets can claim a source of return that is structurally independent of the economy.
A low realised loss rate over decades. The cumulative loss rate from 1997 to 2023 was 2.69 per cent of all notional issued. Across 27 years including multiple record hurricane seasons, under three cents of every dollar issued was lost to a trigger. The tail is real, but the base rate of loss has been low.
Floating-rate structure removes duration risk. Because the collateral sits in Treasury money-market funds, the coupon rises with short rates and the price barely moves when rates rise, the opposite of the long bonds that lost double digits in 2022. In a high-rate world, that is a structural tailwind ordinary fixed income cannot match.
A durable, growing demand base. Insured losses trending at 5-7 per cent real per year toward ~$186bn by 2030 mean insurers will keep needing more cover than their own balance sheets can hold. The record $25.6bn of 2025 issuance and 15 new sponsors are the demand made visible. The people who win here are the specialists who read triggers and models better than the market prices them, which is why assets concentrate with managers like Fermat.
XVII. The risks
A trigger can wipe your principal. This is the risk that dwarfs the rest. The World Bank Class B investors lost 100 per cent. A single bad landfall on the wrong bond does not merely dent your return; it ends it for that layer, and the coupon does nothing to cushion a trigger.
The spread you are paid has compressed. The 2.56 per cent spread over expected loss in late 2025 was the tightest since November 2019. Three record years drew in record capital, and more buyers chasing the same risk means a thinner premium. Buy at a tight point in the cycle and you are underpaid for the tail.
Model risk is rising with climate change. Pricing rests on catastrophe models, and the perils driving losses, the 92 per cent secondary-peril share of 2025, are exactly where the models have the least history. If the models understate a changing climate, the 2.33 per cent expected loss you are paid against is too low.
Concentrated tail exposure. The 1-in-10 chance of a $300bn insured-loss year is a scenario where many bonds referencing the same peak peril lose together, defeating the diversification within a cat bond portfolio itself.
Wrapper and liquidity risk. The ETF carries a 1.58 per cent fee and, per John Seo, possibly too much frictional cost to run cleanly. UCITS funds can gate or slow dealing in a stressed market, exactly when you might want out.
XVIII. The Alternative Fortune verdict
Cat bonds are one of the few assets whose central promise, a return uncorrelated to the economy, is both true and proven under live stress. The ~0.24 equity correlation, the 2.69 per cent cumulative loss rate over 27 years, and the floating-rate structure that sidestepped the 2022 bond rout are real advantages, not marketing. What you give up for them is severe: you are selling insurance on the worst days imaginable, a trigger can wipe a tranche to zero, the models pricing the risk are being tested by a changing climate, and after three record years the spread you are paid has compressed to its tightest in six years.
Weighed against the alternatives, cat bonds are a genuine diversifier and a poor core holding. Versus investment-grade bonds, they offer higher income and no duration risk but a fatter, uglier tail. Versus high-yield credit, they offer comparable income with a completely different risk driver, weather rather than the credit cycle, which is where their value sits. On balance the diversification case is stronger than the yield case. Buy this for what it is not correlated to, and size it for the storm season it can lose in. It suits an investor who already holds equities and bonds and wants a return stream that does not move with either. It does not suit anyone who needs the money to be there regardless of the weather, or who is buying purely because the last three years printed double digits.
Where the edge actually is. The edge is not the headline yield. That is the most competed-away part of the asset, as the compressed spread shows. It sits instead in three specific places. First, the multiple, timed to the cycle: the ~2.1x spread-to-expected-loss is the number to watch, and the real money is made buying when a hard market pushes it wide, typically after a big loss year scares capital away, not after three soft record years. Second, model discernment: because the whole market prices off catastrophe models, the edge belongs to the specialists who can tell an honestly-priced bond from a wishfully-priced one, which is why Fermat and its peers command the assets. Third, the wrapper arbitrage: John Seo’s warning that ETFs may bleed frictional cost means the cheapest-looking access is not always the one that delivers the asset’s return. The edge can lie in paying a specialist manager who nets more after their fee than a cheap tracker does after its frictions.
Questions to ask before you invest, by vehicle:
If you are looking at the ETF (NYSE: ILS): – Does the 1.58 per cent fee plus trading frictions leave enough of the ~8-11 per cent gross to justify the wrapper over a fund? – How does it handle a triggered bond mid-year: does the price gap, and can I stomach that intraday?
If you are looking at a UCITS cat bond fund (Dublin or Luxembourg): – What is the fund’s current portfolio multiple (spread ÷ expected loss), and how does it compare to the market’s ~2.1x? – Which share class and currency (EUR, GBP or USD) suits me, what is the minimum, and what are the total fees net of the collateral yield? – Can the fund gate or suspend dealing in a stressed market, and under what terms? – What is the peril and geographic concentration: how much sits in US peak-peril hurricane?
If you are looking at an offshore specialist fund: – What is the dealing frequency and lock-up, and does my horizon match it? – What is the manager’s track record through an actual major loss year, not just the 2023-24 record run?
Cat bonds have earned their reputation as one of the cleanest diversifiers available to an individual investor, and the wrappers have finally made them reachable. The asset rewards the investor who reads the trigger, watches the multiple, respects the tail, and buys when capital is scarce rather than when returns are advertised. It punishes the one who buys on the strength of the last three years’ numbers. Those three headline years are both the reason to look and the reason to be careful. What you do with that is your call to make.
| Claim | Source | Status |
|---|---|---|
| 2025 issuance $25.6bn, +45%, first $20bn+ year | Artemis | verified ✓ |
| Outstanding market $61.3bn, +24%/$11.9bn | Artemis | verified ✓ |
| 122 transactions; Q4 $7bn/27 deals/45 tranches/$259.6m avg | Risk & Insurance | verified ✓ |
| 15 new sponsors (record) | Artemis | verified ✓ |
| 144A property $57bn; combined 144A $60.7bn | Artemis | verified ✓ |
| Cumulative loss rate 1997-2023 = 2.69% | Morningstar | verified ✓ |
| 2025 index +11.40%; 2024 +17.29%; 2023 +19.69%; 2023-24 ~40% | Artemis | verified ✓ |
| ~0.24 equity correlation, near-zero to bonds | Morningstar | verified ✓ |
| 2022 market ~ -2%, sheets -5-10% (Ian) | Artemis | verified ✓ |
| 2025 insured nat-cat ~$107bn; 2024 $141bn; economic $220bn | Artemis | verified ✓ |
| LA wildfires ~$40bn | Artemis | verified ✓ |
| SCS ~$50bn; secondary perils 92% of insured losses | Swiss Re | verified ✓ |
| Insured losses 5-7% real/yr; ~$186bn by 2030 | Reinsurance News | verified ✓ |
| 1-in-10 chance of ~$300bn peak-loss year | Swiss Re | verified ✓ |
| Structure: $50m-$500m, 1-5yr (3-4 norm), coupon = collateral + spread | Morningstar | verified ✓ |
| Late-2025: yield ~8.70%, EL 2.33%, spread 2.56% (tightest since Nov 2019) | Artemis | verified ✓ |
| Multiple ~2.1x; ~8.7% all-in | Artemis | verified ✓ (arithmetic re-checked below) |
| ETF NYSE: ILS, Apr 2025, TER 1.58%, ≥80% cat bonds | Businesswire / etf.com | verified ✓ |
| UCITS sector $19.12bn, +$5.3bn/~39% | Artemis | verified ✓ |
| Twelve $4.55bn; Schroder GAIA $4.05bn; Leadenhall ~$2.0bn; Fermat UCITS +$1.79bn/238% | Artemis | verified ✓ |
| Fermat ~$11bn; offshore NAV $2.3bn/+170%; UCITS NAV $2.6bn | SWI swissinfo | verified ✓ |
| World Bank pandemic bonds paid $195.84m; Class B 100% loss; Class A 16.67% | Artemis / CGD | verified ✓ |
| Hurricane Milton: 2-15% early → max ~4%/single digits; $20bn-$60bn | Insurance Journal | verified ✓ |
| Seo “pretty breathtaking” / “far from over” / ETF frictional-cost quote | SWI swissinfo | verified ✓ |
| Wrosch Milton quote | Insurance Journal | verified ✓ |
| Grollimund SCS quote | Artemis | verified ✓ |
| Tax: coupon = ordinary income; SPV Bermuda/Cayman; UCITS Ireland/Lux | Morningstar / Securis | verified ✓ |
| Bermuda leading ILS domicile; BMA 36 ILS registrations in 2025 | Artemis | verified ✓ (globalisation add) |
| BSX ~92% of outstanding cat bonds; $65.2bn ILS listings end-2025 | Artemis | verified ✓ (globalisation add) |
| UK ILS regime; Lloyd’s London Bridge Risk PCC, PRA/FCA-approved Jan 2021 | Artemis | verified ✓ (globalisation add) |
| Twelve Securis Zurich-based (Zurich/London/Munich/Bermuda/Tokyo); GAM Star Cat Bond; Schroders overtook GAM | Twelve Securis / Artemis | verified ✓ (globalisation add) |
Arithmetic re-check (Section VII): Collateral ~4.0% plus insurance spread (expected loss 2.33% plus spread over EL 2.56% = ~4.9%) reconciles to the stated market yield of ~8.70%; on $100k that is ~$8,700 gross. Multiple = spread ÷ expected loss = 2.56 ÷ 2.33 ≈ 1.10 as a spread multiple, but the market convention “multiple” = (spread over EL plus EL) ÷ EL = 4.89 ÷ 2.33 ≈ 2.1x, matching the brief’s stated ~2.1x. Figures rounded for readability.
Spelling / framing: British spelling verified (favour, capitalise, per cent, modelled, centre). Jurisdiction-neutral: no “how to invest from [country]”; peril geography spans North America, Japan/APAC, Europe and emerging/multilateral; market geography spans Bermuda, London, Zurich, Dublin and Luxembourg; tax framed as “ask your adviser.” Category pillar linked: https://alternativefortune.com/investments/private-credit.
Note on softened figures: The Treasury collateral yield (~4% in late 2025) and the ~8-11% gross-yield range are framed as approximate context rather than pinpoint sourced figures, since the brief gives the ~8.70% market yield as the precise, dated data point and describes the collateral component qualitatively. No other figure was softened; all specific numbers carry their source link.