Over the past several years deploying capital in the insurance capital markets, I’ve found that most coverage rarely goes beyond surface-level headlines, outcomes, and broad narratives. 

This newsletter takes a different approach - a perspective from inside the market - the pricing cycles, the capital flows, what’s moving markets, and what’s behind the headlines.

One of the most sophisticated investors I've spoken to this year told me a story I keep thinking about.

About a decade ago, they invested in a catastrophe bond strategy. Two bad loss years arrived back to back. They believed they could handle the volatility - it turned out they couldn't. They pulled their capital.

Five years later, the pricing environment was the best in a generation. Those that weathered the storm made a return that more than compensated for the losses.

What stays with me isn't the mistake they made in pulling their capital. It's that they made the mistake while fully understanding the asset class.

In insurance, the returns go to whoever is still in the room when pricing resets.

Everyone reading this has been trained by fifteen years of equity markets to buy the dip. The logic is sound: a drawdown is a price fall on the same asset. If the fundamentals are intact, the expected return just went up. Acting on that has been the winning trade since 2009.

In insurance, the reflex misfires - not because the logic is wrong, but because the asset is different.

A bad year in insurance isn’t a price fall; it’s a capital event. A hurricane or a court system consumed the money. And because prices in this market are set by the supply of capital willing to carry the risk (Issue 4), the asset on offer afterwards - next year's renewal contract - pays more because there is less capital left to write it.

Which leaves the investor who redeems after a loss year in a strange position: they aren’t selling the asset that fell in value. They’re declining to buy the policy renewal - a different, better-paying asset - either because a backward-looking account statement told them to, or because they can’t afford it.

The industry's own word for post-event pricing - 'payback' - is a clear manifestation of this.

One distinction before going further. Insurance investments bifurcate into quite different categories - each running on different clocks.

Property catastrophe - the fast clock: Hurricanes, earthquakes, wildfires. The event comes, the bill arrives within a couple of quarters, and the market reprices at the next renewal. This is the world of catastrophe bonds (and the strategy in the story above).

Casualty - the slow clock: Injuries, negligence, product failures - the things that end in court. The loss is an estimate that develops over time. An insurer books a reserve today against lawsuits that will settle in years and, until they do, results are whatever the reserve assumptions say they are. The actual result arrives years later as "reserve strengthening" or “reserve releases” - the industry's terms for discovering that old losses were bigger or smaller than the money set aside for them.

The mechanism is the same in both categories - capital gets consumed, price follows - just at different speeds. The two categories are currently trending in opposite directions. Property catastrophe - the fast clock - is softening (rates going down). Casualty - the slow clock - is hardening (rates going up).

A single catastrophe, however large, almost never turns a softening market on its own. At most, it interrupts a softening trend. What turns the market is accumulation - successive loss years grinding down capital and patience - compounded by a shock that hits the whole market.

2017 was the costliest catastrophe year on record at the time - $144bn of insured losses, led by the $92bn loss from Harvey, Irma and Maria.¹ The market's response at the next renewal was a 6.1% price increase - the first increase in six years. Then the softening immediately resumed.

The losses were "largely within model parameters, and overall industry capital did not decline."² The models had priced in the losses - capital survived them - and investors backfilled any gaps within weeks.

Fast forward to 2022. Hurricane Ian was smaller than the 2017 trio, but it landed on a market that had already absorbed three loss years, a pandemic, and two further years of old losses creeping past what the models had projected - and it came in the same twelve months that inflation repriced every insured asset while rising interest rates cut the value of reinsurers' bond portfolios, taking industry capital down 15% in a single year.³

That combination; i) losses, ii) eroded belief in the models after several years of losses, and iii) a hit the balance sheet, produced the 2023 reset: the sharpest price increase since 2006.⁴

The 2023 reset paid those that stayed around. The catastrophe bond market returned 19.7% in 2023, 17.3% in 2024, and 11.4% in 2025 - the first three consecutive double-digit years in the benchmark's history.⁵

But staying was harder than it looked. The largest investor in the market - PGGM, the Dutch pension manager - was down 13.3% in 2017, and after five more years of grinding results was still underwater at the end of 2022. Then +18.9% and +15.1% in 2023 and 2024.⁶

Whether you could be present was largely decided at entry. Through the bad years, the average fund in this market lost 5.6%, then 2.9% - painful but recoverable.⁷ But investors in the more leveraged corners of the market watched losses get re-estimated upward for years while their collateral sat frozen against them - "trapped" - in the market's vocabulary.

At the end of 2019, roughly one dollar in six of the investor-funded market - the capital investors put in directly, as opposed to reinsurers' own balance sheets - was locked up this way, and the era's largest leveraged vehicles lost most of their value and wound down.⁸ For those investors, staying wasn't on the menu.

So the framework question is not "can I tolerate the volatility”. The question is whether the position was built to be held: instruments you can keep through the bad years, liquidity terms that don't force a decision at the bottom, and an evaluation clock that runs longer than the immediate payoff cycle. For investors that pulled at the bottom, the failure was usually installed on the way in, and the bad year just revealed it.

Which brings us to August 2026 - the story is running again.

Catastrophe bond issuance just set another all-time record.⁹ In a May survey of ILS investors, 60% said they intend to increase their allocation to this market within two years.¹⁰ The money is arriving on the strength of the 2023-25 record - and it’s arriving into a market that has been repricing downward for two straight years.

The multiple of expected loss paid by new catastrophe bonds averaged 4.5x in 2023, the peak year; last quarter it was 2.3x.¹¹ One research firm's modelled expectation for the market's 2026 return: about 6%.¹²

None of this is a prediction that prices fall from here. There are real reasons this softening may run shallower than past ones: the structural gains of 2023 have mostly held, and several of the largest reinsurers are shrinking their books and returning capital to shareholders rather than defending market share - evidence of discipline.

But the same logic that caused an investor to pull capital in 2019 - a three-year account statement overruling a repriced market - is behind many investors allocating today based on the 2023-25 numbers - just with the sign flipped.

While record capital chases the fast clock's visible three-year record into falling prices, casualty - where the bad years are surfacing and pricing is tightening in response - generates headlines about reserve holes. Money is moving away from wherever losses are showing, toward wherever gains are showing.

The standing invitation holds. If you’ve been sitting on a question - about a product, a structure, an idea - send it over.

Until then.

Charlie Evans, CAIA
Chief Investment Officer, Zephyros Investments

¹ Swiss Re Institute, sigma 1/2018 - record 2017 insured catastrophe losses of $144bn; Harvey, Irma and Maria ~$92bn combined. https://www.artemis.bm/news/insured-disaster-losses-hit-record-144bn-in-2017-swiss-re/

² Guy Carpenter Global Property Catastrophe Rate-on-Line Index, 1 January 2018. https://www.artemis.bm/news/global-property-catastrophe-rate-on-line-index-up-6-1-guy-carpenter/

³ Aon - global reinsurer capital declined 15%, or $100bn, in 2022, driven by unrealised investment losses. https://www.reinsurancene.ws/global-reinsurer-capital-declined-100bn-yoy-in-2022-aon/

⁴ Guy Carpenter index +27.5% at 1 January 2023, the largest rise since 2006 (+36.6%). https://www.reinsurancene.ws/global-property-cat-rates-on-line-up-27-5-at-jan-1-renewals-guy-carpenter/

⁵ Swiss Re Global Cat Bond Performance Index - the series Issue 5 cited; 2025 completed the first-ever three consecutive double-digit years. https://www.artemis.bm/news/swiss-re-global-cat-bond-performance-index-returns-11-40-for-2025/

⁶ PGGM/PFZW ILS portfolio returns: -13.3% (2017) through -0.9% (2022), then +18.9% in USD terms (16.5% in euros) in 2023 and +15.1% (2024). https://www.artemis.bm/news/pggm-pfzw-ils-portfolio-delivered-18-9-return-usd-in-2023/ · https://www.artemis.bm/news/pggm-pfzw-ils-portfolio-returns-15-1-in-2024-aum-flat-in-q4-at-us-8-68bn/

⁷ Eurekahedge ILS Advisers Index, average of constituent ILS funds, 2017 and 2018. https://www.opalesque.com/671884/Eurekahedge_ILS_Advisers_Index_ended_2018_down188.html

⁸ Aon - alternative capital flat at $93bn in Q3 2019 with ~$15bn trapped; AM Best - ~20% of 2018 capacity trapped. https://www.artemis.bm/news/alternative-capital-flat-in-q3-2019-15bn-still-trapped-growth-to-return-aon/ · https://www.artemis.bm/news/around-20-of-2018-ils-capacity-remains-trapped-a-m-best/

⁹ Artemis - record H1 2026 catastrophe bond issuance of $17.98bn; record $65.6bn outstanding. https://www.artemis.bm/news/catastrophe-bond-market-records-that-were-broken-in-h1-2026/

¹⁰ Gallagher Securities institutional investor survey, May 2026. https://www.artemis.bm/news/60-of-institutional-investors-intend-to-increase-ils-allocations-gallagher-securities-survey/

¹¹ Artemis - average cat bond multiple of expected loss 4.54x across 2023; 2.29x in Q2 2026. https://www.artemis.bm/news/cat-bond-investors-secured-higher-risk-adjusted-returns-in-2023-as-multiple-rises/ · https://www.artemis.bm/wp-content/uploads/2026/07/catastrophe-bond-ils-market-report-q2-2026.pdf

¹² Lane Financial, via Artemis - modelled expected cat bond market return of ~6% for 2026. https://www.artemis.bm/news/catastrophe-bond-market-expected-total-return-just-6-for-2026-lane-financial/