Global reinsurance capital closed March 31, 2026 at a record $790 billion, property catastrophe rates on line fell 16% through the July 1 renewal, and global reinsurance demand rose more than 10%. For a cedant actuary those are not buyer's-market talking points. They are inputs to an optimal retention recalculation, and most programs are still carrying structures whose attachment points were set by hard market price rather than by risk appetite.
Key Takeaways
- A 16% ROL decline cuts layer cost by the same proportion while expected recovery is unchanged, which moves marginal layers across the economic threshold rather than simply making the tower cheaper.
- North America property cat fell 20% to 25% at mid-year, steeper than the 16% global composite, so US towers repriced hardest in the hard market gain most.
- Aggregate covers and earnings protection are being written again at attachment levels where they function, after being effectively unavailable through the hard market.
- Alternative capital stood at $141 billion, up 4% from year-end, and committed multi-year ILS structures do not withdraw at a renewal the way traditional capacity does.
- Lowering the attachment moves the reserve, not just the premium. Development patterns fitted to the old structure overstate net IBNR at the new one.
The NPV Math Behind a 16% Decline
Through 2022 and 2023 the governing question in retention analysis was what minimum retention kept the cat program economically defensible, given what protection cost per unit of expected recovery. Many cedants placed attachment points well above where appetite would have put them, because buying down meant paying rates on line above the present value of the protection.
A 16% decline changes that arithmetic precisely. Layer cost falls by the same proportion on a present-value basis while the expected losses in the layer are unchanged. A layer at 4.0% ROL in 2023, $40 million a year on $1 billion of occurrence limit, costs about $33.6 million now. If the cat model put annual expected recovery in that layer at $30 million, the expected loss ratio against the old cost was 75%, just under a typical break-even. Against $33.6 million it is 89%.
| Illustrative Layer ($1B Limit) | 2023 Hard Market ROL | Current ROL (approx.) | Prior ELR vs. $30M Expected | Current ELR vs. $30M Expected |
|---|---|---|---|---|
| Cat Layer A | 4.0% ($40M) | 3.36% ($33.6M) | 75% (uneconomical) | 89% (near break-even) |
| Cat Layer B | 3.0% ($30M) | 2.52% ($25.2M) | 100% (break-even) | 119% (clearly economic) |
| Cat Layer C | 2.0% ($20M) | 1.68% ($16.8M) | 150% (already economic) | 179% (strongly economic) |
That is a threshold effect rather than a saving. Any marginal layer sitting between roughly 75% and 90% expected loss ratio against prior pricing flips from decline to buy, and the effect compounds across a multi-layer tower. Holding expected losses constant and refreshing only the ROL inputs moves the efficient frontier materially toward more protection at any given capital target. North America fell 20% to 25% at mid-year against the 16% global composite, so the shift is largest for the cedants whose towers were repriced hardest.
A cedant running its optimization against last year's ROL curve is generating structures for 2023 market conditions.
Retention Is a Risk Appetite Question Again
The recurring error at a soft market transition is to treat the decline as a saving on a fixed structure rather than as the removal of a constraint. Risk appetite lives in the ERM framework, in board-stated earnings volatility tolerance, and in regulatory capital requirements. What the market charged in 2023 was in none of those. It was a binding constraint, and it has loosened.
The exercise is specific. Take the current ROL curve by layer, run expected recovery per layer through the portfolio optimization, and find the attachment where expected recovery per ROL dollar meets the internal hurdle rate. Compare that to current net retention. Where they diverge, the answer is restructuring rather than repricing the existing tower.
Two mid-year developments widen the solution space beyond price. Reinsurers became willing to write aggregate covers and earnings protection, which attach across the frequency distribution rather than against single events and cannot be replicated by a per-occurrence tower. Josh Knapp of Gallagher Re noted aggregate protection is "making a resurgence, especially for US cedants, when the program is appropriately structured and has a practical attachment point for everyone involved." The qualifier carries the weight: an aggregate at an unrealistic attachment supplies basis risk without recovery, and what changed is that these are now written where they are operationally useful.
Parametric supplements sit in the same reopened space, and the case is strongest for secondary perils where cat model parameter uncertainty is highest. Cat bond issuance reached almost $18 billion in the first half of 2026, with $11.3 billion in the second quarter alone, and alternative capital closed mid-year at $141 billion, up 4%, competing at parametric layers as well as traditional ones. A cedant that rejected a parametric severe convective storm supplement in 2023 on price is looking at a different number now.
The capital base behind all of it does not reprice in a quarter absent a major loss. Total dedicated reinsurance capital was $648 billion at year-end 2025, up 11%, reaching $790 billion with alternative capital included by the first quarter of 2026.
Moving the Attachment Moves the Reserve
Buying down has a consequence on the other side of the balance sheet that the pricing analysis does not surface. A lower net retention cedes a higher proportion of the attritional cat loss distribution, so net cat IBNR falls on an expected basis. The arithmetic is simple and the documentation is not.
The prior reserve analysis used a net cat expected loss and a recovery pattern calibrated to a specific attachment, and both change when the attachment moves. Carrying forward loss ratios and development patterns fitted to the old structure overstates net IBNR at the new one, because the frequency losses that used to develop in the retained column now develop in the ceded column. Net development patterns from historical data stop being a valid projection basis at the point the structure changes.
Counterparty exposure changes shape at the same time. Collectability assumptions calibrated to upper tower layers do not describe credit exposure at lower attachment layers, where the recoveries are more frequent and the counterparties may differ.
The discount is the piece most easily carried forward untouched. A lower net retention shrinks the pool of net IBNR earning investment income across the development period, and at current yields that offset is material on patterns running beyond two or three years. Holding the prior discount factor and applying it to a smaller retained reserve pool misstates the economics of the structure the cedant just bought.
Parametric layers need their own line in that decomposition rather than a share of the indemnity treatment. A parametric layer reduces net IBNR only to the extent triggers are expected to fire, and a layer that does not trigger on an actual event reduces nothing for that event whatever the expected frequency implied. The basis risk assumption has to be quantified and stated, and at the next year-end the net loss has to be split into triggered coverage, untriggered coverage lost to basis risk, and retained risk.
Further Reading
- Third-Party Reinsurance Capital Growth Halves to 6% in 2026
- Drought Puts $8.4B Through Crop Insurers' Reinsurance Math
- Reinsurance's $648B Capital Record Is Retained, Not New
- Agentic AI Faces Its First Real Test at the July 2026 Reinsurance Renewal
- Casualty Cedants Held Retentions Flat as Midyear XL Rates Fell 5 to 10 Percent
- Property Cat at -23% from Peak: Reinsurer ROE and the 2027 Cost-of-Capital Horizon
- Parametric Reinsurance for Secondary Perils: Basis Risk, RBC Credit, and the Actuarial Certification Gap
- Cat Bond H1 2026 Targets $17B as European Sponsors Reshape the ILS Market
- Cat Bonds Hit $18B in H1 2026: What the Records Actually Mean
- RenRe Lifts 2026 Reinsurance Demand Forecast 50% to $15B as Mid-Year Rates Fall
- Reinsurance Illiquidity: Why Record Capital Still Costs Too Much
- The 43-Point Gap: How Reinsurers Priced Casualty Portfolio Quality at the July 2026 Renewal
- London's 10-Day Cat Bond Push Tests Bermuda's 90% ILS Grip
- Everest Q2 2026: An 88.5% Treaty Combined Ratio in a Softening Reinsurance Market
- Markel's $205M State National Loss Exposes a Fronting Collateral Gap
Sources
- Aon, “Record $790bn Reinsurance Capital Underpins Softer Mid-Year Renewals,” Reinsurance News, July 2026
- Aon, Reinsurance Market Dynamics Midyear 2026 Renewal Report, Actuarial Post, July 2026
- Guy Carpenter, “Global Property Cat Rates Down 16% as Softening Extends into July Renewals,” Reinsurance News, July 2026
- Gallagher Re, “Record Capital Drives Softer Reinsurance Pricing at July Renewals,” Insurance Business, July 2026
- Gallagher Re, First View: Options and Opportunities, GallagherRe.com, July 2026
- Gallagher Re, “Property Aggregate Reinsurance Making a Resurgence,” Reinsurance News, 2026
- Artemis, “Cat Bond Market Shows High Bars Are Set to Be Broken, Records Fall Again in H1 2026,” Artemis.bm, July 2026
- Aon, “Alternative / ILS Capital Rises to $141bn, Drives Reinsurance Market Growth in Early 2026,” Artemis.bm, 2026
- The Insurer, “Parametric Structures Seen in Lower Layers and Retro at Mid-Year Reinsurance Renewals,” TheInsurer.com, June 2026