AM Best: Big Four Kept Cat Appetite by Raising Attachments and Exiting Aggregate Covers

AM Best credits Europe's Big Four with holding their property catastrophe appetite through a softening market, but the report attributes that appetite to structure rather than price: right-sized portfolios, higher attachment points, and a move away from aggregate covers and working layers (Best's Market Segment Report, August 20, 2026). Hannover Re grew its June and July book 12.3% while risk-adjusted prices fell 4.5%.

Volume held up because the layers being sold sit further from the ground. Whatever falls below them belongs to the cedent, and in the first half of 2026 that included $26 billion of severe convective storm loss out of $46 billion of insured catastrophe losses worldwide (Gallagher Re, July 2026).

Key Takeaways

• 12.3% volume growth against a 4.5% risk-adjusted price decline at Hannover Re's June and July renewals is the clearest sign the Big Four bought growth with structure rather than price, after several years of raising attachment points across the book.
• $26 billion of the $46 billion in first-half insured catastrophe losses came from severe convective storm, the peril class whose individual events sit below the per-occurrence attachments reinsurers have spent three renewal cycles pushing upward.
• A 9.1% volume cut at Munich Re, against 12.3% growth at Hannover Re and 11% at Swiss Re, shows the four disagree sharply on how much price they will trade. None of them gave back the attachment point.
• Only about 36% of the group's natural catastrophe budget was consumed in the first half, and Hannover Re booked EUR 784.7 million of large-loss expenditure against a EUR 1,024.6 million budget. Cedents publish no equivalent cushion.
• Risk-adjusted reductions of up to 25% at 1 June, against 14.7% at 1 January, are what cedents spent on the higher retention: a saving booked in one accident year against frequency that arrives across many.

What AM Best Counted as Appetite

The report covers Swiss Re, Munich Re, Hannover Re and SCOR, and its headline finding is continuity through a second consecutive softening season. "The European 'Big Four' are maintaining their ambitious profit targets for 2026, in spite of a softening of rates and the increase in uncertainties tied to the geopolitical environment," said Morgane Hillebrandt, associate director at AM Best (AM Best, August 20, 2026).

What sits under that continuity is a book that has already been rebuilt. AM Best ties the surviving property catastrophe appetite to three completed changes: right-sizing of portfolios, increases in attachment points, and a retreat from aggregate covers and working layers. Those are the three levers that determine which losses reach a reinsurance treaty at all.

The volume responses underneath the single appetite statement diverge more than the headline suggests.

Reinsurer	Mid-year 2026 volume	Price change on renewed business
Hannover Re	+12.3% (1 June and 1 July)	-4.5%, inflation and risk adjusted
Swiss Re	+11% premium volume	Not stated on a comparable risk-adjusted basis
Munich Re	-9.1% (July)	-5.5%, risk adjusted
SCOR	+1.3% traditional reinsurance EGPI	-2.4% gross

Sources: AM Best market segment report as reported by Reinsurance News, August 2026; Hannover Re half-year release, August 2026.

Price moved a great deal around them. Guy Carpenter's global property catastrophe rate-on-line index fell roughly 16% across the 2026 renewals, the steepest annual decline since the late 1990s, though the index still sits about 32% above its 2017 soft-market floor (Guy Carpenter index via Artemis, July 2026). Howden Re measured risk-adjusted reductions of up to 25% at 1 June, against 14.7% at 1 January. The attachment point did not move with it.

How a Higher Attachment Redistributes Frequency

Raising an attachment point is not a price term. It transfers a slice of the loss distribution back to the cedent, and that slice is the highest-frequency, lowest-severity part of the program. Expected ceded loss falls modestly. Retained variance rises, because the events now retained are the ones that recur several times in a season.

Aggregate covers were the instrument that capped that accumulation. Working layers were the instrument that absorbed the individual mid-size event. Withdrawing both at once raises the mean of the cedent's retained annual aggregate and, more consequentially, fattens its body at the return periods that govern quarterly earnings rather than the one-in-200 tail that governs capital. A programme rebuilt this way can leave solvency coverage untouched while materially degrading the stability of reported results.

The first half showed what that looks like in practice. Only 11 events exceeded $1 billion of insured loss against a ten-year average of 16, and severe convective storm supplied more than half the total (Gallagher Re, July 2026). Goldman Sachs read the quarter the same way: "We expect primary insurers will bear the brunt of this quarter's loss events, as the size of each event is not overly material, and SCS losses are historically retained by primary insurers" (Goldman Sachs via Reinsurance News, July 2026).

The pricing consequence is a classification problem before it is a rate problem. Many cedents set a catastrophe-designation threshold above the size of the individual events they now retain, so those losses land in the attritional pick rather than the catastrophe load. A book whose attritional loss ratio drifts upward because its reinsurance attaches higher will produce a rate indication that attributes the deterioration to underlying frequency or severity trend.

The correction is a structure-on-level adjustment to the historical net experience: restating prior accident years on the current retention, the net-side analogue of on-levelling premium. Where the aggregate has been dropped, that restatement also has to strip out the recoveries it would have generated in the loss-heavy years, which is where most of the distortion sits. A three-year net loss ratio built on years when an aggregate was in force understates what the cedent now carries.

Munich Re cut mid-year volume 9.1% on a 5.5% risk-adjusted price decline while Hannover Re and Swiss Re grew 12.3% and 11%. The disagreement is entirely about how much price to trade for share. On where the cover attaches, there was no disagreement, and Moody's forecast the same in January: "Reinsurers are likely to stand firm on high attachment points for insurers' excess of loss (XoL) treaties for 2026 renewals," adding that insurers "will continue to retain a large proportion of losses from secondary perils and will remain exposed to earnings volatility" (Moody's Ratings via Artemis, January 2026).

What a $46 Billion Half Conceals

The retained-frequency position has not yet been tested. The $46 billion first-half total ran 28% below the $64 billion ten-year average and was the lowest first half since 2018, closing a fifth consecutive quarter with no single insured loss above $10 billion (Gallagher Re, July 2026). The structure that pushes frequency to cedents has been running through the quietest stretch in the modern record.

The reinsurers' cushion is visible in their accounts. Hannover Re booked EUR 784.7 million of large-loss expenditure against a EUR 1,024.6 million half-year budget and reported an 83.2% property and casualty combined ratio, improved from 88.4% (Hannover Re, August 2026). Across the four, roughly 36% of the aggregate catastrophe budget was consumed (Fitch Ratings via Insurance Business, August 2026). A reinsurer booking a full budget against a fraction of actual losses carries the difference forward. A cedent absorbing the same quiet half inside its attritional pick books it as experience, then trends off it.

Cedents funded the higher retention out of ceded premium savings, and the two sides of that trade sit in different time frames. A 16% index decline reduces the current accident year's ceded cost. The retained frequency arrives whenever the weather does. It is a one-year cash saving set against a multi-year loss distribution, and the saving is spent long before the distribution is observed.

A counterparty split is opening underneath the structure. Howden Re reported that at 1 June reinsurer appetite for aggregate features, prepaid reinstatements, second-event covers and top-and-drop combinations broadened, in some cases reversing several years of constrained supply. AM Best says the Big Four moved the other way. A cedent restoring aggregate protection in 2026 buys it from a different set of counterparties than the one writing its top layers, with different ratings and a supply that reopened because returns are compressing.

Howden Re's own read is that another round of reductions on this scale would push large segments of industry returns below their cost of capital by 2027. Reinsurers can answer that at the next renewal by repricing or by withdrawing the structures again. A cedent that has already spent its 2026 savings on a higher retention cannot re-cede a season of frequency it has run.

Further Reading on actuary.info

• Fitch: Big Four Reinsurers' Record 21.5% ROE Masks Price Cuts Still to Earn Through - the earnings side of the same four companies, and where the 36% cat-budget figure comes from.
• Cheaper Reinsurance Puts P&C Pricing Actuaries in a Bind - what a falling ceded cost does to a primary rate indication.
• AM Best: Record Capital Tests Reinsurance's 87% Combined Ratio - the wider AM Best series this segment report sits inside.
• Property Cat ROL Falls 16%: Repricing the Net Cost of Reinsurance - the ceded-cost provision a cedent rewrites when it buys a higher attachment.
• Travelers Locks Cat Retentions as Reinsurance Softens - one large cedent holding its retention flat instead of spending the savings.
• Verisk: Property Claim Severity and Frequency Diverge - the primary-side frequency signal cedents now retain in full.
• Bermuda Reinsurers' 85.3% Combined Ratio Rests on a 9% Premium Decline - the Bermuda market's version of trading volume against price.
• Mexico's M7.3 Near Miss and the Basis Risk in Parametric Cat Bonds - the other structure where a cedent keeps loss it thought it had ceded.