Property Cat ROL Falls 16%: Repricing the Net Cost of Reinsurance

Global property catastrophe rate-on-line fell 16% at the July 1, 2026 renewals, with APAC down 19% and Guy Carpenter recording the steepest annual decline since the late 1990s (Guy Carpenter, July 2026). That headline is not the number that belongs in a primary insurer's rate filing. The net cost of reinsurance, ceded premium plus expected reinstatement premium minus expected ceded recovery, moves by a materially different amount, and the gap is where the filing earns its credibility.

The net cost of reinsurance and the retained catastrophe load are two different provisions, and pricing actuaries who treat them interchangeably misstate both. The retained load is what the insurer keeps: modeled losses below the attachment point, plus any coinsurance share above it. The net cost of reinsurance is the frictional charge for transferring everything else, what the cedent pays the reinsurance market above and beyond what the market is expected to pay back. Apply a single blended percentage, drawn from the headline rate-on-line index, to either provision without re-deriving the underlying loss expectation, and the filed number will be wrong in a direction that invites a data call.

Building the Provision, Layer by Layer

For a single reinsurance layer, ceded premium equals rate-on-line multiplied by limit. Expected ceded recovery is the layer's modeled average annual loss: the integral of the occurrence loss distribution between the attachment and exhaustion points, adjusted for the frequency of multiple events and any reinstated limit within the term. Expected reinstatement premium is the share of that recovery expected to trigger a reinstatement clause, priced at the reinstatement's stated percentage of the original rate. Net cost of reinsurance, for that layer, equals ceded premium plus expected reinstatement premium minus expected ceded recovery. It is a provision for the cost of transferring tail volatility, not a provision for the tail losses themselves, and the two should never be netted against each other before they reach the rate indication.

A Worked Layer: Why a 2-Point ROL Cut Does Not Buy a 2-Point Net-Cost Cut

Take a $50 million layer attaching at $50 million, with a modeled average annual loss to the layer of $3.0 million, a loss-on-line of 6%. At an expiring rate-on-line of 12%, ceded premium is $6.0 million and the modeled ceded loss ratio, expected recovery divided by ceded premium, is 50%. If the renewal rate-on-line falls to 10%, in line with this cycle's softening, ceded premium drops to $5.0 million, a 16.7% cut that matches the market average almost exactly. Expected recovery does not move: the loss distribution above the attachment is unchanged by a lower price for the same limit. Net cost of reinsurance for the layer, before reinstatement, falls from $3.0 million ($6.0M minus $3.0M) to $2.0 million ($5.0M minus $3.0M), a 33% decline, roughly double the rate cut.

The amplification is systematic, not an artifact of this example. For a layer where expected loss is a fixed share r of ceded premium (the modeled ceded loss ratio), net cost equals premium times (1 minus r), and a proportional change d in rate-on-line moves net cost by d divided by (1 minus r). At r = 50%, a 16.7% rate cut becomes a 33% net-cost cut. At r = 10%, typical of a thin, high-excess layer far above the attachment, the same rate cut moves net cost by roughly 18.5%, close to the headline. The practical rule for a pricing actuary building a filing memorandum: the lower and more loss-exposed the layer, the more its net-cost provision overshoots the market's own rate-on-line index, in either direction. Applying one blended percentage across an entire tower will misstate every layer except the one whose loss-on-line happens to match the program average.

Reinstatement Premium and the RPP Overlay

Expected reinstatement premium compounds the effect for lower layers. Most property cat treaties reinstate at 100% of the original rate, pro rata as to the amount of limit exhausted, so the expected reinstatement cost is approximately rate-on-line multiplied by expected loss to the layer, since a partial fill triggers a proportional charge. In the example above, expected reinstatement premium falls from $360,000 (12% of $3.0M) to $300,000 (10% of $3.0M) at renewal, a smaller percentage decline than the base premium because it scales with the unchanged loss expectation rather than the rate-on-line. Folding that in, net cost for the layer moves from $3.36 million to $2.30 million, a 31.5% decline. Reinstatement premium protection, a rider that converts the stochastic reinstatement cost into a fixed, pre-paid premium, removes that variable from the net-cost calculation but adds its own line item. Reinsurers were more flexible on structure this cycle, including reinstatement terms, according to Gallagher Re's July 2026 renewal report; primary buyers of low, loss-exposed layers should price any RPP purchase explicitly rather than folding it into a blended treaty average.

Threading the Provision Into the Rate, Without Double-Counting the Capital Relief

The net cost of reinsurance enters the direct rate indication in the numerator: indicated rate level equals the loss and loss adjustment expense ratio, including the net-of-reinsurance catastrophe provision, divided by one minus the expense ratio minus the profit and contingency provision. A lower net cost of reinsurance reduces that numerator directly. It can also justify a lower profit and contingency provision in the denominator, because ceding tail volatility to a well-capitalized reinsurance market reduces the capital the insurer must hold against the retained book. The Actuarial Standards Board's revised ASOP No. 30, retitled "Profit Provisions, Contingency Provisions, and the Cost of Capital in Property/Casualty Risk Transfer and Risk Retention" and open for comment through July 1, 2026 (ASB, December 2025), formalizes exactly this link between retained risk and the capital charge built into rates. The trap is applying both effects to the same dollar of ceded volatility: lowering the numerator for the reinsurance cost saving and separately lowering the profit provision for reduced retained risk, when the retention has not actually moved because the insurer chose to keep the savings as premium relief rather than buy down its attachment point.

The Retention Decision: Buy Down the Attachment or Bank the Savings

Every dollar of reinsurance-cost relief this cycle goes to one of two places: it reduces the ceded premium line at an unchanged attachment point, or it funds a lower attachment point at roughly the prior premium spend, retaining less volatility for the same reinsurance budget. Aon's midyear 2026 report puts global reinsurer capital at a record $790 billion as of March 31, 2026, with double-digit demand growth absorbed without pricing tension and average reinsurer first-quarter return on equity at 14.1%, comfortably above cost of equity (Aon, July 2026). "A stable, well-capitalised and competitive reinsurance market provides insurers with an opportunity to align capital more closely with their risk strategies," said George Attard, Aon's Chief Strategy Officer for Reinsurance (Aon, July 2026). Buying down the attachment genuinely reduces retained volatility and can support a lower ASOP No. 30 risk margin. Banking the savings as pure premium relief leaves the retained book exactly as volatile as it was in 2025, and any accompanying reduction in the profit provision is unsupported by the underlying risk. The filing memorandum should state which choice the company made, and the capital relief claimed in the profit provision should match it.

What a Filed Net-Cost-of-Reinsurance Provision Has to Survive at the DOI

State insurance departments test a declining catastrophe provision against the standard reasonableness question: not excessive, not inadequate, not unfairly discriminatory. A net-cost-of-reinsurance provision invites two specific lines of inquiry. The first is timing: whether the filed provision reflects the actual July 2026 renewal terms or a stale, higher-cost placement carried over from an earlier treaty year, a lag that would leave the filed rate overstated relative to the carrier's current cost of risk transfer. Benign first-half catastrophe activity supports the case for a genuine cost decline: Gallagher Re put global insured catastrophe losses at $46 billion for the first half of 2026, 28% below the 10-year average of $64 billion and the lowest first-half total since 2018 (Gallagher Re, July 2026). The second line of inquiry, sharper for groups with a captive or affiliated reinsurer in the program, is arm's-length pricing: regulators reviewing ceded-to-affiliate arrangements want documentation that the ceded rate-on-line and reinstatement terms mirror what an unaffiliated market would charge for the same layer, not a rate set to move margin off the admitted entity's books. Citing the current Guy Carpenter and Aon renewal benchmarks directly in the filing memorandum, layer by layer, is the most direct way to demonstrate that an affiliated or third-party cession was priced at the market's own July 2026 terms rather than an assumption convenient to the indicated rate.

Insurers that document the layer math, not just the headline percentage, are the ones whose catastrophe provision survives review this cycle.

Sources