The share of primary public company D&O renewals taking a price cut has nearly halved in three quarters, falling from 43% in the fourth quarter of 2025 to 25% in the quarter ended July 31, 2026 (Aon, August 2026). The share taking an increase moved only from 19% to 28%, and the median renewal never left zero. The turn in this line came from the decrease bucket draining, not from underwriters winning increases.

Key Takeaways

  • Up 0.6% year on year on primary policies renewing at the same limit and the same deductible, which Aon identifies as the first primary increase in sixteen straight quarters. Median and mode both held at zero.
  • Down 1.9% on cost per $1 million of limits in that same release. The blended series and the like-for-like series have now disagreed on direction for three consecutive quarters, twice in the other direction.
  • 85% of primary policies renewed at both the same limit and the same deductible, and 96% with the same carrier, so the constant-structure figure describes most of the book rather than a self-selected remnant.
  • 75% to 78%: TransRe's projected commercial D&O loss ratio for 2026 against a roughly 70% breakeven after acquisition and internal expense. A 0.6% rate turn does not close a gap that size.

What the 1.9% Decline Is Measuring

Aon's release carries two separate price series, and the desk consequence turns entirely on which one feeds the rate monitor.

The first is the average cost of $1 million in limits, down 1.9% against the prior-year quarter. That is a unit cost computed across renewals as they were actually placed, so it absorbs every change in what buyers purchased: additional limit, restructured towers, movement in attachment point and deductible. It answers what clients spent per unit of cover, which is a procurement question rather than a rate question.

The second series holds structure fixed. On primary policies renewing at the same limit and the same deductible, the average price change was up 0.6%, with median and mode both at zero. Programs renewing at the same total limit were up 0.5%. Aon flagged the primary figure as the first increase in sixteen straight quarters, which places the start of the decline in the second quarter of 2022.

The constant-structure cohort is not a fringe of the book. Aon reports that 98% of primary policies renewed at the same limit, 87% at the same deductible, 85% at both, and 96% with the same carrier. That disposes of the usual objection to like-for-like series, which is that the accounts holding structure constant are the ones with nowhere else to go. Here they are the market, and the roughly two and a half points separating the two figures need an explanation that is not base rate.

The Turn Came From Shallower Cuts

Reading the three most recent prints together is what exposes the blended series. It has disagreed with the like-for-like primary series in both directions, which is difficult to reconcile with any story in which it tracks rate.

Aon printIndex levelCost per $1M vs prior yearPrimary, same limit and deductibleDecrease / flat / increase
Q4 20251.27+0.8%-1.9%43% / 38% / 19%
Q1 20261.24+10.7%-0.2%27% / 42% / 31%
Q2 20261.04-1.9%+0.6%25% / 47% / 28%

In the fourth quarter of 2025 the blended series was positive while primary was down 1.9%. In the first quarter of 2026 it printed +10.7% while primary was down 0.2%. In the second quarter both signs flipped. Across those two quarters the blended measure travelled 12.6 points and the like-for-like measure travelled 0.8. Base rate on a mature financial line does not move 12 points in six months, and the index level dropping from 1.24 to 1.04 inside the same calendar year says the same thing.

The renewal distribution shows how the primary mean crossed zero. The decrease bucket fell from 43% to 27% to 25%. The increase bucket peaked at 31% in the first quarter and then receded to 28%. Accounts renewing flat rose each quarter, from 38% to 42% to 47%. The mean turned positive because the remaining cuts got smaller, not because more accounts were pushed up.

That distinction limits what the number can be asked to do. A mean driven by a thinning left tail reverses if a handful of large accounts win double-digit reductions next quarter, and it carries no information about an underwriter's ability to hold an increase once quoted. TransRe noted that accounts renewing down in 2025 did so at an average of -10%, and that too few had achieved increases since the end of 2021.

The excess layers are the residual, and the disclosed figures bracket them. Aon's constant-total-limit change of +0.5% sits below the constant-structure primary change of +0.6%, so inside towers that did not change size the layers above primary were still dragging the program average down. Broker commentary describes the behaviour directly: capacity remains abundant in excess, and buyers have used improved pricing to rebuild towers cut back during the hard market (PropertyCasualty360, April 2026).

A second monitor points the same way on primary. Marsh's Q2 2026 index put global financial and professional lines down 3%, with the US the only region in positive territory at +1%, a print the site read in detail in its line-by-line coverage of that index. For on-leveling, the choice between the two series is the whole exercise: substituting the blended cost-per-million series into a chain carrying sixteen quarters of compounded decreases introduces a bias that compounds in the wrong direction in every quarter where the two disagreed on sign, and three of the last three qualify.

A 0.6% Turn Against a 75% Loss Ratio

The turn is real and it is arithmetically trivial against the loss side. TransRe assumes the line breaks even near a 70% loss ratio after 15% acquisition costs and 15% internal expense, and projects commercial D&O loss ratios reaching 75% to 78% by 2026. Recovering five to eight points of loss ratio at 0.6% a year is not a trajectory that closes.

The loss side is not waiting. Cornerstone Research counted 121 securities class action filings in the first half of 2026, up 30% on the prior six months, with its Disclosure Dollar Loss index at $529 billion, a 77% increase. Technology filings went from 9 to 24. "AI-related cases represented a modest share of total filings but an outsized share of alleged investor losses," said Joseph Grundfest, the Stanford law professor and former SEC commissioner, in the same release, which put AI-related DDL at $385 billion.

Settled severity is moving with it. Cornerstone recorded 39 settlements totalling $2.2 billion in the half, an average of $56.4 million against $42.2 million for full-year 2025, with the median at $20 million against $17.6 million (The D&O Diary, July 2026). Defence cost inflation compounds the indemnity trend: TransRe puts the cumulative rise in legal fees since 2020 at 44%, which falls hardest on the excess layers where defence erodes underlying limit before indemnity attaches.

Sector concentration sharpens it further, with healthcare companies drawing a median settlement of $32.2 million against a four-year overall median of $15.48 million (The Baldwin Group, May 2026). The asymmetry lands hardest on excess and Side A writers, and TransRe reports mega cap and large cap prices sitting 13% and 17% below their peak respectively, which are the accounts carrying the deepest towers.

The reserving position is the uncomfortable part. TransRe's observation that Schedule P "2023 is already ringing alarm bells" describes accident years written in the middle of a sixteen-quarter rate decline, against a filing count and a settlement median that have accelerated since those years were priced. One positive primary quarter does nothing for those vintages, and the rate history attached to them is the blended history unless someone rebuilt it.

Further Reading

Sources