Commercial Property Gives Back 6.3% to Fund Auto and Umbrella

Commercial property premiums fell 6.3% in the second quarter of 2026, the steepest line-level move CIAB tracks, while commercial auto rose 4.5% and umbrella rose 5.3% for a 35th consecutive quarterly increase (CIAB, August 20, 2026). CIAB's own read is explicit: the property give-back is partly funding what auto and umbrella still require. A blended 2% average premium decline hides that divergence, and a portfolio managed to the average is being managed to the wrong number.

The Give-Back, Line by Line

The Council of Insurance Agents & Brokers' Q2 2026 Commercial Property/Casualty Market Survey found average premiums down 2% across all accounts, the sixth straight quarter of softening and, more notably, the second consecutive quarter in which every account size declined. Ten of the twelve lines CIAB tracks fell: commercial property, cyber, workers compensation, D&O, employment practices liability, business interruption, construction risks, flood, marine, and terrorism. Cyber and workers compensation each fell 3.2%. Only commercial auto and umbrella moved the other way, and CIAB attributed the property line's decline in part to carriers "seeking to offset increases they asked for in lines like commercial auto and umbrella" (CIAB, August 20, 2026). That is not incidental color. It is the survey's own account of the mechanism: property capacity is being priced to subsidize the two lines whose loss cost trajectories have not turned.

The pace on each side of that trade is also moving. Commercial auto rose 5.8% in the first quarter and decelerated to 4.5% in the second; umbrella rose 4.8% in the first quarter and accelerated to 5.3% in the second (Risk & Insurance, citing CIAB's Q1 2026 index). A pricing desk reading only the Q2 print would miss that auto is decelerating off a structurally negative loss position while umbrella is accelerating into one, two different trajectories that a single "casualty is up" label flattens.

Where the Room to Give Back Comes From

The property give-back is a capital story before it is a loss-cost story. Guy Carpenter's U.S. Property Catastrophe Rate-On-Line Index is down roughly 16% for 2026 through the mid-year renewals, the steepest annual fall since the late 1990s, on dedicated global reinsurance capital Guy Carpenter estimated at $663 billion in 2025, up 9% year over year, with alternative capital up 15% (Guy Carpenter via Artemis.bm, 2026). That capacity flows straight into primary appetite: 75% of CIAB's Q2 respondents reported increased property underwriting capacity, up from 72% in the first quarter. AM Best data shows the effect on the loss side too, with the commercial property calendar-year loss ratio improving to 85% at year-end 2025 from 87.9% at year-end 2024 (AM Best data via Risk & Insurance, May 2026). Exposure has not softened at the same rate. Global insured natural catastrophe losses ran $42 billion in the first half of 2026, well below the $66 billion long-run trend and the lowest first-half total since 2020 (Swiss Re Institute, 2026). Swiss Re's own framing of that result, "rising heat, growing exposure, changing hazards," pairs the benign number with a warning that modeled exposure keeps climbing even when realized losses do not. Giving back 6.3% of property rate against record dedicated capital is defensible underwriting in a benign-loss period; it is also the fastest way to erode adequacy the moment loss experience reverts, and that reversion risk sits with whichever accident quarter draws the next active storm season.

A Pattern That Shows Up Outside the US Retail Channel Too

The CIAB survey covers US retail broker placements, but the same property-versus-casualty split shows up in Marsh's placement data for the identical quarter, a separate index built on large-account global business. Marsh reported global commercial rates down 6% in the second quarter, the eighth straight quarterly decline, with global property down 12% and catastrophe-exposed programs above $1 million in premium down a full 20%, twice the rate cut on non-catastrophe programs under $1 million. Global casualty, the only major line still rising, was up just 2%, though US casualty ran hotter (Insurance Journal, citing Marsh's Q2 2026 Global Insurance Market Index, August 17, 2026). Marsh president John Donnelly said current conditions "likely persist absent severe storms or major catastrophes," a caveat that lands on exactly the reversion risk Swiss Re's exposure framing points at. actuary.info has covered the Marsh index's own property/casualty divergence and US ex-workers'-comp casualty print in a separate breakdown, and the two datasets corroborate rather than duplicate each other: CIAB shows the mechanism inside a US multiline retail book across all account sizes, while Marsh shows the same split holding across a global large-account placement channel. A cross-subsidy visible in two independently surveyed populations is a market condition, not a broker's editorial framing.

The Account-Size Split the Average Hides

CIAB's account-size breakdown adds a second layer the blended 2% conceals. Large accounts, those generating more than $100,000 in annual commissions and fees, fell 3.7% in the second quarter after a 2.7% decline in the first, the deepest and most consistent softening of any segment. Medium accounts held steady at a 1.9% decline in both quarters. Small accounts are the segment worth watching most closely: they rose 1.1% in the first quarter, still a hard-market print, then flipped to a 0.5% decline in the second (CIAB, Q1 and Q2 2026 surveys). That is a full swing from positive to negative rate change in a single quarter for the segment that had, until now, been the last holdout of rate adequacy. If small-account softening is starting rather than stabilizing, the "large accounts compete hardest, small accounts hold" assumption that has anchored account-tier plan loss ratios for the past two years needs to be re-tested with the Q3 print, not assumed to still hold.

What the Cross-Subsidy Does to the Book

For a multiline commercial pricing desk, the practical consequence is that the blended book rate change has stopped being a usable proxy for portfolio margin trajectory. A book running flat to slightly down on the headline number can simultaneously be losing adequacy on property, where capacity is being deployed faster than the exposure base is shrinking, and gaining adequacy on auto and umbrella, where structurally negative loss ratios are finally catching a rate response. Commercial auto's loss ratio has exceeded 100% every year since 2014 except 2021, with net underwriting losses above $5 billion in both 2023 and 2024 and calendar-year loss ratios of 109.2% and 107.2% respectively (AM Best data via actuary.info's coverage of the fourteenth straight unprofitable year). A 4.5% increase against that backdrop is rate catching up to loss cost, not rate running ahead of it, and treating it as comparable to the 6.3% property decline in a blended average mismeasures both lines' margin trajectory. Umbrella carries its own severity story layered on top: nuclear verdicts totaled 135 in 2024, a 52% increase over 2023, with awards exceeding $31 billion, and first-layer towers are now commonly capped at $5 million to $10 million as carriers ration capacity at the attachment points most exposed to that severity (IMA Financial Group, 2026). That capacity constraint, not just the loss trend, is why umbrella's rate increase has accelerated for 35 straight quarters even as most of the rest of the book softens: buyers needing higher limits are layering more carriers to fill towers that used to clear with fewer, and each added layer negotiates its own rate independent of the composite print. actuary.info has covered the ILF repricing mechanics behind that severity in a prior piece on nuclear-verdict-driven excess repricing and the reserve-side consequence in Hartford's Q2 excess and umbrella reserve strengthening. Neither dynamic reverses because property is soft; if anything, a soft property market gives brokers more room to negotiate concessions elsewhere in the account, which increases pressure on the auto and umbrella lines to hold their ground on price.

The Desk-Level Read Into Q3

The plan loss ratios set against a hardening-across-the-board assumption two years ago now need re-anchoring line by line, not at the portfolio level. A pricing team holding blended rate flat while property gives back 6.3% and umbrella takes 5.3% is not holding the book stable; it is letting the property indication drift toward inadequacy while the casualty indications finally close a gap open since 2014. The give-back deserves a harder look at how much of the quoted decrease reflects treaty capacity passed through by the carrier versus an actual change in that account's loss experience, a question worth putting directly to the broker or reinsurer at the next renewal rather than accepting the market movement at face value. CIAB's quarterly cadence, surveys released in February, May, and August so far this year, points to a Q3 release around mid-November, which will show whether the small-account flip to negative territory in Q2 was a one-quarter blip or the start of the softening the large-account segment has run for a year. Guy Carpenter's January 1, 2027 renewal outlook, typically published in the fourth quarter, will show whether the capacity funding this give-back keeps expanding or pulls back if the back half of 2026 runs above the benign pace Swiss Re reported for the first half. Until then, the safer planning assumption is that the property give-back is a capacity-driven concession, reversible on short notice, while the auto and umbrella increases are catching up to loss cost that has not moved.