Q2 2026 Casualty Reserves: Chubb and Travelers Release While Everest and CNA Build

Travelers converted $578 million of prior-year reserves into current profit in the second quarter of 2026, a 4.5-point combined-ratio benefit (Travelers, July 17, 2026). Chubb added $283 million more of its own (Chubb, July 21, 2026). In the same reporting season, Everest Group booked close to $200 million of North America casualty reserve strengthening and CNA Financial took a $77 million after-tax mass-tort charge, its second consecutive year doing so (CNA Financial, August 3, 2026).

Four large-cap carriers, reading the same 2021 through 2024 accident years, landed in opposite places within a three-week window. That split is not noise. It is the visible surface of a reserving argument that has been building since commercial casualty rates decoupled from the rest of the market: US casualty rates excluding workers' compensation rose 11% in the second quarter even as the global commercial composite fell 6%, the eighth straight quarterly decline (Marsh Global Insurance Market Index, July 2026). Pricing actuaries keep pushing rate on casualty because they do not believe the loss trend is settled. Reserving actuaries at four different companies just gave four different answers about whether it is.

The Releases: Chubb and Travelers Find Redundancy, Mostly Outside Casualty

Travelers' $578 million of net favorable prior-year development spread across all three segments: $319 million in Business Insurance, $75 million in Bond & Specialty, and $184 million in Personal Insurance. Within Business Insurance, workers' compensation contributed more than $200 million and commercial property roughly $80 million, and CFO Dan Frey told analysts the company saw no pressure in casualty lines such as umbrella or commercial auto (Travelers Q2 2026 earnings call, July 17, 2026). The Bond & Specialty release traced mainly to management-liability coverages within general liability rather than to primary commercial casualty. Travelers' consolidated combined ratio came in at 83.6%, improved 6.7 points year over year, with 2.5 points of that improvement attributable to the larger reserve release and another 3.6 points to lower catastrophe losses.

Chubb's $283 million pre-tax favorable development, up from $249 million a year earlier, pushed its consolidated P&C combined ratio to 83.8%. But the segment detail complicates the headline. North America Personal P&C's combined ratio improved 6.2 points, with 2.4 points of that coming from higher favorable prior-period development, and Overseas General Insurance improved 1.4 points on the same driver. North America Commercial P&C, the segment carrying the bulk of Chubb's casualty exposure, moved the other way: its combined ratio worsened to 85.4% from 83.5% a year earlier, driven partly by a higher current-accident-year loss ratio excluding catastrophes. Chubb's release does not break out prior-year development by segment in dollar terms, but the directional pattern is legible without it: the redundancy driving the consolidated number down is concentrated in personal lines and international, not in the North America commercial casualty book where rate has been running hardest.

That pattern, headline improvement funded by short-tail or personal-lines release while the commercial casualty segment quietly worsens, is the same mechanism actuary.info found in Hartford's Q2 2026 print, where a four-point sequential combined-ratio improvement masked $46 million of adverse general liability development and $26 million of adverse commercial auto liability development, both concentrated in the 2023 and 2024 accident years, offset by $110 million of favorable workers' compensation development through the first half (The Hartford, Q2 2026 earnings release, July 23, 2026). Three different carriers, three different consolidated combined ratios, and the same underlying seam: whatever is releasing is not commercial casualty, and whatever is commercial casualty is not releasing.

The Builds: Everest, CNA, and a Reserving Actuary Who Would Not Wait

Everest Group's version of the same seam came with a name attached. CFO Elias Habayeb told analysts the company "decided to react now to what we're seeing in the data and not wait till the reserve studies are done next quarter," describing the roughly $200 million of North America casualty treaty reserve strengthening booked in the quarter (Elias Habayeb, Everest Group Q2 2026 earnings call, July 30, 2026). The charge was driven by elevated loss trends and higher-than-expected loss emergence concentrated in older accident years, and management chose to reflect the signal across most accident years rather than isolate it, without changing the 2026 accident-year casualty loss picks themselves; the annual long-tail reserve review is scheduled to complete in the third quarter. Everest also absorbed a separate $55 million charge tied to the 2024 Baltimore bridge collapse in the same quarter (Everest Group, July 29, 2026).

The reserve action sits inside a broader retreat from the line. Everest has withdrawn from more than $1 billion of casualty business over the past several quarters as pricing, structure, or expected returns fell short of its bar, a deliberate underwriting decision rather than a shortage of available premium, management said on the same call (Coverager, July 2026). Even against that pullback, Everest still posted $585 million of operating income, a 14.9% annualized operating return on equity, a 90% core combined ratio, and book value per share of roughly $408, up 12% year over year, so the reserve charge and the volume retreat both read as a company acting from a position of margin, not distress.

CNA Financial's build is smaller in absolute terms but pointed at a narrower, more troubling source: legacy mass tort. The $77 million after-tax charge marks the second consecutive year CNA has taken a reserve action of similar size for the same exposure category, an unusual repeat that suggests the first charge understated the ultimate cost rather than resolved it. CNA's headline P&C combined ratio was 96.5%, worsened from 94.1% a year earlier, with the release citing 2.3 points of catastrophe impact and no net prior-period development impact companywide this quarter, meaning the mass-tort charge was offset elsewhere in the book rather than showing up as a standalone unfavorable print (CNA Financial, August 3, 2026). That offsetting is itself a version of the Chubb and Hartford pattern in miniature: a casualty-specific problem, absorbed inside a blended number that looks stable.

$578M

Travelers net favorable PYD, Q2 2026, mostly workers' comp and property

$283M

Chubb favorable PPD, while NA Commercial P&C combined ratio worsened to 85.4%

$200M

Everest North America casualty treaty reserve strengthening, older accident years

$77M

CNA after-tax mass-tort charge, second consecutive year of similar size

Same Accident Years, Different Conclusions

Four carriers looking at overlapping 2021 through 2024 casualty accident years should not, in principle, reach opposite conclusions about reserve adequacy from the same industry loss environment. That they did points to three mechanical differences rather than a genuine disagreement about the underlying tort trend.

The first is book mix and venue concentration. Everest's strengthening sits specifically in its North America casualty treaty book, reinsurance exposure that aggregates losses across many cedents and is therefore more sensitive to a handful of adverse jurisdictions or large-account claims than a single primary carrier's diversified commercial portfolio. CNA's charge is narrower still, tied to legacy mass tort rather than the current casualty book. Travelers' CFO could credibly say umbrella and commercial auto showed no pressure precisely because Travelers' mix and the specific accounts in its book had not yet produced the adverse signal that showed up in Everest's treaty layer or Hartford's general liability line.

The second is timing philosophy. Everest's management explicitly chose to act on emerging data mid-cycle rather than wait for its scheduled annual long-tail review, a decision that front-loads the correction into Q2 rather than deferring it to Q3. Most carriers run formal reserve studies on an annual or semiannual cadence, which means a company that happens to complete its long-tail review in a quarter when adverse signals are emerging will report a build, while a company whose scheduled review falls in a calmer quarter will not, even if both are tracking the same underlying loss emergence with a lag of only a few months.

The third is the redundancy funding source. Workers' compensation has been the industry's most reliable release since 2017, as claim frequency has run persistently below pre-pandemic baselines. A carrier with a large WC book, as Travelers and Hartford both have, can post a favorable headline number even while its casualty lines strengthen, because the WC release is large enough to more than offset the casualty charge in the consolidated figure. A carrier without a comparably sized WC cushion, or one where the casualty charge outruns the available redundancy, reports the build as a net unfavorable or a standalone item instead. Reserve adequacy across the group is less divergent than the headline combined ratios suggest; what is genuinely divergent is how much offsetting redundancy each company happens to be carrying elsewhere in the same quarter.

Carrier	Q2 2026 Headline CR	Consolidated PYD	Casualty-Specific Signal
Travelers	83.6%	$578M favorable	No pressure cited in umbrella/commercial auto
Chubb	83.8% (consolidated)	$283M favorable	NA Commercial P&C worsened to 85.4% from 83.5%
Hartford	91.4% (Business Insurance)	$111M favorable	$46M adverse GL, $26M adverse commercial auto
Everest	90% (core)	~$200M unfavorable (NA casualty)	Older accident years, reflected proactively
CNA	96.5%	No net impact companywide	$77M mass-tort charge, second year running

What the Split Means for 2026 Accident-Year Reserving

A soft market with a hard casualty exception is precisely the setup actuarial reserving literature associates with delayed recognition of adverse development. Loss development factors selected off recent, benign-looking triangles tend to understate ultimate losses when severity is being driven by a small share of large claims and litigation-funded verdicts rather than by claim frequency, because a moderate accident-year window can look stable right up until a handful of large settlements or verdicts land and reveal the true trend all at once. That is the mechanism behind Everest's and Hartford's Q2 disclosures, and it is also the mechanism a carrier like Travelers or Chubb is implicitly betting against when it reports no casualty pressure in the same quarter.

The practical read for a reserving actuary on a comparable commercial casualty book is that four large, sophisticated carriers examining overlapping accident years in the same six-week window is not confirmation that reserves are adequate industrywide; it is closer to a natural experiment showing how much a company's reported reserve position depends on book mix, review timing, and available offsetting redundancy rather than on a shared read of the underlying loss trend. A pricing actuary setting 2026 casualty loss-trend assumptions should treat the Marsh-reported 11% ex-comp rate increase as the market's own answer to that question: carriers are still pushing rate hardest on the lines where Everest, CNA, and Hartford just found adverse development, which is not the pricing posture of a market that considers the casualty trend settled. A reserving actuary at a carrier currently reporting favorable casualty development has a narrower and less comfortable task: showing why its book is different from the three that are not, rather than assuming the favorable read will hold simply because the current quarter's number came in green.

What to Watch Next

AIG reports second-quarter 2026 results on August 6, and the metric worth isolating is not the consolidated combined ratio but the segment-level development detail once the 10-Q publishes: whether favorable development, if any, sits in short-tail lines like US Property while North America Commercial casualty moves the other way, the same split now visible at Chubb. Everest's own annual long-tail reserve review completes in the third quarter and will show whether the roughly $200 million Q2 action was sufficient or a down payment. CNA's repeat mass-tort charge, two years running, is the one item on this list that reads less like a market-cycle signal and more like a legacy liability that has not yet been fully reserved; a third consecutive charge in 2027 would be difficult to describe as anything else. For every carrier in this group, the number that matters is not the headline combined ratio investors trade on, but the accident-year-level development table buried in the 10-Q, where the actual direction of casualty reserve adequacy, as opposed to its consolidated appearance, is the only place it can be seen.