The Stop-Loss Turnaround Is Rate, Not Trend: A Q2 2026 Carrier Scorecard

The question hanging over the medical stop-loss market since late 2024 has been simple: when does the repricing catch the claims? With second-quarter 2026 results now filed by every major publicly traded writer, the answer is finally legible in the disclosures rather than in broker commentary. The turnaround is real, it is being bought entirely with rate, and it is not finished. Voya's reported stop-loss loss ratio has fallen from a 115% peak in Q4 2024 to 85.4% this quarter, Sun Life says its book never left its mid-70s target range and is growing sales 86%, and Cigna told investors its margin-recapture program on the industry's largest book is on schedule (Voya Q2 2026 investor supplement, August 4, 2026; Sun Life Q2 2026 earnings call, August 7, 2026; Cigna Q2 2026 earnings call, July 30, 2026). Set against that, QBE booked an onerous-contract provision on its accident and health book after claims inflation ran roughly 30% against the 20% it had priced, and the underlying medical trend that created the problem is still accelerating: PwC's Health Research Institute now projects 2027 group medical cost trend at 9.0%, the highest in 17 years (QBE H1 2026 earnings call, August 13, 2026; PwC HRI, June 2026).

The Book That Broke

Medical stop-loss direct premium reached roughly $39.9 billion in 2024, up 12.4% year over year and more than double the 2017 market of $16.9 billion, according to the Oliver Wyman and Guy Carpenter statutory-data analysis that serves as the segment's de facto annual report (Oliver Wyman/Guy Carpenter, Fall 2025). Nearly all of that growth was price and claims mix rather than lives: premium per member per month rose from $33.63 in 2019 to $54.07 in 2024 while enrollment grew only 0.3% per year. The same report puts the market-wide 2024 gross loss ratio at 86.0%, the worst on record in its series, up from 80.0% in 2023.

The damage was not evenly distributed. The large national medical carriers (Cigna, UnitedHealth, CVS/Aetna) ran a 95.6% loss ratio on their combined $14.0 billion of 2024 stop-loss premium, while the large direct employer-stop-loss writers (Sun Life, Tokio Marine HCC, Voya, Symetra) ran 80.5% and the specialty and life cohort ran 75.2% (Oliver Wyman/Guy Carpenter, Fall 2025). That 15-point gap between integrated and independent books is the single most useful sorting variable for reading the 2026 results: the carriers repricing hardest today are, for the most part, the ones whose 2024 books deteriorated furthest past their pricing.

What broke the books was severity above the attachment point, not underlying utilization. The Oliver Wyman/Guy Carpenter claims data shows the frequency of claims above $2 million nearly doubled from 7.8 to 15.3 per million members between 2019 and 2023, and claims above $5 million tripled. Leveraged trend, the amplified inflation rate experienced by a fixed attachment point when underlying costs rise, averaged 10.0% per year at a $1 million attachment across 2018 to 2023, then spiked to 20.9% in the single year from 2022 to 2023; at $2 million the one-year figure hit 30.7%. Gene and cell therapies alone trended at 31.4% per year at the $1 million attachment. Sun Life's 2026 high-cost claims study, drawn from more than 70,000 high-dollar claims, found $1 million-plus claims up 46% over its four-year window, with blood cancers averaging $5.45 million per claim in 2025 and the Duchenne gene therapy Elevidys averaging $3.6 million per administration (Sun Life, May 21, 2026). Tokio Marine HCC's 2026 market report counts claims above $2 million up 213% since policy year 2020 and finds children under ten now account for 39% of claims spending above $1 million, more than triple any other ten-year age cohort (Tokio Marine HCC, June 2026). actuary.info has covered the diagnosis-level severity concentration in QBE's 2026 stop-loss severity report and the leveraged-trend mechanics behind these numbers.

Who Writes the Market

A note on vintage before the league table: stop-loss market share is only published from NAIC statutory Accident & Health exhibit data, which files in the spring and gets compiled into the Oliver Wyman/Guy Carpenter ranking each fall, so 2024 remains the most recent full-year ranking available as of August 2026; the 2025 edition lands this autumn. The rank order moves slowly (the top five are unchanged since 2019, though UnitedHealth compounded at nearly 23% per year over that span and Humana all but exited), but the books themselves have moved a great deal since 2024. The right-hand column pairs each ranked writer with its freshest public 2026 signal.

Rank	Carrier group	2024 premium (latest full-year statutory)	Where the book stands, mid-2026
1	Cigna	$5,623M	Tracking to $8B+ FY2026 premium (management basis); repricing two-thirds complete, stop-loss on plan in Q2
2	UnitedHealth	$5,335M	No stop-loss-specific 2026 disclosure
3	CVS Health (Aetna)	$3,058M	No stop-loss-specific 2026 disclosure
4	Sun Life Financial	$2,985M	Stop-loss sales +86% YoY in Q2 2026; loss ratios in mid-70s target range
5	Elevance (Anthem)	$2,468M	No stop-loss-specific 2026 disclosure
6	Tokio Marine (HCC)	$2,277M	Employee Benefits NPW +21.8% YoY (Apr-Jun 2026) on stop-loss rate increases
7	HCSC	$1,870M	No stop-loss-specific 2026 disclosure
8	Voya Financial	$1,810M	$1,537M annualized in-force at 6/30/26, ~17% below the mid-2024 peak
9	Highmark (HM Insurance Group)	$1,231M	$107M operating loss on the 2025 book; stop-loss businesses repositioned
10	Sumitomo (Symetra)	$1,136M	Majority of block ceded via 80% funds-withheld coinsurance; per-claim cession threshold lowered to $4M
11	W. R. Berkley	$705M	Berkley Accident and Health crossed $1B annualized premium (Feb 2026)
12	Swiss Re	$679M	No stop-loss-specific 2026 disclosure
13	QBE	$635M	Onerous-contract provision in H1 2026; 20%+ rate increases; book in deliberate temporary contraction
14	BCBS South Carolina	$607M	No stop-loss-specific 2026 disclosure
15	Fairfax Financial (Crum & Forster)	$580M	Hiring stop-loss underwriting and sales leadership through mid-2026 to expand

Ranking and 2024 premium: Oliver Wyman/Guy Carpenter Stop Loss Market Update, Fall 2025, from NAIC statutory filings; the top 15 hold roughly 79% of the market. Mid-2026 status: Q2/H1 2026 company disclosures cited throughout this article. Note Cigna separately describes its stop-loss book as roughly $8 billion of annual premium; the statutory-entity figure uses a narrower classification basis. Skyward Specialty sits below the 2024 top 15 but is the fastest mover, with A&H premium up 58% in Q2 2026.

Voya: The Deepest Cycle, the Cleanest Disclosure

No carrier gives actuaries more to work with than Voya, whose investor supplement publishes a policy-year loss-ratio development table. It shows exactly how a stop-loss book breaks and how it gets repaired. The 2024 policy year was originally booked at an 81% ultimate loss ratio as of mid-2024; by year-end 2024 the pick had blown out to 94%, driving a reported calendar-quarter loss ratio of 115% in Q4 2024 as the catch-up ran through earnings. That cohort has since developed favorably back to 88.5%. The 2025 policy year, priced with a 21% average January rate increase, is now picked at 89%, improved from 90% at year-end. The 2026 policy year, priced with a further 24% average increase, is booked at 87% (Voya Q2 2026 investor supplement, August 4, 2026).

The calendar-period arithmetic is moving the right way: Q2 2026's reported stop-loss loss ratio of 85.4% brings the trailing twelve months to 86.2%, against 92.0% for the prior-year period, and trailing-twelve-month stop-loss underwriting gain rose to $208 million from $128 million. But the repair has been bought with book shrinkage as well as rate. Annualized in-force premium of $1,537 million at June 30 stands roughly 17% below its mid-2024 peak of $1,845 million, and net premium is down 8.3% on a trailing basis (Voya Q2 2026 investor supplement). Management told investors the 2026 cohort is emerging with lower claim frequency and fewer high-severity claims than either 2024 or 2025, while cautioning that the cohort is only 15% to 20% developed and that the firmer read comes at Q4 (Voya Q2 2026 earnings call, August 5, 2026). Two numbers frame how much work remains: the 2026 book is priced to an 87% ultimate, and Voya's stated long-term target range is 77% to 80%. Management has consistently described full margin restoration as a 2027 event, after a third consecutive repricing cycle (HR Executive, February 13, 2026).

Sun Life: The Book That Did Not Break Leans In

Sun Life's US stop-loss business, the largest independent book at roughly $3 billion of premium, is the counterexample that shows the 2024 shock was survivable with ordinary-course repricing if the starting point was priced adequately. Its worst full-year stop-loss loss ratio of the cycle was 74% in 2024, against a pricing target of about 73%, after a fourth-quarter severity spike that then-US president Dan Fishbein attributed to later-stage cancers, premature births, and hospital price increases (Sun Life Q4 2024 earnings call, February 2025). The company answered with a 14% average increase at the January 2025 renewal and 17% at January 2026, priced at target margins (Sun Life Q4 2025 earnings call, February 12, 2026).

The Q2 2026 result is a carrier playing offense in a hard market. US underlying net income rose 15% to US$164 million on stop-loss revenue growth, stop-loss sales rose 86% year over year, and management characterized loss ratios as stable within the mid-70s target range. The January 2025 cohort is 97% complete and in line with expectations; the January 2026 cohort is about 15% complete and tracking to pricing (Sun Life Q2 2026 earnings call, August 7, 2026; Q2 2026 shareholders' report). The read-through for competitors is uncomfortable: the market's rate correction is generous enough that the carrier with the least margin repair to do is converting it into share instead, at close ratios management attributes partly to proprietary risk-selection analytics. One caveat for modelers: Sun Life does not disclose a precise quarterly stop-loss loss ratio, and it acknowledged an unfavorable, unquantified update to the January 2025 cohort in Q3 2025 before the cohort stabilized, so the public record on the exact path is thinner than Voya's.

Cigna: Two-Thirds Through Repricing an $8 Billion Book

Cigna is the market's largest writer on any measurement basis, and its stop-loss margin problem, disclosed with its Q4 2024 results, is being worked off on a stated two-year schedule. On the Q2 2026 call, CEO Brian Evanko said the company continues to make progress on stop-loss margin recapture, is tracking toward more than $8 billion of stop-loss premium this year, and, notably for pricing actuaries, has seen stable frequency in high-cost claimants across multiple attachment levels, consistent with both expectations and prior-year levels. CFO Ann Dennison added that stop-loss was not a variance driver in the quarter (Cigna Q2 2026 earnings call, July 30, 2026). Management has separately indicated the repricing is roughly two-thirds complete in 2026, with 2027 the final year of margin recovery and price increases expected to moderate afterward (Becker's Payer Issues, May 2026). Cigna Healthcare's medical care ratio of 84.5% and its drivers are covered in actuary.info's Q2 Cigna analysis; the stop-loss-specific point is that the largest book in the market is repricing on the same 2026-to-2027 calendar as everyone else, which is a large part of why nobody's rate action is being competed away.

Tokio Marine HCC: On Plan, and Planning for More

Tokio Marine discloses its US stop-loss results inside an "Employee Benefits" line that combines TMHCC's accident and health book with group life at Reliance Standard, so the disclosure is coarser, but the direction is unambiguous. Employee Benefits net premiums written rose 21.8% year over year in the April-to-June quarter (10.3% excluding currency), which the company attributes primarily to solid rate increases in medical stop-loss, and the line's combined ratio of 94.3%, though up 2.9 points from the unusually good prior-year quarter, is running exactly on the full-year plan of 94.3% (Tokio Marine Holdings, Overview of 2026 1Q Results, August 12, 2026). In other words: deterioration was planned, priced for, and is arriving on schedule.

TMHCC's stop-loss chief executive Jay Ritchie has been the most explicit voice in the market on duration, telling Insurance Business in June that the tightening will persist through at least 2027 and describing the cycle as three quiet years followed by "three years of crush" running 2024 through 2027. He also put numbers on the pricing miss that started the cycle: January 2024 trend ran 5 points above the five-year average, and January 2025 ran 9.2 points above (Insurance Business America, June 18, 2026).

QBE: The Carrier Still Paying for 2025 Quotes

QBE's first-half result is the clearest evidence that the severity wave outran the quoting calendar. The group booked an onerous-contract provision on its North America accident and health book under AASB 17, effectively pulling an expected full-year loss on the line into the half; management did not quantify the provision but said it drove the North America ex-crop combined ratio above 100% and should not repeat in the second half (QBE H1 2026 earnings call, August 13, 2026). CEO Andrew Horton's explanation was a timing story: claims inflation emerged at roughly 30% against the roughly 20% the book had been priced for, and it surfaced mainly late in the third and fourth quarters of 2025, after much of the business had already been quoted. QBE is now pushing accident and health rate increases above 20%, against high-single-digit rate across its North America portfolio, and described a temporary contraction of the book rather than an exit; management also observed that competitors' stop-loss results are not dramatically different and that the market is repricing in unison (QBE H1 2026 earnings call; Insurance Business America, August 13, 2026). The half validates a mechanism actuaries should internalize: in a 12-month-policy market, a severity shock that lands in Q4 poisons two policy years, the one on the books and the one already quoted.

Skyward and Berkley: Capacity Runs Toward the Hard Market

Not every public carrier is retrenching. Skyward Specialty's accident and health division, focused on smaller self-funded groups and medical-cost management, grew gross written premium 57.8% year over year in Q2 to $95.5 million, with first-half growth of 51.6%, inside a group combined ratio of 89.5% (Skyward Q2 2026 earnings release, August 4, 2026). CEO Andrew Robinson told analysts the growth inflected after the launch of a group-captive offering and characterized Skyward's 2024 loss ratio, roughly 71% in NAIC data by his description, as among the five best of the top 50 writers. He also pointed to reinsurer withdrawals, driven by losses on poorly performing MGA-fronted business, as adding fuel for direct writers with clean books (Skyward Q2 2026 earnings call, August 5, 2026). Skyward's own accident and health growth chief has published the sharpest version of the bull case: an 86% industry gross loss ratio in 2024, 2025 tracking above 90%, and roughly $1.5 billion of reinsurance capacity exiting (Skyward Specialty, "Medical Stop-Loss Approaches an Inflection Point," 2026).

W. R. Berkley does not break out its Berkley Accident and Health unit in the 10-Q; stop-loss sits inside a short-tail-lines bucket whose net earned premium rose 7.0% in the quarter. The company did mark a milestone in February: Berkley Accident and Health crossed $1 billion of annualized premium, describing itself as the fastest-growing independent stop-loss carrier in 2024 NAIC data (Berkley Accident and Health, February 3, 2026). Fairfax's Crum & Forster, the number 15 writer, is likewise silent on the line in quarterly disclosure but has been hiring stop-loss underwriting and sales leadership through mid-2026, signaling expansion rather than retreat.

The Casualty List: Symetra's Cessions, Highmark's Loss, RGA's Exit

The carriers without earnings calls tell the harsher half of the story. HM Insurance Group, Highmark's stop-loss arm and the market's ninth-largest writer, posted a $107 million operating loss on $1.3 billion of operating revenue for 2025, which the parent attributed to increased frequency and severity of high-dollar claims, adding that the stop-loss business has been strategically repositioned (Highmark Health, March 24, 2026). Symetra, celebrating its fiftieth year in stop-loss, has visibly de-risked through reinsurance: its 2025 statutory statements disclose that for business written after October 2024 it cedes claims above $4 million per individual, maintains an 80% coinsurance funds-withheld arrangement covering the majority of the block, and beginning in 2025 added a further morbidity cession on certain specific and aggregate policies (Symetra Life statutory financial statements, 2025).

Above the carriers, the reinsurance layer is contracting outright. RGA is exiting stop-loss and US medical reinsurance entirely, citing extraordinary losses that began in late 2024, with its withdrawal filings approved in April 2026. Gallagher Re counts three major reinsurers withdrawing from US medical reinsurance in recent months while framing the episode as restructuring rather than systemic distress (Gallagher Re, "Crisis or Opportunity?", February 2026). The withdrawal is already visible in cedent results: AXIS Capital's Q2 10-Q attributes part of its higher current accident-year loss ratio to elevated employer stop-loss experience in its accident and health book (AXIS Capital 10-Q, Q2 2026). For self-funded plan sponsors, less reinsurance behind the market means the hard-market toolkit, lasering, higher specifics, decline-to-quote, persists even as carrier loss ratios improve; Mercer reports rising decline-to-quote rates and more frequent lasering alongside January 2026 renewal increases that averaged 23%, up from 18% the prior cycle, with even well-performing groups averaging 15% (Mercer, June 3, 2026). Segal's national survey, which skews toward public-sector and multiemployer plans, puts the average 2026 premium increase at 12.7%, up from 9.7% (Segal, 2026 stop-loss survey).

So Is Excess Medical Inflation Actually Turning?

Separate the two questions the disclosures answer. Are carrier loss ratios turning? Yes: Voya's trailing ratio is down almost six points in a year, Sun Life is at target, Cigna calls its book on plan, TMHCC is on plan, and even QBE frames its provision as a one-half event already repriced. Is the underlying claims environment turning? No. PwC's Health Research Institute restated its 2026 group medical trend upward to 9.0% and projects 2027 at 9.0%, the highest in 17 years; the 2026 Milliman Medical Index rose 7.9%, its largest non-COVID increase in over a decade; and Segal projects 2027 medical trend at 9.9% with prescription drugs at 11.5% (PwC HRI, June 2026; Milliman, May 20, 2026; Segal 2027 trend survey). Run any of those through the leverage arithmetic at a fixed $1 million attachment and the loss-cost trend a stop-loss book must absorb remains in the high teens to low twenties, which is precisely what the 17% to 24% carrier rate actions and 23% average renewals are chasing.

What has plausibly changed is the second derivative of severity. Cigna reports stable high-cost-claimant frequency across attachment points; Voya sees fewer high-severity claims in its 2026 cohort than in either prior year; Sun Life's maturing cohorts are landing on pricing. The million-dollar-claim frequency surge of 2023 through 2025 appears to be flattening at its new, higher plateau rather than compounding at the same pace. That is the difference between a market that can be repriced and one that cannot: rate can catch a level shift, but only if the level holds still long enough. The gene and cell therapy pipeline is the standing threat to that assumption, with approved one-time therapies now listing between $2.1 million and $4.25 million per administration (Oliver Wyman/Guy Carpenter, Fall 2025; see actuary.info's cell and gene therapy pricing analysis).

The Desk

For pricing and reserving actuaries, four working conclusions fall out of the quarter. First, distinguish reported calendar-period loss ratios from policy-year ultimates before declaring victory: Voya's 85.4% quarter includes favorable development on prior cohorts, while its current cohort is still booked seven points above target, and every carrier's 2026 cohort is at most 20% developed. The honest statement of the turnaround is that January 2026 business was priced closer to adequacy than January 2025 business, and the confirming data arrives at Q4. Second, the QBE half is the cautionary tale on trend recognition lag: a severity signal that emerges in Q4 lands on a book that is already quoted, so 2027 rate indications should be set against the possibility that late-2026 severity is not yet visible in paid data, not against the emerged experience alone. Third, the loss-ratio gap between integrated and independent writers (95.6% versus 80.5% in 2024) says the remaining repricing burden is concentrated in the largest books, which supports the consensus that market-wide rate firmness holds through the January 2027 renewal even as individual carriers reach target sooner. Fourth, for plan-sponsor-side actuaries, the binding constraint is shifting from price to terms: with reinsurance capacity out and lasering up, contract provisions, rate caps, no-new-laser guarantees, aggregating specifics, are where the next renewal's economics will actually be decided. The scoreboard to watch: Voya's Q4 read on its 2026 cohort, Sun Life's January 2027 renewal increase, and whether Cigna confirms 2027 as the final repricing year. If all three land as guided, the 2028 market softens; if the gene-therapy tail or another Q4 severity surprise intervenes, Ritchie's three years of crush get a fourth.