Stop-Loss Specific Claims Trend Hit 18% in 2025: The Attachment Adequacy Deficit Running Into 2027 Renewals

The specific claims trend for self-funded employer stop-loss ran at 18% in 2025, 9.5 percentage points above the 2019 to 2023 historical baseline and 4.2 points above 2024's already-elevated 13.8% (Tokio Marine HCC, BenefitsPro, June 2026). Tokio Marine HCC, one of the largest stop-loss carriers in the U.S. market, stated it was "undeniably surprised" by the magnitude. Attachment points set before the mid-2024 surge are now systematically underinsured against the claim distribution that emerged in 2025.

A 4.2-Point Acceleration Sitting Well Above a Five-Year Baseline

The comparison frame matters. In January 2024, monthly specific trend data from Tokio Marine HCC was running 5 percentage points above the prior five-year average. By January 2025, that gap had widened to 9.2 percentage points (Tokio Marine HCC, June 2026). The full-year 2025 result of 18% reflects a structural shift in what constitutes a high-severity claim, not a temporary frequency blip, and the word "undeniably" in the carrier's own characterization points to how far the result departed from internal pricing assumptions.

The large-claim frequency data confirms the distribution has moved. Claims exceeding $2 million rose 213% from policy year 2020 through 2025 (Tokio Marine HCC). Claims exceeding $1 million now affect 49% of the 1,268 self-funded plan sponsors surveyed in the 2025 Aegis Risk/IFEBP Medical Stop-Loss Premium Survey, compared with 23% in the prior survey period. The same survey describes the result with a phrase that should inform every 2026 and 2027 renewal: "$3 million claims are the new $1 million claim" (Aegis Risk/IFEBP, 2025). A benchmark widely used to define a catastrophic claim has effectively doubled in less than four years.

Approximately two-thirds of large U.S. employers with coverage self-insure their health plans (NAIC). Stop-loss is the structural mechanism that prevents a single catastrophic policy year from destabilizing a plan sponsor's benefit budget. When stop-loss carriers are repricing faster than employers anticipated and raising specific deductibles to restore loss ratio adequacy, the financial viability of self-funding for mid-to-large employer groups is the question behind the trend number.

Cigna and Sun Life confirmed consistent findings. Both characterized their 2025 stop-loss experience as running in line with what they observed since the surge initiated in mid-2024 (BenefitsPro, June 2026), which removes any interpretation that the 18% figure is idiosyncratic to Tokio Marine HCC's book. Three major carriers, independent portfolios, converging on the same trend reading.

Specific Deductibles, Leverage, and the Attachment Adequacy Gap

Specific stop-loss covers individual claims above a contractually defined specific deductible (SD). The carrier pays each dollar of covered-member expense that exceeds the SD within a policy year. The actuarial price for that coverage derives from the excess loss factor: expected losses above the SD expressed as a percentage of total expected plan costs. The ELF is not a linear function of the SD level. Because the tail of the claim severity distribution has lower probability mass and higher variability, a given rightward shift in the underlying distribution raises the ELF more steeply at higher attachment points. A 10-point shift in the severity distribution might raise the ELF at a $200,000 SD by 20%, but it raises the ELF at a $500,000 SD by 50% or more. This leverage is why specific trend errors compound so quickly at high attachment levels, and why 18% specific trend in a single year is not simply twice as bad as 9% specific trend.

At each annual renewal, carriers propose an SD based on a group's demographics, historical experience, and trend assumptions. A group that negotiated a $175,000 SD in 2022 using an 8% specific trend assumption was implicitly projecting an equivalent 2025 SD of about $220,000 under three years of compounded growth. Under 18% specific trend in 2025 alone, the actuarially equivalent SD would sit closer to $260,000. If the carrier raises the SD to $260,000 at renewal to restore pricing adequacy, the employer must absorb all claims between $220,000 and $260,000 that stop-loss would have covered under the original pricing. That $40,000 per-claimant gap, multiplied across a group with several high-cost members, is the attachment adequacy deficit the Tokio Marine HCC admission describes in practical terms.

The IFEBP premium data illustrates the leverage operating at the carrier level. Average monthly specific stop-loss premiums at the $100,000 deductible rose 8.8% in 2025, reaching $229.40 per covered employee. At $500,000, the increase was 10.1%. At $1,000,000, premiums jumped from $13.84 to $17.69 per covered employee per month, a 27.8% increase in a single year (Aegis Risk/IFEBP, 2025). The premium gradient confirms that carriers are absorbing the heaviest corrections at the highest attachment levels, where ELF leverage magnifies both the loss impact and the required premium response.

The claim frequency data by threshold reveals the same pattern from a different angle:

Specific Deductible Threshold	Percentage-Point Change in Claim Frequency
$200,000	+46 percentage points
$500,000	+75 percentage points
$1,000,000	+112 percentage points
$2,000,000	+247 percentage points

Source: Tokio Marine HCC, Insurance Business America, June 2026. Percentage-point changes in large-claim frequency relative to prior baseline.

The frequency acceleration is not proportional across thresholds. Claims above $200,000 rose 46 points; claims above $2 million rose 247 points. That divergence is the signature of a distribution that has thickened in the tail rather than simply shifted upward, which is the key reason why applying a uniform trend factor to an ELF table calibrated on pre-2024 data produces systematically wrong answers at high attachment points. The tail shape has changed, not only the location.

No Single Cause: Why Multi-Driver Convergence Breaks Trend Models

When Tokio Marine HCC enumerated the conditions driving 2025 specific trend, no single category dominated. The carrier's analysis cited increased claim frequency across all severity thresholds, concentrated in cancer, neonatal and infant care, and transplant cases, with secondary contributions from Medicare Advantage reimbursement cut-throughs, trade tariffs on medical supplies, and physician supply constraints. That structure is the actuarial problem. A single-driver trend model, calibrated to one category's rate of change, systematically underprices the residual.

Cancer was cited as a catastrophic claim driver by 92% of plan sponsor respondents in the IFEBP survey, up from 83% the prior year (Aegis Risk/IFEBP, 2025). Cancer and cardiovascular disease together accounted for over 48% of total stop-loss claim costs in 2024, compared with 44% in 2021 (Tokio Marine HCC). The channel is not primarily utilization volume. Checkpoint immunotherapy drugs like pembrolizumab (Keytruda) run roughly $200,000 annually at commercial rates. CAR-T cell therapies for hematologic malignancies carry wholesale acquisition costs from $373,000 to $475,000 per course, with total episode costs including inpatient management routinely exceeding $1 million. A self-funded plan with five or six members starting CAR-T or high-dose checkpoint therapy in a single policy year can produce specific-layer claims that no ELF table calibrated on 2019 to 2022 data will price correctly.

Neonatal and infant cases operate at the opposite end of the age distribution with equivalent severity. Infants under one year old average $1.37 million in claim severity for cases exceeding the $500,000 threshold (Tokio Marine HCC, 2026). Children under ten account for 39% of stop-loss spending above $1 million, more than three times the share of any other age group. Premature birth, congenital conditions, and neonatal intensive care do not have a prior authorization lever that a stop-loss underwriter can apply at the group level. The claims arrive when they arrive, and they are immune to the utilization management tools that contain oncology cost at the individual plan level.

Medicare Advantage reimbursement cuts add a cost-shift channel that does not appear as a utilization event in any single claim category. As CMS revised MA benchmark rates and risk adjustment factors through 2024 and 2025, hospital systems responded by increasing negotiated commercial rates. Self-funded employer plans priced at commercial rates absorbed higher charge-based claim costs without any corresponding change in utilization. This channel is particularly difficult to isolate in aggregate trend data, which is part of why the 2024 specific trend surprise was not fully visible in population-level indices until 2025 plan year experience developed.

GLP-1 medications round out the convergence, but in the aggregate layer rather than the specific layer. At $15,000 to $25,000 annually per patient at gross cost, GLP-1 drugs fall below the specific deductible for most employer groups. Their impact flows through total plan cost and aggregate stop-loss exposure rather than individual specific-layer claims. The overlap with the oncology and neonatal surge means that aggregate expected claims are rising faster than specific expected claims, compressing the corridor buffer simultaneously with the specific-layer pressure. Carriers setting 2026 and 2027 aggregate attachment points must account for both effects, or their aggregate loss ratio will deteriorate independently of their specific claims experience.

Aggregate Stop-Loss: The Corridor Under Recalibration

Aggregate coverage protects the plan against total claims exceeding a corridor above expected, typically set at 120% to 130% of expected net retained claims. The aggregate attachment in absolute dollars derives from the expected claims projection at renewal. When specific trend is higher than the assumption embedded in that projection, expected claims are too low, and the corridor in dollar terms triggers sooner than the carrier priced.

Carriers are responding through corridor factor widening at 2025 and 2026 renewals. A plan that purchased aggregate at 125% of a $2.0 million expected claims projection in early 2024 may find its carrier offering only 130% of a revised $2.3 million expected figure at the 2025 renewal, raising the aggregate attachment from $2.5 million to $2.99 million. The employer absorbs more retained aggregate risk simultaneously with higher specific deductibles, compressing self-funded plan finances on both stop-loss layers at the same time. Plans that were close to their aggregate corridor in 2023 and 2024 may find the 2025 recalibration pushes their aggregate attachment beyond their retained claims, eliminating aggregate recoveries they had incorporated into their benefit budget projections.

The 2027 Renewal Trajectory

Jay Ritchie, President and CEO of Tokio Marine HCC's A&H Group, provided the clearest public market signal available: "I think the market is going to continue tightening...through 2027" (Tokio Marine HCC, June 2026). The carrier's framework characterizes the current period as three years of market correction, from 2024 through 2027, following a pandemic-era period of suppressed claim frequency that allowed pricing to drift below actuarial adequacy. That framing implies the 2026 and 2027 renewal markets will not offer meaningful relief from the SD escalation and premium increases that defined the 2025 cycle.

The correction mechanism operates with a 12 to 18 month lag. Full-year 2024 stop-loss claims developed into 2025 loss experience that confirmed the prior-year signal. The pricing adjustments visible in 2025 renewals, 8.8% to 10.1% increases at standard SD levels (Aegis Risk/IFEBP, 2025), are correcting for one to two prior years of below-trend premium. Achieving full adequacy at the market level requires 2026 and 2027 renewals to sustain similar or higher premium adjustments while continuing to escalate SDs at a pace that outpaces base medical trend until the excess loss factors are properly calibrated to the post-2024 severity distribution.

Sun Life's 43% increase in medical stop-loss sales through May 2026 (BenefitsPro, May 2026) suggests some market concentration as smaller carriers with less capital depth limit new business while established carriers continue writing. Groups with documented oncology histories, recent large neonatal claims, or members on high-cost specialty pharmacy regimens are the most exposed to laser provisions, named-individual exclusions, and aggregate-layer adjustments that can produce effective coverage cost increases of 30% to 50% from their 2022 to 2023 baseline.

Actuarial Modeling in a Non-Stationary Specific Trend Environment

Trend extrapolation from pre-2024 data does not produce defensible 2026 and 2027 stop-loss pricing. The structural break in claim frequency above every major threshold reflects categorical changes in the high-cost care mix: a generation of oncology biologics that were not material in the 2018 to 2021 data that underlie most carrier ELF tables, a growing gene and cell therapy population with per-event severity that does not fit historical severity distributions, and a Medicare Advantage cost-shift channel that amplifies commercial claim costs without appearing in utilization trend data. An ELF table built from a lognormal distribution fitted on 2019 to 2022 experience will systematically underestimate the excess loss factor at high attachment points because it underestimates the tail thickness that the current mix of oncology biologics and CGT claims produces. A spliced model with a Pareto or generalized Pareto tail, refitted on the most recent two to three policy years, better captures the observed exceedance probabilities at $1 million and above.

Stop-loss actuaries with access to employer-group clinical data have several tools that reduce the lag between the claim distribution's evolution and the pricing assumption. Pharmacy benefit manager runout reports, available with a 30 to 60 day lag, reveal drug spend committed but not yet adjudicated, providing a forward look at the specific-layer exposure before claims fully develop. Utilization management authorization data, from prior auth approvals for high-cost oncology or specialty pharmacy regimens, carries similar predictive power. Disease registry data, where self-funded plans partner with specialized case management programs, can identify members at elevated transplant or NICU risk before catastrophic claims materialize. None of these tools eliminates model uncertainty. They reduce the effective lag between the claim distribution's current state and the actuary's inputs, which is the primary source of the attachment adequacy gap that emerged in 2024 and widened through 2025.

A group with three members actively on CAR-T therapy in policy year 2026 should not be priced at the population-average oncology specific trend rate, because its specific-layer ELF is already a deterministic function of committed drug spend, not a probabilistic function of population trend. The carrier that incorporates forward clinical signals into its group-level ELF adjustment produces renewal terms that match the group's actual risk profile. The carrier that does not will need to recalibrate retroactively on the next renewal cycle, with the employer absorbing the SD escalation in the interim. That is the operational version of what "undeniably surprised" means at the portfolio level.