KFF's 2027 ACA Median Lands at 15%, Splitting Trend From the Subsidy Cliff

PwC pegs the 2027 individual-market medical trend at 8.5% (PwC, 2026), the baseline for how much care simply costs more each year. KFF's newly completed 50-state count of 276 insurer filings puts the median proposed 2027 marketplace increase at 15% (KFF, August 2026). The 6.5-point gap is not noise: carriers are now filing it as a distinct, quantified subsidy-cliff morbidity line, separate from trend.

Decomposing the Filed Median: Trend, Morbidity, and the Residual

KFF's August 3 update, which extended its rate-filing review to all 50 states and the District of Columbia, found a median proposed increase of 15% across 276 insurers for 2027 (KFF, August 2026). That follows a median proposed increase of 18% for 2026, which settled at a final median of 20% after state rate review, itself the highest in nearly a decade. A 15% median is a step down from 20%, but KFF's own framing puts it in context: it is still the second-highest requested rate change since 2018, a year when marketplace pricing had been comparatively flat for several years (KFF, August 2026).

Set against PwC's 8.5% individual-market trend, the arithmetic leaves roughly 6.5 percentage points of the national median unexplained by the cost of care alone. That residual is not a single number with one cause; it blends the subsidy-cliff morbidity effect, market-wide adjustments for risk-adjustment transfers and exchange user fees, and in a subset of states, cost-sharing reduction loading that must now be separately documented in the actuarial memorandum under the finalized portion of the 2027 Payment Notice (see actuary.info's coverage of the CSR loading requirement). But the morbidity component is the piece insurers are naming explicitly, and it is the piece with independent third-party corroboration. The Congressional Budget Office models how much lower 2027 gross benchmark premiums would run if enhanced premium tax credits had been extended rather than allowed to expire: 5.7% in 2027, widening to 9.0% in 2028 before narrowing to 3.3% in 2029 as the market re-stabilizes around a smaller, sicker enrolled base (CBO, 2025). CBO's 5.7-point estimate was modeled before a single 2027 rate was filed. That it lands within a point of the actuarial residual implied by KFF's filed median minus PwC's trend estimate is a rare case of a top-down macro model and bottom-up filing data converging on roughly the same number for roughly the same reason.

How Carriers Are Quantifying the Morbidity Load

Enhanced premium tax credits, in place since the American Rescue Plan Act and extended through 2025, removed the income cap on subsidy eligibility and reduced required contributions across every income band. Their expiration at the end of 2025 restored the pre-2021 subsidy structure, which raises net premiums most sharply for enrollees near or above 400% of the federal poverty level and reintroduces a hard subsidy cliff at that threshold. Twenty-two million of the roughly 24 million marketplace enrollees in 2025 received a premium tax credit; KFF had earlier estimated that a currently-subsidized enrollee could see a net premium increase of about 114% if the credits lapsed entirely (KFF, October 2025). A net premium move of that size is not uniform across the risk pool. Enrollees with lower expected claims costs are more price-elastic and disproportionately likely to drop coverage when their net premium jumps; enrollees managing chronic conditions have less flexibility to go uninsured and are more likely to absorb the increase. The remaining pool skews sicker purely as a function of who left, independent of any change in underlying morbidity within either group.

The Georgetown University Center on Health Insurance Reforms' review of early 2027 actuarial memoranda found three carriers that quantified this effect directly rather than folding it into an undifferentiated trend assumption. Fallon Community Health Plan in Massachusetts reported that "3.3 percentage points of its 25.7% proposed increase is due to morbidity impacts from policy changes, including expired ePTCs" (Fallon Community Health Plan filing, via Georgetown CHIR, June 2026). ConnectiCare Benefits attributed a full 5 percentage points of its Connecticut filing to the scheduled expiration of enhanced premium tax credits (Georgetown CHIR, June 2026). MVP Health Plan in Vermont went further, expressing the effect as a claims-cost figure rather than a rate-multiplier: an additional $48 per member per month in projected claims tied to the loss of enhanced credits (Georgetown CHIR, June 2026). None of these three filings is typical in its precision, most 2027 memoranda still cite morbidity and rising costs together without separating them, but they establish that the morbidity load is measurable at the carrier level and that at least some actuaries are choosing to file it as its own line rather than absorb it into a single blended trend factor.

The Dispersion the Median Hides

A national median of 15% describes the center of a distribution that runs from mid-single digits to nearly 23% by state. Vermont's filed statewide average, at 6.5%, sits closest to underlying medical trend; Iowa and Utah followed close behind. At the other end, Alaska's filed average reached 22.9%, with Washington close behind at 22.4%, and Georgia, New York, and Rhode Island each in the 20% to 21% range (healthinsurance.org, July 2026).

State	Proposed 2027 Increase
Vermont	6.5%
Iowa	6.7%
Utah	6.8%
Rhode Island	20.1%
Georgia	20.7%
New York	20.7%
Washington	22.4%
Alaska	22.9%

The spread tracks recognizable structural variables more than it tracks any single national policy input. States with their own reinsurance programs or state-funded subsidy wraps, Vermont among them, dampen the net-premium shock that drives the morbidity spiral, because fewer enrollees face the full net-premium jump that pushes healthier members out. States that never expanded Medicaid, a group that includes Georgia, run marketplaces that already skew toward lower-income enrollees for whom the subsidy cliff is a binary affordability cutoff rather than a marginal cost increase, which concentrates the morbidity effect in a smaller, more price-sensitive population. Washington and Alaska both carry thin, low-competition individual markets where a single carrier's exit or repricing decision moves the statewide average disproportionately. None of this is visible in the 15% headline; it only appears once an actuary pulls the state-level filings apart.

The Feedback Loop CBO's Multi-Year Path Implies

CBO's own three-year path, a 5.7-point premium effect in 2027 that widens to 9.0 points in 2028 before narrowing to 3.3 points in 2029, is itself evidence that the morbidity effect does not resolve in a single pricing cycle (CBO, 2025). actuary.info's earlier tracking of the 2027 filing cycle documented the mechanism driving that widening middle year: healthier enrollees who exit before submitting a single claim are invisible to the risk-adjustment transfer formula, which only recognizes cost through diagnosis codes generated by claims, so a first-month lapse shifts pool composition without generating any of the data the formula uses to compensate carriers for it (see actuary.info's earlier analysis of the compounding morbidity spiral and the eight-state early filing data that first flagged the pattern). Each year the risk-adjustment formula recalibrates to a prior-year pool that was already worse than the formula's coefficients assume, and each year's shortfall layers on top of the last rather than resetting. Wakely's separate analysis of 2026 first-month enrollment found paid stayers carrying 10.2 percentage points higher morbidity than enrollees who never paid their first premium, translating into an estimated 2.9% to 6.5% pool-level morbidity shift for 2027 assumptions (see actuary.info's coverage of the Wakely data), a range that brackets the CBO estimate and the Fallon and ConnectiCare filings from two independent angles.

The practical read for a pricing actuary is that 2027's filed morbidity load is not a corrective, one-time reset following the 2026 subsidy cliff. It is the first full-year data point in what CBO's own model treats as a multi-year adjustment path, with the largest single-year effect still ahead in 2028 under current law. A 2027 filing that treats this year's morbidity assumption as the ceiling, rather than as a point partway along a rising curve, is pricing against last cycle's dynamic rather than the one CBO's model and Wakely's claims data both describe.

From the Filed Percentage to the Index Rate

The 15% median is a rate-level statistic, but it does not translate one-for-one into the index rate that underlies every plan's premium build. Under the Unified Rate Review Template, a carrier first projects allowed claims per member per month for its full experience period, then applies trend to bring that projection forward, then layers market-wide adjustments, most significantly the risk-adjustment transfer, the exchange user fee, and, where applicable, CSR reconciliation, to arrive at the index rate. Only after the index rate is set does the carrier apply plan-level factors, including the age curve, geographic rating area, tobacco surcharge, and actuarial-value adjustment for metal tier, to produce the premium an individual enrollee actually sees. A morbidity load filed as a distinct line, the way Fallon and ConnectiCare structured theirs, sits inside that allowed-claims-PMPM projection rather than in the market-wide adjustment section; that placement matters because it means the load compounds with trend rather than being a fixed dollar add-on, and it is why a carrier that under-projects morbidity in the base claims estimate cannot fully correct for it downstream through the market-wide adjustments alone. The 2027 Notice of Benefit and Payment Parameters also reduced the federal exchange user fee percentage that feeds into that market-wide adjustment step for 2027, a change that pulls in the opposite direction of the morbidity load and that actuary.info has covered separately in the context of the user-fee cut's effect on 2027 rate filings. A carrier reconciling a 15%-range headline increase against a smaller net premium change for a specific plan should expect the user-fee reduction, not a miscalculation, to explain part of the gap.

What a Repricing Actuary Should Stress-Test if Congress Acts Late

Every 2027 filing locked its subsidy-cliff morbidity assumption on the premise that enhanced premium tax credits stay expired for the full 2027 plan year. That premise has already proven unstable once: the credits were extended repeatedly on a short timeline before their 2025 lapse, and several bills, including the Lower Health Care Costs Act, remained active in Congress as 2027 filings were being finalized. A late extension, arriving after rates are filed but before or during the 2027 plan year, would leave carriers holding a morbidity load priced for a sicker pool that partially fails to materialize, because some price-elastic enrollees who would otherwise have exited stay covered once their net premium drops back down. A repricing actuary should model that scenario explicitly rather than treat the filed assumption as fixed. That means running a sensitivity table on the morbidity load itself, showing projected loss ratio and margin under full expiration, partial mid-year extension, and full retroactive extension; documenting, in the actuarial memorandum, the specific enrollment and morbidity assumptions that would need to reverse for a late extension to require a rate correction, so that any subsequent request to state regulators for a mid-year adjustment has a pre-built justification; and separately tracking the CSR and risk-adjustment market-wide adjustments, which move on a different timeline than a legislative subsidy change and should not be re-opened by the same trigger. States vary in whether they permit any mid-year rate correction at all, which means the more consequential decision for some carriers is not the sensitivity analysis itself but how much of the subsidy-cliff morbidity load to build into the base 2027 rate versus reserve for justified but unfiled contingency, a call that has to be made now, months before Congress's timeline on any extension becomes clear.