2027 ACA marketplace filings are tracking double digit for a second consecutive year. Early statewide averages run from 6.5% in Vermont to 22.4% in Washington, according to Georgetown CHIR's June 2026 review of states that publish filings early.

The 2026 spike was widely read as a one-time correction for enhanced premium tax credit expiry. The 2027 filings say otherwise. The first shock is compounding, because 14% of enrollees never paid a January premium and left the pool before submitting a claim.

Key Takeaways

  • 22.4% in Washington against 8.5% individual-market medical trend leaves 13.9 percentage points of the statewide average that is morbidity and enrollment composition rather than cost.
  • 14% of marketplace enrollees did not pay their first January 2026 premium, with state-level variation from below 7% to above 18%, per Wakely's April 2026 issuer analysis.
  • 10.2% higher morbidity among paying stayers than among non-payers translated into a pool morbidity increase of 2.9% to 6.5% above the pre-cliff baseline.
  • 3.3 percentage points of Fallon Community Health Plan's 25.7% request is attributed explicitly to federal policy morbidity; MVP Health Plan carries a $48 per member per month morbidity load.
  • 650,000 people are affected by six carrier exits across one third of states, concentrating migration into the carriers that stay.

What the Early 2027 Filings Actually Show

The 2026 national median proposed increase was 18%, the highest in nearly a decade, and it was read at the time as a corrective spike tied to subsidy expiry. That reading carried a forecast: with the enrollment shock absorbed, 2027 filings would revert toward the 6% to 7% trend rates of the pre-cliff market.

The early filings do not revert. Washington's statewide average is 22.4%. Fallon Community Health Plan in Massachusetts is requesting 25.7% overall and attributes 3.3 percentage points of it explicitly to morbidity from federal policy changes. MVP Health Plan in Vermont is carrying a $48 per member per month morbidity load tied to subsidy expiry effects on the population it still has. These are quantified primary drivers filed by carriers working from a full year of post-cliff claims, not residual adjustments layered on a stable base.

The number behind the headline is the spread, not the average. PwC puts underlying individual-market medical cost trend at 8.5% for 2027. Washington's 22.4% sits 13.9 percentage points above it, and that gap is the morbidity and composition load.

Vermont's 6.5% and Washington's 22.4% are therefore not a distribution around trend. They are a distribution of how much composition damage each state's pool absorbed in a single year, filed on a base that had already priced one round of adverse selection.

Why the 14% Never Reaches the Risk Adjustment Formula

Wakely's April 2026 analysis, drawing on 75-plus issuers representing roughly 80% of the ACA individual market, found that 14% of enrollees did not pay their first January 2026 premium, with state-level rates from below 7% to above 18%.

The pricing consequence turns on what separates a first-month non-payer from a mid-year lapse. A member who lapses in June has submitted claims. Those claims establish an HCC diagnosis history, and the risk adjustment transfer formula acts on it. A member who never pays January's premium submits zero claims, carries no diagnostic codes, and exits without generating anything the formula can compensate.

That matters because the exits were not random. Wakely found paid stayers carried 10.2% higher morbidity than non-payers, 6.2% on a state-average basis after adjusting for plan mix, and estimated 2026 pool morbidity rose 2.9% to 6.5% above the pre-cliff baseline. KFF projects average monthly effectuated enrollment falling to about 17.5 million in 2026 from 22.3 million in 2025.

So the morbidity moved before the formula could see it move. Fallon's 3.3 points and MVP's $48 per member per month are the visible form of that residual: the part of the deterioration a carrier has concluded risk adjustment will not hand back, priced into the premium instead. A 2027 filing that treats the risk adjustment receivable as a full morbidity offset is pricing a shift the formula was never given the claims data to recognize.

The Calibration Lag Compounds, and the Bronze Cushion May Not

The HHS-HCC formula is calibrated to the prior year's national pool. In 2027 those coefficients reflect 2026 experience, which means they see the post-exit pool but not the exit event, because the 14% left no claims behind. For a single-year shock that lag is manageable and largely corrects in year two. Two consecutive years of adverse selection layer the lag on top of a 2026 settlement that was itself partly short.

Concentration makes the layering uneven. Six carriers have announced 2027 exits affecting roughly 650,000 people across one third of states, and Cigna's ACA enrollment fell 17% year over year. Migrating members skew toward higher utilization, because healthier enrollees are more likely to drop coverage than to shop for a replacement plan. Texas closed 2026 at a 34.7% final gross increase, silver up 35.1% and gold up 31.9%, in a non-expansion state where the subsidy cliff is an affordability cutoff rather than a relative-cost shift.

The projected 4.8 million person decline in nationwide effectuated enrollment does not spread evenly. States whose pre-cliff books were concentrated in the band just under 400% of the federal poverty level lose the most members per carrier, so an identical national morbidity shift arrives at very different sizes depending on where a carrier writes.

The one offsetting signal is fragile. Bronze selections rose to 9.2 million in 2026 from 7.3 million, and average deductibles reached a record $3,786 from $2,759, which suppresses induced utilization. But bronze also concentrates healthier members in the tier that cross-subsidizes least, and those are the members second-year erosion takes first. A carrier holding a large bronze book in 2027 is relying on a cushion whose composition is the thing most likely to move.

Further Reading

Sources