2027 ACA marketplace increases are tracking double-digit for a second consecutive year. Early state filings show statewide averages from 6.5% in Vermont to 22.4% in Washington (Georgetown CHIR, June 2026). The mechanism is not a second enrollment shock; it is the first shock compounding: first-month premium nonpayment reached 14% in January 2026, removing healthier members before any claim was submitted, and shifting the morbidity index before risk adjustment transfers even ran (Wakely, April 2026).

Early 2027 Filings: What the Numbers Say About Year Two

The 2026 national median proposed premium increase came in at 18%, the highest in nearly a decade, and it was widely framed as a corrective spike tied to enhanced premium tax credit expiry. The corrective-spike framing carried an implicit forecast: if 2026 rates absorbed the enrollment shock, 2027 filings would revert toward the 6% to 7% trend rates that characterized the pre-2026 market. The early 2027 data invalidates that forecast.

Georgetown CHIR’s June 18, 2026 analysis of states that release rate filing data early shows 2027 proposed statewide averages clustered in the same double-digit band as 2026. Washington’s statewide average is 22.4%. In Massachusetts, Fallon Community Health Plan is requesting a 25.7% overall increase and attributes 3.3 percentage points of it explicitly to morbidity from federal policy changes (Georgetown CHIR, June 2026). MVP Health Plan in Vermont is carrying a $48 per member per month morbidity load tied directly to enhanced premium tax credit expiry effects on its remaining enrolled population (Georgetown CHIR, June 2026). These are not residual pricing adjustments on top of an otherwise stable base. They are primary drivers, quantified explicitly, filed by carriers that have a full year of post-cliff claims experience to work from.

PwC estimates the individual market’s underlying medical cost trend at 8.5% for 2027 (PwC, 2026). The gap between 8.5% medical trend and the 22.4% statewide average Washington filed is 13.9 percentage points. That spread is the morbidity and enrollment composition load. It did not stabilize between 2026 and 2027; it is re-accumulating on a base that already priced one round of adverse selection. The corrective-spike frame missed the compounding dynamic entirely.

How the 14% Nonpayment Rate Moves the Morbidity Index Before Claims Are Filed

Wakely’s April 2026 analysis, drawing on data from 75-plus issuers representing approximately 80% of the ACA individual market, found that 14% of marketplace enrollees did not pay their first January 2026 premium, with state-level variation from below 7% to above 18% across markets (Wakely, April 2026). The dollar and health implications flow from what distinguishes a first-month non-payer from a mid-year lapse.

A member who lapses in June has submitted several months of claims. Those claims establish an HCC diagnosis history. The risk adjustment transfer formula captures that history when computing risk scores and transfer payments. A member who never pays January’s premium submits zero claims, carries no diagnostic codes into the HCC system, and exits without generating any data the transfer formula acts on. The formula never compensates the carrier for that member’s expected cost contribution, because the member never entered the claims pipeline.

Wakely found that paid stayer members carried 10.2% higher morbidity than members who did not pay their first premium, and on a state-average basis, accounting for plan mix differences, the differential was 6.2% (Wakely, April 2026). Translating that into a pool-level morbidity shift, Wakely estimated 2026 morbidity increased 2.9% to 6.5% above the pre-cliff baseline. Effectuated enrollment, the members who actually paid and received coverage, is projected to fall 17% to 26% below 2025 levels, with KFF projecting average monthly effectuated enrollment dropping to approximately 17.5 million in 2026 from 22.3 million in 2025 (KFF, May 2026).

The actuarial problem for 2027 is that the 14% first-month exit happened before the risk adjustment cycle ran, which means the formula’s 2026 calibration reflects a pool that was already compositionally worse than the pre-cliff market, but calibrated to coefficient weights derived from a period when 24 million healthier enrollees were in the system. The error is baked into the base. It does not self-correct without explicit methodological intervention.

Risk Adjustment’s Calibration Lag Across Two Consecutive Cycles

The HHS-HCC risk adjustment transfer formula is calibrated to the prior year’s national pool distribution. In 2027, those coefficients will reflect 2026 experience, including the post-first-month-exit composition of a pool from which 14% of healthier members had already departed. But those departures, because they happened before any claims were filed, are not fully legible to the formula. The recalibration sees the post-exit pool, not the exit event itself, and the pre-exit comparison class has been shrinking for two years.

For a single-year shock, this lag is manageable. The formula corrects in year two, and retroactive settlement partially closes the gap. For two consecutive years of compounding adverse selection, the lag accumulates. A carrier that correctly identified the 2026 calibration-lag undercompensation and built it into its 2026 residual premium load now faces a 2027 filing where the base itself reflects the already-deteriorated 2026 pool. The lag does not restart from the pre-cliff baseline; it layers on top of a residual that was partially undercompensated in the prior year’s settlement.

Fallon Health’s explicit 3.3 percentage point morbidity attribution and MVP Health’s per-member monthly load are notable as filing evidence that individual carriers are quantifying the residual morbidity that risk adjustment does not absorb. A carrier that treats its 2027 risk adjustment receivable as a complete morbidity offset, without running an explicit calibration-lag analysis, is applying the single-year framework to a two-year compounding problem. The 2026 settlement will not have made it whole on the full accumulated lag, and the 2027 bid should not assume otherwise.

The Case for Separate Morbidity Assumption Sets in 2027 Filings

The standard ACA pricing approach treats the enrolled pool as a uniform cohort with a single aggregate morbidity adjustment factor. That approach works in a market where selection dynamics are stable and the remaining population changes composition gradually. The 2026 and 2027 individual market satisfies neither condition.

Two structurally distinct sub-populations are now identifiable within any carrier’s enrolled base. The stayer population, members who maintained continuous coverage through the 2025 open enrollment into 2026, has now been through two consecutive open enrollment cycles at materially higher net premiums. Its morbidity is measurable in 2026 claims experience and HCC distribution data. The Wakely evidence on the 10.2% morbidity differential between payers and non-payers is the clearest quantification of how this population sorts: those with chronic conditions and high utilization stayed; those with lower expected costs were more price-elastic and more likely to exit before or shortly after the coverage start date.

The second sub-population is newly enrolled or newly migrated members, including those who shifted from exiting carriers. At least six carriers announced exits from ACA marketplaces for 2027, affecting approximately 650,000 people across one-third of states (Georgetown CHIR, June 2026). Cigna, whose ACA enrollment dropped 17% year-over-year, provides a concrete example: Brian Evanko, Cigna’s president, stated the company sees no "clear path" to scaling its ACA business in the current market environment (Georgetown CHIR, June 2026). When Cigna-enrolled members migrate to remaining carriers, the migration skews toward higher-utilization enrollees, because healthier members are more likely to exit coverage entirely when their carrier leaves than to actively seek an alternative plan.

Applying a single morbidity factor to both sub-populations conflates two different risk gradients. The stayer population’s morbidity is well-characterized in 2026 claims data; the newly-migrated population’s morbidity is identifiable in a retroactive HCC score comparison on the cohort; the newly enrolling population requires an assumption calibrated to the post-cliff demographic composition rather than the pre-cliff enrollment pattern. Building separate projections and blending them at a carrier-specific weight, derived from actual 2026 enrollment composition, is the methodologically defensible path. A single aggregate trend multiplier applied to the 2026 net premium load treats a pool that has demonstrably sorted into distinct sub-populations as if it had not.

Carrier Strategic Exposure from the Concentration Effect

Texas reached a final gross rate increase of 34.7% for 2026, with silver plan rates up 35.1% and gold up 31.9% (ACA Signups, 2026). Texas is disproportionately exposed to the enrollment contraction because it never expanded Medicaid, so the ACA marketplace serves lower-income enrollees for whom the subsidy cliff creates a hard affordability cutoff rather than a relative-cost shift. The concentration effect matters for remaining carriers in states like Texas: the 4.8 million person decline in projected nationwide effectuated enrollment does not distribute evenly across markets. States with larger pre-cliff enrollment concentrated in the 400% FPL band face steeper per-carrier morbidity deterioration when those members exit.

Nationally, 9.2 million marketplace enrollees selected bronze plans in 2026, up from 7.3 million in 2025 (KFF, May 2026), a 26% increase in bronze plan share in absolute terms. Bronze migration works on both sides of the morbidity equation. Bronze enrollees carry higher deductibles that suppress induced utilization, which is a favorable cost signal. But bronze self-selection also concentrates relatively healthier members in a plan tier that generates less cross-subsidization for the risk pool as a whole. Average marketplace deductibles reached a record $3,786 in 2026, up from $2,759 in 2025 (KFF, May 2026). Carriers with high bronze concentration should model whether the induced-utilization reduction persists into 2027 or whether second-year enrollment erosion strips the healthier bronze enrollees disproportionately, leaving a bronze cohort that is compositionally sicker than the first-year bronze migration implied.

Three Methodology Decisions That Define the 2027 Filing

Working through multiple ACA rate filing cycles, the risk pool segmentation assumptions that function in a stable-enrollment environment become systematically wrong once the selection dynamic turns adversarial, and 2027 filings are being built in an adversarial environment. Three concrete methodology choices separate a defensible 2027 filing from one that repeats the 2026 estimation error.

Build the 2027 morbidity load from the sub-population structure, not from a single trend multiplied by the 2026 net premium load. The stayer population’s morbidity is measurable from 2026 claims; the migrant population’s morbidity is identifiable in a retrospective HCC score comparison on the migration cohort; the newly enrolling population requires an explicit assumption grounded in the post-cliff demographic composition. Blending all three at carrier-specific weights is slower and more document-intensive than applying a single aggregate factor, but the single-factor approach has no defensible basis when the 2026 experience demonstrates non-uniform selection across identifiable sub-populations.

Run the calibration-lag analysis explicitly for the two-year compounding case. Quantify how much of the true 2026 morbidity the formula is expected to undercompensate in 2027 settlements, add that to the residual from the 2026 settlement shortfall, and build the combined amount into the 2027 premium rather than relying on retroactive correction. The correction timeline does not align with the filing deadline in most states, and the single-year correction framework does not account for second-year accumulation.

Document the assumption that the 2026 base experience represents a fair starting estimate for the 2027 enrolled pool. If administrative enrollment barriers from pending federal rule changes will further shift the composition beyond what 2026 experience shows, that assumption requires explicit support: an enrollment projection, a morbidity index sensitivity, and a stated basis for the delta between 2026 experience and the 2027 projection. The 2017 to 2019 CSR defunding cycle showed that carriers who built one-year stabilization assumptions into their filings consistently underestimated actual morbidity in year two and beyond. The stabilization that eventually arrived came from insurer exits concentrating the remaining pool, not from genuine risk pool improvement. That is the same dynamic playing out now, without the silver-loading offset that partially contained the 2018 and 2019 financial impact. The second year of double-digit increases is not an anomaly to be priced through. It is the baseline from which the 2028 filing problem will be built.

Further Reading

Sources