Paid enrollment fell in every state but New Mexico, insurers want a median 15 percent more for 2027, and the enrollees who stay use more behavioral health.
Key Takeaways
- 2.6 million fewer payers: Roughly 2.6 million people stopped paying for marketplace coverage over the past year, and KFF projects the total could fall another 2 million by the end of 2026.
- Another double-digit year: Carriers want double digits again next year, the second such rise running, and they attribute about four points of it directly to the subsidy expiration.
- Who stays, and what they use: AM Best reports that marketplace utilization is starting to resemble Medicaid, with heavier use of emergency and behavioral health services.
- Where it reaches providers: Autism mandates and parity rules attach to these plans, so ABA and behavioral health providers face both lost commercial volume and larger patient balances.
New Mexico is the only state in the country where more people are paying for marketplace coverage this year than last. Enrollment there grew 14 percent, according to a KFF analysis of federal effectuated enrollment data, and it is the one state that fully replaced the expired federal enhanced premium tax credits with a state-funded subsidy. Every other state lost ground.
Nationally the figures are blunter. Effectuated enrollment, which counts people who actually paid a premium rather than those who selected a plan, fell from 21.8 million in February 2025 to 19.2 million in February 2026. The share of people who selected a plan and then kept it slipped from 90 percent to 83 percent nationally, and to 61 percent in Mississippi. For behavioral health and autism services organizations, that arithmetic describes a commercial book that is contracting, repricing and growing sicker at the same time, with a second open enrollment period starting November 1.
What the Marketplace Enrollment Data Shows
The 2026 open enrollment period produced 23.1 million plan selections, down from 24.2 million the year before. CMS’s national snapshot counted 3.4 million consumers new to the marketplaces and 19.6 million returning. The gap between those selections and the 19.2 million who were still paying by February is the part that matters operationally, because a patient with a lapsed plan looks identical to a covered patient until the claim is denied.
KFF projects enrollment could settle between 16.5 million and 17.5 million across 2026, depending on how many people keep up with premiums. The enhanced credits that had been in place since 2021 expired at the end of 2025, restoring the subsidy cutoff at 400 percent of the federal poverty level and removing assistance entirely from households above it. Roughly 87 percent of enrollees still qualified for some subsidy during the 2026 open enrollment period, but the average net premium payment rose sharply and average deductibles climbed to a record.
Losses were not evenly distributed by age. Adults between 18 and 34 accounted for a disproportionate share of the drop, which is the pattern insurers had predicted and the one that drives everything downstream. A risk pool loses its healthiest members first when the net price rises, because those are the members for whom the coverage was closest to optional.
A Second Consecutive Year of Double-Digit Premium Increases
For 2027, insurers are asking for more. KFF’s analysis of filings from 276 insurers across all 50 states and the District of Columbia found a median proposed increase of 15 percent, which the organization called the second consecutive year of double-digit hikes after a 20 percent median finalized increase for 2026. Proposed increases range from under 7 percent in Vermont, New York, Iowa and Utah to roughly 29 percent in Arizona. If the requests hold, typical marketplace premiums will have risen by more than a third between 2025 and 2027.
Insurers attribute most of the increase to medical prices, inflation and labor costs, the same drivers they cite every year. What is specific to this market is the morbidity shift. Carriers tied roughly four percentage points of the 2027 request directly to the expiration of the enhanced credits and the resulting change in who remains enrolled. Final rates are approved by state regulators in the fall, and the open enrollment window for 2027 coverage runs from November 1 through January 15 on HealthCare.gov.
Employer coverage is repricing on a similar curve. KFF’s parallel review of small group filings found nearly 300 insurers proposing a median 14 percent increase for 2027. For a behavioral health organization with a mixed book, that removes the obvious hedge. There is no commercial segment absorbing the pressure on behalf of the others.
Behavioral Health’s Rising Share of a Shrinking Pool
Composition change is not neutral across service lines. In a commentary issued as the credits were expiring, AM Best wrote that marketplace utilization patterns are becoming similar to Medicaid, with higher use of emergency room and behavioral health services, and that the members remaining in the pool carry higher morbidity. Younger and healthier enrollees drop coverage first when the net premium rises; people managing a chronic condition, including a psychiatric or substance use condition, are likelier to keep paying.
That produces an unfamiliar combination for providers. The number of commercially covered patients falls while the behavioral health share of the remaining population rises. Organizations treating co-occurring conditions feel it first, since those patients touch multiple service lines and the split billing structures separating mental health and addiction claims already complicate collection. Autism services sit in the same current. All 50 states require commercial plans to cover autism treatment, as the study Acuity reported on private equity’s expansion into autism therapy documented, and those mandates are part of what made ABA a predictable commercial revenue line in the first place.
Mental health and substance use treatment also sit inside the essential health benefits that individual-market plans are required to cover, which is why marketplace enrollment translates almost directly into covered behavioral health lives. A contraction in this market is therefore not a general insurance story that happens to touch behavioral health. It is a reduction in the population holding plans that are obligated to cover these services at all, arriving alongside the state-level coverage fights Acuity has tracked in places like California, where Medi-Cal’s draft all plan letter would tighten utilization management on the public side at the same time.
Deductibles, Patient Balances and the Collection Problem
Rising cost-sharing changes the economics of high-frequency services more than episodic ones. A patient who sees a psychiatrist quarterly encounters a deductible differently than a child receiving 20 or more hours of ABA a week, where family responsibility accumulates through the first months of every plan year. Providers absorb that difference as patient balances, slower collections and, eventually, bad debt.
The commercial market was already an uncomfortable place for behavioral health margins. Acuity’s reporting on the Discovery Behavioral Health debt default described a business built on commercial reimbursement that climbed more slowly than labor, real estate and compliance costs. Thinner commercial volume at higher patient cost-sharing tightens the same screw. The reimbursement gap between what evidence-based behavioral health care costs and what payers will pay does not improve when the payer has fewer, sicker members.
Network composition is the other variable to watch. The number of insurers participating per state declined in 2026 and several carriers have announced further exits for 2027. Fewer carriers in a market means fewer contracts to hold, less leverage in rate negotiations and a higher chance that a given patient’s new plan does not include a given provider. Practices that built commercial volume across four or five issuers may find that book concentrated into two.
Verification, the Subsidy Cliff and the November Window
Coverage verification deserves more attention this year than last. The gap between plan selections and effectuated enrollment means a meaningful share of patients will present with coverage that lapsed in the first months of the plan year, and providers who verify once at intake will find it in the accounts receivable aging instead.
Knowing which patients sit near the 400 percent cliff is the other practical step, since those families lose subsidy entirely rather than partially. Some will move to Medicaid, which carries its own complications given the community engagement requirements taking effect under the 2025 reconciliation law and the implementation timelines states are publishing, including California’s. Others will go uninsured, which shifts volume toward organizations with sliding-scale capacity or cost-based payment rather than commercial contracts. Integrated primary care practices running collaborative care arrangements face a version of the same problem, since that model depends on stable enrollment to sustain the caseload math.
Geography will determine how sharply any given organization feels this. New Mexico replaced the federal credits outright and grew; several states funded partial subsidies and lost less ground than their neighbors; states that did nothing absorbed the full effect. A provider operating across state lines should expect its payer mix to diverge by market this year in ways it has not before, which complicates everything built on a single national assumption, from staffing models to revenue cycle benchmarks.
Final 2027 rates arrive this fall and the enrollment picture will not be clear until February data lands. Two consecutive years of double-digit increases and a smaller, sicker commercial pool are already on the record. Behavioral health’s share of what remains rises every time a healthy enrollee walks away.
