ACA Subsidy Cliff 2026: What Employers Must Know
Enhanced ACA subsidies expired, and the 400% FPL cliff is back for 2026. Here is what small employers need to know about premiums, income limits and ICHRA.
The short version
- Enhanced ACA premium tax credits expired at the end of 2025. For 2026 coverage, the 400% FPL subsidy cliff is back: cross that income line by $1 and the premium tax credit drops to $0.
- KFF projected a 114% jump in average premium payments (from $888 to $1,904 a year) if nobody changed plans. The realized increase, after plan downgrades, was 58% (from $113 to $178 a month).
- Marketplace enrollment fell from about 24.2 million to 23.1 million between the 2025 and 2026 open enrollment periods, per CMS, the sharpest single-year drop since the exchanges launched.
- A new 2026 rule removes the cap on subsidy repayment, raising the stakes for employees with variable income who estimate their income too low.
- An ICHRA's affordability test runs on W-2 wages, not household income, so it has no equivalent cliff, but the 2026 affordability percentage jumped to 9.96%, the highest on record, so contribution math needs a fresh look.
What the subsidy cliff actually is
The ACA subsidy cliff is back for 2026 marketplace coverage because the enhanced premium tax credits that eliminated it expired at the end of 2025. In plain terms: if a household's income lands at or below 400% of the Federal Poverty Level (FPL), they can qualify for a premium tax credit that shrinks as income rises. The moment household income crosses 400% FPL by even a single dollar, the credit does not taper toward zero, it disappears entirely, for the whole year. That all-or-nothing drop is the "cliff."
This is not a new invention. Before 2021, the ACA always capped premium tax credit eligibility at 400% FPL. The American Rescue Plan Act removed that cap for 2021 and 2022, and the Inflation Reduction Act extended the removal through 2025, replacing the hard cutoff with a sliding scale that capped required contributions at a percentage of income no matter how high that income went. Those enhanced credits were the reason roughly 22 million of the roughly 24 million people enrolled in ACA marketplace plans in 2025 received a subsidy, according to KFF. Congress did not extend the enhancement past 2025, so the original 400% FPL structure governs 2026 coverage, and the cliff is back exactly where it was before 2021.
Why this matters to an employer, not just an employee
An employee's individual-market premium is usually invisible to their employer, until it is not. A subsidy that disappears turns into a take-home pay problem, a benefits question at your next all-hands, or a resignation you did not see coming. Employers who understand the mechanics before employees start asking questions are in a much better position to respond.
How much premiums actually rose
There are two honest, correctly sourced answers to "how much did premiums go up in 2026," and conflating them is the most common mistake in coverage of this topic. Before the 2026 open enrollment period closed, KFF modeled what would happen if no one changed their plan selection: the average marketplace enrollee's premium payment would rise 114%, from $888 a year in 2025 to $1,904 a year in 2026, an increase of roughly $1,016. That is a projection, not a measured outcome, and it assumes static behavior.
People did not hold static. Once enrollment closed and KFF measured what actually happened, the realized increase in average monthly premium payments was 58%, from $113 to $178 a month, because a large share of enrollees responded to sticker shock by downgrading to cheaper plans. KFF's data show Bronze plan selection rose from roughly 30% to 40% of enrollees, while Silver selection fell to a record low of 43%, down from 57% the year before. The average deductible rose 37%, from $2,759 to $3,786, the steepest deductible increase KFF has recorded in this market. In other words, the realized premium increase looks smaller than the 114% headline number largely because people bought worse coverage to afford it, not because the underlying cost pressure was smaller.
2025 to 2026 ACA marketplace enrollment, by state, open enrollment period
Percent change in plan selections, 2025 OEP to 2026 OEP. Source: KFF and Newsweek, based on CMS data.
Enrollment fell in 41 states overall between the 2025 and 2026 open enrollment periods. New Mexico was a notable exception, up 18%, due to a state-funded supplemental subsidy.
Why so many people downgraded to Bronze
The gap between the 114% projected increase and the 58% realized increase is not a rounding error, it is a behavior change worth understanding on its own, because it previews what your employees are likely to do if you do not intervene. Faced with a benchmark Silver premium that suddenly cost hundreds of dollars more a month, a large share of marketplace shoppers did not absorb the increase, they downgraded. KFF's data show Bronze plan selection climbing from roughly 30% of enrollees to 40% between the 2025 and 2026 plan years, while Silver selection fell to a record low of 43%, down from 57% the year before.
Bronze plans carry lower monthly premiums in exchange for materially higher out-of-pocket costs, and the market-wide effect shows up clearly in KFF's deductible data: the average marketplace deductible rose 37% in a single year, from $2,759 to $3,786, the steepest one-year deductible increase KFF has recorded since it began tracking the marketplace. In practice, this means a meaningful share of employees who look "fine" on paper, because their premium increase was modest, actually took on thousands of dollars in additional exposure the first time they need real care. A lower premium bought with a much higher deductible is not the same thing as affordable coverage, and it is worth asking, rather than assuming, which trade-off your employees actually made this year.
This is also where an ICHRA's design advantage over a defined single group plan becomes concrete. An employee funding a Bronze plan through an ICHRA allowance made that trade-off themselves, with visibility into the full menu of Silver and Gold alternatives at their real, post-subsidy price. An employee whose employer sponsors one Silver-tier group plan has no such choice: the plan design is fixed regardless of whether that employee would rather have a lower premium, a lower deductible, or a completely different carrier network.
2026 FPL cliff thresholds by household size
For 2026 marketplace coverage, subsidy eligibility is measured against the 2025 HHS poverty guidelines, published in January 2025, not the guidelines HHS publishes in January 2026 (those apply to 2027 coverage). The 100% FPL base for a household of one in the 48 contiguous states and DC is $15,650, rising by $5,500 for each additional household member. Multiplying by four gives the 400% FPL cliff threshold, the line above which the premium tax credit disappears entirely for 2026.
| Household size | 100% FPL | 400% FPL cliff (2026) |
|---|---|---|
| 1 | $15,650 | $62,600 |
| 2 | $21,150 | $84,600 |
| 3 | $26,650 | $106,600 |
| 4 | $32,150 | $128,600 |
| 5 | $37,650 | $150,600 |
| 6 | $43,150 | $172,600 |
Two things matter about this table for an employer. First, these are household-income figures, not individual wages, so an employee's spouse's income counts, even if the employer never sees it. Second, the cliff is a single-dollar event: an employee whose household lands at $128,601 for a family of four gets nothing, while a neighbor at $128,599 can still receive a meaningful subsidy. There is no phase-out zone. That is the mechanic that makes 2026 conversations with employees near this line materially different from prior years.
22M
Marketplace enrollees who had an enhanced subsidy in 2025 (KFF)
9.96%
2026 ICHRA/ACA affordability percentage, the highest on record (IRS Rev. Proc. 2025-25)
23.1M
2026 marketplace enrollment, down from 24.2M in 2025 (CMS)
719
Counties where ICHRA currently beats small group in our dataset, 2026 plan year
Who is hit hardest, state by state
The steepest percentage declines in 2026 marketplace enrollment landed in North Carolina, down 22%, the largest drop in the country, followed by Ohio at 20%, West Virginia at 17%, and Indiana, Delaware, and Arizona each down 16%, according to CMS data reported by Newsweek and KFF. Enrollment fell in 41 states overall between the 2025 and 2026 open enrollment periods. New Mexico was a notable exception, up 18%, because the state runs its own supplemental subsidy program layered on top of the federal credit.
Florida tells a different story: not the steepest percentage drop, but the largest absolute exposure. Florida entered 2026 with a record 4.7 million marketplace enrollees, more than one in five residents, the largest marketplace population of any state. The Florida Policy Institute, citing actuarial modeling from Wakely and Health Management Associates, projects that 1.1 to 1.9 million Floridians, a central estimate of roughly 1.5 million, could lose coverage as a direct result of the 2026 subsidy changes, and that the state's uninsured rate could climb from a historic low of 10.7% in 2023 to 16.7%, the highest level since before the ACA passed.
Our own county-level dataset adds a useful cross-check. We map all 719 US counties where 2026 individual-market benchmark premiums are lower than small-group benchmark premiums, verified against CMS public-use marketplace files. Overlaying that dataset onto the states above shows an uneven picture: Ohio, one of the states with the steepest enrollment decline, has 88 qualifying counties in our dataset, led by Morrow County at a 62.3% rate advantage, worth an estimated $6,561 a year per employee. Indiana has 75 qualifying counties, led by Fayette County at 38.4%. North Carolina, despite the country's steepest decline, has only 13 qualifying counties, topped by Davidson County at 23.4%. West Virginia, Arizona, and Florida currently have zero qualifying counties in our dataset, meaning small-group coverage still wins on price in those states even as marketplace subsidies shrink, at least for plan year 2026.
| State | 2025→2026 OEP enrollment change | Qualifying ICHRA counties in our dataset | Best county-level estimate |
|---|---|---|---|
| North Carolina | −22% (steepest in the US) | 13 | Davidson County, 23.4% / ~$1,261 per employee/yr |
| Ohio | −20% | 88 | Morrow County, 62.3% / ~$6,561 per employee/yr |
| Indiana | −16% | 75 | Fayette County, 38.4% / ~$2,392 per employee/yr |
| West Virginia | −17% | 0 | Small group still wins on price in 2026 |
| Florida | Largest absolute exposure (~1.5M projected coverage loss) | 0 | Small group still wins on price in 2026 |
Georgia was not confirmed among the steepest-decline states in this research, but it is worth noting on its own terms: Georgia has 130 qualifying counties in our dataset, more than any other state we track, topped by Gordon County at a 54.9% rate advantage worth an estimated $4,431 a year per employee. Whatever is happening to marketplace subsidies there, the county-level ICHRA math is already favorable for a large share of Georgia employers.
The new repayment risk employers should flag
A change that has gotten far less attention than the cliff itself: starting with 2026 coverage, a 2025 budget reconciliation law eliminated the cap on premium tax credit repayment. Previously, if a household underestimated its income and received more subsidy than it turned out to qualify for, the amount they had to repay at tax time was capped at a fixed dollar figure that scaled with income. For 2026, that cap is gone. Marketplace enrollees who receive advance premium tax credits and then earn more than they estimated must repay the full excess amount when they file, with no ceiling.
Who this actually affects
This change matters most for employees with variable income: commissioned salespeople, seasonal or tipped workers, gig and 1099 contractors, and anyone whose year-end income is hard to predict in January. An employee who estimates conservatively to keep a subsidy, then has a strong year, can now owe the entire excess credit back, not a capped amount. W-2 employees on a fixed salary do not carry this specific risk, which is one more reason predictable, wage-based ICHRA contributions look different from marketplace subsidies this year.
Why this is an employer problem, not just a personal one
It is tempting to treat marketplace subsidies as a household finance issue that has nothing to do with a business that does not sponsor a group plan. The data says otherwise. Per HRA Council's 2025 report, more than 83% of employers who started offering an ICHRA in 2025 had never offered any group health coverage before. Their employees were, by definition, buying individual coverage on their own, which means a large and growing share of the small-employer workforce is directly exposed to exactly the premium increases and subsidy losses described above, whether or not the employer has ever thought about benefits at all.
The HRA Council also reports small-employer ICHRA adoption up 52% year over year among its member cohort, large-employer adoption up 34%, more than a million Americans now covered through an HRA arrangement, and employer renewal rates reported around 92%. That growth did not happen in a vacuum. Coverage of the 2026 subsidy cliff has been almost entirely consumer-facing, aimed at individuals shopping HealthCare.gov, not employers. Employee Benefit News, one of the few trade outlets covering the employer angle directly, has urged brokers to shift employer conversations from "does our company offer coverage" to "how exposed are our people to this year's marketplace changes," and to run affordability reviews that identify which employee segments, often in hospitality, retail, logistics, and food service, are most reliant on marketplace coverage and therefore most exposed to the cliff.
The subsidy cliff is not an abstraction for your payroll. If your people buy their own coverage, their take-home cost of staying insured changed this year, whether you sponsor a plan or not.
Mike MooreHow an ICHRA sidesteps the cliff
An Individual Coverage HRA does not eliminate the subsidy cliff, but it changes which cliff an employee is exposed to. Marketplace premium tax credits are tested against total household income, which is why a spouse's raise, a bonus, or a good freelance year can push a family over 400% FPL and wipe out a subsidy with no warning. An ICHRA's affordability test works on a completely different basis: it is measured against the employee's own W-2 wages or rate of pay, using one of several IRS safe harbors, not household income.
That distinction has two practical consequences for an employer weighing whether to fund an ICHRA this year. First, an employee's ICHRA allowance can be affordable and dependable regardless of what a spouse earns or where total household income ends up relative to the FPL table above, so it does not carry the same one-dollar, all-or-nothing exposure. Second, because the employer sets the contribution, it is a number the business controls and can budget, unlike a marketplace subsidy that depends on federal policy the employer has no influence over.
The trade-off employees actually face is the same one covered in our ICHRA vs small-group decision guide: if the ICHRA allowance is deemed affordable under the IRS test, the employee must waive the premium tax credit to use marketplace coverage funded by the ICHRA. If it is unaffordable, the employee can decline the ICHRA and try to claim a subsidy instead, assuming their household income still qualifies under the table above. Full ICHRA eligibility rules, including employee classes, are worth reviewing separately. Modeling that affordability test correctly, for every employee, before setting contribution amounts, is the single most important step in an ICHRA rollout, and it is more important in 2026 than in any year since the rule was written.
| Dimension | Marketplace premium tax credit | ICHRA affordability test |
|---|---|---|
| What it is measured against | Total household income (MAGI) | Employee's own W-2 wages, one IRS safe harbor |
| Behavior at the top of the range | Hard cliff: $1 over 400% FPL = $0 credit | No cliff; affordability is a fixed percentage test |
| Who sets the amount | Federal formula, can change with legislation | The employer, within IRS limits |
| Spouse's income counted? | Yes, full household MAGI | No, only the employee's own wages |
| 2026 key number | 400% FPL: $62,600 single / $128,600 family of 4 | 9.96% of W-2 wages (up from 9.02% in 2025) |
The 2026 ICHRA affordability number: 9.96%
Per IRS Revenue Procedure 2025-25, published July 18, 2025, the ACA affordability percentage for plan years beginning in 2026 rises to 9.96%, up from 9.02% in 2025, the highest figure recorded since the affordability requirement was created. In practice, this is the ceiling: an employer's ICHRA contribution is treated as affordable for an employee if it leaves that employee's required monthly contribution for the lowest-cost silver plan on the individual market at or below 9.96% of their W-2 wages (using the W-2 safe harbor, one of several IRS- approved methods). Employers running non-calendar-year plans continue using the 2025 rate, 9.02%, until their next plan year begins.
A higher affordability percentage sounds like good news for employers, and in one narrow sense it is: it gives a business slightly more room to set a lower contribution and still clear the affordability bar. But it cuts the other way for employees, because it also means an employer can shift more of the premium cost onto workers while remaining compliant. An ICHRA sized to have cleared the affordability test at 9.02% in 2025 may not automatically clear it again at 9.96% in 2026 if wages have not moved, so this is not a "set it and forget it" number. Re-run the affordability math for every employee class before your 2026 plan year begins, not after.
A worked example: one employee, three outcomes
Consider a single, 55-year-old employee earning $63,000 a year in W-2 wages, $5,250 a month. That income sits $400 above the 400% FPL threshold for a household of one, $62,600. The following is an illustrative example built to show the mechanics, not a quote or a promise of any specific premium; individual results depend on age, county, plan selection, and carrier participation.
No employer benefit, no ICHRA
- Household income: $63,000, just over the $62,600 cliff
- Premium tax credit: $0, the full cliff applies
- Illustrative benchmark silver premium: ~$950/month
- Employee pays the full premium out of pocket
$11,400Illustrative annual premium, unsubsidized
Employer funds a $500/month ICHRA
- Affordability ceiling: 9.96% of $5,250/mo ≈ $523/mo
- Employee's net cost after the allowance: ~$450/month
- Falls under the $523 ceiling, so the ICHRA is affordable
- Employee must waive the premium tax credit to use it
$5,400Illustrative annual employee cost, after a $6,000/yr allowance
A third outcome is worth naming even though it is not in the grid above: if the same employer instead funded a smaller ICHRA, say $150 a month, the employee's net cost would land above the $523 affordability ceiling, making the ICHRA unaffordable under the IRS test. In that case the employee could decline the ICHRA and would still be shut out of the premium tax credit anyway, because their household income already crossed the 400% FPL line, leaving them worse off than either outcome above. That is the scenario a real affordability model, run per employee before setting contribution amounts, is designed to catch before it becomes a payroll-week surprise.
Run your own counties before you model contribution amounts
The example above uses an illustrative premium. Your actual number depends on where your employees live. Ten minutes on the savings map tells you whether your counties are among the 719 where individual-market rates currently run below small-group rates, which changes how much ICHRA allowance it takes to make coverage affordable in the first place.
Edge cases worth understanding first
The FPL table and the worked example above cover the central case, but real employee situations rarely land on a clean, round number. A few edge cases come up often enough to be worth understanding before you set contribution amounts, rather than discovering them mid-year.
An employee sits exactly at the line. Because the cliff is a single-dollar event, two employees earning $62,590 and $62,610 for a household of one can have wildly different outcomes: one keeps a meaningful subsidy, the other gets nothing. If you know an employee's household income is close to their household-size threshold in the table above, that is precisely the employee an affordable ICHRA allowance helps most, because it offers a stable alternative that does not depend on which side of the line they land on.
A mid-year raise or bonus. An employee who receives a raise, a bonus, or picks up overtime partway through the year can cross the 400% FPL line without realizing it, because marketplace subsidies are trued up annually, not per paycheck. Combined with the newly uncapped repayment rule described above, an employee who does not update their marketplace income estimate after a raise can end up owing back an entire year of subsidy at tax time. Encouraging employees to update their HealthCare.gov income estimate after any significant pay change is a low-cost, high-value piece of employee communication.
A spouse loses or changes a job. Household income for FPL purposes includes a spouse's income, so a spouse's job loss can push a household back under the 400% line mid-year, restoring subsidy eligibility, while a spouse's new job or raise can push it over. An ICHRA sidesteps this specific volatility because its affordability test only looks at your employee's own wages, not what is happening in a spouse's separate job.
Medicaid expansion states. In states that expanded Medicaid, employees whose household income falls low enough may qualify for Medicaid rather than a marketplace subsidy, which is a separate program with separate rules from everything described in this article. If you have hourly or seasonal employees near the bottom of the income range, do not assume the subsidy cliff mechanics above apply to them; a benefits advisor can help identify who falls into Medicaid-eligible territory instead.
Employees under 26 on a parent's plan. An employee who is still covered under a parent's plan is not shopping the marketplace at all, and an ICHRA allowance offered to them may go unused unless they choose to enroll in their own individual coverage. It is worth confirming coverage status during enrollment rather than assuming every employee needs, or wants, the same allowance treatment.
What to do before open enrollment closes
2026 federal open enrollment opened November 1, 2025, with a January 15, 2026 standard deadline for most states, and December 15, 2025 as the cutoff to have coverage effective January 1. Several state-based exchanges, including California, New York, New Jersey, Pennsylvania, Massachusetts, Virginia, Rhode Island, and DC, run extended deadlines into late January. Whichever window applies to your state, the practical planning steps are the same regardless of exactly when your local deadline falls.
- Identify who is exposed. Pull a rough sense of which employees currently buy individual marketplace coverage and are likely to be near the 400% FPL line for their household size.
- Check your counties on the rate spread, not a national average. Whether an ICHRA is worth funding at all is a county-level question; run your census against the savings map before assuming a number.
- Model affordability at 9.96%, not last year's 9.02%. Re-run the math for every employee class using 2026 wages and the new percentage before you set contribution amounts.
- Decide your contribution with the cliff in mind, and follow our step-by-step ICHRA setup timeline. An allowance sized to clear the affordability test gives affected employees a stable alternative to a marketplace subsidy that may have just disappeared.
- Communicate before employees find out on their own. A short, plain-English note about what changed and what support is available beats employees discovering a $1,000-plus premium increase alone during open enrollment.
- Flag the repayment-cap change to variable-income employees. Anyone on commission, tips, or 1099 income who kept a marketplace subsidy should understand that under-estimating 2026 income now carries uncapped repayment risk.
Questions employers are asking
What is the ACA subsidy cliff?
It is the point at which household income crosses 400% of the Federal Poverty Level (FPL) and premium tax credit eligibility drops from a meaningful subsidy to zero, all at once, rather than phasing out gradually. Enhanced tax credits passed in 2021 removed this cliff through 2025 by capping required contributions at a percentage of income with no upper income limit. Those enhanced credits expired at the end of 2025, so for 2026 marketplace coverage the original 400% FPL cliff is back in force.
What are the 2026 400% FPL income limits by household size?
For 2026 marketplace coverage, eligibility is measured against the 2025 HHS poverty guidelines. The 400% FPL cliff sits at $62,600 for a household of one, $84,600 for two, $106,600 for three, and $128,600 for a family of four, rising by $22,000 for each additional household member. Cross that line by even one dollar and the premium tax credit falls to zero for the year.
How much did ACA marketplace premiums actually go up in 2026?
Two honest numbers exist, and they measure different things. KFF projected before open enrollment that the average enrollee premium payment would rise 114%, from $888 to $1,904 a year, if nobody changed plans. Once open enrollment closed and people had a chance to react, KFF's realized data show average monthly premium payments rose 58%, from $113 to $178, because a large share of enrollees downgraded to cheaper, higher-deductible plans to control the increase.
Does offering an ICHRA protect employees from the subsidy cliff?
Indirectly, yes. ICHRA affordability is tested against an employee's W-2 wages, not household income, so it does not have a household-income cliff the way marketplace subsidies do. An employee whose ICHRA allowance is deemed affordable can use it regardless of what their spouse earns or where their total household income lands relative to 400% FPL. It is a different mechanism entirely, and for employees at risk of losing subsidies entirely, that stability can matter more than the raw dollar amount.
Can an employee use an ICHRA and still shop the ACA marketplace?
Yes. An employee offered an ICHRA still buys an individual marketplace plan; the ICHRA simply reimburses part or all of the premium tax-free. What changes is the subsidy math: if the ICHRA is affordable under the IRS test, the employee must waive the premium tax credit to use it. If it is unaffordable, they can decline the ICHRA and claim a subsidy instead, assuming they are income-eligible.
What is the 2026 ICHRA affordability percentage?
Per IRS Revenue Procedure 2025-25, the ACA affordability percentage for plan years beginning in 2026 is 9.96%, up from 9.02% in 2025, and the highest figure since the requirement was created. An employer's ICHRA contribution is treated as affordable if it leaves the employee's required contribution for the lowest-cost silver plan at or below that share of their wages.
What happens if an employee's income changes mid-year and they already claimed a subsidy?
Starting with 2026 coverage, a 2025 budget reconciliation law removed the cap that used to limit how much excess premium tax credit a household had to repay if their actual income came in higher than estimated. Previously, repayment was capped at a fixed dollar amount depending on income tier. For 2026, marketplace enrollees must repay the full excess credit when they file taxes, which raises real exposure for employees with variable or commission-based income who estimate too low.
Is my small business actually affected by the subsidy cliff if I do not offer coverage today?
Very possibly. Per HRA Council data, more than 83% of employers who started offering an ICHRA in 2025 had never offered any group coverage before. If your employees currently rely on individual marketplace coverage and enhanced subsidies, the cliff is likely to show up as a retention or pay conversation this year whether or not you sponsor a plan, because their take-home cost of staying insured just changed.
Sources
- KFF, "ACA Marketplace Premium Payments Would More than Double on Average Next Year if Enhanced Premium Tax Credits Expire" (2025)
- KFF, "What We Know So Far About 2026 ACA Marketplace Enrollment, Premiums, and Deductibles" (May 2026)
- KFF, "8 Things to Watch for the 2026 ACA Open Enrollment Period" (October 2025)
- CMS, "Exchange Coverage Remains Near Record High: 23.1 Million Enroll for 2026" (March 2026)
- CMS, "Over 24 Million Consumers Selected Affordable Health Coverage" (January 2025)
- healthinsurance.org, "Federal Poverty Level" glossary and 2026 FPL cliff explainer
- Newsweek, "The States Seeing the Biggest Obamacare Enrollment Drops" (June 2026)
- Florida Policy Institute, "Raising the Alarm on the Oncoming Tidal Wave of Health Care Coverage Loss for Florida" (2025)
- HRA Council 2025 report, via PR Newswire and remodelhealth.com
- IRS Revenue Procedure 2025-25 (2026 ACA affordability percentage), via HUB International compliance bulletin
- Employee Benefit News, "The post-subsidy ACA reality: What it means for benefits and the evolving broker role"