When Congress let the enhanced premium tax credits lapse on December 31, 2025, almost every headline framed it as a story about individual buyers watching their marketplace premiums jump. That is real. The quieter story, and the one that matters for any company with 20 to 500 employees, is what the same week did to the value of the health plan you already offer.

TL;DR
  • The expanded ACA subsidies expired on December 31, 2025, and the 400 percent poverty cliff returned.
  • In the same quarter, the IRS raised the 2026 employer affordability threshold to 9.96 percent, the highest figure since the rule existed.
  • Combine the marketplace data with the employer data and one number appears: for a worker earning just above $62,600, your health plan is now worth roughly $6,000 a year in avoided premium, up about $2,000 from a year earlier.

Two policy changes landed in the same three months

The first change was the expiration of the enhanced premium tax credits. These were the richer subsidies created in 2021 and extended through 2025. A Senate bill to push them to 2028, the Lower Health Care Costs Act, could not reach the 60 votes it needed in December 2025, so the expansion ended on schedule. The older, thinner subsidy structure snapped back into place on January 1.

The second change came from the IRS. Revenue Procedure 2025-25, released in July 2025, set the 2026 affordability percentage for employer coverage at 9.96 percent of income. That is up from 9.02 percent in 2025 and is the highest the figure has ever reached. It sounds like a footnote. It is actually the lever that decides how much of the premium an employer can ask a worker to carry while still passing the Affordable Care Act test.

On their own, each change generated its own trade-press coverage. Put side by side, they point the same direction, and that is the part almost nobody has written down.

What the 2026 marketplace actually looks like

Start with the individual market, because that is the alternative your employees are quietly comparing you against. KFF puts the average 2026 benchmark silver premium at $625 a month before any subsidy. Insurers raised marketplace rates about 26 percent on average for the year, roughly 17 percent in states that run their own exchanges and near 30 percent on HealthCare.gov.

The sticker increase is only half of it. Because the richer subsidies are gone, the amount enrollees actually pay out of pocket is climbing far faster than the sticker. KFF projects the average subsidized enrollee will see their share rise about 114 percent, with the typical monthly payment moving from roughly $888 to roughly $1,904. Same coverage, more than double the cost to the person buying it.

The 400 percent cliff is back, and it lands at $62,600

The single most important structural detail for employers is the return of the subsidy cliff. Under the expired formula, subsidies phased out gradually with income and never fully disappeared. In 2026 they cut off entirely above 400 percent of the federal poverty line. For a single person in the continental United States that line sits at $62,600. For a family of four it is $128,600. Alaska and Hawaii sit higher, near $78,200 and $71,960 for a single person.

Cross that line by a dollar and the marketplace subsidy goes to zero. A worker at $62,700 pays the full $625 benchmark with no help, while a coworker at $62,500 still gets something. This is a cliff, not a ramp.

Do not confuse the 26 percent with the 114 percent

Two very different numbers get quoted about 2026, and mixing them up understates the shift. The 26 percent is the rate hike insurers filed, the change in the gross premium before any help. The 114 percent is the change in what a subsidized enrollee actually pays after the thinner subsidy is applied. The second number is larger because the subsidy used to absorb most of the premium and now absorbs far less. When you compare your plan against a worker's outside option, the 114 percent figure is the one that matters, because it reflects the check they would really write. Anchoring on the 26 percent makes the alternative look four times cheaper than it is.

The band nobody is pricing correctly

Look at who actually left the marketplace when the rules changed. Enrollees between 400 and 500 percent of poverty were only about 3 percent of 2025 sign-ups, but they accounted for 27 percent of the drop into 2026. Sign-ups in that band fell 44 percent, more than 321,000 people gone in a single year.

Hold that group in your mind, because it is not an abstract slice of the population. A single worker earning $63,000 to $78,000, or a married couple with kids in the low six figures, is a mid-career employee at exactly the kind of firm reading this. When their subsidy vanished, most did not find a cheaper option. They found their employer.

Now the employer side of the ledger

Here is where the 9.96 percent figure earns its keep. To pass the affordability test in 2026, the cost of your lowest single plan cannot exceed 9.96 percent of a worker's income, measured through one of three safe harbors: W-2 wages, rate of pay, or the federal poverty line. Under the poverty-line safe harbor, a plan is affordable as long as the employee share stays at or below $129.90 a month for single coverage in the continental states.

What do employers actually charge? KFF's 2025 Employer Health Benefits Survey pegs the average worker contribution for single coverage at $1,440 a year, about $120 a month, which is 16 percent of the single premium. Family coverage runs $6,850 a year. Smaller and mid-size firms, those with 10 to 199 workers, ask more from workers on family plans than large firms do: $8,889 versus $6,227.

So the typical single worker pays around $120 a month for employer coverage, and the ceiling before the plan becomes legally unaffordable is $129.90.

Putting the two datasets together

This is the calculation the rest of the web has not run. Take one worker, a single employee earning $62,700, one notch above the cliff, and price both of their options in 2026.

Option for a single worker at $62,700 2025 2026
Marketplace benchmark silver, out of pocket about $5,320/yr $7,500/yr
Employer single coverage, worker share about $1,440/yr $1,440 to $1,559/yr
Value of the employer plan to that worker about $3,900/yr about $6,000/yr

In 2025, with the richer credits still active, that worker's marketplace cost was capped near 8.5 percent of income, roughly $5,320 a year, so your plan saved them close to $3,900. In 2026 the cliff strips the subsidy to zero, the full $625 benchmark becomes $7,500 a year, and the same plan now saves them about $6,000. The employer coverage advantage for this one worker grew roughly $2,000 in twelve months, and none of it came from anything the employer did.

Why the two forces compound instead of cancel

The reason the number moves so much is that both policy changes push the same way. The affordability threshold rising to 9.96 percent lets an employer shift more premium onto workers while still clearing the legal bar. At the very same time, the loss of those subsidies made the outside option, the individual marketplace, dramatically more expensive for the mid-wage band. One force widens what you can charge. The other force weakens the alternative your employees could flee to. They stack.

A year ago these two markets sat close enough that a healthy single worker could shrug at losing group coverage and pick up a subsidized silver plan for a similar net cost. That arbitrage is gone for anyone near or above the cliff, and the closest alternative they could buy now costs far more.

What this means for retention math

Benefits leaders talk about coverage as a cost center. The 2026 numbers turn part of it into a retention asset with a measurable dollar value. For every employee sitting in the 300 to 500 percent poverty band, your plan is now worth several thousand dollars a year more than the closest option they could buy on their own, and that gap is largest at smaller firms where workers already pay a bigger share of the premium than their peers at large employers.

The catch is that the value is invisible unless someone surfaces it. A worker who has never priced a marketplace plan has no idea their fallback got 114 percent more expensive. The retention dividend only pays out if the employee understands what they would lose by leaving.

Families sit on a steeper version of the same cliff

The single-worker math is the clean case. Family coverage makes the gap wider and, for smaller employers, wider still. The family subsidy cliff in 2026 lands at $128,600 for a household of four. A family that clears it by a little loses every dollar of marketplace help at once, and family benchmark premiums run far above the single figure.

Meanwhile KFF's employer survey shows the average worker pays $6,850 a year toward family coverage, and at firms with 10 to 199 employees that number climbs to $8,889. That is real money out of a paycheck. Even so, an unsubsidized family silver plan on the exchange in most markets runs well past $20,000 a year in gross premium. A worker paying $8,889 through work against a $20,000-plus unsubsidized alternative is looking at an employer plan worth more than $11,000 a year to their household. The cliff turned a close call into a lopsided one.

The safe harbor you pick changes the ceiling

Affordability is measured against income, and the 9.96 percent test runs through one of three yardsticks. The poverty-line safe harbor caps the single employee share at $129.90 a month regardless of what the worker earns, which is the friendliest option for lower-wage staff. The rate-of-pay safe harbor uses hourly rate times 130 hours, so a worker at $25 an hour has a monthly income figure of $3,250 and an affordability ceiling near $324 a month. The W-2 safe harbor keys off Box 1 wages. A firm with a wide wage range often mixes these, using the poverty-line method for its hourly tier and W-2 for salaried staff. Getting the choice right is the difference between a clean filing and a penalty exposure under section 4980H.

Three moves that turn the shift into a plan

Re-run your affordability math against the new threshold

The 9.96 percent figure and the $129.90 poverty-line safe harbor change what qualifies as affordable. Confirm your lowest single plan still clears the test under the safe harbor you use, then decide whether the extra headroom is worth using or holding. The ACA compliance checker lets you test each safe harbor against your actual wage data instead of guessing, and the 2026 affordability rules guide walks through the contribution limits in detail.

Map your workforce against the cliff

Pull a wage distribution and mark everyone between roughly $47,000 and $78,000 in single income, plus married workers in the low six figures. Those are your cliff-exposed employees, the ones for whom your plan quietly became far more valuable. This is also the group most at risk if a competitor offers richer coverage, because they can no longer backstop themselves cheaply on the exchange.

Communicate the real replacement cost

At open enrollment, show the number. A worker paying $120 a month for single coverage is looking at $625 a month for a comparable unsubsidized silver plan if they walk. Put the two figures next to each other in plain language. The level-funded calculator and the health funding projector help you model what your own plan costs to run, so the savings message rests on your real numbers.

Where ICHRA fits after the cliff

Individual coverage HRAs got more interesting and more delicate at the same time. An ICHRA lets you fund employees to buy their own marketplace plan, and if the reimbursement makes that plan affordable under the 9.96 percent rule, the worker cannot also claim a premium tax credit. In 2025 that trade-off stung, because the credit was generous. In 2026 the credit is worth far less for the mid-wage band and nothing at all above the cliff, so pairing an ICHRA with a solid contribution now competes better than it did. For a firm with a wide geographic spread or a young workforce, it is worth modeling side by side with a traditional group plan rather than assuming last year's answer still holds. The ICHRA calculator and the ICHRA versus group cost comparison let you run both structures against your census before you commit.

What could change this, and what probably will not

Two caveats keep this honest. First, Congress could still act. Extension bills keep getting filed, and a future deal could restore some version of the richer credits mid-year, which would soften the cliff for the affected band. Nothing in the December 2025 votes suggested that is close, but benefits planning should stay flexible enough to absorb a reversal. Second, a handful of states have stood up their own premium assistance to fill part of the gap. Most are thinner than the federal help that lapsed, and they vary widely by state, so a worker in one state may have a cushion a worker next door does not. Check your own states before you assume the cliff bites evenly across a multi-state workforce.

Neither caveat changes the core direction for 2026 as it stands. The employer affordability ceiling is at a record high and the marketplace fallback is materially weaker for mid-wage workers. Plan for the world you are in, and adjust if Washington changes it.

The bottom line

Two datasets, one from the IRS and one from KFF, tell a single story when you read them together. The employer affordability threshold hit a record 9.96 percent while the individual marketplace lost the subsidies that made it a viable substitute. The result is a specific, countable shift: for a mid-wage employee just over the $62,600 line, the health plan you already offer is worth about $6,000 a year, roughly $2,000 more than it was in 2025. That value is sitting on your benefits statement right now. Whether it does anything for retention depends on whether you measure it and say it out loud.