Health cover has a cliff in it again, and one strong December can push you over

If you buy your own cover, the premium rise is the small half of this story. The big half is a line at 400% of the poverty line that you either stay under or do not, with nothing in between, and a good final quarter can put you over it without you noticing until April. Here is where the line sits, what it is worth in dollars, and which levers move it without costing you at sale.

One line decides more than the price does

For 2026 cover, a household of one crosses it at $62,600. That is four times the $15,650 federal poverty guideline for one person. A household of four crosses it at $128,600.

Stay under and there is a ceiling on what you can be asked to pay. The IRS set that ceiling for 2026 at 9.96% of your income for anyone between 300 and 400 percent of the guideline. On $62,600 that works out at $6,234.96 a year, about $520 a month, for the benchmark plan. Benchmark plan means the second cheapest silver plan sold where you live; the credit is always measured against that one, whichever plan you actually buy. Everything above the ceiling is paid by the premium tax credit.

Go over the line and the credit does not shrink. It stops. You pay what the plan costs.

That ceiling did not exist from 2021 to 2025, because Congress had temporarily lifted it. The lift ran out on 31 December 2025. The House voted 230 to 196 on 8 January 2026 to put it back for three years. The Senate has not passed it and nothing has been signed, so the rule you plan around is the one on the books.

The line itself moves with the poverty guidelines each year. The set published on 15 January 2026 puts four times the one-person figure at $63,840 and four times the four-person figure at $132,000. The 9.96% ceiling is a 2026 figure and the 2027 version has not been published yet, so treat the cap as roughly that shape rather than exactly that number.

What insurers have actually filed for next year

Rates for 2027 had to be in by 15 July 2026, so they are public before you have to choose anything.

Across 276 insurers in all 50 states and DC, the median proposed increase is 15%. Most of it is ordinary medical inflation, a median of about 10 percentage points. Insurers put roughly 4 points down to the expiry of the enhanced credits itself: healthier people left when prices rose, which leaves a sicker pool, which raises the price again. Several filed morbidity adjustments between 4.7% and 7.7% on that reasoning alone.

Now attach a dollar figure. CMS reported that for 2026, across every plan selection on the Exchanges, the average monthly premium was $619 before any credit and $178 after it. A 15% rise on $619 is about $93 a month, roughly $1,114 a year. That is the number every article this month is about.

The gap between $619 and $178 is $441 a month, $5,292 a year. That is what disappears the moment you cross the line.

The cliff is worth nearly five times the rate rise. Both of those are national averages used as an illustration, not a quote for your plan, but the ratio is the point: almost all of the advice you will read this autumn is pointed at the smaller number.

The part that is your problem and not your customers'

Someone on a salary knows what they will earn next year. You do not, and that is the whole difficulty.

Here is how it goes wrong. In November you enrol for 2027 and estimate $60,000. The Marketplace pays an advance credit straight to the insurer every month, so cover feels cheap all year. Then December is strong, a corporate order lands, a quiet January invoice gets paid early, and the year closes at $64,000.

You were over the line the whole time and had no way to know. At tax time Form 8962 reconciles what was advanced against what you were entitled to, which is zero.

Then the sharp bit. Under the line there is a cap on what you have to hand back. Over it there is none. The IRS instructions to Form 8962 limit the repayment to $375 for a single filer under 200 percent of the guideline and $1,625 between 300 and 400 percent, then state plainly that at 400 percent or more "there is no repayment limitation."

On the averages above, that extra $4,000 of profit brings about $5,292 of repayment with it, on top of the income tax and self-employment tax you owe on the $4,000 itself. The last few thousand dollars you earned cost you more than they paid.

No accountant will ever show you that as a tax rate, and it traces back to a number you guessed eleven months earlier.

Four levers that move the figure the Marketplace looks at

The credit is measured on household income, which is built on your adjusted gross income. Anything that legitimately reduces AGI reduces the number that decides which side of the line you land on.

• Deductible retirement contributions. A SEP-IRA or a solo 401(k) contribution comes off before AGI. For most owner-operators this is the largest single lever available, and it turns a credit you were about to lose into money that is still yours.
• HSA contributions, if your plan qualifies. From 2026 every bronze and catastrophic Marketplace plan is HSA-eligible in every county on HealthCare.gov, so the pairing is now available everywhere rather than in a handful of plans. The contribution is another adjustment to income.
• The health insurance deduction itself. Premiums for you, your spouse and your dependents come off on Schedule 1 as an adjustment to income, capped at your earned income from the business. Because that reduces AGI, it reduces the income the credit is measured on, and because the credit reduces the premium, each one moves the other. IRS Publication 974 gives an iterative method and a simplified method for computing the two together. This is worth an hour of your accountant's time rather than a guess from software.
• Not moving revenue into January. It works arithmetically, it is the lever owners reach for first, and it is the only one here with a price tag attached. See the last section.

If you are anywhere near the line, redo the estimate in early December, when you can see the year, not in November when you enrol. Open enrollment starts 1 November 2026 and 15 December 2026 is the last day to enrol for cover starting 1 January, so there is a fortnight in there where you know most of your year and can still change the plan you sit in.

If you have people on the payroll

There is a version of this that helps your team without you running a group plan.

A qualified small employer health reimbursement arrangement (QSEHRA) lets you reimburse staff tax free for premiums and certain costs, up to $6,450 for an employee on their own and $13,100 for one with family cover in 2026. To use one you need fewer than 50 full-time employees, no group health plan or FSA, and the same terms offered to all full-time staff.

Two things to be clear-eyed about. It reduces the premium tax credit those employees can claim, so it is not stacked on top of their subsidy, it partly replaces it. And if you are a sole proprietor you are not an employee of your own business, so it does not solve your own cover.

There is also a date on it. Current employees need 90 days of written notice before the plan year starts. For a 1 January 2027 start, that notice goes out by 3 October 2026.

What a buyer does with all of this

Your own premiums are a personal cost, and a buyer treats them as one. When they rebuild your earnings into SDE, owner health cover is a standard add-back. So paying more for cover does not lower your price, and paying less does not raise it. That whole fight is invisible at sale.

One lever is not invisible, and it is the one people pull hardest.

A buyer prices the trailing twelve months they can verify. Hold $10,000 of revenue out of the year they read, and at a 3x multiple you have spent roughly $30,000 of price protecting a credit worth about $5,000 once. Do it for three years running and you have taught your own accounts to understate the business.

The distinction that matters is which levers survive diligence. Retirement contributions and owner premiums are recognised add-backs, because they are visible, documented and clearly discretionary. Deferred invoices and soft spending are not, because nobody can prove what the business would have earned without them, and an add-back a buyer cannot verify is an add-back a buyer refuses.

So the rule is simple. Manage the line with the levers a buyer already adds back. Never manage it with timing. If you want to see the difference on your own numbers, the SDE calculator and what is my business worth do the same arithmetic a buyer will.