The ACA Subsidy Cliff Is Back: A 2026 Playbook for Early Retirees
For four years, early retirees got a reprieve. The enhanced premium tax credits passed in 2021 removed the income cap on ACA subsidies — earn what you wanted, and your benchmark premium was still capped as a share of income. That reprieve ended on December 31, 2025. In 2026, the subsidy cliff is back, and it came back at a bad time: insurers raised marketplace premiums roughly 26% on average this year, and KFF reports the average monthly premium payment among marketplace enrollees jumped 58%, from $113 to $178.
If you retired before 65 — or plan to — this is now one of the most important lines in your financial plan. Here is where the line sits, what pushes you over it, and how to plan around it.
Where the line is in 2026
Subsidy eligibility for 2026 coverage is based on the 2025 federal poverty guidelines. In the lower 48 states, 400% of the federal poverty level works out to about $62,600 for a single person and $84,600 for a couple. (Alaska and Hawaii use higher guidelines.)
Under the line, your benchmark silver premium is capped on a sliding scale set by the IRS for 2026 (Rev. Proc. 2025-25): from about 2.1% of income at the bottom of the scale up to 9.96% for households between 300% and 400% of FPL. A couple with $84,600 of income pays at most about $8,400 a year for the benchmark silver plan, no matter what that plan actually costs.
One dollar over the line, the cap disappears entirely. Not a phase-out — gone. That couple now pays the full sticker price of their plan, which for two people in their late fifties or early sixties routinely runs to a multiple of the capped amount, varying widely by state and age. That is why it's called a cliff: the last dollar of income can cost you thousands.
Why this is specifically an early-retiree problem
Most working people never see the cliff — they have employer coverage. Most retirees never see it either — Medicare starts at 65. The cliff lives in the gap: the years between the last paycheck and Medicare, which is exactly where early retirees spend five, ten, or twenty years.
But here's the part that makes it a planning opportunity rather than just a penalty: unlike a salary, retirement income is substantially chosen. You decide how much to withdraw, from which account, and when to realize gains. Which means your MAGI — the number the cliff is measured against — is, within limits, a dial you control.
Know what counts toward MAGI
ACA subsidies use modified adjusted gross income: your AGI, plus tax-exempt interest, untaxed Social Security benefits, and excluded foreign income.
Counts toward the line: traditional IRA and 401(k) withdrawals, Roth conversions, realized capital gains, dividends, interest, rental income, and pension income.
Doesn't count: qualified Roth IRA withdrawals, spending down cash savings, and the return-of-basis portion when you sell taxable investments — only the gain counts, not the proceeds.
That asymmetry is the whole playbook. A couple spending $90,000 a year can easily show $60,000 of MAGI — or $110,000 — depending entirely on which accounts the money comes from. Levers that pull MAGI down: spend from cash and Roth, sell high-basis lots first, harvest losses against gains, and contribute to an HSA if you're on an HSA-eligible plan (that deduction reduces MAGI directly).
The Roth conversion collision
Last week's post made the case for filling low tax brackets with Roth conversions in the gap years — for a married couple in 2026, up to $133,000 of gross income before any dollar is taxed above 12%.
If you're pre-65 on a marketplace plan, stop and reread the number above: the cliff for that same couple is $84,600 of MAGI. The bracket math and the subsidy math point in opposite directions, and the subsidy math usually wins. Converting "to the top of the 12% bracket" blows through the cliff, and the marginal cost of that crossing isn't 12% — it's 12% plus the entire lost subsidy, concentrated on the dollars that pushed you over.
Two clean ways to resolve the collision:
Subsidy-first. Keep MAGI comfortably under the line during your ACA years and defer conversions to the window between 65 and your RMD age, when Medicare has replaced the marketplace. One caution if you go this route: IRMAA looks back two years, so conversions from age 63 on can raise your Medicare premiums later — a smaller penalty than the cliff, but a real one.
Deliberate cliff years. If the conversion math is compelling, bunch it. Take one intentional over-the-cliff year — convert a large amount, pay full freight for insurance that year — then run tight, fully subsidized years around it. Ten modest conversions that each drift $10,000 over the line cost you the subsidy ten times. One big year costs it once.
What doesn't work is the middle path: being slightly over the line every year, paying full premiums annually for the privilege of small conversions.
The same collision applies to tax-gain harvesting. The 0% long-term capital gains bracket reaches $98,900 of taxable income for joint filers in 2026 — but every harvested gain counts toward MAGI. Before 65, harvesting "tax-free" gains can cost more in lost subsidy than it saves in future tax.
If you'll be over the line anyway
Some households can't reasonably get under $84,600 — pensions, rental income, or a portfolio that throws off too much yield. If that's you, the cliff stops being a planning target and becomes a shopping problem: without a subsidy there's nothing special about the benchmark silver plan, so compare bronze and HSA-eligible options on total expected cost, and check your state's actual 2026 prices before assuming the worst. And if you're near the line but over, check again in December — one deductible HSA contribution can be the difference.
Run it as one system
Notice how many systems one withdrawal decision just touched: the tax bracket, the subsidy cliff, the 0% gains bracket, IRMAA two years out, and the long-term health of the portfolio. This is the recurring lesson of the gap years — the bracket chart is never the whole picture.
Tally models your withdrawals, conversions, and realized gains as what-if scenarios inside a lifetime, tax-aware forecast — federal plus all 50 states — so you can see each scenario's income and tax picture next to the cliff numbers in this post before you move a dollar, and see what the decision does to your runway over decades, not just this April.
Start your 14-day free trial at thetally.io — no credit card required.