From 2021 through 2025, temporary legislation removed the 400% FPL cutoff entirely and made the subsidy formula more generous at every income level. Nobody, regardless of income, paid more than a capped percentage of their income toward the benchmark plan premium. That provision expired at the start of 2026 and Congress did not renew it, so the marketplace is back to the original ACA rules written in 2010.

The most consequential change for early retirees is the return of the hard cliff at 400% of the federal poverty level. Below that line, there's still a premium tax credit, just a smaller one than in 2021-2025. At or above it, in 2026 that's roughly $84,600 for a two-person household, higher for larger households, the credit doesn't shrink to zero gradually. It simply stops. A household at 405% of FPL can pay meaningfully more for the identical plan than a household at 395% of FPL.

This hits early retirees disproportionately because the ACA marketplace is often the only coverage option between leaving employer insurance and reaching Medicare at 65, and retirement income, pension payments, taxable IRA withdrawals, Roth conversions, capital gains, is one of the more controllable income streams a household has. Unlike a W-2 employee, a retiree often has real ability to manage MAGI in a given year specifically to stay under the 400% FPL line.

The math is not symmetric. Giving up a dollar of additional taxable income to stay under the 400% FPL threshold can be worth many multiples of that dollar in preserved premium tax credit for a household that would otherwise fall off the cliff entirely. This makes the marketplace income threshold one of the more important numbers in an early retiree's annual tax and withdrawal plan, arguably more consequential than the next ordinary income tax bracket.

The fix, where one exists, uses the same toolkit as other MAGI-sensitive thresholds: sequencing withdrawals from taxable versus tax-deferred versus Roth accounts, timing Roth conversions in years further from Medicare enrollment, harvesting capital losses instead of gains in a subsidy-sensitive year, and being deliberate about the difference between gross withdrawal needs and the taxable income actually generated.

If you're retired and on an ACA marketplace plan before Medicare eligibility, run your actual 2026 premium and subsidy numbers rather than assuming last year's math still applies. The rules underneath your plan changed even if your income didn't.

If your household income sits anywhere near 400% of the federal poverty level, treat that threshold with the same seriousness as an IRMAA bracket or a capital gains bracket, it deserves its own line item in year-end tax planning, not an afterthought.

If a Roth conversion, a larger-than-usual withdrawal, or a capital gain would push you over 400% FPL in a given year, run the actual subsidy loss before assuming the conversion is still worth doing that year. It may still be the right call, but the cost side of that decision just got bigger.

• Assuming this year's ACA premium subsidy will be roughly the same as last year's without rechecking the math against the reverted 2026 rules.
• Not realizing the 400% FPL cutoff is now a hard cliff again, not a gradual phase-out.
• Doing a Roth conversion or taking a larger withdrawal without first checking whether it pushes household income over the 400% FPL subsidy cliff.
• Overlooking that gross withdrawal needs and MAGI-counted taxable income aren't the same number, and only one of them matters for the subsidy calculation.