The ACA Subsidy Cliff Is Back, and It Put a Step in Your FIRE Number

If you plan to retire before 65 and buy your own health insurance, your target moved this year, and not because the market did anything.

The enhanced premium tax credits that governed the last four open enrollments expired on 1 January 2026. Eligibility for the premium tax credit now ends at 400% of the federal poverty level. Below that line the credit tapers. Above it the credit is zero, with nothing in between.

That single change converts health insurance from a smooth cost curve into a step function, and a step in your annual spending becomes a step in the portfolio that has to fund it. For the illustrative household below, the step is about $217,000, or roughly 14% of the target. There is also a range of portfolio values, about $217,000 wide, where having more money leaves you with less to live on.

What actually changed

The American Rescue Plan removed the 400% ceiling in 2021 and the Inflation Reduction Act extended that removal through 2025. During those years, a household above 400% of FPL still received a credit, because the benchmark silver premium was capped at 8.5% of income no matter how high income went. That cap is gone.

Congress has not restored it. The House passed a multi-year extension in January 2026 by 230 to 196, and the Senate has not advanced it, with competing proposals still under negotiation as of that reporting. Check the current position before you act on any of this, because it is the one input here that could change with a single vote.

The effect on behaviour is already visible. Households between 400% and 500% of FPL were about 3% of 2025 marketplace enrollees and accounted for 27% of the drop in sign-ups, a decline of roughly 44% in that band. Meanwhile insurers filing 2027 rates have proposed a median increase of 14% across 77 insurers in 16 states and DC, a second consecutive double-digit year, partly because the enrollees who left were the healthy ones.

The cliff in units that matter to a retirement plan

Premium increases are the wrong frame for someone building a FIRE number. What matters is how much portfolio it takes to fund a given standard of living. So price the cliff that way.

An illustrative household. Single filer, age 60, 2026 coverage, retiring with everything in a traditional IRA so that every withdrawn dollar lands in Marketplace MAGI. The 2025 poverty guideline for a household of one is $15,650, which puts the 400% line at $62,600. Assume a full-price benchmark silver premium of $14,900 a year, and an applicable percentage near 9.96% in the top eligible band. Premiums vary enormously by age, state, and plan, so treat that $14,900 as a placeholder and get your own figure from the KFF subsidy calculator.

Withdraw exactly $62,600 and you sit on the line, still eligible:

Withdrawal, and therefore MAGI	$62,600
Premium after credit (9.96%)	$6,235
Left to live on	$56,365

Now withdraw one more dollar.

Withdrawal, and therefore MAGI	$62,601
Premium, no credit	$14,900
Left to live on	$47,701

That dollar cost $8,664.

The dead zone

The interesting consequence is not the single dollar. It is what happens across a whole range of portfolio sizes. Hold the withdrawal rate at 4% and vary the portfolio:

Portfolio	Withdrawal (MAGI)	Premium	Left to live on
$1,565,000	$62,600	$6,235	$56,365
$1,570,000	$62,800	$14,900	$47,900
$1,600,000	$64,000	$14,900	$49,100
$1,650,000	$66,000	$14,900	$51,100
$1,700,000	$68,000	$14,900	$53,100
$1,750,000	$70,000	$14,900	$55,100
$1,781,626	$71,265	$14,900	$56,365

Every portfolio between $1,565,000 and $1,781,626 funds a worse life than $1,565,000 does. Someone who retires with $1.7 million and mechanically draws 4% ends up about $3,265 a year behind someone who retired with $135,000 less.

The extra $216,626 buys back exactly what the cliff took, and nothing more. That is the step in the FIRE number: about $217,000, or 13.8% of the $1,565,000 target, to hold the same standard of living.

Two honest limits on that figure. It assumes the full-price premium stays flat in dollars as income rises, which it does, since an unsubsidized premium is a sticker price rather than a function of income. And it assumes every withdrawn dollar counts as MAGI, which is the case here only because the household holds nothing but a traditional IRA.

That second assumption is where the escape route is.

The fix is asset location, not a bigger number

You do not have to buy your way across the dead zone. The cliff is indexed to MAGI, and MAGI is not the same thing as spending.

Qualified Roth withdrawals do not appear in Marketplace MAGI. Neither does the return of your own basis in a taxable account, though the realised gain does. A household with meaningful Roth and taxable balances can spend $70,000 while reporting MAGI well under $62,600, which is not available to a household whose money is all pre-tax.

So the cliff does not really raise the FIRE number for everyone. It raises it for people whose assets are concentrated in traditional accounts, and it raises the value of Roth and taxable basis for anyone planning to bridge from an early retirement date to 65. If you are still working and still accumulating, that is the lever, and it is a lever you pull years in advance.

What this does to the conversion decision

We have argued before that Roth conversions during the bridge years are a genuine tradeoff against subsidies, and that cost-sharing reductions are the piece people forget. The cliff sharpens that argument in a specific direction.

Conversions done before you leave work, while an employer plan covers you, cost you nothing in subsidies. Conversions done during the ACA years now carry a potential $8,000-plus penalty if they push MAGI across 400% of FPL, on top of the tax. The value of front-loading conversions into your final working years went up in January, and the value of large bridge-year conversions went down for anyone near the line.

For a household comfortably below 400% of FPL, very little has changed, and the cost-sharing thresholds at 150%, 200%, and 250% remain the binding constraints. The cliff is a problem for people planning to spend somewhere between roughly $60,000 and $90,000 a year from pre-tax money.

What would change this conclusion

Congress restoring the 8.5% cap would erase the step entirely and make most of this article moot for coverage years after the change. That is a live possibility, not a remote one, and it is the reason to model both cases rather than rebuild your plan around one of them. If you are within a year or two of pulling the trigger, run your number with the cliff and without it, and make sure the decision survives both.

Two things that will not change: MAGI, not spending, is what the marketplace measures, and the account your money sits in determines how far apart those two numbers can be.

Checking your own line

Stoke's FIRE headroom planner computes your MAGI against a guardrail you select, from 150% to 400% of FPL, using the prior-year poverty table that applies to your coverage year. Set the guardrail at 400% and it will show what is left before the credit ends. The Roth conversion and capital gains headroom planner does the same for a specific conversion, showing the tax ceiling and the guardrail ceiling side by side so you can see which one binds first.

The figures above are reproducible. The worksheet is in the repository at frontend/blog/worksheets/aca-subsidy-cliff-fire-number.py.

If you are near the line for 2027 coverage, the useful window is now. Final rates land before open enrollment begins on 1 November 2026.

This article is general information, not individualised financial, tax, or legal advice. Poverty guidelines, applicable percentages, and subsidy rules are set annually and the enhanced credit rules have changed repeatedly. Confirm current figures for your own coverage year and household, and consider professional advice before acting on a conversion or withdrawal plan.