The ACA Bridge: Health Insurance from Early Retirement to Medicare

Retire at 50 and you have fifteen years of health insurance to buy before Medicare. Under 2026 law, what you pay depends almost entirely on one number you control: your MAGI.

---

Ask anyone who actually pulled the trigger on early retirement what almost stopped them, and health insurance comes up before sequence risk does. The math of "do I have 25x expenses" is well-trodden. The math of "what does a 52-year-old pay for family coverage until 65" is murkier — and it changed materially in 2026.

This post walks through the actual rules: the contribution curve, the return of the 400% FPL cliff, the below-100%-FPL trap, and why managing your Modified Adjusted Gross Income is the single biggest lever an early retiree has. Then it covers how retireclarity models all of this, including where the model simplifies.

The 2026 Rules: The Cliff Is Back

From 2021 through 2025, the American Rescue Plan's enhanced subsidies capped everyone's benchmark premium at 8.5% of income, with no upper income limit. Those enhancements expired at the end of 2025. For 2026, the law reverted to the original ACA structure: subsidies exist only between 100% and 400% of the Federal Poverty Level, and the expected-contribution percentages are steeper.

The FPL numbers that govern 2026 coverage (published by HHS in 2025):

Between those bounds, your share of the benchmark Silver plan is capped at a percentage of income, per the IRS 2026 applicable percentage table (Rev. Proc. 2025-25):

Your subsidy is the benchmark premium minus your expected contribution. The structure to internalize: below 400% FPL, the premium you pay depends on your income, not your age. A 60-year-old couple's benchmark plan costs about $31,848 a year unsubsidized, but at $60,000 of MAGI (277% FPL) they pay about 9.3% of income — roughly $5,600 — and the subsidy covers the rest.

At 400% FPL plus one dollar, the subsidy is zero. Not reduced — zero. For a 64-year-old couple whose benchmark premium runs around $35,208 a year, crossing $86,560 of MAGI by a single dollar costs roughly $26,600 in lost subsidy (the full premium minus the ~$8,621 they were already contributing). That's the cliff, and it's the harshest discontinuity in the entire tax code for early retirees.

MAGI Is the Lever

Here's why this is a FIRE superpower rather than just a hazard: early retirees have unusual control over their MAGI. ACA MAGI is your adjusted gross income plus untaxed Social Security and tax-exempt interest. What counts:

• Traditional IRA/401(k) withdrawals — full amount
• Roth conversions — full amount, even though you're not spending it
• Capital gains — only the gain, not the proceeds
• Dividends, interest, rental income
• Total Social Security (including the untaxed portion)

What doesn't count:

• Roth withdrawals — contributions, seasoned conversions, and qualified earnings
• Return of basis when you sell taxable investments
• HSA withdrawals for medical expenses
• Cash you already have

So two households spending an identical $80,000 a year can have wildly different MAGIs. One funds spending from traditional IRA withdrawals: MAGI $80,000, close to the couple's cliff. The other spends $50,000 of Roth basis and sells $30,000 of taxable shares with $12,000 of embedded gain: MAGI $12,000... which is a problem in the other direction.

The Trap Below 100% FPL

Under 2026 law, an income below 100% FPL makes you ineligible for marketplace subsidies entirely. The ACA assumed Medicaid would catch this group — and in the 40 states (plus DC) that expanded Medicaid, it does, up to 138% FPL. In non-expansion states, there's a coverage gap: too little income for subsidies, potentially too many assets or too much "wrong-shaped" income for Medicaid.

This bites exactly the people who optimized hardest. A retiree living entirely on Roth basis and cash reports near-zero MAGI and lands below $15,960 (single) — no premium tax credit for them. The fix is deliberate: realize income on purpose. A Roth conversion sized to land your MAGI between 100% and 200% FPL costs little in tax (you're in the 10–12% brackets with a standard deduction), makes progress on your conversion ladder, and buys heavily subsidized coverage. Manufacturing income to qualify for a subsidy sounds backwards, but under this law it's often the correct move.

Conversions vs. Subsidies: The Real Tension

The same mechanism cuts both ways. Roth conversion ladders — the standard FIRE tool for penalty-free early access — generate MAGI with every rung. Convert $60,000 a year as a couple and you've consumed most of the room below the cliff before you've paid for groceries. Every conversion dollar in a bridge year is a dollar of cliff capacity spent.

There's no universal answer. Big-picture: the more years you have before 65 and the higher your benchmark premium (premiums roughly double from age 50 to 64), the more the subsidy is worth protecting. The larger your traditional balance and the closer you are to Medicare, the more conversions tend to win. It's a year-by-year optimization, which is why eyeballing it fails.

How retireclarity Models This

The projection engine computes each pre-65 year's ACA MAGI properly — ordinary income plus capital gains plus total Social Security — and applies the 2026 rules: the Rev. Proc. 2025-25 contribution curve, the hard cliff at 400% FPL, and ineligibility below 100% FPL. Years below 100% FPL are modeled as estimated Medicaid at $0 premium, and labeled as such.

The Roth Conversion Optimizer treats subsidies as real dollars — and only as real dollars. There's no special weight and no cliff fence. A lost subsidy is a premium you have to pay out of your own portfolio, so it drains balances inside the projection and shows up where every other effect shows up: in the money you keep at the end. Every candidate schedule is measured against a no-conversion baseline on exactly that number, which means a schedule that crosses the 400% line has already absorbed the full cost of crossing it, and can only win by out-earning that cost outright. The schedule marks each year whose conversion was held down by the cliff, so you can see where the line is doing the work.

Known simplifications, per the Modeling Notes:

• Benchmark premiums are estimates. If you don't enter your own quote, premiums come from a 2026 national-average benchmark (KFF's age-40 US average scaled by the federal standard age curve, ages 50–64, priced at the primary person's age and doubled for couples), grown at your medical inflation setting. Geographic spread is wide, so this is an estimate, not a quote. Younger retirees and couples with big age gaps should enter their actual healthcare.gov quote.
• FPL is projected with your inflation setting, not actual future HHS guidelines.
• The Medicaid estimate assumes an expansion state. The non-expansion coverage gap is not modeled.
• Married Filing Separately is modeled at full price (statutorily ineligible for subsidies; the narrow exceptions aren't modeled).

The point of the model isn't to predict your exact premium in 2034 — nobody can. It's to make the cliff, the curve, and the conversion trade-off visible in your own numbers, so a $1 mistake doesn't cost you $15,000. The ACA subsidy calculator puts those bridge years on one page, seeded with a sample early retiree if you don't have a plan of your own yet.

---

Related Articles

• The ACA Cliff: Why a $1,000 Raise Can Cost You $20,000 — A deeper look at cliff scenarios and when subsidies beat conversions.
• The Roth Conversion Ladder, Modeled Honestly — The other half of the bridge-years puzzle.
• Inside the Roth Conversion Calculator: How the Math Actually Works — How the optimizer weighs subsidies against the money you keep.

From the Guide: ACA Settings · ACA-Aware Optimizer · Modeling Notes