Roth Conversions and the ACA Subsidy Cliff

The trap in one sentence

The years between retirement and Medicare at 65 are the best years for Roth conversions — low income, wide-open brackets — but they're also the years you're most likely to be buying health insurance on the ACA marketplace, where the same income that makes a conversion cheap on taxes is the income that determines your health-insurance subsidy. Push it too high and the subsidy shrinks or disappears, sometimes costing more than the conversion saves.

How ACA premium tax credits work

If you buy a marketplace (Obamacare) plan, the government may pay part of your premium through a Premium Tax Credit (PTC). The size of that credit is based on your household MAGI relative to the federal poverty level (FPL) — the lower your income, the bigger the subsidy.

The key input is MAGI, and here's what bites: a Roth conversion adds directly to it. So does an RMD, a capital gain, or interest income. Every dollar you convert raises the income figure the marketplace uses to decide how much help you get.

Two rule regimes — know which one applies

This is where it gets confusing, because the rules have changed and may change again:

• The classic "cliff" (400% FPL rule). Under the ACA's original design, subsidies phase out gradually as income rises — until you hit 400% of FPL. One dollar over that line and all subsidy vanishes at once. This is a true cliff, and it can be brutal: a 60-year-old couple can lose $10,000–$20,000+ of annual premium credit by going a few hundred dollars over.
• The enhanced subsidies (ARPA / Inflation Reduction Act). Temporary rules removed the hard 400% cliff and instead capped benchmark premiums at 8.5% of income with no upper income limit. Under these rules there's no vertical cliff — but subsidies still shrink steadily as income rises, so a conversion still has a real cost, just a smoother one.

A worked example (the hard cliff)

A couple, both 61, retired, living on cash and taxable brokerage. They have a $900k Traditional IRA they'd love to shrink before RMDs. 2025 federal poverty level for a household of two is about $21,150, so 400% of FPL is roughly $84,600.

Baseline income before conversion:     $30,000
400% FPL ceiling (household of 2):     ~$84,600
"Fill the 12% bracket" would convert:  ~$97,000  (AGI ≈ $127,000)

Conversion that stays under the cliff: ~$54,000
Conversion that fills the bracket:     ~$97,000

The extra $43k converted pushes MAGI from ~$84,000 to ~$127,000.
That crosses 400% FPL → PTC drops from ~$18,000/yr to $0.

So "filling the 12% bracket" converted an extra $43k at a nominal 12% (~$5,200 of federal tax) — but it also cost $18,000 in lost premium subsidy. The real marginal cost of that slice of conversion was north of 50%. The bracket math said "cheap"; the cliff said otherwise.

How this differs from IRMAA

The ACA cliff and the IRMAA cliff are cousins — both are income-triggered penalties layered on top of income tax — but they hit at different life stages:

• ACA applies before 65, while you're on a marketplace plan, and reacts to your income in the same year (with the reconciliation happening at tax time).
• IRMAA applies from 65 onward, on Medicare, and uses your income from two years prior.

The awkward overlap: a conversion at age 63 can trigger the ACA effect now and set up an IRMAA surcharge at 65. Both cliffs can be live in the same year. This is exactly the kind of multi-constraint sizing problem that's miserable by hand — the topic of our post on how much to convert each year.

What to do about it

1. Find your FPL ceiling for your household size, then compare it to where "fill the bracket" would land your MAGI. If the bracket ceiling is above 400% FPL, the cliff — not the bracket — is your real limit while you're on ACA.
2. Convert less now, more at 65. Many early retirees deliberately keep conversions modest during ACA years, then convert aggressively once they're on Medicare and only IRMAA constrains them.
3. Weigh subsidy lost vs. tax saved. A conversion is still worth doing if the future tax it avoids exceeds the subsidy it costs — but you have to put both numbers on the table, not just the bracket rate.
4. Mind capital gains too. Realized gains and dividends also raise ACA MAGI, so a "free" 0% capital-gains harvest can still cost subsidy. The cliffs compete for the same headroom.

How RothHelper handles it

The RothHelper calculator models ACA premium tax credits with both the 400% FPL cliff and the enhanced (8.5%-of-income) rules, and shows a net out-of-pocket premium column year by year. Because the projection tracks the age at which you move from marketplace coverage to Medicare, it applies the ACA cliff in pre-65 years and switches to IRMAA afterward automatically.

The Advisor tab's lifetime-cost ranking includes ACA premium changes — so a strategy that trips the subsidy cliff is correctly penalized, not just measured on federal tax. Look at the ACA columns in the Projections table to see exactly which conversion sizes keep your subsidy intact.