RetirementAdvanced7 min read

The decade-long Roth conversion plan: bracket-filling, IRMAA, and the ACA cliff

One-off conversion advice misses the point. The real optimization is a coordinated multi-year schedule threading tax brackets, Medicare surcharges, and subsidy cliffs.

Roth conversion articles usually answer a single-year question: 'should I convert this year, and how much?' But conversions are not a one-year decision — they're a decade-long scheduling problem. The typical window runs from retirement day to age 73-75, when required minimum distributions begin, and the goal is to move money from traditional to Roth at the lowest average rate across all those years combined. That reframing changes the answers: sometimes the optimal move is converting into a higher bracket than feels comfortable, and sometimes it's converting nothing for two years to protect a health insurance subsidy worth more than the bracket savings.

The core arithmetic: average down the mountain

Picture your traditional balance as a mountain that must eventually pass through the tax system — via conversions, RMDs, or your heirs' withdrawals. The multi-year question is how to route it through the cheapest years. A $1.5 million IRA left alone at 6% growth becomes roughly $2.5 million by age 73, forcing a first-year RMD near $95,000 — which, stacked on Social Security, can push a married couple into the 24%+ brackets for life and both spouses' Medicare premiums into surcharge territory. Converting $80,000-$120,000 annually through a decade of low-income years can flatten that mountain into the 12-22% range instead. The prize isn't avoiding tax; it's rate arbitrage across your own timeline.

ThresholdApprox. MAGI/incomeCost of crossing
Top of 12% bracket~$123,500 taxable + deductionNext dollars taxed at 22%
ACA subsidy erosion (pre-65)Scales with incomeEffective +10-15% marginal on conversions
First IRMAA tier (65+)$206,000 MAGI≈ $1,600-$2,000/yr per couple in Medicare surcharges
Top of 22% bracket~$201,000 taxable + deductionNext dollars at 24%
NIIT interaction$250,000 MAGI3.8% on investment income above line
The obstacles a conversion schedule must thread (married filing jointly, approximate 2024 thresholds)

IRMAA: the two-year time bomb

Medicare premiums are means-tested against your MAGI from two years prior — so the conversion you do at 63 sets the premium you pay at 65. IRMAA is a cliff, not a slope: one dollar over a tier boundary triggers the full surcharge for both spouses for the entire year. But it's also smaller than people fear: the first tier costs a couple roughly $1,600-$2,000 a year. That's a real cost worth avoiding when you're near a boundary — convert $205,900 instead of $206,100 — but it should almost never veto an otherwise-sound conversion. Paying one year of tier-one IRMAA to move $60,000 out of a future 24% bracket into today's 22% is frequently a winning trade. Compute it as a dollar cost, not a taboo.

The ACA cliff: the pre-65 dominator

For early retirees buying marketplace insurance, ACA premium subsidies are usually the binding constraint — more than brackets, more than IRMAA. Subsidies phase out as MAGI rises, effectively adding 10-15 cents of lost subsidy to every conversion dollar, and in years when the 400%-of-poverty-line cliff applies, one extra dollar of MAGI can erase $8,000-$15,000 of subsidy at once. The schedule implication is stark: conversion dollars are expensive at 60-64 and cheap at 65-72. Households on subsidized coverage often do token conversions (or none) pre-65, then convert aggressively in the Medicare-but-pre-RMD window.

A ten-year schedule in dollars
A couple retires at 62 with $1.4 million traditional, $250,000 taxable, and Social Security of $60,000/year planned at 70. Ages 62-64 (ACA years): convert only $30,000/year — enough to fill the standard deduction and 10% bracket — costing ~$1,000/year in tax while preserving ~$14,000/year of subsidies. Ages 65-69 (Medicare, no wages, no Social Security yet): convert $110,000/year, filling the 12% and part of the 22% bracket, paying an average ~15% and deliberately staying $5,000 under the IRMAA tier; total moved: $550,000. Ages 70-72 (Social Security running): convert $40,000/year in the 22% bracket. By 73, the traditional balance is roughly $500,000 instead of $2.3 million; the first RMD is ~$19,000 instead of ~$87,000. Lifetime tax on the converted ~$700,000 averaged about 14% versus the 24-28% (plus IRMAA, plus survivor-bracket risk) it faced as future RMDs — roughly $80,000-$110,000 of household wealth preserved.

Second-order reasons that shift the schedule

  • The survivor's brackets: after the first death, the survivor files single — same income, roughly half the bracket widths. Converting while both spouses are alive is a hedge against the widow's tax; this alone justifies more aggressive conversion for couples with large age or health gaps.
  • Heirs' tax rates: the 10-year inherited-IRA rule forces beneficiaries to drain traditional accounts during their own peak earning years. If your kids are surgeons, your 22% conversion beats their 35% withdrawal.
  • Paying tax from taxable funds: conversions are dramatically better when the tax bill comes from the taxable account, effectively moving extra wealth into the Roth wrapper. A schedule should not outrun the taxable cash available to pay it.
  • Market crashes are conversion sales: converting after a 25% drawdown moves 33% more shares per tax dollar. Keep opportunistic room in the schedule to pull a future year's conversion forward into a crash.
Conversions are irrevocable — model before you move
Recharacterization was abolished in 2018: once converted, there's no undo. A December conversion sized on estimated income can be blown up by a surprise capital gain distribution or a late bonus, cascading into IRMAA tiers and lost credits. Convert in tranches — part mid-year, the final precision tranche in December when your MAGI is nearly known — and model the whole decade in planning software or with a CPA before year one, because the optimal size of this year's conversion depends on every future year's.

Running the plan year by year

  1. 1
    Map the window

    List every year from retirement to RMD age with expected income: wages, pensions, Social Security start, taxable dividends. The low-income years light up on their own.

  2. 2
    Pick the target rate

    Estimate the marginal rate your future RMDs (and survivor, and heirs) will face. Convert in every year where today's rate is clearly below it.

  3. 3
    Set each year's ceiling

    Fill brackets up to the cheapest binding constraint that year: bracket top, ACA subsidy math pre-65, IRMAA tier at 63+.

  4. 4
    Execute in tranches

    Convert 60-70% mid-year, true up in December against actual MAGI. Pay the tax from taxable funds.

  5. 5
    Re-optimize annually

    Tax law, balances, and health change. The decade plan is a living document — redo it every January, not once at 62.

The bottom line

Roth conversion is a scheduling problem disguised as a yes/no question. The retiree's task is to route a mountain of pre-tax money through a decade of tax years at the lowest blended cost, respecting three different obstacle courses: brackets always, ACA subsidies before 65, IRMAA tiers after 63. Convert lightly in subsidy years, heavily in the Medicare-to-RMD window, opportunistically in crashes, and always with tax paid from outside the IRA. Done over ten years, the same dollars that would have been taxed at 24-28% — or at a widow's single-filer rates — pass through at 12-15%. Few retirement moves are worth six figures; for large pre-tax balances, this one routinely is.

Check your understanding

1 of 3
For an early retiree buying marketplace insurance, which constraint is usually the binding one on conversions before 65?

Not quite — try again.

The Worth letter

Get smarter about money every week

One email, no spam — practical guides and Worth updates. Unsubscribe anytime.

Put this into practice

Worth tracks your accounts, budgets, and goals — so the concepts in this article aren't just theory.

Start free trial