Partial Roth Conversions: When They Work and When They Don’t

Partial Roth conversions work when you have low-income “gap years” before RMDs, a modest IRA (roughly under $500,000 to $750,000), or charitable intent that absorbs future RMDs. They fall short when a 7% return outpaces the yearly conversion, so the IRA never shrinks, or when a decade-long plan holds you above the 2026 IRMAA threshold ($109,000 single, $218,000 joint MAGI) longer than a compressed sequence would.

A partial Roth conversion moves only part of a traditional IRA into a Roth IRA in a given year, usually calibrated to “fill the bracket” without crossing into the next one, or to stay below a Medicare IRMAA tier. An aggressive conversion instead completes larger amounts over a compressed 3 to 6 year window to substantially lower the balance before RMDs begin at age 73.

A partial Roth conversion works best in three situations: you have low-income “gap years” between retiring around age 60 and the start of RMDs or Social Security, your traditional IRA is modest (roughly under $500,000 to $750,000), or you plan significant charitable giving that will absorb RMD pressure. In each case a slow, bracket-managed pace usually captures the tax-free growth without overpaying today.

Gap years are the low-income window between retiring (often around age 60) and the start of RMDs at age 73 (age 75 if you were born in 1960 or later, first affected in 2035) plus delayed Social Security. With wages gone and RMDs not yet started, taxable income can be low enough to convert inside the 12% or 22% bracket, making these the prime partial-conversion years.

For households with roughly $500,000 to $750,000 or less in traditional IRAs, projected RMDs often stay small enough to avoid the bracket-and-IRMAA cascade that justifies an aggressive plan. A partial sequence, paced to fill a moderate bracket, can lower future RMDs enough without paying tax at the 32%, 35%, or 37% rates. The slow approach fits because the underlying RMD problem is manageable.

Two more situations favor a partial pace. If you give to charity, Qualified Charitable Distributions (available from an IRA, not directly a 401(k), starting at age 70.5) can satisfy RMDs with no taxable income, reducing the need to convert aggressively. If you lack outside cash to pay the conversion tax, a smaller yearly amount keeps the bill affordable without selling assets or dipping into the IRA itself.

Partial Roth conversions fall short in two ways that competitors rarely model. First, on a large IRA, growth can outpace the conversion: 7% on $1.5 million adds roughly $105,000 a year, more than a $60,000 to $80,000 conversion, so the balance never shrinks. Second, a decade-long slow plan can hold a household above an IRMAA tier far longer than a compressed sequence would.

A slow conversion fails to shrink the IRA when annual growth exceeds the converted amount. At a 7% return, a $1.5 million traditional IRA grows by about $105,000 in a year. A partial conversion of $60,000 to $80,000 does not keep up, so the account can be larger three or four years in than when conversions began, even though the household paid conversion tax every year.

Each conversion year that pushes modified adjusted gross income (MAGI) above the 2026 IRMAA threshold ($109,000 single, $218,000 joint) raises Medicare Part B and Part D premiums two years later, because IRMAA uses a 2-year MAGI lookback. A slow 10-year plan can hold you above a tier for that entire stretch, while a compressed 4-year sequence crosses the tier briefly, then settles back below it.

IRA millionaires usually need a more aggressive sequence because the relevant constraint is not this year’s bracket, it is the size and bracket of RMDs for the next 20 to 30 years. When future RMDs will sit in the 32%, 35%, or 37% bracket for decades, capping conversions at 24% today only defers a much larger tax. Converting more in fewer years addresses the actual problem.

To tell which strategy fits, answer five questions: your projected IRA balance at RMD age, your first-year RMD and its bracket, whether that RMD lifts you into a higher IRMAA tier or increases Social Security taxation, the after-tax difference between a 4-year and 10-year sequence, and your heir-tax exposure under the SECURE Act 10-year rule. The answers, not a single-year bracket cap, decide the pace.

The most common mistakes with partial Roth conversions are defaulting to “partial” because it feels safer, stretching a plan over 10 to 12 years and paying more cumulative IRMAA than a shorter one would, ignoring IRA growth so the plan looks better on paper than in reality, and treating a single-year bracket cap as a multi-year rule. Each error can quietly undo a plan’s intended tax savings.

A partial Roth conversion moves only a portion of a traditional IRA or 401(k) into a Roth IRA in a given year, usually sized to fill the current tax bracket without crossing into a higher one. The converted amount is taxable ordinary income, is uncapped, and is irreversible. Over a multi-year sequence the cumulative total can reach a large share of the balance.

Not always. For households with modest IRA balances or limited future RMD pressure, partial conversions are often the better fit. For IRA millionaires whose future RMDs will sit in the 32% to 37% bracket for decades, a compressed, more aggressive sequence frequently produces materially better lifetime results by lowering the balance before RMDs begin. The right answer depends on a multi-year projection.

Many households size each conversion to fill a bracket, for example up to the 2026 22% ceiling of $100,800 (joint) or the 24% ceiling of $403,550 (joint, which is where the 32% bracket begins). Higher-balance households may deliberately go further to shrink future RMDs. The right amount depends on projected RMDs, IRMAA thresholds, and available cash to pay the tax.

Yes, converted dollars are no longer subject to required minimum distributions for the original owner, so conversions reduce future RMDs. The catch is scale: if a 7% return grows a $1.5 million IRA by about $105,000 a year and the conversion moves only $60,000 to $80,000, the balance keeps rising and RMDs are barely reduced. Reducing RMDs meaningfully requires converting faster than the account grows.

A conversion raises modified adjusted gross income, and if MAGI exceeds the 2026 IRMAA threshold ($109,000 single, $218,000 joint), Medicare Part B and Part D premiums rise two years later under a 2-year lookback. The standard 2026 Part B premium is $202.90 per month, with surcharges added per tier and per person. Because of the lookback, the last conversion year that does not affect a premium is age 62.

No. Once you reach RMD age (73, or 75 if born in 1960 or later), you must take your full RMD before converting, and the RMD itself cannot be converted to a Roth. The RMD is taxable and stays outside the Roth. Only amounts above the RMD can be converted, and the conversion must be completed by December 31 of the year.

Many households convert during low-income “gap years” between retiring around age 60 and the start of RMDs and Social Security, when taxable income is lowest. A market downturn is also favorable, because the same tax moves more shares into the Roth to recover tax-free. Converting before age 63 also avoids the 2-year IRMAA lookback affecting Medicare premiums at age 65.

Yes. There is no limit on how many years you convert or how much, since conversions are uncapped and separate from annual contribution limits. Many plans convert every year for 4 to 10 years, with each year’s amount recalibrated against current income, deductions, balance, and IRMAA thresholds. Completed conversions are permanent and cannot be undone.

The choice between a partial and a more aggressive conversion depends on far more than this year’s bracket: it turns on projected RMDs, IRMAA exposure, cash flow, charitable intent, and estate priorities. The only reliable way to see which pace fits your household is a full multi-year projection built around your own numbers rather than a single-year bracket cap.