A 58 Year Old With $2.2 Million in a 401(k) Has Twelve Years to Execute the Roth Conversion That Saves $315,000 in Lifetime Taxes

The scenario lands in a lot of inboxes. A 58-year-old engineer or executive, household income north of $300,000, sees $2.2 million sitting in a traditional 401(k) and a planned retirement at 63. The instinct is to keep deferring. The math says the opposite. The next five years are the wrong time to convert, and the twelve years after that are a closing window that determines whether a future RMD schedule costs roughly $163,000 or roughly $451,000 in federal tax.

The difference, about $315,000, is the entire game.

Why the next five years favor waiting

At 58 and still earning a salary in the 24% bracket, converting traditional 401(k) dollars to Roth is mostly a bad trade. Every converted dollar stacks on top of W-2 income. For a married couple filing jointly in 2026, the IRS places the 24% bracket from $211,400 to $403,550, and a high earner is already deep inside it. A $100,000 conversion today costs $24,000 in federal tax, and state tax often rises in parallel.

Wait until the paycheck stops, and the same conversion runs through the 12% bracket (up to $100,800 for MFJ) and the 22% bracket (up to $211,400) instead. That arithmetic is the foundation of bracket-filling strategies outlined by retirement researchers including Wade Pfau and Michael Kitces. The One Big Beautiful Bill Act, signed July 4, 2025, locked the TCJA bracket structure into law permanently, removing the sunset-date uncertainty that once complicated long-range Roth conversion planning.

The twelve-year window from 63 to 75

A common planning mistake surfaces here. Most articles frame the pre-RMD window as ending at 73, when distributions begin for many retirees. Under SECURE 2.0, anyone born in 1960 or later faces an RMD starting age of 75, not 73. A 58-year-old today was born around 1968, squarely inside that cohort. The window from retirement at 63 to the first required distribution at 75 stretches to twelve years, two years longer than the older framing assumes.

The target plan: convert roughly $80,000 per year across those twelve years, totaling about $960,000. Stacked on minimal other income, those conversions fill the 12% bracket and spill into the lower half of the 22% bracket. The blended federal rate lands near 17%, for a total conversion tax of about $163,000. That is the number to hold against the do-nothing alternative.

The do-nothing path looks very different. The untouched $2.2 million compounds for seventeen years and arrives at age 75 considerably larger. The first RMD uses the IRS Uniform Lifetime Table divisor of 24.6 at age 75, producing a meaningful five-figure withdrawal that grows every subsequent year. Layer Social Security income on top, and the marginal rate climbs into the 24% to 32% range. Cumulative federal tax on RMDs alone runs toward $451,000.

One planning wrinkle worth flagging: the OBBBA also created a Senior Bonus Deduction of $6,000 per eligible taxpayer aged 65 and older, running from 2025 through 2028. A married couple where both spouses qualify can claim up to $12,000 combined. The deduction phases out above $150,000 of MAGI for joint filers, so retirees early in the conversion window who keep income modest may capture it, slightly widening the bracket-filling room in those years.

The IRMAA trap nobody plans for

Medicare eligibility starts at 65, and the premium surcharge formula uses a two-year lookback on MAGI. The conversion done at age 63 sets the Part B premium at age 65.

For a married couple in 2026, IRMAA tier one begins at $218,000 of MAGI and adds $81.20 per month per person on top of the $202.90 base Part B premium. Critically, IRMAA is a cliff system: crossing a tier by even one dollar triggers the full surcharge, not a prorated amount. An $80,000 conversion that keeps MAGI under $218,000 is clean. A $150,000 conversion pushing MAGI to $275,000 can add $200 to $400 per person per month in surcharges two years later. Sizing each year’s conversion to the IRMAA threshold matters as much as sizing it to the tax bracket.

The detail that breaks the plan

The conversion tax has to come from a taxable brokerage account. Using 401(k) dollars to pay the bill defeats the math entirely, because those dollars come out taxable and shrink the amount that ever reaches the Roth. A reader without roughly $180,000 in outside savings to cover twelve years of tax payments has to either build that cushion now or scale the conversion schedule down.

Three things to do this year

1. Build the tax-payment bucket. Redirect new savings above the 401(k) match into a taxable brokerage account for the next five years. That cash funds the conversion tax from 63 to 75 without touching the retirement account itself.
2. Model the conversion ladder against the IRMAA cliff. Run each year’s planned conversion through the 2026 MFJ brackets and the $218,000 IRMAA threshold. The correct annual figure is whichever ceiling arrives first.
3. Decide on Social Security timing before age 63. For those born in 1960 or later, full retirement age is 67. Delaying benefits past FRA and up to age 70 grows the benefit by 8% per year of deferral, a 24% total boost for waiting the full three years. Claiming early compresses the bracket-filling room and erases a large share of the Roth conversion savings.

The five-year clock is the piece that is easiest to miss. Every year of working past 58 in the 24% bracket is a year unavailable for converting. Every year past 75 is a year of mandatory withdrawals at an unfavorable marginal rate. The twelve years in between are where the $315,000 lives.

Editor’s note: This pass clarified that the OBBBA Senior Bonus Deduction is $6,000 per eligible spouse (up to $12,000 combined for a qualifying couple), added that the Social Security delayed retirement credit of 8% per year applies past full retirement age (67 for those born in 1960 or later), and noted that IRMAA operates as a cliff system rather than a phase-in. The 2026 bracket thresholds, IRMAA figures, and RMD starting age of 75 for the 1960-or-later cohort were all confirmed against current IRS and CMS data.

Contact [email protected] for any questions or corrections.