When not to do bracket filling
By Mario Bailey · Published · Last reviewed
Of the four canonical scenarios the calculator tests, do-nothing wins for a meaningful subset of inputs. There is no algebraic guarantee that converting helps; the strategy only creates value when the marginal rate paid today is lower than the marginal rate avoided later, and several common situations break that condition. The four cases below each have a specific worked example you can run to verify the result.
Small traditional balance
RMDs are proportional to the account balance. A $300,000 traditional IRA at age 75 generates a first-year RMD of roughly $12,200 (using the Pub. 590-B Appendix B Table III Uniform Lifetime divisor of 24.6). For a MFJ couple with $100,000 of other income, that $12,200 RMD barely moves the needle on marginal rates. The bracket space at 22% and 24% was already consumed by the other income; there is no RMD-rate escalation problem to solve.
Running the calculator for this case: MFJ, born 1962, $300k traditional IRA, $200k Roth, $100,000 of other income (interest and distributions), $3,000/month Social Security at FRA, retire at 65, no state tax, 6% return, 2.5% inflation, life expectancy 92. The four scenarios: do-nothing costs $422,934 lifetime; fill-12 costs $407,289; fill-22 costs $406,188 (recommended); fill-24 costs $419,169. Fill-22 saves $16,746 versus do-nothing, compared to $119,811 in the flagship $1.4M example. The conversion does produce a modest benefit here, but it is a second-order effect. If the couple has higher-priority uses for the liquidity that would go toward covering the conversion tax, do-nothing is a reasonable decision. The calculator surfaces this honestly: it recommends fill-22, but the dollar figure is close enough to baseline that the recommendation should be weighed against the cost of the conversion taxes paid in the window years.
High-income still working
Bracket filling assumes unused bracket space. A filer earning $400,000 per year has exhausted the 10%, 12%, 22%, and 24% brackets entirely. Any Roth conversion lands in the 32% or 35% marginal bracket, which is almost always higher than the projected RMD rate after wages stop. The strategy does not create value; it destroys it by accelerating tax at an unfavorably high rate.
Running the case: MFJ, born 1965, $400,000 wages, $1M traditional IRA, no Roth, retire at 70, $3,500/month Social Security at FRA, no state tax, 6% return, 2.5% inflation, life expectancy 90. Fill-22 produces zero conversions in every working year from age 61 to 69 because the $400,000 wage income already consumes the 22% bracket ceiling. Fill-22 lifetime cost: $998,845. Do-nothing lifetime cost: $998,845. The calculator correctly produces identical results because no conversions actually occur during the working years under fill-22. Fill-24 is far worse at $1,136,126 because it converts aggressively into the 24% bracket starting at retirement age 70, only three years before the RBD at 73, at a rate that is not demonstrably lower than what RMDs would have cost on the same income. Fill-12 does save $19,535, but only because it converts conservatively after retirement at 70 when the 12% bracket is briefly available.
The takeaway for working professionals in the 32%-35% bracket: the answer to "how much should I convert this year" is zero until the year wages stop. That is not a failure of the strategy; it is the strategy working correctly.
Very large traditional balance
For balances large enough that RMDs alone will push ordinary income well into the 24% or higher bracket regardless of what you do before the RBD, the question shifts. Converting at 22% before age 75 to avoid 24% after age 75 on those same dollars is sound arithmetic. But converting at 22% on a balance that will generate 24%-bracket RMDs forever is a partial win at best, because you cannot convert the full balance in any reasonable number of years.
Running the case: MFJ, born 1962, $5M traditional IRA, $200k Roth, $100k pension, $3,000/month Social Security at FRA, retire at 65, no state tax, 6% return, 2.5% inflation, life expectancy 90. Under do-nothing, first-year RMDs at age 75 are $385,833 on a balance that has compounded for eleven years. The pension plus RMD plus Social Security puts the couple deep into the 24% and 32% brackets from the first RMD year onward. Lifetime tax plus IRMAA: $2,475,851.
The calculator recommends fill-24 with a lifetime cost of $1,956,869, saving $518,982. Fill-22 costs $2,178,025, saving $297,826. The difference between fill-22 and fill-24 is $221,156 in favor of fill-24, because for a $5M account the conversion work needed to move the needle on future RMDs requires converting at the 24% rate to make meaningful progress on the balance. Converting only to the 22% ceiling moves roughly $145,000-$211,000 per year; at that pace and at 6% growth, the traditional account continues growing faster than conversions can drain it. Fill-24 accelerates the conversion, paying slightly more today to avoid paying significantly more on a much larger balance later.
The "when not to do bracket filling" version of this case is the filer who sees a large balance and assumes aggressive converting at 24% is always right. For this cohort it is, but the correct frame is that the strategy is converting at 24% to avoid 32% and 35% on RMDs, not converting at 22% to avoid 24%. If your balance is large enough that the RMD trajectory lands in the 32% bracket regardless, fill-22 will underperform fill-24 by a substantial margin.
IRA designated for charity
A traditional IRA left entirely to a qualified charity passes tax-free: the charity is a tax-exempt entity and owes no income tax on the distribution. Converting that account to Roth and paying income tax today only reduces the amount the charity ultimately receives. The conversion provides no benefit because the income tax liability that bracket filling is designed to reduce will never be owed by anyone. A partial charitable designation scales the same logic: if 40% of the IRA is earmarked for charity and 60% for taxable heirs, only the 60% fraction is a candidate for conversion analysis. The full argument is laid out in What is bracket filling?.
Related reading
The flagship worked example where fill-22 does beat do-nothing by $119,811 is in What is bracket filling? How to size the conversion in years where the window is open and bracket space is available is covered in How much should I convert to Roth this year? The window's length by cohort and how SECURE 2.0 changed the right edge for people born after 1959 is covered in The Roth conversion window.