How Much to Convert to Roth: Fill the Bracket, Not Over

Margaret and David, both 66, are married filing jointly in Ohio. They retired last year with a $1.2M traditional 401(k) rolled into an IRA. This year their taxable income — a small pension plus interest, after the $31,500 standard deduction — lands at $60,000. They keep asking the same question: “How much should we convert to Roth?” The answer is not a feeling. It is arithmetic. If they want to fill the 22% bracket, they convert $146,700 this year (the 2026 MFJ 22% top of $206,700 minus their $60,000). If they want to stay conservative and fill only the 12% bracket, they convert $36,950 (the $96,950 top minus $60,000). The decision is which bracket to target. The dollar amount is just subtraction.

The three-step bracket-fill method

Every annual conversion decision reduces to three steps. Run them every January, because your income changes year to year.

1. Estimate this year’s taxable income before any conversion. Add pension, taxable interest, dividends, the taxable portion of Social Security, and any other ordinary income. Subtract the 2026 standard deduction ($31,500 MFJ, $15,750 single, plus the age-65 add-on of $1,250 each for MFJ or $1,600 for single). That is your starting taxable income.
2. Pick a target bracket ceiling. Most retirees aim for the top of the 12% bracket (a minimize-lifetime-tax posture) or the top of the 22% bracket (a drain-the-IRA-before-RMDs posture).
3. Convert the gap. Subtract your step-1 income from the bracket ceiling. That difference is your conversion room. Convert up to it — not a dollar over.

For Margaret and David at $60,000, filling the 22% bracket means $206,700 − $60,000 = $146,700. Every one of those converted dollars is taxed at 22% or less. The dollar after $206,700 would be taxed at 24% — so they stop.

2026 bracket ceilings you fill against

These are the 2026 ordinary-income bracket tops (IRC §1; IRS Rev. Proc. 2025-32). A conversion is ordinary income, so it stacks against these edges.

Note how compressed the single-filer brackets are. A single retiree fills the 22% bracket at $103,350 — barely half the MFJ edge. This is why widows and widowers face a brutal squeeze: the year a spouse dies, the survivor often files single the following year, and the same income that fit in the 12% MFJ bracket now spills into 22% or 24%. Front-loading conversions while both spouses are alive is one of the strongest cases for filling a higher bracket early.

Worked example: Margaret and David’s two choices

Same couple, same $60,000 starting income. Here is what each target bracket produces.

The 22% blended figure reflects that the first $36,950 of the conversion still fills the 12% bracket before the remainder taxes at 22%: roughly ($36,950 × 12%) + ($109,750 × 22%) = $4,434 + $24,145 = about $28,579 (rounding aside). The headline contrast is the drain timeline. At the 12% pace, the IRA is still mostly intact when RMDs begin at 73 — and those forced distributions could land the couple in the 24% or 32% bracket. At the 22% pace, the account is essentially emptied before RMDs ever start. That is the entire strategic trade.

There is no annual conversion limit — and that confuses everyone

People conflate two completely different numbers. The $7,500 figure (plus a $1,000 catch-up at 50+, IRC §219) is the annual contribution limit — how much new money you can put into an IRA. The Roth conversion rules (IRC §408A) have no dollar cap and no income limit. You can convert $10,000 or $1,000,000 in a single year. The recharacterization escape hatch was eliminated by the TCJA, so once you convert you cannot undo it — but the amount itself is unlimited.

This is why “how much can I convert” is the wrong question. The IRS will let you convert everything. The real question is how much you should convert before the marginal cost of the next dollar — a higher bracket, an IRMAA surcharge, a lost ACA subsidy — exceeds the benefit. The bracket-fill method answers that by putting a self-imposed ceiling where the tax cost jumps.

One more deadline trap: the conversion deadline is December 31 of the tax year, not April 15. Unlike a contribution, you cannot make a prior-year conversion in the spring. If you want a 2026 conversion, the money has to move out of the traditional IRA and into the Roth by year-end.

Two execution rules decide whether a bracket-filling conversion actually pays off. First, pay the conversion tax from outside the IRA. If Margaret and David convert $146,700 and withhold the roughly $28,579 of federal tax from the conversion itself, only about $118,000 lands in the Roth and the withheld portion is gone forever — and if either spouse were under 59½, that withheld amount would also trigger the 10% early-withdrawal penalty (IRC §72(t)). Paying the tax from a taxable brokerage account instead lets the full $146,700 compound tax-free. Second, each conversion starts its own five-year clock (IRC §408A(d)(2)): converted principal withdrawn before five years have passed — and before age 59½ — gets hit with the 10% penalty. For a 66-year-old couple this is moot, but for an early retiree filling brackets in their 50s it means converted dollars are not truly liquid for five years.

The reward for getting all of this right is durable. Every dollar moved into the Roth grows tax-free, never triggers an RMD during the owner’s lifetime (IRC §408A(c)(5)), and passes to heirs who can let it compound across their own 10-year inherited-account window without owing income tax on the distributions. A bracket-filling conversion strategy is really a multi-decade tax-rate arbitrage: pay a known 12% or 22% now to spare yourself — and your survivor — an unknown, likely higher rate later.

The guardrails that lower your real ceiling

The raw bracket top is rarely your true limit. Four thresholds can make the next converted dollar cost far more than its stated rate. Check each before you finalize the amount.

For a couple on Medicare, IRMAA is usually the binding constraint long before the 24% bracket. Margaret and David’s 22%-fill target of $206,700 taxable income would push their MAGI past the $206,000 first IRMAA tier — so the disciplined move is to back the conversion down to keep 2028 MAGI under that tier, or to consciously accept one surcharge tier as the price of draining the IRA faster. That is a deliberate choice, not an accident.

What most people miss: the conversion stacks under your capital gains

Here is the trap almost nobody models. Long-term capital gains and qualified dividends sit on top of ordinary income on your return, but ordinary income — including a Roth conversion — fills the brackets first. The 2026 MFJ 0% LTCG bracket tops out at $96,700 of taxable income (IRC §1(h)).

Suppose Margaret and David also have $30,000 of long-term capital gains they were planning to harvest at the 0% rate. With $60,000 of ordinary income, they have $36,700 of room left under the $96,700 LTCG threshold — so $30,000 of gains would be tax-free. But the moment they layer a $146,700 conversion on top, their ordinary income alone blows past $96,700, and every dollar of those capital gains now taxes at 15%. That is a hidden $4,500 cost the bracket-fill math misses if you look only at the ordinary brackets.

The lesson: in a year you plan a large conversion, you generally cannot also harvest gains at 0%. The two strategies fight for the same low-bracket space. Decide which one wins this year, and sequence them across different years.

Aggressive vs. conservative: pick by your expected future bracket

The whole decision collapses to one comparison: your tax rate now versus your expected marginal rate later, when RMDs begin (age 73 for those born 1951–1959, age 75 for those born 1960 or later, SECURE 2.0 §107).

• Conservative — fill the 12% bracket. Choose this when you expect future income to stay low, your traditional balance is modest, or you value certainty over speed. You pay the lowest possible rate on each converted dollar, but you drain slowly and risk leaving a large balance to face RMDs and the survivor-single-filer squeeze.
• Moderate — fill the 22% bracket. The most common retiree target. You convert meaningfully without touching the 24% rate, and for a couple with a $1M–$1.5M traditional balance, this pace usually empties the account before RMDs force higher-rate distributions.
• Aggressive — fill the 24% bracket. Justified only when you expect RMDs plus Social Security plus pension to land you in the 32%+ bracket later, or you anticipate the survivor will file single at a much higher rate. Paying 24% now to avoid 32% later is a clear win; paying 24% now to avoid 22% later is a loss.

The decision lever

The lever is not “how much can I convert” — the IRS imposes no ceiling. The lever is which bracket you choose to fill, and you set it by comparing today’s rate against the rate your future self will pay once RMDs at 73 or 75 force money out whether you want it or not. Estimate your income, pick the bracket ceiling, subtract, and convert the gap — then re-run the three steps every January, because next year’s income, the inflation-adjusted bracket tops, and your IRMAA lookback will all have moved.