How Many Years It Takes a Roth Conversion to Pay for Itself

The honest answer is that “how many years” is the right question in one scenario and the wrong question in the other. If you pay the conversion tax out of the IRA itself — withholding it from the amount you convert — there is no break-even year. The conversion either wins on day one or loses forever, and which one depends only on whether your tax rate today is lower than the rate you’d have paid on that money later. Time horizon doesn’t enter the math at all. If you pay the tax with outside cash from a taxable account, then a genuine break-even period exists, and under common assumptions it lands anywhere from immediately to several decades out.

For the case where a break-even period does exist, the driver is not “how long tax-free growth needs to work.” It’s the gap between your conversion rate and your future withdrawal rate, plus how much annual tax drag the outside money would have suffered if you’d left it invested in a brokerage account instead. With a 24% conversion rate, an assumed 22% future rate, a 6% return, and roughly half a percentage point per year of drag on the taxable alternative, the break-even in the worked example below is about 18 years. Change the drag to a full 1.5 points per year and the same conversion breaks even in about 6. Same conversion, same rates — a three-fold difference in the answer, purely from an input most calculators bury. Everything below is US federal rules; state treatment varies and other countries have nothing equivalent.

Why the source of the tax payment changes the question

A Roth conversion moves money from a pre-tax account (traditional IRA, SEP, SIMPLE, or a 401(k) rollover) into a Roth IRA. The converted amount lands on your Form 1040 as ordinary income for that tax year. There is no penalty on the conversion itself at any age, but the tax bill is real and due for that year.

You can settle that bill two ways, and they are mathematically different animals.

Withhold from the conversion. Convert $100,000, have $24,000 withheld for federal tax, and $76,000 arrives in the Roth. Note that under current rules, if you’re under 59½, the withheld $24,000 is a distribution you did not convert — it’s taxable income and generally carries the 10% additional tax on early distributions, roughly $2,400 in this example, unless an exception applies.

Pay from a taxable account. Convert $100,000, all $100,000 lands in the Roth, and $24,000 comes out of your brokerage or savings. You’ve effectively moved $24,000 of taxable-account money into tax-sheltered space without using a contribution limit.

Case 1: tax paid from inside the account — the break-even year doesn’t exist

Run the algebra and the result is uncomfortable for anyone who’s been sold on a break-even timeline.

Start with a pre-tax balance P, a return r, n years, a conversion rate t_c, and a future withdrawal rate t_w.

• Leave it traditional: P(1+r)ⁿ, taxed at the end → P(1+r)ⁿ(1−t_w)
• Convert, paying tax from the balance: P(1−t_c) grows tax-free → P(1−t_c)(1+r)ⁿ

Multiplication is commutative, so n cancels out of the comparison entirely. If t_c = t_w, the two paths produce identical after-tax dollars whether you wait 3 years or 40. If t_c < t_w, the conversion is ahead the moment it’s done and stays ahead by the same percentage forever. If t_c > t_w, no holding period ever repairs it — waiting longer just makes the shortfall bigger in dollar terms.

This is the part most break-even charts obscure. When the tax comes out of the account, a conversion is a bet on tax rates, not a bet on time. “Give it fifteen years and it’ll come out ahead” is not how the arithmetic behaves.

Case 2: tax paid from outside money — now years genuinely matter

Here the comparison is between two portfolios, not two balances.

• Convert: the full P sits in the Roth, growing untaxed. Your taxable account is down by P·t_c.
• Don’t convert: P stays pre-taxed, and the P·t_c you didn’t spend stays invested in the brokerage, growing at a lower after-tax rate r_a because dividends and realized gains get taxed along the way.

Set the two after-tax totals equal and solve for n:

n = ln(t_c / t_w) / ln((1 + r) / (1 + r_a))

Three things fall out of that expression.

If t_c ≤ t_w, n is zero or negative. Converting at a rate no higher than your future rate, with outside money, is ahead immediately. No waiting period to describe.

If t_c > t_w, the break-even is driven by the drag gap, not by the size of the return. The denominator is the annual advantage of sheltered compounding over taxable compounding. A broad index fund held in a taxable account leaks maybe 0.3–0.6 points a year to qualified dividend taxes at typical rates. A high-turnover fund, a bond ladder throwing off ordinary interest, or a resident of a high-income-tax state can leak far more. That single input dominates the answer. (If you want the mechanics of why reinvested distributions still generate a tax bill even when you never see the cash, we covered that in how DRIPs work in a taxable account.)

Worked example, with stated assumptions. Assume a $100,000 conversion, tax paid from a brokerage account, a 24% marginal rate at conversion, an assumed 22% rate at withdrawal, 6% nominal returns, and a taxable alternative earning 5.5% after drag. Under those assumptions:

Conversion rate → future rate	Taxable drag 0.5 pt/yr	Taxable drag 1.5 pt/yr
24% → 24%	Break-even at year 0	Break-even at year 0
24% → 22%	~18 years	~6 years
32% → 24%	~61 years	~20 years
22% → 24%	Ahead immediately	Ahead immediately

Those are outputs of the formula above under those specific inputs, not a forecast of your result. The model also ignores the embedded capital-gains tax you’d eventually owe on the taxable side account, which pushes the true break-even earlier than the table shows — so treat these as conservative-long estimates rather than precise years.

The rate you plug in is marginal, and rarely the headline bracket

Both t_c and t_w in that formula are marginal rates on the specific dollars involved. Several mechanisms make the effective marginal rate on conversion income diverge from the number printed on the bracket table.

Social Security’s provisional income formula. Under current rules, up to 85% of benefits can become taxable, with the statutory thresholds at $25,000/$34,000 (single) and $32,000/$44,000 (joint) — figures written into the statute and not indexed. Inside the phase-in range, each extra dollar of conversion income can drag up to $0.85 of benefits into taxable income. A dollar taxed at a nominal 22% can therefore carry an effective 40.7% (22% × 1.85). This is why conversions done before benefits start behave very differently from conversions done after.

IRMAA’s two-year lookback. Medicare Part B and Part D surcharges are set from the modified AGI on your return from two years prior. Conversion income at 63 shows up in your premium at 65. The tiers are cliffs, not phase-ins — one dollar over a threshold moves you into the higher tier for the entire year — and the thresholds are re-set annually, so you’d check the figures published for the year in question rather than using an old number.

ACA premium tax credits before 65. If you’re on a marketplace plan, conversion income raises the MAGI used to size your credit. The clawback can exceed the nominal tax on the conversion itself.

Capital gains stacking and NIIT. Conversion income is ordinary income stacked underneath long-term capital gains, so it can push gains that would have been taxed at 0% up into 15%. Conversion income is not itself net investment income, but it does raise MAGI, which can drag existing investment income over the $200,000/$250,000 NIIT thresholds and trigger the 3.8% surtax on it.

State tax. Converting while resident in a state that taxes income and withdrawing later in one that doesn’t raises t_c and lowers t_w, which stretches the break-even out — sometimes past any realistic horizon. The reverse compresses it. Some states also exempt a portion of retirement income specifically, which affects only t_w.

The pro-rata rule. If you hold any pre-tax money across your traditional, SEP and SIMPLE IRAs, you can’t convert only after-tax basis. Form 8606 pro-rates the taxable share using the aggregate balance measured on December 31 of the conversion year. Someone with $10,000 of basis in a $200,000 aggregate IRA balance converting $10,000 will find 95% of it taxable, not 0%. That directly changes t_c — and therefore the break-even.

Two clocks that can delay access, separate from break-even

These don’t change the arithmetic above, but they determine when the money is reachable without extra tax.

The per-conversion five-year clock. Each conversion starts its own five-year period, backdated to January 1 of the conversion year — a December 2026 conversion is treated as starting January 1, 2026. Touch the converted amount before that period ends while under 59½ and the 10% additional tax generally applies to the taxable portion, even though the income tax was already paid. After 59½, this clock stops mattering for the penalty.

The five-year clock on earnings. Separately, growth inside a Roth IRA isn’t qualified until five tax years have passed since January 1 of the first year you funded any Roth IRA. Distributions come out in a fixed order — contributions, then conversions oldest-first, then earnings — so earnings are the last thing touched, but this clock is real for someone opening a first Roth IRA late.

And one thing that no longer exists: recharacterization of conversions was repealed for conversions after 2017. A conversion cannot be undone if the market falls or your income comes in higher than projected.

What compresses the horizon, and what stretches it past usefulness

Several mechanisms make the effective payoff period shorter than the simple two-account formula suggests.

RMDs on the pre-tax side. Traditional IRAs carry required minimum distributions from age 73 under current law, rising to 75 later this decade under SECURE 2.0’s schedule. Roth IRAs have no RMDs for the original owner, and since 2024 designated Roth accounts in employer plans no longer have them either. Forced distributions can push you into a higher bracket in years you didn’t need the money, which raises the realistic t_w well above what a naive projection assumes.

The single-filer transition. A surviving spouse generally moves to single filing thresholds the year after the death, with roughly the same income compressed into narrower brackets. That mechanically raises t_w for the surviving spouse.

Heirs and the 10-year rule. Most non-spouse beneficiaries must empty an inherited account within 10 years under the SECURE Act. An inherited traditional IRA distributes as ordinary income at the heir’s marginal rate — often their peak earning years. An inherited Roth also has to be emptied within 10 years, but the distributions aren’t taxable income. If the money is realistically going to a beneficiary, the relevant horizon isn’t your life expectancy, and the relevant t_w isn’t yours.

Pushing the other way: a conversion done at a rate meaningfully above the future rate, tax paid from inside the account, never breaks even at all — that’s the Case 1 result. Bracket-straddling has the same effect, since converting an amount that spills into the next bracket produces a blended conversion rate higher than the bracket you started in.

Running it on your own numbers

The mechanism, stripped down: identify whether the tax comes from inside or outside the account; if inside, the question is purely whether t_c < t_w and years are irrelevant; if outside, the break-even is ln(t_c/t_w) / ln((1+r)/(1+r_a)), and it’s governed by the rate gap and the taxable-account drag rather than by the return assumption. Then adjust t_c and t_w for the second-order effects — Social Security’s 1.85 multiplier, IRMAA’s two-year lookback and cliff structure, premium tax credits, gains stacking, NIIT, state residency, and the pro-rata rule — because those are what typically separate the modelled answer from the real one.

Every number in this article is a labelled assumption in a worked example, not a projection of what any specific conversion would produce. Tax rules described here are current US federal rules and are set by statute, which can change. Conversions interact with your full return, your filing status, your state, and your beneficiaries in ways a formula can’t capture, and they can’t be reversed once done — worth having a CPA or a fee-only advisor model your specific year before anything is executed.