Paying Roth Conversion Tax From Your IRA: What It Actually Costs

Ask almost anyone whether you should pay the tax on a Roth conversion out of the IRA you are converting, and you will get the same sentence back: don’t. It is the most repeated rule of thumb in retirement tax planning, and it is correct. Paying from outside cash is better.

What the sentence does not tell you is how much better, and that turns out to be the question people actually have. A retiree whose savings are almost entirely in a traditional IRA, with no brokerage account and no meaningful cash, is not choosing between two funding methods. They are choosing between a self-funded conversion and no conversion. The rule of thumb answers a question they cannot ask.

So here is the arithmetic instead of the slogan. Every figure below comes from running the scenario through a projection engine, not from a simplified rate multiplication, and the mechanics that make self-funding expensive turn out to be more interesting than the rule suggests.

The loop that makes it expensive

Start with the obvious part. You convert $150,000. That is ordinary income, so you owe tax on it. If you pay from a checking account, the IRA gives up exactly $150,000 and the Roth receives exactly $150,000. Clean.

Now take the tax out of the IRA instead. Suppose the bill is $44,639. You withdraw $44,639 to pay it. But that withdrawal is also ordinary income, so it generates tax of its own. Now you owe more than you withdrew. You withdraw the shortfall, which generates a little more tax, which requires a little more withdrawal.

That loop is called the gross-up, and it converges: each round is smaller than the last. But it does not converge anywhere near the naive number. Funding $44,639 of tax in this example required pulling $69,751 out of the account, because the withdrawal had to pay the tax on itself.

What that 47 cents actually is

It would be easy to read $1.47 as 47 cents burned, and that is not what happens. A conversion is taxable however you pay for it. Of the $69,751 pulled out to fund the bill, $44,639 is tax that would have come out of a checking account instead. It is money leaving either way; self-funding only changes which pocket it leaves from.

The genuinely extra cost is the rest: $25,111 of additional wealth given up compared with paying from cash, which is about 17 percent of the conversion rather than 47. That is the number to weigh against the benefit of converting, and it is still large enough to change plenty of decisions. It is simply not the scariest number available, and this article is not in the business of quoting those.

One wrinkle in the comparison, kept here rather than buried: the additional tax is $26,846, slightly more than the $25,111 of additional wealth given up. The $1,735 difference is a Medicare surcharge that is not billed until 2028, so it has not left the household in the conversion year. More on that below.

Meet Dianne: $150,000, California, age 66

Dianne is 66, single, and retired in California. She has $1,000,000 in a traditional IRA and essentially nothing outside it: no brokerage account, no Roth, no cash reserve worth naming. Her Social Security benefit would have been $2,500 a month at full retirement age, but she claimed at 66 rather than 67, so it pays $28,000 a year. She is past 59 and a half, so the early-distribution penalty is not in play, and she is already on Medicare.

She is a composite, not a client, and she exists to make one thing concrete: she is exactly the person the rule of thumb cannot help, because she has no outside cash to pay with. She wants to convert $150,000.

The tax bill is not 60 percent larger because the tax rate changed. It is larger because self-funding added $69,751 of ordinary income that would not otherwise have existed, and that income is taxed at the top of the stack: federal, California, and a Medicare surcharge tier.

Notice what the last two rows do. The conversion is supposed to shrink the traditional balance, and it does. But self-funding shrinks it by $69,751 more than the conversion itself, and none of that extra lands anywhere useful for Dianne. It is the one part of the transaction with no offsetting benefit.

State tax is not a footnote here

Move Dianne to Florida, changing nothing else, and the picture shifts by a lot. The same $150,000 conversion, self-funded, requires a $45,469 withdrawal rather than $69,751, and total tax lands at $47,205 rather than $71,486.

The gap between $1.30 and $1.47 is entirely state income tax compounding through the gross-up loop. If a move across state lines is anywhere in your plan, the order of operations matters more than the conversion size.

See the gross-up on your own numbers

RetireSmartIRA models the tax-funding withdrawal, not just the conversion

Choose whether each year's conversion tax is withheld, funded from your accounts, or paid from outside money, and see what actually reaches the Roth. Free through 2026.

The effect nobody mentions: your Social Security

The gross-up is the part people can at least imagine. This one is more surprising, and at ordinary conversion sizes it matters more.

How much of your Social Security is taxable depends on your provisional income. The withdrawal you take to pay conversion tax is ordinary income, so it raises provisional income, so it pulls more of your benefit into taxation, so the tax bill grows, so the withdrawal has to grow. It is a second loop running inside the first.

Give Dianne a more cautious year. Same person, same Florida address, but she converts $40,000 instead of $150,000, which is the size most people actually pick when they are being careful.

Read the last two rows together. Dianne withdrew $5,125 and it cost her $891 in additional tax. That is 17.4 percent on dollars belonging to someone whose statutory bracket is 12 percent. The extra 5.4 points are not a bracket. They are $2,300 of her Social Security being dragged into taxable income by the act of paying the conversion tax. Nothing on her return will label it that way.

Withholding is a different trade, not a cheaper one

Custodians will happily withhold federal tax from the distribution, and the totals make it look like a bargain. In the California example, a 24 percent withholding election produces $51,980 of total tax against $71,486 for funding from your accounts. That gap is not a discount. It is a smaller conversion: $36,000 is withheld on the way through, so only $114,000 reaches the Roth instead of $150,000.

So the real choice is not cheaper against dearer. Funding from your accounts keeps the conversion whole and drains more of the IRA. Withholding leaves the extra withdrawal alone and quietly shrinks what you converted. Both are defensible. What matters is knowing which one you actually chose, because the custodian’s form will not tell you.

What is unambiguously bad is electing a rate far above what you owe. At 24 percent, the $40,000 conversion above sends $9,600 to the IRS and lands $30,400 in the Roth. The tax actually owed was $4,234. The $5,366 of over-withholding comes back as a refund the following spring, but the Roth space does not come back at all. You cannot retroactively convert it. Custodian defaults vary, and the number on the form is rarely the number you owe, so check it rather than accepting it.

What has to be true for self-funding to make sense

Everything above is a cost. None of it settles the decision, because the alternative is not free either: leaving the money in a traditional IRA means required minimum distributions later, at whatever rates then apply, possibly to a surviving spouse filing single, possibly to heirs under a ten-year deadline. Self-funding is expensive. Doing nothing has its own bill.

Four things have to hold before the expensive version is still worth doing:

• •A durable rate gap, not a one-year one: The rate you pay now has to be meaningfully below the rate the money would face later. A 2-point gap does not survive a 17-percent incremental cost. A gap created by a genuine low-income window, before Social Security starts or between retirement and RMDs, can.
• •Enough runway for the Roth to earn back the loss: You are deliberately shrinking the portfolio to change its tax character. That trade needs years of tax-free growth to repay. Late in life, with a short horizon, it usually does not.
• •Age 59 and a half, cleanly past it: The converted amount escapes the 10 percent early-distribution penalty. The separate withdrawal you take to pay the tax generally does not. Under 59 and a half with no outside cash, the numbers rarely work.
• •Enough left afterward that the plan still stands: Check the balance after the funding withdrawal, not after the conversion. Dianne's IRA drops to $780,249, and the question is whether that still funds the next thirty years. If it does not, the conversion is not a tax decision, it is a solvency decision.

Which brings it back to Dianne, and to how common she is. For a retiree whose entire net worth sits in a traditional IRA, “pay from outside money” is not advice. It is a description of somebody else’s situation. Her real choice is a smaller self-funded conversion, sized so the funding withdrawal does not cross a Medicare tier or push her benefit to its taxable ceiling, against not converting at all. Sometimes the answer is still no. It should be reached by arithmetic rather than by repeating the maxim at her.

The rule of thumb, restated honestly

Pay from outside cash if you have it. That part was never in dispute. But “never fund it from the IRA” is a rule that stops exactly where the hard cases begin, and the households most likely to benefit from a conversion window are often the ones with no cash to pay for it. The cost of self-funding is not a mystery and it is not a rule of thumb. It is $1.30 or $1.47 or some other number that depends on your state, your benefit, your age, and how much you convert, and it can be calculated before you decide rather than discovered afterward.