The survivor files single on much the same portfolio — half the brackets, half the thresholds.
Figures on this page are for the 2026 tax year. Rules that are fixed in law rather than indexed — the Social Security thresholds, for instance — are noted where they come up.
This one is uncomfortable to plan for, which is probably why so few calculators model it. It's also, for most married couples, the single strongest argument for converting earlier rather than later.
When one spouse dies, the survivor's income barely changes. Their tax treatment changes completely.
The surviving spouse can file jointly for the year of the death. After that — unless they have a dependent child, which most retirees don't — they file as single.
Compare the two, on 2026 figures:
| Married filing jointly | Single | |
|---|---|---|
| Standard deduction | $32,200 | $16,100 |
| Top of the 12% bracket | $100,800 | $50,400 |
| Top of the 22% bracket | $211,400 | $105,700 |
| Social Security taxable at 85% above | $44,000 | $34,000 |
| First IRMAA threshold | $218,000 | $109,000 |
| 0% capital gains band ends | $98,900 | $49,450 |
Nearly everything halves. Meanwhile:
So the survivor is taxed as a single filer on something close to the same investment income. Every threshold arrives sooner. The Social Security torpedo bites at a lower level. IRMAA starts at $109,000 instead of $218,000.
In testing one scenario, modelling this added roughly $104,000 of lifetime tax and about tripled the measured benefit of converting early. That is one scenario, not a general result — your numbers will differ. The direction won't.
The usual comparison is "my rate now" against "my rate later". The widow's penalty says the "later" rate has a decent chance of being a single-filer rate on joint-sized assets.
That pushes in one direction: convert more, earlier, while both spouses are alive and the joint brackets are available. Bracket space that exists today may not exist for whoever is left.
It also sharpens a decision people otherwise treat as neutral. Filling the 22% bracket as a couple looks aggressive next to a 12% alternative — until you weigh it against the survivor paying 22% or 24% on the same money with half the room, plus IRMAA surcharges that the joint thresholds would have absorbed.
The widow's penalty and required minimum distributions compound.
RMDs are calculated from the traditional balance. That balance is unaffected by a death — the survivor typically rolls the deceased spouse's IRA into their own. So the survivor takes RMDs on the combined balance, as a single filer.
And RMDs grow as a percentage over time: the divisor is 26.5 at 73 and 24.6 at 75, falling every year after. A survivor in their eighties can be taking forced distributions well above 5% of a large balance, taxed in single brackets, with the resulting income setting IRMAA surcharges two years out.
Converting during the joint years attacks this at the root, because it permanently shrinks the balance the RMDs are calculated from.
There's a second-order version worth knowing. Under the SECURE Act, most non-spouse beneficiaries must empty an inherited traditional IRA within 10 years. Those withdrawals are ordinary income to them — often landing in their peak earning years, stacked on top of their own salaries.
So the "future rate" for money you never spend may not be your rate at all. It may be your children's marginal rate in their forties or fifties, which is frequently higher than yours in retirement. A Roth inherited instead comes out tax-free, and the 10-year rule still applies but costs them nothing.
This is why the calculator asks for a "future tax rate on what's left" rather than assuming one. It's the benchmark the whole verdict is judged against, and it's worth setting deliberately rather than accepting a default.
The calculator has an optional refinement for this. You set a year — the projection switches that year's filing status to single, adjusts the Social Security to the survivor's share, and shades the affected rows in the year-by-year table so you can see exactly where the tax jumps.
Two ways to use it that don't require predicting anything:
As a sensitivity test. Run it at year 10 and year 20 and see how much the conclusion moves. If converting is worth doing under both, the decision is robust and you can stop thinking about it.
As a bracket check. Simply switch the filing status to Single with everything else unchanged. That's a five-second way to see what the survivor's tax picture looks like, and it's usually more startling than people expect.
None of that says convert as much as possible. It says the "rate later" side of the comparison is probably worse than the naive version, and a plan built on the naive version under-converts.