Five-Year Roth Conversion Plan

This is a composite. The couple isn’t real. The five-year window is, and so is how easy it is to waste it.

Meet Frank and Susan. He’s 63 and just retired from engineering. She’s 61 and stepped back from her own career last year. They have $3.9 million, most of it, around $2.9 million, in traditional IRAs and his old 401(k). No pension. They’ve delayed Social Security. Frank, being an engineer, had already read about Roth conversions and arrived with a question that sounded smart: should we just convert it all now and be done?

No. And the reason why is the whole point.

Why the window exists

A Roth conversion means deliberately paying income tax on traditional IRA money now, so it grows tax-free afterward and never triggers a required withdrawal. The catch is that the conversion itself counts as income in the year you do it. Convert too much at once and you pay top-bracket rates on the back half, which defeats the purpose.

Frank and Susan had stumbled into something valuable without seeing it. From his retirement at 63 until RMDs begin, the IRS-mandated withdrawals from tax-deferred accounts, they would have years of unusually low income. No salaries, no Social Security yet, no forced withdrawals. For a household their size, that’s roughly a five-plus-year stretch of empty space in the low tax brackets.

That empty space is the most valuable thing they own right now, and it’s completely invisible on a tax return. It only shows up when you look forward instead of back.

The cost of converting all at once

Run Frank’s “do it all now” idea and it falls apart fast. Dumping $2.9 million of conversions into a year or two would stack income sky-high and tax the bulk of it at the 37% federal rate, which for 2026 starts at $768,700 of taxable income for a married couple. You’d voluntarily pay the highest rate in the code to avoid a future rate that’s almost certainly lower. That’s not a strategy. That’s a panic with a spreadsheet.

The opposite mistake is just as real: convert nothing, enjoy the low-tax years, and let the balance compound until RMDs at 75 force out enormous taxable withdrawals and spike Medicare premiums through IRMAA, the income surcharge. Doing nothing has a price too. It just hides better.

The plan, with a ceiling

So we built a multi-year ladder with a hard ceiling. Each year we convert traditional IRA money up to the top of a target bracket and then stop. The discipline is in the stop.

We anchored it to the 24% bracket, which for 2026 runs up to $403,550 of taxable income for a married couple. That gave us a wide, sane lane: convert enough to make real progress, never so much that we tipped into the 32% rate or jumped a Medicare IRMAA tier two years down the road. Every December we recheck. A year with a big capital gain means we convert less. A quiet year means we convert more. The ceiling moves the plan; the plan doesn’t blow the ceiling.

We also paid the conversion tax from their taxable brokerage account, not from the IRA itself. That matters. It lets the full converted amount keep compounding inside the Roth, which is the whole engine of the thing.

The outcome

Year one looked boring. It’s supposed to. There’s no fireworks in a well-run conversion, just a tax return that’s a little higher on purpose.

Across the five years, we expect to move the majority of that $2.9 million into Roth at middle-bracket rates, cut the RMDs that would have hammered them at 75, flatten their lifetime tax, and leave their kids tax-free Roth dollars instead of a taxable IRA.

Frank wanted to solve it in one heroic move. The real solve was patience with a number attached. Five quiet years, a ceiling, and the discipline to stop. The window was always there. Most people just never look up in time to use it.