The best years to convert are the ones right after the paychecks stop.

The shape of a retirement tax bill.

When I map out someone's taxes from today to age 90, it almost never looks like a straight line. It looks like a valley.

The middle of that table is the opportunity. If you do nothing in the gap years, for many people taxes stay low for a few years and then rise for the rest of their lives. If you use them, you pay some tax on purpose now, at a rate you choose, so the IRS doesn't get to choose it for you later.

Why the valley doesn't last.

Two things close the window, and both are on a calendar.

Social Security. Once both checks are on, a large part of your bracket is already filled before you convert a dollar.

Required minimum distributions. The IRS makes you start taking money out of your pre-tax accounts at 73, or at 75 if you were born in 1960 or later. The amount is set by an IRS table, it's taxed as ordinary income, and it grows as a share of the account every year. You don't get a vote. I call it the freight train, because by the time you can see it, it's hard to get out of the way. Here's what it looks like on a growing account →

Here are the ages worth circling:

Filling the bracket, on purpose.

A Roth conversion adds to your taxable income for the year. So the question is never "should I convert?" in a vacuum. It's "how much room do I have in the bracket I'm willing to pay?"

In practice I usually show people two or three sizes side by side, for example converting to the top of the 12% bracket, the 22% bracket or the 24% bracket, with the tax bill for each. And I always add one more column: not converting this year. Then they choose. It's your money and your plan; my job is to show you what each choice costs and what it buys.

The top of a bracket is a reference point, not a target. A conversion can fit the bracket perfectly and still be too big for your cash flow, because once the paycheck stops, your savings are paying for your life and the tax at the same time. More room in the bracket doesn't mean more money to pay the bill.

The things that quietly shrink the window.

This is where people get surprised. Plenty of income shows up in the gap years that nobody counted:

• The year you retire. Retire in June and you still have half a year of wages, maybe a bonus or a vacation payout. It's rarely the low-income year people picture. The first full year without a paycheck usually is.
• A pension. Even a small one fills the bottom of the bracket every year it pays.
• Part-time work or consulting. Great for the plan, but every dollar of it is a dollar less room.
• Interest on cash. Several hundred thousand dollars in savings at today's rates is real taxable income.
• Selling stock or company shares. Capital gains stack on top of everything else.
• Medicare's two-year lookback. From about age 63, a big conversion year can raise both spouses' Medicare premiums two years later.
• Health insurance before 65. If you retire early and buy coverage on the exchange, conversion income can reduce the help you get with premiums. Check that cost alongside the bracket.
• State tax. In Connecticut, IRA and pension income is fully exempt when a married couple's federal AGI is under $100,000, and the exemption phases out up to $150,000. Land a conversion in the middle of that band and the state can take a surprisingly large bite of the last dollars. I usually size conversions to land below it or clear of it, not inside it.

What this looks like in real life.

Here's a pattern I see often. A couple retires in their late fifties, a few months apart. They have saved well, almost all of it in 401(k)s and IRAs, with very little in a Roth. For the next several years, until Social Security starts, their income is a little consulting, some interest and maybe a small pension. Most of them think of those years as "living off savings." I think of them as the best conversion years they will ever get.

The decisions that matter are practical ones: how much to convert each year, which bracket to stop at, where the tax money comes from, and when each of them claims Social Security. Claiming early is a perfectly good choice for some people, but every year a check is turned on is a year with less room to convert. These things have to be planned together, not one at a time.

One more thing: what you move matters as much as how much.

I think of the Roth as the best account you'll ever own, because qualified withdrawals, growth included, are tax-free. So when you convert, move your growth investments first and leave your bonds in the IRA. Accelerate the Roth. Decelerate the IRA. If all your money sits in one blended target-date fund, that takes extra work, and I explain why in One fund got you here. Spending it takes a plan.

Still working? Don't wait for the valley to start.

The gap years are usually the best time for the big conversions. They aren't the only time to build a Roth. If you're still working and you believe in the Roth, it doesn't make sense to keep adding to the pre-tax bucket you plan to convert later. You can switch your 401(k) contributions to the Roth option, contribute to a Roth IRA if you qualify, and even do a small conversion in a working year with room in your bracket. Every year you wait is a year of growth that happens in the account you're trying to shrink. I cover this in If you believe in the Roth, why are you still feeding the IRA?