The Roth conversion ladder

How the ladder works, the five-year rule per conversion, and the three costs — ACA cliff, IRMAA, and the tax torpedo — that decide what a conversion really costs.

A Roth conversion ladder is one of the few genuinely clever things in retirement planning. It is also one of the easiest to get wrong, because the rule everyone quotes is not quite the rule that exists, and the cost everyone calculates is not the cost that lands. This spoke sits under the Retirement hub.

What it is

You move money from a traditional IRA or 401(k) into a Roth IRA, a slice at a time, one year at a time. Each slice is ordinary income in the year you convert it, so you pay tax on it then. Five tax years later, that slice can be withdrawn without the 10% early-distribution penalty — even if you are nowhere near 59½.

Do it every year and the slices start maturing every year. You have built a rolling, penalty-free income stream out of money that was otherwise locked up until 59½.

Who it is actually for

Two groups, wanting different things from it.

Early retirees bridging to 59½. Someone who stops working at 50 has nearly a decade before they can touch a traditional 401(k) without a penalty. The ladder is the standard way across, and it is where the five-year rule does real work.

Anyone in the low-income window. Someone who retires at 62, has not claimed Social Security, and has years before RMDs start is living through the only period of their adult life in which they set their own taxable income. For this group the penalty rules are mostly irrelevant — they are already past 59½ — and the cost rules below are everything.

The five-year rule — per conversion

This is the thing readers get wrong, so here it is in the IRS's own words. Publication 590-B:

A separate 5-year period applies to each conversion and rollover.

Every conversion starts its own clock. There is no single ladder-wide clock. And the clock starts on the first day of the tax year in which you converted — not on the date of the conversion itself.

That second detail is worth money. Convert on 15 December 2026 and the five-year period is deemed to have begun on 1 January 2026. It runs through the end of 2030, and that tranche is free of the 10% penalty from 1 January 2031. A late-December conversion buys you almost a full year of the clock for free.

The other five-year rule, which is not this one

There is a second five-year rule, it does something different, and conflating them is the most common error on this topic.

The practical traps fall out of the table.

If you are already over 59½, the conversion clock does not bind you. Reaching 59½ is itself listed as an exception to the 10% additional tax. Your converted principal was taxed on the way in and comes back out free. This is why building a ladder at 57 to "unlock" money at 62 is busywork — you would have been 59½ anyway.

If you have never had a Roth IRA before, the earnings clock does bind you. Open your first Roth at 60 with a conversion and your earnings are not qualified until five tax years have passed, at 65. Your converted principal is available immediately. A small Roth IRA opened years earlier, and left alone, starts that clock cheaply.

The ladder, year by year

Say a couple retires at 50 in 2026 and converts an illustrative $40,000 a year. (These dollars are an example, not a recommendation.)

Read the gap. The first rung does not mature for five years. Between 2026 and 2030 this couple converts steadily and can spend none of it without penalty.

So the ladder does not solve the early-retirement income problem on its own. It solves years six onward. Years one through five have to be funded from somewhere else: a taxable brokerage account, cash, or Roth IRA contributions you already made — which, under the ordering rules, come out first and are always tax and penalty free. A ladder without five years of other money behind it is not a plan, it is a countdown.

Bracket filling, and where it goes wrong

The sizing rule is simple: convert up to the top of a bracket, and stop. Filling your current bracket takes the cheapest dollars available. Spilling one dollar past it prices the overflow at the next rate up, for no more benefit than the dollar before it.

That is the textbook answer. Before 65, it is frequently the wrong one.

A worked example. A married couple, both 62, retired, no Social Security yet, buying health insurance on the marketplace. Household of two, with an illustrative $30,000 a year of other income from a taxable brokerage account. The inputs are an example; the thresholds and brackets are the site's 2026 constants.

Their ACA subsidy cliff — 400% of the federal poverty line for a household of two — sits at $84,600. From $30,000 that leaves $54,600 of headroom. The top of their 12% bracket is far further out: filling it would take a conversion of about $103,000.

Two options, run through the same 2026 brackets the conversion cost checker uses:

The extra $48,400 of conversion costs $5,808 in federal tax — a 12% marginal rate, exactly as the bracket-filling logic promises. And then it forfeits the household's entire premium tax credit for the year, because one dollar over the cliff is enough.

How much is that credit worth? It depends on where you live and how old you are, because marketplace premiums are age-rated. As a national-average reference point, the benchmark silver plan for a 60-year-old runs about $15,914 a year unsubsidised, and KFF's illustration of a 60-year-old at $65,000 of income puts the annual increase after the enhanced credits expired at about $10,389. Against numbers like that, a $5,808 tax saving on the bracket-filling logic is not a saving at all.

The bracket was never the binding constraint. The cliff was, by a factor of two. Bracket-filling advice written for a 45-year-old, applied at 62, is how people lose five figures while believing they are being efficient.

Three costs that no one prices together

Everything above is table stakes. Here is what most ladder guides leave out, and why a conversion that looks smart on brackets can be a bad trade.

1. The ACA subsidy cliff — any year you are on a marketplace plan before 65

The enhanced premium tax credits expired on 31 December 2025. For 2026, household ACA MAGI one dollar above 400% of the poverty line means a premium tax credit of zero. Not tapered. Zero.

For a single person that line is $62,600; for a couple, $84,600. It is built from the poverty guideline of $15,650 for one person plus $5,500 for each additional member.

Two things make this worse than it sounds:

• A conversion counts in full toward ACA MAGI. So do traditional withdrawals, capital gains, pensions and all of your Social Security benefits — including the untaxed portion. A Roth withdrawal from an account you already own does not.
• Below the cliff, a cap still exists. The enhanced credits did not just remove the cliff, they also softened the sliding scale beneath it. Both changes reversed. Do not read "the cliff is back" as "there is no help below it" — there is, it is simply narrower and harsher.

And there is a floor, too. Below 100% of the poverty line you generally get no marketplace subsidy at all. Someone crushing their income to duck the cliff can overshoot into that. The ACA bridge tool shows where you land.

The cliff applies at any pre-65 age, not just 60–64. But the dollars at stake grow with age, because premiums do.

2. IRMAA — from 63

Medicare surcharges are set from your income 2 years earlier. That is the whole planning point: the tax return you file for age 63 decides your Part B and Part D premiums at 65.

Cross the first tier — $109,000 single, $218,000 joint — and it costs roughly $1,148 for that year, per person on Medicare. If both spouses are enrolled, each pays it. Higher incomes reach higher tiers that cost considerably more.

It is a cliff, not a ramp: a dollar over the threshold triggers the whole tier. So a conversion sized to land just under it is worth real money, and one sized without checking is a self-inflicted wound with a two-year fuse.

3. The tax torpedo — once Social Security starts

This one is the most misunderstood, so here is the mechanism rather than the slogan.

How much of your Social Security is taxed depends on combined income: your other income, plus tax-exempt interest, plus half your benefits. Below $25,000 (single) or $32,000 (joint), none of the benefit is taxed. Above it, up to 50% becomes taxable. Above a second, higher threshold, up to 85% does.

Now add a conversion dollar. Inside the 85% band, that dollar does two things at once: it is taxable itself, and it drags 85 cents of your Social Security benefit into taxable income with it. One dollar of conversion produces $1.85 of taxable income.

Run that through the brackets:

• In the 12% bracket: 1.85 × 12% = a real marginal rate of about 22.2%.
• In the 22% bracket: 1.85 × 22% = about 40.7%.

You are being taxed at nearly double the rate on your bracket table, and no line on your return says so.

Be honest about the shape of it, though: this is a band, not a permanent surtax. Once 85% of your benefit is already being taxed — the statutory maximum — the multiplier drops back to 1 and your marginal rate returns to the ordinary bracket rate. The torpedo is a stretch of income you pass through. But a conversion sized on the bracket table, done in the years after you claim, can be priced through that stretch without you ever noticing.

These thresholds have never been indexed. They are the same nominal dollars set in 1983 and 1993. Every year of inflation pulls more retirees into the band — which is also why the senior deduction is not "no tax on Social Security": it is a deduction, it expires after 2028, and it did not touch these thresholds at all.

The cleanest way out of the torpedo is to do your converting before you claim. Which brings us to the deadline.

The deadline that creates the whole problem: RMDs

You do not get to leave the money alone forever. Required minimum distributions start at 73 for those born 1951–1959, and 75 for anyone born in 1960 or later — which is most people reading this. Note that it is not a flat 73, whatever the article you read last said.

At that point the choice ends. A large pre-tax balance produces a large forced withdrawal, on top of Social Security, every year for the rest of your life. That withdrawal drags the benefit into tax, pushes you up the brackets, and can hold you in an IRMAA tier permanently.

Convert too little in the quiet years and you inherit that. Convert too much and you pay tax you never needed to. Neither error is recoverable, which is the honest reason this decision is hard.

Roth accounts, by contrast, have no lifetime RMDs at all — not the IRA, and since 2024, not the designated Roth 401(k) either.

Two rules that will cost you if you skip them

Pay the tax from outside the conversion. If you convert $100,000 and have 20% withheld for tax, only $80,000 reaches the Roth. The $20,000 sent to the IRS is a distribution you did not convert — so under 59½ it takes the 10% early-distribution penalty on top of the income tax you already owed. You have paid $2,000 for the privilege of having less money in your Roth. Pay from a taxable account. If you cannot, that is a strong signal the conversion is too big.

The pro-rata rule. The IRS does not look at the one account you converted from. It aggregates all of your traditional, SEP and SIMPLE IRAs into a single pot and spreads your after-tax basis across it. So a conversion comes out proportionally taxable no matter which account the money physically left. Publication 590-B: "Until all of your basis has been distributed, each distribution is partly nontaxable and partly taxable." Form 8606 does the sums. If you have a big pre-tax rollover IRA, do not assume a conversion of after-tax money is tax-free — it will not be.

And you cannot undo it. Recharacterisation of conversions ended for conversions made in 2018 or later. Publication 590-A says so directly. Convert in January, watch the market fall 30% in March, and you still owe tax on January's value. That risk alone argues for converting in smaller slices, and later in the year, when you can see your actual income.

The window

Between the day you stop working and the day Social Security and RMDs start, you are the person who decides what your taxable income is. That is a handful of years, it never comes back, and it is the only period in which a conversion is cheap. That is the whole opportunity.

The years are also crowded. Before 65 the ACA cliff is watching. From 63 Medicare is already recording. Once you claim, the torpedo is live. At 75 the choice is taken away from you.

There is a real answer in there, and it is specific to your balances, your state, your health insurance, and when you plan to claim. It is also a decision where being wrong by one dollar can cost five figures and nothing can be undone. This is exactly the decision worth paying a professional for. Our conversion cost checker shows you the shape of it — tax, cliff and surcharge in one place, which is more than most free calculators do — but it is an educational estimate, federal only, and it is not a plan.

Key takeaways

• Each conversion has its own five-year clock, and it starts on 1 January of the year you convert. That is a different rule from the single five-year clock on Roth earnings.
• Past 59½, the conversion clock stops binding you — so many people build ladders they never needed.
• Your first rung matures in year six. You need five years of spending money from somewhere else, or the ladder is a countdown, not a plan.
• Fill a bracket, do not spill past it — but before 65 the ACA cliff usually binds long before the bracket does, and one dollar over forfeits the whole credit.
• Pay the tax from outside the conversion, mind the pro-rata rule, and remember you cannot undo it.
• The window between retiring and 75 is the whole opportunity. Convert too little and a forced RMD drags your Social Security into tax anyway.

Educational only — not financial or tax advice.