Roth Conversions in RetireOdds: A Complete Guide

A conversion moves money from a pre-tax account into a Roth and pays the tax now. The Roth Conversions page prices exactly that trade, year by year, against your own plan.

A Roth conversion takes money out of a pre-tax account — a traditional IRA or 401(k) — and moves it into a Roth account. The IRS treats the converted amount as ordinary income in the year you convert, so you owe tax on it now, at whatever bracket your income lands in that year. In exchange, the money is done paying tax forever: it grows in the Roth and comes out tax-free later, and it's no longer part of the balance the IRS forces you to draw down through Required Minimum Distributions (RMDs).

The Roth Conversions page models a ladder — a planned sequence of conversions over several years rather than one lump sum — and shows you, year by year, what it costs today and what it's worth later. This guide walks through every control on that page, what the chart and KPIs actually mean, how the app tracks Roth basis and the two different five-year clocks, what conversions can cost you in ACA subsidy or Medicare IRMAA, how the expat angle works, and what the optimizer panel does and doesn't search. Nothing here is tax advice — it's a description of what the model computes so you can read its output correctly.

What a conversion is, and the problem a ladder solves

Every dollar in a traditional IRA or 401(k) carries an unpaid tax bill. You can pay it on the IRS's schedule — via RMDs starting at age 73 (or 75 if you were born in 1960 or later, under SECURE 2.0 — both ages are modeled) plus ordinary withdrawals — or you can choose to pay it earlier, in years when your bracket is low, by converting. A conversion doesn't reduce the tax owed on that money; it changes when and, if your bracket differs between now and later, how much you pay.

The opportunity most people are trying to capture is a low-income window: the years after you stop earning wages but before Social Security and RMDs start filling your tax return back up. In that gap, your taxable income can be low enough that a meaningful conversion still fits inside a low bracket. Convert in those years and you permanently move money out of a balance that would otherwise be taxed later — often at a similar or higher marginal rate once RMDs and Social Security stack on top of each other (the model prices this stacking explicitly, including the Social Security "tax torpedo," where taxable Social Security itself rises as other income rises).

That example is illustrative, not a projection — your own numbers depend on your balances, brackets, and the years you choose. The ladder exists to turn "should I convert, and how much" into a concrete, year-by-year projection instead of a guess.

Setting up a ladder, step by step

1. Choose the first question to preview

If you have not saved a Roth strategy before, the page does not run one automatically. It first asks you to choose a 12%, 22%, or 24% federal bracket ceiling. That action creates a temporary comparison with no conversions; it does not choose a strategy for you or change the Ledger. After the first preview, the full controls below become available and assumption changes update the comparison after a short pause.

2. Where the balances come from

The ladder doesn't ask you to re-enter numbers. It sums your existing accounts by tax treatment: everything tagged Tax-deferred becomes the trad balance, everything tagged Tax-free becomes the Roth balance, and Taxable or Cash accounts fund the taxable bucket that pays each year's conversion tax. Accounts tagged Mixed are deliberately excluded — the model won't guess which slice of an unclassified account is pre-tax.

3. Pick a strategy type

• Fill to a bracket. Converts however much is needed to bring that year's taxable ordinary income up to the top of a bracket you choose (12%, 22%, or 24% on the page's quick-pick buttons; 10% and 32% are also accepted — the optimizer can recommend either). The bracket's dollar ceiling is expressed in today's dollars and indexed with inflation each projection year, so a 22% ceiling means the 22% bracket in that year's inflation-adjusted terms, not a frozen dollar figure.
• Keep MAGI under a ceiling. You enter a MAGI figure in today's dollars (the field defaults to $212,000, and you can change it); the ladder converts up to that ceiling instead of a bracket. MAGI here is built from wages, RMD/ordinary withdrawals, the conversion itself, and taxable Social Security — the same figure IRMAA and ACA calculations key off of. This mode is what you'd use to plan around a specific MAGI cliff (an IRMAA tier or the ACA subsidy cutoff) rather than a bracket line.
• Fixed dollars per year. Converts the same nominal amount every year in the window, capped by whatever's left in the trad balance.

All three strategies share a start age, end age, and an optional annual cap — a hard per-year ceiling layered on top of whichever rule computes the year's amount. The window defaults to your retirement age through age 73, and both ends are editable.

4. Growth: leave the field blank for AUTO

The Growth % field is blank by default, and blank means something specific: the server resolves it, in this order — an explicit number you type wins; otherwise your profile's expected-return override wins; otherwise the projection uses an allocation-anchored auto model, the same one the rest of the app uses (a historical geometric-return blend for your portfolio's equity percentage, compounded with your inflation assumption). Typing 0 is a valid explicit choice and is treated differently from leaving the field empty — the model never silently substitutes a default for an entered zero. When AUTO is active, the field's placeholder shows the resolved percentage so you can see the number actually being used.

5. "Include in my Chance of Success"

This toggle is off by default. Off, the ladder is a standalone projection — it doesn't touch your dashboard odds. Switching it on persists the strategy to your plan and folds it into every Chance-of-Success run: the simulation moves the same trad-to-Roth dollars on each Monte Carlo path and charges that year's conversion tax (and any NIIT) as a real outflow, so the conversions are genuinely priced into your retirement odds rather than illustrated separately.

Reading the outputs

The header strip above the chart shows four numbers: lifetime tax saved (today's dollars, this ladder vs. converting nothing), after-tax wealth at your life expectancy (the delta between the two runs, each converted to a single after-tax figure with the same progressive federal/state engine used everywhere else in the app — not a flat haircut), total converted (with a "before move" sub-total if you're an expat using the concentrate control), and RMDs reduced (the percentage cut in total RMD dollars taken over your whole horizon, comparing the with-ladder and no-ladder runs).

The chart plots two solid lines — your traditional balance and your Roth balance — plus a dashed line showing what your traditional balance would have been with no conversions, so you can see the balances physically cross over as the ladder drains one into the other. A vertical marker reading "RMDS BEGIN AT <age>" appears at the first age either run has an RMD. At the end of the plan, the chart labels the final Roth balance "TAX-FREE" and the final no-conversion traditional balance "STILL TAXABLE" side by side. Clicking anywhere on the chart (or a row in the table) selects that year for a detail readout; a "Full table →" toggle switches between milestone rows (first year, every fifth year, the RMD-start year, any year a conversion starts or stops, and the last year) and every year. The table itself shows, per age: the conversion amount, its tax, your Trad and Roth balances, the no-conversion Trad balance for comparison, your taxable balance, and that year's RMD.

Basis and the two five-year clocks

Roth basis — your regular contribution basis and the taxable/nontaxable split of every conversion, by tax year — is entered on the Tax page, under "Roth IRA basis and five-year clocks," and is shared automatically with the Roth ladder; you only enter it once. If you don't track it, the ladder still runs, but any pre-existing Roth balance is conservatively classified as unproven "earnings" — the last dollars out, and the most exposed to tax and penalty if withdrawn early.

When money comes out of a Roth before age 59½, the model follows the statutory ordering from IRS Publication 590-B: regular contribution basis first, then conversion lots oldest-first (the taxable part of a lot before its nontaxable part), then earnings last. Two separate five-year rules sit on top of that ordering:

• Each conversion's own recapture clock. The taxable portion of a specific conversion lot carries the 10% early-withdrawal penalty if it comes out within five tax years of that conversion and you're under 59½ with no exception — tracked lot by lot, not as one plan-wide clock.
• The separate qualified-distribution test. Whether your Roth earnings come out tax-free depends on five tax years since your very first Roth contribution or conversion, plus a qualifying event (59½, death, disability, or a first-time home purchase up to a $10,000 lifetime limit).

That's why entering your conversion lots matters: without them, the model can't tell how much of a withdrawal is basis you already paid tax on versus taxable earnings, and it can't start the qualified-distribution clock. If you haven't entered a first-contribution year and you have no conversion lots either, the ladder treats that clock as never satisfied — every non-qualified distribution is priced as if it fails the five-year test, which is the conservative direction. If you do enter conversion lots but leave the first-contribution year blank, the Tax page infers it as your earliest lot's tax year automatically; if you have basis or lots but no year and nothing to infer it from, the page blocks saving and asks you to enter it.

The cost side: ACA and IRMAA

A conversion is ordinary income, so it raises your MAGI in the year you convert — and MAGI is exactly what two other costs key off of. Before 65, a higher MAGI can shrink or eliminate an ACA marketplace premium tax credit; the app prices the advance credit against your estimate and reconciles it the following tax year against your actual MAGI. At 65 and later, Medicare's IRMAA surcharge uses a statutory two-year MAGI lookback — a conversion at 63 can raise your premiums at 65. The Roth ladder uses the same healthcare and tax assumptions as the rest of your plan (see the Roth FAQ), so these interactions aren't a separate estimate; the base ladder projection already runs conversion tax, NIIT, and any entered Form 8606 IRA basis through the same shared annual tax ledger the rest of the app uses. For the full breakdown — separate IRMAA and lost-ACA-subsidy dollar totals per candidate — see the optimizer panel below. For more on how the underlying tax and healthcare rules work, see the tax guide and the healthcare guide.

Converting before you move abroad

If your profile has a non-US destination country set, the Roth Conversions page shows an extra "Concentrate before I move" checkbox and a move-age field. Turning it on restricts the ladder's conversion window to strictly before your move age — still bounded by whatever start and end ages you've chosen — on the logic that your US tax brackets today may be a better deal than however your destination treats Roth withdrawals once you're a resident there.

That "however your destination treats it" question resolves to one of three states in the model, per destination country: exempt (the treaty is modeled as preserving the Roth's tax-free character — France is the one supported country currently modeled this way), taxable (the host country taxes withdrawals as ordinary pension-like income — the conservative default for most supported destinations, including Portugal, Spain, Italy, Greece, Mexico, Thailand, and India), and taxed as account (an explicit election you make yourself, never a default: the host taxes withdrawals and taxes the account's nominal growth every year at its passive-income rate — the worst case).

It's important to be precise about where that host treatment actually applies: the Roth Conversions page itself always prices conversions under US tax only — the numbers on this page don't change based on your destination. The host treatment is applied when you turn on "Include in my Chance of Success" and run the full simulation. For supported destination countries, the saved ladder is coordinated between US and host tax; for an unsupported destination, the saved conversions are not applied to your Chance of Success at all, because the simplified estimate used for unsupported countries can't price a conversion's coordinated US-and-host obligation. Even for a supported destination, the coordinated pricing requires your accounts to be classified into taxable/tax-deferred/Roth buckets — without that, the plan refuses to apply the ladder rather than misprice it. In every case, the conversion amounts are still chosen by the same US-bracket ladder or optimizer described in this guide; only the tax consequence of each conversion year is priced under the coordinated obligation. For the destination-specific mechanics, see the expat taxes guide.

The optimizer panel: what it actually searches

The optimizer is a deterministic search, not a Monte Carlo simulation: it evaluates a fixed grid of candidate schedules — fill-to-bracket at 10/12/22/24/32%, plus fill-to-MAGI at your first three Medicare IRMAA tier thresholds and, if ACA is active and you retire before 65, the 400%-of-federal-poverty-line ACA cliff — crossed against a handful of start/end age windows (through age 63, through your Social Security claim age, through your RMD start age, and a long window to age 75 or your life expectancy), all at a single growth-rate assumption.

You choose an objective — maximize after-tax wealth at your life expectancy, minimize lifetime cost (tax plus IRMAA plus lost ACA subsidy, in today's dollars), or a weighted balance of wealth, cost, and how smooth the conversion amounts are year to year. Optional "practical guardrails" let you set a minimum taxable-cash cushion, a maximum tax bill in any single conversion year, an IRMAA or ACA MAGI ceiling, or a latest acceptable break-even age; any candidate that violates one is rejected and listed, not silently dropped. Doing nothing competes on equal footing — if no candidate schedule beats the no-conversion baseline under your objective and guardrails, the optimizer says so instead of forcing a recommendation.

The result includes the recommended schedule's own year-by-year table, up to three near-miss runner-ups, and a sensitivity check: the same search re-run at growth ±1.5 percentage points, showing whether the same strategy still wins. "Apply to my plan" writes the recommended strategy into your saved settings, turns on Include-in-Plan, and re-runs your Chance of Success. Running the optimizer requires an existing tax-deferred balance — there's nothing to convert without one.

Common workflows

• Retired early with a gap before Social Security and RMDs. A fill-to-bracket strategy from your retirement age through the year before RMDs start uses exactly that low-income window — watch the RMD-reduction KPI and the chart's end-of-plan annotation to see how much of the traditional balance moved before it would have started forcing distributions.
• Trying to stay under an ACA subsidy cliff. A fill-to-MAGI strategy with the ceiling set at (or the optimizer's candidate set at) the 400%-FPL threshold keeps conversions from pushing you out of marketplace subsidy range in a given year — check the optimizer's ACA-subsidy-lost cost chip to see the trade-off in dollars.
• Working down a large pre-tax balance ahead of RMDs. A longer window at a higher bracket ceiling converts more, faster; the trade-off shows up directly in the KPI strip's lifetime-tax-saved and after-tax-wealth numbers, and in how much smaller the RMD line becomes on the chart.

Honest limitations

• This is a planning estimate, not tax advice — federal and state constants are pinned to the 2026 tax year and indexed forward with your own inflation assumption, not against actual future law changes.
• The standalone ladder and the optimizer are deterministic — one projected path at one growth rate. Only turning on "Include in my Chance of Success" puts the conversions through the Monte Carlo simulation, where market volatility and sequence-of-returns risk are actually priced. See the Monte Carlo guide.
• Traditional, Roth, and taxable balances are modeled with equal total return — the ladder can't represent holding a different fund mix in each bucket.
• Social Security's own taxability thresholds are deliberately left un-indexed in the model (correctly mirroring current law), while federal brackets and deductions are indexed with your inflation input — an intentional asymmetry, not an oversight.
• Conversions are irreversible under current law (recharacterization was eliminated in 2018) — the app has no "undo" for a conversion once it's part of your saved plan, so it models forward from a made decision rather than a reversible one.

For the tax mechanics behind every number on this page, see the tax guide; for how healthcare costs interact with a conversion year, see the healthcare guide. If you're weighing a conversion against a move abroad, the expat taxes guide covers the destination side in depth.