Withdrawal Order in Retirement: Brokerage Account vs. Roth IRA

The scenario

You're retired with two accounts holding the exact same investment mix — say a total-stock-market index fund — in two different wrappers:

1. A taxable brokerage account funded with after-tax dollars. Your basis is whatever you paid; everything above that is an unrealized capital gain.
2. A Roth IRA funded with after-tax dollars that now grow and come out completely tax-free (assuming you're over 59½ and past the 5-year rule, which a retiree typically is).

You need to pull living expenses from somewhere. Your instinct, and the conventional advice, is: deplete the brokerage account entirely before touching the Roth. Are you right?

For most retirees in this exact setup, yes — Roth-last is the correct default. But "default" is doing a lot of work in that sentence. The same-asset, two-wrapper framing hides several second-order effects — the step-up in basis at death, IRMAA surcharges, the 0% capital-gains bracket, and ACA subsidies for early retirees — that can flip or fine-tune the answer. Let's build the core intuition first, then walk through every place it breaks.

Why Roth-last is the right default

The cleanest way to see it: both accounts hold the same asset earning the same return, so the investment return is a wash. The only thing that differs is the tax drag on each wrapper while the money sits there and when it comes out.

• Roth IRA: zero tax drag. No tax on dividends, no tax on rebalancing, no tax on withdrawal. Every dollar of growth is yours.
• Taxable brokerage: leaky. You pay tax on dividends every year whether you want the income or not (a total-stock fund throws off roughly 1.3–1.5% in dividends annually), and you pay capital-gains tax when you sell to fund spending.

Because the Roth compounds with no leak, every year you leave a dollar in the Roth instead of the brokerage, that dollar grows tax-free instead of tax-drag-reduced. Spending the leaky account first and preserving the perfectly sheltered account longest maximizes the total after-tax value of the portfolio over time.

A simple worked example. Suppose you need $40,000 of spending and you have $40,000 sitting in each account, all of it earning 7%. You spend the brokerage $40,000 this year and leave the Roth alone. Next year the Roth is worth ~$42,800, and all of that growth is tax-free forever. Had you spent the Roth instead, that $2,800 of growth would have happened inside the brokerage account, where its future dividends and eventual sale are taxable. Over a 20–30 year retirement, repeatedly favoring the tax-free compounding account is worth a meaningful amount.

There's a second, subtler reason Roth-last wins that pure "tax drag" framing misses: the brokerage account gets a step-up in basis at death, and the Roth is already tax-free — so for heirs, draining the brokerage first is also usually fine, as we'll see next. The default holds up coming and going.

The capital-gains step-up at death changes the brokerage math

Here's the wrinkle that makes spending the brokerage first even more attractive than the simple tax-drag story suggests.

When you sell appreciated stock in your brokerage account to fund spending, you realize capital gains and pay tax on them. That's the cost of using the brokerage. But if you die holding that appreciated stock, your heirs receive a step-up in basis — the cost basis resets to the fair market value on your date of death, and the entire embedded gain is wiped out, tax-free, under current law (2026).

This creates a genuine tension:

• Spend the brokerage first → you realize gains and pay capital-gains tax now, but you avoid being forced to realize them, and you preserve the Roth.
• Hold the brokerage until death → the embedded gains escape income tax entirely via step-up.

For most retirees who will actually consume most of their portfolio, the step-up is a tie-breaker, not a strategy — you can't both spend the money and let it step up. You spend the brokerage, you preserve the Roth, and whatever brokerage assets remain at death step up anyway.

The step-up matters most for retirees with estate-planning intent — those likely to die with assets unspent. For them, the ideal bequest mix is: a Roth IRA (tax-free to heirs, though subject to the 10-year drain rule for most non-spouse beneficiaries) plus brokerage assets with a fresh stepped-up basis (also tax-free). Both are great inheritances. The asset you do not want to leave behind is a large traditional IRA, which lands on heirs as fully taxable ordinary income inside a 10-year window — but that's not in this scenario, since this retiree has no traditional/pre-tax account.

So in the two-account case here, the step-up reinforces the default: spend the brokerage, leave the Roth (and any leftover stepped-up brokerage shares) to heirs.

The 0% long-term capital-gains bracket: spend the brokerage almost for free

If Roth-last is the default, the 0% LTCG bracket is the reason it can be nearly costless — and the reason you should be thoughtful about how much brokerage you draw each year.

For 2026, long-term capital gains and qualified dividends are taxed at 0% until your taxable income reaches $49,450 (single) or $98,900 (married filing jointly). Long-term gains stack on top of ordinary income, so the room you have in the 0% bracket is whatever's left after your ordinary income fills the bottom.

Worked example, single retiree (2026):

• Spending need: $50,000/year, funded entirely from the brokerage.
• Suppose $30,000 of the sale proceeds is return of basis and $20,000 is long-term gain. Only the $20,000 gain is income.
• This retiree has no other ordinary income (no pension, not yet claiming Social Security, no traditional IRA). After the standard deduction of $16,100 (2026), taxable income is well under the $49,450 (2026) 0% LTCG threshold.
• Federal tax on the withdrawal: $0. The gains fall entirely in the 0% bracket.

That's the magic of this setup: a retiree spending from a brokerage account with a healthy basis, and little other income, can often pull tens of thousands of dollars a year and pay nothing in federal tax. Draining the brokerage first isn't just tax-efficient sequencing — for many early retirees it's tax-free sequencing.

This also argues for gain harvesting: in years when you're under the 0% threshold and not spending the full amount, you can sell appreciated shares and immediately rebuild basis (rebuy them — there's no wash-sale rule on gains) to soak up the free 0% bracket. It quietly reduces the embedded gain in the brokerage so future, larger withdrawals don't spill into the 15% bracket.

When Roth-first (or Roth-partial) actually makes sense

The default assumes a fairly smooth, moderate spending profile. Several situations justify pulling from the Roth earlier than "last."

1. Avoiding an IRMAA cliff. Medicare Part B and Part D premiums jump in steps based on your modified AGI from two years prior. For 2026, the standard premium applies at MAGI ≤ $109,000 (single) / ≤ $218,000 (MFJ); the first surcharge tier (2026) begins the dollar you cross $109,001 / $218,001. IRMAA is a cliff, not a phase-in: one dollar over the line and the full tier's surcharge applies for the whole year. If a large brokerage sale (think a concentrated position, a home sale, or a big rebalance) would push you just over an IRMAA threshold, funding part of that spending from the Roth instead keeps your MAGI down, because qualified Roth withdrawals don't count toward MAGI at all. Here the Roth's best use isn't compounding — it's as a tax-invisible spending source that smooths you under a cliff.

2. Very large future RMDs — not applicable here, but worth flagging. The classic Roth-earlier argument is for retirees with big traditional IRAs facing forced Required Minimum Distributions. RMDs now begin at age 73 (born 1951–1959) or 75 (born 1960 or later) (2026). A retiree who lets a large pre-tax IRA grow untouched can get hit with RMDs so large they spike into high brackets, trigger IRMAA, and make more Social Security taxable — so spending or converting pre-tax money earlier (and leaning on tax-free Roth less, or doing conversions) can lower lifetime tax. This retiree has only Roth and brokerage, so there are no RMDs (Roth IRAs have no RMDs for the original owner), and this exception doesn't apply. It's the single biggest reason real-world withdrawal-order advice gets complicated — but it's absent from this specific two-account case.

3. Bequest and bracket management. If you expect to leave the Roth to heirs and have years of unusually low income, you might intentionally not drain the brokerage to zero, instead spreading realizations to stay in the 0% LTCG bracket and preserving stepped-up-basis shares. This is fine-tuning the default, not reversing it.

The Roth conversion ladder: the missing middle (and a caution for this scenario)

Most discussions of withdrawal order assume your accounts are fixed. They're not — you can move money between wrappers, and the Roth conversion ladder is the tool.

A conversion ladder means deliberately converting traditional/pre-tax IRA money to Roth during low-income early-retirement years, paying tax at today's low brackets, so the money is never subject to RMDs and grows tax-free thereafter. It's the strategic complement to withdrawal ordering: you spend the brokerage for cash flow and use the headroom in your low brackets to convert pre-tax money to Roth at, say, 10% or 12%, rather than letting it balloon into RMDs taxed at 22%+ later.

Important caveat for this specific scenario: the retiree here has no traditional/pre-tax account — only a taxable brokerage and a Roth. There is nothing to convert. A conversion ladder is therefore not applicable to this exact two-account case. We include it because (a) it's the natural next question once you understand sequencing, and (b) the brokerage account interacts with conversions in a crucial way: you want to pay the conversion tax with brokerage (or cash) dollars, not by withholding from the conversion itself, so the full converted amount lands in the Roth. If you ever add a pre-tax account to this picture, draining the brokerage first does double duty — it funds living expenses and the tax on Roth conversions while you're in low brackets.

For readers who do have a traditional IRA alongside, the rough hierarchy becomes: spend the brokerage for cash, fill your low brackets with conversions, and preserve the Roth — which collapses back to the same "Roth-last" instinct, just with an extra moving part.

ACA subsidies: the early retiree's hidden 0% (or huge) marginal rate

If you retire before 65 (pre-Medicare) and buy health insurance on the ACA marketplace, your withdrawal order can be worth far more than the income tax on it — because ACA premium subsidies are based on your MAGI, and qualified Roth withdrawals don't count toward MAGI.

For 2026, the enhanced subsidies have expired and the 400%-of-federal-poverty-level cliff has returned: roughly $63,840 (single) / $132,000 (family of 4) (2026). Below the cliff you get premium tax credits; one dollar above it and you lose the entire subsidy — potentially $10,000–$20,000+ of premium credits for a family, gone.

This radically changes the calculus for a 55–64-year-old in our scenario:

• Spending from the brokerage generates MAGI (the realized gains and dividends count), which eats into your subsidy and can push you over the cliff.
• Spending from the Roth generates zero MAGI, so it's invisible to the ACA calculation.

Worked example, a 60-year-old single early retiree (2026):

• They want to spend $55,000 this year.
• Their brokerage sales would realize $30,000 of MAGI-counting income, landing them comfortably under the $63,840 cliff — fine, the brokerage-first default works.
• But suppose a one-time need (new roof, car, helping a kid) pushes desired spending to $80,000. Funding all of it from the brokerage might realize $50,000+ of gains and blow past the cliff, vaporizing the subsidy.
• The fix: fund the baseline spending from the brokerage to stay under the cliff, and fund the extra lumpy amount from the Roth, keeping MAGI flat. The Roth's tax-free, MAGI-free nature is precisely what you want for the marginal dollars near a subsidy cliff.

So for early retirees, the refined rule is: brokerage-first up to the point where it threatens an ACA cliff (or IRMAA later), then Roth for the rest. The Roth isn't just the last account to spend — it's your tool for managing MAGI around the cliffs.

Putting it together

For the retiree in this scenario — taxable brokerage plus Roth IRA, same 100%-stock mix, no pre-tax account — the original instinct is correct: deplete the brokerage first, preserve the Roth. Three forces all point the same way:

1. Tax-free compounding. Sheltering the leak-free Roth longest maximizes lifetime after-tax wealth.
2. Step-up at death. Brokerage assets reset basis at death; the Roth is already tax-free — so the bequest is clean either way, and spending the brokerage first never wastes the Roth's advantage.
3. The 0% LTCG bracket (2026). A retiree with modest other income can often pull brokerage gains at a 0% federal rate, making brokerage-first nearly costless.

The exceptions are about smoothing your income around cliffs, not reversing the order: pull from the Roth to duck under an IRMAA threshold ($109K single / $218K MFJ, 2026) at 65+, or under the ACA 400% FPL cliff ($63,840 single / $132,000 family of 4, 2026) before 65. The big classic exception — managing oversized RMDs — doesn't apply here because there's no traditional account, and for the same reason a Roth conversion ladder has nothing to convert. Both remain essential to understand the moment a pre-tax account enters the picture.

Run your own numbers — basis, expected income, the cliffs that apply to your age — with the roth-vs-traditional and roth-conversion calculators to see exactly where your brokerage-first plan stays free and where a Roth dollar earns its keep.