Mega-Backdoor Roth vs Brokerage: $46K, 30 Years

The decision, resolved with one couple’s numbers

Priya and Daniel are a married-filing-jointly couple in Austin, Texas. Combined W-2 income: $340,000. They already max the $24,500 each in 401(k) employee deferrals, both Roth IRAs through the backdoor, and the family HSA. They have $46,000 of additional cash to invest every year and one open question: should it go into the mega-backdoor Roth their employer plans support, or into a plain low-turnover taxable brokerage account they already own?

Their advisor-cousin told them “just use the brokerage — the step-up wipes out the tax anyway.” That is the myth this article corrects. For money you actually spend in retirement, the step-up never triggers, and the mega-backdoor Roth wins by roughly $445,000 over 30 years. The brokerage only wins in two narrow cases. Here is the worked decision.

Where the $46,000 comes from

The mega-backdoor Roth uses the gap between the employee deferral limit and the total 401(k) limit. For 2026 (IRC §402(g) and §415(c)):

The mechanic: Priya contributes $46,000 of after-tax (non-Roth, non-deferral) dollars to her 401(k), then immediately converts them to the Roth sub-account via in-plan Roth conversion or in-service withdrawal to a Roth IRA. Because the conversion happens before the money earns anything, there is no tax on the conversion. From that point, all growth is tax-free under IRC §408A. The same $46,000 in a brokerage is invested in already-taxed dollars and grows in a taxable account.

The 30-year head-to-head

Same inputs for both: $46,000 contributed at the end of each year, 7% nominal return, 30 years. The pre-tax ending balance is identical in both accounts — about $4.34M. The entire difference is the tax you pay to use the brokerage money.

The brokerage figure already gives the taxable account two breaks. First, it assumes a low-turnover index strategy that defers almost all gain to the end — a high-turnover or dividend-heavy account would owe annual tax along the way and end far lower. Second, it taxes the embedded gain at the favorable long-term rate, not ordinary income. Even with both breaks, the Roth’s zero-tax growth wins by about $445,000. The true gap is wider for any brokerage that throws off taxable dividends or realizes gains during the accumulation years (a 2% annual tax drag on a 7% return knocks roughly 15%–20% off the brokerage’s ending balance over 30 years).

The capital-gains math the brokerage can’t escape

When Priya and Daniel sell the brokerage position in retirement to fund spending, the embedded $2.96M gain is taxed. Their bracket depends on the year’s realized income, but a high-net-worth couple realizing six figures of gain lands in the 15% or 20% long-term bracket, plus the 3.8% Net Investment Income Tax on the amount over the $250,000 MFJ MAGI threshold (IRC §1411, stats.md §2). At a 15% blended effective rate on the gain — spreading sales across years to stay mostly in the 15% band and keep MAGI below the NIIT line — the lifetime tax bill runs around $445,000. Realize it all in fewer years and the 3.8% NIIT (and even the 20% bracket) push the bill past $550,000.

The Roth pays none of that. Once Priya is past 59½ and any conversion has cleared the 5-year rule, every withdrawal — principal and 30 years of growth — comes out at zero federal tax (IRC §408A(d), stats.md §12). There is no NIIT on a qualified Roth distribution. That is the structural reason the Roth wins for money you spend.

The two cases where the brokerage actually wins

Case 1: You die holding the position (step-up under §1014)

If Priya and Daniel never sell and pass the brokerage account to their kids, IRC §1014 resets the cost basis to the date-of-death fair market value. The entire $2.96M of unrealized gain vanishes — the heirs can sell the next day and owe nothing. Texas is a community-property state, so the survivor also gets a full step-up on the deceased spouse’s half plus the survivor’s half (stats.md §6 lists the nine community-property states: AZ, CA, ID, LA, NV, NM, TX, WA, WI).

Compare that to the inherited Roth. A non-spouse heir of a Roth IRA must drain the account within 10 years of death under the SECURE Act’s rule (IRC §401(a)(9)(H), stats.md §5). The withdrawals are still tax-free, but the tax-free compounding stops — the money has to leave the shelter on a 10-year clock. As a pure estate asset to leave behind untouched, the stepped-up brokerage account is the better wrapper. Because the step-up erases the entire embedded gain, the brokerage’s full ~$4.34M passes tax-free at death — no LTCG bill, no 10-year clock.

Case 2: You realize gains inside the 0% LTCG bracket

For 2026, long-term capital gains are taxed at 0% when taxable income is at or below $48,350 (single) or $96,700 (MFJ) (stats.md §2). A retiree with low ordinary income — living off cash, a small pension, or deferring Social Security — can sell appreciated brokerage shares each year up to that ceiling and pay zero federal tax on the gain. In those years the brokerage matches the Roth’s tax treatment and keeps full pre-59½ access. This is the genuine tie, and it is why a brokerage bridge account next to the Roth is smart sequencing, not a mistake.

The liquidity catch before 59½

The Roth’s weakness is access. Your converted after-tax basis comes out tax- and penalty-free at any age once each conversion clears its own 5-year clock (IRC §408A(d)). But the earnings — the part that makes the Roth worth so much — cannot be touched before 59½ without a 10% penalty plus ordinary tax. A brokerage has no age gate, no 5-year rule, no penalty. You can sell on a Tuesday and have cash Thursday.

For Priya and Daniel, this argues for splitting the $46,000 by time horizon:

• Money they won’t touch before 59½ → mega-backdoor Roth. Maximum tax-free compounding, no downside since they won’t need early access.
• Bridge money for early retirement (age 50–59½) or emergencies → taxable brokerage. The liquidity and the 0%-bracket harvesting opportunity are worth the small tax drag.

The decision is not Roth or brokerage. It is which dollars go where. Most high earners maxing the mega-backdoor still want a taxable brokerage running alongside it for exactly the access reasons above.

What most people get wrong about the step-up

The cousin’s advice — “the step-up wipes out the tax, so use the brokerage” — is the most common error in this comparison. The step-up is real, but it only applies if you die holding the asset. The moment you sell to fund a car, a home, a child’s wedding, or your own retirement spending, the step-up is irrelevant and you pay the full capital-gains bill. For the typical accumulator who plans to spend a meaningful share of the portfolio in retirement, the step-up never triggers on the spent dollars, and the Roth’s zero-tax growth wins.

The second misconception: that the favorable 15% capital-gains rate makes the brokerage “tax-efficient enough.” A 15% rate is lower than ordinary income, but it is not zero — and the Roth is zero. Add the 3.8% NIIT for high earners and the brokerage’s “efficient” rate is closer to 18.8% on every realized gain. On a $2.96M gain, that is real six-figure money. “Lower than ordinary” is not the same as “free.”

The third: ignoring direct indexing’s limits. Direct indexing inside the brokerage harvests losses to offset gains and can recover part of the gap — but a portfolio held 30 years drifts to almost entirely gains, so the harvestable losses dry up after the early years. It might recover $50,000–$100,000 of the $445,000 gap, mostly front-loaded. Useful as a bridge-account enhancement; not a reason to skip the Roth.

Decision checklist

1. Confirm your plan supports it. The mega-backdoor only works if your 401(k) allows after-tax contributions and in-plan Roth conversion or in-service withdrawal. No feature, no strategy — route the $46,000 to the brokerage by default.
2. Earmark by time horizon. Dollars you won’t need before 59½ go to the Roth. Dollars you might need earlier go to the brokerage.
3. Convert immediately. Convert after-tax contributions to Roth the same week you make them, so no taxable earnings accrue in the after-tax bucket before conversion.
4. Track each conversion’s 5-year clock if you might tap converted basis before 59½ (IRC §408A(d)).
5. Plan the 0%-bracket harvest. In low-income early-retirement years, realize brokerage gains up to $96,700 MFJ to capture the 0% LTCG rate before larger income sources (Social Security, RMDs at 73–75) push you out of the band.

The lever

For Priya and Daniel, the answer is settled: send the bulk of the $46,000 to the mega-backdoor Roth and keep a smaller brokerage sleeve for liquidity and 0%-bracket harvesting. Over 30 years that choice is worth about $445,000 of tax they will never pay — unless their actual plan is to die holding the position and leave it to heirs, in which case the stepped-up brokerage is the better estate wrapper. The lever is not the rate of return; it is matching each dollar to the wrapper that taxes it least when you actually use it. Decide whether each dollar is “spend in retirement” money (Roth wins) or “leave behind untouched” money (step-up wins), and the allocation falls out automatically.