This article is educational and not tax, legal, or investment advice; unless labeled otherwise, all figures reflect 2026 tax year rules and CMS 2026 Medicare amounts. Consult a CPA or CFP before executing a withdrawal sequence.
TL;DR — Quick Verdict
- Bracket-filling — withdrawing from tax-deferred accounts up to the top of the 12% bracket, then switching to Roth or taxable — beat the conventional taxable-first order by roughly $94,000 in our 25-year model on a $1.4 million portfolio.
- The conventional advice (spend taxable, then tax-deferred, then Roth) leaves the 10% and 12% brackets unused for a decade, then forces $80,000+ RMDs at age 73 taxed at 22% or higher.
- Morningstar’s 2025 research puts the base-case starting safe withdrawal rate at 3.9% for a 30-year horizon at 90% success — not 4%, and flexible strategies reach 5.7%.
- Crossing the 2026 IRMAA tier-one threshold of $218,000 in modified adjusted gross income costs a married couple $2,297 per year in Medicare surcharges — triggered by one dollar of excess income.
- Recommendation: model your age 73 required minimum distribution first, then work backward. If projected RMDs push you past the 22% bracket or an IRMAA cliff, start tax-deferred withdrawals or Roth conversions in your 60s.
A retiree with $900,000 in a traditional 401(k) will face a required minimum distribution of roughly $34,000 in the first RMD year, because the IRS Uniform Lifetime Table sets the age-73 divisor at 26.5. That is not optional income. It arrives whether the retiree needs the cash or not, and it stacks on top of Social Security, pension payments, and dividends. Fidelity, Vanguard, and Schwab all publish withdrawal-order calculators that default to the same sequence — taxable first, tax-deferred second, Roth last — and that default is wrong for a large share of the people using it.
This analysis compares four withdrawal sequences head to head using 2026 federal tax parameters from IRS Revenue Procedure 2025-32, the Uniform Lifetime Table from IRS Publication 590-B, and the CMS 2026 Medicare premium schedule. We model a specific couple across 25 years, show the arithmetic behind each strategy, quantify the two cliffs most retirees hit without seeing them coming, and identify the household profiles where the conventional order actually wins.
The Four Withdrawal Sequences, Defined Precisely
Withdrawal strategy means the order in which a retiree draws from three account types with three different tax treatments: taxable brokerage accounts (capital gains rates on realized gains only), tax-deferred accounts such as traditional 401(k)s and IRAs (ordinary income rates on every dollar), and Roth accounts (no federal tax on qualified distributions).
Four sequences dominate practice. Taxable-first exhausts the brokerage account, then tax-deferred, then Roth — the sequence most brokerage tools default to. Proportional draws from all three each year in proportion to their balances, holding the tax character of the portfolio constant. Bracket-filling withdraws from tax-deferred accounts up to a chosen bracket ceiling, then covers remaining spending from taxable or Roth assets. Roth-first spends the tax-free bucket earliest, which almost never wins but occasionally makes sense for someone expecting a large one-year income spike.
The distinction matters because tax-deferred balances are not really yours. A $1,000,000 traditional IRA belongs partly to the IRS, and the government’s share is determined by the bracket you are in when the money comes out — a bracket you partially control through timing. Understanding your Roth versus traditional IRA tax treatment is the foundation, but the withdrawal decision is where that theory gets tested against real money.
One more variable overlays all four: Social Security timing. Delaying benefits creates a low-income window in the early 60s that bracket-filling exploits. That interaction is covered in our analysis of Social Security claiming age and lifetime income.
2026 Tax Parameters That Drive Every Withdrawal Decision
Every strategy comparison depends on the exact bracket thresholds, because the whole game is deciding which dollars fall into which bracket in which year. The 2026 figures below come from IRS Revenue Procedure 2025-32.
Source: Internal Revenue Service, Revenue Procedure 2025-32 inflation adjustments for tax year 2026.
Three deductions widen the low-tax runway for retirees specifically. The 2026 standard deduction is $32,200 for married couples filing jointly and $16,100 for single filers. Taxpayers 65 and older add $1,650 each if married, or $2,050 if single. The One Big Beautiful Bill Act layers on a temporary senior deduction of $6,000 per qualifying person for tax years 2025 through 2028, phasing out above $75,000 of modified adjusted gross income for individuals.
Stack those for a 66-year-old couple: $32,200 plus $3,300 plus $12,000 equals $47,500 of deductions before a dollar of tax. Add the 12% bracket ceiling of $100,800 in taxable income, and the couple can realize $148,300 of gross ordinary income while never leaving the 12% bracket. Most retirees never use that room.
Modeling a Real Couple: $1.4 Million Across Three Buckets
Consider David and Maria, both 65 in 2026, retiring with $600,000 in a traditional 401(k), $500,000 in a taxable brokerage account with a $350,000 cost basis, and $300,000 in a Roth IRA. They need $80,000 per year in after-tax spending and plan to claim Social Security at 70, expecting a combined $62,000 annually at that point.
Under taxable-first, they spend the brokerage account from 65 to 70, realizing modest capital gains taxed at 0% or 15%. Their ordinary income is near zero for five years, meaning the 10% and 12% brackets go completely unused — roughly $148,300 of low-tax capacity wasted each year, or $741,500 across the window. Meanwhile the 401(k) compounds at an assumed 5.5% nominal, growing from $600,000 to roughly $784,000 by age 73.
Applying the Uniform Lifetime Table divisor of 26.5, their first RMD lands near $29,600. Add $62,000 of Social Security, and up to 85% of those benefits become taxable because provisional income sits far above the $44,000 married-filing-jointly threshold. Their gross income clears $100,000 and pushes them into the 22% bracket permanently.
Under bracket-filling, David and Maria withdraw $70,000 annually from the 401(k) between 65 and 70, pay roughly 10% to 12% on it, and cover the rest of their spending from the brokerage account. By 73 the 401(k) has been drawn down to roughly $460,000, producing an RMD near $17,400 instead of $29,600. The 22% bracket never arrives.
Our 25-year model puts the lifetime federal tax difference at approximately $94,000 in favor of bracket-filling — a figure that assumes constant real bracket thresholds and no state income tax. State treatment changes the arithmetic substantially and is modeled separately in our work on RMD calculation and tax costs.
Taxable-First vs Bracket-Filling: Which Is Better for a $1M+ Portfolio?
Cost comparison across the two dominant strategies, using David and Maria’s numbers:
Modeled by Real Cost Report using 2026 bracket thresholds from IRS Revenue Procedure 2025-32 and the Uniform Lifetime Table in IRS Publication 590-B (verify at irs.gov). Assumes 5.5% nominal growth and constant real thresholds; results are modeled, not measured.
Front-loading tax is counterintuitive. Bracket-filling costs David and Maria $46,100 more in tax during their 60s, and that expense feels like a mistake in year one. The payback arrives from 73 onward, when a smaller tax-deferred balance produces smaller mandatory distributions taxed at a lower rate for two decades.
Verdict
Bracket-filling wins decisively for households with tax-deferred balances above roughly $500,000 and a gap between retirement and RMD age. Taxable-first wins for two profiles: retirees whose total portfolio is small enough that RMDs will never exceed the standard deduction, and retirees who expect to leave assets to charity, since tax-deferred dollars pass to a qualified charity untaxed. If your projected age-73 RMD plus Social Security keeps you under $100,800 of taxable income, the conventional order is fine and simpler to administer.
The Two Cliffs: IRMAA and Social Security Taxation
Marginal rate tables understate what retirees actually pay, because two federal provisions create step functions rather than smooth slopes.
Medicare’s income-related monthly adjustment amount is the harsher of the two. CMS set the 2026 standard Part B premium at $202.90 per month, with surcharges beginning at $109,000 of modified adjusted gross income for single filers and $218,000 for joint filers, measured from the tax return filed two years earlier. Cross the joint threshold by one dollar and both spouses pay tier-one surcharges for the full year — roughly $2,297 in combined Part B and Part D cost for the couple. At the top tier, beginning at $500,000 single and $750,000 joint, the total monthly Part B premium reaches $689.90.
Because the lookback runs two years, a Roth conversion executed in 2026 sets the 2028 Medicare premium. Retirees planning conversions should read our detailed treatment of the IRMAA surcharge impact on retirement income before committing to a conversion amount.
Social Security taxation creates a second, subtler cliff. Provisional income — adjusted gross income excluding benefits, plus tax-exempt interest, plus half of benefits — determines how much of a benefit is taxable. Those thresholds sit at $25,000 and $34,000 for single filers and $32,000 and $44,000 for joint filers, and they have never been indexed to inflation since Congress set them. Inside the phase-in range, each additional dollar of IRA withdrawal can make an additional $0.85 of Social Security taxable, producing effective marginal rates of 22.2% for someone nominally in the 12% bracket.
That interaction is the single strongest argument for doing tax-deferred withdrawals before benefits begin. A $70,000 IRA withdrawal at 66 with no Social Security costs 12%. The identical withdrawal at 71 alongside benefits can cost 22.2%. Retirees still earning wages face a third overlay covered in our guide to earned income effects on Social Security benefits.
What Most People Get Wrong
Five errors recur across withdrawal plans, and each carries a quantifiable cost.
Mistake 1: Treating 4% as the rule. Morningstar’s 2025 State of Retirement Income research, using data as of September 30, 2025, puts the base-case starting safe withdrawal rate at 3.9% for a 30-year horizon at a 90% success probability — and just 3.3% for a 40-year horizon. Consequence: an early retiree drawing 4% on a 40-year horizon materially raises depletion risk. Correct action: match the rate to your horizon, and note that Morningstar found flexible spending methods supported starting rates up to 5.7%.
Mistake 2: Waiting until 73 to think about RMDs. By the time the first distribution is due, the tax-deferred balance is fixed and the divisor is fixed. Consequence: a permanently higher bracket. Correct action: project the age-73 balance in your late 50s using retirement savings benchmarks and catch-up costs as a reference point.
Mistake 3: Converting to Roth without checking the IRMAA tier. A conversion that pushes joint modified adjusted gross income from $217,000 to $219,000 adds $2,297 of Medicare cost two years later — an effective marginal rate above 100% on that $2,000. Correct action: size conversions to land just under the tier boundary, using the framework in our Roth conversion costs and timing analysis.
Mistake 4: Selling the wrong lots in the taxable account. Highest-cost-basis lots produce the smallest gain per dollar withdrawn, yet most brokerage defaults sell first-in-first-out. Consequence: unnecessary capital gains that inflate modified adjusted gross income and can trigger the cliffs above. Correct action: set specific-lot identification with your custodian before the first withdrawal.
Mistake 5: Ignoring sequence risk in the first five years. Withdrawing a fixed dollar amount during a market decline permanently locks in losses. Consequence: portfolios that fail despite an adequate average return. Correct action: hold one to two years of spending in cash or short Treasuries and pair it with the guardrails logic described in our coverage of sequence-of-returns risk in early retirement.
Who Should Change Their Withdrawal Order — And Who Should Not
Bracket-filling is not universally superior. The decision turns on four conditions, and the honest answer for many households is that the conventional order is adequate.
Change your order if your tax-deferred balance exceeds roughly $500,000 and you have at least three years between retirement and your required beginning date. Change it if you are delaying Social Security past full retirement age, which creates precisely the low-income window bracket-filling needs. Change it if you expect a surviving spouse to file as single, since the single brackets are half as wide and the survivor’s effective rate jumps sharply on the same income.
Keep the conventional order if your combined retirement accounts fall below roughly $400,000, because projected RMDs will stay inside the standard deduction and additional complexity buys nothing. Keep it if you plan large charitable bequests from tax-deferred accounts. Keep it if you are still building the balance rather than drawing it down, in which case the more relevant question is your 401(k) contribution limit and maxing-out value.
Household structure changes the calculus too. Couples who divided accounts through a qualified domestic relations order often hold asymmetric balances that make proportional withdrawal easier to administer than bracket-filling; our analysis of divorce impact on retirement accounts and QDROs covers that split. Retirees with a defined benefit stream face a different starting point entirely, since the pension already fills the lower brackets — see defined benefit pension value versus a 401(k).
Frequently Asked Questions
Can I skip an RMD if I do not need the money?
No. The excise tax on a missed required minimum distribution is 25% of the shortfall under Internal Revenue Code Section 4974, reduced to 10% if corrected within a two-year window. You may withdraw more than the required amount at any time, and after age 70½ you can satisfy the requirement through a qualified charitable distribution, which excludes the amount from adjusted gross income entirely.
Does the RMD age move to 75?
Yes, but not yet for most current retirees. Under SECURE 2.0, individuals born from 1951 through 1959 begin required distributions at age 73. Those born in 1960 or later begin at 75, starting in 2033. For 2026, the applicable age is 73, and the Uniform Lifetime Table divisor at that age is 26.5, implying a first-year withdrawal floor of about 3.77% of the prior year-end balance.
Do Roth withdrawals count toward IRMAA?
Qualified Roth distributions do not count toward modified adjusted gross income, so they do not affect Medicare surcharges or Social Security taxation. Roth conversions do count in the year performed. Because CMS applies a two-year lookback, a conversion completed in 2026 determines your 2028 premium against thresholds that begin at $109,000 single and $218,000 joint in 2026 terms.
How much of my Social Security will be taxed?
It depends on provisional income, not benefit size. Married couples filing jointly owe nothing below $32,000, face up to 50% inclusion between $32,000 and $44,000, and up to 85% above $44,000. Single filers use $25,000 and $34,000. These thresholds are statutory and have never been indexed to inflation, so cost-of-living increases steadily pull more retirees into taxable territory.
How We Researched This Article
Tax parameters were drawn directly from Internal Revenue Service Revenue Procedure 2025-32 and the accompanying IRS announcement of 2026 inflation adjustments, which supplied bracket thresholds, the standard deduction amounts of $16,100 single and $32,200 married filing jointly, and the additional age-65 deduction amounts. Required minimum distribution mechanics, including the age-73 Uniform Lifetime Table divisor of 26.5 and the 25% excise tax on shortfalls, come from IRS Publication 590-B, Appendix B, Table III. Medicare figures — the $202.90 standard 2026 Part B premium, the $109,000 and $218,000 first-tier thresholds, and the $689.90 top-tier monthly premium — come from the Centers for Medicare & Medicaid Services 2026 Parts A and B premium release published November 14, 2025 (verify at cms.gov). Social Security taxation thresholds come from Internal Revenue Code Section 86 as documented in IRS Publication 915. Safe withdrawal rate figures are from Morningstar’s State of Retirement Income: 2025, with data as of September 30, 2025.
The David and Maria scenario is a deterministic model, not a Monte Carlo simulation and not measured outcomes. It assumes 5.5% nominal annual growth applied uniformly across account types, constant real bracket thresholds, no state income tax, no long-term care event, and no change to federal law during the projection period. Each of those assumptions is a real limitation. State income tax alone can shift the strategy ranking, since several states exempt Social Security and retirement income entirely while others tax both. The senior deduction created by the One Big Beautiful Bill Act is scheduled to expire after tax year 2028, which materially affects the low-bracket runway in later years of the model. Lifetime tax totals are rounded to the nearest thousand dollars and should be treated as directional comparisons between strategies, not as forecasts of any individual household’s liability. Research was last conducted July 2026. All figures were verified against named primary sources before publication.