Withdrawal Strategy

A withdrawal strategy answers a narrow, high-value question: given money spread across taxable brokerage accounts, tax-deferred accounts like a traditional IRA or 401(k), and tax-free Roth accounts, which one do you spend from this year? The withdrawal amount is set elsewhere — that is the safe withdrawal rate question, and the research behind it lives on that page. The overall process of turning savings into income belongs to retirement income planning, and how the portfolio itself is segmented belongs to the bucket strategy. What this page owns is sequencing, and the reason it deserves its own page is that the same dollar of spending can generate wildly different tax bills depending on where it comes from — not just this year, but for the rest of retirement.

The conventional default and why it is often wrong. The standard advice is to spend taxable accounts first, then tax-deferred, then Roth last, on the logic that money should stay sheltered as long as possible. It gets one thing right — Roth generally does belong last, because tax-free growth is the most valuable kind and Roth IRAs have no lifetime required minimum distributions. But followed rigidly it produces a specific and expensive failure: several years of near-zero taxable income in early retirement, while an untouched traditional IRA compounds into a balance that later forces large distributions on top of Social Security, at higher rates, with no flexibility left. The saver paid almost no tax for a decade and then paid too much for twenty years.

The better framing is a target taxable income, not a fixed sequence. Ask what your total taxable income should be this year — usually enough to fill the low brackets you will never see again — and then draw from whichever accounts produce that number. In practice that means taking more from tax-deferred accounts in low-income years than a spend-taxable-first rule would, often alongside a Roth conversion, and topping up spending from taxable accounts. The years between the last paycheck and the start of Social Security and required minimum distributions are the cheapest tax years most people will ever have, and they do not come back.

The interaction effects are what make sequencing high-stakes. Four of them matter:

• The taxation of Social Security. How much of your benefit is taxable depends on provisional income, which includes tax-deferred withdrawals and even tax-exempt interest — but not qualified Roth distributions. In the range where each extra dollar of income drags another fraction of a Social Security dollar into taxable income, the effective marginal rate on an IRA withdrawal can be far above the bracket rate. This is often called the tax torpedo, and it is the single strongest argument for moving money out of tax-deferred accounts before benefits begin.
• Medicare premium surcharges. Higher-income beneficiaries pay an income-related adjustment on Part B and Part D premiums, based on the tax return from two years earlier. A large one-off withdrawal or conversion at 65 can raise premiums at 67, which makes the timing of big taxable events a two-year-lagged decision. The thresholds are set annually, so check the current figures before pushing income near one.
• Capital gains rates. Long-term gains are taxed at 0%, 15%, or 20% depending on total taxable income, so an ordinary-income withdrawal can push otherwise 0%-rate gains into the 15% band. The breakpoints are indexed each year. This cuts both ways: a low-income year is also an opportunity to realize gains deliberately at 0%.
• Required minimum distributions. From your required beginning age — 73, and 75 for anyone born in 1960 or later — the tax code forces withdrawals from traditional accounts whether you need the money or not. Every dollar moved out earlier at a low rate is a dollar not forced out later at a higher one, which is why the planning window closes at a known date.

The tools that implement it. Roth conversions in the gap years, sized to fill a bracket deliberately rather than by accident. Qualified charitable distributions, available from IRAs from age 70 1/2, which satisfy required minimum distributions without entering income at all — the most tax-efficient way for a charitable retiree to give. Specific-lot selection in taxable accounts, so a sale draws mostly on basis rather than gain, plus the step-up in basis that makes low-basis holdings worth keeping for heirs. And a cash reserve, so a bad market never forces you to sell an investment at the wrong time to fund a withdrawal — the account-order plan and the portfolio-structure plan have to be compatible.

What this page deliberately leaves alone. How much you can withdraw each year is the safe withdrawal rate question. How to segment the portfolio by time horizon is the bucket strategy. Which assets to hold in which account type in the first place is asset location, a related but distinct decision. And sequence of returns risk — the danger of bad early returns — is a portfolio risk, not a sequencing one, though a cash reserve helps with both.

A hypothetical example that shows why order matters more than it looks.

Carla retires at 66 with $1,400,000: $300,000 in a taxable brokerage account, $1,000,000 in a traditional IRA, and $100,000 in a Roth IRA. She needs $70,000 a year and is delaying Social Security to 70.

The conventional route. She spends the taxable account first. Because much of a withdrawal from a brokerage account is her own basis coming back, her taxable income for four years is close to nothing and her federal tax bill is close to zero. It feels like a win. Meanwhile the IRA compounds untouched, and at 70 her Social Security starts, and at her required beginning age the minimum distributions arrive on a balance that has grown substantially. From then on she has a large forced income she cannot control, a chunk of her Social Security pulled into taxable income by it, and no low-bracket years left to work with.

The bracket-filling route. In each of those four low-income years she takes a hypothetical $60,000 from the IRA — enough to fill the 10% and 12% brackets, whose top edge the IRS indexes annually — and tops her spending up from the taxable account. Over four years that moves $240,000 out of the IRA at a blended rate of roughly 11%, costing about $26,400 in federal tax. Left in the IRA, those same dollars would plausibly have come out later in the 22% bracket, costing about $52,800 — and that comparison still ignores the extra Social Security dragged into income alongside them. The difference on those dollars alone is roughly $26,400, and her eventual required distributions are smaller for good measure.

She can go further in the same years: convert some of that $60,000 to a Roth rather than spend it, so the dollars leave the IRA permanently and grow tax-free afterward. And because her taxable income is low, she may be able to realize long-term gains in the brokerage account at a 0% rate at the same time — provided the conversion does not push her past that breakpoint. Those two moves compete for the same bracket space, which is exactly the sort of trade-off a sequencing plan exists to resolve.

(All figures hypothetical. Real answers depend on your actual brackets, state tax, charitable intent, and what you expect to leave to heirs — a plan built on someone else's numbers is not a plan.)

Pros (of planning the order deliberately)

• Can reduce lifetime tax substantially without changing how much you spend or how you invest — it is one of the few genuinely free improvements available.
• Shrinks future required minimum distributions by moving money out of tax-deferred accounts while rates are low.
• Uses the gap years before Social Security and required distributions, which are time-limited and cannot be reclaimed later.
• Manages the second-order effects — Social Security taxation, Medicare surcharges, capital gains rates — that a simple sequence rule ignores entirely.
• Preserves Roth assets for the years when flexibility is worth the most, including as the account you draw from without raising taxable income at all.

Cons

• It requires annual attention and a tax projection, not a rule you set once.
• It depends on assumptions about future tax law and your own future income, both of which can change.
• Paying tax earlier than strictly necessary is a real cost if you turn out to be in a lower bracket later, or if you die earlier than expected and heirs would have received a step-up in basis on taxable assets.
• The interactions are genuinely complicated, and a small mistake near a Medicare or capital-gains threshold can cost more than the strategy saved that year.
• It is easy to over-optimize taxes at the expense of the plan's actual goals, including simplicity, which has value of its own.