Retirement withdrawal strategies

Saving for retirement has a scoreboard: the balance either grows or it does not. Spending it down is harder to judge, because the test runs for the rest of your life and the grade arrives too late to fix. You do not know how long you will live, what markets will return, or what inflation will do, yet you have to pick a number to withdraw this year anyway.

A withdrawal strategy is a set of rules for that decision: how much to take out each year, how to adjust when markets move, and which accounts to tap in what order. The rules exist to manage two opposite failures. One is running out of money in your 80s or 90s. The other, more common in practice, is dying with a large unspent balance after decades of needless scrimping. Every strategy in this article strikes a different balance between a steady paycheck and a safe portfolio, and each suits a different temperament.

The stakes compound because early mistakes are the expensive ones. A too-high withdrawal rate in a falling market does damage that later gains cannot fully repair, while a too-cautious rate quietly costs you trips, gifts, and help to family you could have afforded. This article walks through the research, the main strategy families, the tax layer, and the role of guaranteed income.

The 4 percent rule#

Nearly every modern withdrawal rule descends from a 1994 paper by financial planner William Bengen, "Determining Withdrawal Rates Using Historical Data," published in the Journal of Financial Planning ¹. Bengen tested a portfolio of half US large-company stocks and half intermediate-term Treasury bonds against every rolling 30-year retirement starting from 1926. His question: what is the highest first-year withdrawal rate that never exhausted the portfolio, even for the unluckiest retiree? The answer, driven by the person who retired in 1966 into stagflation, was just over 4 percent ¹.

The mechanics matter, because the rule is often misquoted. You withdraw 4 percent of the portfolio in the first year only. Each year after, you ignore the balance and raise the previous year's dollar amount by inflation. A $1 million portfolio supports $40,000 in year one; if inflation runs 3 percent, year two pays $41,200 whether markets rose or crashed. The portfolio, not your spending, absorbs every market move.

The 1998 Trinity study, by three professors at Trinity University, reframed the same history as success rates and made the rule famous ². Withdrawing an inflation-adjusted 4 percent from portfolios holding 50 to 75 percent stocks succeeded in at least 95 percent of historical 30-year periods; at 5 percent, success rates fell sharply, and 6 percent failed roughly a third of the time ².

It helps to be clear about what the rule is. It is a research finding about worst cases in one country's market history, not a spending plan. It ignores taxes and investment fees, assumes you never once adjust to circumstances, and covers exactly 30 years. Almost nobody should or does follow it mechanically. Its real use is as a benchmark: a quick answer to "roughly what income can this portfolio defend?"

Sources for this section: [1] [2]

What later research says#

The 4 percent figure has been pushed from both directions since 1994, and the disagreements come from assumptions, not arithmetic.

Bengen himself now considers 4 percent too conservative. In a 2025 book, A Richer Retirement, he raised his worst-case rate to 4.7 percent, with most of the gain coming from broader diversification: the updated work spreads the portfolio across seven asset classes instead of the original two ³. He emphasizes that 4.7 percent remains a worst-case floor, and that retirees in ordinary markets could often have spent more ³.

Morningstar's annual State of Retirement Income research argues for starting lower. Its edition published in December 2025 put the highest safe starting rate for 2026 at 3.9 percent, assuming a 30-year retirement, a 90 percent chance of money remaining at the end, 30 to 50 percent in stocks, and forward-looking return forecasts rather than US history ⁴. That was up from 3.7 percent the year before. The same research found that retirees willing to let spending vary could start as high as roughly 5.7 percent, and that simply accepting the well-documented decline in retiree spending with age, instead of assuming inflation-matched spending forever, raises the starting rate by about a percentage point ⁴.

Read together, the honest range for rigid, inflation-adjusted spending over 30 years is roughly 3.5 to 5 percent. Where you land inside that range depends on how diversified you are, how much failure risk you accept, how long your horizon runs (early retirees need lower rates; see the FIRE movement), and above all whether you can cut spending when markets demand it.

Sources for this section: [1] [2] [3] [4]

Fixed versus dynamic approaches#

Strategies differ mainly in where market risk lands: on your balance, on your paycheck, or split between them.

A fixed real strategy is the 4 percent rule as designed: identical purchasing power every year, with the portfolio absorbing all volatility. It suits retirees whose budgets are mostly fixed obligations, and it is the approach most likely to end in one of the two extremes, depletion or a large surplus.

A fixed percentage strategy withdraws a set share, say 4 or 5 percent, of whatever the balance is each January. The portfolio can never hit zero, but income swings with markets, and a deep bear market can cut your paycheck by a third within a couple of years. Few retirees tolerate that in practice.

Guardrails strategies split the difference and have become the mainstream middle ground. The best-known version comes from planner Jonathan Guyton and professor William Klinger: start at a higher rate than a fixed rule would allow, then apply decision rules ⁵. If markets fall far enough that your current withdrawal rate drifts 20 percent above the initial rate, you cut the year's spending by 10 percent; if markets rise enough that the rate drifts 20 percent below, you give yourself a 10 percent raise; in losing years you also skip the inflation increase. Later work, notably by planning researcher Michael Kitces, criticizes the classic rules for forcing repeated cuts in long bear markets and proposes risk-based guardrails that adjust more gently ⁵. The family resemblance is what matters: small, rule-driven spending changes made early prevent large forced ones made late, which is why Morningstar's flexible-strategy rates run so much higher than its fixed-rate estimate ⁴.

Sources for this section: [4] [5]

Sequence of returns risk#

Two retirees can earn the same average return over 30 years and end in different places, because for a portfolio under withdrawal the order of returns matters as much as their size. Imagine two $1 million portfolios paying out $40,000 a year, each averaging 6 percent over three decades. The one that hits a 35 percent decline in years one and two must sell shares at depressed prices to fund spending, and those shares are gone when the recovery comes. The one that hits the same decline in years 28 and 29 barely notices. Same market, different lives.

This is sequence of returns risk, and it is the reason withdrawal research obsesses over worst cases rather than averages. It concentrates in the years just before and after the retirement date, when the balance is largest and the remaining horizon longest. The standard defenses all reduce selling into weakness: a lower starting rate, a cash or short-term bond reserve, spending rules that trim withdrawals after losses, part-time earnings in the early years (see working in retirement), and an income floor that keeps essentials off the portfolio entirely.

Note: Sequence risk is highest in the five years on either side of your retirement date. A severe bear market in that window is the one scenario worth planning for explicitly, because withdrawals lock losses in; the same crash ten years earlier or later does far less harm.

The bucket strategy#

The bucket approach organizes the portfolio by when the money will be spent rather than by asset class. Planner Harold Evensky pioneered the idea in the 1980s with a simple cash-reserve design, and Morningstar's Christine Benz popularized a three-bucket version for retirees ⁶. Bucket one holds one to two years of planned withdrawals in cash. Bucket two holds roughly the next five to eight years in bonds. Bucket three holds the remainder in stocks. Spending comes from bucket one, which is refilled in good years by selling from whichever bucket has grown ⁶.

Analytically, a bucket portfolio is just an asset allocation wearing a costume: two years of cash plus eight of bonds on a $1 million, $40,000-a-year plan is a 60/40-ish portfolio however you label it, and studies find no return advantage over plain rebalancing ⁶. Its value is behavioral. Retirees who know the next several years of groceries sit in cash are measurably more willing to leave the stock bucket alone through a crash, and panic selling, not withdrawal math, is how many plans actually die. For people who found 2008 or 2022 sleepless, the costume can be worth its modest cost in cash drag.

Sources for this section: [6]

Which accounts to draw down first#

Once you know how much to withdraw, the tax layer decides how much you keep. The conventional order is taxable accounts first, tax-deferred accounts (traditional 401(k) and IRA) second, and Roth IRA money last ⁷. Taxable withdrawals are the cheapest, since only the gains are taxed and long-term capital gains rates run from 0 to 20 percent; meanwhile the tax-advantaged accounts keep compounding, and models suggest the sequence can add years of portfolio life ⁷. Roth money goes last because it grows tax-free, has no lifetime withdrawal mandate, and is the best account to leave heirs.

The rigid version, though, wastes a major opportunity: the low-bracket gap years. Someone who retires at 63, delays Social Security to 70, and faces no required distributions until 73 may spend a decade in the 10 or 12 percent brackets. Filling those brackets deliberately, either with traditional-account withdrawals or with Roth conversions, moves money out of future high-tax years at today's low rates ⁷. The countervailing details live in taxes in retirement: extra income in a given year can push capital gains out of the 0 percent bracket, make more of a Social Security benefit taxable, or cross an income-related monthly adjustment amount (IRMAA) threshold that raises Medicare premiums two years later.

In practice most careful plans blend all three account types every year: enough traditional-account income to use the low brackets, taxable or Roth dollars for spending above that line, and Roth reserved for income spikes such as a roof or a car. Who benefits from which order depends on bracket now versus bracket later, state taxes, health, and heirs, which is why this is the one part of withdrawal planning where personalized advice most often pays for itself.

Sources for this section: [7]

RMDs and your plan#

At 73, the government adds its own withdrawal rule. Required minimum distributions force money out of tax-deferred accounts each year, starting near 3.8 percent of the prior December 31 balance and rising with age ⁸. Two clarifications keep them in perspective. RMDs are a tax event, not a spending mandate: nothing stops you from reinvesting the after-tax proceeds in a taxable account. And they are not a withdrawal strategy endorsement, though an "RMD-style" rule (each year, divide the balance by remaining life expectancy) is a legitimate dynamic strategy in its own right, one that automatically adapts to both markets and age.

For retirees with large tax-deferred balances, RMDs are better treated as a planning deadline. The bracket-filling withdrawals and Roth conversions described above, made in the 60s and early 70s, shrink the balance that RMD percentages will later apply to, and from age 70 1/2 qualified charitable distributions can route IRA money directly to charity without it ever appearing in your income. The annual caps, rules, deadlines, penalties, and inherited-account provisions are covered in required minimum distributions.

Sources for this section: [8]

An income floor, annuitized#

Every strategy above leaves your paycheck partly at the market's mercy. A different school of thought, often called flooring, says essentials should not be negotiable in the first place: add up housing, food, insurance, and utilities, then cover that amount with income that arrives regardless of markets, and let the portfolio fund only the wants above it.

Floors are built from Social Security, a pension if you have one, and, when those fall short, an income annuity: a contract that converts a lump sum into a monthly check for life. For most people the cheapest floor upgrade is not a product at all but a claiming decision, since delaying Social Security toward 70 buys more inflation-protected lifetime income per dollar forgone than any commercial annuity; when to claim Social Security works through that math. Beyond it, a single premium immediate annuity can close a specific gap, and a qualified longevity annuity contract, a deferred annuity bought inside an IRA with payments starting as late as 85, doubles as late-life insurance while shrinking the balance RMDs are calculated on. Current payout rates, fees, guaranty protections, and the sales pitfalls are covered in the annuities article.

The cost of a floor is flexibility: annuitized dollars are usually gone from your estate and cannot be re-liquidated, and level payments lose purchasing power to inflation. The benefit is that the rest of the plan gets easier. A retiree whose essentials are covered can run a higher withdrawal rate, tolerate guardrail cuts, and ride out bear markets without touching necessities.

Choosing among the approaches#

No single strategy wins on every axis, and the research is genuinely unsettled between the 3.9 percent camp and the 4.7 percent camp ³⁴. The practical questions are narrower than the debate suggests. How much of your budget is fixed? Fixed-heavy budgets argue for a floor and a conservative rate; flexible ones can start higher with guardrails. How long is your horizon? A 55-year-old needs a lower rate than a 70-year-old, full stop. Would you rather risk cutting spending later or risk underspending for thirty years? Dynamic rules trade the second risk for the first. And who manages this at 88? Simplicity, automatic rules, and guaranteed income all age better than a strategy that needs a sharp analyst at the wheel.

Whatever you choose, the plan deserves an annual review rather than a one-time setting, and spending targets belong next to a real budget, as described in budgeting in retirement. Withdrawal rates fail in the literature; in real households, what fails is usually the absence of any rule at all.

Sources for this section: [3] [4]

RetiredWiki. (2026, July 6). Retirement withdrawal strategies. https://retiredwiki.com/article/retirement-withdrawal-strategies