Skip to content

The 4% Rule: The Math Behind Withdrawal Rates

Where 4% comes from, the verified arithmetic of inflation-adjusted withdrawals, a two-path example showing why average returns lie, and the guardrail rules practitioners actually run.

By the editorial team at The Long View How we check our guides 10 min read
A retirement planning scene with a monthly budget notebook, a compass, and a small plant on a bright desk

The 4% rule is the most quoted number in retirement planning: retire with a portfolio, withdraw 4% of it in year one, raise the dollar amount with inflation each year, and - according to decades of historical backtests - the money lasts 30 years in most market sequences. It is a genuinely useful starting point, and it is also widely misquoted in both directions: as a guarantee (it is not) and as an obsolete relic (it still tells you something). This guide shows where the rule comes from, runs the actual math with verified numbers, demonstrates with a two-path example why identical average returns can produce opposite outcomes, and explains the refinements practitioners actually use today.

Where the 4% comes from

The rule descends from two research traditions. In 1994, financial planner William Bengen tested U.S. stock and bond market data back to the 1920s, looking for the highest initial withdrawal rate that survived every 30-year retirement window in his dataset - including retirees who happened to start right before the Great Depression. For a portfolio split roughly between stocks and bonds, that rate came out to about 4% (4.15% on his final dataset). A few years later, the "Trinity study" (Cooley, Hubbard, and Walz, 1998, updated several times) reached a similar conclusion using different data: a 4% initial withdrawal, inflation-adjusted annually, succeeded over 30-year horizons in roughly 95% of historical rolling periods for a stock-heavy portfolio.

Read that sentence carefully, because it contains the rule's two big ideas and its one big limitation. The ideas: the withdrawal is a percentage of the initial portfolio, held constant in dollars and adjusted only for inflation; and the test asks about survival across all historical sequences, not just average ones. The limitation: the past is one dataset, not a law of nature. The 95% figure means 95% of the windows studied would have worked - not that there is a 95% chance your future will work.

The mechanics, with real numbers

Take a $1,000,000 portfolio and a 4% starting rate. Year one, you withdraw $40,000. From then on, the withdrawal amount is not tied to the portfolio's balance - it is last year's dollar amount plus inflation. At 3% inflation, the withdrawal schedule looks like this:

Year Withdrawal Calculation
1 $40,000 4% of initial portfolio
2 $41,200 $40,000 x 1.03
5 $45,020 $40,000 x 1.03^4
10 $52,191 $40,000 x 1.03^9
20 $70,140 $40,000 x 1.03^19
30 $94,263 $40,000 x 1.03^29

Notice what this implies: the effective withdrawal rate rises over time even though the rule is called "4%." By year 20 you are pulling $70,140 - if the portfolio happens to be at $1,000,000 again by then, you are effectively withdrawing 7% of it. The rule survives because good markets, historically, grew portfolios faster than inflation grew withdrawals. It breaks when markets refuse to cooperate for long enough, which brings us to the part most explanations skip.

Why average returns are not the point: a two-path example

A natural reaction to the 4% rule is: "my portfolio should average 7% a year, and I am only taking 4% plus inflation, so the math obviously works." The error is that withdrawals compound sequence, not just average. Here is a demonstration with deliberately simple numbers, computed exactly.

Two retirees, each with $1,000,000, each withdrawing 4% of the initial portfolio and raising the amount 3% a year. Over 30 years, both experience the same set of annual returns - 15 years of +12% and 15 years of -2%, an average of +5% per year in both cases - just in opposite order:

Path Return order Outcome over 30 years
A +12% years 1-15, then -2% Survives; ends with ~$2.42M
B -2% years 1-15, then +12% Depleted in year 25

Same average return, same withdrawal rule, opposite endings. Retiree B ran out because early losses hit a portfolio that withdrawals are simultaneously draining: every dollar withdrawn in a down year is a dollar that never participates in the later recovery. Retiree A banked the good years first, so the flat decade barely registered.

This is why Bengen's backtest found 4% and not 7%: the rate has to survive the worst historical sequences - retirements beginning in 1929, 1966, 1973 - not the pleasant ones. The worst-case paths are the whole game. (For a deeper treatment of this mechanism, see our guide to sequence-of-returns risk in retirement.)

Sensitivity is brutal. On that same bad-early path, the initial withdrawal rate decides survival almost by itself: 4.0% depletes in year 25, 3.5% in year 19, 3.2% in year 21 in this simplified model - while the good-early path at 4% ends with $2.4 million. A rule tested only on average returns will mislead you precisely when it matters. This is also why withdrawal plans need in-retirement adjustment rules, which the refinements section covers.

What the rule quietly assumes

Every assumption below is defensible but worth stating, because each one shifts the answer if your situation differs:

A 30-year horizon. The backtests ask "does it last 30 years?" Retiring at 50 with a 45-year horizon is a different problem; common practice scales the initial rate down for longer horizons (often toward 3.0-3.5%), because more time means more chances to meet a hostile sequence.

A specific portfolio. Bengen used roughly 50-60% stocks with the rest in bonds (he later found intermediate-term Treasuries worked better than long bonds). The 4% figure is not portfolio-agnostic: all-bond and all-cash portfolios historically supported lower sustained rates, and the Trinity study showed the pattern varying with stock allocation.

U.S. market history. The dataset is dominated by a century of exceptional U.S. equity returns. International datasets and forward-looking simulations built from current bond yields generally produce lower "safe" rates - a recurring critique since the 2010s, and one reason many planners now start planning conversations at 3.5% rather than 4%.

Gross of costs and taxes. The studies largely ignore fund fees and taxes. A portfolio paying 1% in all-in costs has, in effect, a 3% rule. Tax treatment depends on account type and withdrawal order - our Roth vs traditional tax math guide covers why the same dollar of withdrawal can carry different tax bills from different accounts.

The refinements people actually use

Nobody serious runs a fixed rule blind for 30 years. Three refinements dominate:

Constant-percentage withdrawals. Withdraw a fixed percentage of the current balance each year - say 4% of whatever the portfolio is worth. The portfolio mathematically can never reach zero, but the income swings: $40,000 the year the market is up, $24,000 after a 40% crash. In our bad-early example above, a retiree withdrawing 4% of current balance ends with about $1.25M - but with income that fell from $40,000 to roughly $20,000 in the early flat years. Lower volatility of survival, higher volatility of lifestyle.

Guardrails (Guyton-Klinger style). Start at a rate near 4-5%, then adjust: if the current withdrawal rate drifts more than about 20% above its starting value (because markets fell), cut spending by around 10% until it drifts back; if it drifts well below (markets boomed), allow a raise. The rule trades small, planned, temporary spending cuts for a much lower probability of catastrophic failure. The cost is that you must actually execute the cut in a bear market - which is harder emotionally than it sounds on paper.

Floor plus upside. Cover non-negotiable expenses with guaranteed income (Social Security claiming strategy, pensions, or a bond/TIPS ladder sized to the gap), and let portfolio withdrawals fund discretionary spending that can flex in bad years. This is less a formula than a structure, but it directly attacks sequence-of-returns risk: the money that must be spent is not sitting in assets that get sold at the worst time. Our TIPS real-yield guide explains how inflation-protected bonds anchor such a floor.

How to use the 4% number without being burned by it

Three practical habits turn the rule from a slogan into a planning tool:

Use it as a sizing anchor, not a withdrawal script. "4% of initial" is most useful in reverse: if you need $50,000 a year from a portfolio, the rule says a $1,250,000 portfolio historically supported that - which makes the savings target concrete. The number helps you plan decades before you retire; the day-to-day withdrawals are where guardrails belong.

Track your effective rate, not the starting one. After year ten, compute withdrawal divided by current balance. Above roughly 6%, you are in the danger zone where a sustained downturn forces action; below 4%, you have room for a raise. This one number tells you which regime you are in, whichever rule you follow.

Pre-commit the adjustment rules now. Decide, while markets are calm and you are not yet retired, what you will cut and by how much if your effective rate breaches a threshold. Sequence risk is unfixable in the moment; the only protection is rules made in advance and followed under stress. Retirees who keep even one flexible lever - working a year or two longer, trimming discretionary spending, pausing inflation raises in bad years - materially change the survival math without touching the portfolio.

One more integration point: withdrawal sequencing interacts with rebalancing. Selling from whichever asset class is overweight (rather than forced selling of whatever is down) is the quiet hero of withdrawal plans; the mechanics are in rebalancing bands: when to rebalance. And for the broad frame of how a stock/bond mix affects both the growth and the drawdown years, see asset allocation: choosing a stock and bond mix and the inflation and real return math.

The honest summary

The 4% rule is a historically calibrated starting point that says: with a balanced portfolio, a withdrawal of about 4% of the initial portfolio, inflation-adjusted, survived nearly all 30-year U.S. market sequences - including the worst ones - but not all of them, and not necessarily under future conditions. Its real lessons are the two things the headline number hides: survival depends on return sequence more than return average, and fixed rules need adjustment mechanisms to be safe in practice. Plan with the number, monitor the effective rate, pre-commit the guardrails, and treat any claim that the rule is either a guarantee or worthless as a signal that the speaker has not run the math.

This article is for general educational purposes only and is not investment, tax, or financial advice. Historical success rates do not predict future results, and individual circumstances - horizon, taxes, other income sources, health - change the appropriate plan. Consider consulting a fee-only fiduciary financial planner before acting on any withdrawal strategy, and never make decisions based on a single article, including this one.