The Four Percent Rule

Contents

1. The man with the spreadsheet
2. What the rule actually says
3. The revisions
4. The critics
5. What retirees do
6. The two enemies
7. The strongest evidence

Brent O’Mara has celebrated anniversaries and retirements with clients, and the question every retirement eventually asks is the one this entry’s number was invented to answer. The measurement is a withdrawal rate, and its unit is a percentage of the balance a person holds on the day retirement begins. Four percent of that starting balance is withdrawn in the first year; thereafter the dollar amount, not the percentage, is adjusted for inflation each year, much as Social Security benefits receive a cost of living adjustment.¹ The second unit is time. The rule was built to produce a paycheck that lasted 30 years, and it was tested against every 30 year stretch of American market history from 1926 onward.³

The man with the spreadsheet

In the early 1990s Bill Bengen, then a financial planner in southern California, had several anxious clients asking how much they could spend without running out of money, and found that his textbooks offered no guidance rooted in facts.³ He built the answer on his own, from a spreadsheet and a book of historical figures, and published the result in the Journal of Financial Planning in October 1994.¹ The New York Times later called it the most widely referenced rule of thumb in the field, and Bengen, an aeronautical engineer by training who retired in 2013, warned from the start that it was not a law of nature and that a worse case than any in the record could someday arrive.³

What the rule actually says

Bengen’s original assumptions were specific: savings held in a tax advantaged account, a fixed mix of stocks and bonds reset to its target once a year, a 30 year horizon that could end with a zero balance, and withdrawals that started at a set percentage and rose with inflation.² Under those assumptions no retiree in the historical record ran out of money over 30 years at a first year rate just above 4 percent, which rounded to the rule.²¹ The commonest misreading, he told CNBC in 2025, is to withdraw 4 percent of the current balance every year; the rule fixes only the first year.¹ The Times summarized the mechanism in 2025: even across the three worst decades for stocks on record, which began in October 1968, a retiree who followed the rule still had money left after 30 years.⁷

Figure

How the rule is applied

1. 01
   Year oneWithdraw 4 percent of the starting balance.
2. 02
   Later yearsAdjust that dollar amount for inflation each year, much as Social Security applies a cost of living adjustment.
3. 03
   VariantsSome retirees instead take a fixed percentage each year, or more early and less later.
4. 04
   ContextThe horizon, the account types, any inheritance goal, the mix of holdings and the timing of withdrawals all bear on the rate.

The revisions

The author has revised the number more than once. With a more diversified mix of holdings he later raised it to 4.5 percent.³ In 2025, with a book expanding the original research, he set a new default of 4.7 percent for a 30 year retirement, a figure he calls the historical maximum safe rate for all retirees, and about 4.2 percent for a 50 year retirement; he added that the average across all the historical starting points was about 7.1 percent, that anyone expecting to leave money to heirs must cut the rate, and that the historical maximum is based on past results and carries no promise about future retirees.¹² The 2025 figure also rests on a different mix than the 1994 study, with cash added and a broader set of stock holdings.²

The critics

Criticism has come from three directions. The first concerns conditions. Wade Pfau, a professor of retirement income, argued in 2015 that historically low interest rates combined with high stock valuations put new retirees in uncharted waters, and his analysis found that a retiree who wanted assurance of never cutting spending below a floor would need to start at 2.85 to 3 percent, at the cost of usually dying with most of the money unspent.³ Michael Kitces answered that the rule was built around some rather horrific past bear markets and saw no reason to assume the next would be worse, and Jonathan Guyton’s rules allowed nearly 5 percent at the start in exchange for cuts in losing years.³ The second concerns what the rule leaves out: investment costs, taxes, horizons other than 30 years and the fact that few retirees spend in a straight line.³ Jack Guttentag, writing in Forbes, called its principal weakness that the retiree does not know and cannot easily control the risk involved, from an asset return lower than assumed or a life longer than expected, and proposed a replacement that makes age, inflation, horizon and assumed return explicit inputs.⁴ The third concerns the number itself. Anthony Webb, an economist, estimated that 3 percent is a safer thumbnail if the goal is to lower the odds of running out.⁵

Figure

The rate Bengen has published over time

• 1994 study, 30 year retirement4.1
• 2025 book, 30 year retirement4.7
• 2025 book, 50 year retirement4.2

What retirees do

Most retirees follow no rule at all. About half of retirees have no method of any kind for drawing down savings and only 22 percent follow a plan, according to a survey the Times reported in 2025, and almost two in three Americans fear outliving their money more than they fear death.⁷ A 2023 survey described by the Center for Retirement Research found retirees split between two extremes, withdrawing chunks out of necessity or withdrawing little or nothing out of fear; nearly half withdrew 3 percent or less, and most of those took under 1 percent, which the center judged too cautious to enjoy.⁵ One in four of those retirees had left work earlier than planned.⁵ The IRS table that governs required minimum distributions is itself a rival method, and the Stanford Center on Longevity found that a strategy built on it, with a well timed Social Security claim, compared favorably with 292 others.⁶ What rate fits a household’s accounts and tax position depends on personal circumstances, and settling it is the business of qualified professionals.

The two enemies

Bengen names two things that shorten the life of a retirement portfolio: a bear market early in retirement, which drains the balance at the same moment withdrawals are drawing from it, and high inflation, which he calls the greatest enemy of retirees.²¹ He created the rule during a stretch of low, stable inflation, and he noted that the record 8.7 percent cost of living adjustment of 2023 was mostly claimed by higher prices.¹ The first enemy has its own entry under the sequence of returns and the second under inflation across a long retirement.

Figure

What retirees actually do

of retirees take no systematic approach to withdrawing their savings

about half

follow any kind of plan

22 percent

of Americans say they fear running out of money in retirement more than death

nearly two thirds

The strongest evidence

The firmest ground is the narrowest. On the mechanism there is no dispute: a fixed real withdrawal from a shrinking balance in the early years is what the rule was designed to survive, and the historical test it passed covered every 30 year period from 1926 at the original rate.³⁷ On the exact number the evidence is weaker, because each revision has rested on a different mix of holdings and a different reading of the same past, and because the strongest version of the rule is a statement about history that its author declines to extend to the future.² The Stanford analysis made the same point from the other side: using historical assumptions produced dramatically higher projected incomes than forward looking ones, so the choice of assumption can decide what a retiree spends.⁶ Where pensions once made the spending problem simpler, the rule is the clearest evidence of what replaced them.