Where the 4% Rule Came From

The 4% rule gets quoted so often that it can sound like a law of nature. It is not. It is a research finding with a specific history, specific assumptions, and specific limits — and knowing that history is the difference between using the rule well and being misled by it.

Before 1994: planning on averages

For decades, the standard way to answer "how much can I spend?" was to reason from long-run average returns. If a balanced portfolio had averaged, say, 7% a year — a number chosen here purely for illustration — it seemed to follow that a retiree could spend something close to that average and let the portfolio replenish itself forever.

The flaw in that logic is ordering. Averages are computed over decades, but a retiree lives through one specific sequence of good and bad years. Someone who hits a deep bear market or a burst of inflation in the first years of retirement is selling assets at depressed prices to fund withdrawals, and the portfolio may never recover even if returns later revert to a perfectly normal average. Two retirees can experience identical average returns and end up in completely different places. This is sequence-of-returns risk, and pre-1994 rules of thumb simply ignored it.

Bengen's 1994 study: ask about the worst case

William Bengen, a financial planner with an engineering background, published "Determining Withdrawal Rates Using Historical Data" in the Journal of Financial Planning in 1994. His method was simple and, in hindsight, obviously right: instead of asking what a retiree could spend on average, he asked what a retiree could have spent in the worst historical case. He replayed hypothetical retirements beginning in each historical year in his data, each one withdrawing a fixed percentage of the starting portfolio and then adjusting that dollar amount for inflation every year afterward.

The result: with a portfolio holding a substantial allocation to stocks — roughly half or more — an initial withdrawal rate of about 4% survived every 30-year retirement period in his historical record, including retirements that began on the eve of the Great Depression and at the start of the 1970s stagflation. Higher rates failed in the bad cases; the 4% figure was set by the unluckiest starting points, not the typical ones. Bengen also found that portfolios that held too few stocks fared worse over long horizons, because they lacked the growth needed to outrun inflation.

The Trinity study: success rates across the grid

In 1998, three finance professors at Trinity University — Cooley, Hubbard, and Walz — extended the same historical approach into a grid. Rather than reporting a single worst-case rate, they computed the historical success rate of many combinations of withdrawal rate, stock/bond allocation, and retirement length. Their broad conclusion matched Bengen's: an inflation-adjusted 4% withdrawal rate, paired with a portfolio holding roughly 50-75% stocks, survived nearly all historical 30-year periods. The Trinity study is where the now-familiar framing of "success rates" comes from, and its tables are the ancestors of tools like the ones on this site.

What the rule actually assumes

Both studies embed assumptions that are easy to forget once the headline number detaches from the research:

• A 30-year horizon. The classic result is about 30-year retirements. Longer horizons — early retirees, for instance — face tougher arithmetic.
• Mechanical inflation-adjusted spending. The simulated retiree raises withdrawals with inflation every year and never cuts back, no matter what markets do.
• No fees, no taxes. The studies used index-level returns. Real investors pay fund expenses, possibly advisory fees, and taxes — all of which reduce what is actually sustainable.
• U.S. market history. The evidence is drawn entirely from American stocks and bonds during the American century.

The standard criticisms — taken seriously

The rule has attracted decades of criticism, much of it fair:

• The U.S. had an exceptionally good century. Researchers who ran the same backtests on other countries' market histories — countries that experienced wars fought on their soil, hyperinflations, and market closures — found that historically safe withdrawal rates were often meaningfully lower. American history may be a lucky draw, not the base case.
• Nobody actually behaves mechanically. A real retiree watching a portfolio collapse would cut spending, and a retiree in a boom would loosen up. The rigid rule is a research convention, not a description of human behavior — which is why flexible strategies exist.
• Starting valuations may matter. Some researchers argue that when stocks are historically expensive at the retirement date, prospective safe withdrawal rates are lower than the historical record suggests, and that the starting rate ought to reflect that. Others note that valuation-based forecasts have a mixed record of their own. The debate is unresolved.
• Thirty years is not everyone's horizon. A long-lived couple, or someone retiring early, needs the money to last longer than the classic studies tested.

Bengen's own second thoughts — upward

Interestingly, Bengen himself came to view the original figure as conservative. His later work added more asset classes beyond the original two — slices of the market such as smaller-company stocks — and found that broader diversification supported somewhat higher worst-case withdrawal rates. He has said in interviews that he considers the popular fixation on "4%" a simplification of what was always a moving research frontier. The direction of his revision is worth noting: the loudest criticisms push the number down, while the rule's own author pushed it modestly up.

What our data shows

This site re-runs the classic experiment against 1,128 months of real market data covering 1928-2021 — actual stock returns, actual bond yields, actual inflation, month by month. The findings are consistent with the original research: inflation-adjusted 4% withdrawals at a 60/40 mix succeeded in 96% of the 30-year historical periods since 1928 (100% for fixed-dollar withdrawals), and the worst-case sustainable inflation-adjusted rate was 3.83%, set by the December 1965 retiree — the person who walked straight into the stagflation of the 1970s. You can explore the worst-case rate for any allocation and horizon on the break-even analysis page, and the full data and method are documented on the methodology page.

What the rule does not claim

The 4% rule is not a guarantee — it summarizes what survived one country's past, and the future is under no obligation to stay inside history's range. It is not a plan — it says nothing about taxes, fees, Social Security, spending that changes with age, or the flexibility every real household has. What it is, and what it has always been, is a research benchmark: a disciplined, evidence-based answer to "what withdrawal rate survived the worst the past ever delivered?" Used that way — as a yardstick to compare strategies against, not a promise to lean on — it remains one of the most useful ideas in retirement planning.

Official sources

• Investor.gov — the SEC's plain-language investor education site
• ConsumerFinance.gov — the CFPB's retirement planning resources

This article is educational only and is not financial, investment, tax, or legal advice. Rules and limits change; verify details with official sources or a qualified professional.