The 4% Rule Was Built for 1994, and Retirements Now Stretch Past 30 Years

Ask ten people how much they need to retire and at least six of them will say the same thing: save 25 times your annual spending, withdraw 4% a year, and you're set for life. It's a tidy rule, it fits on a napkin, and it's been repeated so often that it's hardened into something people treat as settled math. It isn't settled, and the researcher who came up with it has spent thirty years revising his own answer.

Bill Bengen published the original 4% rule in 1994, based on historical U.S. market data, back when 10-year Treasury bonds paid close to 8% and a balanced portfolio could lean on bonds to do real work. The 10-year yields about 4.7% today. That single shift, bonds doing roughly half the job they used to, is a large part of why the number everyone still quotes has quietly stopped being one number.

Bengen's original research asked a specific question: across every rolling 30-year period in U.S. market history, what's the highest withdrawal rate that would have let a portfolio survive without running out of money? His answer was 4%, and it became one of the most repeated numbers in personal finance because it was simple, it was backed by real data, and for a long time it held up.

The Trinity Study followed in 1998 and largely confirmed it, at least for portfolios holding 50% or more in stocks over a 30-year horizon. For two decades, 4% was close enough to gospel that most financial planning conversations stopped there.

Two things broke the clean version of this rule, and neither is a mystery.

Bond yields fell. When Bengen ran his original numbers, a retiree's bond allocation was pulling in a strong income stream, often 6% to 8% a year. Today's 10-year Treasury sits near 4.7%, which means the portfolio has to lean harder on stocks to hit the same target, and leaning harder on stocks brings more volatility right when a retiree can least afford it.

People are also living longer. Bengen modeled a 30-year retirement. A healthy 65-year-old today has a real chance of needing that portfolio for 35 years or more, and Morningstar's research shows what the extra stretch costs: extending the drawdown period from 30 to 35 years drops the safe starting withdrawal rate from 3.9% to 3.5% (a 40-year horizon drops it to 3.2%). Five extra years of retirement costs almost half a percentage point of spending power, every single year, for the rest of your life.

Morningstar publishes an updated safe withdrawal rate every year, using forward-looking return forecasts instead of Bengen's historical averages, and the number has moved as market conditions shifted: 3.3% in the 2021 study, 3.8% in 2022, 4.0% in 2023, 3.7% in 2024, and 3.9% in the latest edition, the base case for someone retiring in 2026 with a 30-year horizon, 30% to 50% in stocks, and a 90% probability the money lasts.

Bengen himself has moved too, in the other direction. In his 2025 book he raised his worst-case estimate, what he calls the SAFEMAX, to 4.7% for a more diversified portfolio, and wrote that a SAFEMAX of 5.25% to 5.5% "seems like a reasonable, conservative estimate for current retirees." The gap between Morningstar's 3.9% and Bengen's 5.5% isn't a contradiction so much as a reminder that this was always a range dressed up as a single number.

The scarier problem isn't the average return over 30 years. It's the order the returns show up in. A portfolio that earns a strong average over three decades can still fail if the first few years happen to be bad ones, because withdrawals during a downturn lock in losses the portfolio never gets a chance to recover from. Morningstar's simulations put a shape on it: nearly 70% of the retirement plans that failed involved portfolios that took losses in the first five years of retirement.

This is the piece a napkin calculation can't capture. Two retirees with identical portfolios and identical 30-year average returns can end up in completely different places, purely because of when the bad years landed.

Here's where the withdrawal-rate debate turns into a real dollar figure. Say you need $80,000 a year from the portfolio, on top of Social Security. The nest egg required swings dramatically with the rate you plan around.

That's a swing of more than $1 million on the exact same $80,000 spending target, depending entirely on which rate you plan around and whether your spending can flex with the market. The rate you pick isn't a rounding error. It's the single biggest lever in the entire retirement number.

The retirees who come out ahead in Morningstar's research aren't the ones who found a smarter fixed number. They're the ones using flexible spending, cutting back a little in bad years and spending more freely in good ones, instead of locking in one rate on day one and never adjusting. That flexibility, combined with moves like a guardrails approach, is what pushes the supportable starting rate from 3.9% toward 5.7% in Morningstar's modeling. A rigid rule can't adapt to a bad sequence of returns. A retiree who's willing to adjust can. I've written about how a guardrail spending plan works in practice.

Delaying Social Security does similar work from a different angle. Every year you wait adds a larger, inflation-protected, guaranteed income floor underneath the portfolio, which takes real pressure off the withdrawal rate you need from your investments.

Start with your real spending, not a guess, and separate the fixed costs from the ones you could trim in a rough year. Then pick a starting withdrawal rate that matches your actual horizon: a 3.9% base case if you're planning for 30 years and want to mostly set it and forget it, something more conservative if you're retiring early or expect to live past 95, and room to flex upward in good years if you're comfortable adjusting when markets cooperate. Stress-testing your retirement date against a bad early sequence is the single most useful version of this exercise.

The 4% rule isn't wrong so much as incomplete. It was a genuinely useful answer to a question asked in 1994, under bond yields and life expectancies that no longer describe most retirements. Treat it as a starting point worth stress-testing against your own timeline, not a number worth building your entire retirement around.