Bengen’s original 1994 rule was simple enough to fit on a cocktail napkin, which is part of why it stuck. Start retirement by spending 4% of savings in year one, then raise that dollar amount each year with inflation. The research behind it suggested that a retiree following that path could avoid running out of money over a 30-year retirement, based on U.S. stock and bond data going back to 1926, USA Today reported this week.

The published 4% number was not the exact output. Bengen’s math came out to 4.15%, then got rounded down, a conservative move that helped the rule escape academic papers and become retirement folklore. That shorthand is now so familiar that many people treat 4% as if it were a law of nature. It never was.

Bengen’s updated research changes the setup. He now works with seven asset classes instead of two, including stocks of large, mid and small companies, international stocks, bonds and Treasury bills, USA Today reported. His current calculations assume a portfolio that is 55% stocks, 40% bonds and 5% cash. He says the shift reflects one thing: his research has gotten more sophisticated.

That broader mix supports a higher safe starting rate. In Bengen’s newest work, the 4.7% figure is the SAFEMAX, the worst-case starting withdrawal rate for a 30-year retirement under his assumptions, CNBC reported in December. For a 50-year retirement, the figure falls to about 4.2%, which tells you something the slogan never could: time horizon still rules the game.

There is one point of tension in the reporting worth noting. USA Today describes the updated mix as 55% stocks, 40% bonds and 5% cash, while CNBC reports 55% stocks, 45% bonds and 5% cash in the T-bill form. The core point survives either way, but the exact allocation should be pinned to Bengen’s original text before publication.

The split between 4.7% and 3.7% looks dramatic until you look at the machinery behind it. Bengen’s method is historical. It asks what withdrawal rate would have survived the ugliest 30-year stretches in U.S. market history. Morningstar’s method is forward-looking. It starts with current market conditions, then models likely future outcomes from there.

That difference alone changes the answer. Morningstar’s most recent base case for a new retiree is 3.7%, down from 4.0% the year before, according to NAPA-Net in December 2024. Morningstar also said in March 2025 that its estimates have varied every year since it began publishing the research in late 2021, moving from 3.3% in 2021 to 3.8% in 2022, 4.0% in 2023 and 3.7% in 2025.

Those shifts are not a sign that the firm keeps changing its mind for sport. They reflect changing bond yields, changing equity valuations and changing assumptions about what the next 30 years may look like. Morningstar also uses a 90% probability of success, while Bengen’s historical work is essentially asking for full survival across the worst cases. Same broad problem. Different yardsticks.

The market backdrop matters too. Investopedia noted in January that today’s high stock valuations and lower bond yields point to more modest returns than the past delivered. That is precisely the sort of environment a forward-looking model is designed to capture, even if it produces a lower starting percentage than the old rule.

Sequencing risk is the other reason the numbers are so sensitive. Bengen has said a substantial bear market early in retirement can pull down the safe withdrawal rate because losses land just as withdrawals are leaving the portfolio, CNBC reported. A bad first few years do more damage than a bad decade later on. Retirement, annoyingly, is not only about averages.

The attraction of a rule like this is obvious. It turns a foggy problem into a clean number. Rob Williams of Charles Schwab put it neatly in USA Today: the rule survives because it is memorable and makes a complicated human problem feel manageable.

The trouble starts when the number gets treated like a contract. Douglas Ornstein of TIAA Wealth Management told USA Today that retirement spending is dynamic, not static. People often spend more early on travel and activity, then less in the middle years, then more again later as healthcare climbs. A flat inflation adjustment does not capture that shape.

Healthcare is the hardest part to ignore. Investopedia reported in January that Fidelity estimates a 65-year-old who retired in 2025 will spend $172,500 on healthcare over retirement, and that the cost is rising faster than general inflation. That is not a neat little line on a spreadsheet. It is a wobble that can show up years after retirement starts, when flexibility is already more valuable than certainty.

That is why more flexible withdrawal methods keep showing up in the literature. A guardrails strategy sets upper and lower limits, then adjusts spending when the portfolio crosses a threshold. A bucket strategy separates near-term cash, medium-term bonds and long-term equities so market swings do not force bad decisions at the wrong time. Some retirees simply start lower, around 3% to 3.5%, then adjust with Social Security, pensions or part-time income as life unfolds, Investopedia reported.

Morningstar’s own framing points the same way. In Morningstar’s March note, the firm said it is not suggesting retirees should change their withdrawal amount every year to mirror the latest estimate. The idea is to set a reasonable starting point, then revisit the plan as conditions change. That is less dramatic than a rule, but far more useful.

The debate over 3.7% versus 4.7% assumes there is enough money for the difference to matter. For a lot of households, that is already a stretch. USA Today reported that the typical American household ages 55 to 65 has about $185,000 in retirement savings, based on the 2022 Survey of Consumer Finances. At 4%, that works out to about $7,400 a year.

That is not a retirement plan. It is a supplement.

Amy Arnott, a portfolio strategist at Morningstar, told USA Today that many families have no retirement savings at all. For them, the important questions are earlier in the chain: when to claim Social Security, whether part-time work is possible, how much housing equity can be tapped, and whether other benefits can be optimized. Withdrawal rates matter, but they are hardly the first problem on the list.

That said, the debate is not a luxury item for everyone. On a $1 million portfolio, the difference between 3.7% and 4.7% is $10,000 a year. That is real money, and it can change the shape of retirement. It just does not solve the deeper problem of whether the portfolio is large enough to begin with.

Bengen’s revised 4.7% figure should be read as a floor for a specific portfolio design, not a universal instruction. It comes from a broader asset mix and a more favorable historical result than the original rule captured. Move the portfolio, fees or tax treatment, and you are no longer in the same experiment.

Morningstar’s work points to a different lesson: time horizon matters as much as starting percentage. Morningstar says a retiree with a 20-year horizon can reasonably spend more than 5% of a balanced portfolio, while a 40-year horizon pushes the starting rate down to 3.1%. Retirement age is only half the story. Longevity is the other half, and it refuses to stay in the background.

The useful way to read all this is not as a duel between 4% and 4.7%, but as a reminder to treat any rule of thumb as a first draft. Start with a percentage if it helps frame the problem. Then look at your time horizon, your spending pattern, your market exposure and your willingness to adjust. Investopedia put it plainly in January: regularly reviewing the plan and adjusting withdrawals over time may matter more than picking the perfect percentage at the start.

Bengen is still refining the same question after three decades, which is a decent sign that retirement planning is alive, not settled. The 4.7% rule for retirement is a useful update, and probably a better floor than the old shorthand gave him credit for. It just does not get the last word.