The 4% Rule Is Dead. Here's What Replaced It

Where the 4% rule came from

In 1994, financial planner William Bengen published a study called "Determining Withdrawal Rates Using Historical Data" in the Journal of Financial Planning. He ran 50-year rolling historical periods of U.S. stock + bond returns and asked: what's the highest withdrawal rate from a 50/50 portfolio that never depleted in 30 years across any historical starting point?

The answer was 4.15%. Round it down to 4% for safety. That became the foundation of nearly every retirement income discussion for the next 30 years.

Bengen's analysis was honest, rigorous, and useful for its time. It was also bounded by assumptions that have all weakened:

• U.S.-only data from the most-successful equity market in history. Other countries had decades where 4% was nowhere close to safe.
• 50/50 allocation. Many retirees today hold more or less than this.
• 30-year retirement. A 65-year-old in 1994 had average remaining life expectancy of ~17 years. A 65-year-old today can expect about 19.5 more years on average (CDC, 2023 data), and a real share will reach 95 or older.
• Higher starting bond yields. In 1994, the 10-year Treasury yielded 7%+. The fixed-income half of the portfolio carried much more weight. Today's lower yields shift that math.
• It only passed the U.S. past. Bengen tested history, not the future. Newer studies run thousands of simulated markets and usually accept a 90% to 95% success rate, which still means 1 in 10 to 1 in 20 retirees runs short.

What the updated research says

Multiple major studies have re-examined Bengen's work over the past decade. The findings have converged:

• Wade Pfau (American College). Using international data and 2010s yield curves, the safe withdrawal rate dropped to ~2.8-3.0% for a 30-year horizon.
• Morningstar (annual studies). Its base case (30 years, 90% chance of success) was 4.0% in the 2023 study and fell to 3.7% in the December 2024 study, on higher stock valuations and slightly lower bond yields (Morningstar, December 2024).
• Vanguard. Models suggest 3.0-3.5% for moderate allocations and 30-year retirements.
• Schwab. Their Center for Financial Research recommends a "spending budget" range of 3.5-4.0% with explicit adjustment downward in market drawdowns.

The center of professional consensus is now roughly 3.0-3.5% as the new "safe baseline" for a 30-year retirement with a moderate stock/bond allocation. A 25% reduction from the old rule.

What this means in dollars: a $1,000,000 portfolio that used to "safely" support $40,000/year of inflation-adjusted spending now safely supports $30,000-$35,000/year. To replicate the old $40,000 income level safely, you need closer to $1,150,000-$1,300,000.

Why the math actually got worse

Three converging trends pushed the safe number down:

1. Lower bond yields

Bengen's 1994 analysis assumed 10-year Treasury yields would average something resembling history. Yields cratered in the 2010s. While they've recovered modestly to ~4% in 2026, the fixed-income half of a 50/50 portfolio is doing meaningfully less heavy lifting than it did in Bengen's analysis. The portfolio's real-return engine is more concentrated in equities, which carry more sequence risk.

2. Longer expected lifespans

The U.S. life expectancy at 65 has risen by 2-3 years since 1994. More importantly, the tail probabilities have shifted: the probability of a 65-year-old living to 95 has roughly doubled. A retirement plan that survives 30 years isn't enough anymore. It needs to survive 35+ years for the longest-lived spouse in a couple.

3. Higher expected equity valuations

U.S. equity Shiller P/E ratios have been historically elevated for the past decade. Most academic research suggests forward returns from elevated starting valuations are lower than long-term historical averages. Lower forward equity returns + lower bond yields = a smaller real return engine to draw from.

None of these are temporary. They reflect structural shifts in demographics, monetary policy, and market valuation. The 4% rule worked for Bengen's specific historical window. It is unlikely to be the right number going forward.

The income-floor alternative

The most sophisticated modern retirement income research has moved away from "what % can I withdraw" toward "how do I cover my essential spending with guaranteed income."

The framework: identify your essential spending floor (housing, food, utilities, healthcare, insurance, minimum lifestyle). Cover that floor entirely with guaranteed lifetime income sources:

• Social Security (especially if delayed to 70)
• Pensions, if any
• A Treasury or TIPS ladder sized to cover any remaining gap for a set number of years

Once the floor is covered, the remaining portfolio is "upside money." You can invest it aggressively (70-100% equities) because you don't need it for survival. You can withdraw discretionary amounts in good years, less in bad years. The 4% / 3% question becomes moot because you're no longer relying on the portfolio for life support.

This framework has the strongest mathematical backing in the post-2010 research. By moving essential spending off the portfolio, retirees achieve:

• Lower sequence-of-returns risk
• Higher long-term expected returns on the discretionary portfolio
• Larger estates passed to heirs in 70-80% of simulated scenarios
• Less behavioral risk (people don't panic-sell during downturns when their essentials are covered)

What this means for your portfolio sizing

If you're still working and planning your retirement target, the old "save 25x your annual spending" rule (the inverse of 4%) has become "save 28-33x your annual spending" (the inverse of 3.0-3.5%).

Worked example: a couple targeting $80,000/year of inflation-adjusted retirement income (excluding Social Security):

• Old 4% rule: Need $2,000,000 portfolio.
• New 3.3% baseline: Need $2,420,000 portfolio.
• Income-floor approach: Delaying the higher earner's Social Security to 70 can add $15K-$25K/yr of lifetime, inflation-adjusted income, which shrinks the portfolio's job and the nest egg it needs.

The income-floor approach often gets you to a target with a smaller total nest egg and dramatically lower sequence risk. The savings spent bridging to a later Social Security claim aren't "lost." They buy a larger lifetime check, which is exactly what essential-spending coverage requires.

What to actually do

Three actions, depending on where you are:

If you're 10+ years from retirement

Update your retirement-savings target. If you've been targeting "25x spending," shift to "28-33x" OR start thinking about how an income-floor will be funded. Both are reasonable. The savings rate implications are meaningful, typically 2-4% more of income per year, but compoundable.

If you're 5-10 years from retirement

This is the right window to start thinking concretely about income floor structure. What does your essential spending look like? What will Social Security cover at your planned claiming age? What's the gap? Your Social Security claiming age and the savings you set aside to bridge to it are the two biggest levers.

If you're already retired

Test your current withdrawal rate against the 3.0-3.5% baseline. If you're at 4%+ and your portfolio is mostly equities, you're carrying meaningful sequence risk, especially in a year when markets have been strong (because valuations are elevated). Consider either reducing the withdrawal rate, building an income floor by delaying Social Security, or both.

The workshop linked below walks through income-floor sizing for different essential-spending targets.

Hans Goldstein, EA

Founder & Executive Director · Retirement Literacy Foundation, a 501(c)(3) non-profit
Admitted to practice before the IRS

Want more information on this? Put your name below and I’ll send it over in plain English, at no cost. If your situation has a wrinkle, reply to the email and tell me what it is; I read them.

You’re reading: The 4% Rule Is Dead. Here's What Replaced It

First name Email Send me more information

Free education. The Foundation sells nothing. By submitting you agree to receive free educational emails. We never sell or share your information and you can unsubscribe from any email.