The 4% rule in 2026: what the research really says.

Few numbers in personal finance are as famous — or as misunderstood — as the 4% rule. Withdraw 4% of your portfolio in your first year of retirement, adjust for inflation each year after, and your money should last 30 years.

Thirty-two years after financial planner William Bengen published the research behind it, the rule is still the starting point for nearly every retirement spending conversation. It's also under more scrutiny than ever. Here's what the research actually says, what it doesn't, and how to find a number that fits your plan.

Where did the 4% rule come from?

In October 1994, financial planner William Bengen published "Determining Withdrawal Rates Using Historical Data" in the Journal of Financial Planning — the paper the entire rule traces back to. His method was straightforward: take a portfolio of 50% large-cap U.S. stocks and 50% intermediate-term Treasury bonds, assume a retiree withdraws a fixed percentage in year one and adjusts that dollar amount for inflation each year after, and test every 30-year retirement starting in each year from 1926 to 1976 — 51 hypothetical retirees in all, one per starting year.

His finding: a 4% starting withdrawal never exhausted the portfolio within 30 years. The worst case — the retiree who needed the lowest rate to survive — started in October 1968, at a market peak just before a 14-year bear market and a run of double-digit inflation. Bengen called that worst-case rate SAFEMAX: 4.15%.

Four years later, the "Trinity Study" by three finance professors confirmed the result with a different bond index: for portfolios holding at least 50% stocks, 4% inflation-adjusted withdrawals succeeded in 95% or more of 30-year periods. The rule of thumb was born.

Three things about the original research are worth remembering:

• It assumed no fees or taxes. Real portfolios pay both.
• It assumed a 30-year horizon — roughly age 65 to 95.
• It was a worst-case finding, not a recommendation. Bengen was answering "what survived the worst period on record," not "what should everyone spend."

The 4% rule is a worst-case number, not an average

This is the most misunderstood part of the whole story. Researcher Michael Kitces later extended Bengen's work across more starting years and found that the average safe withdrawal rate across all historical 30-year periods was about 6.5%. The 4% figure only binds in the handful of worst sequences — 1965 through 1969, with 1966 the worst of all.

Kitces puts it this way: the 4% rule is really a statement that there's a roughly 96% chance you'll finish 30 years with all of your starting principal still intact. Not just solvent — untouched. In the median historical scenario, a 4% retiree died with nearly three times their starting balance in real terms.

So the rule has two failure modes, and only one gets attention. Yes, spending too much risks running out. But spending 4% when history would have supported 6.5% risks something quieter: working years you didn't need to work, trips you didn't take, gifts you didn't make. Both are real costs.

Bengen's own updates

Bengen never treated 4% as carved in stone. In research published in his 2006 book, he added small-cap stocks to the mix and raised his worst-case number to 4.5%. Later, adding micro-caps, mid-caps, international stocks, and T-bills, he arrived at 4.7% — which he still calls the worst-case number, suitable only for the most conservative planners.

He's also been clear about what actually breaks a plan. In interviews, Bengen has said two things drive the safe rate down: a major bear market early in retirement, and high inflation during retirement. Of the two, he considers inflation the retiree's worst enemy. Both the 2000 retiree (dot-com crash) and the 2007 retiree (financial crisis) — each of whom suffered brutal early bear markets — appear to be tracking fine at 4.5%, he has said. But a decade or more of sustained high inflation, he warns, could change the math.

His practical advice has also evolved. Bengen now recommends against set-it-and-forget-it withdrawals: review your spending every couple of years against a template of where your plan should be. If markets recover after an early bear market — as they usually do — riding it out beats panic-cutting. If inflation spikes, cut spending early and preserve capital.

For longer retirements, he calculated lower worst-case rates: about 4.3% for 35 years, 4.2% for 40 years, and 4.1% for 45 years. Early retirees planning for 50-year horizons are outside the territory the original research covered.

What does Morningstar recommend for 2026?

Morningstar publishes an annual "State of Retirement Income" report that re-estimates a safe starting withdrawal rate using forward-looking return assumptions rather than pure history. For 2026, their base case is 3.9% — up from 3.7% in 2025 — for a 30-year retirement with a 90% probability of not outliving the money, assuming 20% to 50% in stocks with the rest in bonds and cash.

The year-to-year moves are instructive: 3.3% in 2021, 3.8% in 2022, 4.0% in 2023, 3.7% in 2024, 3.7% in 2025, 3.9% in 2026. The number moves because expected returns move. Morningstar explicitly says retirees should not adjust their own withdrawals to each year's new number — you use the rate from your retirement year and stick with it.

The more interesting finding is what happens with flexibility. Morningstar tested dynamic spending strategies and found they could lift the starting rate substantially — up to 5.7% with the most flexible approaches. A guardrails-style strategy supported a 5.2% starting rate on a 40% stock / 60% bond portfolio. The price of the higher number is variability: your income moves with the market, and cuts land during crashes.

Why 2026 makes people nervous — and what to make of it

Three things fuel today's 4%-skepticism: valuations, yields, and longevity.

Valuations are extreme. As of September 2026, the Shiller CAPE ratio — the S&P 500's price divided by ten years of inflation-adjusted earnings — sat above 40. That's the second-highest reading in over 150 years of data, behind only the dot-com peak of 44.2 in late 1999, and more than double the long-run average near 17. When CAPE has exceeded 30 in past cycles, subsequent 10-year annualized returns averaged below 4%. High valuations don't predict a crash — CAPE is a long-horizon signal, not a timing tool, and it has been elevated for much of the past decade — but they do suggest lower expected returns ahead.

Bond yields have normalized from historic lows, which actually helps retirees: bonds once again provide meaningful income, one reason Morningstar's 2026 number ticked up. The 2021 low of 3.3% was largely a story of near-zero yields — and the Fed's September 2026 hike pushed that normalization further still, for cash and new bonds at least.

Retirements are longer. A 65-year-old couple today has a meaningful chance that at least one spouse reaches 90. The 30-year assumption that underpins the 4% rule is fine for a 65-year-old planning to 95 — but thin for someone retiring at 55, and overly cautious for someone retiring at 72.

The honest synthesis: 4% was built on the worst 30-year sequence in U.S. history. Today's starting conditions look challenging on valuations, which argues for caution — but the rule already assumes you retire into something like 1966. For the rule to fail going forward, the next 30 years would have to be worse than anything in the historical record. That's possible. It's just a stronger claim than "valuations are high" gets you to on its own.

The rule's blind spots

Even taken on its own terms, the 4% rule leaves a lot out:

• Fees. A 1% annual advisory fee is a 25% surcharge on a 4% withdrawal. The research assumed zero costs.
• Taxes. Withdrawals from a traditional IRA are taxed as ordinary income; Roth withdrawals aren't. Two retirees with the same portfolio and the same 4% withdrawal can have very different after-tax spending power depending on account mix and sequencing — see our free Roth conversion calculator for whether shifting some of that mix pays off.
• The 30-year box. Retire at 60 and you may need 35+ years. Retire at 70 and 4% is likely too conservative.
• No response mechanism. The rule says keep withdrawing the inflation-adjusted amount no matter what. No real retiree does this — and the research on dynamic strategies suggests you shouldn't.
• Required minimum distributions. Under SECURE 2.0, traditional IRA and 401(k) owners born 1951–1959 must begin RMDs at 73, and those born 1960 or later at 75. Late in retirement, RMDs can easily exceed 4% of the portfolio — forced withdrawals the rule never contemplated.

Dynamic alternatives that beat the rule

The most credible challengers to the 4% rule aren't lower numbers — they're smarter systems.

Guardrails (Guyton-Klinger, 2006). Pick a starting rate — their research supported 5.2% to 5.6% with 65%+ in equities. Each year, compare your current withdrawal rate to the starting rate. If it's drifted 20% above (markets fell), cut spending 10%. If it's drifted 20% below (markets rose), raise spending 10%. Otherwise, change nothing. You start higher and adjust as reality unfolds. The honest caveat, raised by researchers including Derek Tharp and Justin Fitzpatrick in 2024, is that the cuts can be steep and poorly timed for retirees without spending flexibility — the strategy demands slack in the budget and works best with Social Security as a floor.

Ratcheting (Kitces). Start at a safe baseline like 4%, but give yourself a 10% raise (above inflation) whenever the portfolio grows 50% above its starting value, limited to once every three years. The 1966 retiree never gets a raise; the 1982 retiree gets several. It converts the "die with triple your money" problem into lifetime spending.

Both approaches share a philosophy: decide your response rules in advance, then let the plan breathe. That's strictly more informative than a single fixed number.

Worked examples: what the rates mean in dollars

On a $1,000,000 portfolio at retirement:

The gap between 3.9% and 5.2% is $13,000 a year — $390,000 over 30 years before inflation adjustments. That's the price of rigidity. A retiree with flexible spending, a cash buffer, and guaranteed income covering essentials can reasonably plan around a higher starting number than a retiree whose entire budget is fixed costs.

And remember the other direction: at 4%, history says you'll probably die with far more than you started with. If your plan shows a 98% success rate and a median ending balance of $2.5 million, the question isn't whether your money will last — it's whether you're spending enough.

How do you find your personal number?

Rules of thumb are starting points, not answers. Your safe rate depends on your horizon, your asset mix, your account types and their tax treatment, your Social Security timing, your spending flexibility, and how you feel about the tradeoff between "never run out" and "don't underspend your life."

That's what Runway's max sustainable spending analysis is for. Instead of handing you 4%, it works backward from your actual plan — your accounts, your Social Security strategy, your Roth conversion schedule, your time horizon — and finds the highest spending level your plan supports across 2,000 Monte Carlo simulations. Then the Historical Backtest checks that number against 125 real starting years, and the full Social Security claiming-age grid shows you how delaying benefits changes the answer.

The free tier includes the 200-simulation Monte Carlo, the 30-year projection table, and withdrawal sequencing, so you can see the range of outcomes for your own numbers before deciding what "safe" means for you.