Under the 4 percent rule, bonds barely change how often a retirement portfolio runs out of money; what they change is how badly it fails. In 1,509 U.S. retirements starting between 1871 and 1996, a 4% inflation-adjusted withdrawal ran dry within 30 years in 2.1% of cases with 100% stocks and 2.6% with 60/40. But the worst all-stock retirement, which began in October 1929, was empty after 16.5 years, while the worst 60/40 retirement lasted 25.7 years. Holding 20% to 50% in 10-year Treasuries lifted the worst-case safe withdrawal rate from 3.07% to between 3.70% and 3.80%; putting more than half the portfolio in bonds made both typical and worst outcomes worse.

Thirty-two years after it was born, the 4% rule is still the most quoted number in retirement planning. Financial planner William Bengen introduced it in the October 1994 issue of the Journal of Financial Planning, after testing U.S. retirements from 1926 onward with portfolios of large-company stocks and intermediate-term Treasuries. He advised keeping 50% to 75% in stocks. The 1998 Trinity study by Philip Cooley, Carl Hubbard and Daniel Walz used high-grade corporate bonds instead and found that a 50/50 portfolio withdrawing 4% a year survived 95% of the 30-year periods from 1926 to 1995.

The debate has never settled. In his 2025 book A Richer Retirement, Bengen raised his own figure to 4.7%, using a broader mix of 55% stocks, 40% bonds and 5% cash. Morningstar went the other way: its December 2025 State of Retirement Income report put the safe starting rate for 2026 retirees at 3.9%, with the best results from portfolios holding only 30% to 50% in stocks, partly because bonds now yield more. Both camps hold bonds. So we isolated one question: with the same retirement and the same withdrawals, what does each extra slice of bonds actually do?

What the 4 Percent Rule Says, and Where Bonds Come In

The rule itself is simple. In your first year of retirement, withdraw 4% of your portfolio. Every year after that, withdraw the same amount adjusted for inflation, whatever the markets do. On a $1 million portfolio, that means $40,000 in year one, rising with the cost of living.

For any retirement start date, the highest starting rate that would have lasted the full 30 years is that date’s safe withdrawal rate. The 4% rule is, in effect, a bet that your retirement won’t be worse than the worst one in the historical record.

Bonds come in because of sequence-of-returns risk: a crash early in retirement forces you to sell stocks at low prices to pay the bills, and the portfolio may never recover. Bonds are meant to be the part you sell instead while stocks recover.

How We Tested the 4% Rule With Different Stock-Bond Mixes

• Data: U.S. stocks (the S&P 500 and its predecessors, dividends reinvested), 10-year U.S. Treasury yields and U.S. consumer prices, monthly from 1871 to September 2026.
• Retirements: 1,509 monthly start dates from February 1871 to October 1996, each lasting 30 years. Later start dates have not completed 30 years yet.
• Portfolios: 11 mixes from 0% to 100% in stocks, in steps of 10 points, with the rest in a constant-maturity 10-year Treasury whose return we calculate from its yield (coupon income plus price change). Every portfolio is rebalanced monthly.
• Spending: a fixed share of the starting balance, from 3% to 5% a year, withdrawn at the start of each month and raised with inflation every month.
• Failure: the money runs out before 30 years. No fees, taxes, pensions or Social Security.
• Robustness checks: withdrawing and rebalancing once a year, as Bengen did, gave almost the same results: a worst-case safe rate of 3.73% for 60/40, against 3.76% in the monthly version. We also swapped in T-bills, 5-year and long-term Treasuries and Aaa-rated corporate bonds over shorter periods (see below).

Two caveats. Before 1989 the stock data are monthly averages, and bond yields before 1953 are smoothed annual figures, which makes the early record look a little calmer than it was. And the data are American because they are the longest continuous record; the logic carries over to other markets, but the exact numbers will not.

4 Percent Rule Backtest: Failure Rates for Every Stock-Bond Mix

1,509 U.S. retirements starting February 1871 to October 1996, 30 years each. Bonds = 10-year Treasuries; portfolios rebalanced monthly; withdrawals raised with inflation; before fees and taxes. Worst safe rate = highest starting withdrawal that survived every start date. Left per $100 = median balance after 30 years at 4%, after inflation.

Highest starting withdrawal rate, raised with inflation, that lasted 30 years: worst and median of 1,509 U.S. retirements starting February 1871 to October 1996, by share of stocks (the rest in 10-year Treasuries), rebalanced monthly, before fees and taxes. Source: S&P 500 (S&P Dow Jones Indices, Robert Shiller); U.S. Treasury/Federal Reserve (FRED); U.S. Bureau of Labor Statistics; Bondfish analysis.

Three things stand out.

1. Between 50% and 100% stocks, bonds barely changed how often 4% failed. Every mix in that range ran out in between 1.6% and 3.4% of retirements. The 4% rule’s safety record comes mainly from stocks’ long-run growth, not from the bonds alongside them.
2. Bonds’ real job is the worst case. Moving from 100% stocks to 70/30 lifted the worst-case safe withdrawal rate from 3.07% to 3.80%, within a whisker of the 4% rule itself. Almost all of that gain came from one disaster: the 1929 crash.
3. Too many bonds hurt more than too few. Below 40% stocks, failures multiplied: 10.9% of retirements ran out at 30/70, 24.9% at 20/80 and 48.4% with all bonds. Inflation slowly ate a bond-only portfolio.

The price of bonds is what you leave behind. The typical (median) all-stock retiree withdrawing 4% ended 30 years with $318 for every $100 they started with, after inflation. A 60/40 retiree ended with $133, and an all-bond retiree with just $4.

Bonds Don’t Stop the 4% Rule From Failing; They Delay the Failure

Failure rates hide how badly a plan fails. Running out of money in year 29 and running out in year 16 are very different retirements.

• 100% stocks: 32 of the 1,509 retirements ran out at 4%. The earliest, starting in October 1929, was empty after 16.5 years.
• 60/40: 39 retirements ran out, slightly more, but the earliest lasted 25.7 years and the typical failure came in year 28.
• At a 4.5% withdrawal rate the gap widens: the earliest all-stock failure came after 13.8 years, against 20.5 for 60/40.

For someone retiring at 65, that is the difference between running short at about 81 and at about 90.

Where Bonds Earned Their Keep: Deflation in 1929, Not Inflation in 1966

Bonds helped in far fewer retirements than most investors assume. A 60/40 portfolio supported a higher safe withdrawal rate than all stocks in only 117 of the 1,509 start dates (7.8%); in 1,330 of them, all stocks would have been the best mix. But the few start dates where bonds won are the ones that define the 4% rule.

Lowest safe withdrawal rate among the retirements starting in each calendar year, 30 years, withdrawals raised with inflation. Bonds = 10-year Treasuries. Source: S&P 500 (S&P Dow Jones Indices, Robert Shiller); U.S. Treasury/Federal Reserve (FRED); U.S. Bureau of Labor Statistics; Bondfish analysis.

• 1929–1932, deflation: U.S. stocks lost 82% of their value, including dividends, from September 1929 to June 1932, while consumer prices fell 27% by March 1933. Ten-year Treasuries returned 12.5% a year after inflation from October 1929 to June 1932. An all-stock retiree from October 1929 could have safely withdrawn only 3.07%; with 60/40 the figure was 4.29%, and with 40/60, 4.71%.
• 1941–1951, inflation with low yields: the 10-year yield was 1.95% in January 1941, and consumer prices rose 88% by the end of 1951. Bonds lost 3.85% a year after inflation. An all-bond retiree from February 1941 could safely withdraw only 2.40%, the worst case in our data, and at 4% ran out after 18.2 years. A 60/40 retiree from the same month could withdraw 5.75%.
• 1966–1981, the Great Inflation: the 10-year yield climbed from 4.65% at the end of 1965 to 15.84% in September 1981, and consumer prices nearly tripled. Bonds lost 2.95% a year after inflation, more than stocks (0.97%). For the January 1966 retiree, every dollar moved into bonds made things worse: the safe rate was 3.91% with all stocks, 3.80% with 60/40 and 3.32% with all bonds.

In other words, regular bonds are insurance against one kind of disaster: a deflationary bust, in which stocks crash and consumer prices fall, so each fixed bond payment buys more. They do little, and can be a drag, when the enemy is inflation, which is why every 60/40 retirement that ran out at 4% began between 1964 and 1969. It is the same reason how bonds react to inflation matters as much as how many of them you hold.

Real Retirees Since 2000: Bonds Rescued One Group and Failed Another

Recent history shows both sides. None of these retirements has finished 30 years yet, so the table shows where each stood in September 2026.

Real portfolio value at the end of September 2026 as a share of the starting balance (start = 100), withdrawing 4% a year raised with inflation. Bonds = 10-year Treasuries.

A retiree who started in January 2000 with 100% stocks hit the dot-com crash and then 2008 within the first decade. Withdrawing 4%, they had just 5% of their starting balance left in real terms by September 2026 and were on course to run out before year 30. A 60/40 retiree still had 55%. From 2000 to 2002, stocks lost 37.6% in total while 10-year Treasuries gained 42.5%.

Portfolio value after inflation and after withdrawals of 4% of the starting balance a year, raised with inflation, January 2000 to September 2026. Bonds = 10-year Treasuries, rebalanced monthly. Source: S&P 500 (S&P Dow Jones Indices, Robert Shiller); U.S. Treasury/Federal Reserve (FRED); U.S. Bureau of Labor Statistics; Bondfish analysis.

Notice the all-bond line. It stayed ahead of 60/40 for more than 23 years, then fell from 73 at the end of 2021 to 46 by October 2023, as yields jumped and inflation pushed up the withdrawals. The 2022 retiree got the opposite of the 2000 one. In 2022, stocks fell 18.1% and 10-year Treasuries 16.4%, while consumer prices rose 6.5%. Bonds offered no cushion: by September 2026, an all-bond 2022 retiree had 55% of the starting balance left in real terms, against 87% for 60/40 and 115% for all stocks. The November 2007 retiree, who started a month after the pre-crisis stock market peak, shows that stocks can also win a bad start: 2008 hurt, but the long bull market that followed more than made up for it, leaving an all-stock portfolio at 194% of its starting value in real terms against 117% for 60/40.

Which Bonds Help Most With the 4% Rule? Short Beats Long in the Worst Case

Bengen used intermediate-term Treasuries; the Trinity study used high-grade corporate bonds. Comparing bond types needs data for all of them, which narrows the test to retirements starting from 1962 to 1996. That window includes the 1966 disaster but not 1929.

60/40 portfolio with each type of bond, U.S. retirements starting February 1962 to October 1996, 4% withdrawals raised with inflation, rebalanced monthly. Long-term Treasuries and Aaa corporates modeled as 25-year bonds. All bonds = a portfolio of nothing but that bond.

60/40 portfolio using each type of bond, U.S. retirements starting February 1962 to October 1996, 4% withdrawals raised with inflation, rebalanced monthly. Long-term Treasuries and Aaa corporates modeled as 25-year bonds. Source: S&P 500 (S&P Dow Jones Indices, Robert Shiller); U.S. Treasury/Federal Reserve (FRED); Moody’s (via FRED); U.S. Bureau of Labor Statistics; Bondfish analysis.

• Short bonds gave the best worst case. A 60/40 portfolio with T-bills or 5-year Treasuries could have safely withdrawn 3.95% and 3.93%, against 3.76% with 10-year Treasuries and 3.48% with long-term Treasuries. In this window, 10-year Treasuries did not improve on all stocks’ worst case (3.79%) at all. Shorter bonds lose less when inflation drives yields up, the core idea behind bond duration versus maturity.
• Long bonds left more behind in normal times. The median 60/40 retiree ended with 1.6 times the starting balance in real terms with T-bills, against 3.2 times with long-term Treasuries and 3.5 times with Aaa corporates.
• T-bills only work as a partner to stocks. A portfolio of nothing but T-bills ran out at 4% in 56% of these retirements, because bill yields barely kept up with inflation.
• Corporate bonds behaved like long-term Treasuries. Aaa corporates gave a similar worst case (3.50% against 3.48%) and a little more left over thanks to their extra yield. Our series is built from Moody’s yields on seasoned Aaa bonds, which ignore defaults and downgrades, so if anything it flatters corporates.

No single bond won every disaster. Ten-year Treasuries were the hero in 1929; in the 1966 inflation, shorter bonds held up better. (Our T-bill data start in 1934, so we cannot test them on 1929.)

The 4 Percent Rule With Bonds Yielding 5%: What Changes Today

Starting yields matter far more for bonds than for stocks. Across our 1,509 retirements, the correlation between the 10-year yield on retirement day and the safe withdrawal rate was 0.66 for an all-bond portfolio, 0.47 for 60/40 and only 0.20 for all stocks.

On October 29, 2026, the 10-year Treasury yielded about 5.0% and the 30-year about 5.3%. That is the backdrop for the case for holding more bonds. But a high nominal yield is not enough on its own. In February 1966 the 10-year also yielded 5.02%, and that month began the worst 60/40 retirement among all start dates with a yield of 5% or more, because inflation then ran far above that yield.

What is different now is that retirees can lock in a real, after-inflation yield. Treasury Inflation-Protected Securities (TIPS) maturing in 20 and 30 years yielded about 3.1% above inflation on October 29. In month-end data back to 2004, the 30-year TIPS yield has been higher only in October 2008 and September 2026. A ladder of TIPS maturing each year for 30 years, bought at those yields, would pay about 4.9% of the starting amount a year, raised with inflation and backed by the U.S. government, if held to maturity. In November 2021, when real yields were negative, the same ladder would have paid only about 3.0%.

That compares well with the 4% rule, but it is not the same thing. A TIPS ladder spends the money down to zero in year 30, so it leaves nothing for heirs or for a longer life, and TIPS are U.S. securities with their own tax rules. Investors elsewhere can use inflation-linked government bonds in their own currency, which France, Italy and the U.K., among others, issue. A bond ladder calculator helps plan the rungs, and you can screen government bonds by maturity and yield on Bondfish.

How to Use Bonds With the 4% Rule: Five Lessons From the Backtest

1. Hold some bonds, but not most of your money in them. Keeping 20% to 50% in bonds gave the best worst case (3.70% to 3.80%). Below 40% stocks, failures multiplied.
2. Know what your bonds insure. Regular bonds protected against deflationary crashes like 1929 and bear markets like 2000–2002, not against 1966-style inflation. Inflation-protected bonds cover that gap.
3. Use shorter bonds for the safety role. Short and intermediate Treasuries gave 60/40 its best worst case from 1962 to 1996. Long bonds and corporates left more behind on average but failed more often.
4. Your withdrawal rate matters more than your mix. For 60/40, raising the withdrawal rate from 4% to 5% lifted the failure rate from 2.6% to 24.0%. No change in the stock-bond mix between 50% and 100% stocks moved it by more than 1.8 points.
5. Consider today’s real yields. TIPS did not exist until 1997, so the 4% rule’s creators never had this option. At current yields, a 30-year TIPS ladder can fund about 4.9% a year in real terms with no market risk if held to maturity.

The Bottom Line

Bonds are not what makes the 4 percent rule work. In 155 years of U.S. data, stocks did most of the work, and between 50% and 100% stocks the failure rate at 4% barely moved. What bonds bought was protection against the worst sequence on record, the 1929 crash: they lifted the worst-case safe withdrawal rate from 3.07% to about 3.80% and pushed the earliest failure at 4% from 16.5 years to 25.7 years. They did nothing against inflation, and holding too many of them led to slow failure. With real yields near 3%, the best bonds for a 4% retiree today may be the ones protected against inflation.

Data and sources

S&P 500 total return: Robert Shiller’s U.S. stock market data (monthly averages, 1871–1989) and month-end S&P 500 total return index levels (from October 1989). 10-year Treasury yields: Robert Shiller (to 1961), U.S. Treasury and Federal Reserve via FRED (from 1962). T-bill, 5-year and long-term Treasury yields: U.S. Treasury and Federal Reserve (FRED). Corporate bond yields: Moody’s seasoned Aaa corporate bond yield (via FRED). U.S. consumer prices: Robert Shiller (to 1912) and U.S. Bureau of Labor Statistics (from 1913). Current Treasury and TIPS yields: U.S. Treasury market yields at the October 29, 2026 close. Bond returns, portfolio returns, withdrawal simulations and the TIPS ladder payout (real yields interpolated across maturities) are Bondfish’s own calculation.

Research on the 4% rule

• Wikipedia: 4% rule (Bengen, Journal of Financial Planning, 1994, and later updates)
• AAII Journal: Will Your Savings Last? What the Withdrawal Rate Studies Show (on the Trinity study)
• Morningstar: The State of Retirement Income 2025 (December 2025)
• Advisor Perspectives: Bill Bengen Boosts the “4% Rule” to 4.7% (August 2025)

Data

• Robert Shiller: Online Data
• FRED: 10-Year Treasury Constant Maturity Rate (DGS10)
• FRED: 3-Month Treasury Bill Secondary Market Rate (TB3MS)
• FRED: Moody’s Seasoned Aaa Corporate Bond Yield (AAA)
• U.S. Bureau of Labor Statistics: Consumer Price Index
• U.S. Treasury: Daily Treasury Par Real Yield Curve Rates, October 2026

This article is for general information only and is not investment advice. Investing involves risk, including possible loss of principal. Past performance and historical backtests do not guarantee future results. Withdrawal rates depend on your own circumstances, taxes and other income; consider consulting a licensed financial professional before making retirement decisions.

Author

Ivan Govakov

Emerging Markets Bond Analyst, ex-Santander

Fresh news

1. The 4% Rule: How Much Do Bonds Actually Help?
2. Weekly bond digest: October 19 – October 25, 2026
3. Risk Parity for Retail: Stocks and Bonds at Equal Risk
4. Barbell vs Bullet vs Ladder: Three Ways to Own Bonds
5. Weekly bond digest: October 12 – October 18, 2026