Warren Buffett’s 90/10 rule got an unwelcome stress test this week. The S&P 500 fund dropped 4.4% on Thursday, and the broad index now sits 5.6% below its October 6 record. For retirees weighing Buffett’s famous allocation, the pullback raises a practical question: is a portfolio that is 90% stocks really safe when you are drawing income from it?
The answer depends less on the stock percentage than on the size of the bond cushion. For a couple with $2 million, a 10% bond sleeve equals $200,000, or three years of withdrawals at $66,000 a year. That was enough for the COVID crash and the 2022 decline. It was not enough for the 2007–2009 financial crisis, when the wait to recover stretched to 65.1 months and withdrawals totaled $358,050. The $200,000 reserve would have run out about 29 months before the old high was regained.
What Buffett Actually Said
Buffett’s instruction comes from his 2013 shareholder letter. He directed the trustee for his wife’s money to place 10% in short-term government bonds and 90% in a very low-cost S&P 500 index fund. He also said he expected that trust to outperform most investors who pay high fees for professional management. The rule was written for a trust with one beneficiary and no fixed annual withdrawal, a detail that matters when applying it to a retiree’s budget.
How the Math Works for a $2 Million Couple
Consider Gary, 66, and Linda, 63, a retired couple with $2 million split across an IRA, a taxable account, and a Roth. They spend $120,000 a year, receive $54,000 in Social Security, and draw $66,000 from savings. That is a 3.3% withdrawal rate, which is relatively modest because Social Security covers 45% of their spending.
Under Buffett’s 90/10 rule, they would hold $1,800,000 in the stock fund and $200,000 in short-term government bonds. Thursday’s 4.4% drop would have trimmed about $79,365 from the stock side, leaving $1,720,635. From the October 6 peak to Friday’s close, the decline is roughly $101,195. None of that forces a sale, because their next year of spending is already parked in the bond sleeve.
Three Crashes, One Conclusion
Looking at three recent drawdowns shows the limits of a 10% reserve. In 2020, the S&P 500 fund fell 34.1% and recovered in 5.9 months. Withdrawals during the wait were $32,450, easily covered. In 2022, the fund fell 25.4% and took 24.5 months to recover, with withdrawals of $134,750. The $200,000 reserve covered that too, leaving $65,250.
The 2007–2009 crash was different. The fund fell 56.5%, a paper loss of $1,017,000 on the $1.8 million stock position. The wait to regain the old high was 65.1 months, and withdrawals totaled $358,050. The $200,000 bond sleeve lasted about 36 months, leaving the couple to sell stocks for roughly 29 months while the fund was still underwater.
Sizing the Bond Sleeve by Years, Not Percentages
The better approach is to measure the bond reserve in years of withdrawals rather than a fixed percentage. Multiply your annual withdrawal by the longest recovery you want to withstand. For Gary and Linda, $66,000 times 65.1 months divided by 12 equals $358,050, or about 18% of the $2 million portfolio. That would mean an 82/18 split: $1,641,950 in stocks and $358,050 in bonds.
In a 56.5% decline, the stock side would still fall about $927,269, but the couple would never have to sell into the downturn. A more conservative 70/30 split, with $600,000 in bonds, would cover 9.1 years of withdrawals. It fits someone drawing closer to 5% of savings, where the $200,000 Buffett sleeve would last only two years.
The Caveats
The index figures used here are price only and exclude dividends. With dividends reinvested, each recovery would have come sooner, so the reserve math is cautious. The bond sleeve is also assumed to hold its value and earn nothing, though short-term government bonds do pay interest, which would stretch the reserve further.
Three crashes are a small sample, and a wait longer than 65.1 months is possible. Taxes are another omission: most of this money sits in an IRA, so a $66,000 withdrawal is less than $66,000 to spend. Refilling the reserve by selling in the taxable account could trigger capital gains.
Buffett’s rule assumes an investor can watch a seven-figure paper loss and keep spending from bonds without touching the stock fund. For Gary and Linda, the stock weighting is not the problem. The bond sleeve is. At $200,000 it covers three years of withdrawals; at $358,050 it covers the longest recent recovery. Thursday’s drop changes none of that. It is precisely the kind of day the reserve exists for.