Since 1946 the S&P 500 has dropped 10% or more about once every 1.8 years on average, and 20% or more about once every six. Each decline feels unprecedented while it is happening, which is why the record is worth knowing in advance. This guide lists every S&P 500 bear market since 1946 with how far the index fell, how long the slide lasted and how long it took to climb back, shows what bonds and cash did in the worst years, works through the arithmetic of recovering from a loss and ends with a checklist of what to do and avoid. For the basics of owning index funds, start with the guide to investing in stocks.
What is a bear market, and how is it different from a correction?
A bear market is a fall of at least 20% from a previous high, while a correction is a fall of 10% to just under 20%.
Both are usually measured on closing prices, from a peak to the lowest close before the index turns up. Neither term has an official definition, but the 10% and 20% thresholds are the convention most market historians use, including Yardeni Research, whose tables this guide relies on. A bear market ends at its low, which can only be identified afterward, and the next bull market is dated from that low.
Thresholds on closing prices can split hairs. In 2025 the S&P 500 fell 18.9% from its February 19 close to its April 8 close, although on April 7 it briefly traded 21.3% below the February closing high. Because no close crossed the 20% line, that slide counts as a correction and does not appear in the bear market table below. In 2026 the deepest decline through October was 9.1%, from a record 6,978.60 on January 27 to 6,343.72 on March 30, just short of a correction; the index set a new record of 7,818.93 on October 6 and closed at 7,720.39 on October 29.
How many bear markets has the S&P 500 had since 1946?
Thirteen, counting declines of 20% or more on closing prices, from the 26.6% slide of 1946 to the 25.4% drop of 2022.
| Peak close | Lowest close | Decline | Days from peak to low | First close back at the old peak | Months from low to old peak | Gain needed from the low |
|---|---|---|---|---|---|---|
| May 29, 1946 | Oct 9, 1946 | −26.6% | 133 | Jun 9, 1950 | 44.0 | 36.3% |
| Jun 15, 1948 | Jun 13, 1949 | −20.6% | 363 | Jan 9, 1950 | 6.9 | 25.9% |
| Jul 15, 1957 | Oct 22, 1957 | −20.7% | 99 | Sep 16, 1958 | 10.8 | 26.0% |
| Dec 12, 1961 | Jun 26, 1962 | −28.0% | 196 | Sep 3, 1963 | 14.3 | 38.8% |
| Feb 9, 1966 | Oct 7, 1966 | −22.2% | 240 | May 4, 1967 | 6.9 | 28.5% |
| Nov 29, 1968 | May 26, 1970 | −36.1% | 543 | Mar 6, 1972 | 21.4 | 56.4% |
| Jan 11, 1973 | Oct 3, 1974 | −48.2% | 630 | Jul 17, 1980 | 69.5 | 93.1% |
| Nov 28, 1980 | Aug 12, 1982 | −27.1% | 622 | Nov 3, 1982 | 2.7 | 37.2% |
| Aug 25, 1987 | Dec 4, 1987 | −33.5% | 101 | Jul 26, 1989 | 19.7 | 50.4% |
| Mar 24, 2000 | Oct 9, 2002 | −49.1% | 929 | May 30, 2007 | 55.7 | 96.6% |
| Oct 9, 2007 | Mar 9, 2009 | −56.8% | 517 | Mar 28, 2013 | 48.6 | 131.3% |
| Feb 19, 2020 | Mar 23, 2020 | −33.9% | 33 | Aug 18, 2020 | 4.9 | 51.3% |
| Jan 3, 2022 | Oct 12, 2022 | −25.4% | 282 | Jan 19, 2024 | 15.2 | 34.1% |
Peak and low dates, declines and day counts follow Yardeni Research's tables, which use S&P 500 closing prices; days are calendar days. Recovery dates are the first close at or above the old peak, taken from daily closes. For 1946 and 1948 the old peaks sat below the 1929 record, which the index did not regain until September 22, 1954. All figures are price only: reinvested dividends would have shortened every recovery.
How long do bear markets last, and how long do recoveries take?
Declines have typically been shorter than recoveries: a median of about 9 months from peak to low, then about 15 months back to the old high.
| Measure across the 13 bear markets | Median | Average | Range |
|---|---|---|---|
| Decline from peak to low | 28.0% | 32.9% | 20.6% in 1948–49 to 56.8% in 2007–09 |
| Time from peak to low | 282 days (9.3 months) | 361 days (11.8 months) | 33 days in 2020 to 929 days in 2000–02 |
| Time from low back to old peak | 15.2 months | 24.7 months | 2.7 months in 1982 to 69.5 months in 1974–80 |
| Time from peak back to old peak | 23.2 months | 36.5 months | 5.9 months in 2020 to 90.2 months in 1973–80 |
Depth and duration usually travel together. The three deepest declines, in 1973–74, 2000–02 and 2007–09, each took more than four years to recover from the low, while the three shallowest, all under 23%, recovered in under 11 months. Depth is only a rough guide, though: the 26.6% fall of 1946 needed 44 months to recover from its low, while the 33.9% crash of 2020 needed under five.
Most corrections stop short of a bear market. Yardeni counts 44 declines of 10% or more from February 1946 through the October 2023 low; adding the 18.9% slide of 2025 makes 45. Thirteen became bear markets and 32 did not, though several came within a point: 19.4% in 1976–78, 19.9% in 1990, 19.3% in 1998, 19.4% in 2011 and 19.8% in 2018.
Recessions make declines worse but are not required. Eight of the 13 bear markets overlapped a recession dated by the National Bureau of Economic Research, including the 18-month downturn from December 2007 to June 2009. The five that did not were 1946, 1961–62, 1966, 1987 and 2022.
What happens to bonds and cash in a bear market?
Cash earned a positive return in every losing year for stocks since 1946, and high-quality government bonds usually rose, but not always.
The table shows each calendar year since 1946 in which the S&P 500 lost money, including dividends, with the same year's return on three-month Treasury bills, 10-year Treasury bonds and Baa-rated corporate bonds, from Aswath Damodaran's data at NYU Stern.
| Year | S&P 500 | 3-month T-bills | 10-year Treasuries | Baa corporate bonds |
|---|---|---|---|---|
| 1946 | −8.43% | 0.38% | 3.13% | 2.51% |
| 1953 | −1.21% | 1.89% | 4.14% | 1.62% |
| 1957 | −10.46% | 3.22% | 6.80% | −0.72% |
| 1962 | −8.81% | 2.77% | 5.69% | 6.50% |
| 1966 | −9.97% | 4.86% | 2.91% | −3.45% |
| 1969 | −8.24% | 6.67% | −5.01% | −2.03% |
| 1973 | −14.31% | 7.04% | 3.66% | 4.32% |
| 1974 | −25.90% | 7.85% | 1.99% | −4.38% |
| 1977 | −6.98% | 5.26% | 1.29% | 9.95% |
| 1981 | −4.70% | 14.04% | 8.20% | 8.46% |
| 1990 | −3.06% | 7.75% | 6.24% | 5.65% |
| 2000 | −9.03% | 6.00% | 16.66% | 9.38% |
| 2001 | −11.85% | 3.48% | 5.57% | 8.60% |
| 2002 | −21.97% | 1.64% | 15.12% | 12.06% |
| 2008 | −36.55% | 1.40% | 20.10% | −3.44% |
| 2018 | −4.23% | 1.97% | −0.02% | −3.18% |
| 2022 | −18.04% | 2.09% | −17.83% | −15.23% |
Treasury bills returned more than zero in all 17 years, and 10-year Treasuries gained in 14, averaging 4.63% a year across the 17 while the S&P 500 averaged a loss of 11.98%. Corporate bonds were less dependable: Baa bonds, the lowest investment-grade tier, lost money in 7 of the 17 years, including 2008, because worries about defaults rise when stocks fall.
The exceptions share a cause. In 1969 and 2022, inflation and rising interest rates pushed stock and bond prices down together, and 2022 was the worst year for 10-year Treasuries in Damodaran's data back to 1928. Why bond prices fall when rates rise explains the mechanism. Cash behaves differently: it cannot fall in price, but what it earns depends on short-term rates at the time, from 0.38% in 1946 to 14.04% in 1981. The Treasury bill calculator shows what a bill pays at current rates. Even funds built for people near retirement fell hard in 2008, as the guide to target-date funds shows.
Why do the best days often come during bear markets?
Because the sharpest rebounds tend to follow the sharpest drops, the market's best and worst days arrive close together, usually in the same turbulent weeks.
Over the 20 years from October 30, 2006 to October 30, 2026, the S&P 500's price rose 8.90% a year, turning $10,000 into $55,011 before dividends. Missing only the 10 best of those 5,031 trading days would have cut the result to 4.58% a year, or $24,494.
Every one of those 10 best days fell inside a bear market or within two weeks after a major low: four in October and November 2008, two in March 2009, three in March and April 2020, and April 9, 2025, the day after the low of that year's 18.9% slide. The two largest were gains of 11.58% on October 13, 2008 and 10.79% on October 28, 2008. Nine of the 10 worst days fell inside the 2008 and 2020 bear markets, led by an 11.98% loss on March 16, 2020; the exception was August 8, 2011, during that year's 19.4% correction.
The pattern matters most for anyone tempted to sell after a run of bad days and wait for calm: the rebounds have often come before the calm did. Current prices, returns and the deepest recent drop for an S&P 500 fund are on the SPY data page.
How much does a portfolio need to gain to recover from a loss?
More than it lost, because the recovery starts from a smaller base: a 20% loss needs a 25% gain, and a 50% loss needs 100%.
Worked example: How long a recovery takes at steady returns
The 2020 bear market cut the S&P 500 by 33.9%, so recovering needed 1 ÷ (1 − 0.339) − 1 = 51.3%. At a steady 6% a year that takes ln(1.513) ÷ ln(1.06) = 7.1 years; at 8%, 5.4 years; at 10%, 4.3 years. In fact the index regained its February 2020 peak on August 18, 2020, less than five months after the low, which shows how uneven rebounds can be. The 2007–09 decline of 56.8% needed a 131.3% gain and took 48.6 months from the March 2009 low.
Worked example: A saver who kept buying through 2007 to 2009
An investor put $20,000 into an S&P 500 index fund at the October 9, 2007 close and kept adding $500 at each month-end close. At the March 9, 2009 low, after the initial $20,000 and 17 monthly purchases, $28,500 in all, the account was worth $13,619 on a price-only basis; the original $20,000 alone would have been worth $8,645. Because the later purchases bought shares at lower prices, the account was back above the money put in by January 14, 2011, worth $39,570 against $39,500 invested with the index at 1,293.24. The index itself did not regain its 2007 peak of 1,565.15 until March 28, 2013.
Worked example: Why withdrawals deepen a decline
A retiree has $500,000 in stock funds when prices drop 25%, leaving $375,000. Taking out $20,000 for living costs leaves $355,000, which needs a 40.8% gain to get back to $500,000, against 33.3% with no withdrawal. Shares sold near the low cannot take part in the rebound, which is the logic behind keeping some spending money in cash or short-term bonds, as covered in safe investments for retirees.
What should you do, and not do, during a bear market?
Most of the useful steps are set up before a decline starts; during one, the main job is avoiding forced or panicked decisions.
During a bear market
- Do keep an emergency fund outside stocks. Several months of expenses in cash means a job loss does not force a sale at low prices; how much emergency fund covers sizing it.
- Do keep automatic contributions running if the plan already includes them. New money buys at lower prices, and pausing turns a schedule into a guess.
- Do rebalance by rules set in advance, such as a calendar date or a band around the target mix, rather than by headlines. Asset allocation by age explains target mixes.
- Do consider harvesting losses in taxable accounts. Realized losses offset gains plus up to $3,000 a year against ordinary income ($1,500 if married filing separately), but the wash-sale rule disallows a loss if a substantially identical security is bought within 30 days before or after the sale (a 61-day window). See tax-loss harvesting and the wash sale rule.
- Do write down how the decline feels. A note made at a 25% loss is a better guide to the right stock allocation than a questionnaire filled in during a rally.
- Don't sell everything with a plan to buy back later. The best days of the past 20 years came during declines or soon after the lows, and that plan has to be right twice.
- Don't borrow against falling investments. A margin loan can force sales at the worst prices; margin accounts explained shows how margin calls work.
- Don't move near-term money into stocks because prices look cheap. A down payment or tuition bill due soon may not leave time for a recovery; where to keep a down payment covers the alternatives.
- Don't raid retirement accounts to cover losses elsewhere. Early withdrawals usually bring income tax and often penalties, and the money stops compounding.
- Don't assume bonds will offset every stock loss. In 2022, 10-year Treasuries lost 17.83% in the same year the S&P 500 lost 18.04%.
A windfall that arrives during a decline raises its own question, covered in lump sum vs dollar cost averaging.
What are the risks of relying on past recoveries?
The US record shows every bear market since 1946 recovering, but it is one market's history, and the averages hide long waits and losses that never came back.
- Very long waits. An investor who bought at the September 1929 peak waited until September 22, 1954, 25 years, for the index to regain it, price only. After the January 1973 peak the wait was seven and a half years.
- Other markets have fared worse. Japan's main index needed until February 2024 to top its 1989 peak, as the BBC reported; international stocks for US investors covers spreading money across countries.
- Single stocks have no recovery mechanism. An index keeps replacing shrinking companies with growing ones, while one company can fall to zero and stay there. Index funds explained covers how broad funds spread that risk.
- Inflation stretches the real wait. The recovery dates in the table are in nominal dollars; after 1973–74, when inflation ran high, it took far longer for the index to regain its old peak in purchasing power.
- Leverage can end the story early. Margin loans and leveraged funds can wipe out an account before the market recovers.
- Crypto declines run deeper. Bitcoin and other crypto assets have fallen much further and faster than stock indexes; crypto market cycles covers those swings.