A count of every 1% down day in the S&P 500 index since 1951: how many a normal year has, how 2026 compares so far, and why these days cluster in the same years as the big up days.
Is the market backdrop risk‑on or risk‑off? Get the daily read in Exent Market intel →
In a typical year, about 21 to 29 times. Exent’s data on S&P 500 index closes shows a median of 21 down days of 1% or more per full year from 1951 to 2025, and 28.5 since 1990. A year has about 252 trading sessions, so that is roughly one 1% drop every two weeks.
The average is higher than the median because a few crisis years pull it up. The range is wide: from 3 drops in 1964 to 75 in 2008. A 1% drop here means a close at least 1% below the prior close, using the index price without dividends.
| Period | Full years | Median 1% drops | Mean | Fewest | Most |
|---|---|---|---|---|---|
| 1951–2025 (ex 1970) | 74 | 21 | 25.2 | 3 (1964) | 75 (2008) |
| Since 1990 | 36 | 28.5 | 30.7 | 4 (1995, 2017) | 75 (2008) |
| Since 2000 | 26 | 30 | 33.9 | 4 (2017) | 75 (2008) |
| Since 2010 | 16 | 25 | 28.2 | 4 (2017) | 63 (2022) |
21 times in 2026 through Oct 22, in 203 trading sessions. Two of those drops reached 2% or more. None reached 3%. Over the same stretch the index also rose 1% or more 23 times.
October has been busier than the summer. The three most recent drops came within seven sessions: Oct 14 (−1.11%), Oct 16 (−1.09%) and Oct 22 (−1.21%).
| 2026 through Oct 22 | Count |
|---|---|
| Trading sessions | 203 |
| Drops of 1% or more | 21 |
| Drops of 2% or more | 2 |
| Drops of 3% or more | 0 |
| Gains of 1% or more | 23 |
Not by this measure. By Oct 22, the median year since 1990 had 22 drops of 1% or more, and the mean was 25.5. 2026 had 21. Of the 36 full years since 1990, 16 had fewer 1% drops by this date and 20 had more.
So 2026 sits close to the middle, slightly on the calm side. It is busier than very quiet years like 2017, which had only 4 all year. It is far from 2008, 2002 or 2022. Against the longer 1951–2025 record, the median by Oct 22 was 17, so the longer view reads a little busier. Since the 1990s, daily moves of 1% have simply been more common than in the 1950s and 1960s.
A count of 1% days is one rough gauge. It says nothing about how far the market sits from its high. For that, see how risky the stock market is right now and what a normal drawdown looks like.
It swings with the market cycle. The table shows every year since 2000, with 1% drops, 1% gains and 2% drops side by side. The busiest years line up with bear markets: 2000–2002, 2008–2009, 2011, 2020 and 2022.
| Year | 1%+ drops | 1%+ gains | 2%+ drops |
|---|---|---|---|
| 2026 (to Oct 22) | 21 | 23 | 2 |
| 2025 | 29 | 26 | 7 |
| 2024 | 19 | 31 | 4 |
| 2023 | 28 | 36 | 1 |
| 2022 | 63 | 59 | 23 |
| 2021 | 21 | 34 | 5 |
| 2020 | 45 | 64 | 25 |
| 2019 | 15 | 22 | 5 |
| 2018 | 32 | 32 | 15 |
| 2017 | 4 | 4 | 0 |
| 2016 | 22 | 26 | 5 |
| 2015 | 31 | 41 | 6 |
| 2014 | 19 | 19 | 4 |
| 2013 | 17 | 21 | 2 |
| 2012 | 21 | 29 | 3 |
| 2011 | 48 | 48 | 21 |
| 2010 | 37 | 39 | 10 |
| 2009 | 55 | 62 | 28 |
| 2008 | 75 | 59 | 41 |
| 2007 | 34 | 31 | 11 |
| 2006 | 13 | 16 | 0 |
| 2005 | 17 | 13 | 0 |
| 2004 | 20 | 21 | 0 |
| 2003 | 37 | 45 | 5 |
| 2002 | 72 | 53 | 29 |
| 2001 | 54 | 51 | 13 |
| 2000 | 54 | 48 | 19 |
| Median since 1990 (full years) | 28.5 | 32.5 | 5 |
Source: Exent’s data on S&P 500 index closes, price only. 2001 had 248 sessions because markets closed from Sep 11 to Sep 14 after the attacks; that gap is real, not missing data. The busiest year before 2000 was 1974, with 67 drops of 1% or more.
Because both are signs of the same thing: a nervous market. When uncertainty rises, prices move more in both directions. In 2008 the S&P 500 had 75 drops of 1% or more and 59 gains of 1% or more. In 2020 it had 45 drops and 64 gains. In 2017, the calmest recent year, it had 4 of each.
This is why a run of 1% drops is not, by itself, a sign that a crash has started. Many of the largest up days in history came in the middle of bear markets. Missing them by stepping out after a bad week was a common cost for investors. How often real crashes have come is in how often the stock market crashes, and how to tell routine moves from meaningful ones is in market noise vs signal.
In dollars it can feel large, even though it is routine. Illustrative: on a $750,000 portfolio that tracks the index, a 1% drop is $7,500 in one day. A typical year has more than 20 days like that, and about as many days of 1% gains. The dollar figure grows with the portfolio, but the frequency does not change.
Investors often find a few simple checks more useful than watching each day (education, not advice):
For an index fund view, see the SPY data page and the S&P 500 member list. What has tended to hold up on down days is in what goes up when stocks go down.
The count uses daily closes of the S&P 500 index from Jan 3, 1950 to Oct 22, 2026, price only, without dividends. A drop of 1% or more means a close at least 1% below the previous session’s close; 2% and 3% drops and 1% gains are counted the same way. The series was checked against Exent’s own index table for 2016 to 2026 with no mismatches.
Full‑year statistics cover 1951 to 2025. 1970 is left out because our series is missing Jan 2 to Mar 19, 1970, which would undercount that year. Year‑to‑date comparisons use each year’s sessions through Oct 22. These are past counts. They describe how the market behaved, not how it will behave for the rest of 2026.
From 1951 to 2025 the S&P 500 closed 1% or more lower in a median of 21 sessions a year, and 28.5 a year since 1990. That is roughly one drop every two weeks, with far more in bear markets.
About 21 to 29, depending on the period. The median was 21 for 1951 to 2025 and 25 since 2010. Calm years had as few as 3 or 4, and 2008 had 75.
In 2026 through Oct 22, the S&P 500 fell 1% or more 21 times in 203 sessions. Two of those drops were 2% or more, none reached 3%, and the index rose 1% or more 23 times.
By the count of 1% drops, no. The median year since 1990 had 22 by Oct 22 and 2026 had 21, so this year is slightly calmer than typical so far. Past counts do not predict the rest of the year.
Much less often. Since 1990 the median full year had 5 drops of 2% or more, but the range ran from 0 in calm years to 41 in 2008. 2026 had 2 through Oct 22.
Market Intelligence reads the market regime and tells you whether the backdrop favors taking risk or protecting capital.
Educational content, not investment advice. Examples are illustrative; past performance does not guarantee future results.