Stock Market Correction, Explained: Every 10% Decline Since 1950

A stock market correction is a decline of 10% or more from a recent high, usually measured on closing prices of a broad index like the S&P 500, that stops short of 20%. Past 20% the same decline is called a bear market. Those thresholds are conventions, not rules: there is no regulator that declares a correction, which is why two news stories can disagree on the day one began. The numbers that matter more than the label are how deep corrections go, how long they last, how often they arrive, and how many of them turn into something worse. We measured all of that on every S&P 500 decline of 10% or more since 1950.

The definitions first, because the searches ask for them. A pullback is a drop of roughly 5% to 10%, common enough to happen several times a year. A correction is 10% to 20%. A bear market is 20% or more. A crash is a word for speed, not depth: a very large drop in a day or a few days, like October 19, 1987 or March 2020, which may or may not be inside a bear market. Schwab's explainer and Fidelity's both note that the 10% line is widely used but not official, and that the measurement runs from a closing high to a later closing low.

Every 10% decline in the S&P 500 since 1950, sorted into corrections and bears

We took daily closes of the S&P 500 index from Yahoo Finance from January 1950 through November 6, 2026, tracked the record closing high, and marked an episode each time the index closed 10% or more below that record. The episode's depth is the lowest close before the index made a new record, and it ends the day a new record is set. Declines that happened while the index was still recovering from an earlier, larger fall are counted inside that larger episode; the 19% slide of 2011, for example, sits inside the 2007 to 2013 recovery.

Three readings fall out of the table. First, 42% of the declines that reached 10% kept going to 20%: 11 of 26. A correction in progress is not a coin flip toward a bear market, but it is not a rare path either. Second, the ones that stopped were quick by market standards: a median of three months down and three months back, with the whole round trip under a year in every case but one. Third, bears are a different animal in time, not just depth: the median bear took fifteen months to bottom and another fifteen to recover.

How deep the 15 corrections went

Corrections cluster at the two ends of their range. Five of the fifteen stopped between 10% and 11%, barely qualifying. Four went past 18.9%, within a point or two of the bear line, including the 19.9% of 1990, the 19.8% of late 2018 and the 18.9% of early 2025. The middle is thin: six episodes between 12% and 15%. In other words, once a decline passes about 15%, history has more often carried it toward 20% than stopped it there.

How often they come

Twenty-six episodes in 76 years works out to one 10% decline roughly every three years, but the spacing is uneven. The 1950s and 1960s each had four or five, the 1970s one enormous one, the 2000s two that were both bears, and the 2010s three. The 2020s have had three in six years: the 2020 crash, the 2022 bear and the spring 2025 correction. Long calm stretches, like 2003 to 2007 or 2009 to 2015, are as normal as clusters.

The last twelve

Is the stock market in a correction right now?

As of the close on Friday, November 6, 2026, the S&P 500 stood 5.65% below its record close of October 6, 2026. That is a pullback by the conventional labels, not a correction. A correction would begin on a close roughly 10% below that October high, and a bear market on a close 20% below it. This paragraph describes the index on one date; it says nothing about where it goes next, and we will not pretend otherwise.

What causes corrections

The honest answer is that the trigger is usually named after the fact. The 2018 corrections were pinned on rate fears and a trade dispute, 2025's on tariff announcements, 1998's on a currency crisis and a hedge fund failure. What the episodes share is less a cause than a setup: valuations that had run ahead of earnings, positioning that was crowded on one side, and a piece of news that made a lot of people want to leave at once. Corrections are the market's ordinary way of repricing; the question for anyone with positions is not why this one started but what their plan says to do while it runs.

What a plan does with this history

Two facts from the table carry directly into rules. A typical correction takes about three months to bottom and gives back roughly 14% at the index level, which means individual stocks fall far more; and about four in ten of these declines keep going. Rules-based accounts handle that with settings they already have, rather than with a view on the market.

• A stop loss on every position, sized so that an ordinary down day does not trigger it. How often the index drops 1%, 2% or 3% in a day is the reference for that, and the basics of stop losses cover the mechanics.
• A trailing stop that follows a winner up and lets a correction take profit rather than erase it. Our test of 1% to 8% trailing stops on every 10% mover of October 2026 shows the trade-off between tight and loose.
• A daily loss brake, a dollar amount at which the account stops opening new positions for the day. Corrections are made of clusters of red days, and the brake is what keeps a bad week from compounding.
• Position sizing from cash, so that a 14% index decline, which can be 25% in a single name, is a small share of the account rather than a large one.

JorgAI members set all four of these themselves and the software carries them out inside their own brokerage account; the auto-trader also stops opening positions when the daily brake is hit and only re-enters on its own rules. None of that changes what the market does; it changes how much of a correction one account absorbs. The longer-horizon version of this question, how far an account should ever be allowed to fall, is the maximum drawdown guide, and the calendar context for when declines have clustered is in stock market seasonality since 1950.

Quick answers

What is a stock market correction?

A decline of 10% or more from a recent high, measured on closing prices, that does not reach 20%. The 10% line is a convention used by exchanges, brokers and the press, not an official rule.

What is the difference between a correction and a bear market?

Depth. A correction is a 10% to 20% decline from a recent high; a bear market is 20% or more. In the S&P 500 since 1950, 11 of the 26 declines that reached 10% went on to become bear markets.

How long does a stock market correction last?

In the S&P 500 since 1950, the 15 declines that stayed between 10% and 20% took a median of 91 days to reach their low and 94 more days to regain the record, about seven months in total. The longest took 543 days (2015 to 2016), the shortest 59 (October 1997).

How often do stock market corrections happen?

A decline of 10% or more from a record high has arrived about once every three years since 1950, with long quiet stretches and clusters. The 2020s have had three so far: 2020, 2022 and 2025.

What percentage of corrections become bear markets?

About 42% in the S&P 500 since 1950: 11 of the 26 declines that passed 10% went on to pass 20%.

Is the stock market in a correction in November 2026?

Not as of the November 6, 2026 close, when the S&P 500 was 5.65% below its October 6, 2026 record. A correction would start on a close about 10% below that high.

What is the difference between a pullback and a correction?

Size. A pullback is usually a 5% to 10% dip and happens several times in most years; a correction is 10% to 20%.

Is a correction the same as a crash?

No. A crash describes speed, a very large drop over a day or a few days. A correction describes depth over any length of time. March 2020 was both a crash and the start of a bear market; the spring of 2025 was a correction that took seven weeks to bottom.

How we put this together. The statistics come from Yahoo Finance daily closes of the S&P 500 index (ticker ^GSPC) from January 3, 1950 to November 6, 2026, pulled on November 9, 2026. An episode starts the first day the index closes 10% or more below its record closing high and ends the day it sets a new record; depth is the lowest close in between; days are calendar days. Declines inside a longer recovery are counted within the larger episode. Definitions follow Schwab and Fidelity, read November 9, 2026. The current-distance figure is as of the November 6 close and will be stale the day after you read it.

To see how a stop, a trailing stop and a daily brake behave through a down week before connecting a broker, start on the free plan and run the simulator.

JorgAI is software that carries out the rules you set in your own brokerage account. It is not investment advice and does not predict or promise results. Index history is descriptive only. Trading involves risk of loss.