Learn
What is a market correction? The 10% drop every investor meets
The Nasdaq fell more than 10% from its record high this week — the textbook definition of a correction. What the word means, how often it happens, and what history says comes next.
On Thursday 5 November, the Nasdaq Composite fell more than 10% below its record closing high of 27,599.89, set on 6 October — putting it on course to confirm a correction, its second of 2026. The word was everywhere in the headlines within hours. But what does "correction" actually mean, and should an ordinary investor be worried? The short answer: it is a normal, frequent event, and history says most corrections end with the market back at a new high. The longer answer is worth knowing before the next one arrives.
The definition: 10% from the high
A market correction is a decline of 10% or more in a broad index — usually the S&P 500 or the Nasdaq Composite — measured from its most recent peak. There is no official definition; 10% is simply the line the market has agreed to use. Fidelity puts it this way: "While there is no official definition, a stock market correction is generally considered a drop of at least 10% from recent market highs."
The numbers around it form a ladder:
- Pullback: a drop of less than 10%. Small, frequent, barely newsworthy.
- Correction: a drop of 10% up to just under 20%.
- Bear market: a drop of 20% or more.
The labels describe how far the index has fallen, not why it fell and not what happens next. A correction can be caused by anything — an election result, an earnings miss, a rate scare — and the same size of drop has meant very different things in different years. That is the first thing to understand: the word tells you the size of the fall, nothing more.
How often do corrections happen?
More often than most people expect. According to Fidelity, since 1980 the S&P 500 has fallen 5% or more at some point in 93% of calendar years, and 10% or more in 48% of calendar years — in other words, a correction has shown up in nearly every other year.
Counting by episodes rather than by years gives a similar picture. The Motley Fool counted 24 declines of 10% or more from a record close in the S&P 500 since 1950, or about one every three years on average — but never on a schedule. Some years have none. Some years have two. (Different studies count differently depending on whether they measure from record closes or from any peak, so treat any single count as approximate.)
The useful point is not the exact number. It is that a correction is not a rare disaster. It is a routine feature of owning stocks, like turbulence on a flight: uncomfortable, sometimes sharp, and expected often enough that a plan should assume it will happen.
How long do they last?
Most corrections are surprisingly short. Fidelity cites Yardeni Research data showing past corrections have averaged about 115 days. The Motley Fool's analysis of the 24 declines since 1950 found that all of them eventually ended with a new record close, and the median wait from the bottom to that new high was around five months. For the declines that stayed below 20% — true corrections, not bear markets — the median wait was around three months, and none took more than about six and a half months.
But averages hide the range. The quickest recoveries took around a month; the slowest took almost six years, and those were bear markets, not corrections — the drops that ended in 1974, 2002 and 2009 each took about four to six years to come back. There is no way to know in advance whether a 10% drop will stop at 12% or keep going to 30%.
There is one more number worth holding on to. Fidelity notes that the S&P 500 has had an average decline of about 14% from peak to low within a calendar year, based on data from 1980 through 2025 — yet its average calendar-year return over the same period was 13.3%. The year can be a good one even with a big dip inside it. The dip and the year's result are two different things.
A concrete example: this week
The Nasdaq's slide this week shows how a correction actually arrives. On Wednesday 4 November, the Dow had its best day since April 2025, up 3.03%, after the midterm election result. By Thursday afternoon, the Nasdaq was down 4.6% on the day and more than 10% below its October record — a correction, confirmed in the space of a day, on the same news that had pushed the market up the day before.
Nothing about the companies had changed that fast. What changed was the story investors told about them: on Wednesday, a divided Congress meant gridlock and less policy risk; on Thursday, the same Congress meant closer scrutiny of the AI data-centre buildout, and chip stocks fell nearly 11% at their lows. The correction label describes the size of the move. The reason for it is a separate question, and it can change overnight.
What beginners get wrong
Treating 10% as a magic line. It is a convention, not a law of nature. An index down 9.9% is not safe and one down 10.1% is not doomed. The line matters because everyone watches it, and that watching can itself move prices — but it predicts nothing on its own.
Assuming a correction means a crash. Most corrections do not become bear markets. The majority end with the index recovering to a new high within months. The fear in the moment is real, but the base rate is on the side of recovery.
Selling at the bottom, then waiting for "confirmation" to buy back. This is the expensive mistake. The bottom is only visible afterwards, and the recovery often comes fast — the quickest of the post-1950 declines took around a month from trough to new record. An investor who sells during the fall and waits until the headlines turn positive has usually missed the sharpest part of the rebound.
Checking the portfolio every hour. A correction feels worse the more often you look at it. For a long-term investor, the daily moves during a correction are noise around a position that was chosen for years, not days.
What to watch next
A correction is confirmed by the close, not by an intraday low, so the first thing to watch is whether the Nasdaq actually closes more than 10% below its record. After that, the questions are simple: does the selling broaden beyond chips and AI names, or does it stay concentrated? Does the VIX, the fear gauge, keep climbing or settle? And does the October jobs report, due Friday morning, change the story again?
The deeper question is whether the drop stops at a correction or deepens past 20% into a bear market. No indicator answers that in advance. What history does say is that corrections are common, usually brief, and that every one of the S&P 500's 10% declines since 1950 eventually ended at a new record high. That does not guarantee the next one will. It does mean panic is the one response the record does not support.
This article is for education only, not financial advice.