Correction (Market)

A market correction is a decline in the price of a stock, market index, commodity or other asset of roughly 10% or more from its most recent peak, but less than the 20% fall that is conventionally used to mark a bear market. The term implies that prices had risen too far, too fast, and are being "corrected" back towards a more sustainable level. Corrections can last from a few days to several months.

The 10% and 20% thresholds are rules of thumb used by financial media and analysts rather than formal definitions. Corrections are a normal part of market cycles: broad stock indices such as the S&P 500 have historically experienced them every year or two on average, and most did not turn into bear markets. In crypto markets, where prices are far more volatile, falls of 20% or 30% are often still described as corrections within a longer uptrend.

Picture a ball thrown up a staircase. It bounces up several steps, then rolls back down one or two before carrying on. A correction is that roll back down: a noticeable drop after a strong run, without the ball falling all the way to the bottom of the stairs.

For investors, a correction feels uncomfortable because a tenth or more of their gains disappears, often quickly. But it is very common, and it can let some of the excess optimism out of prices. The real question, which nobody can answer in advance, is whether the drop is a short pause or the start of a much longer slide.

A correction is measured as the percentage fall from the most recent peak, usually based on closing prices:

• $P_{\text{peak}}$ is the highest closing price reached before the decline.
• $P_{\text{current}}$ is the current or lowest closing price since that peak.

For example, if an index peaks at 5,000 and falls to 4,400, the decline is (5,000 − 4,400) ÷ 5,000 × 100% = 12%, which counts as a correction. If it falls further to 3,900, the decline is 22%, and the episode would be called a bear market. The same calculation, applied to a portfolio, gives its drawdown. Because the peak is only known after the fact, a correction is usually confirmed once the 10% threshold has been crossed.

Corrections are usually triggered by a change in expectations or risk appetite rather than a single cause. Common triggers include:

• Interest rates. Rising rates or signs that central banks will tighten monetary policy reduce the present value of future profits.
• Economic data. Weak growth, high inflation or fears of recession can prompt investors to cut risk.
• Valuations. After long rallies, high valuations leave markets sensitive to disappointment.
• Shocks. Geopolitical events, financial failures or pandemics can cause sudden falls.
• Leverage and positioning. When many investors are borrowing to buy, falling prices can force selling through margin calls.

A correction is defined by size; a crash is defined mainly by speed, such as a fall of 10% or more in a day or two. A bear market is a deeper and usually longer decline of 20% or more, often linked to recessions. A correction can become a bear market if selling continues, and a market downturn can include several corrections and rallies. Some investors treat corrections as buying opportunities, while others reduce exposure; evidence suggests that trying to time them consistently is difficult. The opposite of a correction in a falling market is a rally or "relief rally".

Crypto assets regularly experience much larger swings than shares. During the bull runs of 2017, 2021 and 2024, Bitcoin fell by 20% to 30% several times before reaching new highs, and smaller tokens often fell by more. Traders therefore use looser definitions, and corrections can be sharpened by leveraged derivatives, which can trigger cascading liquidations. A prolonged decline of 70% or more, as in 2018 and 2022, is usually called a crypto winter rather than a correction.

Because the thresholds are arbitrary, a 9.8% fall is not technically a correction while a 10.1% fall is, though the economic difference is negligible. Measuring from closing prices or intraday highs can give different answers, and a correction in an index can hide much larger falls in individual stocks. The label describes what has happened rather than what will happen, so it is of limited use for forecasting. The behaviour of investors during corrections is studied as part of volatility and loss aversion research.