Herd behaviour and the stock market crash: why crowds move as one

What is herd behaviour in investing, and why does it matter in a crash?

Herd behaviour describes investors who copy others instead of acting on their own analysis. Observers have long linked large market trends to it: Wikipedia notes that big trends often begin and end with periods of frenzied buying, called bubbles, or frenzied selling, called crashes. Herding therefore sits at both ends of a cycle, and the same mechanism that lifts prices can also push them down.

Researchers in behavioural finance, including Vernon Smith, Amos Tversky, Daniel Kahneman, and Robert Shiller, are associated with work on herding as a form of investor irrationality. John Maynard Keynes listed greed and fear among his animal spirits. The parent guide on panic selling looks at the selling side. This page looks at the common root, which is the choice to follow others, and at what that choice does to prices.

How does an information cascade make herd behaviour spread?

An information cascade starts when a person learns from others’ choices and sets aside private signals. The standard picture is a diner choosing between 2 empty restaurants: later diners pick the busier one, which then gets more business. Banerjee published a simple model of herd behaviour in 1992, and Bikhchandani, Hirshleifer, and Welch modelled fads and informational cascades in the same year (Wikipedia).

In markets the choices are trades. A holder who sees heavy selling infers that sellers hold bad news, and sells too, which makes the next holder’s inference stronger. Cascades can rest on very little information, because each follower adds a choice but no new information. That is why a few early sellers can start a wave that later sellers join without news of their own.

Which motives push investors to follow the herd?

3 motives explain most following. The first is learning: others may know more, so copying is a sound guess. The second is safety in numbers, since an error made with everyone else is easier to explain than an error made alone. The third is momentum: a rising price attracts buyers, and a falling price attracts sellers, whatever the cause.

Each motive is reasonable for one person, and the sum is a market in which many people hold the same position. Hu and colleagues found in 2021 that shoppers on a television shopping network were more likely to buy when prior purchases were displayed, which shows the pull of visible behaviour outside finance too (Wikipedia). Screens that show prices and volumes display the same signals to traders.

How does herding inflate a bubble before a stock market crash?

Herding builds a bubble when buying copies buying. The EQOS Asia 2008 back-test (2026) records a rise of 124% in US house prices from 1997 to 2006, a price-to-income ratio of 4.6 times against an average of 3.2 times, and a subprime share of lending of 23.5%, up from 8% in 2003. Each number grew while the crowd agreed that the trend would continue.

The study measures the agreement. It sets resonance at 0.65, the market vibrating in unison around the idea that real estate never falls, and transparency at 0.15. The ratio of 4.3:1 sits above its 3:1 threshold, which it labels a bubble anomaly. A crowd that agrees this strongly holds the same position, and the same position sells at the same moment.

Why does herding reverse into a stock market crash?

The mechanism that lifts prices also pulls them down, because it depends on the crowd’s direction and not on value. When a trigger breaks the shared belief, the first sellers become the signal for the rest. The study records trust falling from 0.55 to 0.10 within days in August 2007, when the TED spread rose from 20bp to 200bp after BNP Paribas froze 3 funds.

Speed comes from the same copying that built the bubble. On 19 October 1987 the Dow fell 22.6% in one session (Wikipedia), and on 28 and 29 October 1929 it fell 12.82% and 11.73%. In each case, visible sales invited more sales. The lesson for a reader is that crowd agreement is a state to measure, because it marks how many holders can move together.

How can you measure herding in markets?

4 measures give a practical view. Concentration of holdings shows how many investors own the same shares. Correlation across stocks shows whether prices move together more than usual. Order-flow balance shows whether trades run in one direction. And dispersion of analyst views shows whether opinions have converged. Research on institutional herding, such as Froot and colleagues’ Herd on the Street (1992), shows that the topic has an academic literature.

Record each measure with a date and compare it with its own history. A reading that rises while prices rise suggests a crowded trade, and a reading that rises while prices fall suggests a crowded exit. The study’s resonance to transparency ratio, 4.3:1 in 2006 and 2007, is one structural way to combine these ideas, with 3:1 as its threshold.

How can you respond to herd behaviour as an investor?

Write your reasons down before you act, so that you can compare them with the crowd’s reasons. Check whether a view rests on a source, a date, and a calculation, or only on the fact that many people hold it. Warren Buffett’s advice to be fearful when others are greedy, and greedy only when others are fearful, points the same way: the crowd’s mood is data about the crowd (Wikipedia).

The EQOS Asia Λ simulator tests a case in 10 minutes, and a 30-minute pre-diagnostic applies the scoring to your exposure. This page is general information, not investment advice, as of October 2026, and it describes behaviour without recommending a trade. The sister guide on the fear and greed index covers one gauge of crowd mood. Review your plan each quarter.

Frequently asked questions

About the author

Stéphane Jambu

President, EQOS Forecast Asia

Stéphane Jambu leads EQOS Forecast Asia and works as a consultant in e-commerce, search visibility, and AI. LinkedIn profile