Herd behavior

Herd behavior describes decisions shaped by what other people appear to think or do, rather than by an independent assessment of the underlying situation. In financial markets, this dependence can make rising prices look like proof that a stock is desirable, until imitation displaces the diversity of judgment that would otherwise help investors question the boom. The same threshold dynamic also illuminates crowds: collective action can build gradually, but once a crowd crosses into violence, its most extreme members can shape what follows.

Investor Behavior and Its Costs

The book places herding among several recurring departures from rational investment. Investors may prefer the apparent safety of acting with others; give disproportionate weight to recent, prominent news; mistake a run of good results for proof of a manager’s skill; cling to losing stocks because realizing a loss is painful; and trade too much because they overestimate their ability to beat the market. Herd behavior is part of this broader pattern, not a claim that every investor always follows the crowd.

In the account of the bowling-stock bubble, investors saw the rising prices of AMF and Brunswick as evidence that other people believed bowling was the next big thing. That apparent consensus made the shares more attractive, and buying them helped drive their prices higher still. Investors could treat AMF as a seemingly safe purchase because they expected another buyer to take the shares off their hands. As the prices climbed, the incentive to investigate whether the bowling boom made sense diminished. The resulting single-mindedness displaced the diversity of opinion on which a healthy market depends: everyone’s enthusiasm became evidence for everyone else’s enthusiasm.

A bubble and a crash are presented as opposite expressions of dependence. In a bubble, investors expect prices to keep rising; in a crash, they become convinced prices will keep falling. In either case, they can become less concerned with a stock’s real value than with whether it can be resold. The book notes that major crashes have often seemed disproportionate to their immediate causes, comparing the uncertainty of their timing to a sandpile that holds its shape as it grows until one more grain sends it tumbling.

News and the Feedback of Expectations

The stock market contains both independent and dependent decision-making. Some investors select companies they believe are strong and trust that others will eventually recognize their value. Others try to choose stocks they expect other investors to find attractive. The mix is usually changing but relatively stable; bubbles and crashes occur when it shifts too far toward dependence. The problem is not that investors consider one another at all, but that their expectations can become the main evidence on which they act.

Financial news can strengthen this feedback by making other people’s apparent views unusually visible. Before CNBC, traders had to infer what others thought largely from the ticker, leaving some distance between an individual and the market. CNBC instead reported continually on what traders were saying and how the market was feeling. In the book’s Keynesian beauty-contest analogy, this is like hearing all day who other people think is attractive, making it harder to judge independently.

The consequences are clearest when markets are already falling. A broadcast that frames a day as a panic can make each investor’s decision reflect the alarm displayed on screen; attention turns from the value of an investment to what everyone else is doing. A study of CNBC’s Midday Call segment gives a measure of the speed of such reactions: prices rose within the first fifteen seconds after a positive report, and trading volume sextupled in the first minute. The account argues that this response was too fast to show careful evaluation of the report’s content. Investors were reacting to the expectation that others would trade on the news, and the immediate contest became one of speed.

More information, however, does not necessarily produce better judgment. In experiments conducted with business students at MIT in the late 1980s, Paul Andreassen compared a group that saw only stock-price changes with another that received those changes along with a constant stream of financial news. The less-informed group did better. Read alongside the CNBC example, the finding distinguishes information that helps people assess value from information that mainly tells them how others may react.

Crowds, Thresholds, and Extremes

The book connects market herding with the formation of violent crowds through the work of sociologist Mark Granovetter. A crowd need not turn violent all at once. Some people are ready to riot, while others will not do so; most lie between those poles, and their willingness depends on how many others are already rioting. As participation grows, people whose threshold has been reached join in, encouraging further participation. What looks from the outside like one organism acting with one mind can therefore emerge through a sequence of individual decisions.

That escalation is not inevitable. If a crowd includes only a few instigators and many people who will act only when a sizable share of the group does, the disturbance may never begin. For violence to take hold, there must be instigators, people with low thresholds for violence, and a larger group that can be swayed. Once the crowd crosses its threshold, however, its behavior is shaped by its most violent members. Unlike an average judgment that might express collective wisdom, the judgment of a mob is extreme.

The comparison also clarifies what can restrain collective escalation. Granovetter’s model includes people who will not riot under any conditions; their decisions do not depend on what the crowd does. Their presence can make others less likely to join. The parallel with markets is direct: investors who refuse to buy merely because other people are buying help prevent a bubble from inflating. The more independent judgments remain in circulation, the less likely the market is to become a chorus of imitation.

Dependence Beyond the Market

Herding is one risk within the book’s wider account of how groups coordinate. Distributed knowledge can be valuable, but it is not enough for individuals to possess useful information if a system cannot bring it together. Conversely, centralizing decisions can impair the independence and specialization that allow people close to a problem to contribute what they know. The challenge is to make local knowledge collectively useful without eliminating its independence.

The same tension appears in organizational and public cooperation. A television-industry rating system may be costly to replace even if networks, stations, and advertisers would collectively benefit from better information: no single participant gains enough to justify paying for the change, while each can wait for others to act. Public goods create a related incentive to free ride, since people can benefit even when they do not contribute. In these cases, the obstacle is not simply a crowd following a trend; it is a system in which individual incentives can frustrate a collectively beneficial outcome.

The distinction matters because the book does not equate all group behavior with irrationality. Decentralized workers can coordinate without constant orders, and markets can combine independent and dependent decisions without becoming bubbles. Herd behavior names the point at which dependence overwhelms that balance. Its opposite is not isolation from other people, but enough independent judgment to keep shared expectations from becoming self-confirming and extreme.