Introduction
Imagine you're walking down an unfamiliar street, looking for somewhere to eat dinner. You pass two restaurants. One has a line out the door, excited chatter spilling onto the sidewalk. The other sits empty, a bored host scrolling through her phone.
You know nothing else about either place. No reviews, no recommendations, no menu information.
What do you do?
Most people join the line. And that decision—entirely reasonable, entirely rational—contains the seed of every market bubble, every panic, every episode of collective madness in financial history.
The line contains information. Those people chose this restaurant for some reason. Maybe they know something. Maybe the food is exceptional. Following their lead isn't stupidity—it's intelligent social learning. Humans have survived and thrived by observing others and learning from their choices.
But here's the trap: if everyone makes this calculation, the line becomes self-perpetuating regardless of food quality. The third person in line didn't taste the food—they just followed persons one and two. The tenth person followed the first nine. By the hundredth customer, the line has become entirely self-referential. No one is evaluating the restaurant anymore; they're evaluating the line.
This is herding. It's rational at the individual level. It's often catastrophic at the collective level. And understanding this paradox is essential for any investor who hopes to navigate markets successfully.
The Three Mechanisms of Herding
Herding isn't a single phenomenon. It emerges from three distinct mechanisms, each operating through different logic but producing the same result: crowds moving together, often in the wrong direction.
Informational Herding is the restaurant-line phenomenon. When others seem to have information you lack, following them is rational. If Warren Buffett buys a stock, his action reveals something about his analysis. Ignoring that signal would be foolish.
The problem arises when everyone follows the signal without verifying the underlying information. The signal becomes detached from its source. People follow people following people, and eventually no one is looking at the actual investment anymore—they're just watching each other.
Reputational Herding operates through a different logic entirely. Consider a professional fund manager choosing between two investments. Investment A is a contrarian bet—if it works, she looks brilliant; if it fails, she looks reckless. Investment B is what everyone else is buying—if it works, she matches her peers; if it fails, she fails with the crowd.
The asymmetry is stark. Being wrong alone threatens her career. Being wrong with everyone else is forgivable—"no one could have predicted it." Given this calculus, following the herd isn't just psychologically comforting; it's professionally rational.
This explains a finding that puzzled economists for years: professional fund managers herd more than retail investors, not less. David Scharfstein and Jeremy Stein documented this in their landmark 1990 paper. The smart money isn't immune to herding—in many cases, the smart money is the most susceptible because their careers depend on relative performance.
Payoff Herding completes the picture. In markets, other people's actions literally change your payoffs. If enough investors buy a stock, the price rises, which makes existing holders wealthier, which attracts more buyers. Your decision to buy isn't just following others—it's creating conditions where following you becomes profitable.
This creates momentum: past returns predict future returns, at least for a while. Momentum trading—buying what's going up—works precisely because it's self-fulfilling. Of course, it works until it doesn't, and the reversal can be violent.
Information Cascades: When Private Knowledge Gets Ignored
The most insidious form of herding happens when rational individuals ignore their own private information to follow the crowd. This is the information cascade, and it can lead entire markets astray.
Here's how cascades work in practice.
Suppose a company announces a new product. Investor A has access to industry sources suggesting the product will succeed. She buys the stock. Investor B sees A buying, and also has some positive information—maybe less reliable, but still positive. He buys too.
Now Investor C comes along. He has private information suggesting the product might fail—perhaps a contact at a competing company who says the technology isn't as good as advertised. But he also observes that A and B bought.
Rationally, C faces a calculation. Two informed investors saw something positive enough to buy. His own single negative signal might be wrong. On balance, the weight of evidence—two positive signals from others versus one negative signal of his own—favors buying. So he does.
The cascade has begun.
Investor D observes three buyers. Even if D has negative information, the apparent consensus overwhelms it. D buys. Then E, then F, then hundreds of others. Private information gets ignored in favor of the increasingly strong signal from the crowd.
The terrifying part: the cascade can be completely wrong. If A's positive signal was mistaken, the entire edifice rests on a false foundation. B followed A's error. C and D followed B's following of A's error. By the time the truth emerges—when the product actually fails—the stock has been bid up far beyond any reasonable valuation, and the collapse is spectacular.
This isn't theoretical. Economists Sushil Bikhchandani, David Hirshleifer, and Ivo Welch formalized cascade theory in their 1992 paper, showing mathematically how rational behavior can lead to fragile consensus that shatters when new information arrives.
Case Study: The Dot-Com Bubble
The dot-com bubble of the late 1990s offers a textbook illustration of all three herding mechanisms operating simultaneously.
Informational herding drove the early stages. When Netscape went public in August 1995 and more than doubled on its first day of trading, investors took note. Here was information: this internet thing might be big. When Amazon, Yahoo, and eBay followed with spectacular IPO performances, the signal grew louder. Each successful investment revealed apparent knowledge about the future of technology.
Reputational herding amplified the mania among professionals. Fund managers who avoided internet stocks in 1998 underperformed badly. Their clients fled to competitors who rode the wave. Career self-preservation demanded participation, even for managers who privately doubted the valuations. Julian Robertson, the legendary value investor who ran Tiger Management, saw his assets shrink from $23 billion to $6 billion as clients abandoned his fund for tech-heavy alternatives. He closed Tiger in March 2000—essentially the market peak—having been right about valuations but destroyed by herding dynamics.
Payoff herding made the bubble self-reinforcing. Each purchase pushed prices higher, which enriched early buyers, which attracted more buyers, which pushed prices higher still. Momentum strategies—buying what had gone up—generated enormous returns through 1999. Not following momentum wasn't just psychologically difficult; it was financially punishing.
The cascade eventually broke. It always does. When companies started reporting actual earnings (or lack thereof), when cash burn rates became unsustainable, when the disconnect between prices and reality became too grotesque to ignore, the cascade reversed.
And here's the crucial point: the reversal was just as violent as the advance. People didn't gradually reassess their views. They herded in the opposite direction. The same reputational dynamics that punished skeptics in 1999 punished believers in 2001. Information cascades reversed: selling begat selling, which revealed that others were selling, which convinced still more to sell.
The NASDAQ lost 78% from peak to trough. Five trillion dollars of paper wealth evaporated. The very stocks that everyone "knew" would change the world became symbols of collective folly.
Case Study: GameStop and the Meme Stock Phenomenon
Twenty years later, a different kind of herd emerged—one that turned the traditional herding dynamic on its head.
In January 2021, shares of GameStop, a struggling video game retailer, rose from around $20 to nearly $500 in a matter of weeks. The catalyst was a community of retail traders on Reddit's WallStreetBets forum who identified that hedge funds held massive short positions in the stock.
The institutional short position was itself a form of herding. So many hedge funds had piled into the same bearish bet—reputational herding made shorting GameStop feel safe—that they'd sold short more shares than actually existed. When retail traders coordinated a buying campaign, the shorts faced catastrophic losses and were forced to buy shares at any price to cover their positions.
GameStop wasn't about fundamental value. It was a collision between two herds: institutional money herding into shorts versus retail money herding into longs. The retail herd won—at least temporarily—because their coordination was tighter and their social pressure ("diamond hands") more intense.
But notice: both sides were herding. The retail traders were following each other just as surely as the hedge funds had followed each other into short positions. The difference was timing and direction, not the underlying dynamic.
Many retail traders who bought GameStop at $300 lost money when the price eventually collapsed. They'd joined the line without evaluating the restaurant. The information cascade that drove the stock up eventually reversed, as all cascades do.
How Herding Creates Technical Patterns
Here's where crowd psychology meets price charts. The patterns that technical analysts identify—trends, support levels, resistance levels, breakouts—aren't mystical. They're the footprints of herding behavior.
Trends emerge because herding creates momentum. Buying attracts more buying; selling attracts more selling. This isn't random—it's the logical consequence of informational and payoff herding operating together.
Support and resistance levels form because of psychological anchoring. Investors who bought at $50 watch that level obsessively. If the stock falls below and then approaches $50 again, they're eager to get out even, creating selling pressure. That resistance is the collective memory of past decisions.
Breakouts are violent because they represent cascade restarts. When price finally pushes through resistance, all those anchored sellers are proven wrong simultaneously. The cascade that had formed against them reverses, and new buyers pile in.
Reversals are sharp because herds don't gradually disperse. They stampede. When the cascade breaks, everyone tries to exit at once, and the resulting price action is violent and rapid.
Understanding herding doesn't make these patterns disappear. But it explains why they exist, which is far more valuable than simply memorizing shapes on a chart.
Protecting Yourself from the Herd
If herding is so pervasive and so often destructive, what can an individual investor do?
The answer isn't to become a reflexive contrarian. The crowd is right more often than it's wrong—that's why Galton's ox-guessers were so accurate. Betting against the herd every time is just as mindless as following it every time.
Instead, the goal is to develop awareness—to recognize when you're following the crowd because it might know something, versus following the crowd because everyone else is following the crowd.
Some practical approaches:
Pre-commit to your analysis. Before you see what others are doing, form your own view. Write it down. Then—and only then—check the consensus. This prevents your private information from being overwhelmed by social signals before you've even evaluated it.
Identify your herding triggers. When do you feel most compelled to follow? Usually it's during periods of strong momentum, when it feels painful to miss out. Recognizing these emotional states—and treating them as warning signs rather than buy signals—is a skill that can be developed.
Diversify your information sources. Cascades form when everyone watches the same signals. If your research process looks identical to everyone else's—reading the same analysts, watching the same CNBC segments, following the same Twitter accounts—you're maximizing your exposure to herding dynamics. Seek out variant perspectives.
Maintain a long time horizon. Herding is most destructive in short timeframes. The same stock that swings wildly day to day tends to reflect fundamentals over years. By extending your holding period, you reduce the damage that short-term herding can inflict on your returns.
Conclusion: The Paradox of Social Intelligence
Herding is deeply human. It's how we survived as social animals, learned from each other, and built civilizations. The instinct to follow others isn't a flaw—it's a feature that has served our species extraordinarily well.
But in financial markets, this feature can become a bug. The same social learning that helps us navigate unfamiliar cities leads us astray when applied to asset prices. The same reputational concerns that maintain professional standards also create correlated errors among supposed experts.
Understanding this dynamic is the first step toward investing with independence. Not rebellious contrarianism—that's just another kind of herd behavior, just a smaller herd. But genuine independence: the ability to use social information intelligently while maintaining the discipline to trust your own analysis when it matters.
In our next part, we'll explore prospect theory—how humans process gains and losses asymmetrically—and how this psychological quirk creates the support and resistance levels that herding makes meaningful.
Key Insights
- Herding is individually rational but collectively dangerous. Following others makes sense when they might know something you don't. The problem arises when everyone follows everyone.
- Professional investors herd more than amateurs. Career risk creates reputational herding that can overwhelm analytical judgment.
- Information cascades can be completely wrong. When early actors are mistaken, everyone follows a bad signal, and private information never corrects the error.
- Technical patterns are the footprints of herding. Trends, support, resistance, and breakouts aren't mystical—they're what herding behavior looks like on a chart.
- Independence is a skill, not a personality trait. It can be developed through deliberate practice: pre-committing to analysis, recognizing emotional triggers, diversifying information sources, and extending time horizons.
Sources
- Bikhchandani, Sushil, David Hirshleifer, and Ivo Welch. "A Theory of Fads, Fashion, Custom, and Cultural Change as Informational Cascades." Journal of Political Economy, 1992.
- Scharfstein, David S., and Jeremy C. Stein. "Herd Behavior and Investment." American Economic Review, 1990.
- Lefèvre, Edwin. Reminiscences of a Stock Operator. 1923.
- Surowiecki, James. The Wisdom of Crowds. Doubleday, 2004.