Herding Behaviour: Why We Follow the Crowd With Our Money — and Why It Costs Us

In the early 1950s, psychologist Solomon Asch ran a series of experiments on conformity that revealed something disturbing about how people respond to social consensus. Participants were shown a line and asked to identify which of three comparison lines matched its length. The answer was obvious — the lines differed clearly in length. But when confederates in the room unanimously gave the wrong answer, approximately 75% of participants conformed at least once, and 32% conformed consistently — giving answers they knew were wrong rather than contradict the apparent group consensus. When it comes to financial decisions, where the correct answer is far less obvious than line length, the pull toward group consensus is even stronger and far more costly.

Why We Herd: The Social Learning Logic

Herding behaviour in financial markets is not purely irrational — it has a rational information-processing component. When you observe that many other investors are buying a particular asset, their collective behaviour contains information: presumably they know something, or have collectively processed information in ways that justify their purchases. Following the crowd is a way of aggregating distributed information that you don’t individually possess. This social learning mechanism is genuinely useful in many contexts — observing that other people are avoiding a neighbourhood, a restaurant, or a product often contains reliable information about its quality or safety.

The problem in financial markets is that everyone is herding simultaneously, so no one is actually anchored to independent fundamental analysis — they’re all following each other in a circular fashion that produces collective price movements detached from underlying value. When enough investors buy a rising asset because other investors are buying it, the price rise itself becomes the justification for further buying — a self-reinforcing dynamic that can drive prices far above any reasonable fundamental value. This is the mechanism of speculative bubbles, and it operates because each individual investor’s decision to buy contains what looks like useful social information (others are buying) without revealing that those others are themselves herding rather than making independent fundamental assessments.

Herding in Retail Investor Behaviour

The most direct financial cost of herding for individual investors is performance chasing — buying assets after they’ve already risen significantly, driven by the social information that many other investors are buying them. Mutual fund flow data documents this pattern with exceptional consistency: money flows into equity funds at market peaks (when everyone is excited and buying) and out of equity funds at market troughs (when everyone is panicking and selling). The aggregate retail investor systematically buys high and sells low, not through any deliberate bad strategy, but through the herding mechanism that makes buying into rising assets feel prudent (everyone else is doing it) and selling falling assets feel urgent (everyone else is doing that too).

The meme stock phenomenon of 2021 — GameStop, AMC, and similar assets — was herding behaviour in a particularly pure form: assets whose prices rose not because of improving fundamentals but because retail investors were coordinating their purchases through social media, explicitly acting as a herd to generate price momentum. The social coordination was the mechanism, not any independent assessment of value. Participants who bought early and sold into the momentum profit; the majority who buy near the peak because the social signal is strongest exactly when the price is highest — and who subsequently sell in panic when the momentum reverses — experience the full downside of herding without the upside of early adoption.

How Financial Media Amplifies Herding

Financial media and social platforms serve as herding amplifiers — they broadcast the social signals of what other investors are buying and selling at a scale and speed that significantly exceeds what any individual could observe through their personal network. When CNBC covers a hot sector, when Twitter is full of posts about a particular stock, when podcasts are discussing a specific investment theme — these signals broadcast what the crowd is doing to millions of investors simultaneously, coordinating herding behaviour at market-wide scale. The media’s coverage is itself driven by what’s popular (because popular stories attract audiences), creating a feedback loop where trending investments attract media coverage that creates more investor awareness that generates more buying that produces more trending performance that attracts more coverage.

The timing of this media coverage relative to investment return tends to be precisely backwards for investors acting on it: media coverage of an investment theme peaks when the theme is widely discussed and therefore already priced in, not when it represents an undiscovered opportunity. By the time a sector, asset class, or individual stock is the subject of mainstream financial media attention, the investors who profited from the original trend identification have already captured most of the gain. The investors who buy in response to the media attention — following the herd at its maximum size — are buying at the highest prices from those early investors who are now selling.

Contrarianism: The Opposite Mistake

The awareness of herding behaviour can produce a different error: reflexive contrarianism, the tendency to assume that whatever the crowd is doing is wrong and to take the opposite position. This is also a mistake. Markets are efficient enough that simply betting against popular assets doesn’t reliably produce outperformance — the crowd is right more often than not about the direction of asset prices in the medium run, and contrarian strategies have their own systematic failure modes. The bubble can keep inflating long after it appears overvalued to any individual observer, wiping out premature short-sellers even when their eventual call proves correct.

The appropriate response to herding behaviour is not to systematically bet against the crowd but to systematically avoid following the crowd — maintaining a pre-committed investment strategy that doesn’t respond to social signals about what other investors are buying and selling. This is the deep logic of passive index investing: it explicitly abandons the attempt to infer information from other investors’ collective behaviour (which produces herding) and instead accepts the market’s aggregate price as the best available estimate of value, investing in that aggregate rather than trying to identify which specific corners of the market other investors are over- or under-valuing.

Building Herd Resistance Into Your Investment Process

The most effective defence against herding in personal investment is pre-commitment to a strategy that specifies investment decisions in advance rather than in response to market conditions and social signals. An investment policy statement — a written document specifying your target asset allocation, rebalancing rules, and the conditions under which you will and won’t change your strategy — provides a reference point that doesn’t move with market sentiment. When the crowd is buying enthusiastically and social pressure to join the herd is strongest, the policy statement provides an explicit counter-commitment: my strategy is to maintain my target allocation and rebalance when it drifts, not to add exposure to recently appreciated assets.

Reducing exposure to financial media and investment social platforms during market extremes — when herding pressure is highest in both directions — is a practical structural intervention that removes the social signals driving herding before they can influence decisions. The investor who doesn’t know that everyone else is panic-selling in a market crash is less likely to panic-sell themselves. The investor who doesn’t know that everyone is excitedly buying a particular sector is less likely to chase that sector at its peak. Information control — deliberately limiting exposure to the social signals that drive herding — is an underrated but effective component of maintaining contrarianism-agnostic investment discipline through market cycles.

The Institutional Evidence Against Herding

The evidence against herding is perhaps most compelling in studies of institutional investor behaviour, where the herding costs should be smallest — professional investors with research resources, analytical tools, and incentive structures theoretically aligned with performance. Yet institutional herding is well-documented: fund managers systematically increase exposure to recently popular sectors and reduce exposure to recently unpopular ones, driven by the same career risk logic (it’s safer to be wrong in a crowd than right alone) that drives individual investor herding. Institutions that herd into popular assets and out of unpopular ones experience the same buy-high, sell-low return drag as retail investors, if typically at smaller magnitude. The data on institutional herding is useful for individual investors because it confirms that herding pressure is not primarily a product of financial naivety — it’s a feature of social decision-making in environments of uncertainty that affects sophisticated professionals as well as novice individuals. The antidote is structural — pre-committed rules that don’t require resisting social pressure in the moment — not increased financial sophistication that somehow immunises against the pull of consensus.

The investor who maintains their allocation, ignores market noise, and continues buying consistently through both euphoric peaks and panicked troughs is not doing nothing — they’re actively resisting one of the most powerful social forces in financial markets. That resistance, systematically maintained through pre-committed rules rather than moment-by-moment willpower, is one of the most reliable competitive advantages available to patient long-term investors in markets dominated by herding behaviour.