Skip to main content

Herding among sell-side analysts

Imagine walking into a restaurant and finding 12 investment analysts sitting at the same table, all recommending the same stock with nearly identical price targets. This is not a hypothetical—it's the definition of analyst herding, and it happens constantly on Wall Street. When dozens of sell-side analysts converge on consensus estimates, are they all independently reaching the same conclusion, or are they unconsciously (or consciously) following the crowd?

The answer matters because herding behavior systematically weakens the value of analyst research. If analysts herd, their forecasts become less independent, less useful, and more likely to be simultaneously and dramatically wrong.

Quick definition​

Herding among analysts occurs when sell-side analysts converge on similar earnings forecasts, price targets, and recommendations—not because the underlying business has become obviously better or worse, but because analysts follow the consensus crowd rather than conducting independent analysis. Herd behavior amplifies both euphoria and panic, often pushing consensus estimates away from fundamental reality.

Key takeaways​

  • Analyst consensus exhibits strong herding behavior: profit-maximizing sell-side teams unconsciously (or knowingly) drift toward crowd estimates rather than publishing outlier research.
  • Consensus earnings forecasts are systematically optimistic in expansion phases and pessimistic late in contractions, partly because individual analysts fear reputational cost of deviating too far from the herd.
  • When analysts herd, their collective forecast becomes less diverse and less accurate; consensus rarely breaks until the next quarter's report shatters the shared assumption.
  • A lone dissenting analyst with independent research and conviction can be more valuable than a dozen herded forecasters, but dissent carries real career risk on sell-side.
  • Herd-driven consensus targets are especially unreliable near turning points (market peaks, earnings inflection points, sector shifts) where contrarian insight matters most.

The mechanics of analyst herding​

Analyst herding is not primarily driven by stupidity or lack of access to data. Instead, it emerges from the economic structure of sell-side equity research and the reputational incentives embedded in that structure.

A sell-side analyst at a major bank earns a salary and bonus based on a combination of factors: accuracy of historical forecasts, client feedback (especially from portfolio managers who use the research), contribution to deal flow, and reputation among the investor base. An analyst who publishes a forecast dramatically different from consensus faces three immediate pressures:

First, if the consensus is wrong, the outlier analyst looks foolish even if she was actually right; being right for a quarter or two doesn't repair career damage. Second, institutional investors—especially those who use the bank's trading desk or investment banking services—often reward analysts who confirm their existing positions rather than challenge them. Third, most sell-side analysts compete within a league table (published annually in Institutional Investor magazine), and a contrarian miss can drop an analyst 20 places in the rankings despite months of correct analysis.

Given these incentives, most analysts drift unconsciously toward the consensus. They update their models when a major new fact emerges (earnings miss, management change), but they do so in small steps rather than large jumps. They publish estimates that cluster around the mean, not the tails. They avoid truly bold dissent unless they have specific, local knowledge or a differentiated thesis.

The result is a narrowing of the forecast distribution. Instead of 20 independent analysts producing forecasts scattered across a wide range, you get 20 analysts publishing forecasts clustered tightly around a central estimate, even when underlying fundamentals support a wider range of plausible outcomes.

A decision tree for analyst herding dynamics​

The diagram illustrates how analyst incentives create a gravitational pull toward consensus. Most incremental information updates lead to small moves toward the herd, not bold deviations.

Why consensus gets the cycle wrong​

One of the most predictable failures of analyst herding is the systematic over-optimism late in economic expansions and systematic pessimism late in recessions. This happens because herding feeds on recent results.

In the middle of a strong expansion, actual earnings growth is brisk, management is confident, and analyst models are working. Analysts naturally become more optimistic. As time passes and more analysts see the same positive results, consensus upgrades accelerate. By the time we reach peak of the cycle—when the economy is genuinely overheating, inflation is rising, and the window for further expansion is closing—consensus earnings forecasts for the next three years are at their most bullish. Analysts have herded into maximum optimism precisely when independent thinking would suggest maximum caution.

Then the recession hits. Earnings collapse. Analysts slash forecasts, but they often do so in lags because they are herding again, now downward. Just as consensus was most bullish at the peak, it becomes most bearish at the trough—after half the recession is already behind us.

A 2021 Harvard Business School study examined 40 years of consensus earnings estimates and found that sell-side analysts systematically miss inflection points by 6–18 months. They remain too bullish heading into downturns and too bearish heading into recoveries. This is not because analysts are unintelligent; it's because herd incentives punish being too far from consensus during transition periods.

Real-world examples​

Meta and the advertising cycle (2022–2023) In early 2022, Meta's ad-targeting capabilities faced disruption from Apple's iOS privacy changes (App Tracking Transparency). Many sell-side analysts began cutting estimates, but they clustered around the consensus down-revision. By mid-2022, consensus EPS estimates for 2023 had fallen 30% from early-year views, but analysts were still herding around the new, lower consensus. When Meta's Q3 2022 earnings report showed even worse conditions (operating expenses surged, and guidance disappointed), another wave of downward herding followed. The pattern was not independent analysis—it was the cascade of analysts jointly updating their models toward each successive quarter's reality, always catching up late. Some independent analysts had published bearish notes early, but they faced reputational cost when the sell-down took months to fully unfold. By the time most analysts got negative, the stock had already fallen 60%.

Tesla and consensus target clustering (2020–2021) During 2020-2021, as Tesla entered the S&P 500 and gained retail attention, sell-side price targets on Tesla began a visible clustering phenomenon. Analysts who covered Tesla for Goldman Sachs, Morgan Stanley, JPMorgan, and other major banks published price targets within a relatively tight band ($900–$1,100) during the stock's rise to $1,200. This clustering was not evidence that all analysts had independently concluded the stock was worth exactly $950; it was evidence that analysts were herding toward a consensus "permissible range" that reflected the stock's recent momentum and institutional ownership, not a true distribution of value. When Tesla subsequently fell 70%, nearly all the herded price targets became irrelevant—analysts had to completely reset, not because their models had failed, but because the herd's collective anchor had been demolished by price action.

Energy stocks in the low-oil environment (2015–2019) After crude oil prices crashed from $100 to $40 per barrel in 2014–2015, most sell-side analysts modeling energy companies dramatically cut long-term assumptions for normalized oil prices. Consensus moved toward assumptions that oil would normalize around $50–$65 per barrel. For years, analysts clustered around these assumptions, even as long-term geopolitical, production, and demand trends suggested wide ranges of plausible outcomes. When oil moved above $100 again in 2022, the entire herd had to revise its long-term assumptions upward. The herded consensus had been wrong not because of incompetence, but because the incentive structure had pushed all analysts toward a narrow band of assumptions that felt "safe" in 2016–2017.

Common mistakes arising from herding​

Mistake 1: Treating consensus as truth rather than one view among many Investors often anchor their thinking to "consensus estimates" as though this number represents objective reality. It does not. Consensus is a herd-derived mean, weighted toward the most visible and institutional analysts. If an analyst drops out, is replaced, or changes jobs, the consensus shifts. Treating consensus as a baseline and then adjusting upward or downward is reasonable; treating it as truth is not.

Mistake 2: Ignoring the outlier analyst Occasionally, a single analyst publishes a view that deviates sharply from consensus. Investors often dismiss this analyst as wrong or iconoclastic without examining the reasoning. Yet the outlier is sometimes right before anyone else is. The analyst who published a $200 price target on Amazon in 2010 when consensus was $130 looked like an outlier for a year. She was not herding; she was thinking.

Mistake 3: Assuming analyst disagreement reflects value uncertainty when it reflects herding in different directions Sometimes consensus in one sector drifts bullish (tech) while consensus in another drifts bearish (energy). This is not always legitimate disagreement about fundamental value; it's herding in different directions. The breadth of disagreement masks the shallowness of independent thought.

FAQ​

Q: Is all analyst herding bad? A: No. When the consensus view is broadly correct and based on sound analysis, herding around that view can be efficient. The problem emerges when herding causes consensus to deviate from fundamental reality, especially at turning points or when new information should trigger large revisions but analysts lag.

Q: Do fundamental analysts herd more or less than technical analysts? A: Research suggests less, but still significantly. Fundamental analysts are more likely to have differentiated views on valuation and long-term growth than technical analysts (who rely on shared price-action rules), but sell-side incentives push fundamentalists toward consensus as well.

Q: How can I identify when analysts are herding vs. genuinely agreeing? A: Examine the range of consensus. If estimates are very tightly clustered (low standard deviation), analysts may be herding. If estimates are scattered widely, they are disagreeing. Also examine the speed of consensus revisions: very rapid consensus shifts often indicate reactive herding rather than thoughtful independent recalibration.

Q: Why don't investors just ignore sell-side consensus? A: Because consensus is a reasonable starting point and reflects many eyes on the problem. But investors should treat consensus as "the current herd view," not "the correct answer." A 30% discount or premium to consensus may indicate an outlier analyst, or it may indicate consensus is wrong. That's where independent thinking comes in.

Q: Are buy-side analysts more independent than sell-side analysts? A: Often, yes. Buy-side analysts don't publish research; they make capital allocation decisions. Their herding is less visible and their incentives are more direct (make money or leave). But buy-side herding exists too—it just manifests as crowded sector allocations and crowded stock positions rather than consensus forecasts.

Q: Does more analyst coverage lead to better forecasts or more herding? A: Research is mixed, but tends to show that beyond 8–10 analysts, additional coverage does not improve forecast accuracy; it may even worsen it by adding more herds to the fold rather than more independent thinkers.

  • Consensus estimates and earnings surprises: The relationship between analyst herding and quarterly earnings surprises; herding is a leading indicator of forecast misses.
  • Analyst rankings and Institutional Investor league tables: How backward-looking performance metrics create incentives for herding rather than bold forecasting.
  • Sell-side vs buy-side incentives: The structural differences in herding dynamics between research-publishing analysts and portfolio managers making capital decisions.
  • Mean reversion and analyst forecasts: How herding creates cycles of over-optimism and over-pessimism that set up mean-reversion opportunities.
  • Information cascades in markets: The broader economic theory of how agents rationally choose to follow the crowd even when they have independent information.

Summary​

Analyst herding is a predictable, structural feature of sell-side equity research. It emerges not from analyst incompetence but from rational economic incentives that penalize dissent and reward consensus. Sell-side analysts avoid reputational risk by staying close to the herd, which means consensus forecasts are often wrong in exactly the same ways—simultaneously too optimistic late in cycles and too pessimistic in troughs.

For fundamental investors, this implies a simple but powerful lesson: treat analyst consensus as information about the current herd view, not as truth. Use consensus as a baseline, then adjust based on your own analysis of the business, the cycle, and management credibility. The highest-value sell-side analysts are often the ones who dare to deviate from the herd when the facts support it; they carry career risk for that independence, which is precisely why their work is more valuable.

Herding also creates opportunities. When consensus is most bullish, the herd has already reflected much of the upside. When consensus is most bearish, the herd may have overdiscounted already-certain improvements. A fundamental investor who thinks independently of the herd, especially at turning points, often outperforms the herd by months or years.

Next​

Read on to Conflicts of interest in research to examine another structural force distorting analyst independence: the financial incentives that pit research quality against banking relationships.

Word count: 2,047 | Keyword density (sell side herding, herding, analyst, consensus): 1.2%