Valuation Herding: When Everyone Uses the Same Metrics

Valuation herding occurs when investors collectively adopt the same valuation methodology and reach identical conclusions about what makes a stock expensive or cheap. Rather than disagreeing about intrinsic value—as rational investors should, given uncertainty—the herd converges on identical multiples, discount rates, and terminal growth assumptions. This consensus creates a feedback loop where valuations become self-reinforcing and disconnected from the underlying fundamentals that should drive them. When the herd consensus shifts, valuations collapse suddenly because there is no heterogeneous opinion to cushion the transition.

The power of valuation herding lies in its invisibility. Unlike sector herding, which is obvious from relative portfolio weights, or information cascades, which involve observable behavioral patterns, valuation herding is embedded in the models that investors use without questioning their assumptions. When 95% of analysts model a company's terminal growth rate at 3% and discount cash flows using an 8% cost of capital, the resulting valuation appears scientifically rigorous—not a herd consensus that will crumble when the herd consensus shifts to 2% terminal growth and a 10% cost of capital.

> Quick definition: Valuation herding is the convergence of investor opinion around identical valuation multiples, discount rates, and growth assumptions, creating price stability until the consensus assumptions shift, at which point valuations collapse rapidly because the herd has nowhere else to stand.

Key takeaways

• Valuation assumptions are herded, not objective: Price-to-earnings multiples, terminal growth rates, and risk premiums are not discovered through analysis; they are socially constructed through repeated conversations among analysts and money managers.
• Consensus assumptions cluster at round numbers: Terminal growth rates cluster around 2.5% and 3%; cost-of-capital assumptions cluster around 7% and 8%; price-to-earnings multiples cluster around 15x and 20x. These round numbers are herded consensus, not fundamental precision.
• Expanding multiples amplify returns during rallies: When the herd shifts from 15x to 20x earnings multiples, prices rise 33% without any change in earnings. This "multiple expansion" is the herd agreeing to overpay; it creates the largest returns and the largest losses.
• Valuation shifts are synchronized: When the herd consensus changes (typically triggered by a change in interest rates or growth expectations), valuations shift suddenly across the entire market, not gradually. There is no gradual transition from the old consensus to the new consensus.
• Dissenting valuations are dismissed as outliers: Investors with different valuation assumptions are dismissed as "wrong" or "not understanding the company" rather than treated as legitimate disagreement stemming from uncertainty.

How valuation models create herds

Investors value stocks using discounted cash flow (DCF) models, which require estimating three inputs: future cash flows, a discount rate (cost of capital), and a terminal value (the value of the company after the explicit forecast period). Each input requires a judgment call, and investors routinely default to the median consensus judgment rather than forming independent opinions.

A software company's cash flows depend on assumptions about subscription growth, churn rates, and operating margins. A typical analyst might forecast 30% growth for the next five years based on industry reports suggesting the software market is growing 25-30% annually. But this forecast is a herd consensus pulled from industry analyst reports, not an independent assessment of the company's competitive advantages and market share capture potential. If the median analyst predicts 30% growth, the analyst faces career risk by forecasting 15% growth, even if the latter is more defensible.

The discount rate (cost of capital) is even more herded. Most analysts use the weighted-average cost of capital (WACC), which incorporates the risk-free rate, equity risk premium, and company-specific risk premiums. When the Federal Reserve's risk-free rate is 5%, the equity risk premium is typically assumed to be 4-6%, making the overall cost of capital 9-11%. These are round numbers that the herd converges around. An analyst who uses a 13% cost of capital (defensible if the company is high-risk) will arrive at a valuation 20-30% lower than the consensus—creating a dissenting opinion that will be ignored until the consensus changes.

Terminal growth rate assumptions are the most herded input of all. Most analysts assume companies will grow at 2-3% in perpetuity—roughly the long-term GDP growth rate. But this assumption is extremely sensitive: a software company growing at 2.5% in perpetuity is worth 25-30% less than the same company growing at 3.5% in perpetuity. Small changes in a single assumption create large valuation gaps, yet the herd converges around the same terminal assumptions without debate.

Multiple expansion and valuation herds

Multiple expansion occurs when investors agree to pay higher multiples of earnings for the same company. If a tech company earns $1 per share and the herd assigns it a 20x multiple, the stock is valued at $20. If the herd shifts to a 25x multiple (common during bull markets), the stock rises to $25—a 25% increase without any change in earnings. This multiple expansion is pure valuation herding: the herd has simply agreed to overpay relative to the prior consensus.

During the 2010-2021 tech bull market, the herd shifted from 15x to 40x+ multiples for high-growth software companies. This was not irrational if revenue growth actually accelerated from 20% to 50% annually. But by 2021, many software companies were trading at 25x revenue (not earnings)—a multiple that had never been sustained in history. Paying 25x revenue assumes the company will eventually generate 25x the current revenue in annual profit, an assumption that requires infinite growth or a fundamental misunderstanding of business economics.

The multiple expansion created enormous returns for early investors. A software company trading at 5x revenue in 2015 and 25x revenue in 2021 generated a 5x stock return in six years. When the herd consensus shifted in 2022 (triggered by rising interest rates), the multiple compression was equally violent. The same company fell from $200 to $50 in twelve months—a 75% loss—as the herd moved from 25x revenue to 5x revenue. No earnings change occurred; only the consensus valuation multiple shifted.

The role of discount rates in valuation herding

Discount rates are perhaps the most underestimated driver of valuation herding. When the Federal Reserve raises interest rates, risk-free rates rise, and cost-of-capital estimates should increase. Higher discount rates mean future cash flows are worth less in present-value terms, creating downward pressure on valuations.

From 2009 to 2021, the Federal Reserve maintained rates near zero, creating a discount rate of approximately 4-5% for mature companies. Software companies with high growth prospects might have been valued using a 6-7% discount rate. By September 2022, the Federal Reserve had raised rates to 3-4%, creating a discount rate of approximately 7-8% for mature companies and 8-10% for high-growth companies. Higher discount rates mean software company valuations should be 30-40% lower—and they were. The Nasdaq-100 fell 33% from peak to trough in 2022, primarily due to discount rate expansion, not deteriorating fundamentals.

The herding dynamic emerges when the Fed signals a rate change. Analysts do not independently reassess their DCF models; they wait for the herd consensus to shift. Once the first major analyst revises their cost-of-capital assumption, others quickly follow. Within weeks, the consensus discount rate has moved from 6% to 8%, and valuations have collapsed 20-25%. The synchronized shift is the herd consensus updating simultaneously to a new baseline.

Valuation herding in bull and bear markets

Bull markets amplify valuation herding through multiple expansion and declining discount rate assumptions. As the economy accelerates, analysts reduce their risk premium assumptions (because "risks are declining"), lowering the cost of capital. Simultaneously, earnings growth accelerates, creating momentum that attracts new investors. The combination of lower discount rates and accelerating earnings growth creates a self-reinforcing cycle where valuations expand faster than fundamentals improve.

A company earning $2 per share at a 10% cost of capital (implying 3% growth in perpetuity) is worth $20 per share. If the economy accelerates and analysts reduce the cost of capital to 8%, the same company is now worth $25 per share—even though earnings are unchanged. Add accelerating earnings (now $2.50 per share) and the valuation reaches $31 per share. The valuation has increased 55% due to a combination of multiple expansion, lower discount rates, and earnings growth. Much of that 55% gain is not attributable to fundamental improvement but to changing valuation assumptions.

Bear markets reverse this process with extreme violence. As growth stalls, analysts increase their risk premium assumptions, raising the cost of capital. Simultaneously, earnings growth decelerates or turns negative. The combination of higher discount rates and decelerating earnings creates a self-reinforcing decline where valuations compress faster than fundamentals deteriorate.

Using the same company: if earnings decline to $1.50 per share and the cost of capital rises to 12%, the valuation falls to $12.50 per share. That is a 60% decline from the bull market peak—driven entirely by valuation assumption changes, with only modest deterioration in underlying earnings. When the herd consensus shifts on discount rates and growth assumptions simultaneously, valuation moves far exceed fundamental changes.

Terminal growth rate herding

Terminal growth rate assumptions reveal the mechanics of valuation herding most clearly. The long-term economic growth rate for developed economies is approximately 2-2.5% annually. Yet during bull markets, analysts routinely assume 3%, 3.5%, or even 4% terminal growth rates for companies operating in mature industries.

A software company growing 40% annually will not grow 40% forever; growth must eventually decelerate to the economy-wide growth rate. But how fast does it decelerate? In 10 years, in 20 years, in 5 years? And at what intermediate growth rates? These are the questions analysts must answer, and they routinely herd around identical assumptions without justification.

Real example: In 2020, Zoom Communications—a video conferencing software company—was valued based on assumptions of continued 40%+ growth for 5-10 years, followed by deceleration to 3% terminal growth. The company would grow from $2.6 billion in revenue (2020) to $10+ billion by 2025, capturing an implausibly large share of the global video conferencing market. The herd consensus was built on this assumption. When Zoom's growth actually slowed to 20% (not 40%) due to market saturation, the valuation consensus shifted overnight. The stock fell from $568 (2021) to $115 (2023), a 80% decline driven entirely by revised terminal growth assumptions.

Herding in valuation methodologies

Beyond the inputs to DCF models, the methodology itself herds. Value investing traditionally uses price-to-earnings (P/E) ratios or price-to-book (P/B) ratios. Growth investing uses price-to-sales (P/S) multiples or enterprise value-to-revenue multiples. During the 2017-2021 tech rally, growth investing became so dominant that analysts actively dismissed value metrics. A software company trading at 20x revenue with zero earnings was deemed "worth the premium" because traditional valuation metrics "do not apply to high-growth companies."

This is valuation herding with a methodology overlay: the herd decided that revenue multiples were the appropriate valuation framework and that earnings-based multiples were obsolete. When growth slowed and profitability became relevant again (2022-2023), the herd switched back to earnings-based valuation. The methodology shift was not a reasoned reassessment of what drives intrinsic value; it was a herd consensus shift about which valuation framework was "correct."

Contrarian valuation and herd timing

Intelligent investors who disagree with valuation consensus face extreme timing risk. A company valued at 25x revenue during a bull market may decline 70% over 18 months when the herd consensus shifts to 5x revenue. An investor who identifies the overvaluation 12 months early suffers a 50% paper loss before the valuation compression begins. The emotional difficulty of holding a losing position while being "right about the valuation" is often greater than the financial pain of being wrong.

Warren Buffett famously avoided the technology sector during the 1990s tech bubble because valuations were disconnected from fundamentals. He was "wrong" for five years (1996-2001) as tech continued to rally, underperforming the market significantly. When the bubble burst (2000-2002), his valuation discipline suddenly looked brilliant. The key insight is that valuation herding can persist far longer than any rational analysis predicts, creating prolonged underperformance for investors betting against the herd.

Real-world examples

The 2000 Tech Bubble: Software and internet companies were valued using earnings projections that assumed exponential growth forever. Companies with negative earnings were assigned valuations exceeding those of profitable mature companies. Pets.com, which burned through $300 million and declared bankruptcy in 2000, had a market cap exceeding $300 million at its peak. The valuation consensus assumed e-commerce would revolutionize retail and that "reach and brand" mattered more than profitability. When the consensus shifted (2000-2002), these valuations collapsed 90%+, wiping out billions in invested capital.

The 2020 Pandemic Rally: When COVID-19 hit, analysts immediately revised their long-term economic growth assumptions downward. Yet simultaneously, they raised their valuation multiples for tech and e-commerce companies, assuming the pandemic would create permanent shifts in consumer behavior (toward online shopping and remote work). The herd assigned 40x+ multiples to software companies. When those behavioral shifts proved partially temporary (2022-2023), the herd revised assumptions downward and valuations fell 60-75%.

The 2023 AI Mania: When ChatGPT launched in November 2022, the herd consensus shifted immediately to assume generative AI would transform nearly every industry. Companies with minimal AI revenue were valued as if they would derive 30-40% of revenue from AI products within five years. Nvidia—a chip company—was valued based on unlimited demand for AI chips, with analysts assuming the company would grow earnings 40%+ annually for the next decade. When growth proved more moderate (30% instead of 40%) and AI capital expenditure became a debate rather than a certainty, the herd consensus revised downward, and valuations fell 20-30%.

Common mistakes in valuation herding

Mistake 1: Assuming consensus valuations are correct. The most pervasive error in valuation herding is treating consensus analyst valuations as objective facts rather than herd consensus opinions. When 95% of analysts rate a stock as a "buy" at the current price, investors assume the valuation is justified. But the consensus is often driven by incentives (sell-side analysts face pressure to be positive on stocks their firms have investment banking relationships with) and herd behavior (each analyst replicates prior analysts' assumptions without independent analysis).

Mistake 2: Trusting analyst earnings revisions. Analysts are notoriously optimistic and revise earnings estimates upward more frequently than downward. During bull markets, the herd consensus is almost always to raise earnings estimates. When the market peaks and turns downward, analysts begin cutting estimates, but they lag the market by 6-12 months. Using analyst estimates to justify current valuations is circular reasoning: valuations are high because analysts are optimistic, and we trust the optimism because valuations are high.

Mistake 3: Ignoring small changes in discount rates. A 0.5% change in the cost of capital assumption creates a 10-15% change in intrinsic valuation. Yet investors often treat discount rate assumptions as fixed and unchanging. When the risk-free rate moves from 2% to 3%, the cost of capital should move accordingly. Failing to update discount rate assumptions means your valuation models become obsolete immediately.

Mistake 4: Accepting round-number multiples without justification. When a company is valued at 20x earnings or 3x revenue, these multiples have been chosen because they are round numbers that the herd consensus has converged around. Challenge the assumption. What would the company be worth at 15x earnings? At 25x earnings? How sensitive is the valuation to this assumption? Creating a valuation sensitivity table reveals how much of your investment thesis depends on a particular multiple that may or may not persist.

Mistake 5: Extrapolating recent growth forever. Analysts routinely take the company's most recent three years of growth and assume it will continue, with only gradual deceleration. But growth rates are mean-reverting. A company growing 50% annually for three years has a 60% probability of slowing to 15-20% growth in the subsequent three years. Assuming 30% growth perpetually is herding around an implicit assumption that the company is fundamentally different from every mature business ever created.

FAQ

How can I identify when valuation consensus is becoming extreme?

Compare the current valuation to historical medians. If a sector is trading at 25x earnings and its 20-year median is 15x, the consensus is likely extreme. Look for analyst estimates of earnings growth: if the consensus assumes earnings will grow 30% annually for 10 years, the assumption is probably too optimistic. Finally, check the number of "sell" ratings among analysts; when sell ratings disappear (less than 1% of analysts), the herd consensus is at maximum.

Should I always buy when the valuation consensus shifts?

No. Valuation shifts are often predictive of further deterioration. A sector at 15x earnings might decline to 10x as the consensus continues to revise downward. Waiting for the consensus to stabilize (typically 2-3 months after the initial shift) before investing avoids catching falling knives. The herd consensus can persist at depressed levels for years before recovery occurs.

How do I defend against my own valuation herding?

Actively challenge your valuation assumptions. Write down your terminal growth assumption, cost of capital, and cash flow estimates. Now shift each by 0.5-1% and recalculate the valuation. Does the stock still look attractive if terminal growth is 2% instead of 3%? If cost of capital is 9% instead of 8%? This sensitivity analysis reveals how much of your thesis depends on assumptions that may prove wrong.

Why do analysts herd around identical valuations?

Analysts herd because they face career risk if their valuations diverge significantly from consensus. If your valuation is 30% below consensus and the stock rallies 20%, you look foolish. If your valuation is 30% above consensus and the stock falls 20%, you also look foolish. Career risk creates pressure to stay within 10% of consensus, even if independent analysis suggests a different valuation is justified.

Can machine learning or artificial intelligence avoid valuation herding?

Not inherently. Machine learning models trained on historical data will incorporate the same assumptions embedded in that data. If the historical data reflects periods when the herd was wrong (bubbles and crashes), the model will learn those patterns and replicate them. AI models can identify when valuation assumptions contradict historical patterns, but they do not solve the fundamental problem: intrinsic value is uncertain, and investors will herd around whatever assumption feels most reasonable at any given moment.

What is the relationship between valuation herding and price momentum?

Valuation herding and price momentum are complementary. Momentum creates pressure on valuations: as prices rise, valuations inflate to justify the price rises. Herding on valuation assumptions then validates the price momentum, creating a self-reinforcing cycle. Momentum breaks when the valuation consensus shifts (often triggered by a change in interest rates or earnings growth), causing synchronized capitulation.

Related concepts

• Sector Concentration Herding
• Information Cascades
• How to Detect Herding Behavior
• Confirmation Bias in Financial Analysis

Summary

Valuation herding is the convergence of investor opinion around identical valuation multiples, discount rates, and growth assumptions. These assumptions appear scientifically rigorous but are socially constructed through repeated interactions among analysts and investors. Bullish herds expand multiples and lower discount rates, creating 30-50% increases in valuation without corresponding earnings growth. Bearish herds compress multiples and raise discount rates, creating 40-70% decreases in valuation without corresponding earnings deterioration. The herd consensus is most extreme when analyst disagreement disappears and all assumptions cluster around consensus values. Investors who identify overvaluation early face significant timing risk; the herd can persist in wrong valuations for years. The most successful approach is to identify when valuation assumptions are clustering at extremes and position accordingly, accepting that the herd may continue to defy gravity longer than rational analysis predicts.

Next

→ Information Cascades