Anchoring Bias in Stock Valuation — How Your Entry Price Distorts Judgment
You bought Apple at $180. Today it trades at $220. Your brain says: “expensive”. You bought Bayer at $95. Today it trades at $25. Your brain says: “it has to come back”. Both judgments are biased — they measure today’s price against a random anchor (your purchase price) instead of intrinsic value. Here are the mechanics, five common traps and a checklist that switches the anchor off.
What anchoring is
Tversky/Kahneman 1974 showed: people pull numerical estimates unconsciously toward a random reference. In the classic experiment they spun a wheel of fortune with numbers 1-100 and then asked for the share of African UN member states. Those who saw a higher wheel number estimated significantly higher — even though the number had nothing to do with the question.
In stocks your most common anchor is the purchase price. Other popular anchors: the 52-week high, a psychological round price ($100, $1,000), a pre-crisis price level (e.g. “the COVID high”) or the all-time high.
The anchor pulls your judgment toward itself — even when you know it is irrelevant. Studies show distortions of 30-60% of anchor distance. In stocks: with a $100 cost basis and $70 current price, subjects estimate fair value about $10-18 higher than uninvolved observers do.
Five anchor traps that hit every investor
- The cost-basis anchor: “I sell once I’m back at break-even.” Cost basis is tax information — not valuation information. The market does not know what you paid and does not behave accordingly.
- The 52-week-high anchor: “It’s 30% off the high, so it’s cheap.” Depending on the cycle, the old high is irrelevant. Tech 2021 was an all-time high for many names — and four years on, still not reclaimed.
- The round-number anchor: $100, $500, $1,000 act like thresholds. But Berkshire B at $460 and at $540 was at the same valuation point — splits move the anchor, not the value.
- The pre-crash anchor: “Before COVID it was 30% higher.” The March 2020 price has no link to today — earnings differ, rates differ, the business model differs. The COVID-era price is one data point, not a benchmark.
- The analyst-target anchor: “Goldman sees $250.” Sell-side targets are continually revised toward consensus and on average too high. Anyone using the target as an anchor systematically holds too long.
How much does the anchor cost you?
| Study / source | Finding | Return drag |
|---|---|---|
| Genesove/Mayer 2001 (housing) | Sellers anchored on cost basis realise lower prices and hold 9 months longer | 3–5% on sale price |
| Mussweiler/Strack 2000 | Even pros (lawyers, traders) show anchoring — effect ~70% of laypeople | 1–2% p.a. on trading returns |
| Jegadeesh 1990 (mean reversion) | Pure mean-reversion buys (just because 30% off the high) underperform market | 1–4% p.a. negative |
| BMI investor data 2024 | 76% of sell triggers are anchor-related (cost, high, round level), not thesis-related | Structural, hard to single-quantify |
The anti-anchor test in 5 questions
- Q1: Would I buy this stock today? Neutralises the cost-basis anchor. If no → sell, regardless of where you bought.
- Q2: Which fundamental metrics support the current price? P/E, EBITDA margin, cashflow, growth. Fair value is derived here — not from the chart.
- Q3: What are three comparable alternatives? If others offer a better risk-reward profile, anchor binding is irrelevant — you switch.
- Q4: What new information since purchase has changed the thesis? New earnings, new competitors, new regulation. If none → re-check anchor, not thesis. If yes → re-check thesis, drop anchor.
- Q5: If a friend gifted me this stock, would I keep it? Neutralises both anchor and endowment effect.
Example: BMW before and after the anchor reset
In anchor mode BMW feels “underwater”. In fair-value mode BMW is undervalued. Both can be true — but only the second is an investment decision.
Special case: round-number anchors at the threshold
Stocks around $100, $500 or $1,000 frequently see clear technical resistance — not because the round number means anything fundamental, but because many participants stack stops and take-profits there. Knowing the anchor lets you turn it into your own execution edge: limit buys a few cents below the round number fill more reliably than limits exactly at it (the crowded-order effect).
Pros & cons: using the anchor on purpose
- Other participants have anchors — you can anticipate their reactions
- Round-number levels give technical triggers
- 52-week lows often signal forced-selling phases
- Your own cost basis is always to be ignored — irrelevant to fair value
- Pre-crash highs are mostly obsolete because fundamentals shifted
- Sell-side targets are systematically too high and constantly moving
Rule of thumb: ignore your own anchors (cost basis, dream-entry price) and use external anchors (round-number triggers, 52-week range) only as technical signals. Not the other way round.
Common questions
Is “buy the dip” anchoring?
Partly yes. Buying just because a stock is 20% off its 52-week high is anchor-driven. Buying because valuation relative to current earnings is attractive is fundamentals-driven. The first loses money in falling knives, the second is repeatable.
How does a stock split change the anchor?
Strongly, even though it changes nothing fundamentally. Berkshire B at $470 feels “normal”, Berkshire A at $700,000 feels “unreachable” — same substance, different anchor. NVIDIA’s 10-for-1 split in 2024 made the round-number effect at $100 and $200 visible. Splits move anchors, not value.
Does DCF help against anchoring?
Yes, when inputs are current. Discounted-cash-flow forces you to work with cashflows, growth rates and discount rates — none of them anchors. Weakness: DCF is highly sensitive to assumptions, easily gamed. More robust: P/E vs peer median, EV/EBITDA spread, dividend yield vs bonds.
What about Bitcoin and round prices?
Crypto is heavily round-number driven. $10,000, $50,000, $100,000 each carry strong psychological weight — and correspondingly stacked stop-loss and take-profit orders. Trading limits a few percent below the round number optimises fill probability noticeably.
Does indexing help against anchoring?
Indirectly yes. With an ETF there is no per-stock cost-basis anchor — you buy market weights. But ETF investors still anchor on index levels (S&P 5,000 as a threshold, DAX 18,000). With broadly diversified savings plans the effect almost disappears — with active ETF rotation it returns.
Anchoring and tax optimisation?
Here the anchor genuinely helps — as cost logic, not as a behavioural bias. For tax-loss harvesting your cost basis IS the relevant number. But it stays a tax input, not a valuation input. Always separate: valuation decision (ignore anchor) vs tax optimisation (use anchor).
Related Hubs: Investor Glossary | Legendary Quotes
Try TradingView Free for 30 Days
Plus get a $15 discount on your first subscription through this link.