What it is
Anchoring is the cognitive bias in which an initial piece of numerical information — the anchor — has a disproportionate influence on subsequent estimates and judgments, even when the anchor is arbitrary or irrelevant. People adjust from the anchor, but they do not adjust enough. The final estimate is systematically pulled toward the starting point.
The bias does not require the anchor to be plausible or credible. Random numbers work almost as well as informative ones. The anchor does not even need to be in the same domain as the estimate.
The classic example
Kahneman and Tversky ran an experiment where subjects spun a wheel of fortune rigged to stop at either 10 or 65. They then asked: "What percentage of African countries are in the United Nations?" Subjects who saw the wheel land on 65 gave estimates averaging 45%. Subjects who saw 10 gave estimates averaging 25%. The correct answer is around 68%. The randomly generated number — clearly irrelevant to the question — shifted estimates by 20 percentage points.
In a more applied study, experienced judges were asked to roll a pair of dice before determining a sentence for shoplifting. Judges who rolled high (and knew it was random) handed down longer sentences than those who rolled low. The anchor distorted expert professional judgment even under full awareness that it was meaningless.
In markets and analysis
Analyst price targets are one of the most consequential expressions of anchoring in finance. When a stock trades at $100, a price target of $120 feels bold. A price target of $200 feels extreme. The current price anchors the range of "reasonable" targets, regardless of what the fundamentals actually imply.
This is why sell-side price targets cluster near the current price. Not because analysts believe prices are already right, but because the current price provides the anchor from which adjustments are made. Studies show that when stocks move substantially, targets follow — adjusted in the direction of price, but not enough to maintain any consistent relationship to fundamentals.
In M&A, the acquisition premium relative to the prior-week price is the anchor for deal discussions. Both sides know this anchor is arbitrary — the stock could have been at a different price last week for reasons unrelated to value — but it structures the negotiation anyway. The anchoring is so powerful that companies time large transactions to follow periods of depressed stock prices to reduce the nominal premium.
In personal finance, people anchor to purchase prices. An investor who bought a stock at $100 and sees it fall to $60 feels the loss relative to $100 — not relative to its current fair value. The purchase price has no future relevance, but it anchors the emotional experience and often the decision of whether to sell.
The right way to think about it
The corrective is to generate an independent estimate before seeing the anchor. In negotiation, this means researching and forming your own view of fair value before the other party names a price. In financial analysis, it means building a model from fundamentals before looking at where the market prices the security.
Once the anchor is visible, it is extremely difficult to ignore — the research suggests awareness of anchoring does not fully eliminate its effect. The best defense is structural: create your own estimate first, write it down, commit to it, and only then engage with the anchor. The written commitment creates a second anchor that partially counteracts the first.
Where it shows up
In real estate: the listing price anchors the buyer's sense of value. In salary negotiations: the first offer determines the range. In fundraising: the lead investor's valuation anchors subsequent rounds. In legal settlements: the initial demand sets the range for eventual resolution.
Each domain has the same structure. The entity that sets the first number has a significant advantage, because the other party adjusts from that number rather than from their own independent estimate.
One thing most people get wrong
De-anchoring is not achieved by reminding yourself that the anchor is arbitrary. Research shows that subjects who are explicitly told "the next number you see is random and should be ignored" still anchor on it. The bias operates below the level where rational self-awareness corrects it. The only reliable defense is process: generate your number independently before the anchor appears, and commit to it in a form that persists — a written model, a stated position, a documented prior belief. Once anchored, adjust aggressively and deliberately, knowing you will almost certainly under-adjust.