Anchoring Bias

Anchoring bias is a systematic distortion in numeric judgment in which an initial value, the anchor, biases a subsequent estimate toward itself. Tversky and Kahneman described it in 1974 as anchoring and adjustment: people start from whatever number is available and adjust from there, and the adjustment typically stops short. What makes the effect notable rather than obvious is that it operates even when the anchor is plainly uninformative, and even when people are told it is arbitrary and asked to disregard it. In personal finance the anchors are rarely arbitrary and rarely accidental. They are list prices, opening offers, prior account balances, and the first salary figure named in a negotiation.

The replication record splits along a line that changes what can honestly be claimed, and this is the part most consumer articles get wrong.

The presented anchor, where a person is shown a number and asked to judge something relative to it, holds up well. Jacowitz and Kahneman's anchoring task, published in Personality and Social Psychology Bulletin 21(11) in 1995, was one of the thirteen effects tested in the Many Labs project (Klein and colleagues, Social Psychology 45(3), 2014), which ran four of its anchoring items across 36 samples. Ten of the thirteen effects "replicated consistently," and the project named all three that did not do so, none of which was anchoring. The replications in fact recovered a larger anchoring effect than the original study had. This is also the form Tversky and Kahneman's original wheel-of-fortune demonstration took: the number was uninformative, but it was explicitly offered for comparison.

The incidental environmental anchor, where a number is merely present in the surroundings rather than offered for comparison, is a different matter. Critcher and Gilovich's 2008 finding of that kind was among the studies replicated in Many Labs 2 (Klein and colleagues, Advances in Methods and Practices in Psychological Science 1(4), 2018), a very high-powered multi-site project across 125 samples, and it was not supported. That project as a whole found replication effect sizes substantially smaller than the originals.

So the defensible claim is narrower and more useful than the popular one: an anchor you are invited to consider moves your estimate; a number that merely happens to be in the room does not reliably do so. Articles asserting that any nearby number will shape your spending are describing the version that did not replicate. That distinction also tells you where to look for real exposure, which is at the moment someone deliberately puts a number in front of you.

The effect operates in three steps, and only the first is conscious.

1. A number arrives and becomes the starting point, whether or not anyone chose it as a reference.
2. The mind adjusts away from it in the direction it believes correct.
3. The adjustment stops early, usually as soon as the answer becomes plausible rather than when it becomes right.

Because step three is where the error lives, the fix is not to adjust harder. It is to avoid starting from someone else's number.

Four anchors that do real damage in household finance:

• A list price. It is the seller's number, set before any negotiation, and it frames every subsequent offer as a discount from it rather than as a valuation.
• A prior portfolio peak. A balance that was once higher becomes the reference point for whether the current one is a loss, even though the peak was a moment rather than a value.
• The first salary figure named. Whoever names a number first sets the range the conversation happens in.
• A purchase price you paid. What a holding cost is a fact about the past, not evidence about what it is worth or what it will do, but it anchors the decision to sell.

A hypothetical, resolving a process question rather than an amount. Two buyers look at the same house. The first reads the listing, then estimates what the house is worth. The second pulls three comparable sales and writes down a number before opening the listing. Both may end up bidding, and the second may even bid more. What differs is that the second buyer has a number that did not come from the seller, so they can tell whether the list price is high or low rather than only how far from it they are willing to move. Setting your own figure first is the whole countermeasure, and it works because it changes what the starting point is rather than trying to adjust from someone else's.

Pros

• Anchoring is one of the easiest biases to plan around, because the countermeasure is a single concrete step taken before exposure.
• Naming it makes the structure of a negotiation visible, and explains why who speaks first matters.
• It identifies a specific class of financial moment worth slowing down for, which is any moment where a number is being presented to you.

Cons

• The concept is routinely over-applied. Not every number that precedes a decision is an anchor, and treating them all as one is not useful.
• Awareness alone is a weak defense. The literature on warning people about anchoring and expecting them to correct for it is not encouraging.
• The incidental-anchor version that circulates widely in popular writing is the version that failed to replicate, so advice built on it is built on sand.
• An anchor can be informative. A list price does carry some information about a market, and disregarding it entirely is its own error.