Anchoring

Anchoring is the tendency to let a reference point—often an arbitrary number or a stock’s current price—shape later decisions. The book distinguishes the bias of an individual judgment from the performance of a system that aggregates many judgments: anchoring can constrain how a trader responds to new information, yet the resulting market can still approach an optimal price if individual errors do not all push in the same direction.

Definition and Operation

In financial decisions, the current price of a stock can serve as an anchor. Rather than assessing a company’s prospects afresh, investors may adjust their estimates from the price they already see or from the price at which they bought. The adjustment is responsive to new information, but not fully independent of the starting point. Anchoring therefore does not require a person to ignore evidence; it can work by making evidence count only as a modification of an established reference.

The book places anchoring among behavioral quirks that matter to markets only under particular conditions. Overconfidence or attention to recent news may affect individual investors, but those tendencies do not necessarily distort prices systematically: investors can be overconfident in opposing positions, and the same information can mean different things to different people. The danger arises when errors align, producing a directional bias rather than canceling one another.

Robot Traders and Market Aggregation

Economists Karim Jamal and Shyam Sunder tested anchoring with robot investors using a straight anchor-and-adjust strategy. The simulated traders began from a particular reference point and related new information back to it, rather than evaluating each new piece freely. Their behavior was deliberately unlike the fully independent reassessment predicted by idealized economic theory.

Even so, their buying and selling eventually converged on a price very near the optimum. The result does not show that anchoring is harmless in every setting; it shows that individually biased decision rules need not prevent a market from solving a problem well. In the book’s account, aggregation can turn imperfect judgments into a useful collective verdict when errors are not systematic. When they are systematic, markets do a poorer job.

Contrast with Group Judgment

The book’s account of crowd wisdom clarifies why the robot-trader result matters. A large group can make a good estimate when its members bring diverse information, judge independently, rely on decentralized or local knowledge, and have a mechanism for aggregating their judgments. Averaging errors can leave the information those judgments contain. But cancellation cannot rescue a group whose errors all point in the same direction, nor can aggregation create information where none exists.

This distinction separates anchoring as an individual influence from a market-wide outcome. The central question is not whether every participant reasons perfectly, but whether participants’ limitations collectively impose a common bias. The robot experiment offers a case in which anchored behavior remained compatible with convergence; the book’s broader argument leaves open the possibility of periods when emotion or other systematic distortions make market prices wrong.

A Contrast in Organizational Judgment

The Columbia disaster provides a separate illustration of how judgments formed under uncertainty can become resistant to correction. After foam struck the shuttle’s left wing during ascent, engineers in the Debris Assessment Team estimated the foam’s size and speed and used the Crater algorithm to estimate how deeply it could penetrate the wing’s thermal-protection tiles. The team had not reached a conclusion, and the available analysis was limited; a request for on-orbit pictures had not been approved.

In the Mission Management Team’s discussions, managers treated the possibility of serious damage as largely settled rather than beginning with the evidence. They rejected requests for images, including on the basis of assumptions about image resolution that they were not cleared to verify and did not check with the Department of Defense. Later, the engineers presented five possible scenarios and stressed that their conclusion that the shuttle was likely safe depended on poor information and tools not designed for an impact of that scale. Management emphasized the reassuring conclusion instead of its uncertainty. This organizational episode is not the robot-trader experiment: it shows how a prior judgment can shape which evidence is sought and how cautionary evidence is heard, rather than demonstrating that every such bias will be canceled by aggregation.

Limits and Significance

The book’s treatment makes the effect of anchoring conditional, not absolute. A reference point can influence an individual without determining the collective result; conversely, a group process can fail when a shared assumption directs attention away from missing evidence or competing explanations. The distinction is important because market efficiency does not require perfectly rational participants, but it does depend on the conditions under which their errors combine.

Anchoring thus sits at the boundary between personal bias and collective intelligence. The robot traders show that a market can converge despite a constrained decision rule; the Columbia account shows the cost of a decision process that narrows inquiry before uncertainty has been resolved. Together, these cases make the decisive question not simply whether judgments are biased, but whether the structure in which they are made allows independent information and correction to survive.