Shared Noeis wiki

Behavioral Economics & Decision-Making

Six linked ideas from the psychology of judgment and choice, each with the original studies and the replications that test them.

Loss Aversion

Loss aversion is the tendency to dislike a loss more than you like a gain of the same size.^[1] Kahneman and Tversky measured gains and losses against a reference level, which is often simply the status quo, rather than against total wealth.^[1]

Richard Thaler used loss aversion to explain the endowment effect, in which people value the same item more highly once they own it.^[2] A 2024 meta-analysis pooling hundreds of published estimates put the average loss aversion coefficient at 1.955, which would mean a loss counts roughly twice as much as an equal gain.^[3]

The idea is less settled than its fame suggests. Reviewing the evidence, Gal and Rucker concluded that losses do not, on balance, tend to be any more impactful than gains.^[4] They argue for a more contextual view of when losses outweigh gains, rather than treating loss aversion as a general law.^[4]

Prospect Theory

Prospect theory is a model of how people actually choose between risky options, published by Daniel Kahneman and Amos Tversky in Econometrica in 1979.^[1] The Nobel committee that honored Kahneman in 2002 described it as an alternative to standard theory that better accounts for observed behavior.^[2]

Its central move is to assign value to gains and losses rather than to final levels of wealth.^[1] People are more sensitive to how an outcome deviates from a reference level than to the outcome itself, so where that reference point sits shapes the choice.^[3] The same sensitivity to starting points runs through Anchoring.

The value function is "generally steeper for losses than for gains," which is the formal root of loss aversion.^[1] It is usually concave for gains and convex for losses, and people underweight outcomes that are merely probable compared with outcomes that are certain.^[1]

Anchoring

Anchoring is the pull an initial number exerts on a later estimate: people start from a value and adjust it to reach an answer, and different starting points yield different answers.^[1] In Tversky and Kahneman's 1974 study, groups given the arbitrary starting numbers 10 and 65 gave median estimates of 25 and 45 percent for the share of African countries in the United Nations.^[1]

The classic effect holds up well. The 2014 Many Labs project, which reran 13 effects across 36 samples, found anchoring robust and concluded that the original studies had underestimated its size.^[2]

Some offshoots have not held up. A 2020 replication found no significant evidence that numbers merely present in the environment bias price estimates, while standard anchoring was robust in the same study.^[3] Why anchoring works is also unsettled: a key test of the selective accessibility account failed to replicate across several experiments.^[4]

Read next: Base Rates, Availability Heuristic.

Availability Heuristic

The availability heuristic is judging how frequent or likely something is by the ease with which examples come to mind.^[1] Tversky and Kahneman argued that people judge frequency and probability using a limited number of heuristics, and set out availability at length in a chapter on exactly that question.^[2]

Like the other heuristics, it is economical and usually effective, but it leads to systematic and predictable errors.^[1] The Nobel committee summarized this line of work as showing that human judgment may take shortcuts that depart from basic principles of probability.

A 1978 study of how people judge the frequency of causes of death found a tendency to overestimate small frequencies and underestimate larger ones.^[3] The authors traced these biases partly to disproportionate exposure, memorability and imaginability, and participants could not correct for them even when told to.^[3]

Base Rates

A base rate is how common something is in the relevant population before you learn anything about the particular case. The base-rate fallacy is the tendency to ignore base rates in favor of case-specific information.^[1]

In Tversky and Kahneman's engineer-and-lawyer study, people were told the proportions of each profession in a group, yet those proportions were effectively ignored once a personality description was added.^[2] In a similar task, most subjects judged a person interested in politics to be a member of parliament, even though salespeople made up a much larger share of the population.^[3]

In Maya Bar-Hillel's version, a problem about a cab and a witness, only about 10 percent of respondents came close to the correct Bayesian answer of 41 percent.^[4] She argued that people weigh information by its perceived relevance, and that information feels more relevant the more specifically it relates to the case being judged.^[1]

Read next: Availability Heuristic, Anchoring.

Hyperbolic Discounting

People value rewards less the later they arrive: a thousand kronor next year is perceived as worth less than a thousand kronor today. Hyperbolic discounting is a way of modeling this that makes preferences dynamically inconsistent, so the choice that looks best in advance can stop looking best when the moment comes.^[1]

David Laibson showed that such preferences give people a motive to constrain their own future choices.^[1] Richard Thaler described the same tension as one between a long-term planning self and a short-term, pleasure-focused self, noting that it is often easier to promise to save more in the future. His Save More Tomorrow program with Shlomo Benartzi built on that, and average saving rates among participants rose from 3.5 to 13.6 percent.^[2]

A major review of three decades of research on choices over time catalogued the empirical regularities that conflict with the standard discounted utility model, along with the alternative models proposed to address them.^[3] Its authors stress separating pure time preference from the many other considerations that shape such choices.^[3]