Loss Aversion

frame 02 / 10

Provenance

Where it was first filed

Daniel Kahneman & Amos Tversky, 1979 — Prospect Theory: An Analysis of Decision under Risk. The concept that 'losses loom larger than gains' was articulated as a core component of prospect theory's value function, which is steeper for losses than for gains. The riskless-choice formalization and a reference-dependent model came later in Tversky & Kahneman (1991, QJE). Kahneman received the 2002 Nobel Memorial Prize in Economics for this body of work (Tversky had died in 1996).

frame 03 / 10

Tested variation 1

Prospect theory value function (founding claim)

Choices implied a value function that is concave for gains, convex for losses, and markedly steeper for losses than for gains - the original statement of loss aversion.

Daniel Kahneman, Amos Tversky, 1979 · yield not reported as a single lambda in the 1979 paper; loss aversion expressed qualitatively as a steeper loss limb of the value function · groups of respondents (often students and faculty); problem-level Ns reported per item, not a single overall N

frame 04 / 10

Tested variation 2

Reference-dependent riskless choice / lambda estimate

Losses and disadvantages weigh more than commensurate gains; the 1991 QJE paper formalizes this qualitatively for riskless choice. The widely cited lambda ≈ 2.25 figure was estimated in the subsequent Tversky & Kahneman (1992) paper 'Advances in Prospect Theory: Cumulative Representation of Uncertainty' (Journal of Risk and Uncertainty 5(4):297-323, N=25), not in the 1991 QJE paper.

Amos Tversky, Daniel Kahneman, 1991 · yield lambda approx 2 in the broad literature; the specific estimate of 2.25 derives from Tversky & Kahneman (1992, Journal of Risk and Uncertainty 5(4):297-323), not from the 1991 QJE paper; treat as canonical-textbook value, not a precise single result from the 1991 work · not a single fixed N; the 1991 QJE paper draws on choice experiments; the 1992 CPT estimate used N=25

frame 05 / 10

Tested variation 3

Endowment effect (Cornell mugs)

Owners demanded roughly twice what buyers would pay (median WTA around $5.25 vs WTP around $2.25-$2.75), so far fewer mugs traded than the Coase theorem predicts - taken as evidence of loss aversion over the endowed good.

Daniel Kahneman, Jack L. Knetsch, Richard H. Thaler, 1990 · yield WTA/WTP ratio approximately 2:1; trading volume well below the ~50% Coasean prediction · Cornell University undergraduates across multiple sessions

frame 06 / 10

Tested variation 4

Neural basis of loss aversion (fMRI)

As potential losses increased, activity in reward-related regions (e.g., ventral striatum, ventromedial PFC) decreased more steeply than it increased for equivalent gains ('neural loss aversion'), and this neural asymmetry tracked individual behavioral loss aversion.

Sabrina M. Tom, Craig R. Fox, Christopher Trepel, Russell A. Poldrack, 2007 · yield correlation between neural and behavioral loss aversion reported as significant; exact r not extracted here · 16 participants

frame 07 / 10

Method

How the judgment forms

In prospect theory the carrier of value is the change from a reference point, not the final wealth state; the value function is steeper in the loss domain than the gain domain, so a given deviation hurts more as a loss than it pleases as a gain. This reference dependence is hypothesized to drive downstream phenomena: the endowment effect (giving up an owned good is coded as a loss), the disposition effect (selling at a loss realizes the painful loss), and conservative effort to avoid falling below a salient reference (par, the status quo). Neuroeconomic accounts locate an asymmetry in reward-circuit responses to potential losses vs gains.

Critics argue the data are better explained without a special loss-specific weighting. Gal & Rucker (2018) and Yechiam (2019) contend that apparent loss aversion often reduces to status-quo/inertia bias, risk aversion, or attentional asymmetries, and that the effect vanishes for small-to-moderate stakes; on this view 'losses loom larger' is a context-dependent regularity, not a universal psychological constant. A 2025 re-meta-analysis argues loss aversion appears mainly when the design makes losses larger than gains or uses ordered payoffs.

frame 09 / 10

On record

Documented in the wild

• Disposition effect: investors reluctant to realize losses · 1998

  Terrance Odean analyzed trading records for 10,000 accounts at a large discount broker (1987-1993) and found investors realized gains at roughly a 50% higher rate than losses (proportion of gains realized ~14.8% vs proportion of losses realized ~9.8% outside December; PGR/PLR ≈ 1.51), holding losing stocks too long - a pattern interpreted via prospect theory / loss aversion.

• Loss aversion among PGA Tour golfers (incl. Tiger Woods) · 2011

  Pope & Schweitzer analyzed ~2.5 million putts from 239 PGA tournaments (2004-2009, 421 pros). Players sank par putts about 2 percentage points more often than equidistant birdie putts, because missing par is coded as a loss relative to the salient reference point (par) - persisting despite experience, competition, and high stakes.

• Endowment effect in lab markets (Cornell mugs) · 1990

  Owners of randomly assigned mugs demanded roughly double what buyers would pay (median WTA ~$5.25 vs WTP ~$2.25-2.75), so trade volume fell far short of the Coase-theorem prediction - a documented market-level manifestation attributed to loss aversion over the endowed good.