Microfiche reel · crank through the variations
Disposition Effect
frame 01 / 10
Reel · disposition-effect
Disposition Effect
The empirically documented tendency of investors to realize gains on assets that have appreciated ("winners") at a higher rate than they realize losses on assets that have declined ("losers"), even when doing so is sub-optimal after tax and rebalancing considerations.
“The disposition to sell winners too early and ride losers too long.”
frame 02 / 10
Provenance
Where it was first filed
Hersh Shefrin & Meir Statman, 1985 — The Disposition to Sell Winners Too Early and Ride Losers Too Long: Theory and Evidence. Shefrin and Statman coined the term 'disposition effect' and framed it as the joint product of prospect theory, mental accounting, regret aversion, and self-control, against tax-loss-selling incentives that should produce the opposite pattern.
frame 03 / 10
Tested variation 1
Odean (1998) — seminal field test
Investors realized gains at a markedly higher rate than losses; the effect reversed in December consistent with tax-loss selling. The disposition behavior was not explained by rebalancing, transaction costs, or subsequent returns, and lowered after-tax returns.
Terrance Odean, 1998 · yield PGR ~14.8% vs PLR ~9.8% outside December (per Wikipedia summary of the paper); paper states a stock that was up was almost 60% more likely to be sold than one that was down. In December PGR fell to ~10.8% and PLR rose to ~12.8%. · 10,000 brokerage accounts, 1987-1993
frame 04 / 10
Tested variation 2
Weber & Camerer (1998) — lab experiment
Subjects sold winners and held losers contrary to Bayesian optimization; the disposition effect was greatly reduced when shares were automatically sold each period, implicating the act of voluntary realization.
Martin Weber, Colin F. Camerer, 1998 · yield not reported (direction reported; precise magnitude not retrieved) · Laboratory subjects (exact n not retrieved)
frame 05 / 10
Tested variation 3
Frazzini (2006) — disposition effect and price drift
Post-announcement drift is most severe when the sign of the capital gain/overhang matches the sign of the news, consistent with disposition-driven underreaction.
Andrea Frazzini, 2006 · yield Event-driven long/short strategy yielded monthly alphas over ~200 basis points (~2.4% per month spread) · U.S. equities; mutual fund holdings data
frame 06 / 10
Tested variation 4
Ben-David & Hirshleifer (2012) — mechanism challenge
Selling probability is V-shaped in profit (big losers sold at short horizons), with little upward jump at zero profit; argues against a simple realization-preference (prospect-theory) explanation and favors belief-revision-driven trading.
Itzhak Ben-David, David Hirshleifer, 2012 · yield not reported as a single statistic; reported as a V-shaped selling-probability function · Individual investor trading records (large brokerage panel)
frame 07 / 10
Method
How the judgment forms
The classic account (Shefrin & Statman 1985) combines prospect theory — concave-over-gains, convex-over-losses value function defined relative to the purchase price as reference point — with mental accounting: selling a winner closes a mental account at a gain (pleasurable), while selling a loser locks in a loss (painful), so investors gamble to recover. Realization utility (Barberis & Xiong) formalizes this as a burst of utility experienced at the moment of sale.
Barberis & Xiong (2009) show prospect theory does NOT robustly predict the disposition effect under reasonable parameterizations and propose realization utility instead; Kaustia (2010) shows field selling patterns are inconsistent with prospect-theory predictions; Ben-David & Hirshleifer (2012) find V-shaped selling and argue belief revision (speculative trading on expected returns), not realization preference, drives the data. Rational explanations (mean reversion beliefs, informed trading) have also been advanced.
frame 08 / 10
Verdict · replication
mixed
The aggregate behavioral pattern (selling winners more readily than losers) replicates very broadly — across retail brokerage panels (Odean 1998; Dhar & Zhu 2006, ~50,000 investors), lab experiments, social-trading platforms, and even mutual funds (loser funds ~1.7x more likely to realize a gain than a loss, ~8% PGR-PLR gap per Frazzini). A 2024 meta-analysis of disposition-effect experiments (Cheung) reports investors are roughly 10% more willing to sell winners than losers at baseline, with little selective-reporting bias in baseline estimates. However, the MECHANISM is genuinely contested: prospect-theory explanations are challenged (Barberis & Xiong 2009; Kaustia 2010), and Ben-David & Hirshleifer (2012) argue the realization-preference interpretation is not supported by V-shaped selling. So the phenomenon is robust; its standard causal story is not.
frame 09 / 10
On record
Documented in the wild
Odean (1998) discount-brokerage trading records · 1998
Across 10,000 accounts (1987-1993), investors realized a higher proportion of paper gains than paper losses outside December, reversing in December for tax-loss selling; the behavior reduced after-tax returns.
Genesove & Mayer (2001) Boston condo market · 2001
Boston condominium owners facing nominal losses set asking prices 25-35% above the loss-implied price, achieved selling prices 3-18% higher, and had much lower sale hazards — a real-estate analogue of reluctance to realize losses.
Frazzini (2006) U.S. equity post-news drift · 2006
Documented that unrealized capital gains/losses (a disposition overhang) predict the magnitude of post-earnings/news price drift in U.S. stocks; an event-driven strategy earned monthly alphas over ~200 bps.
Locke & Mann (2005) CME floor futures traders · 2005
Professional floor futures traders held losing positions significantly longer than winning ones (a disposition pattern), yet the most successful, disciplined traders held losers least long, linking lower disposition behavior to trading success.
frame 10 / 10
Final frame · the counter-move
Evaluate each position purely on forward-looking expected return, ignoring the purchase price (a sunk reference point); pre-commit to rules such as automatic rebalancing or stop-losses, and deliberately exploit tax-loss harvesting, which pushes in the corrective direction. Weber & Camerer found that removing the voluntary act of realization (automatic selling each period) sharply reduced the bias.
Weber & Camerer (1998) showed the disposition effect was greatly reduced when holdings were automatically sold each period; Odean (1998) showed the December tax-loss-selling reversal, indicating the bias is suppressible by salient incentives/rules.
← scrub the reel → · Disposition Effect
Adjacent reels
- Loss AversionProspect theory's loss aversion (losses loom larger than gains) is the most-cited driver; the disposition effect is one of its market-level manifestations.
- Prospect TheoryReference-dependence around the purchase price plus the S-shaped value function is the classic explanation — though its sufficiency is disputed.
- Sunk Cost FallacyBoth involve refusing to abandon a losing position; the disposition effect specifically concerns realizing paper losses vs gains on tradable assets.
- Mental AccountingClosing a mental 'account' at a gain feels good while closing at a loss feels bad; framing each holding separately produces the asymmetry.
- Realization UtilityBarberis & Xiong's formalization: utility arises at the moment of sale, generating disposition behavior without relying on prospect theory over paper gains.
- Tax-Loss SellingRational tax incentives push toward realizing losses (especially in December), the opposite of the disposition effect — which is why the December reversal in Odean is diagnostic.