Imagine an investor holding two stocks. One is up 30 percent since purchase; the other is down 30 percent. Cash is needed, and one position has to go. Most people sell the winner. It feels prudent. Lock in the profit, and give the loser time to come back.
This instinct is so common that finance researchers gave it a name: the disposition effect. Hersh Shefrin and Meir Statman coined the term in 1985 to describe the tendency to realise gains quickly while letting losses linger. It shows up in retail brokerage accounts, among professional futures traders, in housing markets, and even in the behaviour of fund managers.
The puzzle is that this pattern runs against both tax logic and the evidence on stock returns. Selling winners triggers taxable gains, while selling losers creates deductions. And the stocks investors cling to tend to keep underperforming. Seeing why intelligent people systematically do the opposite of what maximises returns reveals something fundamental about how markets process information. It also shows how individual portfolios quietly leak value, one well-intentioned decision at a time.
The Evidence: A Bias Written in Trading Records
The landmark study came from Terrance Odean in 1998. He examined trading records from 10,000 accounts at a US discount brokerage between 1987 and 1993 and compared how often investors realised available gains versus available losses. The proportion of gains realised was 14.8 percent. The proportion of losses realised was 9.8 percent. On any given day, investors were roughly 50 percent more likely to sell a winner than a loser.
The cost was measurable. The winners investors sold went on to outperform the losers they kept by about 3.4 percentage points over the following year. That figure is before taxes, which widen the gap further, because the pattern meant paying capital gains tax early while deferring deductible losses. Tellingly, behaviour shifted in December, when tax-loss selling briefly overrode the instinct.
The finding has replicated widely. Mark Grinblatt and Matti Keloharju found it across the comprehensive Finnish shareholder registry. Martin Weber and Colin Camerer reproduced it in laboratory experiments, even when subjects were told how prices were generated. Studies of Chicago futures traders and Israeli brokerage clients showed that professionals are not immune, though experience tends to soften the effect. In housing, David Genesove and Christopher Mayer found that Boston homeowners facing a nominal loss set higher asking prices and waited longer to sell.
The effect also leaves fingerprints on market prices. Andrea Frazzini showed that when good news arrives for stocks widely held at a gain, disposition-prone investors sell into the rally, dampening the immediate price reaction. Full adjustment arrives slowly, contributing to the well-documented drift after earnings announcements. A private bias, aggregated across millions of accounts, becomes a market-wide inefficiency.
TakeawayA bias that costs one investor a few percentage points can, when shared by millions, slow the speed at which prices absorb news. Individual psychology and market efficiency are not separate subjects.
The Psychology: Prospect Theory and the Pain of Closing a Loss
The theoretical backbone comes from Daniel Kahneman and Amos Tversky's prospect theory, published in 1979. Its central insight is that people evaluate outcomes as gains and losses relative to a reference point, rather than as final levels of wealth. For investors, that reference point is almost always the purchase price, a number the market neither knows nor cares about.
Two features of prospect theory's value function produce the disposition effect. For gains, the function is concave, so people become risk-averse and prefer a certain gain to a gamble. Selling a winner locks in that sure thing. For losses, the function is convex, so people become risk-seeking and will gamble to avoid crystallising a loss. Holding a loser is that gamble. Loss aversion deepens the asymmetry, since losses feel roughly twice as painful as equivalent gains feel good.
Richard Thaler's work on mental accounting explains why each position is judged in isolation. Investors open a separate mental account for each stock, and closing one at a loss means formally admitting a mistake. Regret aversion compounds this. A paper loss can still be reversed, but a realised loss becomes a permanent record of a bad decision. Pride works in the opposite direction, rewarding the act of booking a gain.
More recent work by Nicholas Barberis and Wei Xiong frames this as realisation utility. In this view, investors get a burst of pleasure from the act of selling at a profit, independent of what happens to their total wealth. Brain-imaging studies have found neural activity consistent with the idea. The implication is uncomfortable. The disposition effect is not ignorance of finance. It is a feature of how humans experience outcomes.
TakeawayThe purchase price is the most emotionally charged number in a position and the least informative one. Markets price the future, but investors keep score against the past.
Breaking the Pattern: Rules, Reframing, and Portfolio Thinking
Because the disposition effect is rooted in emotion rather than misunderstanding, knowing about it is rarely enough. Research suggests that awareness alone does little to reduce it. The countermeasures that work change the decision environment instead of relying on willpower in the moment.
The first tool is rules-based selling, set before emotions attach. Predetermined stop-loss levels, scheduled reviews, or position-size limits remove discretion at exactly the point where bias does its damage. Trend-following traders have long repeated the maxim to cut losses short and let profits run, because it inverts the natural inclination. A rule written in calm conditions acts as a commitment device against a future self under stress.
The second is reframing. A useful exercise is the clean-slate test: if this position were currently cash, would you buy it at today's price? The purchase price is irrelevant to future returns, and experiments by Cary Frydman and Antonio Rangel found that making purchase price less salient weakened the disposition effect. Some investors simply hide the cost-basis column. Tax-loss harvesting offers another reframe. Selling a loser becomes an action that produces a concrete benefit, turning a defeat into a gain.
The third is evaluating at the portfolio level. Systematic rebalancing to target weights forces decisions based on allocation rather than on gain or loss. A decision journal that records the original investment thesis helps separate two different events: the price fell, and the thesis broke. The goal is not to sell every loser. It is to make each decision to hold a fresh choice, rather than a refusal to close an account.
TakeawayThe best defence against a predictable bias is a decision made before the bias arrives. Rules and reframes work because they are set while the purchase price still means nothing to you.
The disposition effect is one of the most robust findings in behavioural finance. It has been documented across countries, asset classes, and levels of sophistication. It persists because it is built on how people experience gains and losses, not on gaps in knowledge.
For market observers, it helps explain why prices sometimes underreact to news and why momentum can persist. For individual investors, it represents a quiet, recurring cost: paying taxes early, holding laggards, and trimming the very positions that may continue to work.
The purchase price feels like the most important number in a position. To the market, it means nothing. Learning to see a portfolio the way the market does, forward-looking and indifferent to history, may be one of the more valuable adjustments an investor can make.