The disposition effect is one of the most consistently documented behavioural patterns in retail investing. In every dataset ever studied, retail investors show a systematic tendency to sell their winning positions faster than statistical expectations would predict and to hold their losing positions longer. The gap between the two behaviours is large, remarkably consistent across cultures and time periods, and directly costly to portfolio returns.
The classic experiment
Terrance Odean's 1998 paper "Are Investors Reluctant to Realise Their Losses?" analysed roughly 10,000 discount-brokerage accounts over several years. His central finding was that investors sold winning positions 1.7 times as frequently as losing positions — despite the fact that, on average, the winning positions continued to outperform after being sold, while the losing positions continued to underperform after being held. In other words, investors were making the wrong side of the decision consistently.
Subsequent studies have replicated the finding in Israeli, Chinese, Finnish, and many other datasets. The pattern is not culture-specific. It appears in every retail investor population that has been studied.
The mechanism
The disposition effect operates through two behavioural biases interacting. Loss aversion — the pain of losses being felt roughly twice as strongly as the pleasure of equivalent gains — makes selling a losing position painful in the moment. Selling locks in the loss and activates the full 2x weighting. Holding it keeps the loss in the "not yet realised" mental category, which somehow feels less painful even though the mathematical exposure is identical.
Simultaneously, mental accounting — the tendency to treat gains and losses on different positions as separate psychological events — makes selling a winner feel like a discrete gain event to be enjoyed. The gain is realised and added to the mental "wins" column. The pleasure of the realised gain outweighs the (asymmetric) potential pain of possibly giving it back.
The two biases work together to produce the specific disposition pattern. Cutting winners locks in the certain gain that feels good. Holding losers defers the loss that would feel bad. Neither decision is rational; both feel right in the moment.
The cost to portfolio returns
The empirical cost is substantial. Retail investors who exhibit strong disposition effects underperform buy-and-hold benchmarks by several percentage points per year on average, according to multiple studies. The gap comes from three sources: winning positions typically continue to outperform after being cut (so selling them early forgoes future gains); losing positions typically continue to underperform after being held (so holding them extends losses); and the taxable events created by selling winners early produce tax drag that is not present in longer holding patterns.
For a retail investor with a typical size portfolio and typical turnover pattern, the disposition effect can easily cost 2–4 percentage points of annual return. Over decades, this compounds to a portfolio value substantially below what buy-and-hold would have produced with the same underlying stock selection.
Why professional investors show less of the effect (but not zero)
Studies of institutional trading behaviour show a smaller but non-zero disposition effect. Institutional traders are trained to think in terms of forward expected value rather than realised P&L, and their compensation structures typically reward absolute performance rather than realised gains specifically. But the underlying psychological pull is not fully absent even in professional contexts.
Institutions with the smallest disposition effects tend to be systematic quantitative funds where trading decisions are made by algorithms without human intervention. Discretionary traders, even sophisticated ones, still show measurable disposition patterns — smaller than retail but not zero.
Practical implications for a retail investor
Three practices measurably reduce the disposition effect.
Separate the sell decision from the profit-and-loss framing. Ask "based on today's information, would I buy this position today at the current price?" If the answer is yes, holding is justified. If the answer is no, selling is justified — regardless of whether the position is currently profitable or unprofitable. This framing removes the anchor to cost basis.
Set exit rules that are not price-anchored. "I will sell if my thesis is invalidated" is a rule not anchored to price. "I will sell if the stock drops 20%" is a rule anchored to your entry price. The first is a discipline; the second is disposition-effect thinking dressed as one.
Rebalance systematically. Rules-based rebalancing (returning positions to target weights at fixed intervals or when they drift beyond thresholds) forces the discipline of trimming positions that have grown large — some of which are winners. It also forces the discipline of adding to positions that have shrunk — some of which are losers. Neither decision is made under the emotional pressure of the moment; both are made mechanically.
The tax complication
There is one specific case where holding losers longer than "pure" expected-value analysis would suggest is defensible: tax-loss harvesting. In taxable accounts, realising a loss creates a tax benefit that offsets other realised gains. The specific mechanics of holding losses to appropriate periods, harvesting them at strategic moments, and avoiding wash-sale rules can produce meaningful after-tax benefits.
But tax-loss harvesting is different from the disposition effect. The tax-loss harvester holds losers because of a specific tax benefit that offsets the mathematical cost of holding losing positions. The disposition-effect investor holds losers because of a psychological aversion to realising the loss, without any offsetting benefit. The two behaviours look similar externally but come from very different sources, and the second is much more costly to portfolio returns over time.
The rule to internalise
The tendency to sell winners too early and hold losers too long is not a moral failing; it is a well-documented pattern that most retail investors exhibit and that costs them measurably. The path to reducing the effect is not intellectual awareness alone — knowing about disposition effect does not eliminate it — but the construction of decision frameworks that force position sizing and exit decisions to be made independently of the current P&L. Rules-based rebalancing, thesis-based exit criteria, and periodic zero-based valuation of every position are the practices that produce the largest improvement in the aggregate. None of them removes the psychological pull; all of them make it less consequential when it happens.
Educational content only. Not investment advice.