The Disposition Effect: Selling Winners and Keeping Losers

*Investors sell the holdings that have risen and hold on to the ones that have fallen. They do this consistently, across countries and decades, and it costs them.*

Among the most robustly documented patterns in investor behaviour is one that sounds almost too simple to be interesting. When investors decide to sell something, they are considerably more likely to sell a holding that has risen than one that has fallen. They lock in their gains and they retain their losses. This tendency has a name, the disposition effect, and it has been observed so consistently, across so many populations and so many markets, that its existence is not seriously disputed.

The pattern was named and described by Hersh Shefrin and Meir Statman in the 1980s, and it was subsequently documented in actual brokerage records by Terrance Odean, who examined thousands of accounts and found that investors realised their gains at a substantially higher rate than their losses. The finding has since been replicated in numerous markets, including among professional participants who might be expected to know better, and among investors in countries with very different cultures and tax arrangements.

The behaviour is puzzling on its face. There is no reason why the price at which an investor happened to purchase something should bear on whether it is worth owning now. The company does not know what you paid. Any analysis of whether to continue holding should concern current prospects relative to the current price, and the investor's own cost is irrelevant to that question.

And yet the purchase price exerts an enormous psychological force, because it functions as a reference point against which the investor judges themselves. A holding above the purchase price represents a decision that has been vindicated, and selling it converts that vindication into something permanent and undeniable. A holding below the purchase price represents a decision that appears to have been wrong, and selling it converts a paper disappointment into an acknowledged error. The reluctance is not about the security. It is about the admission.

This explains why the effect resists correction by knowledge. An investor who understands its logic, and intends not to succumb, will still find that closing a losing position feels materially worse than closing a winning one, and the feeling does not respond to the argument. It is not a failure of understanding. It is a failure of the understanding to reach the part of the mind that is actually making the decision.

The cost is real and has been measured. Because the effect causes investors to retain their weakest holdings and dispose of their strongest, it works against them in markets where recent performance carries any tendency to persist, which the evidence suggests it does over some horizons. Studies examining the subsequent performance of the securities investors sold and those they retained have generally found that the sold holdings went on to outperform the retained ones, which is precisely the wrong way round.

There is a tempting objection that deserves an answer, since it is the rationalisation the effect uses to protect itself. An investor may argue that holding a fallen security is simply the correct application of a value discipline: the price has declined, the business is unchanged, and the security is therefore more attractive than before. This can be true. But the disposition effect can be distinguished from genuine value reasoning by a simple test. The value investor would happily purchase more at the current price. The investor in the grip of the disposition effect would not; they simply cannot bring themselves to sell.

The distinction is worth applying rigorously, because it is available to anyone and it cuts through a great deal of self-deception. When examining a losing holding, an investor might ask whether, holding no position at all and encountering this security today at this price, they would choose to buy it. If the answer is yes, then holding is coherent. If the answer is no, then the only thing sustaining the position is the purchase price, which is a fact about the investor's history and not about the security.

The effect also interacts unpleasantly with taxes. Where realised gains are taxed and realised losses may offset them, the rational sequence would frequently be the opposite of what investors actually do. The disposition effect therefore causes many to pay more tax than necessary, in the course of behaviour that was already costing them.

The structural defences against the effect are the ordinary ones and they are not exotic. Deciding in advance, in writing, what conditions would cause a holding to be reconsidered, removes the decision from the moment when the purchase price is exerting its pull. Reviewing a portfolio by asking what one would buy today, rather than by examining gains and losses against cost, reframes the question in the only terms that are actually relevant. Neither defence requires the investor to stop feeling the pull. They require only that the decision be made elsewhere.

It is worth noting how the effect interacts with the passage of time, because the interaction makes it worse rather than better. A losing position that is retained does not become easier to close as the loss grows; it becomes harder, because the admission required grows in proportion. An investor who could not bring themselves to accept a modest loss will find a substantial one considerably more difficult to accept, and the position accordingly becomes more entrenched the more it deteriorates. The result is that the holdings an investor is least willing to sell are frequently the ones they should have sold soonest, and the reluctance intensifies precisely as the case for acting strengthens. Nothing about this dynamic corrects itself, and an investor who intends to deal with a losing position later has usually decided, without knowing it, never to deal with it at all.

At VESTFY™ the disposition effect is presented as the clearest available example of a bias that understanding alone does not fix. The investor who knows about it still feels it, and the knowledge is useful not because it dissolves the impulse but because it allows them to recognise the impulse for what it is and to build arrangements that do not depend on resisting it.