The Disposition Effect: Why You Sell Winners Too Early and Ride Losers Down
Open any long-running brokerage account — maybe your own — and a pattern tends to appear: the winners are gone and the losers are still there. The stocks that went up were sold, politely, for modest gains. The stocks that went down remain, year after year, “waiting to get back to even.”
Finance has a name for this: the disposition effect — the disposition to sell winners too early and ride losers too long. It is one of the most robustly documented behaviors in individual investing, and unlike most trading folklore, it comes with a body count of real account data.
What the evidence actually shows
The landmark study is Odean (1998), which examined the trading records of 10,000 accounts at a large discount broker. The design is elegant: on any day an investor sells something, look at everything else in the account that could have been sold, and ask whether the position sold was sitting at a gain or a loss relative to purchase price. Investors realized their gains at a substantially higher rate than their losses — except in December, when tax-loss selling briefly reverses the pattern, which is itself evidence that people can overcome the reluctance when a deadline concentrates the mind.
The bitter twist in Odean’s data is that the behavior wasn’t secretly wise: the winners investors sold went on to outperform the losers they kept over the following year. Selling the winner and keeping the loser was, on average, exactly backwards — and that’s before counting the tax cost of realizing gains while deferring deductible losses.
A fair boundary on the claim: this is evidence about individual investors at one broker in one era, and the effect’s size varies across studies, countries, and investor sophistication. The direction, though, replicates stubbornly.
The math underneath: reference points
Why would anyone prefer keeping a loser? Kahneman and Tversky’s prospect theory supplies the machinery. People don’t evaluate outcomes as final wealth; they evaluate gains and losses relative to a reference point — and the value function is concave for gains (each extra dollar of gain feels smaller) but convex for losses (a deeper loss barely feels worse than the current one). Concavity in gains makes you eager to lock a sure win. Convexity in losses makes the gamble of holding feel cheap — the extra pain of falling further is muted, while closing the position makes the loss final, and finality is what hurts.
For a stockholder, the reference point is almost always the purchase price — which is why your average cost basis, a number with real accounting meaning, doubles as an emotional anchor. The market does not know your cost basis. The position’s value today, and its prospects from today, are the same whether you bought at $50 or $150. But the break-even price sits on the screen like a promise: just get back there and this never happened.
Seeing it in your own trade list
The disposition effect leaves fingerprints that arithmetic can surface. Load a position into the stock calculator — every lot, including the ones that are underwater — and look at two numbers side by side: realized P/L (what your selling has actually harvested) and unrealized P/L (what your holding has actually accrued). A long-running pattern of small positive realized gains sitting on top of a large negative unrealized figure is the effect’s signature, written in your own data.
The research ends where advice would begin, and so does this article: none of this says what any position of yours is worth holding or selling — that depends on facts about the company and your situation that no bias study addresses. What the evidence supports is narrower and more useful: know that the pull exists, know which direction it pulls, and be suspicious of any decision whose main argument is the price you happened to pay.