Investor psychology Lesson 3 of 15
Articles · Behavioral finance
The disposition effect: selling winners and holding losers
What the disposition effect is, how Odean measured it in 10,000 brokerage accounts, what it costs after tax, where the evidence is disputed and what helps.
On this page
The disposition effect is the tendency to sell investments that have gone up and keep investments that have gone down. Hersh Shefrin and Meir Statman gave it the name in 1985: investors are disposed to sell winners too early and ride losers too long.
It is the best-documented bias in real brokerage records, and it is unusual among biases in that it has a price tag. In a taxable account it works against the tax code, which rewards the opposite behavior. This lesson shows how the effect was measured, what a single instance costs in dollars, what researchers still argue about, and what the evidence says reduces it. The overview in cognitive biases in investing places it among the other biases.
How it was measured
The obvious test, counting how many winners and losers investors sell, does not work. In a rising market most positions are winners, so most sales will be winners whatever investors prefer. Terrance Odean's 1998 solution is worth understanding because it is the measure every later study uses.
He obtained the trading records of 10,000 accounts at a US discount brokerage from January 1987 to December 1993. On every day that an account sold something, he looked at the whole portfolio and sorted each position into four boxes: gains that were sold (realized gains), gains that were not (paper gains), losses that were sold and losses that were not.
Across the whole year PGR was 0.148 and PLR was 0.098. A position showing a gain was about 50% more likely to be sold than one showing a loss. From January to November the gap was wider (0.152 against 0.094). In December it reversed: PGR 0.108, PLR 0.128. Selling a loser before the year ends captures a tax deduction, and with that reason in view investors did sell losers. They were able to. For eleven months they chose not to.
Odean checked the innocent explanations. Investors were not rebalancing, because the pattern held when he looked only at sales of an entire position. They were not avoiding high trading costs on low-priced losers. And they were not right about the future, which is the finding that gives the bias its cost.
Over the following 252 trading days, the winners that had been sold beat the market index by 2.35% on average. The losers that had been kept trailed it by 1.06%. The belief that a loser "will come back" was, on average, wrong by about 3.4 percentage points in the first year.
A laboratory test points the same way. Martin Weber and Colin Camerer ran an experimental market in which prices followed a known random process, so there was no reason to prefer losers. Their subjects still sold clearly more shares that were above their purchase price than shares that were below it.
What one instance costs: a worked example
An illustrative investor, Priya, needs $10,000 from a taxable US account. She holds two positions, each worth $10,000 and each held for more than a year:
| Position | Cost | Value now | Gain or loss |
|---|---|---|---|
| A | $6,000 | $10,000 | +$4,000 |
| B | $14,000 | $10,000 | −$4,000 |
She has no view that either company will do better than the other from here. Her long-term capital gains rate is 15%, and she has other gains this year.
- Sell A, the winner. She owes 15% of $4,000, which is $600. She keeps $9,400 and still holds B with its unused loss.
- Sell B, the loser. She owes nothing on the sale, and the $4,000 loss offsets $4,000 of other gains, saving $600. She is $10,600 better off in cash terms and still holds A.
The difference is $1,200 on a $10,000 sale, before anything the two stocks do next. Most people in Priya's position sell A. Selling A feels like a success being banked. Selling B feels like signing a confession.
Two details keep the example honest. The tax on A is deferred, not escaped: if she sells A later she will owe it then, so part of the $1,200 is a timing benefit. And US rules limit the other side. A net capital loss can offset only $3,000 of ordinary income a year ($1,500 if married filing separately), with the rest carried forward, and a loss is disallowed if she buys the same or a substantially identical security within 30 days before or after the sale (the wash sale rule). These are the IRS rules for tax year 2025; tax-loss harvesting covers them in full. In a retirement account none of the tax applies, and the cost of the bias is only what the stocks do next.
Why it happens: three explanations and an argument
Prospect theory. Shefrin and Statman's original account used the value curve from loss aversion. Above the purchase price an investor is in the zone of gains, where people avoid risk, so they lock the gain in. Below it they are in the zone of losses, where people take risks to avoid a certain loss, so they hold on and hope.
Mental accounting and regret. Each stock sits in its own mental account, opened at the purchase price. Selling at a loss closes the account in the red and makes the mistake final; see mental accounting. Shefrin and Statman also named regret, weak self-control and tax considerations.
Belief in reversal. Investors may simply expect losers to bounce and winners to fade. Odean's data show the belief was mistaken on average, but it is a belief, not a preference.
The argument is over which of these is true, and it is not settled.
- Nicholas Barberis and Wei Xiong worked through the prospect-theory explanation formally and found that it does not reliably produce the effect. In some versions of their model it predicts a disposition effect and in others the opposite.
- Itzhak Ben-David and David Hirshleifer found that the probability of selling is V-shaped: investors are more likely to sell after a large gain or a large loss than after a small move. There is no jump in selling as a position crosses from loss to gain, which is what a pure dislike of realizing losses would produce. They read the pattern as trading on beliefs.
- Tom Chang, David Solomon and Mark Westerfield found a reverse disposition effect for actively managed funds. Investors sell losing funds readily and hold winning ones. Their explanation is blame: with a fund there is a manager to fault, so selling does not mean admitting your own error.
What survives the argument is the behavior and its cost. Individual investors in stocks sell gains more readily than losses, in several countries and decades of data, and it lowers their after-tax returns.
Who shows it, and who does not
The effect is an average. Ravi Dhar and Ning Zhu found that 19.7% of the investors in a large brokerage sample showed no disposition effect or the reverse, and that wealthier investors and those in professional occupations showed less. Lei Feng and Mark Seasholes followed investors in China and found that sophistication and trading experience together eliminated the reluctance to realize losses and cut the eagerness to realize gains by 37%, without removing it.
The habit also leaves a mark on prices. Andrea Frazzini used mutual fund holdings to estimate how many of a stock's holders were sitting on gains or losses, and found that prices react slowly to news when the two line up: good news for a stock whose holders have gains is met by their selling, so the price drifts up instead of jumping.
What helps
The evidence here is thinner than the evidence for the bias, and none of it promises a fix.
- Take the decision out of the moment. In Weber and Camerer's experiment, when shares were sold automatically at the end of each period and subjects had to actively buy back what they wanted, the disposition effect was greatly reduced. The real-world versions are a stop level or a written exit condition set at purchase. See stop-loss placement.
- Ask the cash question. "If I had this amount in cash today, would I buy this stock?" removes the purchase price from the decision. It is the same test used against the sunk cost fallacy.
- Review winners and losers on one list with one standard. The bias lives in treating the two groups differently. A scheduled review that asks the same written questions of every position, whatever its color, removes the asymmetry from the procedure.
- Use December's logic all year. Investors sold losers when a tax reason was in front of them. In a taxable account that reason exists in every month.
How Finabase helps
A Finabase portfolio lists every position with its average cost, market value, gain and gain percent on one screen, and its report page counts realized and unrealized wins by period. That makes the comparison in this lesson, what you sold against what you kept, something you can read rather than remember. Paper trading lets you practice the uncomfortable half of the skill, selling a position that is below its cost, with no money at stake. An alert at a price you chose in advance turns an exit into a notification instead of a debate.
Frequently asked questions
What is the disposition effect in simple terms?
It is the habit of selling investments that are up and keeping those that are down. Investors do it because taking a gain feels good and taking a loss feels like admitting a mistake, not because of anything about the investments' prospects.
Is the disposition effect the same as loss aversion?
No. Loss aversion is a feature of how people weigh gains and losses. The disposition effect is a trading pattern that loss aversion was first used to explain. Later research showed that loss aversion alone does not predict it, and beliefs, regret and mental accounting also play a part.
Should you always sell losers and keep winners?
No. The research describes an average tendency and its average cost. A falling stock can be a good holding and a rising one can be overpriced. The lesson is that the purchase price should not decide it, and that in a taxable account a realized loss has a value that a paper loss does not.
Does the disposition effect apply to funds?
For actively managed funds the pattern reverses: investors sell losing funds more readily than winning ones. Researchers attribute this to being able to blame the manager. For individual stocks the ordinary disposition effect holds.
Summary
The disposition effect is the tendency to sell winners and hold losers. Odean measured it in 10,000 accounts: investors realized 14.8% of available gains against 9.8% of losses, the winners they sold went on to beat the losers they kept by about 3.4 percentage points in a year, and only in December, with tax in view, did they sell losers more readily. In a taxable account the habit also pays tax early and leaves deductions unused, about $1,200 on the $10,000 sale in this lesson's example. The explanation is contested and the effect reverses for managed funds, but the behavior in stocks is robust. What reduces it is experience, rules set before the price moves, and judging every position by the same question: would I buy it today?
Check your understanding
5 questions on this lesson. Each answer comes with the reason, and nothing is scored.
In a rising market, 80% of an investor’s sales are of stocks that went up. Why does that not prove a disposition effect?
- Because most positions in a rising market are winners, so most sales would be winners with no preference at all
- Because rising markets cancel behavioral biases
- Because only December sales count
Show the answer
A. Because most positions in a rising market are winners, so most sales would be winners with no preference at all Odean’s ratios fix this by dividing by opportunity: the gains realized as a share of all gains that could have been realized, compared with the same share for losses.
Odean found PGR of 0.148 and PLR of 0.098 for the year, but PGR 0.108 and PLR 0.128 in December. What does December show?
- Investors cannot bring themselves to sell losers in any month
- Investors can sell losers when a clear reason, the tax deduction, is in front of them
- Stocks fall every December
Show the answer
B. Investors can sell losers when a clear reason, the tax deduction, is in front of them The reversal shows the reluctance in other months is a choice. The tax value of a realized loss exists all year; only the deadline makes it visible.
An investor must raise $10,000 from a taxable account. Position A shows a $4,000 long-term gain, position B a $4,000 loss, and she expects the same from both. At a 15% rate with other gains to offset, how much better off is she selling B?
- No difference
- About $600
- About $1,200
Show the answer
C. About $1,200 Selling A costs $600 of tax now. Selling B saves $600 by offsetting other gains. The difference is $1,200, part of which is the timing of the tax on A.
Which finding is hardest to explain if the disposition effect were simply loss aversion?
- Investors sell winners more often than losers
- Investors sell losing actively managed funds more readily than winning ones
- The effect is larger from January to November
Show the answer
B. Investors sell losing actively managed funds more readily than winning ones A pure dislike of realizing losses would apply to funds too. The reversal suggests that admitting one’s own mistake, not the loss itself, is what investors avoid.
Which habit attacks the asymmetry at the heart of the disposition effect?
- Checking prices more often
- Reviewing every position, up or down, against the same written question: would I buy this today with cash?
- Selling whatever has risen most
Show the answer
B. Reviewing every position, up or down, against the same written question: would I buy this today with cash? The bias is treating winners and losers by different standards. One question asked of every position removes the purchase price from the decision.
Exercise
One task to do in Finabase. Nothing is graded or saved.
Open Paper trading (signed out, it shows a sample account; signed in, use your own). In Transaction history find a sale. Compare its fill price with today’s Price for the same symbol in Positions. Then read the Gain % column and decide which position you would sell if you had to raise cash. Write down the number you looked at to decide.
Open Paper tradingWhat you should find
If the number you used was Gain % or Avg cost, your choice was driven by the purchase price, which says nothing about what the stock does next. Check whether the sale you found was of a position that showed a gain at the time: most sales are. The useful question for each row is whether you would buy it today at Price.
Sources and further reading
- Shefrin and Statman — The disposition to sell winners too early and ride losers too long: theory and evidence, Journal of Finance 40(3), 1985
- Odean — Are investors reluctant to realize their losses? Journal of Finance 53(5), 1998
- Weber and Camerer — The disposition effect in securities trading: an experimental analysis, Journal of Economic Behavior and Organization 33(2), 1998
- Feng and Seasholes — Do investor sophistication and trading experience eliminate behavioral biases in financial markets? Review of Finance 9(3), 2005
- Dhar and Zhu — Up close and personal: investor sophistication and the disposition effect, Management Science 52(5), 2006
- Frazzini — The disposition effect and underreaction to news, Journal of Finance 61(4), 2006
- Barberis and Xiong — What drives the disposition effect? An analysis of a long-standing preference-based explanation, Journal of Finance 64(2), 2009
- Ben-David and Hirshleifer — Are investors really reluctant to realize their losses? Trading responses to past returns and the disposition effect, Review of Financial Studies 25(8), 2012
- Chang, Solomon and Westerfield — Looking for someone to blame: delegation, cognitive dissonance, and the disposition effect, Journal of Finance 71(1), 2016
- IRS — Topic no. 409, Capital gains and losses
