The Disposition Effect Is Selling Winners and Keeping Losers
Investors realise gains far more readily than losses. It is one of the most consistently documented patterns in individual trading data, and it is backwards on both economics and tax.
The Finding
Study individual brokerage records and a consistent pattern appears. Given a portfolio containing both winners and losers, investors sell winners at a substantially higher rate.
This is the disposition effect, and it has been documented across countries, asset classes, decades, and among professional as well as individual investors. It is one of the more replicable findings in behavioural finance.
Where It Comes From
The mechanism follows directly from prospect theory. A position is evaluated relative to its purchase price. A winner sits in the gain region, where people are risk averse and prefer to lock in a certain outcome. A loser sits in the loss region, where people become risk seeking and prefer the gamble of holding on.
There is a second, simpler component. Selling a loser converts a paper loss into a realised one, which requires admitting the decision was wrong. Holding it preserves the possibility that it was merely early.
An unrealised loss is a hypothesis that has not been tested yet. Selling closes the question, and closing the question is the part people avoid.
Why It Is Backwards
The behaviour is wrong on two separate grounds.
On tax, in a taxable account, selling winners triggers a capital gains liability while selling losers generates a deductible loss. The tax optimal behaviour is the exact opposite of the observed behaviour, and the cost of getting it backwards compounds annually.
On returns, the evidence suggests the sold winners subsequently outperform the retained losers. Momentum, whatever its cause, means recent relative strength has tended to persist over intermediate horizons. Selling what is working to keep what is not runs directly against that.
| Action | Tax effect | Evidence on subsequent returns |
|---|---|---|
| Sell winner | Triggers taxable gain | Tends to keep performing |
| Hold loser | Forgoes deductible loss | Tends to keep lagging |
The Round Trip Illusion
A related distortion is the fixation on getting back to even. A position down 40 percent requires a 67 percent gain to recover, which is arithmetic many holders never work out explicitly.
More importantly, breaking even is not an investment objective. The relevant question is whether this capital, at its current value, is better deployed here or elsewhere. The entry price has no bearing on that comparison, and organising a decision around it means allocating capital according to a personal historical accident.
Where It Appears Professionally
Portfolio managers show a muted version of the same pattern, and there is an additional institutional driver. Realised losses appear in reports, while unrealised ones sit quietly in the holdings. A manager approaching a reporting date has a visible incentive to defer recognition.
The related phenomenon of dressing up a portfolio before a reporting period, selling embarrassing positions and buying respectable ones, is the same instinct expressed for an audience.
The Practical Corrections
Rebalance on a schedule rather than on judgement. A calendar rule sells what has grown and buys what has shrunk without asking how you feel about either.
Harvest tax losses deliberately, treating them as an asset to be captured rather than a defeat to be avoided.
Set exit conditions when entering, expressed as changes in the business or thesis rather than as price levels relative to your cost.
And apply the ownership test to every holding: would you buy this today, at this price, with cash. Positions that fail that test are being retained by the entry price rather than by analysis.
The Bottom Line
The disposition effect is the well documented tendency to sell winners and hold losers, driven by risk seeking in the loss region and by reluctance to confirm a mistake. It is wrong on tax and appears to be wrong on subsequent returns. The fixes are mechanical: scheduled rebalancing, deliberate loss harvesting, and exit rules written before the position was opened.