The disposition effect
Investors sell their winners and hold their losers, in a measurable ratio, in hundreds of thousands of real accounts — and the pattern reverses in December for a reason that shows the rest of the year is a mistake.
Chapter 3 · Beginner
Chapter 2's value function predicts something specific: investors should be readier to sell what is up than what is down. This chapter is the test, and it is a good example of how a behavioural claim gets established.
How you measure it
You cannot simply count sales of winners, because people hold more winners in a rising market. The measure has to be a rate.
Proportion of gains realised — realised gains divided by realised gains plus paper gains. Of all the winning positions you could have sold, what share did you sell?
Proportion of losses realised — the same for losing positions.
If the purchase price were irrelevant, the two rates would be roughly equal.
What the data say
Odean examined 10,000 accounts at a large discount brokerage from 1987 to 1993.
| Gains realised | Losses realised | |
|---|---|---|
| Entire year | 0.148 | 0.098 |
| January–November | 0.152 | 0.094 |
| December | 0.108 | 0.128 |
For the full year, winners were realised at about one and a half times the rate of losers, on 13,883 realised gains against 11,930 realised losses, with 79,658 paper gains and 110,348 paper losses behind them. The difference is 0.050 with a t-statistic of 35 — not a marginal result.
Then December reverses it. Losses are realised more often than gains, and the sign of the effect flips.
Why December matters more than the headline
The reversal is the most informative row in the table, because it rules out a whole class of explanations at once.
In a taxable account, realising a loss can reduce tax. The incentive is strongest as the tax year closes. So in December a financial reason to sell losers appears, and investors act on it.
That tells you they can sell losers when they have a reason to. The reluctance for the rest of the year is therefore not an inability, a constraint, or a liquidity problem. It is a preference that a tax incentive is strong enough to override — which means for eleven months of the year investors are leaving a real benefit on the table. Odean's abstract puts the consequence plainly: for taxable investments the behaviour "is non-optimal and leads to lower after-tax returns".
Indian investors should read that mechanism rather than the specific tax rule, since the rates and the set-off rules here are their own subject. The structural point travels: a realised loss can have value, and an unrealised one has none.
Working the problem
Four innocent explanations, and how each fails.
Rebalancing. If a position has grown, selling some of it restores your intended weights, which would produce exactly this pattern with no psychology involved. Ruled out because rebalancing predicts partial sales of appreciated positions; the effect persists when you look at positions sold entirely. It also predicts buying more of what has fallen, which is not what is observed.
Trading costs. Losers may have fallen to low prices, where percentage costs are higher, so avoiding selling them could be rational. Ruled out in the paper directly — the behaviour is not explained by the higher trading costs of low-priced stocks.
Skill. Perhaps investors keep losers because those losers are about to recover. Ruled out by what happens next: the behaviour is "not justified by subsequent portfolio performance". The losers they kept did not go on to outperform the winners they sold.
Liquidity. Needing cash forces some sale, but it does not explain which. Faced with one winner and one loser and a need for cash, the prediction is that the winner goes — and that is the finding, not an escape from it.
What survives is the reference-point account: the purchase price divides holdings into two categories that the asset itself does not recognise.
Why it is expensive
It inverts the question you should be asking. The useful question is which holding has the worst prospects. The disposition effect answers a different one — which holding is below what I paid — and the two have no reliable relationship.
It concentrates your portfolio in your mistakes. Sell winners, keep losers, repeat, and over years the portfolio drifts toward the positions you were most wrong about.
It feels like discipline. "I don't sell at a loss" has the cadence of a rule, which is why it survives. It is a rule about your own accounting, not about the investments.
The point
Across 10,000 accounts, investors realised gains at 0.148 and losses at 0.098 — winners sold about half again as often as losers — and the pattern reversed in December when tax gave them a reason to sell losers. That reversal is what makes it a finding rather than an artefact: they could sell losers all along. Rebalancing, trading costs, liquidity and skill each predict something extra that the data do not show, leaving the purchase price as the thing doing the sorting.
Check yourself
4 questions. Every answer is explained afterwards, including the ones you get right — guessing correctly is not the same as knowing. Score 70% or more and the chapter is marked done.
Question 1 of 4
0 of 4 answered. You can submit with questions unanswered — they simply score zero.
Now do it with your own numbers
You find that investors realise gains about one and a half times as often as losses. Before concluding anything about psychology, list the innocent explanations that could produce this pattern, and say how you would rule each one out.
Rebalancing, trading costs, and genuine skill at picking which to keep are all candidates. Each makes a different additional prediction.