Open any brokerage’s data on retail behavior and one pattern appears with embarrassing regularity: investors sell their winning positions far more readily than their losing ones. Odean’s classic study of ten thousand accounts put numbers on it: the stocks people sold went on to outperform the losers they kept. Peter Lynch had already named the behavior: cutting the flowers and watering the weeds.
This is the disposition effect, and it is worth dissecting because, unlike some biases, it produces a paper trail you can audit in your own account.
Why the losing position is so hard to sell
Three gears mesh here. The first is loss aversion’s accounting rule: a loss on paper is pain pending; a loss realized is pain booked. Selling the loser converts a maybe into a fact, and the brain will pay real money to keep a fact from existing. Hence the universal prayer of the underwater investor: “I’ll sell when it gets back to what I paid.”
Look at that sentence closely — it is the second gear. Your purchase price is a number that exists in exactly one place in the universe: your records. The market does not know it and cannot care. A stock sitting at 60 after your buy at 100 has no memory, no obligation, and no tendency to revisit 100 because you’d find that emotionally tidy. Anchoring on cost turns a random historical accident (the day you happened to click) into the reference point for all future decisions.
The third gear is the sunk-cost instinct: having “invested so much” (money, time, public conviction), abandoning the position feels like waste. But the money is spent either way. The only question any position ever answers is forward-looking: knowing what I know today, would I buy this, at this price? If no, holding it is buying it, every single day, with extra steps.
Why selling the winner feels so good
Symmetrically, the winner offers a treat: sell, and pride is booked — a win made permanent, safe from future reversal. So the position with momentum, the one whose thesis is working, gets harvested early to lock in the feeling, while capital reallocates toward the position whose thesis is failing. Run that policy for twenty years and you have systematically transferred money from your successes to your failures. Taxes twist the knife: in most systems, realizing gains triggers tax while realizing losses offsets it, so the disposition effect has you paying extra for the privilege of underperforming.
The audit, and the fix
Run the audit once. It takes fifteen minutes and three columns:
- List your last ten sells from your broker’s history — ticker, date, and whether the position was green or red on the day you sold.
- Count the colors. An unbiased seller’s ledger should look roughly like their portfolio; most people find eight greens, two reds.
- For each red you kept instead of selling, write the reason you gave yourself at the time. If the phrase “when it gets back to” appears more than once, you have met your own disposition effect in its natural habitat.
Most people find a lopsided ledger and, behind it, the honest motive: sells timed to harvest feelings, not to serve a plan.
The fix is structural, as always in this column. Index-and-policy investors largely amputate the problem: a broad portfolio has no individual weeds to water, and its only sell rules — rebalancing and eventual spending — reference targets and dates, never your cost basis. If you do hold individual positions, write sell conditions at purchase (the kill condition again), and grade every holding by one question with the purchase price surgically removed: buy it today, at today’s price, or not?
Whatever you paid is history’s business. Your portfolio’s only business is tomorrow.