The Disposition Effect: Why You Sell Your Winners and Hold Your Losers
The disposition effect is the well-documented tendency to sell profitable positions too early and hold losing ones too long — the opposite of what's optimal for returns or taxes. What the research actually found, why it happens, and how to catch it in your own portfolio.
The disposition effect is the tendency to sell winning investments too early and hold losing ones too long — realizing gains quickly while letting losses sit open, unrealized, sometimes for years. It was named and documented by finance professors Hersh Shefrin and Meir Statman in a 1985 paper in The Journal of Finance, and it remains one of the most replicated findings in behavioral finance. If you’ve ever taken quick profit on the one position that worked while a pile of red ones sat untouched in the same account, you’ve done this.
It matters more than most cognitive biases covered in this space because it’s not really about feeling — it’s about a specific, measurable action: which positions you click “sell” on. That makes it unusually easy to check against your own trade history, which is what most of this article is for.
The study behind it
The most cited empirical test of the disposition effect is Terrance Odean’s 1998 paper, “Are Investors Reluctant to Realize Their Losses?”, published in The Journal of Finance. Odean analyzed real trading records from roughly 10,000 accounts at a large U.S. discount brokerage between 1987 and 1993, and measured two numbers for each investor: the proportion of available gains they actually realized (sold), and the proportion of available losses they actually realized.
The result: investors realized about 14.8% of their available gains, but only 9.8% of their available losses — meaning they were roughly 50% more likely to sell a position that was up than one that was down, holding position size and other factors constant. The one clean exception was December, when tax-loss selling temporarily reversed the pattern as investors harvested losses to offset gains before year-end — itself a telling detail, since it shows the behavior isn’t fixed, it’s overridden the moment a stronger, more concrete incentive (an actual tax deadline) shows up.
Odean also checked whether the held losers were quietly the better bet — whether investors were, in effect, right to hold them. They weren’t: the losing positions investors kept underperformed the winning positions they’d sold, over the following year. The behavior wasn’t informed patience. It cost money.
Crypto shows the same pattern — with a twist
This isn’t just an old finding about 1990s stockbrokers. A 2023 study published in Digital Finance examined on-chain Bitcoin transaction data from 2013 through 2021, using transfers to exchange wallets as a proxy for sell decisions, and found the same disposition effect present in Bitcoin trading — and that it grew markedly more pronounced after the 2017 boom-and-bust cycle brought in a wave of newer, less experienced investors.
The twist worth knowing if you trade crypto: some Bitcoin-specific research has found the pattern isn’t perfectly symmetric with stocks. In strongly bullish stretches, some studies observe traders doing the opposite — riding winners too long, hoping for more — while the classic pattern (dump the winner, cling to the loser) shows up more reliably once the market turns bearish. Either direction is the same underlying failure: letting the feeling of a position, rather than its forward-looking merit, decide when you sell.
Why it happens
Three mechanisms compound to produce this, and none of them require you to be a bad investor — they’re default wiring:
Mental accounting. Your brain treats each position as its own closed storyline with a beginning (your entry price) and, ideally, a happy ending (a profit). Selling at a loss forces that storyline to end unhappily and permanently. Selling at a gain lets you close the book on a win. You’re not weighing the two positions against each other on their merits going forward — you’re managing two separate narratives, and one of them you want to finish and one you don’t.
Regret aversion. Selling a loser locks in a concrete, undeniable mistake with your name on it. Holding it costs nothing today — the loss already happened on paper regardless — but selling converts a vague, deferrable regret into a specific, dated one. Holding feels like avoiding the regret. It’s actually just postponing the admission.
A reference point that won’t update. Your purchase price becomes an anchor that the position is measured against long after it’s stopped being informative about the asset’s future. (This is covered in more depth in Anchoring Bias: Why You Can’t Stop Comparing the Price to What You Paid.) As long as the position is below that anchor, closing it feels like a loss; the moment it crosses back above, closing it feels like a win — and that switch happens at your cost basis, a number the market has no idea you’re using.
A worked example
You hold two positions, each originally $3,000. Position A is now worth $4,200 (up 40%). Position B is now worth $1,800 (down 40%). You need $3,000 in cash for an unrelated expense.
The disposition effect makes the “obvious” move to sell A: lock in the win, keep believing in B. But nothing about needing cash has anything to do with which position has better forward prospects. Run the same test used elsewhere in this framework — would you buy each one today, at today’s price, for its own merits? If A still looks like the better forward bet, selling it anyway just to fund the cash need while keeping B is the disposition effect making the decision, not you. The correct move is to evaluate A and B identically, ignoring which one is currently green and which is currently red, and sell whichever one you’d be less willing to buy today.
What the disposition effect actively costs you
| Where it shows up | What the bias makes you do | What’s actually optimal |
|---|---|---|
| Portfolio composition | Winners get sold off, losers accumulate — your holdings drift toward your worst ideas | Hold based on forward thesis, not current P&L |
| Taxes (most jurisdictions) | Realizing gains (often taxable) while sitting on unrealized losses that could offset them | Realize losses first when tax-loss harvesting is available — see the tax-loss harvesting basics below |
| Returns | Held losers underperformed the sold winners in Odean’s data | Same evaluation standard for every position, independent of its current color |
| Risk | Portfolio concentration quietly increases in your weakest, most underwater positions | Position sizing driven by conviction and risk, not by which ones you can’t bear to sell |
The tax row is the sharpest irony: the disposition effect pushes you toward the exact opposite of what tax-loss harvesting asks you to do. Harvesting wants you to sell losers to realize a deductible loss and keep winners running; your instincts want to do the reverse. If you haven’t looked at this from the tax side, Tax-Loss Harvesting for Crypto: The Basics Nobody Explains Simply walks through how the mechanics actually work.
A five-minute self-audit
- Pull your last 10 closed positions (sold, not just currently held).
- Mark each one: was it sold at a gain or a loss relative to your cost basis?
- For the gains, note how long you held before selling. Do the same for the losses.
- Compare: are your winners closing faster than your losers, on average?
- For anything still open and red, apply the buy-today test from the worked example above — not the “it has to come back” test.
If your losers are open noticeably longer than your winners were, that’s the disposition effect showing up in your own numbers, not a hypothetical. The fix isn’t a new rule for every trade — it’s holding both a winner and a loser to the same standard: would you buy it today, at today’s price, on its own merits, independent of what it’s already done for or to you.
FAQ
Is the disposition effect the same thing as loss aversion? They’re related but not the same mechanism. Loss aversion describes how a loss feels worse than an equivalent gain feels good — it’s about the emotional weighting of an outcome. The disposition effect is the specific selling behavior that loss aversion (combined with regret aversion and mental accounting) tends to produce: closing winners quickly to lock in the good feeling of a realized gain, while holding losers open so the loss never has to become official. You can think of loss aversion as one of the engines and the disposition effect as the behavior it drives at the moment you decide what to sell.
Doesn’t it make sense to sell winners and let losers run if you think the loser will recover? That’s the rationalization, and sometimes it’s even true for a specific position — but the research tests this directly, and it doesn’t hold up on average. Odean’s study found that the losing positions investors held on to underperformed the winning positions they sold, over the following year. If holding losers were driven by genuine, well-founded conviction that they’d outperform, you’d expect the opposite result. The pattern looks much more like an unwillingness to realize a loss than a considered bet that laggards are about to lead.
How do I know if I’m doing this, versus just holding for a legitimate reason? Look at your last 5-10 closed positions. If you sold the ones that were up and are still sitting on the ones that are down — especially if the down ones have been open noticeably longer than the up ones — that’s the pattern. The honest test from the framework in this article: would you buy each losing position today, at today’s price, for the reason you’re currently giving for holding it? If the reason keeps circling back to “it has to come back” rather than a live thesis, it’s the disposition effect, not conviction.