Loss Aversion and the Disposition Effect

Losing $100 hurts about twice as much as gaining $100 feels good. That asymmetry isn't a personal weakness, it's a documented feature of human cognition, and it quietly runs a lot of portfolios.

Last verified 2026-07-16

The asymmetry, and where it was measured

Kahneman and Tversky's prospect theory established that people don't evaluate outcomes against zero — they evaluate them as gains and losses from a reference point, and losses loom roughly twice as large.

That single fact explains a great deal of otherwise puzzling behaviour. It's why a portfolio down 10% feels like an emergency while one up 10% feels like it was always going to happen. Same magnitude, very different weight.

The reference point is the part worth noticing, because it's arbitrary. Your brain treats the price you happened to pay as the line between winning and losing — but the market has never heard of that number, and nothing about a company's future depends on what you paid for it.

What it does to real accounts

Odean tested this against actual brokerage records rather than laboratory choices, and found investors were significantly more likely to realise gains than losses. Selling a winner confirms you were right; selling a loser makes the loss real.

So the pattern is: winners get sold early, losers get held indefinitely. That's the disposition effect, and it's exactly backwards on two counts.

It's backwards on tax, because in a taxable account realising a loss can offset gains while realising a gain triggers tax you could have deferred. And it's backwards on evidence, since nothing about a stock's future is encoded in whether you're currently up or down on it. The unsold loser isn't a loss you've avoided — it's a loss you're still holding.

Why knowing this doesn't fix it

The uncomfortable part of this literature is that awareness barely helps. Kahneman said he continued to make these errors after decades of studying them, and there's no reason to think reading one lesson does what a career of research didn't.

What does work is deciding in advance, when nothing is at stake, and then removing yourself from the moment of decision. Automatic contributions, a written policy for what you hold and why, scheduled rebalancing on a date rather than a feeling.

Those aren't discipline in the willpower sense — they're the opposite. They're an admission that willpower at the moment of maximum emotion is not something to rely on, so the choice gets made while you're calm and then executed by something that isn't you.