Mistake Lab: Selling Winners Too Early

By Swoopr Editorial Team Published 2026-09-17 Research-assisted analysis. Verify sources before acting.

Key Takeaways

• Odean's 1998 study (10,000 accounts, 1987 to 1993): sold winners outperformed held losers by 3.4% over the next year, confirming the disposition effect imposes a measurable return cost.
• Behavioral roots: prospect theory (Kahneman and Tversky) predicts this pattern; gains feel less good than equivalent losses feel bad, so investors lock in small gains and delay realizing small losses.
• Tax consequence: selling a winner early triggers a taxable gain; if the gain is short-term (held under 1 year), it is taxed at ordinary income rates (up to 37%) vs. long-term capital gains rates (0%, 15%, or 20%). Selling too early often maximizes the tax burden.
• Most damaging for momentum-driven assets: if a stock is rising because of a genuine improvement in business fundamentals, selling early cuts off the right tail of the return distribution when the fundamentals are still improving.
• Process-based antidote: make sell decisions based on a pre-defined rule (target price, valuation threshold, changed thesis) rather than the current gain percentage. The current gain is irrelevant to the future expected return.

Why Investors Sell Winners Early

Prospect theory (Kahneman and Tversky, 1979) shows that people experience gains and losses asymmetrically: a $1,000 gain produces less psychological satisfaction than a $1,000 loss produces pain. This asymmetry creates a strong pull to lock in gains (certainty of a known gain is appealing) and defer losses (avoiding the psychological pain of realizing a loss). The mental accounting framework reinforces this: investors tend to evaluate each position in isolation rather than as part of a portfolio, so a 20% gain feels like 'mission accomplished' independent of whether the underlying thesis has played out. The emotional impulse is to 'take profits' before the gain disappears, even when no rational argument supports selling.

The Return Cost of Early Selling

Odean's data showed that the average sold winner returned 2.35% over the 252 trading days following the sale; the average held loser returned negative 1.06% over the same period. The difference (3.41%) represents the return the investor sacrificed by selling the winner and holding the loser. This cost appears across multiple studies in multiple countries (Germany, Finland, China, Taiwan) with similar magnitudes. The cost is not uniform: it is largest for stocks with strong recent momentum (where the thesis for continued appreciation is strongest) and smallest for mature, dividend-paying companies where mean-reversion is more likely.

The Tax Amplification Effect

Selling a winner early has a compounding tax cost beyond the direct return sacrifice. In a taxable account: selling a position held less than one year triggers short-term capital gains tax (taxed as ordinary income); if held over one year, the long-term capital gains rate applies (significantly lower). An investor selling a $10,000 position with $5,000 gain after 11 months in the 32% marginal bracket pays $1,600 in taxes (32% of $5,000). Waiting one more month reduces the tax to $750 (15% of $5,000 long-term rate). Beyond the long-term threshold, the tax argument for selling a winner still further weakens: each year of additional holding defers the tax liability and allows that deferred tax dollar to compound.

Building a Sell Discipline

The antidote to early selling is not resolving to 'be more patient' (emotional resolutions are overridden by the disposition effect in the moment) but installing process-based rules that remove discretion. Effective sell rules: thesis invalidation (if the reason you bought the stock no longer holds, sell regardless of current gain or loss); target price reached (set a rational price target when buying based on valuation; if met, evaluate whether the target should be raised given new information rather than automatically selling); position sizing rebalance (if a position grows beyond a target allocation percentage, trim mechanically to target). Crucially, the rule should not include 'I've made X% so I should sell' -- the gain percentage is not a criterion for selling.