The sunk cost fallacy in investing: why past losses should not drive future decisions

The sunk cost fallacy is the tendency to continue a course of action because of resources already committed to it, rather than because of its future prospects. In investing, it manifests as holding a losing position because the investor does not want to realise the loss—even when the forward-looking expected value of the position is negative and the capital would be better deployed elsewhere. The cost is already sunk; it cannot be recovered by continuing to hold. Yet the psychological pull of past costs on future decisions is powerful and well-documented.

What the sunk cost fallacy is

The economic principle is unambiguous: a rational decision-maker should consider only future costs and benefits when evaluating whether to continue an activity. Past expenditures that cannot be recovered are irrelevant to the expected value of future actions. A company should abandon a project if its remaining expected value is negative, regardless of how much has already been invested. An investor should sell a position if the forward-looking case no longer holds, regardless of what the position originally cost.

Thaler (1980) formalised the sunk cost effect as a departure from this rational benchmark, noting that individuals systematically allow unrecoverable past costs to influence future behaviour. The mechanism is closely related to loss aversion—the tendency, documented by Kahneman and Tversky (1979), to feel the pain of a loss approximately twice as strongly as the pleasure of an equivalent gain. Selling a position at a loss makes the loss concrete and irreversible; continuing to hold preserves the possibility—however slim—that the position might recover. Loss aversion makes the certain loss from selling feel worse than the uncertain future of continuing to hold.

How it manifests in investing

The most common manifestation is the reluctance to sell losing positions. An investor buys a stock at USD 100. The stock falls to USD 60 following adverse news about the company's fundamentals. The rational question is: given what the stock is worth now and what I expect it to return in the future, would I buy it today at USD 60? If the answer is no, the position should be sold. But many investors instead hold because they do not want to realise the loss, implicitly asking a different question: how long until it gets back to USD 100?

A related pattern is doubling down—investing additional capital into a losing position to reduce the average cost. "Averaging down" can be rational if new information has improved the investment case. It is irrational if the motivation is purely to lower the cost basis so that the position returns to break-even sooner. In the second case, the investor is using new capital not to pursue the best available opportunity but to fight a psychological battle against an irrelevant number.

Shefrin and Statman (1985) identified what they called the disposition effect: the tendency of investors to sell winners too early and hold losers too long—which is the sunk cost fallacy expressed as an asymmetric holding pattern. Odean (1998) confirmed this in a study of 10,000 individual brokerage accounts over 1987–1993, finding that investors were significantly more likely to sell positions at a gain than positions at a loss, even when accounting for tax incentives. The stocks that investors held (the losers) subsequently underperformed the stocks they sold (the winners), confirming that the disposition effect reduced returns.

The cost

The quantifiable cost of sunk cost thinking in investing is not limited to holding underperforming positions. It also includes the opportunity cost of capital deployed suboptimally—capital tied up in a position held for sunk cost reasons is capital unavailable for better-returning alternatives. Odean (1998) found that, on average, stocks investors sold outperformed stocks they held by approximately 3.4 percentage points over the following year, suggesting that the disposition effect—and the sunk cost reasoning underlying it—had a direct and measurable negative impact on returns.

What helps

The most effective structural defence against sunk cost thinking is to remove the cost basis from the decision frame entirely. Pre-committed exit rules—sell if the asset falls more than X% from peak, or sell if it underperforms its benchmark for Y consecutive months—define exit criteria before the position is entered and on a basis that is independent of the purchase price. A stop-loss rule triggered by a price level rather than a loss percentage relative to cost achieves this: it makes the exit condition forward-looking rather than cost-anchored.

Systematic, rules-based investing addresses the sunk cost fallacy structurally. When allocation decisions are made by an algorithm that evaluates each asset's current momentum, carry, and value characteristics—without any reference to historical cost—the cost basis cannot enter the decision. The algorithm neither knows nor cares what the asset was purchased at; it evaluates the forward-looking case based only on current information. This is one of the clearest advantages of systematic strategies over discretionary ones: the exit decision is made before the emotional pressure of a loss arises.

Related articles

• Loss aversion in investing: why losses hurt twice as much as gains help
• The disposition effect: why investors sell winners too early and hold losers too long
• Systematic vs discretionary investing: rules, flexibility, and the evidence on which wins
• Prospect theory: how gains and losses are actually valued in investment decisions