Sunk Cost Fallacy

Core Idea

The sunk cost fallacy is the tendency to continue investing in a project, relationship, or course of action because of previously invested resources (time, money, effort) rather than based on future expected value. Standard economics dictates that sunk costs — costs already incurred that cannot be recovered — should be irrelevant to forward-looking decisions. Yet people routinely "throw good money after bad," continuing failing projects to justify past expenditures. The fallacy is driven by loss aversion (abandoning the investment makes the loss "real"), mental accounting (wanting to close the mental account in the positive), waste aversion (reluctance to "waste" invested resources), and self-justification (admitting the investment was wrong threatens self-concept). It produces escalation of commitment in business, government, and personal decisions.

Explainer

"We've come too far to stop now." This sentiment — utterly natural, viscerally compelling, and economically irrational — captures the sunk cost fallacy. The logic of rational choice is unequivocal: past expenditures that cannot be recovered should have no influence on future decisions. Only future costs and future benefits matter. Yet the pull of sunk costs is powerful enough to drive individuals to finish terrible movies ("I paid for the ticket"), companies to continue failed projects ("we've invested too much to walk away"), and governments to persist in failing policies ("the sacrifices so far would be wasted").

The connection to loss aversion is direct. Abandoning an investment makes the loss concrete and undeniable — the mental account is closed in the red. Continuing the investment preserves the possibility, however remote, that the account will eventually show a positive balance. Loss aversion makes the certain loss of stopping feel worse than the expected loss of continuing, even when expected-value analysis favors stopping. This is compounded by the way mental accounting frames the decision: it is not "should I invest another $1 million in this project?" (the economically correct framing) but "should I write off the $10 million I've already spent?" (the psychologically natural but irrelevant framing).

Waste aversion — the reluctance to "waste" resources — is a related but distinct driver. People who have paid for a gym membership feel compelled to go to the gym even when they would prefer to rest, not because they will get the money back by going, but because not going makes the expenditure feel "wasted." The money is equally gone whether they go or not — but going maintains the narrative that the money was well spent. This drive to avoid the feeling of waste is psychologically real even when it leads to economically irrational behavior.

Escalation of commitment is the organizational manifestation of the sunk cost fallacy. Classic case studies include the Concorde (continued despite clear evidence that it would never be commercially viable, earning the sunk cost fallacy the alternate name "Concorde fallacy"), the Vietnam War (arguments against withdrawal centered on "honoring the sacrifices already made"), and numerous corporate product launches that continued long past the point where market feedback indicated failure. Staw's research on escalation showed that the effect is stronger when the decision-maker is personally responsible for the initial investment, when the decision is public, and when there are organizational norms against "quitting."

The practical remedy is straightforward in principle but difficult in practice: separate the evaluation of future prospects from the accounting of past expenditures. Pre-commitment to decision criteria (before investing, define what evidence would trigger stopping), external review (bringing in evaluators who have no stake in the original decision), and organizational norms that celebrate cutting losses (rather than stigmatizing "quitters") can help. Some organizations formally assign a "devil's advocate" whose role is to argue for termination of projects, specifically to counterbalance the natural tendency toward escalation. The challenge is that these measures require organizations to design against human psychology — which is possible but requires deliberate effort.