The Sunk Cost Fallacy: Why We Can't Stop Throwing Good Money After Bad

You are 90 minutes into a film you are not enjoying. You know it won't improve. Yet you keep watching. After all, you've already invested the time. Or: you've been in a relationship for five years that is clearly not working. You know it in your bones. Yet you stay, because five years is so much to walk away from. Or: your company has spent $50 million on a software project that the engineers now tell you will never work. The rational response is to stop. Instead, you authorise another $20 million, because to stop now would mean admitting the $50 million was wasted. In each case, the past is making the decision for the future — and the future will be worse for it.

What Is the Sunk Cost Fallacy?

A sunk cost is any past investment of time, money, effort, or emotion that cannot be recovered — regardless of what you do next. The sunk cost fallacy is the error of allowing these irrecoverable past investments to influence future decisions, when only future costs and future benefits should be relevant.

The formal economic principle is clear: sunk costs are irrelevant to rational decision-making. A cost that is already paid and cannot be refunded has the same value whether you continue the current course or abandon it. The only question that matters is: given where I am now, which option — continuing or stopping — will produce the better outcome going forward? Past investment doesn't change that calculation. But psychology consistently interferes.

The fallacy was formally described by economists Hal Arkes and Catherine Blumer in a 1985 paper in Organizational Behavior and Human Decision Processes, which demonstrated the effect across multiple experimental scenarios and named it explicitly. The finding was not new to practitioners — engineers, military planners, and project managers had long recognised the problem — but Arkes and Blumer provided the experimental grounding that anchored it in the behavioural economics literature.

The Season Ticket Experiment

Arkes and Blumer's most vivid demonstration: participants were asked to imagine they had accidentally bought two ski trip tickets — one for $50 and one for $100 — and then discovered that the $50 trip was likely to be more enjoyable than the $100 trip. The trips are on the same weekend; you can only go to one. Which do you choose?

Rationally, you should choose the more enjoyable trip — the $50 one. The $100 is already spent; you cannot recover it whether you go or not. The fact that you spent more on the worse trip is irrelevant to which trip will make you happier going forward. Yet most participants chose the $100 trip. They were paying, in future enjoyment, for a past expenditure that was already irrecoverable.

The effect is robust. In a real-world version of a similar experiment, researchers gave participants at a university sports complex either a full-price or a discounted season ticket to athletic events, then tracked attendance across the season. Those who paid full price attended significantly more events than those who received discounted tickets — even though their actual enjoyment of the events showed no difference. The higher payment created a higher psychological investment that drove continued attendance regardless of the objective value of each individual event.

The Psychology Behind the Fallacy

Why does the sunk cost fallacy persist even in people who understand the economics? Several psychological mechanisms are implicated:

Loss Aversion

Abandoning a project means recognising a loss — and according to prospect theory, developed by Kahneman and Tversky, losses loom psychologically larger than equivalent gains. To stop a failing project is to crystallise and acknowledge a loss. To continue it, however irrationally, preserves the possibility — however remote — that the investment might yet be recovered. The continuing project exists in a kind of psychological superposition: the loss isn't quite real yet. Stopping collapses the superposition. The anchoring bias compounds this — the original investment becomes an anchor against which continuations are psychologically compared, distorting the framing of the forward-looking decision.

Commitment and Identity

Substantial investments become entangled with identity and commitment. If you have dedicated three years to a project, some portion of your identity has been invested in that project's success. To abandon the project is not just to lose the investment; it is to accept that you made a wrong decision for three years, that your judgment was flawed, that you pursued the wrong thing. This is psychologically costly in ways that go beyond the financial. The sunk cost fallacy is partly a face-saving mechanism — it allows the investment to continue justifying itself rather than demanding explicit recognition of error.

Waste Aversion

The feeling that stopping constitutes "waste" is deeply culturally embedded. In most human cultures, frugality and not wasting resources are moral virtues. The sunk cost fallacy exploits this. Stopping feels like waste — walking away from something you've paid for, abandoning work you've done. Continuing feels like honouring the prior investment. The moral framing overrides the economic framing, and the waste aversion that would ordinarily prevent bad spending decisions ends up causing them in sunk cost situations.

Wars, Escalation, and the Concorde

The sunk cost fallacy has driven some of the most costly collective decisions in history. The Concorde supersonic airliner — developed jointly by Britain and France from the late 1960s — is perhaps the canonical large-scale example. Engineers and analysts identified early that the project would not be commercially viable. The economic case for abandonment was clear. But both governments continued funding it, year after year, partly because so much had already been invested that stopping felt more expensive than continuing. The plane eventually flew commercially for 27 years before being retired in 2003, having never turned a profit. The "Concorde fallacy" is now a synonym for sunk cost escalation in business and policy contexts.

Military escalation follows the same logic. During the Vietnam War, decision-makers at multiple levels argued against withdrawal partly on the grounds that the investment of American lives and treasure could not be "wasted" by an abandonment of the mission. The logic was: we cannot let the 50,000 who died have died in vain. But the number of deaths incurred is a sunk cost — it cannot be altered by future decisions. The question was whether continued engagement would produce better outcomes than withdrawal, not whether it would retrospectively justify previous losses. The answer, as events showed, was no. Yet the sunk cost logic contributed to several more years of conflict and approximately 20,000 additional American combat deaths, along with vastly greater Vietnamese casualties.

The pattern of "throwing good money after bad" appears wherever large organisations — governments, corporations, militaries — make substantial commitments. The larger the initial commitment, the stronger the psychological pressure to continue, and the more catastrophic the eventual compelled recognition that continuation was always wrong. This is sometimes called "escalation of commitment" — the documented tendency of groups and individuals to increase their investment in a failing course of action in response to negative feedback.

Personal Relationships

The sunk cost fallacy is ubiquitous in personal relationships. People stay in relationships that make them unhappy, that are clearly not working, that both parties know are wrong — because of the time and emotional investment they have made. "We've been together for eight years" is not a reason to continue a relationship that has no future; eight years is a sunk cost. The relevant question is: will the next years of this relationship be better or worse than the alternatives? But the weight of history makes the question feel different — as if the past has earned a future it is not actually likely to provide.

This is compounded by the identity entanglement described above. Long-term relationships become constitutive of identity. Leaving means not just losing the relationship but revising the self-narrative that the relationship has been part of. The psychological cost of this revision is real — but it is a cost of accurately perceiving the situation, not a reason to avoid the perception.

Business and Project Management

In business contexts, the sunk cost fallacy produces the pattern sometimes called "zombie projects" — initiatives that everyone involved knows are failing or failed, but that continue to receive resources because the accumulated investment makes formal cancellation feel too costly. Studies of corporate project management have found that failed projects frequently continue for months or years after the point at which internal assessments identified them as non-viable, consuming resources that could have been redeployed to better uses.

The costs of this pattern extend beyond the project itself. The organisation's attention, talent, and capital that remain locked in the zombie project are unavailable for alternatives. The employees who know the project is doomed but are prevented from saying so develop cynicism and distrust of organisational decision-making. The eventual forced recognition of failure is more damaging than early cancellation would have been.

Good project governance explicitly addresses sunk cost bias by separating the question of past investment from the question of forward-looking viability. Stage-gate processes — which require explicit re-approval at defined milestones based on current prospects rather than past investment — are institutional designs intended to counteract the fallacy. Pre-mortems, in which teams are asked to assume the project has failed and explain why, are another. The challenge is implementing these processes against the organisational dynamics that reinforce sunk cost logic: careers, reputations, and political capital are staked on projects in ways that make admitting failure personally costly even when it is organisationally rational.

The Rational Correction

The correction to sunk cost thinking is, in principle, simple: when evaluating whether to continue a course of action, ask only about future costs and future benefits. The past is irrelevant. The money spent is gone; the question is what the future looks like from here.

In practice, several approaches help:

• Ask: if I were starting fresh today, would I begin this project/relationship/commitment? If the answer is no, that is evidence that sunk costs are distorting your evaluation of a situation that the fresh-start framing reveals more clearly.
• Separate the "waste" question from the forward decision. Yes, stopping means the past investment was "wasted." But continuing doesn't un-waste it — it adds more waste to the pile. The amount already spent cannot be recovered either way.
• Name the sunk cost explicitly. When you notice yourself thinking "but I've already invested so much," identify that as a sunk cost consideration and deliberately exclude it from your decision calculus. Not because it doesn't feel important — it does — but because it has no logical bearing on future outcomes.
• Create accountability structures that reward stopping. In organisations, this means explicitly praising timely project cancellations rather than treating them as failures. In investment, it means setting stop-loss rules in advance, when you are not in the grip of loss aversion about a specific position.

The sunk cost fallacy is ultimately about the past trying to control the future. The rational response is to refuse to let it — not by denying what has been invested, but by recognising that the past is over, and only the future remains to be decided.

Sources & Further Reading

• Arkes, Hal R., and Catherine Blumer. "The Psychology of Sunk Cost." Organizational Behavior and Human Decision Processes 35, no. 1 (1985): 124–140.
• Kahneman, Daniel, and Amos Tversky. "Prospect Theory: An Analysis of Decision under Risk." Econometrica 47, no. 2 (1979): 263–291.
• Staw, Barry M. "Knee-Deep in the Big Muddy: A Study of Escalating Commitment to a Chosen Course of Action." Organizational Behavior and Human Performance 16, no. 1 (1976): 27–44.
• Thaler, Richard H. Misbehaving: The Making of Behavioral Economics. W. W. Norton, 2015.
• Kahneman, Daniel. Thinking, Fast and Slow. Farrar, Straus and Giroux, 2011.
• Wikipedia: Sunk cost