Argument from Sunk Cost: "We've Come Too Far to Stop Now"

In 1962, the British and French governments signed an agreement to develop the Concorde supersonic airliner. As costs mounted and the commercial prospects dimmed, both governments repeatedly faced the same decision: continue or stop? Each time, a powerful argument was marshalled for continuing: too much had already been invested to walk away. By the time Concorde entered service in 1976, it was clear it would never be commercially viable. The argument that had sustained it for years — we've put too much into this to quit — was, in retrospect, the reason the project should have been stopped years earlier. The Concorde fallacy now names the broader phenomenon. But calling it a fallacy doesn't fully explain why the argument is so powerful, so prevalent, and so difficult to resist in the moment.

The Argument Structure

The argument from sunk cost, as a pattern of reasoning, takes the following form:

• Investment premise: We have already invested X (time, money, effort, reputation) in course of action C.
• Loss premise: If we stop now, the investment X will have been wasted.
• Continuation conclusion: Therefore, we should continue with C.

This is sometimes stated explicitly — "We've spent $2 billion on this project; we can't pull out now" — and sometimes encoded in seemingly reasonable framings: "We owe it to those who've already given their lives." "After everything we've been through, we can't give up." "The only way to justify what we've spent is to see it through." The investment varies; the logical structure is the same.

The argument is almost always fallacious in its basic form. A sunk cost is, by definition, a cost that has already been paid and cannot be recovered regardless of what happens next. Whether you continue or stop, the $2 billion is gone. Whether you stay in the relationship or leave, the five years are past. The future decision should be made entirely on the basis of future costs and future benefits — the past investment is simply irrelevant to that calculation. For the cognitive bias dimension of this error, see: sunk cost fallacy.

Why the Argument Is Persuasive

The argument from sunk cost derives its persuasive force from several interlocking psychological mechanisms, each of which has been studied extensively in behavioural economics and cognitive psychology.

Loss aversion. Daniel Kahneman and Amos Tversky's prospect theory established that losses feel approximately twice as painful as equivalent gains feel pleasurable. Framing a decision to stop as "writing off" a past investment activates the loss framing — the potential loss looms disproportionately large. See: loss aversion.

Completion pressure. We are motivated to complete things we've started — a tendency studied under the label of the Zeigarnik effect and the endowment effect. Unfinished projects create cognitive and emotional tension; completion provides relief. This drive toward closure can override rational assessment of whether completion is worth the cost.

Identity investment. When we have publicly committed to a course of action, continuing it is not just about recovering costs — it is about maintaining a consistent self-image and avoiding the social cost of having been wrong. Stopping means admitting the original decision was a mistake. The argument from sunk cost often functions as face-saving: "We must continue" is sometimes a coded version of "I cannot admit I was wrong."

Waste aversion. The prospect that a past investment will have been "for nothing" triggers a specific form of discomfort — waste aversion. We are conditioned to avoid waste, and the argument from sunk cost exploits this by framing continuation as preventing waste. But waste has already occurred, or will have, regardless of what happens next. Continuing a bad project does not un-waste the resources already spent; it only adds new waste.

The Argument in Political and Military Contexts

The argument from sunk cost is especially consequential in political and military decision-making, where the invested resources include human lives. The Vietnam War is the paradigmatic case. By 1967, it was clear to many analysts that the US could not achieve its stated objectives. But the death toll — already tens of thousands of Americans, hundreds of thousands of Vietnamese — was repeatedly invoked as a reason to continue. "We cannot let these men have died in vain." This is the sunk cost argument in its most tragic form: past deaths used to justify future deaths, when the only thing the continuation accomplished was adding to the total.

The same structure appears in: colonial projects maintained past their viable point because of "the civilising mission" we'd already sacrificed for; failed states propped up because of aid already delivered; geopolitical commitments maintained past their purpose because of alliances already formed. In each case, the past investment — not the future prospect — is doing the work in the argument.

The argument from sunk cost interacts with practical reasoning here in an important way: the sunk cost is used to substitute for a genuine means-end analysis. Instead of asking "Will continuing achieve the goal?" the argument asks "Have we invested enough that stopping would be wrong?" These are entirely different questions, and conflating them is how wars, projects, and policies outlive their justifications.

The Escalation of Commitment

Psychologist Barry Staw studied what happens when decision-makers are explicitly told that a previous investment was theirs to make and has not been working. His 1976 study, "Knee-Deep in the Big Muddy," found that personal responsibility for the initial investment led to greater escalation of commitment compared to situations where the initial decision was made by someone else. People invested more heavily precisely because they had been the ones who made the original commitment — apparently to justify their past decision to themselves.

This finding has been replicated across domains: managers, investors, students, and government officials all show greater commitment to courses of action they personally initiated, even when presented with evidence that those courses are failing. Staw and Ha Hoang (1995) showed the same effect in professional basketball: NBA teams that drafted players in the first round played those players more even when their performance was inferior to later-round picks, presumably to justify the draft investment.

The Schemed Version: When Past Investment Is Legitimately Relevant

The textbook verdict — sunk costs are irrelevant; ignore them — is right as a first approximation, but it needs qualification. There are situations in which something like the sunk cost argument has genuine force:

Learning investment. If you have invested heavily in acquiring expertise, equipment, or institutional knowledge that will be useful in future projects, that investment is not purely sunk — it has created a capital stock that informs future decisions. The question is whether the investment has created lasting value, not whether it's been paid. This is different from the pure sunk cost.

Reputation and credibility. Organisations and states sometimes need to demonstrate that they will see commitments through, because the credibility of future commitments depends on it. NATO's commitment to defend member states is worth only as much as the demonstrated willingness to bear costs in defending them. Withdrawing from a difficult situation may impose costs on future credibility that are real and relevant — though these are future costs, not past ones, and should be explicitly weighed as such.

Contractual and relational obligations. Having made a commitment to another party creates obligations that don't simply dissolve because it turns out to be costly to honour them. These obligations are real — but they are better understood as future costs (costs of breaking trust, legal liability, reputational damage) than as the sunk cost argument properly understood.

The key distinction: when past investment has created ongoing value — learning, credibility, relationships — that value is a current asset and genuinely relevant to future decisions. The pure sunk cost argument points to investment that creates no such ongoing value. The practical skill is distinguishing these cases honestly rather than using the "ongoing value" framing as a rationalisation for what is really just loss aversion. See: status quo bias.

Identifying the Argument in Practice

The argument from sunk cost can be recognised by certain linguistic markers: "after everything we've invested," "we've come too far to stop," "we can't let this have been for nothing," "all those years," "we owe it to the people who gave their lives." These phrases are almost always symptoms of sunk cost reasoning.

The productive counter-move is not to dismiss the past investment as meaningless — that can seem callous, especially when the investment includes human suffering — but to reframe the decision in terms of the future. "I know we've invested enormously. Given where we are now, and looking forward: will continuing achieve the goal we care about? What will it cost? What are the alternatives? What would we advise someone else to do in this situation, knowing nothing about our past commitment?" That last question is especially powerful: our judgment of others' sunk costs is typically much less distorted than our judgment of our own.

Sources & Further Reading

• Arkes, Hal R., and Catherine Blumer. "The Psychology of Sunk Cost." Organizational Behavior and Human Decision Processes, 35(1), 1985.
• 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(1), 1976.
• Kahneman, Daniel. Thinking, Fast and Slow. Farrar, Straus and Giroux, 2011.
• Thaler, Richard H., and Cass R. Sunstein. Nudge. Yale University Press, 2008.
• Walton, Douglas. Sunk Costs and Argumentation Schemes. In: Argumentation Schemes. Cambridge University Press, 2008.
• Wikipedia: Sunk cost
• Wikipedia: Escalation of commitment