Mental Models a field guide

Trap/Cognitive Bias/No. 1003

Sunk Cost Fallacy

The sunk cost fallacy is the tendency to continue an activity because of money, time or effort already spent, even when future costs outweigh expected benefits. Named in economics by Richard Thaler, it is also called the sunk cost effect or Concorde fallacy.

Also called Sunk Cost Effect · Concorde Fallacy

a trap: easy to walk into
Evidence
Well established
Read
5 min
Links
13 connections
Useful when
Deciding under uncertainty · Habits and self-control · Money and investing · Running projects

01You've seen this when…

  1. in life

    Forty minutes into a film you’re not enjoying, you stay to the end, because leaving now would mean the first forty minutes were wasted.

  2. at work

    A project is eight months in and the numbers have stopped adding up. Someone in the review says, “We can’t stop now. Look how much we’ve put in.”

  3. out in the world

    A council keeps funding a half-built ring road that no longer makes sense, because cancelling would “throw away” the money already spent.

02The idea

Money, time or effort that you’ve already spent and can’t get back is a sunk cost. Whatever you decide next, it stays spent. So it shouldn’t count either for or against carrying on.

The only question that matters looks forward: from here, is what’s still to come worth what it will still cost? The fallacy is letting the past spending answer that question for you. “We’ve come this far” feels like a reason, but it only describes how you got here.

A quick test: imagine you’d inherited the situation this morning with nothing invested. Would you choose to start? If the answer is no, the past investment is the only thing keeping you there.

03Why it happens

  • Stopping turns a cost into a loss. As long as you continue, the spending still feels like an investment. Stopping makes it a loss, and losses sting more than equal gains please. That’s loss aversion.
  • Quitting looks wasteful. Arkes and Blumer found that much of the effect comes from the wish to avoid appearing wasteful. Abandoning something feels wasteful even when continuing would waste more.
  • Stopping admits a mistake. Walking away says the original decision was wrong. For whoever championed it, carrying on protects their judgment and reputation. Repeated over time, this becomes escalation of commitment.
  • Effort makes things feel valuable. The more something cost us, the more we believe it must be worth. That’s effort justification.

04A worked example

A product team has spent nine months building a recommendation engine. The pilot shows a lift of 0.3%, which is within the noise. Finishing it properly would take another four months. In the review, the lead says: “We’re seventy percent done. Stopping now throws away nine months of work.”

What it looks like Prudence. Finishing what you started, protecting an investment.

What’s actually going on The nine months are gone whichever way the team decides. The choice on the table is four more months on this or four months on the next-best thing. Framed that way, the question becomes: would we start a four-month project today that promises a lift we can’t yet detect? Almost certainly not. That’s the opportunity cost the sunk cost was hiding.

What would have helped Agreeing on stop criteria at the start (“if the pilot lift is under 1%, we stop”), so that stopping is simply following the plan. And asking someone outside the original decision to make the call.

05How to spot it

06What to do instead

  • Ask the clean-slate question. If we were starting today, with what we know now, would we choose this?
  • Compare against the best alternative. Stopping frees the budget for something else, so every further month here is a month taken from the next-best use.
  • Set stop rules in advance. Write them while you’re still neutral: “If X hasn’t happened by date Y, we stop.”
  • Separate the decision from its author. Someone with no stake in the original call sees the remaining costs more clearly.
  • Count what stopping gives you. The freed-up budget and attention belong on the scale too.

07When honoring the past isn’t a fallacy

Persistence can be the rational choice. The past can carry information, and quitting has costs of its own:

  • The remaining cost is small. Finishing the last tenth of something can pay off precisely because only a tenth is left. That’s forward-looking reasoning.
  • Quitting has its own price. You might break a promise or damage your reputation. Going to the concert because your friends are counting on you is about your friends, not the ticket.
  • The spend was a signal. A high price can tell you something about quality you’d otherwise have to guess.

Economists McAfee, Mialon and Mialon argue that many apparent sunk-cost cases can be explained this way. The fallacy is specifically when the amount already spent does the arguing.

08Roots

The advice is older than the name. Economics textbooks have long told students to ignore costs they can’t recover. Richard Thaler gave the mistake a place in behavioral economics in 1980, and Hal Arkes and Catherine Blumer showed in 1985 how readily ordinary people make it.

Biologists named it first. In 1976 Richard Dawkins and Tamsin Carlisle called it the Concorde fallacy, after the supersonic airliner whose backers kept paying because they had already paid so much. They were warning their own field: it’s tempting to assume an animal guards a nest harder because of what it has already put in, but natural selection only cares about what’s still to gain.

That raises a fair question: is this an old instinct we share with other animals? Arkes and Peter Ayton reviewed the evidence in 1999 and found that animals rarely seem to commit it. Their suggestion was that people over-apply a rule that usually serves them well, don’t waste, to cases where carrying on is the bigger waste. If they’re right, the fallacy is a good habit stretched too far.

09How solid is this?

ContestedMixedUsefulEstablished

Shows up reliably when people answer hypothetical scenarios, and in field data on theater tickets and car use. In lab experiments with real money the effect is often smaller and less consistent than its reputation, and some persistence is perfectly rational.

10Connections

confused withconfused withconfused withconfused withcountered bycountered bycountered bycountered bycountered byfollows fromSunk CostFallacyEscalationof CommitmentPath DependenceStatus Quo BiasSwitching CostsCommitmentDeviceCost-BenefitAnalysisMarginalAnalysisOpportunityCostOptimalStoppingLoss Aversion

+ 3 more in the list

11Origin and sources

Named in economics by Richard Thaler (1980). Hal Arkes and Catherine Blumer (1985) made the classic psychological case. Biologists Richard Dawkins and Tamsin Carlisle (1976) called it the Concorde fallacy.

  1. [1]Thaler, R. (1980). Toward a positive theory of consumer choice. Journal of Economic Behavior & Organization, 1(1), 39–60.
  2. [2]Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124–140.
  3. [3]Dawkins, R., & Carlisle, T. R. (1976). Parental investment, mate desertion and a fallacy. Nature, 262, 131–133.
  4. [4]Arkes, H. R., & Ayton, P. (1999). The sunk cost and Concorde effects: Are humans less rational than lower animals? Psychological Bulletin, 125(5), 591–600.
  5. [5]Friedman, D., Pommerenke, K., Lukose, R., Milam, G., & Huberman, B. A. (2007). Searching for the sunk cost fallacy. Experimental Economics, 10(1), 79–104.
  6. [6]McAfee, R. P., Mialon, H. M., & Mialon, S. H. (2010). Do sunk costs matter? Economic Inquiry, 48(2), 323–336.
  7. [7]Ho, T.-H., Png, I. P. L., & Reza, S. (2018). Sunk cost fallacy in driving the world's costliest cars. Management Science, 64(4), 1761–1778.

Suggest an edit· Updated 2026-10-01