Sunk Costs

Cover of Sunk Costs, Evidence Review No. 13

Evidence Review No. 13 · 18 pages · 22 sources

Is the sunk cost fallacy real?

Yes, but it is more conditional than often claimed. Theatregoers charged full price went to more plays early in the season, and 85 per cent would finish a doomed project already paid for against 17 per cent who would start it. But one large review found sunk costs alone had a small effect; they pull hardest when a project seems nearly finished.

What the report found

  1. The classic findings are striking. Theatregoers randomly charged full price for a season ticket went to more plays in the first half of the season than those given a discount. Asked whether to finish a plane already 90 per cent paid for, 85 per cent said yes; with no money spent, 17 per cent did.
  2. It appears with real stakes. Health club members attended most in the months just after paying, and less as the next payment approached.
  3. Its size is disputed. One review of 100 effect sizes found a moderate effect; another of escalation studies found sunk costs alone had a small effect that could not be told apart from zero.
  4. Training helps, a little. Economists reported walking out of bad films and meals more often than other academics, and a short lesson improved students' reasoning, though not much their behaviour.
  5. The effect is real but conditional. It is strongest when a project seems nearly finished, when people feel responsible, and when they are young; it fades with time since payment.

What is inside

  1. The idea and where it came from
  2. The original studies, explained
  3. What happened when others tested it
  4. Where researchers disagree
  5. What we can say with confidence

Plus a timeline of the key studies, two debates set out side by side and a table grading each common claim.

Who it is for

For managers, investors, product teams, anyone who sells prepaid plans, and anyone deciding whether to keep going with something.

How we work

Evidence Reviews follow one well known idea from its origin to today: the original studies, what happened when others tested them, where researchers disagree and what we can say with confidence. Every claim is graded, and both sides of each debate are set out before we say where we stand.

Cite this report

The Why We Project (2026) Sunk Costs: Why we keep paying for what has already gone. Evidence Review No. 13. Dubai: Why We Tech Ltd. Available at: thewhyweproject.com