Sunk cost fallacy
Continuing to invest money, time or loyalty because of what has already been spent rather than because of the expected return from here on.
What drives it
Avoids admitting waste, which feels like admitting a mistake. Stopping makes the loss real; continuing keeps alive the chance that the original decision will be vindicated.
How it shows
Work
Keeping an underperforming senior hire because of the search fee, the relocation and what you told the board about them. 'We've put two years into this platform.' The person who made the original call escalates more than a successor would (Staw's finding). Early backers of a project become its least objective judges.
Relationships
'I've given him ten years.' Friendships or marriages defended by their history rather than their present. Continuing to rescue an adult child because of how much has already gone in.
Self
You cite what has gone in rather than what will come out. When defending someone you backed, check whether the defence rests on their current performance or on how much you have invested in being right about them.
What tells it apart
Commitment and consistency: Commitment and consistency is social: people stick with what they publicly said; sunk cost is about prior spend and persists even when nobody else knows about the original decision.
Cognitive dissonance: Dissonance changes how much the person values the thing; sunk cost can operate with the person openly unhappy about the investment and still unable to stop.
Trauma bonding: Trauma bonding holds a person through cycles of harm and intermittent warmth, with intense longing when apart; sunk-cost staying is calmer and argued in terms of what has been invested.
A test
Ask 'starting fresh today with nothing invested, would I choose this person or project?'; if the answer is no and the commitment continues, sunk cost is driving it.
Questions to carry
- Starting fresh today with nothing already spent, which part of this commitment would still be chosen?
- What would stopping now make me admit about the decision I made at the start?
How strong the research is
moderate. The effect replicates in human decision studies, including Arkes and Blumer's scenarios, but it is smaller and more context-dependent than its reputation, and some apparent cases reflect reasonable concerns about reputation or information.
Where it comes from
Barry Staw, 1976 (escalation of commitment); Richard Thaler, 1980; Hal Arkes and Catherine Blumer, 1985
Sources
- Leon Festinger, A Theory of Cognitive Dissonance (1957)
- Daniel Kahneman and Amos Tversky, Prospect Theory: An Analysis of Decision under Risk (1979)