Sunk-Cost Fallacy
A decision-making bias where prior unrecoverable investments — money, time, effort — are allowed to influence current choices, even though they should not.
Hal Arkes and Catherine Blumer's 1985 Organizational Behavior and Human Decision Processes paper made the sunk-cost fallacy a core finding of behavioral economics. Their experiments showed people would sit through a movie they were not enjoying because they had paid for the ticket, finish a meal they did not want because they had ordered it, and continue funding failing projects because of resources already spent. The rational rule is straightforward: only future costs and future benefits should affect a forward-looking decision; money already spent is gone regardless of what you choose next. But several psychological forces fight that rule — loss aversion (abandoning the project converts a paper loss into a realised one), commitment consistency (we want our future selves to validate our past selves), and a fairness intuition that effort should be rewarded with completion. The fallacy compounds across organizations: the more visibly invested a project is, the harder it becomes to kill, even when the remaining return on investment is clearly negative.
Frequently Asked Questions
How is the sunk-cost fallacy different from loss aversion?
Loss aversion is about how strongly future losses are weighted relative to gains. The sunk-cost fallacy is about letting past, unrecoverable losses bias forward decisions — keeping a doomed project alive because killing it would "waste" what was spent. Loss aversion is one of the engines that drives sunk-cost behaviour, but they are not the same thing.
What is the cleanest way to escape sunk-cost reasoning?
Reframe the decision as if you were arriving fresh: given the current state, with no history of investment, would you start this project today? If the answer is no, the past spend should not change the present choice. Bringing in an outside reviewer who never approved the original decision is the human-scale version of the same trick.
Do non-human animals show the sunk-cost fallacy?
Some studies suggest pigeons and rats show analogous patterns under certain conditions, while others fail to replicate. The current consensus is that the fallacy in its full form — driven by self-justification and commitment to a public choice — is largely a human phenomenon, even if simpler perseveration shows up more broadly.