# Sunk Cost Fallacy The **sunk cost fallacy** is allowing costs that are already irrecoverable to influence decisions about future investment — honoring the spent rather than evaluating the spendable. Rational-choice theory says a sunk cost is decision-irrelevant by definition: whatever you do next, the money, time or lives already spent stay spent. Humans systematically disagree, and they disagree *more* the larger the prior investment, the more public the commitment, and the more personally responsible they were for the original decision. That last variable is the one that matters for this vault. ## The research Richard Thaler named the effect in 1980 as one of the anomalies that a "positive theory of consumer choice" would have to explain. Arkes and Blumer (1985) supplied the canonical experiment: Ohio University theatre-season subscribers were randomly sold tickets at full price, $2 off or $7 off; the full-price buyers attended significantly more performances in the first half of the season — the ticket's cost was sunk in all three cases, but the people who had paid more felt more obliged to use it. Staw (1976) had already shown the organizational version, **escalation of commitment**: managers who had personally allocated funds to a failing division allocated *more* to it than managers who inherited the decision. And Dawkins and Carlisle (1976) gave it its other name, the **Concorde fallacy**, after the British and French governments' continued funding of the supersonic airliner long after it had no economic case — the example the folk model Throwing Good Money After Bad had described for centuries. The cognitive substrate is two other entries here: [[Loss Aversion]] (abandoning the project converts a paper loss into a real one) and [[Inconsistency-Avoidance Tendency]] (quitting indicts one's own earlier judgment). The escalation literature adds a third: sunk costs are strongest when the decision-maker's *identity* is bound up with the original choice, which is why the bias is a leadership and institutional pathology more than an individual one. ## Where the vault uses it - **War.** The argument that a war must continue so that those already killed "will not have died in vain" is the sunk-cost fallacy in its purest and most consequential form; it is the rhetorical engine of [[Forever Wars]] and appears in every escalation the vault records from the [[Vietnam War]] to the [[Iraq War]] and [[Afghanistan War]]. - **Belief.** [[Cult Logic]] and [[Conspiracy Theory Logic]] both run on it: the more a member has sacrificed for the belief, the more expensive doubt becomes, and disconfirmation produces *increased* commitment (Festinger's *When Prophecy Fails* is the classic case). [[Conformity]] pages describe the same dynamic in groups. - **Institutions.** [[Lock-In]] and [[Increasing Returns]] are the economic versions — technologies and policies persist because switching would write off what was invested, whether or not the investment was wise. The [[Enron Scandal]] page shows executives doubling down on a failing structure; [[Public Choice Theory]] explains why programs are easier to start than to stop. - **Economics.** [[Classical Economics]] and [[Neoclassical Economics]] assume the fallacy away — which is one reason their models of firm and voter behavior miss what the vault's case studies show. ## See Also - [[Decision Science Hub]] · Folk Models · Throwing Good Money After Bad - [[Loss Aversion]] · [[Inconsistency-Avoidance Tendency]] · [[Lock-In]] · [[Increasing Returns]] - [[Forever Wars]] · [[Cult Logic]] · [[Conspiracy Theory Logic]] · [[Conformity]]