Behavioral Decision-Making
Sunk-Cost Fallacy
Why people throw good money, time, and effort after bad
The sunk-cost fallacy is the tendency to continue an endeavor because of previously invested resources — money, time, or effort — that should rationally be irrelevant to forward-looking decisions. Hal Arkes and Catherine Blumer's 1985 paper "The Psychology of Sunk Cost" established the experimental paradigm with their famous ski-trip vignette: 54% of participants chose a $100 Michigan ski trip over a more enjoyable $50 Wisconsin trip simply because more money had been spent on Michigan. Subsequent research linked the bias to loss aversion, commitment escalation, and self-justification. It surfaces in finance, military strategy, relationships, and product development.
- Foundational paperArkes & Blumer (1985)
- Classic vignetteMichigan vs Wisconsin ski trip
- Vignette result54% chose less-enjoyable but more-paid trip
- MechanismLoss aversion, commitment, self-justification
- Related conceptEscalation of commitment (Staw 1976)
- Animal evidenceMagalhães & White (2016) — rats and pigeons show it
Interactive visualization
Press play, or step through manually. The visualization is yours to drive — try it before reading on.
Watch the 60-second explainer
A condensed visual walkthrough — narrated, captioned, under a minute.
Why the sunk-cost fallacy matters
- Project management. Failing projects continue past their economic justification.
- Investing. Holding losing positions to "make it back" exemplifies the bias.
- Relationships. Years invested make leaving harder than the future justifies.
- Military strategy. Wars escalate when withdrawal would crystallize prior losses.
- Product development. Companies ship features no one wants because they were built.
- Education choices. Sticking with a major after losing interest because of credits earned.
- Career decisions. Specialization sunk costs trap professionals in declining fields.
Common misconceptions
- Continuing always means falling for it. Continuing can be rational when past investment supplies forward value.
- Only humans show it. Rats, mice, and pigeons exhibit similar patterns.
- Awareness defeats it. Knowing about sunk costs barely reduces the bias in lab tasks.
- Money matters most. Time and effort produce equally strong sunk-cost effects.
- It's purely irrational. Reputation and commitment signaling sometimes justify persistence.
- It is the same as loss aversion. Loss aversion is general; sunk cost specifically weights past investment.
Frequently asked questions
What was the Arkes and Blumer ski-trip experiment?
Participants imagined buying a $100 nonrefundable ski trip to Michigan, then later finding a $50 nonrefundable trip to Wisconsin they expected to enjoy more. The trips overlap so only one can be used. Roughly 54% chose Michigan despite predicting less enjoyment, citing the $100 already spent. Rationally the $100 is sunk; the choice should hinge on which trip will be more enjoyable. The result, published in Organizational Behavior and Human Decision Processes (1985), framed the paradigm.
How does it differ from escalation of commitment?
Barry Staw (1976) introduced escalation of commitment for ongoing courses of action — managers throwing more resources at failing projects rather than abandoning them. Sunk-cost fallacy and escalation overlap heavily. Sunk-cost is the static error of weighing past investment; escalation is the dynamic process of doubling down. Most real-world cases involve both — past investment makes new investment feel justified.
Why do people commit the fallacy?
Multiple mechanisms. (1) Loss aversion — abandoning the project crystallizes the loss; continuing keeps the loss potential. (2) Self-justification — admitting waste damages self-image; continuing supports a coherent narrative. (3) Commitment to consistency — Cialdini's research shows people honor prior decisions to stay consistent. (4) Hope of recovery — additional investment may eventually make the prior worthwhile. (5) Reputation — stopping signals failure to others.
Do animals show sunk-cost behavior?
Recent evidence says yes. Sweis, Abram, Schmidt, et al. (2018) found rats, mice, and humans all showed sunk-cost behavior in a foraging task — willingness to wait for a reward grew with time already invested. Magalhães and White (2016) found the same in pigeons. The convergence suggests sunk-cost may have an adaptive evolutionary basis — perhaps a heuristic shortcut for tasks with high uncertainty about completion.
How does it appear in business?
Boeing's 737 MAX continued through engineering compromises partly because earlier design commitments locked in the platform. Concorde — the canonical example — flew commercially despite never recovering R&D costs because cancellation would acknowledge the prior loss. Software projects famously over-run because rewriting feels like wasting prior code. Mergers continue past the point of value because executives have publicly committed.
Are there exceptions?
Yes. Continuing an investment can be rational when (1) prior investment provides genuine forward-looking value (infrastructure, learning, partial completion), (2) abandoning incurs new costs, (3) information is asymmetric and prior investment signals private knowledge of value, (4) reputational benefits of completion outweigh costs. The fallacy occurs only when past costs are weighted independently of forward-looking value.
How can you debias?
Several techniques help. (1) Ask "if I were starting today with current information, would I begin this project?" — Buffett's "if you find yourself in a hole, stop digging." (2) Separate decisions about past spending from decisions about future spending. (3) Use independent reviewers without prior commitment. (4) Pre-commit to decision criteria — kill criteria for projects. (5) Reframe abandonment as redirecting toward better opportunities, not admitting failure.