What Is a Sunk Cost?
A sunk cost is any past expenditure that stays spent no matter which path you choose going forward. Research and development on a shelved product, a nonrefundable deposit, marketing dollars for a campaign that already ran, and years invested in a failing project all qualify. Because these outlays are identical across every future option, economics says they should carry zero weight in the decision itself.
The contrast with relevant costs is the key distinction. When deciding whether to finish a half-built factory, the $50 million already poured into the foundation is sunk. The decision should rest entirely on whether the additional $30 million needed to complete it will generate more than $30 million of value, judged as if you were evaluating the project fresh today.
The Sunk Cost Fallacy
The sunk cost fallacy is the tendency to keep committing resources to something because of what has already been invested, rather than because of its future prospects. Behavioral economists trace it to loss aversion and the desire to avoid admitting a mistake: abandoning a project makes the earlier loss feel real, while pressing on preserves the hope of vindication. The classic corporate example is the Concorde supersonic jet, which governments kept funding for years after it was clearly uneconomic.
Investors fall into the same trap through anchoring on purchase price. Refusing to sell a losing stock "until it gets back to what I paid" treats a sunk cost as decision-relevant. The market does not know or care what you paid; the only question is whether the stock is attractive at today's price versus the alternatives available for that capital.
Why It Matters in Practice
Capital budgeting frameworks are built to screen sunk costs out. NPV and IRR analysis considers only incremental future cash flows, so a feasibility study that cost $2 million last year is excluded when evaluating whether to proceed, even though it feels like part of the project. Interviewers in banking and consulting frequently test this: a case that mentions money already spent is usually inviting you to state explicitly that it should be ignored.
Disciplined organizations institutionalize the lesson. Pharmaceutical companies kill drug candidates after hundreds of millions in trials when the data turns, private equity firms write off failed platform investments rather than doubling down, and good managers reframe the question as "would we start this today?" Pairing sunk cost thinking with opportunity cost, the value of redeploying resources elsewhere, is what separates rigorous capital allocators from sentimental ones.
