A company has spent $5 million developing a software product that market research now shows has minimal demand. The question is whether to spend another $2 million to complete and launch it. The $5 million is gone — it cannot be recovered regardless of what happens next. The relevant decision is: does investing $2 million more yield at least $2 million in future value? If yes, proceed. If no, stop. The $5 million already spent should not factor in either direction — it is a sunk cost.
In plain terms
A sunk cost is a cost that has already been incurred and cannot be recovered regardless of future decisions. Because it cannot change based on what you choose next, it is economically irrelevant to any forward-looking decision.
Sunk costs include: the purchase price of a non-refundable concert ticket, the development costs of a canceled product, the down payment on a house you're considering selling, the education investment in a career you're reconsidering, and the time already spent on a project you're considering abandoning.
The relevant costs for any decision are only incremental future costs — what will be spent going forward — and the opportunity cost of those resources. The Congressional Budget Office's project evaluation framework explicitly excludes sunk costs from forward-looking program assessments — only prospective costs and benefits are included in cost-benefit analyses.
Why it works this way
The logic is mechanical: a sunk cost is already paid. It represents the same deduction from your net position regardless of what you decide to do from here. If spending an additional $2 million on a failing project will yield only $500,000 in revenue, the decision to continue produces a net loss of $1.5 million going forward — and the $5 million already spent doesn't change that math. Continuing "to justify" the $5 million produces a final loss of $6.5 million instead of $5 million.
This is the sunk cost fallacy — continuing a losing course of action because of past irrecoverable investment. It is irrational in strictly economic terms and deeply understandable in psychological terms. The behavioral version of sunk cost is treated separately in the glossary; this entry concerns the economic concept and its correct handling in business and policy decisions.
A real example
The U.S. government's continuation of the F-35 fighter jet program despite repeated cost overruns provides a real-world example of sunk cost tension. By any point in the program's development, billions had already been spent. The rational question at each subsequent decision point was: given current expected costs and capabilities, does proceeding generate net value greater than cancellation? The Government Accountability Office's defense acquisition reports document this ongoing evaluation — though political and industrial commitment dynamics make pure sunk-cost logic difficult to apply in practice.
Why it matters
Sunk cost awareness is one of the highest-value disciplines in business decision-making. Firms that ignore sunk costs and focus on forward-looking marginal costs and benefits — cutting failing projects, exiting unprofitable markets — outperform firms that continue investing in losers to avoid acknowledging past mistakes. The discipline is simple to state and psychologically difficult to apply, which is why it produces ongoing business value for teams that internalize it.
