The sunk-cost fallacy is a cognitive bias in which an irrecoverable investment of money, time, effort, or other resources improperly influences a decision about what to do next. It commonly appears as a reluctance to abandon an activity because doing so would seem to waste what has already been invested. In economics, past expenditures that cannot be recovered should not independently determine a choice between future alternatives: the relevant comparison concerns the benefits and costs that can still change.
Economic meaning
A sunk cost is a cost already incurred that cannot be recovered through the decision currently under consideration. Whether a cost is sunk depends on the available options, not simply on whether payment occurred in the past. Money spent on a nonrefundable ticket is generally sunk; the resale value of equipment is not, because selling the equipment remains an alternative.
The fallacy arises when the irrecoverable expenditure itself becomes a reason to continue. Statements such as “too much has been spent to stop now” express its characteristic reasoning, although they do not by themselves establish that continuing is a mistake.
A simplified decision rule illustrates the distinction. Suppose:
- is an irrecoverable expenditure already made;
- is the expected future benefit of completing a project;
- is the additional cost of completion;
- is the net future value of the best alternative, including any recoverable value from stopping.
Under a simple monetary expected-value criterion, continuing is preferable when
The sunk expenditure does not enter this comparison because it is the same regardless of the choice now made. Adding to both alternatives does not change their ranking.
This does not mean that all decisions should maximize money. Enjoyment, obligations, risk, and other consequences may matter. The principle is that an unrecoverable historical cost should not, by itself, change the relative attractiveness of future options.
Examples
Consumer decisions
A person buys a nonrefundable ticket to a performance but later expects the evening to be unpleasant. Attending solely to “get the money’s worth” exemplifies sunk-cost reasoning: the ticket price cannot be recovered by attending. Attendance can nevertheless be reasonable if the performance is still preferable to the available alternatives.
Similarly, finishing an unenjoyable book because many hours have already been spent reading it can reflect the fallacy. Finishing because the remaining chapters are expected to be rewarding is a forward-looking reason.
Business projects
A company has spent 10 million monetary units developing a product. Completing it would cost another 2 million and generate expected future receipts of 1 million, with no other relevant benefits or exit costs. The earlier expenditure does not make completion financially attractive: completing the project would add an expected loss of 1 million.
Conversely, a project can remain worth finishing even if it will never recover its total historical cost. If expected future receipts exceed the remaining costs and the value of alternatives, completion may be the better choice.
Time and effort
The bias is not restricted to money. A person may persist with an ineffective procedure because learning it required substantial effort, or continue a lengthy task because stopping feels like admitting that previous work was pointless. Time already spent is unrecoverable, while the opportunity cost of spending additional time concerns activities that could still be pursued instead.
Research and terminology
The economic distinction between sunk costs and decision-relevant future costs predates the psychological research on the fallacy. Its experimental study became prominent within behavioral economics and organizational psychology during the 1970s and 1980s.
Barry Staw’s 1976 research on commitment to a chosen course of action examined how personal responsibility for an earlier decision could encourage further investment after negative feedback. Hal Arkes and Catherine Blumer’s 1985 paper, The Psychology of Sunk Cost, helped establish the sunk-cost effect as a subject of experimental investigation. Their studies examined how prior investments could influence subsequent choices, including decisions about paid activities.
The expression Concorde fallacy refers to the Anglo-French supersonic aircraft project, which became a familiar illustration of the argument that substantial prior expenditure justifies further spending. Applying the label to a particular historical decision, however, requires distinguishing sunk expenditures from prospective commercial, contractual, strategic, or political consequences.
Researchers sometimes distinguish the sunk-cost effect, an observed influence of previous investment on behavior, from the sunk-cost fallacy, the judgment that this influence violates an appropriate decision standard.
Psychological explanations
No single mechanism explains every instance. Several processes may contribute.
Avoidance of waste
People commonly learn that wasting resources is undesirable. This rule is useful before resources are committed, but it can be misapplied afterward. Continuing an unwanted activity may feel like using the investment rather than wasting it, even when continuation consumes additional resources without sufficient benefit.
Self-justification
Abandoning a project can imply that the initial decision was mistaken. Continuing may help preserve a favorable view of one’s judgment. This explanation connects sunk-cost behavior with cognitive dissonance, particularly when a person was personally responsible for the initial commitment.
Confirmation bias can reinforce persistence if decision-makers favor evidence supporting eventual success and discount evidence supporting withdrawal.
Reference points and losses
Prospect theory describes how choices can depend on gains and losses relative to a reference point. A decision-maker may treat stopping as accepting a definite loss and continuing as a chance to avoid that loss. This framing can make additional risk attractive.
Loss aversion is therefore relevant to some explanations, but it is not synonymous with the sunk-cost fallacy. Sensitivity to losses need not involve past irrecoverable investments.
Mental accounting
Mental accounting describes how people organize expenditures and outcomes into psychological categories. A prepaid activity may remain an “open” account until it has been consumed. Participation can then feel necessary to balance the account, even though the payment is already unrecoverable.
Relationship to escalation of commitment
Escalation of commitment is a broader pattern in which people or organizations maintain or increase commitment to a course of action despite unfavorable feedback. Sunk-cost reasoning can contribute to escalation, but the concepts are not interchangeable.
Escalation may also reflect reputational concerns, organizational incentives, competition, or a desire to justify earlier decisions. Some of these influences involve real future consequences for the decision-maker, even when they encourage choices that are costly for an organization. Conversely, a sunk-cost effect can occur in a small, one-time consumer decision without a prolonged cycle of escalating investment.
Limits and interpretation
Persistence is not sufficient evidence of a fallacy. Several distinctions are important:
- Past spending may provide information. Earlier work can reveal technical feasibility, customer demand, or likely completion costs. This information is relevant even though the expenditure itself is sunk.
- Previous investment may create valuable assets. Skills, equipment, intellectual property, and partially completed work can change future payoffs. Their present usefulness should not be confused with their historical acquisition cost.
- Stopping can have future costs. Cancellation payments, lost salvage value, and effects on future relationships can legitimately enter the comparison.
- Uncertain projects may justify further investigation. A limited additional investment may produce information valuable for a later decision.
- Outcomes do not establish decision quality. A successful result does not prove that sunk-cost reasoning was sound, just as failure does not prove that continuing was unreasonable.
Experimental findings also depend on how choices are presented, the type of investment, responsibility for earlier decisions, and participants’ understanding of the task. A stronger demonstration holds future prospects constant while varying only the irrelevant past investment.
Decision procedures
Methods studied or used to limit sunk-cost influence include separating historical expenditures from forecasts, specifying stopping criteria before investment, and obtaining assessments from people who did not make the initial commitment. A common diagnostic question is whether the same future opportunity would be chosen if the previous investment had not occurred.
These procedures have limits. Independent reviewers may lack relevant knowledge, and rigid stopping rules may fail when circumstances change. Their purpose is to make the reasons for continuation explicit and distinguish recoverable value, future consequences, and new information from the desire to vindicate an irreversible past choice.