The sunk cost fallacy is the tendency to continue investing time, money, or effort in a course of action because of what has already been invested, even when continuing is no longer the best use of remaining resources. From an economic standpoint, prior expenditure is irrelevant to a current choice — only future costs and future benefits should matter. Psychologically, however, prior investment exerts a powerful pull. People finish bad movies, stay in dying relationships, defend losing strategies, and pour additional billions into failed projects, not because the road ahead looks promising but because the road behind feels too costly to walk away from.
Hal Arkes and Catherine Blumer documented this pattern in their 1985 paper "The Psychology of Sunk Cost," establishing it as a robust phenomenon distinct from rational economic behavior. The fallacy turns out to be linked to loss aversion, personal responsibility for the initial choice, identity, social signaling, and the asymmetry between losses we have already absorbed and losses we feel we are about to take. Understanding the mechanism is a first step toward making decisions that respect the past without being trapped by it.
Key Facts About the Sunk Cost Fallacy
- Foundational experiments by Hal Arkes and Catherine Blumer (1985)
- Defined as continued investment driven by prior, unrecoverable expenditure
- Closely related to but distinct from escalation of commitment
- Strongly amplified by personal responsibility for the original decision
- Rooted in loss aversion — losses loom roughly twice as large as equivalent gains
- Documented at individual, organizational, and national-policy levels
- Sometimes called the Concorde fallacy after the supersonic-jet program
- Reduced by separating the new decision from the historical investment
Understanding the Sunk Cost Fallacy
A Definition That Distinguishes Past From Future
A sunk cost is any expenditure of money, time, effort, or emotion that cannot be recovered. Standard decision theory holds that sunk costs should not influence forward-looking decisions because they are unaffected by what you do next. Only marginal costs and benefits — the additional resources required and the additional value to be received — should determine the choice. The sunk cost fallacy is the empirical observation that human decision-makers systematically violate this principle: they let what has been spent influence what they decide to do next.
The fallacy is not about valuing past investment for its informational content. It is rational to learn from how something has gone so far. The fallacy is about using past expenditure as a reason in itself to keep going, even when the lesson the past would teach is to stop.
The Difference From Escalation of Commitment
Escalation of commitment, a closely related concept developed by Barry Staw beginning in the 1970s, refers specifically to the pattern of increasing one's investment in a failing course of action over time. Sunk cost effects can drive escalation, but escalation involves additional dynamics: self-justification, social pressure, organizational politics, and a desire to recover prior expenditure through a dramatic turnaround. In short, all escalation involves sunk costs, but not all sunk cost effects are escalations.
Why Past Investment Feels Like Evidence
For most people, abandoning a project after sustained investment feels like an admission that the original decision was wrong. If the initial choice was a mistake, then everything spent since was waste. The mind resists this conclusion because it implicates the self. Continuing offers the (often illusory) prospect of redemption: if the project ultimately succeeds, the prior investment will look like patience rather than folly. The fallacy is thus partly a strategy for protecting self-image.
Cross-Cultural Robustness
The sunk cost effect has been demonstrated across cultures, age groups, and economic contexts. Some studies show variation in magnitude — for instance, the effect appears weaker in early childhood and somewhat stronger when personal responsibility for the original choice is high — but the basic phenomenon is widespread. Interestingly, non-human animals show much less evidence of sunk cost reasoning, suggesting that the bias may be partly tied to the human capacity for self-narrative and justification.
The Research Foundation
Arkes and Blumer (1985)
The foundational paper presented a series of scenarios demonstrating that prior expenditure shifted decisions in directions inconsistent with expected-utility theory. In one well-known item, participants were asked to imagine they had bought a non-refundable ticket to a ski trip for $100. They then learned of a better trip for $50 on the same date, also non-refundable, which they bought. They were told they could go on only one trip and that the cheaper trip would be more enjoyable. A majority chose the more expensive, less enjoyable trip — a clear violation of rational choice, since the money was gone either way and the cheaper trip was, by stipulation, more fun.
Across many such scenarios, Arkes and Blumer showed that the effect was reliable and substantial. They also demonstrated that providing accounting information about prior expenditure could trigger fallacious reasoning even when participants were instructed to make forward-looking decisions.
Staw and Escalation
Barry Staw's research from 1976 onward examined why people commit additional resources to failing courses of action. In his classic experiment, business-school students who had made an initial investment decision allocated more money to the same division after learning it was performing poorly than students who had not made the initial choice. Personal responsibility for the original allocation was a powerful driver — the self-justification motive sharpened sunk cost reasoning.
Thaler and Mental Accounting
Richard Thaler's work on mental accounting helped explain why sunk costs feel real. People do not treat money as fungible; they assign it to mental accounts. Prior expenditure on a project opens an account that the mind tries to close in the black. Abandoning the project closes the account in the red — a felt loss. Continuing offers at least the possibility of breaking even, even if expected value says otherwise.
Loss Aversion and Prospect Theory
Kahneman and Tversky's prospect theory, articulated in 1979, established that losses are felt roughly twice as strongly as equivalent gains. Sunk cost behavior is consistent with prospect theory: cutting losses crystallizes a loss, whereas continuing keeps open the possibility of a gain. People become risk-seeking in the domain of losses, sometimes accepting unfavorable gambles to avoid locking in a sure loss.
Modern Replications and Extensions
Subsequent work has examined when sunk cost effects appear and disappear, individual differences in susceptibility, and the role of cognitive reflection. The effect is robust but not universal; certain framings, decision aids, and personality factors moderate it.
How It Works
Loss Aversion as the Engine
The first ingredient is loss aversion. Walking away from a project crystallizes a loss; continuing keeps the outcome ambiguous. Because losses feel disproportionately painful, people prefer the gamble of continuation to the certainty of cutting losses. This is why sunk cost effects strengthen as the magnitude of prior investment grows — the loss being crystallized is larger.
Self-Justification
The second ingredient is the desire not to have been wrong. The original decision was made for a reason; abandoning the project implies the reason was insufficient or the judgment was poor. Continuing protects the integrity of the original choice. Staw's research is most direct on this point: when the person making the current decision is also the person who made the original one, sunk cost effects sharpen considerably.
Identity and Public Commitment
Decisions become part of identity. A founder who has built a startup, a researcher who has spent years on a question, a parent who has steered a child toward a particular career — each has an identity-level stake in the project succeeding. Public commitments make this worse. Visible reversal looks like failure; quiet persistence looks like resilience.
Optimism About Recovery
People who continue often believe, sincerely, that one more push will turn the project around. The belief is not always wrong, but it is systematically biased in the direction of past investment. The greater the sunk cost, the more confident the projection.
The Endowment of Effort
Effort itself appears to create attachment. Studies on the "IKEA effect" and related phenomena show that people value objects they assembled or built more than equivalent objects they did not. Similar dynamics apply to projects: the work invested generates a felt ownership that complicates dispassionate evaluation.
Asymmetric Regret
Anticipated regret runs in only one direction. People anticipate the regret of abandoning a project that might have succeeded, but they discount the regret of persisting in a project that fails after even more investment. The anticipated regret asymmetry tilts choices toward continuation.
Everyday Examples
The Bad Movie
You are an hour into a film you are not enjoying. The rational question is whether the next hour will be worth your time. The ticket cost and the hour already spent are gone. People nonetheless stay because they paid for the ticket and "might as well finish."
The Long-Term Relationship
Couples sometimes describe staying together because of how much they have built, even when the relationship is no longer working. Years invested, friends in common, shared property, and joint identity make ending feel like negating the past. This is sunk cost reasoning in one of its most consequential forms.
Higher Education and Career Path
A student deep into a degree they dislike often continues because they have completed three years. A mid-career professional locked into a field they hate stays because they spent a decade getting there. In both cases, prior years are sunk and ought not to determine whether the remaining years are best spent the same way.
Home Renovations
Projects that begin with a modest budget routinely escalate. Each new problem reveals additional work; abandoning the project after partial renovation feels worse than completing it, regardless of whether the marginal investment makes financial sense.
Stock Portfolios
Investors notoriously hold losing stocks longer than they should, hoping to "get back to even." The disposition effect — selling winners too soon and holding losers too long — is a market-level manifestation of sunk cost reasoning combined with loss aversion.
All-You-Can-Eat Buffets
Many diners report eating past comfort because they paid a fixed price. The price is sunk; comfort and health are forward-looking. Eating more food does not redeem the cost — it adds an additional cost in physical discomfort.
Where It Shows Up
Large Infrastructure Projects
Megaprojects routinely exceed budgets and timelines while continuing because abandonment is politically and financially costly. The Concorde supersonic-jet program is the archetype — British and French governments continued investing for years after it was clear the project would not be commercially viable. The name "Concorde fallacy" is sometimes used interchangeably with sunk cost fallacy in policy circles.
Military Engagements
War, perhaps the most consequential domain of sunk cost reasoning, is shaped by phrases like "their sacrifice must not be in vain." The argument that prior casualties justify continued engagement is structurally a sunk cost argument. Historians have documented this pattern across many conflicts.
Corporate Strategy
Failing product lines, acquired companies that never integrated, and strategic bets that did not pan out often persist long past the point where the numbers justify continuation. Leadership transitions are frequently the moment when sunk cost projects are finally killed, because new leaders are not personally responsible for the original commitment.
Research and Development
Scientists and engineers face strong pressure to continue lines of work in which they have built expertise, equipment, and reputation. Funding agencies can amplify this by reviewing projects in light of prior investment rather than prospective value.
Personal Finance
Home buyers stretch to complete renovations on properties they should have walked away from. Subscribers continue services they no longer use because they have paid for the year. Members keep gym memberships unused because cancellation feels like wasting the months already paid.
Public Policy
Programs that have produced few results continue because of prior investment, political constituencies, and the difficulty of admitting that earlier commitments were unwise. Sunset clauses and structured reviews are policy responses to this dynamic.
Real-World Consequences
Persistent Misallocation of Resources
Time, money, and attention spent continuing failing projects are unavailable for better uses. The cost is not only the additional resources poured in but the opportunity cost — the better projects that were not pursued because attention was committed elsewhere. Across an organization, the cumulative drag from sunk cost projects can be enormous.
Catastrophic Outcomes
In high-stakes domains, sunk cost reasoning has contributed to disasters. Engineering decisions to proceed with launches, surgeries, or missions in the face of new disconfirming information have repeatedly been linked, in retrospective analysis, to dynamics that include sunk cost reasoning and escalation of commitment.
Emotional and Relational Damage
People who stay in unfulfilling relationships, dead-end jobs, or stalled creative projects often experience erosion of well-being. The damage is rarely a single bad day; it accumulates as years invested in a path that no longer fits.
Distorted Strategic Reasoning
Strategy is forward-looking by definition. When sunk costs intrude, strategy becomes about salvaging the past rather than positioning for the future. Organizations that allow sunk cost reasoning to dominate strategy review tend to underperform organizations that protect strategic decisions from historical accounting.
Erosion of Trust
Leaders who refuse to abandon failing initiatives often lose credibility with the people who depend on them. Paradoxically, the move that would protect credibility — admitting the situation has changed and adjusting — is the one the sunk cost effect blocks.
How to Recognize It in Yourself
Listen for Past-Tense Justifications
Phrases like "we've come too far to stop," "I can't waste everything I've already put in," "after all I've done," and "we need to make this worth it" are signals. Each grounds a future decision in past expenditure rather than future value.
Run the Stranger Test
Ask what a competent stranger arriving fresh to your situation, knowing only the current state and future prospects, would recommend. If the stranger would walk away and you would not, sunk cost reasoning is likely operating.
Notice Defensive Reasoning
If you find yourself building elaborate arguments for why this time will be different, or focusing on the worst-case if you stop while glossing the worst case if you continue, the bias is likely active.
Check for Personal Responsibility
You are more vulnerable when you made the original decision yourself. Staw's research shows that personal responsibility magnifies the sunk cost effect substantially. If you are the one whose judgment is implicitly on trial, expect to feel pull.
Look for the Avoided Question
The question being avoided is usually the right one: "If I were starting from scratch today, with my current resources and the current state of this project, would I begin?" If the answer is no, continuation needs a different justification than past expenditure.
How to Counter It
Separate the New Decision From the Old
The cleanest mental move is to treat the current decision as if you were arriving fresh. Imagine that the prior investment was made by someone else. What would you choose now, given only forward-looking costs and benefits? This reframing isolates the rational variable and is the single most effective debiasing technique reported in the literature.
Use Prospective Hindsight
Imagine yourself one year from now, having either continued or stopped. From each future vantage point, what does today's decision look like? Prospective hindsight, developed by Mitchell, Russo and Pennington and popularized by Gary Klein's premortems, forces a balanced consideration of both branches.
Decision Pre-Commitment and Kill Criteria
Before beginning a project, articulate the conditions under which you will stop. "If revenue is below X by Y, we close the line." "If the trial does not show clinical benefit at the interim analysis, we end recruitment." Pre-commitment removes some of the discretion that sunk cost reasoning exploits.
Stage Gates and Structured Reviews
Many organizations use stage-gate processes that require explicit forward-looking evaluation at predetermined points. Decisions made by people who were not part of the original commitment are particularly effective. New leadership, outside boards, and rotating review panels are practical mechanisms.
Reframe the Loss
Abandonment is not the moment of loss — the moment of loss already passed when the investment stopped producing value. Recognizing this can defang the felt finality of stopping. "We are not losing the investment by quitting; we lost it when the project stopped working. Quitting prevents additional loss."
Build a Culture That Praises Smart Quitting
Organizations that publicly recognize teams who kill bad projects build a culture in which quitting is not a career-ending move. When the incentives align with rational reallocation, sunk cost effects soften.
Use Decision Journals
Keeping a record of why a project was started, with the expected outcomes and the assumptions behind them, separates evaluation from current emotions. Reviewing the journal at decision points yields cleaner forward-looking judgment.
Limits of Debiasing
Sometimes Persistence Is Right
Not every cost that has been incurred is sunk; not every project deep in the red should be abandoned. Many endeavors — research programs, businesses, learning curves — produce returns only after a long period of investment. Distinguishing rational persistence from sunk cost reasoning requires asking whether the path forward is justified by future expected value alone, including option value and learning effects, rather than by past expenditure.
Option Value and Learning
Continuing a project sometimes preserves valuable options: the ability to pivot, the accumulation of knowledge that will pay off in unexpected ways, the development of capabilities that have downstream value. Genuine option value belongs in the forward-looking calculation. The fallacy is to invoke option value rhetorically when the real argument is "we have already spent so much."
Awareness Helps but Doesn't Eliminate the Effect
Even people who know about the sunk cost fallacy continue to display it. Research on debiasing suggests that structural interventions — pre-commitment, stage gates, kill criteria, outside review — outperform personal vigilance. Knowing is not enough; the environment must make rational reallocation easier than the default.
Emotional Costs Are Real
The argument that "sunk costs are gone, so ignore them" is correct economically but incomplete psychologically. Letting go of a long-term project, relationship, or identity often involves grief that is itself a forward-looking cost. The point of debiasing is not to numb that grief but to keep it from masquerading as a rational reason to continue.
Identity and Meaning
Some commitments are about meaning, not value. Caring for a chronically ill family member, completing a creative work, or honoring a promise may be the right thing to do even when a cold cost-benefit analysis says otherwise. The sunk cost fallacy describes a specific kind of mistake — letting past expenditure drive future decisions — not a general indictment of perseverance.
Conclusion
The sunk cost fallacy is one of the most expensive cognitive biases people make, both individually and at scale. Every dollar, hour, or emotion already spent is gone whether or not the project continues. The only question worth answering is whether the remaining resources are better spent here than anywhere else. Yet because losses loom large and self-justification is powerful, the mind treats prior investment as if it were a reason to keep going.
The research foundations are robust. Arkes and Blumer, Staw, Thaler, Kahneman and Tversky, and decades of subsequent work have shown that the effect is real, broad, and consequential — appearing in personal life, organizations, finance, military strategy, infrastructure, and public policy. It is closely tied to loss aversion, escalation of commitment, mental accounting, and identity.
Debiasing is possible but partial. The most reliable corrections are structural: pre-commitment, kill criteria, stage gates, outside review, and decision journals. At the individual level, the most useful move is the simple act of separating the new decision from the old — asking what you would do if you arrived fresh, and treating prior expenditure as information about what has happened rather than a vote about what should happen next.