The Day Britain and France Could Not Stop
On November 29, 1962, the governments of the United Kingdom and France signed one of the most consequential industrial agreements of the postwar era. A supersonic passenger aircraft would be built jointly — a machine that would cross the Atlantic in three and a half hours, transforming civil aviation forever. The project was called Concorde.
Within three years, the economics had curdled. By 1965, internal Treasury assessments in Britain were circulating quietly through Whitehall corridors, carrying projections that the aircraft would never recoup its development costs. The numbers were unambiguous: development was running massively over budget, the projected market of 400 aircraft had been revised downward, and most airlines — when pressed — were not interested in buying a plane that burned more fuel per seat mile than any existing jet. The British government commissioned a study. The study confirmed what the Treasury suspected.
And yet the project continued.
The reason was stated plainly by a minister during a 1968 parliamentary debate. Withdrawing, he argued, would mean writing off everything already spent. "We cannot simply walk away," he told the House. "The investment already made demands that we see this through." That phrase — the investment already made — carried all the weight of the decision. It was not a strategic argument. It was not a forward-looking calculation about returns and risks. It was a backward-looking appeal to money that was already gone.
By the time Concorde entered commercial service in January 1976, development costs had exceeded £1 billion (equivalent to roughly £10 billion today), shared between the two governments. Only 20 aircraft were ever built for commercial use — all sold to British Airways and Air France at heavily subsidized prices. The aircraft never turned a commercial profit on its manufacturing. The British and French governments had spent more than a decade pouring resources into a project whose economics they privately doubted, in part because stopping felt like admitting that everything already spent had been wasted.
What happened in Whitehall and Paris in those years was not unusual. It was a precise, government-scale demonstration of one of the most robustly documented errors in human decision-making: the sunk cost fallacy.
What the Sunk Cost Fallacy Actually Is
The sunk cost fallacy is the tendency to continue an endeavor because of previously invested resources — money, time, effort, or emotional capital — rather than on the basis of expected future outcomes. A sunk cost is any cost that has already been incurred and cannot be recovered regardless of future decisions. Rational decision theory holds that such costs should be irrelevant to forward-looking choices: only future costs and future benefits should bear on whether to continue or quit. The fallacy occurs when sunk costs influence decisions they should not.
This definition sounds clean, but the phenomenon sits within a cluster of related but distinct cognitive and behavioral errors. Understanding where the sunk cost fallacy ends and adjacent biases begin clarifies both the phenomenon and why it is so hard to escape.
| Bias | Core definition | When it fires | Characteristic error |
|---|---|---|---|
| Sunk cost fallacy | Continuing an endeavor because of past investment | When facing the choice to continue or quit | Treating irrecoverable past costs as relevant to future decisions |
| Loss aversion | Losses feel roughly twice as painful as equivalent gains feel good | When evaluating any prospect involving potential loss | Overweighting downside relative to upside in expected value calculations |
| Escalation of commitment | Increasing investment in a failing course of action after setbacks | When a project or decision has already produced negative outcomes | Doubling down specifically because of failure, to prove earlier commitment was justified |
| Status quo bias | Preference for the current state of affairs regardless of alternatives | In any choice where inaction is an option | Treating the default option as inherently superior to alternatives |
These four biases are related but separable. Loss aversion is the broader psychological substrate from which sunk cost fallacy partly grows: because losses feel so much worse than gains of equivalent magnitude, the thought of a realized loss — confirming that earlier spending was wasted — triggers a kind of psychological pain that people will take real economic action to avoid. Escalation of commitment shares the sunk-cost structure but adds a specific social and self-justificatory dimension: we escalate especially when we were publicly responsible for the initial commitment and when reversing would require acknowledging that our earlier judgment was wrong. Status quo bias is more general still — a preference for not-changing that applies even when no prior investment is at stake.
The sunk cost fallacy is the most precisely bounded of the four: it is specifically the use of past, irrecoverable investment as justification for present or future action.
The Cognitive Science of Why This Happens
The foundational study of sunk cost effects in a controlled setting was published in 1985 by Hal Arkes and Catherine Blumer in the journal Organizational Behavior and Human Decision Processes. Their design was elegant. Students at Ohio University who had purchased season tickets to the university theater were tracked through the season. Some had paid full price; others had been randomly offered a discount. The students who paid full price attended significantly more performances, including performances during bad weather and on inconvenient evenings.
This is striking because it shouldn't matter. Once you've paid for the ticket, the money is gone whether you go or not. The rational calculus for any given evening should be identical regardless of what you paid: would you rather attend this specific performance, or do something else? But the students who paid more felt a stronger pull toward attendance. The sunk cost was shaping behavior that the sunk cost could not logically justify.
Arkes and Blumer documented the same effect across a series of experiments involving hypothetical investments in research and development projects, ski trips, and business decisions. In each case, subjects given information about prior investment increased their willingness to continue failing projects relative to subjects given the same forward-looking information without the investment history.
To understand why this happens, we need Daniel Kahneman and Amos Tversky's prospect theory, published in Econometrica in 1979. Their work demonstrated that people evaluate outcomes not in terms of absolute levels of wealth or utility but in terms of gains and losses relative to a reference point — and crucially, the value function is asymmetric: a loss of $100 generates roughly twice the psychological impact of a gain of $100.
This loss aversion creates a direct driver for sunk cost thinking. If you have spent $5,000 on a failing project, stopping means psychologically realizing a loss of $5,000. Continuing keeps that loss in a kind of suspended, unrealized state — the project hasn't failed yet, the money hasn't formally been wasted yet. So continuing is not just about future hope; it is about the present pain of mentally closing the books on a loss. The sunk cost fallacy is, in part, a strategy for deferring the experience of loss.
But loss aversion alone does not fully explain the tenacity of sunk cost effects. Three additional psychological mechanisms are typically operating. Waste aversion is a near-universal intuition that waste is morally wrong — a useful heuristic in most contexts, but one that misfires when applied backward in time: stopping a failing project does not waste the money already spent, but the psychological grammar of "waste" constructs stopping as the wasteful act. Self-justification operates through the need to maintain a consistent, positively-valued self-image: if you were the person who decided to launch the project, stopping is not just an economic act but a biographical one, requiring revision of the story you tell about your own judgment. Consistency bias — the need to behave consistently with past behavior and stated beliefs — compounds self-justification, because reversing a major decision requires publicly acknowledging that the reasoning offered to justify it was flawed.
At the neurological level, research using fMRI has identified the anterior insula as a region particularly involved in the experience of loss and regret. When subjects anticipate closing out a losing position, the anterior insula shows heightened activation — an aversive signal that produces a pull toward avoidance. The rational prescription to ignore sunk costs requires overriding these systems, and overriding them is cognitively demanding.
Four Case Studies in the Same Error
Concorde, 1962-2003
The Anglo-French Concorde program illustrates sunk cost logic operating at the scale of nation-states. As detailed above, the project continued through at least a decade of unfavorable economic projections. By the early 1970s, when oil prices quadrupled and the economics of fuel-hungry supersonic flight worsened further, internal analyses in both governments confirmed what outside observers were saying. The rational course would have been termination.
What kept Concorde alive was partly a bilateral treaty that contained no withdrawal clause — a structural mechanism that converted sunk cost logic into a legal obligation. The two governments had built the irrationality into the contract. The phrase "the Concorde fallacy" entered academic literature as a synonym for sunk cost fallacy, partly from this episode. Commercial service ended in 2003, not because of economics but because of the Air France crash in 2000 and the post-9/11 collapse in business travel demand. The airplane flew its last passenger service on November 26, 2003 — forty-one years after the treaty that committed two governments to build it.
Vietnam, 1965-1973
By late 1967, Defense Secretary Robert McNamara had concluded privately that the United States was not winning in Vietnam and could not win on its current trajectory. His classified memorandum to President Johnson in May 1967 stated directly that continued escalation would not produce victory. McNamara himself was gone by February 1968.
What followed was not disengagement but further escalation. The logic expressed in internal administration discussions and public statements repeatedly invoked what had already been sacrificed: the lives, the resources, the political capital. To withdraw, as President Nixon would frame it years later, would be to abandon those who had already given everything. "We cannot dishonor their sacrifice" is sunk cost reasoning applied to human life — the most emotionally overwhelming version of the fallacy. The war continued for six more years after McNamara's resignation. The additional deaths on all sides were the consequence of decisions shaped in large part by the weight of what had already been lost.
Motorola Iridium, 1987-1999
In 1987, Motorola engineers conceived of a satellite phone network that would provide global mobile coverage. The concept was elegant: 66 low-earth-orbit satellites would allow users to make calls from anywhere on Earth's surface. By the mid-1990s, the economics had already deteriorated: cellular networks were expanding rapidly, coverage gaps were closing, and the per-device cost of Iridium hardware was projected at $3,000 per handset with per-minute charges of $3 to $8.
When Iridium LLC launched commercial service in November 1998, the cellular market it had been designed to serve had largely been captured by cheaper competitors. The company had spent approximately $5 billion developing and launching its satellite constellation. The rational analysis was clear: the market for $3,000 phones with $3/minute charges in a world of expanding cellular coverage was too small to service the debt load. Iridium LLC filed for bankruptcy in August 1999 — nine months after launch. A later buyer acquired the entire constellation for $25 million. Five billion dollars had produced an asset that sold for one-half of one percent of its cost.
Robert Falcon Scott's Antarctic Expedition, 1912
Scott's Terra Nova Expedition illustrates sunk cost effects at the most personal scale. When Scott's five-man polar party reached the South Pole on January 17, 1912, they found Roald Amundsen's Norwegian flag already planted more than a month earlier. The primary objective had been preempted. The return march of 800 miles began in deteriorating conditions.
The expedition's entire structure had been built around reaching the pole. The investment of two years — the ship, the preparations, the journey, the sacrifice of careers — had been predicated on that goal. When the goal was found already achieved by others, there was no rational reconfiguration of the remaining journey in light of that new reality. The team pressed on in the manner the original plan required, in service of a mission that had already, in the relevant sense, failed. All five men died on the return journey. The sunk cost here was not primarily financial but temporal and existential — and it was no less binding for that.
Sunk Costs Across Human Experience
The fallacy operates across the full spectrum of human investment.
In personal relationships, the years spent in a partnership constitute a sunk cost. Staying in a relationship that no longer serves either party because of "how much we've invested in this" is structurally identical to continuing to fund Concorde. Relationship researchers have documented that length of relationship predicts likelihood of staying even when controlling for relationship quality — partly through exactly this mechanism.
In investing, the disposition effect — the well-documented tendency of retail investors to hold losing stocks too long and sell winning stocks too soon — is the sunk cost fallacy in financial form. Investors anchor to the price they paid. A stock bought at $100 that has fallen to $60 is evaluated not on its forward-looking expected return but on the gap between $100 and $60, which feels like a loss that must be recovered before selling can be contemplated. The relevant question is whether the stock at $60 represents a better use of capital than available alternatives — not whether it will return to $100.
In business strategy, sunk cost thinking is one of the primary reasons companies continue failing product lines, maintain legacy systems, and persist with failed strategies longer than economic analysis would justify. Barry Staw's research on escalation of commitment — documented in his 1976 paper "Knee-Deep in the Big Muddy" — showed that organizations increase investment in failing initiatives after negative feedback, rather than reassessing.
In consumption and daily life, the sunk cost fallacy appears in the reader who forces herself to finish a bad book because she has read 200 pages, in the cinema patron who stays through a terrible film because the ticket cost $15, in the restaurant diner who overeats because the prix-fixe was expensive. These are trivial examples individually but illustrate how pervasive the fallacy's logic is. The rational prescription in each case is identical: evaluate the next action on its own merits, independent of what has already been paid.
The Intellectual Lineage
The formal study of sunk costs as a cognitive phenomenon has a relatively recent but well-defined intellectual history.
Richard Thaler's 1980 paper "Toward a Positive Theory of Consumer Choice" in the Journal of Economic Behavior and Organization introduced mental accounting — the way people categorize and track financial outcomes in psychologically distinct mental accounts rather than treating money as fungible. Sunk cost effects are, in Thaler's framework, a consequence of people keeping open the mental account for a prior investment and feeling compelled to "close" it with a positive outcome before the account can be settled.
Kahneman and Tversky's prospect theory (1979) provided the deeper psychological substrate — demonstrating that losses loom larger than gains and that people are risk-seeking in the domain of losses, which explains why accepting and closing a loss is so aversive.
Barry Staw's research, beginning with "Knee-Deep in the Big Muddy" (1976) and continued in "The Escalation of Commitment to a Course of Action" (Academy of Management Review, 1981), moved the analysis from individual cognition to organizational dynamics. Staw documented that decision-makers personally responsible for initial commitments escalated further than those who had inherited the decision — a finding with profound implications for how organizations structure accountability.
Arkes and Blumer's 1985 study provided the cleanest empirical demonstration of the pure sunk cost effect in controlled conditions, distinguishing it from related phenomena and establishing the effect's robustness across multiple experimental contexts.
What the Research Shows
The Arkes and Blumer theater ticket study shows the effect most cleanly: students who paid full price for season tickets were measurably more likely to attend performances than discounted-price purchasers. The effect was most pronounced for early-season performances — suggesting that the sunk cost pull diminishes somewhat over time as the original payment becomes psychologically "older."
H. Garland's 1990 study, "Throwing Good Money After Bad" (Journal of Applied Psychology), examined sunk cost effects in project management. Subjects told that a larger percentage of the budget had already been spent were significantly more likely to recommend continued funding even when the forward-looking expected return was explicitly negative.
Chip Heath's 1995 study in Organizational Behavior and Human Decision Processes refined the picture: sunk cost effects are moderated by whether the investment has been mentally "used up." People who had gotten substantial enjoyment from a prior investment were less subject to sunk cost pull than those who felt they hadn't received "their money's worth" — suggesting the fallacy is not simply about monetary amounts but about the degree to which the mental account has been satisfied.
Staw's studies of organizational escalation showed that the effect is amplified, not dampened, by negative feedback. Organizations often increase commitment after receiving bad news rather than decreasing it — the "knee-deep" effect — explained by the combination of self-justification and the need to avoid publicly admitting prior error.
When Continuing Is Actually Rational
The existence of the sunk cost fallacy does not mean that continuing projects in the face of adversity is always irrational. This limit on the analysis is important.
Switching costs are real. Abandoning a partially completed system often requires writing off investment in learning, infrastructure, and relationships, while the replacement incurs those same costs fresh. Commitment value is also real in strategic contexts: in competitive environments, credible commitment — the demonstrated willingness to follow through on investments regardless of short-term costs — can itself be strategically valuable, as Thomas Schelling's work on commitment in bargaining theory demonstrated. Option value may argue for continuation when circumstances are expected to change and reversal would eliminate future opportunities.
The danger of being too trigger-happy in abandoning projects is real. The distinction lies in the quality of the prospective reasoning: is the decision to continue or stop being made on the basis of future costs and benefits, or on the basis of what has already been spent?
The Concorde fallacy names the error; it does not prohibit continuation. It prohibits continuation for the wrong reasons. What Concorde needed, and never got, was a decision-maker willing to say in 1965 or 1968 or 1971: the money already spent is gone. What should we do from here?
References
Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124-140.
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263-291.
Thaler, R. H. (1980). Toward a positive theory of consumer choice. Journal of Economic Behavior and Organization, 1(1), 39-60.
Staw, B. M. (1976). Knee-deep in the Big Muddy: A study of escalating commitment to a chosen course of action. Organizational Behavior and Human Performance, 16(1), 27-44.
Staw, B. M. (1981). The escalation of commitment to a course of action. Academy of Management Review, 6(4), 577-587.
Garland, H. (1990). Throwing good money after bad: The effect of sunk costs on the decision to escalate commitment to an ongoing project. Journal of Applied Psychology, 75(6), 728-731.
Heath, C. (1995). Escalation and de-escalation of commitment in response to sunk costs: The role of budgeting in mental accounting. Organizational Behavior and Human Decision Processes, 62(1), 38-54.
Tversky, A., & Kahneman, D. (1991). Loss aversion in riskless choice: A reference-dependent model. Quarterly Journal of Economics, 106(4), 1039-1061.
Staw, B. M., & Ross, J. (1987). Behavior in escalation situations: Antecedents, prototypes, and solutions. Research in Organizational Behavior, 9, 39-78.
Henderson, P. D. (1977). Two British errors: Their probable size and some possible lessons. Oxford Economic Papers, 29(2), 159-205.
McNamara, R. S. (1995). In Retrospect: The Tragedy and Lessons of Vietnam. Times Books.
Frequently Asked Questions
What is the sunk cost fallacy?
The sunk cost fallacy is the tendency to continue an endeavor because of previously invested resources — money, time, effort, or emotional capital — rather than on the basis of expected future outcomes. A sunk cost is any cost that has already been incurred and cannot be recovered regardless of future decisions. Rational decision theory holds that such costs should be irrelevant to forward-looking choices.
What is the most famous example of the sunk cost fallacy?
The Concorde supersonic aircraft is the defining example. By 1965, three years after the Anglo-French treaty, internal Treasury assessments projected the aircraft would never recoup development costs. The project continued for another decade, eventually costing over 1 billion pounds. Only 20 aircraft were built for commercial use, all sold at heavily subsidized prices. The phrase the Concorde fallacy became a synonym for sunk cost fallacy in academic literature.
Why do people commit the sunk cost fallacy?
Three main mechanisms drive it. Loss aversion (Kahneman and Tversky 1979) makes stopping painful because it crystallizes an acknowledged loss. Waste aversion is the near-universal moral intuition that waste is wrong, which misfires when applied backward in time — stopping a project does not waste money already spent. Self-justification operates through the need to maintain a positive self-image: if you publicly committed to a project, stopping requires admitting your earlier judgment was wrong.
What did the Arkes and Blumer study find?
Hal Arkes and Catherine Blumer published the foundational controlled study in 1985 in Organizational Behavior and Human Decision Processes. Ohio University students who paid full price for theater season tickets attended significantly more performances than students who received discounted tickets — even in bad weather and on inconvenient evenings. Since the money was spent regardless, rational decision theory predicts identical attendance rates. The study confirmed that sunk costs systematically influence behavior they should not.
How does the sunk cost fallacy relate to escalation of commitment?
Escalation of commitment (Barry Staw 1976) is sunk cost fallacy with a social and self-justificatory dimension: organizations increase investment in failing initiatives specifically because of failure, to prove earlier commitment was justified. Staw showed that decision-makers personally responsible for initial commitments escalated further than those who inherited the decision, and that organizations often increase commitment after receiving bad news rather than reassessing.
When is it rational to continue despite sunk costs?
Switching costs are real — abandoning a partially completed system forces you to incur startup costs again on a replacement. Commitment value matters in strategic contexts where a reputation for following through is itself valuable (Thomas Schelling applied this to deterrence theory). Option value may argue for continuation when circumstances are expected to change. The key distinction is whether the decision is based on future costs and benefits, or primarily on what has already been spent.
How can you avoid the sunk cost fallacy?
The most practical technique is to explicitly ask: if I were starting fresh today, with no prior investment, would I begin this project? If the answer is no, that is diagnostic of sunk cost reasoning. A related exercise is to separate the question of whether past investment was worthwhile from the question of whether future investment is worthwhile — these are different questions with different answers. Structurally, organizations can counter it by having decisions reviewed by people who were not responsible for the initial commitment.