The sunk cost fallacy costs organizations billions of dollars annually and destroys countless careers, yet nearly every executive will fall prey to it at some point. This cognitive bias—the tendency to continue investing in a failing course of action because of resources already committed—explains why 70% of mergers fail, why 66% of IT projects end in partial or total failure, and why companies like Kodak, Nokia, and Blockbuster continued investing in dying business models while more agile competitors captured their markets.

The phenomenon is so pervasive that behavioral economists named it the "Concorde Fallacy" after the supersonic airliner whose development continued for decades despite clear evidence it would never be commercially viable, ultimately costing British and French governments over $2.8 billion with no possibility of return.

Understanding and combating this bias matters more today than ever. The company renamed itself around an unproven concept, illustrating how personal identity fusion with strategic decisions can amplify sunk cost thinking to catastrophic levels. Meanwhile, research consistently demonstrates that 90% of clinical drug development fails, IT project continuation decisions waste $50-150 billion annually in the United States alone, and managers spend 17% of their time managing underperforming employees they should have terminated months earlier.

The good news: four decades of research in behavioral economics, cognitive psychology, organizational behavior, and neuroscience have revealed precisely why our brains fall into this trap and, crucially, what leaders can do about it. This report synthesizes the foundational academic research, examines how the fallacy manifests across every major business domain, provides research-backed frameworks for counteracting sunk cost thinking, and explores the related psychological biases that create self-reinforcing escalation cycles.

The Scientific Origins of Sunk Cost Research

The academic study of sunk cost phenomena began in earnest with two parallel research streams in the 1970s that converged to create our modern understanding of the bias. The first emerged from Daniel Kahneman and Amos Tversky's groundbreaking work on Prospect Theory, which demonstrated that humans systematically deviate from rational economic models in predictable ways. The second came from Barry Staw's organizational behavior research on escalation of commitment, which documented how managers throw good money after bad in corporate settings.

Kahneman and Tversky's 1979 paper "Prospect Theory: An Analysis of Decision Under Risk," published in Econometrica, fundamentally reshaped economics and earned Kahneman the 2002 Nobel Prize in Economics. Their core discovery—that losses loom approximately twice as large as equivalent gains—provides the theoretical foundation for understanding sunk cost behavior.

This "loss aversion coefficient" of approximately 2.0 has been replicated across cultures; a 2023 global study spanning 19 countries and 13 languages confirmed a 90% replication rate for Prospect Theory's core predictions.

The companion concept of reference dependence further explains why prior investments distort future decisions. Value is not evaluated in absolute terms but relative to a reference point—typically the status quo. Once an organization has invested substantial resources in a project, that investment becomes incorporated into the reference point. Any decision to abandon now registers as a deviation downward, triggering loss aversion even though the rational economic analysis shows those costs are irrecoverable regardless of the continuation decision.

The Arkes & Blumer Experiments

Hal Arkes and Catherine Blumer's 1985 paper "The Psychology of Sunk Cost" in Organizational Behavior and Human Decision Processes provided the definitive experimental demonstrations of this bias. Their "radar-blank airplane" scenario presented participants with a company president's decision: having already invested $10 million in developing a stealth aircraft, with only $1 million needed to complete it, should the president continue even after learning a competitor has built a faster, more economical version?

The researchers also conducted a field experiment at Ohio University's campus theater, where customers randomly received full-price or discounted season tickets. Those who paid full price attended significantly more performances (4.11 plays versus approximately 3.3 plays) in the first half of the season, demonstrating that higher sunk costs motivated greater utilization even though all ticket holders had equal access to identical performances.

Arkes and Blumer proposed that the effect stems from "the desire not to appear wasteful"—a deeply ingrained social norm instilled from childhood. This insight connects to subsequent neuroscience research showing that the dorsolateral prefrontal cortex, involved in implementing social rules, activates during sunk cost decisions.

Escalation of Commitment

Barry Staw's escalation of commitment research, beginning with his 1976 paper "Knee-deep in the big muddy" in Organizational Behavior and Human Performance, extended these individual-level findings to organizational contexts. His central finding was counterintuitive: persons committed the greatest amount of resources to previously chosen courses of action when they were personally responsible for negative consequences.

A 2012 meta-analysis by Sleesman and colleagues, synthesizing 35 years of escalation research in the Academy of Management Journal, confirmed that personal responsibility significantly increases escalation tendency. Their analysis identified multiple overlapping mechanisms: self-justification (protecting ego), prospect theory (loss aversion), and agency theory (managers acting in their own rather than shareholders' interests).

What Neuroscience Reveals About Sunk Cost Processing

Recent advances in neuroimaging have localized sunk cost processing to specific brain circuits, revealing why this bias is so difficult to overcome through willpower alone. The research shows that sunk cost decisions involve an interaction between emotional and cognitive processing systems, with emotional circuits often overriding rational evaluation.

The insula, a region associated with negative emotional processing and anticipatory affect, plays a central role. Fujino and colleagues' 2016 study in Scientific Reports used functional magnetic resonance imaging (fMRI) with 32 participants completing a modified sunk cost task. They found that left insula activation correlated positively with individual differences in sunk cost susceptibility.

Research by Haller and Schwabe (2014) published in NeuroImage identified another critical finding: previous investments reduced the contribution of the ventromedial prefrontal cortex (vmPFC) to current decision-making. The vmPFC is essential for value computation—calculating expected future value based on costs and benefits. When sunk costs were high, this rational value-computation region showed decreased activity, while the amygdala (emotional processing), anterior cingulate cortex (conflict monitoring), and dorsolateral prefrontal cortex showed increased activity.

Zeng and colleagues (2013) reported in Brain Research that higher sunk costs increased activity in lateral frontal and parietal cortices involved in risk-taking behavior. Critically, no overlapping brain areas responded to both sunk cost and incremental cost—suggesting these are processed by entirely different neural systems. This helps explain why simply knowing that sunk costs should be ignored does not eliminate their influence.

A 2024 Oxford study combining neuroimaging with lesion studies found that patients with damage to key sunk-cost-processing regions were more flexible about switching to better goals. Research by Sweis and colleagues published in Science in 2018 demonstrated sunk cost sensitivity in mice, rats, and humans, with all three species more likely to persist with suboptimal options after time had been invested.

The implications for leaders are profound. Overcoming this bias requires structural interventions that change the decision environment rather than simply exhorting people to "be more rational."

Who Falls Hardest: Individual Differences in Susceptibility

Not everyone is equally susceptible to the sunk cost fallacy. Research has identified personality traits, cognitive factors, demographic variables, and emotional states that predict who will be most affected—and, importantly, who might be most effective as "debiasers" in organizational settings.

Personality research reveals a counterintuitive finding: those who most strongly internalize social rules are most vulnerable. Fujino's 2016 study found significant correlations between sunk cost susceptibility and agreeableness (r = 0.51) and conscientiousness (r = 0.36). People high in these traits more strongly absorb social norms including "don't be wasteful," making them more likely to continue failing investments to honor prior expenditures. Neuroticism, contrary to intuition, showed no significant correlation.

Age and Experience Effects

Age differences are particularly relevant for organizational design. Multiple studies, including research by Strough and colleagues (2008) and Bruine de Bruin and colleagues (2007), have established that older adults are less susceptible to sunk cost fallacy than younger adults. Older adults focus less on negative information generally (a phenomenon called the "positivity bias") and may also have accumulated more experience recognizing sunk cost situations.

Indeed, general cognitive ability does not appear to reduce susceptibility. Research by Haita-Falah (2017) found that raw intelligence fails to protect against sunk cost bias. However, Ronayne, Sgroi, and Tuckwell (2021), publishing in the Journal of Economic Behavior and Organization, found that cognitive reflection—the tendency to override intuitive responses with more careful analysis—does predict resistance.

Childhood and Emotional Factors

Childhood socioeconomic status predicts susceptibility in surprising ways. Jhang's 2023 study in Psychology & Marketing found that individuals from lower childhood SES backgrounds are more susceptible, with the effect mediated by perceiving loss of prior investments as more wasteful. Critically, childhood SES is a stronger predictor than current wealth.

Emotional and affective states also matter. Wong, Yik, and Kwong's research showed that state orientation—the tendency to ruminate about past events—increases susceptibility, while action orientation—the tendency to let go of the past and focus on new actions—decreases it.

How Sunk Costs Destroy IT Projects and Megaprojects

Nowhere does the sunk cost fallacy inflict more damage than in large-scale projects where massive upfront investments create enormous psychological pressure to continue regardless of evidence. The statistics are sobering.

The 1994 CHAOS Report established the baseline: only 16.2% of IT projects were completed on-time, on-budget, with full features. An additional 31.1% were canceled entirely, while 52.7% were "challenged"—over-budget, over-time, or delivering fewer features than promised. Projects averaged 189% of original cost estimates. Large companies fared worst, with only 9% of large company projects succeeding.

The economic impact is staggering. Gallup estimates the U.S. economy loses $50-150 billion annually due to failed IT projects. A 2020 CISQ report calculated $260 billion in unsuccessful development project costs among U.S. firms. McKinsey found that 17% of large IT projects go so badly they threaten the company's existence—not merely project failure, but existential organizational risk.

Infrastructure projects show similar patterns. Swedish research published in ScienceDirect in 2024 found that cost escalation during planning stages is substantial and highly skewed, with a "long right tail" of catastrophic overruns. More troubling, project decisions are "effectively locked in before projects' costs and benefits have been thoroughly assessed."

The Sydney Opera House, approved in 1957 at AUD $7 million with a four-year timeline, was completed in 1973 at AUD $102 million—fourteen times over budget and ten years late. The Canadian Firearms Registry, projected at CAD $2 million, ultimately cost CAD $2 billion—a thousand-fold overrun.

The Completion Imperative: How Sunk Costs Poison M&A Decisions

Mergers and acquisitions represent among the highest-stakes decisions executives make—and among the most susceptible to sunk cost contamination. The failure statistics are remarkable.

The escalation dynamics in M&A are particularly powerful because deal-making creates multiple layers of sunk cost investment. Due diligence alone typically costs millions and requires months of executive attention. Legal fees, advisory costs, and internal resources accumulate. Teams invest enormous effort developing integration plans, conducting culture assessments, and building the case for synergies. By the time serious problems emerge, organizations have invested too much to feel comfortable walking away.

Research published in the Strategic Management Journal documented this dynamic directly. In fixed-exchange-ratio stock mergers, cost shocks during the deal period strongly predicted post-acquisition commitment: an interquartile cost increase reduced subsequent divestiture rates by 8-9%. Critically, these distortions were concentrated in firm-years where the acquiring CEO remained in office—supporting the hypothesis that personal responsibility drives irrational continuation.

The AOL-Time Warner merger remains the largest M&A failure in history, with the combined entity reporting a $100 billion net loss in 2002. Bidding wars represent a particularly dangerous dynamic—competition between pharmaceutical giants Johnson & Johnson and Boston Scientific for Guidant led to dramatic overpayment as each bidder focused on winning the auction rather than value creation.

McKinsey analysis of 2,500 deals from 2013-2018 confirmed that larger transactions fail at higher rates, with 70% of companies overestimating expected synergies. The due diligence investment trap is real: teams who have invested months developing deal rationales become advocates rather than objective evaluators.

Why We Keep Bad Hires Too Long and Bad Strategies Too Late

The sunk cost fallacy extends beyond major capital investments to everyday organizational decisions about people and strategy. In both domains, the pattern is consistent: leaders invest substantial resources, encounter evidence of problems, and respond by investing more rather than cutting losses.

The Bad Hire Problem

Hiring decisions are particularly susceptible because they involve explicit human investment decisions. The U.S. Department of Labor estimates bad hires cost up to 30% of first-year salary; other estimates reach $240,000 per bad hire when productivity losses, management time, and team disruption are included. CareerBuilder research found that 74% of companies report making at least one bad hire annually.

Managers retain underperforming employees for classic sunk cost reasons. Recruiting costs average $4,129 per hire according to SHRM data, and training, onboarding, and ramp-up investments add substantially more. Managers view these expenditures as investments that must be "recouped" through continued employment—even though the investment cannot be recovered regardless of the retention decision. CFO research found supervisors spend 17% of their time managing underperforming employees.

Strategic Persistence

Strategic persistence shows identical patterns. Kodak exemplifies strategic sunk cost thinking at scale. The company invented the digital camera in 1975 and controlled over 80% of the film market at its peak. When film sales first dropped in 2001, executives blamed the 9/11 attacks and invested in marketing to preserve the film business. The company was "addicted to profits from photographic films"—financial lock-in from successful legacy products created inability to reassign resources to emerging technologies. By 2012, Kodak filed for bankruptcy.

Nokia's failure to pivot to smartphones tells a similar story. The company held over 50% global smartphone market share when the iPhone launched in 2007 and had invented the world's first smartphone in 1996. Heavy investment in the Symbian operating system created sunk cost pressure to continue investing rather than adopting Android. By 2013, market share had collapsed to 3%.

The Concorde Fallacy and Other Cautionary Tales

The Concorde supersonic airliner program provided behavioral economics with its most memorable example—so definitive that "Concorde Fallacy" became synonymous with sunk cost thinking in academic literature.

Development discussions began in England in 1956. By the early 1960s, British and French governments had committed to joint development. Initial cost estimates projected approximately £1.5 billion in today's currency. The first commercial flight launched in January 1976, with total development costs having ballooned to approximately £9.43 billion—more than six times original estimates.

Political and legal issues—particularly treaty obligations between Britain and France—made withdrawal extremely difficult. Both governments had invested so heavily they felt unable to abandon the project despite clear financial unviability. Only 20 aircraft were ever built, and the program never recouped its development costs.

Successful Escapes

Intel's 1985-1986 exit from memory chips illustrates how leaders can engineer psychological escape from sunk cost traps. Founded in 1968 as a memory company, Intel faced devastating competition from Japanese manufacturers. One manager said abandoning DRAM was "tantamount to Ford deciding to exit the car business."

The breakthrough came through what CEO Andy Grove later called the "new CEO" thought experiment. Grove asked co-founder Gordon Moore: "If we got kicked out and the board brought in a new CEO, what do you think he would do?" Moore answered: "He would get us out of memories." Grove replied: "Why shouldn't you and I walk out the door, come back in, and do it ourselves?"

The Web of Biases: How Loss Aversion, Status Quo Bias, Confirmation Bias, and Groupthink Amplify Sunk Cost Thinking

The sunk cost fallacy does not operate in isolation. It exists within a network of related cognitive biases that interact to create self-reinforcing escalation cycles.

Loss Aversion

Loss aversion provides the foundational mechanism. Kahneman and Tversky established that losses are psychologically approximately twice as powerful as equivalent gains. Abandoning a project means accepting certain loss; continuing means accepting risk with the possibility of eventual recovery. The asymmetry systematically biases toward continuation.

Status Quo Bias

Status quo bias, formally described by Samuelson and Zeckhauser (1988), refers to the preference for the current state that leads to resistance to change even when better alternatives exist. Critically, continuing a failing project IS the status quo. UCL neuroscience research found that the more difficult a decision, the more likely people are to accept the status quo.

Confirmation Bias

Confirmation bias compounds these effects by shaping how decision-makers process information. Leaders who have invested in projects actively filter out negative signals, seek evidence supporting the original decision, and construct narratives that justify continuation. Kodak's management "discounted the potential threat of digital photography" partly through confirmation bias about analog technology superiority.

Ego and Identity Protection

Ego and identity protection operate through self-justification processes. Mark Zuckerberg renamed his entire company around the metaverse—any acknowledgment of error threatened not just a strategic bet but personal identity. Cognitive dissonance creates psychological discomfort when actions conflict with beliefs. To reduce dissonance, leaders rationalize: "If I just invest more, it will work."

Groupthink

Groupthink, identified by Irving Janis (1972), describes how cohesive groups develop pressure toward unanimity that overrides realistic appraisal. Groups can amplify sunk cost fallacy through several mechanisms: social pressure makes no one want to kill a project; collective rationalization reinforces positive framing; diffusion of responsibility means "we all decided."

Breaking this cycle requires structural interventions at multiple points.

Practical Frameworks for Escaping Sunk Cost Traps

Four decades of research have produced validated frameworks that organizations can implement to counteract sunk cost thinking. These approaches work by restructuring decision environments rather than simply exhorting people to be more rational.

Zero-Based Thinking

Zero-based thinking (ZBT), developed by Brian Tracy, provides a powerful reframing technique. The core question is: "Knowing what I now know, would I still make the same decision today?"

This approach starts from a "zero point"—a blank slate perspective—and analyzes current circumstances independent of past investments. If the answer is "no," the leader immediately identifies and pursues alternate courses regardless of previous investment.

The "New CEO" Thought Experiment

The "new CEO" thought experiment operationalizes similar logic for leadership teams. Imagine a new CEO or fresh decision-maker has been appointed with no emotional attachment to past decisions. What would they decide? This technique works because a different part of the brain activates when thinking about the future—the same part that activates when thinking about strangers.

Pre-Mortem Analysis

Pre-mortem analysis, developed by psychologist Gary Klein approximately 30 years ago, represents one of the most effective debiasing techniques available. The method has been endorsed by Nobel laureates Kahneman and Thaler. A 1989 study found that "prospective hindsight"—imagining an event has already occurred—increases ability to correctly identify reasons for future outcomes by 30%.

Stage-Gate Processes

Stage-gate processes, developed by Robert G. Cooper, are used by Procter & Gamble, 3M, Emerson Electric, and countless other enterprises. The structure alternates stages (where project activities occur) with gates (decision points for Go/Kill/Hold/Recycle decisions).

The critical success factor is what practitioners call "Gates with Teeth"—tough Go/Kill decisions are among the top drivers of successful Stage-Gate implementation. Gates are decision meetings, not status reports. Cross-functional gatekeepers who own resources and report directly to executive level—rather than project sponsors—provide structural independence.

Kill Criteria

Kill criteria established upfront remove emotion from termination decisions. The principle: establish threshold criteria at project outset defining when a project falls short of expectations. For example, if a new product is expected to deliver $5M by year-end, determine the threshold below which investment doesn't make sense—perhaps $4M—accounting for alternative investment opportunities.

Upfront kill criteria force deeper analysis of expected benefits, identify projects that never had a chance, improve benefit prediction, and separate kill decisions from the emotional moment of termination.

Creating Organizational Conditions for Rational Abandonment

Beyond individual decision frameworks, organizations must create structural conditions and cultural norms that make rational abandonment possible. This requires addressing the psychological safety deficit that causes employees to continue failing initiatives rather than face the stigma of project termination.

Psychological Safety

Amy Edmondson's research at Harvard Business School provides the foundation. Psychological safety is "the belief that one will not be punished or humiliated for speaking up with ideas, questions, concerns, or mistakes." Edmondson's hospital studies revealed a paradox: better teams reported higher error rates—not because they made more errors, but because they were more willing to talk about them.

Google's Project Aristotle found psychological safety was the #1 factor distinguishing high-performing teams—"even extremely smart, high-powered employees needed a psychologically safe work environment to contribute the talents they had to offer."

De-Stigmatizing Project Termination

De-stigmatizing project termination requires explicit counter-messaging. Organizations should reframe termination as "learning" rather than "failure," celebrate "intelligent failures" (failures from well-designed experiments), and reward the stop decision by recognizing managers who terminate failing projects early.

Some organizations create explicit "failure funds"—budget allocations for expected project terminations that normalize stopping as part of the innovation process.

Structural Separation

Structural separation of advocates from evaluators provides another critical safeguard. Research shows that evaluation reports commissioned by operative units are systematically more positive than those from central evaluation units. For major projects, gatekeepers should be cross-functional senior groups—not project sponsors. External evaluators provide views considered more objective because they are not part of the organization's power structure.

Diagnostic Questions and Implementation Roadmap

Leaders seeking to implement these insights should begin with diagnostic questions that can identify sunk cost thinking in themselves and their organizations.

Red Flag Phrases

Red flags in decision-making language signal sunk cost thinking:

• "We've already invested too much to stop now."
• "We can't let that investment go to waste."
• "We're so close—we just need to push through."
• "We've come this far..."
• "After all the work we've put in..."
• "It would be a shame to abandon it now."

When these phrases appear in project reviews, strategic discussions, or personnel decisions, they warrant immediate examination.

Distinguishing Legitimate from Illegitimate Reasons

Implementation Roadmap

Immediate (0-30 days): Introduce the zero-based thinking question in the next major project review. Conduct a pre-mortem on one current high-stakes project. Train the leadership team on psychological safety concepts.

Short term (30-90 days): Establish kill criteria for all major projects. Implement stage-gate review processes with cross-functional gatekeepers. Create decision journal templates for executives.

Medium term (3-6 months): Separate project advocates from evaluators structurally. Establish an independent project review board. Develop company-specific diagnostic questions for sunk cost detection.

Long term (6-12 months): Build culture metrics around psychological safety. Create rewards for intelligent project termination. Implement regular post-decision audits.

Conclusion: The Discipline of Prospective Decision-Making

The sunk cost fallacy represents one of the most costly cognitive biases in organizational life—responsible for failed projects, destroyed shareholder value, ruined careers, and missed opportunities measured in the billions of dollars annually. Yet the bias persists not because leaders are ignorant or lazy but because it emerges from evolutionarily ancient neural circuits, is reinforced by deeply ingrained social norms against waste, and is amplified by a web of related biases.

The executives who build these capabilities into their organizations will not eliminate sunk cost thinking—the neural circuits are too deeply embedded for that. But they will create conditions where the bias can be recognized and counteracted, where rational abandonment is possible and even rewarded, and where resources flow to opportunities with genuine future value rather than to efforts sustained only by the weight of past investment.

In a business environment where 70% of mergers fail, 66% of IT projects end in partial or total failure, and 90% of drug development efforts never reach patients, this discipline of prospective decision-making represents a significant and sustainable competitive advantage.

That discipline—the courage to evaluate investments on prospective rather than retrospective terms—remains the essential leadership capability that separates organizations that thrive from those that continue investing in the past while the future passes them by.