And Amos Tversky The Sunk Cost Fallacy Experiments – Hal Arkes and Catherine

In the formal architecture of classical decision theory, human rationality is presumed to operate through forward-looking optimization. Agents are modeled as calculating entities that evaluate future states of the world, weigh contingent utilities against their corresponding probabilities, and discard historical expenditures as inert artifacts of the past. Within this normative framework, unrecoverable investments—whether denominated in capital, physical labor, temporal duration, or psychological anguish—exert zero marginal impact on normative optimization. The economic past is mathematically irreversible, fixed in the ledger of spent resources, leaving only prospective marginal costs and prospective marginal benefits to govern rational choice. Yet, across centuries of statecraft, corporate enterprise, consumer behavior, and personal relationships, human decision-makers systematically violate this foundational prescription. They persist in unviable endeavors precisely because they have already invested in them, compounding past loss with prospective ruin under the compelling psychological conviction that abandonment retroactively invalidates their historical commitment.

This persistent systemic departure from neoclassical economic orthodoxy is recognized today as the sunk cost fallacy, alternatively termed the sunk cost effect. While the intuitive manifestation of this behavioral pathology had long been recognized by observational social critics and institutional economists, its empirical deconstruction and theoretical formalization required a revolution in cognitive psychology. The intellectual scaffolding was erected through two monumental breakthroughs during the late twentieth century. First, the cognitive revolution spearheaded by Amos Tversky and Daniel Kahneman overturned the descriptive validity of expected utility theory through the introduction of prospect theory and cognitive framing paradigms. By demonstrating that human beings evaluate outcomes through subjective reference points and experience an asymmetric, nonlinear aversion to losses, Tversky provided the analytical machinery necessary to comprehend why unrecovered past expenditures alter the risk profile of future decisions.

Second, in a seminal 1985 paper published in Organizational Behavior and Human Decision Processes, psychologists Hal R. Arkes and Catherine Blumer subjected this theoretical framework to rigorous, multifaceted empirical isolation. Operating from Ohio University, Arkes and Blumer engineered a brilliant battery of hypothetical industrial scenarios, hedonic consumer choice dilemmas, and a landmark double-blind randomized field experiment at a university theater box office. Their empirical findings decisively proved that sunk costs distort human choices across diverse domains, driven fundamentally by an internalized, overgeneralized heuristic: the powerful “desire not to waste.” Together, the foundational cognitive architecture articulated by Tversky and the definitive empirical operationalization executed by Arkes and Blumer exposed the profound psychological mechanisms that divorce real-world human judgment from sterile economic axioms. The following treatise provides an exhaustive analysis of this intellectual convergence, tracing the theoretical foundations, empirical breakthroughs, cognitive drivers, mathematical dynamics, and systemic ramifications of the sunk cost fallacy.

1. Theoretical Foundations of Rational Choice and Normative Economics

1.1 The Neoclassical Axiom of Sunk Cost Irrelevance

The conceptual bedrock of neoclassical economics rests upon the foundational premise that economic agents act as forward-looking utility maximizers. Within this axiomatic structure, a sunk cost is formally defined as an expenditure—whether of financial capital, human labor, material resources, or time—that has already been incurred and cannot be recovered through any prospective course of action. Because these outlays inhabit the unalterable past, their mathematical value in forward-looking decision vectors is effectively zero. In marginalist economic theory, as formulated by figures such as Léon Walras, William Stanley Jevons, and Alfred Marshall, optimal resource allocation is dictated strictly by the condition that marginal revenue equals marginal cost ($MR = MC$). In evaluating whether to continue, expand, or terminate an ongoing endeavor, the decision-maker must weigh solely the incremental costs required to complete or sustain the project against the prospective marginal revenues or direct utility streams generated by its completion.

The mathematical irrelevance of sunk costs is formalized within the expected utility framework established by John von Neumann and Oskar Morgenstern in their 1944 treatise, Theory of Games and Economic Behavior. The Von Neumann-Morgenstern utility axioms—specifically the axioms of completeness, transitivity, continuity, and independence—presuppose that preferences are evaluated over ultimate future wealth states rather than historical paths traversed to reach those states. Let $W_0$ represent the initial wealth state of an agent, $I_s$ the unrecoverable sunk investment made in an ongoing project, $C_f$ the prospective forward-looking marginal cost required to complete the project, and $R_f$ the expected prospective return. A fully rational normative agent evaluates the choice to proceed based entirely on whether the utility of the terminal wealth state $U(W_0 – I_s – C_f + R_f)$ exceeds the utility of the abandonment state $U(W_0 – I_s)$. Because the historical expenditure $I_s$ enters both sides of the inequality identically, it cancels out as an invariant constant: $R_f > C_f$. Neoclassical doctrine thus commands an unambiguous decision rule: if the anticipated marginal benefit of continuing does not surpass the anticipated marginal cost, the venture must be immediately abandoned, regardless of whether $I_s$ is one dollar or one hundred billion dollars.

This forward-looking marginalism forms the core pedagogical foundation of modern corporate finance, cost accounting, and normative microeconomics. Textbooks ubiquitously mandate that unamortized past outlays, depreciation write-downs, and historical development investments be treated as irrelevant historical records. They represent balance sheet artifacts devoid of normative authority over future operational choices. Rationality, in the strict neoclassical vernacular, is fundamentally teleological and consequentialist. It demands a dispassionate, prospective orientation wherein the past represents an immutable constraint rather than an actionable parameter. To permit past outlays to dictate future allocations is considered an error in elementary logic, a systemic cognitive pathology known as the sunk cost fallacy.

1.2 Historical Emergence of Behavioral Anomalies

Despite the formal elegance and prescriptive clarity of the marginalist axiom, an acute dissonance between theoretical optimization and practical managerial execution began to emerge across industrial, corporate, and public finance literature throughout the mid-twentieth century. Project appraisers, capital expenditure analysts, and management consultants routinely observed that actual decision-makers consistently refused to execute the prescribed marginalist calculus. When multi-year industrial initiatives, resource extraction ventures, or engineering developments faced unanticipated cost escalations, structural shifts in market demand, or catastrophic technical obstacles, executives did not coolly evaluate prospective marginal returns against alternative uses of capital. Instead, they exhibited a pervasive tendency to authorize subsequent rounds of funding, justifying the ongoing capital drain by citing the staggering quantities of money, time, and reputational capital already expended.

This empirical deviation was documented with particular clarity within capital budgeting literature. Studies analyzing corporate investment cycles repeatedly discovered that projects possessing negative net present values ($NPV < 0$) on a prospective basis were frequently preserved and infused with additional capital if substantial prior investments had already been committed. Rather than reallocating finite capital toward competing projects with demonstrably superior prospective internal rates of return, corporate boards acted as if the past expenditures retained an intrinsic entitlement to vindication. The historical capital commitment seemed to exert a gravitational pull, dragging executive boards into deeper financial entrapment under the psychological imperative to avoid writing off the asset as a total loss on corporate ledgers.

This organizational pathology achieved international prominence through the infamous public investment saga of the Anglo-French supersonic passenger airliner: the Aérospatiale/BAC Concorde. Initiated in 1962 under a joint treaty between the British and French governments, the Concorde project was envisioned as a revolutionary leap in commercial aviation. However, development costs ballooned far beyond initial projections, expanding by several hundred percent while commercial prospects plummeted in the wake of the 1973 oil crisis, noise pollution restrictions, and severe operational range limitations. Long before the aircraft achieved commercial certification, rigorous prospective financial models demonstrated that the aircraft could never achieve operational profitability. Marginal revenues would fail to recoup even ongoing marginal operational costs, let alone amortize historical development expenditures. Yet, successive British and French administrations continued to funnel billions of pounds and francs into the project. The primary justification offered to parliaments and the public was that the two nations had already invested too heavily to abandon the endeavor. The project’s monumental historical commitment was transformed into the central argument for its perpetuity, coining the term the “Concorde Fallacy” within public administration, industrial economics, and evolutionary biology.

1.3 Psychological Vulnerabilities in Standard Economic Models

The widespread persistence of the Concorde Fallacy and related capital budgeting anomalies highlighted the severe psychological deficits inherent in standard economic models. Neoclassical frameworks operated under the assumption of Homo economicus—an idealized agent endowed with unbounded computational capacity, flawless emotional detachment, and an unyielding commitment to axiomatic consistency. This theoretical construct possessed no cognitive mechanisms for regret, pride, social self-presentation, or cognitive dissonance. It assumed that economic agents perceive the world through objective probability distributions and evaluate outcomes across a unified, frictionless wealth horizon.

The first major structural critique of this hyper-rational ideal came from Herbert A. Simon, whose pioneering work on bounded rationality exposed the profound divergence between objective optimization and human cognitive reality. Simon asserted that real-world decision-makers possess neither the computational capacity nor the information-processing resources required to calculate global maxima across complex, uncertain state spaces. Instead, human judgment operates under strict cognitive constraints, relying on heuristics, simplified search procedures, and satisficing criteria. While bounded rationality provided an invaluable macro-framework for understanding cognitive limitations, it did not directly resolve why human beings systematically prioritize unrecoverable historical outlays over optimal prospective gains. Expected utility theory remained fundamentally inadequate because it treated economic choice as an emotionally sterile valuation of final static states, entirely divorced from the historical, emotional, and social trajectory that led to those states.

Human decision-makers are not passionless computational algorithms; they develop profound emotional attachments to their investments, harbor an acute aversion to public failure, and experience genuine psychological trauma when forced to acknowledge that expended resources have yielded zero utility. Standard economic models treated decision-makers as unburdened by the history of their own choices. In reality, human beings are deeply entangled with their historical commitments. The theoretical failures of neoclassical marginalism demonstrated that standard economic theory could not be salvaged by minor adjustments to its mathematical margins. What was required was a systematic, empirically grounded revolution that integrated cognitive psychology, psychophysics, and behavioral experimentation directly into the mechanics of economic judgment.

2. Amos Tversky’s Architecture: Prospect Theory and Cognitive Framing

2.1 The S-Shaped Value Function and Loss Aversion

The intellectual paradigm shift capable of explaining sunk cost entrapment materialized through the groundbreaking collaboration between cognitive psychologists Amos Tversky and Daniel Kahneman. In their historic 1979 paper, “Prospect Theory: An Analysis of Decision under Risk,” published in Econometrica, Tversky and Kahneman dismantled the descriptive validity of expected utility theory and formulated an alternative descriptive model of human decision-making under uncertainty. At the heart of prospect theory sits the mathematical and cognitive architecture of the S-shaped subjective value function, designated as $v(x)$. Unlike neoclassical utility functions, which evaluate absolute states of total wealth ($W$), the prospect theory value function evaluates outcomes exclusively as deviations—gains ($+x$) and losses ($-x$)—relative to a neutral subjective reference point ($r_0$), such that $x = W – r_0$.

The value function possesses three distinct structural properties that directly illuminate the psychological mechanics of the sunk cost fallacy:

• Reference Point Dependency: Utility is not determined by aggregate wealth but by subjective changes in wealth relative to an active reference state.
• Diminishing Marginal Sensitivity: The function is concave in the domain of gains ($v”(x) < 0$ for $x > 0$), producing risk-averse behavior for profitable prospects. Crucially, it is strictly convex in the domain of losses ($v”(x) > 0$ for $x < 0$), inducing risk-seeking behavior when decision-makers face adverse outcomes.
• Loss Aversion: The function is steeper for losses than for corresponding gains. Formally, $v'(-x) > v'(x)$ for all $x > 0$. Empirical estimations typically reveal a loss aversion coefficient $\lambda \approx 2.0 \text{ to } 2.5$, establishing that “losses loom larger than gains.” The psychological pain of forfeiting $1,000 is more than twice as acute as the subjective pleasure of acquiring$1,000.

The implications of this geometric architecture for the sunk cost effect are profound. Consider an economic actor who has committed an unrecoverable expenditure of $10 million to a struggling commercial initiative. If this historical outlay is not psychologically integrated into the baseline reference p\oint, the actor is operating within the convex, loss-averse region of the value function. In this negative domain, the subjective difference between a guaranteed loss of$10 million and an escalated, probabilistic loss of $11 million is psychologically marginal, because the slope of the convex curve flattens as negative values deepen ($v(-10) – v(-11) ll v(0) – v(-1)$). Conversely, an uncertain, long-shot prospect that holds even a marginal probability of breaking even (restoring the state to zero) offers a massive subjective leap in utility, rescuing the actor from the excruciating domain of realized loss. Consequently, past expenditures push the decision-maker into the deeply risk-seeking territory of the value function, compelling them to gamble substantial additional resources on failing projects to avert a definitive, emotionally intolerable loss.

2.2 Framing Effects and Reference Point Shifting

In parallel with the development of the value function, Amos Tversky led pioneering investigations into the cognitive mechanics of framing effects. In their canonical 1981 paper, “The Framing of Decisions and the Psychology of Choice,” Tversky and Kahneman demonstrated that mathematically identical decision problems systematically generate radically divergent choices when formulated through alternative semantic or structural frames. Human preferences are notoriously unstable and context-dependent; they shift unpredictably based on how an outcome is cognitively presented—whether as a foregone gain or as an outright loss relative to an arbitrary reference point.

Tversky’s framing paradigms revealed that historical expenditures establish entrenched, subjective reference points in human consciousness. In classical economic analysis, the only rational reference point is the present, current asset position. However, in human psychology, historical investments continuously anchor the baseline reference point to the past, before the capital was expended. When a venture encounters distress, the decision-maker does not frame abandonment as the sensible conservation of remaining capital. Rather, the act of formal termination forces an explicit, transparent frame: it converts what was an ambiguous, floating expenditure into an irreversible, undeniable, realized loss. By contrast, persisting with the failing project maintains a hidden, prospective frame where the loss remains psychologically unrealized and ostensibly recoverable.

This dynamic operates through what Tversky characterized as the psychological refusal to close an “open mental ledger” at an explicit deficit. When capital is committed to an initiative, a specialized mental tracking account is established. If the endeavor is aborted, the ledger must be closed, forcing the decision-maker to register a massive debit against their psychological balance sheet. This generates an acute state of subjective distress. To prevent this ledger closure, agents alter their risk preferences, desperately committing further capital in an attempt to manipulate reality until the account can somehow be balanced at zero. Sunk cost escalation is therefore fundamentally an artifact of cognitive framing: agents are not chasing prospective utility; they are fleeing the acute psychological agony of formalizing a catastrophic loss.

2.3 Mental Accounting in Collaborative Paradigm Construction

The cognitive framing of past expenditures was further expanded and formalized through the intellectual collaboration between Amos Tversky and behavioral economist Richard Thaler. Thaler developed the theory of mental accounting, which posits that individuals and organizations do not treat financial assets as globally fungible, as neoclassical theory mandates. Instead, human minds organize, categorize, and evaluate economic activities within discrete, compartmentalized mental accounts. These cognitive ledgers possess internal budgets, specialized reference points, and unique rules regarding when transactions can be opened, amortized, or closed.

Thaler and Tversky established a vital structural distinction between topical and comprehensive mental accounting. In comprehensive accounting—the mode mandated by rational choice theory—an economic transaction is evaluated across the individual’s total aggregate wealth and lifetime earning trajectory. Under this framing, an unrecoverable loss of $1,000 on a vacation or an industrial prototype is simply absorbed as a negligible historical shift in total lifetime assets. In contrast, human cognition relies overwhelmingly on topical mental accounting. Under topical accounting, the financial outlay is evaluated strictly within the narrow, isolated boundaries of the specific context in which it occurred. The expenditure is inextricably tethered to the focal transaction.

Within a topical mental account, unamortized sunk costs remain intensely active, functioning as psychological debits that demand corresponding consumption or future revenue to achieve parity. Thaler demonstrated that individuals experience profound transaction utility—the perceived merit, fairness, or psychological validation derived from the financial terms of a deal—which operates alongside standard acquisition utility. When an individual purchases an expensive membership, equipment, or capital asset, the mental account registers an acute negative balance. The account can only be psychologically balanced if the asset is consistently used, consumed, or maintained until its perceived utility matches or exceeds the initial cash outlay. To abandon the endeavor prematurely is to leave the topical mental ledger permanently imbalanced, inflicting a sustained penalty of transaction disutility. Thaler’s mental accounting framework provided the indispensable conceptual bridge connecting Tversky’s mathematical loss aversion to the lived experience of consumer and executive decision-making, setting the stage for direct empirical isolation.

3. The 1985 Landmark Collaboration: Hal Arkes and Catherine Blumer

3.1 Context and Objectives of the Ohio Experiments

By the early 1980s, behavioral economics possessed a formidable theoretical foundation in the work of Tversky, Kahneman, and Thaler, as well as compelling organizational case studies in Barry Staw’s literature on the escalation of commitment. However, empirical psychology lacked a unified, controlled, experimental demonstration that could definitively isolate the sunk cost effect from competing, confounding economic variables. Skeptical neoclassical economists persistently argued that observed organizational persistence was not an irrational psychological fallacy at all. Instead, they claimed it reflected rational information signaling: ongoing project continuation could simply reflect superior private information held by managers, reputation preservation strategies, or rational calculations regarding salvageable learning curves and real options value.

Recognizing this critical methodological impasse, experimental psychologist Hal R. Arkes and his doctoral student Catherine Blumer at Ohio University initiated an ambitious, meticulously designed research project. Their operational objective, culminated in their historic 1985 paper titled “The Psychology of Sunk Cost,” was to construct controlled, experimental environments where all neoclassical economic confounds were systematically eliminated. By stripping away private information asymmetries, prospective revenue advantages, future technological spillover values, and reputational mechanisms, Arkes and Blumer aimed to expose the raw, irreducible psychological machinery of the sunk cost effect.

The intellectual mandate was uncompromising: if experimental subjects were presented with scenarios where marginal costs unequivocally exceeded prospective marginal benefits, any systematic preference to persist predicated solely on historical financial investment would constitute unassailable proof of cognitive irrationality. Arkes and Blumer sought to demonstrate that the sunk cost fallacy was not merely a rare pathology of dysfunctional corporate boards or eccentric geopolitical alliances, but an ingrained, pervasive cognitive vulnerability embedded across ordinary human judgment. Their experimental battery was designed to interrogate consumer decision-making, industrial capital allocation, and physical, real-world monetary behaviors in the wild.

3.2 Defining the ‘Desire Not to Waste’ Phenomenon

The central theoretical contribution of Arkes and Blumer’s 1985 treatise was the identification and formal characterization of the “desire not to waste” phenomenon. Prior analyses had often attributed persistence to sophisticated psychodynamic motives, such as pride, hubris, or public impression management. Arkes and Blumer proposed a more parsimonious, yet profoundly pervasive, psychological engine: human beings are governed by an internalized, overgeneralized social rule that proscribes wastefulness. From early childhood, individuals are heavily socialized across cultures to avoid wasting food, squandering money, abandoning tasks midstream, or neglecting tools and investments.

In standard daily life, this anti-waste heuristic functions as a highly effective behavioral rule of thumb. It fosters discipline, encourages resource conservation, and enforces follow-through. However, the tragedy of human judgment lies in the cognitive inability to recognize the structural boundaries where this heuristic ceases to be applicable. Arkes and Blumer identified a fundamental cognitive conflation between actual economic waste and the superficial appearance of waste:

• Economic Reality: True economic waste occurs the exact instant resources are committed to a project, asset, or experience that yields less prospective value than the alternative use of those resources. Once spent, the capital is gone forever; continuing to throw additional resources after an unviable endeavor constitutes a secondary, compounding act of true economic waste.
• Psychological Perception: Human intuition operates in reverse. The decision-maker feels that the historical expenditure only becomes “wasted” at the precise psychological moment the project is officially terminated, abandoned, or discarded.

To ordinary human psychology, terminating a failing $10 million project without achieving its objective is perceived as “wasting$10 million.” Conversely, authorizing an additional $2 million to finish a useless, unviable prototype feels like a noble, diligent effort to “salvage” the historical$10 million. Arkes and Blumer demonstrated that this psychological dread of appearing wasteful—not only to external observers (social self-presentation) but, crucially, to oneself (private self-justification)—compels agents to commit fresh resources to doomed trajectories. Human beings prioritize the psychological fiction of resource reclamation over the cold mathematical reality of marginal efficiency.

3.3 Experimental Methodologies: Hypothetical Scenarios and Field Experiments

To establish unassailable empirical proof, Arkes and Blumer constructed a methodological program combining high-control laboratory vignettes with a pioneering, randomized field experiment. The methodological architecture was designed to bridge the gap between experimental internal validity and real-world ecological validity. In their laboratory studies, the researchers engineered stylized, mathematically unambiguous dilemmas administered to large cohorts of university students and institutional participants.

These vignettes were structured through strict between-subjects and within-subjects experimental designs. In each paradigm, the forward-looking economic calculus was held absolutely constant across conditions, while the presence, magnitude, or structure of the historical sunk cost was systematically manipulated as the independent variable. By contrasting conditions where identical future paths were paired with either large historical investments or zero prior expenditures, Arkes and Blumer could measure the precise causal impact of historical outlays on decision outputs.

Recognizing that laboratory vignettes might be critiqued as abstract, costless exercises in hypothetical decision-making, Arkes and Blumer expanded their empirical assault by designing an unprecedented randomized field intervention in a naturalistic commercial setting. Utilizing the season ticket subscriptions of the Ohio University Theater, the researchers introduced direct financial stakes, unannounced price randomizations, and longitudinal behavioral tracking spanning six months. This integration of controlled hypothetical paradigms and unobtrusive, double-blind real-world experimentation established a gold standard for empirical behavioral economics, successfully isolating cognitive drivers that had eluded social scientists for decades.

4. The Radar-Invisible Bomber Experiment: Industrial Escalation

4.1 Experimental Design and Scenario Structure

The most iconic industrial vignette in the Arkes and Blumer battery is universally recognized as the “Radar-Invisible Bomber” experiment. Designed to mirror corporate research and development (R&D) capital allocation, the study evaluated how historical investment commitments warp strategic decision-making in high-stakes technological environments. The scenario presented participants with a stark capital allocation dilemma, structured across two distinct experimental conditions.

In the Sunk Cost Condition, participants assumed the role of the chief executive officer of an aviation manufacturing enterprise, the Baxter Corporation. The scenario prompt was structured as follows:

“You are the president of the Baxter Corporation, an aerospace company. You have invested $9 million of the company’s money into a research project to build a radar-invisible plane—a bomber t\hat cannot be detected by conventional radar. When the project is 90% completed, another firm begins marketing a plane t\hat cannot be detected by radar. In addition, it is apparent t\hat their plane is much faster and far more economical than your plane would be if it were completed. The question is: should you invest the last 10% of your research funds—$1 million—to finish your radar-invisible plane?”

In the Control Condition, participants were presented with an identical technological, competitive, and forward-looking economic landscape, but with the historical sunk investment stripped completely from the scenario. These participants were instructed:

“You are the president of the Baxter Corporation, an aerospace company. It has been suggested that you invest $1 million of the company’s money into a research project to build a radar-invisible plane—a bomber t\hat cannot be detected by conventional radar. However, another firm has just begun marketing a plane t\hat cannot be detected by radar. In addition, it is apparent t\hat their plane is much faster and far more economical than your plane would be if it were built. The question is: should you invest$1 million of your research funds into this project?”

The forward-looking economic calculus in both scenarios is mathematically identical: the decision-maker must choose whether to allocate $1 million of finite corporate capital into developing an aircraft that is definitively obsolete upon arrival, technologically inferior, economically uncompetitive, and destined to capture zero viable market share. Normative economic theory dictates that the decision in both conditions must be an unequivocal, resounding rejection of the investment.

4.2 Statistical Divergence and Escalation Patterns

The empirical results generated by Arkes and Blumer revealed a staggering, statistically overwhelming divergence that stunned classical economists. In the Control Condition, where no prior investment was present, participants acted with unimpeachable economic rationality. Only 16.7% of participants (a tiny minority) recommended investing the $1 million in the inferior aircraft, while an overwhelming 83.3% rejected the proposal, opting to preserve corporate capital for superior commercial opportunities.

In stark, dramatic contrast, within the Sunk Cost Condition, the mere presence of the historical $9 million outlay completely inverted participant behavior. A massive 85.1% of participants insisted that the company should authorize the final $1 million to complete the aircraft, with only 14.9% advocating abandonment. The statistical significance of this divergence reached extraordinary thresholds ($p < 0.001$), demonstrating that the historical expenditure was not simply a minor cognitive consideration, but the overwhelmingly dominant determinant of corporate policy.

Participants in the sunk cost condition explicitly recognized that the final product was commercially and technologically dead in the water. Nevertheless, their subjective reasoning was paralyzed by the psychological dread of leaving a $9 million endeavor unfinished. To terminate the project at 90% completion felt like a catastrophic, self-inflicted admission t\hat$9 million had been utterly squandered. Completing the aircraft, even if it was functionally worthless, allowed the decision-maker to preserve the psychological illusion of task completion, thereby avoiding the devastating cognitive debit of writing off the $9 million as a total failure.

4.3 Implications for Corporate Capital Allocation

The findings of the radar-invisible plane experiment provided a definitive psychological foundation for Barry Staw’s organizational observations regarding the escalation of commitment. The experiment revealed that corporate boards and executive leadership teams do not fail primarily because of intellectual incompetence or unpredictable market volatility; they fail because their cognitive architecture is structurally ill-equipped to handle the psychology of unrecoverable investments. Executives routinely fall prey to “throwing good money after bad,” trapped in self-reinforcing spirals of capital escalation.

Within complex enterprise environments, this cognitive distortion induces severe systemic information filtering. When multi-million-dollar R&D initiatives face emerging competitive threats or fatal design flaws, managerial hierarchies consistently misinterpret ongoing project updates. Information signaling imminent failure is reframed as an unexpected, temporary engineering hurdle that merely requires one final injection of capital to resolve. The sunk cost effect weaponizes the sunk capital against the organization: the more capital a firm commits to a failing enterprise, the more psychologically agonizing it becomes to pull the plug, which in turn dramatically increases the likelihood of subsequent, catastrophic capital infusions.

This organizational pathology is heavily institutionalized within enterprise budgeting cycles. Corporate incentives frequently reward managers for “bringing projects across the finish line” rather than terminating doomed initiatives early to safeguard capital. The Baxter Corporation vignette mirrors the real-world histories of ill-fated enterprise initiatives, from catastrophic enterprise resource planning (ERP) software implementations to bloated pharmaceutical drug developments. In each case, executives prioritize the preservation of their historical investments over the prospective solvency and opportunity costs of the enterprise.

5. The Ski Trip Dilemma: Utility Versus Expenditure Misalignment

5.1 The Two-Ticket Problem Paradigm

Having exposed the mechanisms of industrial escalation, Arkes and Blumer engineered a brilliant laboratory scenario to interrogate the sunk cost effect within pure, hedonic consumer choice. While industrial decisions can be clouded by complex organizational assumptions, consumer choices strip away professional accountability, exposing the naked interaction between financial expenditures and personal experiential utility. This paradigm is immortalized in decision science as the “Ski Trip Dilemma,” or the Two-Ticket Problem.

Participants in this experiment were instructed to imagine the following sequence of events:

“Imagine that you have spent $100 for a ticket for a week\end ski trip to Michigan. Several weeks later, you buy a$50 ticket for a weekend ski trip to Wisconsin. You think you will enjoy the Wisconsin ski trip much more than the Michigan ski trip. As you are putting your just-purchased Wisconsin ticket into your wallet, you notice that the two trips are for the exact same weekend! It’s too late to sell either ticket, and you cannot get your money back for either ticket. You must use one ticket and not the other. Which ski trip will you go on?”

The analytical purity of this experimental manipulation is extraordinary. Consider the structural variables established within the prompt:

• Both financial outlays ($100 for Michigan,$50 for Wisconsin) are completely, definitively sunk, non-refundable, and non-transferable. The $150 total expenditure is irrevocably eliminated from the agent’s asset horizon.
• The temporal choice is strictly mutually exclusive: the participant can only attend one trip; the other must be forfeited.
• Crucially, the scenario explicitly establishes a definitive subjective utility hierarchy: the participant knows with absolute psychological certainty that they will enjoy the $50 Wisconsin trip far more than the$100 Michigan trip ($U(\text{Wiscon\sin}) > U(\text{Michigan})$).

Under the axioms of normative decision theory, the past expenditures are completely irrelevant. The individual is faced with a pristine choice between two forward-looking experiential states: an inferior weekend in Michigan versus a superior weekend in Wisconsin. Rational choice theory commands that the agent must select Wisconsin, maximizing their prospective hedonic utility. To choose Michigan would mean knowingly choosing a worse experience while achieving zero financial recovery, an explicit violation of the core tenet of rational utility maximization.

5.2 Empirical Results and Rational Choice Violation

When Arkes and Blumer compiled the experimental choices of their participants, the results struck at the very heart of normative consumer theory. A decisive majority—54% of participants—deliberately chose to embark on the $100 ski trip to Michigan, voluntarily consigning themselves to a weekend that they explicitly knew would be significantly less enjoyable. Only 46% of participants made the economically rational choice to travel to Wisconsin and maximize their experiential satisfaction.

This result constitutes one of the most stark, undeniable empirical violations of transitivity and utility maximization ever captured in experimental psychology. When a consumer consciously chooses to endure a worse experience solely because they spent more money on it, they are actively sacrificing their real-time welfare to appease an arbitrary financial ledger. The participants demonstrated a willingness to pay an emotional tax—subjecting themselves to suboptimal leisure—to avoid confronting the psychological agony of “wasting” a more expensive ticket.

The internal qualitative reasoning articulated by participants who chose Michigan was deeply revealing. They reported that while choosing Wisconsin might provide more pure enjoyment, going to Wisconsin would mean throwing away a $100 ticket, whereas going to Michigan meant only throwing away a$50 ticket. The psychological pain of forfeiting $100 was perceived as vastly more severe than the pain of forfeiting$50. The fact that both tickets were already bought and paid for was completely ignored by their cognitive accounting systems. The participants actively subordinated their prospective experiential happiness to minimize the subjective debit registered within their open mental accounts.

5.3 Reconciling the Findings with Tversky’s Value Function

The irrational outcome of the Ski Trip Dilemma provides a perfect empirical manifestation of Amos Tversky’s prospect theory and subjective value function. To understand why an individual chooses an inferior vacation, one must map the decision onto Tversky’s cognitive coordinates:

When the individual purchases both tickets, two discrete topical mental accounts are opened:

• Account A (Michigan): -$100 cash debit, anticipating Michigan hedonic utility.
• Account B (Wisconsin): -$50 cash debit, anticipating superior Wisconsin hedonic utility.

Because the trips overlap, one account must be immediately closed at a total loss of acquisition utility. If the individual chooses Wisconsin, Account B delivers superior positive hedonic consumption, but Account A must be closed at a complete, catastrophic deficit of –$100. Due to Tversky’s loss aversion parameter ($lambda approx 2.25$), the psychological pain of writing off the$100 ticket is magnified into a profound negative sensation ($v(-100)$). Conversely, if the individual chooses Michigan, Account B is closed at a loss of only –$50, while Account A is technically “redeemed” through consumption, sparing the agent from having to register an out\right -$100 loss.

Because losses loom substantially larger than gains, the differential psychological distress between closing a $100 debit versus a$50 debit ($v(-100) – v(-50)$) completely overwhelms the positive differential in prospective hedonic utility between the two trips ($v(\text{Utility}_{\text{Wiscon\sin}}) – v(\text{Utility}_{\text{Michigan}})$). The consumer makes the tragic, paradoxical choice to suffer through a miserable vacation in Michigan because it protects them from the acute cognitive distress of registering a triple-digit financial loss on an unclosed mental ledger. Arkes and Blumer’s two-ticket paradigm demonstrated that consumer behavior is governed not by the dispassionate pursuit of happiness, but by the frantic avoidance of financial remorse.

6. The Ohio University Theater Field Experiment: Real-World Validation

6.1 Experimental Protocol and Price Randomization

While the Radar Bomber and Ski Trip vignettes delivered undeniable evidence within controlled laboratory conditions, Arkes and Blumer recognized the imperative to demonstrate that the sunk cost fallacy was not an artifact of hypothetical survey designs. Critics from neoclassical economics frequently argued that when real money, tangible assets, and authentic personal commitments were at stake, the disciplines of market reality would wash away these supposed cognitive anomalies. To dismantle this critique, Arkes and Blumer engineered one of the most elegant, methodologically pure field experiments in the history of behavioral science, utilizing the patron base of the Ohio University Theater.

At the beginning of the 1982–1983 theater season, ordinary patrons approached the university box office to purchase their standard season subscription tickets. Unbeknownst to the patrons, Arkes and Blumer had implemented an unobtrusive, double-blind randomized pricing intervention. A total of 60 season ticket purchasers were randomly assigned across three distinct pricing tiers at the exact point of sale:

• Tier 1: Full-Price Condition: Patrons in this control cohort purchased the standard season ticket subscription at the full retail price of $15.00 (ten tickets for five distinct theatrical productions at $1.50 per ticket).
• Tier 2: Minor Discount Condition: Patrons in this experimental cohort were unexpectedly informed at the box office window that, due to an unannounced university arts promotional campaign, they had randomly been selected to receive a $2.00 discount, purchasing the season subscription for $13.00 ($1.30 per ticket).
• Tier 3: Major Discount Condition: Patrons in this experimental cohort were similarly informed that they had randomly been selected to receive an aggressive $7.00 promotional discount, purchasing the entire season subscription for a dramatically reduced price of $8.00 ($0.80 per ticket).

The experimental execution was rigorously double-blinded. Patrons had arrived at the box office with the explicit intention, willingness, and capital to purchase the tickets at the full $15.00 retail price. The discounts were awarded purely through a pre-randomized envelope system handled by box office staff who were blind to the overarching psychological hypotheses. Thus, all three cohorts shared an identical baseline of demographic composition, cultural interest, and initial motivation to attend the theater productions. The only variable that differed between the groups was the absolute magnitude of the unrecoverable financial capital sunk into the tickets.

6.2 Longitudinal Attendance Monitoring and Ticket Usage

Following the random assignment at the box office, Arkes and Blumer initiated a six-month longitudinal tracking protocol. Over the course of the theatrical season, five distinct plays were staged: Equus, A Midsummer Night’s Dream, The Sea Gull, The Children’s Hour, and Candide. To measure patron behavior without introducing observer bias or Hawthorne effects, theater personnel systematically collected and recorded the randomized ticket stubs as patrons passed through the turnstiles for each production.

The normative economic prediction for this field experiment was absolute and unambiguous: once the tickets were purchased, the price paid represented an unrecoverable sunk cost. The marginal financial cost to attend any given performance was precisely zero across all three cohorts ($MC = 0$). Prospective marginal benefits—experiencing the artistic, cultural, and recreational value of the plays—were presumed equivalent across cohorts due to the identical pre-experimental baseline motivations. Therefore, neoclassical economics dictated that attendance rates across the Full-Price ($15), Minor Discount ($13), and Major Discount ($8) groups must be statistically indistinguishable throughout the season.

The empirical field data utterly shattered the neoclassical prediction. During the first half of the season (spanning the initial theatrical productions), attendance frequencies diverged sharply as a direct function of the initial financial outlay:

• Patrons in the Full-Price ($15) Tier attended an average of 4.11 performances.
• Patrons in the Minor Discount ($13) Tier attended an average of 3.00 performances.
• Patrons in the Major Discount ($8) Tier attended an average of only 2.50 performances.

The differences were highly statistically significant. Patrons who had sunk the greatest financial investment into their subscriptions demonstrated a markedly superior compulsion to physically attend the performances. Those who had received a major, unexpected financial discount felt significantly less pressure to attend, frequently allowing their tickets to go unused. Arkes and Blumer had captured the sunk cost effect in the wild: the magnitude of an unrecoverable, historical expenditure exerted an active, coercive force over real-world physical behavior, directly dictating whether individuals showed up at an auditorium.

6.3 The Decay of the Sunk Cost Effect Over Time

While the overall attendance divergence provided definitive real-world proof of the sunk cost effect, the most profound and unexpected finding of the Ohio University Theater field experiment emerged from its longitudinal dimension. Arkes and Blumer did not simply evaluate aggregate season attendance; they meticulously analyzed attendance trajectories across the temporal sequence of the theatrical productions.

When the data was segmented between the first half of the theatrical season (Plays 1 through 3) and the second half of the season (Plays 4 and 5), a striking temporal decay dynamic materialized. The intense statistical divergence in attendance observed during the early months completely vanished as the season progressed. By the time the final two productions were staged, attendance rates between the full-price patrons and the deeply discounted patrons converged into statistical equivalence. The coercive power of the sunk cost had decayed.

This empirical discovery demonstrated that the sunk cost effect possesses a psychological “half-life.” When an expenditure is fresh, recent, and emotionally salient, the unclosed mental ledger is vibrant, generating severe psychological pressure to consume the asset and validate the expenditure. However, as chronological time elapses, the cognitive salience of the historical financial transaction naturally attenuates. The mental account gradually amortizes, fades into background memory, and loses its capacity to manipulate forward-looking behavioral choices. Arkes and Blumer’s field experiment not only proved the existence of the sunk cost fallacy in real-world economics, but it also revealed that its cognitive grip is dynamically tethered to the neurobiological mechanisms of memory, salience, and temporal decay.

7. Psychological Drivers: Deconstructing the Sunk Cost Trap

7.1 The Role of Cognitive Dissonance and Self-Justification

To fully deconstruct why human beings fall prey to sunk cost escalation, one must examine the foundational socio-psychological architecture of cognitive dissonance, first articulated by Leon Festinger in 1957. Cognitive dissonance theory posits that human beings harbor an innate, powerful drive to maintain psychological consistency among their beliefs, attitudes, self-image, and behaviors. When an individual experiences an explicit clash between their self-perception as a competent, rational, and discerning decision-maker and external real-world feedback demonstrating that an ongoing project is a catastrophic failure, profound psychological tension is generated.

In the context of the sunk cost fallacy, abandoning a project or walking away from an investment forces an immediate, unmediated confrontation with cognitive dissonance. To pull the plug is to formally validate an excruciating psychological reality: “I made a catastrophic mistake. I allocated millions of dollars, or months of my life, toward an objective that was flawed, unviable, and fruitless.” The ego cannot easily tolerate this structural assault on its competence. As Elliot Aronson and Carol Tavris have comprehensively demonstrated, human cognition relies heavily on post-decisional self-justification mechanisms to defend self-worth.

Instead of updating their beliefs to align with negative external reality, decision-makers execute aggressive cognitive gymnastics. They systematically reframe incoming feedback: negative signals are dismissed as anomalies, intermediate milestones are artificially celebrated, and the perceived viability of the failing project is inflated. Committing additional sunk capital serves a vital self-protective function: it staves off the admission of failure, effectively kicking the psychological reckoning down the road. The decision-maker gambles further resources not because they believe in the project’s objective economic fundamentals, but because persistence serves as a psychological shield, protecting their fragile self-concept from the excruciating trauma of admitting a monumental, irreversible error.

7.2 The ‘Waste Not’ Heuristic and Moral Overgeneralization

While cognitive dissonance explains the self-defensive maneuvers of the ego, Hal Arkes and Catherine Blumer placed their primary explanatory emphasis on a broader, deeply socialized cognitive flaw: the overgeneralization of the “Waste Not” heuristic. Across nearly all modern human civilizations, proverbs such as “waste not, want not,” “finish what you start,” and “clean your plate” are systematically hammered into the developing cognitive architecture of children. This socialization is rooted in profound survival logic: in resource-scarce ancestral environments, wasting gathered food, abandoning fabricated tools, or squandering tribal energy carried catastrophic survival penalties.

However, modern economic civilization operates within a highly complex, symbolic, and counter-intuitive financial environment. In this modern context, the primitive rule against waste becomes dangerously overgeneralized. Decision-makers apply the rule to non-fungible, historical economic transactions where it is structurally inapplicable. As Arkes and Blumer argued, human beings develop a profound moralization of financial expenditures. To terminate a project feels morally reprehensible, synonymous with throwing food in the trash or burning paper money.

This moral overgeneralization blinds the agent to the profound philosophical and operational distinction between abandoning an asset and wasting capital. When a machine is non-functional, an R&D venture is rendered obsolete, or a theater ticket holds no experiential value, abandoning it does not “waste” anything—the waste occurred when the capital was originally squandered on a bad idea. To force oneself to sit through a terrible movie, eat food past satiety, or pump millions into an unviable bomber out of a warped moral obligation to avoid “waste” is fundamentally absurd. The individual is not conserving resources; they are simply executing a performative ritual of self-punishment to appease a malformed moral heuristic.

7.3 Impression Management and Accountability

Beyond internal psychological dissonance and internalized moral rules, the sunk cost trap is intensely magnified by external sociopolitical dynamics, specifically impression management and social accountability. Human decision-makers do not operate in social vacuums; they inhabit complex organizational, political, and cultural hierarchies where reputation, career advancement, and perceived leadership competence dictate personal survival.

In his influential social-psychological analyses, Barry Staw proved that the structural drive to escalate commitment to a failing course of action is exponentially heightened under conditions of public accountability. When an executive or political leader champions a high-profile initiative, their professional reputation becomes inextricably fused with the project’s perceived trajectory. If the leader abruptly cancels the initiative, the organizational hierarchy does not interpret the cancellation as a brilliant, dispassionate application of neoclassical marginalism. Instead, peers, subordinates, board members, and political adversaries interpret the termination as an overt confession of strategic incompetence, executive failure, and lack of leadership conviction.

This generates a severe principal-agent problem. For the individual manager (the agent), the personal career risks of terminating a project are immediate, acute, and career-ending. Conversely, the risks of continuing the project are diffused across corporate balance sheets and delayed into the future. By pouring additional sunk capital into the endeavor, the manager preserves their reputation for tenacity and decisiveness, holding out hope that a market turnaround, an external intervention, or a miraculous technological breakthrough will rescue the project before the bill comes due. Experimental research has repeatedly demonstrated that when complete anonymity is introduced—completely shielding decision-makers from reputational penalties—the magnitude of the sunk cost effect plummets dramatically. Escalation is often a brutally cynical, locally rational survival strategy enacted by self-interested agents willing to incinerate corporate wealth to protect their personal professional standing.

8. Evolutionary, Developmental, and Cross-Species Perspectives

8.1 Ontogenetic Development of the Sunk Cost Fallacy

One of the most fascinating and counter-intuitive frontiers in sunk cost research involves its ontogenetic development—the study of how this cognitive bias emerges across the human lifespan. If the sunk cost fallacy were a primitive, hardwired biological reflex, one would expect to observe its most intense, unvarnished manifestations in very young children, who lack the sophisticated abstract reasoning capacities of adults. However, empirical investigations conducted by Hal Arkes and Paul Ayton revealed a discovery that completely upended developmental assumptions.

When young children (typically between the ages of four and seven) are presented with operational sunk cost dilemmas—such as choosing whether to continue playing with a broken, frustrating toy simply because they spent their allowance on it, or choosing which movie to watch after paying for both—they demonstrate greater normative economic rationality than fully educated adults. Young children consistently ignore historical expenditures with breathtaking ease. If an activity stops being fun, children simply drop it immediately, shifting their behavioral energy toward whatever alternative pursuit offers superior immediate utility, entirely unburdened by the financial or temporal capital they have abandoned.

Susceptibility to the sunk cost fallacy does not decrease with age and education; it increases as cognitive maturation advances. The fallacy only begins to emerge systematically during late childhood and early adolescence, precisely when the brain develops the executive capacity for abstract rule construction, socialized moralization, and complex metacognition. As children are taught generalized heuristics regarding waste, thrift, social accountability, and personal consistency, they begin to overapply these rules across inappropriate contexts. The sunk cost fallacy is not an evolutionary flaw of our primitive animal minds; it is an over-intellectualized cognitive trap produced by advanced human symbolic socialization.

8.2 Comparative Cognition: Non-Human Animal Studies

The developmental discovery that children are immune to the sunk cost trap catalyzed profound scientific curiosity regarding comparative cognition: Do non-human animals fall prey to the sunk cost fallacy? For decades, evolutionary biologists debated whether animals exhibit the “Concorde Fallacy” when defending territories, incubating eggs, or maintaining constructed nests. If an animal has spent weeks building a nest or incubating a clutch of eggs, does it defend that investment more fiercely when faced with a lethal predator than if the clutch were newly laid?

Pioneering experiments across diverse species—including pigeons, rodents, and non-human primates—revealed complex behavioral nuances. When classical conditioning paradigms were constructed to test whether animals would commit more behavioral effort (such as lever presses or key pecks) to obtain a reward simply because they had already committed substantial effort toward it, animals generally conformed to optimal foraging models. In natural environments, non-human animals operate under evolutionary decision rules governed by Eric Charnov’s marginal value theorem. When a foraging patch is depleted and the marginal rate of energetic return drops below the average return rate of the wider environment, animals abandon the patch immediately. They do not linger in an empty berry bush out of respect for the calories they already burned searching for it.

Recent neuro-computational studies in rodents and macaque monkeys conducted by researchers such as Brian Sweis and colleagues have revealed subtle behavioral escalation patterns when animals are placed in structured “offer-delay” foraging tasks. Animals do occasionally show an increased reluctance to abandon a delayed reward queue if they have already waited a substantial period. However, behavioral neuroscientists have demonstrated that these animal behaviors are fundamentally governed by simple neurobiological mechanisms of time-delay discounting and local reinforcement learning, rather than the rich, narrative-driven, self-justifying cognitive frameworks observed in humans. Non-human animals lack the symbolic self-consciousness required to construct an ego that worries about looking foolish or feeling like a failure. The full, tragic architecture of the sunk cost fallacy remains a uniquely human cognitive affliction.

8.3 Adaptive Logic of Persistence

To understand why this cognitive vulnerability became so deeply entrenched within our species, evolutionary psychologists examine the adaptive logic of behavioral persistence in ancestral environments. The human brain was not engineered in high-finance trading rooms or corporate boardrooms evaluating multi-million-dollar discounted cash flows. It was forged in Pleistocene hunter-gatherer habitats characterized by physical volatility, caloric scarcity, and communal interdependence.

In an ancestral foraging environment, unwavering persistence—what modern psychologists term grit, tenacity, and steadfastness—was overwhelmingly adaptive. Hunting a large mammalian prey animal, constructing shelter, or mastering the fabrication of flint tools required grueling, uninterrupted investments of energetic capital. If an ancestral human dropped their hunting spear and walked away the moment fatigue set in or the moment an animal broke into an unexpected sprint, they would quickly perish from starvation. In the ancestral ecological ledger, abandoning an endeavor midway was almost always fatal. A hardwired bias toward behavioral escalation—a refusal to quit once energy was expended—served as a vital evolutionary mechanism to bridge the gap between intermediate exhaustion and long-term survival payoff.

Furthermore, in tight-knit ancestral bands, behavioral consistency and dependability were paramount components of social capital. An individual who frequently abandoned cooperative hunting ventures, switched allegiances midstream, or exhibited erratic, fluctuating commitments was viewed as fundamentally untrustworthy. Tenacity, even when bordering on stubborn irrationality, functioned as a powerful, honest signal of social reliability and character fortitude. The modern sunk cost fallacy represents a classic evolutionary mismatch: an ancient, hardwired psychological heuristic engineered to maximize caloric extraction and social cohesion in the wild is tragically misapplied to complex, counter-intuitive abstract financial instruments, public infrastructure projects, and corporate balance sheets.

9. Mathematical Synthesis: Prospect Theory and Sunk Cost Dynamics

9.1 Formalizing the Reference Point Shift in Sunk Costs

To establish a rigorous mathematical synthesis of how Amos Tversky’s prospect theory drives the empirical findings of Hal Arkes and Catherine Blumer, one must formally construct the decision utility functions governing an agent confronted with an unrecoverable past expenditure. Let an economic agent possess an initial wealth level $W$. Suppose the agent is engaged in a project wherein an unrecoverable capital expenditure of magnitude $S$ ($S > 0$) has already been irrevocably committed. The agent faces a binary strategic choice: either terminate the venture immediately or commit an additional prospective marginal expenditure $C$ ($C > 0$) to pursue a probabilistic recovery.

Normative expected utility theory evaluates the choice via a linear or concave utility function over absolute terminal wealth:

$$\text{Abandonment Utility:} \quad EU_{\text{abandon}} = U(W – S)$$

$$\text{Continuation Utility:} \quad EU_{\text{continue}} = p \cdot U(W – S – C + R) + (1 – p) \cdot U(W – S – C)$$

Where $p$ represents the subjective probability that the additional expenditure $C$ will yield a gross return $R$, and $(1-p)$ represents the probability of complete failure. Under standard marginalism, because $S$ is invariant across both states, the decision hinges solely on whether $p \cdot R > C$, assuming risk neutrality at the margin.

In stark contrast, prospect theory dictates that utility is not evaluated over total terminal wealth, but over deviations from a subjective reference point $r$. The central behavioral pathology of the sunk cost effect is that the unrecoverable expenditure $S$ anchors the reference point firmly to the pre-investment baseline ($r = W$), rather than adjusting down to the current real asset position ($r = W – S$). Consequently, both abandonment and continuation are evaluated strictly within the negative, loss-averse domain of Tversky’s value function $v(x)$.

Recall Tversky and Kahneman’s parameterization of the value function:

$$v(x) = \begin{\cases} x^\alpha & \text{for } x ge 0 \ -\lambda (-x)^\beta & \text{for } x < 0 \end{\cases}$$

Where $alpha, beta < 1$ capture diminishing marginal sensitivity, and $lambda > 1$ represents the coefficient of loss aversion (empirically estimated at $\lambda \approx 2.25$, with $\alpha = \beta \approx 0.88$). Under this non-linear framework, the subjective decision values diverge radically:

$$\text{Subjective Value of Abandonment:} \quad V(\text{Abandon}) = v(-S) = -\lambda (S)^\beta$$

$$\text{Subjective Value of Continuation:} \quad V(\text{Continue}) = w(p) \cdot v(R – S – C) + w(1 – p) \cdot v(-S – C)$$

Where $w(p)$ represents the non-linear probability weighting function. Because the value function is strictly convex in the domain of losses ($v”(x) > 0$ for $x < 0$), the marginal psychological degradation of compounding a loss from$-S$ to $-(S + C)$ is fundamentally smaller than the initial drop from $0$ to $-S$:

$$|v(-S – C) – v(-S)| ll |v(-C) – v(0)|$$

Algebraically, this structural convexity shifts the agent into a state of extreme risk-seeking. The decision-maker is willing to accept an economically catastrophic gamble—committing fresh capital $C$ to an endeavor where the expected return is deeply negative—because doing so introduces a non-zero probability of breaking even ($R – S – C \approx 0$), thereby entirely wiping out the excruciating loss penalty $-\lambda(S)^\beta$. The sunk expenditure $S$ acts as a massive mathematical wedge, mechanically forcing the agent into irrational, risk-seeking escalation.

9.2 Mental Ledger Amortization Models

To capture the temporal dynamics and real-world behavior identified in Arkes and Blumer’s theater experiment, the static prospect theory value function must be integrated with Richard Thaler’s dynamic mental ledger amortization model. When a transaction is initiated, a dedicated mental account $M_k$ is opened. The unamortized psychological balance of this account at time $t$, designated as $B(t)$, is governed by the initial investment $S$ and an exponential cognitive decay parameter $\delta$:

$$B(t) = S \cdot e^{-\delta t}$$

Where $\delta > 0$ represents the rate of psychological depreciation or memory attenuation over elapsed time $t$. This parameter captures the empirical reality documented by Arkes and Blumer: as time elapses, the cognitive salience of the historical cash outlay inexorably degrades.

The total transaction utility derived from consuming the asset at time $t$ versus allowing it to perish unconsumed can be formalized as:

$$\text{Net Decision Utility}(t) = U_{\text{acquisition}}(t) + \theta \cdot B(t) – C_{\text{effort}}$$

Where $U_{\text{acquisition}}(t)$ represents the genuine experiential pleasure derived from attending the event or using the asset, $C_{\text{effort}}$ represents the transaction friction (e.g., driving through snow to reach the theater), and $\theta$ is an institutional weighting parameter representing the agent’s internal sensitivity to waste. When $t$ is small (early in the theater season):

$$B(t) \approx S implies \theta \cdot B(t) gg C_{\text{effort}}$$

For full-price purchasers ($S =$15$), the value of$B(t)$ is substantial, overpowering minor frictions ($C_{\text{effort}}$) and forcing attendance even if $U_{\text{acquisition}}$ is low. For heavily discounted purchasers ($S =$8$),$B(t)$ is drastically smaller; hence, any increase in effort or marginal drop in acquisition utility leads to immediate, rational ticket abandonment. However, as $t to \infty$ (the second half of the season):

$$\lim_{t to \infty} B(t) = 0 implies \text{Net Decision Utility}(t) to U_{\text{acquisition}}(t) – C_{\text{effort}}$$

As the mental balance decays to zero, the sunk cost effect structurally evaporates. Patrons in all pricing tiers converge toward purely forward-looking marginal decision-making. The combination of prospect theory loss aversion and dynamic mental ledger decay provides a comprehensive mathematical model capable of predicting both the initial escalation of commitment and its eventual temporal dissolution.

9.3 Cumulative Prospect Theory and Probability Weighting

The final layer of mathematical formalization required to understand industrial and public megaproject escalation involves the integration of Cumulative Prospect Theory (CPT), developed by Tversky and Kahneman in 1992. CPT introduced a sophisticated, rank-dependent probability weighting function $w(p)$ that operates over the probability distribution of potential outcomes:

$$w(p) = \frac{p^\gamma}{(p^\gamma + (1 – p)^\gamma)^{1/\gamma}}$$

Where $\gamma \approx 0.61 \text{ to } 0.69$. The fundamental behavioral property of this inverse S-shaped weighting function is that decision-makers systematically overweight low-probability extreme events ($w(p) > p$ for small $p$) and underweight moderate-to-high probabilities ($w(p) < p$ for intermediate and large $p$).

In failing capital investments, industrial R&D ventures, or military conflicts, the probability of a miraculous technological breakthrough, a sudden market turnaround, or an unexpected military victory is typically minuscule ($p < 0.05$). Under standard expected utility theory, a rational executive would multiply this vanishing probability by the potential payoff, immediately recognizing that the expected return does not justify the prospective marginal burn rate.

However, within the architecture of Cumulative Prospect Theory, two catastrophic cognitive biases synergize. First, the historical sunk cost has shoved the decision-maker deep into the risk-seeking, convex domain of the value function ($v”(x) > 0$). Second, the probability weighting function aggressively overweights the tiny probability of miraculous recovery ($w(0.02) gg 0.02$). The corporate executive or state official does not perceive a 2% chance of project vindication; their cognitive architecture inflates this slim possibility into an actionable, compelling lifeline. The compounding interaction between extreme loss aversion and optimistic probability distortion creates an inescapable psychological gravity well, transforming failing initiatives into catastrophic bottomless pits of squandered capital.

10. Systemic Ramifications: Public Megaprojects and Enterprise Decisions

10.1 Megaproject Pathology and Flyvbjerg’s Infrastructure Analyses

While laboratory vignettes and university theater tickets illustrate the foundational cognitive mechanics of the sunk cost fallacy, its most destructive real-world manifestations occur within monumental public infrastructure and defense procurement. In his comprehensive global analyses of mega-infrastructure, Oxford scholar Bent Flyvbjerg revealed an astonishing empirical reality: nine out of ten megaprojects experience massive cost overruns, with overruns of 50% to 100% being standard, and overruns surpassing 500% occurring with shocking frequency across transportation networks, high-speed rail, nuclear power facilities, and hydroelectric dams.

Flyvbjerg demonstrated that megaprojects are governed by a toxic institutional pathology characterized by strategic misrepresentation, optimistic forecasting, and, above all, the deliberate weaponization of the sunk cost fallacy. Project promoters intentionally construct initial budget estimates that are unrealistically low to secure initial legislative approval and public bond financing. Once the construction phase begins and billions of dollars of public capital are irreversibly poured into the earth—tunnels excavated, foundations laid, and specialized contracts signed—the project crosses the proverbial point of no return.

When unexpected geological hazards, engineering errors, or economic realities inevitably trigger multi-billion-dollar cost escalations, political leaders do not conduct rigorous, prospective cost-benefit reviews. Instead, they execute the classic sunk cost playbook: they approach parliaments, congresses, and the taxpaying public, declaring that the hundreds of millions or billions of dollars already spent will be “wasted” if the legislature refuses to approve supplementary emergency appropriations. The monumental historical expenditure is leveraged as an unassailable political weapon to force ongoing funding. Abandoning an incomplete tunnel or a half-built nuclear reactor is politically unviable; it stands as a colossal, visible monument to governmental incompetence. Consequently, democratic states routinely escalate commitment, transforming billion-dollar miscalculations into catastrophic generational financial liabilities.

10.2 Venture Capital and Startup Pivot Reluctance

In the high-stakes, fast-moving ecosystem of venture capital and technology startups, one might assume that the ruthless discipline of market forces and the expertise of sophisticated investors would insulate the sector from sunk cost pathologies. The empirical reality is precisely the opposite: the venture capital landscape is riddled with sunk cost escalation, frequently operating on a staggering scale.

For early-stage technology founders, the sunk cost trap manifests as profound resistance to the strategic pivot. Founders invest years of grueling, eighty-hour workweeks, sacrifice personal relationships, and pour their personal life savings into building a specific proprietary software architecture or hardware prototype. When early customer acquisition metrics decisively demonstrate that the market does not want the product, founders routinely refuse to pivot or liquidate. To abandon the core product is to acknowledge that years of sweat equity, intellectual capital, and personal identity were spent in vain. Founders become pathologically committed to their original architecture, burning through their remaining cash runway in a desperate bid to force product-market fit where none exists.

This dynamic is frequently compounded by institutional venture capital complicity. When a venture fund leads a Series A and Series B round into a startup that subsequently begins to falter, the general partners face profound cognitive and reputational conflicts. Writing off a $50 million investment damages the fund’s internal rate of return (IRR), degrades their institutional reputation with limited partners (LPs), and threatens future fund-raising efforts. Consequently, lead investors routinely construct insider-led “bridge rounds” or follow-on financings, funneling good capital after bad into fundamentally insolvent business models. The catastrophic collapses of heavily capitalized unicorns—such as Theranos, WeWork, and Quibi—stand as modern monuments to corporate escalation, where hundreds of millions of dollars were continually incinerated simply because no one had the courage to close the open mental ledger.

10.3 Software Engineering and Technical Debt

Within the domain of software engineering and digital enterprise systems, the sunk cost fallacy operates through the insidious accumulation and preservation of technical debt. In enterprise technology architecture, organizations routinely find themselves trapped within decaying, highly vulnerable legacy software codebases, antiquated mainframe systems, or obsolete custom frameworks that cost fortunes to maintain and pose severe operational hazards.

The rational engineering imperative is straightforward: when the prospective cost of maintaining, patching, and securing an antiquated software stack exceeds the long-term cost of migrating to a modern, decoupled, scalable architecture, the legacy codebase should be systematically refactored or completely deprecated. Yet, enterprise chief information officers (CIOs) and senior engineering managers consistently resist complete architectural migrations. They point to the millions of dollars and hundreds of thousands of engineering man-hours that have been sunk into designing, testing, and customizing the legacy system over the preceding decades.

Software developers themselves develop intense, personal emotional attachments to the code they have authored. To rewrite a sprawling codebase from scratch feels like a traumatic, wholesale erasure of the intellectual labor, creative problem-solving, and professional identity invested into the system. Engineering teams convince themselves that writing yet another complex abstraction layer, adding another kludge, or paying millions for proprietary legacy support contracts is a sensible way to “protect” the historical investment. This organizational inertia creates enterprise software systems that are brittle, unscalable, and extraordinarily fragile, until a catastrophic systemic outage or massive cybersecurity breach finally forces an unavoidable reckoning.

11. Debiasing Strategies and Institutional Interventions

11.1 Structural Separation of Evaluation and Execution

Because the psychological drivers of the sunk cost fallacy—cognitive dissonance, self-justification, and loss aversion—are deeply hardwired into individual human cognition, attempting to eliminate the bias through pure willpower or simple educational lectures is notoriously ineffective. True debiasing requires the construction of robust, institutional governance architectures that structurally divorce project evaluation from project execution.

The most powerful organizational intervention is the mandated operational separation between Project Initiators and Project Continuators:

• The Agency Flaw: When the exact same manager, team, or executive who conceived, championed, and launched an initiative is granted total budgetary authority over whether to continue, expand, or terminate that project, rational prospective marginalism is impossible. The decision-maker’s personal ego, professional identity, and career survival are inextricably bound to the sunk capital.
• The Structural Solution: Forward-thinking enterprises, advanced military structures, and elite venture capital firms implement independent capital allocation and oversight committees. Once a project reaches a pre-defined evaluation horizon or breaches pre-determined risk thresholds, continuation authority is transferred completely to an objective oversight body composed of individuals who hold zero historical, financial, or reputational connection to the project’s inception.

Furthermore, organizations must establish formalized, binding “Kill Criteria” prior to project launch. Before a single dollar of capital is committed or a line of code is written, leadership teams must explicitly define the prospective operational, financial, and competitive metrics that will automatically mandate project termination. By pre-committing to transparent, unalterable kill criteria while operating in a neutral, pre-investment cognitive state, organizations effectively build an institutional Ulysses contract. When the project later encounters distress, the decision to terminate is no longer an agonizing, subjective dilemma subject to cognitive dissonance; it is the automated, objective execution of an established protocol.

11.2 Cognitive Reframing and Explicit Mental Accounting Shifts

On an individual and operational level, cognitive debiasing can be achieved through systematic mental accounting interventions designed to neutralize Tversky’s framing effects. Decision-makers must be explicitly trained to identify and dismantle the artificial topical mental accounts that anchor their judgments to the past.

The most potent cognitive reframing intervention is the “Fresh Start” Heuristic, alternatively designated as the Clean Slate Technique. When an executive or manager is wrestling with whether to continue funding an embattled, underperforming initiative, they must be forced to confront a radical counterfactual thought experiment:

“If you were unexpectedly hired from the outside to take over this company today, with completely fresh eyes and zero personal history, and you discovered this exact project in its current state—with this remaining cash burn, these technical hurdles, and this competitive landscape—would you authorize the expenditure of your finite capital to launch this project today?”

If the honest prospective answer is “No,” then any inclination to continue funding the project is immediately unmasked as an artifact of the sunk cost fallacy. The historical expenditure is stripped of its coercive power, forcing the manager to confront the prospective reality of the choice.

A complementary debiasing method involves the Semantic Reframing of Abandonment. In standard corporate vernacular, terminating a project is shrouded in the vocabulary of failure, defeat, waste, and retreat. Enlightened organizations intentionally alter their linguistic architecture. Terminating a project is semantically reframed as “resource liberation,” “capital optimization,” or “strategic redeployment.” Project teams that recommend the termination of their own initiatives are celebrated as corporate heroes who have saved the organization millions in prospective losses, rather than being marginalized as failures. By transforming project cancellation into a source of pride, institutional prestige, and professional reward, the acute psychological pain of closing the open mental ledger is completely neutralized.

11.3 Incentive Alignment and Corporate Governance Reform

Ultimately, individual and structural debiasing mechanisms will fail if the underlying corporate incentive structures actively reward sunk cost escalation. In far too many modern corporations, compensation packages, executive bonuses, and promotion tracks are directly tethered to the absolute scale of the budget managed, the total number of headcount supervised, or the physical completion of multi-year enterprise roadmaps, regardless of prospective profitability.

Corporate governance reform requires the radical restructuring of executive performance metrics:

• Decoupling Compensation from Scale: Executive compensation must be tied to enterprise-wide return on invested capital (ROIC) and economic value added (EVA), rather than localized project completion milestones.
• Rewarding Intelligent Termination: Boards of directors must establish explicit corporate protocols that reward managers for killing value-destroying projects early. When a division head conducts a rigorous prospective marginal analysis and shuts down a failing $50 million venture, saving the enterprise$30 million in projected follow-on burn, that executive should receive a direct financial bonus and institutional accolades equivalent to someone who generated $30 million in new commercial revenue.
• Institutionalizing Post-Mortem Learning Reviews: Organizations must replace punitive inquisitions with blameless, psychological-safety-oriented post-mortem reviews. When a project is cancelled, the objective of the post-mortem must be the rigorous extraction of empirical data, organizational learning, and intellectual capital. By proving to employees that the intelligence gained from an abandoned project is actively leveraged across other divisions, the organization systematically disarms the toxic “desire not to waste,” proving that historical investments are never truly wasted if they serve to optimize future prospective decisions.

12. Contemporary Replications, Methodological Debates, and Lasting Legacy

12.1 Replication Initiatives in Modern Behavioral Science

In the wake of the modern replication crisis that has swept through the psychological sciences over the past two decades—challenging the reliability of numerous classic social-psychological phenomena—the foundational empirical canon of Hal Arkes and Catherine Blumer has been subjected to rigorous, large-scale contemporary re-examination. Modern behavioral science initiatives, including the Many Labs replication projects and open-science pre-registered replications, have systematically reassessed the Radar-Invisible Bomber, the Ski Trip Dilemma, and related sunk cost paradigms.

The results of these massive, multi-site contemporary replications have delivered a resounding, unambiguous validation of the original 1985 findings. The statistical divergence between control conditions and sunk cost conditions documented by Arkes and Blumer has replicated with exceptional fidelity, routinely generating massive effect sizes ($Cohen’s d > 0.80$) across diverse demographic cohorts. The fundamental human compulsion to escalate commitment in the presence of unrecoverable expenditures stands as one of the most robust, dependable, and reproducible behavioral phenomena in the entire history of cognitive science.

Furthermore, contemporary cross-cultural replication studies have illuminated how cultural dimensions modulate the magnitude of the fallacy. While the sunk cost effect is universal across the human species, researchers have discovered intriguing cross-cultural variance:

• Individualistic Cultures: In individualistic Western cultures (e.g., the United States, the United Kingdom), the sunk cost effect is heavily driven by personal self-justification, cognitive dissonance, and the internal struggle to maintain self-worth.
• Collectivistic Cultures: In collectivistic cultures (e.g., East Asian and Latin American societies), the sunk cost effect frequently manifests with even greater intensity, but its primary psychological driver shifts toward external social accountability, the avoidance of public shame, and the imperative to protect familial, institutional, or corporate honor.

12.2 Critiques, Boundary Conditions, and Alternative Interpretations

Despite its unassailable empirical standing, the sunk cost paradigm has faced sophisticated intellectual critiques and alternative theoretical interpretations from economists, evolutionary theorists, and legal scholars seeking to establish boundary conditions around what constitutes genuine irrationality.

The most prominent economic counter-critique emerges from Real Options Theory. Prominent financial economists argue that in complex, volatile, and highly uncertain real-world environments, what appears to be an irrational escalation of commitment may actually represent a fully rational investment in information discovery. When a firm injects an additional $1 million into a struggling R&D project, it is not simply trying to salvage the historical$9 million; it is purchasing an “information option.” Continuing the project allows the firm to observe emerging market dynamics, learn critical technical lessons, develop internal human capital, and preserve the optionality of a future breakthrough. To label all persistence as a cognitive “fallacy” ignores the profound strategic value embedded within ongoing experiential learning.

A second major theoretical critique comes from reputational signaling and game-theoretic models. In an imperfect world characterized by severe information asymmetry, an executive’s or politician’s greatest asset is their perceived credibility, resolve, and steadfastness. In strategic competitive environments, an agent who demonstrates a ruthless, unyielding willingness to persist in an endeavor—even when facing astronomical costs—sends an intimidating, deterrent signal to competitors, rivals, and adversaries. In the brutal arena of market dominance or geopolitical conflict, cultivating a reputation as an actor who refuses to back down can be locally rational for the individual, even if it generates global economic inefficiency. These critiques do not invalidate the psychological discoveries of Tversky, Arkes, and Blumer; rather, they enrich our understanding, highlighting the complex boundary conditions where psychological bias intersects with strategic rationality.

12.3 The Unified Legacy of Tversky, Arkes, and Blumer

The convergence of Amos Tversky’s theoretical architecture and Hal Arkes and Catherine Blumer’s empirical mastery represents one of the most triumphant, consequential chapters in the development of modern behavioral science. Prior to their groundbreaking contributions, the study of human decision-making was paralyzed by an unbridgeable chasm between the cold, hyper-rational, mathematically sterile dogmatism of neoclassical economics and the fragmented, descriptive observations of clinical and social psychology.

Tversky provided the revolutionary conceptual scaffolding: the S-shaped value function, loss aversion, cognitive framing, and the dynamics of open mental accounts. He exposed the mathematical geometry of the human mind, explaining why humans perceive reality through the subjective lens of gains, losses, and historical reference points. Hal Arkes and Catherine Blumer took this brilliant theoretical blueprint and transformed it into definitive, unassailable empirical reality. Through the Baxter Corporation bomber, the snowy ski resorts of Wisconsin, and the real-world box office of the Ohio University Theater, they proved that economic rationality is routinely subordinated to an internalized, moralized, and desperate “desire not to waste.”

The unified legacy of Tversky, Arkes, and Blumer extends far beyond academic textbooks. It has revolutionized corporate capital budgeting protocols, reshaped modern public infrastructure policy, informed military strategic doctrine, and transformed modern consumer protection law. By courageously exposing the psychological mechanisms that chain human judgment to the ghosts of past expenditures, their work achieved the ultimate objective of behavioral science: it liberated human decision-makers from the tyranny of the immutable past, providing the cognitive tools and institutional frameworks necessary to embrace a truly rational, prospective future.

Conclusion: Toward an Integrated Science of Decision-Making

The intellectual journey from the neoclassical axiom of sunk cost irrelevance to the contemporary behavioral consensus represents a profound maturation in our understanding of human nature. For over a century, orthodox economic theory clung to an idealized vision of humanity—a vision of calculating automatons effortlessly discarding historical commitments, immune to the sting of loss, and navigating state-contingent futures with clinical detachment. The landmark experiments of Amos Tversky, Hal Arkes, and Catherine Blumer permanently shattered this illusion, revealing an empirical reality that is infinitely richer, more tragic, and deeply human.

Their work proved that human decision-makers do not evaluate the world from an unburdened present. We are creatures intrinsically tethered to the trajectories of our historical choices. We build mental ledgers that resist premature closure; we experience the agonizing asymmetry of loss aversion; we confuse the moral imperative of thrift with the cold calculus of forward-looking marginal efficiency; and we wage desperate psychological wars to defend our fragile self-concepts against the painful admission of catastrophic failure. The sunk cost fallacy is not an isolated intellectual glitch or a minor computational error; it is an intrinsic, systemic vulnerability forged by the intersection of our evolutionary heritage, our developmental socialization, and the fundamental architecture of our cognitive processing.

Yet, in diagnosing the deep psychological anatomy of the sunk cost trap, Tversky, Arkes, and Blumer did not consign humanity to inevitable irrationality. Rather, their profound empirical discoveries provided the indispensable blueprint for our intellectual liberation. By identifying the specific mechanisms of reference point shifting, mental ledger accounting, and the “desire not to waste,” they empowered modern institutions to engineer robust debiasing architectures—from the structural separation of capital evaluation to the linguistic reframing of project termination. The ultimate lesson of their unified legacy is as philosophically profound as it is economically transformative: true rationality does not reside in the vain, backward-looking pursuit of vindicating what is already lost, but in cultivating the cognitive humility, moral courage, and institutional wisdom to let the dead past bury its dead, anchoring our choices solely in the open, unwritten possibilities of the future.

References

• Arkes, H. R., & Ayton, P. (1999). The sunk cost and Concorde effects: Are humans less rational than lower animals? Psychological Bulletin, 125(5), 591–600. https://doi.org/10.1037/0033-2909.125.5.591
• Arkes, H. R., & Blumer, C. (1985). The psychology of sunk cost. Organizational Behavior and Human Decision Processes, 35(1), 124–140. https://doi.org/10.1016/0749-5978(85)90049-4
• Arkes, H. R., & Hutzel, L. (2000). The role of prospective utility in the sunk cost effect. Journal of Behavioral Decision Making, 13(3), 293–306. https://doi.org/10.1016/0040-5809(76)90040-X
• Festinger, L. (1957). A theory of cognitive dissonance. Stanford University Press. https://www.sup.org/books/title/?id=3850
• Flyvbjerg, B. (2009). Survival of the unfittest: Why the worst infrastructure gets built—and what we can do about it. Oxford Review of Economic Policy, 25(3), 344–367. https://doi.org/10.1093/oxrep/grp024
• Flyvbjerg, B., Bruzelius, N., & Rothengatter, W. (2003). Megaprojects and risk: An anatomy of ambition. Cambridge University Press. https://doi.org/10.1017/CBO9781107050891
• Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of decision under risk. Econometrica, 47(2), 263–291. https://doi.org/10.2307/1914185
• Simon, H. A. (1955). A behavioral model of rational choice. The Quarterly Journal of Economics, 69(1), 99–118. https://doi.org/10.2307/1884852
• 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. https://doi.org/10.1016/0030-5073(76)90005-2
• Sweis, B. M., Abram, S. V., Schmidt, B. J., Seeland, K. D., MacDonald, A. W., Thomas, M. J., & Redish, A. D. (2018). Sensitivity to “sunk costs” in mice, rats, and humans offers insights into cross-species economic decision-making. Science, 361(6398), 178–181. https://doi.org/10.1126/science.aar6980
• Tavris, C., & Aronson, E. (2007). Mistakes were made (but not by me): Why we justify foolish beliefs, bad decisions, and hurtful acts. Harcourt. https://www.harcourtbooks.com
• Thaler, R. H. (1980). Toward a positive theory of consumer choice. Journal of Economic Behavior & Organization, 1(1), 39–60. https://doi.org/10.1016/0167-2681(80)90051-7
• Thaler, R. H. (1999). Mental accounting matters. Journal of Behavioral Decision Making, 12(3), 183–206. https://doi.org/10.1126/science.7455683
• Tversky, A., & Kahneman, D. (1992). Advances in prospect theory: Cumulative representation of uncertainty. Journal of Risk and Uncertainty, 5(4), 297–323. https://doi.org/10.1007/BF00122574
• Von Neumann, J., & Morgenstern, O. (1944). Theory of games and economic behavior. Princeton University Press. https://press.princeton.edu/books/hardcover/9780691130613/theory-of-games-and-economic-behavior