Escalation of commitment is a pattern of decision-making in which a person, group, or organization continues or increases its investment in a course of action despite evidence that it is performing poorly. The investment may involve money, time, effort, reputation, or other resources. Studied in social psychology, organizational research, and economics, the concept concerns how earlier commitments can shape subsequent choices independently of the prospective merits of continuing. It is often associated with “throwing good money after bad,” although continued investment after a setback is not necessarily irrational.
Definition and scope
A typical escalation situation has three elements:
- An initial decision commits resources to a particular undertaking.
- Subsequent information indicates disappointing results or threatens the undertaking’s prospects.
- A decision-maker has an opportunity to withdraw, change direction, or commit further resources.
Escalation occurs when commitment persists or increases in the face of this negative feedback. Researchers differ in how strictly they define the phenomenon: some use the term descriptively for continued commitment, while others reserve it for persistence that cannot be justified by an impartial assessment of future costs and benefits.
The concept is closely related to the sunk cost fallacy, but the two are not identical. A sunk cost is an expenditure that cannot be recovered. The sunk cost fallacy occurs when such expenditure improperly influences a present decision. Escalation of commitment describes a broader process unfolding over time, potentially involving repeated decisions, institutional pressures, personal responsibility, and reputational concerns. Sunk costs can contribute to escalation, but they are not its only cause.
Escalation also differs from ordinary perseverance. Continuing a difficult project can be reasonable when the expected future benefits exceed the remaining costs, when new information supports recovery, or when withdrawal would itself impose substantial costs. Poor past performance alone does not establish that further investment is mistaken.
Development of the concept
The modern research literature is strongly associated with organizational psychologist Barry M. Staw’s 1976 article, “Knee-deep in the Big Muddy: A Study of Escalating Commitment to a Chosen Course of Action.” In an experimental investment task, participants who were made personally responsible for an earlier allocation committed more additional resources following negative feedback than participants who did not bear responsibility for that initial decision.
This work connected investment decisions with earlier research on cognitive dissonance and self-justification. An unfavorable outcome can threaten a decision-maker’s belief that the original choice was competent or reasonable. Renewed commitment can help defend that belief by preserving the possibility of eventual vindication.
Later research broadened the explanation beyond individual defensiveness. Studies examined project characteristics, group interactions, organizational incentives, and external pressures. Escalation consequently came to be understood as a multilevel process rather than simply an individual cognitive bias.
The related expression Concorde effect refers to continued support for an undertaking partly because of resources already invested. It is commonly associated with sunk-cost reasoning and overlaps with escalation of commitment, but it does not capture every mechanism studied under the broader concept.
Psychological mechanisms
Self-justification and personal responsibility
People who selected or publicly endorsed a course of action may interpret its failure as evidence against their judgment. Continuing offers a chance to demonstrate that the original decision was sound, whereas withdrawal may appear to acknowledge an error.
This process can involve private self-image as well as public reputation. A decision-maker may defend a choice even when no audience is present. Conversely, reputational pressures can encourage continuation even when the person privately doubts the project.
Personal responsibility is an important influence in many studies, but it is neither necessary nor sufficient. People can inherit troubled projects and still escalate their commitment, while those responsible for an initial choice sometimes withdraw promptly.
Selective interpretation of evidence
Confirmation bias can lead decision-makers to emphasize signs of improvement and discount indications of failure. Motivated reasoning can shape how they explain setbacks: unfavorable results may be attributed to temporary conditions, inadequate funding, or poor implementation rather than weaknesses in the underlying plan.
Negative feedback is therefore not always experienced as a reason to stop. It may instead be interpreted as evidence that the undertaking needs more resources.
Loss framing and risk-taking
Prospect theory provides another account. People evaluate outcomes relative to reference points, and some choices framed as losses can encourage greater willingness to take risks. Ending an unsuccessful project may make a loss definite, while continuation preserves an uncertain possibility of recovery.
Loss aversion and risk-taking in the domain of losses are related but distinct features of this account. Neither implies that people invariably choose risky continuation. Behavior depends on the available options, their probabilities, and how the decision is framed.
The framing effect is relevant because an identical decision can be presented as either accepting a loss or reallocating resources toward a better future use.
Proximity to completion
Commitment can intensify when a project appears close to completion. Decision-makers may feel that stopping would waste an almost-finished achievement, especially when substantial effort has already been expended.
However, proximity to completion can also be economically relevant. A nearly completed asset may require little additional expenditure before generating benefits. The analytical question is whether continuation is supported by remaining costs and likely benefits, rather than merely by a desire not to abandon past effort.
Organizational and social influences
Escalation is not adequately explained by individual psychology alone. Organizational arrangements can make continuation easier, safer, or more rewarding than withdrawal.
Relevant influences include:
- Incentives and careers: managers may benefit if recovery occurs while bearing only part of the cost of additional failure.
- Public commitments: visible sponsorship can make reversal appear inconsistent or damaging to authority.
- Organizational identity: a project may become associated with the organization’s mission, prestige, or legitimacy.
- Resource dependencies: staff, suppliers, and other participants may rely on continuation and advocate further support.
- Fragmented decisions: a series of small extensions may avoid the scrutiny that a single large commitment would attract.
- Information structures: unfavorable evidence may be delayed, filtered, or reported by people whose interests favor continuation.
These factors connect escalation with corporate governance and organizational accountability. They also distinguish what is rational for a particular manager from what is beneficial for the organization. A manager may have personal incentives to continue a project even when the organization would gain from ending it.
Groups do not invariably correct individual escalation. Discussion can expose weak assumptions, but it can also reinforce shared justifications and commitments.
Economic interpretation
A forward-looking evaluation compares the consequences of the options available now. Irrecoverable past expenditures should not, by themselves, determine which option is preferable. Relevant considerations include future revenues, additional expenditure, uncertainty, withdrawal costs, and the opportunity cost of resources that could be used elsewhere.
For example, suppose a firm has already spent 10 million units on a project. Completion requires another 2 million, and the completed project is expected to generate 4 million in net receipts before accounting for that additional expenditure. Continuing may be justified despite the project’s overall loss. The earlier 10 million cannot be recovered by cancellation.
If completion instead requires another 5 million for the same expected receipts, the past expenditure does not itself justify proceeding. A claim that the firm has “spent too much to stop” substitutes historical investment for a comparison of future alternatives.
This illustration abstracts from uncertainty, financing constraints, contractual obligations, and strategic effects. In practice, these can matter greatly. A decision may also have learning value: limited continuation can generate information that improves later choices. Such benefits must be assessed rather than assumed.
Research methods and limitations
Researchers investigate escalation through controlled experiments, simulations, surveys, and organizational case studies. Experiments can vary responsibility for the original decision, the size of prior investment, the framing of feedback, and the availability of alternatives. Outcomes often include additional resources allocated or willingness to continue.
Field studies examine how actual commitments evolve. They can reveal institutional pressures that simplified experiments omit, but identifying the cause of continuation is more difficult. A project may persist because of self-justification, contractual constraints, new information, or several influences acting together.
An important limitation is causal interpretation. Observing that a failing project received additional funding does not establish that escalation bias caused the decision. Likewise, a project’s eventual failure does not prove that continuing was unreasonable given the information available at the time.
Studies also differ in what they measure. The amount of prior investment, responsibility for an earlier choice, and the number of repeated commitment decisions are related but distinct variables. Findings about one should not automatically be treated as findings about all forms of escalation.
De-escalation and decision design
Research on de-escalation examines conditions under which decision-makers reduce or end an established commitment. Proposed and studied approaches include independent review, explicit performance milestones, consideration of alternative uses of resources, and separating project evaluation from personal defense of the original decision.
These mechanisms have limits. A new evaluator may lack essential knowledge; rigid stopping rules may terminate promising projects during temporary setbacks; and accountability can increase defensiveness if it primarily demands justification of past choices. Accountability focused on the quality of the current decision may operate differently from accountability focused on defending the original commitment.
De-escalation is therefore not equivalent to immediate withdrawal. It can involve narrowing a project, changing its objectives, replacing its implementation strategy, or making further investment conditional on new evidence. Its defining analytical task is to distinguish the future value of continuation from the psychological and institutional weight of what has already been committed.