> ## Documentation Index > Fetch the complete documentation index at: https://meta.niceshare.site/llms.txt > Use this file to discover all available pages before exploring further. # Escalation of Commitment > Escalation of commitment is the tendency to increase investment in a failing course of action because of prior sunk costs and identity stakes—not because evidence improved. **Category**: Effects
**Type**: Behavioral Pattern / Decision Bias
**Origin**: Organizational psychology; named and tested in Staw (1976) and synthesized in later reviews (e.g., Brockner, 1992)
**Also known as**: Commitment escalation; Concorde fallacy (informal) **Quick Answer** — Escalation of commitment is the pattern where decision makers double down on a struggling project or strategy after earlier resources and reputations are on the line. Barry Staw’s foundational experiments showed people can allocate *more* to a failing path when they feel personally responsible for the initial choice. The key insight is to separate “learning” from “saving face,” and to treat prior spending as information, not a mandate to continue. ## What is Escalation of Commitment? Escalation of commitment is a dynamic trap: the more you have invested—money, time, public promises—the harder it becomes to stop, even when new information says the path is unpromising. It is not the same as patient persistence; persistence responds to improving odds, while escalation often responds to *prior* commitment and self-justification. > Past effort is a sunk story; escalation turns it into a future mandate. It overlaps with `sunk-cost-fallacy`, but escalation highlights **sequences of decisions** in organizations and teams, where incentives, accountability, and identity amplify the pull to “stay the course.” In groups, `groupthink` can suppress exit talk, while `confirmation-bias` helps leaders notice data that flatters the original plan. ### Escalation of Commitment in 3 Depths * **Beginner**: If the main reason to continue is “we already spent so much,” treat that sentence as a red flag. * **Practitioner**: Run periodic **decision resets** that ask what you would fund today if none of the past spending had happened. * **Advanced**: Design governance so project owners are not the sole judges of their own prior promises—separate advocacy from adjudication. ## Origin **Barry M. Staw** formalized escalation in management research with experimental paradigms showing increased commitment to a previously chosen course after negative feedback, especially under self-justifying accountability (Staw, 1976). Later work expanded the construct across R\&D, public investment, and military planning contexts. **Joel Brockner** and others synthesized mechanisms—self-justification, prospect-theoretic framing, and organizational pressures—in broad reviews that treat escalation as a predictable failure mode rather than a rare mistake (Brockner, 1992). The informal label **Concorde fallacy** captures the same structure in public megaprojects where political credibility attaches to earlier announcements. ## Key Points Treat “honoring the plan” as a hypothesis to test, not a moral duty to prior budgets. Money already spent should not mechanically justify new money, yet psychologically it often does because stopping feels like admitting error. Leaders escalate partly to protect reputation with boards, voters, or teams who watched them champion the initiative. Narrow success metrics and milestone rituals can hide negative evidence until overruns are large. Independent stage-gates, challenger roles, and pre-set kill criteria reduce the social cost of stopping. ## Applications Use these moves where budgets are large and exit is socially costly. Pair every major initiative with a written “kill question” answered by someone who did not author the roadmap. Separate “return on *incremental* dollar” from pride in the original thesis; track opportunity cost explicitly. Publish ranges and off-ramps early; make continued funding contingent on independent reviews, not prior announcements alone. Reframe quitting a bad master’s plan or job search tactic as data efficiency, not character failure—journal the decision rule in advance. ## Case Study In Staw’s classic administrative simulation, participants who were made to feel personally responsible for an initial product decision later allocated **more** not less—additional development resources to that same product line after performance turned negative, compared with conditions where responsibility was diffuse. The measurable pattern is the *direction* of continued funding after bad news, not mere random noise. The lesson maps to real organizations: the combination of ownership and public justification predictably inflates follow-on investment. ## Boundaries and Failure Modes Escalation is not the same as legitimate learning-by-doing. **Boundary 1: True option value**\ Some projects look troubled early yet contain real options; the defense is evidence of convexity, not sunk-cost rhetoric. **Boundary 2: Coordination costs of stopping**\ Sometimes continuing is cheaper than unwinding contracts; that is an economic comparison, not automatic escalation. **Common misuse**: Labeling any long project “escalation” to force a stop—without distinguishing evidence from identity—destroys trust and hides real breakthrough paths. ## Common Misconceptions Stopping early is not the same as being “weak.” **Reality**: Normatively they should not matter; psychologically they routinely do—especially with audiences watching. **Reality**: Extra metrics can be captured to support the narrative unless review power is independent. **Reality**: Experts escalate in domains where their identity is tied to the original forecast—expertise does not immunize. ## Related Concepts Combine these lenses when designing exits and reviews. Why past spending pulls emotional weight even when it should not affect forward-looking choices. How labor and ownership inflate perceived value—fuel for escalation in things you built yourself. Why ending a failing bet can feel like locking in a loss, even when cutting losses is rational. ## One-Line Takeaway Decide the next dollar as if the previous dollars were never yours—then pay the social cost of stopping with governance, not guilt.