Economists hold that a rational decision-maker, choosing whether to fund a project's next phase, should weigh only the prospective costs and benefits of that phase; resources already spent are sunk and irrecoverable, and should not bear on the choice. Yet managers routinely commit additional funds to ventures that, judged prospectively, no longer merit them. This pattern, termed escalation of commitment, is not adequately explained by mere optimism, since it persists even when managers privately acknowledge the venture is unlikely to succeed. Researchers locate the driver in self-justification. A manager who authorized the original investment has a personal stake in vindicating that decision; abandoning the project amounts to a public admission that the initial commitment was an error. Continued funding defers that admission and preserves the possibility, however slim, that the venture will yet succeed and retroactively validate the original choice. Crucially, this motive operates on the decision-maker who made the prior commitment, not on the resources themselves. A testable implication follows. If escalation springs from self-justification rather than from the sunk cost as such, then a manager who inherits a troubled project from a predecessor, and who therefore bears no personal responsibility for the original decision, should prove markedly more willing to terminate it, even though the sunk cost confronting the two managers is identical. Studies in which responsibility for the initial decision is experimentally separated from the choice to continue have borne this out.