Escalation of Commitment

What it is

The rational response to bad news about an investment is to reassess it. Money already spent is gone whatever happens next, so the only sensible question is whether the next pound is well spent, judged from here. What Staw found is close to the opposite. Told that a course of action was going badly, people put more into it, and the ones who put in the most were the ones who had chosen that course themselves.

The study is "Knee-Deep in the Big Muddy: A Study of Escalating Commitment to a Chosen Course of Action", published in Organizational Behavior and Human Performance in 1976. Two hundred and forty business students played a financial vice-president allocating research and development funds between divisions of a company. Some chose which division to fund; others were told the choice had already been made. Then they received results. Those who had personally chosen a division that subsequently performed badly allocated the most additional money to it in the second round.

That design is what makes the paper matter, because it isolates the variable. If escalation were simply about sunk resources, the participants handed someone else's failing division would have escalated just as hard. They did not. Personal responsibility for the original decision is the amplifier, and once you can see that, the phenomenon changes shape. The thing being defended is not the money. It is the judgment of the person who committed it, which is why escalation is worst in exactly the situations where it is most expensive: senior people, public decisions, projects with a name attached.

Two further features follow from this. Escalation is asymmetric with respect to feedback. Good news confirms the decision and requires nothing; bad news threatens it and demands a response, and increasing the commitment is one available response that does not involve conceding anything. And it compounds, because each additional commitment is itself a decision the person is now responsible for, which makes the next piece of bad news slightly harder to accept than the last.

The everyday name for the neighbouring idea is the sunk cost fallacy, and the two are related without being identical. The sunk cost fallacy is the reasoning error: treating unrecoverable past spending as a reason to continue. Escalation of commitment is the behaviour, and Staw's contribution is showing that the behaviour is driven by self-justification rather than by arithmetic. A person can understand the sunk cost argument perfectly and still escalate, because the argument addresses the money and the motive is about the decision.

Loss aversion supplies part of the pull. Once a commitment is made, the reference point moves to include the thing committed to, so abandoning it registers as a loss rather than as a decision not to continue. That connection is well described and it should be handled with the caveat the loss-aversion literature now carries: the asymmetry is real and context-dependent, and the familiar two-to-one coefficient does not survive as a universal law.

A note on the evidence. This page is tagged primary-cited, which in our vocabulary means the primary citation is present and correct and nobody here has assessed how well the wider literature has held up. Staw's 1976 paper has been read in full and the description above comes from it. The escalation literature that grew out of it, spanning decades of organisational research, has not been audited by us, so treat the 1976 study as a solid, well-attributed demonstration rather than as a summary of what fifty years of research has concluded.