Beyond Rational Choice: What Behavioral Economics Missed About Escalation

The field of behavioral economics has spent three decades documenting how people deviate from rational choice theory—how we anchor on irrelevant numbers, fall prey to loss aversion, and make decisions that contradict our own stated preferences. Yet in focusing so intently on individual irrationality, behavioral economists have largely overlooked a phenomenon that sits at the intersection of psychology and systems: escalation of commitment, and more specifically, why people continue investing in failing courses of action even when the rational calculus screams otherwise.

The standard behavioral explanation treats escalation as a cognitive bias—sunk cost fallacy, status quo bias, or ego protection. A person throws good money after bad because they cannot psychologically afford to admit failure. This framing is not wrong, exactly. It is just incomplete. It mistakes the symptom for the disease.

What behavioral economics has missed is that escalation is not primarily about individual psychology at all. It is about the structure of choice itself.

Consider the executive who commits $10 million to a product launch. Six months in, market signals suggest the product will fail. The rational response is to cut losses. But the executive faces a choice architecture that makes this nearly impossible. Admitting failure means confronting the board, explaining the sunk costs, and accepting a reputational hit. More subtly, it means acknowledging that the original decision—made with the same information-processing capacity and judgment that got them promoted—was wrong. The escalation that follows is not irrational in the psychological sense. It is rational within a constrained choice environment where the alternatives to escalation carry costs that extend far beyond the immediate financial decision.

The behavioral literature treats this as a bias to be corrected through better framing or decision procedures. But this misses the deeper insight: escalation persists because the choice architecture itself creates incentives for it. When organizations structure decisions such that admitting error carries organizational or reputational penalties, escalation becomes a rational response to that structure, not a deviation from rationality.

This distinction matters because it changes what we should actually do about escalation. If the problem is cognitive bias, the solution is debiasing—better information presentation, pre-commitment devices, or training. If the problem is choice architecture, the solution is structural: redesigning how decisions are made, how failure is communicated, and how organizations distribute the cost of being wrong.

The most effective organizations do not solve escalation by making people think more clearly. They solve it by making it easier to change course than to persist. They separate the person who made the original decision from the person evaluating whether to continue it. They build in explicit decision gates where the default is to stop, not continue. They create cultures where changing direction is treated as evidence of good judgment, not weakness.

Behavioral economics has been invaluable in showing us that human decision-making is not a simple optimization problem. But it has often stopped there, treating deviations from rationality as individual failures rather than systemic features. Escalation reveals the limits of this approach. The person escalating commitment is not necessarily making a cognitive error. They are navigating a choice environment that was never designed to make the rational choice easy.

The real insight is not that people are irrational. It is that rationality is always embedded in context. Change the context—the incentives, the information flow, the social costs of admitting error—and behavior changes without requiring any shift in individual psychology at all. This is not a minor correction to behavioral economics. It is a fundamental reorientation toward understanding that choice is not something individuals make in isolation. It is something systems produce.