Sunk Cost Escalation: The Point Where We Stop Updating

## The Setup

The sunk cost fallacy in its basic form is well understood: past investments that can't be recovered shouldn't influence decisions about the future. Rationally, only expected future value matters. In practice, what we've already spent, built, or committed shapes what we do next in ways that consistently diverge from what the expected-value calculation would recommend.

What's less discussed is *escalation of commitment* — the pattern where the more we've invested, the more we continue investing, even as evidence accumulates that the outcome is deteriorating. Sunk cost fallacy describes a single irrational weighting. Escalation describes a dynamic process where that weighting gets stronger over time.

## What the Research Actually Found

Barry Staw's 1976 paper "Knee-deep in the Big Muddy" introduced the concept using a business investment scenario. Participants who had made the initial investment decision themselves allocated significantly more resources to a failing project than participants who inherited the situation from someone else. The personal responsibility for the initial decision amplified the reluctance to cut losses.

Subsequent research showed the effect isn't purely irrational. Brockner's 1992 review noted that some degree of commitment-increasing behavior under uncertainty is functionally adaptive — persistence sometimes does lead to eventual success, and abandoning projects too quickly is also a real failure mode. The research doesn't say "always cut your losses immediately." It says the persistence/abandon tradeoff gets systematically skewed by prior investment.

The important variable in escalation isn't just how much has been invested, but **who made the original decision and how that decision was publicly framed**. When people have staked reputational position on an outcome, the cost of reversing isn't just the sunk investment — it's the admission of error. Arkes and Blumer's 1985 theater ticket study showed that people who paid more for tickets attended more performances in bad weather, demonstrating the effect even in low-stakes consumer contexts. But the organizational studies show the effect scales dramatically with seniority and public commitment.

## The Mechanism

Escalation is sunk cost fallacy plus overconfidence plus loss aversion running in sequence. Here's the cycle:

1. An investment is made and publicly committed to. A reference point is set.
2. Evidence begins suggesting the investment won't pay off. But this evidence gets processed through confirmation bias — people look for signs of eventual success and interpret ambiguous signals as consistent with the original thesis.
3. The option to exit is evaluated against the sunk cost. Loss aversion weights the "loss" of cutting off as psychologically larger than the gain from preserving capital.
4. The escalation decision gets made — more resources, more commitment. This creates additional sunk costs and deepens the public commitment.
5. The new investment becomes its own reference point. Repeat.

What Staw and later Ross (1987) found is that the escalation dynamic is most powerful when **exit is framed as a reversal of a prior decision**. When teams can reframe continuation or exit as a fresh decision about future expected value, rather than as a verdict on the past decision, escalation rates drop significantly. The framing does work.

## What This Means in Practice

Escalation of commitment shows up predictably in technology project management (the classic enterprise software implementation that's two years late, three times over budget, and still gets more funding), in venture portfolios (the founder who raises a down round rather than shutting down), and in interpersonal decisions (staying in situations that clearly aren't working because of what's already been invested).

The structural fix that organizations have found most useful isn't "remind people about sunk costs" — awareness of the fallacy doesn't reliably reduce it. What helps:

**Pre-mortem analysis**: before committing, explicitly imagine the project has failed and work backward to what caused it. This surfaces assumptions and creates early warning signs rather than framing exit as admission of failure.

**Third-party review at decision points**: the research is clear that decision-makers who didn't make the original investment are better at evaluating whether to continue. Building in external review at milestone checkpoints — not as oversight but as structure — changes the decision framing.

**Explicit off-ramp criteria defined in advance**: when conditions that would trigger a stop decision are defined before investment rather than during, escalation rates drop. The criteria were created by the same person who made the original commitment, which removes the reputational-reversal framing from the exit decision.

## The Takeaway

Escalation of commitment is sunk cost fallacy with momentum. Each investment becomes a justification for the next, and the deeper the investment, the more the option to exit gets weighted against the accumulated prior commitment rather than against future expected value. The exit is framed as an admission rather than a decision — and that framing, more than the investment level itself, is what drives the pattern.

Sunk Cost Escalation: The Point Where We Stop Updating

// COMMENTS

ON THIS PAGE