The Sunk Cost Fallacy's Dangerous Cousin: Escalation of Commitment in High-Stakes Decisions

Most people have heard of the sunk cost fallacy. You've paid for a concert ticket, the night arrives, you feel terrible, and you drag yourself out anyway because you already spent the money. The logic is broken, everyone agrees, move on.

Photo by Pok Rie on Pexels.

But there's a harder version of this problem, one that brings down product lines, military campaigns, and entire companies. It's called escalation of commitment, and it operates at a scale where the irrationality becomes nearly invisible.

Here's the basic pattern: a decision-maker invests resources in a course of action. Early results are discouraging. Rather than reassessing, they invest more. Results deteriorate further. They invest more still. By the time the failure is undeniable, the losses dwarf what an early exit would have cost.

This isn't just stubbornness.

Researcher Barry Staw first documented this phenomenon systematically in the 1970s, using a study where participants were asked to allocate R&D funds to a struggling division. When participants were told they had personally made the original investment decision, they consistently poured more money into the failing project than participants who were told someone else had made the initial call. Personal responsibility for the original choice was enough to distort subsequent judgment.

What makes escalation so treacherous is how rational each individual step looks. You've already committed significant resources. Abandoning now means confirming the loss. Staying in gives you a chance of recovery. Every individual decision to continue feels like reasonable risk management. The trap is that you're applying future-oriented logic to justify past-oriented behavior.

Four forces tend to drive escalation, and they rarely act alone.

Completion pressure. Projects nearing a milestone feel almost done. "Almost done" is a cognitive mirage. It's how infrastructure projects balloon to three times their original budget and still get funded through the next phase.

Self-justification. Reversing course feels like an admission of error, and for many decision-makers (especially those in public or hierarchical roles), admitting error carries real social costs. The decision to continue gets tangled up with ego protection.

Prospect theory effects. When you're already in a loss position, the psychology of risk tolerance flips. People who would never accept a gamble from a neutral starting point will accept enormous risk to avoid locking in a loss. Daniel Kahneman and Amos Tversky documented this carefully. Losing organizations tend to make increasingly reckless bets precisely because they're trying to climb back to zero.

Political and organizational momentum. Inside large institutions, a committed project develops its own gravity. People's careers are attached to it. Vendors are contracted. Reporting structures form around it. Stopping requires overriding not just a decision but an entire social ecosystem that has grown up to support continuity.

Here's a diagram of how the cycle compounds:

graph TD
    A[Initial Investment Decision] --> B(Disappointing Early Results)
    B --> C{Reassess or Continue?}
    C --> D[/Justification: "We're too far in"/]
    D --> E[Additional Resources Committed]
    E --> F(Further Deterioration)
    F --> C
    C --> G((Exit Point Missed))

The loop is self-sealing. Each pass through it raises the psychological cost of exit and lowers the probability that anyone will call a stop.

So what actually breaks the cycle? A few things have empirical support.

One is structural separation: the person who evaluates whether to continue should not be the person who made the original commitment. This is why independent audits, outside boards, and red teams exist. The problem is that organizations resist bringing in outside reviewers precisely when they're most needed, because leadership senses the scrutiny would be uncomfortable.

Another is prospective framing. Ask not "how much have we invested?" but "if we were starting today with full knowledge of current conditions, would we choose to invest in this?" This reframe doesn't always work, but it can interrupt the past-orientation that keeps escalation going.

Pre-defined exit criteria also help significantly. Before a project begins, decide what conditions would trigger a reassessment and write them down. When those conditions arrive, you're not admitting failure; you're honoring a prior commitment to clarity. The decision gets made when emotions are calm, not when the pressure to continue is highest.

None of this is easy. Organizations that build genuine off-ramps into their decision processes are rare. The ones that do tend to fail less expensively, which is a form of success that rarely gets celebrated the way it should.

The map runs out when the territory keeps changing and your last known position is a sunk cost you can't let go of. Recognizing the loop is the first step. Building institutions that can break it is the harder, longer work.