The Sunk Cost Fallacy: How to Decide Whether to Keep Going

I started at UH Mānoa as a computer science major because it felt like the responsible choice. I was on a computer all day anyway. Then the algorithm-heavy math arrived, along with a grade in ICS 202 I was barely holding above water. What kept me there was the semesters I had already put in. Leaving felt like admitting those semesters were wasted. A banner for UH's Academy for Creative Media eventually got me into a counselor's office with a Change of Major form. I switched to film and haven't looked back.

Those semesters were spent either way. The only thing still on the table was the next one.

Psychologists Hal Arkes and Catherine Blumer ran a field study at the Ohio University Theater. Sixty people buying season tickets for the 1982 to 1983 season were randomly charged $15, $13, or $8 for the same tickets. Over the first half of the season, the full-price group showed up to more plays than either discounted group. Same seats, same shows. The only difference was what they had paid.

Arkes and Blumer published the result in 1985 and traced the pull to a reluctance to appear wasteful. Paying more made skipping a show feel like throwing money away, even though the money was gone whether they went or not.

One caution: a 2007 team of experimental economists ran lab experiments and found a sunk cost effect that was smaller and less consistent than the theater study suggests. Treat it as a tendency worth checking for in your own reasoning.

Why it works

Sunk cost thinking sneaks in through the framing. "I've already put in two years" makes the past the subject of the sentence. Rewording the question so the past is zeroed out forces you to judge the thing on what it offers from here.

How to run it

• Write the project, commitment, or habit at the top of a page.
• Under it, answer in writing: "If I had put in no time and no money so far, would I start this today?"
• Give yourself four minutes. A yes or no with one line of reasoning is enough.
• If you catch yourself writing about what you've already spent, cross it out and answer again.

The honest limitation

The question can hide real switching costs, and some of those are legitimate. If leaving means a penalty, a broken contract, or restarting a credential, those are future costs and belong in the decision. The question only removes the costs you can't recover either way.

Why it works

Sunk costs and future costs feel like one big pile of effort. Writing them in separate columns shows you how little of that pile is still up for decision.

How to run it

• Draw a line down the page. Left column: everything already spent. Time, money, reputation, the semesters.
• Right column: what continuing costs from today, and what it's likely to return.
• Make the decision using only the right column. The left column is a record, and records don't vote.
• Keep the left column anyway. What you learned there comes with you whatever you choose.

The honest limitation

Your right column is a forecast. You're estimating effort and payoff for something you've been struggling with, which makes a confident estimate hard to come by. Write a range instead of a single guess.

Why it works

Plenty of people stay stuck because the only alternative they can picture is quitting outright, which feels final and a little humiliating. A no to the zero-based question opens more doors than that.

How to run it

• Continue. A yes is a fine answer. Plenty of hard things pass the test, and a rough stretch on its own doesn't mean the project is wrong.
• Pause. Set it down with a date to revisit. Table it, don't trash it.
• Shrink. Cut it down to the part you'd still start today and drop the rest.
• Hand it off. If the work matters and you're the wrong person for it, find the right one.
• Quit. If none of the above fits, close it cleanly and send the time somewhere better.

The honest limitation

Pausing can become a polite way of never deciding. If something has sat paused through two revisit dates, that's your answer. Make it official.

Some commitments are supposed to outlast your enthusiasm: caring for someone, a promise to a team mid-project, a lease you signed. The zero-based question still helps you see clearly, and other people's stake in the decision still counts. If you're in a low stretch from burnout or grief, almost everything will fail the "would I start this today" test. Wait for steadier ground before making a big call.

• Time and money already spent are gone whether you stay or leave.
• Ask in writing: "If I had put in no time and no money so far, would I start this today?"
• Four minutes, one page, one line of reasoning.
• Split the page: spent on the left, next on the right. Decide from the right.
• Real switching costs count. Unrecoverable ones don't.
• A no gives you options: pause, shrink, hand off, or quit.
• A yes is a fine answer. Hard and wrong are different things.
• Recognizing a wrong path quickly is a skill.