Decision making · Article

Sunk Cost Fallacy: Examples and How to Know When to Quit

What the sunk cost fallacy is, what the classic studies found, everyday examples from degrees to projects, when past investment really does matter, and five questions that help you decide whether to keep going or quit.

Sunk Cost Fallacy: Examples and How to Know When to Quit

Rohan is a physical therapist in Denver. Sixteen months ago he started a part-time master's in data analytics, because he wanted a desk job with better hours. He has paid $18,000 in tuition and given up most of his evenings. This fall he started an internship, and he's learned something he can't unlearn: he doesn't like the work. He misses his patients.

Over dinner, he tells his sister Meera he's going to finish anyway. "If I quit now, the whole $18,000 is wasted."

Meera pours the tea and asks one question. "Forget what you've paid. If someone offered you the last eight months of this degree today, for $9,000 and all your evenings, would you sign up?"

Rohan opens his mouth, then closes it. He wouldn't. The only reason he's staying is money he can never get back. That pull has a name: the sunk cost fallacy. This guide covers what it is, what the classic studies found, why it's so hard to resist, when past investment really does matter, and five questions that help you decide whether to keep going or quit.

What is the sunk cost fallacy?

A sunk cost is time, money or effort you've already spent and can't get back. The sunk cost fallacy is letting those past costs push you to keep going with something, even when the future costs outweigh the future benefits. Instead of asking "Is the next hour or dollar worth it?", you ask "How do I avoid wasting what I've already put in?"

The logic is simple once you see it. The $18,000 is gone whether Rohan finishes or quits. It can't be a reason to choose either path, because it's the same on both. Only the future differs: eight more months, $9,000 more, and a career he may not want.

It's entry 11 in our guide to cognitive biases. You'll also see it called the sunk cost effect, the Concorde fallacy, or "throwing good money after bad." When each new investment makes the next one feel more necessary, researchers call it escalation of commitment.

What the classic studies found

In economics, it's a basic rule that sunk costs shouldn't affect future decisions. The psychology of why people break it took off in 1985, when Hal Arkes and Catherine Blumer of Ohio University published "The Psychology of Sunk Cost" in Organizational Behavior and Human Decision Processes.

The ski trip

One of their questions asked people to imagine they had paid $100 for a ski weekend in Michigan and $50 for one in Wisconsin, then realized both trips were on the same weekend. Both were already paid for, and they expected to enjoy the Wisconsin trip more. Which would they go on?

Economic logic says Wisconsin: all $150 is spent either way, so pick the better weekend. But 33 of the 61 people asked chose Michigan, the trip they expected to enjoy less, because it had cost more.

The theater tickets

The more striking study happened at a real box office. The first 60 people who came to buy season tickets for the Ohio University Theater's 1982-83 season were randomly sold them at full price or at one of two discounts. Those who had paid full price went to more plays in the first half of the season. By the second half, the difference had faded.

The company that kept investing

Before that, in 1976, the organizational psychologist Barry Staw ran an experiment with 240 business students at the University of Illinois. Each played the financial vice president of a fictional company, the Adams & Smith Company, and had to give $10 million in research money to one of two divisions. Then they learned what happened over the next five years and were given $20 million more to split however they liked.

Students who had picked a division themselves and then watched it decline put an average of about $13 million of the $20 million back into it, more than any other group. Students told that another executive had picked the failing division didn't pile on in the same way. Staw's explanation, in a paper titled "Knee-Deep in the Big Muddy," was self-justification: we keep investing to prove the first decision was right.

Why it's called the Concorde fallacy

In 1976, the biologists Richard Dawkins and T. R. Carlisle published a short paper in Nature titled "Parental investment, mate desertion and a fallacy." They argued that an animal deciding whether to keep caring for its young should weigh the future benefits, not how much it had already invested. They named the mistake after the supersonic Concorde airliner, whose development went on long after it was clear the project wouldn't pay for itself.

Here's the twist. In a 1999 review in Psychological Bulletin titled "The Sunk Cost and Concorde Effects: Are Humans Less Rational Than Lower Animals?", Arkes and Peter Ayton concluded that there were no clear cases of animals actually committing the fallacy. They also reported studies in which young children seemed less prone to it than older children and adults.

Their explanation: adults learn a sensible rule, "don't waste," and apply it where it doesn't fit. Abandoning something we paid for feels like waste, even when staying wastes more.

Sunk cost fallacy examples

You'll recognize most of these. In each, the past is quietly casting a vote on the future.

Where What it sounds like What's actually sunk The forward question
Education "I'm halfway through this degree, I can't quit now." Tuition and semesters already done Do I want what the rest of the degree leads to?
Projects at work "We've spent two years on this system, we have to launch it." Budget and hours already spent Would we fund the remaining work if it were a new proposal?
Money "I'll sell the stock once it gets back to what I paid." The purchase price Would I buy this stock today at today's price?
Relationships "We've been together seven years, I can't just walk away." Years together Do I want the next seven years to look like this?
Everyday "The movie's terrible, but I paid for the ticket." The ticket Is the next hour better spent here or somewhere else?

The stock example often comes with loss aversion: selling makes a loss feel "real," so people wait to break even. Annie Duke, a former professional poker player and author of Quit: The Power of Knowing When to Walk Away, put the right question in one line in an interview with David Epstein: "would you use that $40 to buy that stock today?"

Debate it on Debatly“Should you finish a degree you no longer want just because you're more than halfway through?”

Why the sunk cost fallacy is so hard to resist

Knowing the definition doesn't make you immune.

  • The "don't waste" rule. As Arkes and Ayton argued, it's a good rule that gets overapplied. Leaving feels like admitting the money was wasted, even though it was spent either way.
  • Loss aversion. Losses tend to feel worse than equal gains feel good. Quitting turns a hoped-for recovery into a definite loss.
  • Self-justification. Staw's students escalated most when the original decision was theirs. Quitting means admitting a mistake, to yourself and to others.
  • Time already spent. In a 2018 study in Science, Brian Sweis, David Redish and colleagues at the University of Minnesota gave mice, rats and humans the same kind of waiting game: rodents waited for food in a maze, people waited for short videos behind a download bar. In all three species, the longer they had already waited, the less likely they were to quit. Time spent deciding whether to get in line didn't seem to count. The pull started once they had committed.

Duke sums up the result bluntly. In an interview with The Motley Fool, she said that "almost all the cognitive forces that are working inside of your head are going to get you to quit too late."

When past investment does matter

Not every "I've already put so much into this" is a fallacy. The money itself shouldn't count, but what it bought you often should. The test: does the past change your future costs or benefits?

Sounds like sunk cost But it can be a real reason Why it counts
"I've already done 90 percent of the work." The remaining cost is small The next step is cheap, so finishing may be the best deal on the table
"I've spent two years learning this." You now have skills and knowledge Your future odds are better than a beginner's
"We've already tested three versions." You've learned what doesn't work Information changes your forecast
"I promised the team I'd see it through." Reputation and trust Breaking a promise has a future cost
"Switching now would be expensive." Real switching costs The cost of changing is in the future, not the past

So "I'm 90 percent done" is fine if the last 10 percent is small and still worth doing. It's a fallacy when that 10 percent is really half the work, or the finished thing is no longer worth having.

To sort it out, translate every backward-looking reason into a forward-looking one. "I've spent $18,000" says nothing about the future. "My finished courses already count toward a certificate" does.

Idea What happens Example
Sunk cost fallacy Past, unrecoverable costs drive a future choice Finishing a bad meal because you paid for it
Escalation of commitment Repeatedly adding more resources to a failing course, often to justify the first choice Funding year three of a project that missed every goal in years one and two
Loss aversion Losses feel bigger than equal gains Refusing to sell a stock below the purchase price
False dilemma Seeing only "quit everything" or "keep going exactly as before" Ignoring a smaller version, a pause or a pivot

The last row matters most. Sunk cost thinking often hides behind an all-or-nothing frame. There's almost always a third option: scale it down, pause it, hand it off, or keep the useful part and drop the rest.

5 questions to decide whether to keep going

You don't need to fight your feelings. You need questions that make the future visible.

1. Would I start this today?

Imagine you had never begun. Knowing what you know now, would you sign up for the remaining cost, time and effort to get the remaining benefit? This is Meera's question. If the honest answer is no, the past is the only thing keeping you in.

2. What does the next hour or dollar buy?

Duke put it this way in an interview with CNBC: "What really matters is whether the next minute or the next dollar or the next bit of effort that you put into that project is worthwhile." Write down what the next step costs and what it gets you, and leave the past column blank.

3. What else could I do with it?

Every hour you put in here is an hour you can't put in somewhere else, which economists call opportunity cost. Name your best alternative specifically. A vague "something else" always loses to the concrete plan you already have.

4. What would a newcomer decide?

Staw's students escalated most when they had made the first choice themselves. Borrow that finding: picture a smart person taking over tomorrow with no ego in the original decision. What would they do in their first week? Ask someone who wasn't there at the start, or argue the other side yourself (see playing devil's advocate).

5. What would make me stop?

The best time to decide when to quit is before you're deep in. Duke calls these kill criteria: conditions, written down in advance, that trigger a stop. Speaking with David Epstein, she added that "it's not just a state of the world that you need to be identifying, it's actually a date." For example: "If I don't have three paying clients by June 30, I close the side business."

If you've already started without them, set them now. A pre-mortem helps: imagine the project failed a year from now and ask what the warning signs were. Those signs make good kill criteria.

Debate it on Debatly“Is "I've come too far to quit" ever a good reason to keep going?”

Worked example: Rohan's degree

Rohan spent the weekend on the five questions, and Meera made him write the answers down.

Question Rohan's answer
Would I start this today? No. I'd never pay $9,000 and eight months of evenings for a job I now know I don't enjoy.
What does the next dollar buy? A full degree for a career I don't want, plus some data skills I could use anyway.
What else could I do with it? Ask my clinic about a senior role with better hours, which I'd been avoiding asking.
What would a newcomer decide? Keep what's useful, stop paying for what isn't.
What would make me stop? Not relevant anymore, but next time: decide by the end of the internship.

Then he looked for a third option, and found one. His program gave students who left early a graduate certificate for the courses they had finished. His new skills weren't sunk: his clinic was starting to track patient outcomes and needed someone who could read the data.

He left with the certificate and asked about the senior role. Six weeks later he had a version of it, with fewer evening shifts and the clinic's outcome reports. The $18,000 wasn't "wasted." It just stopped deciding for him.

How to practice this with AI

The sunk cost fallacy is hard to catch alone, because the person who made the first decision is the one judging whether to continue. A debate brings in voices with no stake in your past choices. On Debatly, several AI models and characters argue different sides of your question over several rounds, you can step in at any point, and a judge gives a verdict with reasoning.

  • Frame it as a fresh choice. In the debate room, don't ask "Should I keep going with my degree?" Ask "Starting today, should I spend $9,000 and eight months on the rest of this degree, or take the certificate and go back to clinical work?" The framing alone removes the sunk cost from the question.
  • Bring in a newcomer. Ask Petra, the devil's advocate, to argue for quitting as hard as she can. Add James, the CFO character, to look only at future costs and returns. Neither of them made your original decision, so neither needs to defend it.
  • Ask for third options. Have the debaters list ways to keep the useful part and drop the rest: a smaller version, a pause, a hand-off.
  • Read the verdict, then the reasoning. Judge Alex explains which arguments carried the round. If the winning side relied on "you've already invested so much," be suspicious.

When Rohan tried his question as a debate, the side arguing for leaving asked in its first round whether the program had an early exit option. He hadn't thought to check.

Debate it on Debatly“Should a company cancel a project that's 80 percent finished if the market has changed?”

An honest limit

AI models don't know your finances, your program's rules or how you feel on a Monday morning. They also lean toward what you seem to want, so a single chatbot may happily tell you to keep going (see why AI agrees with you). Use the debate to surface forward-looking arguments and missed options, then check the facts and make the call yourself.

Let the past stop voting

The sunk cost fallacy turns "I've already spent this" into a reason to spend more. But what's spent is gone either way. What should decide: what the next step costs, what it gets you, and what else you could do instead.

This week, pick one thing you're continuing mostly because you started it, and ask Meera's question: would you sign up for the rest of it today? If it's a big one, run it as a debate on Debatly and let someone with no stake in your first decision argue the other side. Give it a try. Plans are on the pricing page.

Debate it on Debatly

“Sunk Cost Fallacy: Examples and How to Know When to Quit”

FAQ

It's continuing with something because of what you've already put into it, rather than because of what you'll get out of it from here. The time, money or effort already spent can't be recovered, so it shouldn't decide what you do next. Finishing a degree you no longer want because you've paid half the tuition is a classic example.
The biologists Richard Dawkins and T. R. Carlisle used the name in a 1976 paper in Nature, after the supersonic Concorde airliner. Its development continued long after it was clear the project wouldn't pay for itself. The name is used mostly in biology, while psychologists usually say sunk cost effect.
The fallacy itself is, because spent costs are the same whichever way you choose. But past investment can change your future: you may have gained skills, learned something useful, made promises, or have only a little work left. Those are forward-looking reasons and perfectly fair to weigh.
Ask whether you would start the project today, knowing what you know now. Look only at the next step's cost and benefit, name your best alternative, and imagine what a newcomer with no stake would decide. Setting stop conditions with a date before you begin, sometimes called kill criteria, makes the decision much easier later.
The sunk cost fallacy is letting past costs influence a single decision. Escalation of commitment is the pattern that can follow: repeatedly adding more resources to a failing course, often to justify the original choice. Barry Staw's 1976 study found people escalated most when they had personally made the first decision.
#sunk-cost-fallacy #cognitive-biases #decision-making #critical-thinking #quitting