The Sunk Cost Fallacy Keeps People Investing in Failing Decisions

Most people quit a bad decision too late, and it's rarely about the decision itself. A 2018 study on mice, rats, and humans shows why sunk costs are so hard to walk away from.

Most people who abandon a failing plan do it too late, and the reason usually has nothing to do with the plan itself. It's the money, time, or effort already spent that keeps them going, a pattern psychologists call the sunk cost fallacy.

A 2018 study found the same behavior in mice, rats, and humans, which suggests this isn't a glitch in modern reasoning. It may be older than reasoning itself, wired into how brains everywhere decide when to quit.

What Actually Counts as a Sunk Cost

A sunk cost is any resource already spent that cannot be recovered no matter what you choose next. The money paid for a concert ticket, the years spent on a failing degree, the funding poured into a stalled project: none of it comes back.

That's true whether you continue or stop, and no future decision can retroactively change it. The only thing left to weigh is what happens from this point forward.

Rational choice theory says only future costs and benefits should guide the next decision. In practice, people weigh the past investment as if it were still on the table, treating money already gone as a reason to spend more.

Economist Richard Thaler formalized this gap between rational models and real behavior in 1980. His work built on research that later fed into Daniel Kahneman and Amos Tversky's prospect theory, which showed people feel losses more sharply than equivalent gains.

The 1985 Study That Gave the Bias Its Name

Psychologists Hal Arkes and Catherine Blumer ran the study that made the sunk cost effect a formal concept. In their best-known scenario, participants imagined buying a $100 ski trip to Michigan, then a $50 trip to Wisconsin they expected to enjoy more.

Both trips fell on the same weekend, and only one ticket could be used. Most participants chose the worse trip anyway, reasoning that skipping the pricier ticket would waste more money overall.

Arkes and Blumer traced this to what they called the don't waste rule, a social norm against appearing wasteful that people apply even when nobody else is watching. The same researchers ran a field test at an Ohio University theater.

Some patrons received full-price season tickets, others received a discount, with equal access to every performance. Full-price ticket holders attended noticeably more plays across the season on average, showing the bias playing out in real theater seats rather than a hypothetical survey scenario dreamed up in a lab.

The Bias Shows Up in Mice and Rats, Not Just People

In 2018, a University of Minnesota team led by neuroscientist Brian Sweis tested mice, rats, and humans on parallel versions of a foraging task, publishing the results in the journal Science. Rodents chased flavored food pellets through a maze of timed feeding stations.

Human subjects, meanwhile, hunted for entertaining videos under a limited time budget. All three species showed the same pattern once they committed time to waiting for a reward, they grew less willing to walk away, even as the wait became clearly unproductive.

Senior author David Redish concluded that sensitivity to sunk time appears conserved across species. That points toward a shared cognitive bias rooted in old decision-making circuitry, not a uniquely human failure of logic layered on top of modern life.

The rodents had no language, no social image to protect, and no spreadsheet of past spending, yet they still hesitated to abandon a bad wait once they were already committed to it. That timing detail turned out to matter most: the hesitation only appeared after the initial commitment, not before it, in every species tested.

Why the Brain Falls for It

Two forces seem to drive the pattern in humans specifically, on top of whatever older mechanism the rodent studies point to. The first is loss aversion, the tendency to feel a loss roughly twice as intensely as an equivalent gain.

That makes walking away from a sunk investment feel like actively losing it, rather than simply not gaining more later, even though the two outcomes leave a person with the same balance sheet.

The second force is image and self-protection, the discomfort of admitting that earlier effort was wasted, especially in front of other people who watched the original decision get made.

A 2020 study in Frontiers in Psychology tested this directly, tying the two forces together with real data instead of just theory. People scoring higher on loss aversion measures were also more prone to the sunk cost effect for money invested.

The pattern was weaker and less consistent for time or effort spent. The researchers noted this split matters for anyone designing a workplace policy around the bias, since a rule built for money may not transfer cleanly to hours logged on a doomed project.

Researchers Are Now Questioning How Solid the Bias Really Is

Here is what most articles on this topic skip entirely, and it complicates the tidy story told so far. The 2018 mice-and-rats study, widely cited as proof the bias is built into animal brains, has since been formally challenged by other scientists working from the same published dataset.

A 2022 paper in the Proceedings of the National Academy of Sciences argued the original result could reflect a statistical artifact called attrition bias, not genuine sunk cost sensitivity. The authors built a model reproducing the same behavior without invoking sunk costs at all.

A separate 2025 study from the University of Wrocław went further, testing the classic vignettes researchers have relied on for four decades, including Arkes and Blumer's own scenarios. It found these vignettes often fail to reliably measure the same underlying trait between sessions.

That raises real questions about what decades of sunk cost research on cognitive bias actually captured. The bias itself is not in serious doubt, but the tools used to measure it are proving shakier than most textbooks suggest.

None of this means people are free of the trap. It means science is still sorting out exactly how it works and how consistently it shows up from one person, and one study, to the next.

How to Actually Catch Yourself Doing It

The fix researchers keep landing on is a single reframe, one simple enough to run through in the middle of an argument or a budget meeting. Ask what you would choose if you were starting today with zero prior investment, knowing only the current costs and the current odds of success.

If the answer is no, the money or time already spent should not change it, however strongly it feels like it should. This reframe works better in private decisions than group ones, where the stakes of admitting a mistake feel higher.

Reversing course in a group can look like admitting a public mistake, which is one reason overconfidence and sunk cost thinking often travel together inside failing projects.

Naming the decision as a sunk cost out loud, before discussing next steps, appears to weaken its pull in several debiasing studies built on Arkes and Blumer's original psychological framework from the 1980s.

Conclusion

The sunk cost fallacy survives because it disguises itself as loyalty, discipline, or simple thrift, not as an error. What the newest research adds isn't a bigger warning about the bias.

It's a more honest picture of it: real, measurable in decisions from theater tickets to corporate budgets, yet less mechanically simple than the tidy lab stories suggest.

The next time a bad investment feels too far along to quit, remember that the money already gone is the one number that should carry no weight at all.

Frequently Asked Questions

What is the sunk cost fallacy in simple terms?

It's the tendency to keep investing in a decision because of money, time, or effort already spent, even when the current evidence says stopping would be smarter. The past cost cannot be recovered either way, so it shouldn't affect the choice.

Is the sunk cost fallacy the same as loss aversion?

They're related but not identical. Loss aversion is the broader tendency to feel losses more strongly than equivalent gains, and researchers believe it partly explains why sunk costs feel so hard to walk away from.

Do animals really experience the sunk cost fallacy?

A 2018 Science study found mice, rats, and humans all showed increased persistence after investing time in a task. That finding has since been challenged by researchers who argue it may reflect a statistical artifact rather than true sunk cost sensitivity.

Why do businesses keep funding failing projects?

Escalation of commitment, the organizational version of the sunk cost fallacy, often stems from managers not wanting to admit a prior investment failed, especially when their own reputation is tied to the original decision.

What is the best way to avoid the sunk cost fallacy?

Evaluate the decision as if starting fresh today, using only current costs and expected future outcomes. If the past investment were zero, would you still choose this path.

Does time count as a sunk cost the same way money does?

Not always as strongly. Research on investment type suggests the sunk cost effect is more consistent and better documented for money than for time or effort, where results across studies have been mixed.