The Sunk Cost Fallacy
A reader's summary of the gap between what economics says about sunk costs and how people actually decide — the theater-ticket experiments that gave the effect its name, its 1970s “Concorde fallacy” nickname, and why continuing to invest in something because of what's already been spent shows up everywhere from unfinished projects to bad relationships.
The term at a glance
The sunk cost fallacy names the tendency to keep investing time, money, or effort in a decision because of what has already been spent, rather than because of what the choice is actually worth going forward. Standard economics treats sunk costs as irrelevant to any rational forward-looking decision — money spent is gone regardless of what happens next — but people routinely let past investment justify continuing anyway, often phrased as not wanting the earlier cost “to have been for nothing.”
Origin
The underlying principle is old in economics — cost theory has long held that only future, avoidable costs and benefits should factor into a decision, since sunk costs are the same size no matter what choice is made next. The behavioral name comes later: psychologists Hal Arkes and Catherine Blumer gave the effect its modern documentation in a 1985 paper, “The Psychology of Sunk Cost,” reporting a series of experiments — including one where people who had already paid for a discounted theater season ticket were far more likely to attend a specific play than people who had bought the same ticket at full, non-discounted price, purely because more money was already on the line. A parallel nickname, the “Concorde fallacy,” comes from evolutionary biologists tracking the Anglo-French supersonic jet program: the British and French governments kept funding Concorde well past the point commercial viability made sense, seemingly because so much had already been committed.
History and context
Arkes and Blumer's 1985 paper set off decades of follow-up work situating the sunk cost effect inside the broader behavioral-economics picture built by Daniel Kahneman and Amos Tversky's prospect theory, which had shown a few years earlier that losses register more heavily than equivalent gains — a plausible mechanism for why abandoning a costly project feels like locking in a loss, while continuing preserves the fantasy that the investment might still pay off. The “Concorde fallacy” label, coined independently by biologists studying whether animals escalate commitment to a losing strategy, briefly suggested the bias might be a general feature of decision-making under investment rather than a uniquely human, culturally learned habit — a claim later research complicated considerably.
Main ideas
Rational choice says sunk costs shouldn't count
Standard microeconomic theory treats only marginal, forward-looking costs and benefits as relevant to a decision — a cost already paid, and unrecoverable regardless of what happens next, carries no information about whether continuing is worthwhile.
The theater-ticket experiments made the bias visible
Hal Arkes and Catherine Blumer's 1985 studies varied only how much people had already paid for otherwise identical options, and found spending decisions shifted anyway — a purely sunk, already-spent amount changed what people chose to do next, contrary to the rational-choice prediction.
Concorde gave the effect its most-cited real-world case
Britain and France kept funding the supersonic Concorde program well after cost overruns and thin commercial prospects were clear, a pattern biologists later used as shorthand for any actor persisting with a losing course of action because of prior investment.
Loss aversion offers a mechanism, not just a label
Daniel Kahneman and Amos Tversky's prospect theory found people weigh losses roughly twice as heavily as equivalent gains, which helps explain why walking away from a sunk cost — an admission the money or time is simply gone — feels worse than continuing to hope it eventually pays off.
Whether animals show it at all is contested
Early comparative work suggested other species escalate commitment the way humans do, but later, more carefully controlled experiments — particularly with pigeons and mice — have struggled to replicate an animal version of the effect, leaving open whether it requires the kind of narrative self-justification only humans engage in.
Critique
- The lab studies don't always generalize. Many classic demonstrations use small, one-off stakes and forced-choice setups; some later replications with real money and real projects find the effect shrinks or disappears once people can gather new information before deciding.
- Continuing isn't always irrational. Sunk costs are sometimes bundled with real information: quitting early can signal something about a project to investors, partners, or a market, so persistence can be a rational reputational or informational choice mislabeled as pure bias.
- “Fallacy” flattens a more mixed picture. The same behavior gets called sunk cost fallacy when someone quits too early and called grit or commitment when someone persists and later succeeds — the label is frequently applied only in hindsight, once the outcome is already known.
Impact
The sunk cost fallacy now sits alongside the Dunning-Kruger Effect and the Backfire Effect as one of the small set of named cognitive biases invoked constantly in ordinary online argument — about a game that has stopped being fun after fifty hours played, a relationship kept going past the point either side is happy, or a founder unwilling to shut down a product because of the years already sunk into it. Its most frequent modern use is as an accusation: telling someone their reasoning is “just sunk cost” has become a fast way to end a debate about whether to keep going, whether or not the situation actually matches Arkes and Blumer's original experiments.