Your critical thinking gym
Home
The Commandments
Join the Club
Your critical thinking gym
● Cognitive bias
← From "sunk cost" (economics term, 20th century), first documented experimentally by psychologists Hal Arkes and Catherine Blumer, 1985.
The pull to keep going with something because of what you've already put into it, money, time, effort, instead of what it's actually worth continuing from here.
Also called: sunk cost effect · sunk cost bias · "throwing good money after bad." Not the same as escalation of commitment or loss aversion (Section 01).
The tendency to continue investing in a decision, relationship, or project based on the cumulative resources already committed to it (money, time, or effort) rather than on an unbiased assessment of the future costs and benefits of continuing from this point forward.
Plain version
Your brain treats what you've already spent as a reason to keep spending, even though that money, time, or effort is gone either way. The only thing that should actually matter for the decision in front of you is what happens next, not what already happened.
Aliases
Sunk cost effect and sunk cost bias are used interchangeably with sunk cost fallacy in most research. The "Concorde fallacy" is an older, less formal nickname for the same pattern, named after the supersonic jet whose development continued for years after it was clear it would never turn a profit.
Not the same as escalation of commitment
Sunk cost fallacy is the general judgment error: treating an irrecoverable past cost as relevant to a forward-looking decision. Escalation of commitment is the specific organizational-behavior term researcher Barry Staw coined in 1976 for this pattern playing out in institutional settings, often compounded by pressures sunk cost alone doesn't require: public accountability, personal responsibility for the original call, and pressure to save face in front of others. Every escalation-of-commitment case involves sunk cost thinking; not every sunk cost decision involves the extra organizational pressure that defines escalation of commitment.
Not the same as loss aversion
Loss aversion is the broader finding that losses are felt roughly twice as strongly as equivalent gains, in almost any context. Sunk cost fallacy is one specific, common misapplication of that general tendency: treating "stopping now" as though it were the loss, when the resources were actually lost the moment they were spent, regardless of what happens next. See Section 09 for how the two interact.
Scope
Gets confused with
Gambler's fallacy: opposite logic, distinguished in Section 09.
There's usually more of a felt signal here than with a lot of biases on this site: a real pang about "wasting" what you've already put in. But that pang gets misread as caring about not wasting things, not flagged as bias, and a completely reasonable-sounding cover story is always sitting right there ("I'm just seeing it through"), while quitting means admitting the money, time, or effort already spent bought you nothing.
Why your brain does this
"Don't waste things you've worked for" is a genuinely useful rule most of the time. Quitting the instant something gets hard would mean abandoning plenty of efforts that were about to pay off: most projects, relationships, and skills go through a rough patch before they work. A rough general instinct to push through, rather than re-litigating every setback from scratch, saves a lot of promising efforts from premature abandonment. This isn't a broken brain. It's a fast one, tuned for a world where quitting too early is a real and common mistake too.
Where it misfires
The shortcut breaks down once a decision genuinely should be made on future terms alone, when the honest answer to "would I start this today, knowing what I know now?" is no. At that point, "don't waste what I've put in" stops protecting you from premature quitting and starts protecting you from ever admitting the earlier investment isn't coming back, no matter what you do next.
The live academic debate, briefly
Psychologist Gerd Gigerenzer's ecological-rationality research treats persistence heuristics like this one as reasonable tools misapplied outside the conditions where they work well: in an uncertain world where you rarely know in advance whether "one more push" would have paid off, a general bias toward finishing what you start is a defensible strategy on average, even though it produces clear losses in hindsight cases like this one. The competing framing, from psychologists Daniel Kahneman and Amos Tversky's prospect theory, treats this squarely as an error against a rational standard: people weigh a certain, immediate loss (admitting the investment is gone) so much more heavily than an uncertain future one (more resources possibly wasted later) that they'll accept worse expected outcomes just to delay facing the first one. This page doesn't take a side between these two views; both are live in the research, and the practical lesson is the same either way: the pull to continue isn't stupidity, it's a heuristic built for stopping too early, being run in a situation that actually calls for the opposite.
How it activates, step by step
1. You commit resources (money, time, effort) to a decision, project, or relationship.
2. New information arrives suggesting the future prospects are worse than expected, or no longer worth it.
3. Stopping gets framed, internally, as making the prior investment "wasted," a loss you have to accept right now.
4. You continue investing to avoid confirming that loss, even when the future costs and benefits, weighed on their own, say to stop.
Sunk cost fallacy isn't only something that happens to you privately. Several propaganda techniques are built specifically to trigger it in an audience.
Dangles an eventual payoff, "just a little more and it'll all be worth it," at the exact moment an audience might otherwise cut its losses. Your own reluctance to have invested for nothing does the rest, making the promised future reward feel like the only thing that could redeem the resources already spent.
"Don't let their sacrifice have been for nothing" reframes stopping a costly policy, conflict, or campaign as dishonoring what's already been paid in lives, money, or effort, recruiting the audience's own sunk cost aversion in service of continuing something that might otherwise be reconsidered on its merits.
False urgency / crisis framing
Pairs a ticking clock with reminders of what's already been invested, so that stopping now looks like it would waste both the resources already spent and a narrow, closing window to act, doubling the pressure to keep going rather than reassess.
The more publicly and personally a follower has invested in a leader (money donated, relationships strained, reputation staked), the higher the psychological cost of admitting that investment was a mistake. Leader-centered rhetoric can lean on that directly, treating any doubt as a betrayal of what the follower has already given.
Outside politics
Gym memberships and timeshare sales rely on this passively: once an initiation fee is paid, canceling feels like a waste even for a membership never used. Recovery scams use it actively and predatorily, targeting people who've already lost money once (Section 05).
What increases it
Strong: personal responsibility for the original decision. Researcher Barry Staw's foundational 1976 research (Section 06) found escalation was strongest specifically when the person deciding whether to continue was also the one who made the initial choice.
Strong: strong negative emotional reaction to the potential loss. A 2019 study by researchers Koen Dijkstra and Ying-yi Hong found the size of the sunk cost effect tracked closely with how much negative feeling a scenario provoked, not just its dollar value.
Moderate: younger age. Several studies (researchers Strough et al., 2008; Bruine de Bruin et al., 2007) found younger adults more susceptible than older adults, who were more willing to cancel a failing plan even after heavy prior investment.
What does NOT clearly predict it
Unlike some biases on this site, the research on general intelligence here is genuinely mixed rather than a clean "no effect." Some studies (researcher Haita-Falah, 2017) found no protective effect from broad cognitive-ability measures. Others found an effect only for a narrower skill called cognitive reflection, the ability to override an instinctive first answer, rather than for raw intelligence itself (researchers Ronayne, Sgroi and Tuckwell, 2021). The honest summary: being "smart" in the general-IQ sense doesn't reliably protect you here; a specific habit of pausing to double-check your own gut reaction might.
State-dependent factors
Yes: the size of the effect rises with negative emotion in the moment (Dijkstra and Hong, 2019), consistent with why it can feel harder to think clearly about quitting something precisely when quitting would sting the most. Some research has also found a correlational link between elevated rumination and higher sunk-cost susceptibility (researchers Jarmolowicz et al., 2016), a finding about a general research sample, not something that says anything about any individual reader.
The pattern across all four: the size of the sunk cost doesn't have to be large, and the person doesn't have to be careless or uninformed. What predicts the effect is whether stopping would require openly treating the earlier investment as gone, in a theater seat, a national war, a lab experiment, or a second scam payment alike.
First documented
Psychologists Hal Arkes and Catherine Blumer published the first systematic experimental demonstrations in "The Psychology of Sunk Cost" (Organizational Behavior and Human Decision Processes, 35(1), 124-140, 1985).
The landmark studies
Arkes and Blumer's paper combined hypothetical-scenario experiments, including a case where participants decided whether to keep funding a failing aircraft project with or without being reminded how much had already been spent, with the real-world Ohio University theater ticket field experiment described in Section 05. Together, the two approaches showed the effect held both in the lab and in people's actual money-spending behavior.
Two parallel research threads
Researcher Barry Staw's 1976 paper "Knee-Deep in the Big Muddy", which coined "escalation of commitment," approached the same behavior from organizational psychology, emphasizing personal responsibility and self-justification rather than pure loss-avoidance. Around the same time, psychologists Daniel Kahneman and Amos Tversky's 1979 prospect theory supplied the underlying mechanism many researchers now use to explain both: losses are weighted roughly twice as heavily as equivalent gains, which economist Richard Thaler later connected to "mental accounting," the finding that people track sunk costs in a separate mental ledger from future costs, rather than ignoring them the way standard economic theory says they should.
An open dispute worth flagging
A 2018 Science paper by researcher Brian Sweis and colleagues reported that mice, rats, and humans all showed a similar time-investment-based pattern in a foraging task, suggesting a shared, ancient mechanism across species. A later methodological critique argued the same behavioral pattern can arise from ordinary statistical reasoning about expected reward timing, without needing to invoke genuine sunk cost sensitivity at all, a genuinely unresolved dispute in the animal-cognition literature, not a settled finding either way.
See Section 02 for how the German (Gigerenzer) research tradition frames why this shortcut exists at all.
Score: 2 out of 5 (see the breakdown in the card after Section 01).
The in-the-moment question
"If I hadn't already spent what I've spent so far, and I were deciding today from zero, would I still choose to start this?" If the honest answer is no, that's the signal, not whatever discomfort comes with asking the question.
Why this is hard to catch
There's usually a real felt pang here, which gives this bias more introspective visibility than some others on this site. But that pang gets misread as caring about not wasting things, rather than flagged as bias, and a completely reasonable-sounding story is always available ("I'm just seeing it through," "I'm not a quitter"). Meanwhile, stopping means openly admitting the earlier investment bought nothing, exactly the loss the whole shortcut exists to help you avoid confronting. A faint internal alarm is up against a strong, ready-made cover story and a real emotional cost to listening to it.
Debiasing research note
The evidence here is real but modest. Brief mindfulness-based interventions have shown a measurable reduction in lab settings (Hafenbrack, Kinias & Barsade, Psychological Science, 2014), and bringing in an uninvolved evaluator reliably reduces escalation in organizational simulations, but neither eliminates the effect, and results that hold up in a controlled lab task don't always transfer cleanly to a real, high-stakes, emotionally loaded decision.
Commonly co-occurs with
Loss aversion compounds with this one directly but doesn't yet have its own Filter page to link to.
Often mistaken for
Sounds related but runs the opposite direction: it's the mistaken belief that independent random events will "balance out," a coin "due" for heads after a run of tails. Sunk cost fallacy isn't about random independent events at all; it's about treating a real, non-random past investment as relevant to a decision it has no actual bearing on. One is a mistake about probability, the other is a mistake about what counts as evidence for a decision. This site doesn't yet have a Filter page for gambler's fallacy to link to.
Compounding effect
Loss aversion typically comes first and feeds sunk cost fallacy directly: it's what makes "stopping now" get framed as a loss to avoid, rather than as simply declining to spend more. Once the decision to continue is locked in, overconfidence and self-serving bias often follow, generating a "this will still pay off, I know what I'm doing" story that further insulates the decision from being reversed later.
Primary sources
Accessible reading