The sunk cost fallacy is the error of continuing a failing course of action because of resources already invested that cannot be recovered. Rational decision theory holds that past costs should be irrelevant to future choices. In practice, people allow them to dominate. The correct mental move is this: given where you are now, ignoring everything already spent, what is the best path forward? That question, asked honestly, bypasses the fallacy.
Where this came from
Economists Hal Arkes and Catherine Blumer published the defining study in 1985. In one experiment, participants were asked to imagine they had bought two ski trip tickets: a $100 trip to Michigan and a $50 trip to Wisconsin. They then discovered the trips overlapped and could not be rescheduled. Which trip do they attend? The rational answer is whichever trip they expect to enjoy more. Most participants said they would take the more expensive Michigan trip, even though the $100 had already been spent regardless of which they chose. The prior cost, not the expected enjoyment, drove the decision.
Arkes and Blumer ran multiple variants of the experiment and found the pattern consistently. The more a person had invested in a failing course of action, the more committed they became to continuing it. They called this the sunk cost effect and connected it to loss aversion: people continue investing not to gain value but to avoid the psychological pain of acknowledging a loss.
The Concorde supersonic jet remains the canonical institutional example. The UK and French governments poured money into the programme for decades after it became clear the aircraft would never achieve commercial viability. Government officials repeatedly cited previous investment as justification for continued funding. When the programme finally ended in 2003, the total cost was estimated at over three billion pounds in 1970s money. The decision to continue had been driven by prior costs rather than future prospects at nearly every review point.
How it works
Sunk costs are costs that have already been incurred and are non-recoverable regardless of what you decide next. The money spent on a degree you have already earned, the years invested in a career, the capital deployed in a failing business, the time given to a relationship that is not working: none of these can be returned. Standard economic theory says they should play no role in your next decision. Only future costs and benefits are relevant.
The human brain, however, registers those past costs as ongoing obligations. Stopping feels like wasting what has already been spent. Continuing feels like giving those prior investments a chance to pay off. This is an illusion. The past investment is gone whether you continue or not. Continuing only adds more cost to the loss.
Several mechanisms amplify the effect. Loss aversion, documented by Kahneman and Tversky, makes the pain of a confirmed loss feel roughly twice as intense as the pleasure of an equivalent gain. Admitting a project has failed means taking that loss. Continuing the project keeps the outcome uncertain and the loss technically unrealised. Ego threat adds another layer: if you were the person who championed the investment, stopping means publicly admitting your judgment was wrong. Public commitment amplifies this further: the more visibly you committed to a course of action, the more stopping looks like a personal failure rather than a rational correction.
The result is a predictable pattern in which the people most qualified to stop a failing project, those who know it most deeply, are also the people most psychologically committed to continuing it.
When to use it and when not to
Recognising the sunk cost fallacy is the tool. The relevant question is which situations create the highest risk of it operating unchecked. The answer: any decision where you have significant prior investment and the current trajectory is negative. Failing businesses, underperforming hires kept too long, technology projects past their original scope, real estate held past its rational exit point, and any relationship where the primary reason for staying is "we have been together for X years."
One important distinction: past investment sometimes does carry legitimate weight. If you have built expertise, relationships, or infrastructure through prior effort, those are real current assets, not sunk costs. The question is whether the past investment has created something of value that persists, or whether it is simply a psychological anchor. Expertise you acquired through years of work is a current asset. The $50,000 you spent on a business model that no longer works is a sunk cost.
Escalation of Commitment
Escalation of commitment is the pattern in which decision-makers increase investment in a failing course of action specifically because it is failing. Unlike the basic sunk cost fallacy, escalation involves an active doubling down: more resources, more public statements of confidence, more resistance to exit options. It is driven by social and ego pressure. Once a person has publicly committed to a course of action and staked their reputation on it, each piece of negative evidence creates a new incentive to prove the commitment was correct rather than reassess it. This is why institutional disasters, whether business or political, often involve not just persistence in error but acceleration toward it.
Run this framework on your actual decision.
DecisionsMatter.ai walks you through a structured 5-step analysis. Your first analysis is free.
How to apply it in practice
The single most useful exercise is the clean-slate reframe. Take the decision you are facing and write it out as if you are encountering it for the first time today, with no prior investment. "Should I start this project, given its current expected return and cost to complete?" not "Should I continue this project, given what I have already spent?" The question looks nearly identical but the mental framing is entirely different.
A second technique is to identify your stopping criteria in advance, before you are in a loss position. Before beginning any significant project or commitment, write down the specific conditions under which you would stop. What level of financial loss? What timeline slippage? What shift in market conditions? When those conditions arise, you are not making a new emotional decision under duress. You are executing a prior rational commitment.
When reviewing an ongoing investment, list only the future costs and future expected benefits. Explicitly exclude historical costs from the analysis. If the future-only calculation does not support continuation, the case for continuing rests entirely on the sunk cost, which is not a case at all.
If you find yourself using phrases like "but we have already put so much into this" or "we cannot walk away now after everything we have invested," treat those as diagnostic signals. They indicate the sunk cost fallacy is operating. The correct response is to acknowledge the loss, accept it as gone, and evaluate only what comes next.
This is one model from the upcoming Decisions Matter book.
30 mental models. 40 cognitive biases. A 5-step decision system. Get the relevant chapter in your inbox when it publishes.
Frequently asked questions
What is the sunk cost fallacy?
The sunk cost fallacy is the tendency to continue investing in a decision, project, or relationship because of resources already committed, even when the rational move is to stop. Sunk costs are costs that have already been incurred and cannot be recovered. Rational decision theory holds that they should be irrelevant to future choices. In practice, people find them very difficult to ignore. Economists Hal Arkes and Catherine Blumer documented the fallacy experimentally in 1985, showing that people in loss positions consistently over-invested to try to justify prior spending.
Why do smart people fall for the sunk cost fallacy?
Several forces compound each other. Loss aversion, documented by Kahneman and Tversky, means the pain of a loss feels roughly twice as intense as the pleasure of an equivalent gain, which makes accepting a loss feel disproportionately bad. Ego threat compounds this: admitting a project has failed means admitting the person who championed it made a mistake. Public commitment amplifies both effects. When a decision was announced publicly, retreating from it feels like a social failure on top of a financial one. Together, these forces make the irrational choice feel emotionally superior to the rational one.
How do you overcome the sunk cost fallacy?
The core mental move is to reframe the decision from its beginning rather than from where you are. Ask: "If I were starting fresh today, with no prior investment, would I choose to begin this project, continue this relationship, or take this path?" If the answer is no, the sunk cost fallacy is likely driving continued investment. Separating the decision from the investment history is cognitively difficult but becomes easier with practice. Writing down the reframed question explicitly, rather than reasoning through it silently, has been shown to improve decision quality.
What are the most common real-world examples of the sunk cost fallacy?
The Concorde supersonic jet is the canonical institutional example: the UK and French governments continued funding it for decades despite clear evidence it would never be commercially viable, partly because of the billions already invested. In business, failed product lines continue to receive funding because of development costs already spent. In personal life, people stay in careers they dislike because of the years invested in a degree, and remain in relationships that are not working because of the years already given. In investing, traders hold losing positions far longer than winning ones because selling a loss forces the admission that the original thesis was wrong.