You are staying in something that is clearly costing you more than it is giving you. A job, a relationship, a project, a strategy. The data is available. You can see it.
And you can't leave. Because you've put too much in.
This is the sunk cost fallacy — the tendency to continue an investment in a course of action because of what has already been spent on it, rather than because of what the future is likely to produce. Economists call the past investment "sunk" because it is gone regardless of what you do next. The rational decision should look only forward. The human brain rarely manages this.
Why the Fallacy Persists Despite Being Well-Known
The sunk cost fallacy is one of the most widely taught cognitive biases in economics, psychology, and business education. It is also one of the most persistent — knowing the name does not reliably produce immunity to it.
The reason is that the fallacy is not primarily a reasoning error. It is an emotional error that reasoning is then deployed to justify. When you've invested significantly in something, withdrawal creates multiple overlapping emotional costs: loss aversion (the pain of losing what you've put in), identity threat (the investment is entangled with who you are and what you've committed to), social exposure (others know about the investment and withdrawal produces visible failure), and grief (the recognition that the future the investment was supposed to produce is not coming).
The sunk cost fallacy is often not about refusing to use good reasoning. It is about using sophisticated reasoning to justify what the emotional system has already decided: that the pain of stopping now is unbearable.
Understanding the mechanism — "I'm engaging in motivated reasoning because the losses feel intolerable" — is more useful than the economic framing of "past costs are irrelevant to future decisions," because the economic frame doesn't address the emotional system that's driving the behavior.
Loss Aversion as Amplifier
Daniel Kahneman and Amos Tversky's research on prospect theory established that losses produce approximately twice the psychological impact of equivalent gains. Losing $100 feels roughly twice as bad as gaining $100 feels good. This loss aversion asymmetry means that the prospect of abandoning an investment — crystallizing the loss as real and permanent — is psychologically more costly than continuing the investment, even when continuing is the objectively worse outcome.
Sunk costs interact with loss aversion to produce a specific trap: as the investment grows, the potential loss from abandonment grows, which means the emotional cost of rational exit increases over time. The longer you stay in a bad investment, the harder it becomes to leave — not because the rational case for leaving has weakened, but because the emotional case for staying has strengthened.
This is counterintuitive in the extreme. You would expect the recognition that something is failing to produce an exit impulse. Instead, it often produces increased investment — a phenomenon economists call "escalation of commitment."
The Escalation Phenomenon
The escalation of commitment is the sunk cost fallacy's active form. Rather than simply failing to exit a bad investment, the person increases their investment in response to evidence of failure. The psychological logic: "If I invest more, the prior investment will be justified by eventual success." The economic logic: there is no prior investment to justify. Each new investment should be evaluated only on its own expected future return.
In organizations, the escalation of commitment has been documented in project management (the project is clearly failing but resources are added rather than withdrawn), military campaigns (attrition continues past the point of rational strategy), financial markets (averaging down in a declining asset), and relationships (staying in damaging situations because of how much has already been given).
Barry Staw, the organizational psychologist who named the escalation phenomenon in the 1970s, found that the people most likely to escalate commitment were those who were personally responsible for the original decision. The investment of the self — not just the investment of resources — intensified the trap.
What Actually Helps
The standard advice — "ignore sunk costs, only evaluate future costs and benefits" — is correct and nearly useless in practice because it doesn't address the emotional processing that needs to happen before rational evaluation becomes accessible.
Decoupling identity from investment. Sunk cost escalation is most powerful when the investment is identity-entangled — when continuing is not just about the project or relationship but about who you are and what your choices say about you. Separating your self-evaluation from the success or failure of a specific investment requires the cognitive work of distinguishing between "I made a decision that isn't working out" and "I am the kind of person who fails." The first is a recoverable data point. The second is an identity statement and the nervous system will protect it with sunk cost logic.
Pre-mortems and exit criteria. The most reliable way to exit bad investments is to define exit criteria before you enter them. "I will exit this investment if X has not occurred by Y date" — set in advance, before the emotional investment has accumulated. Pre-commitment of this kind bypasses the escalation mechanism because the decision rule was established when the decision-maker was not defending a prior choice.
Grief work as precondition. The loss aversion driving sunk cost behavior is about fear of loss. But the loss has already occurred — the sunk investment is gone. What the emotional system is protecting against is the experience of grief: the acknowledgment that what was invested is not coming back. Allowing the grief — processing the real loss — is often the precondition for rational exit, because once the grief is metabolized, the loss aversion diminishes. You are no longer avoiding the pain of loss. The loss has been felt.
The Protocol
-
Write the forward-looking analysis without reference to what has already been spent. This is the economist's move: describe only the costs and benefits from this moment forward. What will this cost going forward? What is the realistic probability of achieving what you're hoping for? What are the alternatives? Do this analysis in writing, so your reasoning is externalized and can be examined.
-
Identify the identity stake. What does abandonment say about you that you are afraid to have said? Name the specific narrative: "If I leave this, I am the kind of person who gives up / made a bad decision / wastes resources / fails publicly." The identity stake, once named, can be examined rather than defended. It is rarely as fixed as it feels.
-
Define your exit criteria before you need them. For every significant ongoing investment — a project, a relationship, a strategy — define the conditions under which you would exit. Write them down. The pre-commitment exists outside the emotional context that makes exit feel unbearable at the moment it's needed.
-
Allow the grief. Name the loss explicitly. Write about what you are giving up, what you hoped for, what was real and good about the investment. This is not wallowing — it is metabolizing. The grief that is processed cannot drive behavior the way avoided grief can.
The sunk cost fallacy's deepest trap is that it operates by making your past decisions the justification for your future ones. Your past self made a decision. Your present self is not required to honor it indefinitely. The best tribute to a past investment that isn't working is not continued investment. It is the freedom to build something that does.
