You buy a concert ticket for $100. Concert night arrives, and you are exhausted. You have no desire to go. You would prefer to rest at home. By any forward-looking logic, the ticket is a sunk cost — it is spent, gone, irretrievable. The only decision that matters is: what would you rather do tonight: go to the concert or rest? If rest > concert, skip the concert.1
But the $100 haunts the decision. You think, "I already paid $100, I cannot waste it." You drag yourself to the concert despite low utility, driven by the feeling that not attending would mean "wasting" the $100. This is the sunk cost fallacy: the tendency to factor irretrievable past expenditures into forward-looking decisions. Rationally, the $100 should be irrelevant (it is already gone). Behaviorally, it is dispositive (it drives the decision).1
The sunk cost fallacy is not stupidity or irrationality. It is a sophisticated adaptation to a world where commitment matters. If people ignored sunk costs completely, they would abandon projects constantly, never follow through on commitments, and create chaos. The sunk cost fallacy is the mechanism that makes people honor commitments, stick with projects long enough for them to work out, and not constantly bail on uncertain ventures. The fallacy is adaptive in some contexts, maladaptive in others, and humans have not developed the ability to toggle between contexts.1
Sunk cost effects operate through three overlapping mechanisms:
1 — Loss Aversion and Waste Aversion A sunk cost activates loss aversion. The $100 is a loss that has already occurred. Not attending the concert feels like compounding the loss — you lose both the $100 and the experience, wasting the money. Attending the concert feels like at least recovering something from the loss. The pain of the loss is already incurred; the decision is between "loss + waste" (skip) and "loss + some recovery" (attend). Loss aversion makes the second option feel better, even though it objectively produces lower utility (the concert is worse than rest, but at least you attended).1
2 — Entitlement and Obligation A sunk cost creates a psychological entitlement: you paid, you are owed. Not claiming the owed experience (not attending the concert) feels like a violation of the entitlement. The money spent creates an obligation to use the ticket, to get your money's worth, to make the purchase "pay off." This obligation is emotional, not logical, but it is powerful. The mind keeps a ledger, and the $100 is an obligation that can only be settled by attending.1
3 — Identity and Commitment Signals A sunk cost becomes incorporated into identity: "I am the person who bought concert tickets." Not attending feels like an admission of bad judgment (you bought tickets you do not want to use). Attending, conversely, feels like justifying the purchase, making it a good decision retroactively. Sunk costs are tied to identity and the need to be seen as consistent and committed. Abandoning the concert feels like admitting failure; attending feels like following through.1
Sunk cost effects appear across all domains and all populations. Experts fall prey to them just as novices do. A business leader who has invested $5 million in a failing project finds it harder to cut losses than a leader facing the same project from scratch. The $5 million invested creates obligation and loss aversion that makes liquidation feel intolerable. The project continues to receive funding not because expected returns justify it (they do not) but because the sunk cost creates forward-looking bias.1
The effect is particularly strong when:
The original decision was public. A person who bought concert tickets and mentioned it to friends feels more obligated to attend (honoring the public commitment) than someone who bought tickets secretly.
The investment is personal. A business venture you founded and funded with your own money triggers stronger sunk cost effects than a venture someone else started.
The opportunity cost is high. If not attending the concert means a boring night at home, you feel the loss. If not attending means doing something you value more (important work, time with someone you love), sunk cost effects are weaker.
The reference point makes loss salient. If you think of the $100 as "money wasted if I don't go," sunk costs are stronger. If you reframe it as "a ticket I can no longer use," the effect weakens.
Time has passed. Sunk cost effects are stronger if the purchase was made long ago and memory of it has faded, because the purchase feels more like a committed obligation than a recent decision. The longer ago you paid, the more the payment feels like a binding commitment.1
The stock market is a domain where sunk cost effects have massive consequences. An investor buys a stock at $100. The stock has fallen to $60. The investor knows the stock is unlikely to recover. But the investor holds, refusing to sell. Why? Because selling locks in the $40 loss (makes it real and concrete), while holding preserves the hope of recovery (keeps the loss theoretical). The sunk cost — the $100 invested — is gone regardless. But the investor cannot ignore it. The loss is emotionally salient, and holding onto the position feels like "not yet admitting defeat."1
This creates a portfolio problem: investors hold losing positions too long (sunk cost effects) and sell winning positions too early (trying to "lock in" gains before they evaporate). The result is a bias toward loss-locking and gain-realizing that is opposite of the tax-efficient strategy. The sunk cost of the original investment drives this backwards behavior.1
Sunk costs affect relationship decisions as powerfully as they affect financial ones. A person in an unsatisfying relationship may stay because of the sunk cost of years invested, shared history, emotional deposits. The years invested are gone, irrelevant to whether the relationship will improve going forward. But loss aversion makes the thought of "wasting" those years intolerable. Staying feels obligatory because the past investment demands justification.1
This is particularly pernicious because it traps people in low-utility situations (the relationship is unsatisfying) to avoid the loss frame of abandonment (years invested with no return). The sunk cost fallacy makes people stay in relationships not because they believe they can improve but because the past investment feels irretrievable and thus worth defending. The mechanism is identical to the concert ticket: loss aversion + entitlement + identity + public commitment.1
Sunk cost effects can be weakened by reframing. A person who thinks, "I already paid, I should attend the concert" is dominated by sunk cost effects. A person who thinks, "What do I actually want to do tonight?" is freed from sunk costs. The reframe works by redirecting attention from past investment (the $100) to forward-looking utility (what would I rather do now).1
However, reframing is difficult under emotional arousal. When loss aversion is active (the thought of wasting $100 is emotionally painful), reframing to "ignore the sunk cost" feels wrong because it acknowledges the loss. The reframe requires accepting the loss psychologically, which is hard. A more effective reframe is to shift the reference point: instead of "I wasted $100 if I don't attend," reframe as "I spent $100 in the past, that is done, and now I have a choice." The past decision is sealed; only future choices matter. This reframe acknowledges the loss but closes it, freeing the forward-looking decision from its gravitational pull.1
Not all cases where past investment influences future decisions are sunk cost fallacies. Sometimes past investment should influence the future. If you invested $100 in concert tickets, and the concert is in two hours, the relevant decision is not "should I use the tickets" (sunk costs, irrelevant) but "has my preference genuinely changed since I bought them?" If your preference has changed (you are exhausted), attending is irrational. But if your preference has not changed (you still like the artist, the venue is fine), attending is rational — the past investment reflects a genuine preference, and the current situation supports acting on it.1
The key distinction is between:
The difference is whether your actual preference has changed. If it has, honoring sunk costs is fallacy. If it has not, past commitment reflects genuine preference and is rational to honor.1
Psychology: Loss Aversion — Sunk cost fallacy operates through loss aversion: the $100 spent is experienced as a loss, and the mind seeks to minimize the total loss by at least "using" the ticket. The fallacy would be weaker or absent if losses did not hurt so much relative to gains. Loss aversion is the emotional mechanism that makes sunk costs binding.
History: Strategic Patience and Calibrated Retreat — Military leaders and strategists struggle with sunk cost fallacy at scale. A general who has invested enormous casualties in a war theater finds it hard to retreat, because abandoning the position feels like wasting the soldiers already fallen. The sunk cost (casualties incurred) drives the decision to continue, even when retreat is militarily optimal. Understanding when to cut losses and when to honor past commitment is a central strategic skill, and sunk cost fallacy is the psychological mechanism that makes cutting losses emotionally intolerable.
Cross-Domain: Bias as Adaptive Heuristic — Sunk cost effects appear to be a fallacy (you should not let past spending influence future decisions), but they are actually adaptive in a world where commitment is valuable. If people ignored sunk costs completely, they would abandon projects constantly, undermining trust and consistent effort. Sunk cost effects enforce commitment and follow-through. The fallacy becomes harmful only when context changes (preferences shift, new information arrives) but the commitment mechanism does not adjust. The adaptive value of sunk cost effects is maintaining consistency; the harm is maintaining consistency even when consistency is no longer optimal.
The Sharpest Implication: If sunk costs drive decisions forward-looking logic says they should not touch, then much of what looks like "commitment" is actually loss aversion applied to past investment. A person who says, "I am committed to this project" might mean "I genuinely believe in this project" (rational commitment based on expected value), or might mean "I have invested so much that abandoning it feels unbearable" (sunk cost fallacy dressed as commitment). The two are indistinguishable from the outside, but they have opposite implications: the first says continue; the second says cut losses. The practical implication is that you cannot trust your sense of commitment to reliably distinguish between genuine belief and sunk-cost emotional binding. You must create external rules: before deciding to continue an investment, write down your expected return without considering past investment. If you would not make the investment today at current cost, stop. This bypasses sunk cost effects by creating a decision rule that eliminates past investment from the calculus.
Generative Questions: