Sunk Costs: Why the Past Cannot Be Recovered
Money already spent cannot be recovered by any future choice, so sunk costs belong in your history, not in your decision.
Reading
0%
Time left
~18 min
Quiz score
0/4
What this means
You paid $100 for a concert ticket months ago. A family emergency now makes it impossible to attend. Someone offers you $40 for the ticket. Should you take it?
Almost everyone's first instinct is to refuse, because $40 "loses" $60. But look carefully at what is actually still on the table. The $100 left your account long ago. It is gone whether you sell for $40, sell for $80, or let the ticket expire worthless in a drawer. No available choice returns it. That makes the $100 a sunk cost, and sunk costs are identical across every option you are comparing.
Sound decisions rest on marginal analysis. Strip away everything the options share and ask what genuinely differs between them. Here the comparison is stark: sell and hold $40, or do not sell and hold $0. Forty dollars is greater than zero. Sell.
Notice what the principle does not claim. It does not say the $100 was wasted, or that you were foolish to buy the ticket. Buying it was probably a good decision given what you knew at the time. The principle is narrower and stricter than that: past expenditures carry no information about which future option is better, because they attach equally to all of them.
The mirror image also holds for benefits. Rewards you have already received are equally sunk. An athlete who has already won a championship should not stay in a punishing training program purely because of what that program delivered in the past. The relevant question is what the next season of training costs and produces.
Economists call the failure to apply this reasoning the sunk cost fallacy. It reliably shows up when the prior investment was large, public, or personally chosen, because abandoning it feels like admitting the earlier decision was wrong.
Why it matters
The fallacy is expensive precisely because it feels responsible. Finishing a film you are not enjoying, staying in a major you have come to dislike because you already completed four courses in it, continuing to repair a car whose repair bills now exceed its value — each is the same error wearing different clothes. In every case the sensible question is identical: from right now, forward only, does continuing produce more benefit than the best alternative use of my time and money?
Organizations fall into it at enormous scale. A large infrastructure project years behind schedule and far over budget generates powerful arguments to continue built entirely on what has already been spent. Those arguments are precisely the ones that should carry no weight. Whether to continue depends only on remaining costs versus remaining benefits.
Real-world example
Look up a major public infrastructure project in your state or country that has run substantially over its original budget, then read how supporters argued for continued funding. You will usually find a version of "we have already invested too much to stop now." That sentence is the sunk cost fallacy stated out loud. The defensible version of the argument is entirely different and sounds like this: the cost to finish from today is less than the value the finished project will deliver from today. Notice which argument the coverage actually makes, and notice how rarely the two are distinguished.
Try it
- Work the concert ticket case in writing. List every option available to you now: sell at $40, hold out for a higher offer, give the ticket away, let it expire. For each, record only what you gain and what you give up from this moment forward. Deliberately leave the $100 out of every row.
- Now add a row for the $100 in each column. Confirm that it is the same number everywhere, and explain in one sentence why a quantity identical across all options cannot change which option ranks highest.
- Complicate it. Suppose a buyer offers $40 today, but you believe there is a reasonable chance of finding a buyer at $70 tomorrow, and holding out risks the ticket becoming worthless after the event begins. Identify which figures in this version are genuinely relevant, and which are still sunk. Note that the original $100 is sunk in this version too.
- Design a case where the price you paid IS relevant, then explain why. Refunds and resale floors are the interesting territory here. If the venue offers a full refund, the $100 is no longer sunk — it is recoverable, and the refund becomes an actual competing option worth $100.
- Interview someone about a decision they stuck with longer than they should have. A subscription, a project, a hobby with expensive equipment. Ask what made stopping difficult, and classify their reasons as forward-looking or backward-looking.
- Find a current news story about a project, product line, or program facing pressure to be cancelled. Extract every argument for continuing and sort them into two lists: arguments about resources already committed, and arguments about future benefits. Assess which list the case actually rests on.
- Write a short response to the following objection: "If we ignore sunk costs entirely, people will never learn from bad decisions." Explain how past outcomes can inform future judgment without being counted as costs of the current choice.
Teacher note
The single most common student response to the ticket problem is that selling at $40 means "losing $60," and this framing is worth confronting directly rather than correcting quickly. Ask the student to state what they hold under each option, in dollars, right now. The answer is $40 or $0, and stating it aloud usually does more work than any explanation you can offer. A second and subtler misconception is that the principle means the original purchase was irrational; students hear "irrelevant" as "mistaken." Separate the two explicitly, because a decision can be perfectly sound when made and still produce a cost that is sunk afterward. Step 4 is the step that distinguishes genuine understanding from a memorized rule, because it forces students to notice that sunk-ness is a property of recoverability, not of the past — a refundable expenditure sitting in the past is not sunk at all. Watch for students who apply the rule mechanically and conclude that any past expense should be ignored even when a refund is available. The most stubborn cases are personal ones involving effort rather than money, since students readily ignore sunk dollars but defend sunk hours of practice or study fiercely; press on those, because the logic is identical and the emotional resistance is where the learning lives. A student has it when they can explain that the $100 is irrelevant because it is the same under every option, rather than simply asserting that sunk costs should be ignored.
Check yourself
You paid $100 for a concert ticket you can no longer use, and the venue offers no refund. A buyer offers $40. What is the correct economic reasoning?
Which of the following is NOT a sunk cost?
A city has spent a large sum on a transit project that is now far over budget. Which argument for completing it is economically valid?
Why does the sunk cost principle apply to benefits already received as well as to costs already incurred?
Money and effort you cannot get back are the same under every option you are choosing between, so they can never tell you which option to pick.