Sunk Cost Fallacy vs Confirmation Bias
Sunk Cost Fallacy and Confirmation Bias are two Behavioral Economics concepts in AP Economics that students often mix up. The sunk cost fallacy is continuing an endeavor because of money or effort already spent, even when it is no longer worthwhile. Confirmation bias is the tendency to seek, notice and remember evidence that supports what you already believe, while discounting evidence that does not. Here is how they compare side by side.
Rational decisions should ignore sunk costs (which can't be recovered) and weigh only future costs and benefits. People fall into this trap because of loss aversion and a reluctance to 'waste' past investment.
A belief you already hold changes how you handle information at three stages: which sources you go to, how you read what you find, and what you later recall. Evidence that fits gets accepted at face value, while evidence that conflicts gets picked apart for flaws. Because the filter is directional, gathering more information can leave two people further apart rather than closer together. In economics it helps explain why forecasters cling to a prediction after the data turns, and why investors hold a losing position by reading every bad quarter as temporary. Note the contrast with the availability heuristic: availability distorts a judgment through what is easy to remember, with no prior conclusion required, while confirmation bias needs a conclusion to defend.
Sunk Cost Fallacy vs Confirmation Bias: Two Separate Engines Behind One Escalation
| Sunk Cost Fallacy | Confirmation Bias | |
|---|---|---|
| Which half of the decision it corrupts | The cost side, by counting money already gone | The evidence side, by filtering what gets noticed |
| Needs a prior outlay? | Yes, with nothing spent there is nothing to honor | No, an inherited project or a stated opinion is enough |
| Needs a prior belief? | No, the spending does the work by itself | Yes, the belief comes first and steers the search |
| What unambiguous evidence does to it | Nothing, the pull survives a clear verdict | Disarms it, since there is nothing left to filter |
| Corrective question | Would I fund this today from a standing start? | What result would make me stop, and did I go looking for it? |
| Fix that actually bites | Hand the call to someone with no stake in the spending | Assign someone to build the case against |
| Signature in the record | More money committed right after bad news | Only supportive evidence ever gets collected or recalled |
Cut the spending or cut the ambiguity, and only one engine is still turning
Two manipulations separate these cleanly, and each one kills exactly one of them. Start by cutting the outlay. An incoming editor inherits a print run they never paid for, having argued in public that students still want paper. Ambiguous sales figures come in and they read every one as a sign of loyal demand. Nothing was sunk by that editor, so what is running is the belief filtering the evidence. Now cut the ambiguity instead. Tell the original editor flatly that the printer closes next month and the replacement quote is double, a fact with no room for interpretation. If the pull to continue survives that, it is not coming from a reading of the evidence at all, it is coming from what has already gone in. The diagnostic difference is what each one touches. Confirmation bias distorts which facts arrive on the desk, so you catch it by auditing what was searched for and what was ignored. The sunk cost fallacy leaves the facts intact and distorts which figures get admitted into the comparison, so you catch it by asking what would happen if the past spending were erased from the file. Pair it with /glossary/overconfidence-bias and the filtering gets worse, because a confident forecaster searches less.
The school paper shows exactly how much optimism the bias has to manufacture
Put numbers on it and the two engines stop blurring. A club has spent 210 dollars on layout software and a banner, then 96 dollars printing 40 copies at 2 dollars 40 cents each. Twelve copies sold at 3 dollars, bringing in 36 dollars. The live question is whether to print another 40. That run costs 96 dollars again, and at the observed rate it returns 36 dollars, losing 60. For the next run to break even the club needs 32 of the 40 copies sold, because 32 times 3 dollars is 96. So the belief required to justify printing is not mild optimism about a 12 copy result. It is a jump from 12 to 32, close to a tripling, and confirmation bias has to manufacture the whole gap by treating one enthusiastic reader as a trend and the unsold 28 as bad weather. The sunk cost engine needs none of that arithmetic. It accepts 12 as the honest rate and still argues for printing, because 306 dollars is already committed and stopping makes that number feel wasted. One engine rewrites the forecast, the other rewrites which costs count, and a club can be running both at once.
Frequently asked questions
Are the sunk cost fallacy and confirmation bias the same thing?
Escalation of commitment usually runs on both, which is why they get merged, but they enter through different doors. The sunk cost fallacy pulls unrecoverable spending into a comparison where it does not belong. Confirmation bias distorts the forecast by collecting and remembering supportive evidence while discounting the rest. One can operate with a perfectly accurate forecast, the other with no money spent at all, so neither is a special case of the other.
How can you tell which one is driving a decision to continue?
Change one input at a time. Remove the past spending from the story, give the project to someone who inherited it, and see whether the enthusiasm survives, which points to a belief problem. Alternatively remove the ambiguity by supplying a verdict nobody can reinterpret, and see whether the urge to continue holds anyway, which points to the spending. Whichever manipulation fails to change the decision names the engine that is actually turning.
Does bringing in an outsider fix both biases?
An outsider helps with each one for a different reason, so the brief matters. Against sunk costs, what you need is someone with no stake in the money already spent, asked one question: would you start this today at these prices? Against confirmation bias, what you need is someone assigned to argue the opposite case and to name in advance the result that would end the project. A reviewer who only rechecks the arithmetic catches neither.
Related comparisons
Get AP Econ exam tips in your inbox
Occasional emails with study tips, new interactive graphs, and exam-season reminders. Free, no spam.
No spam. Unsubscribe anytime. Read our privacy policy.
Keep track of what you have studied
A free EconLearn account adds progress tracking, your quiz history, and achievements. Studying here is free either way, and there is nothing to pay for as a student.
Create a free accountAlready have one? Sign in
Last updated