The Sunk Cost Fallacy: Why You Finish Bad Movies and Stay in Bad Jobs

You're ninety minutes into a three-hour movie and it's bad. Not "give it a chance" bad — actively, structurally bad. You watch the rest anyway. Not because you expect it to improve. Because you've already put in ninety minutes, and walking out would make those ninety minutes "wasted." Congratulations, you've just paid ninety more minutes to protect ninety minutes that were already gone.

That's the sunk cost fallacy: treating money, time, or effort you can't get back as a reason to keep going, instead of judging the decision on what's left to gain. It's one of the most well-documented errors in behavioral economics, and it runs your life more often than you'd like to admit (mine included — I once finished a 900-page biography of a man I'd stopped caring about by page 200, purely out of spite for my own past decisions).

Where the fallacy got its name

Hal Arkes and Catherine Blumer ran the study that put this on the map in 1985. In one experiment, people who'd already paid for a ski trip were more likely to go on it — even after learning a better, more enjoyable trip had become available for the same weekend. The money was gone either way. It shouldn't have mattered which trip they picked. It mattered anyway.

Their broader finding: people don't just weigh future costs and benefits. They weigh how much they've already invested, as if abandoning a project retroactively makes the investment "wasted" — rather than accepting it was already spent the moment it left your hands. Economists have a blunt name for the correct approach: sunk costs are irrelevant to rational decisions. Nobody's brain got that memo.

Why quitting feels like losing twice

The mechanism runs through loss aversion, first formalized by Daniel Kahneman and Amos Tversky in their 1979 prospect theory paper. Losses hurt roughly twice as much as equivalent gains feel good. Walking away from a sunk cost means admitting a loss right now, in a moment you can point to. Continuing lets you postpone that admission indefinitely, even if it guarantees a bigger loss later.

There's also a self-image problem sitting underneath the math. Quitting a project you defended for two years means telling yourself — and everyone who watched you defend it — that you were wrong. Continuing lets you avoid that conversation for one more week. Then another week. This is how "one more week" becomes eighteen months (ask anyone who's ever renovated a bathroom).

Escalation of commitment: the sunk cost fallacy at scale

Give the fallacy a boardroom and a budget, and it grows into something researchers call escalation of commitment. Barry Staw's foundational 1976 study found that decision-makers who caused a failing outcome were more likely to pour additional resources into it than people brought in fresh — specifically because they felt personally responsible for making it work.

The clearest real-world case study is the Concorde. Britain and France kept funding the supersonic jet program for years after internal projections showed it would never turn a profit, partly because too much had already been spent to "waste" by stopping. The plane flew for decades. It also lost money on nearly every route it ever operated. Economists sometimes call this specific pattern the Concorde fallacy, which is either a fitting tribute or the least flattering thing you can name after a supersonic jet.

Corporate versions of this look identical to the ski-trip study, just with more zeros. A failing product line gets another funding round because of the R&D already spent. A doomed merger gets pushed through because of the legal fees already paid. The spent money is gone in both cases — the only question that should matter is whether the next dollar in produces a return, and that question gets buried under the last dollar spent.

My opinion: relationships are where this fallacy does the most damage, and it's not close

Here's the strong claim: sunk cost thinking wastes more human years in relationships than in every bad investment combined, because the "already invested" ledger includes things you can't even put a number on — years, shared friends, a joint lease, the version of yourself you built around being someone's partner. A 2007 study by Sabrina Bruyneel and colleagues found sunk cost effects extend beyond money into emotional and identity investment, which is exactly the territory relationships live in.

"I've put five years into this" is not a reason to put in a sixth. It's a description of the past tense. The only question worth asking is whether year six, evaluated on its own, is worth having — and if the honest answer is no, the five years are already spent whether you leave on a Tuesday or wait for year eleven.

Where this doesn't apply: distinguishing a bad match from a rough patch. Every long relationship, PhD, and business has genuinely hard stretches that resolve with persistence rather than an exit. The fallacy is specifically about decisions driven by what you've already lost, not decisions where the actual forward-looking evidence still says stay. If your honest forecast for what's ahead is good, staying isn't a sunk cost error — it's just staying.

The one question that actually works

Awareness of the fallacy barely helps on its own — plenty of behavioral economists who've published on sunk costs still finish bad books. What works is a specific reframe: ask "if I were starting fresh today, knowing everything I now know, would I choose this path?" Not "have I already put in too much to quit." A different question, a different, more honest answer.

This works because it deletes the sunk cost from the equation by construction. It doesn't ask you to ignore your history — it asks you to evaluate the decision the way an outsider would, someone with no investment to protect and no reputation riding on the outcome. That outsider doesn't care what you already spent. Neither should the version of you making the next choice.

Write the answer down before you talk yourself out of it. If the honest answer is no, the two years, the tuition, the renovation budget — none of it becomes less "wasted" by adding more to the pile. It just becomes a bigger pile.

The money's gone. The time's gone. The only thing actually up for a vote is what happens next — so vote on that, and only that.