The Sunk Cost Trap: Why You Keep Investing in Losing Positions

You already know the rule. The money you’ve spent is gone, the time is gone, and nothing you do next brings any of it back, so the only thing worth weighing is whether the future spend earns its future return. You could explain that to anyone. And you’ll still sit in a review eighteen months into a project that was meant to take twelve, listening to the room argue about how to recover what’s already been put in, and feel the pull yourself.

That’s the part the textbook version doesn’t prepare you for. Knowing the rule doesn’t protect you, because the thing keeping you in the losing position was never really the arithmetic. The number on the page is just the cover story. What’s actually holding you there is that you chose this path, people watched you choose it, and stopping now means standing up and saying the choice was wrong. That’s a different kind of cost, and it doesn’t show up in the budget.

The evidence

Start with how clean the basic effect is, because it’s among the most reliably reproduced findings in decision research. Give two groups the same theatre tickets, but charge one group full price and let the other have them at a steep discount, and the full-price group attends more shows, including ones they don’t enjoy, purely to feel they got their money’s worth. Same seats, same plays, same value going forward. The only thing that differs is what was paid in the past, and the past spend should be irrelevant. It shapes the choice anyway. This is the sunk cost effect, and it holds across food people keep eating because they paid for it, investments they hold because they’re down, and projects they refuse to kill in rough proportion to what’s already gone in.

So far that’s just a quirk you could correct with a moment’s thought. The reason it survives in senior people, where it does the most damage, is a second finding that sits underneath it. When researchers tracked who escalates hardest into a failing course of action, the worst offenders weren’t the people with the most money at stake. They were the people who’d personally made the original call. Hand someone a struggling project they inherited, and they’ll cut it with a clear head. Hand the same project to the person who launched it, and they’ll pour in more. This is escalation of commitment, and it follows authorship rather than the money. When you stop, you write off more than the spend, you write off the judgment of the person who approved it, and that person is usually the one in the room with the power to stop.

How it works

The engine under all of this is loss aversion: a loss of a given size hurts more than a gain of the same size feels good. While the project runs, the money already spent sits in a kind of limbo. It isn’t a loss yet, because the story isn’t finished and you can still tell yourself it might come good. The instant you stop, that limbo collapses and the loss becomes real and final, with a date on it. Continuing is partly a way of postponing that moment. You keep the account open because closing it means booking the loss, and a part of you would rather carry an unresolved bet than a settled defeat.

Layer the authorship problem on top and you get something stronger than a simple miscalculation. To stop is to declare that your earlier self was wrong, in front of the people who backed you. So you reach for reasons to keep going, and once you’re looking for them, you find them. The small good signals get amplified into proof it’s turning around. The bad ones get filed as temporary noise. That’s confirmation bias riding shotgun: the sunk cost keeps you committed, and the selective reading keeps the commitment feeling justified. The longer it runs, the more the story hardens, because every month you continue is another month of choices you’d now have to disown.

Underneath the budget argument is a harder one: stopping means admitting, in front of the people who watched you choose, that you chose wrong.

How to use it

The basic move is the reset: write down what you’ve put in, cover it with your hand, and ask whether you’d choose this path fresh today with none of it invested. Run that first. But notice what it asks of you. It asks the person who is most personally exposed to judge their own call honestly, and that’s exactly the person the evidence says can’t. So when the stakes are real, do two more things the question on its own won’t do.

First, take yourself out of the chair. Don’t ask whether you’d start this today. Build the case from today forward, in writing, with no history in it: here’s the remaining spend, here’s the expected return, here’s the chance it works. Then hand that document to someone who had nothing to do with the original decision and ask them, cold, whether they’d fund it as a new request. Their answer isn’t tangled up in defending a past call, so it’s the closest thing you have to the truth. If you can’t bring in an outsider, write the case as if you were inheriting the project from someone who’d just left. The inheritor in you decides more cleanly than the author does.

Second, treat the language as a smoke alarm. “We’ve come too far to stop now.” “We can’t waste everything we’ve already put in.” “We just need one more quarter to get there.” None of those are arguments about future return. Every one of them is pointing backward, at the spend, which is the one thing that can’t matter. When you hear it from the room, or catch it in your own mouth, name it out loud as what it is. Saying “that’s a sunk cost argument” doesn’t settle the decision, but it drags the real basis of the resistance into the open, where it can’t keep working on the room unnoticed.

Why it matters

The reason this one is worth the effort is that the costliest version of the trap doesn’t look like a single bad meeting. It looks like a strategy defended for three years past its point, a product line propped up because killing it would mean admitting the launch was a mistake, a senior hire kept on long after it stopped working because letting them go would reopen the question of why you brought them in. In each case the spend was never the real anchor. The anchor was that someone in charge had attached their name to the call and couldn’t afford, in front of everyone, for it to have been wrong.

Which means the rarest and most valuable thing isn’t the discipline to push through adversity. Most cultures already reward that, often to a fault, treating persistence as character and quitting as weakness. The harder skill is the reverse: looking at something you championed, accepting that the past is gone no matter what you do next, and stopping it without needing the stop to be framed as anything other than the right call now. The reset question gets you to the answer. Getting yourself, or the person who made the original bet, willing to act on it is the part that takes a structure, not just a moment of clarity.