The Sunk Cost Fallacy: Why Smart People Keep Paying for Bad Decisions

Of all the reasoning traps in this series, sunk cost is the one I’d most like every reader to catch in themselves. I think it drains more money, time and careers than any other mistake smart people make with their eyes open.

What is the sunk cost fallacy?

A sunk cost is any resource you’ve already spent and can’t get back. Money, hours, reputation, years.

The fallacy is letting that spent resource influence what you do next. The correct question is always what the next dollar or the next hour buys. The question people ask instead is what the last hundred dollars or the last five years would have been for.

Economists have been clear about this for a long time and it changes nothing, because the fallacy doesn’t live in the reasoning. It lives in the part of a person that can’t stand waste.

Consider a project eighteen months in and clearly failing. The rational analysis looks only forward. What does finishing cost from here, and what’s finishing worth. The eighteen months appear nowhere in that calculation, and in the meeting they’re the only thing anyone talks about.

That meeting is a waste of everyone’s time. Everyone at the table can read the forward numbers, and the conversation keeps circling back to what’s already gone because nobody wants to be the one who says it out loud. The people who pay for that silence are the ones assigned to the doomed project for another year.

Why do intelligent people fall for sunk cost?

Because intelligence isn’t the faculty being tested.

Loss aversion is what drives it. (More on Kahneman’s Thinking, Fast and Slow) People feel a loss roughly twice as strongly as they feel an equivalent gain, so abandoning something converts a vague ongoing situation into a definite loss with a number attached. Continuing keeps the loss theoretical.

Public commitment adds to it. Somebody argued for this project, staffed it, defended the budget. Stopping it means standing in front of the people who watched all that and saying the money is gone.

Identity comes in on top of those. A person who has spent nine years becoming something doesn’t experience quitting as a decision about the next nine years. They experience it as a statement about who they’ve been.

None of that makes a person stupid. Ordinary human responses produce an expensive answer, and I think the organizations that punish people for admitting a loss carry most of the blame, because they make continuing the safest move for everyone in the room.

How does sunk cost show up in business decisions?

Four places, and every one of them is expensive.

Failing projects get finished because stopping would waste what was already spent. The waste already happened. Finishing adds to it. Bad hires stay because of what recruiting and training cost. The training is gone whether the person stays or leaves, and every additional month is a fresh decision to keep paying.

Software gets maintained past the point of sense because the build was expensive. The build is done. The only live question is what maintaining it costs against what replacing it costs.

Vendor relationships continue because of the integration work already completed. That work is spent. The switching cost is real and belongs in the analysis. The original integration doesn’t.

The pattern in all four is the same. A number from the past is being entered into a calculation about the future, where it doesn’t belong.

What is the difference between sunk cost and persistence?

Separating the two is the hard part, and anyone who says the distinction is obvious hasn’t run anything.

Persistence and the sunk cost fallacy look identical from the outside. Both involve continuing when things are difficult. Both feel like commitment. Plenty of successful ventures survived because somebody refused to quit during a period that looked exactly like failure. The distinction lives in the reason. The behavior looks the same either way.

Persistence says the plan is still sound, the conditions still favor it, and the evidence hasn’t changed. Sunk cost says look how much we’ve put in.

A useful test: if this project landed on your desk today, with the current state and no history, would you fund it. If yes, continuing is persistence. If no, the only thing holding it up is the money already spent. Most people can’t run that test on their own work. Somebody else has to ask.

How do writers use the sunk cost fallacy in fiction?

The sunk cost fallacy is among the most useful engines available for a character who needs to make a bad decision the reader still believes. Readers reject characters who behave stupidly for no reason. They accept characters trapped by reasoning that feels correct from inside, and sunk cost feels extremely correct from inside.

A detective who has built a case against the wrong suspect can’t easily abandon it, because eight months of work argue against every piece of contrary evidence. He’s being human, and the reader can see the trap closing while the character cannot. I’d take that character over one the author made dumb for the plot’s convenience every time, and writers who lean on plain stupidity to move a story are cheating their readers.

The same engine drives a marriage nobody wants to leave after twenty years, a business that swallows a family’s savings one decision at a time, and a soldier who can’t accept that the position is lost because of what taking it cost.

What makes it work in fiction is that the audience adds up what the character won’t. That gap is tension, and it costs nothing to build.

How do you write a character escaping a sunk cost trap?

The escape has to cost something visible, or the reader won’t believe it.

A character who simply realizes the truth and walks away hasn’t escaped anything. Realization is cheap. The trap demands that the character accept the loss publicly, in front of people who watched them commit. Write the moment where they say the number out loud. Eight months. Nine years. Every dollar we had. The saying is the price, and it’s what makes the exit land. Then let the loss stay lost. A story that rewards the character with immediate compensation has implied that leaving was free, and leaving is never free.

What question breaks the sunk cost fallacy?

One question, and it works on projects, relationships, hires, and drafts.

If I were starting from here today, knowing nothing about what it took to get here, what would I do next?

The question strips the past out of the calculation. That’s where the past belongs. Everything already spent stays spent regardless of the answer. Writers need this question more than most. Authors spend years polishing a manuscript that stopped working in chapter three, because walking away would mean the years were wasted. Those years are gone whatever they decide, and I’d much prefer to see them start the next book than bury another year in the last one.

Two things make it work in practice. Ask it out loud to somebody who didn’t make the original decision, because the person who made it can’t hear their own answer. And ask it on a schedule and not in a crisis, since a project reviewed every quarter never accumulates the emotional weight that makes stopping unthinkable.

The fallacy grows with time and silence, and a regular review starves it. I think any leader who lets a failing project run another year to dodge an awkward meeting owes the people doing the work an apology, because they’re the ones spending their days on something nobody believes in anymore.

Frequently Asked Questions