Sunk Cost in a Twelve-Year Marriage

"Don't fall for the sunk cost fallacy" is the rationalist's standard advice for someone considering whether to end a long relationship. It's also incomplete and, applied naively, often wrong.

**What sunk cost actually means**

In economics, sunk cost is money or effort already spent that cannot be recovered. The fallacy is to factor that unrecoverable investment into a forward-looking decision. The textbook example: if you've spent $50 on a movie ticket and after 30 minutes the movie is bad, you should leave — the $50 is gone whether you stay or leave. Staying to "get your money's worth" is irrational.

The advice that follows: don't stay in a bad relationship just because you've already invested twelve years. The twelve years are gone whether you stay or leave.

**Where the analogy breaks down**

Two ways. First: a relationship is not a movie. The investment you've made is not "spent and gone." It's stored. Shared history, mutual knowledge, financial entanglement, friend networks, family relationships, children, the accumulated infrastructure of a life together — these are not sunk costs. They are *present assets* that exist now, that you would lose if you leave, and that are not automatically recoverable in a new relationship.

The cost of leaving is not just emotional. It includes losing those assets. That cost is real, and weighing it in your decision is not a fallacy. It's correct reasoning.

Second: relationships have non-linear payoffs. A movie at minute 30 is probably a representative sample of the rest of the movie. A relationship at year twelve is not necessarily a representative sample of years thirteen to fifty. Karney and Bradbury's longitudinal work shows real variability — some couples improve substantially after a difficult middle stretch. Treating year twelve as a fixed forecast for the rest of the relationship's life is statistically wrong.

**The version that is relevant**

The sunk cost fallacy does still apply in one specific way: do not stay because you cannot bear to acknowledge that the previous years were wasted. That feeling — the desperate refusal to admit a long investment didn't pay off — is the genuine fallacy. It pushes people to stay in unmistakably bad relationships out of sheer reluctance to call the years a loss.

The clean separation is: weighing the assets you would lose by leaving is rational; refusing to leave because you can't tolerate the loss to your self-image is not.

**The honest question**

Are you considering staying because the assets you'd lose are genuinely valuable to you and you're choosing them on net, or because you cannot face the identity loss of admitting the relationship didn't work?

The first is legitimate. The second is the fallacy. Most people who probe this honestly can tell which one is operating.

**A practical filter**

Sit with this counterfactual: if you had only been in this relationship for one year, and everything else was the same — same patterns, same incompatibilities, same dynamic — would you stay? If the honest answer is no, your length-of-investment is doing real work in your decision, and it's worth examining whether that work is legitimate (real assets you don't want to lose) or fallacious (identity protection).

**What the literature actually says about long-marriage divorces**

Hetherington's longitudinal data on post-divorce outcomes shows that those who leave long marriages do not, on average, end up worse off five years later — but they go through a hard transition that's longer than the popular narrative suggests, particularly above age 50. The financial cost is real. The social-network disruption is real. Those costs do not make leaving wrong. They make it important to understand what you're choosing.