The Sunk Cost of a Long Relationship
Twenty years is not a reason to stay another twenty.
“We cannot change the cards we are dealt, just how we play the hand.”
— Randy Pausch
When people explain why they have stayed in a relationship that finished some time ago, the reason offered most often is the length of it. Twenty five years. Three children. A house. A shared history that nobody else has. Every one of those is real and none of them is an argument about the future, which is the only thing a decision can actually affect. Three children, a house and a history nobody else has are all real and none is forward looking.
This is the sunk cost fallacy, familiar from economics and rarely applied to the place where it does the most damage. Money already spent should not influence whether to spend more, because it is gone either way. Years already spent should not influence whether to spend more, for exactly the same reason. Understanding that does not make the decision easy, and it does remove one argument that was never valid. Economics named this decades ago and almost nobody applies it here.
Why the Years Feel Like an Argument
Human beings are strongly averse to waste, and ending something after two decades feels like declaring those decades wasted. That feeling is powerful and it rests on a mistake. The years were not an investment awaiting a return; they were the life you were living at the time, and they were good or bad on their own terms regardless of what happens next. Declaring two decades wasted is what the feeling insists upon.
A relationship that was genuinely good for fifteen years was good for fifteen years. That is not cancelled by an ending, any more than a job you enjoyed becomes retrospectively worthless when you leave. The framing of waste imports a logic from investments that does not apply to lived experience, and it keeps a very large number of people in situations they would not otherwise choose. A job you enjoyed does not become worthless when you leave it.
What the Question Actually Is
The only question a decision can address is what the next period should look like. Given where things actually are, and given a realistic assessment of whether they can change, is continuing better than the alternative. That is a forward looking question and the twenty years contribute nothing to answering it except as evidence about what the pattern has been. History as evidence is legitimate. History as investment is the fallacy.
That last point is the one exception worth noting. History is useful as data: a couple who have tried repeatedly to change something and failed have learned something genuine about the likelihood of change. But that is history as evidence, which is legitimate, rather than history as investment, which is the fallacy. The two are easily confused and produce opposite conclusions. Repeated failed attempts at change are genuine information about likelihood.
The Other Costs That Are Not Sunk
It is worth being clear that not every reason for staying is a fallacy. Financial consequences are real and forward looking. Effects on children are real, though the evidence suggests conflict matters more than separation. Practical enormity is real. Fear of being alone is a genuine consideration about the future rather than about the past. Conflict appears to matter more than separation where children are concerned.
These deserve to be weighed properly rather than dismissed, and the useful discipline is to separate them from the sunk cost. Somebody staying because they cannot afford to leave is making a coherent decision about their circumstances. Somebody staying because of the years already spent is making no decision at all, and the difference matters considerably for what happens next. Constraints are worth naming as constraints rather than as doubts.
Asking It the Other Way Round
A useful reframing removes the history from the question. If you met this person today, in these circumstances, with this relationship as it currently is, would you choose it. The answer is frequently instant and frequently uncomfortable, and it strips out the accumulated weight that makes the ordinary version so hard to answer. Removing the history is the entire trick of the question.
It is not a decision procedure and it is a diagnostic. Somebody who answers yes immediately has learned something reassuring. Somebody who answers no and then produces a list of reasons why the question is unfair has also learned something, since the reasons are usually about the past or about fear rather than about the relationship. Neither answer settles anything and both are considerably more informative than another year of wondering. Neither answer is binding on anything at all.
💡 Key Takeaway
The length of a relationship is the most commonly cited reason for continuing and is not an argument about the future. Years spent were the life you were living, not an investment awaiting a return. History is legitimate as evidence about whether change is likely, and not as a reason in itself. Asking whether you would choose it today is unusually revealing.