Sunk cost in careers

"I've already put eight years into this" is the most common sentence in career indecision, and it is doing something strange. It is offering the past as evidence about the future.

Economically it should carry no weight at all. A cost that cannot be recovered is irrelevant to what you should do next; only the remaining costs and benefits count. Psychologically it carries enormous weight, and the research on why is solid. But the careless version of this advice — ignore the sunk cost, the years are gone — is also wrong when applied to a career, because unlike a theatre ticket, eight years of work leaves behind things that are still assets. The useful skill is telling the two apart, and that is what this entry is for.

Hal Arkes and Catherine Blumer demonstrated it cleanly in 1985 with season theatre tickets. Buyers who had randomly received a discount attended fewer plays than those who paid full price. The plays were identical; the only difference was how much had already been spent, which cannot logically affect whether tonight's play is worth going to. It affected it anyway.

Their other scenarios have the same structure. People continue a doomed project because a large share of the budget is already committed, and abandon an identical one where it is not. The pattern is robust and easy to replicate.

Two mechanisms are generally offered. Loss aversion, from Kahneman and Tversky's prospect theory: closing out a position makes the loss real and final, while continuing keeps it notional and preserves the chance it turns around. And self-justification: stopping is a public admission that the earlier decision was wrong, so continuing protects the decision-maker rather than the decision.

Barry Staw's 1976 study — Knee-deep in the big muddy — put the second on the map. Business students allocating research funding put more money into a division that had performed badly when they had personally made the original allocation than when someone else had. Being the author of the earlier choice was what predicted throwing good money after bad. Sleesman and colleagues meta-analysed decades of this work in 2012 and found the determinants sorting into recognisable families: characteristics of the decision-maker, of the project, of the social setting, and of the situation. Escalation is strongest where identity and visibility are involved — which is precisely the position you are in with your own career.

Here is the correction that most sunk-cost advice skips.

Genuinely sunk: the time, the tuition, the opportunity cost of the roles you did not take, and the self-image you built around the choice. None of these come back whatever you do next, and none of them are arguments for continuing.

Not sunk, and wrongly filed as sunk by people in a hurry to be rational: transferable skill, which is an asset you still hold and which changes the economics of the next move; relationships and reputation, which are usually more portable than they feel and less portable than optimists assume; credentials with real gate value, where the licence or qualification opens doors that are genuinely closed otherwise; and domain knowledge, which is often worth more one step sideways than it is worth where it was earned.

So the question is not "should I ignore the eight years?" It is: which parts of the eight years are still on the balance sheet, and do they point where I am going? A doctor moving into health policy is not discarding a decade; they are deploying it. A doctor moving into landscape gardening is, and that can still be the right call — it is just an honest one rather than a reframe.

The reliably bad reasoning is the pure form: I must continue because I have continued. The tell is that the argument never names a benefit that lies in the future.

Three features make career decisions worse than the laboratory version.

The decision-maker is the author. Staw's central finding. Nobody else chose your degree, and stopping means revising a story you have told for years — including to the people who paid for it, which is the second feature.

The audience is real. Escalation intensifies under observation. A career change is watched by parents, partners, colleagues and a professional network that has categorised you. The cost of visible reversal is not imaginary, and pretending it is does not help anyone decide.

Identity is inside the position. This is why the same argument is so much harder here than in a failing investment. When the course of action is who you are, abandoning it reads as self-annihilation rather than reallocation, which is the machinery the work and identity entry describes. Escalation of commitment and identity fusion are the same problem wearing two vocabularies.

One more thing is true and should be said plainly: sometimes persistence is correct. Most careers contain a dismal stretch in years two to five where the work is hard and the competence is not yet built. The evidence in that case is not "I've invested a lot" — it is "the curve is still moving." Mistaking a competence valley for a wrong career is a real and expensive error in the other direction.

Four, in the order that works.

The stranger test. Someone with your exact skills, savings and obligations arrives today with no history in this field and is offered your job. Do they take it? This removes the sunk cost while keeping everything real.

The forward accounting. From here, what does staying cost and pay over three years, and the alternative? If you cannot answer without referring to the past, the argument for staying is sunk cost in a costume.

The direction of the curve. Is it getting better or have you been saying "next year" for four years? A valley has a floor. A decline does not, and the difference is visible in the last twelve months rather than in your feelings about them.

The named benefit. Ask what specifically improves by continuing. If the only honest answer is "I avoid admitting it was wrong," that is not a benefit; it is the fallacy identifying itself.

A closing correction to the tone of most advice on this. The years are not wasted merely because you leave — they produced the person who now knows this is not it, and that is information nobody gets for free. The Career Values test reads what you are actually optimising for, which is the input the forward accounting needs, and the Career Interest Match is worth taking before rather than after the decision. If what is keeping you in place is dread of the conversation rather than the maths, personal boundaries is a more relevant entry than this one.