My parents divorced when I was fourteen. Among the things that experience taught me early: that people do not always behave rationally even when they can see clearly what is happening, that knowing something and acting on it are separated by a distance that logic alone cannot close, was the observation that my father stayed in a job he openly disliked for eleven years after he could have left.
He was not trapped financially by the point I was a teenager. He was trapped by something else, something he could not name and therefore could not move against.
I spent a significant portion of my career in digital marketing watching the same pattern play out at scale. Smart, capable people: people who could diagnose exactly what was wrong with their situation and articulate it with precision, staying put in roles that were making them quietly miserable, sometimes for years.
When I started writing seriously about behavioural science, it was one of the first things I wanted to understand properly, because the standard explanation — inertia, fear, laziness — had never seemed adequate to me. It was too vague to be useful and too blaming to be accurate.
The psychology underneath it is considerably more specific than inertia. And specific problems, unlike vague ones, have traction.
The sunk cost fallacy is doing more damage than most people realise
The sunk cost fallacy is one of the most well-documented phenomena in behavioural economics, established through decades of research by Daniel Kahneman and Amos Tversky in their work on how humans make decisions under conditions of loss.
The core finding is this: people systematically overweight what they have already invested — time, money, effort, identity — when making decisions about whether to continue a course of action. The rational approach, as any economist will tell you, is to treat sunk costs as irrelevant to future decisions. What you have already spent cannot be recovered. Only what lies ahead should factor into the choice.
Human beings do not do this. We treat past investment as a reason to continue regardless of whether continuation serves us, because stopping feels like admitting that the investment was wasted.
The mind reframes exit as loss rather than as course correction, and loss aversion — another of Kahneman and Tversky’s foundational findings, which showed that losses feel roughly twice as painful as equivalent gains feel good — means that the perceived pain of leaving outweighs the anticipated relief, even when the rational calculation clearly favours going.
Apply this to someone who has spent eight years building expertise in a field they no longer care about, who has accumulated a salary level that took time to reach, who has professional relationships and a title and an identity that is legible to the people around them. Leaving is not just a logistical decision. It is a decision that the mind frames as surrendering all of that — as losing, rather than as redirecting.
And the nervous system treats losing as something to be avoided with considerable force.
Why identity is the part that gets least attention
The sunk cost mechanism is well known enough that most people who have read anything about behavioural economics can name it. What gets considerably less attention — and what I think is the more operationally important factor in why people stay — is the degree to which professional identity becomes fused with self-concept in ways that make exit feel like a kind of self-erasure.
This is not metaphorical. Research on self-concept and role identity consistently finds that the longer a person occupies a professional role, the more deeply that role becomes integrated into their sense of who they are at a foundational level.
After eight or ten or twelve years in a particular kind of work, the question “what do I do?” and the question “who am I?” have, for many people, become functionally the same question. Leaving the job does not feel like changing what you do. It feels like changing what you are.
I spent over a decade in digital marketing, and when I finally admitted to myself that writing was what I had wanted to do all along and began making the actual move toward it, the resistance I encountered was not primarily practical.
The practical obstacles were real but manageable. What was harder was the sustained period of not knowing how to answer the question of who I was in the gap between identities — between the person the career had made me and the person the writing might eventually make me. That gap is genuinely uncomfortable in a way that most career advice dramatically underestimates.
The person who stays in the job they hate is often not staying because they have not thought about leaving. They are staying because leaving requires them to inhabit that gap — to be, for a period of uncertain length, someone whose identity is not yet formed enough to be explained to other people. That is a specific kind of psychological discomfort, and it is fixable, but only if it gets named as what it is.
The present bias problem
There is a third mechanism operating alongside the sunk cost effect and the identity fusion, and it is the one that explains the timing problem — why people who have decided to leave still do not leave for months or years afterward.
Behavioural economists call it present bias: the documented tendency to overweight immediate costs relative to future benefits, even when the future benefits are large and the immediate costs are comparatively small.
In the context of staying in a bad job, present bias means that the discomfort of the job search, the awkwardness of resigning, the uncertainty of the transition, and the temporary income disruption all feel larger and more immediate than the benefits of eventually being in a better situation — even though the rational calculation, extended over time, overwhelmingly favours leaving.
Present bias is why people say “I’ll start looking properly after the holidays” and then find another reason after the holidays, and another reason after that. It is not weakness. It is a well-documented feature of how human temporal perception operates, and it responds predictably to specific interventions: making the future benefit more concrete and immediate, and making the first step smaller and more specific than the generalized decision to “do something about it.”
What actually creates movement
The implication of all three of these mechanisms is that the person stuck in a job they hate does not need to be told to take the leap or reminded that life is short or encouraged to believe in themselves. They need a different set of conditions — conditions that change the psychological calculation rather than simply exhorting them to override it.
For the sunk cost problem: naming it explicitly. Most people feel immediate relief when they understand that their sense of obligation to past investment is a cognitive pattern with a name rather than a rational argument for staying. You are not being irrational for feeling it. You are being human. But the feeling is not evidence about what to do next.
For the identity fusion problem: starting to build the new identity before leaving the old one. Not waiting until after resignation to begin writing, consulting, practising the new skill, spending time in the community of the thing you want to move toward. The gap between identities is less destabilising when there is already something forming on the other side of it.
For the present bias problem: making the first step genuinely small. Not “start a job search” but “update one section of the resume this week.” Not “figure out what you want to do” but “have one conversation with one person working in the field you are curious about.” The brain’s present bias is not defeated by willpower. It is defeated by reducing the immediate cost of the first action until it is smaller than the immediate discomfort of staying still.
None of this is framed to suggest that leaving a bad job is simple, or that the practical constraints people face are not real. They are real, and they matter. But the thing keeping most people in situations that are clearly not working is not a failure of courage or a deficiency of ambition. It is a set of well-documented cognitive patterns that operate reliably on almost everyone and that respond, with more consistency than most people expect, to being understood rather than overridden.
My father eventually left. He was in his mid-fifties by then, and the decade he spent staying was a real cost. What I have come to think, though, is that the cost was not primarily about the time. It was about the years he spent believing that the staying was irrational without having the framework to see that it was, in fact, entirely predictable — and therefore something he could have worked with, rather than simply against.
