What makes a job good or bad, and why does society pay some jobs so much more than it honors them?

The number nobody can quite justify

Nearly everyone who works has, at some quiet moment, run the same private calculation. The paycheck arrives, and behind it sits a question that feels as though it should have a clean answer and never quite does: why this number, and not one considerably larger? The work is real. The hours are long. The tasks would not do themselves if you stopped. And yet the figure at the bottom of the slip seems to have been set somewhere far away, by no one you can name, according to a rule no one ever explained. Ask why the job exists at all and the answer is usually clear enough; someone needs the thing done and is willing to pay to have it done. Ask why it pays what it pays, and the ground gets soft.

The soft ground is worth standing on for a whole chapter, because the answer turns out to be the key to most of the rest of this book. The four questions that come later in this volume, about whether a raised minimum costs you your shift, about the gap between your sister’s pay and her male colleagues’, about whether the new arrivals push your wage down, about the chief executive who takes home three hundred times the median, all read differently once the wage-setting question is answered plainly. This chapter sets out that answer, and it runs against the way pay is usually justified. Your pay is not a verdict on your worth. It is the price of an exchange, and like any price it is settled by the position of the two parties who strike it.

Pay is a bargain, and bargains turn on leverage

A wage is what an employer pays to have work done and what a worker accepts to do it, and it settles somewhere in the range between the least the worker would take and the most the employer would give. Where in that range it lands is decided by leverage: by how easily each side could walk away and do something else instead. The two things that move a wage, more than any judgment about the value of the work, are how replaceable the worker is and how credibly the worker can leave. When many people can do your job and would be glad to, an employer who loses you simply hires the next one, and your wage sits near the floor of the range. When few people can do what you do, or when you could plausibly take your labour elsewhere tomorrow, the employer has to bid to keep you, and your wage climbs toward the ceiling. Everything else in this chapter is a variation on those two forces.

This is meant as a description of how wages are set, not a complaint about how they ought to be. Calling pay a function of bargaining position is no more a grievance than calling the tide a function of the moon. An earlier chapter treated your labour as a thing you sell, with the same logic as any other sale (why the party in the stronger position captures the larger share of the value), and the wage is simply the price that sale fetches. It helps to notice what the framing leaves out. It says nothing about whether the work is honourable, needed, or good. A wage measures scarcity and leverage; it was never built to measure worth, and one way to see that is to line the two up and watch them come apart.

Read the figure for its broad shape and it looks reassuring: the least honoured work pays least, the most honoured pays most, and pay and standing seem to travel together. They do, up to a point, because a job that demands rare training tends to earn both respect and money. The revealing part is the vertical spread. Fix the level of standing near the middle of this figure and the pay still ranges over a factor of two. A firefighter and a software developer sit at almost the same rung of public esteem and take home wages that are not close. A financial manager, less esteemed than the nurse who may hold a life in her hands, is paid roughly twice as much. If pay were a reading of how much a job is respected, or of how much it is needed, these points would line up. They do not line up, and the reading this chapter defends is that pay is answering a different question: how hard is this worker to replace, and how easily could this worker leave. A handful of plotted occupations cannot prove that reading; what they show is the gap it has to explain, and the rest of the chapter assembles the evidence. On that reading, the developer and the financial manager have outside options the firefighter and the teacher do not, and the wage records the difference.

The worst work, paid the worst, where the options are fewest

A direct test of whether pay rewards the unpleasantness of work is to look at its most extreme edge, the risk of being killed doing it. Economic theory has a tidy prediction here, old and intuitive: dangerous jobs should pay a premium, a bit extra to compensate for the chance of not coming home, because otherwise no one would take them over the safe alternative. The premium is real in places, and careful studies that hold skill and education constant do find that riskier work pays somewhat more than equivalent safe work. What the raw picture shows, though, before any such adjustment, is how weak the compensation is where it should be strongest, and why.

The pattern is the thesis in miniature. Danger commands a premium only when the dangerous worker has somewhere else to go. A commercial pilot is expensive to train and hard to replace, so an airline that wants pilots has to pay for them, and the pay is high; the danger rides along, but the skill is what is being bought. A logger or a roofer is, in the brutal language of the labour market, easier to replace, and a worker who can be replaced cannot bargain for hazard pay, because the reply to the demand is that someone else will take the risk for the going rate. So the burden of danger falls heaviest on the workers least able to charge for bearing it. The market does not fail to notice the danger. It notices, and it discovers that it does not have to pay for it.

Scarcity made to order

If pay tracks scarcity, then anything that manufactures scarcity raises pay, and the most widespread machine for manufacturing it is the credential. A licence or a required degree does two things at once. It can genuinely certify skill, protecting the public from the incompetent surgeon or the untrained electrician, which is the reason usually given for it. And it limits how many people are allowed to do the work, which raises the wages of those already inside the gate, whether or not the training the gate demands is truly needed for the job. The two effects are hard to separate, and both are always present. This is the puzzle a great many workers feel personally, the requirement to hold a degree for a job that never seems to use it. The degree is partly a signal that its holder can be trained, and partly a fence.

The scale of the fencing has grown remarkably. Around a quarter of American workers now need a state licence to do their jobs, and the share is higher once local and federal licences are counted, up from fewer than 5 percent in the early 1950s. Some of that growth tracks a shift toward work that genuinely warrants certification. A large part of it does not; licences have spread to florists, hair braiders, interior designers, and auctioneers, occupations where the safety case is thin and the effect on the wage is not. Licensed workers earn something on the order of 10 to 15 percent more than otherwise similar workers without the licence, a premium that has little to do with any difference in the work and a great deal to do with the barrier at the door. The credential converts a skill anyone might learn into a right only some are permitted to exercise, and the wage rises to meet the restricted supply. That is scarcity made to order, and it is one of the few forms of leverage a worker can acquire rather than be born into.

Seen this way, a job’s quality is largely a question of how much leverage it hands the person doing it, and leverage has components beyond the wage. Whether the work is secure or can be ended at a week’s notice, whether the schedule is yours or handed to you, whether you are easily monitored or trusted to work unwatched, all of these track the same underlying thing, the balance of power between the worker and the firm. Ownership shifts that balance too, which is why a later question in the framework volume asked what changes when workers own the enterprise (why employee-owned firms tend to offer more stable and more humane jobs, and why they nonetheless stay rare). A good job, in the end, is one where you hold some of the cards. A bad job is one where you hold none, and the pay, the hours, and the dignity all follow from that.

The part of the paycheck you never see

One more feature of pay stays hidden from the person receiving it, and it grows more important the more you earn. What an employer spends to employ you is not the same as the wage that reaches your account. On top of the cash sits a bundle of other costs the employer pays on your behalf, health insurance, retirement contributions, paid leave, and the taxes the law requires employers to pay toward your future benefits. You never negotiated these line by line, and most workers could not say what they are worth, yet together they make up a large and rising slice of the true cost of employing anyone.

The bundle matters for the argument in two ways. It means the true inequality between jobs is wider than the wage gap alone, because the better-paid also collect more of the benefits bundle, from tax-favoured insurance and retirement contributions to paid leave, and more of the security it buys. And it means a real part of your compensation was set collectively and quietly, by an employer’s benefits policy and by the tax code, rather than by anything you bargained for. The worker who imagines the whole of the deal sits on the pay stub is missing as much as a third of it, and the missing third leans toward whoever already had the leverage to command the larger wage.

What the vocabulary is for

The point of settling the wage-setting question first is that the rest of the volume can now be read through it. If pay tracks replaceability and exit rather than worth, then the fights over pay that fill the news are, underneath, fights over leverage. A minimum wage is an attempt to set a floor under the workers with the least of it. The gap between what men and women earn is partly a question of which jobs each ends up in and how the hours are structured, which is to say a question of bargaining position as much as of any single employer’s choice. The worry that immigrants drive down wages is a worry about the supply of replaceable labour. The chief executive’s enormous package is what happens at the far end of the scale, where a worker is treated as nearly irreplaceable and the leverage runs entirely one way. Each of these later chapters returns to the vocabulary set out here, and each is easier to think about clearly once pay is understood as the price of a bargain rather than a measure of desert.

There is one more thread the wage cannot capture at all, and it is the subject of the next chapter. If pay is set by leverage, and leverage has drifted toward employers over the past two generations, then the growth an economy produces need not reach the people who produce it. The paycheck can stall even as the output it comes from keeps rising, and the felt experience of that gap, of working in a richer country every year and not feeling any richer, is where the argument turns next.