From one wage to many
In the last guide we built a market for a *single* kind of labour: a wage is just a price, set where the demand for workers meets their supply. That gave us a clean story — but it also flattened the world. In reality there is no single labour market; there are thousands, one for cardiac surgeons, one for warehouse pickers, one for violinists. Each has its own supply and its own demand, so each clears at its own wage. The puzzle of *why pay differs* is really the puzzle of why those separate markets land in such different places.
Keep one anchor from last time: an employer's willingness to pay for a worker traces back to that worker's marginal revenue product — the extra revenue one more hour of their work brings in. A wage can sit high either because the *demand* side is fierce (each worker generates a lot, or there are few who can do it) or because the *supply* side is thin (few people are willing or able to do the job). Most of what follows is just naming the forces that push these two curves apart from market to market.
Human capital: paying for what you can do
The single biggest, most studied reason wages differ is human capital — the stock of skills, knowledge, and experience a worker carries. The phrase is deliberate: just as a firm invests in a machine that will earn returns for years, a person invests in schooling, training, and practice that raise what an employer will pay. A trained nurse can do things an untrained helper cannot, so her marginal revenue product is higher, so her wage is higher. Pay differences across jobs are, to a large extent, differences in accumulated human capital.
Treating education as an *investment* explains a puzzle: why does anyone pay tuition and forgo years of earnings? Because the higher wage that follows is a stream of future money, and we can value it the way we value any future payment. Suppose a degree costs 80,000 in fees plus forgone pay, but lifts your salary by 8,000 a year for the rest of your career. Discount that long stream back to today — exactly the present value tool from the finance rung — and if its value today beats the 80,000 outlay, the investment pays. That is the cool arithmetic behind a deeply personal choice.
Compensating differentials, scarce talent, and discrimination
Skill is not the only thing wages reward. Two jobs needing identical training can still pay differently if one is nastier, riskier, or less pleasant. To lure people into the unpleasant one, it must offer extra pay — a compensating differential. The night-shift oil-rig welder, the deep-sea diver, the sewage technician are paid partly for the danger and discomfort, not just the skill. Notice this works through *supply*: fewer people willingly offer their labour to a grim job, the labour-supply curve sits higher, and the compensating differential is the gap that clears it. (Genuinely fun jobs run the other way — they can pay *less*, because people queue to do them.)
Then there is raw talent. A handful of people can do something almost no one else can — sing like that, hit a ball like that, prove theorems like that — and when technology lets one performer reach millions at once, the few at the very top capture enormous pay. The wage here is far above what is needed just to keep them working; that excess is economic rent, a payment owed to scarcity itself rather than to any cost they bear. Superstar pay is the loudest example, but the same logic gives a modest rent to anyone with a hard-to-replace knack.
Finally, the uncomfortable one: discrimination. Two workers with the same skill, doing the same job equally well, are sometimes paid differently because of race, sex, or other traits irrelevant to the work. Economists distinguish *taste-based* discrimination — an employer simply preferring one group — from *statistical* discrimination, where an employer who cannot observe a worker's true ability leans on group averages, which is still harmful and still unjust to the individual. Markets push against taste-based bias (a rival who hires the underpaid talent profits), but the push is weak and slow, and decades of evidence show real, persistent gaps remain after skills are accounted for.
The skill premium and why it grew
Across many rich countries, the gap between what high-skill and low-skill workers earn — the skill premium — has widened markedly since roughly the 1980s. A degree was always worth something; over those decades it grew worth a great deal more, relative to no degree. Why? The leading explanation is *skill-biased technological change*: computers, software, and automation are complements to skilled, abstract work (they make a good analyst far more productive) but substitutes for routine work (they replace the clerk, the line operator). When demand for skill rises while demand for routine labour falls, the price of skill climbs.
Why the skill premium widens: a supply-and-demand story SKILLED workers LOW-SKILL / ROUTINE workers ---------------- --------------------------- tech is a COMPLEMENT tech is a SUBSTITUTE -> demand for them RISES -> demand for them FALLS -> wage pushed UP -> wage pushed DOWN skill premium = skilled wage - low-skill wage (widens) Also at work, pulling the OTHER way: globalization, weaker unions, the supply of graduates, and minimum-wage / institution changes. Shares disputed.
Be honest that technology is not the lone culprit. Globalization moved routine production toward lower-wage countries, trimming demand for routine workers at home. The decline of the bargaining power once held by unions, the eroding real value of the minimum wage in some places, and shifts in the *supply* of graduates all matter too. Economists broadly agree the skill premium rose; they argue, often sharply, over how much weight each cause deserves. Anyone who tells you it is *all* technology, or *all* trade, or *all* policy is selling a story simpler than the evidence.
Productivity, luck, and power: how much is really earned?
The tidy version says wages equal what you add: pay tracks your marginal product, so everyone gets what they contribute. There is real truth in it — across millions of workers, more productive ones do tend to be paid more, and the model predicts a great deal. But it would be dishonest to stop there, because at least three things drive a wedge between what you produce and what you take home.
First, luck. The country you were born in, the family that funded your schooling, the boom or bust the year you graduated, being in the right team when it took off — these swamp individual effort more than we like to admit. Much of what looks like a reward for talent is partly a reward for circumstance. Second, bargaining power. When one employer dominates a local labour market — a monopsony, the buyer-side mirror of monopoly — workers have nowhere else to go, and pay can sit *below* marginal product. The same logic in reverse explains why unions, or a binding minimum wage, can sometimes raise pay without killing jobs: they offset employer power rather than overriding a perfect market.
Third, measurement honesty. For a top executive, a star trader, or a viral creator, individual marginal product is fiendishly hard to measure, which leaves wide room for pay to reflect rent and clout rather than crisp contribution. The grown-up conclusion is a balance: productivity *is* a powerful force on wages and the supply-and-demand frame explains a great deal — but it is not the whole truth, and treating every paycheck as a precise verdict on a person's worth confuses a useful model with a moral scoreboard. We will pick this thread up directly when we measure inequality in the next guides.