The textbook verdict — and why people doubt it
In the last two guides you treated a worker's pay just like any price: the wage is set where the labour market clears, where the number of people willing to work meets the number of hours firms want to hire. Now the government steps in and declares a minimum wage — a legal floor below which no one may be paid. Apply what you already know about a binding floor in any market, and the prediction looks brutal. A price floor set *above* the clearing price creates a surplus: more sellers want in than buyers want to buy. In a labour market, that surplus has a name — unemployment.
The logic flows straight from the demand for labour. Recall that a firm hires one more worker only while that worker's marginal revenue product — the extra revenue the worker brings in — exceeds the wage. Raise the legal wage above some workers' marginal product, and the firm should, in theory, let them go: they now cost more than they add. So the clean diagram predicts that a higher minimum wage prices the least-productive workers out of jobs, hurting precisely the people it was meant to help. For a long time that was the near-consensus, and it is still the right starting intuition. The trouble is that the data have been stubborn.
What the evidence actually shows
Here is where honesty matters more than a tidy story. Starting in the 1990s, economists compared employment in neighbouring areas where one side raised its minimum wage and the other did not — fast-food jobs across a state border, for instance. If the textbook were the whole truth, the high-wage side should have shed jobs. Again and again, the measured job losses came out small, sometimes statistically indistinguishable from zero. This does *not* mean a minimum wage is a free lunch, and it does not mean you can raise it to any level without harm. It means the simple supply-and-demand diagram is missing something about how real labour markets work.
There are several honest ways to reconcile the diagram with the data. Firms can absorb a higher wage through slightly higher prices, lower profits, fewer hours rather than fewer heads, or — strikingly — through lower turnover: better-paid workers quit less, so the firm saves on hiring and training. Each channel softens the blow on employment. But the deepest reason the simple model can mislead is that it assumes the firm is a *wage-taker*, helpless to set pay, just as a competitive seller is a price-taker. Drop that one assumption and the whole prediction can flip. That assumption has a name.
Monopsony: a monopoly on the buying side
You already met monopoly — a single *seller* who can choose its price. Monopsony is its mirror image: a single (or dominant) *buyer* who can choose the wage it pays. The classic picture is a company town — one big factory or hospital that is the only serious employer for miles. Workers cannot easily walk next door, because there is no next door. That gives the employer real market power over wages, the same way a monopolist has power over prices. Pure monopsony is rare, but a milder version — a few large employers, workers who find it costly to move, switch, or retrain — is everywhere, which is exactly why it matters.
Here is the twist that makes minimum wages behave unexpectedly. A monopsonist who wants one more worker usually has to raise the wage to attract them — but it must then pay that higher wage to *everyone*, not just the newcomer (you cannot easily pay the new hire more than loyal old-timers). So the true cost of an extra worker is more than just their wage: it includes the raise given to all the workers already there. Sound familiar? It is the exact mirror of the monopolist's marginal revenue lying *below* price. A monopsonist's marginal cost of labour sits *above* the wage. To keep that cost down, it deliberately hires *too few* workers and pays them *less* than their marginal product.
Hiring one MORE worker when you must raise the wage for ALL wage workers total wage bill marginal cost of labour ---- ------- --------------- ----------------------- $10 3 $30 -- $11 4 $44 +$14 <- not just +$11! $12 5 $60 +$16 $13 6 $78 +$18 Going from 4 to 5 workers: the 5th worker earns $12, but you also raise the other 4 by $1 each (+$4). Marginal cost of labour = 12 + 4 = $16, well above the $12 wage.
Now watch what a minimum wage does to *this* employer. By law, the firm must now pay at least the floor — but for any worker up to that floor, the extra cost of hiring one more is simply the floor itself, no longer the bigger marginal-cost figure, because it can no longer hold the wage down by hiring fewer people. The penalty for expanding has been removed. A well-chosen minimum wage can therefore push a monopsonist to hire *more* workers, not fewer, while also raising their pay. This is the startling result: in a monopsonised labour market, a minimum wage can raise both wages and employment at once — the opposite of the textbook floor.
Unions: workers bargaining as one
If a dominant employer can have power over wages, can workers fight back by gaining power of their own? That is the idea behind a labour union: instead of each worker negotiating alone — weak against a big firm — workers band together and bargain as a single voice, a process called *collective bargaining*. A lone worker who demands more can be replaced; an entire workforce that can withhold its labour all at once, by threatening to strike, holds something the firm cannot easily ignore. In effect, a union tries to create a kind of seller-side market power to offset the buyer-side power on the other end of the table.
When does a union raise wages, and at what cost? The answer depends on what it is bargaining against. Facing a monopsonist, a union can be the second half of a remarkable bargain: just like a minimum wage, it can pull pay up toward workers' true marginal product *and* push employment up — two market powers partly cancelling out, a situation economists call *bilateral monopoly*. But facing an otherwise competitive industry, a union that wins wages above the market rate can shrink employment, just as the textbook floor warned. The same union tactic can help or hurt depending on the structure it pushes against — which is why blanket claims that "unions are good" or "unions kill jobs" are both too crude.
Why this case is the real lesson
Step back and notice what just happened. The supply-and-demand model is one of the most powerful tools you own, and it remains your first instinct for any labour market question — for good reason. But this guide is the place where you learn to ask, before trusting the diagram, *is anyone here a price-taker, or does someone hold power?* The whole textbook prediction rested on a hidden assumption — competitive, wage-taking employers — and when that assumption fails, the conclusion can reverse. A model is only as honest as its assumptions, and the mark of real understanding is knowing when to reach past the first diagram, not just how to draw it.
- Reach for supply and demand first — it is right far more often than not.
- Then check the hidden assumptions: is each side small and powerless, or does one side dominate?
- If one side has market power — a monopsony employer, or a union — redo the analysis, because the simple prediction can flip.
- Finally, let the evidence judge: when careful data disagree with your diagram, fix the model — do not ignore the data.