The old answer and the shock
The textbook model is simple: a price floor above the market-clearing wage reduces the quantity of labour demanded. Raise the minimum wage enough and some workers price themselves out of jobs. For decades that was taught as settled.
In 1994 David Card and Alan Krueger compared fast-food restaurants in New Jersey, which raised its minimum wage, with neighbouring counties in Pennsylvania, which did not. Employment did not fall. The paper was fiercely attacked and widely replicated, and it helped move the empirical consensus.
The modern version of the argument is not whether the effect is exactly zero but how it varies: small increases relative to local wages appear to have little employment effect, while larger ones — and those far above the local median — do.
What the better data show
Cengiz, Dube, Lindner and Zipperer (2019) used a continuous measure of how much each wage bracket was affected and found that jobs paying below the new floor were lost, but the total number of low-wage jobs was roughly unchanged: employers reshuffled hours and pay rather than cutting headcount.
The Congressional Budget Office's 2019 assessment of a $15 federal floor projected both effects at once — raises for tens of millions of workers, and a loss of jobs in the low hundreds of thousands, concentrated among the least skilled. Both halves of that projection are usually quoted alone.
Seattle's 2017 study found hours fell when the wage rose to $13, while a later reanalysis of the same data found smaller losses; the disagreement is about measurement and specification, not about ideology.
Settled: a moderate increase, in a strong labour market, does not produce the large job losses the simple model predicts, and it raises pay for people who keep their jobs.
Contested: the effect of large increases, the effect in weak labour markets and rural areas, and the long-run effect on automation and on prices.
Worth noting: the policy question is distributional as much as aggregate. Even a policy that costs a few thousand jobs may be judged worth it if it raises the incomes of millions — that is a political choice, not an economic finding.
The archive labels this live debate: the naive version (any minimum wage destroys jobs) is dead, and the strong version (minimum wages have no costs) is not established either.
Why it is in the archive
Because it is the clearest case in economics where a famous prediction failed a natural experiment, and where the failure was absorbed rather than denied — which is how a discipline is supposed to work.
Related: universal basic income, technological unemployment.
Sources for this page
- Wikipedia, “Minimum wage” — theory and evidenceen.wikipedia.org
- Cengiz, Dube, Lindner & Zipperer (2019), QJE — the effect of minimum wages on low-wage jobsdoi.org
- Card & Krueger (1994), NBER working paper 4509 — the New Jersey fast-food studywww.nber.org
- Congressional Budget Office (2019), the effects of a $15 federal minimum wagewww.cbo.gov