Economics focus

Debating the minimum wage
Feb 1st 2001
From The Economist print edition

When governments put a regulatory floor under wages, does that destroy jobs? An update on a long-running dispute.

THE story so far. For many years economists took it for granted that a compulsory minimum wage, set much above the floor that emerges in an unregulated labour market, would reduce employment. Young or unskilled workers would be unable to find work at the mandatory minimum: the bottom of the demand curve would be chopped off. Of course, if the minimum were so low as to impinge on none of the wage bargains actually being struck, there would be no effect. The higher the minimum wage, once it starts to bind, the bigger the loss of jobs.

The possibility that a minimum wage would not reduce employment, and might even increase it, was acknowledged. This could happen through market failure of some kind. Suppose, for instance, employers have monopsony power as buyers of labour: they will curb their demand, to drive wages lower. A minimum wage could remedy this, raising wages and expanding employment at the same time. But in a normal, competitive labour market, this should not arise.

Then along came David Card and Alan Krueger, two of America's most distinguished labour economists. They looked at the effect of a big increase in New Jersey's minimum wage, in 1992, on employment in the local fast-food industry - and they discovered that it pushed employment up. This attracted wide attention. The claim featured prominently in their widely cited book "Myth and Measurement". It influenced the debate over raising the minimum wage in the United States; and when Britain introduced a minimum wage of its own, their findings were cited again.

Next came a paper by David Neumark and William Wascher. They went back to the New Jersey case, using different and (they argued) better data and methods. They found that the rise in the minimum wage had reduced employment, after all, much as one might have expected.

For those who follow this intriguing quarrel, there are new developments to report. In the current issue of the American Economic Review, after the protracted delay normally associated with that esteemed journal,both sides publish revised and polished versions of their earlier positions. They have reviewed each other's work and made adjustments. The difference has narrowed, but remains. In essence, Messrs Card and Krueger defend the validity of their earlier work; in essence, Messrs Neumark and Wascher still think it wrong. But the range of their respective estimates has shifted - enough so that they now touch, just, in the middle. Messrs Card and Krueger no longer insist that the higher minimum wage pushed employment up; they have settled for saying that (contrary to the standard model) it "probably had no effect". Messrs Neumark and Wascher have lightened the emphasis on falling employment, emphasising instead their conviction that (contrary to what Messrs Card and Krueger had first claimed) employment did not go up.

Second inning

Just as the discussion seemed about to fizzle out, two new contributions have arrived. A note by Thomas Michl suggests a different compromise: maybe both of the earlier positions were correct. After examining the data already collected, Mr Michl suggests the following possibility: that the minimum-wage increase left the overall number of workers employed roughly the same, but reduced their hours. (Not implausible, given that most workers in the fast-food business are part-timers.) Then it would be true that the wage rise reduced the demand for (hours of) labour, as the standard model says; but at the same time it could also be true, as advocates of the minimum wage say, that the incomes of the affected workers went up, thanks to the combination of fewer hours at work and the higher wage rate. In which case the policy could be judged a success, even though it had "reduced employment".

Still there? Hold on, because yet another new idea, much more hostile to minimum wages, has now been put forward. Peter Tulip, an economist at the Federal Reserve, asks a different but very interesting question: could a high minimum wage raise the equilibrium economy-wide rate of unemployment (or NAIRU)?

This directs attention away from the "affected workers" in fast-food restaurants or wherever, the focus of the earlier research, to the labour market as a whole. Suppose a high minimum wage, by pressing on the structure of pay differentials, raises wage growth and hence inflation across the economy. Higher unemployment would then be necessary to stop inflation accelerating. Employment might not fall among "affected workers", but it would have to fall in some other part of the economy. The damage would be subtler, but no less real.

Mr Tulip finds, on crunching his numbers, that this is what happens. So strong is this indirect effect, on his calculations, that the gradual fall in the relative value of America's minimum wage over the past 20 years is capable of explaining 1.5 percentage points of the fall in the country's equilibrium rate of unemployment over the same period. In Mr Tulip's view, a good part of the difference between the low equilibrium rate of unemployment in America (and Britain) and the much higher rates in continental Europe can be attributed to Europe's higher minimum wages.

That should stir things up. It will be interesting to see how other specialists respond. You will probably be able to read about it in the American Economic Review in, say, three or four years' time.