The new consensus on minimum wages

Coming in a bit late, I have the opportunity to survey a range of blogospheric discussion of the topic of minimum wages, which largely supports the view (not surprising to anyone but an economist) that minimum wages are good for low-income workers. The traditional view among economists was that minimum wages reduced employment and thereby harmed workers, but this view has been overturned, or heavily qualified, by empirical evidence, beginning with the work of Card and Krueger.

The debate kicked off with a piece by Stephen Landsburg in Slate, noting that recent US econometric studies had failed to find economically and statistically significant negative effects on employment resulting from higher minimum wages. This was surprising, in view of a range of earlier studies which found right-signed effects, but were statistically weak because of small samples. Landsburg argues that this might be an example of publication bias, in which studies with no statistically significant results tend to get discarded. He concludes

Now that we’ve re-evaluated the evidence with all this in mind, here’s what most labor economists believe: The minimum wage kills very few jobs, and the jobs it kills were lousy jobs anyway. It is almost impossible to maintain the old argument that minimum wages are bad for minimum-wage workers. In fact, the minimum wage is very good for unskilled workers. It transfers income to them

Landsburg then goes on to argue against the minimum wage on the curious ground that it’s a less transparent alternative to policies such as an Earned Income Tax Credit. Brad de Long responds, endorsing the EITC, but arguing that minimum wages are also an effective policy instrument for transferring income to the poor.

There are quite a few interesting responses. Steve Verdon develops Landsburg’s argument, pointing out that a minimum wage increase which raises the general cost of goods and services is like a consumption tax and has an associated deadweight loss. That’s true, but it’s also true of whatever tax may be used to finance the EITC. Robert Waldmann suggests changing the structure of payroll tax, but as he himself points out, his argument disregards the point that the budget is already in deficit. Tyler Cowen observes that increases in wages may be offset be reductions in working conditions. Interestingly, no-one seems to have defended the traditional view on empirical grounds.
An interesting and important question is whether these results can be transferred to other countries like Australia, where the minimum wage is higher relative to average weekly earnings. In the survey of the literature we did for the National Wage Case, Steve Dowrick and I concluded that, although there might be some reduction in employment and some leakage to low-wage workers in high-income households, the evidence showed that minimum wages help low-income workers . Our study is here (PDF file)

Overall, my view is close to that of Brad de Long. Minimum wages are a useful policy instrument, but by no means the only or most important one, to improve the position of low-income workers.

Update Jacob Levy asks, reasonably enough

If, as Landsburg claims, the published studies are “all in agreement” about the direction of the effect, then the underlying distribution of studies can’t be as he describes it, can it? Publication bias in favor of significant findings, superimposed on an actually-neutral relationship ought to generate equal numbers of ostensibly-significant findings in each direction.

Actually, the Card and Krueger study found weak positive impacts of minimum wages on employment using a data set where most of the obvious sources of bias had been removed. There may have been earlier studies with similar results, but they would almost certainly have been discarded, on reasonable grounds of weak statistical significance or omitted variable bias. By contrast, studies with similar weaknesses, but with the expected sign would have been published.

10 thoughts on “The new consensus on minimum wages

  1. Sir,

    In your first paragraph you write, “The traditional view among economists was that minimum wages reduced unemployment and thereby harmed workers….”

    Either I completely misunderstand, or this should instead say that minimum wages “reduced employment.”

  2. John,

    I think it’s difficult to read too much from the Card Kruger results for Australia since in the US minimum wages are low and the increase they examined was small.

    On your survey, I thought your appendix and critique and Lewis and McDonald was particularly clear. However, it was about estimating the effect on employment of increases in aggregate real wages, not minimum wages, so is only partly relevant to this discussion. More important, you conclude that that the output-constant elasticity of aggregate demand for labour with respect to the nominal wage is only -0.18. However, this assumes the labour share stays constant at 0.6. If the labour share of output inceases, so does the negative effect of wages on employment.

    As your own Figure 6 shows, in the mid 70s, there was a big increase in the labour share as real wages increased. On your analysis, this would have led to a big decrease in the demand for labour.

    You’re too smart to say that labour demand is never affected by real wages, but a lot of people, citing Card Krueger and a other studies they don’t really understand, fall into that trap. That’s only true if the labour share of ourput doesn’t change, but if the labour share of output was constrained to not change, there wouldn’t be any arguments about wages in the first place.

  3. Some of the shallower studies ask the wrong question. What counts most isn’t whether minimum wages harm the workers but whether they harm the workers and potential workers, i.e. the unemployed. It’s a survivor bias thing; tautologically, any constant rate of unemployment causes no harm whatsoever to workers in aggregate, so an increase in minimum wages must help the survivors once any effect on those losing work washes out.

    So it comes down to whether the studies made proper allowance for this survivor bias – and the fact that people losing work might not be slap bang in the middle of sectors being studied. Even following individuals over time wouldn’t bring out the true cost, if those individuals simply dropped a rung and yet other persons dropped off the end. You would measure the cost of one rung dropped rather than of falling off completely.

    I think I’ll give another plug to Professor Kim Swales of the University of Strathclyde, whose approach to unemployment amounts to raising minimum wages without causing an increase in the marginal cost of labour (in fact, a reduction). It’s a Pigovian thing, which I go into at my publications page linked via the URL I posted along with this.

  4. The paper an interesting survey, the Appendix hard. Taking output as fixed you get conditional input demands given prices. You then suppose output is determined by an aggregate demand function not by a supply function where input prices have a role. Strange if you are building up a model based on price-taking firms.

    Why not suppose firms maximise profits unconditionally so input demands depend only on input prices? This captures effects of wages both on labour demands directly and also effects of wage increases in reducing supply and thereby reducing input demands. This will presumably make estimated labour demands more wage elastic.

    George Stigler once wrote a paper “All Demands are Inelastic”. The Card/Krueger results an instance of Stigler’s Law. The observed inelasticity a reflection of the difficulty of the pursuit or because elasticity aint there?

  5. Milton and Harry are both correct in pointing out that excessive growth in the wage share can reduce employment – this can be more more important than the substitution effect we usually focus on, and was important in the 1970s and 1980s. But the wage share doesn’t seem excessive at the moment, and it’s important to consider the other side of the coin, the effect of increased wages on aggregate demand.

    Harry, can you give me the Stigler reference?

  6. The Stigler paper I read about 20 yrs ago in some collected essays of his. I am sorry I can’t recall specific title.

    The title was obviously intended to be provocatively humorous. The general line was that economists were always finding good reasons for supposing demands were inelastic contrary to basic demand theory.

  7. Based on the official govt. figures from the UK the effect of a minimum wage policy seems to have been one of stimulating ever higher numbers of people in employment. How believable this is depends on how much faith you have in the govt’s honesty with number I guess.

    Last month seems to have reversed by 6000 jobs but let’s not panic yet.

    However I thinkmany recently graduated well qualified young people are in no danger of having to repay their student loans. It appears that employers see the minimum wage as also being very close to the maximum wage, which suggests that for a large number of jobs of different types the recompense for performing the task may be reducing in real terms. Everything reducing to the lowest level allowed.

    Have any studies considered that possibility?

Comments are closed.