Monday 9 May 2011

Minimum wages - addressing the more sensible critiques

Here's the best case I can make against my position on minimum wages in New Zealand. If Tyler has Tyrone, here's Ehrlich:
Economists' opposition to minimum wages is largely faith based. Sure, demand curves slope down. But nobody would believe that a move from a minimum wage of $0.01 to a minimum wage of $0.02 would have any effect on employment. It's all in the elasticity at the relevant point on the curve. And, the evidence there just isn't great. The US studies often find no effect; those that do find an effect may be subject to pretty strong selection and publication bias. And the trend in evidence isn't in your favour either. Earlier studies found greater effects of minimum wages on employment, but the more recent evidence shows very little or no effect. If the employment effects are weak, then minimum wages can't be that bad a policy: gains to those keeping their jobs have to outweigh losses to those losing theirs if not that many people actually lose their jobs. And appropriate division of surplus between employers and employees is largely a matter of taste if the efficiency tradeoff is small.
Here's my best response, cribbed somewhat from the worthwhile bloggers over at WCI.

First, I'd grant that at low levels, minimum wages are exceedingly unlikely to be binding. Or, rather, that any instance in which they are binding is going to be swamped by other noisiness in the data such that effects can't be seen. A minimum wage of $3/hour might be binding on sheltered workshops for the mentally disabled, but that doesn't really show up in aggregate employment statistics. At higher levels, they're more likely to be binding. That's why nobody argues for $50/hr minimum wages, not even Unite. But have a look at the time path of the real value of the US minimum wage.
Is it much of a surprise that earlier studies would have found greater effects? Stephen Gordon over at Worthwhile Canadian Initiative traces the minimum wage as a percentage of average hourly wages. In 1980, they approached 48% of the average wage; by 1990, they'd dropped to around 34% of the average wage. It would be surprising if studies on earlier data didn't find bigger effects given the time path of the minimum wage relative to the average wage. Here's Stephen's summary of some work by Pierre Fortin:
The actual state of knowledge of the impact that the minimum wage has on employment in North America, and especially in Qu├ębec, leads to the conclusion that a minimum wage that is greater than 50% of the average wage is harmful to small wage earners and that a minimum wage that is less than 45% has very little risk for this group of workers. Between these limits, the area of 45% to 50% would represent an increasing danger to employment.
I'm happy to grant that opposition to minimum wages are largely ideological where minimum wages are down around a third of average wages. Yes, some people will still be pushed out of work by them. But not many, or at least not many in the short run. Choices of whether the marginal bit of money goes to employers or employees in the absence of employment effects is more a matter of taste than anything else. I'd still oppose minimum wages, but with my "libertarian freedom of contract" hat on rather than my "pluralist but largely utilitarian economist" hat.

You then have to read statements by economists about the minimum wage in the context of the relative bindingness of the minimum wage then prevailing. When Whaples conducted his survey of labour economists on the minimum wage, only 13% disagreed with the proposition that minimum wages hurt the employment prospects of young and low skilled workers; memories of very high relative minimum wages were fresh in those labour economists' minds. When Fuller and Geide-Stevenson conducted theirs, opposition rose to about a quarter of economists, but the prior decade had seen a relatively lower real value of the minimum wage. And so it was less binding.

The relative strength of the labour market will matter as well. The best New Zealand study of the effects of minimum wages on youth employment, conducted by Dean Hyslop and Steve Stillman, found that changes in the youth minimum wage in the early 2000s had no effect. It was a great study, but it's also true that the study spanned a period in which overall unemployment was at record lows and businesses were desperate for any workers. The increase in the youth minimum wage over that period just wasn't particularly binding.

Another source of contrary findings can be industry choice. New Zealand labour unions like to cite the recent and excellent paper by Dube and coauthors showing no effect of minimum wages on workers in the American restaurant industry. But it's a lot harder to replace waitresses with machines than it is to outsource call centre work. That may not be true for the kinds of places with low paid wait staff for much longer, but it would be true in the periods studied. Effects on the restaurant industry needn’t be representative of effects for the country as a whole.

Further, we'd want to pay close attention to the relative level of the minimum wage in the States before recommending increases here. Dube et al's work spans 1990 to 2006 in the US. The average dollar value of the minimum wage in their sample is $4.84 per hour. The average wage in the US in 1998, the midpoint of their sample, was about $13. So changes in the minimum wage didn't make much difference in the restaurant industry when the minimum wage was around 37% of the average wage. I'd believe that. The minimum wage wasn't getting into the range where we'd expect large overall effects and the study's of an industry where it's hard to substitute other things for minimum wage workers.

New Zealand’s minimum wage is thirteen dollars per hour. New Zealand's average total hourly earnings are about $26 per hour. And where the US numbers cited above on hourly wages are for production and nonsupervisory employees, the NZ numbers are for all earnings. Making the figures comparable would make the real value of the NZ minimum wage a much higher fraction. We're well into the range where we expect bad employment outcomes from minimum wages. It's bad enough applying those minimum wages to regular workers; it borders on insane to apply them to sixteen year olds. And unless you want to call Chris Dillow a raving neoliberal, you can't call objections to increasing the minimum wage when it's in this kind of range ideological. Reducing the minimum wage to about 40% of the average wage is sound economics. Reducing it to zero - there's room for folks to disagree because the employment effects down in that range probably aren't very big.

I'd also be pretty nervous about folks citing petitions by American economists favouring increases in the minimum wage as having any relevance in New Zealand. They advocated a slow move of the minimum wage from $5.15 - less than a third of the then average wage - to $7.25. By the time the higher level would have hit, average wages would have risen to keep the higher minimum wage still below 40% of the average wage. I'd be surprised if many of those economists would also have signed on for increases in the minimum wage to levels anywhere near 50% of average wages; I'd be shocked if they thought it a good idea to apply those minimums to sixteen year olds. And recall that while 600 AEA members and 5 Nobel Laureates signed the petition, 19,400 members of the AEA and 29 Nobel Laureates did not.


  1. On why economists support the minimum wage see:

    "Reasons for Supporting the Minimum Wage: Asking Signatories of the “Raise the Minimum Wage” Statement" by Daniel B. Klein and Stewart Dompe, 'Econ Journal Watch', Volume 4, Number 1, January 2007, pp 125-167.

  2. The most sane reason of the lot, if I recall correctly, was that it was a second best policy when constrained against making other policy moves that benefit the working poor, and that at low rates relative to the average wage, employment effects were small.

    I'm not sure why economists with those preferences don't lobby directly for a guaranteed minimum income or a negative income tax.

  3. Ehrlich is a horrible name for an alter-ego. It's the last name of a prominent environmentalist doom-monger. Erin would be much better. Or you could go with Rick or Elrod.

  4. You got a mention on News Talk ZB tonight when larry was interviewing fill in phil at about 6:15pm. Might be worth a listen as I'm not sure you'll like the way larry framed what you said.

    Have you been interviewed on this topic by Larry sounds like he'd be interisted in doing so.

  5. This makes me wonder whether there are studies that take as their main independent variable the ratio of the minimum to the average hourly wage. Also, given that siginificance tests are arguably inappropriate in the kinds of samples this research uses, which way do the nonsignificant coefficients point?

  6. @TGGP: Ehrlich is the opposite of Eric. I thought it fit. And Erin is my sister.

    @Anon: Missed it, thanks for the heads up. Haven't talked to Larry about it, or at least not this go round. I was on his show in February, but about Rugby World Cup. Larry's characterisation isn't bad. He took out a caveat or two, but that's where I'd be putting my money were there a market on it - youth unemployment would be 5-8 points lower (8 points by this quarter's figures, about 12,000 kids) had they not gotten rid of the lower differential youth minimum wage. The more quarters pass, the harder it is to point to anything else that could have driven the wedge between youth and adult unemployment rates. You need something that starts in 08 and persists through present that plausibly makes the wedge between adult and youth unemployment.