When the minimum wage bites

The 'old minimum wage research' - summarised here - found that the effect of an increase in the minimum wage on employment was significant. The 'new minimum wage research' - à la Card and Krueger - found that the effect was not significant. When the idea of increasing the Quebec minimum wage to 10$/hr was floated a few years ago, UQAM's Pierre Fortin took a look at the literature and asked himself: Why? He came up with what I think is a very plausible conjecture, one that explains many things, and offers some insight about the various minimum wage proposals that are floating around these days.

A good place to start is this graph:

Before 1980, the minimum wage fluctuated around 40-45% of the average hourly wage,  and researchers found that the minimum wage had a measurable effect on employment. Then came Reagan's minimum wage freeze, and that ratio fell to 33%. When the minimum wage was finally revised upwards in the early 1990's, the effect on employment was insignificant.

The same thing happened in Quebec, although I'm not able to show it to you: Statistics Canada, in its infinite wisdom, did a major data revision a few years ago, and (once again!) didn't think it necessary to update the historical series. But since Pierre was working with the old data, his survey has a graph that shows that the Quebec minimum wage was well over 55% of the average until the late 1970's, and declined to about 40% by the mid-1980's. The same pattern repeated itself: as the ratio fell, the estimated relationship between minimum wages and employment weakened.

The conjecture is that we shouldn't simply toss aside the old studies and conclude that labour demand is not elastic with respect to changes in the minimum wage. Instead, we should conclude that it is not isoelastic. When the minimum wage is low, the effect on employment is small. But as it rises, its effect also increases.  And the effect isn't constant over time, either: the long-run effects are stronger than in the short run.

An internal working group at the Quebec government kindly offers us a translation for how Fortin summarises the argument (the actual paper is written in French, and as far as I know, no electronic version exists - I got my copy by fax):

The actual state of knowledge of the impact that the minimum wage has on employment in North America, and especially in Québec, 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.

As Chris Dillow never tires of pointing out, one country where current research finds a significant effect of the minimum wage on employment is the UK. The current minimum wage there is something like 49% of the average, so it may be that this conclusion can be extended beyond North America as well.

So how do the proposals before us fit into this framework? The well-known petition to increase the US minimum wage puts it this way:

The value of the 1997 increase in the federal minimum wage has been fully eroded. The real value of today’s federal minimum wage is less than it has been since 1951. Moreover, the ratio of the minimum wage to the average hourly wage of non-supervisory workers is 31%, its lowest level since World War II. This decline is causing hardship for low-wage workers and their families.

We believe that a modest increase in the minimum wage would improve the well-being of low-wage workers and would not have the adverse effects that critics have claimed. In particular, we share the view the Council of Economic Advisors expressed in the 1999 Economic Report of the President that "the weight of the evidence suggests that modest increases in the minimum wage have had very little or no effect on employment." While controversy about the precise employment effects of the minimum wage continues, research has shown that most of the beneficiaries are adults, most are female, and the vast majority are members of low-income working families.

As economists who are concerned about the problems facing low-wage workers, we believe the Fair Minimum Wage Act of 2005’s proposed phased-in increase in the federal minimum wage to $7.25 falls well within the range of options where the benefits to the labor market, workers, and the overall economy would be positive.

The first part describes what we see in the graph: expressed as a percentage of the average, the current minimum wage is well below the threshold where we'd expect to see significant employment effects. But the increase to 7.25$/hr would bring that ratio to 43%, two percentage points higher than what was produced by the two previous revisions, and within sight of the 'danger zone' in which the effects on employment become significant. Although the current level is clearly so low that minor changes will not affect employment, it's by no means clear that a 40% increase can still be considered a minor change. But if the increase is phased in slowly enough so that the rise in the average wage can keep the ratio at or below the 41% produced by the previous two revisions, then there's good reason to think that the employment effects would be small.

On to Ontario. Currently, the minimum wage is about 38% of the average, so a moderate increase in the minimum wage is likely to have a negligible effect. But going from 7.75$/hr to 10$/hr is not a moderate increase: it would bring the minimum wage to 48% of the average, well into the range in which the employment effects are significant.

Of course, even if the effects on employment are small, that doesn't mean it'll do much to reduce poverty. It's instructive to play around with a few numbers here. In a study using Canadian data from 1993, Nicole Fortin and Thomas Lemieux find that 26.4% of minimum wage workers were from households in the bottom two income deciles (there's no official poverty line in Canada, so I'm using that). That's an over-representation, so an increase in the minimum wage that doesn't affect employment will have a progressive redistributional effect. That's a good thing, but the effect will be very small. If we suppose that this ratio held in 2005, then we can use the fact that 4.3% of workers earned minimum wage and that the employment rate was 62.7% to find that the proportion of people in the bottom two deciles who are minimum wage workers is about 4%. The vast majority of those in poverty would not benefit from an increase in the minimum wage.

So what have we learned from all this?

• When minimum wages are 'low' - say, less than 40% of the average hourly wage - then moderate increases won't have a significant short-run effect on employment.
• When minimum wages are around 45% of the average, they significantly reduce employment.
• No-one has been able to find any evidence to suggest that increasing the minimum wage has a measurable effect on reducing poverty.

Comments

[thwap]

Here's a Fortin study from 1998: http://www2.arts.ubc.ca/cresp/irpp1.pdf

[Stephen Gordon]

That's the one I mentioned in the second-last paragraph - the one by Nicole Fortin and Thomas Lemieux.

[thwap]

Ah, .. I misread. I thought the version you were talking about hadn't been translated. never mind.

[Fred]

thus I suggest bringing back the minimum wage to 0 for everyone in order to make disappear poverty

Some asks why economists are always wrong. The answer is easy: just read one of their studies