The Non-Simple Economics of the Minimum Wage

The National Post ran a little pro-and-con debate on minimum wages in today’s paper.  I was the “pro” side; my argument was excerpted from a longer paper on “What determines wages and income distribution” that is available on the CAW’s web site. The “con” side was written by two economists at the Fraser Institute.

The Fraser argument was centred in the traditional neoclassical claim that intervening in markets only hurts those you are trying to help.  How compassionate of them to care so much about the poor!

Interesting that the Ontario government — which led the way in Canada, the first province to achieve a $10 minimum wage — just last week announced the wage would be frozen this year.  Bankers aren’t getting their pay frozen (bonuses at the Big Six were up to a record $9 billion in the fiscal year ended October 31 2010).  But the hard-working people slinging coffee at Timmie’s are.

I recommend the longer CAW paper, if I do say so myself, as an accessible introduction to heterodox thinking about wage determination and income distribution.

Here is my side of today’s debate:

Like most economics majors, I was taught early on at university that minimum wages screw up an otherwise efficiently-functioning marketplace for labour.  You see, there’s a demand curve for labour, and it slopes down.  There’s a supply for labour, and it slopes up.  The two lines cross in the middle, at the sweet spot where supply equals demand.

Now draw the minimum wage: a horizontal line, positioned above the cross.  It’s plain as day.  Too much supply, too little demand, too much unemployment.  Well-meaning but foolish bureaucrats should leave the market alone to perform its autonomous, masterful balancing act.

The story is simple.  It’s elegant.  And it’s wrong.  But you have to progress far beyond Economics 101 to find out why.  And in the meantime, that simplistic supply-and-demand diagram gets deeply imprinted on too many impressionable minds.

No employer hires labour just for the sake of having workers around.  So the fact that labour is cheaper, in and of itself, never guarantees that more will be hired.  Why do employers hire workers?  To work: that is, to produce something.  Employment is a derived demand, dependent on sales of whatever good or service workers produce.  Their employment depends mostly on whether there’s enough demand for their output, so that their employers can profitably produce it.  That, in turn, depends on a whole stable of economic variables, macro as well as micro – not least including whether working families have the purchasing power to buy back the stuff they produce.

So, contrary to Economics 101, there’s not really an independent demand curve for labour.  And the supply for labour cannot be depicted simplistically, either.  In fact, labour supply often declines as wages increase (since workers can then afford more leisure time).  The two lines don’t reliably and stably cross.  In fact, it’s not clear they cross at all: even where wages are unregulated, unemployment is a normal, permanent feature of the economy.  Except under very unusual circumstances (like World War II), labour supply never actually equals demand.

Nevertheless, that Economics 101 story was digested holus bolus by a generation of economists, and minimum wages fell out of fashion.  From the late 1970s to the mid 1990s, Canadian minimum wage levels (after inflation) declined by about one-third.

But then a whole new spate of research threw the conventional wisdom deeply into question.  Led by U.S. economists David Card and Alan Kruger, the research found no reliable empirical evidence that higher minimum wages produce unemployment.  But they found several benefits of minimum wages, including a “trickle-up” effect (whereby many non-minimum-wage workers also got raises); reduced turnover; and higher productivity.

Indeed, in a more complete economic model, there are even circumstances in which higher minimum wages can lead to higher employment.  If aggregate demand conditions are chronically weak, the purchasing power created by higher wages can spark a positive and circular process of demand generation, spurring job creation.  Similarly, under conditions of “monopsony,” where powerful employers can influence wages (rather than accepting a going rate), minimum wage laws can induce higher employment.  Employers can then no longer suppress wages (which are now set by law), and hence stop artificially curtailing their hiring.

In practice, the effect of minimum wages on employment is probably a wash.  Gradual increases in minimum wages, within reasonable bounds, have virtually no impact on employment at all, in either direction.  So long as levels are set realistically relative to productivity and profitability, minimum wages can be increased with no measurable damage to employment.

Perhaps influenced by this recent sea-change in economists’ attitudes, policy-makers in most provinces have begun to revitalize minimum wages.  After years of stagnation, the real purchasing power of minimum wages has increased markedly since 2005, boosting incomes across the lower tiers of Canada’s labour market.  With business profits simultaneously reaching their highest share ever of Canadian GDP, it could hardly be argued that these modest but important increases squeezed out private sector activity.  On the other hand, those higher minimum wages contributed notably to the first real wage gains enjoyed by Canadian workers in a generation.

This successful policy trend should be continued.  Minimum wages (now at or near $10 per hour in most provinces) should be increased gradually but steadily in the years to come.


  • I remember McGuinty musing about freezing the minimum wage a couple of years ago.

  • All these people who are so concerned about the poor, but who fear that the minimum wage is too blunt an instrument to help them, should put their money where their mouth is and join the call for a national poverty reduction strategy with legislated targets and timelines. In addition, groups like the Fraser Institute and the Canadian Federation of Independent Business who operate in the Western provinces should publicly advocate for provincial poverty-reduction plans. I can assure them that such plans will be more comprehensive and effective than simply increasing the minimum wage.

  • Iglika,

    All those groups would probably take up your idea if and only if it were funded through increases in consumption taxes.

  • Hi all,

    Newfoundland and Labrador’s minimum wage has increased by 67% in five years. We reached $10 in July of 2010. Big campaign here by labour movement and progressive partners and a government willing to listen. See the below news release where higher minimum wages are linked to poverty reduction.

    Also please note in 2010 when the minimum wage increase by $1 (50 cents in Janaury and 50 cents in July) employment levels in the province continued to rise.

    As we all know higher minimum wages will not solve poverty, but they are part of the solution – an important part. They are also critical to how we share the wealth and narrow the gap between the rich and the rest.

    Lana Payne
    NLFL President

  • What about the argument that raising minimum wages will not reduce poverty because most minimum wage earners are not living in poor households?

  • That would seem to be irrelevant. The question isn’t whether most minimum wage earners live in poor households, it’s whether most poor households have minimum (or near-minimum) wage earners. Presumably most of the working poor aren’t making high wages.

  • Jim Stanford’s column in G and M Feb 24,points out another problem, although he doesn’t comment. Tamils hire Tamils ( at his Esso station). Even low paying jobs are hard sought after by the unemployed, and this sort of discrimination should not be allowed. It is common,among many ethnic groups, but comes under the radar.

  • I suppose that might be considered a a problem, but is it an economic problem? Someone who was unemployed becomes employed. If that’s a Tamil rather than someone else, because the Tamils are networking, I don’t see quite what economic impact it has. This is especially true since the data are strong that these nonwhite ethnic groups make less money than everyone else anyway, so them helping each other can only reduce economic disparities.

    Surely networking among rich white guys who went to Upper Canada College is a bit more economically relevant, since it’s more likely to increase stratification and reduce economic mobility by keeping the lower classes out of the good jobs.

  • “Someone who was unemployed becomes employed. If that’s an X rather than someone else, because the Xs are networking, I don’t see quite what economic impact it has.”

    But Mr. Library Guy, that logic could be used to defend any and all discrimination in hiring.

    Also, I think we need to distinguish between personal experiences and nationwide averages. If you are an unemployed white guy trying to get a job at a gas station, you gain nothing from the economic success of unrelated “rich white guys who went to Upper Canada College” (with Marc Lee) pulling up the white-guy average.

  • I didn’t say it wasn’t discrimination, or that discrimination was OK. I just said it wasn’t an economic problem. It isn’t. Economically if anything it decreases inequality. It’s the way immigrant populations have tried to move out of poverty for all of history. So to the extent it’s a problem, it’s not one that relates to the economics of the minimum wage. Lots of groups network and practice in-group preferential treatment; among them all, the only one where I can see a defensible claim for economic impact would be the upper class white guys.

    As to whether I have a problem with it as discrimination–well, it’s much like racist jokes: I tend to figure that while racist jokes are generally not OK, jokes about oppressed groups are always more not OK than jokes about dominant groups. Whites of all income layers network, whites hire preferentially, and whites according to the stats still get better jobs and get paid better for the same work. Same goes for men vs. women. I’ll worry hard about the Tamils when they reach par with the dominant population. There are of course individual white males who don’t successfully network and don’t get good jobs, or jobs at all–but then, I don’t believe Tamil preferential hiring has resulted in zero Tamil unemployment either.
    But, sure, it’s kind of annoying; I remember one time my stepdaughter tried to apply for a job at a Chinese restaurant and they literally pushed her out the door. I thought that was very rude, even if they didn’t want to hire someone non-Chinese. If a white-owned restaurant had done the same to an Asian there could have been a major incident, and rightly so. Go figure, every race has jerks, and there are downsides to the cohesion of ethnic communities. But until they don’t need that cohesion to survive I’m not gonna squawk too much.

  • Surely we want to make a distinction between mutual aid societies established within subordinate classes and those established by the dominant class to maintain their class dominance(and sell access to their class eg. UCC).

    The exclusive hiring practices of the Sikh community in Abottsford for example can perhaps be forgiven in the face of the exclusive hiring practices of other ethnic groups in the valley. We also might forgive the Sto:lo nations for having exclusive hiring preferences.

    And what about unions: many jobs are internally posted. Once all positions are filled by existing members then new positions are posted to the public. It is discriminatory but I am not sure it is a problem.

    Politically of course all we have established is that solidarity among subordinate classes is much more difficult to maintain than the solidarity among the dominant class. Before Olson we all knew this; and the ruling class better than us. There is reason why solidarity is the watchword of progressives: it is the most difficult thing to achieve and then maintain.

Leave a Reply

Your email address will not be published. Required fields are marked *