Economist Harry Holzer pinged me yesterday re this piece by Ben Spielberg and me on the dust up in Birmingham, AL around their local minimum wage. Harry felt we gave much too short shrift to the the research showing job loss impacts from minimum wage increases. So, him being a reasonable guy, and me being the same, I invited him to write down his thoughts about this.
I offer some responses below.
Do Local Minimum Wage Increases Never Reduce Employment?
Harry J. Holzer, Georgetown University
My friend Jared Bernstein and I have largely similar views on most economic issues, especially regarding the labor market. For instance, when it comes to raising the minimum wage, we both support increases at the state or federal levels up to $10 an hour, and usually oppose much larger increases to $15. (I am also skeptical of going as high as $12, though I am not sure Jared shares this concern).
So I was surprised by Jared’s column (with Ben Spielberg) in yesterday’s Washington Post, which seems to argue that increases at the city or county level never lead to employment losses there.
Bernstein and Spielberg’s main point is to criticize a proposed new law, passed by the legislature in Alabama, that would forbid any city to raise its minimum wage or impose other mandates on employers, like providing modest paid sick leave to its workers. Like them, I would find such a law both foolish and offensive.
The Alabama law was passed in response to a vote by the Birmingham city council to raise the city’s minimum wage up to $10 an hour over two years. This is not such an egregiously large increase – though I would want to know about a bit more about average wages and skill levels in Birmingham before fully endorsing it. Nevertheless, even if it caused some modest shift in employment over time out of Birmingham and towards other localities in Alabama, it is hard to see how such an increase would be so disruptive to business there as to merit such a draconian response by the legislature.
Yet, in making this argument, Bernstein and Spielberg also overstate their case, and throw caution to the wind. For one thing, they cite just two studies of the minimum wage literature – one by Card and Krueger (on fast food restaurants in New Jersey and Pennsylvania) and one unnamed (by Arin Dube and his coauthors). But David Neumark and William Wascher have written well-known and credible critiques of both studies, and found larger negative employment effects in their work. In this research, small tweaks to the methods used can generate very different findings – and we need to present all of them (and to use an average of their estimated effects) in any honest discussion of the likely effects of a specific minimum wage increase.
There are other studies that are relevant here as well. Aaron Yelowitz of the University of Kentucky has analyzed the effects of large local minimum wage increases in Santa Fe and found sizable impacts on unemployment there. A new paper by Peter Brummund and Michael Strain, recently posted at the American Enterprise Institute, studies employment effects of minimum wage increases across county lines and finds that the size of effects often depend on several different circumstances – such as whether or not the new minimum is indexed against inflation (which worsens effects on employment) and the average wage level in the county at the outset (with lower levels tending to worsen employment losses).
And some recent research suggests that the effects of minimum wage increases are fairly nuanced, and therefore harder to discern in the traditional studies, which look only at short-term employment effects on broad groups. For instance, Jonathan Meer and Jeremy West find that minimum wage increases do not cause large short-term employment losses, but rather slower employment growth over a longer time period for low-wage groups where they occur. And Jeffrey Clemens of UC San Diego finds substantial negative effects on employment of very specific groups more directly affected by the minimum, like young high school dropouts. In fact, he finds a nearly 6-percentage point drop in the employment of dropouts below the age of 30 as a result of the federal minimum wage increase in 2007-09, though effects on broader groups are much more modest.
A denial of the true research evidence on minimum wage increases has led some cities to increase their minimum wages up to $15 an hour, as Bernstein and Spielberg note. Perhaps Seattle and the Sea-Tac region in Washington state can afford such increases without creating major employment losses; but similar increases in Los Angeles and Washington DC (where voters in a recent referendum voted to go to a $15 minimum before a local judge put it on hold) could have serious negative effects on the employment of large groups of very unskilled workers in these cities. Piling on with additional measures such as extremely generous paid family leave and other worker protections, as the DC Council is trying to do, would likely drive many employers over the river to Arlington VA over time (or cause DC employers to economize on their hiring, perhaps by mechanizing more of their operations).
Yet nothing in the Bernstein-Spielberg piece seriously acknowledges these likelihoods. There are, of course, large benefits in minimum wage increases that also must be weighed against these losses. But much of the wage increases will accrue to middle class youth or part-time second earners in middle class families, while the high school dropouts suffer the largest job losses.
In short, the debate over the effects of the minimum wage needs less absolutism on both sides, and a more nuanced discussion of their pros and cons. Like Bernstein and Spielberg, I condemn a thoughtless new law in Alabama that needlessly ties the hands of local city councils to enact moderate wage increases. But I hope Jared and Ben will also be more careful in their future writing to acknowledge a broader range of research evidence on this issue, and to urge more moderation in the discussion.
[Back to me: Many reasonable points here. Speaking for myself, my goal in the WaPo piece was to point out the egregious intervention by the legislature to block the local council’s initiative, one that I judge, based on the full spate of the research, including that cited by Harry above, is likely to have its intended effect of helping the majority of low-wage workers affected by the raise. I have, as HH notes, argued that $15 is too high a min wg in certain parts of the country–and thus too high for a federal level, at least w/out a very long phase in.
Min wg expert Alan Krueger recently wrote, in this spirit: “Research suggests that a minimum wage set as high as $12 an hour will do more good than harm for low-wage workers, but a $15-an-hour national minimum wage would put us in uncharted waters, and risk undesirable and unintended consequences.”
The challenge in assessing the empirical min wg literature is that while there are papers that find negative results, as HH correctly notes, there are many that do not. Moreover, the magnitude of the negatives matter. Negative elasticities (i.e., job loss effects) well below ‘1’ imply many more beneficiaries–“more good than harm” as Alan K puts it–than those hurt by the increase. So, as is so often the case in empirical economics, we have to humble about our knowledge.
I think it is fair to summarize this huge and contentious body of research as follows: moderate increases in minimum wages have their intended effects. What’s moderate? History suggests increases that capture less than around 20% of the low-wage workforce in their sweep fit this definition. But others may draw defensible lines in other places, both lower and higher.
As I read the state wage data for AL, the Birmingham proposal is moderate by this criterion, as HH suspects. And I’m very glad to hear him agree with me and Ben re the oppressive preemption of the state legislature.]