In the Washington Examiner, Tim Worstall explains why a new well-researched paper on the minimum wage will be misunderstood and then used to “prove” things it doesn’t actually say:
None of this changes the standard intuition that when there’s a heavy such bite then there will be ill effects. What it does do is then lead us to trying to calculate what is a wage that does have that snarl, that bite? What is a minimum wage that is “too high” in the sense of having an excess of those ill effects upon employment? This is where I predict — no, not fear, not posit, nor surmise, but predict — this paper will be misused.
Our thinking is that the effects come from the relationship between the minimum and median wages. If we insist that wages cannot be lower than more than we already pay half the people, then we really are going to have problems. A minimum wage of 100% of the median wage isn’t going to work, that is. That ratio is called the Kaitz Index. This paper shows us that there are few to no such bad things happening up to 0.59 on that Kaitz measure. We can have the minimum wage at 59% of the median wage and know that we’ll have the good effects and only trivial amounts, at worst, of the bad.
You can see what’s going to happen next, can’t you? The Economic Policy Institute tells us that the median wage is about $22 this year, and 59% of that is $13. A bit of rounding and some aspiration, and why not go for a $15 minimum wage?
Except there are two median wages. Part-time and seasonal wages tend to be lower than full-year and full-time ones. The Economic Policy Institute is using that higher full-time one. The one for all jobs is quite a bit lower, $18.58 per hour. Take 59% of that and you get a rather lower level of $10.95 an hour. That’s around and about what McDonald’s, Walmart, and similar establishments pay as entry-level wages, which does seem about right, doesn’t it?
So, the new research paper, from esteemed researchers, published in the world’s top English language economics journal, tells us that minimum wages up to a certain level cause few to no problems. They’ve shown this for up to 59% of median wages. But which median do they mean? Dube himself told me they mean that lower one — specifically, the “median wage of all workers, not just for full time.”
But we all know how this is going to be used, don’t we? As proof that $15 an hour won’t cause any problems — which isn’t what the paper shows at all. Rather, it says that a $10.95 an hour minimum wage shouldn’t cause any problems of note.
The new paper is good empirical work. The fault is in what people will argue it says, not what it does.
