Failure to correctly summarize minimum wage research

In the original version of “Economics Is a Science, part 3”, I used this illustration of minimum wage research from Doucouliagos and Stanley (2009):

I used this chart to describe a consensus estimate of a negative effect on employment. This is not correct. The authors specifically made this chart to uncover any publication bias in minimum wage research, and indeed, it shows that there’s a bias toward negative results. Adjusted for this bias, the effect is still near zero but positive, not negative. Nevertheless, it’s true that economists tend to think the effect is small and negative, though there’s a lot of disagreement. The consensus, as I described in the original version, is that the effect tends to be near zero in practice and is definitely not always positive or negative, and there are weird edge cases.

Having repeated this mistake in a post I cooperated on with Terence Highsmith, John Ruf pointed it out, and I fixed it immediately. I’ll have to wait for Terence to fix the post Ruf was restacking, though, since I don’t have editing privileges for it.

The real mistake was using minimum wage research to try to show economists can come to a consensus on something. It’s like stepping on a landmine to jump higher.