Tuesday, March 12, 2013

Bryan Caplan on the minimum wage


I think any people who have doubted me should take Bryan's word for it that Card and Krueger's work is "well done". He continues to disagree with the finding, though, because of strong priors about a downward sloping demand curve, and a few other concerns. First, the downward sloping demand curve:
"Part of the reason is admittedly my strong prior. In the absence of any specific empirical evidence, I am 99%+ sure that a randomly selected demand curve will have a negative slope. I hew to this prior even in cases - like demand for illegal drugs or illegal immigration - where a downward-sloping demand curve is ideologically inconvenient for me. What makes me so sure? Every purchase I've ever made or considered - and every conversation I've had with other people about every purchase they've ever made or considered."
Sensible, right? Here's the problem with that - Card and Krueger think the demand curve is downward sloping too. You should not be under the impression that people who think the minimum wage has a negligible effect (which is different from minimum wage advocates) are suggesting that the demand curve isn't downward sloping. Although there are more complicated search theoretic arguments, remember the explanation provided is that firms have some degree of market power over low skill labor. This makes sense, right? Think of when you were a low skill worker (maybe you still are, of course). Could you walk up to the boss and make demands (politely, of course) with any expectation of getting what you wanted? Of course not. Your ability to do that sort of thing grows with your skill level because you're a differentiated commodity which means you're acquiring some market power of your own.

So think about the traditional monopoly set up (it's the flip side of market power in the labor market but I think it's easier for people to think about). Does the monopoly set up rely on a horizontal monopolist supply curve? Sometimes it's horizontal for exposition (constant costs), but you also see upward sloping supply curves. It's not the supply curve that drives the behavior of the monopoly case - it's the demand curve. Similarly it's not the demand curve the drives the behavior of the monopsony case - it's the supply curve. We still assume that the demand curve is downward sloping.

I agree with Bryan's skepticism of empirical research, in the sense that I agree that our data is riddled with endogeneity and selection bias such that any given study can only be thought of as a rough approximation of the answer. The answer, I think, is to do a lot of different kinds of study on different data exploring the question from different angles, and then see where the center of gravity of the results are. This is why things like that CEPR lit review I shared and the Neumark and Wascher study are so important. If people were just going off of Card and Krueger I think you'd see more people saying the minimum wage did good. People are actually going off of a whole suite of results and that's why they say it's a negligible effect. You still have to explain why there's a negligible effect rather than a negative effect, and in that respect Card and Krueger's explanation of the result is still a pretty popular one.

Bryan then goes on to say that findings on immigration and European labor legislation should influence our thoughts on the minimum wage.

On immigration - at some point I should post on low skill immigration. I'm kind of in the same place here as I am with the minimum wage. The negative effect seems to be small, probably because of complementarities and the fact that immigrants aren't just suppliers of labor - they're demanders of output. I actually think the Borjas design is somewhat stronger than the Card design in this case. Card's natural experiment is fine, but his comparison groups aren't nearly as strong as in Card and Krueger, and that's my concern. As a precaution - I formulated my take on Borjas/Card as a master's student in the estimable Prof. Parsons's PhD labor class five or six years ago. I should probably revisit the papers sometime soon. All my recent work on immigration has been of the high-skill variety and it hasn't been nearly as technical (in the sense that I haven't been estimating any native impacts so much as thinking about general policy issues). To circle back to my initial point - this shouldn't necessarily sway you even if you believe Borjas that the labor demand curve is downward sloping because as I said - I think it's downward sloping for low wage labor too.

I agree with Bryan on the European labor regulations. I think most people who like the minimum wage would agree with that too, and know that there's a point where even in a monopsony case you get too high a minimum wage.

I'll need to revisit the Keynesian argument Bryan makes. This doesn't seem very relevant at all, but perhaps I'm missing something. It might be more relevant for demand shocks and the minimum wage.

A reminder - my policy position is that we probably shouldn't have a federal minimum wage, but it's not a bad idea for states to experiment with it. It's not a bad idea because I am convinced that the Card/Krueger monopsony effects are real, but also because I'm usually pretty flexible with state-level experimentation with policy. The federal minimum wage imposes high real wage floors on the poorest states.


  1. We could call it productivity investment.

  2. Just a warning, his link to the supposed Keynesians who think the empirical evidence overwhelmingly supports their position is just another post of his making up a bunch of shit about "Keynesians".


All anonymous comments will be deleted. Consistent pseudonyms are fine.