Bryan Caplan links to a relatively new working paper by Kleiner and Krueger on occupational licensing. Studies show that it is associated with a 15% wage premium. What Bryan is interested in here is that licensing doesn't seem to inflate the returns to education very much (often getting a license requires a college degree). Bryan is surprised... I'm not sure we should be that surprised. Bryan is a skeptic about how useful degrees really are, but of course if you think they are useful, then this isn't an artificial inflation of the returns to education at all, and it's probably a pro forma requirement for formal training that people working in this line of work would have pursued anyway.
Morris Kleiner - a nationally known expert on licensing and the co-author of this paper - is one of the chapter authors for the engineering book I'm contributing to (his chapter will also be on the licensing of engineers). He's a very nice, down to earth guy. I shared a cab with him from NBER back to the airport, and he had lots of interesting thoughts to share on what academic life was like.
I think Kleiner and Caplan are generally right about these sorts of supply restrictions, but we need to be careful. There are real principal-agent problems and information asymmetry problems in professions that use licensing, which a reasonable licensing standard could actually help with. You can't just point to the 15% wage premium and chalk it up to protectionism. A chunk of that may be genuine productivity gains. Licensing like this emerges naturally from self-policing by different professional groups. Of course it can be overdone, but we shouldn't just instinctively turn up out nose at it. You've got to be careful with that sort of libertarian social engineering instinct, or you'll wreck some good emergent order!
Where I think licensing probably does the most damage is in lower skill occupations that really aren't plagued with principal-agent problems, and don't need the self-policing. Those can be real barriers to employment for low skill workers. Matt Yglesias used to post on this a lot - he was good at bringing attention to this problem.
Another great observer of American social problems highlighted the risks to low income workers posed by licensing as far back as the sixties (when licensing was less common). Does anyone know who he was?
Dan,
ReplyDeleteI know you think this is wonderful and important work, but when I read this stuff my eyes just role back in my head.
A statement like 15% wage premium is just meaningless. Who are we talking about? Barbers in a bad part of Baltimore or brokers for Goldman Sachs?
Moreover, having a license reduces the income of many people because having a license limits what services you can provide and what charges you can impose.
I practice law, which restricts what other services I can provide. For example, investment bankers make far more than lawyers who could, but for their license restrictions, provide the same service.
The same is true for healthcare.
IOW, seems to me you are making gross generalizations at best and bad ones as well
This is part of the reason why I pointed out that that 15% could be a mix of causes.
DeleteI need to read the paper itself - I was wondering exactly how they're identifying the model. I can't imagine licensing for lawyers varies a lot over time and across states (I'm sure the minutae do and that has an impact on you guys - but from the researcher's perspective there can't be all that much variation). This is probably true of a lot of these "traditional" licensed professions.
Actually, this was one hesitation I had about his chapter for our book - there wasn't a whole lot of variation in engineering licensing requirements, so it was hard to tell exactly what was demonstrated.
However - that's not true of all fields. At first glance, my intuition is to take this as a local impact for relatively lower skilled occupations whose licensing requirements vary widely across states. That's still a mixed bag, of course, but you do what you can do. Different types of education have different returns to them as well. A degree from one school is not worth the same thing as a degree from another. The dispersion of these estimates is just something that we have to make the best of.
But I do agree this nuanced thinking about exactly what it is we're looking at is good.