Paul Romer explains why we need more science PhDs than we think we do

It's always fun when a future Nobel-prize-winner comments on your blog. It's even better when the comment can be recycled as a post of its own.

First, a little context. I was annotating James Pethokoukis's list of "five ideas that both liberals and conservatives might agree on that also could actually improve economic growth" (itself a response to my cover story on the next president and the economy). The fifth of his ideas was to:

Give universities incentives to create more science geeks, and offer grad students free-floating fellowships to choose whatever field they see as the best market match for their skills.

My take was that this idea was "cool, and very Paul Romer," because the Stanford economist has been pushing this idea for more than a decade. (Romer reports that this is the first known use of his name as an adjective.) Anyway, commenter TomP wrote:

Do we really need more "science geeks"? My impression is that, for many years now, we've been producing far more science Ph.D.s than the scientific research community (academic and private) can support.

To which Romer responded:

It is a common mistake to assume that the goal of policy should be to produce a number of PhDs that is equal to the number of new positions available in universities.

Note first that the number of new positions in universities is not fixed but exhibits the usual downward sloping demand we see in any labor market. What people usually mean is to limit supply to the number of available positions at current wage rates for tenure-track faculty.

People in the university community with PhDs argue for this approach because they do not want wages for PhDs to fall. From the point of view of narrow self-interest, this is perfectly understandable, but keeping wages high for professors is not a good basis for public policy.

Most innovation takes place outside of universities. Most new PhDs start work outside of universities. If we subsidize the supply of graduate students, more people will be doing innovative work outside the university. Everything else equal, an increase in the supply of newly trained graduate students would reduce their wages and lead to more hiring by private firms. However, different forms of training would give students skills that are better matched to the work that private firms are doing. This could offset the supply effect and lead to higher wages.

(More after the break)

The way to get this better match is to provide those subsidies to graduate education as portable fellowships that put the students in control. They could choose types of degrees (perhaps not always PhD) and areas of interest (perhaps entirely new areas) that would prepare them for the many exciting lines of investigation that organizations other than universities will pursue.

People in universities often argue that the government should subsidize what they do, but not anything that goes on in the for-profit sector. Why do I think it would be a good idea to provide an indirect subsidy to innovation there? Because intellectual property rights (IPR) are always incomplete and partial, as they should be. Iron-clad IPRs would create strong monopolies that would stifle downstream innovation. So if we settle (as we should and do) for IPRs that are partial and incomplete, the private return to innovative activity is less than the social return. Whenever the social return to some activity is greater than the private return, there is a clear justification for subsidizing the activity.

From the perspective of political economy, a key practical advantage of providing these subsidies via portable fellowships is that they do not require that some government agency pick winning lines of investigation or favored firms. Nor do they create a private sector special-interest group that will capture the subsidies and direct them in ways that may not serve the public interest.

In contrast, the R&D tax credit, which also avoids the need for the government to pick winners, has created a big recurring lobbying effort, which naturally focuses first on minimizing corporate tax bills, not on maximizing innovation and the rate of growth. This has bad effects on both sides. The congress has an incentive to do short extensions of the credit because this means that members can raise funds from the interest group each time the credit comes up for renewal. Firms make contributions and lobby for provisions that reduce their tax bills but may not actually increase their innovative effort. We'd probably be better off as a nation if we scrapped the entire R&D tax credit and put all the additional tax revenue we'd collect into portable fellowships for grad students. This is something that liberal and conservative citizens might agree on, but that politicians will avoid like the plague. Like the farm lobby, the special-interest group around the R&D tax credit is now too big for any individual politician to challenge.

Hmmm, doesn't sound promising. But maybe if Pethokoukis and I started a "Romer Club" to push portable fellowships, we could shift the political balance. Because, you know, Greg Mankiw's Pigou Club has been so successful in getting politicians to support higher gas taxes.