Sunday, March 24, 2013

Tyler Cowen shares David Sinky on Scientific Labor Markets

Insofar as I wade into policy preferences, my view of the scientific labor market is that the problems are principally demand side problems rather than supply side problems so if we should be doing anything as public policy we ought to buy science (or the derivatives of science) rather than supply more and more scientists to the market. At an NBER conference I once characterized this as the Field of Dreams strategy: "if you build it, they will come" (that made Richard Freeman laugh - a great triumph and endorsement, IMO). If you make investments in the scientific outputs you feel are lacking, the scientific labor market appears to be pretty good at getting the bodies where they need to be to make that output. The interesting market failures are on the demand side.

However, Tyler Cowen shares thoughts from David Sinky this morning presenting an alternative view with some interesting arguments. I am bolding the ones I particularly like:
"Since it seems that the supply of talented researchers in any specific area is likely fairly inelastic in the short term, to what extent do you see cash funding (as opposed to supply of talent) as a major constraint to specific scientific research in either the short and medium terms?

Even if we believe that this funding will lead to a proportional increase in clean-tech research, I suspect that returns may be quite low since the impacts of this funding would seem to be:

1. Pulling smart people from their private sector efforts into publicly funded research

2. Funding marginal projects by lower quality researchers where returns are likely to be significantly lower than average returns to research funding (which may be quite low already)

3. Increasing the funds available to established, high-status labs and researchers. If a large percentage of a lab’s output is due to the abnormally high human capital of its lead researchers, the binding constraint is their time and mental resources rather than cash so the returns on additional cash would not be very high.

4. Allowing institutions that were already going to fund this sort of research to direct additional funds to other priorities such as undergraduate academics (stem or otherwise), student amenities or other unrelated research initiatives. I suspect much of this logic also applies to donations to “cancer research charities” which I believe may be one of the single least efficient use of charitable dollars.

In general, I am disappointed that neither the right nor the left seems interested in trying to estimate the return to marginal government spending on research (either in aggregate or for specific programs).

The points above lead me to suspect it is quite low in aggregate but I’m open to being convinced otherwise if you think there is good evidence to do so."
I don't know if I agree that number three is true, but I really couldn't say. This is something for someone like Paula Stephan - who has spent a lot of time thinking about the economics of labs specifically - to weigh in on.

The first two I think are widely recognized concerns. Of course what you're funding has to be of value and the whole problem with market failures in science (and anything) is that you know there is high social value work to be done but knowing that doesn't tell you exactly what that work is.

Four is an empirical question.

I think that in theory the measurement of returns to research is important but it's very difficult, for at least two reasons. First the benefits can be far off into the future if we're thinking about basic research. Think about the economic returns to relativity. Yes, I mean Einstein - we're going back that far. The most obvious answer today is GPS and cell phones right? Relativity has had an enormous payoff but you had to wait a hundred years for the really interesting stuff. One answer to that is that after discounting the benefits are minimal in the long run, but that raises the second problem that we've seen with climate change: what's the right discount rate? Who should we care about - just ourselves or future generations? Because of the way we experience time we're stuck with a sort of temporal autarky. I can't go back 100 years and compensate someone for discovering relativity. That leads to less science than there ought to be. Should this problem be remedied or not? Do I have standing in the policy decisions of 100 years ago if I am affected by those policy decisions (as I certainly am). Do people 100 years from now have standing in today's policy decisions that affect them - or should we discount them according to our preferences?


  1. So much of the return has little to do with what is funded. We spent our efforts perfecting color tv while the Soviets developed Sputnik. It certainly wasn't beyond our capabilities to develop them, we didn't think about why we should or what the payoff would be and overestimated the cost and underestimated the benefits of doing so. Was color tv really more valuable than satellites? When the race to the moon spurred integrated circuits and computers, was the money spent on the race, money that could have been better spent on the latter, or, much more likely in my estimation, any money not spent on the race would have meant even less spent on the latter. Basic research is an ultimate social good and almost pure externality.

  2. "Since it seems that the supply of talented researchers in any specific area is likely fairly inelastic in the short term"

    The time to double the number of researchers in any area with just modest incentives is probably on the order of five years. I would not consider that inelastic. If you wanted to do a crash program and throw money at a problem the doubling time will be a lot shorter - perhaps as short as a month or two.

  3. If he is saying some research areas are over saturated and the money could be better spent in other areas, this is almost undeniable as some areas will always be more salient to the public mind but even then this may not be wasted but generate spinoffs. To argue research resources are finite and zero sum, or that market resaarch is inherently superior to charitable or government research, is quite ludicrous. It will often be duplicative and unshared, or focus on sales or trivialities, or won't be pursued at all as deemed too far afield.


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