In his latest post on Angry Bear, pgl seems to want to run a cost-benefit calculation on R&D spending:
Economists are guilty, however, of wanting to weigh the prospective benefits of R&D against their costs....here is the key question – how does the allocation of public funds weigh the opportunity cost of publicly funded R&D against the prospective benefits.
In the past, he's also raised the question of whether there was too much private R&D spending
R&D in the internet, communications, and technology (ICT) sector as well as the life science (pharmaceutical and biotechnology) sector may have risen. But are the higher levels welfare enhancing?
Pgl's suggestion to applying a cost-benefit model to R&D--especially to basic and early stage applied R&D--is simply unworkable. The reason is simple: R&D has a very high, very uncertain payoff. The benefits of a successful innovation--such as the Internet, or, say, a new method of cheaply converting sunlight into electricity, or a cure for lung cancer--go to everyone alive today, and to their descendants.
So if pgl wants to do cost-benefit calculations on R&D, he's in the position of making some truly heroic calculations. What is the value of the Internet? What is the social discount rate (so we know how to value the benefits to our descendants)? And how do we even think about social risk aversion?
I think the argument is much simpler. We have an obligation to our children to continue funding science and technology--especially the early stage stuff--to the degree possible. We don't want to apply cost-benefit calculations to R&D, just like we don't want to apply them to education. This is the true seed corn for the future.