Wednesday, 21 November 2012

Is there a relation between growth and R&D? And why it doesn’t matter


A lot of interesting tweets last night from the Royal Society Innovation Debate. (Twitter mark: #innovationdebate). I couldn't go, so I am relying on the tweets I received that raised some interesting points on R&D and innovation. (There were some annoying and, IMHO opinion wrong, comments about economists which I shall return to at the end of this post).



One point raised was by @JackStilgoe "Annoyingly, national R&D spending does not correlate with economic growth. See Edgerton, To which I replied "well yes, but growth is determined by other things than R&D. Taking those into account there is a correlation IMHO". And @GordonBrianR chimed in with "R&D also needed for absorptive capacity - to stay on the knowledge frontier. There is a Red Queen effect here too.".


So what do economists know about all this?


The assertion of the non-relation between growth and R&D refers to the article by David Edgerton. Now, I consider myself a friend of David and his books and expertise in this area are unmatched. On this issue however, I beg slightly to differ. Actually, I think he's mostly talking about the correlation between growth and public R&D. So let's break it down.


There is of course no simple bivariate correlation between growth and R&D since, as mentioned above, one has to control for other things, R&D takes time to come through etc. etc. One very thorough study on this at the company level is Foray, Hall and Mairesse, Cemi-Working paper-2007-003. They criticize the assertion of no relationship made by Booz Allen Hamilton, Winter 2006, issue 45, "Strategy and Business" and show there is just such a relation at firm level as long as one is careful with data construction, accounts for the correct lags etc. An study at the country level is Griffith, Redding and van Reenen.


But let me put out another thought. Who cares if there is no relation?


To be fair, many economists are interested in the correlation between R&D in country/industry/firm A and growth in country/industry/firm A. But most are interested in an even more interesting correlation: the relation between R&D in country/industry/firm A and growth in firm B. That is to say, one might not find any correlation at all between one entity's R&D and its growth, for it might all be relying on using R&D from another place. Indeed, part of the Royal Society's honourable tradition is to foster that very information flow, by encouraging open science and communication. And if there is such a relation then we are off to the races, for the free market might under-provide research, there might be role for subsidies and public provision, mobility of scientists and absorbtive capacity might affect such transfer etc. etc. So the lack of a relation actually makes the policy issues more urgent and not less.


So, I shouldn't really say that the lack of relation does not matter, but that its not right to cite that lack of relation as an indication that policy is impotent.


If interested, here are some additional comments from an earlier post, "Innovation: A guide for the Perplexed".
Additional comment
There were some rather negative tweets about economics which may or may not have accurately recorded what the participants said. I can only say economists never sit around at Economics conferences and complain that scientists just spend their time watching apples fall from trees. Let's base criticism of other disciplines on what they actually do, not what people seem to think they do.