Those of you following the literature on the economic effects of R&D may have seen a paper by Van Elk et al. on the economic effects of investment in public R&D. Those who really follow that literature well may have seen both the working paper version, and the officially published one, and have noticed a slight change in authorship between those two versions. More specifically, I was an author (named Al) in the working paper version, but in the journal version, I was just a line in a footnote. Let me explain here how that came about.
Back in 2014, I got involved in a Dutch debate on the economic value of public investments in science and universities. A main issue in that debate was whether the government’s economic think tank CPB should incorporate R&D, and specifically public R&D, into the models that they use to evaluate economic policy (plans) and political programs. Being more of a scholar than a politician, I ended up doing a research study on the economic effect of public R&D investment, together with some nice folks at CPB itself.
Our aim was to test a whole battery of econometric approaches and compare the evidence. And we managed to do that very well: even if it took us a while to do the very many regressions that were implied, we had a cool paper at the end. But as it turned out, the hardest part of that paper was the concluding section. The main conclusion that I drew from the results can be summarized as follows: “If we assume that the economic effect of public R&D investment is the same for every country in the broad set of core OECD members, we will not find many significant effects, either with a positive or negative sign. But if we use an econometric approach that assumes that every country has a distinct effect, we find positive and significant effects for most countries in the sample”.
My CPB co-authors were inclined to a somewhat different formulation, in which they wanted to stress that the results were inconclusive altogether, and that the econometric evidence cannot support any positive effect of public R&D investment on economic performance. Fair enough, we reached agreement on the text, and at the time I felt sufficiently happy with what we wrote, to go ahead and publish the working paper.
Then the paper got presented and talked about in policy audiences. And right away I realized that the nuance that I wrote (and read) into the conclusions, was completely lost. What policymakers and opinion makers picked up from the paper was pretty much that there was no evidence at all of a positive impact of public R&D. Which, incidentally, was pretty much how my co-authors would have formulated the conclusions in the first place.
At the same time, we were trying to get the paper published in a journal, and while I was getting increasingly uneasy with the way that people in policy circles talked about that paper, we were also getting a few rejections from the journals. Of course, what one does in those cases is to submit to another journal. But at one point in that whole process, I was getting fed up with all of it, and I decided that I wanted to let the paper rest (unpublished).
As the reader may already suspect, that decision did not go well with my co-authors. And frankly, I don’t blame them for that (or for anything else), it was a personal decision that was clearly connected to my position in the policy debate (and to the happy fact that I no longer need to worry about one less paper on my cv). The obvious thing was that they would go on and publish the paper on their own, and I would go my way, which in this case meant new research.
Because from my conclusion from that paper follows one very natural follow-up: if the effect of public R&D differs country-by-country, a pure time series approach applied to one country at a time makes a lot more sense than a standard panel approach. So this is what I did, together with two home-field co-authors, Luc Soete and Thomas Ziesemer, who turned out to be like-minded. Just this week we had our paper on the Netherlands (showing sizeable positive economic impact of public R&D) accepted for publication in Economics of Innovation and New Technology. So it was back on an author line after all…
Lesson learned: don't be too nuanced in policy debates, and if you need that nuance, you better stay out of the debate.