Thread by @Noahpinion

1/So, I want to write a thread about this paper, which was interesting and had some good history, but which I ultimately found unconvincing.

The paper claims that the decline of corporate labs is (partially?) responsible for the productivity slowdown.

nber.org/chapters/c1425…

2/I thought there was going to be some evidence linking the productivity slowdown to the decline of corporate labs.

Instead, it's just the most hand-wavey thing imaginable.

3/In fact, the timeline doesn't even match up, since the paper identifies the 1980s as the time when corporate labs declined, while productivity slowed down in the 1970s (then accelerated in the 90s, then declined again in the 00s)!

4/Second, the article defines the decline in corporate research in terms of *percentages* - of sales, of R&D spending, of # of companies, etc. But the absolute *number* of corporations publishing in scientific journals seems to have increased...

5/If ideas are really nonrival (and remember, these are papers published in journals, not patented secrets!), shouldn't we expect this raw increase to mean corporations are doing *more* to boost productivity via lab research now, even though the percentages have declined?

6/Next, the paper suggests that corporate lab science just didn't pay off for the corporations that did the science (Xerox, Bell, etc.), because the ideas were nonrival.

This is exactly what basic economic theory would predict.

7/Basic economic theory says corporations won't do much scientific research because the ideas will leak out and the corporations won't be able to capture the value of their investments in science. Hence, science needs to be done by the government.

8/The story of the rise and fall of corporate labs seems to follow exactly this narrative. As information technology increased and dissemination of scientific knowledge became easier and quicker, corporate labs became less valuable to corporations. Hence, they declined.

9/Because of this, economists recommend having universities do the science and then transfer their insights to companies to do the product innovation - basically the pipeline that @MazzucatoM lays out.

10/The paper repeatedly claims that this university-to-industry pipeline doesn't work very well.

But it gives no evidence to back up that claim.

11/Finally, the paper deals only with the United States.

But the productivity slowdown of the last 15 years has been global in nature.

Is a substantial part of that really due to the decline of corporate labs in the U.S.? Seems doubtful!

12/So although the paper has a lot of interesting historical tidbits and some useful theoretical musings about the strength of corporate scientific research, it ultimately fails to make its case that corporate labs are economically better than university-based research.

(end)