Killing the Golden Goose? The changing nature of corporate research, 1980-2007 [pdf]
faculty.fuqua.duke.edu
faculty.fuqua.duke.edu
In my own experience, a large part of this is explained by the increasing use of trade secret protections rather than patents for the products of corporate research. This is particularly noticeable around computer science, where the academic literature in many areas is years behind the state-of-the-art in corporate research labs that is never published. The research is still being done, it is just intentionally kept invisible. The declining ability to enforce IP rights globally as a practical matter and the rise of virtualized infrastructures and services, which makes reverse-engineering more difficult, has made this an attractive research monetization strategy.
The other major trend is letting startups drive research activity and taking an equity position in the interesting ones, often as cheaper money than more traditional VCs. It is less expensive for large companies to make a venture capital investment in startups solving a problem they need solved than to do the research themselves. 40% of all venture capital is now corporate VC, which reflects this trend. It reduces R&D risk while also putting those companies much closer to potential acquisition targets.
Basically, this new structure is more efficient and of the corporate research that is being done, much less is being publicly disclosed due to the practical limits on IP protection.
"Finding that European firms display similar reductions in investment in science as American firms is not consistent with the idea that specific regulatory changes in American institutions drive the results of this paper."
This is a clever way to get at what effect SOX and Bayh-Dole have had on private large-enterprise investment in science, but the evidence and argument are a bit opaque. What counts as a "European" vs. "US" firm? Most corporations are global at the scale of interest to this argument. US regulations might still have had an effect, even on "European" firms.
Logically the best they can argue here is that the evidence is consistent with their hypothesis, not that the evidence is inconsistent with an alternative hypothesis that SOX and Bayh-Dole had an effect.
" It may well be that other organizations -- smaller firms and universities -- are making up the shortfall in investment in research. According to this interpretation, what is happening is a reallocation of research from large corporate labs to more efficient cient organizations."
The suggestion that universities might be more efficient than private corporations at anything is almost comic.
But this isn't an error. There is actually something quite profound at work. It is indeed the case that "smaller firms and universities" are more efficient at commercializing new technology than most large corporations. But it's not the smaller firms and universities independently; it's the smaller firms and universities together as one super-entity.
In Silicon Valley, the grad students quit and found a startup. Their advisor joins the Board. The university creates the IP. The startup commercializes it.
The puzzle for me ten years ago was this: Where is the licensing revenue for the universities? If they're creating the IP, then how come they're not getting anything back?
The answer is that they are, but in the form of alumni donations. :-D
It's not a priori implausible to me. Why shouldn't there be things universities do better than corporations? They're different entities with different structures; it makes perfect sense that they'd have different strengths and weaknesses.
All IMHO of course, but based on personal experience in both areas.