The changing structure of American innovation: Some cautionary remarks for economic growth

new NBER theorizes that moving research off corporate balance sheets to"...Spinoffs, startups, and university licensing offices...." quote"....has not fully filled the gap left by the decline of the corporate lab...".lots of good factoids

"...NSF data indicate that share of research (both basic and applied) in total business R&D in the U.S. fell from about 30 percent in 1985 to below 20 percent in 2015 (figure 2).The figure also shows that the absolute amount of research in industry, after increasing over the 1980s, barely grew over the 20 year period between 1990 to 2010. Other data show the same decline...Utilizing data on scientific c publications, Arora et al. (2018) show that the number of publications per rm fell at a rate of 20 percent per decade from 1980 to 2006 for R&D performing American listed firms. The authors also find that the drop is even more dramatic for established firms in high quality journals. For articles within the top quartile of Journal Impact Factor scores, the magnitude of the drop increases to over 30 percent. Large firms' withdrawal from science can also be gleaned from the list of R&D 100 awards winners. Fortune 500 firms won 41 percent of the awards in 1971, but only 6 percent in 2006 (Block and Keller, 2009). Over the same period, total industry R&D and patenting grew steadily, as did university performed research (see figure 6 below). This evidence points to the emergence of a new division of innovative labor, with universities focusing on research, large firms focusing on development and commercialization, and spinoffs, startups, and university technology licensing offices responsible for connecting the two.....Anecdotal evidence suggests that neural network research published by Google brain has been implemented by follow-own research at firms. It is now standard practice among researchers to test their algorithm's performance against Alexnet or LSTM | both of which were re fined at Google. We find that ML papers published by large firms are cited more often in patents than other ML papers: large firms published 12 percent of the papers in KDD, AAAI, ICML, IJCAI, and NIPS between 2011 and 2018, but accounted for 32 percent of the papers that are cited by patents.Bikard (2015) finds corporate publications to be 23 percent more likely to be cited than university publications one the same scientific discovery. We add wider correlational evidence in support of this prediction by comparing the likelihood of a U.S. utility patent issued between 1980 and 2006 citing a corporate scientific publication versus a university counterpart in its non-patent literature section.Using a linear probability model, we estimate that corporate publicationsare on average 11 percent more likely to be cited as a university publications. We control for the possibility that these results are driven by lower-quality universities, "applied" journals, or industry level differences in scientific quality, and find that the results hold. Panel (A) of Figure 12 visualizes the citation likelihood differences between these two groups, while panel (B) shows that corporate publications first order stochastically dominate university publications in terms of the number of citations they receive from patents...we suggest that this division of innovative labor has not, perhaps, lived up to its promise. The translation of scientific knowledge generated in universities to productivity enhancing technical progress has proved to be more difficult to accomplish in practice than expected.Spinoffs, startups, and university licensing offices have not fully filled the gap left by the decline of the corporate lab.Corporate research has a number of characteristics that make it very valuable for science-based innovation and growth. Large corporations have access to significant resources, can more easily integrate multiple knowledge streams, and direct their research toward solving specific practical problems, which makes it more likely for them to produce commercial applications. University research has tended to be curiosity-driven rather than mission-focused. It has favored insight rather than solutions to specific problems, and partly as a consequence, university research has required additional integration and transformation to become economically useful. This is not to deny the important contributions that universities and small firms make to American innovation.Rather, our point is that large corporate labs may have distinct capabilities which have proved to be difficult to replace....During the so-called Golden Age of American Capitalism, large corporate labs were important loci of research, and important sources of scientific and technical advances. At the start of the period, the university research sector was small (certainly compared to the current period) and uneven in quality. Over time, university research grew, bolstered by significant support from the federal government. This period also coincided with (and perhaps this was more than a coincidence) incumbent firms enjoying significant market power but restrained by aggressive anti-trust actions. Despite the apparent successes, corporate research, and the large corporate labs in particular, fell out of favor with investors, and eventually, also with managers. The focus shifted to university research, and startups, often venture funded, that aimed to capitalize on the scientific and technical advances in university labs. Corporations turned to sourcing ideas and inventions from the outside, hoping to combine it with their downstream development and commercialization abilities. These hopes have not been fully realized, at least not yet. Even as this division of innovative labor has progressed, so have the challenges it faces become more evident. University research is different from corporate research: it is less likely to be mission-driven. Its smaller scale and greater disciplinary focus mean that university research typically produces insights which then need further development and integration to produce commercializable inventions. This requirement of converting insight to product has proved more onerous and challenging than commonly appreciated. It seems unlikely that corporate research will rediscover its glory days. The boost in employment of data scientists, machine learning experts, and even economists, in large firms would appear to prognosticate a different future. We disagree. For some time, quick wins from low-hanging fruit (such as optimizing auction or advertising formats) may cover up the problem, but the fundamental challenge of managing long-run research inside a for-profit corporation remains a formidable one. Put differently, although there are significant efficiency gains that companies have realized from hiring data scientists and economists, there are only a handful of cases of significantly new markets created from such efforts, and incumbent firms continue to rely on outside inventions to fuel their growth. In the longer run, therefore, university research will remain the principal source of new ideas for such inventions. And therefore the ongoing economic experiments of discovering efficient ways to translate scientific insights in universities into technical advances that eventually manifest in productivity growth will remain crucial to our future prosperity...."

Ashish Arora, Sharon Belenzon, Andrea Patacconi, and Jungkyu Suh, "The changing structure of American innovation: Some cautionary remarks for economic growth," National Bureau of Economic Growth, May 22, 2019, https://www.nber.org/chapters/c14259.pdf