Science and the Market for Technology
- Date Posted:
- Is Database:
- Database
Federal R&D spending reallocation from defense to fields like computer science, oceanography, and biology has led to significant shifts in research output and patent value.

Ashish Arora, Sharon Belenzon and Jungkyu Suh, "Science and the Market for Technology," National Bureau Of Economic Research, March 2021, https://www.nber.org/papers/w28534
New NBER using the collapse of the USSR as a natural experiment in terms the response of/the shock to American scientific spending as the USG reallocated Federal R&D spending away from defense (so physics, chemistry, and electrical engineering) and towards computer science, oceanography and biology (See Figure A2) Relative to the counterfactual paper finds (intuitively) more funding resulted in more papers, and less funding leads to fewer papers. Using patent citations/and tradability as a measure for usefulness the paper finds the share of patents citing scientific papers was correlated with the uptick/downtick in the number of paper by field (so driven by federal R&D spending) This in turn produced higher quality innovation in terms of the patents likelihood to be traded by the firm that owns it, "..Patents that reference a scientific article are 16-23% more likely to be traded than patents that do not reference science..."
Good factoid to demonstrate innovation has become more science intensive"...At the same time, the reliance of inventions on science has increased, as indicated by the rise in the share of patents citing science from 4% to 28% of all U.S. utility patents between 1980 and 2015..."
Core findings, first “…a patent that cites a scientific publication has a 23% higher probability of being traded compared to a patent that does not cite science. After wecontrol for quality using number of claims, triadic patenting status, (and stock market value for a subset of patents matched to listed firms), the magnitude of the science-MFT coefficient is reduced by a third, but still positive and statistically significant. We interpret this as suggesting that reliance on science is related to MFT in other ways, not just through quality…”
Second, “…consistent with the view that science increases gains from trade,we find that the relationship between science and MFT is four times larger for small firms compared to large firms. Since larger firms already have high commercialization capabilities (and hence derive lower value from selling), reliance on science will have a smaller effect on the gains they can reap from selling their invention compared to smaller firms.We also find that the science-MFT relationship is about a third larger for patents using a one standard deviation “newer” combination of technological subclasses (Fleming, 2001). Unfamiliar inventions are likely to be more difficult for buyers to evaluate, but when such inventions are based in science, buyers may gain a better understanding of otherwise unfamiliar technological components. We also find that reliance upon more recent science is more strongly associated with patent trade. This suggests that the principal channel through which science affects technology markets is not by reducing transfer costs, since mature science is more likely to provide a “common language” or background knowledge for market participants (Arora and Gambardella, 1994). Newer science, and more specialized science, on the other hand, is arguably more important in illuminating uncharted applications of inventions, thereby increasing gains from trade…”
Third, “… we present instrumental variable results that are based on an exogenous change in the availability of science, which affects the prevalence of science-based inventions unrelated to trade conditions. After the fall of the Soviet Union, federal funding for research fell by almost 40% for Space and Defense between 1986 and 1992, whereas it doubled in Medical and Energy. These changes led to substantial variation in the supply of new knowledge, and therefore, variation in the reliance on science in different fields of invention.The geopolitical circumstances that precipitated the end of the Cold War are exogenous to MFT. We find that patent classes with a standard deviation increase in the share of science-based patents experience a 6.5% increase in the share of reassigned patents. Our IV estimates are 13% smaller than our OLS estimates, but still positive and statistically significant….”




Ed Comment:“My thinking isn't clear on this but my sense from being on the waters board is that academic research is woefully unproductive relative to commercial research. Yet academies siphone off an enormous amount of talent. So the gov funding more academic research may make the academics more productive in response to the funding than they otherwise would be. But in general, gov funding of academic research makes talented people less productivity because it makes the academies larger than they otherwise would be. Is this a variation of the Simpson paradox?”