In the early-1990s, Kortum (1993) noted that the patent–R&D ratio in the
United States had declined steadily for over 30 years. At that time, some sug-
gested that an exhaustion of technological opportunities had reduced the produc-
tivity of corporate R&D. Others argued that expanding world markets had raised
the value of patents and that growing competition in the research sector had
resulted in greater R&D expenditures per patent. Like Griliches (1990), Kortum
(1993) simply found that rising costs of dealing with the patent system led
researchers to patent fewer of their inventions. While industry data once sup-
ported the inference of a decline in the corporate propensity to patent, more recent
data suggest the opposite. During the 1990s, there was an unprecedented surge in
corporate patenting in the United States. Using both international and domestic
data on patent applications and awards, Kortum and Lerner (1999) showed that
the recent jump in corporate patenting reflects an increase in innovation spurred
by improvements in the strategic management of corporate R&D expenditures.
The use of patent statistics in economic research has been impeded by the
fact that patents vary in their economic importance or value. Hence, simple patent
counts are less than fully informative about the economic value of innovative
output. Trajtenberg (1990) addressed this problem by examining the usefulness of
patent indicators in the context of a particular innovation, Computed Tomography
scanners, one of the most important advances in medical technology of recent
times. As in prior studies, Trajtenberg (1990) found that simple patent counts are
highly correlated with contemporaneous research and development expenditures.
Interestingly, Trajtenberg (1990) also found a close association between citation-
based patent indexes and independent measures of the social value of innovations
in that field. Moreover, the value weighting scheme appears to be nonlinear in the
number of citations, implying that the informational content of citations rises at
the margin.
Albert et al. (1991) sought early validation of citation counts as indicators of
important industrial patents and found a strong association between citation
counts for highly cited U.S. patents and knowledgeable peer opinion as to the
patent’s technical importance. A total of 20 researchers and research managers at
Eastman Kodak Research Laboratories, all of who are working in the area of
silver halide technology, were asked to rate the technical impact and importance
of each patent in overlapping sets of Eastman Kodak silver halide patents. A total
of 77 patents were selected for rating. These patents ranged from those receiving
zero citations to highly cited patents, or those patents which were cited 10 or
more times on the front pages of subsequent U.S. patents issued through 1988.
Among infrequently cited patents, there were no statistically significant differ-
ences between peer and citation ratings. In contrast, highly cited patents were
rated far more important by the evaluators. As in the case of Trajtenberg’s (1990)

Get Tech Stock Valuation now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.