Chapter 8 | Invention, Knowledge Transfer, and Innovation

Knowledge Transfer

Collaborative R&D activities among universities and colleges, businesses, and other parties have taken place in the United States throughout the 20th and early 21st century. And as federal funding of academic research expanded in the post–World War II era, academic administrations became increasingly engaged in patent management (Mowery et al. 2004). The Bayh–Dole Act (Patent and Trademark Act Amendments of 1980, P.L. 96–517) created a uniform patent policy among the many federal agencies that fund research, enabling small businesses and nonprofit organizations, including universities, to retain ownership of inventions made under federally funded research programs. The Bayh-Dole Act has since been engaged by large companies as well. It is widely regarded as having been an important stimulant since its 1980 enactment for academic institutions to pursue technology transfer activities. Other countries implemented policies like the Bayh–Dole Act by the early 2000s, giving their academic institutions (rather than inventors or the government) ownership of patents resulting from government-funded research (Geuna and Rossi 2011).

The Association of University Technology Managers (AUTM) gathers information on the invention and main patent-related activities of its member universities. Invention disclosures filed with university technology management and transfer offices describe prospective inventions and are submitted before a patent application is filed. The number of these disclosures grew from 13,718 in 2003 to 22,507 in 2015 (notwithstanding small shifts in the number of institutions responding to the AUTM survey over the same period) (Figure 8-10). Likewise, new U.S. patent applications filed by AUTM university respondents also increased, nearly doubling from 7,203 in 2003 to 13,389 in 2015. As described earlier for all U.S. academic patents, U.S. patents awarded to AUTM respondents stayed flat between 2003 and 2009, before rising to reach 6,164 in 2015 (see Appendix Table 8-26).

Data from AUTM also provide counts of new startups formed and of startups still operating, and these indicators also show an increased growth rate since 2009. New startups reached 950 in 2015 with the number of past startups still operating 4,757 in 2015 (Appendix Table 8-26). Active licenses increased from 18,845 in 2001 to 40,402 in 2015.

While license income is not the dominant objective of university technology management offices (Thursby, Jensen, and Thursby 2001), the 165 institutions that responded to the AUTM survey reported a total of $1.8 billion in net royalties from their patent holdings in 2015. This amount has grown from $754 million in 2001 (Appendix Table 8-26).

The Stevenson-Wydler Technology and Innovation Act of 1980 (P.L. 96–480) directed federal agencies with laboratory operations to become active in the technology transfer process. It also required these agencies to establish technology transfer offices (termed Offices of Research and Technology Applications) to assist in identifying transfer opportunities and establishing appropriate arrangements for transfer relationships with nonfederal parties. Follow-on legislation in the 1980s and through 2000 amending the Stevenson-Wydler Act has worked to extend and refine the authorities available to the agencies and their federal laboratories to identify and manage intellectual assets created by their R&D and to participate in collaborative R&D relationships with nonfederal parties, including private businesses, universities, and nonprofit organizations (FLC 2013).

As indicated in Chapter 4, about 11% of the current U.S. R&D total ($54.3 billion of $495.1 billion in 2015; see Table 4-1 in Chapter 4) is performed by the federal government, through federal agencies’ own research facilities and the 41 federally funded research and development centers (FFRDCs). In response to these longstanding federal policies promoting technology transfer, nearly all the agencies and their associated federal laboratories have become active in recognizing and promoting the transfer of inventions from their own R&D with potential for commercial applications.

As applied in the federal setting, technology transfer can occur through varied channels: commercial transfer (the movement of knowledge or technology developed by a federal laboratory to private organizations or the commercial marketplace), scientific dissemination (publications, conference papers, and working papers distributed through scientific or technical channels, or other forms of data dissemination), export of resources (federal laboratory personnel made available to outside organizations with R&D needs, through collaborative agreements or other service mechanisms), import of resources (outside technology or expertise brought in by a federal laboratory to enhance existing internal capabilities), and dual use (development of technologies, products, or families of products with commercial and federal [mainly military] applications).

The metrics on federal technology transfer continue to primarily track the number of activities—that is, invention disclosures, patent applications and awards, licenses to outside parties of patents and other intellectual property, and agreements to conduct collaborative research with outside parties (Institute for Defense Analyses, Science and Technology Policy Institute 2011). Nonetheless, systematic documentation of the downstream outcomes and impacts of transfer remains a challenge.^​ Also missing (until most recently) for most agencies and their laboratories are comprehensive data on technology transfer through the scientific dissemination mode (i.e., technical articles published in professional journals, conference papers, and other kinds of scientific communications), which remains widely regarded by laboratory scientists, engineers, and managers (federal and private sector) as a key means of transfer. The Department of Commerce’s (DOC’s) most recent Summary Report on federal laboratory technology transfer (with data on FY 2014, published October 2016) is expanded to include a bibliometric analysis of scientific/technical publications originating from federal laboratories (DOC/National Institute of Standards and Technology [NIST] 2016). Additional perspective on this topic is provided earlier in Table 5-25 in Chapter 5, where an original bibliometric analysis conducted for Science and Engineering Indicators contrasts the share of U.S. S&E articles in 2016 for the federal government with that for other performers.

Seven agencies account for most of the annual total of federal technology transfer activities: Department of Defense (DOD), Department of Health and Human Services (HHS), Department of Energy (DOE), National Aeronautics and Space Administration (NASA), U.S. Department of Agriculture (USDA), DOC, and Department of Homeland Security (DHS). (Each of these agencies also conducts more than $1 billion of R&D annually through its intramural facilities or FFRDCs; see Table 4-16 in Chapter 4.) Technology transfer statistics for these agencies for FY 2014 (the latest data year available), with comparisons with FYs 2006, 2009, and 2012, appear in Table 8-3. (Similar statistics for a larger set of agencies, going back to FY 2001, appear in Appendix Table 8-27.) Consistent with the agencies’ statutory annual reports, these statistics mainly cover the activity areas of invention disclosures and patenting, intellectual property licensing, and collaborative relationships for R&D.

As the distribution of the statistics across the activity types in Table 8-3 shows, most of these agencies engage in all the transfer activity types to some degree—although the emphases differ. Some agencies (e.g., DOD, DOE, HHS) are particularly intensive in patenting and licensing activities; others (e.g., DOC, NASA, USDA) are intensive on transfer through collaborative R&D relationships. Furthermore, some agencies have unique transfer authorities (statutory) that can confer practical advantages. NASA, for example, can establish collaborative R&D relationships through special authorities it has under the National Aeronautics and Space Act of 1958; USDA has several special authorities for establishing R&D collaborations other than cooperative research and development agreements; DOE has contractor-operated national laboratories, with nonfederal staff, that are not constrained by the normal federal limitation on copyright by federal employees and can use copyright to protect and transfer computer software. In general, the mix of technology transfer activities pursued by each agency reflects a broad range of considerations such as agency mission priorities, the technologies principally targeted for development, the intellectual property protection tools and policies available, and the types of external parties through which transfer and collaboration are chiefly pursued.

The data for the most recent years in this series (FYs 2012–14) indicate that federal agency laboratories and FFRDCs as a group put forth some 5,100–5,400 invention disclosures annually, 2,400–2,600 patent applications, and receive 1,900–2,200 patent awards. These numbers have generally grown over the years, which is more apparent in the longer time series of data available in Appendix Table 8-27.

Year to year, the intramural or FFRDC laboratories of DOE and DOD consistently account for the largest levels of invention disclosures, patent applications, and patent awards. For example, DOE reported 1,588 invention disclosures in FY 2014, 1,144 patent applications, and 693 patent awards; DOD reported 963 invention disclosures, 916 patent applications, and 670 patent awards (Table 8-4). In contrast, HHS, which is also one of the largest intramural or FFRDC R&D performers, reported 351 invention disclosures in FY 2014, 216 patent applications, and 335 patent awards. Further, NASA reported a high number of invention disclosures in FY 2014 but had low levels of patent applications and awards (146 and 117, respectively). This emphasizes that care must be used in comparing the track of these invention indicators over time and across agencies. Depending on the technologies involved, application areas, type of external development partners, and technology transfer authorities available—all of which vary across the federal government—the priority of attention to patenting as a main mechanism for promoting the transfer and downstream commercial development of federal laboratory inventions can differ among the agencies.

Indicators of economically valuable knowledge reflect only a portion of the knowledge about S&T that is shared. Tacit knowledge, shared through person-to-person exchanges, spreads locally and across networks of people interested in similar topics. This can take place informally and in conferences, through paid consulting and other business services, and through institutions organized for sharing knowledge and technology.

Economically valuable knowledge also spreads through publicly and freely available records, such as scientific publications, patent records, and open-source software, as well as through use of intellectual property, such as licensing of patents, copyrights, software, and trade secrets. Such documents and records are codified, or in some way formalized for transmission between people.

A key feature of knowledge is that many can use it, and it can be used repeatedly without being exhausted. It can spread or spill over to users outside the institutions where the knowledge is created. The ability for knowledge to be reused and shared across users yields great potency in fueling further economic growth (Romer 1986, 1990; Lucas 1988).

Sources of knowledge used in invention and innovation include business R&D, university and nonprofit institution research, the work of federal laboratories, and the experiences of scientists, engineers, and inventors as they create and develop new and useful products and processes. For business product innovation in manufacturing, sources outside of internal R&D labs are pervasive. Arora, Cohen, and Walsh (2016) found that for U.S. manufacturing firms, 49% reported that the invention underlying their most important innovation was external to the firm (see sidebar Open Innovation).

Coauthorship provides a means by which economically valuable knowledge can flow through collaboration with other scientists and engineers to the business sector, leading to the development of new and improved products and processes. Although the great majority of peer-reviewed S&E publications are produced by universities (described in the Chapter 5 section Publication Output, by U.S. Sector), authors with business-sector affiliations produced more than 51,000 publications in 2016 (Table 8-5), over 80% of which were coauthored with academic, government, or foreign researchers. Reflecting the importance of collaboration with academic researchers, almost half (49%) of all business publications were produced with authors from U.S. academic institutions. Government coauthors appear on 13% of all business publications, and foreign coauthors appear on more than a third of all business publications (35%).

In addition to co-authorships, citations of S&E articles in patent documents provide indicators of economically-valuable knowledge as inputs to invention. Patent documents accessed from USPTO provide text citations to earlier patents issued (prior art) and to nonpatent literature (NPL), which includes peer-reviewed research and other published documents.

As an indicator of knowledge transfer, the linkages can be indirect. Earlier patents may be cited by the inventor to demonstrate their difference from prior art or added by the examiner to limit the scope of the patent (IEEE 2010). Citations to NPL are considered stronger indicators of the impact of academic research on business patenting than citations to patents, though both miss flows from private and contract research, as well as flows from basic research (Roach and Cohen 2012).

Almost a quarter (23%) of USPTO patents issued in 2016 cite S&E articles (Table 8-6), with almost 300,000 S&E articles cited. Six fields of science accounted for nearly all (98%) of the citations in USPTO patents granted in 2016 (Appendix Table 8-28). Biological sciences make up the largest share (34%), followed by medical sciences (24%), computer sciences (12%), engineering (11%), chemistry (9%), and physics (8%) (Figure 8-11).

Across fields, the authors of most cited S&E literature in patent documents are from the academic sector. Consistent with its large share of S&E publications and citations overall, the U.S. academic sector received 31% of NPL citations from all USPTO patents in 2016 and 67% of citations from patents granted to U.S. patent owners. Within fields of science, industry publications receive 20% or more of the patent citations in computer sciences, engineering, and physics (Figure 8-12). Articles from other nonacademic sectors receive far fewer citations in patents, but this varies by field. After academia, industry articles capture the next largest share of citations overall, with particularly high citations in computer sciences (27%), physics (23%), and engineering (21%). In medical sciences, industry and nonprofit articles each account for 10% of patent citations. Compared with other fields, federal government S&E articles receive the largest number of citations in biological and medical sciences (each 5%), and FFRDCs receive the largest number of citations in physics (8%).

The globalization of USPTO patents is reflected in the foreign sources of cited articles and in the foreign share of USPTO patents described earlier, in the section USPTO Patenting Activity. In 2016 foreign articles drew more citations in USPTO patents (54%) than U.S. articles (46%) (Table 8-6).

Licensing allows intellectual property developed within firms to be used externally and globally active businesses transfer their intellectual property across national boundaries, exploiting opportunities in external markets. This intellectual property includes the use of proprietary rights—patents, trademarks, copyrights, industrial processes, and designs—and licenses to reproduce and/or distribute intellectual property embodied in produced originals, prototypes, live performances, and televised broadcasts (World Trade Organization 2016).

The export revenues for these types of transactions, known as “charges for the use of intellectual property,” provide a broad indicator of technology flows across the global economy and the value of an economy’s intellectual property in the international marketplace.^​ Receipts from other countries for this trade provide a partial measure of market-based income for the use of intellectual property. International receipts for the use of intellectual property also represent global exports of services, playing an important role in understanding the global balance of trade. However, such receipts are a partial indicator of these flows. The volume and geographic patterns of U.S. trade in royalties and fees have been influenced by U.S.-based multinational companies transferring their intellectual property to low-tax jurisdictions or their foreign subsidiaries to reduce their U.S. and foreign taxes (Gravelle 2010:8; Mutti and Grubert 2007:112).

Global exports (receipts for the use of intellectual property) were $272 billion in 2016 (Appendix Table 8-29). The United States was the world’s largest exporter (45% global share) with a substantial trade surplus (Figure 8-13).^​ However, over several years the U.S. global share has fallen from 54% in 2008 to 45% in 2016.

The EU is the second largest, with a global export share of 24%, but it has a substantial deficit. After falling from 26% to 20% between 2008 and 2012, the EU share rose to reach 24% between 2013 and 2016. Japan, the third largest (14% share), has a substantial trade surplus. Japan's global export share has remained stable between 2008 and 2016. For developing countries, receipts for the use of intellectual property are very low; for example, the global export shares of China and India were less than 0.5% in 2016 (Appendix Table 8-29).