HOW ACADEMIC RESEARCH INFLUENCES INDUSTRIAL INNOVATION
The relationship between academic research and industrial innovation is a topic of great interest to policy makers and has received considerable attention from researchers in recent years. Although this relationship often is (mis)conceptualized as one in which “academia invents and industry invests” in commercialization, in fact the relationship may be better understood as an interactive one that is supported by flows of ideas and people from universities to industry and vice versa. Indeed, industrial research may lead and influence the agenda of academic research in some fields, as was the case in the early stages of research on light-emitting diodes (LEDs) and semiconductors (Mowery, 2011). According to Lécuyer (2005), Provost Frederick Terman of Stanford University encouraged William Shockley to locate his new firm near the university in 1955 to expose Stanford’s engineering faculty to new research in solid-state physics and electronics, and a future dean of Stanford’s Engineering School served an apprenticeship of sorts at Shockley Semiconductor to better understand semiconductor fabrication and design.2 As noted by Mowery (2009b, p. 6):
1The committee was assisted by David C. Mowery in drafting this appendix.
2“[James] Gibbons [future dean of engineering at Stanford], a junior faculty in the electrical engineering department [sic] at Stanford, worked at Shockley Semiconductor on a part-time basis. Frederick Terman, Stanford’s provost, and John Linvill, the head of the
The movement of researchers between industry and academia facilitates this interactive relationship (e.g., the move by Dr. Shuji Nakamura, a pioneering researcher in gallium-arsenide LEDs, from Nichia Chemicals in Japan to the University of California, Santa Barbara [UCSB] in 2000; see Chapter 7 of Mowery et al. , for further discussion).3
Although the movement of researchers between industry and academia in the United States hardly could be described as frictionless, the boundaries between these different institutional venues for research and innovation are relatively porous in the U.S. research system. This represents a source of strength that may well distinguish the U.S. system from that of other industrial economies (although indicators on this point are difficult to develop).
As further described by Mowery (2009b), other studies have examined the influence of university research on industrial innovation. All of these studies (Cohen et al., 2002; Government-University-Industry Research Roundtable, 1991; Levin et al., 1987; Mansfield, 1991) emphasize differences among industries in the relationship between university and industrial innovation. The studies by Levin and colleagues (1987) and Cohen and colleagues (2002) summarize industrial research and development (R&D) managers’ views on the relevance to industrial innovation of various fields of university research. Virtually all of the fields of university research rated by industrial respondents in both surveys as “important” or “very important” for their innovative activities were related to engineering or applied sciences, fields of U.S. university research with a long history of university-industry collaboration. Industry respondents consider few fields of university basic science research, aside from chemistry, as relevant to their innovative activities.4
Solid-State Laboratory, had recently apprenticed Gibbons to William Shockley. They had asked Gibbons to learn the techniques required for the fabrication of silicon devices from Shockley and then transfer these techniques back to the university. This was not the first time that Terman had sought to appropriate process technologies from local firms.” (Lécuyer and Brock, 2006, p. 138).
3The academic research agenda in semiconductors and software at both the University of California, Berkeley, and the University of California, Santa Barbera, for example, benefited from the recruitment by academic departments of faculty from industry in both the United States and Japan. Equally important contributions to academic research flowed from faculty sabbaticals in industry and industry researchers’ sabbaticals at universities (see Kenney and Mowery, forthcoming, for further examples). The OECD study Benchmarking Science-Industry Relationships, which emphasizes the importance of researcher mobility in strengthening such relationships, focuses primarily on flows of researchers from universities to industry in its benchmark measures (OECD, 2002).
4As noted by Mowery and Sampat (2005), however, the absence of such fields as physics, biology, and mathematics in these survey responses should not be interpreted to mean that academic research in these fields makes no contribution to technical advances in industry.
Cohen and colleagues (2002) also examined the relative importance for industrial innovation of different channels of communication linking intrafirm R&D to R&D in government or university laboratories. They found that pharmaceutical executives assigned greater importance to patents and licensing agreements involving universities and public laboratories than did other research managers in other industries. In these other industries, patents and licenses for inventions from university or public laboratories were reported to be less important than publications, conferences, informal interaction with university researchers, and consulting (similar findings are reported by Agrawal and Henderson , Mowery and Sampat , and Nelson ). Pharmaceutical executives assigned greater importance than R&D managers from other industries to patents and license agreements involving universities and public laboratories, but pharmaceutical industry respondents, like those from other industries, rated research publications and conferences as a more important source of information than patents and licenses.
The consistency in the findings of the study by Levin and colleagues and the more recent survey conducted by Cohen and colleagues is striking. These studies also indicate that the relationship between academic research and industrial innovation in the biomedical field differs from that in other knowledge-intensive sectors (Mowery and Sampat, 2005). Academic research rarely produces “prototypes” of inventions for development and commercialization by industry, but academic research does inform the methods and disciplines employed by firms in their R&D facilities. Industrial R&D managers rely on a variety of channels for learning about and exploiting the results of academic research, and the channels rated by industrial R&D managers as most important in this complex interaction between academic and industrial innovation rarely include patents and licenses (Mowery and Sampat, 2005).
The work of Cohen and colleagues (2002) and other scholars singles out six channels through which industrial innovation is influenced by university research:
(1) faculty publishing;
(2) university patenting and licensing;
(3) faculty consulting;
(4) faculty entrepreneurship, including the establishment by faculty
Instead, these results reflect the fact that the effects on industrial innovation of basic research findings in such areas as physics, mathematics, and the physical sciences are realized only after a considerable lag. Moreover, application of academic research results often requires that these advances be incorporated into the applied sciences, such as chemical engineering, electrical engineering and material sciences.
or other university researchers of new firms to commercialize their inventions;
(5) informal interactions between university and industry researchers, including conference presentations and related interactions; and
(6) training and placement of students in industrial positions.
The recognition by scholars of the array of channels through which universities interact with industry has only begun to inform understanding of the relationships among these seemingly parallel channels. A better understanding of these relationships, as well as the links among these different channels of interaction, is essential to the design of policies intended to enhance the contributions of university research to industrial innovation. For example, if an emphasis on patenting and licensing has a chilling effect on faculty’s open disclosure of research results, the overall contributions of university research to industrial innovation could be reduced in the face of expanded patenting and licensing of university research advances. Understanding the relationships among these different channels of interaction is further complicated by the fact that their relative importance differs among fields of technology. The following sections summarize some of the recent scholarly research on several of these different channels of interaction between academic research and industrial innovation.
FACULTY CONSULTING AND ACADEMIC RESEARCH
A number of studies have examined the role of faculty consulting in patenting and technology transfer to industry. A series of papers by Zucker, Darby, and collaborators (Zucker et al., 1998, 2002) focuses on the collaboration with industry of “star scientists” in biomedical fields at universities. The studies examine the effects of such collaboration (such as serving on a scientific advisory board and collaborating with scientists in the firm) on the performance of biotechnology firms, arguing that, because of the “noncodified” nature of essential knowhow in biotechnology, collaboration that is mediated by the labor market for star scientist consultants is far more important than formal channels of technology transfer such as licensing. For Zucker and Darby, faculty consulting and entrepreneurship are more important than patenting and licensing in university-industry technology transfer. But these studies do not directly compare the effects of links between star scientists and firms with the effects on firms of licensing university intellectual property. In many cases, the faculty consultant and the licensed technology are likely to be complements, since the star scientist may be the developer of the technol-
ogy being licensed. But this complementary relationship raises the possibility that at least some of the effects of the consultant relationships highlighted by Zucker and Darby reflect the presence of a technology license.
A study of faculty consulting and patenting by Thursby and Thursby (2007) develops a different perspective on the relationship between faculty consulting and university technology licensing, focusing on the substantial fraction (nearly 38 percent) of a sample of patents that have faculty members at 87 U.S. universities listed as inventors but are not assigned to the faculty’s university employers. The large share of faculty patents that are not assigned to universities reflects in part the considerable variation among U.S. universities in the regulations governing the assignment of patented intellectual property.5 This study also found that the biomedical sciences exhibit the lowest share of nonuniversity-assigned patents, while engineering fields show the highest such share.
Thursby and colleagues conclude from their econometric analysis that the patents not assigned to faculty inventors’ universities are associated with faculty consulting, although their evidence on this point is indirect, based as it is largely on what they describe as the “more incremental” character of the patents not assigned to universities. The results of this study suggest a relationship between faculty consulting and patenting that differs from that emphasized by Zucker and Darby. For Zucker and Darby, faculty patenting and faculty consulting (as well as other forms of paid interaction with firms) are substitutes. The analysis by Thursby and Thursby, however, suggests that consulting and some forms of patenting (more incremental patents that are assigned to firms) may be complements, although these authors did not examine the relationship in this study between different assignment patterns for individual faculty—that is, whether faculty who are listed as inventors on numerous firm-assigned patents also were contributors for large numbers of university-assigned patents.
FACULTY PATENTING AND SCHOLARLY PUBLISHING
A related body of empirical work examines the relationship between patenting and another important channel for university-industry interaction—publishing. Agrawal and Henderson (2002) and Cesaroni and colleagues (2005) examine the relationship between publishing and patenting for individual faculty members at U.S. research universities.
5Some institutions do not require such assignment for intellectual property developed without the use of academic facilities, while others, such as the University of California, assert a right of ownership over all intellectual property developed by faculty, staff, or students, regardless of the extent of use of university facilities in its development.
Agrawal and Henderson examined this relationship for Massachusetts Institute of Technology (MIT) faculty in two engineering departments, while Cesaroni and colleagues looked at a larger sample of faculty from a number of U.S. universities. Both studies found that higher levels of patent productivity do not reduce publication productivity. Agrawal and Henderson found no relationship between the two spheres of productivity, although they concluded that faculty whose publications are more highly cited (i.e., whose research had a greater impact) appear to patent more extensively, and Cesaroni and colleagues found that higher levels of patenting were associated with higher levels of publication productivity.
Of interest, Cesaroni and colleagues found that only university-assigned patents issued to faculty were associated with higher levels of publication productivity, providing additional evidence in support of the contention of Thursby and Thursby (2007) that faculty patents not assigned to universities are associated with consulting. The positive relationship between publication and patent productivity also appears to taper off at higher levels of patent productivity, suggesting that very intensive faculty patenters may indeed be slightly less productive in publishing; however, these diminishing returns are seen at fairly high levels of patenting (the peak occurred at nine patents per researcher).
Overall, this and other evidence (see Azoulay et al., 2009) indicate that faculty patenting and faculty contributions to the open scientific literature are if anything complementary. There is no evidence in this empirical work that faculty patenting below extremely high levels is associated with a diminution in scholarly publishing. The evidence on the impact of faculty patenters’ research (as measured by citations of their work in subsequent papers), however, is more complex: there are at least some indications that patenting may slightly reduce citations of publications describing research advances that are subsequently patented.
FACULTY ENTREPRENEURSHIP AND SCHOLARLY PRODUCTIVITY
Although the above findings on patenting and productivity shed considerable light on the relationship between two important channels for university-industry interaction, an equally important issue, and one on which little research has yet been published, concerns the relationship between faculty entrepreneurial activity (including but by no means restricted to consulting) and scholarly productivity. If anything, the efforts of faculty to assist in the foundation and early-stage growth of firms are likely to impose greater demands on their time and distractions from academic research than is true of patenting, which arguably draws on many of the same skills as publishing.
One of the few attempts to examine the relationship between faculty entrepreneurship and research activity is that of Ding and Choi (2011), who looked at the participation of faculty as founders or members of scientific advisory boards (SABs) for firms in the U.S. biotechnology industry that had successful initial public offerings during 1972-2002. Although Zucker and Darby include faculty participation in SABs as one of their measures of linkage between star scientists and firms, the Ding and Choi analysis includes a much broader sample of firms, faculty, and universities; compares the participation of faculty in SABs with their involvement as firm founders; and examines the relationship between both types of entrepreneurial activity (which in fact differ substantially in content and time demands) and faculty research productivity. The results of this study suggest that SAB participation and faculty involvement as firm founders are if anything inversely related, both in frequency and in terms of the point during faculty career trajectories at which they occur (older faculty at more prestigious universities are more likely to be SAB members).6 Another important and interesting finding is the lack of any strong negative relationship between publication activity and participation by faculty in either form of entrepreneurial activity. Contemporaneous research productivity appears to be more strongly correlated with faculty involvement in the founding of a firm, perhaps as a vehicle for commercial exploitation of an important research advance, but neither type of involvement with firms appears to significantly depress research activity. Indeed, faculty involved as SAB members are more likely to have larger cumulative publication stocks. Like the work of Zucker and Darby discussed above, the Ding and Choi analysis suggests that research and faculty entrepreneurship are complementary.
The links between faculty-founded firms and university licensing are surprisingly unclear in most available data, including data compiled by the Association of University Technology Managers (AUTM) (2001, 2002). The AUTM data suggest that firms founded specifically to commercialize licensed technology account for a minority of university licensees. The AUTM annual reports for 2001 and 2002 indicate that 14-16 percent of university patent licensees in these years were start-up firms founded to exploit the licensed inventions. More than one-half (50-54 percent) of academic licensees during this period were small firms (fewer than 500
6One reason for this observed relationship between seniority and SAB activities among faculty is the likelihood that junior faculty may pursue consulting or other entrepreneurial activities with industry (e.g., participating in the establishment of a new firm or scientific collaboration with a start-up firm).
employees) already in existence, while roughly one-third of licensees (32-33 percent) were large firms.7
The emphasis in recent academic research (DiGregorio and Shane, 2003) on the role of university “spin-offs” in the licensing activities of U.S. universities thus needs to be qualified by recognition that such start-ups are less significant as licensees than large firms in absolute numbers. Surprisingly, in view of the large amount of research in this broad field that has examined academic spin-off firms (Nerkar and Shane, 2003; Shane and Stuart, 2002), little is known about the relationship between technology licensing and the formation, growth, or survival of such firms. Little information exists, for example, on the share of academic spin-offs that are also technology licensees, and the AUTM data suggest that such spin-offs account for a small share of all licensees of university patents.
Indeed, the role of faculty entrepreneurs who establish new firms to commercialize university inventions and the role of university spin-offs in regional economic development both merit critical scrutiny (Mowery, 2011). The analysis of faculty patents not assigned to universities by Thursby and Thursby (2007) suggests that a significant share of the patents of those faculty who are financially involved with these firms are not licensed from their university but assigned directly to spin-off firms. This characteristic of patents not assigned by faculty to their university does not of course preclude the possibility that a foundational patent or patents may have been licensed by the spin-off to which other patents have been directly assigned by the faculty inventor, but it assuredly points to a more complex relationship. Because of the lack of comprehensive data, surprisingly little is known about the origins of the technological innovations that are central to the university spin-offs founded by faculty or about the role of licensed inventions in these firms’ foundation and success or failure, and even less is known about the role of university faculty in the management of these firms (Mowery, 2011).
U.S. UNIVERSITY-INDUSTRY RESEARCH COLLABORATION AND TECHNOLOGY TRANSFER BEFORE THE BAYH-DOLE ACT8
Starting in the earliest decades of the 20th century, university-industry collaboration in the United States was facilitated by the unusual structure
7The AUTM survey data report only the characteristics of licensee firms for all licenses, both exclusive and nonexclusive. These data do not enable examination of the relative importance of exclusive vs. nonexclusive licenses among different types of firms (e.g., start-up, small company, large company).
8This discussion draws on Mowery et al. (2004).
of the nation’s higher education system, which contrasted with those of other industrial economies. As Mowery (2007, p. 165) notes:
The U.S. higher education system was significantly larger, included a highly heterogeneous collection of institutions (e.g., religious and secular, public and private, large and small, and so on); lacked any centralized, national administrative control; and encouraged considerable interinstitutional competition for students, faculty, resources, and prestige (see Geiger, 1986, 1993; Trow, 1979, 1991, among other discussions). In addition, the reliance of many public universities on local (state-level) sources for political and financial support further enhanced their incentives to develop collaborative relationships with regional industrial and agricultural establishments. The structure of the U.S. higher education system thus strengthened incentives for faculty and academic administrators to collaborate in research and other activities with industry—and to do so through channels that included much more than patenting and licensing.
Although a growing number of U.S. universities had adopted formal patent policies by the 1950s (Mowery, 2007), many of these policies, especially those at medical schools, prohibited patenting of inventions, and university patenting was less widespread than was of the case in the post-1980 period. According to Mowery and Sampat (2005, p. 119):
The decade of the 1970s, as much as or more so than the 1980s, represented a watershed in the evolution of U.S. university patenting and licensing. U.S. universities expanded their patenting, especially in biomedical fields, and assumed a more prominent direct role in managing their patenting and licensing activities, supplanting the Research Corporation. Agreements between individual federal agencies and universities [institutional patent agreements (IPAs)] also contributed to the expansion of patenting during the 1970s. Private universities in particular also began to expand their patenting and licensing rapidly during this decade.
Several factors appear to have contributed to the new approach taken by many U.S. universities to managing their intellectual property. Among the most important of these factors during the 1970s was slower growth in federal funding for university research, reflecting reductions in defense-related funding for university research, which particularly affected both MIT and Stanford. For financially pressed universities, reduced growth in federal research funding increased the attractiveness of the potential revenues associated with licensing these research advances. Interest in patent licensing revenues among faculty and administrators grew in parallel with their dissatisfaction with the performance of the leading institutional “agent” charged with responsibility for handling many universities’ patenting and licensing transactions—the Research Corporation.
A second important factor in U.S. universities’ increased interest in patenting and licensing faculty research advances during the 1970s was new scientific discoveries that appeared to hold considerable promise for licensing to industry. Federally funded academic research in the life sciences was an important catalyst for these discoveries. Much of the academic research that generated fundamental scientific advances in the field of molecular biology, laying the foundations for the biotechnology industry, was funded by the National Institutes of Health (NIH) as part of the Nixon Administration’s “war on cancer,” which led to increased funding for biomedical research during the 1970s. Several of the private research universities that were experiencing reductions in federal defense-related research funding (e.g., Stanford) housed academic medical centers that were the locus of significant basic research advances in the field of molecular biology. In contrast to many basic research advances, these discoveries appeared to leading pharmaceutical firms and other enterprises to hold enormous commercial promise, creating a potentially strong and lucrative market for licenses to biomedical intellectual property.
The influence of these factors is revealed in the decision by Stanford University to patent and license the Cohen-Boyer rDNA technique, which was based on research conducted at Stanford and the University of California, San Francisco.9 Stanford had established a technology licensing office in 1970 under the direction of Neils Reimers, who learned of the Cohen-Boyer invention from a New York Times article published in 1978 (Hughes, 2001). Stanley Cohen was initially opposed to patenting the technology, but Reimers convinced him that patenting the discovery would spur industrial application.10 Reimers also argued that “the patent, if granted, might become the impressive royalty generator that the university had thus far never had” (Hughes, 2001, p. 561). Such a “royalty generator” was particularly important in the face of flat growth in overall federal research funding and significant cutbacks in defense-related support for academic research. As another Stanford administrator noted in a letter to Donald Frederickson, then director of NIH: “It is a fact that the financing of private universities is more difficult now than at any time in recent memory…we cannot lightly discard the possibility of significant income [from the invention]” (quoted in Hughes, 2001, p. 564).
9Although the technique was developed jointly by researchers at Stanford and the University of California, Stanford managed the patenting and licensing process because it had an IPA with NIH, which funded the research (see below).
10Nonetheless, in a discussion of the Cohen-Boyer licensing strategy, Reimers argued that “whether we licensed it or not, commercialization of recombinant DNA was going forward. As I mentioned, a nonexclusive licensing program, at its heart, is really a tax. . . . But it’s always nice to say “technology transfer” (Reimers, 1998).
THE BAYH-DOLE ACT OF 1980
As reported by Mowery and Sampat (2005, p. 119),
The Bayh-Dole Patent and Trademark Amendments Act of 1980 provided blanket permission for performers of federally funded research to file for patents on the results of such research and to grant licenses for these patents, including exclusive licenses, to other parties. The Act facilitated university patenting and licensing in at least two ways. First, it replaced a web of IPAs that had been negotiated between individual universities and federal agencies with a uniform policy. Second, the Act’s provisions expressed congressional support for the negotiation of exclusive licenses between universities and industrial firms for the results of federally funded research.
In addition, the act reduced the power of federal funding agencies to oversee the terms of licensing agreements between research performers and licensees.
Lobbying by U.S. research universities was one of several factors behind the passage of the Bayh-Dole Act in 1980. The act is as much an effect as a cause of expanded patenting and licensing by U.S. universities during the post-1960 period (Mowery and Sampat, 2005). The IPA regime, along with similar programs at the Department of Defense, had facilitated growth in university patenting and licensing during the 1970s. Nevertheless, by the late 1970s, many of the U.S. universities active in licensing were concerned about potential restrictions on their licensing policies imposed by federal agencies. In August 1977, the Office of the General Counsel of the U.S. Department of Health, Education, and Welfare (HEW) expressed concern that university patents and licenses, particularly exclusive licenses, could contribute to higher health care costs (Eskridge, 1978). HEW ordered a review of its patent policy, including a reconsideration of whether universities’ rights to negotiate exclusive licenses should be curtailed.11 During the ensuing 12-month review, the agency deferred decisions on 30 petitions for patent rights and three requests for IPAs.
According to Broad (1979, p. 476), in response to HEW’s review of its patent policies, “universities got upset and complained to Congress.” A former Purdue University patent attorney, Norman Latker, who had been an architect of the changes in HEW’s patent policies in 1968 that led to the creation of IPAs, was fired from HEW after denouncing the agency’s
11According to the testimony of Comptroller General Elmer Staats during the Bayh-Dole hearings, the purpose of the HEW review was “to make sure that assignment of patent rights to universities and research institutes did not stifle competition in the private sector in those cases where competition could bring the fruits of research to the public faster and more economically” (U.S. Senate Committee on the Judiciary, 1979, p. 37).
subsequent review of these policies. Latker asked Senator Birch Bayh of Indiana to develop legislation liberalizing and rationalizing federal policy toward university patents on federally funded research.12 At the same time, technology transfer officials at Purdue complained to Bayh about difficulties in obtaining rights to patents funded by the U.S. Department of Energy (Stevens, 2004). Latker, together with these other university licensing officials, aided in drafting portions of what became the Bayh-Dole Act. As described by Mowery (2009b, p. 9),
The passage of the Bayh-Dole Act was one part of a broader shift in U.S. policy toward stronger intellectual property rights.13 Among the most important of these policy initiatives was the establishment of the Court of Appeals for the Federal Circuit (CAFC) in 1982. Established to serve as the court of final appeal for patent cases throughout the federal judiciary, the CAFC soon emerged as a strong champion of patentholder rights. But even before the establishment of the CAFC, the 1980 U.S. Supreme Court decision in Diamond v. Chakrabarty upheld the validity of a broad patent in the new industry of biotechnology, facilitating the patenting and licensing of inventions in this sector.
Any assessment of the effects of Bayh-Dole thus must take into account the effects of the (nearly simultaneous) shift in U.S. policy toward intellectual property rights, as well as the effects of growth in NIH funding in molecular biology and related fields before and after 1980.
THE EFFECTS OF BAYH-DOLE
Noting that it is impossible to separate the effects of Bayh-Dole from those of other developments in policy during the early 1980s, it is interesting to consider how U.S. university patenting has changed since 1980.14
12Latker returned to HEW’s patent office in 1978, after his dismissal was overturned by a civil service review board on procedural grounds. Reporting on these events in Science, Broad (1979, p. 476) noted: “The reinstatement is timely. Support is now building for the Bayh-Dole patent bill, and Latker’s return to the HEW is seen by many university researchers and patent transfer fans, to whom Latker is something of a hero, as a shot in the arm for their cause.”
13According to Katz and Ordover (1990), at least 14 congressional bills passed during the 1980s focused on strengthening domestic and international protection for intellectual property rights, and the Court of Appeals for the Federal Circuit created in 1982 has upheld patent rights in roughly 80 percent of the cases argued before it, a considerable increase from the pre-1982 rate of 30 percent for the federal bench.
14The Bayh-Dole Act sought to accelerate the commercialization of federally funded research across all institutional performers, including federal laboratories. Surprisingly, in light of the flood of empirical studies of university technology transfer since 1980, studies have attempted to examine the response of the U.S. federal laboratories to the act.
Universities increased their share of patenting from less than 0.3 percent in 1963 to nearly 4 percent by 1999, but the rate of growth in this share appears to have accelerated before rather than after 1980 (Mowery, 2007). Also noteworthy is the distribution of university patents among technology fields during the pre- and post-Bayh-Dole periods. While nonbiomedical university patents increased by 90 percent from 1968-1970 to 1978-1980, biomedical university patents increased by 295 percent (Mowery, 2007). The increased share of the biomedical disciplines within overall federal academic R&D funding, the advances in biomedical science that occurred during the 1960s and 1970s, and industry’s interest in the results of this biomedical research all affected the growth of university patenting during this period (Mowery, 2007).
Following passage of the Bayh-Dole Act, universities become increasingly involved in the management of patenting and licensing. As described by Mowery (2007, p. 168),
The share of U.S. research university patenting accounted for by institutions with at least 10 patents issued before 1980 declined from more than 85 percent during 1975-1980 to less than 65 percent by 1992. By contrast, low-intensity pre-1980 patenters (institutions with fewer than 10 patents) increased their share of all academic patents from 15 percent in 1981 to almost 30 percent in 1992. And institutions with no patenting activity during 1975-1980 increased their share of overall academic patenting from zero in 1980 to more than 6 percent by 1992.
The less experienced “entrant” universities received less significant patents in the immediate aftermath of the act’s passage (based on citations to these patents in subsequent granted patent applications), although the gap between the quality of their patents and those of experienced institutional patenters had narrowed by the end of the 1980s (Mowery, 2007). This narrowing of the gap in the quality of patents among different university cohorts after 1980 suggests that the patenting strategies of less experienced academic patenters changed during the 1980s toward a more selective approach. Entrant universities in particular appear to have learned to patent, but the sources and mechanisms of such learning are not well understood. Mowery (2007, pp. 168-169) notes:
This evidence concerning the relatively low quality of the early patents obtained by many entrant institutions also underscores the need for caution in using counts of patents (on their own or relative to R&D spending) as a measure of the productivity of research universities. Patents vary widely in quality: as with academic papers, a great many patents are never cited or actively worked by anyone, and the value of any portfolio of patents typically is dominated by a very small number of patents. Comparisons of patent activity across universities, or (even more questionable) between universities and industry, must incorporate
some adjustment for the quality of patents, for example, through citation-weighting of patents.15
MANAGEMENT OF UNIVERSITY PATENTING AND LICENSING
According to Mowery (2009a, p. 39):
By 2002, according to the Association of University Technology Managers (2003), gross licensing revenues for all U.S. universities exceeded $1.2 billion. Licensing data from the University of California nine-campus system [a tenth campus was opened in 2007], Stanford University, and Columbia University, all of which have ranked among the institutions reaping the highest gross licensing income, show that biomedical patents accounted for more than 66-85 percent of the gross licensing revenues of these academic institutions for much of the 1980s and 1990s (Mowery et al., 2004). Even for these relatively successful academic licensors, however, licensing revenues (especially net licensing revenues that flow to the institution) represent a remarkably small share of overall academic operating budgets.
Very few U.S. universities publish sufficient data to enable estimation of the net revenues that their licensing operations contribute to institutional income. One institution that did publish such data through the early 2000s is the University of California system-wide technology licensing office, whose data cover the entire University of California system. As noted by Mowery (2009a), that system’s annual net licensing revenues after deduction of operating expenses and payments to inventors averaged roughly $30 million during fiscal years 1999-2004—roughly 1 percent of the system’s annual research expenditures of nearly $3 billion and well below the $235 million in industry-sponsored research conducted in the system in fiscal year 2003. Given that the system reported relatively high gross annual licensing revenues (averaging nearly $100 million) during this period, it appears likely that the financial contributions of patent licensing to most university operating budgets are modest at best, and negative for a great many institutions. Moreover, these financial inflows appear to be dwarfed by those associated with industry sponsorship of academic research (Mowery, 2009a).
Revenues are of course not the only motive for university licensing
15An especially misguided use of patent counts for evaluation involves use of such counts for individual university faculty as a measure of professional performance. Among other effects, the creation of incentives for faculty to patent is likely to increase significantly the operating expenses of university technology licensing offices that must bear the costs of prosecuting an expanded flow of patent applications, many of which may cover inventions of limited technological significance and commercial potential.
activities.16 Other important motives include recruiting and/or retention of faculty who wish to see their inventions patented and licensed, the transfer of university inventions to commercialization, regional or state-level economic development, and preservation of the freedom of academic scientists to conduct research. This array of potential goals for patenting and licensing activities, however, creates some challenges for management. These goals are not entirely compatible—for example, support for regional economic development may entail acceptance of lower royalty rates on licenses for firms active in the vicinity of the university. Technology licensing thus will involve some trade-offs among these goals, trade-offs that must be embodied in a coherent institutional policy and clearly communicated to the staff charged with responsibility for these activities. The goals of institutional technology licensing programs also must be consistent with the metrics used by the institution to evaluate and reward licensing office staff. Yet despite these trade-offs, as well as the relatively modest scale of net revenues at many university technology licensing offices, a recent survey revealed that technology licensing officers regard licensing revenues as the most important goal of their activities (Jensen and Thursby, 2001).
INDUSTRY-SPONSORED RESEARCH AND INDUSTRY CRITICISM OF UNIVERSITY LICENSING POLICIES
Despite significant growth in industry-supported research at U.S. universities as a share of total U.S. university R&D after 1980 (see Figure B-1), this share peaked in 1998-1999 at slightly more than 6 percent of total university research funding and has since declined. As of 2008, industry-sponsored research at U.S. universities accounted for a smaller share of university R&D than was the case in the early 1950s. Reasons for the decline in the industry-funded share of U.S. university R&D since 1999 are not well understood.
Industry criticism of U.S. universities’ licensing policies and practices intensified during the early 2000s, particularly in the information technology (IT) sector. As recalled by Mowery (2009a), Dr. R. Stanley Williams of Hewlett Packard, a firm with a long history of close research collaboration with U.S. universities, stated in testimony before the U.S. Senate Commerce Committee’s Subcommittee on Science, Technology and Space:
16Indeed, the National Research Council’s (2010) study of university technology licensing programs states that “patenting and licensing practices should not be predicated on the goal of raising significant revenue for the institution. The likelihood of success is small, the probability of disappointed expectations high, and the risk of distorting and narrowing dissemination efforts is great” (p. 5).
FIGURE B-1 Industry-funded share of total academic R&D, 1953-2008 (excluding federally funded research and development centers).
NOTE: R&D = Research and Development.
SOURCE: National Science Foundation (2010).
Largely as a result of the lack of federal funding for research, American Universities have become extremely aggressive in their attempts to raise funding from large corporations. . . . Large U.S.-based corporations have become so disheartened and disgusted with the situation they are now working with foreign universities, especially the elite institutions in France, Russia and China, which are more than willing to offer extremely favorable intellectual property terms.17
A more sweeping critique was presented at a 2003 conference organized by the Government-University-Industry Research Roundtable at the National Academy of Sciences, as described by Mowery (2007):
. . . the universities’ approach of securing iron-clad protection for intellectual property seems to be yielding diminishing returns, even within the narrow confines of the licensing activity itself. . . . The requisite legal negotiations for IP-that-will-ultimately-prove-to-be-useless are laborious, individualized and negotiated between universities and companies on a case-by-case. The up-front legal negotiations can easily cost more than the total cost of the research project being conducted, and/
17See http://www.memagazine.org/contents/current/webonly/webex319.html [August 2014].
or extent past the time when the company has interest in the technology path being pursued. . . . In summary, the uncertainty of the true value of university-generated intellectual property, combined with a litigious culture, have combined to make the university-industry working relationship—one that has historically contributed greatly to graduate education—unaffordable and nearly unsustainable within the U.S. (Government-University-Industry Research Roundtable, 2003, p. 2).
These critical comments triggered considerable discussion between large industrial firms (many of which are in the IT sector) and U.S. research universities over intellectual property policies and licensing guidelines (Mowery, 2007). In December 2005, four large IT firms (Cisco, Hewlett Packard, IBM, and Intel) and seven universities (Carnegie Mellon University; Georgia Institute of Technology; Renssellaer Polytechnic; Stanford University; University of California, Berkeley; University of Illinois at Urbana–Champaign; and University of Texas at Austin) agreed on a “statement of principles” for collaborative research on open-source software that emphasizes liberal dissemination of the results of collaborative work funded by industrial firms (Mowery, 2009a).18
Relationships between established firms and universities were most often the targets of the critical statements that received press coverage and some attention from policy makers (Mowery, 2009a). Interestingly, the economic interests of established firms with large patent portfolios differ in some ways from those of small start-up firms that are owners or licensees of far fewer patents. Furthermore, a large number of conflicts
18The “Open Collaboration Principles” cover “…just one type of formal collaboration that can be used when appropriate and will co-exist with other models, such as sponsored research, consortia and other types of university/industry collaborations, where the results are intended to be proprietary or publicly disseminated.” According to the “Principles,” “The intellectual property created in the collaboration [between industry and academic researchers] must be made available for commercial and academic use by every member of the public free of charge for use in open source software, software related industry standards, software interoperability and other publicly available programs as may be agreed to be the collaborating parties. . . ” See http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCkQFjAA&url=http%3A%2F%2Fsites.kauffman.org%2Fpdf%2Fopen_collaboration_principles_12_05.pdf&ei=lHNpU4ebJoqPyASi4YKYCw&usg=AFQjCNGTkBvau6MnVz9eEeazakuYxAJBHw&sig2=ywl6B3MeD7wy886l0bMQGA&bvm=bv.66111022,d.aWw&cad=rja [May 2014].
These principles originated in an August 2005 “University and Industry Innovation Summit” in Washington, DC, organized by the Kauffman Foundation of Kansas City and IBM. See http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&cad=rja&uact=8&ved=0CC8QFjAB&url=http%3A%2F%2Fpublic.dhe.ibm.com%2Fsoftware%2Fdw%2Funiversity%2Fcollaborativeresearch%2FUniversityIndustrySummit.pdf&ei=8HNpU8TQHMiZyATj_YLICg&usg=AFQjCNFQCmyFJ7q7FyYKqD2YBpN105McFA&sig2=w1E29mTk59rrsjE-cVobNw&bvm=bv.66111022,d.aWw [May 2014].
have involved nonbiomedical firms, as the value of individual patents in such industries as IT typically is lower than is true of biomedical research. Nevertheless, the current controversies and discussions between U.S. industrial firms (some of which, as was noted earlier, contrast U.S. university intellectual property right policies unfavorably with those of universities outside of the United States) and U.S. research universities appear to have led some U.S. universities to revise their institutional strategies for supporting collaborative research relationships with U.S. industry (Mowery, 2009a).
For example, some leading U.S. research universities have developed licensing strategies that assign greater weight to the goal of using licensing as a tool to increase industry sponsorship of academic research. The director of the Stanford Office of Technology Licensing now oversees the university’s Industrial Contracts Office, which manages sponsored-research agreements with industry while also overseeing materials transfer agreements, which govern the transfer of research tools and materials among researchers. Industrial firms supporting campus research can receive licenses (in some cases, royalty-free) to the results of this research.
A similar trade-off between maximizing licensing revenues and obtaining industry research funding is apparent in the creation in 2003 of the Intellectual Property and Industry Research Alliances Office in the University of California, Berkeley licensing office, which absorbed the established Office of Technology Licensing and a new Industry Alliances Office, charged with overseeing the negotiation of sponsored-research agreements with industry. Moreover, the Berkeley licensing office, along with other University of California technology licensing offices, has implemented a new policy recognizing the differences among industries in the value (and likely licensing income) of patents in different fields of research. In 2000, the University of California President’s Office authorized the negotiation of royalty-free licenses with industrial sponsors of campus research in electrical engineering and computer science.
These initiatives, along with the assignment of responsibility for a broader set of relationships with industry to campus directors of technology licensing offices at both Stanford and the University of California, Berkeley suggest that these leading academic licensors are developing a more nuanced approach to the management of trade-offs within their technology transfer strategies (Mowery, 2007). Two key features of this new approach appear to be an effort to manage patenting and licensing as part of a broader institutional strategy for supporting collaboration with industry, and some effort to tailor patenting and licensing policies to the contrasting economic and technological importance of formal intellectual property instruments among different fields of research. What is less well understood is why it has taken so long for U.S. universities to revise their
approach to managing patenting and licensing. Indeed, the more than three decades since the passage of the Bayh-Dole Act have witnessed a surprising lack of experimentation with new licensing structures and policies among U.S. universities. As U.S. (and non-U.S.) universities expand their use of alternative policies and organizations for managing interactions with industry, better data on what is being done and how well it is accomplishing its goals are badly needed.
U.S. universities have a long tradition of engagement with industry in research and other collaborative activities. This pattern of engagement has benefited from a two-way flow of ideas and people between academic and industrial research settings. Both the historic structure of the national U.S. system of higher education and factors external to U.S. universities (e.g., labor markets that support relatively high levels of domestic interinstitutional mobility of researchers, new-venture financing from various private sources) have contributed to this tradition of collaboration, which has included extensive patenting and licensing of university inventions to industry. But interaction between U.S. universities and innovation in industry throughout the 20th and 21st centuries has relied on a number of different channels, ranging from the training of students to faculty consulting, publication of research advances, and industry-sponsored research, among others. These channels operate in parallel and are interdependent. Moreover, the relative importance of different channels of interaction and information flow between industrial and academic researchers appears to vary considerably among different research fields.
The so-called “Bayh-Dole era” that began in 1980 extended and expanded this engagement, which drew as well on extensive federal support for research, notably in the life sciences. That support produced important advances that sparked growth in university patenting and licensing, increasingly managed directly by U.S. universities, during the 1970s. There is little evidence during the post-1980 period that increased faculty engagement in entrepreneurial activities, including new-firm formation and patenting and licensing of inventions, negatively affected the scholarly productivity of leading researchers. Nonetheless, the intellectual property management policies of at least some U.S. universities sparked criticism from U.S. firms in the early 2000s, especially those in IT. In response to this criticism, some U.S. universities have experimented with new approaches to the management of patenting and licensing that take into account the differences among research fields in the importance of patents as vehicles for information exchange and technology transfer.
Reflecting their complex roles in regional and U.S. national econo-
mies, university technology transfer programs can be used to pursue an array of institutional goals. But these goals are not always mutually consistent or compatible, meaning that policy priorities must be established in technology transfer programs and clearly linked to policies in operation. Revenue-maximizing licensing strategies are shortsighted.
Effective metrics for evaluating the performance of universities in transferring technology and supporting industrial innovation should be aligned with the specific goals of a given university or research institute. They also should attempt to account for the full breadth of channels through which university research influences industrial innovation, including the training and placement of students, faculty research publications, faculty- or student-founded firms, patents, and licenses. Given the lack of data covering these various channels for most U.S. universities, as well as the need for metrics to be tailored to the goals and environment of individual universities, it would appear to be unrealistic and unwise for federal agencies or other government evaluators to impose a single, uniform set of metrics purporting to measure the technology transfer performance of all U.S. universities. The institutional heterogeneity that historically has been a strength of the U.S. system of higher education should be recognized and preserved in any evaluative framework.
This institutional heterogeneity also implies a need for flexibility and variety in the policies used by U.S. universities to support interactions with industry and the commercialization of academic research advances. Although the Bayh-Dole Act and other relevant federal policies do not dictate any single institutional structure for managing patenting, licensing, and related activities in university-industry collaboration, U.S. universities have been slow to experiment with different approaches to managing these activities during the more than three decades since the act’s passage. Such experimentation, combined with efforts to assess the effectiveness of alternative approaches to university policies on collaboration, should be encouraged by federal agencies, industry, and other stakeholders. Nonetheless, no single approach is likely to prove feasible or effective across the numerous and diverse academic institutions and private firms engaged in federally funded research and industry collaboration.
Agrawal, A., and Henderson, R. (2002). Putting patents in context: Exploring knowledge transfer from MIT. Management Science, 48(1), 44-60.
Association of University Technology Managers. (2001). Licensing Survey FY2001. Available: http://www.autm.net/AM/Template.cfm?section=FY_2001_Licensing_Survey&template=/CM/HTMLDisplay.cfm&ContentID=10196 [May 2014].
Association of University Technology Managers. (2002). Licensing Survey FY2002. Available: http://www.autm.net/AM/Template.cfm?section=FY_2002_Licensing_Survey&template=/CM/HTMLDisplay.cfm&ContentID=10195 [May 2014].
Association of University Technology Managers. (2003). Licensing Survey FY2003. Available: http://www.autm.net/AM/Template.cfm?section=FY_2003_Licensing_Survey&template=/CM/HTMLDisplay.cfm&ContentID=10194 [May 2014].
Azoulay, P., Ding, W., and Stuart, T. (2009). The impact of academic patenting on the rate, quality, and direction of (public) research output. Journal of Industrial Economics, 57(4), 637-676.
Broad, W. (1979). Patent bill returns bright idea to inventor. Science, 205(4405), 473-476.
Cesaroni, F., di Minin, A., and Piccaluga, A. (2005). Exploration and exploitation strategies in industrial R&D. Creativity and Innovation Management, 14(3), 222-232.
Cohen, W.M., Nelson, R.R., and Walsh, J.P. (2002). Links and impacts: The influence of public research on industrial R&D. Management Science, 48(1), 1-23.
DiGregorio D., and Shane, S. (2003). Why do some universities generate more start-ups than others? Research Policy, 32(2), 209-227.
Ding, W., and Choi, E. (2011). Divergent paths to commercial science: A comparison of scientists’ founding and advising activities. Research Policy, 40(1), 69-80.
Eskridge, N. (1978). Dole blasts HEW for ‘stonewalling’ patent applications. Bioscience, 28(9), 605-606.
Geiger, R.L. (1986). To Advance Knowledge: The Growth of American Research Universities, 1900-1940. New York: Oxford University Press.
Geiger, R.L. (1993). Research and Relevant Knowledge: American Research Universities since World War II. New York: Oxford University Press.
Government-University-Industry Research Roundtable. (1991). Industrial Perspectives on Innovation and Interactions with Universities. Washington, DC: National Academy Press.
Government-University-Industry Research Roundtable. (2003). Background Paper Articulating the Need for a New Approach to University-Industry Relations. Washington, DC: The National Academies Press.
Hughes, S. (2001). Making dollars out of DNA. Isis, 92(3), 541-575.
Jensen, R., and Thursby, M. (2001). Proofs and prototypes for sale: The licensing of university inventions. American Economic Review, 91(1), 240-258.
Katz, M.L., and Ordover, J.A. (1990). R&D competition and cooperation. Brookings Papers on Economic Activity: Microeconomics, 137-192.
Kenney, M., and Mowery, D.C. (Eds.). (Forthcoming). Regional Economic Development, Public Universities, and Technology Transfer: Studies of the University of California’s Contributions to Knowledge-based Growth. University of California, Davis.
Lécuyer, C. (2005). What do universities really owe industry? The case of solid state electronics at Stanford. Minerva, 43, 51-71.
Lécuyer, C., and Brock, D.C. (2006). The materiality of microelectronics. History and Technology, 22, 301-325.
Levin, R.C., Klevorick, A., Nelson, R.R., and Winter, S. (1987). Appropriating the returns from industrial research and development. Brookings Papers on Economic Activity, 3, 783-820.
Mansfield, E. (1991). Academic research and industrial innovations. Research Policy, 20, 1-12.
Mowery, D.C. (2007). University-industry research collaboration and technology transfer in the United States since 1980. In S. Yusuf and K. Nabeshima (Eds.), How Universities Promote Economic Growth (pp. 165-181). Washington, DC: The International Bank for Reconstruction and Development/The World Bank. Available: https://openknowledge.worldbank.org/bitstream/handle/10986/6631/383330Universi101OFFICIAL0USE0ONLY1.txt?sequence=2 [May 2014].
Mowery, D.C. (2009a). Plus ca change: Industrial R&D in the “third industrial revolution.” Industrial and Corporate Change, 18(1), 1-50.
Mowery, D.C. (2009b). University-Industry Collaboration and Technology Transfer in Hong Kong and Knowledge-based Economic Growth. Working Paper. Berkeley: University of California, Berkeley. Available: http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CCkQFjAA&url=http%3A%2F%2Fwww.savantas.org%2Fcmsimg%2Ffiles%2FResearch%2FHKIP%2FReport%2F1_mowery.pdf&ei=LHFxU-6pLonvoATA9oDABg&usg=AFQjCNHz6wwAqV1reOgHmW9at3u38DDHHQ&sig2=_KTsgdR3Uf5PTbUyqgZFQQ&bvm=bv.66699033,d.cGU [May 2014].
Mowery D.C. (2011). Learning from one another? International policy “emulation” and university-industry technology transfer. Industrial and Corporate Change, 20(6), 1827-1853.
Mowery, D.C., and Sampat, B.N. (2005). The Bayh-Dole Act of 1980 and university-industry technology transfer: A model for other OECD governments? Journal of Technology Transfer, 30(1/2), 115-127.
Mowery, D.C., Nelson, R.R., Sampat, B.N., and Ziedonis, A.A. (2004). ‘Ivory Tower’ and Industrial Innovation: University-Industry Technology Transfer Before and After the Bayh-Dole Act. Redwood City, CA: Stanford University Press.
National Research Council. (2010). Managing University Intellectual Property in the Public Interest. S.A. Merrill and A. Mazza (Eds.). Committee on Management of University Intellectual Property: Lessons from a Generation of Experience, Research, and Dialogue. Board on Science, Technology, and Economic Policy, Committee on Science, Technology, and Law, Policy and Global Affairs. Washington, DC: The National Academies Press.
Nelson, A. (2012). Putting university research in context: Assessing alternative measures of production and diffusion at Stanford. Research Policy, 41(4), 678-691.
Nerkar, A., and Shane, S. (2003). When do start-ups that exploit patented academic knowledge survive? International Journal of Industrial Organization, 21, 1391-1410.
OECD. (2002). Benchmarking Science-Industry Relationships. Paris, France: Author.
Reimers, N. (1998). Stanford’s Office of Technology Licensing and the Cohen/Boyer Cloning Patents. Berkeley: University of California, The Bancroft Library.
Shane, S., and Stuart, T. (2002). Organizational endowments and the performance of university start-ups. Management Science, 48(1), 154-170.
Stevens, A.J. (2004). The enactment of Bayh-Dole. Journal of Technology Transfer, 29, 93-99.
Thursby, M., and Thursby, J. (2007). U.S. Faculty Patenting Inside and Outside of the University. Working Paper No. 13256. Cambridge, MA: National Bureau of Economic Research.
Trow, M. (1979). Aspects of diversity in American higher education. In H. Gans (Ed.), On the Making of Americans (pp. 271-290). Philadelphia: University of Pennsylvania Press.
Trow, M. (1991). American higher education: ‘Exceptional’ or just different. In B.E. Shafer (Ed.), Is America Different? A New Look at American Exceptionalism (pp. 138-186). New York: Oxford University Press.
U.S. Senate Committee on the Judiciary. (1979). The University and Small Business Patent Procedures Act: Hearings on S. 414, May 16 and June 6. Washington, DC: U.S. Government Printing Office.
Zucker, L.G., Darby, M.R., and Armstrong, J.S. (1998). Geographically localized knowledge: Spillovers or markets? Economic Inquiry, 36(1), 65-86.
Zucker, L.G., Darby, M.R., and Armstrong, J.S. (2002). Commercializing knowledge: University science, knowledge capture, and firm performance in biotechnology. Management Science, 48(1), 138-153.