Local Nodes in Global Networks: The Geography of Knowledge Flows in Biotechnology Innovation

Abstract

The literature on innovation and interactive learning has tended to emphasize the importance of local networks, inter‐firm collaboration and knowledge flows as the principal source of technological dynamism. More recently, however, this view has come to be challenged by other perspectives that argue for the importance of non‐local knowledge flows. According to this alternative approach, truly dynamic economic regions are characterized both by dense local social interaction and knowledge circulation, as well as strong inter‐regional and international connections to outside knowledge sources and partners. This paper offers an empirical examination of these issues by examining the geography of knowledge flows associated with innovation in biotechnology. We begin by reviewing the growing literature on the nature and geography of innovation in biotechnology research and the commercialization process. Then, focusing on the Canadian biotech industry, we examine the determinants of innovation (measured through patenting activity), paying particular attention to internal resources and capabilities of the firm, as well as local and global flows of knowledge and capital. Our study is based on the analysis of Statistics Canada's 1999 Survey of Biotechnology Use and Development, which covers 358 core biotechnology firms. Our findings highlight the importance of in‐house technological capability and absorptive capacity as determinants of successful innovation in biotechnology firms. Furthermore, our results document the precise ways in which knowledge circulates, in both embodied and disembodied forms, both locally and globally. We also highlight the role of formal intellectual property transactions (domestic and international) in promoting knowledge flows. Although we document the importance of global networks in our findings, our results also reveal the value of local networks and specific forms of embedding. Local relational linkages are especially important when raising capital—and the expertise that comes with it—to support innovation. Nevertheless, our empirical results raise some troubling questions about the alleged pre‐eminence of the local in fostering innovation.

Keywords:

• Knowledge flows
• biotechnology
• buzz
• inter‐firm collaboration

Acknowledgements

The authors are grateful to Fred Gault and Antoine Rose of the Science, Innovation and Electronic Information Division, Statistics Canada, for their invaluable support and advice.

Notes

1. In a recent overview of the “geography of innovation” literature, Asheim and Gertler (2005: 296–298) note that biotechnology draws primarily upon “analytical” forms of knowledge that, at least in theory, are relatively easily codified and transmitted across long geographical distances and cultural/institutional divides. However, they also note the marked tendency for biotech innovation to cluster geographically, and offer further insights as to why this might be the case, emphasizing the continuing importance of local networks and relations, as well as the growing importance of quality of place in attracting and retaining the highly educated people that typically work in this sector.

2. Star scientists are defined as outstanding scientists who combine scientific productivity (as measured by number of journal articles while affiliated with the firm) with specific knowledge of the new technique for the commercialization.

3. Dyer (2002b: 13) quotes Daniel Vasella, chairman of Novartis, as observing that “the biggest pool of untapped scientists and hospital research was in the Boston area”.

4. These highlights are a product of the Canadian Biotechnology Commercialization Roundtable organized by Ernst and Young in March 2002, which brought together industry members and analysts.

5. The success of biotechnology firms is often measured by their innovative output rather than by their revenues or profits. Zucker et al. (1998a) point out that revenues are generally non‐existent in the young biotech industry and are therefore of limited use as an indicator of firms’ success. Innovative outputs, on the other hand, can indicate the ability of a firm to attract VC that in turn allows a firm to develop and commercialize the fruits of its research. Zucker et al. suggest that innovation can be measured by looking at the number of products in development, the number of products on the market and net growth in employment. Nevertheless, they and others agree that patents are perhaps the most useful predictor of commercial values and future returns, and that patents are key to attracting VC to further develop or commercialize innovative products (Niosi, 2000b). Notwithstanding this consensus, it should also be acknowledged that patents as a measure of innovation are subject to some well‐known limitations. Many patents are never commercialized. For many forms of innovation, patents may offer only partial protection of intellectual property; for others, patents may be wholly infeasible or impractical (see Teece, 1986; Levin et al., 1987; Klevorick et al., 1995).

6. The following are the categories listed in the survey which correspond to each of the variables: Internal—internal sources/staff or parent/subsidiary firm; Codified—academic journals/trade publications, library/literature search, database retrieval services; Tacit/personal—universities/colleges/private training institutes, personal contact with others, other companies, professional/industry associations, conferences/workshops/trade shows; and Government—federal agencies, provincial agencies.

7. The survey asks firms for the location of their collaborative partners by country. Therefore, for the purpose of this analysis, “local” partners were defined as those within Canada. Previous work (Ernst & Young, 2002) confirms that biotech firms are highly concentrated in a few Canadian city‐regions, with Toronto, Montreal and Vancouver dominating. This suggests that there is a high likelihood that a partner in the same country is also located within the same city‐region.

8. Unfortunately, the survey design does not permit us to determine the location of these private/relational sources.

9. It should be acknowledged that, although we are defining “distance” in traditional physical terms here, it is highly likely that cultural or institutional distance is equally relevant to this discussion, as Gertler (2004) has demonstrated.

10. As Malmberg and Maskell (2002) point out, the benefits arising from locating close to other firms in the same industry may have as much to do with competition, monitoring and observation as with collaboration and cooperation.

11. The survey does not allow us to examine directly the relative importance of local vs. non‐local and global research and teaching institutions. Nevertheless, the larger study of which this analysis is a part will pursue these issues through a series of interview‐based case studies of bioscience activity in several different regions of Canada.