Abstract

This article analyzes the texture characteristics of regions that developed substantial technological activities in the biotechnology field. Building on patent and publication-based indicators, we find support for the presence of two distinctive types of regions that are able to develop into “top regions” during the rapid growth phase of the biotech industry (period 1992–1997). In addition, the study highlights differences between top region and other regions in terms of regional texture characteristics (presence of anchor tenant firms, entrepreneurial-oriented scientific actors, and scientific eminence).

1. Introduction

Modern biotechnology has generated important breakthroughs in several industries—from food and agriculture to the chemical industry but most particularly in pharmaceuticals ( Arora and Gambardella, 1990 ; Zucker and Darby, 1997 )—by enabling the creation of entirely new organic materials and, thus, profoundly changing the process of (drug) discovery and product development ( Powell et al. , 1996 ). Consequently, biotechnology is often considered a promising technology that will bring economic growth and welfare to a region. Over the past decades therefore, thriving biotech clusters, such as the San Francisco Bay Area, San Diego and Boston but also other emerging biotech regions, have been widely studied in order to identify the main factors behind the success of biotech regions.

A general consensus exists that well-developed biotech regions, so-called clusters or hot spots, are characterized by the presence of world-class scientific research, high levels of entrepreneurial activity (both academic spin-offs and industrial ventures), high labor mobility and dense social networks, the availability of venture capital, and a dedicated support infrastructure (e.g. Cooke, 2001 ; Owen-Smith et al. , 2002 ; Casper, 2007 ). Concerning the respective roles and importance of public knowledge institutes and private firms in the emergence and early development of biotech regions, various perspectives have been advanced. Case-study research provides evidence that universities and knowledge-generating institutes have played a central and active role in the creation of biotech clusters in the Boston region ( Breznitz et al. , 2008 ) and San Francisco Bay area ( Chiaroni and Chiesa, 2006 ). In contrast, private firms have exercised a prominent role in the development of biotech activities in the regions of Milano (Italy) and Uppsala (Sweden) ( Chiaroni and Chiesa, 2006 ) as well as in Japan ( Bartholomew, 1997 ). As (industrial) biotechnology enters a growth phase ( Lecocq and Van Looy, 2009 ), the question of whether regions can evolve into leading clusters by relying on “distributed” textures or whether the presence and/or emergence of anchor tenant firms ( Agrawal and Cockburn, 2003 ) is a prerequisite becomes increasingly pertinent, both for practitioners and policy makers engaged in regional economic development.

While case-study research provides valuable insights into, and detailed information on, the characteristics and dynamics of individual biotech clusters, previous research typically covers only one, or a limited number of, biotech region. The results of these studies are difficult to compare since they use different (performance) indicators and relate to different regional units of analysis in different time periods. Large-scale empirical studies addressing the texture characteristics of biotech regions are lacking. Building on patent and publication-based indicators, we engage in such a study in the biotechnology field. Our analyses cover 78 regions from North America, Europe, and Asia-Pacific that developed substantial technological activities in the field of biotechnology over the period 1992–1997, an era characterized by rapid growth of the biotech industry. Our study contributes to the existing literature on biotech clusters by introducing a new typology (“concentrated” versus “distributed” regions) and examining different antecedents of the global competitiveness of biotech clusters in terms of regional texture characteristics, according to cluster type.

The article is organized as follows: in the next section, the importance of basic science in molecular biology and the role of knowledge-generating institutes, small dedicated biotech firms, and large established (pharmaceutical) firms in the development of modern biotechnology are discussed. Then, hypotheses are developed with respect to the distinctive industrial texture characteristics as well as the presence of entrepreneurial-oriented knowledge institutes in the top regions during the growth phase of the biotech industry (period 1992–1997). Subsequently, data sources and variables used in the analyses are introduced. In the analysis section, the leading regions worldwide in terms of biotech technology development in the period 1992–1997 are identified. Next, the texture characteristics of those “top” regions in terms of their industrial base and the presence of entrepreneurial-orientated public knowledge institutes are further investigated. In the last part of the analysis, we investigate the texture characteristics that are instrumental in creating a top region in the biotechnology field during the rapid growth phase of the industry. The article concludes with a discussion of the results and some policy implications.

2. The field of modern biotechnology

Modern biotechnology is a complex, fast-changing, knowledge-intensive field that has its origin in academic research in molecular biology. Of primary importance to the development of modern biotechnology was the discovery of the double-helix structure of DNA (1953) by Watson and Crick in the laboratories of the University of Cambridge (UK). The foundation of the modern biotech industry was laid in 1973, when professors Cohen (Stanford University, USA) and Boyer (University of California, USA) discovered the recombinant DNA technique, which facilitated the transfer of the basic science of molecular biology into useful knowledge for a wide range of industrial applications ( Feldman, 2003 ).

Following the discovery of the recombinant DNA technique, the second half of the 1970s and the 1980s was marked by the creation of the first companies dedicated to modern biotechnology, the so-called New Dedicated Biotech Firms (NDBFs). These new biotech companies were often co-founded by, or maintained strong linkages with, academic researchers ( Zucker and Darby, 1996 ). They focused on exploring new technological and scientific research results and translating them into the commercial domain ( Acharya, 1999 ; Galambos, 2006 ). As new scientific knowledge is often characterized by a substantial amount of tacit knowledge, developing an idea from science usually requires close links with the academic inventor(s) ( Rosenberg and Nelson, 1994 ; Zucker and Darby, 1996 ). NDBFs were, therefore, most often established in close proximity to universities or research centers ( Prevezer, 2001 ).

In the USA, small research-intensive biotech firms, set up to further explore and commercialize the results of scientific research, have contributed significantly to the development of biotech clusters. For example, in the Greater Boston area—one of the first biotech clusters in the USA—more than 50 biotech companies spun off from the Massachusetts Institute of Technology and an additional 50 start-ups were founded by academic inventors from the university ( Breznitz et al. , 2008 ). In the USA, first-mover advantage in the growth of small research-intensive biotech firms has been made possible by supportive institutional arrangements such as the presence of venture capital specialized in high technology and the Bayh-Dole Act 1 that facilitated technology transfer between academia and industry ( Prevezer, 2001 ; Owen-Smith et al. , 2002 ; Niosi, 2011 ). Indeed, it was possible for scientists in the USA to become involved in the creation of start-ups while retaining their academic positions ( Prevezer, 2001 ). Furthermore, Owen-Smith et al. (2002) point to the diversity of organizations involved in research activities (universities, research institutes, hospitals, and small firms) and to the support of the National Institutes of Health that enabled the integration of basic science and clinical development.

By the mid-1980s, large established firms in the chemical and, in particular, the pharmaceutical industry increasingly displayed an interest in the biotechnology field. However, as the principal knowledge base (organic chemistry) of these incumbent firms differed significantly from the science base of biotechnology (molecular biology), large firms had difficulty internalizing this new knowledge ( Zucker and Darby, 1997 ). From the late 1980s onward, they entered into the field by setting up strategic alliances with or acquiring small biotech firms ( Arora and Gambardella, 1990 ; Rothaermel, 2001 ; Dominquez-Lacasa, 2006 ; Roijakkers and Hagedoorn, 2006 ; Rothaermel and Hess, 2001 ). Most of these research and joint development agreements with small biotech firms were product-specific, and more market-oriented alliances focused on clinical trials, FDA regulatory management, and marketing and sales activities ( Arora and Gambardella, 1990 ; Rothaermel, 2001 ), while acquisitions were mostly aimed at acquiring NDBF’s specialized knowledge in particular areas of biotechnology research ( Arora and Gambardella, 1990 ).

Several case studies provide evidence that large established firms have played an important role in developing regional activities in the biotechnology field: in the Basel region (Switzerland), the strong presence of a pharmaceutical industry with firms such as Novartis and Roche has contributed to the growth of biotech activities in the region ( Houston, 2003 ); equally, in the Bioregion Rhineland in Germany, the presence of a chemical and pharmaceutical industry is considered to be “an advantage for the creation of an integrated biotech sector from research to production” ( Zeller, 2001 ). In the 1990s, mergers and acquisitions by large established players resulted in an upsurge of entrepreneurial activities in biotechnology in the regions of Milan (Italy), Uppsala (Sweden), and San Diego (South California), leading to the emergence of totally new business structures in the region ( Chiaroni and Chiesa, 2006 ). In the San Diego cluster, for example, nearly 50 industrial spin-offs were created by former employees/scientists of the biotech company, Hybritech, after it was acquired by the pharmaceutical company, Eli Lilly.

A large strand of literature shows that further technology advances in the biotech industry rely, to a large extent, on inter-organizational collaborations between organizations with complementary resources—namely, universities and public research centers plumbing the source of new scientific knowledge; large pharmaceutical and chemical firms with the capabilities to market products (including experience of clinical testing, engineering know-how about manufacturing, and access to commercial markets); and NDBFs often considered the nexus between academia and large established firms ( Arora and Gambardella 1990 ; Gertler and Vinodrai, 1996; Powell et al. , 1996 ; Gambardella et al. , 2000 ; Mangematin et al. , 2003 ). Collaboration on an international scale appears to be of particular relevance to technology development since it introduces new knowledge and skills to a region ( Cooke, 2001 ; Zeller, 2001 ; Lecocq et al. , 2009 ). Over the period 1975–1999, Roijakkers and Hagedoorn (2006) found evidence of changing patterns in inter-firm R&D-partnering networks in the pharmaceutical biotechnology sector: from a few, isolated clusters of cooperating firms in the early years (1975–1979) to increasingly dense research networks in the 1980s, with small dedicated biotech firms playing important bridging roles between different networks. By the 1990s, inter-firm R&D networks had become very large and densely interrelated, and some large, incumbent firms developed into important network players.

On a global scale, the number of contractual R&D partnerships between firms has grown over time, especially since the 1980s, with a clear dominance of the pharmaceutical sector (including pharmaceutical biology) and other high-tech industries (information technology, and aerospace and defense) from the mid-1980s onward ( Hagedoorn, 2002 ). The need for learning and flexibility in a highly competitive landscape are advanced by Hagedoorn (2002) as major reasons for the increase in joint R&D projects in high-tech industries. Powell et al. (1996) point to the specific skills and know-how required to translate scientific advances into commercial applications, which, in an industry characterized by a regime of rapid technology development and a complex knowledge base such as biotechnology, do not usually reside within a single organization but have to be acquired through networks of learning.

Figure 1 presents the (worldwide) evolution of biotech technology over the period 1978–1999, measured by the number of EPO patent applications. The figure shows a steady, linear increase in the number of patent applications in the early phase of the biotech industry (period 1978–1990), followed by an exponential growth in the number of patent applications from the early 1990s onward ( Lecocq and Van Looy, 2009 ). This study focuses on the period 1992–1997, characterized by rapid growth in biotech technology development.

Figure 1.

Evolution of patenting in the field of biotechnology (EPO Patents, period 1978–1999, worldwide)

Figure 1.

Evolution of patenting in the field of biotechnology (EPO Patents, period 1978–1999, worldwide)

3. The regional clustering of biotech activities: toward hypotheses

Aforementioned studies (e.g. Bartholomew, 1997 ; Prevezer, 2001 ; Zeller, 2001 ; Owen-Smith et al. , 2002 ; Houston, 2003 ; Chiaroni and Chiesa, 2006 ; Breznitz et al. , 2008 ), providing evidence of different types of actors leading the process of cluster emergence in the biotechnology field, seem to suggest the existence of different pathways of regional cluster formation. Part of these differences in texture may be related to life-cycle dynamics, with universities playing a major role in the early stages of the industry (period 1978–1989), while industrial capabilities became more important after a “dominant design” was embedded ( Utterback, 1994 ) and technologies (products) were commercialized (period 1990–1999). In order to profit from the take-off of economic activities in the growth phase of biotech, regions may benefit from a different configuration—in terms of the blend of (entrepreneurial) research universities and assorted industrial companies (NDBFs and more established firms)—than in the early days of the industry. The question whether regions can evolve into leading clusters in the growth phase of the biotech industry by relying on a “distributed” texture where both private firms and public knowledge institutes contribute significantly to regional biotech technology development or whether the presence and/or emergence of an anchor tenant firm is a prerequisite becomes pertinent.

3.1 Industrial texture characteristics

By their nature and core raison d’être, firms are best placed to identify market needs, translate technological opportunities into prototypes and commercial products, and bring these new products to the market. Even in science-intensive fields such as biotechnology, private firms remain the major player in the marketplace. When industries are evolving and technologies are becoming more mature, relations clearly become more market-based, and the competition among firms intensifies ( Baglieri et al. , 2012 ). In regions with a critical mass of activities directed at market exploitation and commercialization, firms have greater opportunities to interact and learn from high-quality suppliers, demanding (industrial) customers, and other innovative firms producing similar or complementary goods and services ( Porter, 2000 ). Taken together, the result is enhanced innovation dynamics in the region.

The concentration of innovative activities in larger, R&D-intensive firms may be particularly relevant to the development of a new industry because of their scale and access to larger financial resources compared to new and/or small firms ( Gray and Parker, 1998 ). By creating local niches and/or intermediary markets, larger firms may also encourage entrepreneurial activity in the region and attract high-quality suppliers that, without the anchor firm, would either be absent or of lower quality ( Agrawal and Cockburn, 2003 ). Therefore, we propose that:

Hypothesis 1a: Regions in which technology development activities are, to a greater extent, located in firms are more likely to become a leading biotech region in the growth phase of biotech.

Hypothesis 1b: Regions with higher levels of concentration of regional biotech technology development activities (by private firms) in an anchor tenant firm are more likely to become a leading biotech region in the growth phase of biotech.

3.2 Science and entrepreneurial-orientated knowledge institutes

For firms active in complex, science-intensive fields such as biotechnology, searching for and acquiring highly specialized scientific knowledge from outside the boundaries of the organization is essential ( Powell et al. , 1996 ). As the field of biotech further develops, knowledge is codified in routine procedures or commercially available equipment such as the automatic DNA sequencer ( Rothaermel and Thursby, 2007 ), and the relevant scientific knowledge is spread on a more global scale. The diffusion of knowledge is further enhanced by the numerous international collaborations between public knowledge institutes and private firms (e.g. Cooke, 2001 ; Zeller, 2001 ; Lecocq and Van Looy, 2009 ). Therefore, the geographical proximity of a strong science base and the presence of entrepreneurial-oriented knowledge institutes in the region may become less important in later stages of the technology life cycle. This leads to the following two hypotheses:

Hypothesis 2a: The science intensity of a region, measured by the number of publications per population, is no longer instrumental in becoming a leading biotech region during the growth phase of biotech.

Hypothesis 2b: The entrepreneurial orientation of scientific actors, measured by their involvement in technology, is no longer instrumental in becoming a leading biotech region during the growth phase of biotech.

4. Data

To identify the leading clusters globally in terms of biotech technology development and study the texture characteristics of biotech regions quantitatively, we draw on the data set with Web of Science publications and EPO patent applications in the field of biotechnology created by Glänzel et al. (2004) . In this study, biotechnology publications were extracted from the Science Citation Index Expanded database based on a set of journals assigned by the Thomson Reuters to subject categories related to biotechnologies. 2 Biotechnology patents were identified based on IPC codes using the traditional classifications of OECD, Fraunhofer and NBER, complemented with other classes and groups (validated by field experts) to include both modern and more traditional forms of biotechnology.

The use of patent and publication data has several advantages ( Pavitt, 1985 ; Jaffe, 1989 ; Griliches, 1990 ). They are an important source of information about the time and location of technological and scientific inventions, as well as the organizations and institutions involved. Furthermore, patent and publication data have a global coverage and facilitate the adoption of a (technology) field-specific perspective. Prior research has established patent counts as a valid indicator of novel technological activities at the level of regions ( Acs et al. , 2002 ) and firms ( Narin and Noma, 1987 ; Hagedoorn and Cloodt, 2003 ), and this is certainly true for the field of biotechnology, which is characterized by a high propensity to patent (Arundel and Kabla, 1998). Some research points out that patented inventions may vary in technical and economic value ( Mansfield, 1986 ; Gambardella et al. , 2008 ) and, therefore, suggests weighing patent counts by the number of forward citations received ( Trajtenberg, 1990 ; Harhoff et al. , 1999 ; Hall et al. , 2005 ). Indeed, recent research by Hagedoorn and Cloodt (2003) indicates that patent counts and patent citations represent somewhat different aspects of innovation performance; patent counts are more strongly related to R&D input, while patent citations show a stronger link with new product development. Based on a study of European, US, and Japanese multinational firms, Criscuolo (2006) provides evidence that triadic patents 3 are more highly cited, and thus of higher economic and technological value, than patents that are not extended in the other two major patent offices. In addition, Criscuolo (2006) found that OECD triadic patent data, in contrast to USPTO or EPO, do not show any bias toward the home country of application, making triadic patent counts a valid indicator to compare the technological performance of regions worldwide.

4.1 Regional allocation of patent and publication data

For the purpose of this study, all biotech patents and biotech publications with applicant or author addresses in Australia, Canada, Europe (EU-15), Japan, the USA, and Switzerland have been withheld. Together, these countries represent more than 97% of all patents in the field of biotechnology. We focus on the time frame 1992–1997, a period of rapid growth in the biotech industry.

In a first step, all patents and publications have been allocated to their respective regions based on the address information of applicants (patents) and authors (publications) following the “ patent allocation methodology ” developed by Lecocq et al. (2011) . Table 1 shows, for every country, the regional level of analysis selected in this study. In order to obtain regional units of analysis which are comparable across continents, institutional subdivisions of countries were chosen for which local authorities have administrative or policy competencies (e.g. states in the USA, prefectures in Japan, and government regions in Germany). In addition, the regional level of analysis was selected to provide most comparable units of analysis in terms of population: NUTS 2 for the larger countries and NUTS 1 for the smaller European countries 4 ; the largest US state, 5 California, was split into North and South California, as the state covers two large and distinct biotech clusters. Only those regions that developed a substantial amount of biotech activity over the period 1992–1997 (minimum 30 patent applications, i.e. on average five patents/year) are retained for the analyses. 6Appendix A provides an overview of these 78 biotech regions.

Table 1.

Regional level of analysis

Country, regional level  Population (in ‘000) a
 
Mean Median Max Min 
North America 4803 3412 20946 27 
 Canada: provinces and territories ( n  = 13)      
 United States: states ( n  = 51) b     
Europe (EU-15 and Switzerland) 1954 1593 11046 26 
 Denmark and Luxemburg: country     
 Austria, Belgium and Greece: Nuts 1 regions     
 Finland, France, Germany, Ireland, Italy, Netherlands, Portugal, Spain, Sweden, Switzerland, and United Kingdom: Nuts 2 regions     
Asia-Pacific 2657 2005 12064 196 
 Australia: states and territories ( n  = 8)      
 Japan: prefectures ( n  = 47)      
Country, regional level  Population (in ‘000) a
 
Mean Median Max Min 
North America 4803 3412 20946 27 
 Canada: provinces and territories ( n  = 13)      
 United States: states ( n  = 51) b     
Europe (EU-15 and Switzerland) 1954 1593 11046 26 
 Denmark and Luxemburg: country     
 Austria, Belgium and Greece: Nuts 1 regions     
 Finland, France, Germany, Ireland, Italy, Netherlands, Portugal, Spain, Sweden, Switzerland, and United Kingdom: Nuts 2 regions     
Asia-Pacific 2657 2005 12064 196 
 Australia: states and territories ( n  = 8)      
 Japan: prefectures ( n  = 47)      

a Average population in the period under study (1992–1997).

b The state of California was split into North and South California.

4.2 Sector allocation of assignees

The “ sector allocation methodology ” developed by Du Plessis et al. (2011) allows us to identify which type of actor (private firms, public universities and research centers, research hospitals, and/or persons) applied for the patent. Based on the “ name harmonizing method ” of Magerman et al. (2011) , we identify the firm and/or other actor with the largest number of biotech patents in the region. In the study, we refer to those firms and other actors in each region as the “lead company” and the “lead actor” respectively. The “lead company” in the region is further classified as “New Dedicated Biotech Firm” (NDBF), “Established Firm” (EF), or “Other Firm” according to the definitions in Table 2 . This classification of firms relies on information on the industry(ies) in which the firm is (primarily) active, its year of establishment and the location of the headquarters 7 retrieved from company Web sites and other sources 8 such as reports on merger and acquisition activities, reports on new products and technologies in the field of life science, market research reports, and company profiles.

Table 2.

Refinement of the typology of firms

New Dedicated Biotech Firm (NDBF) Firm primarily active in the field of biotechnology and established after 1974. 
Established Firm (EF) Firm primarily active in fields other than biotechnology (e.g. pharmaceutical, chemical, food, and other industries) and established before 1974. 
Other Firm Firm active in the field of biotechnology but not as a product or research company (e.g. regional technology transfer offices, venture capitalist, regional industrial agency). 
New Dedicated Biotech Firm (NDBF) Firm primarily active in the field of biotechnology and established after 1974. 
Established Firm (EF) Firm primarily active in fields other than biotechnology (e.g. pharmaceutical, chemical, food, and other industries) and established before 1974. 
Other Firm Firm active in the field of biotechnology but not as a product or research company (e.g. regional technology transfer offices, venture capitalist, regional industrial agency). 

4.3 Regional texture and performance variables

Table 3 provides an overview of the regional texture variables used in the analyses. All variables have been constructed based on the characteristics (assignee type and address information) of the biotech patent applications and biotech publications. The industrial texture characteristics of regions imply the share of patents in the region owned by companies (“ Share of company patents ”), the count of firms in the regions that are active in biotech technology development (“ Number of firms ”) as well as the degree of concentration of industrial biotech technology development in the region’s leading firm (measured by the concentration ratio, “ Company concentration index ”). As a measure of the scientific capabilities of regions, the number of publications normalized by population (“ Science intensity of the region ”) is used. This measure includes both publications from scientific actors and company publications. The ratio of the total number of patents owned by knowledge institutes to the total number of publications in the region (“ Entrepreneurial orientation of knowledge institutes ”) is used as an indicator of the entrepreneurial attitude of knowledge institutes in the region.

Table 3.

Regional texture variables

Variable Description 
Share of company patents Share of company-owned biotech patents over the total number of biotech patents in the region. 
Number of firms Number of companies in the region active in biotech patenting. 
Company concentration index Ratio of the number of biotech patents of the leading firm in the region to the total number of company biotech patents in the region. 
Science-intensity of the region Number of biotech publications in the region per 1000 inhabitants. 
Entrepreneurial orientation of knowledge institutes Ratio of the total number of biotech patents applied by public knowledge generating institutes in the region to the total number of biotech publications in the region. 
Variable Description 
Share of company patents Share of company-owned biotech patents over the total number of biotech patents in the region. 
Number of firms Number of companies in the region active in biotech patenting. 
Company concentration index Ratio of the number of biotech patents of the leading firm in the region to the total number of company biotech patents in the region. 
Science-intensity of the region Number of biotech publications in the region per 1000 inhabitants. 
Entrepreneurial orientation of knowledge institutes Ratio of the total number of biotech patents applied by public knowledge generating institutes in the region to the total number of biotech publications in the region. 

To measure the performance of regions in terms of novel biotech technology development activities (“ Technological performance ”), only triadic patents, i.e. EPO applications with patents of the same family filed at both the USPTO and JPO patent offices were retained. 9 The regions with the highest count of triadic patents (based on assignee addresses) in biotechnology are considered as leading regions in biotech globally (“ top region ”).

4.4 Control variables

Previous research (e.g. Jaffe, 1989 ; Jaffe et al ., 1993 ; Anselin et al. , 2000 ; Acs et al. , 2002 ; Leten et al. , 2014 ) has shown that regions benefit in terms of technological and innovative performance from knowledge spillovers emanating from universities, public and private R&D in the region. For science-intensive industries, also positive spillover effects from neighboring regions were found. Knowledge spillovers from the best performing, top biotech regions are expected to be strongest. Therefore a variable “Neighbouring to top region” is inserted in the analyses, which takes the value “0” if the region is not situated along a top region in year t, the value “1” if the region is neighboring to one of the top biotech region in year t, and a value “2” if the region is neighboring to multiple top biotech regions in year t.

Next, given the importance of inter-organizational collaboration in the field of biotechnology (e.g. Arora and Gambardella, 1990 ; Gertler and Vinodrai, 1996; Powell et al. , 1996 ; Roijakkers and Hagedoorn, 2006 ), measures for international R&D collaboration based on co-patenting between assignees from different countries will be included in our analyses: based on the type of foreign assignee, we distinguish between (i) international technology collaborations with firms (“ International collaboration with firms ”) and (ii) international technology collaborations with knowledge institutes (“ International collaboration with knowledge institutes ”). It should be noted that not all international technology collaborations will be captured with such co-patenting indicators as not all collaborative R&D efforts result in a patent application. In addition, specific IP arrangements between the collaboration partners involved might result in patent applications from only one partner. This is particularly the case for collaboration between public knowledge-generating institutes and industry, where patent rights are often transferred to the industrial partner ( Meyer-Krahmer and Schmoch, 1998 ; Saragossi and van Pottelsberghe de la Potterie, 2003 ; Breschi et al. , 2007 ). The international collaboration indicators are based on assignees type and location, and therefore do not encompass interactions between inventors from different countries within the same local assignee, between local inventors and foreign assignees, or local assignees and foreign inventors. The international collaboration variables used in the analyses should therefore be viewed as a conservative estimation of international collaboration, with actual levels being higher.

Finally, a population variable (expressed in thousands population) controls for differences in the size of the regions.

Table 4 provides descriptive statistics and correlation coefficients of the regional performance, the texture, and control variables in the study. In general, the large standard deviation in relation to the mean value of the variables indicates the presence of large differences between the regions under study in terms of regional performance, industrial texture characteristics, science intensity, and the entrepreneurial orientation of the knowledge institutes in the regions. The technological performance of a region is principally correlated with the number of firms in the region active in biotech technology development. None of the indicators used as explanatory variables in the regression analyses show excessively high correlations.

Table 4.

Descriptive statistics

 Variable (468 observations) Mean SD 
Technological performance 16.182 22.577 1.000         
Share of company patents 0.659 0.270 0.117 1.000        
Number of firms 7.577 10.253 0.838 0.191 1.000       
Company concentration index 0.509 0.267 −0.332 0.034 −0.494 1.000      
Science-intensity of the region 0.217 0.136 0.175 0.017 0.128 −0.064 1.000     
Entrepreneurial orientation of knowledge institutes 0.006 0.006 0.391 −0.565 0.233 −0.143 0.100 1.000    
International collaborations with knowledge institutes 0.321 0.848 0.374 0.230 0.308 −0.093 0.072 −0.001 1.000   
International collaborations with knowledge firms 0.316 0.824 0.439 0.193 0.405 −0.154 0.131 0.043 0.293 1.000  
Population (‘000) 5416 4114 0.491 −0.209 0.489 −0.377 −0.377 0.291 0.057 0.146 1.000 
 Variable (468 observations) Mean SD 
Technological performance 16.182 22.577 1.000         
Share of company patents 0.659 0.270 0.117 1.000        
Number of firms 7.577 10.253 0.838 0.191 1.000       
Company concentration index 0.509 0.267 −0.332 0.034 −0.494 1.000      
Science-intensity of the region 0.217 0.136 0.175 0.017 0.128 −0.064 1.000     
Entrepreneurial orientation of knowledge institutes 0.006 0.006 0.391 −0.565 0.233 −0.143 0.100 1.000    
International collaborations with knowledge institutes 0.321 0.848 0.374 0.230 0.308 −0.093 0.072 −0.001 1.000   
International collaborations with knowledge firms 0.316 0.824 0.439 0.193 0.405 −0.154 0.131 0.043 0.293 1.000  
Population (‘000) 5416 4114 0.491 −0.209 0.489 −0.377 −0.377 0.291 0.057 0.146 1.000 

5. Analyses

5.1 Top regions in biotech

Biotechnology development activities are highly concentrated in a few regions or clusters worldwide ( Feldman and Florida, 1994 ; Audretsch and Feldman, 1996 ). Our data provides evidence that, in the rapid growth phase of the biotech industry (period 1992–1997), the 15 leading biotech regions worldwide—in terms of biotech technology development measured by the count of triadic patents—account for 60% of all biotech patent activity. Table 5 shows that eight top regions are located in the USA, e.g. North California (San Francisco region), Massachusetts (Boston), and South California (San Diego region). Japan has two top regions in biotech: Tokyo and Osaka. Europe counts five top biotech regions, of which the three largest are Île-de-France (Paris region, France), Denmark, and London (UK). Neighboring top biotech regions are mostly found in the USA, in particular on the east coast of the USA (Massachusetts-New York-New Jersey, Pennsylvania-Maryland).

Table 5.

Leading biotech regions

Rank Region, country Triadic patents 1992–1997 (rank) Triadic patents 1978–1990 (rank) 
North California, USA 667 (1) 271 (4) 
Massachusetts, USA 495 (2) 277 (2) 
South California, USA 430 (3) 185 (9) 
Tokyo-TO, Japan 422 (4) 441 (1) 
New Jersey, USA 374 (5) 223 (6) 
New York, USA 358 (6) 275 (3) 
Île-de-France, France 325 (7) 187 (8) 
Maryland, USA 291 (8) 57 (23) 
Pennsylvania, USA 259 (9) 71 (16) 
10 Denmark, Denmark 195 (10) 68 (19) 
11 Osaka-FU, Japan 194 (11) 264 (5) 
12 Karlsruhe, Germany 173 (12) 204 (7) 
13 Inner London, UK 157 (13) 162 (12) 
14 Illinois, USA 151 (14) 108 (14) 
15 Nordwestschweiz, Switzerland 145 (15) 120 (13) 
Rank Region, country Triadic patents 1992–1997 (rank) Triadic patents 1978–1990 (rank) 
North California, USA 667 (1) 271 (4) 
Massachusetts, USA 495 (2) 277 (2) 
South California, USA 430 (3) 185 (9) 
Tokyo-TO, Japan 422 (4) 441 (1) 
New Jersey, USA 374 (5) 223 (6) 
New York, USA 358 (6) 275 (3) 
Île-de-France, France 325 (7) 187 (8) 
Maryland, USA 291 (8) 57 (23) 
Pennsylvania, USA 259 (9) 71 (16) 
10 Denmark, Denmark 195 (10) 68 (19) 
11 Osaka-FU, Japan 194 (11) 264 (5) 
12 Karlsruhe, Germany 173 (12) 204 (7) 
13 Inner London, UK 157 (13) 162 (12) 
14 Illinois, USA 151 (14) 108 (14) 
15 Nordwestschweiz, Switzerland 145 (15) 120 (13) 

Triad patents, period 1992–1997, based on assignee addresses.

For the 78 biotech regions under study, we find further evidence of a high correlation ( r = 0.84) between the number of triadic patents in the early phase (period 1978–1990) and the number of triadic patents in the growth stage (period 1992–1997). Likewise, a high correlation is found between the ranking of regions in terms of biotech patents in both periods ( r = 0.82), suggesting the presence of important early mover advantages at the regional level in the development of biotech activities.

5.2 Towards a typology of (leading) biotech regions

The history of the biotech industry illustrates that different types of actors, ranging from private firms (NDBFs and EFs) to public knowledge institutes (universities and public research centers) and research hospitals, are involved in biotech technology development. In this part of the analysis, we investigate whether, during the growth phase of biotech, technology development activities in “top” regions are to a larger extent driven by firms compared to the other biotech regions (Hypothesis 1a).

Figure 2 shows the technological performance of regions and the share of biotechnology development activity undertaken by private firms for the 78 regions under study. The figure again confirms the strong geographical concentration of biotech technology development. Overall, no obvious, linear relationship can be discerned between the performance of regions and the share of technology development undertaken by private firms. On the one hand, it is clear that there are no “top regions” when the share of companies (in terms of technology development) is below 40%. On the other hand, when looking at the leading regions only, we note that, in some regions, technology development activities are highly concentrated within firms (share of company patents above 75%), while in other regions, technology development is much more distributed over private firms and other types of actor (where the share of company-owned patents is approximately 50%). These results indicate that, to become a leading biotech region in the growth phase of biotech, regional technology development activities do not need to be driven primarily by private firms and, thus, Hypothesis 1a only holds in part.

Figure 2.

The technological performance of regions and the share of biotech technology development activities undertaken by private firms (period 1992–1997, 78 biotech regions). Notes. The top 15 leading regions in biotech technology development over the period 1992–1997 are 1. North California, USA; 2. Massachusetts, USA; 3. South California, USA; 4. Tokyo-TO, Japan; 5. New Jersey, USA; 6. New York, USA; 7. Île-de-France, France; 8. Maryland, USA; 9. Pennsylvania, USA; 10. Denmark; 11. Osaka-FU, Japan; 12. Karlsruhe, Germany; 13. Inner London, UK; 14. Illinois, USA; 15. Nordwestschweiz, Switzerland.

Figure 2.

The technological performance of regions and the share of biotech technology development activities undertaken by private firms (period 1992–1997, 78 biotech regions). Notes. The top 15 leading regions in biotech technology development over the period 1992–1997 are 1. North California, USA; 2. Massachusetts, USA; 3. South California, USA; 4. Tokyo-TO, Japan; 5. New Jersey, USA; 6. New York, USA; 7. Île-de-France, France; 8. Maryland, USA; 9. Pennsylvania, USA; 10. Denmark; 11. Osaka-FU, Japan; 12. Karlsruhe, Germany; 13. Inner London, UK; 14. Illinois, USA; 15. Nordwestschweiz, Switzerland.

Table 6 shows, for each of the 15 main biotech regions, the “lead actor(s)” in the region, where “lead actor” is defined as the organization with the largest number of biotech patent applications in the years 1992 to 1997. For the leading biotech regions where technology development is highly concentrated in private firms, the leading organization in the region is always an EF, primarily active in pharmaceuticals. In the leading biotech regions where technological activity is more widely distributed over private firms and other actors, the leading organizations in the region consist of a combination of public research institutes (university, research center, or research hospital) and private firms (NDBF or EF).

Table 6.

Leading organizations in the top biotech regions

Region Organization name Organization type 
1. North California, USA Genentech Inc. New Dedicated Biotech Firm 
Incyte New Dedicated Biotech Firm 
University of California University 
2. Massachusetts, USA General Hospital Corporation Hospital 
Genetics Institute New Dedicated Biotech Firm 
3. South California, USA Amgen New Dedicated Biotech Firm 
Gen-Probe Incorporated New Dedicated Biotech Firm 
Scripps Research Institute Research Center 
4. Tokyo-TO, Japan Ajinomoto Co., Inc. Established Pharmaceutical Firm 
Kyowa Hakko Kogyo Co., Ltd. Established Firm 
5. New Jersey, USA Becton Dickinson & Co. Established Pharmaceutical Firm 
Merck Established Pharmaceutical Firm 
6. New York, USA Bristol Myers Squibb Co. Established Pharmaceutical Firm 
Johnson & Johnson Established Pharmaceutical Firm 
Ludwig Institute for Cancer Research Research Center 
New York University University 
7. Île-de-France, France Institut National de la Sante et de la Recherche Medicale (INSERM) Research Center 
Institut Pasteur Research Center 
Rhone-Poulenc AG Established Firm 
8. Maryland, USA Department of Health and Human Services Research Center 
Human Genome Sciences, Inc. New Dedicated Biotech Firm 
9. Pennsylvania, USA BAYER AG Established Firm 
Smithkline Beecham Established Pharmaceutical Firm 
University of Pennsylvania University 
10. Denmark Novo Group Established Pharmaceutical Firm 
11. Osaka-FU, Japan Ono Pharmaceutical Co., Ltd. Established Pharmaceutical Firm 
Sumitomo Electric Industries, Ltd. Established Firm 
Suntory Limited Established Firm 
Takeda Chemical Industries, Ltd. Established Firm 
12. Karlsruhe, Germany Roche Diagnostics Established Pharmaceutical Firm 
13. Inner London, UK Cancer Research Campaign Technology Limited Other Firm 
Medical Research Council Research Center 
Unilever Established Firm 
Zeneca Established Pharmaceutical Firm 
14. Illinois, USA Abbott Laboratories Established Pharmaceutical Firm 
15. Nordwestschweiz, Switzerland F. Hoffmann-La Roche AG Established Pharmaceutical Firm 
Novartis Established Pharmaceutical Firm 
Region Organization name Organization type 
1. North California, USA Genentech Inc. New Dedicated Biotech Firm 
Incyte New Dedicated Biotech Firm 
University of California University 
2. Massachusetts, USA General Hospital Corporation Hospital 
Genetics Institute New Dedicated Biotech Firm 
3. South California, USA Amgen New Dedicated Biotech Firm 
Gen-Probe Incorporated New Dedicated Biotech Firm 
Scripps Research Institute Research Center 
4. Tokyo-TO, Japan Ajinomoto Co., Inc. Established Pharmaceutical Firm 
Kyowa Hakko Kogyo Co., Ltd. Established Firm 
5. New Jersey, USA Becton Dickinson & Co. Established Pharmaceutical Firm 
Merck Established Pharmaceutical Firm 
6. New York, USA Bristol Myers Squibb Co. Established Pharmaceutical Firm 
Johnson & Johnson Established Pharmaceutical Firm 
Ludwig Institute for Cancer Research Research Center 
New York University University 
7. Île-de-France, France Institut National de la Sante et de la Recherche Medicale (INSERM) Research Center 
Institut Pasteur Research Center 
Rhone-Poulenc AG Established Firm 
8. Maryland, USA Department of Health and Human Services Research Center 
Human Genome Sciences, Inc. New Dedicated Biotech Firm 
9. Pennsylvania, USA BAYER AG Established Firm 
Smithkline Beecham Established Pharmaceutical Firm 
University of Pennsylvania University 
10. Denmark Novo Group Established Pharmaceutical Firm 
11. Osaka-FU, Japan Ono Pharmaceutical Co., Ltd. Established Pharmaceutical Firm 
Sumitomo Electric Industries, Ltd. Established Firm 
Suntory Limited Established Firm 
Takeda Chemical Industries, Ltd. Established Firm 
12. Karlsruhe, Germany Roche Diagnostics Established Pharmaceutical Firm 
13. Inner London, UK Cancer Research Campaign Technology Limited Other Firm 
Medical Research Council Research Center 
Unilever Established Firm 
Zeneca Established Pharmaceutical Firm 
14. Illinois, USA Abbott Laboratories Established Pharmaceutical Firm 
15. Nordwestschweiz, Switzerland F. Hoffmann-La Roche AG Established Pharmaceutical Firm 
Novartis Established Pharmaceutical Firm 

Lead actor based on biotech EPO patent count, period 1992–1997.

Analysis of the texture characteristics of the top biotech regions thus provides evidence for the presence of two types of region: regions in which technology development is mainly located or concentrated in private firms—hereafter called “ concentrated regions ”—and regions where technology development is not dominated by the private sector but more broadly shouldered or “distributed” between private firms and entrepreneurial universities and/or research centers/hospitals, hereafter referred to as “ distributed regions ”. Figure 2 shows that both a distributed and a concentrated texture can give rise to a leading technology cluster in biotech.

The Wilcoxon–Mann–Whitney test statistics on the refined texture variables in Table 7 reveal some distinct features of leading “distributed” versus leading “concentrated” biotech regions. Leading “concentrated” regions are characterized by a higher share of technology development activities in private firms. Technology development activities by private firms is also much more concentrated in the region’s leading firm than is the case in the leading “distributed” regions. On average, however, the leading “distributed” regions have a greater number of private firms active in biotech technology development than the leading “concentrated” regions. Leading “distributed” regions are further characterized by a higher science intensity in the region, measured by the number of publications per population, as well as the presence of universities and research centers with a more entrepreneurial orientation.

Table 7.

Texture characteristics of leading ‘distributed’ versus ‘concentrated’ regions

  ‘Distributed’ regions ( n  = 8)  ‘Concentrated’ regions (n = 7) z -valueWMW test  
Share of company patents 0.581 0.913  −7.938 *** 
Number of firms 23.813 20.071 2.481** 
Company concentration index 0.273 0.419 −2.447** 
Science intensity of the region 0.314 0.215  3.453 *** 
Entrepreneurial orientation of knowledge institutes 0.014 0.004  7.312 *** 
  ‘Distributed’ regions ( n  = 8)  ‘Concentrated’ regions (n = 7) z -valueWMW test  
Share of company patents 0.581 0.913  −7.938 *** 
Number of firms 23.813 20.071 2.481** 
Company concentration index 0.273 0.419 −2.447** 
Science intensity of the region 0.314 0.215  3.453 *** 
Entrepreneurial orientation of knowledge institutes 0.014 0.004  7.312 *** 

Mean value, Wilcoxon–Mann–Whitney test, based on yearly figures, period 1992–1997.

***, **, *Represents statistical significance at the 1%, 5%, and 10% level, respectively.

5.3 What differentiates leading regions from other biotech regions?

During the growth phase of biotech (period 1992–1997), some biotech regions catch up while other regions fall back in the ranking of (leading) biotech regions. In this section, we analyze which texture characteristics differentiates leading regions from other biotech regions by means of logit regression models with the following functional form:  

P(yit=1/xit),where xitcontainstheexplanatoryandthecontrolvariables
The analyses comprise all 78 biotech regions in our study, with the dependent variable taking the value “1” if region i is among the top 15 biotech regions in year t, and the value “0” if region i is not a top biotech region in year t. Random effects are used to control for the unobserved heterogeneity of regions. The explanatory variables are the refined texture variables presented in the data section ( Table 3 ). They are related to science and entrepreneurial oriented research institutes in the regions, as well as the industrial texture characteristics of regions. We further include a spatial variable in the model indicating whether a region is neighboring to one or more top biotech region(s) in year t, and international R&D collaboration variables. A population variable controls for differences in the size of the regions, while the US dummy variable is added to control for other non-observed regional differences between the USA and other regions, such as institutional and cultural differences. Finally, year dummy variables are added in the regression models to control for time-specific effects.

Table 8 shows the results of the logit regression models. First, the regression is run for all 78 biotech regions (Model 1). As prior results in this article showed that leading regions have different texture characteristics, separate analyses are also run for regions with a “distributed” texture ( n  = 50, Model 2), and regions with a “concentrated” texture ( n  = 28, Model 3), where the latter have been defined as those regions in which technology development activities are predominantly located in private firms (share of company patents ≥ 0.75), and the leading player in the region (period 1992–1997) is an EF.

Table 8.

Random Effect Logit models

 Model 1 All regions Model 2 ‘distributed’ regions Model 3 ‘concentrated’ regions 
Science-intensity of the region 11.7855**  79.2463 *** 12.9222 
(5.6426) (26.4937) (8.6062) 
Entrepreneurial orientation of knowledge institutes 176.4081** 1272.0633** 384.6875** 
(87.7313) (574.8714) (179.2008) 
Number of firms  0.6510 *** 3.0694** 0.4323** 
(0.1647) (1.4093) (0.2198) 
Company concentration index  7.8224 *** 31.6114 11.9934** 
(2.9752) (23.4744) (5.5524) 
Neighboring to top regions 1.1404 1.4101 1.5538 
(0.7737) (3.9083) (1.2695) 
International collaboration with knowledge institutes 0.8532* 6.1586 1.0848* 
(0.5077) (5.6847) (0.6051) 
International collaboration with firms −0.228 −0.3491 0.0009 
(0.4527) (3.6987) (0.5157) 
US dummy 0.1395 11.5101* −1.6593 
(1.5308) (6.7333) (3.001) 
Population 0.0003 0.001 0.001 
(0.0002) (0.0008) (0.0006) 
Constant  −18.3525 ***  −100.8265 ***  −21.5144 *** 
(4.4994) (34.7372) (8.042) 
Year dummies Yes yes Yes 
Observations 468 300 168 
Log likelihood −61.2552 −12.2199 −34.2609 
P 0.0409 0.0746 0.716 
 Model 1 All regions Model 2 ‘distributed’ regions Model 3 ‘concentrated’ regions 
Science-intensity of the region 11.7855**  79.2463 *** 12.9222 
(5.6426) (26.4937) (8.6062) 
Entrepreneurial orientation of knowledge institutes 176.4081** 1272.0633** 384.6875** 
(87.7313) (574.8714) (179.2008) 
Number of firms  0.6510 *** 3.0694** 0.4323** 
(0.1647) (1.4093) (0.2198) 
Company concentration index  7.8224 *** 31.6114 11.9934** 
(2.9752) (23.4744) (5.5524) 
Neighboring to top regions 1.1404 1.4101 1.5538 
(0.7737) (3.9083) (1.2695) 
International collaboration with knowledge institutes 0.8532* 6.1586 1.0848* 
(0.5077) (5.6847) (0.6051) 
International collaboration with firms −0.228 −0.3491 0.0009 
(0.4527) (3.6987) (0.5157) 
US dummy 0.1395 11.5101* −1.6593 
(1.5308) (6.7333) (3.001) 
Population 0.0003 0.001 0.001 
(0.0002) (0.0008) (0.0006) 
Constant  −18.3525 ***  −100.8265 ***  −21.5144 *** 
(4.4994) (34.7372) (8.042) 
Year dummies Yes yes Yes 
Observations 468 300 168 
Log likelihood −61.2552 −12.2199 −34.2609 
P 0.0409 0.0746 0.716 

78 biotech regions, yearly figures, period 1992–1997.

***, **, *Represents statistical significance at the 1%, 5%, and 10% level, respectively.

The regression results in Table 8 (Model 1) reveal that a stronger entrepreneurial orientation in the region’s knowledge institutes as well as increasing numbers of firms active in biotech technology development contribute to becoming a “top region” in biotech. These results hold for both “distributed” (Model 2) and “concentrated” regions (Model 3) and indicate that, in science-intensive industries such as biotechnology, Hypothesis 2b does not hold: despite the greater diffusion of scientific knowledge globally, the entrepreneurial orientation of scientific actors in the region remains important in becoming a leading region during the growth phase of biotech. Equally, the results indicate that the creation or attraction of companies active in biotech technology development is also instrumental.

Next, the regression results reveal that “distributed” regions (Model 2) benefit from higher levels of science intensity, while no significant impact is found in the “concentrated” regions (Model 3). Thus, the results indicate that in “distributed regions”, in contrast to Hypothesis 2a, the continuous development of a strong science base remains essential—and this holds true in the growth phase of the technology as well.

The analyses (Model 3) further reveal that leading biotech regions with a “concentrated” texture not only benefit from increasing numbers of firms active in biotech technology development but also higher levels of concentration within an “anchor-tenant” firm are instrumental in developing a leading position as well. These results confirm that, for regions with a “concentrated” texture, Hypothesis 1b holds true: regions with higher concentrations of regional biotech technology development activities in an anchor-tenant firm are more likely to become a leading biotech region in the growth phase of biotech.

The co-patent data reveals that top regions are engaging substantially in different kinds of international R&D collaborations: collaborations between firms, between firms and knowledge institutes, and between knowledge institutes. The regression results provide evidence that “concentrated” regions (Model 3) also benefit from international technology collaborations with knowledge institutes. A large part of these international collaborations consist of collaborations of large established pharmaceutical firms in the top “concentrated” regions with leading research centers, often located in top “distributed” regions. For the “distributed” regions (Model 2), no similar effect is found in terms of international collaboration. The results also reveal no significant impact from international technology collaborations with firms.

For the “distributed” regions, the US dummy variable is positive and significant, although the result is not very outspoken (significant only at the 10% level). This suggests that being located in the USA appears to increase the likelihood of becoming a leading region with “distributed” texture characteristics. This result differs from a possible size-effect resulting from the fact that the US regions in our sample are on average larger than the other regions in the analyses. Indeed, the absence of a significant effect of the population variable in Model 2 indicates that regions are not more likely to attain “top” status when their population size is larger. The absence of a region size effect also holds for the “concentrated” regions.

The spatial proximity variable, indicating whether a region is neighboring to one or more top biotech regions, is positive but not significant. Hence, neighboring a top biotech region does not seem to contribute directly in becoming a top biotech region, after controlling for regional texture variables related to science, entrepreneurial orientation of knowledge institutes, and the composition of the industrial texture in terms of number of firms active in biotech and presence of an anchor firm.

All the variables used in the model are contemporaneous, hence the results show “what differentiates leading regions from other biotech regions” but do not imply any causality.

Robustness checks with the top 13 and top 17 leading regions in biotech as dependent variable show that the results for all regions still hold. When engaging in more refined analyses for the “distributed” and “concentrated” regions separately, we find that results remain robust when constraining the definition of top regions from 15 to 13 regions, while the results related to science and entrepreneurial attitude of knowledge institutes are less outspoken when increasing the number of top regions from 15 to 17 regions. This pattern should come as no surprise, as extending the number of top regions implies more variety (within the group of top regions). As such, the findings for the top 13 regions are re-assuring: the observed differences—between top and “average” regions—are not driven by idiosyncrasies.

We also verified the results of the analyses for different thresholds for being a “concentrated” region. When lowering or raising the threshold of the share of regional patents by companies to 0.70 and 0.80 respectively, the results for both the “distributed” and the “concentrated” regions remain robust.

6. Discussion and conclusions

In this article, the texture characteristics of regions (industry composition, presence of entrepreneurial-orientated scientific actors) are studied in relation to their technological performance in the field of biotechnology. Our analyses comprise 78 regions in North America, Europe, and Asia-Pacific that developed a substantial amount of biotech activity over the time period 1992–1997. The period under study corresponds with an era of rapid growth in the biotech industry in which industrial capabilities clearly became more important.

Our results confirm that biotech technology development activities are highly concentrated in a limited number of top regions worldwide ( Feldman and Florida, 1994 ; Audretsch and Feldman, 1996 ). Evidence is provided for the existence of two types of leading biotech regions: “concentrated” regions in which technology development is mainly located in private firms and “distributed” regions where both private firms and entrepreneurial universities and/or research centers/hospitals significantly contribute to regional biotech technology development. The high rank order correlation—in terms of the technological performance of biotech regions in the early and growth phases of the biotech industry—suggests the presence of important early mover advantages at the level of regions in new and emerging science-based fields.

Using random effect logit models, we further analyze which texture variables differentiate leading regions from other biotech regions. The empirical analyses indicate that regions with “concentrated” texture characteristics benefit, in terms of overall technological activity, from increased concentrations of technology development activities in a leading firm, thereby supporting the anchor-tenant hypothesis proposed by Agrawal and Cockburn (2003) . Further research reveals that the “anchor” firm(s) in the leading “concentrated” biotech regions is (are) large, R&D-intensive firms primarily active in the pharmaceutical, chemical, food, and other industries, and established well before the creation of the first dedicated biotech firms in the second half of the 1970s. Our analyses suggest that these large EFs, which have extensive industry experience and important access to (internal) financial resources, have been of particular importance for the development of regional biotech technology activities in the growth phase of the biotech industry. Following Agrawal and Cockburn (2003) , such large R&D-intensive firms, by creating local niche and/or intermediary markets, may have played an important role in breeding regional entrepreneurial initiatives in the biotechnology field and attracting high-quality suppliers to the region. Our results also indicate that regions with a “concentrated” texture benefit from international technology collaborations between large EFs in the region and worldwide leading scientific actors. In science-based industries such as biotechnology, also the presence of entrepreneurial-oriented knowledge institutes in the region matter.

While the role of science and entrepreneurial-orientated universities and research centers is widely acknowledged as important in the early incubation phase of new, science-based technologies, our study shows that, in the growth phase of the biotech industry, the orientation and contribution of scientific actors in terms of technology development differentiates top “distributed” regions from other biotech regions. Indeed, our results show that top “distributed” regions benefit, in addition to acquiring an excellent science base, from a more entrepreneurial orientation by their knowledge institutes. Leading regions “distributed” texture also create industrial activities in biotechnology by generating new entrepreneurial activities or attracting new firms active in biotech into the region. Additionally, continuous investment in a strong science base remains important in the growth phase of science-based industries.

Our results have important implications for policy makers wishing to foster regional economic activities in science-based technology fields. While previous research has shown the importance of institutional settings such as the legal environment, the availability of venture capital, and a dedicated support infrastructure in the development of regional biotech activities, our results point to the importance of the existing texture characteristics of regions in terms of the presence of entrepreneurial-oriented knowledge institutes and/or large EFs. Our analyses show that large, EFs have been able to adopt the new technological opportunities offered by a disruptive technology and, in several leading regions, dominant anchor firms have played a key role in the development of the new technology, as has been the case in the fields of nanotechnology ( Rothaermel and Thursby, 2007 ; Baglieri et al. , 2012 ; Genet et al. , 2012 ) and microelectronics ( Genet et al. , 2012 ). However, to become a leading region in the growth phase of biotech, regional technology development activities do not necessarily need to be driven by private firms for the most part: entrepreneurial-oriented knowledge institutes continue to play a major role. In those regions characterized by a more “distributed” texture, large incumbent firms do not appear to exercise the dominant role in technology development as they did in the nanotech and microelectronics fields, where they formed an important bridge between public research and the industry itself ( Genet et al. , 2012 ).

The existence of two different avenues for becoming a leading biotech region calls for public policies consistent with the existing texture characteristics of the region. At the same time, our results demonstrate that both entrepreneurial-oriented knowledge institutes and a strong industrial texture consisting of many competing firms is of crucial importance in sustaining innovation and growth in regions during the rapid growth phase of science-based industries. Indeed, as technologies continue to mature and become more specialized, and as the industry as a whole moves through its life cycle and becomes more standardized, further development activities in technology clusters tend to become more homogeneous and increasingly co-aligned, thereby affecting the innovative potential of firms inside the cluster (Pouder and St. John, 1996). To prevent regional lock-in ( Porter 1998 ; Gertler and Vinodrai, 2009 ; Baglieri et al. , 2012 ), entrepreneurial-oriented research institutes and intense technology competition between local firms are essential to stimulate new knowledge creation and the opening of new technological trajectories, and to sustain the innovation dynamics in science-based technology clusters.

Further research into regional development paths during the more mature stages of the biotech industry seems relevant to better understand path dependence and regional lock-in as place-dependent processes ( Martin and Sunley, 2006 ). In this study, the focus is on regional technological development in the biotechnology field. Since biotechnology is a very broad field with applications in many industries, further refinement of our results could take into account the extent to which regions specialize in one or more niches in biotech or, alternatively, develop a wide range of biotech activities, and investigate how different levels of specialization are associated to cluster development (see e.g. Owen-Smith et al. , 2002 ; Gertler and Vinodrai, 2009 ). In this respect, refining the level of analysis toward cities and urban regions might become an interesting venue for further research, as regions still host a lot of “within variety” which might further inform our understanding of—differentiated—growth trajectories (see e.g. Feldman and Audretsch, 1999 ; Boschma et al. , 2014 ).

1 The Bayh-Dole Act (1980) allows—and even encourages—US universities to appropriate the results of publicly funded research through patenting.
3 Triadic patents are patents of the same family filed in all three of the major patent offices (USPTO, JPO, and EPO) and thus protecting the same invention in North America, Europe, and Asia.
4 Note that for some smaller European countries (Denmark, Ireland, Luxembourg, Switzerland), the NUTS1 level is missing or consists of a division between the mainland/continent and the island(s) (Finland, Portugal). Therefore the country level was used as level of analysis in the case of Denmark and Luxembourg, and NUTS 2 level for all other small European countries.
7 Information on the (headquarters) location was matched with the address information on the patent to ensure the information retrieved via web sources corresponds with the assignee (company) of the patent application.
8 Since the 1990s has been characterized by not only considerable merger and acquisition activity in the biotechnology field but also a high failure rate among new (biotech) companies, we have had to rely on exhaustive web searches to find company information, especially for companies that have gone out of existence, operate under a different name, or have been acquired in the intervening period.
2 Those subject categories include plant sciences, biochemical research methods, biochemistry and molecular biology, biophysics, biotechnology and applied microbiology, microbiology, cell biology, genetics and heredity, and developmental biology.
5 Some US states are substantially larger than regions in other countries. Only in California, the largest state, evidence was found for the presence of two distinct top biotech clusters. US states are used as level of analysis since further refinement of all US regions to the level of Metropolitan Statistical Areas (MSAs) would bring down the average population of US regions substantially below regions in other countries and disregard nonmetropolitan areas in the USA.
6 Two regions, the Australian Capital Territory (AU) and Washington D.C. (USA) have been further removed from the analyses. Both are a capital territory or district within a federation with special designated legal status. In contrast to the other states in the federation, both regions are very small in terms of population as they only consist of one city, the capital city of the federal state. Because of their special legal status and their small population size, these regions are not comparable with other regions. Note that, based on regional patent counts, these capital territories or districts do not qualify as top region in biotech.
9 For the 78 regions under study, 54% of the EPO patent applications in the domain of biotechnology were also filed at the USPTO and JPO patent offices.

Acknowledgements

The authors would like to thank René Belderbos, Ron Boschma, Bruno Cassiman, Iain Cockburn, Koenraad Debackere, Joep Konings, Jo Reynaerts, Reinhilde Veugelers, and two anonymous referees for their valuable feedback on earlier versions of this research.

References

Acharya
R.
(
1999
),
The Emergence and Growth of Biotechnology. Experiences in Industrialized and Developing Countries
  .
Edward Elgar: Cheltenham, UK
.
Acs
Z.
Anselin
L.
Varga
A.
(
2002
), ‘
Patents and innovation counts as measures of regional production of new knowledge
,’
Research Policy
  ,
31
(
7
),
1069
1085
.
Agrawal
A.
Cockburn
I.
(
2003
), ‘
The anchor tenant hypothesis: exploring the role of large, local, R&D-intensive firms in regional innovation systems
,’
International Journal of Industrial Organization
  ,
21
(
9
),
1227
1253
.
Anselin
L.
Varga
A.
Acs
Z.
(
2000
), ‘
Geographic and sectoral characteristics of academic knowledge externalities
,’
Papers in Regional Science
  ,
79
(
4
),
435
445
.
Arora
A.
Gambardella
A.
(
1990
), ‘
Complementarity and external linkages: the strategies of the large firms in biotechnology
,’
The Journal of Industrial Economics
  ,
38
(
4
),
361
379
.
Arundel
A.
Kabla
I.
(
1988
), ‘
What percentage of innovations are patented? Empirical estimates for European firms
,’
Research Policy
  ,
27
(
2
),
127
141
.
Audretsch
D.
Feldman
M.
(
1996
), ‘
R&D spillovers and the geography of innovation and production
,’
The American Economic Review
  ,
863
,
630
640
.
Baglieri
D.
Cinici
M. C.
Mangematin
V.
(
2012
), ‘
Rejuvenating clusters with ‘sleeping anchors’: the case of nanoclusters
,’
Technovation
  ,
32
(
3–4
),
245
256
.
Bartholomew
S.
(
1997
), ‘
National systems of biotechnology innovation: complex interdependence in the global system
,’
Journal of International Business Studies
  ,
82
(
2
),
241
266
.
Boschma
R.
Balland
P.
Kogler
P. F.
(
2014
), ‘
Relatedness and technological change in cities: the rise and fall of technological knowledge in US metropolitan areas from 1981 to 2010
,’
Industrial and Corporate Change
  ,
24
(
1
),
223
250
.
Breschi
S.
Lissoni
F.
Montobbio
F.
(
2007
), ‘
The scientific productivity of academic inventors: new evidence from Italian data
,’
Economics of Innovation and New Technology
  ,
16
(
2
),
101
118
.
Breznitz
S. M.
O’Shea
R. P.
Allen
T. J.
(
2008
), ‘
University commercialization strategies in the development of regional bioclusters
,’
Journal of Product Innovation Management
  ,
25
(
2
),
129
142
.
Casper
S.
(
2007
), ‘
How do technology clusters emerge and become sustainable? Social network formation and inter-firm mobility within the San Diego biotechnology cluster
,’
Research Policy
  ,
36
(
4
),
438
455
.
Chiaroni
D.
Chiesa
V.
(
2006
), ‘
Forms of creation of industrial clusters in biotechnology
,’
Technovation
  ,
26
(
9
),
1064
1076
.
Cooke
P.
(
2001
), ‘
Regional innovation systems, clusters and the knowledge economy
,’
Industrial and Corporate Change
  ,
10
(
4
),
945
974
.
Criscuolo
P.
(
2006
), ‘
The “home advantage’ effect and patent families: a comparison of OECD triadic patents, the USPTO and the EPO
,’
Scientometrics
  ,
66
(
1
),
23
41
.
Dominquez-Lacasa
I.
(
2006
), ‘
Capturing the changes in the knowledge base underlying drug discovery and development in the 20th century and the adjustment of Bayer, Hoechst, Schering AK and E. Merck to the advent of modern biotechnology
,’
Scientometrics
  ,
66
(
2
),
345
364
.
Du Plessis
M.
Looy
B. Van
Song
X.
Magerman
T.
(
2011
), ‘
Sector allocation
,’
in Patent Statistics at Eurostat: Methods for Regionalisation, Sector Allocation and Name Harmonisation. Eurostat Methodologies and Working Papers
  .
Luxembourg: European Union
.
Feldman
M.
(
2003
), ‘
The locational dynamics of the US biotech industry: knowledge externalities and the anchor hypothesis
,’
Industry and Innovation
  ,
10
(
3
),
311
328
.
Feldman
M.
Florida
R.
(
1994
), ‘
The geographic sources of innovation: technological infrastructure and product innovation in the United States
,’
Annals of the Association of American Geographers
  ,
84
(
2
),
210
229
.
Feldman
M.
Audretsch
D. B.
(
1999
), ‘
Innovation in cities: science-based diversity, specialization and localized competition
,’
European Economic Review
  ,
43
(
2
),
409
429
.
Gambardella
A.
Harhoff
D.
Verspagen
B.
(
2008
), ‘
The value of European patents
,’
CEPR Discussion Paper No 6848
  .
Gambardella
A.
Orsenigo
L.
Pammolli
F.
(
2000
), ‘
Global competitiveness in pharmaceuticals: a European perspective
,’
Report prepared for the Enterprise Directorate-General of the European Commission
  .
Luxembourg: Belgium
.
Galambos
L.
(
2006
), ‘
Innovation and industry evolution: a comment
,’ in
Mazzucato
M.
Dosi
G.
(eds),
Knowledge Accumulation and Industry Evolution: The Case of Pharma-biotech
  .
Cambridge University Press
:
Cambridge, UK
.
Genet
C.
Errabi
K.
Gauthier
C.
(
2012
), ‘
Which model of technology transfer for nanotechnology? A comparison with biotech and microelectronics
,’
Technovation
  ,
32
(
3–4
),
205
215
.
Gertler
M. S.
Vinodrai
T.
(
2009
), ‘
Life sciences and regional innovation: one path or many?
,’
European Planning Studies
  ,
17
(
2
),
235
261
.
Glänzel
W.
Meyer
M.
Schlemmer
B.
Du Plessis
M.
Thijs
B.
Magerman
T.
Debackere
K.
(
2004
), ‘
Domain study “Biotechnology” – an analysis based on publications and patents
,’
Report Published By ECOOM. www.ecoom.be
Gray
M.
Parker
E.
(
1998
), ‘
Industrial change and regional development: the case of the US biotechnology and pharmaceutical industries
,’
Environment and Planning
  ,
30
(
10
),
1757
1774
.
Griliches
Z.
(
1990
), ‘
Patent statistics as economic indicators – a survey
,’
Journal of Economic Literature
  ,
28
(
4
),
1661
1707
.
Hagedoorn
J.
(
2002
), ‘
Inter-firm R&D partnering in pharmaceutical biotechnology since 1975: trends, patterns and networks
,’
Research Policy
  ,
35
(
3
),
431
446
.
Hagedoorn
J.
Cloodt
M.
(
2003
), ‘
Measuring innovative performance: is there an advantage in using multiple indicators?
,’
Research Policy
  ,
32
(
8
),
1365
1379
.
Hall
B.
Jaffe
A.
Trajtenberg
M.
(
2005
), ‘
Market value and patent citations
,’
Rand Journal of Economics
  ,
36
(
1
),
16
38
.
Harhoff
D.
Narin
F.
Scherer
F.
Vogel
K.
(
1999
), ‘
Citation frequency and the value of patented inventions
,’
Review of Economics and Statistics
  ,
81
(
3
),
511
515
.
Houston
S.
(
2003
), ‘
Switzerland - Swiss pharma strength spurs growth in biotech
,’
Chemical Week
  ,
165
(
4
),
35
39
.
Jaffe
A. B.
(
1989
), ‘
Real effects of academic research
,’
American Economic Review
  ,
79
(
5
),
957
969
.
Jaffe
A. B.
Trajtenberg
M.
Henderson
R.
(
1993
), ‘
Geographic localization of knowledge spillovers as evidenced by patent citations
,’
Quarterly Journal of Economics
  ,
108
(
3
),
577
598
.
Lecocq
C.
Looy
B. Van
(
2009
), ‘
The impact of collaboration on the technological performance of regions: time invariant or driven by life cycle dynamics? An explorative investigation of European regions in the field of Biotechnology
,’
Scientometrics
  ,
80
(
3
),
847
867
.
Lecocq
C.
Looy
B. Van
Vereyen
C.
(
2011
), ‘
Regionalisation of patent data. in: patent statistics at eurostat: methods for regionalisation, sector allocation and name harmonisation
,’
Eurostat Methodologies and Working Papers
  .
Luxembourg: European Union
.
Leten
B.
Landoni
P.
Looy
B. Van
(
2014
), ‘
Science or graduates: how do firms benefit from the proximity of universities?
,’
Research Policy
  ,
43
(
8
),
1398
1412
.
Magerman
T.
Peeters
B.
Song
X.
Grouwels
J.
Callaert
J.
Looy
B. Van
(
2011
), ‘
Name harmonisation. in: patent statistics at eurostat: methods for regionalisation, sector allocation and name harmonisation
,’
Eurostat Methodologies and Working Papers
  .
Luxembourg: European Union
.
Mangematin
V.
Lemarié
S.
Boissin
J. P.
Catherine
D.
Corrolleur
F.
Coronini
R.
Trommeter
M.
(
2003
), ‘
Development of SMEs and heterogeneity of trajectories: the case of biotechnology in France
,’
Research Policy
  ,
32
(
4
),
621
638
.
Mansfield
E.
(
1986
), ‘
Patents and innovation: an empirical study
,’
Management Science
  ,
32
(
2
),
173
181
.
Martin
R.
Sunley
P.
(
2006
), ‘
Path dependence and regional economic evolution
,’
Journal of Economic Geography
  ,
6
(
4
),
395
437
.
Meyer-Krahmer
F.
Schmoch
U.
(
1998
), ‘
Science-based technologies: university-industry interactions in four fields
,’
Research Policy
  ,
27
(
8
),
835
851
.
Narin
F.
Noma
E.
(
1987
), ‘
Patents as indicators of corporate technological strength
,’
Research Policy
  ,
16
(
2–4
),
143
155
.
Niosi
J.
(
2011
), ‘
Complexity and path dependence in biotechnology innovation systems
,’
Industrial and Corporate Change
  ,
20
(
6
),
1795
1826
.
Owen-Smith
J.
Riccabonni
M.
Pammolli
F.
Powell
W. W.
(
2002
), ‘
A comparison of US and European university-industry relations in the life sciences
,’
Management Science
  ,
48
(
1
),
24
43
.
Pavitt
K.
(
1985
), ‘
Patent statistics as indicators of innovative activities: possibilities and problems
,’
Scientometrics
  ,
7
(
1
),
77
99
.
Porter
M.
(
1998
), ‘
Clusters and the new economics of competition
,’
Harvard Business Review
  ,
76
(
6
),
77
90
.
Porter
M.
(
2000
), ‘
Location, competition, and economic development: local clusters in a global economy
,’
Economic Development Quarterly
  ,
14
(
1
),
15
34
.
Powell
W. W.
Koput
K. W.
Smith-Doerr
L.
(
1996
), ‘
Interorganizational collaboration and the locus of innovation in biotechnology
,’
Administrative Science Quarterly
  ,
41
(
1
),
116
145
.
Prevezer
M.
(
2001
), ‘
Ingredients in the early development of the US biotechnology industry
,’
Small Business Economics
  ,
17
(
1–2
),
17
29
.
Roijakkers
N.
Hagedoorn
J.
(
2006
), ‘
Inter-firm R&D partnering in pharmaceutical biotechnology since 1975: trends, patterns, and networks
,’
Research Policy
  ,
35
(
3
),
431
446
.
Rosenberg
N.
Nelson
R. R.
(
1994
), ‘
American universities and technical advance in industry
,’
Research Policy
  ,
23
(
3
),
323
348
.
Rothaermel
F. T.
(
2001
), ‘
Complementary assets, strategic alliances, and the incumbent’s advantage: an empirical study of industry and firm effects in the biopharmaceutical industry
,’
Research Policy
  ,
30
(
8
),
1235
1251
.
Rothaermel
F. T.
Hess
A. M.
(
2001
), ‘
Building dynamic capabilities: innovation driven by individual-, firm-, and network-level effects
,’
Organization Science
  ,
18
(
6
),
898
921
.
Rothaermel
F. T.
Thursby
M.
(
2007
), ‘
The nanotech versus the biotech revolution: Sources of productivity in incumbent firm research
,’
Research Policy
  ,
36
(
6
),
832
849
.
Saragossi
S.
van Pottelsberghe de la Potterie
B.
(
2003
), ‘
What patent data reveal about universities: the case of Belgium
,’
Journal of Technology Transfer
  ,
28
(
1
),
47
51
.
Trajtenberg
M.
(
1990
), ‘
A penny for your quotes: patent citations and the value of innovations
,’
Rand Journal of Economics
  ,
21
(
1
),
172
187
.
Utterback
J. M.
(
1994
),
Mastering the Dynamics of Innovation
  .
Harvard Business School Press
:
Boston
.
Zeller
C.
(
2001
), ‘
Clustering biotech: a recipe for success? Spatial patterns of growth of biotechnology in Munich, Rhineland and Hamburg
,’
Small Business Economicsm
  ,
17
(
1–2
),
123
141
.
Zucker
L. G.
Darby
M. R.
(
1996
), ‘
Star scientists and institutional transformation: Patterns of invention and innovation in the formation of the biotechnology industry
,’
Proceedings of the National Academy of Sciences of the United States of America
  ,
93
(
23
),
12709
12716
.
Zucker
L. G.
Darby
M. R.
(
1997
), ‘
Present at the biotechnological revolution: transformation of technological identity for a large incumbent pharmaceutical firm
,’
Research Policy
  ,
26
(
4–5
),
429
446
.

Appendix A. Biotech regions

North America ( n = 32)      Population average 7.6 mio, max 19.8 mio, min 0.7 mio  
Canada: Alberta, British Columbia, Ontario, Quebec 
United States: Arizona, North California, South California, Colorado, Connecticut, Delaware, Florida, Georgia, Illinois, Indiana, Iowa, Maryland, Massachusetts, Michigan, Minnesota, Missouri, New Jersey, New York, North Carolina, Ohio, Oregon, Pennsylvania, Tennessee, Texas, Utah, Virginia, Washington, Wisconsin 
Europe ( n = 36)         Population average 3.2 mio, max 10.9 mio, min 0.9 mio  
Austria: Ostösterreich; Belgium: Vlaams Gewest; Switserland: Région Lémanique, Nordwestschweiz, Zürich; Germany: Karlsruhe, Oberbayern, Berlin, Darmstadt, Gießen, Braunschweig, Düsseldorf, Köln, Rheinhessen-Pfalz; Danmark; Spain: Comunidad De Madrid; Finland: Etelä-Suomi; France: Île De France, Alsace, Rhône-Alpes; Italy: Lombardia, Lazio; Netherlands: Gelderland, Noord-Holland, Zuid-Holland, Limburg; Sweden: Stockholm, Östra Mellansverige: United Kingdom: Greater Manchester, East Anglia, Inner London, Outer London, Berkshire, Buckinghamshire and Oxfordshire, Surrey, East and West Sussex, Gloucestershire, Wiltshire And North Somerset, Eastern Scotland 
Asia Pacific ( n = 10)       Population average 6.3 mio, max 11.9 mio, min 1.9 mio  
Japan: Aichi-Ken, Hyogo-Ken, Kanagawa-Ken, Kyoto-Fu, Okayama-Ken, Osaka-Fu, Saitama-Ken, Tokyo-To 
Australia: New South Wales, Victoria 
North America ( n = 32)      Population average 7.6 mio, max 19.8 mio, min 0.7 mio  
Canada: Alberta, British Columbia, Ontario, Quebec 
United States: Arizona, North California, South California, Colorado, Connecticut, Delaware, Florida, Georgia, Illinois, Indiana, Iowa, Maryland, Massachusetts, Michigan, Minnesota, Missouri, New Jersey, New York, North Carolina, Ohio, Oregon, Pennsylvania, Tennessee, Texas, Utah, Virginia, Washington, Wisconsin 
Europe ( n = 36)         Population average 3.2 mio, max 10.9 mio, min 0.9 mio  
Austria: Ostösterreich; Belgium: Vlaams Gewest; Switserland: Région Lémanique, Nordwestschweiz, Zürich; Germany: Karlsruhe, Oberbayern, Berlin, Darmstadt, Gießen, Braunschweig, Düsseldorf, Köln, Rheinhessen-Pfalz; Danmark; Spain: Comunidad De Madrid; Finland: Etelä-Suomi; France: Île De France, Alsace, Rhône-Alpes; Italy: Lombardia, Lazio; Netherlands: Gelderland, Noord-Holland, Zuid-Holland, Limburg; Sweden: Stockholm, Östra Mellansverige: United Kingdom: Greater Manchester, East Anglia, Inner London, Outer London, Berkshire, Buckinghamshire and Oxfordshire, Surrey, East and West Sussex, Gloucestershire, Wiltshire And North Somerset, Eastern Scotland 
Asia Pacific ( n = 10)       Population average 6.3 mio, max 11.9 mio, min 1.9 mio  
Japan: Aichi-Ken, Hyogo-Ken, Kanagawa-Ken, Kyoto-Fu, Okayama-Ken, Osaka-Fu, Saitama-Ken, Tokyo-To 
Australia: New South Wales, Victoria