Clustering of life-science
firms in the Netherlands
Amsterdam Business School
Part-time MBA thesis
Submission date: August 11, 2013 Author: Richard Hibbert Student number: 10297375
Course: Master of Business Administration
Contact Details: Ottho Heldringstraat 31B, 1066 XT Amsterdam, The Netherlands.
Email: rick.hibbert@gmail.com
1
Executive Summary
Related industries are often geographically clustered. These clusters provide competitive advantage by promoting knowledge transfer. However, the phenomenon is surprising given the globalization of modern business and the abundance of methods of inexpensive communication. As a result, imperfections in knowledge transfer that are overcome by the clusters can be considered to be an economic externality and it is of interest for economists to understand the effect, and of policy makers and managers to exploit it to derive competitive advantage.
Existing research on clusters has focussed upon a few dominant methodologies. Studies have followed small groups of “star” scientists or engineers, analysed patent and other citations to understand the movement of knowledge and human capital, or have analysed average levels of human capital in a given geographical area. Such studies have confirmed the existence of a number of clusters. However, the factors that are important in cluster formation are poorly understood and empirical studies that identify and characterize clusters are lacking.
Bioscience parks provide an appealing setting for such empirical studies since they are designed to provide the benefits of clusters to bioscience firms. In this thesis, a set of biotechnology firms in Leiden Bioscience Park, a successful Dutch science park, were selected together with a comparison set from outside the park. The networking site LinkedIn allowed analysis of the movement of employees between the firms to identify clusters and identify factors that lead to their formation.
The data were analysed against the primary hypotheses that network effects within the park would be stronger than those outside it. Surprisingly, employees were no more likely to find new jobs within the Leiden Bioscience Park than outside it, despite the close geographical proximity and shared resources of these firms. However there was more movement of people between small and large firms than would have been expected. Furthermore, the movement of people between “End-product” and “Service” firms was statistically significant. Consequently this study provides a basis for understanding clusters in terms of movement of human capital and alternative methodologies for further academic studies.
2
Contents
I. Introduction ... 3
II. Research Question ... 8
III. Research Setting ... 10
IV. Data and Methods ... 17
V. Statistical hypothesis testing ... 21
VI. Discussion ... 24
VII. Limitations ... 28
VIII. Summary, conclusions and recommendations ... 30
A. Summary ... 30
B. Conclusions ... 30
C. Recommendations ... 31
3
I. Introduction
Formal inter-firm alliances and networks have become increasingly popular in modern business administration because of potential benefits such as reduced costs, increased responsiveness to environmental change and greater efficiency. Examples of firms utilizing such networks include US biotechnology firm Genentech and the drug distributors McKesson (Lorenzoni and Baden-Fuller, 1995).
Alliances and networks have been extensively studied in the academic literature. For example, the questions of where to position oneself in a network to gain benefits in terms of power, access to resources and innovation has been reviewed at the individual, intragroup and intergroup level (Brass et al. 2004). Two dominant themes are centrality within a network, a proxy for Network Prominence and the number of “structural holes” (Burt, 1992) that one controls, an indication of an entrepreneurial position (Brass et al. 2004). Koka and Prescott (2008), build on this theme in their study of the effect on network position on performance in the steel industry. They conclude that firms which try to occupy both entrepreneurial and prominent positions perform worse than those who do not. Entrepreneurial positions are favourable when environmental conditions change, but both network forms can be detrimental when the change is particularly radical. Conversely, network prominence can be advantageous for firms pursuing an ‘Analyser’ strategy (Miles and Snow, 1978): firms that are intermediate between conservative ‘Defenders’ and first mover ‘Prospectors’.
Within formal networks, the concept of a “Strategic Centre” has been defined as a vital firm in the network that drives the communication and innovation and technology throughout the network such that value is created (Lorenzoni and Baden-Fuller, 1995). Unlike subcontracting, this strategic outsourcing requires firms to be strategic partners that initiate projects. Multinational pharmaceutical firms seem like likely candidates for this strategic centre position within the life sciences industry, as a result of their global reach, their vast resources and their strong brands and control of down-stream processes. Conversely, many smaller research-driven firms are viable candidates as strategic partners.
4
Firms, however, do not necessarily need to form formal networks of alliances to benefit from other firms in the same industry. Michael Porter (1998) observes that many industries tend to be geographically clustered. These clusters can provide access to human resources, innovation and competition, as well as promoting the formation of new enterprises that enhance the cluster further. Such clusters can vary in size and can cross political borders, as exemplified by a chemical cluster in Germany that crosses into Switzerland (Porter, 1998, pg. 79). Studies have shown that geographical clustering of businesses can be associated with improved productivity (Paci and Usai, 2000) and more innovation (Baptista and Swann, 1998). Porter (1998) believes that it should be en explicit aim of governments to work with market mechanisms to promote clusters where they arise. Publically funded institutions such as universities, research institutes and university hospitals provide likely focus points for such regional clusters, and Science Parks and Bioscience Parks are attractive instruments for policy makers attempting to take advantage of clustering effects.
A study of regional clusters analyzed the criteria required for a regional innovation system (Cooke, 2001). This study highlighted that successful regional economies tend to have a successful mix of infrastructure, institutions, firms and policies. The study concluded that the innovation gap between Europe and the US was the result of excessive public intervention in Europe, and that better institutional and private sector support provided the solution (Cooke, 2001).
Much of the academic literature on the benefits of technological clusters has focussed on Silicon Valley in the US, as an example of where clustered technological firms can benefit from each other and gain competitive advantage. Saxenian (1994) addressed why technological firms in the Silicon Valley region thrived whereas those in the Boston region did not, despite their similar histories and industries. The study concludes that informal ties between technical professionals assisted the diffusion of technical innovation from the perspective of the firms, whereas employees could mitigate the career risks of innovative projects, since if one venture fails then they could move to a neighbouring company.
5
In another landmark study, Jaffe et al (1993) studied the geographic localization of patent citations. This study provided evidence that technical knowledge could be geographically localized, since patent citations tend to belong to the same geographical region or state. Furthermore, this situation was quite stable with time, since the localization of patent citations diminished only slowly. The study provides strong evidence that despite the intangible nature of knowledge and ideas, they flow much less freely that would be would be expected. Practitioners have described the localization of knowledge as an externality (see, for example, Almeida and Kogut, 1999); hence it is of interest to economists to understand the basis of clusters and of interest for managers and policy makers to exploit clustering effects to derive competitive advantage.
Almeida and Kogut (1999) performed a follow-up of the Jaffe study, also using patent citations applied to design of semiconductor devices, to understand the localization of knowledge. The study found evidence for knowledge localization within Silicon Valley and to a lesser extent other regions. Furthermore they studied the inter-firm mobility of the engineers who were named upon the patents. Importantly, the study found that ideas were spread by inter-firm mobility of patent holders and there was evidence for regional labour networks within the successful region of Silicon Valley (Almeida and Kogut, 1999).
Zucker, et al. (1994) also studied the effect of human capital upon innovation in the biotechnology sector. The authors find a correlation between scientific discoveries and the establishment of biotechnology firms. Their findings place a particular emphasis upon academic “star scientists” at universities who are proposed to be the scarce resource required to drive innovation in a particular region.
The work of Zucker et al. places an emphasis upon key individual scientists and engineers, whereas the analyses of patent data includes a slightly larger but still well-defined group of (technical) professionals. However, there is another school of thought upon the link between human capital and knowledge localization that looks at bulk or average human capital in a given labour market (reviewed by Feldman, 1999, pg. 12-13). For example, Lucas’ landmark 1988 study on economic development suggests a link between innovation and average levels of education and ability of the
6
labour force. A possible interpretation is that “star” individuals are required to establish a firm that takes advantage of a new idea or technology, but a broader set of professional is required for the firm to grow and prosper (see Feldman, 1999, pg. 13).
Despite the clear existence of an externality that includes knowledge diffusion, clearly demonstrated geographically localized clusters and evidence that the clusters provide competitive advantage, the field is not sufficiently developed that either economists or policy makers can use the theories in a predictive manner. For example, international policy makers would like to recreate the success of Silicon Valley, but the theory does not yet provide sufficient guidance of how to achieve this. Van Geenhuizen and Reyes-Gonzole (2007, pg. 1) are particularly critical of policy makers who try to utilize the existing academic theories, stating that “no other concept in regional economic and policy seems to be so often used and so poorly understood on the basis of systematic research than the cluster concept”. They highlight that Michael Porter’s theories are only one of several that are used to rationalize the basis for competitive advantage from clusters; others include small-firm centric theories that build upon Schumpeter’s theories of innovation, or the “Triple Helix Model”, that emphasises relationships between universities, industry and governments (Etzkovitz and Leydesdorff, 2000). One reason why the cluster concept remains poorly defined and understood is a lack of empirical studies that accurately define the size of a cluster and the factors that contribute to there being a cluster at all.
This thesis builds upon the theories of knowledge diffusion being driven by inter-firm mobility of knowledge workers. The geographical mobility of employees is studied both within a successful Dutch science park, Leiden Bioscience Park, and nationally. The social networking site LinkedIn is used to search for employees who have moved either between firms in the park, or between firms outside the park and those within it. Hypothesis-based methods are used to study whether a science park can constitute a cluster of inter-firm human capital, and what factors contribute to inter-firm movement of employees.
7
The study shows that Leiden Bioscience Park does not constitute a defined cluster; however employees do show a propensity to move between firms of different sizes and between end-product and service firms. The thesis addresses the vital question: what exactly is a cluster? The findings are important to the field from a methodological perspective, since they demonstrate an analysis of bulk employees, without the restraint of analysing only highly visible “star” employees. Finally, the findings on knowledge diffusion and clusters may be helpful to policy makers, who wish to use national policy instruments to promote innovation.
8
II. Research Question
The primary research question aims to use empirical methods to address whether the largest bioscience park in the Netherlands forms a distinct cluster in terms of inter-firm movement of human capital: are the networks of movement of human capital between firms within the bioscience park stronger than such networks between bioscience firms nationally.
It would be expected that strong clustering effects are evident in Science Parks: national policy instruments have been used to form such parks because it is believed that the firms within them will benefit from clustering effects via strong links and collaborative projects between the firms and other organixations. As a result, transfer of employees between the firms within the park is expected to be amplified. There may also be more movement of employees within the park for purely geographical reasons, since employees may commonly look locally for new opportunities.
Hypothesis H1 (primary hypothesis): Employees are more likely to move from one firm in Leiden Bioscience Park to another firm in the park, than between firms in the park and outside it.
Secondary research questions address factors that can contribute to such clusters. Variables including firm size, whether the firms produce products or services and if they primarily deal in tangibles or intangibles are included in the analysis.
Overall, the link between retention or employee turnover and firm size has not been fully addressed (Cardon, 2004). The academic literature has shown, however, that job-seekers have strong preferences with regards to firm size, with a group of prospective employees preferring larger firms and another distinct group preferring small firms (Barber, 1999). Building upon these data, one would expect employees to transfer between firms of a similar size, hence providing a rationale for the hypothesis that employees will move between firms of different sizes less than would be expected by chance.
Hypothesis H2: Movement of employees between firms of different sizes is non-random.
9
A bioscience park has been selected as the setting for this study since the academic literature has shown that firms within the same industry are more likely to benefit from clustering effects (Porter, 1998), and this would be expected to be applicable to the movement of human capital. However, other factors that give rise to clusters are not well understood. Here, the business models of the firms are assessed to understand if they contribute to clusters of exchange of human capital. Specifically, the study includes firms that are involved in producing end pharmaceutical products or intermediate products and services, and firms that are involved in producing tangible goods or those that deal purely in intangibles.
Hypothesis H3a: Movement of employees between end-product orientated firms or service orientated firms is different than would be expected by chance. Hypothesis H3b: Movement of employees between firms dealing in tangibles or intangibles is different than would be expected by chance.
10
III. Research Setting
Life Sciences industries are defined as comprising medical technology, medical biotechnology and pharmaceuticals (see Life Sciences in the UK, 2010). According to the IMAP Pharma Report (2011), the global pharmaceutical market is expected to pass $1 trillion in the near future, with double digit growth rates predicted. Ernst & Young report that the total market capitalization of European publically-listed Biotechnology firms was $72 billion in 2011, with revenues of $18.9 billion. Of public firms listed in the Netherlands, the total market capitalization was $3.3 billion and the revenues were $1.1 billion (Ernst & Young - Beyond borders, 2011; 2012). Global private equity investment in healthcare industries was estimated to be between $30 and $44 billion in 2011 (Bain & Company, 2012; Yang, 2011). Europe represents about 30% of publically listed biotechnology firms, second only to the US (Huggett and Lähteenmaki, 2012). According to a Battelle/BIO State Bioscience Industry Development report (2012), the Life Sciences sector grew significantly over the past decade, outperforming most other knowledge based industries in terms of job creation.
Not surprisingly, development of Life Sciences industries are priorities for many countries including the US (National Bioeconomy Blueprint. April, 2012. The White House, Washington; Waltz, 2012) and the UK (Life Sciences in the UK, 2010). For example, the Cameron government introduced an “innovation booster” in the UK in 2012, including tax breaks, a $280 million fund to help biotechnology SMEs (Moran, 2012) and $380 million in new funds for fundamental research (Nasto, 2012).
Large pharmaceutical companies are a dominant and important force for the Life Sciences industry. In 2011, the biggest pharmaceutical companies were Johnson and Johnson, Pfizer and Novartis (Arrowsmith, 2012). Such companies have considerable expertise in down-stream processes such as manufacturing, marketing and regulatory affairs. In practice they increasingly collaborate with biotechnology firms who often have expertise in specialised technologies (Enzing et al, 2004).
Recently the pharmaceutical industry has faced a number of problems, including a recent “patent cliff”, where many blockbuster drugs lost their patent protection, declining productivity, increasing globalization (IMAP pharma report,
11
2011; Bhidé, 2009) and a more uncertain regulatory climate (Pammolli and Riccaboni, 2004). Scannell et al. (2012) suggest several problems, including cautious regulators, increasing competition of off-patent drugs with new ones and tendencies for companies to overinvest in speculative R&D. The uncertainty in the industry had been accompanied by a surge in the number of mergers and acquisitions (M&A; The Economist, 2007), with M&A deals worth over $200 billion in 2009 and 2010 alone. These pressures are leading pharmaceutical companies approach their R&D with an increasingly external focus (Barrett et al. 2009; O’ Hagen and Farkas, 2009).
Some of the particular challenges that pharmaceutical firms face stem from the role of innovation and new technologies in the industry. Biotechnology firms provide increasingly important solutions to these problems in research and innovation. Coriat et al. (2003) consider biotechnology to be a new model of a science-based firm because of the way in which biotechnology firms conduct science, rather than just use it. Interestingly, biotechnology firms do not need to actually produce their own products to be considered successful; many are loss-making but function through alliances with large pharmaceutical firms. Such alliances have allowed pharmaceutical companies to develop new technologies, but the alliances remain important as new technologies evolve (Galambos and Sturchio, 1998). It is estimated that most innovative drugs brought to market in the next 5 years will involve such partnerships (Berggren et al. 2012). Recent growth in the biotechnology sector has been driven by recombinant vaccines, monoclonal antibodies and hormones (Aggerwal, 2011) and Europe and the US remain dominant forces in such innovation (Friedmann; 2010).
McNamee and Ledley (2012) analyse some recent patterns of innovations in the Life Sciences in terms of the product life cycle literature and the concept of “disruptive innovations”, introduced by Clayton Christensen et al. (2001). This study concludes that the links biotechnological products have with both innovative basic science (which is aligned with early stages of a product life cycle) and the stringent regulatory demands of healthcare products (developed from mature technologies) presents its own challenges. The study found evidence of disruptive innovations, such as in the development of humanized monoclonal antibodies used as
12
therapeutics to treat a range of diseases. However, critics of the “biotech revolution’ suggest that knowledge diffusion remains incremental and slow (Nightingale and Martin, 2004).
Khilji, et al (2006) apply the innovation literature to the Biotechnology sector and conclude that too much emphasis is placed on technological “push”, with too little appreciation from market mechanisms and for the complexities of the different development phases of healthcare produces. A recent Avance study (2009) into success rates of biotechnology products in clinical trials, finds a figure of approximately 20% for biotechnology firms, well above the 10% figure for pharmaceutical firms. However, many more products fail at a late stage for biotechnology firms: an indication that failing products are often pursued for too long.
The Life Sciences Industry in the Netherlands
According to the Dutch life sciences outlook (Life Sciences Health – The Dutch Life Sciences & Health Sector, 2011), the Dutch Life Sciences sector has approximately 314 companies employing 24,500 people, with total revenues of €18.7 billion from 104 products. The public investment is €250 million, with €620 million of private investment. This represents major growth from 2001 when there were 15,000 people in 100 firms, and from 1994 when there were only 18 firms (Enzing et al, 2004).
Much of this rapid growth was the result of Dutch governmental policies. In 1998 the Dutch Ministry of Economic affairs noted that several critical success factors were missing from the Netherlands, including venture financing, the presence of incubators and university policies. This resulted in the investment of hundreds millions euros in Biotechnology research, from schemes such as the “Life Sciences Action Plan”, which aimed to use a budget of €45 million to establish 75 Biotechnology start-up firms (2000-2004), or the Innovation Orientated Research Program (IOP) for industry-orientated academic research, with a budget of €20 million (2000-2004). Three programs were also established in this period to promote
13
the commercialization of biotechnology: BioPartner, which administered the Life Sciences Action Plan, Mibiton and STIGON (Enzing et al, 2004).
Mergers and acquisitions have also profoundly affected the landscape of the Netherlands Life Sciences Industries, particularly with respect to pharmaceutical giants. DSM acquired the Dutch antibiotics supplier Gist-Brocades in 1998, and US-based Catalytica Pharmaceuticals in 2000. Organon acquired Kanebo, a Japanese pharmaceutical company, in 1999 and Covance Biotechnology Services in 2001, but sold its diagnostics division Organon Teknika to the BioMerieux in the same year (Enzing et al, 2004). Akzo Nobel proposed to demerge Organon Biosciences in 2007, but within days of the completion date it was instead sold to Schering-Plough for $14.5 billion (http://news.bbc.co.uk/2/hi/business/6441423.stm), who were later taken over by US pharmaceutical giant Merck, Sharp and Dohme (MSD). It is noteworthy that one of the leading Dutch Life-Sciences firms found itself on the periphery of an international giant within such a short time: this final episode highlights one of the risks of the increasingly internationalized pharmaceutical industry.
Leiden-based Crucell was taken over by its collaborator Johnson and Johnson in 2010 for $2.3 billion (Ernst & Young, 2011), but this deal was treated in a more positive light by analysts than the Organon deals (see for example, a report by NautaDutilh, 2012), possibly because Johnson and Johnson had a clearer motive of investing in the firm in the future. Johnson and Johnson also took over two further firms within Leiden Bioscience Park: Centricor and Jansen Bioscience, to form the Jansen Network of Excellence in the Netherlands.
Problems that have been identified as limiting the growth of the Netherlands Life-Sciences industry include a lack of human resources (Enzing et al, 2004; Haverman, 2009), with employees combining scientific and managerial skills a particular problem, and a lack of entrepreneurship, with academics reluctant to leave University positions to pursue applied research (Enzing et al, 2004). An Open Method of Coordination (OMC) report on the Netherlands notes that private expenditure on research and development (R&D) is only about 1% of GDP in the Netherlands (Eriksson, 2007), with a total R&D intensity of about 1.8%, compared with an EU average of 2% (Eurostat pocketbook, 2012). The average growth rate of
14
R&D expenditure (2005-2010) was among the lowest in Europe. A recent study by Scientific American (Scientific American - WorldView), placed the Netherlands in 7th place globally in terms of R&D intensity, but 17th place in terms of Enterprise support: the available collection of resources in a particular country. The relatively small presence of large pharmaceutical firms in the Netherlands is also a potential problem, since such firms may prefer to collaborate with biotechnology companies in close proximity to themselves, whereas biotech firms rely on pharmaceutical companies for their specialization in down-stream processes (Enzing et al, 2004). The Netherlands does benefit from strong academic research (see Peng, 2011 for a summary of biotechnology publications) but lower than average levels of venture capital funding and R&D expenditure (Rabo Innovation Index. 2011).
The Netherlands uses a whole range of incentives to promote business. Ernst & Young (2012) highlight national strengths in its infrastructure of communications and transport, which would be expected to benefit many industries. The Netherlands also provides a wide range of tax incentives for investment, collaboration and innovation, such as tax exemption via an innovation box”, which is likely to be of particular benefit to Biotechnology companies (Ernst & Young, 2012).
Van Geenhuizen (2009) is critical of governmental Life Sciences policies in the Netherlands, noting that the rapid growth in the number of firms left many firms small and vulnerable. The report also criticises the short time-frame of the Life Sciences Action Plan at only 4-5 years, and of a lack of national coordination. A new policy for Life Sciences, introduced in 2008, built on the original initiative and filled some of the investment gap for these small firms and promoted international cooperation (Van Geenhuizen, 2009). The ministry of economic affairs announced a new policy for all sectors including life sciences, in 2011. The policy aimed to support small and medium enterprises (SME) via a €500 million investment, to provide solutions for a lack of technical employees and to reduce taxation on R&D costs. It also reiterated an ambition to be a top 5 knowledge economy by 2020 (Life Sciences Health – The Dutch Life Sciences & Health Sector, 2011).
15
The science park model and Leiden Bioscience Park.
The first Science Parks were established in the 1970, but they have rapidly grown in number, such that there were 123 University Science Parks in the United States by the early part of the 21st century, with nearly 1000 science incubators in the US and Europe and over 200 in Asia. The rapid growth was driven, at least in part, by national governmental policies. However, academic studies into best practices are still in their infancy (Phan et al. 2005).
According to Felsenstein’s (1994), the aim of Science Parks is “nurturing the development and growth of new, small, high-tech firms, facilitating the transfer of university know-how to tenant companies, encouraging the development of faculty-based spin-offs and stimulating the development of innovative products and processes’’ and to generate regional economic development to promote economic growth. Similarly, Chan and Lau (2005) note that the International Association of Science Parks lists three criteria for science parks: links to a university or other higher education institution, a design to encourage the formation and growth of knowledge-based businesses and a management function which is actively engaged in the transfer of technology and business skills to the organisations on site.
One form of Science Park is concerned Life sciences, in so called Bioscience Parks. Internationally the largest bioscience parks are in Boston and San Francisco in the US, Munich, Germany and Cambridge, UK. The analysis of Lecocq (2012), which was conducted in terms of patents, places the Dutch province of South Holland as the 20th largest global biotechnology cluster, also as a result of governmental policies. In fact, precisely defining the size of biotechnology clusters in the Netherlands is not trivial since it depends upon definitions of firms’ characteristics and geography. Within the densely populated Randstad region of the Netherlands, which includes Amsterdam, Leiden, Rotterdam, Utrecht and Delft, there were 62 biotechnology firms in 2004, using a stringent definition. The highest concentration in this region was in the Leiden Bioscience Park, with 24 firms (Van Geenhuizen, 2009). The Leiden Bioscience Park is the oldest Bioscience Park in the Netherlands. It was founded in 1984, relatively late in the Science Park revolution. However it was rapidly successful, not least because of the entry and growth of a number of firms
16
and institutes including US multinational Centocor (now part of Johnson and Johnson) and Japanese pharmaceutical company Yamanouchi (now part of Astellas) as well as the sites of Leiden University and Leiden Academic Medical Centre among other academic institutes, and TI Pharma (van Geenhuizen, 2009).
By 2005, 2,689 people were working in private companies in Leiden Bioscience Park, approximately evenly split between “endogenous” companies that were set up with links to institutions already in the park, and “exogenous” firms, which were set up independently of institutions in the park. However, most of the growth in the number of companies is dominated by formation of new exogenous companies, whereas growth in the numbers of employees is dominated by relocations or the establishment of new divisions (Jousma, 2009). Van Geenhuizen and Reyes-Gonzales (2007) find higher levels of innovation in the park compared with other locations.
Leiden Bioscience Park is currently the largest Bioscience Cluster in the Netherlands and in the top five in Europe (Smailes, 2011). The Leiden University technology transfer office, LURIS commissioned a study by BiGGAR Economics, Midlithian, UK (2011), into the direct and indirect economic impacts of Leiden University. This study concluded that the Leiden University and Leiden Bioscience Park were responsible for €1.3 billion in gross added value to the Dutch economy, and 18,000 jobs, with 15,700 people working directly on the Park in 2010 and 128 products in development (Development of Leiden Bioscience Park, 2010; 2011).
This thesis applied clustering methodologies to Leiden Bioscience Park, to further understand if forms an identifiable cluster in terms of labour mobility, and what factors may contribute to it being a cluster.
17
IV. Data and Methods
For the study, the movement of employees is used as a proxy for the formation of a cluster in the Biotechnology sector.
1. Selection and characteristics of the firms in the study
The firms within Leiden Bioscience Park were identified via the Park’s website (http://www.leidenbiosciencepark.nl). They were assessed against the following selection criteria:
1. The firms should operate primarily from only a single site within the Netherlands. 2. Only a single firm from a corporation should be included.
3. The firms should display a mixture of sizes, such that small, medium and large firms are represented.
These criteria mean that several of the larger firms from the science were excluded. For example the Japanese Pharmaceutical corporation Astellas, which has a second site in Meppel, was excluded because of concern about ambiguity as to whether or not employees worked within the Science Park. Of the three firms in the Science Park that are part of the US Corporation Johnson and Johnson: Janssen Biotechnology, Centricon and Crucell, only Crucell was included since potentially high exchange of employees between these three firms could confound the analysis.
Eight firms were selected from Leiden Bioscience Park against the selection criteria. These include large firms such as Crucell, part of the Johnson and Johnson family of companies, medium sized biotechnology firms such as HAL Allergy and Octoplus and smaller biotechnology start-up firms such as to-BBB. The complete list is: Batavia Bioservices, BioFocus (a services division of Galapagos), Pharming, Xendo, OctoPlus, HAL Allergy, Crucell and to-BBB (Table I).
Next, the firms outside the Leiden Bioscience Park were selected against the selection criteria for comparison. Ten firms from outside the park were selected in total. These include larger biotechnology firms such as the TEVA Pharmachemie firm from the TEVA Corporation, medium firms such as Genmab and Synthon, and
18
smaller academic spin-off biotechnology firms such as Pepscan and Agendia. The complete list of included firms outside Leiden Bioscience Park is: Progress-PME, Kiadis Pharma, Pepscan, Agendia Inc., Bilthoven Biologicals, Genmab, TEVA Pharmachemie, Synthon, Yxion, Synco Biopartners.
The company LinkedIn pages of these firms were used to obtain information about their location, size and the types of products and services that they provide (Table I).
Company Location Size Product/Service Tangible/Intangible Batavia Bioservices LBSP <200 Service Tangible Galapagos Biofocus LBSP <200 Service Tangible
Pharming LBSP <200 Product Tangible
Xendo LBSP <200 Service Intangible
Octoplus LBSP <200 Service Tangible
HAL Allergy LBSP <200 Product Tangible
Crucell LBSP >200 Product Tangible
To-BBB LBSP <200 Product Tangible
Progress-PME Hoofdorp <200 Service Intangible Kiadis Pharma Amsterdam <200 Product Tangible
Pepscan Lelystad <200 Service Tangible
Genmab Utrecht >200 Product Tangible
TEVA Haarlem >200 Product Tangible
Synthon Nijmegen >200 Product Tangible
Yxion Amsterdam <200 Service Intangible
Agendia Amsterdam <200 Product Tangible
Bilthoven Bio. Bilthoven <200 Service Tangible Synco
Biopartners
Amsterdam <200 Service Tangible
19
2. Identification of employees moving between the firms
The movement of people between these firms was assessed using public profiles from the social networking site LinkedIn. The website was searched in the first week of April, 2013, against the names of the 18 firms for former employees of one of these firms that transferred to another of the firms. These hits were manually searched to remove duplicates and false positives. The most common source of such false positives was consultants who had worked for more than one of the firms.
To F rom B ata v ia B ioF oc us P ha rm in g X en do O c toP lus HA L A llergy Cr uc el l to -B B B P rog res s -P M E K ia di s P ha rma P ep s c an A ge nd ia I nc . B iltho v e n G en ma b T E V A S y nth o n Y x ion S y nc o Batavia – 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 BioFocus 0 – 0 0 1 0 2 0 0 0 0 0 0 0 1 0 0 0 Pharming 4 0 – 2 5 2 9 2 0 4 0 0 0 1 3 2 0 2 Xendo 8 0 2 – 5 1 10 0 0 1 0 0 0 1 3 2 3 0 OctoPlus 0 1 1 3 – 9 21 0 2 0 0 0 0 2 9 2 0 1 HAL 0 0 1 2 1 – 0 0 2 0 0 0 0 0 1 0 0 0 Crucell 9 3 5 4 11 2 – 1 2 3 0 1 1 15 9 1 3 5 to-BBB 0 0 0 0 1 0 1 – 0 0 0 0 0 0 1 0 0 0 P-PME 0 0 0 1 0 0 1 0 – 0 0 0 0 0 1 0 0 0 Kiadis 0 0 1 0 4 0 3 0 1 – 0 0 0 0 1 1 0 0 Pepscan 0 0 0 0 0 0 2 0 0 0 – 0 1 0 0 0 0 0 Agendia 0 0 0 0 0 0 0 0 0 0 0 – 0 0 0 0 0 0 Bilthoven 0 0 0 0 0 0 0 0 0 0 0 0 – 0 0 0 0 0 Genmab 0 0 0 0 2 1 5 1 0 0 1 0 1 – 1 3 0 0 TEVA 0 1 0 2 2 1 5 0 2 0 0 0 0 0 – 0 1 5 Synthon 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 – 1 0 Yxion 0 0 0 1 0 1 0 0 0 0 0 0 0 0 0 0 – 0 Synco 4 0 0 0 2 1 7 0 4 0 0 1 0 1 4 0 0 –
Table II. Matrix summarizing the job changes between the 18 firms included in the study.
20
The results are directional, such that employees who moved from Firm A to Firm B are listed separately from those who moved from Frim B to Firm A. In total, 294 job changes between the 18 firms were found. This was considered to be sufficient to provide statistical power to the study. The results are summarized in Table II.
3. Statistical tests
The directional nature of the data facilitates the testing of the hypotheses using a test for categorical variables. For each hypothesis, the LinkedIn data matrix in Table II is reduced into categories that allow the hypotheses to be tested statistically by means of a Chi-squared test. The data were analysed in Excel and the observed data, the expected data and the p value are quoted.
21
V. Statistical hypothesis testing
Hypothesis H1 (primary hypothesis): Employees are more likely to move from one firm in Leiden Bioscience Park to another firm in the park, than between firms in the park and outside it.
Observed: Expected:
To To
LBSP Other Total LBSP Other Total
Fr o m LBSP 130 83 213 Fr o m LBSP 131 82 213 Other 51 30 81 Other 50 31 81 Total 181 113 294 Total 181 113 294
Significance test: p=0.76. Statistically not significant.
The null hypothesis, that movement of employees is unaffected by whether the firms are inside or outside Leiden Bioscience Park, cannot be rejected.
Hypothesis H2: Movement of employees between firms of different sizes is non-random.
Observed: Expected:
To To
Large Small Total Large Small Total
Fr o m Large 39 74 113 Fr o m Large 51 62 113 Small 93 88 181 Small 81 100 181 Total 132 162 294 Total 132 162 294
22
Significance test: p=0.0047. Statistically significant.
The null hypothesis, that movement of employees is unaffected by firm size is rejected.
Upon closer examination, it is evident that in our sample there is more exchange of human resources between small and large firms, and this comes at the expense of exchange between firms in the same size category.
Hypothesis H3a: Movement of employees between end-product orientated firms or service-orientated firms is different than would be expected by chance.
Observed: Expected:
To To
Prod Serv Total Prod Serv Total
Fr o m Prod 87 83 170 Fr o m Prod 101 69 170 Serv 87 37 124 Serv 73 51 124 Total 174 120 294 Total 174 120 294
Prod = firms whose business model is to produce their own end pharmaceutical products. Serv = firms that primarily provide services to other firms.
Significance test: p=0.0011. Statistically significant.
The null hypothesis, that movement of employees is unaffected by whether the firm produces its own end-products, or provides services to other firms, can be rejected.
Upon closer examination, there is less movement of people between firms that produce end-products, or between firms that provide services, than would be expected. Consequently, there is a compensatory increase in movement between these groups.
23
Hypothesis H3b: Movement of employees between firms dealing in tangibles or intangibles is different than would be expected by chance.
Observed: Expected:
To To
Tang Intan Total Tang Intan Total
Fr o m Tang 219 34 253 Fr o m Tang 219 34 253 Intan 36 5 41 Intan 36 5 41 Total 255 39 294 Total 255 39 294
Tang = firms that primarily deal in tangible products. Intan = firms that primarily deal in intangibles.
Significance test: p=0.83
The null hypothesis, that movement of employees between companies that deal in tangible products or intangibles cannot be rejected. However, the number of employees moving between firms that deal in intangibles is small, which may make this result unreliable.
24
VI. Discussion
The aim of this study is to contribute to the academic literature that has studied potential benefits of geographical clustering of related industries, and specifically to identify factors that can contribute to the formation of a cluster. Rather than using established methodologies that study “star” scientists, patent holders or average human capital levels, this study used a more novel approach applied to inter-firm movement of employees of Dutch lifsciences firms using data from the e-networking platform LinkedIn.
The primary hypothesis was that movement of employees between the firms studied would show evidence that Leiden Bioscience Park formed a distinct cluster, since movement of employees within the park or to/from the park was statistically different than would have been expected by chance. There was no statistical evidence for this hypothesis based upon these data; indeed not even a trend towards significance was observed.
The rationale for the primary hypothesis was that if the clustering and network effects within the park are strong, one would expect strong links and collaborative projects between the firms. Consequently, employees would be expected to be more likely to move between the firms within the park. Even in the absence of a formal network, one may expect more movement of employees within the park for purely geographical reasons. As employees ran out of opportunities in one firm, they may have been expected to look to a neighbouring firm for new opportunities (see, for example Porter, 1998). However, this was not supported by empirical data.
This lack of evidence for clustering effects within Leiden Bioscience Park using movement of human capital as a data source is consistent with several other critical studies of science parks, as reviewed by Dabrowska (2011). These studies use a number of different data sources, but show little evidence for performance-enhancing clustering effects.
There were reasons to think that studying Leiden Bioscience Park would show evidence for a cluster, even though other studies on science parks have not. The park is specialized towards biosciences, unlike other Dutch Science Parks. One may
25
have envisaged that this specialization could serve to promote clustering effects, since the firms share a common industry. Consistent with this, Dijkstra (2010), found evidence for a local network in LBSP in terms of human capital, whereas van Geenhuizen and Reyes-Gonzales (2007) found the park to show higher levels of innovation compared with other locations. However, evidence building on these data was not provided by this study. Further empirical studies are required to explain the factors that contribute to the formation and growth of a cluster and science parks such as Leiden Bioscience Park may be appealing settings for such studies.
Van Geenhuizen (2009) found that smaller biotechnology firms tended to have local/national networks, whereas larger ones were more globally focussed. However, since the more established firms were larger, approximately 92% of knowledge workers were in global rather than local networks according to this study. It would be interesting to perform a similar study in terms of movement of human resources. Overall, one may assume that individuals would be less willing to change countries for a new job than to form an international collaborative project in this age of cheap and easy international communication. Consequently, one may expect networks of movement of human capital to be more local than those of collaborative projects. Further studies could be used to understand if more established firms have more international networks of movement of human capital. It is interesting that several of the biotechnology start-up firms in Leiden Bioscience Park established additional sites, such as Octoplus (Van Geenhuizen, 2009), which may be consistent with growth leading to a more international focus in human capital movement.
One can question whether the recruitment and movement of employees is a valid metric to study clustering effects. Bioscience firms clearly require in-depth knowledge and expertise of their employees, so human resources are almost certainly important. Almeida and Kogut, (1999) analysed inter-firm movements of engineers from patent data to show that knowledge was localized in Silicon Valley, California, and that this affected knowledge transfer. Since the firms in this study are likely to rely on transfer of tacit knowledge to compete, it is likely that movements of human capital are of importance for knowledge transfer.
26
Secondary hypotheses explored how firms’ size and whether they deal in end-products or services and tangibles or intangibles affected movements of employees. These secondary hypotheses in particular should be treated as preliminary and should be validated with a data set that contains more firms and data points. The limitations are discussed in detail below. Unlike the primary hypothesis, these hypotheses did not benefit from a control group so it is difficult to exclude that people’s LinkedIn behaviour differed between the groups. Nonetheless, two of the secondary hypotheses were statistically significant, and hence further research to validate the findings may be of interest.
The employees in this study not only were willing to move between firms of different size, but seemed to move between them more than would be expected by chance. Almeide and Kogut (1999) describe the engineers in their study having more of an association with their craft than with the firms in which they work. It may be that transferring between firms of different sizes is part of career progression in the Biotechnology sector. If such a finding could be validated then it would be of interest to policy makers and economists alike: employees actively exchanging between firms of different size could help to promote the growth of small firms as they move from start-up to a more mature phase, or alternatively, could promote new ideas and innovation in larger firms.
Although the effect of firm size on the transfer of human capital is not well understood (Cardon, 2004), job-seekers do have preferences with regards to firm size, with a some prospective employees preferring larger firms and others preferring small firms (Barber, 1999). Conversely, however, the employees in this study were happy to move between firms of different sizes.
Williamson (2002) notes that smaller firms face recruitment problems relative to larger firms. This is analysed in terms of organizational knowledge and organizational legitimacy (Williamson, 2002). Our data suggest that employees are happy to move between firms regardless of size, and may even have a preference to move between large and small firms. It is possible that the cluster effects in play in the Netherlands negate these problems with knowledge and legitimacy of a firm,
27
since employees within the life sciences sector will be highly aware of the other local firms in the same industry.
Furthermore, the putative finding that employees are highly comfortable moving between end-product and service firms suggest that Life Sciences clusters should be designed so that they contain both types of firm, as well as firms of different sizes. Such an exchange of people between different firms, if it could be properly utilized, surely supports the idea of clusters in Life Sciences where firms can benefit from geographical proximity (see Porter, 1998).
Further studies using human capital as a metric with which to study clustering effects will be of interest to further understand what drives the formation and growth of successful clusters. “Big data” sources such as LinkedIn facilitate such studies and may be applied to lending empirical support to explain imperfections in knowledge transfer and successful application of clustering theories.
28
VII. Limitations
The advantage of the LinkedIn analysis is that it provides a large amount of quantitative information that can be analysed using statistical methods. Indeed, of several ideas for desk-research for this thesis, including analysis of patents, publications and collaborations for evidence of collaborations, LinkedIn was the only data source that was explored that provided the required number of data points. However, the approach suffers from a number of limitations.
Firstly, use of the social networking site potentially introduces a number of biases, since the data are incomplete. LinkedIn has a penetration of about 50% in the Netherlands. It is possible that the individuals who join the site, or the way in which they share the data reduces the validity of the study, or introduces biases. However the high penetration reduces that chance of this compared with, for example, a survey with a low response rate.
The second challenge with LinkedIn is the accuracy of the information that is contained within it. For example, if an individual chose to fabricate an employer, it is unlikely that it would have been detected in the validation stage of the study. Historically, however, many online resources are relatively free of errors (O’Reilly, 2007). LinkedIn does not benefit from crowd-sourcing, but the public nature of the information may serve to increase the validity of the data.
There are also potential biases introduced by the choice of companies that were included in the study. The exchange of human resources will always be influenced by strategic decisions of a company. For example, collaborations between firms or decisions to invest in or close a business unit will inevitably have a large effect on any findings. Desk research was performed to try to rationalize such trends, but any findings would have to be retested in another independent study in order to be considered reliable. For example, Xendo announced a collaboration with Batavia Biosciences in 2011 (Bloomberg business week, 26th August 2011), whereas Crucell was bought out by Johnson and Johnson in 2011, following a strategic alliance. Both of these will have had a large effect on the exchange of personnel.
29
Furthermore, the LinkedIn derived data design does not sufficiently take the histories of the firms into account. For example, HAL Allergy moved from Haarlem to Leiden Bioscience Park in 2009; but the timing of the move was not taken into account in the data analysis section of the study. Finally, the definition of firms as, for example, either end-product or service orientated may be a little narrow since in practice many of the firms in this study are involved in both activities. The LinkedIn profiles and websites of the companies themselves were used to categorize the companies, and the two findings that were significant seems to be robust with respect to the less certain data, however, further validation of the findings is clearly required.
30
VIII. Summary, conclusions and recommendations
A. Summary
Academic studies have shown that firms in related industries are often geographically clustered, and that such clusters can provide competitive advantages in terms of more efficient knowledge transfer. This study used the information from the social networking site LinkedIn to explore clustering and networks effects in life sciences firms in the Netherlands, with a focus on Leiden Bioscience Park. The exchange of human capital between eight firms within the park, or a further ten firms outside the park was analysed, and the data was tested against several hypotheses: whether Leiden Bioscience Park forms a distinct cluster in terms of exchange of human capital, and whether transfer between firms of different size, those dealing in tangibles and intangibles or end-products and services are statistically different that would be expected by chance.
The primary finding of this study is a negative one: based on a study of 294 job changes from the social networking site LinkedIn, people were no more likely to move between firms in the Leiden Bioscience Park than would be expected by chance. This is consistent with the Science park literature, which provides little evidence for performance-enhancing clustering effects within Science Parks. Two secondary findings were that movement between firms of differing size was greater than would have been expected, whereas movement between end-product and service based firms was also statistically significant.
B. Conclusions
From a methodological standpoint, this study is interesting for two reasons. Firstly, analysing cluster formation directly in terms of exchange of human capital is an interesting deviation from studies that follow citation data, high profile individuals or averaged metrics within a geographically-defined population. Consequently, the approaches may offer insight over and above existing studies. Secondly, since the data used in the study is in the public domain, the approach is scalable. A study with more extensive resources than this one could analyse a very large number of
31
employees on either a national or international level to understand on what levels clusters can be identified, and what factors lead to the formation of such clusters.
This study did not show Leiden Bioscience Park to be a distinct cluster in terms of movement of human capital. It may be that clusters of human capital among Dutch life-sciences professionals cover a larger geographical area than a single bioscience park, but such a hypothesis would require further validation. The study did show that Life Sciences personnel seem to be particularly willing to move between firms of different sized and between end-product and service orientated firms: something that could assist future policy-making. Further studies, however, are required to validate these findings.
C. Recommendations
The conclusions of this study lead to recommendations for further academic studies, and recommendations for managers and policy makers interested in promoting life-sciences in the Netherlands. The use of LinkedIn to understand network and clustering effects was an interesting approach because of the large amount of data that it can provide. Such ‘big data’ sources are set to become more important in the future in many aspects of business life. It would be interesting to expand this study to include far larger networks and clusters, since this could fill gaps in the academic literature related to clusters, and explore what networks in the Life Sciences really look like on a national or international level. Specifically, one could envisage that the actual clusters of exchange of human capital cover the entirety of the Randstad, or beyond.
The suggestion that employees look to move between firms of different sizes and between end-product and service firms supports the idea that clustering effects can be beneficial in the Life Sciences industries, since such exchanges could help firm to gain the knowledge that is required to grow and prosper. This suggests that further exploration of such potentially beneficial effects from Science Parks and Life Sciences clusters would be worthwhile and it would be interesting to perform additional empirical studies to support these findings.
32
This and other studies are yet to show clear performance-enhancing effects of science parks; this study did not find evidence for Leiden bioscience Park forming a distinct network or cluster in terms of movement of human capital. This is despite the close geographical proximity of the firms in Leiden that may facilitate movement of jobs between local firms, the potential for more collaborative projects within the park, and a designated recruitment website for Leiden Bioscience Park. Historically, six biotechnology clusters have been supported by national policy instruments in the Netherlands; five of which are in the close geographical proximity of the Randstad in Leiden, Utrecht, Rotterdam, Delft and Amsterdam. Such a division centred each cluster upon city-bound universities, hospitals and research institutes. That this study did not find evidence for a science-park centred cluster, consistent with the literature, suggests that this division is artificial since the actual Bioscience networks and clusters are larger than being confined to a single city. Policy makers may be advised to apply policy instruments to a region that encompasses a larger geographical region including several science parks if their intervention is to have maximum effect.
33
References
Aggerwal, S. (2011). What’s fueling the biotech engine—2010 to 2011. Nature biotechnology 29:1083-1089.
Almeida, P and Kogut, B (1999) Localization of knowledge and the mobility of engineers in regional networks. Management Science 45:905-17.
Arrowsmith, J. (2012). A decade of change. Nature Reviews Drug Discovery 11:17-18.
Avance biotech success rates (2009). 7 March. Available from: www.avance.ch/newsletter docs avance biotech success rates.pdf Bain global healthcare private equity report (2012). Available from:
http://www.bain.com/publications/articles/global-healthcare-private-equity-report-2012.aspx
Balconi, M. et al. (2004). Networks of Inventors and the role of academia: an exploration of Italian patent data. Research Policy 33:127-145.
Baptista and Swann (1998). Do firms in clusters innovate more? Research Policy 27:525–540
Barber, A.E., et. al., (1999). A tale of two job markets: organizational size and its effects on hiring practices and job search behavior. Personal Phycology. 52:841-68. Barrett, C.W., et al. (2009). Upgrading R&D in a downturn. The McKinsey Quarterly, February. 1-3.
BBC news (2007). Akzo Nobel in 11bn-euro sell-off. Available from http://news.bbc.co.uk/2/hi/business/6441423.stm
Berggren, R. et al. (2012) Outlook for the next 5 years in drug innovation. Nature Reviews Drug Discovery. 11:435-6
Bhidé, A. (2009). Where innovation creates value. The McKinsey Quarterly, February. 1-6.
BiGGAR Economics (2011). Economic Impact of Research & Commercialisation at Leiden University & Leiden University Medical Centre. A report to Leiden University Research and Innovation Services. Available from:
http://media.leidenuniv.nl/legacy/biggar-report-aug-2011.pdf
Biotechnology Industry Organization (BIO) press release, June 19, 2012. Available from www.bio.org.
Bloomberg business week. Batavia Bioservices and Xendo BV Announce Business Collaboration. 26th August 2011. Available from:
http://investing.businessweek.com/research/stocks/private/snapshot.asp?privcapid=5 2247300
Brass, D.J. et al. (2004). Taking stock of networks and organizations: a multilevel perspective. Academy of management journal 47:795-817.
Burt, R.S. (1992). Structural Holes: The Social Structure of Competition. Harvard University Press: Cambridge, MA.
34
Cardon, M.S. and Stevens, C.E. (2004) Managing human resources in small organizations: What do we know? Human Resource Management Review 14:295– 323
Chan and Lau (2005). Assessing technology incubator programs in the science park: the good, the bad and the ugly. Technovation 25:1215-28.
Christensen, Clayton M., Michael Raynor and Matt Verlinden (2001), Skate to Where the Money Will Be, Harvard Business Review, (November):73-81.
Cooke, P. (2001). Regional innovation systems, clusters and the knowledge economy. Industrial and Corporate Change.10:945-74.
Coriat B., et al. (2003). Does biotech reflect a new science-based innovation regime. Industry and Innovation 10:231–253.
Dabrowska J. (2011). Measuring the success of science parks: performance monitoring and evaluation. Full Paper for the XXVIII IASP World Conference on Science and Technology Parks.
Development of Leiden Bioscience Park in 2010. Leiden Bioscience Park foundation. Available from:
http://www.leidenbiosciencepark.nl/uploads/downloads/lbsp_jaarverslag_2010.pdf Development of Leiden Bioscience Park in 2011. Leiden Bioscience Park
foundation. Available from:
http://www.leidenbiosciencepark.nl/uploads/downloads/lbsp_jaarverslag_2011.pdf Dijkstra, J. (2010). Knowledge diffusion and labor mobility: A Leiden Bioscience Park case study. Erasmus University Rotterdam. MSc Thesis. Economics & Business. Enzing, C., et al (2004). OECD Case Study on Innovation: The Dutch Pharmaceutical and Food Biotechnology Innovation Systems. TNO report, April, 2004.
Eriksson, P., et al. OMC Policy Mix Review Report - Country Report for The Netherlands April 2007.
Ernst & Young (2011). Beyond borders. Global Biotechnology Report 2011. Available from: http://www.ey.com/GL/en/Industries/Life-Sciences
Ernst & Young (2012). Beyond borders. Global Biotechnology Report 2012. Available from: http://www.ey.com/GL/en/Industries/Life-Sciences
Etzkovitz, H. and Leydesdorff, L. (2000). The dynamics of innovation; from national systems and ‘Mode 2’ to a triple helix of university-government-industry networks. Sci. Public Policy 29:115-28.
Eurostat pocketbook. Science, technology and innovation in Europe. 2012 edition. Feldman, M.P. (1999): The New Economics of Innovation, Spillovers And
Agglomeration: A review Of Empirical Studies, Economics of Innovation and New Technology, 8:5-25
Felsenstein, D. (1994). University-related Science Parks – “seedbeds” or “enclaves” of innovation. Technovation 14:93-110.
Friedmann, Y. (2010). Location of pharmaceutical innovation: 2000–2009. Nature reviews of Drug Discovery. 9:835-836.
35
Galambos and Sturdio (1998). Pharmaceutical firms and the transition to
biotechnology: a study in strategic innovation. Business history review 72:250-78. Haverman E. (2009). The Netherlands in the fast lane - KIA photograph. 27 February 2009. Available from:
http://www.enid-europe.org/winter/summer_school/Edo_Haveman_presentation.pdf
Huggett, B. and Lähteenmaki R. (2012). Public biotech 2011—the numbers. Nature Biotechnology 30:751-7.
IMAP. Pharma Report. 2010. Available from http://www.imap.com
Jaffe, A.B. et al, (1993). Geographic localization of knowledge spillovers as evidenced by patent citations. Quarterly Journal of Economics 108:577-598. Jousma, H., et al. (2009). Framework for Analyzing the Growth of University Research Parks Applied to the Bioscience Park in Leiden, the Netherlands. Paper presented at the 7th Triple Helix Conference in Glasgow, UK, on June 18, 2009 Khilji, S.E., et al. (2006) From Invention to Innovation: Toward Developing an Integrated Innovation Model for Biotech Firms. J. Prod. Innov. Manag. 23:528–540 Koka, B.R. and Prescott, J.E. (2008). Designing alliance networks: the influence of network position, environmental change and strategy on firm performance. Strategic management journal 29:639-61.
Lecocq, C., Leten B., Kusters J. and van Looy B. (2012). Do Firms Benefit from being Present in Multiple Technology Clusters? An Assessment of the Technological
Performance of Biopharmaceutical Firms. Regional studies 45:1107-1119.
Life Sciences Health – The Dutch Life Sciences & Health Sector (2011). Available from www.lifescienceshealth.com
Life Sciences in the UK – Economic analysis and evidence for ‘Life Sciences 2010: Delivering the Blueprint’. BIS Economics Paper, January, 2010.
Lorenzoni, G. and Baden-Fuller, C (1995). Creating a strategic center to manage a web of partners. California Management Review 37:146-63.
Lucas (1988). On the mechanisms of economic development. Journal of economic development. 22:3-42.
McNamee L.M., and Ledley, F.D. (2012). Patterns of technological innovation in Biotech. Nature Biotechnology 30:947-43.
Miles and Snow (1978). Organizational strategy, structure and process. New York: McGraw-Hill.
Moran, N. (2012). UK government unveils innovation booster. Nature Biotechnology 30:124-125.
Nasto, B. (2012). UK’s life sciences pitch. Nature Biotechnology 30:574. National Bioeconomy Blueprint. April, 2012. The White House, Washington. NautaDutilh (2012) Life Sciences Outlook 2012 – Dutch biotech companies from start-up to exit.
Nightingale, P. and Martin, P. (2004). The myth of the biotech revolution. TRENDS in Biotechnology 22:564-9.
O’ Hagen P and Farkas C (2009). Pharma companies can increase innovation—and improve results—through a new approach to R&D. Bain and Company. Available from www.bain.com
36
O’Reilly, T. (2007). What is Web 2.0: Design patterns and business models for the next generation of software, Communication and strategies, 65:17-37.
Paci and Usai (2000). Externalities, Knowledge Spillovers And The Spatial Distribution Of Innovation," ERSA conference papers ersa00p104, European Regional Science Association.
Pammolli F and Riccaboni M (2004). Market Structure And Drug Innovation. Health Affairs, 23:48-50.
Peng, W. (2011). Trends in biotech literature 2009–2010. Nature biotechnology 29:781.
Phan, P.H., Siegel, D.S. and Wright, M. (2005). Science parks and incubators: observations, synthesis and future research. Journal of Business Venturing 20:165-82.
Porter, M.E. (1998). Clusters and the new economics of competition. Harvard Business Review. November-December:77-90.
Rabo Innovation Index. How can the Netherlands boost its innovation capability? October 2011.
Saxenian, A., (1994). Regional Advantage: Culture and Competition in Silicon Valley and Route 128. Harvard University Press, Cambridge, MA.
Scannell, J.W., et al (2012) Diagnosing the decline in pharmaceutical R&D efficiency. Nature Reviews Drug Discovery 11:191-9
Scientific American. WorldView – A Global Biotechnology Perspective. Available from www.saworldview.com
Smailes, R (2011). Made in Leiden. EBR Bioregional. Available from:
http://www.leidenbiosciencepark.nl/uploads/nieuws_bijlagen/bestand_46.pdf The economist. Mergers and acquisitions. 11th January 2007.
Van Geenhuizen, M. (2009). Life Sciences in the Netherlands: Weakness in a Strong National Policy. European Planning Studies, 17:No. 1.
Van Geenhuizen, M and Reyes-Gonzales, L. (2007) Does a clustered location matter for high-technology companies’ performance? The case of biotechnology in the Netherlands. Technology forecasting and social change 74:1681-96.
Williamson, I.O., Cable D.M. and Aldrich, H.E. (2002). Smaller but not necessarily weaker: How small businesses can overcome barriers to recruitment. In Jerome A. Katz, Theresa M. Welbourne (ed.) Managing People in Entrepreneurial Organizations (Advances in Entrepreneurship, Firm Emergence and Growth, Volume 5), Emerald Group Publishing Limited, 83-106.
Yang, W. (2011). Biotech on pace for record year. Nature biotechnology 29:780. Zucker, Darby and Brewer (1994). Intellectual capital and the birth of US
