Master thesis MSc Entrepreneurship
Location, foundation,
agglomeration, knowledge creation
Does location and source of foundation matter?
A study on local knowledge spillover effects on
Dutch life sciences ventures
Author Sybren Beneder
Student No. 11428244 Universiteit van Amsterdam 2586807 Vrije Universiteit
Program MSc Entrepreneurship (Joint Program) Supervisor mw. dr. N. Ugur
Date 01-07-2017
Preface
I would like to thank all people that were involved in writing this thesis. These include fellow students, friends, and family that advised me while writing this thesis. In particular I would like to thank my supervisor, mw. dr. N. Ugur who guided me throughout the process.
I hereby state that the copyright rests with the author. The author is solely responsible for the content of the paper, including mistakes.
Abstract
Purpose – Location and source of foundation are seen as important indicators of growth for ventures. University spinoff and clustering is believed to have a positive effect on growth of ventures. This study gives insight in how these traits influence the performance of Dutch life sciences ventures. This sector is a mature industry that is primarily located in clusters throughout the country. This study provides a comparative analysis to what extent this influences spinoffs, ventures located in clusters, also it looks into the differences between three clusters: Amsterdam, Groningen, and Oss.
Method – By combining data from the Dutch life sciences database and the chamber of commerce, a dataset of 104 Dutch ventures is examined. This dataset includes cross-sectional data about: year of foundation, source of foundation, location, no. of employees, assets, and sector. The study data that is collected evaluates growth over a four-year period, from 2011 to 2015. Distinct groups of life sciences ventures are compared and a regression model was made what the effect of certain variables is on growth.
Findings – Results indicate that spinoffs benefit from higher growth of employees compared to independently founded ventures. There is a positive effect on growth of employees for spinoffs. For other groups that were investigated for this research no significant findings were discovered. Hence, groups that show no significant differences do show different characteristics, further research could help to explain this phenomenon.
Implications – Learning and collaboration is seen as an important means for growth, therefore this should be addressed appropriately. Universities are seen as an important driver of learning for these knowledge intensive ventures, increased collaboration is expected to benefit both universities and ventures.
Table of content
Preface ... 2 Abstract ... 3 Table of content ... 4 1 Introduction ... 6 1.1 Introduction ... 61.2 Practical and theoretical contributions ... 7
1.3 Problem statement and research question ... 8
1.4 Outline thesis ... 9
2 Theoretical framework ... 9
2.1 Life sciences ... 9
2.2 Local knowledge Spillover ... 10
2.3 Spinoffs ... 11
2.4 Spinoff parent learning mechanisms ... 12
2.5 Organizational clustering ... 13
2.6 Local knowledge spillover in Dutch life sciences ... 14
3 Methodology ... 16
3.1 Research approach ... 16
3.2 Sampling ... 17
3.3 Dependent variables ... 17
3.3.1 Firm growth: assets and employees ... 18
3.4 Independent variables ... 18 3.4.1 Source of foundation ... 18 3.4.2 Located in cluster ... 18 3.5 Control variables ... 18 3.5.1 Sector ... 19 3.5.2 Age ... 19 3.5.3 Location ... 19 3.6 Research method ... 19 3.6.1 Checking normality ... 19 3.6.2 Descriptive statistics ... 19 3.6.3 Comparing means ... 20
3.6.4 Linear regression ... 20
3.6.5 Cluster location dummies ... 20
4 Research findings ... 20
4.1 Descriptive statistics for spinoffs and independently founded ventures ... 21
4.2 Descriptive statistics for ventures located in a cluster and ventures not located in a cluster ... 22
4.3 Descriptive statistics for clusters Amsterdam, Groningen, and Oss ... 23
4.4 Independent sample t-test: spinoff and independent foundation ... 24
4.5 Independent samples t-test: ventures located in a cluster and ventures not located in a cluster ... 25
4.6 ANOVA, comparing clusters: Amsterdam, Groningen, and Oss ... 26
4.7 Regression analysis, growth of employees and assets ... 27
5 Discussion and conclusion ... 28
5.1 Theoretical and practical implications ... 29
5.2 Limitations and future research ... 30
5.3 Conclusion ... 31
1 Introduction
The first chapter will give insight in this study and the content of this thesis and points out the context of this study, important literature background, and life sciences industry developments. Furthermore, it introduces the research approach, the research question it aims to answer, and what the practical and theoretical contributions of this thesis are.
1.1 Introduction
Learning is an important means for growth of new and existing firms. The creation of inter-organizational relationships is an important contributor for knowledge creation. Geographical proximity is a facilitator in these relationships that facilitate learning (Sapienza et al., 2004; Zahra, 2002). The importance of local agglomeration for knowledge transfer as a source for entrepreneurial activity has received attention from both geographers and economists (Hervas-Oliver et al., 2017; Boschma and Wenting, 2007). Clustering oftentimes lowers barriers for exchange of tacit knowledge and facilitates networking.
Not only do ventures cluster to facilitate knowledge transfer, but also universities play an important role in the foundation of new life sciences ventures. Patents and technologies that are developed through university research are commercialized trough spinoff ventures. Subsequently, both clustering and spinoff can be seen as important sources for entrepreneurial opportunity. When looking at the various life sciences clusters it is notable that they agglomerate in proximity to universities (Davis and Weinstein, 2002).
This study specifically looks at agglomerations of life sciences ventures; life sciences have been a very dynamic sector in the Netherlands. The term life sciences is used as an umbrella term for biotech, pharmaceutical, and medical technology ventures. With over 500 ventures in the Netherlands that are currently developing high technology products, and a similar number of firms that support this process of knowledge creation by supplying or advising these ventures, life sciences is an important economic motor for the Dutch knowledge economy. In 2016 there was an increase of 88 ventures that are operating, this counts for a 15% increase in ventures in one year (Dutch life sciences database, 2017).
Clustering of life sciences ventures can be seen in the landscape of Dutch life sciences ventures as well, showing large agglomerations of ventures around Dutch universities. This can be explained by benefits for knowledge sharing-and creation. Also, because of locational antecedents, most clusters exist in these locations for a long time. Clusters of Dutch life
sciences ventures can be found in Amsterdam, Groningen, Leiden, to name some of the biggest agglomerations. Not only do they have a high number of life sciences ventures, they also facilitate academic hospitals, research institutes, and accelerators to support and facilitate growth for the ventures (BioPartner, 2017; MedicalDelta, 2017).
This study investigates how location and source of foundation affect growth of Dutch life sciences ventures. Learning is believed to have a positive effect on the growth of these ventures. Specifically, the aim of this research is to find out how university spinoffs and ventures located in a life sciences cluster benefit from these traits. Also, a comparison of three life sciences clusters is made to check if there are differences between the various Dutch life sciences clusters.
1.2 Practical and theoretical contributions
This thesis aims to contribute to the knowledge spillover theory. As cluster and spinoff literature gives a wide range of views of the effect of local knowledge spillover and how this both positively and negatively could impact performance. As the composition of ventures are inherently different from each other in various geographical contexts. The learning effects that are related to local knowledge spillover, spinoff, and clustering in this thesis aim to get more insight in the life sciences sector in the Netherlands, and how this effects ventures.
Additionally, research will be performed on the differences in spinoff and independently founded ventures. Clustering will also be examined for this research, cluster membership, and clusters with various characteristics are assumed to show differences in growth. Research on this subject allows for more insight to what extent the effect of proximity enhances the learning effect of ventures. This research builds on previous work of the local knowledge spillover effects for Dutch life sciences (Weterings and Ponds, 2009; Steen et al. 2010; Geenhuizen and Soetanto, 2010), which gives insight in the determinants of new venture creation, success, and performance indicators.
The practical relevance for this research is to give insight how cluster membership or being a spinoff impacts ventures. This can be interpreted by new ventures, existing ventures, universities and policy makers. For new ventures this means they can use these insights to select location or network, and to leverage growth potential. Universities can use this research to assess the success of their spinoff ventures and gain insight in how they relate to other regions that have high concentrations of life sciences ventures. Furthermore, knowledge spillover is associated with regional growth of employment (Knoben, et al. 2011). This makes
for an interesting topic for policy makers and local governments to regulate and subsidize accordingly.
1.3 Problem statement and research question
Knowledge spillover plays an important role in the foundation of innovative firms (Audretsch and Feldman, 1996). While the field of research of knowledge spillover has emerged over the years, there is still research to be done to reveal new insights about this topic. This thesis focusses on how differences in the foundation of ventures, whether they are a spinoff or independently founded venture, result in differences in growth. Cluster membership will also be investigated, either if it matters for ventures to be located in a cluster, and if this does this impact growth.
The aim of this thesis is to give insight in the learning effects in Dutch life sciences clusters and how this affects ventures. In particular, the research will be focused on the learning effects that the spinoff parent has on the growth of the ventures. This goes the same for ventures that are located in clusters, do they benefit from their location? The literature that is available on these learning affects in agglomerations of ventures in various industries and countries, the Dutch life sciences ventures remain of interest. As the development of knowledge intensive innovations such as in life sciences are subject to different constraints, such as regulations, long time horizons, and high financing needs, compared to other industries (Hwang, 2008). Literature suggests that ventures benefit from local transfer of knowledge by collaborating with ventures that are in proximity or benefit from being a spin-off from university (Paulus et al., 2008).
Research on this subject can generate profound knowledge on how these new ventures and spinoff parents interact in Dutch life sciences clusters. This research aims to answer the following questions that address what the locational and foundational benefits for growth of ventures are: (1) what is the effect on growth of being a spinoff, particularly when they are compared to independently founded ventures? (2) What are the differences in growth of ventures that are located in clusters compared to those that are not operating in proximity of other life sciences ventures? (3) What are the differences in growth between various cluster regions?
1.4 Outline thesis
This thesis guides trough a deductive research study that is done over a period of four months. This started with laying a foundation, the theoretical framework aims to reveal knowledge about spinoffs, clusters, knowledge spillover, and the effect on ventures in knowledge intensive industries that life sciences is associated to. This theoretical framework will also lead to the hypotheses that were deducted out of the literature. Furthermore, the operationalization and methodology is explained in the chapter thereafter, giving insight in how the study is conducted and/or could be replicated. In the next chapter, research findings, these results of the study will be explained and. In the concluding chapter of this thesis the findings will be discussed, theoretical and practical implications will be given, and the author will disclose the limitations of this study.
2 Theoretical framework
In this chapter, an overview of the current literature on local knowledge spillover. Outlining spinoffs, cluster, interaction mechanisms of proximity, and the context of Dutch life sciences. The theoretical framework evaluates how learning in clusters and university spinoff effects growth of life sciences ventures. This results in three hypotheses that will be explained in this chapter.
2.1 Life sciences
Innovation in life sciences is widely regarded as an intricate process, the introduction of novel and innovative products and services is subject to several factors that can influence this process. Financial constraints, regulations, and adoption by industry professionals are factors that can influence the acceptance of new innovations in this domain (Fleuren et al., 2004). Determining what makes innovation in life sciences successful can be difficult, as little is known on the subject. Furthermore, determinants for adoption of the innovation are different in various contexts and vary from organization, city, or country.
When looking at the landscape of development of innovations in life sciences there are geographical areas that can be pointed out that are known for high concentrations of organizations in this sector. Cities like San Francisco are well known for their offspring of ventures in the biotech sector. In Europe cities like Paris, London, and Cambridge are showing high concentrations of ventures in the medical sector. These geographical concentrations of
certain economic activities can be referred to as clusters (Carpinetti, 2007). Ventures that are within these clusters seem to profit from sharing firm level knowledge (Cooke, 2001), collective learning, shared infrastructure, and financial benefits (such as lower rent). These benefits can result in higher viability and growth of ventures.
In these clusters, Local Knowledge Spillover seems to be in place, where not only knowledge, but also ideas are shared within the cluster (Caniëls & Romijn, 2005). This results in the creation of new businesses within the cluster, as there is a surplus of ideas that cannot be pursued within the boundaries of current organizations, leading to further growth of innovation clusters.
2.2 Local knowledge Spillover
Local agglomeration of ventures results in higher exchange of knowledge, this can be seen as a source for entrepreneurship, for knowledge intensive ventures the selection of location can result in better access to human capital, network, and resources. These entrepreneurial opportunities show in the initiatives of policymakers that push the development of regional entrepreneurship through subsidizing university spinoff’s, incubators, and science parks (Audretsch and Keilbach, 2007). These institutes generate high amounts of knowledge that allow to be commercialized by entrepreneurs. The high amount of knowledge and ideas generated in these locations are seen as ‘sources of opportunities’ (Shane and Venkatamaran, 2000). When high amounts of opportunities occur, it makes it more likely for individuals to discover, evaluate and exploit them (Venkatamaran, 1997). Thus, locations where high amounts of knowledge generation occur, are typically richer in entrepreneurial activity, and have more extensive entrepreneurial activities (Alvarez and Barney, 2007).
Local agglomeration of knowledge generation and entrepreneurship often show historical determinants that result in regional differences (Brenner and Fornalh, 2008). This can be explained by characteristics that make these regions more suitable for firms to locate in a specific place, such as local labor-market or natural resources. Once locational determinants are established, it is hard to break this spatial pattern (Davis and Weinstein, 2002). In the Netherlands, we can see these spatial patterns in the life sciences sector resulting in the clustering of knowledge intensive ventures mostly around universities.
The knowledge transfer that happens within these clusters is seen by scholars as ‘ubiquitous’, meaning that being within an organizational cluster means you will be exposed to new knowledge without the actor specifically being aware of this (Bathelt, 2005). Being
aware of the latest advancements, rumors, and industry specific news reports, results in an ‘industrial atmosphere’ that results in higher knowledge flows in these areas than those that have a lower concentration of ventures that operate in the same industry (Grabher, 2002).
2.3 Spinoffs
Knowledge can be transferred outside the boundaries of the firm and spring of in to a new venture. These spinoffs can come out of a variety of factors that attribute to the process of knowledge spill-over to the creation of new ventures (Zahra and Wright, 2011). Spinoffs occur both in academic or corporate context, and can result in non-profit and for-profit ventures (Fryges and Wright, 2013).
The process of entrepreneurial spinoff is characterized by either a transfer of tacit key-knowledge about marketing, production, or technology (Sapienza et al., 2004.) Knowledge transfer can also be more formal, such as the transfer of a patent or license that is used or commercialized by the spinoff. Not only is the process of spinoff defined by knowledge, there can also be a transfer of human capital, in which a team or members thereof are joining the entrepreneur in the spinoff process.
Triggers for spinoff can come from various individuals or groups within the spinoff parent and can result in various types of knowledge transfer. Spinoffs can be initiated by the founders of the spinoff parent, or function as a substitute for broadening the activities of the organization and managing it as an independent entity (Clarysse et al., 2007). Also, spinoffs are triggered by opportunity that is discovered by employees, leading to the commercialization of an opportunity that is, most often, not pursued by the parent company. Typically, spinoffs occur ‘horizontal to the parent organization, meaning they operate in the same industry domain (Franco and Filson, 2006). This allows for arrangements within the industry to spur development. Also, it is possible that spinoffs use knowledge in a different industry, benefitting a new combination of resources and spurring industry growth as well (Muendler et al., 2012).
The quality of spinoffs can largely be divided by two main phases; the formation thereof and their performance after the founders of the spinoff have left their parent firm. Both can be explained by reasons that make this process more or less successful. The founders of knowledge intensive firms that have left their parent organization are more likely to be successful than those from less knowledge intensive firms (Klepper and Sleeper, 2005). The other way around, firms that are more knowledge intensive tend to produce more spinoffs (Franco and Filson, 2000). This provides proof that in the creation of spinoff ventures, there is
an inheritance of the parent firm that can include network ties (Furlan and Grandinetti, 2014), key knowledge (Dahl and Pedersen, 2004), and access to resources.
Hence, the process of spinoff can be beneficial to spinoff performance, the characteristics of the parent firm are not solely responsible for the performance of the new venture. Also, firms that are in close proximity to the parent firm, other related experience prior entering, and location contribute to spinoff success (Dahl and Sorenson, 2009).
2.4 Spinoff parent learning mechanisms
The spinoff parent and spinoff have a shared knowledge base, this results in overlapping knowledge between spinoff and spinoff parent. An overlapping knowledge base that is too big or too small hampers the performance of the spinoff (Sapienza et al., 2004). For new ventures, the primary goal is learning, which in turn results in growth (Spender and Grant, 1996). Firm survival and having a competitive advantage thrives on the combination of knowledge and resources. These resources should come partly from the spinoff parent and external stakeholders that provide additional novel information to the spinoff (Zahra and George, 2002).
The spinoff parent provides an important means for knowledge generation. Whereas the spinoff can discard irrelevant knowledge and sustain valuable information (Grant, 1996). This common understanding between parties makes the formation of the spinoff firm more efficient. Especially in the startup process of the new firm, it helps to make more educated decisions what, for example, production process or technology should be used.
The interaction between spinoff’s and their parent organization can have various time-spans which have different effects on performance (Treibich et al., 2013). Whereas some spinoffs only inherit the rights to a certain patent or license and operate independently from thereon. Others are incubated within the parent organization in the startup phase and become more self-sustaining when the firm matures.
The interaction between both parties can be maintained for a multitude of reasons, such as a mutual benefit in generating knowledge, publishing articles, or more informal ways of collaborating (Youti and Shapira, 2008). In this collaboration, there is shared goal of complementing each other, such as the interplay of academic spinoff parent and commercial spinoff (Nowotny et al., 2001). This relationship is also complemented by proximity, personal relationships, and trust that is established by antecedents (Johansson et al., 2005).
Over time, the needs of both the spinoff and the spinoff parent change as activities of both parties’ change. Activities can become more or less aligned with each other, resulting in
a change of interaction (Treibich et al., 2013). Continuous interaction has a positive effect on the relationship but when agendas become less aligned it is difficult to realign both parties again. Also, an uneven distribution of benefits from the relation can harm the relationship, where one party profits more from knowledge generation activities than the other (Zomer, 2010). The above arguments result in the following hypothesis.
Hypothesis 1: university spinoffs have positive impact on growth of [a] employees and [b] assets
2.5 Organizational clustering
Complementing the local knowledge spillover theory, ventures in life sciences are subject to influences of other ventures that are in close proximity. These surrounding firms can be beneficial to the exchange of tacit knowledge, reduce transaction costs, and enhances a network that mitigates innovation and growth within the cluster (Cusmano, 2015). Proximity facilitates an environment that allows for collaboration rather than competition (Lejpras and Stephan, 2011). This act of agglomeration is known as clustering, where ventures that operate in the same industry are in proximity of one another.
Hence, there are thoughts of what the ideal cluster size is that is most beneficial to the agglomeration of ventures. Scholars suggest this is anywhere ranging from 50 to over 100 ventures (Klepper, 2010). Largely dependent on location, industry, and variety of ventures. Spinoffs that occur within clusters tend to spur an increase or decrease of heterogeneity of the ventures that are in the cluster (Menzel and Fornalh, 2009). Heterogeneity can lead to a more diverse knowledge set within the cluster. Homogeneity of ventures within the cluster results in more specialized knowledge and competition among firms, this type of agglomeration can result in rivalry and competition amongst ventures.
Overall, the inclusion of new ventures in clusters is seen as beneficial (Pe’er and Keil, 2013). It allows for startups to attract higher skilled employees, work with more specialized suppliers, and develop more and higher quality customer relationships. When considering the advantage of proximity that influences clusters’ innovation performance, scholars have traditionally focused on the advantage of geographical proximity (Knoben and Oerlemans, 2006). This includes the interplay of actors that exist in a specific place which facilitates accessible face-to-face access, which in turn facilitates the sharing of knowledge and networking (Boschma, 2005). Even though there is a possibility for interaction, collaboration
is not required or is not engaged because of incompatibility among firms (Torre and Rallet, 2005).
The notion of cognitive proximity bridges the gap between ventures that are located in the same geographical area, but are also compatible in their way of perceiving, interpreting, understanding, and evaluating their environment (Nooteboom, 2000). Sharing a common understanding and approach to developing knowledge and innovation leads to accumulating knowledge more efficient and effective. Hence, the collaboration between likeminded firms is assumed to have positive impact on innovation performance, firms need to actively engage in developing a network of complementary and heterogeneous actors.
The interplay of geographical and cognitive proximity facilitates the integration of shared knowledge, creating possibilities to unique possibilities to innovate. Though proximity is a key advantage to innovation performance, it is the intensity of the relationships and the ability to obtain knowledge from other ventures that mitigates this relation (Agrawal et al., 2006). This leads to the following hypothesis:
Hypothesis 2: ventures located in life sciences clusters have higher growth in [a] employees and [b] assets than ventures located outside of life sciences clusters
As mentioned, cluster membership provides more access to knowledge and resources, this leads to the assumption that ventures located in clusters can benefit from the attraction of high skilled employees, capital, and access to resources. Ventures located outside of clusters are assumed to have more boundaries to gain access to growth facilitating contributors.
2.6 Local knowledge spillover in Dutch life sciences
Knowledge is a key indicator of high-income economies, not only on a national level but also on a more local level (Döring and Schnellenbach, 2006). The exchange of knowledge is mostly on a local level, explained by a preference of face-to face contact over more distal communication methods. Though the need for direct communication has been questioned over the years by various scholars (Malmberg and Maskell, 2002), clustering does not immediately imply collaboration or exchange of knowledge. Still, even though modern day communication methods facilitate networking over large distances, proximity facilitates the flow of knowledge more easily and mostly stays within the limits of the region (Breschi and Lissoni, 2003).
Empirical evidence shows this is no different in Dutch life sciences clusters (Weterings and Ponds, 2009). Help-seeking and help-giving relations are of importance to knowledge intensive ventures, as the process of trial-and-error is costly and time consuming, network relations helps to get access to knowledge that is outside of the boundaries of the venture. This makes it more likely, but also important for performance to create industry networks (Laursen and Salter, 2004). These networks are more easily maintained over short distances, which makes it more preferable to establish regional networks. Oftentimes, only when knowledge is hard to find within the regional network, actors will try to obtain information from firms that are located further away. Still, having a regional network can be used to reach more peripheral ties that are on the far outsides of the network.
The location of some of these clusters can be related to an environment of knowledge and universities or access to capital. Though what happens in these clusters and the collaboration of the actors within, remains a black box (Caniëls & Romijn, 2005). Also, the interplay of new ventures and established ones within Dutch life sciences clusters and their learning and innovation influences on each other remains ambiguous. The first milestones that need to be overcome to start the venture are coherent with access to financing to (1) Provide an appealing business plan; (2) Acquire funding to establish the organization; (3) Get access to follow-up funding to remain viable during the startup phase. Though the interaction within cluster is largely unknown, success of these stages is assumed to be dependent on the following factors: parent organization support, quality of the team of the invention, technological strength, IP position, product potential, entrepreneurial experience, and having an external CEO (Steen et al., 2010).
There are several Dutch life sciences clusters defined, though there is a common denominator of open innovation within the various life sciences clusters, inter-organizational knowledge flows are considered beneficial for firm performance (Heirman and Clarysse, 2004). How this takes place within clusters might vary as needs from the included ventures and cluster composition is different. Some ventures might be more dependent on universities when knowledge and access to resources is scarce, whereas others benefit from regional networks to access capital (Vohora et al. 2004). Therefore, network relations should be chosen strategically and timely as they differ for the various stages the venture is in.
Also, the different concentrations of activities, cluster size, and composition, result in the assumption that the interaction within Dutch life sciences clusters is different from each
other. This should result in differences in growth among clusters, which leads to the following hypothesis:
Hypothesis 3: there is a difference in growth of [a] employees and [b] assets between clusters in different locations
For this thesis research, a selection is made of three clusters with differences in size, composition and activity: Amsterdam, Groningen, and Oss. Amsterdam as the largest cluster and capitol city of The Netherlands, Groningen as the cluster with the most spinoff activity, and Oss as the only cluster that is a non-university city with only one spinoff located in this city. This hypothesis should show the differences and variance in clusters, and inforce that differences in cluster characteristics result in differences in growth.
3 Methodology
This research is conducted in a time frame of four months. To allow for generalizable results a quantitative research study is conducted. This quantitative study entails a comparative analysis between cluster and non-cluster ventures, and between spinoff ventures and independently founded ventures. Also, three clusters are compared for differences in growth.
3.1 Research approach
To conduct the research, a robust dataset is needed, it is necessary to have enough data to avoid potential sample selection biases. There are multiple factors that can be a threat to the dataset such as: the different phases of development of the venture, different entrepreneurial attitudes, and venture activity. As this dataset is not readily available, cross-sectional data is gathered to gain the information that is needed to do the analyses. Consequently, the data that is gathered spans over a time period of 4 years, from 2011 (01-01-2011) to 2015 (01-01-2015), this data results in the information of growth for different variables. The dataset is collected from different databases. Firstly, the Dutch life sciences database (dutchlifescience.com) which is part of the international database Biotech Gate, a database that contains up-to-date information of over 40,000 companies around the world that are active in life sciences. This database allows for the discovery of all the firms that are operating in life sciences, and give a description of their activities, location, and if they’re either independently founded or spinoff
venture. Second, the Dutch chamber of commerce (Kamer van Koophandel) provides year reports of the ventures that are in the Dutch life sciences database. In these year reports, information is given about their assets and number of employees for every year that they’re operating. This data is collected manually from both databases and merged together.
According to the Dutch life sciences database, there are 544 organizations that were operating in the industry during the period of the research (Dutch life sciences database, 2017). Due to the labor-intensive work and incompleteness of information of all the organizations a sample was taken. This resulted in gathering the data manually and a selection was made to collect a dataset that was sufficient to answer the research question. Finally, a dataset of 104 ventures was collected to run the tests for the hypotheses with.
3.2 Sampling
In table 3.1 a closer look is given of the ventures that are in the sample, the whole dataset includes 104 ventures. The information that was gathered includes, age, assets 2011, assets 2015, employees 2011, employees 2015, city, sector, if the venture is located in cluster, and if a venture is a spinoff. Having these variables gives the option to investigate the data set.
This shows that from all 104 ventures, 80 of them are located in a cluster. 46 out of the total 104 ventures are spinoff. Hence, ventures can be both in a cluster and be a spinoff at the same time. To gain more insight about Dutch cluster cities, there is a selection of three cities that have distinguishing characteristics: Amsterdam, Groningen, and Oss. Amsterdam as the capital city with a high amount of life sciences organizations (N = 28), and is also known to host a vibrant entrepreneurial ecosystem. Groningen is the cluster that hosts the most spinoff activity (9 out of 15 ventures), and Oss as a cluster with almost no spinoffs that are located in its cluster (1 out of 16 ventures), also it is the only city that does not host a university.
Table 3.1: Life sciences ventures in sample
N Located in cluster Spinoff All ventures 104 80 46 Amsterdam 28 28 10 Groningen 15 15 9 Oss 16 16 1
3.3 Dependent variables
The research in this study investigates the relative performance between spinoffs and independent foundations and ventures that are located in various clusters and those that are not located in cluster based on various measurements.
3.3.1 Firm growth: assets and employees
Growth is a commonly used performance indicator, especially when the dataset includes information of newly founded ventures (Eisenhardt and Schoonhoven, 1990). This is due to the fact that it is hard to obtain profitability information of privately held firms as they are not obliged to disclose these numbers to the public. Also, ventures in their early stage, especially those with knowledge intensive innovations, need more time to become profitable. Hence, high growth numbers usually account for the expectation of profitability at present time or in the future.
For this research, the growth indicators that are used are asset growth and employee
growth. These two indicators are one of the most used measurements (Shepherd and Wiklund,
2009). The formula that is used is to calculate growth is (size t1/size t0), which result in a percentage change from the first date of measurement to the second date of measurement (Coad and Rao, 2010.). In this formula, the variables are defined as t0 = ‘2011’ and t1 = ‘2015’.
3.4 Independent variables
3.4.1 Source of foundation
One of the main independent variables, source of foundation. This nominal independent variable defines whether it is a spinoff venture, or an independently founded venture. Whereas both are ventures, the difference between both is that spinoffs benefit from patents and/or licenses from their spinoff parent. Independent foundations are not based on inheritance of such trait. Hence it is possible that they have advantage from industry networks, access to capital, or tacit knowledge and experience which can be seen as intangible resources but makes them not considered to be a spinoff. In the dataset, these are noted as spinoff (SO = ‘1’) and Independent Foundation (IF = ‘0’)
3.4.2 Located in cluster
Another main variable, the definition of cluster is given if there is a concentration of more than 5 life sciences ventures in the same city. This results in another dummy variable that is noted as outside-cluster (OC = ‘0’) and inside-cluster (IC = ‘1’)
3.5.1 Sector
Life sciences is an umbrella for multiple industries, this dataset includes pharma and biotech. To exclude the assumption that growth within a particular industry could be more beneficial than the learning effects of spinoff or cluster membership industry. Both industries are noted as biotech (BI = ‘0’) and pharma (PH = ‘1’)
3.5.2 Age
The raw data indicates the foundation year of all ventures in the dataset. To operationalize the variable, the moment of foundation is subtracted from 2015, which is latest data point that was used for this study. This results in the following calculation age = 2015 –
foundation date.
3.5.3 Location
To control the variable of location in a multiple regression analysis, location is coded as a dummy variable, creating location dummy 1, and location dummy 2. This variable is made to control if the variance can be explained by being in a certain location. The dummy variable is coded as following: Amsterdam, dummy variable 1 is ‘1’. Groningen, dummy variable 2 is ‘2’. for all other cases, the dummy variable yields ‘0’ ‘0’ for both dummies.
3.6 Research method
To analyze the data, the statistics program SPSS was used. All the data that was collected for this research was aggregated in to a single data set. Then, all the variables were coded as explained in the previous paragraph.
3.6.1 Checking normality
When all data was collected, it was coded as listed in the previous paragraph. The dependent variables were checked for normality using standard Q-Q plots and the Shapiro-Wilk test. This resulted in the observation that the dependent variables ‘employee growth’ and ‘asset growth’ were distributed normally. With the dependent variable ‘employee growth’ having p = .182, and the variable ‘asset growth‘ having p = .377, both variables are not significant on the Shapiro-Wilk test. This means both variables are normally distributed.
3.6.2 Descriptive statistics
Descriptive statistics were calculated in SPSS to gain insight in the statistics of the whole sample, but also of the different groups within the dataset. This
3.6.3 Comparing means
This lead to the selection of the tests per hypothesis. The differences between the three hypotheses were the number of groups. Therefore, the hypothesis with two groups were subject to a t-test and ANOVA was used to compare means for the third hypothesis that uses three groups (Amsterdam, Groningen, and Oss). In the results section, descriptive statistics are also given to get more insight in the various groups that are used to answer the research question.
3.6.4 Linear regression
To explain the variance of the effect of the independent variables on the dependent variables growth of employees and growth of assets, a multiple linear regression will be performed.
3.6.5 Cluster location dummies
To have more specific attention to the various clusters a selection of three clusters is made for this research: Amsterdam (AM = ‘1’), Groningen (GR = ‘2’), Oss (OS = ‘3’), and all other ventures. This selection is made to do a comparative analysis between the 3 cluster cities that hold different characteristics.
4 Research findings
Doing a comparative analysis, tests are done to measure if groups of life sciences ventures with different characteristics have different levels of growth. These different characteristics entail spinoffs, location in clusters, and the differences between three clusters. These tests do not necessarily explain the causal relation of differences in growth, but are used to explain statistically if there are differences in growth between groups. Also, a regression analysis is done to see what the interaction effect is on growth of employees and growth of assets.
First, descriptive statistics are given about the whole sample of life sciences ventures, this includes information about the growth of employees and assets during a period of four years, between 2011 and 2015. Furthermore, the results for spinoffs, clusters, and differences between Amsterdam, Groningen, and Oss will be described. The descriptive statistics are given over table 4.1, table 4.2, table 4.3, and table 4.4, each showing a different variation of the sample. In this table the mean, standard deviation (std. dev.), minimum, and maximum value is given within each sample. Also, each sample includes the N, which indicates the number of ventures in each sample.
Table 3.1, to be found in the previous chapter, shows the descriptive statistics of all life sciences ventures (N = 104), this shows several insights about the industry. Overall, the life sciences in the sample are growing, showing growth of employment of 52 percent and growth of assets of 131 per cent, over the same period. From these statistics, we can assume that most firms in the sample are currently growing. Other statistics that show us a more detailed view of the sample show that spinoffs are the sample that hold both the highest growth in employment (78%, SD = 1.23) and growth in assets (144%, ds = 2.70).
Table 4.1: descriptive statistics for all ventures
Mean Std. Dev. Min. Max. N
All ventures Age 11 7 4 50 104 Employees 2011 20.38 40.15 1 271 104 Employees 2015 22.45 37.99 1 205 104 Growth Employees 1.52 1.04 0.21 6.5 104 Assets 2011 2603.20 7072.90 12 50920 104 Assets 2015 3740.71 10334.42 9 86419 104 Growth assets 2.31 2.99 0.05 18.04 104
4.1 Descriptive statistics for spinoffs and independently founded ventures
Looking at table, 4.2, descriptive statistics are given to show the differences between spinoffs and independently founded ventures, this shows several findings about the sector. Spinoffs compared to independently founded ventures have higher growth in both employees (M = 1.78, SD = 1.23) and assets (M = 2.44, SD = 2.70) than independently founded ventures (M = 1.32, SD = 0.18), (M = 2.21, SD = 3.23). Even though growth was higher for the spinoff group, independently founded ventures on average have more employees and assets than spinoffs.Table 4.2: descriptive statistics for spinoffs and independently founded ventures
Mean Std. Dev. Min. Max. N
Spinoff Age 9 5 4 28 46 Employees 2011 14.15 25.95 1 115 46 Employees 2015 18.65 33.44 1 179 46 Growth Employees 1.78 1.23 0.44 6.5 46 Assets 2011 2249.17 7684.30 23 50920 46 Assets 2015 2492.80 5996.55 13 35308 46 Growth assets 2.44 2.70 0.22 12.80 46 Independent foundation Age 12 8 4 50 58 Employees 2011 25.33 48.22 1 271 58 Employees 2015 25.47 41.29 1 205 58 Growth Employees 1.32 0.81 0.21 6.5 58 Assets 2011 2883.98 6602.96 12 40404 58 Assets 2015 4730.43 12741.58 13 86419 58 Growth assets 2.21 3.23 0.22 17.22 58
4.2 Descriptive statistics for ventures located in a cluster and ventures not
located in a cluster
Table 4.3 shows the descriptive statistics for the different groups, there are several findings to report that shows the differences between ventures located in a cluster and ventures not located in a cluster. First, growth of employees, and growth of assets is relatively similar between the two groups. Resulting in growth of employees of (M = 1.52, SD = 1.07) for ventures located in a cluster and (M = 1.54, SD = 0.95) for ventures not in a cluster. For growth of assets a similar small difference is shown, resulting in (M = 2.28, SD = 2.86) for ventures in a cluster and (M = 2.41, SD = 3.48) for ventures that are not located in a cluster.
Another interesting observation is the high amount of ventures that are in a cluster, from the total sample 80 ventures are located in a cluster and 24 are located outside of a cluster. This is checked if there was a bias within the sample. The raw data of ventures that was collected included 364 ventures of which 244 were located in a cluster. This results in 77% of the sample located in a cluster, and 67% of the raw data to be included in a cluster. Even though we can’t readily explain it with the descriptive statistics
Table 4.3: descriptive statistics for ventures located in a cluster and outside of a cluster
Mean Std. Dev. Min. Max. N
Located in cluster Age 11 7 4 50 80 Employees 2011 19.56 34.19 1 147 80 Employees 2015 21.64 35.06 1 179 80 Growth Employees 1.52 1.07 0.21 6.5 80 Assets 2011 1586.70 2935.91 12 15810 80 Assets 2015 2220.49 3960.47 9 19794 80 Growth assets 2.28 2.86 0.05 17.22 80
Not located in cluster
Age 9 4 4 15 24 Employees 2011 23.13 56.52 1 271 24 Employees 2015 25.17 47.25 1 205 24 Growth Employees 1.54 0.95 0.5 5 24 Assets 2011 5991.54 13373.43 26 50920 24 Assets 2015 8808.13 19737.52 35 86419 24 Growth assets 2.41 3.48 0.41 18.04 24
4.3 Descriptive statistics for clusters Amsterdam, Groningen, and Oss
Looking at table 4.4, this set of descriptive statistics shows the differences of life science clusters Amsterdam, Groningen, and Oss. Several findings came to light while examining the statistics at hand. This shows that Groningen has the highest growth in employees (M = 1.50, SD = 0.84) and growth of assets (M = 2.41, SD = 3.24). It is also the cluster with the highest average age of ventures (M = 15, SD = 9). Considering the sample, Amsterdam is the largest cluster (N = 28. Not only in the sample is Amsterdam the biggest sample, but also looking at the data from Dutch life sciences database (2017) it shows that Amsterdam is the largest cluster (N = 40). The other two clusters are smaller, with Groningen as the second biggest cluster (N = 23), and Oss the smallest (N = 18)
Table 4.4: descriptive statistics for life sciences clusters Amsterdam, Groningen, and Oss
Mean Std. Dev. Min. Max. N
Amsterdam Age 11 9 4 50 28 Employees 2011 31.25 44.46 1 147 28 Employees 2015 33.54 47.62 2 179 28 Growth Employees 1.23 0.69 0.21 4 28 Assets 2011 2531.50 4475.53 18 15810 28 Assets 2015 3375.21 5762.35 9 19794 28 Growth assets 2.41 3.24 0.05 17.22 28 Groningen Age 15 9 4 29 15 Employees 2011 19.40 36.41 1 110 15 Employees 2015 21.13 33.96 2 111 15 Growth Employees 1.50 0.84 0.5 4 15 Assets 2011 1037.93 1441.84 56 4522 15 Assets 2015 1587.13 2649.10 57 9761 15 Growth assets 1.63 1.64 0.42 6.74 15 Oss Age 10 4 5 18 16 Employees 2011 12.38 22.98 2 95 16 Employees 2015 13.00 24.00 1 98 16 Growth Employees 1.14 0.63 0.33 2.5 16 Assets 2011 1515.13 1727.16 12 5044 16 Assets 2015 2026.13 2159.92 20 7130 16 Growth assets 2.32 3.19 0.31 12.80 16
4.4 Independent sample t-test: spinoff and independent foundation
To check if the differences in growth can be explained statistically the independent samples t-test was used to compare groups. First, the group of spinoff and independent foundations was tested on both growth of assets and growth of employees by using an independent samples t-test (table 4.5), this resulted in a significant difference in growth of employment for the spinoff venture (M = 1.78, SD = 1.23) and independent foundations (M = 1.32, SD = 0.81); t(102) = -2.31, p = .023. Hence, the growth of assets did not see a significant result for spinoff (M = 2.31, SD = 2.99) and independent foundations (M = 2.21, SD = 3.23); t(102) = 0.83, p = .633.
Table 4.5: Differences in spinoff and independent foundations
Total sample Spinoff Independent foundation
Growth employees 1.52 (1.04) 1.78 (1.23) 1.32 (0.81) t-test -2.31* Growth Assets 2.31 (2.99) 2.44 (2.70) 2.21 (3.23) t-test -0.39 *p < 0.05
This leads to a partial acceptation of hypothesis 1, there is a difference in growth of employees [h1a] between spinoffs and independently founded ventures. The other part of the hypothesis [h1b] cannot be accepted, as there is no statistical power to prove so.
To look further in to what the effect is of being a spinoff venture a regression analysis will be done in paragraph 4.7.
4.5 Independent samples t-test: ventures located in a cluster and ventures not
located in a cluster
For the second hypothesis, differences between ventures located in clusters and ventures not located in clusters are compared. This is done by using an independent samples t-test with the two groups, the t-test is done twice t-testing both for growth in employees and growth in assets (table 4.6). Both tests reported no significant results, for the growth of employees the test reported for the spinoff venture (M = 1.52, SD = 1.04) and ventures not located in a cluster (M = 1.54, SD = 0.95); t(102) = -0.54, p = .363. For the growth of assets, the independent samples t-test reported (M = 2.31, SD = 2.99) and ventures not located in a cluster (M = 2.41, SD = 3.48); t (102) = -0.08, p = .633.
Table 4.6: differences in ventures located in a cluster and ventures not located in a cluster Total sample Located in cluster Not located in cluster Growth employees 1.52 (1.04) 1.52 (1.07) 1.54 (0.95) t-test -0.54 Growth Assets 2.31 (2.99) 2.28 (2.86) 2.41 (3.48) t-test 0.08
This leads to the rejection of hypothesis 2, that there would be a difference in growth of [h2a] employees and [h2b] assets between ventures located in a cluster, and ventures not located in a cluster.
4.6 ANOVA, comparing clusters: Amsterdam, Groningen, and Oss
The results of the comparison of the three clusters did not show any significant difference of various cluster characteristics on growth of employees (table 4.7) and assets (table 4.8). The test on employees resulted in F(2, 59) = 1.87, p = .16 The test on differences between growth of assets reported F(2, 59) = .243, p = .79, which does not result in any significant result either.
Table 4.7: one-way analysis of growth of employees by cluster cities
Source df SS MS F p
Between groups 2 4.024 2.012 .243 .785
Within groups 59 488.741 8.284
Total 61 492.766
*p = <.05
Table 4.8: one-way analysis of growth of assets by cluster cities
Source df SS MS F p
Between groups 2 2.700 1.350 1.868 .163
Within groups 59 42.631 .723
Total 61 45.331
*p = <.05
This leads to the rejection of hypothesis 3. The assumption that was made, that there were differences in [h3a] growth of employees and [h3b] assets between the three life science clusters, this was not supported by the data.
4.7 Regression analysis, growth of employees and assets
A multiple regression analysis was done to see what factors significantly predict the growth of employees and growth of assets. The result of the regression of growth of employees (table 4.9) indicated one significant predictor of employee growth, the model explained 10.4% of the variance (R2 = .104, F(8,95) = 1.381, p = .215. The significant predictor here is the source of foundation, this describes there is a positive effect of being a spinoff venture to growth of employees (β = .209, p = .047). This finding supports hypothesis 1a, being a spinoff has a positive effect on growth of employees.
Table 4.9: linear regression of growth on employees
B (s.e.) β Constant 1.894 (.470) ln employees 2015 .168 (.107) .217 ln assets 2015 -.092 (.077) -.157 Source of foundation .435 (.216)* .209 Located in cluster .062 (.266) .027 Sector -.095 (.211) -.046 Age -.030 (.017) -.199 Location dummy 1 .180 (.317) .061 Location dummy 2 .024 (.294) .009 N 104 R Squared .104 *p = <.05
The second multiple regression analysis was done to see what indicators were able to predict growth of assets (table 4.10). This resulted in not finding any significant indicators of the variable. The model explained 5.7% of the variance on the dependent variable (R2 = .057, F(8,95) =.718, p = .675.
Table 4.10: linear regression of growth of assets B (s.e.) β Constant 1.916 (1.39) ln employees 2015 -.311 (.318) -.139 ln assets 2015 .341 (.228) .203 Source of foundation -.080 (.638) -.013 Located in cluster -.596 (.788) -.089 Sector -.145 (.625) -.024 Age -.066 (.049) -.152 Location dummy 1 -.674 (.937) -.079 Location dummy 2 -.706 (.871) -.088 N 104 R Squared .057 *p = <.05
5 Discussion and conclusion
By doing a comparison of different groups of life sciences ventures and the learning benefits they have for their venture growth has led to multiple findings that will be discussed in this chapter. In this thesis spinoffs and clusters are examined, also a closer look at three cluster cities was made to check for differences. This research found that spinoffs found higher employee growth than independently founded ventures. Hence, there is a significant difference between the two groups, the causal relationship about this phenomenon cannot be explained solely by using statistics. The assumption that we can make is with what we can find in prior research. By giving higher access to knowledge and problem-solving abilities, universities provide a great resource to new ventures (Sapienza et al., 2004). Therefore, spinoffs that collaborate closely with universities are provided with a safe environment to grow their organization.
The results from this research were less clear on the differences between ventures located in clusters, and ventures not located in clusters. Also, differences in growth between clusters Amsterdam, Groningen, and Oss could not be explained. Even though the assumption was made that there would be differences in growth, these could not be statistically proved. There was no relationship found with higher growth and being in a cluster. Neither was there any significant result in the different cluster locations. being in one of the cities that was chosen for this research, did not show any significant differences.
To elaborate further on the rejected hypotheses; for the former assumption that there would be learning benefits by being in a cluster, the results could not prove that. This assumption was made that proximity to knowledge and access to high skilled employees would be beneficial toward growth of employees. However, this hypothesis could not be proven. What we can derive from statistics is that 77% (table 3.1) of ventures in the sample are located in cluster. This leads to believe that being in a cluster provides advantages that are not shown in statistics. This can be the historical character of a certain region that results in agglomeration of ventures from a related industry (Brenner and Fornalh, 2008).
Hypothesis 3, in which was expected to find a difference in growth of employees and assets could not be confirmed by the results of the analyses. This suggests that there are no attributes that result in differences related to growth within clusters. As Caniëls & Romijn (2005) suggest, inclusion in a cluster is beneficial for performance, but what happens within the cluster is considered a ‘black box’, meaning that clusters are unique from each other, clusters can be widely different and hard to compare. As the composition is different for every cluster, best practices are not suitable to define what is the best way. Whereas some clusters might benefit from a safe environment to develop a product in a university context, others might benefit more from a cluster that offers access to a network of investors.
5.1 Theoretical and practical implications
Learning is an important means for growth. This research addresses the how learning effects growth of clusters and spinoff ventures in life sciences. This results in a broader view how the findings can be interpreted both by scholars, policy makes, and entrepreneurs. Firstly, for scholars this thesis provides insight in the growing industry of life sciences in the Netherlands. While an overall measurement shows the industry is growing, the growth and interaction effects that it is subject to remains an intricate question. We have gained understanding that there are indicators that there are differences within groups of life sciences ventures, particularly spinoff firms.
The practical implications result in our understanding of the important role of universities in the Dutch life sciences sector. Scholars point out the several stages of the firm life cycle that universities can collaborate and profit from industry relations. This can be in the early startup phase but also with ventures that are in a more mature stage. The exchange of knowledge, joint research ability, and access to high skilled employees can result in benefit from both parties. As the data for this research suggests, clusters have different compositions
of ventures and spinoffs, also there are ventures that are not located in clusters at all. For ventures in their startup phase it is interesting to gain understanding of these characteristics and choose location wisely, as benefits of proximity may positively influence their growth potential.
Also, there are implications for policy makers, supporting life sciences ventures can be intricate as this study points out the various characteristics in cluster compositions point out that there is no one size fits all application for ventures. Literature points out though that sharing of knowledge and establishing partnerships is beneficial for the whole industry. Policy makers could play a key role in facilitating these arrangements.
5.2 Limitations and future research
The results of this study are subject to some limitations that should be taken in to account. Firstly, the data used for this study consists of cross sectional data, this data consists of data that was at hand at the time of this research, sourced from the life science data base and the chamber of commerce. This results that numbers about firms that stopped their activities during the period that is examined during the period of research are excluded. For more reliable results panel data would be more beneficial, unfortunately this was not possible for the time period of four months in which this research was done.
Second, the data set consists of ventures that are different in age, resulting in a sample of both startups and established firms. This can result in performance differentials that can be attributed to the firms being in a different life stage. Whereas growth might be higher in the startup phase and dwindle down once they will become more mature. Also, the assumption that is made in this thesis is that proximity and spinoff will positively benefit venture growth. This excludes other variables that perhaps have an influence on firm growth as well.
Also, it can be assumed that spinoffs are only registered with the chamber of commerce once they are deemed to be successful enough to become an independent entity. Those that are considered ‘failures’ will stay within the boundaries of the university and are not subject to this research.
More quantitative research on the subject could result in a research design that includes more variables. This could include deeper understanding of the networking and collaborations that happens inside of clusters, and how it effects life sciences ventures in particular. Also, there are variables that this research has overlooked, variables such as cluster size and venture age could help explain more about the growth of employees and assets of life sciences ventures.
Furthermore, a qualitative approach to learning effects and venture growth could give more insight in the causal relations. Whereas differences between groups are found in this study, we can only explain this in light of the existing literature, possibly overseeing other relations that might be in place.
5.3 Conclusion
This research has given insight in the landscape of Dutch life sciences ventures and the effects of local knowledge spillover. The study covered the effects several traits of the ventures. The source of foundation, whether a firm was founded independently or a spinoff of a university. The location, whether ventures were in proximity to other firms that were also operating in life sciences. Also, characteristics of different clusters were examined. These different groups were subject to a comparative analysis and if there were differences in growth and assets between ventures with the before mentioned characteristics.
Literature led to the assumption that proximity of knowledge would benefit in higher growth of both assets and employees, thus spinoffs and ventures located in clusters should benefit from this. Because of the distinct composition of life sciences clusters an extra assumption was made, that there are differences in growth between the clusters. A selection of the clusters Amsterdam, Groningen, and Oss was made.
With the data set that was used one of the assumptions that was made can be supported. Spinoffs have higher growth than independently founded ventures. This relation is supported in literature by their advantage to operate in a less hostile environment than independent ventures. Also, lower boundaries to collaborate with a knowledge intensive institute such as a university is seen as an advantage. The other assumptions that were made in this thesis could not be supported with this research. No significant relations could be found to explain differences in ventures that are located in a cluster or differences between the clusters that were in the data set that was used for this research.
The understanding that is gained from this research is that the Dutch life science sector have several contributors to growth, these factors are not subject to best practices. Whereas the literature points out the multitude of ways how ventures within clusters interact or support each other. take place in clustering and the process of spinoff, this is likely applicable for Dutch life sciences as well. Therefore, more in-depth studies could help shed light on these differences. This could be done with a quantitative study with a larger sample or a qualitative study with a narrower focus.
6 References
Agrawal, A., Cockburn, I., & McHale, J. (2006). Gone but not forgotten: knowledge flows, labour mobility, and enduring social relationships. Journal of Economic Geography, 6(5), 571-591.
Audretsch, D. B., & Feldman, M. P. (1996). R&D spillovers and the geography of innovation and production. The American economic review, 86(3), 630-640.
Audretsch, D. B., & Keilbach, M. (2007). The theory of knowledge spillover entrepreneurship. Journal of Management Studies, 44(7), 1242-1254.
Bathelt, H. (2005). Geographies of production: growth regimes in spatial perspective (II)-knowledge creation and growth in clusters. Progress in human geography, 29(2), 204-216.
Boschma, R. (2005). Proximity and innovation: a critical assessment. Regional studies, 39(1), 61-74.
Boschma, R. A., & Wenting, R. (2007). The spatial evolution of the British automobile industry: Does location matter?. Industrial and corporate change, 16(2), 213-238. Breschi, S., & Lissoni, F. (2001). Knowledge spillovers and local innovation systems: a critical
survey. Industrial and corporate change, 10(4), 975-1005.
BioPartner. (n.d.). BioPartner. Retrieved from http://www.biopartner.nl
Fornahl, D., & Brenner, T. (2003). Cooperation, networks and institutions in regional innovation systems.
Carpinetti, L. C. R., Gerolamo, M. C., & Galdámez, E. V. C. (2007). Continuous innovation and performance management of SME clusters. Creativity and Innovation
Management, 16(4), 376-385.
Caniëls, M. C., & Romijn, H. A. (2005). What drives innovativeness in industrial clusters? Transcending the debate. Cambridge Journal of Economics.
Clarysse, B., Wright, M., & Van de Velde, E. (2011). Entrepreneurial origin, technological knowledge, and the growth of spin-off companies. Journal of Management
Studies, 48(6), 1420-1442.
Coad, A., & Rao, R. (2010). Firm growth and R&D expenditure. Economics of Innovation and New Technology, 19(2), 127-145.
Cooke, P. (2001). Regional innovation systems, clusters, and the knowledge economy. Industrial and corporate change, 10(4), 945-974.
Cusumano, M. A., Kahl, S. J., & Suarez, F. F. (2015). Services, industry evolution, and the competitive strategies of product firms. Strategic management journal, 36(4), 559-575. Dahl, M. S., & Pedersen, C. Ø. (2004). Knowledge flows through informal contacts in
industrial clusters: myth or reality?. Research policy, 33(10), 1673-1686.
Dahl, M. S., & Sorenson, O. (2009). The embedded entrepreneur. European Management
Review, 6(3), 172-181.
Davis, D. R., & Weinstein, D. E. (2002). Technological superiority and the losses from migration (No. w8971). National Bureau of Economic Research.
Döring, T., & Schnellenbach, J. (2006). What do we know about geographical knowledge spillovers and regional growth?: A survey of the literature. Regional Studies, 40(03), 375-395.
Eisenhardt, K. M. (1989). Building theories from case study research. Academy of management review, 14(4), 532-550.
