By Vartuhi Tonoyan, Robert Strohmeyer & Michael Woywode

Does A Cluster‘s Location Matter for Companies‘ Survival and Growth? Conclusions

from the German Biotechnology Industry

1

By

Vartuhi Tonoyan, Robert Strohmeyer & Michael Woywode

Presentation for the Howe School of Technology Alliance Meeting, November 13

Who We Are… Dr. Vartuhi Tonoyan, Assistant Professor Study of Economics in Armenia and Germany

(graduation with distinction) Dissertation in Management / Entrepreneurship at

the University of Mannheim, Germany (graduation with summa cum laude)Topic: “Corruption, Entrepreneurship and Institutional Environment: A Cross-National Comparison of Emerging and Mature Market Economies”

Head of Entrepreneurship Research Division at the Mannheim’s Institute for SME Research, Germany

1-Year of Post-doc study of Entrepreneurship at the Graduate School of Business of the Stanford University

Joined the Stevens Institute of Technology as Assistant Professor in September of 2011

2

Who We Are…

Robert Strohmeyer, Diploma of Sociology Study of Social Sciences and Statistics at the

University of Mannheim (magna cum laude) PhD-Candidate at the University of Mannheim,

Germany Visiting Research Scholar at the Stevens Institute of Technology Dr. Michael Woywode, Full Professor Head of Institute for Small Business Research

at the University of Mannheim Research Associate at the ZEW Mannheim

3

Research Areas and Selected Projects (1) I. High-Technology Entrepreneurship and Innovation Emergence and Evolution of New High-Tech Industries: Biotechnology Industry in Germany and the US App Industry in the US Strategic Alliances and Collobarations in High-Tech Industries Allocation of Control Rights in Bio-Pharmaceutical Alliances Evaluation of Publicly Funded R&D Projects Impact of Public R&D Funding on Private R&D Investments Project Management Biotechnology Entrepreneurs‘ Decision to Persist with Under-

Performing R&D Projects

4

Research Areas and Selected Projects (2)

II. Corruption and Development Economics Impact of Corruption on Firm Innovativeness in Emerging and

Mature Market Economies

III. Entry and Growth Determinants of Entrepreneurial Companies; Entrepreneurial Decision-Making

Gender gap in entrepreneurship Gender-specific differences in firm performance (such as

employment growth and innovation)

5

Empirical Evidence on Cluster Effects Lack of studies regarding the impact of cluster embeddedness

on firm performance (notable exceptions are Baptista & Swann (1998) for the UK, Dahl & Pedersen (2004) for Denmark, Geenhuizen & Reyes-Gonzalez (2007) for the Netherlands, Whittington et al. (2009) for the US, and Wennberg & Lindquist, 2010, Powell et al. 2011)

Geenhuizen & Reyes-Gonzalez (2007, p. 1683): There is a shortage of studies in which the theoretical and policy claims of the cluster advantages in biotechnology are critically evaluated on basis of a systematic and longitudinal analysis

6

Questions to be answered: Does it pay to locate in a cluster? If so, what cluster(s) to choose? What differentiates good clusters from bad ones?

Answers to these questions should be relevant both for high-tech entrepreneurs and policy-makers !

Definitions of Clusters

7

No ubiquitous definition of clusters

Porter (1998, p. 199) “Cluster is a geographic concentration “of interconnected companies, specialized suppliers and service providers, firms in related industries, and associated institutions (e.g. universities, standards agencies, and trade associations) in particular fields that compete but also cooperate” Other definitions do not assume vertical or horizontal relationships

between companies and/or co-operations between high-tech companies and associated institutions

Still others argue that networking and cooperation should be an important prerequisite for clusters (i.e. geographical concentration of companies is not a sufficient criteria)

Common to all definitions is the geographical concentration of firms in specific regions

No agreement regarding the level of geographical concentration or quorum of organizations or maximum of possible area of expansion of a cluster

Mechanisms of Cluster Effects on Firm Performance (1)

Transportation costs (industries locate close to resources in

order to minimize transportation costs)

Availability of specialized labor force

Intra-industrial specializations (within core industry)

Inter-industrial specializations (specialized suppliers, investors,

end product purchaser)

8

Mechanisms of Cluster Effects on Firm Performance (2)

Transaction cost perspective: Reduces costs for finding business counterparts, defining contract conditions, and monitoring agreements

Knowledge spillover (transfer of explicit and implicit knowledge)

Increases the legitimation of companies

9

Mechanisms of Cluster Effects on Firm Performance (3)

Criticism: The benefits of co-operation can also be reached via social

networking which does not require regional proximity (such as strategic alliances) (Bathelt et al., 2004)

Very strong clusters might produce adverse effects due to congestion and hyper-competition among firms and personnel (vgl. Folta et al., 2006; Sorenson & Audia, 2000)

Lock-in effects: isolation/ freezing of structures withing a cluster

10

Research Questions

Does the cluster embeddedness result in a higher firm performance?

Is there a cluster heterogeneity? What differentiates successful clusters from less successful ones?

11

Multilevel-Analysis with Panel Data

Level 1: Firm-Level Characteristics Firm size and age, business model, venture capital funding

and public R&D subsidies, cooperations (with biotech firms, Big Pharma industry, and also research institutes)

Level 2: Cluster-Level Characteristics Cluster size and cluster age, the level of venture capital

funding and public R&D subsidies, cooperation density, international collaborations

12

Hypotheses (1)Cluster Perspective

Cluster Size

Cluster effects are expected after only a minimum size of a cluster has been reached

Cluster’s Composition and Degree of Maturation

Biotech firms which operate in clusters with a higher percentage of companies focusing on product development and sales will be more successful

Biotech firms which operate in more mature clusters, i.e. clusters with a higher percentage of established companies, will be more successful

Cluster’s Degree of Internationalization Biotech firms which operate in clusters with a higher degree of

internationalization will be more successful

13

Hypotheses (2)Cluster Perspective (cont.)

Network Composition and Existence of Big Players on Cluster Level: Biotechnology firms which operate in clusters with a more

balanced network composition in terms of cooperation with other biotech companies, research institutes and industrial companies will be more successful

Biotechnology firms which operate in clusters with a higher number pharmaceutical companies will be more successful

Cluster’s Level of Funding (Venture Capital Money and Public Subsidies) and Science Dominance

Biotechnology firms which operate in clusters with a more balanced composition in terms of venture capital funding and the science dominance will be more successful

14

Data, Variables and MethodsData Panel data (1998-2008): N= 1064 biotechnology firms in Germany;

Sources: BioCom and Creditreform data

Methods Panel-Econometrics (Wooldrige, 2000) Multilevel-Analysis (Raudenbush & Bryk, 1992)

Dependent and Independent Variables The dependent variable measures bio-technology firm performance

(employment growth) Independent variables at two levels:

Level 1 predictors: firm characteristics (such as age, business models, funding)

Level 2 predictors: cluster characteristics (cluster size, the level of venture capital money and public R&D subsidies at the cluster level, cooperation at the cluster level)

15

16

Multilevel ModelingExample of a simple 2-level model

Level-1 Model (e.g. individuals) Yij = β0j + β1j Xij +rij (1) i represents individuals (i = 1……..n) j represents countries (j= 1……J) rij ~ independently N( 0 , 2 ) Xj : Individual (Level 1) Variable

Level-2 Model (e.g. countries) β0j = γ00 + γ01Wj + u0j (2) β1j = γ10 + γ11Wj + u1j (3) u0j ~ independently N( 0 , 2 ) u1j ~ independently N( 0 , 2 ) Wj : Country (Level 2) Variable

Level-1/2 Model (individuals & countries) Yij = γ00 + γ01Wj + γ10 Xij + γ11Wj Xij + [ u0j + u1j Xij + rij ] (4)

Error Term

Source: Raudenbush & Bryk (2002), p. 128

Descriptive Statistics (1)

17

253

316

393

456483 469 468

497 495 496 501

0

200

400

600

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Figure 1: Number of Dedicated Biotechnology Enterprises in Germany (1998-2008)

Source: ifm/ZEW Bio-Tech Entrepreneurship Panel 2000-2009 (based on Biocom Data)

9,2 9,6

13,7 13,612,5 12,7 13,0

14,2 14,4 14,5

0,0

5,0

10,0

15,0

20,0

1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Descriptive Statistics (2)

18

Figure 2: Number of Employees in Dedicated Biotechnology Enterprises in Germany (1998-2008)

Source: ifm/ZEW Bio-Tech Entrepreneurship Panel 2000-2009 (based on Biocom Data)

Descriptive Statistics (3)

19

0

25

50

75

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008

Red BT Green BT White BT Non-Specific BT IT-Biotech

Figure 3: Share of Enterprises in Biotech Industry By Colors / Fields of Activity

Source: ifm/ZEW Bio-Tech Entrepreneurship Panel 2000-2009 (based on Biocom Data)

Results of Panel Estimation (1)

20

Model 1 Model 2 Model 3

b/t b/t b/t

Firm Level Variables

Age -0.002 -0.002 -0.011

(-0.21) (-0.25) (-1.06)

Business model: R&D only 0.187 0.206 0.187

(1.49) (1.52) (1.38) Business Model: R&D + Services

0.098 0.070 0.077

(1.09) (0.71) (0.78)

Venture Capital 0.347*** 0.384*** 0.290***

(4.54) (4.56) (3.07)

Public Subsidies 0.146** 0.118* 0.132*

(2.21) (1.67) (1.77)

Coop: Industry / Big Pharma 0.117 0.106

(0.80) (0.71)

Coop: Biotech Firms -0.219 -0.233

(-1.55) (-1.64)

Coop: Foreign Research Institute/University

0.533*** 0.542***

(2.75) (2.76)

Results of Panel Estimation (2)Model 1 Model 2 Model 3

b/t b/t b/t

Cluster Level Variables

Cluster size 0.006**

(2.36)

Cluster maturity (mean age) 0.027

(0.92)

Cluster VC (% of Enterp.) 0.368**

(2.23)

Cluster Subsidies (% of Enterp.)

-0.077

(-0.61)

Cluster: Cooperation Density -0.025

(-0.82)

Constant -0.571*** -0.572*** -0.582***

(-17.27) (-17.16) (-17.49)

21

Source: ifm/ZEW Bio-Tech Entrepreneurship Panel 2000-2009 (based on Biocom Data)  M1 is firm level model, M2 = M1 plus cooperation variables, M3 controls for cluster level characteristics

Summary of Results (1)

22

Firm-Level Results: Firm age and business models are not related with employment

growth Venture capital is a strong positive predictor of employment

growth Public R&D subsidies have a positive effect on firm growth Collaborations with big pharmaceutical industry companies have

no effects on employment growth In contrast, collaborations with a foreign research institute or

university strongly positively influence employment growth

Summary of Results (2)

Cluster-Level Results

Cluster size matters: the bigger a cluster in terms of the number of enterprises, the higher the probability of firm growth

But, the degree of cluster’s maturation is not significant in models

Strong VC investments effect on the cluster level: the higher the percentage of firms in a cluster which have received venture capital, the higher the probability of employment growth of firms in this cluster (independent of whether the firm itself has received VC-funding or not)

A high number of collaborations at the cluster level is not related with employment growth

23

Limitations

24

Survivor bias : we have analyzed only those firms in the panel which survived over 2004-2008

Next step: we will implement a selection equation

Reference category: non-clustered firms (a very low N, ca. 30 biotechnology companies)

Implications for Future Research

25

Firm innovativeness (patents) / growth in sales as dependent variables

Panel for 1998-2008 (to analyze exogenous changes in the environment of biotech companies, such as the impact of the financial crisis in 2001 in Germany and/or the world financial crisis in 2008 on firm performance)

Analysis of the institutional environment of clusters and thus rules which regulate the attraction and retention of the venture capital investments , and/or the labor market mobility and flexibility

A US-German comparison of the effects of cluster embeddedness on firm growth

Comparison of performance of firms involved in clusters with the performance of their non-clustered counterparts involved in strategic alliances

26

Thank you for your attention !

Presentation by Vartuhi TonoyanHowe School of Technology Management

Stevens Institute of Technology [email protected]

Backup

27

Variables Description: Business model: R&D only: no services, no production, no sales and

distribution Business Model: R&D & Services: hybrid model, doing both R&D and

biotech services on technology platforms Venture Capital: received VC in year X (dummy) Public Subsidies: received Public subsidies in year X (dummy) Coop: Industry / Big Pharma : dummy Coop: Biotech: dummy Coop: Foreign Reseach Institut /University: dummy

Cluster Level Variables Cluster size : number of BT firms in cluster X Cluster Mature (mean age) : mean of age in cluster Cluster VC (% of Enterp.): %-share of enterprises in cluster X which

received VC Cluster Subsidies (% of Enterp.): %-share of enterprises in cluster X

received public subsidies Cluster: Cooperation Density: mean of number of co-operations in

cluster X

Cluster Identification

28

Usage of MPCluster plugin for MS MapPoint Comparison with Zip-Code method K-Means algorithm Parameters:

max radius: 59 Miles (90 Km), min # of firms: 10

Key Take-Away: 91.67% of the identified clusters are identical with officially

existing clusters