Master Thesis
MSc BA Small Business & Entrepreneurship
“Agglomeration economies in a digital world: The impact of
geographical proximity on SME growth in the ICT era”
by
Fabian M. Brunner
1University of Groningen
Faculty of Economics and Business
2June 2017
Word Count
3: 16.491
Supervisor: Dr. Samuele Murtinu
Co- Assessor: Dr. Florian Noseleit
1 S3141942, f.m.brunner@student.rug.nl; fabian.max.brunner@gmail.com
2 University of Groningen, Faculty of Economics and Business, Nettelbosje 2, 9747 AE Groningen 3
Table of contents
1
Introduction ... 1
2
Research Questions ... 3
3
Literature Review and Development of Hypotheses ... 4
3.1 Agglomeration Economies and SME Growth ... 5
3.2 The Information, Communications and Technology Era... 7
3.3 Advantages of Economic Agglomerations on Firm- Level ... 8
3.3.1 Customer Proximity and Consumer Research ... 9
3.3.2 Information Externalities and Enhanced Access to Knowledge and Innovations ... 9
3.3.3 Cluster Identity, Reputation and Enhanced Legitimacy ... 9
3.3.4 Knowledge Spillovers ... 10
3.3.5 Access to specialized Labor ... 10
3.3.6 Infrastructure Benefits and Cost Advantages ... 11
3.3.7 Enhanced Access to Suppliers ... 11
3.3.8 Enhanced Position to build Social Networks ... 12
3.3.9 Heightened Competitor Awareness ... 12
3.3.10 Enhanced Inter-firm Cooperation ... 12
3.3.11 Increased Labor Productivity ... 13
3.4 Disadvantages of Economic Agglomerations on Firm- Level ... 14
3.4.1 Increased Competition in Input and Output Markets within Clusters ... 15
3.4.2 Congestion in Input and Output Markets within Clusters ... 15
3.5 The ICT, Media and Creative Industries Cluster Berlin ... 18
3.6 The ICT, Media and Creative Industries Cluster and its position in Germany ... 20
3.7 The moderating Effect of the salient Berlin ICT, Media & Creative Industries Cluster on clustered SME Growth in Germany ... 22
4.1 Sample & Data Collection Procedure ... 25
4.1.1 Measures: Dependent Variables ... 27
4.1.2 Measures: Independent Variables ... 28
4.1.3 Measures: Control Variables ... 28
4.2 Analysis of Quantitative Data & Results ... 29
4.2.1 Descriptive Statistics and independent samples t-test ... 31
4.2.2 Data Reduction: Principal Component Analysis ... 43
4.3 Analysis of Qualitative Data & Results ... 45
4.3.1 Customer Proximity & Consumer Research ... 46
4.3.2 Information Externalities & Enhanced Access to Knowledge and Innovations ... 47
4.3.3 Cluster Identity, Reputation and Enhanced Legitimacy ... 47
4.3.4 Knowledge Spillovers ... 47
4.3.5 Access to specialized Labor ... 48
4.3.6 Infrastructure Benefits & Cost Advantages ... 48
4.3.7 Enhanced access to suppliers ... 48
4.3.8 Enhanced Position to build Social Networks ... 49
4.3.9 Heightened Competitor Awareness ... 49
4.3.10 Enhanced Inter-firm Cooperation ... 50
4.3.11 Increased Labor Productivity ... 50
4.3.12 Increased Competition in Input and Output Markets within Clusters ... 50
4.3.13 Congestion in Input and Output Markets within Clusters ... 51
5
Discussion of Results ... 52
6
Conclusion, Contribution, and Managerial Implications ... 55
7
Limitations & Recommendations for Future Research ... 57
Acknowledgments ... 58
References ... 59
Executive Summary
Does geographical proximity impact SME growth, despite technological opportunities of the ICT era? Previous research on agglomeration economies has proposed that geographical proximity between SMEs and its business partners and competitors stand to benefit from the positive externalities that stem from geographical proximity. In contradiction to that, technological opportunities outshine any impact of geographical location on SME growth according to findings of researchers from the technologically- oriented literature field of digital networking in the ICT era. A study with 40 SMEs involved in the ICT, Media or Creative Industry in Germany has been conducted. Four hypotheses derived from previous economic agglomeration literature from an ecology organization perspective were tested if geographical proximity in the ICT era plays a role, and if and how 15 forms of advantages and four forms of disadvantages of geographical proximity are perceived to impact growth. By applying cluster theory to the Berlin ICT, Media and Creative Industries Cluster in Germany, it is further shown how a national salient industrial cluster moderates effects from geographical proximity. Firstly, an independent samples t- test was conducted to prove the data provided by clustered and non- clustered firms to be statistical significant. That followed, a principal component analysis was performed in order to identify which advantages and disadvantages of geographical proximity between an SME and competitors or partners are most applicable and crucial. Finally, a structured interview was conducted with three SME executive managers in order to gain an in-depth understanding how the particular 19 factors from the literature are perceived by the respondents to impact growth. Results of the study stress benefits for the individual company to locate within an economically dense agglomeration of business partners and competitors. For non- clustered SMEs however, geographical location hardly impacts the company in terms of growth. Findings of this thesis form a basis for further research and deliver implications for managers concerning the question whether to locate an SME within an industrial cluster and which advantages or disadvantages are expected to influence the firm´s growth. The high number of words is due to multiple forms of analysis in this thesis.
1 Introduction
The geographical proximity of related businesses in specific locations, so-called economic agglomerations, attracted the attention of scholars already in the 19th century. This concept of clusters of economic activity dates back to Alfred Marshall (1890), at the time of industrial revolution, who was the first scholar to investigate in the concentration of specialized trades with physical goods in particular locations. (Kukalis, 2010). He referred to the benefits on the performance of a firm within regional or local concentrations of trades with tangible goods, services or human capital as a triad of external economies: skilled labor, specialized related trades and specialized firms in different stages of the production cycle (Kukalis, 2010; Marshall, 1890). However, in the late 20th century and early 21st century, many scholars conducted research in the science field of agglomeration economies or cluster economies based on Marshall´s findings from 1890. (e.g. Enright, 1996; Krugman, 1991; Kukalis, 2010; Rosenfeld, 1997; Porter 1998a, 1998b, 1990). The work of Porter has defined the standard theme in the field of the cluster concept:
“A cluster is a geographically proximate group of interconnected companies and related institutions in a specific market, linked by interdependences in providing a related set of products and/ or services.” (Porter, 1998a, p. 197)
opportunities (Ashurst et al., 2012; Kim et al., 2011; Morgan-Thomas, 2016). Consequently, digital technologies represent a key concern for SME managers and policy makers (Jones et al., 2013). Standing in contradiction to one another, traditional cluster theory by Porter (1998a, 1998b, 1990) foregrounds the beneficial factors that geographical proximity brings on the performance of a firm, whereas network research of SMEs in the ICT era implies that physical geographical proximity loses importance regarding influencing a firm´s performance. (Ashurst et al., 2012; Jones et al., 2013 Kim et al., 2011; Kukalis, 2010; Morgan-Thomas, 2016; Porter 1998a, 1998b, 1990).
of an SME in the ICT era in an industry, which mainly deals with the exchange of services or digital, intangible products. By addressing the literature gap, I aim to contribute to the agglomeration and industry cluster research field, by empirically testing the importance of geographical proximity of an SME to its business partners and competitors in the digital era on growth.
2 Research Questions
Based on the identified literature gap and the traditional cluster theory (see Marshall, 1890; Porter, 1990, 1998a, 1998b), I have formulated one research question and four sub-research questions. The missing empirical evidence of the effect of geographical proximity on the growth of SMEs in the ICT era leads to my first research question:
Research Question: Does geographical proximity impact SME growth, despite
technological opportunities of the ICT era?
In addition to that, the question was divided into four sub-research questions:
Sub-Research Question 1: How do advantages derived from agglomeration economies literature influence SME growth in the digital era in the ICT, Media and Creative industries?
Sub-Research Question 2: How do disadvantages derived from agglomeration economies literature influence SME growth in the digital era in the ICT, Media and Creative industries?
Sub-Research Question 3: How do advantages and disadvantages derived from agglomeration economies literature occur?
Economic Affairs (2015), which is responsible for reporting current economical phenomena within the borders of Germany. Kukalis (2010) tested the effects on firm performance from the salient high-technology cluster in the USA, the Silicon Valley and findings of his study confirm the moderating effect of the cluster on performance. Hence, in this thesis, it shall be tested how the salient cluster in Germany moderates the advantages and disadvantages of geographical proximity:
Sub-Research Question 4: How does the specific Berlin Cluster for ICT, Media and
Creative industries moderate the impact of geographical proximity on SMEs growth? The terms “cluster”, “industrial cluster”, “agglomeration economy” and “economic agglomeration” are used as synonyms in this thesis as they have the same meaning in previous literature and thus in the context of this study (see Kukalis, 2010; Porter 1990, Rosenfeld, 1997). The term “competitor” concerns all lateral businesses in the industry that compete in the same output (e.g. sales of products) and input markets (e.g. skilled labor). The term “business partners” concerns suppliers and customers to an equal extent.
3 Literature Review and Development of Hypotheses
In recent years, a large body of literature related to industry clusters has been compiled by scholars (Kukalis, 2010; Marshall, 1890; Porter, 1990, 1998a, 1998b; Saxenian, 1994; Rosenfeld, 1997). Researchers have mainly focused on four prevalent research themes (Kukalis, 2010): the economic characteristics of cluster location; the competitive nature of environments within clusters; the origin, evolution, and development of clusters; and finally the role of clusters in regional development.
disadvantages are likely to occur:”Industrial economists have linked the propensity for innovation-based industries to geographically cluster to the stage of the industry life cycle and suggest that the generation of innovative activities (leading to better firm performance) varies substantially over the course of the industry life cycle” (Kukalis, 2010, p. 7). Advantages, which stem from being inside of a cluster, mainly occur in the early stages of an industry´s life cycle, in the introduction and growth stages. The disadvantages, however, such as increased competition and congestion concerning resources, lead to more dispersion of innovative activities and hence to a disadvantage for clustered businesses compared to other enterprises of the industry outside of the cluster (Audretsch & Feldman, 1996; Kukalis, 2010; Klepper & Graddy, 1990).
From an organization ecology perspective, scholars argued that industry clusters are growth enhancing for the individual firm, only in the early periods of cluster development, when resources are abundantly accessible. Subsequently, with a rising amount of firms forming a dense agglomeration of economic activity, a certain saturation point can be reached, which describes the phenomenon of too many firms competing for the same or similar resources within one limited geographical location. (Baum & Mezias, 1992; Hannan & Freeman, 1989; Lomi, 1995). From that follows the logic that agglomerating of firms from one or related industries leads to overcrowded locations with clustered firms experiencing congestion and competition in input and output markets (Swann & Prevezer, 1996). Hence, clustering also might have besides positive impacts, negative impacts on the performance of a company (see Kukalis, 2010; Pouder & St. John, 1996).
3.1 Agglomeration Economies and SME Growth
theoretical benefits of geographical firm grouping to cluster members compared to non- clustered players in the market. These are advantages derived from the proximate environment of the single firm, such as better access to suppliers and other valuable or scarce inputs, better access to knowledge and innovations, better position to build social networks or a better position to anticipate processes of proximate successful competitors. (Burt, 1992; Saxenian, 1990, 1994; Stuart & Sorenson: Tallman et. al 2004). Other researchers identified the propensity for innovative activity, enhanced inter-firm cooperation, knowledge spillover and resource sharing as benefits that individual firms within an economic agglomeration perceive. (Audretsch & Feldman, 1996; Klepper & Graddy, 1990; Klepper & Miller, 1995; Kukalis, 2010). Moreover, Porter (1998a) viewed clusters as “crucial self- reinforcing systems that strengthen the competitiveness of the cluster firms and consequently the cluster of the cluster itself” (Kukalis, 2010).
Krugman (1991) divided the location of firms in 2 categories: “Periphery” and “Core”. Periphery concerns companies, which are not situated within an economic agglomeration, whereas “Core” regions are geographical areas, which involve a larger number of companies being geographically close to one another. Within a cluster, a core area, as one firm becomes successful, qualified suppliers of inputs, which involve venture capital or skilled workers subsequently locate in the geographical area of economic activities to derive advantages through collaboration for their own purpose. This pool of suppliers promotes the entry for subsequent companies, making a certain location more attractive than other areas. (Pouder & St. John, 1996). The benefits derived by enterprises from economic agglomerations are called agglomeration economies as defined previously by several researchers (Kukalis, 2010, Rauch, 1993; Scott, 1992).
3.2 The Information, Communications and Technology Era
With regard to the specific characteristics of SMEs in the digital era, where digital opportunities could put the physical geographical location in the background, (e.g. Jones et al., 2013; Kim et al., 2011; Morgan-Thomas, 2016) the impact of specific clustering effects on SME growth in the ICT era remains widely unknown.
In this context, it must be differentiated between networks and clusters. According to Rosenfeld (1997: 10), clusters are a “geographically bounded concentration of independent businesses”. Networks, on the other hand, do not necessarily include geographical proximity (Powell, 1990; Rosenfeld, 1997, 2005). Rosenfeld refers clusters as “associative economy” (2005:5), which include vertical and horizontal networks to business partners as business-based interdependencies within a certain geographical area. Therefore, the following table summarizes the differences between a cluster and a network:
Networks Clusters
Networks allow firms access to specialized services at lower cost
Clusters attract needed specialized services to a region
Networks have restricted membership Clusters have open membership Networks are based on contractual
agreements
Clusters are based on social values that foster trust and encourage reciprocity Networks make it easier for more firms to
engage in complex business
Clusters generate demand for more firms with similar and related capabilities
Networks are based on cooperation Clusters take both cooperation and competition
Networks have common business goals Clusters have collective visions
Table 1: Differentiation of Clusters and Networks, Adopted from (Rosenfeld, 1997: 10; Oxborrow, 2012: 9)
transactions, whereas cluster theory from an organization ecology perspective implies advantages derived from the agglomeration of related businesses. Examples of such benefits are increased inter-firm cooperation and a better position to build social networks within a certain geographical area (Audretsch & Feldman, 1996; Burt, 1992;Saxenian, 1994; Stuart & Sorenson; Tallman et. al 2004)
Thus, based on previous research on economic agglomeration theory from an organization ecology perspective, the following hypotheses are tested on SMEs in the digital era:
H1: Geographical proximity between an SME and its business partners and competitors has an impact on SME growth, despite technological opportunities of the ICT era
H1a: Geographical proximity between an SME and its business partners and competitors strongly impacts SME growth in the ICT era for clustered companies H1b: Geographical proximity between an SME and its business partners and competitors hardly impacts SME growth in the ICT era for non-clustered companies Previous research classified effects of agglomeration in supply and demand side economic effects (Swann et. al 1998, Swann & Prevezer, 1996, Cook et al., 2001). However, in this study, it is more appropriate to differentiate in advantages and disadvantages of geographical proximity, as this classification directly aims at answering the research question and the according sub-questions of this thesis. In the next step, the particular advantages of economic agglomerations for the individual enterprise and subsequently the disadvantages are explained, based on previous literature.
3.3 Advantages of Economic Agglomerations on Firm- Level
many subcategories are overlapping (adopted from Pouder & St. John (1996) and Swann & Prevezer, 1996).
3.3.1 Customer Proximity and Consumer Research
According to Munnich et al. (2001) inside of a geographical agglomeration of increased economic activity within a certain industry, companies are not only closer to competitors and business partners but also in immediate proximity to customers. Increased customer proximity leads to decreased consumer research cost, as customers are likely to be close and their existence is to be seen as distinct as stated by Swann et al. (1998). As a consequence of that, clustered companies might achieve a higher level of trust with the customer than the non- clustered contestants in the market and industry. As a result, an increased ability to meet the customer´s needs and a thus competitive advantage regarding sales and subsequent growth occur. (Munnich et al., 2001; Porter 1998a, 1998; Swann et al., 1998).
3.3.2 Information Externalities and Enhanced Access to Knowledge and Innovations
Clustered firms might gather information from other institutions that non- clustered firms might not receive from stakeholders (Porter, 1998a, 1998b; Swann et al. 1998; Kuah, 2002). As a consequence of that, information asymmetries occur that concern every company from the same industry and market (Greenwald & Stiglitz, 1986). Different to knowledge spillovers, which involve benefits derived from interaction with private businesses or suppliers of an individual company, information externalities involve public entities. According to Porter (1990, 1998a), government policies affect the opportunities for upgrading clusters. The “government can motivate, facilitate, and provide incentives for collective action by the private sector” (Porter, 2000: 24). Resulting from those governmental actions, information externalities might positively impact growth due to information asymmetries, which provide a competitive advantage for clustered firms relative to their non- clustered contestants (see Greenwald & Stiglitz, 1986).
3.3.3 Cluster Identity, Reputation and Enhanced Legitimacy
identity and labels the firm from a cluster as special compared to outsiders. Rao et al. (2008) add that a particular location grants internal legitimacy, which consequently grants a firm from a certain region a competitive advantage over non- clustered contestants in the industry. As a specific cluster within a market and industry gains cognitive and sociopolitcal legitimacy, the cluster might attract suppliers of resources, such as financial capital (Aldrich & Fiol, 1994). Furthermore, Swann et al. (1996) describe a reputational effect of clusters, which enhances legitimacy of a company, which is located in a particular geographical area or industry cluster. 3.3.4 Knowledge Spillovers
As proclaimed in previous studies, a very significant benefit of industrial clustering or geographical proximity of competitors or business partners are knowledge spillovers. According to Griliches, (1992: 31) knowledge spillovers are defined as the act "working on similar things and hence benefitting much from each other's research." (Adopted from Feldman, 1999: 7). Jaffe (1986) found that a significant part of the total flow of spillovers that affect a firm's research productivity. originates from companies geographically close to the company itself. These findings show that businesses experience a positive influence from R&D efforts of other firms that are situated close to the enterprise (Feldman, 1999). Henderson (1994) findings show that effects concerning knowledge of geographical proximity are most important for high- tech industries. Hence, it is most appropriate that knowledge spillover effects occur within and high technology industry and less in scale economies, which require the exchange of heavy tangible goods between firms. Maskell (2001) concludes with consequences faced by incumbents of clusters regarding tacit knowledge belonging to categories of information externalities and knowledge spillovers. Maskell (2001) emphasizes the importance of asymmetrical innovation as one of the major advantages, which differentiates clustered firms compared to non- clustered firms as only clustered firms have access to tacit knowledge, which leads to competitive advantage. These information asymmetries are closely related to heightened competitor awareness described by Pouder & St. John (1996) and Porter (1990).
3.3.5 Access to specialized Labor
advantages. According to his thesis, players, who are geographically close to each other, share common resources, such as a pool of skilled workforce within one geographical area. Porter (1990) has developed a diamond model of cooperation and competition of firms, which are geographically close. The model claims that the quality and amount of industry, market and task specific labor might be higher within an economic agglomeration. In addition to that, Feldman and Florida (1994) add that ventures starting up in clusters would be more capable of attracting workers with the expertise that would enable the venture to pursue growth objectives. Hence, it can be stated that companies being situated with an industrial cluster benefit from this economic agglomeration regarding access to specialized labor.
3.3.6 Infrastructure Benefits and Cost Advantages
Based on Porter´s diamond theory from 1990, proximal rivals stimulate the creation of a local infrastructure and extended supply of skilled labor and resources (further adopted by Pouder & St. John, 1996). A group of local competitors works to create an industry's infrastructure and credibility, and further, they derive cost and time savings through their affiliation and proximity (Porter, 1990; Scott, 1989). However, because of recent technological and infrastructural improvements regarding transactions, “the mere existence of spatial proximity does not consistently result in better performance for cluster firms compared with non- clustered firms within the same industry” (Kukalis, 2010: 24). Therefore, the terrestrial environment of a company needs to be adapted to needs and chances of exploiting gaps in the market and industry in which the single firm is situated (Swann et al. 1998). Porter (2000) concludes that dense economic agglomerations of businesses involve institutions, which could benefit the individual firm in its performance and growth such as universities, think tanks, vocational training providers, standards-setting agencies, trade associations that are mostly present in concentrated locations.
3.3.7 Enhanced Access to Suppliers
manufacturers of complementary products or companies related by skills, technologies, or common inputs.”
Therefore, companies located in a consolidated environment of businesses from related industries, experience quicker and cheaper access to valuable and rare inputs according to previous research.
3.3.8 Enhanced Position to build Social Networks
Social interaction is according to Burt (1992) one of the main advantages incumbents of economic agglomerations experience. Information regarding customers and market conditions, the current state of technological possibilities and trends are shared by “informal socialization” as happened in the formative years in the Silicon Valley. (Kukalis, 2010; Saxenian, 1994). According to Saxenian´s (1994) findings, informal information is exchanged within a cluster. As a consequence of that, social networks arise, which provide a relative competitive advantage relative for incumbents of a cluster towards outsiders from non- clustered areas.
3.3.9 Heightened Competitor Awareness
According to Pouder and St. John (1996), the geographical proximity of competitors makes information more accessible to firms, because executives of the firms might be able to better scan the activities of local competitors.
“Frequent social and professional interactions, dependence on a common supplier base, and recruitment from a common, highly mobile professional labor pool would lead to a high level of information exchange among managers.” (Pouder & St. John, 1996, p. 1201).
This leads to heightened information exchange and further to subsequent awareness of aggressive actions in a proximate area within the industry and market. (Porter, 1990; Reger & Huff, 1993; Saxenian, 1994; Scott, 1989).
3.3.10 Enhanced Inter-firm Cooperation
competitive advantage from the economic agglomeration relative to non- clustered contestants in industry and market based on cost advantages through inter- firm cooperation (Kukalis, 2010).
3.3.11 Increased Labor Productivity
According to theory, a tremendous number of definitions concerning the term “labor productivity” could be identified. The OECD Manual (2012) uses a rather basic explanation by defining labor productivity or workforce productivity as the ratio of a volume measure of output to a volume measure of input. Porter (2000) claims that economies with low productivity are characterized by little local rivalry, which implies that the lower the number of firms competing for the same customers and inputs within one specific geographical area, the lower the productivity. However, Porter (1998a, 1998b, 2000) claims that a larger number of competitors and available business partners enhances productivity “not only through the acquisition and assembly of inputs but also through facilitating complementarities between the activities of cluster participants” (Porter 2000: 21). Therefore, it can be argued that labor productivity in geographical dense populated areas is positively influenced by a collaboration of companies based on the exchange of complementary services. Hence, increase labor productivity can be counted to advantages of geographical proximity also in the ICT era based on Marshalls theory from 1890, which applied to companies in the industrial revolution of the 19th century.
All these benefits combined, a positive impact of geographical proximity between an SME and its business partner or competitors is suggested:
H2: Advantages of geographical proximity between an SME and its business partners and competitors stimulates SME growth in the ICT era
H2a: Advantages of geographical proximity between an SME and its business partners and competitors strongly impact SME growth in the ICT era for clustered companies
H2b: Advantages of geographical proximity between an SME and its business partners and competitors hardly impact SME growth in the ICT era for non- clustered companies
In the next step, disadvantages of geographical proximity of competitors and business partners will be explained.
3.4 Disadvantages of Economic Agglomerations on Firm- Level
In contradiction to advantages of geographical clustering, scholars also identified negative factors of SME growth within a specific geographical area of economical activity. The empirical evidence of positive influence on firm growth is mixed. Findings of (Feldman, 1999) stress that being located inside of an industry cluster increases either growth or innovative activity. On the other hand, Swann et al., (1998) and Feldman and Audretsch (1996) discuss congestion and competition effects of economic agglomerations. Evolutionary agglomeration theory implies that the more mature the environment of an individual is in the industry life cycle, the smaller the positive factors of clustering and the larger the disadvantages. (Audretsch & Feldman, 1996; Burt, 1992; Saxenian, 1994; Stuart & Sorenson: Tallman et. al 2004 Klepper & Graddy, 1990; Klepper & Miller, 1995). These findings, viewed from an organization ecology perspective, imply that there are many firms competing for the same scarce resources, trying to reach high sales in a market with only a few demanding customers relative to suppliers (Hannan & Freeman, 1977, 1989) within a densely populated industrial area.
3.4.1 Increased Competition in Input and Output Markets within Clusters Increased competition within a cluster might hinder the clustered firms in achieving higher performance (Porter, 1990). This, in turn, leads to a competitive disadvantage for firms within the cluster to non- clustered contestants. Incumbent competitors restrain themselves in input and output markets and hence deliver a reason to stay outside of a cluster or leave a cluster after a certain period of time (Kukalis, 2010, Pouder & St. John, 1996). Kukalis (2010) describes competition among firms for scarce resources, such as skilled labor in input markets as one of the disadvantages of clustering. Additionally, Audretsch & Feldman (1996) add heightened competitive practices for financial resources (venture capital) to disadvantages for the individual clustered firm resulting from economic agglomeration in terms of input markets. Regarding output markets, scholars suggest that sales figures based on other companies hindering economic actions influence a company´s growth as a negative externality (Pandit et al., 2001). Hence, consequent growth might suffer from increased competition in a dense agglomeration of competing companies (Porter, 1998a, 1998b; Kukalis, 2010).
3.4.2 Congestion in Input and Output Markets within Clusters
In addition to heightened competition within a dense population of businesses involved in the same industry, high resource efficiency within the cluster can progress to a saturation point when there are too many firms competing for the same resources in the same location in input markets, such as skilled labor (Pandit et al. 2001).
the economic agglomeration. Over time, a firm might experience drawbacks of being inside of the economic agglomeration, when resources in output and input markets are scarce (Hannan & Freeman, 1977, 1989).
Opposing to positive factors on a clustered firm based on the theory by Marshall (1890) and further evolved by Porter (1990), the SME might experience disadvantages of being inside an economic agglomeration as cluster generated resource diseconomies and additional insular heightened competition impede performance and decelerate further subsequent growth (Kukalis, 2010; Pouder & St. John 1996).
To put it in a nutshell, and to differentiate congestion and competition from each other: congestion concerns the availability of resources (input markets) and diseconomies in sales (output markets), as there are insufficient buyers relative to a dense and abundant availability of sellers (Swann et al., 1998). Competition, however, concerns the negative influence of one firm on the other firm being active in the same market and geographical area by competing for the same resources and the same customers (Porter, 1990, 1998a, 1998b).
Hence,
H3: Disadvantages of geographical proximity between an SME and its business partner and competitors are restraints to SME growth in the ICT era
Additionally, based on industrial cluster theory from an organization ecology perspective (e.g. Hannan & Freeman, 1989; Marshall, 1890; Porter, 2000, Pouder & St. John, 1996), disadvantages are expected to impact clustered and non-clustered SMEs from the ICT, Media and Creative Industry differently. According to rising congestion and competition proportionally to a rising density of firms involved in the same industry and market, a strong impact derived from geographical proximity is expected to impact clustered firms. On the other hand, due to the absence of heightened competition and increased congestion within a certain location, disadvantages might only scarcely affect non- clustered SME growth.
H3b: Disadvantages of geographical proximity between an SME and its business partners and competitors hardly impact SME growth in the ICT era for non- clustered companies
Advantages of Geographical Proximity between Competitors & Partners
I. Customer proximity
II. Reduced Consumer research costs III. Information externalities
IV. Reputation
V. Knowledge spillovers
VI. Better Access to specialized labor VII. Infrastructure benefits
VIII. Better access to knowledge & innovations
IX. Better access to suppliers and other scarce resources X. Better position to build social networks
XI. Increased competitor awareness XII. Enhanced inter-firm cooperation XIII. Enhanced resource sharing XIV. Cost advantages
XV. Heightened labor productivity
Disadvantages of Geographical Proximity between Competitors & Partners
XVI. Competition in input markets XVII. Competition in output markets XVIII. Congestion in input markets XIX. Competition in input markets
Table 2: Factors of advantages and disadvantages of industrial clustering (Own Figure derived from economic agglomeration literature, 2017)
In the next step, the ICT, Media and Creative Industries Cluster in Berlin will be presented.
3.5 The ICT, Media and Creative Industries Cluster Berlin
From an organization ecology perspective, it is suggested that the more dense the population of competitors, customers and business partners from the same or similar industry is constituted of, the stronger factors from economic agglomeration theory influence performance (Baum & Mezias, 1992; Hannan & Freeman, 1989; Lomi, 1995).
providers of information and communications industry, as well as media and creative industries in the capital region of Germany (Innovationsstrategie der Länder Berlin und Brandenburg, 2014).
In 2014, approximately 36.000 companies, which achieved a total turnover of approximately 30 billion Euro, were located inside of the Berlin ICT, Media and Creative Industries Cluster. 275.000 workers or entrepreneurs were employed in the industry, which represents 21% of the total workforce in Berlin. (Senate Department for Economic Affairs, 2015).
The following table reports the number of companies and the turnover in 2014 and 2015 and further shows the clusters different submarkets:
Table 3: Branches inside of the Berlin ICT, Creative Industries and Media Cluster
Figure 1: Employment in the Berlin ICT, Media and Creative Industry Cluster (SWEB, 2016: p.12)
3.6 The ICT, Media and Creative Industries Cluster and its position in Germany
Figure 2: Funding volumes for high technology start- ups in 2015(EY, 2015, p. 8).
“Berlin is a leading location in Europe´s digital industry, with dynamic startups, innovative SMEs, and globally established players. The ICT industry is growing at an above average rate. For a good reason: The high- quality research landscape, the open innovation culture, the large number of qualified employees and the rich cultural scene are strong divers in the capital.” (SDEA, 2012, p.3).
Thus, based on volume funding and the salience in having a national pioneer position in terms of innovation and competitive advantage in Germany, the Berlin ICT, Media and Creative industries cluster is declared as the largest and densely populated industry cluster, which moderates advantages and disadvantages to have the most substantial impact on growth of the individual SME (adopted from Kukalis, 2010).
3.7 The moderating Effect of the salient Berlin ICT, Media & Creative Industries Cluster on clustered SME Growth in Germany
Based on previous literature, advantages of geographical proximity of related businesses are most evident in terms of a moderate amount of companies within one geographical area, which supports each other regarding decreasing costs by sharing resources, exchanging information or experiencing infrastructure or legitimacy benefits. (e.g. Porter, 2000). In contradiction to that, clustering might have a negative impact due to increased competition and the congestion effect on SME performance if there are too many firms competing in the same input markets for resources and output markets for sales of outcomes (Swann et al., 1998).
Hence, SMEs inside of a cluster would benefit from the effects of clustering to a certain extent, but being inside of a cluster would negatively influence performance in an overcrowded geographical area and consequently would provide a competitive advantage for contestants, which are located outside of the cluster (Kukalis, 2010).Adverse effects of clustering after a certain saturation point of resources would destroy the initial competitive advantage facing non-clustered contestants in the industry of clustered firms in a period of high competition for input and output resources, as resources are not available in abundance. (Klepper & Graddy, 1990; Klepper & Miller, 1995; Kukalis, 2010).
least one of the industries ICT, Media or Creative industries to flourish in terms of growth (InnoBB, 2014). However, it is not known in which stage the industry and the cluster currently are in the industry life cycle based on the model described by Menzel and Fomahl (2010). Hence, this study aims to tackle the issue from an organization ecology perspective. Regarding to Lomi (1995), it can be argued that congestion and competition might be most evident in the salient cluster of an industry within a particular economy, which is the Berlin ICT, Media and Creative Industries Cluster within the borders of Germany. (EY, 2015, 2016).
Thus, this study further tests if being located inside of the Berlin ICT, Media and Creative Industries Cluster has a moderating effect on the impact of geographical proximity on SME growth and further a moderating effect on the disadvantages and advantages derived from traditional cluster theory. The Berlin ICT, Media and Creative Industries cluster, on the one hand, is the largest and most dense agglomeration of competing and cooperating firms in Germany (EY, 2015), which implies that advantages might be most evident in this geographical area of the market and industry from an organization ecology perspective. However, on the other hand, disadvantages of clustering might also be most evident in this specific region, due to the congestion effect within the economic agglomeration and increased competitive practices of contestants for resources and sales in output markets.
Based on cluster theory, (e.g. Marshall, 1890; Porter, 1990, 1998a, 1998b) from an organization ecology perspective (Baum & Mezias, 1992; Hannan & Freeman, 1989; Lomi, 1995), and current information about the Berlin ICT, Media and Creative Industry Cluster within Germany (EY, 2015, 2016; InnoBB, 2014; SDEA, 2015, SWEB, 2016 ), I hypothesize a moderating effect of the salient cluster in the three industries ICT, media and creative industries on SME growth:
H4: The Berlin ICT, Media and Creative Industry Cluster positively moderates the impact of geographical proximity on SME growth in the ICT era
In addition to that, a stimulating effect on the extent of advantages and disadvantages on firm growth is expected:
H4b: The Berlin ICT, Media and Creative Industry strengthens the impact of geographical proximity´s disadvantages on SME growth in the ICT era
The following conceptual model visualizes the variables and hypotheses of this thesis:
4 Methodology
Van Aken, Berends and Van der Bij (2012) argue that the theory testing approach shall be used if the literature streams are already elaborated, which is the case of cluster theory and recent research about the ICT era. However, a literature gap in the theoretical explanations can be identified in the convergence of the two fields of literature on firm-level. Since this is the case for the importance of geographical proximity in the ICT era, a theory testing approach is conducted. The steps proposed by the authors with slight modification will be followed:
(1) Definition of the business phenomenon and identification of the literature gap (2) Generation of conceptual model and hypotheses
(3) Data Analysis
(4) Interpretation of results with the comparison to hypotheses, reaching conclusion and providing implications (van Aken et al., 2012)
4.1 Sample & Data Collection Procedure
By sending out a survey 43 companies could be reached. Using email addresses by SMEs in Germany, which were found via internet research or my personal network, I invited 534 companies to join the research by sending an email. Participants returned completed questionnaires directly to the researcher of this thesis, in order to ensure strict confidentiality. In total, 43 questionnaires were received in return, of which 40 were usable, resulting in a response rate of 7.5 percent. The database contains responses from SMEs, which operate in the ICT, Media or Creative industry in Germany. According to the definition by the European Commission (2013a, 2013b), SMEs can be categorized in three company categories: Medium-sized companies with less than 250 employees, small companies with less than 50 employees, micro enterprises with less than ten employees down to one employee. In this thesis, the three company categories are consolidated as SMEs in according to the definition by the European Commission (2013a, 2013b). Besides that, all firms, whose headquarter is located in Germany and involved in the ICT, Media or Creative Industry are entitled to take part in this study. Hence, the following database contains responses from SMEs, which operate mainly digitally inside or outside of the Start-Up Agglomeration in Berlin, Germany.
The data are collected from a firm-level perspective by individuals who hold CEO or executive managing positions, who responded as the representatives of their respective company. 13 of the studied 40 SMEs´ headquarters are located in Berlin. Ten SMEs are situated inside other economic agglomerations in Germany, and the remaining 17 SMEs are located outside of any cluster. The sample could be classified into three groups of respondents according to Krugman´s (1991) core- periphery pattern with the addition of one further category: Clustered, Non-clustered and Berlin-clustered SMEs being involved in at least one of the industries ICT, Media or Creative Industry. The collected data is treated confidential and the respondents were informed about the confidentiality of this study. The questionnaire is attached as Appendix II. Furthermore, a brochure, which was sent to all participants was attached to in the mails sent. The brochure, which is to be found as Appendix I of this thesis, ensured the participants confidentiality of the conducted research.
interview with open questions was conducted with three respondents from the original sample of 40 companies. It was aimed to increase the validity of the interview by structuring it according to Emans` (2004) eight steps. This part of the data collection was conducted through interviews with the executive managers of three companies from three different locations within an economy according to Krugman (1991): one SME situated in a periphery region, one from a core area of economic activity and one from the Berlin ICT, Media and Creative Industries Cluster. It needs to be added that this was not a random process, but one that was based on theoretical sampling in order to have several perspectives, which are replicable by other researchers (Eisenhardt, 1989). According to the replication logic by Eisenhardt (1989), a comparison across multiple cases has been searched to confirm particular emergent relationships between constructs and thereby improve the validity of the relationship. The main goal of these in-depth interviews was to gather data to answer the research questions of this study (Emans, 2004).
Three respondents were selected on the basis of their geographical location. One interviewee´s SME is situated in Berlin (Snap Track UG) and the second interviewee and its firm is located in Bamberg in Southern Germany (Rakete 7 GmbH), which represents a small industry cluster. The third interview is from a town in Southern Germany, named Kemnath (Passion & Process), which represents an SME from a periphery region according to Krugman´s (1991) classification. The interviewees agreed on recording the interview. The interviews were conducted via phone. However, the interviewees do not want to be contacted after this interview for further interviews concerning the topic by other researchers belonging to the Faculty of Business and Economics of the University of Groningen. Hence, the three respondents were assured of nobody approaching the interviewees after this interview for further questions. The interview guide with particular questions and answers can be found in the appendices of this thesis.
4.1.1 Measures: Dependent Variables
eases answering the questions of the survey (Murphy et al., 1996; McKelvie & Wiklund, 2010; Wiklund & Shepherd, 2005). In addition to that, growth is applicable to virtually all businesses and further increasing sales is seen as a necessity for the growth of other assets, such as employment, by the firm (Weinzimmer et al., 1998; Wiklund & Shepherd, 2005).
4.1.2 Measures: Independent Variables
Since the dependent variable is SME growth, the input variables that influence an SMEs growth are the derived 15 advantages and four disadvantages of geographical proximity on perceived growth. CEO´s or executive managers, who represented their SMEs answered the survey. To measure the perceived impact of geographical proximity of the SME on growth, a 7 point- Likert scale was used to measure the impact of physical geographical proximity to competitors and business partners on growth.
In a second step, the moderating impact of the Berlin ICT, Media and Creative industries cluster on advantages and disadvantages of geographical proximity on SME growth was tested. Again a 7 point- Likert scale was used. The first part consisting of 27 questions was answered by all participants, whereas the second part was only to be answered by firms, which are located inside of the Berlin ICT, Media and Creative industries cluster. The representatives were asked whether to agree or disagree with a strengthening moderating effect of the Berlin cluster in 19 questions consisting of 15 advantages and 4 disadvantages derived from the literature. (1 = Strongly Agree; 2 = Agree; 3 = Agree somewhat; 4 = Undecided; 5 = Disagree somewhat; 6 = Disagree; 7 = Strongly disagree).
Regarding the qualitative data collection and to be in line with the scheme used in the theoretical background of this thesis, the 19 factors of advantages and disadvantages of geographical clustering were asked to the respective respondent in a way that descriptive data regarding the “how” a particular aspect of geographical proximity impacts firm growth. (e.g. “How do knowledge spillovers from partners being geographically close that influence the growth of you occur?”).
4.1.3 Measures: Control Variables
industry and age of the company. Hence, it is ensured that companies from different size, age and industries are scattered within the three different types geographical locations. This, in turn, avoids biases and emphasizes that geographical location of the company as the crucial variable, which stresses that this variable causes SME growth and thus increases the reliability of the results (adopted from Kukalis, 2010). The following table visualizes the categories the respondents could identify in what industry they are involved in, how many workers they employ and how much time has passed since the establishment of the company:
Category Possibilities to indicate
Cluster or not Cluster 1 = “Yes” 2 = “No”
Age of the SME 1 = “Younger than one year” 2 = “Between one and three years” 3 = “Between three and five years” 4 = “Older than five years”
Industry 1 = “ICT”
2 = “Media”
3 = “Creative Industry”
4 = ”More than one of the mentioned industries”
5 = “None of the industries mentioned above”
Amount of employees 1 = “One”
2 = “Between one and ten” 3 = “Between one and 50” 4 = ”Between 50 and 250” 5 = “More than 250”
Table 4: Control Variables indication possibilities (Own table, 2017)
4.2 Analysis of Quantitative Data & Results
Correlations
Control Variables Amount of
employees
Age of the SME
Industry Cluster or not Cluster
-none-a Amount of employees Correlation 1,000 ,393 ,097 ,065
Significance (2-tailed) . ,022 ,587 ,717
Df 0 32 32 32
Age of the SME Correlation ,393 1,000 ,291 ,158
Significance (2-tailed) ,022 . ,095 ,371
Df 32 0 32 32
Industry Correlation ,097 ,291 1,000 ,306
Significance (2-tailed) ,587 ,095 . ,079
Df 32 32 0 32
Cluster or not Cluster Correlation ,065 ,158 ,306 1,000
Significance (2-tailed) ,717 ,371 ,079 .
Df 32 32 32 0
Cluster or not Cluster Amount of employees Correlation 1,000 ,388 ,081
Significance (2-tailed) . ,026 ,654
Df 0 31 31
Age of the SME Correlation ,388 1,000 ,258
Significance (2-tailed) ,026 . ,147
Df 31 0 31
Industry Correlation ,081 ,258 1,000
Significance (2-tailed) ,654 ,147 .
Df 31 31 0
a. Cells contain zero-order (Pearson) correlations.
There is no systematic difference between cluster and non-clustered companies in terms of the control variables.
4.2.1 Descriptive Statistics and independent samples t-test
Firstly, the sample was divided in two groups: Clustered and non- clustered SMEs. The 19 advantages and disadvantages perceived by the respondents in the questionnaire were accumulated, and the mean of the answer given by the respective sample group of N= 23 and N= 17 describe the perceived importance of geographical proximity regarding the particular factor. The 19 advantages and disadvantages are described as factors in this analysis.
Figure 4: Numerical comparison of means (Own figure, 2017)
The increasing vertical numbers on the left side of the diagram represent the means of answers given by the respective participant of the survey. (1 = Strongly Agree; 2 = Agree; 3 = Agree somewhat; 4 = Undecided; 5 = Disagree somewhat; 6 = Disagree; 7 = Strongly disagree). The following two tables report the means, standard deviations and standard errors of the means:
Group Statistics, Disadvantages
Cluster or not cluster N Mean Std. Deviation Std. Error Mean XVI. Congestion in Input
markets
Yes 23 2,52 1,410 ,294
No 17 4,94 1,819 ,441
XVII. Competition in output markets
Yes 23 2,91 1,905 ,397
No 17 5,76 ,903 ,219
XVIII. Congestion in output markets
Yes 23 2,52 1,648 ,344
No 17 5,29 1,448 ,351
XIX. Competition in input markets
Yes 23 2,61 1,469 ,306
No 17 4,63 1,500 ,375
Table 6: Group Statistics, Disadvantages (Own table, 2017)
0,00 1,00 2,00 3,00 4,00 5,00 6,00 7,00 I. II.
Group Statistics, Advantages Cluster or not cluster N Mean Std. Deviation Std. Error Mean
I. Customer Proximity Yes 23 2,35 1,229 ,256
No 17 4,65 1,693 ,411
II. Reduced Consumer Research
Yes 23 2,48 1,442 ,301
No 17 4,76 1,855 ,450
III. Information Externalities Yes 23 2,35 1,265 ,264
No 17 5,18 1,237 ,300
IV. Reputation Yes 23 2,30 1,105 ,230
No 17 5,18 1,380 ,335
V. Knowledge Spillovers Yes 23 2,30 ,926 ,193
No 17 4,41 1,873 ,454
VI. Better access to specialized labor
Yes 23 2,39 1,305 ,272
No 17 5,12 1,317 ,319
VII. Infrastructure benefits Yes 23 2,04 1,296 ,270
No 17 4,88 1,708 ,427
IX. Better access to suppliers and other scarce resources
Yes 23 2,61 1,373 ,286
No 17 4,47 1,772 ,430
VIII. Better access to knowledge & innovations
Yes 23 2,48 1,310 ,273
No 17 5,00 1,768 ,429
X. Better position to build social networks
Yes 23 2,39 1,500 ,313
No 17 4,82 1,551 ,376
XI. Increased competitor awareness
Yes 23 2,30 ,974 ,203
No 17 4,47 1,700 ,412
XII. Increased inter-firm cooperation
Yes 23 2,61 1,469 ,306
No 17 4,76 1,985 ,481
XIII. Enhanced resource sharing
Yes 23 3,13 1,546 ,322
No 17 4,88 1,408 ,352
XIV. Cost advantages Yes 23 2,74 1,389 ,290
No 17 4,94 1,436 ,359
XV. Heightened labor productivity
Yes 23 2,78 1,347 ,281
No 17 4,82 1,629 ,395
Table 7: Group Statistics, Advantages (Own table, 2017)
Secondly, the remaining four disadvantages of economic agglomerations were applied to the two independent groups of clustered and non- clustered SMEs.
The null hypothesis for this study is:
H
0: µ
clustered= µ
non-clusteredwhere µ = population mean and subscripts = the names of the different levels of the independent variable: "Clustered" and "Non- clustered". In words, this states that clustered and non- clustered SMEs share the same perceived importance of the respective aspect of geographical proximity on SME growth
The alternative hypothesis is:
H
A: µ
Clustered≠ µ
non-clusteredIn words, this states that the clustered firms and non- clustered firms mean scores are not equal in the total sample of N= 40, whereas clustered SMEs represent 23
companies and non- clustered SMEs represent 17 enterprises from the total population of this study.
N Number of cases (participants)
M Mean
SD Standard Deviation
SE Standard Error Difference
P p- value of respective aspect (Levene´s
test), sig. (2- tailed)
M Mean Difference of clustered and non-
clustered SMEs (t- test for Equality of Means)
t(df) t- value(degree of freedom)
p p- value (t- test for Equality of Means)
D Cohen´s d
Table 8: Description of codes of the independent samples t- test (Own table, 2017)
Regarding disadvantages of geographical proximity to business partners or competitors, there was homogeneity of variances, as assessed by Levene's test for equality of variances for the three factors XVIII. Congestion in output markets (P = 0.842), XIX. Competition in input markets (P = 0.207) and XIX. Competition in input markets (P = 0.652). If the population variance of both groups is equal, this test will return a p-value greater than 0.05 ,indicating that the assumption of homogeneity of variances is met. However, if the test returns a p-value less than 0.05 (P < .05), the population variances are unequal and assumption of homogeneity of variances are violated.
XVI. Congestion in input markets was more important to clustered firms (M= 2.52, SD = 1.41) than to non- clustered firms (M= 4.94, SD=1.82) and a statistically significant difference with, M = -2.419, SE= 0.510, t (38) = -4.742, p = .00003.
Further, Cohen's d is the appropriate effect size measure if two groups have similar standard deviations and are of similar size. Cohen´s d was calculated for factors, which show homogeneity of variances according to the following formula (Cohen, 1988) in order to evaluate the effect size:
&
As calculated for XVI. Congestion in input markets, d= 1.49, which implies a strong size effect between the two groups regarding congestion in input markets. According to Cohen (1988), Cohen´s d with values larger than 0.8 represent a strong size effect of one group compared to another group.
XVIII. Congestion in output markets was more important to clustered firms (M= 2.52, SD = 1.65) than to non- clustered firms (M= 5.92, SD=1.45) and a statistically significant difference with, M = -2.772, SE= 0.501, t (38) = -5.533, p = .00002, d= 1.81
The assumption of homogeneity of variances was violated, as assessed by Levene's test for equality of variances for XVII. Competition in output markets (P = 0.0047). Therefore, as equal variances are not assumed. XVII. Competition in output markets was also more important to clustered firms (M = 2.91, SD= 1.91) than to non- clustered firms (M= 5.76, SD=0.90) and a statistically significant difference with M= -2.852, SE= 0.5, t (38) = -5.703, p=0.000001, d= 2.1.
There was a statistically significant difference between means (p < .05), and thus, we can reject the null hypothesis and accept the alternative hypothesis for disadvantages of geographical proximity to business partner and competitors.
The same independent Samples test was accordingly applied to 15 advantages from economic agglomeration literature. Homogeneity of variances can be assumed for 13 out of 15 advantages. However with V. Knowledge Spillovers (P = 0.002 < 0.05) and XI. Increased competitor awareness (P = 0.012 < 0.05) equal variances cannot be assumed.
I. Customer Proximity was more important to clustered firms (M = 2.35, SD= 1.23) than to non- clustered firms (M= 4.65, SD= 1.69) and a statistically significant difference with M= -2.299, SE= 0.46, t (38) = -4.982, p=0.000014, d= 0.26.
II. Reduced Consumer Research was more important to clustered firms (M = 2.48, SD= 1.44) than to non- clustered firms (M= 4.76, SD= 1.86) and a statistically significant difference with M= -2.286, SE= 0.521, t (38) = -4.389, p= 0.000088, d= 1.40 .
III. Information Externalities was more important to clustered firms (M = 2.35, SD= 1.27) than to non- clustered firms (M= 5.18, SD= 1.24) and a statistically significant difference with M=-2.829, SE= 0.401, t (38) = -7.056, p= 0.0000092, d= 2.26. IV. Reputation was more important to clustered firms (M= 2.3, SD= 1.11) than to non- clustered firms (M= 5.18, SD= 1.38) and a statistically significant difference with M=-2.872, SE= 0.393, t (38) = -7.31, p= 0.0000083, d= 3.90.
significant difference with M= -2.726, SE= 0.419, t (38) = -6.505, p= 0.000022, d= 1.63.
VII. Infrastructure benefits was more important to clustered firms (M= 2.04, SD= 1.30) than to non- clustered firms (M= 4.88, SD= 1.71) and a statistically significant difference with M= -2.832, SE= 0.481, t (38) = -5.889, p= 0.000012, d= 1.87.
IX. Better access to suppliers and other scarce resources was more important to clustered firms (M= 2.61, SD= 1.37) than to non- clustered firms (M= 4.47, SD= 1.77) and a statistically significant difference with M = 1.862, SE= 0.497, t (38) = -3.747, p= 0.000593, d= 1.18.
VII. Better access to knowledge & innovations was more important to clustered firms (M= 2.48, SD= 1.31) than to non- clustered firms (M= 5.00, SD= 1.77) and a statistically significant difference with M= -2,522, SE= 0.486, t (38) = -5.189, p= 0.000032, d= 2.62.
X. Better position to build social networks was more important to clustered firms (M= 2.39, SD= 1.50) than to non- clustered firms (M= 4.82, SD= 1.55) and a statistically significant difference with M= -2.432, SE= 0.487, t (38) = -4.998, p= 0.000045, d= 1.59.
XII. Increased inter- firm cooperation was more important to clustered firms (M= 2.61, SD= 1.47) than to non- clustered firms (M= 4.76, SD= 1.99) and a statistically significant difference with M= -2.156, SE=0.546, t (38) = -3.952, p= 0.000027, d= 1.26.
XIII. Enhanced resource sharing was more important to clustered firms (M= 3.13, SD= 1.55) than to non- clustered firms (M= 4.88, SD= 1.41) and a statistically significant difference with M=-1.74, SE= 0.486, t (38) = -3.592, p= 0.00000015, d= 3.28.
significant difference with M= -2.041, SE= 0.471, t (38) = -4.334, p= 0.000043, d= 1.40.
The two factors, where equal variances are not assumed based on a Levene´s test for equality of variances within the aspect itself are V. Knowledge Spillovers and XI. Increased competitor awareness.
V. Knowledge Spillovers was more important to clustered firms (M= 2.30, SD= 0.93) than to non- clustered firms (M=4.41, SD= 1.87) and a statistically significant difference with M= -2.10, SE= 0.494, t (38) = -4.270, p= 0.000014, d= 1.43.
XI. Increased competitor awareness was more important to clustered firms (M = 2.30, SD= 0.974) than to non- clustered firms (M= 4.47, SD= 1.70) and a statistically significant difference with M= -2.16, SE=0.460, t (38) = -4.713, p= 0.000044, d= 0.31.
It was resigned in this thesis to generate bar charts or box plots for every aspect discussed, due to complexity of visualization.
Subsequently, the fourth hypothesis was tested on the 13 SME´s from the sample, which are located in Berlin targeting to indicate if the Berlin ICT, Media and Creative Industry Cluster has a moderating effect on SME growth with 1) to 19), which describes the responses of the 13 participating companies from Berlin how the Berlin ICT, Media and Creative Industries cluster stimulates or hampers the impact of factors from geographical proximity accordingly to the factors I. to XIX. on the respective company´s growth. It is not aimed to target if companies from Berlin perceive advantages and disadvantages more than other clustered firms, but to identify if this specific cluster has a fostering or hampering effect regarding the influence of geographical location on SME growth. The descriptive statistical data, gained from the final part might be useful for practicing firms to whether move to Berlin or not. The following data describe the means of indications perceived moderating effect of the Berlin Cluster on growth:
Figure 5: The moderating effect of the Berlin ICT, Media and Creative industries cluster
0 0,5 1 1,5 2 2,5 3 3,5 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) 14) 15) 16) 17) 18) 19) BERM ODER ATIO N Moderating effect on advantages
4.2.2 Data Reduction: Principal Component Analysis
Thirdly, a principal components analysis was run on the 19-question questionnaire to measure the most crucial advantages and disadvantages of geographical proximity on SME growth perceived by the respondents. A principal components analysis (PCA, for short) is a variable-reduction technique that shares many similarities to exploratory factor analysis. Its aim is to reduce a larger set of variables into a smaller set of 'artificial' variables (called principal components) that account for most of the variance in the original variables (Leard Statisitcs, 2015). It was resigned to divide the sample in clustered and non- clustered SMEs, as the sample would be too small to conduct a PCA. Thus, the principal component analysis was conducted with the total sample of N= 40. The suitability of PCA was assessed prior to analysis. Inspection of the correlation matrix showed that all variables had at least one correlation coefficient greater than 0.2. The overall Kaiser-Meyer-Olkin (KMO) measure was 0.901 with individual KMO measures all greater than 0.7, classifications of 'middling' to 'meritorious' according to Kaiser (1974). Bartlett's Test of Sphericity was statistically significant (p < .0005).
The PCA revealed two components that had eigenvalues greater than one and which explained 74.252% and 6.191% of the total variance, respectively. Visual inspection of the scree plot indicated that two components should be retained (Cattell, 1966). In addition, a two-component solution met the interpretability criterion. The two-component solution explained 80,443% of the total variance. Additionally, a Varimax orthogonal rotation was employed to aid interpretability. The rotated solution exhibited 'simple structure' (Leard Statistics, 2015; Thurstone, 1947). The interpretation of the data was consistent with the perceived attitudes towards the impact of geographical clustering on SME growth from the questionnaire. In the rotated component matrix, strong loadings exist on the two components. However, in this study, we emphasize the two strongest loading variables per component in order to make clear statements. “IX. Better access to knowledge and innovations” with a loading of ,903 and “XI. Increased competitor awareness” with a loading of ,845 on component 1 can be observed.
