Innovation Clusters

Table of Contents

List of abbreviations

List of figures

1. Introduction

2. Cluster and location theory
2.1 Defining clusters and cluster participants
2.2 Identifying regional clusters
2.3 Underlying theory
2.3.1 Classical location and agglomeration theories
2.3.2 Contemporary cluster theories
2.4 The ceramic tile cluster in accordance with Porter’s diamond model

3. Clusters as trigger for innovation
3.1 Enhanced innovative performance in clusters
3.1.1 Knowledge spillovers, inter-firm linkages and reduced risks
3.1.2 Innovation through advantages in capital, labor and firm structure
3.2 Cluster policy and practical recommendations
3.3 Failures due to retarded innovation and technological change

4. New approaches through technological advances and globalization
4.1 The death of distance
4.2 Virtual innovation networks
4.3 Globalizing vs. localizing: the location paradox

5. Conclusion and outlook

References

Abstract

Clusters are geographic concentrations of various industrial, scientific and governmental actors, and have been found to trigger and improve the innovative performance of firms inside. This paper gives a review of prevailing cluster theories, as well as several examples from the real economy. Knowledge spillovers, inter-firm linkages and reduced business risks for start-up firms present some of the advantages that foster firms’ innovative activity in clusters. The success of prominent clusters such as the Silicon Valley has encouraged governments to support the formation of clusters; however, technological changes might as well lead to failures of clusters. Despite the advances of a globalized economy, physical proximity in clusters has been shown to transmit input for innovation more successfully than virtual innovation networks.

Zusammenfassung

Cluster sind geografische Konzentrationen bestehend aus verschiedenen industriellen, wissenschaftlichen und staatlichen Akteuren, denen ein positiver Einfluss auf die Innovationstätigkeiten der ansässigen Unternehmen bescheinigt wird. Diese Arbeit gibt einen Überblick über herrschende Ansätze in der Literatur in Verbindung mit Beispielen aus der Realwirtschaft. Wissens-Spillover, Kooperationen und verminderte Risiken für Jungfirmen sind nur einige der Vorteile in Clustern, die die Innovationstätigkeit erhöhen. Der Erfolg von herausragenden Clustern wie das Silicon Valley hat nationale und regionale Regierungen veranlasst, verstärkt in den Aufbau von Clustern zu investieren; drastische technologische Veränderungen können aber auch zum Verfall von Clustern führen. Trotz der Vorteile der Globalisierung für die Wirtschaft, hat sich gezeigt, dass die physische Nähe in Clustern den Input für Innovationen besser übermittelt als virtuelle Innovations-Netzwerke.

List of abbreviations

illustration not visible in this excerpt

List of figures

Figure 1: cluster differentiation;

Figure 2: location quotient

Figure 3: reasons for geographic concentration;

Figure 4: diamond-model of competitive advantage

1. Introduction

Just off the MIT campus at 1 Broadway in Kendall Square, Cambridge, Massachusetts there is a building where start-up entrepreneurs, researchers and venture capitalists share the same offices, cafés and elevators (for whole passage Regalado, 2013). From a total of 9 floors over 600 start-up companies are responsible for a great amount of innovative activity in the U.S.: the Cambridge Innovation Center (CIC) has been starting point for corporations of which some have become big names such as “Amazon”, “Google”, “Twitter”, and “An-droid”. The CIC is an extreme example of an innovation cluster where the proximity of people, ideas and capital are concentrated in one single building. In a promising way the CIC is paving its way of becoming the next Silicon Valley – the textbook cluster in Califor-nia where companies have been pushing technology frontiers since 1960 (Wadhwa, 2013).

Clusters can be found in every advanced economy, with the result that literature on clusters has been overloading in the last two decades (Enright 2003, 99). Theorists from a range of disciplines, including international economics, geography and management studies, continue to investigate the synergistic effects of industrial clustering (Ellison and Glaeser 1997, 890). Firms in highly concentrated industrial areas not only show higher productivity growth rates because of benefits in their supply chains, but also reveal innovative performance that is superior in comparison to isolated firms (Porter 1998, 220). Frequent dialogue among the various cluster participants and fast information flows trigger and enhance innovation in clusters (Enright 1998, 318). Because of this, cluster firms often show up competitive advantages that lead them to dominate the world markets, and reason why the formation of so-called innovation clusters is on the top agenda of most national, federal, and regional governments (McCann and Sheppard 2003, 650).

Given the importance and relevance of the topic, this paper outlines the role that location and regional clusters play in firms’ innovative performance. Overviews on basic theories regarding industrial location and clustering will be given and connected with examples from the real economy. By analyzing beneficial conditions and advantages that come with geo-graphic proximity, the success of clustered firms in innovation will be elucidated. This will be further accentuated by showing up causes for cluster policies’ failures and decays of certain clusters. In the last part it will be reasoned if the new opportunities given through advances in information and communication technology (ICT) have made geographical proximity obsolete, or whether strong localization might has increased in importance in today’s highly connected economy.

2. Cluster and location theory

2.1 Defining clusters and cluster participants

Porter (2000, 15) defines clusters as “geographic concentrations of interconnected companies, specialized suppliers, service providers, firms in related industries, and associated institutions (…) in a particular field that compete but also cooperate”. Their geographic scope occurs at many levels: some clusters are just small districts inside or on the edge of a city; some are large regions that even cross federal borders (Porter 1998, 204). Several parties are involved in a functioning cluster: indirect and direct competitors, firms in related industries, and other institutions which include institutions of research, technical support or trade associations (for whole passage Porter 1998, 199; 204; 207). Usually firms inside clusters act in the same industry but do not directly compete because they focus on different segments. Still, they share the same needs and commonalities. Their firm size varies: in some clusters, like the Italian ceramic tile cluster, small and medium-sized firms are predominant, while others also compromise large firms, like the German chemical cluster. Firms in horizontally and vertically related industries, such as suppliers of intermediate goods, are also based in-side the cluster. In particular research institutions, such as universities or privately funded laboratories, play an essential role for the innovative activity in clusters. Successful clusters like the Silicon Valley (electronics and IT) and Route 128 in Massachusetts (life science) often have their roots in top-educational centers like Stanford University and the MIT.

The difference between industry clusters and industry networks is that networked firms usually are not located in close physical proximity but cooperate via communication technology (Enright 1998, 338). Also agglomerations are a “clustering of economic activity” (Fujita, Krugman and Venables 2001, 1). Some theorists, however, stress that clusters need to be distinguished from the simple co-location in agglomeration areas. In clusters business connections are more sophisticated and industry specialization is much higher than in agglomeration areas, as illustrated in the diagram below (Litzenberger and Sternberg 2005, 266).

illustration not visible in this excerpt

Figure 1: cluster differentiation;

own illustration based on Litzenberger and Sternberg (2005, 266).

Because of the international renown of the Silicon Valley and Route 128, clusters are often associated with high technology industries. However, clusters also occur in more traditional low-tech industries: examples include the carpet cluster in Dalton, Georgia, or the rubber industry on Akron, Ohio (Krugman 1991, 53). In the U.S. typically one cluster exists for one industry, for example the Detroit area for the automotive industry and Wall Street for financial ser-vices (for whole passage Fujita et al. 2001, 290). This type of concentration is called monocentric. In Europe the polycentric concentration is more common, meaning that firms from the same industry form more than one clustering area, such as London and Frankfurt as competing clusters for financial services. This difference is due to the borders among European countries, although growing European integration is likely to lead the European market toward an American-style monocentricity in the future.

2.2 Identifying regional clusters

Clusters can be identified and detected by the elements of the qualitative definition given above but also through applicable quantitative measures like local growth rates, the cluster index (Litzenberger and Sternberg 2005, 267), Krugman’s (1991, 55f) locational Gini coefficient or through the location quotient:

illustration not visible in this excerpt

Figure 2: location quotient

(Isserman 2000, 184; based on Davis 1990)

“The location quotient measures the relative importance of an industry in a place (…) by dividing the industry’s share of jobs locally by its share of jobs nationally” (Isserman 2000, 184f). In 2008 the location quotient for the computer equipment manufacturing industry in the Silicon Valley was 10.4 (U.S. Bureau of Labor Statistics, 2009), meaning that computer equipment manufacturing had an over ten-fold larger share of the regional economy than the rest of the national economy in the U.S. This indicates a high specialization of one industry in a certain area, and therefore a cluster. Surely one could argue that divergence and unevenness of economic activity has no specific reason and is mainly due to chance or trivial accidents. However, Ellison and Glaeser (1997, 907) tested highly concentrated industrial areas against chance and found that the economic structure of the U.S. is concentrated in a way that cannot be explained by chance. This makes it relevant and interesting for economic science to explain why industries concentrate and agglomerations and clusters emerge.

2.3 Underlying theory

2.3.1 Classical location and agglomeration theories

While it “might be clear why some clusters have developed near natural resources, (…) it is less clear why industries with limited dependence on such resources have located in particular places” (Enright 1998, 316f). Several theories have explored this phenomenon and pro-vide a wealth of explanations. The first seminal approach^[1] to explain regional industrial concentrations was made by Marshall (1890) who found three dominant reasons that en-courage firms to co-locate (for the following passages Hagemann, Christ, Rukwid and Erber 2011, 8ff):

Input allocation and lower transaction costs: Because of the proximity, input goods are more specialized and of higher quality. Furthermore, goods are available at market prices due to the competition among the locally concentrated suppliers.

Pooled labor markets: firms have access to highly-skilled workers, workers move to the area because they find new employment easily.

Information flows: through eased communication and internalization of positive externalities knowledge and information can be transmitted with less effort.

On basis of this several neoclassical frameworks were formed that mostly focused on trans-action cost models (McCann and Sheppard 2003, 650; Santos Crus and Teixeira 2007, 2).

2.3.2 Contemporary cluster theories

The internationalization of trade and the developments in ICT in the last decades of the 20th century lead to a renewal of interest in the subject of industrial location. The two most dominant authors who reintroduced the subject in the beginning of the 1990s were Michael Porter and Paul Krugman. Krugman (for whole passage 1991, 10; 14f; 20) formed the “new economic geography”, a new approach which explained the unevenness of spatial industrial activity. The approach focuses on increasing returns, transportation costs and demand that all lead to a cumulative process and a lock-in effect – an effect that makes a location change for the manufacturers inefficient.

In contrast to Krugman’s more formal approach, Porter (1990) was the first to popularize the cluster term and to connect location with innovative activity. Porter (1990, 73) sees the emergence of regional clusters as a consequence of the firms’ need to remain competitive. Clusters promote both competition and cooperation at the same time. Competition is a dynamic process that depends on productivity and innovation – and clusters, in comparison to an isolated location, provide an excellent milieu for both innovation and productivity growth (Porter 1998, 209; 220). Criticism of Porter’s cluster theory comes from theorists who state that the cluster definition is too vague and his assertions are not empirically demonstrated (Laperche, Sommers and Uzunidis 2010, 26).

A simple approach as to why clusters emerge was given by Ellison and Glaeser (for whole passage 1997, 891) who found that two factors reason for geographic concentration of industries: natural advantages are specific forces such as climatic conditions or access to specific resources that for instance make wine growers locate in California, whereas locational spillovers can be divided into physical spillovers (reduction of transportation costs) and intellectual spillovers (knowledge and information transfers).

illustration not visible in this excerpt

Figure 3: reasons for geographic concentration;

own illustration based on Ellison and Glaeser 1997, 891.

2.4 The ceramic tile cluster in accordance with Porter’s diamond model

Porter (1990; 2000) explains the clustering of industries via his diamond-model: the diamond involves four ingredients and conditions that lead a region to competitive advantages. In the following the model shall be clarified by looking at the practical example of the Italian ceramic tile cluster in Sassuolo.

illustration not visible in this excerpt

Figure 4: diamond-model of competitive advantage

(Porter 2000, 20).

The Italian ceramic tile industry represents a textbook cluster in the region of Sassuolo, Emilia-Romagna and has been evolving since the end of World War II (for the following passages Enright and Tenti 1990, 80f). Quickly Italian ceramic tiles became a good that accounted for 30% of world production and 60% of world exports. The success of the cluster lies in the cumulative process boosted by the different elements of the diamond:

Demand conditions: since the local Italian buyers’ market is one of the most sophisticated in the world, producers were always challenged to provide the best possible quality and newest designs. Because of this, producers inside the cluster were always one step ahead of competing foreign producers.

[...]

^[1] Other relevant works of classical location theory include Von Thünen (1826) with focus on the agricultural market, and the transaction cost based approach by Weber (1929).