118
BUSINESS CLUSTERING AND INNOVATION
ARIZONA STATE UNIVERSITY
ENT 305 - PRINCIPLES OF ENTREPRENEURSHIP
WEEK 4
9.1 INTRODUCTION:
The analysis of the geographical neighborhood approach developed from the work of
Marshall (1890) and Becattini (1979, 2002, 2006) who, instead of emphasizing industries -
industrial neighborhoods - emphasized industrial regions and clusters. Evidence of the interest
in geographical environment and cluster studies is shown by the many books published by
economists and sociologists, as well as business scholars (Becattini 2002; Porter 1990; Pyke
and Sengenberger 1992; Saksen 1994; Steiner 1998; Van Dijk and Rabellotti 1997; Weiss
1988), in addition to publications from national and international organizations on the subject
(Observatory of European SMEs 2002; OCED 1996, 1999, 2001; UNIDO 2001; World Bank
2000).
Despite their economic and strategic importance, it was not until the 1990s that
researchers made clusters a new research focus. The seminal work of Porter (1990) and
Krugman (1991) has motivated a growing number of academics to study the empirical
evidence on clusters, their definition, and impact on economic policy and business decision-
making (Sternberg and Litzenberger 2004).In recent years there have been many studies
analyzing the role of clusters in economic activity, both in developed countries, particularly in
high-tech sectors such as biotechnology and electronics, as well as in developing countries,
where clusters are proposed as a tool to improve the revenue and performance of a country's
competitiveness and as a bridge to achieve international positioning (Carlsson 2002).
The majority of research has attempted to link comparative advantage theory to firm
location (Audretsch 1998; Fujita et al. 1999). The underlying idea is that in many industries, it
is not the firm itself but the spatial environment that determines business competitiveness.
Therefore, the emphasis should be shifted from internal economies of scale to external
economies of scale that are local in nature. Regions, from this point of view, are very
important (Soler 2006). In this regard, recent research has essentially focused on studying the
conditions that favor the emergence of clusters in specific regions and countries (Khan and
119
Ghani 2004).
As for the theoretical approaches explaining the emergence of clusters, the most
widely used are economic geography (Krugman 1991), industrial organization theory (Porter
1990), transaction cost theory and innovation systems approach (Muizer and Hospers 2000).
The utilization of clusters has two objectives: on the one hand, to improve the competitiveness
of the small and medium-sized firms that belong to them, by exploiting the advantages
generated by business cooperation and economic agglomeration, and on the other hand, to
revitalize certain regions often witness how deeply rooted traditional industries in their region,
which were the drivers of regional development in the past, are gradually losing their
competitiveness (Mitxeo et al. 2004).
In this way, clusters are assumed to have a positive impact on firm performance, the
economic development of the affected regions, and the level of competitiveness of a country.
As a result, various institutions such as the OECD, UNIDO, World Bank, UNCTAD and the
European Commission, among others, propose the use of clusters as an economic
development tool (Enright and Flowcs-Williams 2001). A review of the literature shows that
clusters and related issues have not been sufficiently analyzed; studies in this field are very
ambiguous in defining what is meant by "cluster" and lack specificity in identifying its main
factors, characteristics and impacts. To achieve these objectives, we first demarcate clusters
starting from the evolution of the concept and its basic characteristics. We then analyze the
relationship between clusters and innovation, directly and indirectly, through their effect on
productivity, as the basic key factor of this type of geographic grouping. Finally, the paper
ends with a brief discussion of the subject.
9.2 CLUSTER CONSTRAINTS: DEFINITIONS AND CHARACTERISTICS:
One of the first things noticed by any researcher embarking on this study, after
conducting a general review of the literature, is the tremendous ambiguity regarding the
definition of what is understood by clusters and the lack of specificity regarding key factors,
characteristics and effects. In this regard, we consider it necessary, before discussing the
relationship between clusters and innovation, to make a theoretical effort to delimit what is
understood by clusters in this paper after assessing the main studies of the subject.
A Brief Journey of Cluster Research:
Marshall's important work on industrial parks was the launching pad for most of the
proposed theories on clusters. The importance of agglomeration economies "can often be
secured by the concentration of many small businesses of a similar character in a given area"
120
(Marshall 1966, p. 230). These economies, which are external to the firm but inherent to the
geographical area in which the firm is located, increase the efficiency of each firm (Rocha
2004).
However, there was a 50-year gap between Marshall's research and the resurgence of
interest in clusters that occurred in the 1970s. This lack of interest can be explained by the
large number of vertically integrated firms between the 1920s and 1960s that utilized
internally generated economies of scale to produce standardized goods for predictable markets
(Amin 2000, p. 149).
In the late seventies, early eighties, interest in industrial parks re-emerged. This was
due to the fact that since the mid-seventies, major economic, technological, institutional and
political changes had taken place that greatly affected the existing model of industrial
organization - the production system. This interest in industrial zones has resulted in a large
body of literature that revolves around three main schools: The Italian stream (Becattini 1979,
1989; Brusco 1992; Pyke and Sengenberger 1992), the Institutional school or flexible
specialization (Piore and Sabel 1984; Sabel and Zeitlin 1985) and the California School (Scott
1988; Storper and Scott 1989).
In the late nineties, there were two contextual factors that enhanced the
importance of clusters: processes of globalization and radical technological change (Rocha
2004). In this context, the literature on clusters is divided into two streams: the economic
stream (Krugman 1991; Porter 1990, 1998, 2001), which highlights the economic externalities
mentioned by Marshall; and the socio-economic and innovative trend, which highlights the
territorial, social, institutional and cultural factors underlying cluster dynamics and is called
the network paradigm (Conti et al. 1995; Powell 1990). This latter stream includes the
innovative neighborhood school (Camagni 1991; Maillat 1996), the Nordic School of
innovation and learning (Lundvall and Maskell 2000; Malmberg and Maskell 1997) and the
geography of innovation approach (Audretsch and Feldman 1996; Zucker et al. 1998a,b).
What is a Cluster?
The word "cluster" means different things to researchers in academia, and even to
politicians (Feser and Bergman 2000). However, the evolution of the cluster phenomenon
shows that clusters have three basic dimensions: geographical proximity, networks between
firms, and networks with organisms and institutions (Rocha 2004).
With this in mind, the most widely accepted definition today is that of Porter (1998):
"a cluster is a group of interconnected firms and related institutions in related industries that
121
are geographically close together". This definition covers an important dimension of clusters
and allows them to be clearly distinguished from other phenomena (Rocha 2004). Thus, many
authors later used Porter's definition in their papers (Carlsson 2002; Iturrioz et al. 2005; Khan
and Ghani 2004; Mitxeo et al. 2004; Rocha 2002, 2004; Rocha and Sternberg 2005; among
others).
The geographical dimension refers to the proximity of firm locations and is the only
cluster dimension included in most quantitative studies (Baptista and Swann 1998). The
geographical aspect is key in defining a cluster. Clusters represent a special case of a
geographically concentrated network of firms1 (Khan and Ghani 2004). The dimension of
inter-firm networks refers to the relationships that exist between firms in the cluster. Inter-
firm networks refer to market-based transactions and informal or non-traded relationships
(Storper 1997) between firms in a cluster. Traded interdependence is the production and
commercial relationship measured by input-output tables and is a key dimension for defining
sectoral clusters (Porter 1990). Non-traded interdependence "takes the form of conventions,
informal rules, and customs that coordinate economic actors under conditions of uncertainty"
(Storper 1997, p. 5).
Finally, the third dimension, institutional networks, refers to the relationships between
firms, non-governmental organizations, and government within the cluster (Aydalot 1986;
Becattini 1979; Saxenian 1994). The institutional network dimension of the cluster includes
both formal and informal relationships. Given the nature of institutional networks public
interest, it is closely linked to the concepts of social capital (Coleman 1990), institutional
embeddedness (Van de Ven 1993) and second and third order networks (Johannisson et al.
2002). In this way, clusters are characterized as a collection of tangible (firms and
infrastructure) and intangible (knowledge, technology, expertise) assets; and institutional
elements such as public administration and training and research centres, which act
interconnected in geographic space.
Clusters therefore represent a new form of spatial organization between markets, on
the one hand, and hierarchies or vertical integration, on the other. Clusters are therefore an
alternative way to organize value chains. Compared to market transactions between buyers
and sellers, the geographical proximity of firms and institutions and the multiplicity of
exchanges between them promote greater coordination and trust. As such, clusters reduce the
problems inherent in relationships without imposing inflexible vertical integration or
management challenges in creating and maintaining formal relationships such as networks or
partnerships. A cluster represents a powerful organizational form that provides advantages in
122
terms of efficiency, effectiveness and flexibility (Porter 1998).
In the earliest literature on this subject, clusters were defined as Industries that are
linked through formal production linkages, regardless of geographical proximity. If the cluster
does not exhibit a high degree of geographic concentration then it is referred to as an
industrial complex (Czamanski and Ablas 1979). Other authors emphasize geographic
concentration and define it as a group of firms geographically located close to each other that
essentially produce or provide the same product or service (Marshall 1890; Arthur 1990).
Porter (1990), for his part, includes in his definition the fact that there are related industries in
the cluster. He also emphasizes the fact that such clusters tend to be located in one space
(Porter 1998). Also very relevant to the definition is the reference to linkages between firms
(Becattini 1989) and with institutions located in the same geographical area (Saxenian 1994).
The lack of consensus in defining clusters means that in some studies, clusters have
been misinterpreted as synonyms for other terms. Specifically, it should be noted that clusters
are often equated with industrial agglomeration, both in theoretical (Glassman and Voelzkow
2001) and empirical studies (Baptista and Swann 1998). Industrial agglomeration is the
concentration of firms from the same industry that enjoy agglomeration economies or external
location economies. This concept originated with Marshall (1920). The basis of agglomeration
economies is the relationship between firms, institutions and other economic agents, which
are located in geographical proximity, yielding advantages in scale and scope (Ivarsson 1999).
However, clusters are more than just industrial agglomerations as industrial agglomerations
are clusters without networks (Rocha and Sternberg 2005).
This is also the case in many studies on industrial estates. The lack of clarity and
specificity in the definition causes clusters and industrial estates to be considered the same,
when in fact they are not. Marshall (1920) proposed the term "industrial estate" to describe the
grouping of small firms with similar characteristics within a region, which seek to increase
productivity as a consequence of the division of labor between them.
Another example of possible confusion with cluster definitions is with Rabellotti
(1995). He proposes four facts to identify an industrial area: (1) a cluster of spatially
concentrated and sectorally specialized small and medium-sized enterprises; (2) a strong and
relatively homogeneous cultural and social background that connects economic actors and
creates common and widely accepted, sometimes explicit but often implicit codes of behavior;
(3) a strong set of backward, forward, horizontal and employment linkages, based on the
exchange of goods, services, information and people both in markets and outside markets; (4)
a network of local public and private institutions that support economic actors in the cluster.
123
To properly define the term, it should be noted that an industrial park is simply a cluster of
small and medium-sized manufacturing firms focused on a dominant type of production
(Becattini 1979, 2006). Therefore, all industrial parks are clusters, but not all clusters are
industrial parks; it is the distinctive social and organizational characteristics of these areas that
distinguish them from a cluster.
Finally, it should be stated that clusters can be classified according to three specific
criteria: the type of cooperation established between cluster members (regional clusters,
industrial clusters, regional innovation networks) (Sternberg and Litzenberger 2004), the type
of activity (clusters based on production value chains and clusters based on other types of
linkages) (Iturrioz et al., 2005) and their origin (spontaneous clusters - informal, organized
and innovative - and constructed clusters) (Mitxeo et al.2004; Mytelka and Farinelli 2000).
Cluster Key Factors:
A review of the literature presents a long list of key factors in the emergence of
clusters: economies of scale and scope, transportation costs (input and output), transaction and
procurement costs, availability of production factors and/or components within a firm-specific
location, abundant knowledge, information and technology, development of innovations,
cooperation between firms or between suppliers and buyers and reduction of uncertainty
(Baptista and Swann 1998; Krugman 1991; Muizer and Hospers 2000; Nelson 1999; Porter
1990).
For Krugman (1991), Marshall (1920) is the main source of the emergence of clusters,
viz: it supports the emergence of specialist labor markets (Almazan et al. 2006; Costa and
Kahn 2001; Diamond and Simon 1990), allows the supply of industry-specific inputs with
greater variety and lower costs (Holmes 1999), and generates technological effects (Jaffe et al.
1993). Krugman (1991) proposes as key factors for clusters the existence of economies of
scale and scope, transportation costs and mobile factors of production.2 A consequence of
labor pooling is that all workers and firms are in the same geographical area. The general idea
is that both firms and workers prefer to be located in clusters, as firms can more easily expand
when they are in the same geographical area are in a market with a large pool of trained labor,
and workers can more easily find new jobs. Firms in this cluster benefit from reduced
sourcing and recruitment costs and from a quality workforce that is also easily available.
Individuals with the necessary skills are attracted to this cluster because of the benefits of
more job opportunities and the reduced risk of relocation if they change jobs. Feldman (1994)
also explicitly asserts that the main factors that lead to the emergence of clusters are the
124
presence of specialized labour and knowledge.
This labor pooling argument can also be extended to the field of capital, which is
It can also be reallocated more efficiently when industry players are geographically close
together. More specifically, locating in a cluster can make it easier for companies to acquire
capital assets when they decide to expand, or more generally to acquire or merge with other
companies to adjust the scale and scope of their business.
Porter (1998) highlights the positive impact of economic agglomeration as an
explanatory factor for the existence of clusters. In particular, he proposed key factors: shared
infrastructure, communication technology and access to input and output markets. In addition,
he also mentions the role of transportation costs (inputs) in the existence of clusters, although
such transportation costs are different from those mentioned by Krugman (essentially output
costs). Although there are some exceptions, many of the factors Krugman uses to explain
clusters relate to firms' input demand, whereas Porter's factors mainly refer to supply, as he
refers to the production process.
Baptista and Swann (1998) and Carlsson (2002) explicitly distinguish between
demand and supply factors. On the demand side, they distinguish four factors: clusters can
emerge in places with strong demand, firms can win market share by being closer to
competitors, the existence of clusters allows for a reduction in client search costs, and firms
located close to them can easily utilize important information provided by the latter. On the
supply side, they cite three factors originally proposed by Marshall: labor mix, availability of
specialized inputs, and technological impact.
Clusters can generate high levels of technological spillover and innovation (Krugman
1991). This is because, first of all, geographical proximity facilitates information flows. All
the cluster characteristics mentioned earlier play an important role in fostering technological
innovation within the cluster. A highly confident environment and easy access to specialist
suppliers increase the number of transactions which, in turn, leads to increased exchange of
technical and technological knowledge among firms in the cluster.
While firms in a cluster often cooperate with suppliers and distributors as they are
external sources of innovation and information, it is interesting to note that little attention is
paid to cooperation as a key factor in clusters. To the extent that some studies show that
clusters emerge close to knowledge institutions (universities, etc.), cooperation with these
institutions is also important, as they serve as a means to disseminate research, provide
services and educate and train future workers.
Khan and Ghani (2004) also highlight trust in clusters as a concept related to social
125
capital. Nahapiet and Ghoshal (1998, p. 243) define social capital as "the sum of actual and
potential resources embedded in, and available through, and derived from networks and
relationships held by individuals or social units". Rousseau et al. (1998) identified
consequences of trust that include a reduction in costly conflicts, decreased transaction costs,
and increased effective response to crises. Adler and Kwon (2002) summarize a series of
studies showing how social capital facilitates inter-firm resource exchange, product
innovation, entrepreneurship, intellectual capital creation, supplier relationships, and regional
production networks (clusters).
Nahapiet and Ghoshal (1998) discuss how networks with high levels of social capital
encourage firms to combine and exchange knowledge. This reduces the amount of time and
investment required to gather information. Saxenian (1994) explains how the high trust
environment in Silicon Valley resulted in the development of a technical community, which
accelerated the diffusion of technological capabilities and knowledge in the region. Krugman
(1991) notes that while most evidence of technology effects is reported for high-tech groups,
similar effects also occur for low-tech groups.
So, while clusters have many key factors, there are four that are particularly important
important: the availability of specialized labor, the presence of dispersed knowledge (e.g., due
to the proximity of universities), the degree of competition and cooperation between firms and
institutions. However, the roots of a cluster are often related to historical circumstances.
Clusters can also emerge due to unusual and sophisticated local demand; and the presence of
supplier industries, related industries, or even entire clusters related to previous clusters
provide new seeds for new clusters. New clusters can also arise due to one or two innovative
firms that stimulate the growth of many other firms (Porter 1998).
In conclusion, it must be shown that government policies can
Institutional support plays an important role in cluster development, and can also play a role
in technology transfer between clusters (Khan and Ghani 2004). Institutional support for
training, funding, and maintenance of new technologies can help reduce the risk of adopting
new technologies (Schmitz 1995a,b). Krugman (1991) argues that local chambers of
commerce and city councils can attract business-less firms to become self-sustaining when a
large number of people have been attracted. On the other hand, faulty regional policies, often
designed to help regions that have experienced difficulties, often end up fragmenting scarce
human and capital resources, thus undermining cluster development (Pouder and St. John
1996).
Government policy plays a dominant role in determining the rules of the game,
126
It is important to design policies that encourage positive innovation rather than negative
innovation. Porter (1998) states that governments, particularly in developing countries, can
play an important role in the development of well-functioning industrial clusters that
encourage positive innovation. Sometimes, government policies unwittingly work against
cluster formation such as restrictions on industrial location and subsidies to invest in affected
areas.
In the case of Spain, Navarro (2003) identifies the following components of
government cluster policy: the provision of certain public assets, such as training,
infrastructure, research, information..., which are specifically adapted to the needs of each
cluster and which the market cannot properly provide; support for the cooperation and
network functioning of its members (firms with firms or with research providers);
construction of a "community" (creation of forums, communication mechanisms, visible
concentration, creation or branding); help compensate for weaknesses or imbalances that may
exist in some parts of the cluster.
9.3 CLUSTERS, PRODUCTIVITY AND BUSINESS INNOVATION:
Most of the effects of being part of a cluster usually come from the cluster's key
factors. Many authors focus on studying cluster key factors but do not explicitly analyze their
impacts. A review of the literature reveals that the main positive impacts of clusters are the
acceleration of economic and technological growth and the development of educational
structures, the enhancement of corporate competitive advantage reflected in increased
productivity and firm performance as well as export levels, and the enhancement of
cooperation and the creation and dissemination of knowledge, and the incentive for firms to
be adventurous. Of the possible impacts brought about by these, in this study we focus on the
impact of clusters on innovation, directly and indirectly, and their impact on firm productivity
and efficiency, variables that explain firm performance and investment growth.
Clusters and Productivity Improvement:
First, the classic consequence of clusters is increased productivity. This increase
occurs as a consequence of greater access to inputs, information, technology and institutions.
Placing a firm in a cluster allows for better coordination with other firms, so that benefits can
be derived from the complementary activities of firms and from an increase in their
negotiating power. That is, supply-side advantages (Prevenzer 1997).
With regard to access to inputs, the ease of accessing workers with specialized skills,
127
both technical and scientific, is noteworthy. Both workers with specialized skills seeking
employment and firms seeking such workers gain access to the general labor market in the
same location (Krugman 1991; Marshall 1920). On the one hand, employees benefit from
access to a new supply of skilled labor from different firms or research centers and
universities. On the other hand, firms have access to available, experienced and skilled
workers without incurring search and transaction costs to recruit them. In this sense, clusters
provide opportunities and reduce relocation risks for employees. In fact, it can be an incentive
to attract talented people from other locations, which is a big advantage in some industries.
The workforce in a cluster is also specialized because they are part of a network and have
knowledge of people and their work. Specialized knowledge of human networks within a
technical or business community is invaluable, especially for new companies. Employee
networks are a mechanism for information exchange.
Companies belonging to a cluster also gain favorable access to other types of
specialized inputs such as information, teams, research instruments, related technologies that
need to be developed and designed for specific new markets. Having a pool of these inputs
nearby will give the company an advantage. Being supplied locally rather than having to use a
distant supplier will reduce transaction costs, in addition to minimizing the need to hold stock,
eliminating costs and delays, and, since it is important to enjoy a good reputation locally,
reducing the risk of the supplier overcharging or not fulfilling its commitments.
In addition, proximity will improve communication making it more likely that
suppliers will provide assistance or support services such as installation of supply ponds and
waste treatment. Another equally important aspect is that local outsourcing is better than
remote outsourcing, especially when the specialized inputs required include technology,
information and services (Porter 1998).
With respect to access to public institutions and assets, investments made by the
government or public administration can increase firm productivity. But private sector
productivity does not only increase because the government provides public assets.
Investments made by firms - training programs, infrastructure, quality centres, laboratories,
etc. also contribute to increased productivity. Such private investments are often made
collectively as cluster participants realize their potential collective benefits. Finally, with
respect to inter-firm complementarities, the greater the number of inter-firm linkages, the
greater the combined value of the cluster relative to the sum of its parts. Since the members of
a cluster are interdependent, the good performance of one firm can drive the success of other
firms.
128
Cluster and Innovation Improvement:
Clusters play a fundamental role in a firm's ability to innovate. Companies in a cluster
have an open window to the market, better than their isolated competitors. Clusters also
provide the capacity and flexibility to act faster. A firm in a cluster can obtain what it needs
more quickly, from its suppliers and partners, so that it can implement its innovations and thus
better meet the needs of its clients (Porter 1998).
Porter (1998) identifies a series of factors that support in-group innovation. The first is
the presence of a sophisticated buying group that is a key source of information on emerging
technologies and markets. Second, cluster firms, thanks to their easy access to a wide range of
specialist suppliers, are highly flexible and able to implement innovations quickly. In addition,
they can experiment with new innovations at lower costs and withhold the necessary
investment until the potential market reaction to the innovation can be verified. Finally, the
high level of competition and pressure that exists among competing firms in this cluster is a
major stimulus to innovate.
In addition, cluster characteristics also play an important role in technological
innovation within the cluster. A high trust environment with easy access to specialist suppliers
increases the number of transactions, resulting in increased exchange of technical knowledge
between cluster firms (Khan and Ghani 2004).
In this case, clusters increase the amount of innovation based on tacit knowledge and
capacity to learn, they generate greater knowledge spillovers and knowledge accumulation
and increase the speed of knowledge diffusion between firms (Krugman 1991). This is mainly
due to close proximity as information flows more easily than over long distances. Based on
the theoretical literature on clusters, clusters should increase the number of innovations,
generate greater knowledge spillovers and knowledge accumulation, and increase the speed of
knowledge diffusion between firms. Since firms are located close to each other, they can
cooperate relatively easily to develop innovations, inform each other of changes in input
specifications, launch new products and detect changes in technology and demand.
In particular, empirical studies have analyzed how clusters enhance innovation,
support small firms, and the development of radically new products or technologies. These
findings confirm the existence of a positive relationship between clusters and innovation
(Baptista and Swann 1998), and between clusters and technological impact (Audretsch and
Feldman 1996; Jaffe et al., 1993).
When cluster phenomena occur in sectors with high technological intensity (e.g.
biotechnology, information technology, new materials) (Bouwman and Hulsink 2002), they
129
are called "technopoles" (Castells and Hall 1994) or "technopolises" (Smilor et al., 1988).
This refers to the "deliberate efforts to plan and promote technologically innovative and
industry-related production in one concentrated area" (Castells and Hall 1994, p. 8). In this
regard, efforts have been made to create and develop technopoles around the world. Such
policies have three objectives: developing new industries through national policies,
regenerating areas of decline or stagnation, and developing an environment of innovation.
These objectives are achieved through increased collaboration between leading research
universities, corporate laboratories, core companies with their subcontractors and spin-offs,
and venture capitalists. Another related concept relevant to understanding dynamic techno-
industrial regions is the "innovation environment", defined by Castells and Hall (1994, p. 9)
as "the social, institutional, organizational, economic, and territorial structures that create the
conditions for the continued creation of synergies, (...) both for the production units that are
part of the environment and for the environment as a whole".
Other empirical studies generally confirm this as well, although most use more indirect
means. In some cases, this can improve the analysis by indicating the nature of the
relationship between innovation and clusters or by indicating the factors that contribute to this
relationship. This explanation comes from the literature on local technology externalities,
sometimes referred to as geographical effects, which analyzes the relationship between
clusters and innovation through technology effects.
In this regard, interest is focused on studying the extent to which research and
innovation are characterized by spatial concentration, and looking for the causes of this
concentration, particularly the localized nature of knowledge transmission. Two types of
approaches can be distinguished in this regard (Feldman 1994): the first approach emphasizes
the coincidence between local growth phenomena and the presence of technological
externalities, while the second approach measures the geographical dimension of runoff
effects.
Jaffe et al. (1993) analyzed the effect of location on innovation impact by comparing
the geographical location of patents. Their findings suggest that clusters generate more
innovation due to local effects. Audretsch and Feldman (1996) built a model that explains the
geographic distribution of innovation in each industry through the geographic distribution of
production, research and development, skilled labor, and university research.
There are also studies that analyze the nature of cluster firms by distinguishing
between the impact of knowledge dissemination on large and small firms. Arc et al. (1993)
assert that the impact stemming from university research has more impact on small firms than
130
large firms, while corporate research and development stimulates innovation in large firms
more than small firms. This is also the finding of Audretsch and Vivarelli (1994) based on
Italian firm data.
Also worth mentioning is the work of Baptista (2000), who used case studies to
analyze the process of technology diffusion in groups. His data shows that the greater the
number of firms in the region that have adopted the technology, the higher the adoption rate.
Hence, geographical location is important in the speed of diffusion of new technologies.
Nonetheless, when examining the way empirical studies analyze the mechanism of
spatial concentration of innovation activities, it is revealed that in geographic spillover
analysis, it is dangerous to infer questions about the location of indicators such as patents, the
number of innovations, or even the relationship between geographic regions and expenditure
on R&D (Anselin et al., 1997). That is why there are authors who argue that traditional
explanations of spillover effects are still unsatisfactory. Economic theory is still uncertain
about the ability of small and medium-sized firms to capture externalities. While conceptual
studies emphasize the concept of absorptive capacity, empirical studies illustrate the
correlation between the intensity of research university presence in a given geographical area
and innovation propensity, regardless of the sector considered. This does not give much
attention to the impact of local externalities.
Audretsch and Feldman (1996) and Audretsch and Stephan (1996) conducted a
specific analysis of the high-tech sector and showed that, in sectors where innovation is based
on science, geographic linkages are weak. 70% of relationships between biotechnology firms
and universities are not based on geographical proximity. However, the study of the
relationship between biotech firms and universities cannot be applied to site selection when
firms are established. In this phase, the relationship between start-ups and firms in the
"natural" network is fundamental and location is often done in the "natural" business
environment. Therefore, it appears that the analysis is very different at the time of start-up
(emergence) when the survival and development of the company depends on a close network
of relationships with its founders, and later when the company is established and builds
relationships in the same scientific, product and technological fields.
In addition, although clusters play an important role in technological innovation, under
certain conditions the opposite may occur, because although clusters initially encourage
innovative behavior, over time firms within the cluster may experience resource diseconomies
and isolated competitive practices, which may adversely affect firm performance and
ultimately stifle innovation (Pouder and St. John 1996).
131
However, despite differences of opinion, it is possible to identify a series of
conclusions with respect to the relationship between clusters and innovation (Lemarie et al.,
2001): (a) innovation in a region is closely related to public and private research invested in
the region (Feldman 1994), including non-research intensive industries (Mangematin and
Martin 1999); (b) innovation in a region is not only related to public and private R&D
investments, but also related to the technology transfer infrastructure throughout the region
(Lemarie et al., 2001) (presence of technology centers, technology transfer institutions, etc.)
(Feldman 1994), so that the presence of complementary activities generates more adverse
effects and reduces the costs and risks associated with firm innovation; and
(c) there is no displacement effect between public and privately invested research and
development. Both are encouraged to create fields of experience (Jaffe et al., 1993).
Therefore, the empirical findings show a consensus in confirming the positive relationship
between clusters and innovation.
Similarly, regional institutions act as vehicles that facilitate knowledge transmission
among firms within a cluster. Public technical centers run by the government offer technical
advisory services and seminars and broadcast information on new technologies and products.
Chambers of Commerce and trade and business associations coordinate activities within
clusters and provide technical information and information on markets, products and new
technologies (Yamawaki 2002).
9.4 FINAL DISCUSSION
The emergence of industrial regions and, as a consequence, clusters, represent a
multidisciplinary research field that combines the perspectives of strategic management,
industrial organization, economic geography and sociology in an attempt to enhance the
formal and informal relationships between firms and institutions in a given region (Feser and
Bergman 2000). This paper conducts an integration to delimit clusters and their main factors
and analyzes the factors that have a fundamental influence on innovation, efficiency and firm
performance.
In this regard, first of all, a review of theliterature onclusters shows that, although this
issue has been extensively analyzed over the past decades, especially from a theoretical point
of view, all the evidence available so far has not allowed a strong general theoretical
framework to be generated so that clusters can be thoroughly analyzed. Although clusters are
studied from different theoretical perspectives economic geography, industrial economics,
resources and capacity it is recognized that the subject of analysis is multidisciplinary;
132
however, an integrating model has not yet been developed.
On the one hand, there is no consensus on the definition of a cluster. The term is often
used synonymously with industrial parks, industrial agglomerations or even industrial
cooperation networks. The lack of unanimity and the absence of strong theoretical models
have complicated efforts to conduct empirical studies. The lack of an accurate methodology to
identify clusters has been one of the main limitations of this research as it precludes
generalization and comparison of the results obtained in different studies (O'Donoghue and
Gleave 2004; Soler 2001). In this paper we have defined clusters from their three basic
dimensions and in accordance with the generally accepted trend of recent years (Porter 1998):
Clusters are groups of interconnected firms and related institutions in related industries that
are geographically proximate.
This aspect also indirectly conditions the creation of an integrating model, as most
authors have analyzed or selected important factors that are closest to the dominant theoretical
perspective in their field of expertise (Paniccia 1998). This again prevents comparison and
generalization of the results obtained. Throughout this paper we have emphasized the
importance of economies of scale and scope, transportation costs related to inputs and outputs,
availability of factors in specific locations, cooperation and the impact of technology.
In the case of Spain, of the various attempts to identify and analyze clusters, the work
of Boix and Galletto (2006)6 and Trullen Thomas (2006) deserve special mention. These
authors, using the ISTAT (Instituto Nazionale di Statistica de Italia) methodology, identified
237 industrial regions in Spain, among which they highlighted the textile industry (with 53
districts), the food industry (52 districts), the furniture, jewelry and toy industry (40 districts),
household products (37 districts) and leather and footwear (30 districts).
To identify these clusters, the authors follow the methodological approach of the
Italian Institute which is based on the idea that, in order to reach the condition of a district, the
detected agglomeration must meet certain sectoral specialization requirements: the percentage
of the population employed in the manufacturing sector must be above the national average,
the percentage of employment provided by small and medium-sized enterprises in the area
must be above the national average, the percentage of employment provided by the main
industry in the local system must be above the national average for that industry, then the
percentage of employment provided by small and medium-sized enterprises in the sector must
be above the national average.
In general, although a large theoretical literature on clusters can be found, empirical
analysis only started to proliferate in the late nineties (Paniccia 1998). In this regard the
133
existing literature, using the case study methodology, has since its inception focused mainly
on explaining the success of specific regions in Europe and North America (Saxenian 1994),
its application to Third World and Asian economies has also been more recent (Cawthorne
1995; Meyer-Stamer 1995). The literature review reveals two main areas of study: the analysis
of the process of cluster formation and its dynamism; and the effect of clusters on business
competitiveness. In this regard, the aim of most of the literature is to explain the creation of
industrial clusters (Arthur 1989; Krugman 1991; Prevenzer 1997) and empirically identify
positive externalities as a consequence of industrial agglomeration (Ausdretsch and Feldman
1996; Jaffe et al., 1993). More recent studies attempt to study the dynamic process of
industrial agglomeration, with the aim of analyzing the key factors that lead to the emergence
of new clusters and the consequences of disagglomeration that some regions and industries
experience in this process (Dumais et al., 2002; Sorenson and Audia 2000).
A review of the literature (Baptista and Swann 1998; Krugman 1991; Muizer and
Hospers 2000; Porter 1990) allows us to assert that the key factor variables in the emergence
of clusters are economies of scale and scope, transportation costs (input and output)
transaction and search costs, availability of location-specific production factors and/or
components, abundant knowledge, information and technology, development of innovations,
cooperation between firms or between suppliers and buyers and reduction of uncertainty.
Much of the literature emphasizes that industry clusters promote firm creation and
enhance competitiveness (Carlsson 2002; Lemarie et al., 2001; Sternberg and Litzenberger
2004; Yamawaki 2002). Much of the literature focuses on the relationship between clusters
and the competitiveness of firms, regions and countries. However, the empirical analysis
tends to be imprecise and the findings inconclusive.
Finally, we should point out that in this paper we have specifically focused on
analyzing two positive aspects of clusters in terms of business efficiency and performance and
innovation. On the one hand, we have shown the benefits that clusters provide to firms due to
better access to inputs, information, technology and institutions. On the other hand, one of the
main characteristics of clusters is their capacity to generate innovative environments that are
based on the complex interaction between cooperation and competition factors (Porter 1990,
1998; Saxenian 1994). Since firms are geographically located close to each other, they can
cooperate relatively easily to develop innovations, inform each other of changes in input
specifications, launch new products and detect changes in technology and demand.
However, cooperation and adverse effects are possible in the long run. In this case,
geographical proximity is especially important for radical technological change, as human
134
contact is essential. It is less important for gradual technological change because this type of
innovation requires more codified knowledge than can be communicated remotely. Clusters
are therefore more important for firms using or developing new technologies than for firms
using or developing incremental innovations.
However, and despite the review and integration carried out in this paper that focuses
on conceptually delimiting clusters on the one hand, and on the other hand, in particular, on
analyzing the role of clusters as direct and indirect generators of innovation - the theoretical
and empirical evidence is still too heterogeneous to be able to establish conceptual boundaries
and generalizable conclusions. All of this points to the fact that it is not only cluster research
that is a field experiencing a resurgence of interest in recent years: this interest still requires
larger and more thorough research to answer the many puzzles and questions regarding what
cluster research actually is. How can it be empirically constrained and what are its main
factors and main impacts on corporate efficiency and economic development.
