Faculty of Economics and Business
The impact of high-tech clusters on the country’s
innovative capacity – A comparative analysis between
Europe and U.S.
AUTHOR: Alina Haita
STUDENT No: 1903861
COURSE NAME: Masters in Business Administration (Strategy and Innovation)
SUPERVISOR: dr. Rene van der Eijk
CO-ASSESOR: dr. T.L.J. Broekhuizen
Executive summary
Innovation through the creation, diffusion and use of knowledge has become a key driver of economic growth for any country. The determinants of innovation performance have changed in a globalizing knowledge-based economy, partly as a result of recent developments in ICT industry. Nowadays, innovation results from increasingly complex interactions at the local, national and world levels among individuals, firms and other knowledge institutions (OECD, 2001). As such, many economies have attempted to gain national economic advanatage from regional clusters of development in information and communications technologies (ICT).
The main purpose of my paper is to show how these high-tech clusters contribute to a country’s innovative capacity measured by the number of patent applications, in countries across Europe and states from U.S.
In the first part of the research, after the introduction, a literature review chapter will describe what innovation is, and how clusters of companies relate to it. Also, the innovative capacity of a country will be defined and its relationship with high-tech clusters will be explained. Furthermore, based on the existing literature on the topic, a conceptual model will be built in order to test the hypothesis regarding the influence high-tech clusters have on the innovative capacity of a country.
The main part of the data used in this research was gathered from statistical websites like the portal of OECD (Organisation for Economic Co-operation and Development) and Eurostat – the statistical office of the European Union, as well as from The Cluster Mapping database which can be found on the European Cluster Observatory official website – (www.clusterobservatory.eu). Additional data for U.S. was collected from the Institute for Strategy and Competitiveness’ web site of the Harvard Business School (data.isc.hbs.edu/isc/).
The results obtained show that there are several sources of differences among countries in the production of visible innovation output and the presence of clusters is one of them. As such, if the number of employees within a cluster has a positive and significant impact on the number of patent applications in both Europe and U.S., the size of a cluster seems to be an important factor influencing the innovation output of a country just in Europe. Also, the tertiary education graduates influence the number of patent applications just in some cases, while BERD (Business Expenditure on Education) plays an important role in the innovation process across countries and states from both Europe and U.S.
TABLE OF CONTENTS
1. INTRODUCTION...4
1.1 PROBLEM STATEMENT...5
1.2RESEARCH QUESTIONS...6
1.3OBJECTIVES AND AIMS...6
1.4RESEARCH OUTLINE...7
2. LITERATURE REVIEW ...8
2.1CLUSTERS AND INNOVATION...8
2.2DEFINING HIGH-TECH INDUSTRIES AND HIGH-TECH CLUSTERS...11
2.3NATIONAL INNOVATIVE CAPACITY...12
2.4CONCEPTUAL MODEL...15
3. METHODOLOGY...19
3.1 RESEARCH STRATEGY AND DATA COLLECTION...19
3.2 SAMPLE SIZE ...20
3.3 METHODS...21
3.3.1 MULTIPLE REGRESSION ANALYSIS...21
3.3.2 VARIABLES...21
4. RESULTS ...23
5. DISCUSSION, LIMITATIONS AND FUTURE RESEARCH ...30
5.1 DISCUSSION...30
5.2 MAIN FINDINGS AND IMPLICATIONS OF THE RESEARCH...32
5.3 LIMITATIONS AND FUTURE RESEARCH...33
6. CONCLUSIONS ...34
7. REFERENCES...36
1. INTRODUCTION
Nowadays nations differentiate themselves not only through endowments of inputs such as labour, natural resources and capital, but prosperity depends more and more on creating a business environment, along with supporting institutions, that fosters innovation and determines productivity growth (Porter, 1990).
Innovation has become perhaps the most important source of competitive advantage in advanced economies, and building innovative capacity has a strong relationship to a country’s overall competitiveness and level of prosperity (Porter and Stern, 2001).
Although less-advanced countries can still improve their productivity by adopting existing technologies or making incremental improvements in other areas, for countries that have reached the innovation stage of development, this is no longer sufficient to increase productivity. Firms in these countries must design and develop cutting-edge products and processes to maintain a competitive edge. This requires an environment that is conducive to innovative activity, supported by both the public and the private sectors. In particular, this means sufficient investment in research and development (R&D) especially by the private sector, the presence of high-quality scientific research institutions, extensive collaboration in research between universities and industry, the protection of intellectual property and the presence of established clusters (Porter, 2008).
In Porter’s (1998) view, competitive success and innovation in so many fields are geographically concentrated. Clusters are seen as geographic concentration of interconnected companies and institutions in a particular field. These clusters refer to companies in industries related by skills, technologies or common inputs. A large variety of clusters, each with specific characteristics, have been identified. As an example, in Europe, watchmakers clustered in Switzerland and fashion designers in Paris. In the U.S., there is the well known cluster in Detroit for the automotive industry, Hollywood for motion pictures, New York City for financial services and advertising and finally Silicon Valley for electronics (Fallah, 2005).
Clusters play a vital role in a company’s ongoing ability to innovate. Companies inside clusters usually have a better window on the market than isolated competitors do; the relationships with other entities within the cluster help companies to learn early about evolving technology, component and machinery availability, service and marketing concepts (Fallah, 2005). Also, a company within a cluster can source what it needs to implement innovations more quickly mainly because all the partners do get closely involved in the innovation process, thus ensuring a better match with customers requirements (Porter, 1998).
environment that facilitates the birth of new businesses. Technology clusters have a higher innovation output mainly because of the knowledge spillovers that exist within the clusters. Companies within such a cluster produce more patents form their R&D investments than companies within other types of clusters (Fallah, 2005). The development of high-tech clusters has also an important impact on a country’s ability to innovate, known as the national innovative capacity.
According to Porter and Stern (2002), the national innovative capacity is the ability of a country – as both a political and economic entity – to produce and commercialize a flow of new-to-the-world technologies over the long term. The innovative capacity of a country is based on three distinct areas of research: ideas-driven endogenous growth theory (Romer, 1990), the cluster-based theory of national industrial competitive advantage (Porter, 1990), and research on national innovation systems (Nelson, 1993). Each of these perspectives identifies country-specific factors that determine the flow of innovation.
In my study I will focus on the impact high-tech clusters have on a country’s innovative capacity, by comparing 31 countries across Europe – Czech Republic, Denmark, Estonia, Greece, Iceland, Ireland, Luxembourg, Norway, Poland, Portugal, Romania, Slovakia, Slovenia, Sweden, Turkey, Finland, The Netherlands, Germany, United Kindom, Austria, Belgium, France, Hungary, Italy, Spain, Cyprus, Switzerland, Bulgaria, Latvia, Lithuania, Malta. The reason I have chosen to conduct an European level analysis is because countries in Europe have many establised ICT clusters with high innovation output like the one from Helsinki in Finland and the ones from Eindhoven and Amsterdam in Holland, or the ICT clusters from Berlin and Stuttgard in Germany as well as the one from Wien in Austria and the Berks, Bucks and Oxon (Oxford) cluster in UK, along with emerging clusters from developing countries like Romania, Bulgaria or Latvia.
Regarding U.S., I have chosen for my analysis the following 39 states: Alabama, Arizona, Arkansas, California, Colorado, Connecticut, District of Columbia, Florida, Georgia, Idaho, Illinois, Indiana, Iowa, Kansan, Kentucky, Maine, Maryland, Massachusetts, Michigan, Minnesota, Missouri, Nevada, New Hampshire, New Jersey, New Mexico, New York, North Carolina, Ohio, Oklahoma, Pennsylvania, South Carolina, Tennessee, Texas, Utah, Vermont, Virginia, Washington and Wisconsin. Here we have many established clusters, with a long history of cooperation regarding the R&D activities, like the famous Silicon Valley.
1.1. Problem statement
Nowadays, innovation results from increasingly complex interactions at the local, national and world levels among individuals, firms and other knowledge institutions (OECD, 2001).
As such, many economies have attempted to gain national economic advanatage from regional clusters of development in information and communications technologies (ICT), but only few succeded. Some successful cases are in Europe: Finland, Germany, France, Italy, Austria, Spain and Belgium and also in U.S., the famous Silicon Valley in California.
It is really intersting to see how clusters influence the innovative capacity in countries that have different economic principles and industrial organization.
1.2. Research Question
In order to find out what impact high-tech clusters have on a country’s innovative capacity, the following question needs to be answered:
• How clusters in high-tech industries contribute to a country’s national innovative
capacity?
In order to answer the main question the following sub-questions will be answered: • How important is the existence of high skilled labor within a cluster? • How important is the level of inputs devoted to innovation?
• How significant is the size of a cluster, measured in number of companies within the cluster, for the innovative output of a country in ICT industry?
1.3. Objectives and aims
While R&D activity takes place in many countries, the development and commercialization of “new-to-the-world” technologies has been concentrated historically in relatively few countries (Stern and Porter, 2000). There are several sources of differences among countries in the production of visible innovation output and the presence of clusters is one of them.
The goal of this research is to measure the impact of high-tech clusters on a country’s ability to innovate and to determine if there are any differences between countries with different cultures and economic principles. The number of patent applications in ICT sector will be used as a measure for a country’s innovative capacity; also, the cluster specific characteristics will be measured by the number of companies and the number of employees within these clusters.
1.4. Research Outline
In the first part of the research, after the introduction, a literature review chapter will describe what innovation is, and how clusters of companies relate to it. Also, the innovative capacity of a country will be defined and its relationship with high-tech clusters will be explained. Furthermore, based on the existing literature on the topic, a conceptual model will be built in order to test the hypothesis regarding the influence high-tech clusters have on the innovative capacity of a country.
In order to better understand the local environment that fosters innovation and cluster formation, an overview of high-tech clustering activities and the national innovation systems in the countries that make the subject of this analysis, will be done.
2. LITERATURE REVIEW
This chapter discusses about what clusters are, about their importance for the growth and competitiveness of any economy and also about their relationship with innovation. It also describes what a country’s innovative capacity is and what contribution high-tech clusters have to its growth. The work of Porter (1990; 1998; 2000) plays an important role in both defining clusters and national innovative capacity. After a detailed overview of the existing literature on this topic, several hypotheses will be formulated and furthermore will be tested in the methodology chapter. 2.1. Clusters and innovation
Innovation has become perhaps the most important source of competitive advantage in advanced economies, and building innovative capacity has a strong relationship to a country’s overall competitiveness and level of prosperity (Porter and Stern, 2002). As such, understanding technical change and innovation is crucial for understanding the dynamics of any knowledge-based or learning economy.
Innovation can be considered an open process, in which many different actors like companies, customers, inventors, universities and other organizations cooperate in a complex way. Ideas move across institutional borders more frequently. Nowadays, the traditional model of innovation with clearly assigned roles for basic research at the university, and applied research in a company R&D centre, in no longer relevant (Europe INNOVA, 2007).
In modern innovation theory, strategic behavior and alliances of firms, as well as interaction and knowledge exchange among firms, research institutes, universities and other institutions, are the heart of the innovation process. Increasing complexity, costs and risks in innovation are enhancing the value of inter-firm networking and collaboration in order to reduce moral hazard and transaction costs. Companies are becoming more dependent on complementary knowledge and know-how, other than their own, with the purpose of applying these new sources of technology to products and production processes (OECD, 1999).
according to the innovative milieu model, geographical proximity and informal relationships between companies facilitate information and knowledge exchange. Therefore collective learning may be a cause of enhanced innovative behavior by companies and has been seen as an uncertainty reducing mechanism in a rapidly changing technology context (Eraydin and Armatli-Köroğlu, 2005).
Academic researches and empirical evidences have showed that clusters enhance regional competitiveness as they increase productivity and efficiency, boost innovation and favor the attraction of new firms and start-ups (Porter, 1998). Opportnities for innovation can be percieved more easily whithin clusters and the skills, assets and capital are more available to pursue them (Stainbock, 2004). Cooperation between companies within a cluster offers them a direct way to improve economic performance, offers the opportunity for learning, enables risks and R&D costs to be shared, and facilitates flexibility (OECD, 1999).
Due to clusters, many European regions have developed competitive advantages in specialized activities such as financial services (London), petrochemicals (Antwerp), flowers (Holland), and biopharma (The Danish-Swedish border region). Successful clusters have also significantly increased their global reach – attracting people, technology and investments, serving global markets, and connecting with other regional clusters that provide complementary activities in global value chains (Europe INNOVA, 2007).
A lot of debate about what a cluster is and how it emerges has been done in the past two decades (Cortright, 2006). The concept of clusters is a modern description of the long observed phenomenon of geographical concentration of economic activities, which is widely believed to be an important factor for economic development. Marshall (1890) described already in the 19th century the advantages of agglomeration of economic activities in terms of availability of a qualified work force and specialization. According to Porter (1998), clusters are “geographic concentrations of interconnected companies, specialized suppliers, service providers, firms in related industries, and associated institutions (for example, universities, standards agencies, and trade associations) in particular fields that compete but also co-operate”. Rosenfeld (2002), on the other hand defines a cluster as "a spatially limited critical mass (that is sufficient to attract specialized services, resources, and suppliers) of companies that have some systemic relationships to one another based on complementarities or similarities”. Also, according to Robinson (2002), clusters are being used to organize local economic development efforts, develop empirical analyses of local economies, and theorize about regional economic growth.
grow in regions that provide specific advantages as a location for companies’ activities in a particular field and they can enable linkages and collaboration; clusters reach their full economic potential if they are well connected to markets and clusters elsewhere and when cluster participants cooperate to strengthen linkages and align decisions that are not designed or allowed to distort the market (Center for Strategy and Competitiveness, 2008).
The emergence of clusters is often a specific result of a certain initiative, based on a national or regional cluster policy. These cluster initiatives can be understood as “organized efforts to increase growth and competitiveness of clusters within a region, involving clusters firms, government and/or the research community (Koker, 2009). There are many reasons that determine the emergence of new technology and innovation centers, like the leverage of multinational corporations, fragmentation of industries, mobilization of knowledge, people and other resources around the world, supported by a suitable cluster policy (Engel, 2009). This policy refers to all those efforts of government to develop and support clusters in a particular area and should be aimed at removing obstacles, relaxing constraints, and eliminating inefficiencies that impede productivity and innovation in the cluster (Porter, 2000a).
What is certain is that, geographic, cultural, and institutional proximity provides companies with special access, closer relationships, better information, powerful incentives, and other advantages that are difficult to tap from a distance.
common raw materials to attract bulk discounts, or by joint marketing. Companies can benefit from sharing knowledge about best practice and reduce costs by jointly sourcing services and suppliers (Saublens, 2007).
One of the most important distinguishing features of clusters is the ease of transfer of the tacit knowledge or know-how that is based on experience and judgment and is not codified. Informal learning, acquiring know-how, and trust building require the face-to-face contact that occurs through social, professional or trade, and business situations (Rosenfeld, 2002). The flow of innovation between companies can be determined by many factors; first, by the presence of a strong common innovation infrastructure which includes a country’s overall science and technology policy environment, the mechanisms in place for supporting basic research and higher education, and the cumulative “stock” of technological knowledge upon which new ideas are developed and commercialized; second, the presence of specific innovation environments founded in clusters; third, the strength of the linkages between the common innovation infrastructure and specific clusters (Porter and Stern, 2002).
Clusters differ substantially in size, geographic span, core and strength of cluster ties. The geographic scope of a cluster can range from a single city to a country or even a network of neighboring countries (Stainbock, 2004). A cluster can contain a small or large number of enterprises that cooperate vertically along the value chain, and horizontally with competitors and sometimes even laterally with complementary but unrelated companies and other organizations like research facilities. Clusters vary widely regarding the number of participants and their degree of organization. They generally contain firms that compete against each other, but cooperation may be achieved on a case-by-case bases. Usually, clustering occurs in all branches of industry, be it high-tech or traditional industries, as well as in agriculture or in the service sector with each cluster being a unique constellation in time and space (OECD, 2005).
2.2. Defining high-tech industries and high-tech clusters
development spending, and a higher than industry average sales growth (National Science Foundation, 1998).
Nowadays, the globalization of the value chain functions has provided the opportunity for many developing regions to focus on high-tech clusters as a means of creating competitive advantage, to attract and maintain high-tech corporations and increase economic development and growth (Fallah, 2005).
One of the most important type of clusters are the technological clusters - geographical concentration of related technology firms including competitors, suppliers, distributors, and customers, usually found around scientific research centers and universities (Fallah, 2005). Industrial (high-tech) clusters are also defined as “a regional network-based industrial system that promotes collective learning and flexible adjustment to changed conditions among specialist producers of complex, related technologies” (Saxenian, 1994). This type of clusters, make use of external networking as a key element in the production and application of new knowledge. Strategic networks potentially provide a firm with access to information, resources, markets and technologies, generate advantages from learning, scale and scope economies, and allow firms to achieve strategic objectives, such as sharing risks and outsourcing value-chain stages and organizational functions (Nachum, 2001). The geographic concentration of technological activity, skilled labor and inter-related industries, confers advantages that can be translated into economic growth and competitiveness.
The development of high-tech clusters has also an important impact on the national economic growth. Firstly, information and communication technology (ICT) - producing industries contribute directly to productivity and growth through their own rapid technological progress; secondly, ICT use, improves the productivity of other factors of production (or inputs); and thirdly, there are ‘spillover effects’ on the rest of the economy as ICT diffusion leads to innovation and efficiency gains in other sectors. Also, the ICT sector is the biggest R&D investing industrial sector and provides other industries with productivity enhancing technologies (Europe’s Digital Competitiveness Report, 2010).
2.3. National innovative capacity
differences in the level of inputs devoted to innovation (R&D manpower and spending), and also an extremely important role is played by factors associated with differences in R&D productivity, like policy choices such as the extent of IP (intellectual property) protection and openness to international trade, the share of research performed by the academic sector and funded by the private sector, the degree of technological specialization, and each individual country’s knowledge “stock”.
In order to be competitive and gain national economic advanatage, each country should offer a favorable environment for innovation by linking private-sector strategies with public-sector policies, a characteristic known as a country’s innovative capacity.
The national innovative capacity is the ability of a country – as both a political and economic entity – to produce and commercialize a flow of new-to-the-world technologies over the long term (Porter and Stern, 2002). According to the authors, differences in national innovative capacity reflect variation in both economic geography (the level of spillovers between local firms) as well as cross-country differences in innovation policy (the level of public support for basic research or legal protection for intellectual property).
Many studies have attempted to measure the level of innovative efforts and innovative outcomes, using different indicators. One such example is the conceptual framework introduced by Liu and White (2001) which describes five different activities of the innovation processes like: research, production, linkage, education and end-use (customers of the product or process inputs). Based on previous work by the OECD (1999), Chang and Shih (2003) introduced another model made up of six elements - R&D expenditure, R&D performance, technology policy, human capital development, technology transfer and the climate for entrepreneurial behavior.
upstream scientific and technical advances can actually diffuse to other countries more quickly than they can be exploited at home. Particularly important linking institutions are a nation’s universities, which can play the role of bridging researchers and companies. The last element that can influence a country’s innovative capacity is the companies’ ability to innovate. Companies must embrace strategies based on innovation and choose supportive operating policies in areas such as R&D spending, customer orientation, recruiting and training.
Also, according to a study made by Europe INNOVA (2007), an initiative for innovation professionals supported by the European Comission, the relationship between clusters and innovation is clearly complex. A comparison between the best performing innovation regions in Europe shows that 7 out of 19 regions having a strong cluster portfolio are among top third most innovative regions. The indicators on which the analysis was conducted include human resources in science and technology, patent applications and employment in medium-high and high-tech manufacturing. As such, the results obtained suggest a strong correlation between the strength of regional cluster portfolios and the regional innovation performance.
In order to measure the differences in the production of innovative input between countries Porter and Stern (2002) suggest the indicators based on patent statistics; these are widely used in order to assess the inventive and innovative performance of a country or a region. The current emphasis on innovation as a source of industrial competitiveness has raised awareness of patents. Patents are used to protect R&D results, but they are just as significant as a source of technical information, which may avoid reinventing and redeveloping ideas because of a lack of information. A patent is an intellectual property right relating to inventions in the technical field. A patent may be granted to a firm, individual or public body by a national patent office. An application for a patent has to meet certain requirements: the invention must be novel, involve an (non-obvious) inventive step and be capable of industrial application (OECD, 2001). A patent gives the right to the inventor for a given period of time to exploit commercial revenues deriving from the application of his own invention. Patents help the entry in markets especially for small and medium sized firms which are less able to protect their innovations in alternative ways, and support investments devoted to the introduction of radical innovations characterized by a high degree of uncertainty, elevated costs and long time lasting between the invention stage and the market introduction of innovation (Crespi, 2004).
Table 2.1 Summary of studies focusing on the impact of clusters on the innovative capacity of a country Author Period of study Countries Ways of measuring Main findings Furman et. al. (2001)
1973-1996 17 Panel data based on 17 cross-sections
When controlling for GDP per capita the following variables were found to have a positive and significant influence on the number of patent applications: Full time equivalent scientists and engineers in all sectors, Aggregate R&D Expenditures, Strength of Protection for Intellectual Property, Share of GDP Spent on Higher Education, Percentage of R&D Performed by Universities Porter and Stern, (2002) 1999-2000 75 Multivariate regression analysis where they included 24 variables, related to 3 main distinct groupings: innovation-related public policy, the cluster innovation environment, and the strength of innovation linkages.
Even if they controlled for the size of the country and the human resources devoted to innovation, 23 of the
variables were positive and statistically significant. Some of the main variables included were: patents in US as a dependent variable and the Intellectual Property Protection, Government R&D Tax Credits, State of cluster
development, venture capital availability, quality of research
institutions, as independent variables. 2.4. Conceptual model
and start-ups (Porter, 1998).
The main interest of the research is to analyse how clusters influence the national innovative capacity of a country. As such, I will compare several countries with different economic environments to see if there are any major differences between them, regarding clustering impact on innovation. In order to conduct my analysis I will build a conceptual model based on the existing literature about clusters and national innovative capacity.
The conceptual model shows how cluster specific characteristics, influence a country’s innovative capacity, measured by the number of patent applications. Patents are used to track the level of diffusion of knowledge across technology areas, countries, sectors, firms, and the level of internationalization of innovative activities. Patent indicators can serve to measure the output of Research & Development (R&D), its productivity, structure and the development of a specific technology/industry (OECD, 2004). They can be considered as an intermediate step between R&D (upstream) and innovation (the invention is used downstream in economic processes). Patents can be obtained at different stages of the R&D process, notably in the case of incremental or cumulative inventions. In this sense, patents can be seen not only as an output of R&D but also as an input to innovation and thus as both inputs and outputs in the invention process. This intermediate character makes patent data a useful bridge between R&D data and innovation data (OECD, 2009).
Factors influencing a country’s National Innovative Capacity
scientific and technological advances, the public policies bearing on innovative activity and the economy’s level of technological sophistication. A country’s innovation infrastructure is based on its scientists and engineers available to contribute to new product development along with investments in R&D from both public and private sectors (Porter and Stern, 2002).
A significant contribution to a country’s innovative infrastructure has the education system; this plays an important role in building competencies for innovation and in deploying human capital on the labor market. Universities have gained a growing importance as providers of useful new knowledge and as trainers of the researchers and other highly skilled workers on which knowledge-based economies relay. Enabling people throughout the economy and society to participate in innovation will provide new ideas, knowledge and capabilities, and enhance the influence of market demand on innovation (OECD, 2010).
Tertiary education sector, which is university education, is responsible for the largest share of a country’s research output. In addition, the sector is responsible for most of the training in research – developing advanced research skills for those entering work as well as producing graduates with skills, knowledge and attributes that enable them to contribute to research and innovation (OECD, 2006). For my study I will use the number of students in the second stage of tertiary education, which is level 6 of education, according to ISCED 1997 (Annex 2). This level is reserved for tertiary programs that lead to the award of an advanced research qualification and it prepares recipients for faculty posts in different institutions as well as research posts in government and industry. This variable is collected as follows: distribution of graduates by country, year, level of education, program destination, program duration, program orientation, field of education and gender. Graduates are those who successfully complete an educational program during the reference year of the data collection. One condition of a successful completion is that students should have enrolled in, and successfully completed, the final year of the corresponding educational program, although not necessarily in the year of reference (OECD, Statistics on Education and Training,
http://stats.oecd.org/Index.aspx?DataSetCode=RGRADAGE). In my analysis, the number of tertiary education graduates refer to those in fields like mathematics and statistics, science and computing, which involves computer science and computer use.
local skilled-labour supply and information spillovers (Marshall, 1920). In my theoretical framework I will use the local skilled-labour, as one factor that can influence the innovative capacity of a country, mainly because the human capital is a key driver for entrepreneurship in a region. A region will need to acquire and maintain the ability to attract people with appropriate skills to develop and maintain an entrepreneurial culture through linkage with schools and other educational establishments as well as a strong focus on workforce development and vocational training. A cluster of firms can attract to a region a rich variety of labour categories specialized to suit the industry in question (Karlsson, 2008) along with investments from both public and private sectors.
Based on the existing literature about clustering and the innovative capacity of a country, it is hypothesized that:
Hypothesis 1: The number of high-skilled employees within a cluster can positively influence a country’s innovative capacity.
3. METHODOLOGY
This chapter will describe the methodology used to conduct my research regarding the influence of high-tech clusters on the innovative capacity of a country. The purpose of this analysis is to determine if clusters can have the same impact on the innovative capacity in countries of different size, industrial structure and economic performance.
3.1 Research strategy and data collection
According to Kumar (2005), research can be classified from three perspectives: application of the research study (pure or applied research), objectives in undertaking the research (descriptive, explanatory and exploratory research) and inquiry mode employed (quantitative or qualitative research).
In my study I will use a quantitative approach, which aims to collect numerical data in order to measure the variation in a phenomenon, situation, problem or issue (Muijs, 2004). The objective of quantitative research is to develop and employ mathematical models, theories and/or hypotheses pertaining to phenomena.
According to Yin (1994), there are two types of data that can be collected for a study: primary data and secondary data. Usually secondary analysis is performing a new analysis over an existing dataset with help of more sophisticated statistical measurements with the goal to test hypotheses and answer questions in a more comprehensive and succinct manner than in the original report (Hakim, 1982).
In order to find out the impact of high-tech clusters on a country’s innovative capacity, I will focus on a quantitative approach, using secondary data gathered through literature research in both scientific and non-scientific fields (e.g. articles, news papers, books), mostly searched for, on the Internet. The main part of my data will be obtained from the statistics portal of OECD (Organisation for Economic Co-operation and Development) and Eurostat – the statistical office of the European Union. I will obtain data also from The Cluster Mapping database which can be found on the European Cluster Observatory official website – (www.clusterobservatory.eu). The Cluster Mapping database is built in the intersection of regions and sectors in Europe. By combining the two dimensions of geography and industry, regional clusters can be statistically traced across Europe. Information about the number of clusters and also the number of emloyees can be found here. Additional data for U.S. was collected from the Institute for Strategy and Competitiveness’ web site of the Harvard Business School (data.isc.hbs.edu/isc/).
application within the three countries of my analysis and the independent and control variables. Before conducting the correlation analysis I will plot a scatterplot to have an overview of the general trend of my data. After testing the correlation, a regression analysis will be conducted in order to predict values of the dependent variable from sevral independent varibles.
The quantitative analysis I will further use must be well constructed to ensure construct validity, internal validity, external validity, and reliability. Construct validity requires the use of correct measures for the concepts being studied (Yin, 1989). Construct validity can be evaluated by statistical methods, like the correlation analysis. In my research, in order to determine the impact of high-tech clusters on a country’s innovative capacity, first I will conduct a correlation analysis to see if there is any linear relationship between my variables and second, I will conduct a multiple regression analysis. In my case, I want to predict the number of patent application in ICT industry by using the number of companies within the clusters and the number of employees within these clusters as predictors along with tertiary education graduates and BERD (Business Expenditure on R&D).
Internal validity demonstrates that certain conditions lead to other conditions and requires the use of multiple pieces of evidence from multiple sources to uncover convergent lines of inquiry (Yin, 1989). The collection of my data is made from multiple evidence sources like the website of OECD and Eurostat, the statistical office of the European Union, as such internal validity is assured. External validity reflects whether or not findings can be generalized. In my case, the study can be further used in order to determine the impact of clusters from other industries on the national innovative capacity of different countries.
Reliability refers to the stability, accuracy, and precision of measurement (Yin, 1989). 3.2 Sample Size
All clusters have ‘geographical borders’, even though in many cases these are not well defined. Many cluster studies simply take regions, provinces or states as relevant cluster region (Van Klink and De Langen, 2001). For my study I will use data available at the cluster level, industry level and also country level for the control variables.
My sample will consist of 31 European countries along with 39 states from U.S and the existing clusters in the ICT (Information and Communication Technologies) industry. In order to see the differences in the innovation output between countries from Europe governed more or less by same regulations and the stats from U.S., I will conduct two analyses and compare the results.
limitations and increasing efficiency in communication (OECD, 2002). The reason I have chosen 2007 as a reference year for my analysis is because this is the most recent year with available data for all my variables.
3.3 Methods
3.3.1 Multiple Regression Analysis
In order to analyze the impact of high-tech clusters on the innovative capacity of a country measured by the number of patent application, I will first conduct a correlation analysis of the variables used in my research. There are two types of correlations: bivariate and partial. A bivariate correlation is a correlation between two variables and the partial correlation is the one that looks at the relationship between two variables while controlling the effect of one or more additional variables (Field, 2000). In my analysis I will conduct a bivariate correlation using Pearson’s correlation coefficient.
Second, I will use a multiple regression model in order to predict values of the patent applications from different independent variables. The multiple regression model, is a linear regression where the dependent variable, yi is a linear combination of the parameters. The equation
for this model, with p independent variables, is the following parabola:
where, xi is the independent variable, βp is the coefficient of each predictor, εi is an error term and
the subscript i indexes a particular observation. In my analysis, the multiple regression model I used based on the method of ordinary least squares (OLS), has the following equation:
Patent appl = b0 + b1Employees in the cluster + b2Companies in the cluster + b3Tertiary
education graduates + b4BERD + εi
In order to better understand the motivation of my choice regarding the variables that I will use in my analysis, furthermore all variables will be described.
3.3.2 Variables
In my study I have decided to measure the innovative capacity of a country by the number of patent applications per million inhabitants. Indicators based on patents provide a good measure of the innovative performance and technology outputs of countries. Patent statistics provide a measure of innovation output, as they reflect the inventive performance of countries, regions, technologies or firms (OECD, 2004).
semiconductors and telecommunications), taking into account the priority date. The priority date corresponds to the first filing worldwide and therefore closest to the invention date. To measure inventive activity, patent should be counted according to the priority date (in the case of patent families, the priority date corresponds to the earliest priority among the set of patents). The priority date will reflect the proper time period of the discovery of both domestic and foreign inventions. For this reason, when compiling patent statistics to reflect inventive activities, it is recommended to use the priority as the reference date (OECD, 2009).
Furthermore, I will consider cluster specific characteristics like the number of companies within the ICT clusters along with the size of ICT clusters, measured by the number of employees, as explanatory variables.
This study incorporates Business Expenditure on R&D measured in millions USD and Tertiary Education Graduates as control variables that may influence the impact of high-tech clusters on a country’s innovative capacity. Choosing several control variables can help predicting in a better way the effects the independent variables have over the dependent variable.
People are at the heart of the innovation process. A first set of indicators focuses on the role education systems play in building competencies for innovation and on how this human capital is actually deployed on the labor market. Enabling people throughout the economy and society to participate in innovation will provide new ideas, knowledge and capabilities, and enhance the influence of market demand on innovation (OECD, 2002).
statistics and computing, which involves computer science and computer use.
Second, I have chosen to use as a control variable BERD (Business Expenditure on Research and Development), mainly because R&D indicators are still the most widely used indicators of innovative activity, and this may be due to a number of reasons. First, R&D subsidies play a central role in national science and technology policies and therefore call attention to R&D-based indicators. Second, R&D data have been considered more reliable, particularly in relation to early innovation survey data. Third, policy makers lack innovation indicators that are as widely accepted and utilized as R&D and therefore find innovation measures less useful. Finally, policy makers may not be fully aware of the innovation data available or its potential uses. (OECD
Science, Technology and Industry Outlook, 2008,
http://browse.oecdbookshop.org/oecd/pdfs/browseit/9208101E.PDF)
Therefore, in my study I will consider that BERD in ICT industry can influence the number of patent applications. This variable includes expenditure on computer and related activities, software consultancy and supply, research and development and telecommunications.
4. RESULTS
Before conducting the correlation and multiple regression analysis a preliminary data analysis was done using a simple scatterplot in order to look at the general trend. Both independent variables, number of ICT clusters and number of employees within the clusters, were taken into account for Europe and U.S.
Figure 1. Scatter-plot of the number of patent applications Figure 2. Scatter-plot of the number of patent applications
against the number of companies within ICT clusters in U.S. against the number of companies within ICT clusters in Europe
U.S. most patent applications come from clusters that have from 1 to 100 companies, in Europe the number of companies is much bigger, from 1 to 5000. Also, in the case of U.S., the average value of patent applications within clusters in ICT is around 100 while in the case of Europe is around 1000. This can mean that companies within clusters from Europe are less productive regarding the innovation output than companies within clusters from U.S. This is the case of Italy, with 29 704 companies in the ICT clusters and only 1103 patent applications and also the case of Poland with 20 752 companies and 48 patent applications.
When looking at the scatter-plot of U.S. we can see one state, California, which is the only state that has a really high number of patent applications and also a high number of companies within the clusters. This may be due to the fact that Silicon Valley is one of the main clusters located in California and most patents come from companies such as Google, Microsoft, EBay and HP.
Figure3. Scatter-plot of the number of patent applications Figure 4. Scatter-plot of the number of patent applications
against the number of employees within ICT clusters in U.S. against the number of employees within ICT clusters in Europe
When looking at Figure 3., the state of California has the highest number of employees (168 699) within the ICT clusters along with the highest number of patent applications (511 per million inhabitants), followed by Texas with 104 156 employees in the ICT clusters and a total number of patent applications of 146. In the case of Europe, Germany has the highest number of patent applications (8249 per million inhabitants) along with a high number of employees (404 938) within ICT clusters followed by France which has 4247 patent applications per million inhabitants and 194 198 employees within the high-tech clusters.
Descriptive Statistics
This chapter describes the empirical results of my study. First, I will provide the descriptive statistics, which are a collection of measurements of two things: location and variability. Location tells about the central value of my variables (the mean is the most common measure). Variability refers to the spread of the data from the center value like the variance and standard deviation (DSS Online Training Section, http://dss.princeton.edu/training/DataPrep101.pdf). Table 4.1 presents the descriptive statistics of my variables for U.S.
Table 4.1. Descriptive statistics U.S. (N= 39 observations)
N Minimum Maximum Mean
Std. Deviation Nb of patent applications(per million inhabitants) 39 1 511 44.64 85.211 Nb of companies in the cluster 39 1 582 52.38 112.324 Nb of employees in the cluster 39 1430 168699 23964.67 31087.337 Tertiary education graduates in Science and Technology
39 2267 89947 16653.95 16586.27
BERD ICT (millions
EUR) 39 265 64187 6716.85 10803.296
GDP per capita (US
dollars) 39 24543 1812968 332308.49 351498.696
According to the descriptive statistics, and more specific the minimum and maximum values, there are significant differences between countries regarding the number of patent applications in 2007. This can be due to differences that exist also between the independent variables. As we can see the main difference is in the number of employees within the clusters, followed by the tertiary education graduates and the expenditure companies have with R&D activities.
In U.S., California seems to be the most innovative state as it has the highest values for all our variables. This can be due to the existence of Silicon Valley, which with its dense and flexible networks of tight relationships among entrepreneurs, venture capitalists, university researchers and highly skilled employees, managed to become the most successful and innovative high-tech clusters (California Regional Economies Project, September 2004).
applications in ICT; this can be due to a low level of BERD, along with a small number of employees within the ICT clusters and a number of tertiary education graduates far below the mean. Also Massachusetts, even though it has a level for BERD over the mean, still their innovativeness in ICT industry is not that significant (59 patent application per million inhabitants). In this case this can be due to a low number of companies within the clusters.
Table 4.2. Descriptive statistics Europe (N= 31 observations)
N Minimum Maximum Mean
Std. Deviation
Nb of patent applications 31 3 8249 767.61 1671.250
Nb of companies within the
cluster 31 44 29704 6707.13 8616.541
Nb of employees within the
cluster 31 869 404938 66234.23
101266.56 4
Tertiary education graduates 31 54 32383 5647.94 8930.795
GBAORD (million EUR) 31 1 3663 419.81 847.283
BERD (millions EUR) 31 6 8993 927.84 1813.875
GDP per capita (US dollars) 31 5163 103823 35089.26 23430.073
Regarding Europe, Germany is the country with the highest number of patent applications (8249 per million inhabitants), followed by France, U.K and The Netherlands, while Bulgaria, Iceland and Slovakia have the lowest number of patent applications. One explanation might be that Germany has the highest value for GBAORD (3663 millions EUR) as well as the highest number of employees within the ICT clusters, while U.K. is in top regarding the number of tertiary education graduates and BERD. While France registered high values for the number of employees within the clusters, The Netherlands has a high number of companies within the ICT clusters as long with a good score for tertiary education graduates. Latvia has the lowest value for BERD (6 millions), which can be a reason also for the low number of patent applications in ICT (only 9 per million inhabitant), Bulgaria and Iceland scored also low values for tertiary education graduates and number of companies within the clusters, which leads to a level of patent applications far below the mean.
companies or the number of employees within the clusters increase, then, the number of patent application will increase as well.
The number of patent applications is positively and significant correlated with the number of employees within ICT clusters with a Pearson’s correlation coefficient of r = 0.913 and there is a less then 0.001 probability of chance that a correlation coefficient this big would have occurred by chance in a sample of N = 39 countries. In the second case the number of patent application is also positively correlated with the number of companies in the ICT clusters with a coefficient of r = 0.787, which is also significant at p < 0.001.
Table 4.3. Correlation matrix U.S. (N=39)
Patent appl. Companies in the cluster Employees in the cluster Tertiary education graduates BERD ICT GDP per capita Pearson 1 Patent appl. Sig. Pearson .787** 1 Companies in
the cluster Sig. .000
Pearson .913** .814** 1
Employees in
the cluster Sig. .000 .000
Pearson .867** .695** .837** 1 Tertiary education graduates Sig. .000 .000 .000 Pearson .925** .729** .859** .861** 1 BERD ICT Sig. .000 .000 .000 .000 Pearson .879** .706** .870** .971** .827** 1 GDP per capita Sig. .000 .000 .000 .000 .000
**. Correlation is significant at the 0.01 level (2-tailed).
Furthermore, Table 4.4. presents the results of the correlation analysis conducted for Europe. In this case I have introduced an extra variable, GBAORD (Government Budget Appropriations or Outlays for R&D by socio-economic objectives), as data was available.
Also in the case of Europe, the values of Pearson’s coefficient indicate a positive relationship between the number of patent applications and both the number of companies and the number of employees within the ICT clusters. In both cases there is a positive and significant correlation between the variables, although, the number of companies in the ICT clusters is less correlated with the number of patent applications (r = .462).
in a sample of N = 31 countries. In the second case the number of patent application is also positively correlated with the number of companies in the ICT clusters with a lower coefficient of r = 0.462, which is also significant at p < 0.009.
When comparing Europe to U.S., we can see a higher correlation between the dependent variable, the number of patent applications, and the control variables: BERD (Business Expenditure on R&D), tertiary education graduates and GDP (Gross Domestic Product) in the case of U.S. As such, these variables have a more significant impact on the number of patent applications in U.S. than in Europe, especially in the case of BERD, which has a Pearson’s correlation coefficient of r = .925 (significant for p < .001).
Table 4.4. Correlation matrix Europe (N=31)
Patent appl. Companies in the cluster Employees in the cluster Tertiary education graduates GBAORD BERD ICT GDP Pearson 1 Patent appl. Sig. Pearson .462** 1 Companies in the cluster Sig. .009 Pearson .832** .726** 1 Employees in the cluster Sig. .000 .000 Pearson .689** .673** .808** 1 Tertiary education graduates Sig. .000 .000 .000 Pearson .717** .360* .655** .493** 1 GBAORD Sig. .000 .047 .000 .005 Pearson .599** .638** .847** .729** .423* 1 BERD ICT Sig. .000 .000 .000 .000 .018 Pearson .160 .059 .124 -.027 .082 .213 1 GDP Sig. .389 .754 .507 .884 .661 .250
**. Correlation is significant at the 0.01 level (2-tailed).
Table 4.5. Results of OLS regression for U.S.
Model Variables
Standardized Coefficients
(Beta) t - value Coefficients Sig. ANOVA Sig. R squared Adjusted
1 (Constant) -1.864 0.070
Nb of companies
within the cluster .129 1.125 0.268 .000a 0.831
Nb of employees
within the cluster .808 7.043 0.000
F =
94.45
2 (Constant) -3.263 0.002
Nb of companies within the cluster .115 1.112 0.274 .000b 0.863
Nb of employees
within the cluster .537 3.962 0.000
F = 81.01 Tertiary education graduates in Science and Technology .338 3.083 0.004 3 (Constant) -2.527 0.016
Nb of companies within the cluster .083 0.957 0.345
Nb of employees
within the cluster .349 2.859 0.007 .000c 0.905
Tertiary education graduates in Science and Technology .120 1.136 0.264 F = 91.55 BERD ICT (millions EUR) .461 4.047 0.000
a. Dependent Variable: Number of patent applications (per million inhabitants)
As we can see from Table 4.5., in the case of U.S., Hypothesis 1: The number of high-skilled employees within a cluster can positively influence a country’s innovative capacity, is supported in all the three models while Hypothesis 2: The number of companies within ICT clusters from a country can positively influence the national innovative capacity, is not. Also, the Tertiary Education Graduates seem to have a positive and significant influence on the number of patent applications but just in the 2nd model; in the 3rd model, after introducing also BERD (Business Expenditure on R&D) in ICT, the number of tertiary education graduates has no longer a significant impact on the dependent variable.
Table 4.6. Results of OLS regression for Europe
Model Variables
Standardized Coefficients
(Beta) t - value Coefficients Sig. ANOVA Sig.
Adjusted R squared 1 (Constant) .051 0.959 Nb of companies
within the cluster -.301 - 2.122 0.043 .000a 0.735
Nb of employees
within the cluster 1.050 7.416 0.000
F = 38.79
2 (Constant) - .201 0.842
Nb of companies
within the cluster -.314 - 2.289 0.030 .000b 0.779
Nb of employees
within the cluster 1.239 5.774 0.000
F = 22.94 Tertiary education graduates in Science and Technology .173 1.072 0.293 BERD ICT (millions
EUR) -.376 -2.146 0.041
3 (Constant) -1.134 0.268
Nb of companies within the cluster -.313 -2.308 0.030
Nb of employees
within the cluster 1.227 5.789 0.000 .000c 0.793
Tertiary education graduates in Science and Technology .226 1.373 0.182 F = 19.15 BERD ICT (millions EUR) -.433 -2.423 0.023
GDP per capita (US
dollars) .125 1.291 0.208
a. Dependent Variable: Number of patent applications (per million inhabitants) 5. DISCUSSION, LIMITATIONS AND FUTURE RESEARCH
In this chapter I will discuss the results of my analysis regarding the influence of ICT clusters on the number of patent applications, along with the implications and limitaions of this research. After analizing the results, some recommandations for future research on this topic will be made.
5.1 Discussion
industry. I also included in my study as control variables the number of tertiary education graduates and BERD (Business Expenditure on R&D) because according to the literature, the pool of human and financial resources devoted to scientific and technological advances are one of the main elements of a country’s innovative capacity (Furman et. al., 2002).
In the case of U.S., the regression results show that the size of a cluster does not have a significant (Prob.= 0.268; t-Statistic=1.125) impact on the national innovative capacity of a country in any of the three models. So, in this case Hypothesis 2 is not supported. On the other hand for Europe, the results show that Hypothesis 2 is supported and the size of a cluster has a positive impact on the number of patent applications (Prob. = 0.043). Even if the correlation between the number of companies within a cluster and the number of patent applications is higher for U.S than for Europe, it seems that in the case of U.S., not always when more companies are involved in the innovation process this will lead to a higher number of patent applications. This could mean that clusters in U.S. are more innovative than the ones in Europe, and can have a similar innovation output from a lower number of companies. As such, the existing literature, which states that cooperation between more companies within the cluster help them learn early about evolving technology (Fallah, 2005), as all the partners do get closely involved in the innovation process (Porter, 1998), is just partially supported and depends on more factors like the age of the cluster, or the way companies within the cluster devote resources to innovation, as well as the individual knowledge “stock” of employees within the specific cluster.
Furthermore, we should take into account also the common innovation infrastructure of a country when analyzing the impact of clusters on the national innovative capacity (Porter and Stern, 2002). A significant contribution to a country’s innovative infrastructure has the education system (OECD, 2010). In my analysis, Hypothesis 1 states that the high-skilled employees within the clusters along with the tertiary education graduates in mathematics, statistics, science and computing both have an impact on a country’s ability to innovate.
When looking at the regression results for U.S., the two variables indeed are significant (Prob. < 0.05); while the number of employees within a cluster is significant for all the three models which supports Hypothesis 1, the number of tertiary education graduates is significant just in the 2nd model, where only the independent variables are introduced. In the 3rd model, when I introduce BERD (Business Expenditure on R&D) as a control variable, which is also significant for Prob. < 0.05, the number of graduates is no longer significant.
graduates has a positive and significant influence on the number of patent applications only when BERD is not considered as a control variable. In the two models where both variables are introduced, only BERD influences the number of patent applications, while the tertiary education graduates is no longer significant. Furthermore, in the analysis for Europe I have introduced GDP per capita as an additional variable in the 3rd model. The reason was to see if this influences in any way my results. Even though this variable has not a significant influence on my dependent variable, introducing it in my model lead to an increase in the significance of the other variables on the number of patent applications compared to the previous models.
The results of my analysis support the existing literature regarding the positive influence that BERD has on a country’s innovative capacity (Porter and Stern, 2002; Furman et. al., 2002; Chang and Shih, 2003), but they do not support the findings of some studies on this topic, like the one of OECD (2006). According to this study, developing advanced research skills for those entering work as well as producing graduates with skills, knowledge and attributes, enable them to contribute to research and innovation. One explanation for this result could be that is not sufficient as a country to have an increased number of graduates in mathematics, statistics and computer science in order to be more innovative and thus have a higher number of patent applications. First, maybe just a small percentage of this number of graduates, actually work within an ICT cluster or maybe the country needs to combine more elements in order to increase its innovative capacity. The extent of IP (intellectual property) protection and openness to international trade, the share of research performed by the academic sector and funded by the private sector, the degree of technological specialization, and its individual knowledge “stock” can be such elements (Porter and Stern, 2002). Maybe if in my analysis I would have included more variables that relate to innovation and patent applications, maybe in that case the influence of tertiary education graduates could have been a positive one. This can be a good starting point for future research.
According to Porter and Stern (2002), countries can vary significantly in their historical ability to produce global innovations and also in their current resource commitment to innovative activity. As we can see from the results, the coefficients vary from one country to another, as well as from one continent to another.
5.2. Main findings and implications of the research
case of U.S., only the number of employees within the high-tech clusters seems to influence the innovative capacity of each state. In this case the results show that, most clusters in U.S. do not have necessary a high number of companies that collaborate, but rather have a higher number of high-skilled scientists and researchers that contribute to new product development.
When looking at the two control variables, Tertiary Education Graduates and BERD (Business Expenditure on R&D), for both Europe and U.S., if the first is significant only when the latter is not included in the regression, BERD seems to have a positive and significant influence on the number of patent applications in all the three models of my regression. Indeed, if companies devote more resources to research and development activities, then they will also be more innovative in terms of patent applications. On the other hand, having a high number of graduates in science and technology at the country level does not always mean that the innovative capacity of the clusters will increase.
Because of the specific industry chosen, ICT (Information and Communication Technologies) the time frame and the chosen varibles, this study comes in support of the existing literature regarding clusters’ role in the ability of a country to be more innovative. Doing a comparative analysis and using data colected for a more recent year (2007), brings new insights to this topic and gives a clearer picture on the clustering activities across continents.
5.3. Limitations and future research
One of the main limitations of my research relates to the lack of data at the cluster level. In order to have more relevant results regarding the effects clusters have on a country’s innovative capacity, it would have been useful to have data about the number of patent applications each cluster had in 2007. Also, relevant data about the amont of investments devoted to R&D by each company or cluster could have changed the results of this study. As this type of data is not available at the cluster level, I took into account the number of patent applications in ICT industry at the country level, considering that most companies in this field are geographically concentrated, so most innovations belong to them. As also the theory suggests, firms rarely innovate in isolation, but rather in networks of production (OECD, 1999). Future research can be done, when the data will be made available.
Europe and U.S., to see if there are significant differences regarding their innovative capacity and the factors that influence it.
A final limitation of my research is that my analysis doesn’t take into account all the possible variables that can influence the number of patent applications like: the expenditure on innovation made by each firm from the specific ICT clusters, ICT investments in each country, the role universities play in the innovation process and within a cluster (co-authorship of scientific articles). This can also be a starting point for future research.
Also another recommendation for future research on this topic is related to the way the skills and competencies required for innovation are measured. Maybe a future study could be made in order to fully exploit available matched firm-worker data to analyze the relation between skills, innovation and performance (Nås and Ekeland, 2009).
In order to obtain more relevant results regarding the impact of high-tech cluster on the innovative capacity of a country, other industries should be analyzed as well, and the panel data should be collected on a longer time period with more cross-sections included.
6. CONCLUSIONS
In this paper I have studied the influence of cluster specific characteristics on the national innovative capacity of a country measured by the number of patent applications. In this analysis it is of great importance also to take into account the common innovation infrastructure of a country. This innovation infrastructure includes all human and financial resources a country devotes to scientific and technological advances, the public policies bearing on innovative activity and the economy’s level of technological sophistication (Porter and Stern, 2002).
For my research I have chosen the ICT industry mainly because it has great potential to increase innovation by speeding up the diffusion of information, favoring networking among firms, reducing geographic limitations and increasing efficiency in communication (OECD, 2002). The time scale for this analysis will consist of one year, 2007, as data was available for all countries and states in my analysis.
Based on the existing literature and the available data, two hypothesis related to the impact cluster specific characteristics (its size or the number of its skilled employees) have on a country’s innovative capacity, were defined.
