Cluster's Influence on Internationalization and Firm Heterogeneity: The case of the Portuguese Footwear Cluster

Cluster’s Influence on Internationalization and Firm Heterogeneity: The case of the Portuguese Footwear Cluster Sandra Micaela Macedo Costa [email protected] Dissertation Master in International Business Supervisor Prof.ª Ana Paula Africano Joint Supervisor Prof. Paulo Teles September 2016 i Biography Sandra Micaela Macedo Costa is a BSc graduate in Economics, from the Faculty of Economics and Management at the University of Porto (FEP), since 2012. Presently she is finishing the MSc in International Business at the same Faculty. During the Bachelor at FEP, Sandra participated in some projects at AIESEC (a global platform for young people to explore and develop their leadership potential, on a nonpolitical, independent and not-for-profit organization that is run by students and recent graduates of institutions of higher education) and several events at FEP. As a 1st year student of the MSc, Sandra spent the second semester studying in Oslo - Norway, at the "BI Norwegian Business School through the "ERASMUS+" programme. In 2013, Sandra start her professional experience at "EXERTUS Consulting", a consultant company. Since 2014, she is working - in the Financial and Administrative areas - at "TYPOGRAPHIA, an international business project, based on a collective of young creative people from all over the world that develops a unique approach to the art of printing and designing t-shirts”. ii Acknowledgments Firstly, I would like to thank to my parents, sister and grandparents for always being by my side and supported me. A special thanks to my supervisor Professor Ana Paula Africano for her guidance in this dissertation and for all the time and work devoted to my investigation. Thanks to my co supervisor Professor Paulo Teles for all his help and guidance in the empirical work. I am grateful to APICCAPS, in particular to his Executive director João Maia, who helped me to get some important information about the footwear sector. I am particularly thankfull to all the firms, of the Portuguese footwear sector, that answered my questionnaire. Without their responses this investigation would have not been possible. I would like to thank to all my friends that supported and encouraged me. Thanks also to my friends from Criar Asas, from my MSc and from my Work "TYPOGRAPHIA". To Lúcia and Carlos, my best friends, thank you for your concern and kindness during this journey. I would like to express my most sincere appreciation to my boyfriend, José, for his support, patience and comprehension. iii Abstract Globalization and localization are two important topics of research with a considerable amount of literature. However, there is an area where the literature is scarce: how industrial clusters influence the internationalization of its firms. According to the existing literature industrial clusters are important facilitators to the internationalization process of firms within a cluster, but how they can help those firms i.e. which characteristics and specific mechanisms are behind the international success of its firms remains unclear, which make this study relevant. The research questions underlying this dissertation are: “How Clusters can influence the internationalization process of its firms? Do all firms take similar advantage of belonging to a cluster?” This study aims to clarify those interconnections, and to understand how firms' heterogeneity may condition their internationalization process. For that purpose this research defines a theoretical framework and a set of hypotheses, based on the literature review. Then the empirical application is a case study about the Portuguese Footwear's Cluster. A questionnaire was applied to a representative sample of firms and the respondents’ data were subject to Factorial analysis and Cluster analysis. The conclusions are that, in fact, the cluster has an influence in the internationalization of its firms through some specific features, and, that firm's heterogeneity influence their process of internationalization which means that firms do not take the same advantage of belonging to a cluster. Keywords: cluster; internationalization; networks; resources; heterogeneity; Portuguese Footwear Cluster. iv Table of Contents Biography ......................................................................................................................... i Acknowledgments ...........................................................................................................ii Abstract .......................................................................................................................... iii Table of Contents ........................................................................................................... iv Index of Tables ............................................................................................................... vi Index of Figures ............................................................................................................. ix Introduction ..................................................................................................................... 1 1. Literature Review .................................................................................................. 3 1.1. Firm’s Heterogeneity ............................................................................................. 3 1.2. Clusters and Competitiveness ................................................................................ 6 1.3. Clusters and Internationalization ........................................................................... 7 1.3.1. Clusters, Networking and Resources ..................................................................... 9 1.4. Review of Empirical Studies ............................................................................... 13 2. Empirical Study on the Internationalization Process of firms within a Cluster and firm’s heterogeneity: The case of the Portuguese Footwear Cluster ................................................................................................................. 19 2.1. Methodology ........................................................................................................ 19 2.2. Hypotheses – Survey Guidelines ......................................................................... 21 2.3. Why the Portuguese Footwear as an empirical base? .......................................... 24 2.3.1. Surveys Procedure ............................................................................................... 30 3. Empirical Results and analysis ......................................................................... 31 3.1. Exploratory analysis – Characterization of firms ................................................ 31 3.1.1. Position in the firm .............................................................................................. 31 3.1.2. Firm Age .............................................................................................................. 32 3.1.3. Portuguese Classification of Economic Activities (CAE) ................................... 33 3.1.4. International activity age ..................................................................................... 34 3.1.5. Geographic location of the firm ........................................................................... 36 3.1.6. Number of employees .......................................................................................... 36 3.1.7. Turnover ............................................................................................................... 37 3.1.8. Number of export markets ................................................................................... 37 3.1.9. Weight of exports in sales .................................................................................... 38 3.1.10. Main export markets ........................................................................................... 40 3.1.11. Percentage of Human Resources with Higher Education ................................... 40 2 (financial resources, human resources, technology, Research & Development, productivity) (Giovannettia, Ricchiutia & Velucchi, 2013). In particular, it seems important to understand the concrete ways in which these networks can contribute to the international development of SMEs and understand why not every firm inside the same Cluster has the same level of internationalization. Following this pattern and with the aim to fulfill the gaps in the literature, this dissertation seeks to address the research question: “How Clusters can influence the internationalization process of its firms? Do all firms take similar advantage of belonging to a cluster?” In doing so this dissertation intends (i) to contribute to the academic discussion about cluster’s , (ii) to understand how industrial clusters influence the internationalization process of firms i.e. which cluster’s characteristics are crucial to facilitate firm’s internationalization, (iii) to analyze within a cluster which type of firms are more likely to internationalize and (iii) to develop an empirical study based on a Portuguese cluster, through questionnaires directed to a sample of firms from the Portuguese footwear cluster. The aim of this empirical study is to elaborate the best practices that a cluster model can have to influence its firm’s internationalization and to analyze which firms are more likely to internationalize. At the same time, if applicable, this cluster could be a role model to others clusters and a contribution to improve the design of public policies. This dissertation is organized as follows: a review of relevant literature, which is presented in chapter 1, important to understand what has been studied and to point out the crucial factors about internationalization of firms and in which way clusters work as facilitators in the internationalization process of its firms. It includes a review of empirical studies very important to analyse the variables that authors have considered in their researches and conclusions. Chapter 2 presents the empirical part of this dissertation which includes the methodology, hypotheses and the reason why the Footwear Cluster is an interesting case to study in this dissertation. Chapter 3 presents the empirical results and analysis of data collected from firms of the Portuguese Footwear Cluster through questionnaires. Finally the main conclusions are presented. 3 1. Literature Review This chapter aims to provide a theoretical framework to this research through literature review and relevant concepts related with clusters and internationalization (Exports and Foreign Direct Investment). The first topic is firm’s heterogeneity and its differences which influence its capabilities of entry into foreign markets even if they are in the same industrial cluster. This topic is very important to understand that not all firms can internationalize, even if they have access to the same resources inside the cluster. The section about “clusters and competitiveness” presents the concepts of industrial clusters and describes clusters as a support for firm’s competitiveness. The last sections emphasize the relationship between clusters and internationalization and how they are related, as well the influence by resources and networking as facilitators of firm’s internationalization within a cluster. The literature review is structured as described above to enable us to respond important questions according to the research question such as: (i) Does the cluster’s influenced all firms despite its heterogeneity? (ii) Which are the facilitators of internationalization that clusters offer? (iii) Which are the main aspects that firms have to improve to successfully internationalize? (iv) Which type of firms influence others to internationalize inside the cluster? 1.1. Firm’s Heterogeneity In this era of globalization and trade liberalization, new literature approached the concept of firm’s heterogeneity, i.e. firms that differ from each other in terms of size, productivity, different levels of technology and firm-specific learning processes (Aiello & Ricotta, 2016). A study by Aiello & Ricotta (2016) measures the influence that location has on firms’ heterogeneity. The results indicate that heterogeneity in firm’s productivity is affected by specific factors internal to the firms. Regarding the influence of location in firms’ heterogeneity, the same study shows that different levels of productivity among firms can be explained by differences across countries (Aiello & Ricotta, 2016). 4 Moreover, in a world of internationally competitive markets not all firms are able to internationalize and regardless of the business environment in which firms operate they differentiate from each other by its internal resources. On the one hand, there are firms that belong to international industries which can enable them to penetrate more foreign markets and, on the other hand, firms may belong to highly domestic industries making it difficult their international growth. But the same can happen inversely, i.e. firms within an international industry that aren’t in foreign markets and international firms that belong to a not internationalized industry (Greenaway & Kneller, 2007). The approach about heterogeneity of these authors is more restricted highlighting the firms’ trade-off between exports and Foreign Direct Investment (FDI). According to Greenaway and Kneller (2007) firms that export and firms that do not export co-exist in the same industries, the market entry mode 1 is not only related with the firm, and its industry, but with the market itself, as well. The question arises, why firms choose to export rather than engage in production in foreign markets. The reason is explained by sunk costs and productivity heterogeneity (Greenaway & Kneller, 2007). We are moving from the new trade theory where all firms export, to one, in which, some firms export and others do not because of their heterogeneity. According to Wagner (2007, 2012) and Greenaway and Kneller (2007), one evidence becomes clear, firms that export or import are more productive than non-exporters and non-importers, not necessarily as a result of exporting but because they have the necessary capabilities to enter in foreign markets through exports, namely to support the associated sunk costs. Melitz (2003) and Bernard (2003) created a model that explain the linkage between firm heterogeneity and industry productivity with exporting being the key factor. The model represents the relationship between firm productivity and probability density of productivity. Firms that export have a higher level of productivity. In opposite, domestic firms have lower productivity. The threshold to export becomes bigger when the probability density of productivity decreases and the firm productivity increases. Others recent authors extended this to consider asymmetries between countries, for instance, in 1 The market entry mode in Greenaway and Kneller (2007) article refer to exports and FDI (Foreign Direct Investment) 5 competition (Melitz and Ottaviano, 2008) and the efficiency (Falvey, Greenaway, & Yu, 2004). This model illustrates, in an indirect way, that micro-heterogeneity influences aggregate outcomes i.e. when trade policy barriers fall, exporting firms with higher levels of productivity will survive and those with lower productivity and that do not export will shrink or exit (Bernard, Jensen, Redding, & Schott, 2012). In other words, due to increased competition “trade liberalization raises average productivity through reallocations of resources across firms within industries” (Bernard et al., 2012, p. 25). In their study, Greenaway and Kneller (2007) show that firms with propensity to export tend to be larger and more productive than non-exporters, however, sunk costs are important to the decision whether firms export or invest in foreign markets. Exports involve lower fixed costs (sunk costs) and FDI (Foreign Direct Investment) involves lower variables costs (Greenaway & Kneller, 2007; Helpman, Melitz, & Yeaple, 2004). Moreover, a study by Giovannettia et al. (2013) analyses the impact of the local’s context-related such as industrial districts and infrastructures on the firms’ internationalization process set in that localization, besides the firms’ specifics characteristics (size; firms’ productivity; Technology; R&D activities: patents and new products; expenditures). The main idea is that few firms are able to compete in international markets, and these firms are more productive and competitive than the domestic ones. “The performance of firms in a globalized world depends on firms’ specific characteristics, on their flexibility to react to market changes but also on the socio-economic environment” (Giovannettia et al., 2013, p. 2666). 6 Table 1 Synthesis of determinants of Internationalization among firms Author Influencers of Internationalization Correlation Aiello & Ricotta (2016) Firm heterogeneity in productivity + Wagner (2007) Wagner (2012) Firm level of Productivity + Bernard, Jensen, Redding, & Schott (2012) Trade Liberalization + (raises productivity through reallocations of resources across firms within industries) Greenaway & Kneller (2007) Firm level of productivity + Sunk Costs - Giovannettia, G., Ricchiutia G. & Velucchi M. (2013) Firms’ specific characteristics - size, productivity, technology, R&D activities (patents and new products), expenditures + Flexibility to react to market changes + Socio-economic environment + Source: Own elaboration based on the literature of this subchapter 1.2. Clusters and Competitiveness After focusing on firm’s heterogeneity (in chapter 1.1.) and how its level of productivity influences its propensity to export (internationalize) is crucial to link that topic with clusters and its impact on firm’s competitiveness and productivity. In the last decades, clusters have been seen as a way to improve the competitiveness, productivity and innovativeness of SMEs, overcoming its size restrictions (Karaev, Koh & Szamosi, 2007). According to Porter (1990) clusters are formed by firms and industries linked through vertical and horizontal relationships located in the same place. In Porter (1998)’s study his definition about clusters is extended by including institutions as universities defining it as “geographic concentrations of interconnected companies and institutions in a particular field” (Porter, 1998, p. 78). Geographical concentration such as clusters are considerate a support to competitiveness of firms, because of its better and faster access to innovation also it enable companies to reduce 7 input-costs and get co-operative relationships which can create competitive advantage (Porter, 1998). Competitiveness depends on firm’s productivity and firms can reach productivity through technology, sophisticated methods and the launch of unique products and services. However, these measures and its leverage are influenced by the local business environment and location of firms. In Porter (1998)’s perspective clusters affect competition “by increasing the productivity of firms; by driving the direction and pace of innovation, which underpins future productivity growth and; by stimulating the formation of new businesses, which expands and strengths the cluster itself” (Porter, 1998, p. 80). Productivity of firms within a cluster increases because they have access to more inputs, suppliers, information, technology and network. Moreover, Porter (1998) also do recommendations concerning industrial policies in which governments must to support and help the growth of firm’ productivity setting the rules of competition and physical infrastructures including all clusters’ sectors, specially the traditional ones such as agriculture because every cluster affects not only the national productivity but also other clusters productivity. Protect intellectual property, enforce antitrust laws, promote cluster formation and upgrade and buildup public goods that have a significant impact on linked business are the procedures proposed by Porter (1998). 1.3. Clusters and Internationalization One of the clusters’ definition by Porter (1990) is that clusters are concentrated in a relatively small area of specialized suppliers, universities, cooperatives, an experienced workforce, domestic and international companies, distribution channels and logistics services that serves international markets. However, as Porter and Ketels (2009, p.174) explained, clusters have different configurations from each other besides the influence that a specific sector has on the cluster performance. Clusters can be distinguished from each other: several are developed from SME networks, others are linked to a central firm who brings out the cluster ensuring basic creation of new businesses or attracting suppliers, and others have developed around universities in which human capital and ideas of researchers led to many spin-offs. Moreover, according to Dubé, Haijuan & 8 Lijun (2015), industrial clusters are a crucial factor to the internationalization process of its firms but it will depend on cluster governance 2 i.e. the cluster composition, the internal network density and the degree of knowledge sharing. The results of this research illustrates that a cluster with higher diversity of internal resources, stronger internal network and deeper knowledge sharing has higher degree of internationalization and consequently firm inside this cluster have access to more resources which can help them in its internationalization process. Clusters can offer opportunities and facilitate the internationalization process to their member firms (Javalgi, Griffith, & White, 2003). According to Libaers and Meyer (2011) “in order to initiate the internationalization process, firms need access to resources, and these resources may reside within the firm and in the immediate external environment i.e. industrial cluster” (Libaers & Meyer, 2011, p. 1433). For example, firms located in clusters have easier access to venture capital and transfer of knowledge, which can increase their technological capacity (Fernhaber et al., 2008). Moreover, subsidiaries of multinational firms are often located in clusters (Birskinshaw & Hood, 2000; Shaver & Flyer, 2000) and co-location with these companies increases the interest of entrepreneurs for opportunities in international markets (Vernon, 1966), as well as, the knowledge of these opportunities (Karagozoglu & Lindell, 1998; Westhead, Wright & Ucbasaran, 2001). A good firm reputation also permits to establish trust with other firms, this confidence is essential to share information and create closer bonding between actors (Saxenian, 1990; Johanson & Vahlne, 2009; Colovic, 2010). These networks can facilitate the internationalization of small firms looking for opportunities. Similarly, a solid cluster reputation can help the internationalization process of firms, particularly small and medium enterprises (Zyglidopoulos, De Martino & McHardy Reid, 2006). Literature suggests that networks can play an important role in the internationalization process of firms, specially SMEs. In addition, the integration of SMEs into territorial networks seems to facilitate the approach to foreign markets because the clustering allows them to access the necessary resources to expand internationally. 2 This topic is deeper explained in chapter “Review of Empirical Studies” in this dissertation 9 Table 2 Synthesis of Clusters and Internationalization Author Influencers and Facilitators of Internationalization within a cluster Libaers and Meyer (2011) Leverage of cluster-based resources Level of Inventive Prowess Fernhaber, Gilbert & McDougall (2008) Venture Capital Transfer of knowledge Technological capacity Vernon (1966) Co-location with multinationals (increases the interest of entrepreneurs for opportunities in international markets) Saxenian (1990) Johanson & Vahlne (2009) Colovic (2010) Firm Reputation (permits to establish trust with other firms and these networks can facilitate the internationalization of small firms) Zyglidopoulos, De Martino & McHardy Reid (2006) Solid Cluster Reputation Dubé, F. N. Haijuan, Y., & Lijuan, H. (2015) Cluster Governance - Cluster Composition - Internal Network Density - Degree of Knowledge Sharing Source: Own elaboration based on the literature of this subchapter 1.3.1. Clusters, Networking and Resources Besides the individual network of firms, researchers examined the potential role of localization within an industrial agglomeration (such as industrial districts 3 , local productive systems, etc) in the internationalization of small firms. Thus, companies located in an industrial district seem to be more “global” (Illeris, 1992) and with higher export results than firms located outside districts (Becchetti & Rossi, 2000; Mittelstaedt et al., 2005). Moreover, Deshais, Joyal & Julien (1992) pointed out that 43.5% of exporting SMEs in three Quebec regions used the resources available in their location, being able to take benefits of their environment to expand internationally. The environment and location of firms seem to play an important role in the 3 Industrial districts was initially introduced by Marshall (1920) in Principles of Economics. 10 internationalization process of these firms. For instance according to Mariotti, Mutinelli, & Piscitello (2008) industrial districts have different structural and behavioural determinants which influence their performance and internationalization initiative. Firms located in an industrial district appear to have well developed links with local firms within the district (Johanisson, 1994) suggesting that the internationalization efforts and their results are disseminated locally and internationalization becomes a collective property. Firms that have achieved high levels of internationalization seem indeed to have significant levels of local networking, particularly in the area of cooperation in Research & Development (R&D), internationalization is embedded on the local network and collaboration in research (Keeble, Lawson, Lawton Smith, Moore & Wilkinson, 1998; Libaers & Mayer, 2011). However, according to size and age, the more the company is small and new, the more it is embedded in the local environment, while larger and more experienced companies are less dependent on local network (Keeble et al., 1998). These results suggest that the industrial agglomerations possess the characteristics of a defined “environment to internationalize” (Torrès, 1999), defined as “the set of actors and factors that facilitate the internationalization process of SMEs and the local business community” (Fourcade & Torrès, 2003, p. 3) or defined as “a sustainable cooperation system in which local actors (SMEs, local authorities, public or semi-public institutions, university research centers, banking systems) are working together to create a dynamic of internationalization” (Torrés, 2003, p. 29). SMEs can indeed develop their competitiveness in global markets from a strong local integration (Torrès, 2003). Thus it would seem that in general, the position of SMEs within industrial agglomeration positively influences internationalization (Belso-Martinez, 2006). In the Uppsala internationalization model (Johanson & Vahlne, 2009) the authors argued that the integration in networks will allow a company to successfully internationalize through trust, learning and development of opportunities in an environment that facilitates the entry in foreign markets. The role of networks is particularly important for the internationalization of SMEs due to their lack of resources and skills to develop internationally (Coviello & Munro, 1995, 1997; Chetty & Blankenburg Holm, 2000; Lu & Beamish, 2001, 2006). Fernhaber et al. (2008) stated 11 that the combination of the resources benefits within a cluster with the importance of resources to the internationalization process suggests that clusters with higher concentration of industry enable easily firms to internationalize its operations because of its higher availability of resources. Moreover, according to Welch & Welch (1996) “the development and utilization of foreign networks is, of course, closely related to the learning process that underlies overall internationalization. Indeed, an important part of a company’s knowledge is often created and maintained through actors in its relevant networks” (Welch & Welch, 1996, p.12). These actors can be foreign intermediaries, customers, alliance partners, suppliers, government officials and others entities. Strong international networks seem well be one of the important characteristics of a global approach (Oviatt & McDougall, 2005). The selection of foreign markets and the participation in international activities emanate from the opportunities created by network contacts, not only from managers’ strategic decisions (Coviello & Munro, 1995). The network of relationships, including those in the country of origin, (Lin & Chaney, 2007; Zhou, Wu & Luo, 2007) trigger and motivate internationalization of firms, influence their market selection and the input mode, help them gain credibility, allow access to other established relationships and channels, help to reduce costs and risks, and finally influence their rhythms and patterns of internationalization (Zain and Ng, 2006). A study by Zen, Fensterseifer, and Prévot (2011) analyzes the resources generated by the cluster and the influence on its firm’s internationalization process, affirming that firms within a cluster have access to more resources that can help them on its internationalization process comparing with those that are not inside the cluster. Other aspect of its conclusions is that managers should consider as well the influence of the cluster’s reputation and the importance of internal network into firm’s internationalization strategy. 18 Table 4 Review of Empirical Studies Authors Topic Empirical Variables Methodology Dubé, F. N. Haijuan, Y., & Lijuan, H. (2015) Level of Internationalization Cluster Governance - Cluster Composition - Internal Network Density - Degree of Knowledge Sharing 1) Cluster measurement tool developed by the European Union Task Force Group (ECA)(TACTICS Reflection Group, 2010). 2) Quantitative methods Giovannettia, G., Ricchiutia G. & Velucchi M. (2013) Firms Heterogeneity Firms Heterogeneity - Firm’s specific characteristics - Flexibility to react to market changes - Socio-economic environment Multilevel Approach Libaers, D., & Meyer, M. (2011) Level of Inventive Prowess Leverage of clusterbased resources Firm Internationalization Firm Performance Firm’s International Intensity Econometric Model Mariotti, S., Mutinelli, M., & Piscitello, L. (2008) Internationalization of production through FDI Structural determinants - Presence of Leader Firms in the Cluster - Role of the Leader Firms within the cluster - Degree of domestic rivalry - Presence of foreign MNCs Behavioural determinants - Cluster’s propensity to export - Internationalization experience of firms in the cluster - Cluster’s innovative capacity Econometric Model Zen, A. C., Fensterseifer, J. E., & Prévot, F. (2011) Influence of cluster resources on the internationalization of clustered companies Knowledge and resources sharing Cooperative relations Cluster reputation Case study through surveys Zyglidopoulos, S. C., DeMartino, R., & Reid, D. M. (2006) Impact of cluster’s reputation on internationalization Cluster Reputation Literature review Source: Own elaboration based on the literature review 19 2. Empirical Study on the Internationalization Process of firms within a Cluster and firm’s heterogeneity: The case of the Portuguese Footwear Cluster This chapter presents the empirical part of this dissertation. The research question of this dissertation forwarded us to an investigation work conception which allowed to find the answers to the questions related with the influence of clusters on its firm’s internationalization and firm’s heterogeneity which can influence its patterns of internationalization. The first section describes the methodology chosen for this dissertation explaining why it is the most appropriate for this dissertation. The following section approaches the hypotheses which are the base for the survey’s questions. The third section explains why the Portuguese Footwear Cluster is the case study chosen for this dissertation. 2.1. Methodology The literature review in chapter 1 and the parallelism among cluster, internationalization, networking and resources allowed to plan precisely the objectives of this study and define the topics to conduct the survey (see Table 5 Theoretical basis for the Research Variables). This subchapter presents the methodology chosen to this dissertation according to the research question and the objectives established. The goal of this dissertation is to understand which cluster’s characteristics and behaviour are crucial to the internationalization process of firms within the cluster, i.e. how clusters can influence the internationalization process of its firms and the relevance of firms’ heterogeneity. In this case the aim of this dissertation is to study and analyze how the Portuguese Footwear Cluster influences the internationalization process of its firms. The methodology chosen is quantitative based on data collected through surveys to the selected firms inside the Portuguese Footwear Cluster. The survey research was considered the most appropriate to analyze how the Portuguese Footwear Cluster influence its firms and which mechanisms and methods are used. 20 “The survey approach refers to a group of methods which emphasize quantitative analysis, where data for a number of organizations are collected through methods such as mail questionnaires, telephone interviews, or from published statistics, and these data are analyzed using statistical techniques” (Gable, 1994, p. 113). Considering the purpose of this investigation by studying a representative sample of firms, which in this case are firms within the Portuguese Footwear Cluster, the survey approach will seek to discover relationships that are common across those firms and therefore to provide generalizable statements about the object of this investigation. Through this type of methodology will be possible to answer the research question of this dissertation: “How Clusters can influence the internationalization process of its firms? Do all firms take similar advantage of belonging to a cluster?” Table 5 Theoretical basis for the Research Variables Empirical Variables (points to inquire) Theoretical basis Cluster Governance - Cluster Composition - Internal Network Density - Degree of Knowledge Sharing Dubé, F. N. Haijuan, Y., & Lijuan, H. (2015) Structural determinants - Presence of Leader Firms in the Cluster - Role of the Leader Firms within the cluster - Degree of domestic rivalry - Presence of foreign MNCs Behavioural determinants - Cluster’s propensity to export - Internationalization experience of firms in the cluster - Cluster’s innovative capacity Mariotti, S., Mutinelli, M., & Piscitello, L. (2008) Knowledge and resources sharing Cooperative relations Clusters Reputation Zen, A. C., Fensterseifer, J. E., & Prévot, F. (2011) Clusters Reputation Zyglidopoulos, S. C., DeMartino, R., & Reid, D. M. (2006) Firms Heterogeneity - Firm’s specific characteristics - Flexibility to react to market changes - Socio-economic environment Giovannettia, G., Ricchiutia G. & Velucchi M. (2013) Source: Own elaboration based on the literature review 21 2.2. Hypotheses – Survey Guidelines The literature review enabled to elaborate hypotheses related to the relationship and the influence of cluster’s characteristics on firm’s internationalization considering the firms heterogeneity as defined below. These hypotheses will be used in the survey’s questions to the selected firms inside the Portuguese Footwear Cluster in which firms will be asked to evaluate the degree of disagreement/agreement in a Likert scale from 1 to 5 4 in some of the questions. The results will allow to understand the correlation of each hypothesis enabling to answer the research question of these dissertation: “How Clusters can influence the internationalization process of its firms? Do all firms take similar advantage of belonging to a cluster?” H1: Cluster Governance 5 has a positive relationship with the internationalization process of its firms H1 a: The share of important and rare resources among firms within the same cluster can ease their internationalization process H1 b: Knowledge share inside the cluster has a positive impact on the internationalization process of its firms H1 c: Cluster’s cooperative relations and networking are positively associated with the internationalization process of its firms H2: The internationalization experience of firms in the cluster enhances the willingness to internationalize of others firms within the cluster H3: The presence of Leader firms 6 in the cluster influence positively the internationalization of others firms in the cluster 4 1 corresponds to “strongly disagree”; 2 to “disagree”; 3 to “neither agree nor disagree”; 4 to “agree”; 5 to “strongly agree” 5 Cluster Governance is defined as the cluster composition, the internal network density and the degree of knowledge sharing. The cluster composition is the internal structure of a cluster and how it influences the firm’s ability to achieve and maintain profitable market positions through access to internal and external resources. The internal network density is the impact that inter-organizational and interpersonal relationships (informal and formal) have on firms’ internationalization. The degree of knowledge sharing is related to the flowing and knowledge shared like the achievements and experiences from others firms and how it can promote firm internationalization. (Dubé, Haijuan, & Lijuan, 2015) 22 H4: The Degree of domestic rivalry among firms inside the cluster influences their internationalization process H5: The presence of foreign MNCs inside the cluster has a positive impact on the internationalization of its firms H6: A cluster’s good reputation facilitates the internationalization process of its firms H7: Cluster’s innovative capacity has a positive impact on its firms and facilitates the internationalization of its firms H8: Heterogeneity of firms is relevant to understand why firms within a cluster have different levels of Internationalization After the hypotheses’ elaboration, the questionnaire was developed and divided into six main parts. The survey starts with questions internal to the firm to analyse its characteristics with the aim to identify the heterogeneity among firms; the second part aims to understand which entities in the geographical proximity have more interaction with the firm, and are more relevant for its business. The following part has questions related to the hypotheses previously elaborated (see Table 6 Hypotheses vs Questions). These questions have the aim to understand the general perception that firms have regarding the influence that the cluster has in their internationalization process. The fourth part, looks for a more personal opinion, of each firm, about the factors that are more important to its internationalization process. The next set of questions aims to know the firms' perception about the role that the cluster has in their internationalization, namely in terms of human and financial resources, productivity improvement among others. The survey ends with the assessment/opinion of each firm on the three main traits/ actions that the cluster should have to facilitate the internationalization process of firms. 6 Leader firms have high growth rates and are the engines of local industrial development as they generate innovation, enlarge and open new markets, and favour human capital spillovers. They also develop international production networks and implement multinational market-seeking strategies. (Mariotti, Mutinelli, & Piscitello, 2008) 23 Table 6 Hypotheses vs Questions Hypotheses Survey’s Question Author H 1 a “A partilha de recursos importantes e raros entre empresas do mesmo cluster facilita o processo de internacionalização das mesmas” Dubé et al. (2015) H 1 b “A partilha de conhecimento dentro do cluster incentive a internacionalização das suas empresas” Dubé et al. (2015) H 1 c “As Relações de Cooperação e Networking dentro do Cluster estão positivamente associados ao processo de internacionalização das suas empresas” Dubé et al. (2015) H 2 “A experiência internacional das empresas dentro do Cluster aumenta o interesse/ vontade das menos experientes em iniciar atividades internacionais” Mariotti et al. (2008) H 3 “A presença de Empresas de Referência dentro do Cluster influencia positivamente a internacionalização de outras empresas pertencentes ao Cluster” Mariotti et al. (2008) H 4 “A intensidade competitiva entre as empresas do Cluster influencia positivamente o processo de internacionalização das mesmas” Mariotti et al. (2008) H 5 “A presença de Multinacionais estrangeiras dentro do Cluster tem impacto positivo na internacionalização das restantes empresas” Mariotti et al. (2008) H 6 1) “A boa reputação internacional do Cluster funciona como incentivo para a internacionalização das suas empresas” 2) “A boa reputação internacional do Cluster facilita o processo de internacionalização das suas empresas” Zyglidopoulos et al. (2011) Zen et al. (2011) H 7 “A capacidade de inovação do Cluster tem um impacto positivo na internacionalização das suas empresas” Mariotti et al. (2008) H 8 “A diversidade das características das empresas dentro do Cluster faz com que cada uma delas tenha diferentes capacidades de internacionalização” Giovannettia et al. (2013) Source: Own elaboration 24 2.3. Why the Portuguese Footwear as an empirical base? The empirical part of this dissertation is a case study about the Portuguese Footwear Cluster based on the study of some of its firms. This Cluster fulfills the essential condition of geographical concentration. The footwear production in Portugal has a strong concentration in two regions, Felgueiras and São João da Madeira being important to the regional economic activity. Moreover, in this two regions and in firms located there, there is complex network, both formal and informal: commercial relations, in particular subcontracting arrangements, and relations with knowledge and information share. The Portuguese Footwear cluster has its own institutional support – Business Association, Technological Center and Professional Training Center (APICCAPS) – whose action is recognized, nationally and internationally. Its production is more than leather, horizontally, the industry produce other shoes that use different raw materials and technologies: safety, sports and others. Vertically, the industry production also extends to the industries of leather goods and footwear components. Over the last decades, the cluster has observed a narrowing of relations with its whole value chain linked to the footwear. Some relationships has been intensifying: equipment industry, fashion accessories industry, other suppliers, footwear distribution firms and firms linked to the world of fashion and design. Footwear Industry Performance Despite the strong economic crisis, the Portuguese footwear industry has improved its value chain especially in international markets. Portuguese firms continued to invest and strengthen their capabilities in innovation, design and fashion, and to invest in international marketing through its presence in international fairs. Some of these firms went forward and created its own brands. At the same time, the Portuguese footwear has changed radically its image, investing on a very bold look and slogans as “designed by the future” and “the sexiest industry in Europe”. (APICCAPS, 2013) The international reputation of Portuguese footwear has changed and nowadays Portugal is seen as the origin of fashion and design. This positive image was reflected in 25 the average price of exported Portuguese shoes that increased from 18 euros, in 2006, to 23 euros, in 2012. The Portuguese footwear has disputed the top of global rankings against countries with stronger traditions of fashion and design. (APICCAPS, 2014) According to APICCAPS, the Portuguese Footwear, Components and Leather Goods Manufacturers’ Association, during 2014 the industry managed to increment sales abroad by 8%. Portuguese footwear industry’s exports continued its dynamic performance in 2014 totaling 1,907.5 million euros and going up from 1,779.1 million euros registered in 2013. Moreover, according to APICCAPS (2014), since 2010 the Portuguese footwear industry has shown a sustained growth tendency and during that period, employment rose by 7.7% and the level of production by 19.6%. At the end of 2013, the industry employed more than 35,000 people and its annual production exceeded 75 million pairs. A good performance in the European Union markets, to where exports grew roughly 7%, is complemented by excellent growth of 12% in markets outside the Union, with a global increase of footwear exports of over 8%. The European Union economies bought 1,622.7 million euros worth of Portuguese footwear in 2014, while other markets purchased Portuguese footwear valued at 1,869.8 million euros. (APICCAPS, 2014) The main destination markets for Portuguese footwear continue to be France, Germany, Netherlands, Spain and United Kingdom, all with positive growth rates. Also it is noticeable the good performance of the Portuguese footwear exports to the American market, with a growth of 51% in the period. According to the Portuguese association: “Since 2009 Portuguese footwear exports increased roughly 54%”. APICCAPS continues to work with its members on a strategy to diversify away the destination markets, as one of the goals of the Portuguese footwear industry is to deepen the presence in new markets, so that extra EU exports represent 20% of the total sales abroad by 2020. 26 Footwear Cluster Capabilities According to APICCAPS (2013), the Portuguese Footwear Cluster’s image is growing and become increasingly recognized, nationally and internationally. Some factors explain why this is happening: the cluster has (1) a diversified industrial base with recognized manufacturing capacity based on flexibility and readiness, (2) a know-how regarding to acting in international markets because of its 20 years of experience and investment in internationalization process, (3) a growing international reputation of Portugal seen as the origin of fashion and design products, and (4) a heritage complicity between industrial identities and institutional identities that support this cluster giving coherence to its action. These facts distinguish the footwear industry from others Portuguese economic sectors. For more than 20 years, the footwear industry follows an international strategy with the aim of became an important player in different markets and prosper its value chain. In the global footwear map, Portugal was seen, 30 years ago, as the location to mass production of low cost. The international trade liberalization enabled new locations with these competitive advantages, especially in Asia, leading to the relocation of some of foreign producers present in Portugal. In this new competitive world, the Portuguese footwear industry wanted to claim its position: the industry has reinvented itself focusing on readiness and flexibility. With the support of Technological Center and others institutional identities, the industry reorganized and reequipped itself being able to respond to any order, no matter how small it was. The industry strengthened its capacity in product development and enhanced their quality standards. Today, the Portuguese footwear industry is recognized for its manufacturing capabilities which is one of its strengths. But this process forced firms to assume new responsibilities and change the market approach. Firms had to do a continuous and persistent work regarding to internationalization and commercial promotion. Hence it resulted in other cluster’s essential asset: an accumulated capital of know-how about acting in international markets. This asset is crucial to the footwear exports be able to achieve 90% of national production. 27 Flexibility, readiness and intensive commercial action aren’t enough competitive factors, firms are sophisticating its offer through the investment in style, design and creation of collections and own brands. The reputation of Portugal as the origin of footwear quality and fashion is nowadays a precious asset to the cluster. However, firms within the cluster are in different stages in this process of evolution on the value chain. Transforming one firm accustomed to be subcontracted by foreign buyers to mass-produce into a firm that produces and sells its own brands it’s a hard and time consuming process which not all firms are capable. Some of these firms are still focused in undervalued products, others doesn’t have the financial means needed to create and develop its own brand or to internationalize. Other problem that some firms have is the lack of management control and a weak organization and human resources without the necessary skills and competences. According to the European Commission, regarding to its analysis about problems that Portugal faces until 2020, this is a general problem in the Portuguese economy and the footwear industry is not an exception. The cluster’s strategy for the coming years must include the firm heterogeneity: there’s no single business model suitable to all firms. But this heterogeneity has its virtues that enable firms to structure networks with different roles and responsibilities. This is a crucial mechanism to the knowledge share inside the cluster. The footwear industry is constituted, almost exclusively, by small and medium enterprises for this reason, APICCAPS, the Portuguese Footwear, Components and Leather Goods Manufacturers’ Association, has been the center of all collective strategies that can be assumed among all firms inside the industry. Strategies that stimulated network and cooperation to overcome the small dimension’s inconvenient. Over time, the daily hard work became into know-how and into a network of complicity between firms and support institutions. This is an unreachable, specific and unique net, hard to imitate and replicate. In other words, the greatest cluster’s patrimony. Footwear Cluster Risks After the strengths and opportunities being addressed, it is essential to know and analyze its weaknesses and threats, as well. The strategy can’t ignore the risks that the 34 3.1.4. International activity age The distribution of the firms’ international activity age (table and plot below) shows the prevalence of ages up to 30 years, with a smaller importance of classes between 10 and 20 years, as it is clear in the histogram. Furthermore, only a few firms show an international activity older than 30 years, with a maximum of 41 years. Therefore, the average international activity age is 18.1 years, next to the median, which is 20 years, the 1st quartile is 7.5 years and the 3rd quartile is 28 years. Ages are concentrated in low and moderate values. This concentration and the existence of some higher ages led to a high dispersion reflected in the coefficient of variation (61.8%). Table 10 International activity age descriptive measures Coefficients Minimum 2 Maximum 41 Average 18.1 1st Quartile 7.5 Median 20 3rd Quartile 28 Skewness Coefficient 0.2 Standard deviation 11.2 Coefficient of variation 61.8% 35 Figure 3 International activity age plot International activity age (years) Density 0.00 0.01 0.02 0.03 0.04 0 5 10 15 20 25 30 41 10 20 30 40 International activity age (years) The following scatterplot shows the existence of a low positive correlation between the firms’ age and their international activity age and consequently the correlation coefficient of Pearson is only 0.39. Since both ages do not have a normal distribution (the p-value of the Shapiro-Wilk test is approximately 0 and 0.015 respectively, rejecting the normality hypothesis), the Spearman coefficient of correlation is 0.37, with a p-value of 0.01. Therefore, we conclude that a significant positive correlation between the two variables exists (although slightly weak). Consequently, in general, when the firms’ age increases (decreases), their international activity age also increases (decreases). 36 Figure 4 Scatterplot of a firm’s age and its international activity age 010 20 30 40 50 10 20 30 40 Company's age (years) International activity age (years) 3.1.5. Geographic location of the firm There are 13 different locations among the respondent firms, predominantly Felgueiras (44.68%) and Pólo S. João da Madeira (36.17%). The remaining locations have few firms and are relatively close to both Polos. Table 11 Location of firms Location n % Pólo Felgueiras 21 44.68 Pólo S. João da Madeira 17 36.17 Santa Maria da Feira 5 10.64 Oliveira de Azeméis 4 8.51 Total 47 100.0 3.1.6. Number of employees The firms that have between 50 and 249 employees prevail with 48.9%, followed by firms that have between 10 and 49 employees (31.9%), less than 10 employees (14.,9%) and those with at least 250 employees (4.3%). Therefore, the SMEs (small and medium enterprise) are predominant. However, there are still some micro firms (less than 10 employees) and a few large firms (250 or more employees). 37 Table 12 Number of employees Number of Employees n % < 10 7 14.9 10 – 49 15 31.9 50 – 249 23 48.9 ≥ 250 2 4.3 Total 47 100.0 3.1.7. Turnover The firms with a turnover between 2 million and 9 million euros prevail with 44.7%, followed by the firms with a turnover less than 2 million euros (36.2%), those between 10 million and 49 million euros (17%) and those with a turnover of at least 50 million euros (2.1%). Therefore, the firms with a lower turnover (up to 9 million euros) strongly prevail with 80.9%. Note that a few firms with higher turnover also exist. Table 13 Turnover Turnover (106 €) n % < 2 17 36.2 2 – 9 21 44.7 10 – 49 8 17.0 ≥ 50 1 2.1 Total 47 100.0 3.1.8. Number of export markets The firms that export to between 2 and 5 markets prevail with 38.3%, followed by the firms that export to between 6 and 10 markets and those that export to more than 10 markets (27.7% each) and finally those that export to just one market (6.4%). A relevant fact is that all the respondent firms export to at least one market, there are no nonexporters. Therefore, firms that export to between 2 and 5 markets prevail with 38.3% but the weight of those that export to at least 6 markets is higher, around 55.4%. So the majority exports to a moderate or high number of markets. 38 Table 14 Number of export markets Number of markets n % 0 0 0.0 1 3 6.4 2 – 5 18 38.3 6 – 10 13 27.7 > 10 13 27.7 Total 47 100.0 3.1.9. Weight of exports in sales The distribution of the weight of exports in sales is strongly negatively skewed (the Fisher asymmetry coefficient is -1.6), as it is clear in histogram and in boxplot below which it means that the higher weights prevail. Therefore, the number of firms with a weight of 70% or less is very small (these firms are very different from the majority) and there are only a few firms with a weight between 70% and 90%. There is also a considerable number with a weight between 90% and 95%, but the majority has a weight higher than 95%, of which a large proportion exports the whole production or almost all of it. The average weight is 81.3% but the median is 90% (which means that half the firms has a weight of 90% or less and that the other half has a weight of more than 90%, which is very high), the 1st quartile is 80% (a quarter of the firms has a weight of 80% or less) and the 3rd quartile is 99% (three quarters of the firms export 99% or less) which also means that a quarter of the firms has a weight between 99% and 100%, representing a very large number of firms, with a maximum of 100% (there are still 10,6% of firms exporting the whole of their sales). Therefore, the export intensity of these firms is extremely high. The weight of exports is concentrated in high or very high values but the existence of low values leads to a moderate dispersion as reflected by the coefficient of variation (30.6%). Table 15 Weight of exports in sales descriptive measures Coefficients Minimum 12 Maximum 100 Average 81.3 1st Quartile 80 39 Median 90 3rd Quartile 99 Skewness Coefficient -1.6 Standard deviation 24.8 Coefficient of variation 30.6% Figure 5 Graphic Representation of the weight of exports in sales plot Weight of exports in sales (%) Desnsity 0.00 0.02 0.04 0.06 0.08 030 60 70 90 100 20 40 60 80 100 Weight of exports in sales (%) The relationships between the weight of exports in sales and on the one hand the age of firms and on the other hand the international activity age are also relevant. The Pearson correlation coefficients of the two pairs of variables are -0.045 and 0.003 respectively, very close to zero, indicating the absence of a linear association. Similarly, the Spearman correlation coefficient is respectively 0.008 with a p-value of 0.96 and 0.089 with a p-value of 0.55, both non-significant. Therefore, it is not possible to say that there is a relationship between the weight of exports in sales and any of the other two variables. In short, firms’ export intensity does not seem to be related with their age or with their international activity age. 40 3.1.10. Main export markets There are 10 missing responses and therefore the table below only displays the results of 37 valid responses. Respondents listed 10 different markets, besides the European Union. The most important are France (24.3%), Germany (18.9%), the United Kingdom (16.2%), Spain and Netherlands (10.8% each). Very few firms export for the remaining markets. Table 16 Main export markets Market n % France 9 24.3 Germany 7 18.9 United Kingdom 6 16.2 Spain 4 10.8 Netherlands 4 10.8 European Union 2 5.4 South Africa 1 2.7 Angola 1 2.7 Dubai 1 2.7 U.S.A 1 2.7 Italy 1 2.7 Total 37 100.0 3.1.11. Percentage of Human Resources with Higher Education The distribution of the percentage of human resources with Higher Education is strongly positively skewed (the Fisher asymmetry coefficient is 2.7), as it is clear in the histogram and in the boxplot below, which means that the lower percentages prevail. Consequently, there is a strong concentration of percentages in the lower classes and, in fact, the class of 1% or less is the most important, being followed by successive falls, with a minimum of zero. There are very few firms with percentages higher than 10% and the number of firms with percentages higher than 50% is extremely low, with a maximum of 100%. Therefore, the average percentage is only 12.6%, higher than the median of 4% (i.e., half the firms has a percentage of 4% or less), the 1st quartile is 1% and the 3rd quartile is only 10%, which means that the percentages are concentrated in extremely low values. This concentration and the existence of some high percentages lead to a very 41 strong dispersion, reflected by the coefficient of variation (187%). In short, most firms have very few human resources with Higher Education available. Table 17 Percentage of human resources with higher education descriptive measures Coefficients Minimum 0 Maximum 100 Average 12.6 1st Quartile 1 Median 4 3rd Quartile 10 Skewness Coefficient 2.7 Standard deviation 23.5 Coefficient of variation 187% Figure 6 Percentage of human resources with higher education plot Percentage of human resources with higher education (%) Density 0.00 0.05 0.10 0.15 0.20 0.25 0.30 010 20 50 80 100 020 40 60 80 100 Percentage of human resources with higher education (%) 42 3.1.12. Attendance at International Fairs The majority of firms (53.2%) has attended international fairs. 3.1.13. Own Brand The majority of firms (70.2%) has their own brand. The distribution of their own-brand sales percentage shows a strong concentration in low percentages, up to 10%, or in very high percentages (100%). The percentages between 10% and 80% are very scattered, with few firms for such a wide range as shown by the histogram. Note also that there are no percentages between 80% and 100% which implies that the last bar on the right in the histogram includes firms with 100% only. Therefore, the average percentage is 50.7%, but it is important to stress that this value has very little meaning because of the concentration in the lower or in the higher percentages mentioned above. The median is 50%, almost coincident with the average, but it also has very little meaning. The 1st quartile is only 10% and the 3rd quartile is 100%, which shows the existence of an important number of firms with a percentage of 100%, i.e., firms who sell their own brand exclusively. The concentration of percentages either in high values or in low values, leads to a high dispersion reflected by the coefficient of variation (77.7%). Table 18 Own-brand sales percentage descriptive measures Coefficients Minimum 1 Maximum 100 Average 50.7 1st Quartile 10 Median 50 3rd Quartile 100 Skewness Coefficient 0.06 Standard deviation 39.4 Coefficient of variation 77.7% 43 Figure 7 Own-brand sales percentage plot Own-brand sales percentage (%) Density 0.00 0.01 0.02 0.03 0.04 0.05 0.06 010 30 50 70 80 95 020 40 60 80 100 Own-brand sales percentage (%) 3.1.14. Outsourcing Only a small minority of firms (21.3%) is outsourced. The percentage of their sales in outsourcing is 1%, 20% (2 firms), 29%, 45%, 80% and 100% (3 firms did not respond). Therefore, the average and the standard deviation are respectively 43.4% and 34.1% (note that dispersion is very high because of the large differences between percentages). 3.1.15. Use of outsourcing The majority of firms (74.5%) uses outsourcing. 3.1.16. Investment in Research and Development (R&D) Almost half of the firms invest in R&D (46.8%). 3.1.17. Investment in Marketing More than half of the firms invest in Marketing (51.1%). 3.1.18. Investment in Design The majority of firms (66%) invests in Design. 50 “A diversidade das características das empresas dentro do Cluster faz com que cada uma delas tenha diferentes capacidades de internacionalização” 61.7% 14.9% 76.6% (High) “A presença de Empresas de Referência dentro do Cluster influencia positivamente a internacionalização de outras empresas pertencentes ao Cluster” 55.3% 17% 72.3% (High) “A intensidade competitiva entre as empresas do Cluster influencia positivamente o processo de internacionalização das mesmas” 63.8% 8.5% 72.3% (High) “A presença de Multinacionais estrangeiras dentro do Cluster tem impacto positivo na internacionalização das restantes empresas” 29.8% 2.1% 31.9% (Low) “A boa reputação internacional do Cluster funciona como incentivo para a internacionalização das suas empresas” 53.2% 27.7% 80.9% (Extremely High) “A boa reputação internacional do Cluster facilita o processo de internacionalização das suas empresas” 51.1% 29.8% 80.9% (Extremely High) “A Capacidade de inovação do Cluster tem um impacto positivo na internacionalização das suas empresas” 55.3% 23.4% 78.7% (Very High) Table 21 Firms’ opinion on Cluster Question Frequency Strongly disagree Disagree Neither agree nor disagree Agree Strongly Agree n % n % n % n % n % 1 0 0.0 1 2.1 12 25.5 30 63.8 4 8.5 2 0 0.0 2 4.3 10 21.3 31 66.0 4 8.5 3 0 0.0 4 8.5 14 29.8 25 53.2 4 8.5 4 0 0.0 1 2.1 9 19.1 30 63.8 7 14.9 5 0 0.0 1 2.1 10 21.3 29 61.7 7 14.9 6 0 0.0 3 6.4 10 21.3 26 55.3 8 17.0 7 0 0.0 2 4.3 11 23.4 30 63.8 4 8.5 8 1 2.1 13 27.7 18 38.3 14 29.8 1 2.1 9 0 0.0 1 2.1 8 17.0 25 53.2 13 27.7 10 0 0.0 1 2.1 8 17.0 24 51.1 14 29.8 11 0 0.0 1 2.1 9 19.1 26 55.3 11 23.4 51 3.2.2. Scale conceptual structure A factor analysis of this questionnaire was run in order to identify the factors underlying the firms’ responses and to validate the questionnaire scale. (See Appendix B) The results of the factorial analysis forced to 5 factors 7 with varimax rotation and Kaiser normalization are displayed in the next table where the factor loadings are shown with the largest loading of each question in bold (note that the questions are ordered according to the factor where they saturate and not according to the order they appear in the questionnaire). Other factor solutions were tried, especially that with 4 factors, but the solution with 5 factors proved to be the most appropriate which means that 5 factors are enough to describe the structure underlying the data (latent structure). Most of the factor loadings are high or very high and only two are acceptable which leads again to the conclusion that the quality of this factor solution is good. The table also displays the communalities, i.e., the proportion of the variance of each question explained by the 5 extracted factors together. That proportion is much larger than 50% for every question (larger than 0.75 in fact) and is high for some questions and very high for others implying again that the results of this factor analysis are reliable. Table 22 Scale of the opinion on the Cluster - Factor structure Question Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Com. 1 0.818 0.038 0.336 0.053 -0.106 0.797 2 0.840 0.317 -0.044 0.022 0.109 0.821 3 0.773 0.411 -0.243 -0.059 0.067 0.834 4 0.609 0.009 0.343 0.457 -0.293 0.783 9 0.112 0.849 0.265 0.122 -0.039 0.819 10 0.352 0.767 0.269 -0.020 -0.087 0.793 11 0.248 0.762 -0.176 0.333 0.093 0.792 6 0.008 0.475 0.589 0.238 0.355 0.755 7 0.085 0.119 0.883 -0.012 0.120 0.816 5 0.014 0.218 0.029 0.928 0.045 0.911 8 -0.011 -0.031 0.175 0.011 0.949 0.932 7 See Technical Details in Appendix B 52 The first factor shows high loadings of questions 1, 2, 3 and 4 and consequently this factor may be called the dimension of “Networking and knowledge and resources sharing among firms”. Question 1 is related to the influence that the share of knowledge among firms have in their internationalization process, question 2 is related to the influence that the share of important and rare resources among firms have in their internationalization process and question 3 approaches networking and cooperation among firms has an incentive for them in their international activities, so these three questions are interlinked by the nature of networking and knowledge sharing among firms. Questions 1, 2 and 3 belong to Hypothesis 1 (H1: Cluster Governance has a positive relationship with the internationalization process of its firms): question 1 is H1a, question 2 is H1b and question 3 is H1c, based on the study by Dubé et al., (2015). Moreover, since the 3 questions have high levels of agreement, H1a, H1b and H1c are confirmed. Question 4 is within the same factor and despite the question may looks different in its approach comparing to the other 3 questions, make sense why they are together. Question 4 is related to the international experience of others firms inside the cluster and how this experience influence other firms to internationalize, this type of question is interconnected with the other 3 questions by its nature of sharing, in this case the international experience sharing among firms. Question 4 is Hypothesis 2 (H2: The internationalization experience of firms in the cluster enhances the willingness to internationalize of others firms within the cluster), based on the study by Mariotti et al. (2008), and since this question has high level of agreement, H2 is confirmed. The second factor shows high loadings of questions 9, 10 and 11 and consequently this factor may be called the dimension of “Cluster’s International Reputation”. Question 9 and 10 are related to the influence that cluster’s good reputation have in the internationalization process of its firms and both are related to Hypothesis 6 (H6: A cluster’s good reputation facilitates the internationalization process of its firms), based on studies by Zyglidopoulos et al. (2011) and Zen et al. (2011). Question 9 refers to how cluster works as an incentive to the internationalization and question 10 refers to how it facilitates the all internationalization process of firms. Question 11 despite having a different aim, it is related with the innovative capacity of the cluster and the 53 impact on the internationalization of firms, so this question leads equally to the importance of the reputation of the cluster but in this case for innovation. This question is Hypothesis 7 (H7: Cluster’s innovative capacity has a positive impact on its firms and facilitates the internationalization of its firms), based on the study by Mariotti et al. (2008). These 3 questions have extremely and very high level of agreement which confirms H6 and H7. The third factor shows high loadings of questions 6 and 7 and consequently this factor may be called the dimension of “The Leader firms’ effect and competitiveness”. Question 6 refers to the presence of leader firms and how it influence other firms to internationalize through imitation and/or know-how that the Leader firm can share as well the international experience, and is related with Hypothesis 3 (H3: The presence of Leader firms in the cluster influence positively the internationalization of others firms in the cluster). Question 7 refers to the impact that competitiveness among firms has on its internationalization process and is related with Hypothesis 4 (H4: The Degree of domestic rivalry between firms inside the cluster influences their internationalization process). It is interesting to observe that H3 and H4 are based on the study by Mariotti et al. (2008) and both are considered Structural Determinants to the author as analyzed in chapter 1.4. Firms who want to have the necessary capacity to compete must internationalize to grow and be capable to overcome other firms. Both questions have high levels of agreement which means that H3 and H4 are confirmed. The fourth factor shows high loadings of question 5 only which means that this question is different from the rest due to its approach, this factor may be called the dimension of “Heterogeneity of firms”. Question 5 refers to the heterogeneity among firms and the impact that the diversity of firms have on its level of internationalization. The positive impact of belong to a cluster and having the same access to know how, knowledge and resources doesn’t mean that all firms will internationalize at the same level due to its heterogeneity. Question 5 refers to Hypothesis 8 (H8: Heterogeneity of firms is relevant to understand why firms within a cluster have different levels of Internationalization) 54 based on the study by Giovannettia et al. (2013). The level of agreement to the question 5, a very high level of agreement, confirms H8. The fifth factor shows high loadings of question 8 only which means that this question is also different from the rest. In fact, recall that the pattern of the responses to this question was distinct from all the others, with the lowest agreement of all and with “Neither agree nor disagree” being the most frequent response. Question 8 refers to the impact of the presence of foreign multinationals inside the cluster on the internationalization process of other firms. These question refers to Hypothesis 5 (H5: The presence of foreign MNCs inside the cluster has a positive impact on the internationalization of its firms) based on the study by Mariotti et al. (2008). Through the questionnaire it is possible to conclude that the presence of Multinationals inside the cluster is not an important factor to the internationalization of other firms, so H5 is not confirmed. The explanation for this is the nature of the internationalization of these firms. The Portuguese footwear sector is an international sector but its foreign activities and networking is through export and not through Foreign Direct Investment (FDI), that’s why MNC’s don’t have an impact in their internationalization process. The quality of the factor model was assessed and the conclusion is that both the total questionnaire and the factors show a good reliability and internal consistency (see appendix C). 3.2.3. Cluster analysis A hierarchical cluster analysis was also run in order to identify homogeneous clusters (groups) integrating firms with similar features and a common profile concerning their opinions on the Footwear cluster. Several distance measures and aggregation indices were tried and the inertia and the R2 coefficient were computed for different numbers of clusters. The comparison of the solutions found led to the choice of Ward linkage with the squared euclidean distance. Clusters were based on the 5 factors previously obtained in the factorial analysis and on the (nonstandardized) factor scores (standardizing the scores would reduce the differences among variables and would assume the same weight for each one). 55 It was necessary to decide on the number of clusters first. Such decision was based on two elements: the inertia (variance) decomposition in within-cluster and between-cluster inertia computed from an analysis of variance with the resulting clusters which also allows the computation of the inertia proportion explained by each considered partition (R2 coefficient) and on the analysis of the dendrogram shown below where the different partitions and their meaning were assessed. Therefore, the next table and plot display the inertia proportions mentioned above, suggesting a solution with 9 clusters because this is where both the decrease of the within-cluster inertia or the increase of the between-cluster inertia start slowing down. The proportions are 32.9% and 67.1% respectively which is acceptable and the dendrogram also suggests the same solution, even though 10 clusters would also be possible. Therefore, the solution with 9 clusters was selected. Table 23 Scale of the opinion on the Cluster - Within-cluster and between-cluster inertia Number of clusters Within-cluster Between-cluster 1 100.0 0.0 2 82.4 17.6 3 70.0 30.0 4 60.6 39.4 5 53.6 46.4 6 47.3 52.7 7 41.9 58.1 8 36.8 63.2 9 32.9 67.1 10 29.6 70.4 Figure 8 Scale of the opinion on the Cluster - Plot of within-cluster and between-cluster inertia 56 Figure 9 Scale of the opinion on the Cluster - Dendrogram – Ward linkage/Squared euclidean distance 57 Cluster membership is displayed in the next table. Table 24 Scale of the opinion on the Cluster - Cluster membership Firm Cluster Firm Cluster Firm Cluster Firm Cluster 1 1 13 7 25 2 37 9 2 1 14 3 26 6 38 2 3 2 15 8 27 2 39 3 4 3 16 2 28 1 40 9 5 4 17 5 29 2 41 2 6 1 18 3 30 4 42 7 7 5 19 3 31 7 43 2 8 6 20 2 32 2 44 6 9 6 21 1 33 1 45 1 10 6 22 7 34 1 46 5 11 3 23 4 35 8 47 4 12 5 24 2 36 7 Cluster interpretation was done next based on the meaning of the factors resulting from the factorial analysis previously run. The mean factor score and standard deviation of each cluster are displayed in the next table and the boxplots show the score distribution for each cluster. Table 25 Scale of the opinion on the Cluster - Cluster mean and standard deviation Cluster Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 1 Mean 3.6 3.4 2.9 2.8 3.0 Standard deviation 0.3 0.4 0.3 0.4 0.6 2 Mean 3.9 2.7 2.7 3.3 3.4 Standard deviation 0.4 0.3 0.5 0.4 0.1 3 Mean 3.8 2.3 2.8 3.1 1.5 Standard deviation 0.4 0.6 0.5 0.4 0.3 4 Mean 3.1 2.0 2.3 2.4 2.5 Standard deviation 0.2 0.1 0.4 0.1 0.0 5 Mean 4.3 3.4 3.3 2.0 1.7 Standard deviation 0.7 0.6 0.7 0.6 0.6 6 Mean 2.4 3.6 2.5 3.3 2.1 Standard deviation 0.6 0.7 0.8 0.1 0.7 7 Mean 3.9 3.7 2.8 3.9 1.9 Standard deviation 0.4 0.3 0.6 0.4 0.7 8 Mean 2.8 1.6 3.5 4.2 2.9 Standard deviation 0.6 0.1 0.5 0.8 0.7 9 Mean 4.1 2.8 0.7 3.4 1.8 Standard deviation 0.7 0.5 0.6 0.2 0.2 58 Figure 10 Scale of the opinion on the Cluster - Cluster boxplots Cluster 1 Cluster 3 Cluster 5 Cluster 7 Cluster 9 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Clusters Factor 1 Cluster 1 Cluster 3 Cluster 5 Cluster 7 Cluster 9 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Clusters Factor 2 Cluster 1 Cluster 3 Cluster 5 Cluster 7 Cluster 9 1234 Clusters Factor 3 Cluster 1 Cluster 3 Cluster 5 Cluster 7 Cluster 9 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Clusters Factor 4 59 Cluster 1 Cluster 3 Cluster 5 Cluster 7 Cluster 9 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Clusters Factor 5 Cluster 1 has high scores on factors 1 and 2 and moderate scores on factors 3, 4 and 5. Firms that are in the Group 1 (8 firms) are firms that agree with the fact that networking and knowledge/resources sharing among firms within the cluster has a positive impact on its internationalization process (Factor 1), and they also agree that the cluster’s good reputation and its innovative capacity have a positive influence on the internationalization process of firms (Factor 2). With moderate scores but still relevant are the factors 3, 4 and 5 that shows that firms inside the group 1 also have a positive opinion about the “Leader firms’ effect and competitiveness” on internationalization; the “Heterogeneity of firms” and its consequences on the level of internationalization; and the positive impact of the presence of foreign multinationals on the internationalization process of other firms. Cluster 2 has high scores on factors 1, 4 and 5, moderate scores on factors 2 and 3. Firms that belong to Group 2 (11 firms) also agree with the fact that networking and knowledge/resources sharing among firms within the same cluster has a positive impact on its internationalization process and also have a positive opinion about the “Heterogeneity of firms” and its consequences on the level of internationalization, as 66  Investir em Design – “Important” represents the majority of the responses (57.4% of the firms), followed by “Very important” (23.4%), “Indifferent” (17%) and “A little important” (2.1%), not existing any “Not important” responses. Therefore, the level of importance is high (note that “Important” and “Very important” jointly represent 80.9% of the firms).  Investir em I&D – “Important” is the most frequent response (46.8% of the firms), followed by “Indifferent” (38.3%), “Very important” (12.8%) and “A little important” (2.1%), not existing any “Not important” responses. Therefore, the level of importance is moderate (note that “Important” and “Very important” jointly represent 59.6% of the firms).  Investir em Inovação – “Important” represents the majority of the responses (63.8% of the firms), followed by “Very important” (21.3%) and “Indifferent” (14.9%), not existing any “A little important” or “Not important” responses. Therefore, the level of importance is high (note that “Important” and “Very important” jointly represent 85.1% of the firms).  Desenvolver novos produtos – “Important” represents the majority of the responses (53.2% of the firms), followed by “Very important” (36.2%) and “Indifferent” (10.6%), not existing any “A little important” or “Not important” responses. Therefore, the level of importance is high (note that “Important” and “Very important” jointly represent 89.4% of the firms).  Conhecer os mercados externos – “Very important” represents the majority of the responses (57.4% of the firms), followed by “Important” (34%) and “Indifferent” (8.5%), not existing any “A little important” or “Not important” responses. Therefore, the level of importance is very high (note that “Important” and “Very important” jointly represent 91.5% of the firms).  Desenvolver uma rede de contactos internacionais – “Very important” represents the majority of the responses (57.4% of the firms), followed by “Important” (29.8%) and “Indifferent” (12.8%), not existing any “A little important” or “Not important” responses. Therefore, the level of importance is high (note that “Important” and “Very important” jointly represent 87.2% of the firms).  Participar em feiras internacionais – “Important” is the most frequent response (44.7% of the firms), followed by “Very important” (31.9%), “Indifferent” (19.1%) and “A 67 little important” (10.6%), not existing any “Not important” responses. Therefore, the level of importance is high (note that “Important” and “Very important” jointly represent 76.6% of the firms).  Ter relações próximas com empresas do Cluster – “Indifferent” is the most frequent response (40.4% of the firms), followed by “Important” (38.3%), “Very important” (14.9%) and “A little important” (6.4%), not existing any “Not important” responses. Therefore, the level of importance is moderate (note that “Important” and “Very important” jointly represent 53.2% of the firms).  Ter relações próximas com Universidades – “Indifferent” represents the majority of the responses (53.2% of the firms), followed by “A little important” (19.1%), “Important” (17%), “Very important” (6.4%) and “Not important” (4.3%). Therefore, the level of importance is low (note that “Important” and “Very important” jointly represent only 23.4% of the firms or, conversely, “Not important”, “A little important” and “Indifferent” jointly represent 76.6%).  Ter relações próximas com centros tecnológicos – “Important” is the most frequent response (40.4% of the firms), followed by “Indifferent” (36.2%), “Very important” (12.8%), “A little important” (8.5%) and “Not important” (2.1%). Therefore, the level of importance is moderate (note that “Important” and “Very important” jointly represent 53.2% of the firms).  Ter relações próximas com instituições de promoção da internacionalização – “Important” represents the majority of the responses (59.6% of the firms), followed by “Indifferent” (21.3%), “Very important” (12.8%), “A little important” (4.3%) and “Not important” (2.1%). Therefore, the level of importance is high (note that “Important” and “Very important” jointly represent 72.3% of the firms). 68 Table 29 Importance for the internationalization process Item Frequencies Not important A little important Indifferent Important Very important n % n % n % n % n % 1 2 4.3 5 10.6 10 21.3 18 38.3 12 25.5 2 1 2.1 2 4.3 4 8.5 18 38.3 22 46.8 3 1 2.1 1 2.1 1 2.1 18 38.3 26 55.3 4 0 0.0 2 4.3 7 14.9 25 53.2 13 27.7 5 0 0.0 3 6.4 11 23.4 28 59.6 5 10.6 6 0 0.0 1 2.1 8 17.0 27 57.4 11 23.4 7 0 0.0 1 2.1 18 38.3 22 46.8 6 12.8 8 0 0.0 0 0.0 7 14.9 30 63.8 10 21.3 9 0 0.0 0 0.0 5 10.6 25 53.2 17 36.2 10 0 0.0 0 0.0 4 8.5 16 34.0 27 57.4 11 0 0.0 0 0.0 6 12.8 14 29.8 27 57.4 12 0 0.0 2 4.3 9 19.1 21 44.7 15 31.9 13 0 0.0 3 6.4 19 40.4 18 38.3 7 14.9 14 2 4.3 9 19.1 25 53.2 8 17.0 3 6.4 15 1 2.1 4 8.5 17 36.2 19 40.4 6 12.8 16 1 2.1 2 4.3 10 21.3 28 59.6 6 12.8 3.3.2. Scale conceptual structure A factor analysis of this questionnaire was also run, for details see appendix D. The results of the factorial analysis forced to 5 factors with varimax rotation and Kaiser normalization are displayed in the next table where the factor loadings are shown with the largest loading of each item in bold (note that the items are ordered according to the factor where they saturate and not according to the order they appear in the questionnaire). Other factor solutions were tried, especially that with 7 factors, but the solution with 5 factors proved to be the most appropriate which means that 5 factors are enough to describe the structure underlying the data (latent structure). 69 Most of the factor loadings are high or at least acceptable which leads again to the conclusion that this factor solution is satisfactory. The table also displays the communalities, i.e., the proportion of the variance of each item explained by the 5 extracted factors together. That proportion is much larger than 50% for every item (larger than 0.7 in most cases) and is high or at least acceptable implying again that the results of this factor analysis are reliable. Table 30 Scale of the importance for the internationalization process - Factor structure Item Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Com. 6 0.575 0.031 0.512 -0.109 0.394 0.760 9 0.779 0.078 0.081 0.197 0.120 0.672 10 0.856 0.086 0.213 0.217 0.020 0.833 11 0.750 0.164 0.272 0.266 -0.041 0.736 4 0.309 0.674 0.164 0.348 -0.186 0.733 13 0.288 0.815 -0.153 0.221 0.136 0.837 15 0.000 0.780 0.204 0.054 0.220 0.702 7 0.299 0.043 0.722 0.087 0.116 0.633 8 0.241 0.123 0.752 0.147 0.027 0.661 16 -0.340 0.453 0.484 0.247 -0.032 0.617 2 0.190 0.118 0.276 0.830 0.193 0.852 3 0.301 0.261 0.090 0.746 0.120 0.738 1 -0.003 -0.066 0.014 0.316 0.821 0.778 5 0.315 0.329 0.264 0.063 0.630 0.678 12 0.312 0.243 -0.277 0.450 0.473 0.659 14 -0.283 0.494 0.224 -0.255 0.547 0.739 The first factor shows high loadings in items 6, 9, 10 and 11. Items 6 and 9 refer to the investment in Design and new products’ development, and items 10 and 11 refer to the knowledge about international markets and the development of an internationalbusiness network. These 4 items together make a lot of sense in terms of firms’ internationalization process. This factor suggest that investments in design and new products are valued as important by firms that are able to export such goods due to their knowledge about international markets and their international business relations. 70 Therefore, this factor may be called as “Design and product development versus international business capabilities”. The second factor shows high loadings in items 4, 13 and 15. Item 4 considers the importance to access to financial resources. Items 13 and 14 refer to the importance of close relationships with institutions inside the footwear cluster, namely with firms and the Technological Centre. This factor reveals that firms that consider the access to financial resources important to their internationalization process also value the intracluster relationships (networking). As seen in the previous section such networking is also associated with accessing resources and knowledge sharing. Thus, we may consider this factor as "Networking with firms and the technological center of the Footwear Industry, plus accessing to financial resources are crucial to the internationalization of its firms”. The third factor shows high loadings in items 7, 8 and item 16, but the latter with a lower value. This factor shows clearly that firms that consider important to invest in R&D and Innovation, for their internationalization process (items 7 and 8), also value the cooperation with institutions to promote their business in international markets (item 16). This is in line with the result found in the previous section and therefore reinforces it. This factor may be designated as “Investing in R&D and Innovation to promote/expand business in international markets. The fourth factor shows high loadings in items 2 and 3. They are extremely important for the internationalization process of the respondents' firm. Thus firms that consider very important to possess qualified human resources (item 2) also attribute high value to the linguistic capabilities (item 3) which makes all sense. This factor may be designated as "Qualified Human Resources and Linguistic capabilities for International Business" The fifth factor shows high loadings in items 1, 5, 12 and 14 and consequently this factor may be called the dimension of “The promotion of own brand in international fairs can initiate or enhance the international activity of firms”. 71 Items 1 and 5 relate, respectively, to own brand and marketing and item 12 relates to international fairs participation, so these 3 are interconnected because marketing is essential to the creation and development of brand recognition, and this is mostly done in international fairs. Item 14 is related with the cooperation with universities, which can be related with the integration of graduates in the firms with the aim to promote the firm’s business and brand. The quality of the factor model was assessed and the conclusion is that both the total questionnaire and the factors show a good reliability and internal consistency (see appendix E). 3.3.3. Cluster analysis A hierarchical cluster analysis with Ward linkage and squared euclidean distance was also run based on the 5 factors previously obtained in the factorial analysis and on the (nonstandardized) factor scores. Selection of the number of clusters was based on the inertia (variance) decomposition in within-cluster and between-cluster inertia computed from an analysis of variance, which also allows the computation of the inertia proportion explained by each considered partition (R2 coefficient), and on the analysis of the dendrogram shown below where the different partitions and their meaning were assessed. Therefore, the next table and plot display the inertia proportions mentioned above, suggesting a solution with 8 clusters because this is where both the decrease of the within-cluster inertia or the increase of the between-cluster inertia start slowing down. The proportions are 33.8% and 66.2% respectively which is acceptable and the dendrogram also suggests the same solution, even though 9 clusters would also be possible (but the 9-cluster solution would lead to several clusters with a very low number of firms). Therefore, the solution with 8 clusters was selected. 72 Table 31 Scale of the importance for the internationalization process - Within-cluster and betweencluster inertia Number of clusters Within-cluster Between-cluster 1 100.0 0.0 2 82.9 17.1 3 70.5 29.5 4 59.7 40.3 5 50.7 49.3 6 44.7 55.3 7 38.7 61.3 8 33.8 66.2 9 30.9 69.1 10 28.1 71.9 Figure 11 Scale of the importance for the internationalization process - Plot of within-cluster and between-cluster inertia 73 Figure 12 Scale of the importance for the internationalization process - Dendrogram – Ward linkage/Squared euclidean distance 74 Cluster membership is displayed in the next table. Table 32 Scale of the importance for the internationalization process - Cluster membership Firm Cluster Firm Cluster Firm Cluster Firm Cluster 1 1 13 1 25 2 37 1 2 2 14 5 26 2 38 1 3 1 15 3 27 3 39 4 4 1 16 6 28 7 40 3 5 3 17 3 29 3 41 3 6 1 18 1 30 1 42 5 7 2 19 1 31 4 43 7 8 1 20 3 32 7 44 2 9 4 21 1 33 3 45 6 10 1 22 1 34 1 46 8 11 2 23 2 35 8 47 7 12 4 24 3 36 6 Cluster interpretation was done next based on the meaning of the factors resulting from the factorial analysis previously run. The mean factor score and standard deviation of each cluster are displayed in the next table and the boxplots show the score distribution for each cluster. Table 33 Scale of the importance for the internationalization process - Cluster mean and standard deviation Cluster Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 1 Mean 2.9 2.7 3.4 3.3 3.7 Standard deviation 0.5 0.5 0.5 0.5 0.3 2 Mean 3.5 3.8 2.5 2.8 2.7 Standard deviation 0.5 0.5 0.4 0.3 0.4 3 Mean 2.0 3.4 3.3 3.2 2.8 Standard deviation 0.5 0.3 0.5 0.5 0.5 4 Mean 3.1 3.6 3.3 3.9 1.5 Standard deviation 0.1 0.7 0.8 0.4 0.4 5 Mean 2.4 1.1 2.8 4.4 3.5 Standard deviation 1.4 1.0 0.5 0.7 0.7 6 Mean 2.7 4.5 4.4 2.9 4.0 Standard deviation 0.2 0.1 0.8 0.2 0.5 7 Mean 2.3 3.2 2.8 1.2 2.6 Standard deviation 0.4 0.5 0.3 1.2 0.6 8 Mean 3.3 1.8 4.6 2.1 0.7 Standard deviation 0.4 0.4 0.5 1.8 0.6 75 Figure 13 Scale of the importance for the internationalization process - Cluster boxplots Cluster 1 Cluster 3 Cluster 5 Cluster 7 1234 Clusters Factor 1 Cluster 1 Cluster 3 Cluster 5 Cluster 7 1234 Clusters Factor 2 Cluster 1 Cluster 3 Cluster 5 Cluster 7 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Clusters Factor 3 Cluster 1 Cluster 3 Cluster 5 Cluster 7 012345 Clusters Factor 4 82 important” (12.8%), not existing any “A little important” or “Not important” responses. Therefore, the level of importance is high (note that “Important” and “Very important” jointly represent 72.3% of the firms). Table 34 Cluster’s importance/contribution Item Frequencies Not important A little important Indifferent Important Very important n % n % n % n % n % 1 2 4.3 1 2.1 12 25.5 25 53.2 7 14.9 2 1 2.1 1 2.1 10 21.3 30 63.8 5 10.6 3 0 0.0 1 2.1 4 8.5 30 63.8 12 25.5 4 1 2.1 0 0.0 7 14.9 33 70.2 6 12.8 5 0 0.0 3 6.4 15 31.9 25 53.2 4 8.5 6 0 0.0 4 8.5 12 25.5 26 55.3 5 10.6 7 1 2.1 0 0.0 19 40.4 21 44.7 6 12.8 8 0 0.0 1 2.1 8 17.0 32 68.1 6 12.8 9 0 0.0 2 4.3 20 42.6 19 40.4 6 12.8 10 0 0.0 1 2.1 14 29.8 24 51.1 8 17.0 11 0 0.0 4 8.5 12 25.5 23 48.9 8 17.0 12 0 0.0 0 0.0 13 27.7 28 59.6 6 12.8 3.4.2. Scale conceptual structure A factor analysis of this questionnaire was also run, for details see appendix F. The results of the factorial analysis forced to 3 factors with varimax rotation and Kaiser normalization are displayed in the next table where the factor loadings are shown with the largest loading of each item in bold (note that the items are ordered according to the factor where they saturate and not according to the order they appear in the questionnaire). Other factor solutions were tried, especially that with 5 factors, but the solution with 3 factors proved to be the most appropriate which means that 3 factors are enough to describe the structure underlying the data (latent structure). 83 Factor loadings are high or at least acceptable which leads again to the conclusion that this factor solution is satisfactory. The table also displays the communalities, i.e., the proportion of the variance of each item explained by the 3 extracted factors together. That proportion is larger than 50% for every item (larger than 0.7 in most cases) with a single exception (that is close to 50%) and is high or at least acceptable implying again that the results of this factor analysis are reliable. Table 35 Scale of the cluster’s importance/contribution - Factor structure Item Factor 1 Factor 2 Factor 3 Com. 1 0.622 0.306 0.293 0.566 2 0.833 0.304 0.065 0.791 3 0.825 0.060 0.341 0.801 4 0.833 0.220 0.210 0.787 7 0.739 0.429 0.018 0.730 5 0.070 0.832 0.384 0.845 6 0.461 0.602 0.297 0.664 8 0.285 0.757 0.146 0.677 9 0.414 0.775 0.132 0.789 10 0.071 0.131 0.898 0.828 11 0.341 0.228 0.495 0.413 12 0.246 0.289 0.804 0.791 Factor 1 shows high loadings in items 1, 2, 3, 4 and 7. Items 1, 2 and 7 refer to the access/acquisition of resources (financial and human) and design capabilities; items 3 and 4 refer to firms’ productivity improvement and to footwear sector’s notoriety. This factor makes sense because those specific resources and professional skills may contribute to the enhancement of both productivity and notoriety. As such, this factor may be designated as “Resources and Skills for Productivity and Notoriety". Factor 2 shows high loadings in items 5, 6, 8 and 9. Items 6 and 8 relate to professional skills and capabilities (Marketing and R&D), and items 5 and 9 refer to the access to Innovation and shared knowledge about international markets. The association among these issues makes all sense because R&D and innovation are normally expensive 84 activities, which firms need recover through commercial competences. Consequently this factor may be called “Innovation and R&D for Marketing in International Markets". Factor 3 shows high loadings in items 10, 11 and 12. Item 10 refers to networking and cooperation and items 11 and 12 refer to the participation in international fairs and activities. This factor shows the Cluster as a mean of cooperation and networking among firms, entities and institutions which, as seen before, facilitates or enhance the willingness to participate in international fairs and activities. So this factor may be called "Networking and Cooperation for International Activities". The quality of the factor model was assessed and the conclusion is that both the total questionnaire and the factors show a good reliability and internal consistency (see appendix G). 3.4.3. Cluster analysis A hierarchical cluster analysis with Ward linkage and squared euclidean distance was also run based on the 3 factors previously obtained in the factorial analysis and on the (nonstandardized) factor scores. The next table and plot display the within-cluster and between-cluster inertia proportions, suggesting a solution with 7 clusters because this is where both the decrease of the within-cluster inertia or the increase of the between-cluster inertia start slowing down. Although this is a good solution, the dendrogram also suggests a solution with 6 clusters which seems preferable because the seventh cluster would be formed by a single firm. Therefore, we selected a solution with 6 clusters and the proportions mentioned above are 41.4% and 58.6% respectively which is acceptable. 85 Table 36 Scale of the cluster’s importance/contribution - Within-cluster and between-cluster inertia Number of clusters Within-cluster Between-cluster 1 100.0 0.0 2 80.9 19.1 3 73.0 27.0 4 63.0 37.0 5 52.3 47.7 6 41.4 58.6 7 25.3 74.7 8 23.8 76.2 9 16.9 83.1 10 15.5 84.5 Figure 14 Scale of the cluster’s importance/contribution - Plot of within-cluster and between-cluster inertia 86 Figure 15 Scale of the cluster’s importance/contribution - Dendrogram – Ward linkage/Squared euclidean distance 87 Cluster membership is displayed in the next table. Table 37 Scale of the cluster’s importance/contribution - Cluster membership Firm Cluster Firm Cluster Firm Cluster Firm Cluster 1 1 13 1 25 2 37 2 2 2 14 1 26 2 38 2 3 2 15 1 27 1 39 3 4 3 16 1 28 2 40 6 5 3 17 2 29 3 41 2 6 2 18 3 30 3 42 1 7 1 19 2 31 4 43 3 8 2 20 2 32 3 44 3 9 4 21 4 33 2 45 6 10 1 22 1 34 2 46 4 11 5 23 1 35 3 47 3 12 6 24 2 36 6 Cluster interpretation was done next based on the meaning of the factors resulting from the factorial analysis previously run. The mean factor score and standard deviation of each cluster are displayed in the next table and the boxplots show the score distribution for each cluster. Table 38 Scale of the cluster’s importance/contribution - Cluster mean and standard deviation Cluster Factor 1 Factor 2 Factor 3 1 Mean 3.2 2.2 4.1 Standard deviation 0.3 0.6 0.5 2 Mean 3.3 3.1 3.4 Standard deviation 0.2 0.2 0.3 3 Mean 3.1 2.2 2.5 Standard deviation 0.5 0.3 0.2 4 Mean 4.1 1.5 3.2 Standard deviation 0.2 0.4 0.5 5 Mean -0.4 3.0 3.5 Standard deviation 6 Mean 4.2 3.8 3.6 Standard deviation 0.3 0.4 1.1 88 Figure 16 Scale of the cluster’s importance/contribution - Cluster boxplots Cluster 1 Cluster 3 Cluster 5 01234 Clusters Factor 1 Cluster 1 Cluster 3 Cluster 5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 Clusters Factor 2 Cluster 1 Cluster 3 Cluster 5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Clusters Factor 3 89 One should notice that most firms in our sample (38) are included in one of the three first clusters, cluster 5 has one single firm whereas clusters 4 and 6 have 4 firms each. Cluster 1 has a very high score on factor 3, high score on factor 1 and a moderate score on factor 2, score values range between 4.1 and 2.2. So, the 11 firms in this cluster agree that the Footwear's Cluster (FC) is very important and contributes highly for factor “Networking and Cooperation for International Markets”. It is also important and contributes for the factor "Resources and Skills for Productivity and Notoriety". Still in their opinion the FC has a small importance/contribution for factor "Innovation and R&D for Marketing in International Markets". Cluster 2 is the biggest one with 16 firms. This cluster registers high scores on all factors (1, 2 and 3) and the lowest dispersion - score's value range between 3.1 and 3.4. These firms agree that the Footwear's cluster is important and contributes for the three factors previously identified: "Resources and Skills for Productivity and Notoriety"; "Innovation and R&D for Marketing in International Markets"; “Networking and Cooperation for International Markets”. However, in this cluster, firms value relatively more factor 2 attributing the second highest score among all clusters. Cluster 3 has a high score on factor 1 and moderate scores on factors 2 and 3, score values range between 2.2 and 3.1. The 11 firms agree that the Footwear's Cluster is important for the acquisition/access of resources (financial and human) and professional skills (Design) which are associated with increased productivity and reputation of the sector. Still, this is a relatively low score compared with the other clusters. The same applies to factor 2. Finally, this cluster, among all, attributes the lowest value to "Networking and Cooperation for International Activities". Cluster 4 has a very high score on factor 1, a high score on factor 3 and a very low score on factor 2 (the lowest among all clusters): score values range between 1.5 and 4.1. The 4 firms in this cluster strongly agree in the importance/contribution of the Footwear's Cluster for the factor: "Resources and Skills for the Productivity and Notoriety". In contrast, they do not recognize the FC as being important to, or contributing to Innovation and R&D, acquisition of Marketing skills, and knowledge share on foreign 90 markets. They have an opinion similar to the other clusters about "Networking and Cooperation for International Activities" Cluster 5 has one single firm that stands out for attributing an extremely low score to factor 1 (the lowest among all factors and clusters). It shows relatively high scores on factors 2 and 3. This firm considers that the FC is not important at all for the acquisition/access of "Resources and Skills for Productivity and Notoriety". But has a relatively positive opinion on the importance, and contribution of FC to factor 2 and 3: "Innovation and R&D for Marketing in International Markets" and "Networking and Cooperation for International Activities". Cluster 6 has a very high score on factor 1 (the highest among all clusters and factors) and high scores on factor 2 (also the highest score among all clusters) and on factor 3 (the second highest): score values range from 3.6 to 4.2. Firms in Group 6 (4 firms) strongly agree that the Footwear Cluster is important and contributes to the acquisition/access of resources (financial and human) and professional skills (Design) which in turn are associated with productivity of firms and reputation of the sector. They also agree on the importance/contribution of the Footwear Cluster to innovation, professional skills in Marketing, R&D and knowledge share about foreign markets. Finally, the Footwear Cluster is a mean of cooperation and networking between firms and various entities and institutions which in turn may ease the participation in international fairs and activities. It is remarkable that firms' heterogeneity is again revealed, now in terms of their assessment on the importance and contribution that the footwear cluster has to the development of critical factors for their internationalization. Despite their differences, one may say that there is a relative convergence of opinions on the importance of the FC for factors 3 and 1, and more dispersion on the importance of the FC for factor 2. 91 3.5. Cluster’s important features/qualities/actions Recall that the same firm may mark several features (all that apply). Therefore, the next table shows the number of features marked and not the number of firms (which explains why the total is not 47). The most important features are Cluster’s good reputation (19.3%), International Sector’s Disclosure (17.9%), Share of International experience among firms (12.4%) and Cluster that invest in Innovation (11%). Less important, but still very relevant, are Share of know-how among firms (9%), Access to Resources (8.3%) and R&D’s knowledge share (6.9%). Little important are Existence of Multinationals producing in Portugal (4.8%), Existence of Networking / Cooperation (4.1%), Domestic competitiveness among firms (3.4%) and Existence of Multinationals outsourcing in Portugal (2.8%). Although exists a large number of different combinations of features (since each firm marks all that apply), the most frequent combinations involve simultaneously three of the most important (i.e., frequent) features mentioned above such as Cluster’s good reputation, International Sector’s Disclosure, Share of International experience among firms and Cluster that invest in Innovation. Table 39 Cluster’s important features/qualities/actions Features/qualities/actions n % Cluster’s good reputation 28 19.3 International Sector’s Disclosure 26 17.9 Share of International experience among firms 18 12.4 Cluster that invest in Innovation 16 11.0 Share of know-how among firms 13 9.0 Access to Resources 12 8.3 R&D’s knowledge share 10 6.9 Existence of Multinationals producing in Portugal 7 4.8 Existence of Networking / Cooperation 6 4.1 Existence of Multinationals outsourcing in Portugal 4 2.8 Total 145 100.0 98 Johanson, Jan; Vahlne, Jan-Erik (2009), “The Uppsala internationalization process model revisited: From liability of foreignness to liability of outsidership”, Journal of International Business Studies, Vol. 40, p. 1411-1431. Kalantaridis, Christos (2004) “Internationalization, Strategic Behavior, and the Small Firm : A Comparative Investigation”, Journal of Small Business Management, Vol. 42, N° 3, p. 245-262. Karagozoglu, Necmi; Lindell, Martin (1998), “Internationalization of small and medium-sized technologybased firms: An exploratory study”, Journal of Small Business Management, Vol. 36, N° 1, p. 4-60. Karaev, A., Koh, S.C.L., Szamosi, L.T. (2007), "The cluster approach and SME competitiveness: a review". Journal of Manufacturing Technology Management 18 (7), 818–835. Keeble, David; Lawson, Clive; Lawton Smith, Helen; Moore, Barry; Wilkinson, Frank (1998), “Internationalisation processes, networking and local embeddedness in technologyintensive small firms”, Small Business Economics, Vol. 11, N° 4, p. 327-342. Leonidou, Leonidas C. (2004), “An Analysis of the Barriers Hindering Small Business Export Development”, Journal of Small Business Management, Vol. 42, N° 3, p. 279302. Libaers, D., & Meyer, M. (2011), "Highly Innovative Small Technology Firms, Industrial Clusters and Firm Internationalization", Research Policy, 40(10), 1426-1437. Lin, Ku Ho; Chaney, Isabella (2007), “The influence of domestic interfirm networks on the internationalization process of Taiwanese SMEs”, Asia Pacific Business Review, Vol. 13, N° 4, p. 565-583. Lu, Jane; Beam ish, Paul (2001), “The internationalization and performance of SMEs”, Strategic Management Journal, Vol. 22, N° 6/7, p. 565-586. Lu, Jane; Beamish, Paul (2006), “Partnering strategies and performance of SMEs’ international joint ventures”, Journal of Business Venturing, Vol. 21, N° 4, p. 461-486. Mariotti, S., Mutinelli, M., & Piscitello, L. (2008), "The Internationalization of Production by Italian Industrial Districts' Firms: Structural and Behavioural Determinants", Regional Studies, 42(5), 719-735. Mayrhofer, Ulrike; Urban, Sabine (2011), "Management international: des pratiques en mutation", Paris, Pearson Education. Melitz, M. J. (2003), "The Impact of Trade on Intra-Industry Reallocations and Aggregate Industry Productivity", Econometrica, 71(6), 1695-1725. Melitz, M. J., & Ottaviano, G. I. P. (2008), "Market Size, Trade, and Productivity". Review of Economic Studies, 75(1), 295-316. 99 Mittelstaedt, John; Ward, William; Nowlin, Edward (2006), “Location, industrial concentration and the propensity of small firms to export – entrepreneurship in the international marketplace”, International Marketing Review, Vol. 23, N° 5, p. 486-503. Oviatt, Benjamin; McDougall, Patricia (2005), “Defining international entrepreneurship and modeling the speed of internationalization”, Entrepreneurship Theory & Practice, Vol. 29, N° 5, p. 537-553. Porter, M. E. (1990), "The Competitive Advantage of Nations". Harvard Business Review, 68(2), 73-93. Porter, M. E. (1995), "The Competitive Advantage: Creating and Sustaining Superior Performance". NY: Free Press. Porter, M. E. (1998), "Clusters and the new economics of competition". Harvard Business Review, 76(6), 77-90. Porter, Michael; Ketels, Christian (2009), “Clusters and industrial districts: Common roots, different perspectives”, dans G. Becattini; M. Bellandi; L. De Propis (sous la direction de), A Handbook of Industrial Districts, Cheltenham, Edward Elgar Publishing, p. 172-183. Prashantham , Shameen; Young, Stephen (2011), "Post-entry speed of international new venture", Entrepreneurship Theory & Practice, Vol. 35, N° 2, p. 275-292. Saxenian, Annalee (1990), “Regional networks and the resurgence of Silicon Valley”, California Management Review, Vol. 33, N° 1, p. 89-111. Shaver, J. Myles; Flyer, Fredrick (2000), “Agglomeration economies, firm heterogeneity, and foreign direct investment in the United States”, Strategic Management Journal, Vol. 21, N°12, p. 1175-1193. Torres, Olivier (2003), "Les PME face à la mondialisation: du management de proximité à la stratégie de glocalisation » dans C. Fourcade, O. Torres (sous la direction de), Les PME entre région et mondialisation: Processus de glocalisation et dynamiques de proximité, Cahier ERFI, Vol. 10, N° 4, p. 22-39. Verdier, Sylvie; Prange, Catherine; Atamer, Turgul; Monin, Philippe (2010), “Internationalization performance revisited: the impact of age and speed on sales growth”, Management International, Vol. 15, N° 1, p. 19-31. Vernon, Raymond (1966), “International investment and international trade in the product cycle”, Quarterly Journal of Economics, Vol. 80, N° 2, p.190-207. Welch, L. S.; Welch, D. E. (1996), "The internationalization process and networks: a strategic management perspective", Journal of International Marketing, 4(3), 11-28. Westhead, Paul; Wright, Mike; Ucbasaran, Deniz (2001), “The internationalization of new and small firms: A resource-based view”, Journal of Business Venturing, Vol.16, N° 4, p. 333-358. 100 Wagner, J. (2007), "Exports and Productivity: A Survey of the Evidence from FirmLevel Data". World Economy, 30(1), 60-82. Wagner, J. (2012), "International Trade and Firm Performance: A Survey of Empirical Studies since 2006". Review of World Economics, 148(2), 235-267. Zain, Mohamed; Ng, Siew Imm (2006), “The impact of network relationships on SME internationalization process”, Thunderbird International Business Review, Vol. 48, N° 2, p. 183-205. Zen, A. C., Fensterseifer, J. E., & Prévot, F. (2011), "Internationalization of clustered companies and the influence of resources: A case study on wine clusters in brazil and france". Latin American Business Review, 12(2), 123-141. Zhou, Lianxi; Wu, Wei-ping; Luo, Xueming (2007), “Internationalization and performance of born-global SMEs: the mediating role of social networks”, Journal of International Business Studies, Vol. 38, p. 673-690. Zimmerman, Monica; Zeitz, Gerald (2002), “Beyond survival: achieving new venture growth by building legitimacy”, Academy of Management Review, Vol. 27, N° 3, p. 414-431. Zyglidopoulos, S. C., DeMartino, R., & Reid, D. M. (2006), "Cluster Reputation as a Facilitator in the Internationalization of Small and Medium-Sized Enterprises". Corporate Reputation Review, 9(1), 79-87. 101 Appendices Appendix A: Survey 102 103 104 105 106 107 Appendix B: Opinion on the Cluster First, in order to check whether these data are appropriate for a factorial analysis, the correlation matrix of the responses to the questions is displayed below showing many moderate and some high correlations. Table 40 Scale of the opinion on the Cluster - Correlation matrix Question Question 1 2 3 4 5 6 7 8 9 10 11 1 1.000 0.630 0.507 0.541 0.101 0.234 0.296 - 0.071 0.268 0.367 0.193 2 0.630 1.000 0.697 0.412 0.146 0.138 0.132 0.046 0.344 0.525 0.458 3 0.507 0.697 1.000 0.367 0.046 0.105 - 0.052 - 0.020 0.353 0.475 0.510 4 0.541 0.412 0.367 1.000 0.328 0.224 0.253 - 0.149 0.238 0.371 0.230 5 0.101 0.146 0.046 0.328 1.000 0.295 0.089 0.041 0.325 0.193 0.406 6 0.234 0.138 0.105 0.224 0.295 1.000 0.501 0.353 0.507 0.431 0.381 7 0.296 0.132 - 0.052 0.253 0.089 0.501 1.000 0.233 0.298 0.303 0.045 8 - 0.071 0.046 - 0.020 - 0.149 0.041 0.353 0.233 1.000 0.032 - 0.003 0.000 9 0.268 0.344 0.353 0.238 0.325 0.507 0.298 0.032 1.000 0.703 0.574 10 0.367 0.525 0.475 0.371 0.193 0.431 0.303 - 0.003 0.703 1.000 0.524 11 0.193 0.458 0.510 0.230 0.406 0.381 0.045 0.000 0.574 0.524 1.000 Furthermore, the Kaiser-Meyer-Olkin measure of sampling adequacy was also computed and is displayed in the next table for each question and for the complete scale. The total value is 0.788 which is good and the values for each question are also good or at least satisfactory. In fact, the latter are all much higher than 0.5 (higher than 0.7 in most cases) which shows that all the questions may be used because they fit in the structure defined by the others. As a result, the factorability of the correlation matrix is good which means that these data are appropriate for a factor analysis. Scale of the opinion on the Cluster 114 fit, nearly good (recall that a good fit occurs for a GFI of at least 0.9) and the Root mean square residual (RMSR) is 0.067 which shows a good fit (it is commonly accepted that an RMSR less than 0.1 shows a good fit). In short, all the coefficients show an acceptable fit. Table 49 Scale of the importance for the internationalization process - Residual matrix Item Item 1 2 3 4 5 6 7 8 1 -0.073 -0.056 0.126 -0.078 -0.032 -0.078 0.024 2 -0.073 -0.018 -0.064 -0.037 0.048 -0.025 -0.039 3 -0.056 -0.018 -0.010 0.033 0.051 -0.028 -0.037 4 0.126 -0.064 -0.010 -0.021 0.064 -0.110 0.037 5 -0.078 -0.037 0.033 -0.021 -0.055 0.010 -0.011 6 -0.032 0.048 0.051 0.064 -0.055 -0.096 -0.118 7 -0.078 -0.025 -0.028 -0.110 0.010 -0.096 -0.092 8 0.024 -0.039 -0.037 0.037 -0.011 -0.118 -0.092 9 0.017 0.002 -0.059 0.016 -0.055 -0.050 -0.110 0.115 10 0.066 -0.004 -0.014 0.040 -0.026 -0.001 -0.026 -0.105 11 0.049 0.006 -0.008 -0.094 -0.049 -0.034 -0.014 -0.059 12 -0.117 -0.044 -0.142 -0.051 -0.024 -0.036 0.169 0.062 13 0.010 -0.002 -0.063 -0.062 -0.009 -0.015 0.088 0.047 14 -0.004 0.059 0.077 0.031 -0.190 0.029 -0.096 0.019 15 -0.052 0.075 0.070 -0.139 0.016 -0.057 0.071 -0.049 16 0.061 -0.077 -0.153 0.022 0.023 0.058 -0.073 -0.121 Table 50 Scale of the importance for the internationalization process - Residual matrix (cont.) Item Item 9 10 11 12 13 14 15 16 1 0.017 0.066 0.049 -0.117 0.010 -0.004 -0.052 0.061 2 0.002 -0.004 0.006 -0.044 -0.002 0.059 0.075 -0.077 3 -0.059 -0.014 -0.008 -0.142 -0.063 0.077 0.070 -0.153 4 0.016 0.040 -0.094 -0.051 -0.062 0.031 -0.139 0.022 5 -0.055 -0.026 -0.049 -0.024 -0.009 -0.190 0.016 0.023 115 6 -0.050 -0.001 -0.034 -0.036 -0.015 0.029 -0.057 0.058 7 -0.110 -0.026 -0.014 0.169 0.088 -0.096 0.071 -0.073 8 0.115 -0.105 -0.059 0.062 0.047 0.019 -0.049 -0.121 9 -0.057 -0.116 0.002 -0.022 0.062 -0.054 0.095 10 -0.057 0.004 -0.056 -0.060 0.030 0.011 0.071 11 -0.116 0.004 -0.040 0.010 0.044 0.040 0.016 12 0.002 -0.056 -0.040 0.039 -0.057 -0.061 0.081 13 -0.022 -0.060 0.010 0.039 -0.058 -0.075 0.006 14 0.062 0.030 0.044 -0.057 -0.058 -0.044 -0.073 15 -0.054 0.011 0.040 -0.061 -0.075 -0.044 -0.168 16 0.095 0.071 0.016 0.081 0.006 -0.073 -0.168 Finally, Cronbach’s Alpha and the composite reliability values are displayed in the next table for each factor. Alfa’s value for the total questionnaire is 0.857 which is high and shows a strong reliability and internal consistency (recall that good reliability is usually considered for a value of at least 0.8). The first and the fourth factors show a high reliability, the second factor shows a good reliability, the third and the fifth factor show an acceptable reliability and the third (recall that, when a dimension includes few items, as the third one, since it includes 3 items only, the value of Alfa is often low but that does not mean a low reliability). Composite reliability is very high for the first, the second and the fourth factors and good for the third and the fifth, showing an appropriate construct reliability (recall that it is generally accepted that a composite reliability of at least 0.7 shows an appropriate construct reliability, even though lower values can still be acceptable). As a conclusion, both the total questionnaire and the factors show a good or at least acceptable reliability and internal consistency. Table 51 Scale of the importance for the internationalization process - Reliability of the questionnaire Factors Alpha CR 1 – Design and product development versus international business capabilities 0.851 0.916 2 – Networking with firms and the technological center of the Footwear Industry, plus accessing to financial resources are crucial to the 0.765 0.859 116 internationalization of its firms 3 – Investing in R&D and Innovation to promote/expand business in international markets 0.579 0.717 4 – Qualified Human Resources and Linguistic capabilities for International Business 0.845 0.913 5 – The promotion of own brand in international fairs can initiate or enhance the international activity of firms 0.645 0.739 Appendix F: Cluster’s importance/contribution In order to check whether these data are appropriate for a factorial analysis, the correlation matrix of the responses to the questions is displayed below showing many moderate and some high correlations. Table 52 Scale of the cluster’s importance/ contribution - Correlation matrix Item Item 1 2 3 4 5 6 1 1.000 0.650 0.584 0.488 0.403 0.457 2 0.650 1.000 0.685 0.682 0.331 0.585 3 0.584 0.685 1.000 0.760 0.282 0.554 4 0.488 0.682 0.760 1.000 0.369 0.596 5 0.403 0.331 0.282 0.369 1.000 0.664 6 0.457 0.585 0.554 0.596 0.664 1.000 7 0.568 0.684 0.514 0.709 0.383 0.525 8 0.463 0.522 0.408 0.392 0.599 0.570 9 0.535 0.535 0.407 0.556 0.674 0.588 10 0.388 0.210 0.323 0.230 0.408 0.396 11 0.276 0.324 0.418 0.448 0.408 0.498 12 0.465 0.348 0.479 0.477 0.543 0.389 117 Table 53 Scale of the cluster’s importance/contribution - Correlation matrix (cont.) Item Item 7 8 9 10 11 12 1 0.568 0.463 0.535 0.388 0.276 0.465 2 0.684 0.522 0.535 0.210 0.324 0.348 3 0.514 0.408 0.407 0.323 0.418 0.479 4 0.709 0.392 0.556 0.230 0.448 0.477 5 0.383 0.599 0.674 0.408 0.408 0.543 6 0.525 0.570 0.588 0.396 0.498 0.389 7 1.000 0.429 0.714 0.199 0.389 0.336 8 0.429 1.000 0.616 0.255 0.331 0.416 9 0.714 0.616 1.000 0.268 0.315 0.513 10 0.199 0.255 0.268 1.000 0.350 0.704 11 0.389 0.331 0.315 0.350 1.000 0.420 12 0.336 0.416 0.513 0.704 0.420 1.000 Futhermore, the Kaiser-Meyer-Olkin measure of sampling adequacy is displayed in the next table for each item and for the complete scale. The total value is 0.826 which is high and the values for each item are also high, very high or at least satisfactory (they are all much higher than 0.5 and are higher than 0.8 in most items). As a result, the factorability of the correlation matrix is very good which means that these data are appropriate for a factor analysis. Table 54 Scale of the cluster’s importance/contribution - KMO measure of sampling adequacy Item KMO Item KMO 1 0.897 7 0.792 2 0.893 8 0.918 3 0.851 9 0.825 4 0.821 10 0.660 5 0.831 11 0.888 6 0.821 12 0.721 Total 0.826 118 Therefore, a factor analysis with factor extraction by principal components was run. In order to select the number of factors, Kaiser’s rule (factors whose eigenvalues are larger than 1) selects 3 factors (a reasonable number), explaining 72.4% of the total variance (a good proportion). Pearson’s rule (proportion of the explained variance of at least 80%) leads to a solution with 5 factors (a number a little large), explaining 84.1% of the total variance. The rule based on the scree plot, also known as Cattel’s rule (factor with the largest decrease of the explained variance) leads to a solution with 5 factors (like Pearson’s rule). Thus, the solutions with 3 and 5 factors were tried and the former was adopted because the results were better both in terms of interpretation and meaning of the factors and in terms of the goodness of fit. Table 55 Scale of the cluster’s importance/contribution - Eigenvalues and variance explained by the factors Factor Eigenvalue % Variance Cumulative % 1 6.238 51.987 51.987 2 1.391 11.596 63.583 3 1.054 8.780 72.363 4 0.813 6.771 79.134 5 0.591 4.925 84.059 6 0.476 3.970 88.028 7 0.441 3.676 91.704 8 0.336 2.803 94.507 9 0.226 1.880 96.387 10 0.175 1.460 97.847 11 0.151 1.261 99.107 12 0.107 0.893 100.000 Appendix G: Factor Model – Cluster’s importance/ contribute In order to assess the quality of the factor model, the next table displays the residual matrix. There are only 28 (42%) nonredundant residuals with absolute value greater than 0.05, which shows a good fit (recall that a proportion of less than 50% shows a good fit). Furthermore, the Goodness of fit index (GFI) is 0.86 showing an acceptable good, almost good (recall that a good fit occurs for a GFI of at least 0.9) and the Root mean square residual (RMSR) is 0.066 which shows a good fnit (it is commonly accepted that an RMSR less than 0.1 shows a good fit). In short, all the coefficients show an acceptable fit. 119 Table 56 Scale of the cluster’s importance/ contribution - Residual matrix Item Item 1 2 3 4 5 6 1 0.020 -0.048 -0.159 -0.007 -0.101 2 0.020 -0.043 -0.093 -0.005 -0.002 3 -0.048 -0.043 -0.013 0.043 0.036 4 -0.159 -0.093 -0.013 0.047 0.016 5 -0.007 -0.005 0.043 0.047 0.016 6 -0.101 -0.002 0.036 0.016 0.016 7 -0.027 -0.063 -0.128 -0.005 -0.032 -0.080 8 0.011 0.044 0.076 -0.044 -0.107 -0.061 9 0.002 -0.054 -0.026 0.013 -0.051 -0.109 10 0.041 0.052 -0.050 -0.047 -0.051 0.017 11 -0.151 -0.062 -0.046 0.010 0.004 0.056 12 -0.012 0.003 -0.016 0.039 -0.024 -0.137 Table 57 Scale of the cluster’s importance/contribution - Residual matrix (cont.) Item Item 7 8 9 10 11 12 1 -0.027 0.011 0.002 0.041 -0.151 -0.012 2 -0.063 0.044 -0.054 0.052 -0.062 0.003 3 -0.128 0.076 -0.026 -0.050 -0.046 -0.016 4 -0.005 -0.044 0.013 -0.047 0.010 0.039 5 -0.032 -0.107 -0.051 -0.051 0.004 -0.024 6 -0.080 -0.061 -0.109 0.017 0.056 -0.137 7 -0.109 0.074 0.074 0.030 0.016 8 -0.109 -0.109 0.004 -0.011 0.009 9 0.074 -0.109 0.019 -0.068 0.081 10 0.074 0.004 0.019 -0.149 -0.073 11 0.030 -0.011 -0.068 -0.149 -0.128 12 0.016 0.009 0.081 -0.073 -0.128 Finally, Cronbach’s Alpha and the composite reliability values are displayed in the next table for each factor. Alfa’s value for the total questionnaire is 0.912 which is very high 120 and shows a very strong reliability and internal consistency (recall that good reliability is usually considered for a value of at least 0.8). The first and the second factors show a high reliability and the third one shows a good reliability. Composite reliability is very high for the first and the second factors and high for the third, showing a very good construct reliability (recall that it is generally accepted that a composite reliability of at least 0.7 shows an appropriate construct reliability, even though lower values can still be acceptable). As a conclusion, both the total questionnaire and the factors show a good reliability and internal consistency. Table 58 Scale of the cluster’s importance/contribution - Reliability of the questionnaire Factors Alpha CR 1 – Resources and Skills for Productivity and Notoriety 0.889 0.940 2 – Innovation and R&D for Marketing in International Markets 0.864 0.920 3 – Networking and Cooperation for International Activities 0.721 0.846