Search

The Evolutionary Economic Geography literature has extensively investigated the impact of local capabilities on innovation and regional diversification. However, the role of external linkages has only received limited attention. The general concern of this thesis is how external linkages impact innovation and regional diversification with an evolutionary approach. This is further elaborated in four research questions. First, what is the impact of returnees on firm innovation? Second, what is the impact of MNEs on regional industrial diversification? Third, through what knowledge spillover channels can MNEs impact local export performance? Last, what is the impact of co-inventor networks on regional technological diversification? The findings of this thesis show that external linkages do matter for innovation and regional diversification both in the context of China and Europe. And external linkages could interact with relatedness and complexity in regional diversification, thus triggering different diversification paths. In more detail, the main content of each chapter could be summarized as follows. Chapter 1 presents the motivation, theoretical background, research questions, and outline of the thesis. Chapter 2 investigates the relationship between the global mobility of Chinese returnees and firm innovation for Chinese public listed firms. We construct a new dataset from CNRDS, PatentsView, CSMAR, and LinkedIn data. Our findings show that, in general, both returnee employees and managers have positive and significant impacts on firm innovation. The roles of returnee employees and returnee managers are complementary, that is, returnee employees as the agents of domestic patenting, and returnee managers as the ones for international patenting. We also find that the heterogeneities of returnees matter for firm innovation. Chapter 3 studies the relationship between foreign MNEs and industry entry in Chinese regions by linking the MNE research with regional diversification literature. The association between industry entry and the presence of MNEs in the same industry and related industries is tested for 331 Chinese prefectures between 1998 and 2007. Our findings indicate that the presence of MNEs in the same industry and related industries is positively and significantly associated with a higher probability of industry entry, and this association also holds for the entry of unrelated and complex industries. The effect of MNEs is larger for knowledge-intensive industries. Chapter 4 focuses on the role of relatedness and forward-backward linkages in the process of MNEs knowledge spillovers to domestic firms in the context of industrial export quantity and quality in Chinese regions. The effect of relatedness and forward-backward linkages is tested for 279 Chinese prefectures between 2000 and 2007. The results in Chapter 4 show that relatedness linkages matter in the context of export quantity and new exporting destination entry, whereas forward-backward linkages matter for the complexity of export products. Chapter 5 explores the role of both internal and external co-inventor networks in regional technological diversification, considering the effect of both relatedness and knowledge complexity with the OECD REGPAT dataset. Our results show that both internal co-inventor network connectivity and external co-inventor network linkages could facilitate regional technology entry and inhibit technology exit. However, too inwardly oriented, or outwardly connected co-inventor networks would hamper patent production. We also find suggestive evidence that both internal co-inventor network connectivity and external co-inventor network linkages could induce related and complex regional technological diversification. Chapter 6 concludes and discusses policy implications. It also discusses the limitations of this thesis and proposes directions for future research.

The Evolutionary Economic Geography literature has extensively investigated the impact of local capabilities on innovation and regional diversification. However, the role of external linkages has only received limited attention. The general concern of this thesis is how external linkages impact innovation and regional diversification with an evolutionary approach. This is further elaborated in four research questions. First, what is the impact of returnees on firm innovation? Second, what is the impact of MNEs on regional industrial diversification? Third, through what knowledge spillover channels can MNEs impact local export performance? Last, what is the impact of co-inventor networks on regional technological diversification? The findings of this thesis show that external linkages do matter for innovation and regional diversification both in the context of China and Europe. And external linkages could interact with relatedness and complexity in regional diversification, thus triggering different diversification paths. In more detail, the main content of each chapter could be summarized as follows. Chapter 1 presents the motivation, theoretical background, research questions, and outline of the thesis. Chapter 2 investigates the relationship between the global mobility of Chinese returnees and firm innovation for Chinese public listed firms. We construct a new dataset from CNRDS, PatentsView, CSMAR, and LinkedIn data. Our findings show that, in general, both returnee employees and managers have positive and significant impacts on firm innovation. The roles of returnee employees and returnee managers are complementary, that is, returnee employees as the agents of domestic patenting, and returnee managers as the ones for international patenting. We also find that the heterogeneities of returnees matter for firm innovation. Chapter 3 studies the relationship between foreign MNEs and industry entry in Chinese regions by linking the MNE research with regional diversification literature. The association between industry entry and the presence of MNEs in the same industry and related industries is tested for 331 Chinese prefectures between 1998 and 2007. Our findings indicate that the presence of MNEs in the same industry and related industries is positively and significantly associated with a higher probability of industry entry, and this association also holds for the entry of unrelated and complex industries. The effect of MNEs is larger for knowledge-intensive industries. Chapter 4 focuses on the role of relatedness and forward-backward linkages in the process of MNEs knowledge spillovers to domestic firms in the context of industrial export quantity and quality in Chinese regions. The effect of relatedness and forward-backward linkages is tested for 279 Chinese prefectures between 2000 and 2007. The results in Chapter 4 show that relatedness linkages matter in the context of export quantity and new exporting destination entry, whereas forward-backward linkages matter for the complexity of export products. Chapter 5 explores the role of both internal and external co-inventor networks in regional technological diversification, considering the effect of both relatedness and knowledge complexity with the OECD REGPAT dataset. Our results show that both internal co-inventor network connectivity and external co-inventor network linkages could facilitate regional technology entry and inhibit technology exit. However, too inwardly oriented, or outwardly connected co-inventor networks would hamper patent production. We also find suggestive evidence that both internal co-inventor network connectivity and external co-inventor network linkages could induce related and complex regional technological diversification. Chapter 6 concludes and discusses policy implications. It also discusses the limitations of this thesis and proposes directions for future research.

This research stresses the role of geography in the formation of network ties that channel ideas, knowledge and technology. It introduces novel aspects to consider while studying the drivers and outcomes of spatial network formation. The thesis establishes that spinoff background matters for tie formation in cluster knowledge networks. It shows that by searching for the drivers of network tie creation and tie persistence separately, we can better understand the evolution of industry clusters. It also provides evidence that co-location and complexity of technologies determine their relatedness ties. Furthermore, this research establishes a connection between forming structurally important linkages and being successful in urban agglomeration of creative industries. Besides extending the literature on spatial networks with new aspects, each of the chapters reinforce that geography is an important determinant of network tie formation.

This research stresses the role of geography in the formation of network ties that channel ideas, knowledge and technology. It introduces novel aspects to consider while studying the drivers and outcomes of spatial network formation. The thesis establishes that spinoff background matters for tie formation in cluster knowledge networks. It shows that by searching for the drivers of network tie creation and tie persistence separately, we can better understand the evolution of industry clusters. It also provides evidence that co-location and complexity of technologies determine their relatedness ties. Furthermore, this research establishes a connection between forming structurally important linkages and being successful in urban agglomeration of creative industries. Besides extending the literature on spatial networks with new aspects, each of the chapters reinforce that geography is an important determinant of network tie formation.

Innovation is a combinatorial process, in which novelty emerges through the combination or recombination of existing knowledge and materials. It implies that the process of knowledge creation is partly embedded in knowledge relations derived from the process of knowledge sourcing. Notwithstanding the straightforward argument that knowledge relations bring about innovation, our understanding of how this process unfolds over time is still limited. This motivates the present PhD thesis, which focuses on the following questions: What are the key driving forces of knowledge transfer? and How do these forces vary between regions and sectors? This PhD thesis includes a number of case studies that provide empirical insights into these research questions. The study in Chapter 2 investigated the impact of space on the likelihood and effectiveness of knowledge exchange. In the analysis, a unique dataset was employed that covers co-inventive patent data over more than 170 years (1836–2010). The empirical findings underline the continuous importance of geographical proximity for the establishment of knowledge ties as well as the increasing importance of geographical proximity for the intensity of knowledge exchange. Chapter 3 focuses on the impact of social contexts on organizations’ knowledge sourcing behaviour. To study the effect of social contexts, we investigated the evolution of the knowledge network in the Berlin biotech cluster (1992–16), where organizations from the former East and West Germany are still embedded in completely different institutional and cultural contexts. By controlling for universal tie formation forces at the dyad (e.g. geographical proximity) and network (e.g., transitivity) levels, the study isolates the effect of context-related factors. Despite a large degree of labour mobility and top-down subsidization of East–West collaborations, we found that in recent years knowledge ties are more likely to have been established among organizations within the same social context. Chapter 4 explores knowledge-sourcing networks in low-tech sectors. The study is based on primary data in the Swarzędz and Kępno clusters in Poland. Contrary to the cluster lifecycle model, the findings show that firms and associated organizations fail to establish extra-regional knowledge ties. Instead, small and medium-sized firms as well as local institutions are the main agents of learning and innovation. The empirical study in Chapter 5 focuses on the importance of sectoral attributes shaping knowledge sourcing processes and how these evolve over time. The investigation used a dataset of R&D projects in Norway (2005–16) to analyse the dynamic impact of a set of proximity dimensions on collaborative tie formation in four sectors. An analysis based on the gravity model revealed that the impact of cognitive proximity and its magnitude vary between sectors and across time periods. Chapter 6 provides a conceptual framework to alleviate issues regarding network-related interventionist policies. Finally, Chapter 7 discusses the main findings of the preceding chapters and point out a number of limitations and unresolved problems. Accordingly, this suggests several possibilities for future research and discusses how these subsequent works will further contribute to our understanding of the driving forces of knowledge sourcing.

Innovation is a combinatorial process, in which novelty emerges through the combination or recombination of existing knowledge and materials. It implies that the process of knowledge creation is partly embedded in knowledge relations derived from the process of knowledge sourcing. Notwithstanding the straightforward argument that knowledge relations bring about innovation, our understanding of how this process unfolds over time is still limited. This motivates the present PhD thesis, which focuses on the following questions: What are the key driving forces of knowledge transfer? and How do these forces vary between regions and sectors? This PhD thesis includes a number of case studies that provide empirical insights into these research questions. The study in Chapter 2 investigated the impact of space on the likelihood and effectiveness of knowledge exchange. In the analysis, a unique dataset was employed that covers co-inventive patent data over more than 170 years (1836–2010). The empirical findings underline the continuous importance of geographical proximity for the establishment of knowledge ties as well as the increasing importance of geographical proximity for the intensity of knowledge exchange. Chapter 3 focuses on the impact of social contexts on organizations’ knowledge sourcing behaviour. To study the effect of social contexts, we investigated the evolution of the knowledge network in the Berlin biotech cluster (1992–16), where organizations from the former East and West Germany are still embedded in completely different institutional and cultural contexts. By controlling for universal tie formation forces at the dyad (e.g. geographical proximity) and network (e.g., transitivity) levels, the study isolates the effect of context-related factors. Despite a large degree of labour mobility and top-down subsidization of East–West collaborations, we found that in recent years knowledge ties are more likely to have been established among organizations within the same social context. Chapter 4 explores knowledge-sourcing networks in low-tech sectors. The study is based on primary data in the Swarzędz and Kępno clusters in Poland. Contrary to the cluster lifecycle model, the findings show that firms and associated organizations fail to establish extra-regional knowledge ties. Instead, small and medium-sized firms as well as local institutions are the main agents of learning and innovation. The empirical study in Chapter 5 focuses on the importance of sectoral attributes shaping knowledge sourcing processes and how these evolve over time. The investigation used a dataset of R&D projects in Norway (2005–16) to analyse the dynamic impact of a set of proximity dimensions on collaborative tie formation in four sectors. An analysis based on the gravity model revealed that the impact of cognitive proximity and its magnitude vary between sectors and across time periods. Chapter 6 provides a conceptual framework to alleviate issues regarding network-related interventionist policies. Finally, Chapter 7 discusses the main findings of the preceding chapters and point out a number of limitations and unresolved problems. Accordingly, this suggests several possibilities for future research and discusses how these subsequent works will further contribute to our understanding of the driving forces of knowledge sourcing.

Broadly speaking, this dissertation studied how technological change operates to determine the fortune of places. Thus linking the global urgency to understand the current level of disparities in terms of wealth and income with the idea that these disparities are, to a large extent, determined by the capacity of individuals to collectively accumulate knowledge and know-how. Such a venture required overcoming two major obstacles. On the one hand, it required developing advanced modellistic tools capable characterizing the mechanisms and channels that relate technological change with economic progress. On the other hand, assessing the impact of the emergence, development, and diffusion of key technologies over time and space required constructing a unique database with disaggregated information on inventive activities of places for the last 150 years. Results show that the development of new technologies is a highly cumulative and path dependent process, in which technological upgrading emerges out of pre-existing knowledge bases and patterns of specialization. The likelihood of diversification into a new technological activity is higher for those domains that are related to countries' existing profile of competences. Countries climb the ladder of technological development rung by rung, as new capabilities have to be built-up gradually. Additionally, technological diversification is more heavily constrained by related indigenous capabilities at early stages of development. This dissertation also documents the existence of a high degree of heterogeneity in terms of how countries distribute their technological capabilities across sectors. More importantly, this heterogeneity appears to be highly correlated with international trade flows. To explain this, Chapter 4 brings forward the crucial role of the allocation of technological capabilities to determine trade patterns by incorporating within-country technological dispersion into an Eaton-Kortum trade model. As a result, this model is able to explain how countries’ overall comparative advantage (which determines overall exports) depend on technological dispersion. Countries with lower input costs and low levels of technological accumulation can benefit (increase their overall exports) by lowering the dispersion of their technological activities. Last but not least, Chapter 5 comes to fill one of the main empirical gaps in the innovation literature, providing comprehensive empirical evidence about the effect of the diffusion of a General Purpose Technology (GPT) over the economy. By relying on historical patents documents, this chapter provides a comprehensive view of the emergence, evolution, development, and diffusion of Electrical & Electronic (E&E) technologies in their historical context. It shows that the behaviour of E&E technologies between 1860 and 1930 is in line with what can be expected from a GPT, namely: Above-average growth rates in patenting activity, the development of a wide variety of innovation complementarities, and a high degree of pervasiveness in the U.S. inventive structure. More importantly, and after combining the database developed in Chapter 3 with economic and demographic data at the county level, this Chapter 5 shows that the adoption of E&E technologies had a positive effect on wages and the income per capita growth of places after the 1900's, generating uneven and durable disparities across counties.

Broadly speaking, this dissertation studied how technological change operates to determine the fortune of places. Thus linking the global urgency to understand the current level of disparities in terms of wealth and income with the idea that these disparities are, to a large extent, determined by the capacity of individuals to collectively accumulate knowledge and know-how. Such a venture required overcoming two major obstacles. On the one hand, it required developing advanced modellistic tools capable characterizing the mechanisms and channels that relate technological change with economic progress. On the other hand, assessing the impact of the emergence, development, and diffusion of key technologies over time and space required constructing a unique database with disaggregated information on inventive activities of places for the last 150 years. Results show that the development of new technologies is a highly cumulative and path dependent process, in which technological upgrading emerges out of pre-existing knowledge bases and patterns of specialization. The likelihood of diversification into a new technological activity is higher for those domains that are related to countries' existing profile of competences. Countries climb the ladder of technological development rung by rung, as new capabilities have to be built-up gradually. Additionally, technological diversification is more heavily constrained by related indigenous capabilities at early stages of development. This dissertation also documents the existence of a high degree of heterogeneity in terms of how countries distribute their technological capabilities across sectors. More importantly, this heterogeneity appears to be highly correlated with international trade flows. To explain this, Chapter 4 brings forward the crucial role of the allocation of technological capabilities to determine trade patterns by incorporating within-country technological dispersion into an Eaton-Kortum trade model. As a result, this model is able to explain how countries’ overall comparative advantage (which determines overall exports) depend on technological dispersion. Countries with lower input costs and low levels of technological accumulation can benefit (increase their overall exports) by lowering the dispersion of their technological activities. Last but not least, Chapter 5 comes to fill one of the main empirical gaps in the innovation literature, providing comprehensive empirical evidence about the effect of the diffusion of a General Purpose Technology (GPT) over the economy. By relying on historical patents documents, this chapter provides a comprehensive view of the emergence, evolution, development, and diffusion of Electrical & Electronic (E&E) technologies in their historical context. It shows that the behaviour of E&E technologies between 1860 and 1930 is in line with what can be expected from a GPT, namely: Above-average growth rates in patenting activity, the development of a wide variety of innovation complementarities, and a high degree of pervasiveness in the U.S. inventive structure. More importantly, and after combining the database developed in Chapter 3 with economic and demographic data at the county level, this Chapter 5 shows that the adoption of E&E technologies had a positive effect on wages and the income per capita growth of places after the 1900's, generating uneven and durable disparities across counties.

The research aims at improving our understanding of the link between economic structure and growth, by tackling a number of open questions. First, it asks whether economic structure – meaning the distribution of production factors among different industries – can significantly explain differences in standards of living across nations and regions. To this aim it proposes a method to decompose cross-country differences in total factor productivity (TFP) into a technological component - a Place effect that depends on the overall productivity of a country - and an allocation component, a Composition effect, which depends on whether factors of productions are allocated to productive or unproductive industries. The analysis estimates that 1/4 of inequality is due to the Composition effect, while 3/4 to the Place effect. Second, it highlights that, theoretically, economic structure is a cause of these differences only if industries have asymmetric externalities. It then documents the heterogeneity across sectors in the impact labor and input-output links have on industry agglomeration. Exploiting the available degrees of freedom in coagglomeration patterns, it estimates the industry-specific benefits of sharing labor needs and supply links with local firms. On aggregate, coagglomeration patterns of services are at least as strongly driven by input-output linkages as those of manufacturing, whereas labor linkages are much more potent drivers of coagglomeration in services than in manufacturing. Third, as the analysis on coagglomeration shows that growth path might result in lock-in, a follow up question is how can countries or region move away from their pre-determined path. The research then looks at the impact of return migration in Mexican city-industries. It shows that the dynamics of migration with the US are such that returnees change the distribution of skills across Mexican cities. Using the exogenous variation in number of city-industry returnees which comes from a) deportation from US authorities and b) the general shock in the late 2000s in the American economy, it estimates an elasticity of employment growth of 9%, while mixed results are found with respect to wage growth. Fourth, it is questioned whether technological change is sufficient for structural change and growth and it is noted that firms in emerging economies not only have to produce high quality, cost-competitive goods, but also win the resistance of consumers in the world market, who are often adverse to purchasing products from countries that yet have to build a reputation. The analysis argues that this country-of-origin bias significantly influences the chances of leadership change. A model that aims at capturing the endogenous dynamics of demand building and leapfrogging is proposed. In sectors with high monopoly power acquiring a superior technology is not sufficient for a latecomer country to become leader, unless a significant share of consumers is aware of the quality of its products. An extension of the model to multiple sectors shows that a latecomer country remains specialized into low-value undifferentiated goods, even after overtaking the technology of the leading country.

The research aims at improving our understanding of the link between economic structure and growth, by tackling a number of open questions. First, it asks whether economic structure – meaning the distribution of production factors among different industries – can significantly explain differences in standards of living across nations and regions. To this aim it proposes a method to decompose cross-country differences in total factor productivity (TFP) into a technological component - a Place effect that depends on the overall productivity of a country - and an allocation component, a Composition effect, which depends on whether factors of productions are allocated to productive or unproductive industries. The analysis estimates that 1/4 of inequality is due to the Composition effect, while 3/4 to the Place effect. Second, it highlights that, theoretically, economic structure is a cause of these differences only if industries have asymmetric externalities. It then documents the heterogeneity across sectors in the impact labor and input-output links have on industry agglomeration. Exploiting the available degrees of freedom in coagglomeration patterns, it estimates the industry-specific benefits of sharing labor needs and supply links with local firms. On aggregate, coagglomeration patterns of services are at least as strongly driven by input-output linkages as those of manufacturing, whereas labor linkages are much more potent drivers of coagglomeration in services than in manufacturing. Third, as the analysis on coagglomeration shows that growth path might result in lock-in, a follow up question is how can countries or region move away from their pre-determined path. The research then looks at the impact of return migration in Mexican city-industries. It shows that the dynamics of migration with the US are such that returnees change the distribution of skills across Mexican cities. Using the exogenous variation in number of city-industry returnees which comes from a) deportation from US authorities and b) the general shock in the late 2000s in the American economy, it estimates an elasticity of employment growth of 9%, while mixed results are found with respect to wage growth. Fourth, it is questioned whether technological change is sufficient for structural change and growth and it is noted that firms in emerging economies not only have to produce high quality, cost-competitive goods, but also win the resistance of consumers in the world market, who are often adverse to purchasing products from countries that yet have to build a reputation. The analysis argues that this country-of-origin bias significantly influences the chances of leadership change. A model that aims at capturing the endogenous dynamics of demand building and leapfrogging is proposed. In sectors with high monopoly power acquiring a superior technology is not sufficient for a latecomer country to become leader, unless a significant share of consumers is aware of the quality of its products. An extension of the model to multiple sectors shows that a latecomer country remains specialized into low-value undifferentiated goods, even after overtaking the technology of the leading country.

The question of why some countries or regions experience higher growth rates and levels of development than others has been debated and studied for long time and remains one of the most important issues in political and economic debates. The aim of this thesis is to investigate some of the factors that contribute to the unevenness of economic performance among European regions, focusing in particular regional industrial composition, spatial and network relations and the technological and institutional characteristics of the regions. The role of industrial structure and composition of regions is assessed in relation to their knowledge and technological features (Chapter 1). Our results indicate that related variety exerts positive effects in only regions characterized by higher technological and knowledge capabilities, whereas no significant effect is found for average or less advanced areas. More generally, we conclude that different levels of knowledge endowment may condition the way knowledge-related externalities affect the local economy. Following this insight, we study spatial and network relations as sources for new knowledge (Chapter 2). In line with the previous literature we find that spatial and trade linkages indeed contribute to channeling new ideas and technologies in the regional economy, with a positive effect on local productivity. However, network connections to most advanced regions – which would potentially grant access to highly valuable knowledge - have no beneficial effect unless the “linking-in” region presents high level of absorptive capacity. Recent literature on institutions has shown how quality of government and social capital are important determinant of economic and innovative performance. Following some contributions in Evolutionary Economic Geography, we investigate how institutional factors interplay with industrial compositions. In Chapter 3, we show how industrial relatedness and bridging social capital are key drivers of regional diversification. Unlike, formal institutions do not directly affect regional branching, though some evidence suggests that they contribute to limiting the potentially negative effects of bonding social capital. The role of industrial relatedness in allowing knowledge and ideas to flow is further studied in relation to the presence, entry and exit of multinational enterprises (MNEs) (Chapter 4). We find that the number of MNEs in related sectors has significant effects on industrial employment only in knowledge-intensive industries in advanced regions. Additionally, the positive relation between employment level and proximity to industries attracting foreign companies provides supportive evidence of short-term knowledge spillovers. However, both intra- and inter-industry spillovers of MNEs are heterogeneous across industries and regions. Positive MNE externalities appear to be stronger in advanced sectors and, more surprisingly, in Eastern EU regions. These regions, however, are also those that suffer the most in the case of MNE exits.

The question of why some countries or regions experience higher growth rates and levels of development than others has been debated and studied for long time and remains one of the most important issues in political and economic debates. The aim of this thesis is to investigate some of the factors that contribute to the unevenness of economic performance among European regions, focusing in particular regional industrial composition, spatial and network relations and the technological and institutional characteristics of the regions. The role of industrial structure and composition of regions is assessed in relation to their knowledge and technological features (Chapter 1). Our results indicate that related variety exerts positive effects in only regions characterized by higher technological and knowledge capabilities, whereas no significant effect is found for average or less advanced areas. More generally, we conclude that different levels of knowledge endowment may condition the way knowledge-related externalities affect the local economy. Following this insight, we study spatial and network relations as sources for new knowledge (Chapter 2). In line with the previous literature we find that spatial and trade linkages indeed contribute to channeling new ideas and technologies in the regional economy, with a positive effect on local productivity. However, network connections to most advanced regions – which would potentially grant access to highly valuable knowledge - have no beneficial effect unless the “linking-in” region presents high level of absorptive capacity. Recent literature on institutions has shown how quality of government and social capital are important determinant of economic and innovative performance. Following some contributions in Evolutionary Economic Geography, we investigate how institutional factors interplay with industrial compositions. In Chapter 3, we show how industrial relatedness and bridging social capital are key drivers of regional diversification. Unlike, formal institutions do not directly affect regional branching, though some evidence suggests that they contribute to limiting the potentially negative effects of bonding social capital. The role of industrial relatedness in allowing knowledge and ideas to flow is further studied in relation to the presence, entry and exit of multinational enterprises (MNEs) (Chapter 4). We find that the number of MNEs in related sectors has significant effects on industrial employment only in knowledge-intensive industries in advanced regions. Additionally, the positive relation between employment level and proximity to industries attracting foreign companies provides supportive evidence of short-term knowledge spillovers. However, both intra- and inter-industry spillovers of MNEs are heterogeneous across industries and regions. Positive MNE externalities appear to be stronger in advanced sectors and, more surprisingly, in Eastern EU regions. These regions, however, are also those that suffer the most in the case of MNE exits.

The contribution of this thesis is to widen the focus of the literature on related variety. So far, this line of research has a strong focus on technological relatedness and cognitive proximity (Boschma et al. 2015). This thesis puts emphasis on three dimensions of relatedness that not yet been at the heart of research: input-output linkages and vertically related variety, the occupational dimensions of relatedness as well as insights from a project level R&D efforts and their relevance for the analysis of technological relatedness. We contribute to the understanding of how the specific composition and the degree of relatedness of economic agents in space shape their ability to generate variety in terms of entry of new occupational specialisations in regions. We enhance the understanding of how related variety affects different regional outcomes such employment growth and innovation and offer insights into the structure of inter-industry relations from a technological perspective and explains how the rise of certain technologies shapes inter-industry relatedness patterns.

The contribution of this thesis is to widen the focus of the literature on related variety. So far, this line of research has a strong focus on technological relatedness and cognitive proximity (Boschma et al. 2015). This thesis puts emphasis on three dimensions of relatedness that not yet been at the heart of research: input-output linkages and vertically related variety, the occupational dimensions of relatedness as well as insights from a project level R&D efforts and their relevance for the analysis of technological relatedness. We contribute to the understanding of how the specific composition and the degree of relatedness of economic agents in space shape their ability to generate variety in terms of entry of new occupational specialisations in regions. We enhance the understanding of how related variety affects different regional outcomes such employment growth and innovation and offer insights into the structure of inter-industry relations from a technological perspective and explains how the rise of certain technologies shapes inter-industry relatedness patterns.