Knowledge flows, externalities and innovation networks

ABSTRACT

Knowledge flows, externalities and innovation networks. Regional Studies. This editorial introduces the articles addressing the multifaceted and composite phenomenon of knowledge flows and externalities within innovation networks. The papers offer a diversity of country contexts, as well as analytical approaches and methods. Starting with methods, they contain econometric analysis as well as social network analysis and combinations of the two. While most of the papers use econometric analysis, the level of analysis ranges from firm and individual level to aggregate analyses at the sectoral and regional level. Moreover, patents are the main starting point, but the proposed contributions show once more how different the approaches can be when using the same variables. The country context is diverse and includes data sources from US and European regions.

摘要

知识流、外部性与创新网络。Regional Studies. 此一编辑特刊引介处理创新网络中知识流与外部性的多重面向与复合现象之文章。这些文章提供多样的国家脉络以及分析取径与方法。这些文章从方法开始，便包含了计量经济分析、社会网络分析，以及两者的结合。虽然这些文章多半使用计量经济分析，但分析的层级却从企业与个人层级，到部门与区域层级的聚集分析。此外，专利是主要的起始点，然而这些文章提出的贡献却再次显示，它们虽使用相同的变因，但却如何採取不同的取径。国家的脉络相当多样，并且包含来自美国与欧洲区域的数据。

RÉSUMÉ

Les flux de connaissances, les effets externes et les réseaux d’innovation. Regional Studies. Cet éditorial présente des articles qui abordent le phénomène multi-dimensionnel et complexe des flux de connaissances et des effets externes au sein des réseaux d’innovation. Les articles fournissent une série de contextes nationaux, ainsi que des approches et des méthodes analytiques. Partant des méthodes, ces dernières comportent non seulement une analyse économétrique, mais aussi une analyse des réseaux sociaux, et des combinaisons des deux. Tandis que la plupart des articles emploient l’analyse économétrique, le niveau d’analyse va de l’entreprise et de l’individu jusqu’aux analyses globales faites aux niveaux sectoriel et régional. Qui plus est, les brevets constituent le principal point de départ, mais les articles en question montrent encore une fois jusqu’à quel point les approches peuvent s’avérer différentes lorsque l’on emploie les mêmes variables. Le contexte national varie et comprend des sources de données provenant des régions situées aux E-U et en Europe.

ZUSAMMENFASSUNG

Wissensströme, Externalitäten und Innovationsnetzwerke. Regional Studies. In diesem Leitartikel werden Beiträge über das facettenreiche und vielschichtige Phänomen der Wissensströme und Externalitäten in Innovationsnetzwerken vorgestellt. Die Artikel bieten ein breites Spektrum von Landeskontexten sowie verschiedene analytische Ansätze und Methoden. Hinsichtlich der Methoden enthalten sie ökonometrische Analysen, soziale Netzwerkanalysen oder eine Kombination beider Methoden. In den meisten Beiträgen kommen ökonometrische Analysen zum Einsatz, doch die Analyseebenen reichen von der Stufe der Firmen und Einzelpersonen bis hin zu Gesamtanalysen auf sektoraler und regionaler Ebene. Der wichtigste Ausgangspunkt sind Patente, doch die vorgestellten Beiträge zeigen erneut, wie die Verwendung derselben Variablen zu den verschiedensten Ansätzen führen kann. Die Landeskontexte sind von großer Vielfalt geprägt und enthalten Datenquellen aus US-amerikanischen und europäischen Regionen.

RESUMEN

Flujos de conocimiento, externalidades y redes de innovación. Regional Studies. En este editorial introducimos artículos que abordan el fenómeno polifacético y complejo de los flujos de conocimiento y las externalidades en las redes de innovación. Los artículos ofrecen una diversidad de contextos nacionales, enfoques y métodos analíticos. Respecto a los métodos, contienen análisis econométricos, análisis de las redes sociales y combinaciones de ambos. Aunque en la mayoría de los artículos se utilizan análisis econométricos, el nivel del análisis va desde empresas y personas hasta un análisis completo en el ámbito sectorial y regional. Moreover, patents are the main starting point, but the proposed contributions show once more how different the approaches can be when using the same variables. El contexto nacional es diverso e incluye fuentes de datos de regiones estadounidenses y europeas.

KEYWORDS:

• knowledge flows
• externalities
• innovation
• networks

关键词:

• 知识流
• 外部性
• 创新
• 网络

MOTS-CLÉS:

• flux de connaissances
• effets externes
• innovation
• réseaux

SCHLÜSSELWÖRTER:

• Wissensströme
• Externalitäten
• Innovation
• Netzwerke

PALABRAS CLAVES:

• flujos de conocimiento
• externalidades
• innovación
• redes

JEL:

• O31
• O33
• R11

Knowledge and innovation are key drivers of economic growth. Their contribution to the wealth of regions and nations has dramatically increased in recent decades due to the rising importance of the so-called ‘knowledge-based’ economy. In view of this, many scholars have endeavoured to offer a better understanding of the dynamics of knowledge production. In this framework, empirical and theoretical studies have documented the role of knowledge externalities, which are understood as ‘untraded interdependencies’ among firms in the Marshallian tradition (Meade, 1952; Storper, 1995; Viner, 1932).

Following seminal studies on the local and systemic dynamics of innovation processes, the role of geographical space in the generation of technological knowledge and the introduction of innovations has been the focus of many empirical and theoretical articles. Within this rapidly expanding field of enquiry, a taxonomy of externalities has been proposed based on the underlying mechanisms through which they contribute to the local accumulation of technological knowledge. Glaeser, Kallal, Scheinkman, and Shleifer (1992) identified three types of knowledge externalities from which the advantages of agglomeration stem. Firstly, Marshall–Arrow–Romer (MAR) externalities emerge out of the concentration of firms within a specific industry. Secondly, Jacobs’ externalities are engendered by the variety of firms and industries clustered in a region. Finally, Porter’s externalities point to the positive effect on local development of local competition between firms clustered in the same industry (for an exhaustive critical review of the empirical literature, see Beaudry & Schiffauerova, 2009; for a survey of the theoretical literature, see Basile & Usai, 2014).

Empirical analysis of knowledge externalities has focused, on the one hand, on documenting positive effects on the economic or innovation performances of co-located firms resulting from the pooling of knowledge sources in a specific place. This line of enquiry advances the avenue of research initiated by Jaffe (1989) and Griliches (1992), adopting an extended production function approach to test the effects of external knowledge sources on firms’ innovation dynamics. On the other hand, following Jaffe, Trajtenberg, and Henderson (1993), an increasing body of literature has investigated the patterns of knowledge externalities by looking at knowledge flow dynamics, proxied by the citations found in patent documents. The basic rationale is that patents are a proxy of codified knowledge, which can more easily flow between organizations. However, understanding the knowledge in patent documents requires access to the tacit knowledge of those who have generated that technology. This enables the absorptive capacity of the receiving organization, increasing the likelihood of successful appropriation of the external knowledge. These dynamics provide an explanation of the spatial concentration of knowledge spillovers proxied by patent documents (Breschi & Lissoni, 2001). More recently, empirical efforts have been directed towards the application of complementary methodological approaches, particularly social network analysis, to test the ‘labour market’ hypothesis to explain the spatial clustering of knowledge externalities (Breschi & Lissoni, 2009).

This collection of papers contributes to this stream of literature and offers a combination of heterogeneous and up-to-date studies focusing on the dynamics of knowledge externalities. They are characterized by different and yet complementary units of analysis, methodological approaches and theoretical underpinnings. They also contribute to the advancement of academic knowledge in three main respects.

Firstly, the application of an evolutionary perspective to the analysis of knowledge flows, especially network structure. The usual static approach to the analysis of innovative networks neglects their typically evolutionary dynamics. This is, however, crucial to assess the bidirectional relationship between innovative networks and regional contexts. The extant literature is much more concerned with the effects of geography of networks than with the differential impact of the latter on regional competiveness and the structure of economic activities. Changing networks can map onto changing relationships in sectors and technologies, in such a way that regional diversification in related and unrelated sectors, is likely to change over time (Boschma, 2017; Frenken, von Oort, & Verburg, 2007; Quatraro, 2010).

Secondly, the conceptualization of knowledge interactions as a means to foster interregional integration. Knowledge exchanges are not only a source of externalities but also reinforce multidimensional proximity among firms and regions (Boschma, 2005). Changing network structures implies the establishment of new links and the entry of new nodes. These dynamics are likely to shape how regions develop diversified interactions with one another. The extent of this process is reflected in local actors’ openness and ability to establish linkages with others at the international level, acknowledging the effects of international activities by multinational corporations on the local economy (Boschma & Iammarino, 2009; Crescenzi & Iammarino, 2017).

Finally, the measurement of local knowledge bases and firm performance. Traditionally measures of regional knowledge stock are on the right-hand side of the equation and firm patents or total factor productivity on the left (Malmberg, Malmberg, & Lundequist, 2000; Raspe & van Oort, 2011). Although informative, this provides a simplistic representation of the dynamics at work, calling for further measurement efforts.

The following papers address these points by tackling the issue of knowledge flows, externalities and innovation networks from different perspectives. They analyse innovation networks with a specific focus on the impact of network properties on regional innovation performances; the impact on firm-level performances of the regional knowledge base and degree of openness; and sectoral differences in spatial patterns of economic activity, with specific attention to localized externalities and the differential effects of proximity across different measures of knowledge flows.

The papers contribute from both a methodological and an empirical point of view. The first three papers are more focused on original research methods and indicators, whilst the subsequent six are more applied in nature, with analyses of knowledge flows dynamics, externalities and collaboration networks in different geographical contexts.

Bottazzi, Gragnolati, and Vanni (2016, in this issue) present an empirical analysis of the functional shape of local externalities. They prove that the usual linear specification is correct, at least when the quadratic form is tested in the context of local labour markets in Italy. They suggest extending the assessment to other scales, where congestion effects may be strong enough to introduce more complex and non-linear forms of externalities.

Antonelli and Colombelli (2017, in this issue) highlight the importance of discriminating between imitation and knowledge externalities to analyse simultaneously the role of external knowledge in both the upstream generation of knowledge and the downstream production of goods. The simultaneous test of the role of spillovers in a system of equations that includes a knowledge production function and a technology production function shows the presence of an integrated and sequential mechanism when applied to a large sample of Italian companies. They find that access to external knowledge plays a key role in increasing knowledge output and has positive indirect and direct effects on the firm productivity.

Bergé, Wanzenböck, and Scherngell (2017, in this issue) suggest a novel measure of regional centrality based on the concept of interregional bridging paths (indirect connections at the regional level). The main advantage of this measure is that it goes beyond the degree of participation in the network by focusing on the relative outward orientation and diversification of research and development (R&D) collaborations. They successfully test this measure with respect to other conventional centrality measures by using the European co-patent network at the NUTS-2 level.

Feldman, Menzel and Brökel (2017, in this issue) contribute to an assessment of the impact of institutional setting on the configuration of network ties in innovation networks. They focus on a co-inventors’ network in the information and communication technology (ICT) domain, located in Research Triangle Park, North Carolina, during the dot.com bubble. They find that the emergence of venture capital led to a large variance in connection patterns during the bubble, which disappeared thereafter, signalling the transition of firms’ strategies from exploration to exploitation.

The importance of institutions is also at the core of the paper by Drejer and Østergaard (2017, in this issue). They investigate patterns of collaboration between firms and academic institutions to show that employee-driven relations, as measured by employees’ and top-managers’ place of education and scientific discipline, strongly influence the likelihood that firms will collaborate with specific universities. The study also confirms the existence of separate and overlapping effects of employee-driven relations and geographical proximity.

The links between academic and firm inventors are also analysed by Coffano, Foray, and Pezzoni (2017, in this issue), who investigate the effects on the regional technological system of the structural characteristics of innovation networks. They show that regional innovation positively depends on a high level of inventors’ degree centrality. Most importantly, their results shed a new light on the mutually reinforcing effect of internal and external linkages: cross-regional linkages are more effective in fostering regional innovation when the inventors connected outside the region are, at the same time, embedded within the local technological community. In particular, they observe that academic inventors are crucial for bridging across regional borders.

The interplay between internal and external forces, at both the firm and local levels, in shaping externalities is the main focus of Grillitsch and Nilsson (2016, in this issue), based on Swedish micro-data from 2004 to 2011. Their estimations provide interesting and novel evidence of a negative relationship between firms’ knowledge intensity and local knowledge spillovers. In other words, knowledge-intensive firms compensate for lack of local knowledge spillovers by building knowledge pipelines with distance partners. At the same time, firms with low internal knowledge have more to gain from local spillovers because they are less affected by negative knowledge externalities (such as knowledge leakage and labour poaching) and less able to exploit global networks.

The complex interaction between geographical distance and other dimensions of proximity is, on the contrary, central to Baruffaldi and Raffo (2017, in this issue). They investigate the relationship between geographical and time proximity in shaping externalities, which take the form of invention replicas. Their results suggest that when it comes to recent and upcoming inventions, competitive incentives are high and localized knowledge flows increase the probability of duplication. Instead, they find that the duplication of less recent inventions is more likely at long distance because of lower awareness of the existence of a technology.

The interplay of distance, contiguity and technological proximity on cross-regional knowledge flows across European regions is the main focus of Quatraro and Usai (2016, in this issue). They distinguish between three types of knowledge flows: co-inventorship, applicant–inventor relationships and citation flows. They observe that, depending on the tacit/codified content of such flows, contiguity and technological proximity have a different moderating role on geographical distance. Moreover, across contiguous countries border effects may play reverse the potentially negative effects of regional remoteness from economic cores.

Overall, the papers in this issue contribute to an understanding of the mechanisms by which geography shapes the exchange of different kinds of knowledge as well as the effect of knowledge externalities on firm performance. They also investigate the impact of the structure of collaboration networks for knowledge production on the development of competitive advantages at the regional level. These works document that geography is likely to shape the structure of such networks, which in turn can affect the performances of geographical units of analysis. In sum, these papers offer a deeper theoretical and empirical understanding of these issues based on different analytical perspectives and on a variety of research methodologies applied to a rich set of geographical and institutional scenarios.

The papers also provide a valuable contribution in terms of policy implications. Although access to knowledge externalities remains a key driver of innovation for firms, the availability of local knowledge sources cannot be considered as a sufficient condition for firms to benefit from spillovers. Firms’ absorptive capacity may in fact shape how they take advantage of external knowledge. Therefore, regional innovation policies should be designed, on the one hand, to promote collaboration among players in the local innovation ecosystem and, on the other hand, to stimulate firms’ innovation efforts to improve their capacity to exploit effectively outsourced knowledge. Moreover, given the importance of geography in knowledge exchanges, the configuration of innovation networks affects the capacity of local agents to take advantage of external knowledge produced in other regions and nations. Regional innovation policies based on global pipelines should hence stimulate local agents to establish numerous connections with external organizations to increase their centrality in the network and facilitate access to extra-regional knowledge.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the editors.

ORCiD

Francesco Quatraro http://orcid.org/0000-0001-5746-2239

Stefano Usai http://orcid.org/0000-0001-5619-1350