Differentiated Knowledge Bases and the Nature of Innovation Networks

Abstract

It is argued in this paper that the nature of innovation networks can vary substantially with regard to the type of knowledge that is critical for innovation. Subject to the knowledge base of an industry, networks between companies can differ in various aspects, such as their geographical configuration, their structure, the type of actors holding a strategic position and the type of relations between actors. The paper comprises a conceptual discussion on social capital theory and networks, followed by a theoretically informed discussion on differentiated knowledge bases and innovation networks, which is subsequently illustrated with empirical material. The empirical analysis is based on social network analysis in association with exclusive data about patterns of cooperation and knowledge exchange in a number of regional industries located in different parts of Europe. The findings suggest that networks in analytical industries are not much constrained by geographical distance; knowledge is exchanged in a highly selective manner between research units and scientists in globally configured epistemic communities. Synthetic industries source knowledge within nationally or regionally configured networks between suppliers and customers, and within communities of practice. Symbolic industries rely on knowledge that is culturally defined and highly context specific, resulting in localized networks that are temporary and flexible in nature.

Notes

The data are drawn from an EU collaborative research project entitled “CRA”. All project partners are gratefully acknowledged for collecting and providing the data.

The term “social capital” occurs first in an article by Loury (1977), who criticises the dominant neoclassical theory to be incapable of taking social context into account.

van Egeraat and Curran (2013) use a similar approach to study knowledge exchange in the Irish biotechnology industry, and find that neither formal R&D networks, nor informal directorship networks do fully reflect the actual scale and scope of knowledge flows between companies.

Studies on the individual cases are published, amongst others, in the European Planning Studies special issue “Constructing Regional Advantage: Towards State-of-the-Art Regional Innovation System Policies in Europe?” (Asheim et al., 2011b).

Degree centrality is a concept in graph theory and calculated using Freeman's (1979) approach. The degree of a vertex is the number of vertices adjacent to that vertex in a graph. For a given graph G := (V, E) with |V| vertices and |E| edges, the degree centrality of vertex ν is defined as C_D (ν) = deg(ν).

A component of a graph is a sub-graph in which any two vertices are connected to each other by paths. The component size of a vertex is the number of vertices directly or indirectly connected to that vertex in a graph (Wasserman & Faust, 1994).

In graph theory, network density reflects the total number of edges divided by the total number of possible edges, that is, the percentage of all possible edges that are actually present in a network (Hanneman & Riddle, 2005).

Numbers in brackets display median values. Median values are used in order to account for the skewed distribution of the network measures.

The space and aviation industries in the Netherlands were excluded from the geographical analysis, as the research design did not distinguish between regional and national collaboration.