Networked research: European policy intervention in ICTs

Abstract

We use social network analysis to evaluate ‘behavioural’ additionality aspects of public programmes supporting research and development (R&D). The paper appraises empirically the partnership and knowledge networks created around the R&D activities of the Information Society Priority of the Sixth Research Framework Programme of the European Community. These emergent, scale-free networks are found to play an important role in generating and, especially, in diffusing knowledge by attracting key industry actors and by strengthening overall network connectivity through public support. Public policy should try to facilitate the development of more European organisations that can be characterised as global network hubs, on the one hand, and to draw larger numbers of the most dynamic small and medium-sized enterprises (SMEs) into these programmes, on the other, to avoid technological lock-ins and mitigate the resistance or network reorientation toward more productive research areas.

Keywords:

• technology and innovation studies
• network analysis
• patent analysis
• information and communications technology
• strategic alliances

Acknowledgements

This paper draws on the report ‘Evaluation of Progress towards a European Research Area for Information Society Technologies’, European Commission, DG Information Society and Media, 2006. The authors acknowledge generous funding by the DG, and its evaluation unit C3, for carrying out the background evaluative study. In particular, Peter Johnston and Frank Cunningham from that unit were instrumental in the success of the study. The authors also wish to thank various participants to workshops in Brussels as well as to the annual conferences of the American Evaluation Association and the Canadian Evaluation Society and of the Association for Policy Analysis and Management in 2005 for useful comments. The authors would also like to acknowledge the financial support of the Italian Ministry of Education, Universities and Research (FIRB, project RISC – RBNE039XKA). The contents of the paper do not necessarily reflect the views and policies of the European Commission. The paper presents the analysis and opinion of the authors who are exclusively responsible for any mistakes and misconceptions.

Notes

The Seventh Framework Programme will run for seven years instead, 2007–2013.

In addition to the instruments listed below, the Commission has introduced in the past few years two other that are quite new and jointly with the FP projects aim at structuring and shaping the ERA, namely the ERANets and the Joint Technology Initiatives (JTIs). The ERANets scheme is about a bottom-up, self-organised coordination and cooperation of national and regional programmes. JTIs have the specific role to implement the research programme of a European Technology Platform (ETP). ETPs help industrial and academic research communities and other stakeholders in specific technology fields to coordinate their research and to tailor it to a common strategic research agenda, which sets out research and development goals, time frames and action plans for technological advances that are relevant to industry and society. Both these instruments are characterised by a bottom-up approach, thus being complementary to more traditional FP funding instruments where the priorities are defined by the commission with a top-down approach. ERANets and JTIs were not part of this research.

European Commission (2005a,b).

J. Stefan Institute (1999) and RAND Europe (2005).

This result and the previous one are not surprising since other studies analysing FP networks have reached similar conclusions, e.g. Breschi and Cusmano (2004).

With reference to the hypothetical sociogram depicted in Figure 1, nodes a, b and c present the highest values of degree centrality being connected to four other organisations.

With reference to the hypothetical sociogram (Figure 1), it is intuitive that node a has a high influence as a network connector. For example, the shortest path between organisations d and g has length 3 and organisations a lies on it. If one takes all possible pairs of organisations (excluding a) and counts the number of shortest paths connecting them, it turns out that organisations a lies on 8 out of 15 of them. The betweenness centrality of organisations a is therefore equal to 8/15=0.53. It is therefore highly influential in mediating knowledge flows taking place among the nodes in the network. By contrast, organisations b lies only on three shortest paths (connecting node d with a, f and g) and is thus characterised by a lower value of betweenness centrality (0.20).

‘If you go for the biggest nodes and take a couple of them out, you can break the system into clusters that don't communicate with each other’ (Albert, Jeong, and Barabasi 2000, 381) This idea developed by Barabasi and co-authors in several studies has been applied in the context of this study. It is enough to focus on the size of the greatest subpart of a network: the so-called giant component (i.e. the greatest set of actors directly or indirectly connected). As one starts deleting the top ranked organisations in terms of centrality as well as their links, the size of the giant component as a percentage of nodes included in it drops dramatically. In the case of the global network, deleting the top 2% of hubs reduces the giant component to one third of its initial size. The 2% cut-off is obviously arbitrary. However, we also considered different values (both higher and lower then 2%) to check for robustness.

The participants in the two workshop were EU officers of different DGs of the Europe Commission (Research, Information Society and Media, etc.), delegates of National or Regional organisations, academic experts and members of organisations participating in the IST projects. We ran also about 10 interviews with representatives of project participants. Only main organisations, both public and private, have been selected. See CESPRI (2006, 91) onwards, for questions and (aggregated) answers.

The percentage values for each type of organisation have been weighted according to the ranking in the overall list of hubs. The rationale for using weighted percentages is that organisations ranking high in the list of hubs are likely to be relatively more influential than organisations ranking low. The weights have been defined in the following way: , where r _i is the ranking of organisation i and max r is the maximum value of the ranking. Please note that the weights sum to 1.

Firms participating in IST projects have been consolidated according to the ultimate parent company. For example, Nokia Italy has been considered as part of the Nokia group. The research labs of large public research organisations (e.g. Fraunhofer Gesellschaft) have been also consolidated. As a robustness check, we have recalculated the list of hubs by considering each subsidiary or research lab as an independent unit. Results are not significantly sensitive to the consolidation of companies and research laboratories.

Company self citations have been excluded.

In this section, IST applications and IST development hubs are examined jointly.

For the empirical analysis leading to these observations see the original report ‘Evaluation of Progress towards a European Research Area for Information Society Technologies’, European Commission, DG Information Society and Media (2006), by the authors of this paper.