Public research intensity and the structure of German R&D networks: a comparison of 10 technologies

Abstract

A rich literature has emerged that analyzes the emergence of inter-organizational networks. However, little is still known about how and why the (global) structures of such networks differ between technologies. Based on a rich database covering subsidized R&D cooperation in Germany, we compare cooperation networks of 10 distinct technologies. A particular focus is hereby on the existence of systematic differences between networks that are dominated by firms and those that primarily connect public research organizations. Among others, we find that the first tend to be larger and less dense with big firms being the most central organizations. In comparison, networks dominated by public actors are stronger centralized and involve more isolates.

Keywords:

• R&D subsidies
• network
• cooperation
• Foerderkatalog
• Germany

Notes

Ter Wal and Boschma (2009) discuss the pros and cons of the various forms of data collection for the application of social network analysis in the context of economic geography.

The classification scheme as well as the assignment of projects to activity classes has been subject to some change over the years (Czarnitzki et al. 2002). For the cross-sectional approach taken in this paper this does however not causes significant biases.

The reason for choosing this time period is that this data has been cleaned in an earlier project as well as merged with some additional information, e.g. the types of actors. We compared the data to other periods but did not find strong variance in terms of project numbers or the share of cooperative projects.

See Broekel and Graf (2010) for a more detailed discussion of the reasons for this mode of aggregation and the consequences thereof.

The latter also includes correlations with a number of alternative statistics, which however do not seem to contain significant additional information.

For the construction of this network we used automatic name matching procedures including fuzzy string matching.