Empirical work shows that networks of research and development alliances are asymmetric, with a small number of firms involved in the majority of partnerships. This article investigates the welfarerelevant effects of such concentrated networks in a model of network formation in an oligopolistic market. We find that concentration is a typical characteristic of a socially efficient network, when the costs of collaborative activity are significant. Moreover, expanding on prior work relating to strategically stable inter-firm networks, we compare the stable and the efficient structures. Our findings suggest that the real-world networks might even exhibit too little concentration.

Empirical work shows that networks of research and development alliances are asymmetric, with a small number of firms involved in the majority of partnerships. This article investigates the welfarerelevant effects of such concentrated networks in a model of network formation in an oligopolistic market. We find that concentration is a typical characteristic of a socially efficient network, when the costs of collaborative activity are significant. Moreover, expanding on prior work relating to strategically stable inter-firm networks, we compare the stable and the efficient structures. Our findings suggest that the real-world networks might even exhibit too little concentration.

No abstract

Research and development (R&D) partnerships are formed to share the risks and benefits of R&D. At the macro level, they result in a globe-spanning network that can be a valuable source of international knowledge spillovers. This network is the subject of a considerable body of literature. Often-made claims are that R&D collaboration is an important activity in a competitive environment, but that the importance of international partnerships has declined over time. Furthermore, it is claimed that collaborations are disproportionally concentrated within the developed economies. However, this literature fails to account for variations in the sizes of underlying firm populations between countries and over time. We argue that these population sizes create an opportunity structure of available collaboration partners for firms, and that ignoring variations in this structure potentially leads to erroneous conclusions about the structure and dynamics of the R&D network. To address this problem, we study the structure and dynamics of the global R&D network on an international and cross-industry scale using longitudinal data for

No abstract