The Effect of Connectivity, Proximity and Market Structure on R&D Networks

In a seminal paper, Goyal and Moraga-Gonzalez (2001) use an undirected network to characterize knowledge flows between firms engaging in research in an oligopolistic market. In their paper, firms are regarded as inhabiting a research joint venture (RJV), if they share the same edge of the network. These firms are allowed an R&D spillover of 1; the outside firms (firms not sharing an edge in the network) are permitted a constant knowledge spillover that is less than one. We begin our paper by showing that this last assumption has important consequences when dealing with R&D networks of size greater than or equal to six firms. We present examples of topologically non-equivalent networks that have the same degree of connectivity and generate identical outcomes in terms of R&D effort, firm profits and total welfare. We then modify their model so that R&D spillovers decrease as the number of shortest paths increases between any two firms. We show that under product differentiated Cournot and Bertrand competition, we have different outcomes for all economic variables. We also show that R&D effort increases with respect to the number of collaborative links if firms are in a weakly competitive market, whereas it declines if firms are in a more competitive market where products are closer substitutes. We also find that in more competitive markets there is a conflict between the stability and the efficiency of RJVs.