R&D network formation with myopic and farsighted firms

We study the formation of R&D networks when each Örm beneÖts from the research done by other Örms it is connected to. Firms can be either myopic or farsighted when deciding about the links they want to form. We propose the notion of myopic-farsighted stable set to determine the R&D networks that emerge in the long run. When the majority of Örms is myopic, stability leads to R&D networks consisting of either two asymmetric components with the largest component comprises three-quarters of Örms or two symmetric components of nearly equal size with the largest component having only myopic Örms. But, once the majority of Örms becomes farsighted, only R&D networks with two asymmetric components remain stable. Firms in the largest component obtain greater proÖts, with farsighted Örms having in average more collaborations than myopic Örms that are either loose-ends or central for spreading the innovation within the component. Besides myopic and farsighted Örms, we introduce yes-Örms that always accept the formation of any link and never delete a link subject to the constraint of non-negative proÖts. We show that yes-Örms can stabilize R&D networks consisting of a single component that maximize the social welfare. Finally, we look at the evolution of R&D networks and we Önd that R&D networks with two symmetric components will be rapidly dismantled, single component R&D networks will persist many periods, while R&D networks consisting of two asymmetric components will persist forever.