Abstract
The first study presents a novel analysis of horizontal R&D networks in the presence of an industry leader that is not only the dominant firm in the market but also the large R&D investor. In particular, it considers a model of an industry leader and n-1 followers. We analyse cost-reducing R&D investment by firms, which can be fully shared through collaborative links with other firms, while spillovers are not perfect when firms are not connected. The existing literature ignores the effects of market leaders or dominant firms on R&D networks. We contribute to the literature by considering a setting in which there is a dominant firm. We find that the follower firms always have incentives to collaborate with the leader firm because they can enjoy large R&D efforts of the leader firm through voluntary R&D knowledge sharing. In addition, the network, in which the industry leader has a large number of links, maximizes industry-profit for almost all values of the parameter. This result suggests that if transfers between firms are allowed, this network architecture is likely to survive potential deviations by firms because the more beneficial firms have the incentive to ‘sponsor’ other firms in order to maintain the network structure. Our result seems to be consistent with the empirical evidence in that we observe firms that invest largely in R&D are the central nodes of R&D networks. However, our result also reveal that the network, in which the leader firm is well-connected, performs badly from the social welfare viewpoint. The last finding shows that private incentives to form R&D collaborations are excessive from a social welfare viewpoint. The second study explores vertical R&D collaborations between firms in a two-tier industry, with n upstream firms and n downstream firms, and exclusive dealings among upstream and downstream firms. In this context, we analyse cost-reducing R&D investment by a firm in a tier, which can be shared through collaborative links with other firms in the other tier while there are no R&D spillovers when firms are not connected. We do not consider R&D collaborations between firms in the same tier. We contribute to the literature by considering the stability and efficiency of vertical R&D networks which have not been studied in the literature on R&D networks. Our results show that the difference between voluntary R&D spillovers of the two directions, from the upstream firm to the downstream firm and vice versa, plays an important role in the stability of vertical R&D collaboration networks. We find that, under certain circumstances, the network in which all vertical R&D collaborations are formed is stable and efficient. Otherwise, we can observe a sequence of networks due to continuously profitable deviations by firms. In addition, our results suggest that the private incentives to establish R&D collaborations are not always adequate from a social welfare point of view. This potential divergence implies that there is room for public policy. For example, R&D subsidy should be considered in order to enhance the collaborative activities when the private incentives to form vertical R&D collaborations are not adequate from a social welfare viewpoint. The third study investigates a model of strategic R&D collaboration networks in the open economy framework with three countries. In each country, there is a firm that can sell in the domestic and foreign markets. Government of each country can initiate bilateral free-trade agreements (FTAs) to abolish the import tariffs of partners, or impose import tariffs on countries with whom it has no FTA, in order to maximizes its national welfare. The set of FTAs between countries creates the FTA network. Firms decide whether and with whom to form R&D collaborations and these R&D collaborations establish the R&D network. This setting builds a double-layer network structure where the FTA network is formed in the first layer and the R&D network is formed in the second layer. In this context, firms invest in cost-reducing R&D, which can be fully shared through collaborative links with other firms, while spillovers between non-collaborating firms are not perfect. Our first finding is that FTAs are beneficial to international R&D collaborations. This result seems to be consistent with the stylized facts that FTAs have grown fast since the early 1990s and international R&D collaborations account for about 60% of newly established R&D partnerships at that time. Our second result shows that the complete FTA networks, in which each country has FTAs with all others, is always stable in the first network layer regardless of the architectures of R&D networks among firms in these countries. In addition, all global social welfare maximizing double-layer networks must contain the complete FTA network as the first layer. Finally, our analysis reveals that private incentives of firms to form international R&D collaborations are excessive from a global welfare point of view.