Collaboration in Research and Innovation Networks
This thesis, consisting of an introduction and four papers, aims to increase the understanding of collaborations in research and innovation networks. It does so by examining several factors that can influence the collaborations observed.
Collaboration is important in research, development and innovation processes, because it allows for transfer of knowledge not available in the market (Ahuja, 2000b), and because it is hard for a single actor to have all the necessary knowledge in a complex innovation processes in- house (Breschi et al., 2009). Collaboration with universities, suppliers or customers, can make firms more innovative (Laursen & Salter, 2014). When collaboration is successful, it “ is greater than the sum of its parts” (Katz & Martin, 1997, p. 15). In line with this, the number of collaborations have increased in industry, between researchers and across different disciplines over the last decades. Policy makers and research councils also support and try to strengthen collaborations between different types of actors, with for example research funding schemes that require collaboration.
Collaboration also involves transactions costs. To reduce transaction cost, actors often choose to collaborate with actors that are similar or proximate to themselves (Boschma, 2005; Powell et al., 2005). Actors also mutually influence each other’s behavior and decisions (Borgatti et al., 2009). This makes it useful to study collaborations from a social network perspective, which views networks as the context of action (Burt, 2004).
In my thesis, I ask three overall questions: (1) Who in the network do actors collaborate with? (2) What makes actors collaborate? (3) How does public policy for collaboration influence networks? My thesis answers these questions by studying collaborations in one innovation network and two research networks. Using interview data I am able to study collaboration in the innovation network (Papers I and II) more in detail. Using register data, I am able to capture larger collaboration networks for the two research networks (Papers III and IV).
In all my papers, I use social network analysis. In Papers II, III, and IV, I use additional econometric models. Papers I and II study collaboration in the subsea industry in Rogaland, Norway. Paper III examines the participation of organizations in Norwegian regions in the EU FP’s green restructuring programmes. Paper IV studies the network of all collaborations in Research Council of Norway funded projects.
I focus on six dimensions that influence the collaboration in the networks observed: (1) The knowledge base of the organization and cognitive proximity between organizations. (2) Where the organization is located and geographical proximity between organizations. (3) Actors being proximate on other dimensions. (4) Actors sharing information and knowledge through other channels. (5) Influence from central actors. (6) Time.
Overall, the findings from my thesis shows that, organizations are more likely to collaborate with other organizations that are proximate to them. Organizations are more likely to collaborate with actors that are cognitively proximate, i.e. in oil related industries in paper I. The effect of cognitive proximity is also positive and with an increasing positive effect over the years in Paper IV.
In the subsea industry, in Papers I and II, firms collaborates to a large extent with geographically proximate actors. In Paper IV I find that the effect of geographical proximity on the likelihood of collaboration is positive and increasing over time. In Paper III, we find regional differences when it comes to who participates in the EU FP’s green restructuring programmes.
Paper IV finds that social proximity measured as previous collaboration has a strong positive effect on the likelihood of collaboration. Paper II examines the association between three different networks among actors: collaboration, recruitment and monitoring. We find that these three networks are highly associated, suggesting that an actor that recruits or monitors another actor is also more likely to collaborate with this actor.
In all the observed networks, some actors collaborate with many other actors, while others only collaborate with a few. In Paper IV, I find that being a central actor early in the network is highly correlated with being a central actor later on. In Paper III, we find that in a region of important industry actors, these industry actors are most central in the region. In region with strong research establishments, the research establishments are the central actors.
In networks funded by public funding agencies, such as through the Research Council of Norway (RCN) with a clear aim of stimulating collaboration between broad groups of actors, I find that geographical, cognitive and social proximities are important factors for the likelihood of collaboration. Geographical and cognitive proximity even have an increasing positive effect over time. In addition to this, the finding in Paper III shows that in regions with strong research establishments firms have low and decreasing participation in programmes focusing on green restructuring. This is a field where firms’ participation is important to do necessary restructuring. The strong influence from proximity dimensions suggests that policy makers must be aware of the risk of path dependency and potential lock-in. My papers also show that in all the networks observed, there is extensive collaboration. The publicly funded networks include a variety of different actors, and the RCN networks increase every year in the number