Network Positions and Knowledge Coordination in Multi-Actor Innovation Systems: Evidence from EIP-AGRI Operational Groups

Collaborative innovation initiatives increasingly rely on multi-actor networks to mobilise and coordinate distributed knowledge. Yet, empirical evidence on how inter-organisational network structures shape innovation outcomes remains limited, particularly in policy-driven contexts. This study examines how partner network configurations influence perceived project outcomes within the European Innovation Partnership for Agricultural Productivity and Sustainability (EIP-AGRI) in Italy. Adopting a relational perspective of the knowledge economy, partner networks are conceptualised as knowledge coordination structures that shape learning processes and innovation effectiveness. Social network analysis is applied to a nationwide dataset of partners from EIP-AGRI Operational Groups to examine network embeddedness, brokerage positions, and the emergence of latent interregional networks. Their association with perceived project outcomes is assessed using a beta regression model with random effects. The findings indicate that embeddedness in the core of the partner network, captured by eigenvector centrality, is positively associated with this outcome measure, whereas brokerage positions characterised by high betweenness centrality are negatively associated with it, pointing to coordination costs in intermediary roles. Membership in latent interregional networks is also negatively associated with perceived project outcomes, suggesting additional coordination challenges in geographically dispersed and weakly institutionalised collaboration structures. Overall, the study shows that the position actors occupy in collaborative networks matters not only for accessing knowledge, but also for managing the coordination burdens that accompany it. These findings have practical implications for the design and governance of EIP-AGRI and similar multi-actor innovation policies.