Spatio-temporal shifts, driving mechanisms, and resilience dynamics: Unraveling the evolution of global innovation networks (2003–2023)

This study examines the spatio-temporal shifts, driving mechanisms, and structural resilience of Global Innovation Networks (GINs) by leveraging transnational patent data from the World Intellectual Property Organization (2003–2023). Through Social Network Analysis (SNA), Temporal Exponential Random Graph Models (TERGM), and Network Resilience Assessment Modeling (NRAM), we deliver a dynamic and multi-level analysis of GINs. Findings indicate that GINs maintain small-world properties and a stable core-periphery architecture, while experiencing a marked eastward shift in influential nodes. The traditional Western-centered core has expanded to incorporate emerging economies such as China, India, and South Korea, signaling a decentralization of global innovation activity. TERGM results reveal multi-level drivers: endogenous structures such as reciprocity and triadic closure guide self-organization; actor attributes exhibit asymmetric effects, where patent protection strength, political stability, and market size attract innovation inflows, whereas economic scale and trade promote outflows; exogenous proximities show cultural similarity fosters connections, while geographic and administrative distances act as barriers. Notably, knowledge distance's constraining role weakens when accounting for structural embeddedness. NRAM assessments show that GIN resilience has strengthened over time, with improved tolerance to both targeted and random disruptions. Yet systemic vulnerability persists through a limited set of core nations (including the US, China, and Germany)—whose failure may trigger broad instability. By incorporating endogenous dynamics, seldom-studied exogenous factors, and resilience into a unified framework, this research advances GIN theory and offers strategic insights for governance and global patent planning amid systemic uncertainties.