The Evolution of Corporate Innovation Networks and Its Impact on Firm Valuation: An Empirical Study Based on Dynamic Graph Neural Networks

Abstract

In an increasingly interconnected and knowledge-driven economy, corporate innovation networks have become critical conduits for competitive advantage. However, prior research has predominantly relied on static or comparatively static analyses, failing to capture the dynamic nature of these inter-firm relationships. This study addresses this gap by investigating how the temporal evolution of a firm's position within its innovation network influences its market valuation. We construct a dynamic network of strategic alliances among U.S. publicly traded firms from 1995 to 2020, using data from the SDC Platinum database, and link it to financial data from Compustat and CRSP. To model the complex, path-dependent nature of network evolution, we employ a Dynamic Graph Neural Network (DGNN), specifically the EvolveGCN architecture. Our empirical results demonstrate that a firm's network trajectory contains significant predictive power for its future valuation, over and above traditional financial controls and static network metrics. Specifically, trajectories characterized by increasing centrality and brokerage capabilities are positively associated with higher firm valuation, as measured by Tobin's Q. These findings contribute to the Knowledge-Based View and network theory by highlighting the strategic importance of dynamic network management capabilities. Methodologically, this study showcases the utility of DGNNs for addressing complex, time-varying relational questions in strategic management and finance.