The Impact of Digital Innovation Network Embedding on Firm Innovation Resilience: Perspective on Knowledge-based Dynamic Capability

Technical Summary: The Impact of Digital Innovation Network Embedding on Firm Innovation Resilience

Problem Statement
In the volatile, uncertain, complex, and ambiguous (VUCA) era, characterized by global trade frictions, geopolitical conflicts, and technological disruptions, the security and stability of industrial and supply chains face severe shocks. While some firms possess "innovation resilience"—the capacity to maintain stability, recover rapidly, and upgrade innovation amidst uncertainty—others face stagnation or exit. Existing literature predominantly focuses on internal resources or traditional innovation networks as drivers of resilience. However, these studies often overlook the transformative role of digital technologies in reshaping open innovation organizations. Furthermore, prior research on digital innovation networks has largely examined static performance outcomes, failing to clarify the dynamic mechanism through which network embedding converts fragmented digital knowledge resources into sustained innovation resilience. This study addresses the gap in understanding how embedding in a digital innovation network influences firm innovation resilience and the specific role of knowledge-based dynamic capabilities in this process.

Methodology
The study employs a quantitative empirical approach using panel data from Chinese A-share listed firms spanning 2012 to 2023.

• Data Construction: The authors constructed a "Digital Innovation Network" using cooperative digital invention patent application data sourced from the National Intellectual Property Administration and the CNRDS database. Patents were screened based on the "Classification of Core Digital Economy Industries and International Patent Classification (2023)."
• Variables:
  • Dependent Variable: Innovation Resilience (IR), measured using a sensitivity indicator based on the deviation of a firm's patent application changes from industry trends.
  • Independent Variables: Digital Innovation Network Structural Embedding (measured by degree centrality) and Relational Embedding (measured by the geometric mean of joint patent application frequency).
  • Mediating Variables: Knowledge-based Dynamic Capability, operationalized through three dimensions: Knowledge Acquisition (backward citations), Knowledge Integration (IPC code combinations), and Knowledge Reconfiguration (new IPC codes in a rolling five-year window).
  • Controls: Firm size, age, ownership nature, financial ratios, and fixed effects for year, industry, and region.
• Analytical Strategy: The study utilizes fixed-effects regression models to test direct effects. It employs a series of robustness checks, including replacing core variables, excluding non-digital networks, and lagging explanatory variables. To address endogeneity concerns (reverse causality, omitted variables, selection bias), the authors applied the Instrumental Variable (IV) method (using Lewbel and Nunn approaches), Propensity Score Matching with Difference-in-Differences (PSM-DID), and the Heckman two-stage method. Heterogeneity analyses were conducted based on executive digital background, industry type (high-tech vs. non-high-tech), and perceived environmental uncertainty.

Key Contributions
The paper makes three primary theoretical and empirical contributions:

1. Framework Integration: It integrates digital innovation network embedding with firm innovation resilience into a unified framework, dissecting how structural and relational embedding modes differentially influence resilience.
2. Mechanism Unveiling: Grounded in dynamic capability theory, the study constructs a "network embedding – knowledge-based dynamic capability – innovation resilience" framework. It elucidates the "black box" of how static, fragmented digital knowledge resources are transformed into dynamic innovation resilience through the mediating roles of knowledge acquisition, integration, and reconfiguration.
3. Contextual Nuance: It provides empirical evidence on the boundary conditions of these relationships, identifying that the resilience-enhancing effects vary significantly based on executive digital knowledge, industry sector, and the firm's perception of uncertainty.

Key Results

• Direct Effects: Both structural embedding (position in the network) and relational embedding (depth of ties) in digital innovation networks have a significant positive effect on firm innovation resilience.
• Mediating Mechanism: Knowledge-based dynamic capability serves as a critical pathway. Structural and relational embedding enhance a firm's ability to acquire, integrate, and reconfigure knowledge, which in turn strengthens innovation resilience.
• Robustness: The findings remain consistent across various robustness checks and endogeneity treatments (IV, PSM-DID, Heckman), confirming the causal relationship.
• Heterogeneity:
  • Executive Background: The positive impact is more pronounced in firms where senior executives possess a digital background.
  • Industry Type: Non-high-tech firms experience a greater enhancement in innovation resilience compared to high-tech firms, suggesting digital networks help bridge the digital-non-digital divide for traditional industries.
  • Uncertainty Perception: Firms with a high perception of external environmental uncertainty derive stronger resilience benefits from network embedding, likely due to their proactive risk-mitigation strategies.

Significance and Claims
The authors claim that this study provides empirical evidence and practical insights for fostering innovation resilience in the context of digitization and openness.

• Theoretical Significance: The research extends the applicability of dynamic capability theory to the digital innovation network context, demonstrating how network resources are converted into firm-level resilience capabilities.
• Practical Significance:
  • For Policymakers: The findings suggest a need for governments to lead in constructing digital innovation infrastructure (e.g., open-source platforms) and establishing regulations for data circulation to lower entry barriers for firms.
  • For Firms: Enterprises are advised to actively cultivate diverse digital partnerships and maintain core collaborative relationships to secure high-value resources. Specifically, non-high-tech firms and those with high uncertainty perception should leverage these networks to overcome traditional innovation constraints.
  • Strategic Implication: The study highlights the importance of developing knowledge-based dynamic capabilities to effectively transform digital network resources into sustainable innovation outcomes, thereby enhancing the security and controllability of industrial and supply chains.

The paper concludes by acknowledging limitations, such as the indirect measurement of innovation resilience and the focus on specific network embedding dimensions, suggesting future research could explore process-based measurements and other network characteristics like structural holes and cognitive embedding.