China’s supply‑chain digitalization pilots strengthened listed firms’ green resilience by spurring green innovation, not by loosening finance or reshaping supplier concentration; gains were largest for firms with higher human capital, green-conscious managers and in tech- and pollution‑intensive sectors.

Main Finding

China's Supply Chain Innovation and Application Pilot Cities (SCD) policy causally improves corporate green resilience among A‑share listed firms (2010–2023). The effect is robust to multiple checks and operates primarily through enhancing firms' green innovation (both quantity and quality), rather than through easing credit constraints or changing supply‑chain concentration.

Key Points

• Policy effect: SCD pilot designation leads to a statistically significant increase in firms' measured green resilience.
• Mechanism: The primary channel is green innovation—firms increase both the number and the quality of green patents/innovations after SCD exposure.
• Non‑mechanisms: Credit availability and supply‑chain concentration were tested but do not significantly mediate the policy effect.
• Robustness: Results hold under propensity score matching, exclusion of COVID‑19 years, inclusion of additional controls, and accounting for other contemporaneous policies.
• Heterogeneity: Stronger SCD effects for firms with
  • higher executive green perception,
  • greater human capital levels,
  • location in first‑ and third‑tier cities,
  • technology‑ and capital‑intensive industries,
  • pollution‑intensive sectors,
  • firms in the growth stage.
• Policy implication drawn by authors: SCD‑oriented institutional reform reshapes firms’ ability to withstand green‑related disruptions and can be used to design targeted sustainability policies.

Data & Methods

• Sample: A‑share listed Chinese firms, panel data from 2010 to 2023.
• Identification strategy: Difference‑in‑differences (DID) exploiting the quasi‑natural experiment of cities selected as Supply Chain Innovation and Application Pilot Cities.
• Outcome: Corporate green resilience (measured by the authors — exact metric not detailed in the abstract).
• Mechanism analysis: Mediation tests examining green innovation (quantity and quality), credit availability, and supply‑chain concentration.
• Robustness checks: Propensity score matching (PSM‑DID), exclusion of COVID‑19 period observations, additional firm/city controls, and controls for other policies active over the sample period.

Implications for AI Economics

• Digitalization and AI as policy levers: The study suggests institutional promotion of supply‑chain digitalization boosts firms' environmental resilience primarily by accelerating green innovation. For AI economics, this indicates AI and related digital tools embedded in supply chains can be important complements to firm‑level innovation that reduces environmental risks.
• Modeling productivity and externalities: When modeling firm productivity or investment responses to digital adoption, include channels for green innovation and resilience to environmental shocks (not just cost or efficiency effects).
• Policy design: Policymakers aiming to foster sustainable transformation should couple digitalization incentives with innovation support (R&D subsidies, IP protection, skills training). AI policy can prioritize applications that directly enable green innovation (e.g., process optimization, design for low emissions).
• Heterogeneity matters: Targeted AI/digitalization programs may be more effective for firms with higher human capital or managerial green awareness, and in certain city tiers and industries. Evaluations of AI policy should therefore allow for heterogeneous treatment effects.
• Measurement and evaluation opportunities: Use AI methods (text analysis, patent classification, satellite imagery, supply‑chain data linkage) to better measure green innovation, resilience, and causal impacts in future work.
• Financial channels: The lack of mediation via credit availability implies digitalization‑driven green resilience may not be primarily financial. AI‑driven interventions should therefore emphasize technological diffusion and human capital rather than only easing finance.
• Future research directions: Explore the specific digital/AI technologies deployed in pilot cities, their adoption diffusion mechanisms, and long‑run productivity vs. environmental tradeoffs; build structural models that incorporate digital adoption, innovation choices, and environmental externalities to inform optimal policy.