Digital and intelligent transformation makes Chinese listed firms more resilient by boosting both risk management and core competitiveness; financing shortages blunt its defensive benefits while amplifying offensive gains, particularly for private firms in competitive sectors.

Main Finding

Digital and intelligent transformation (DIT) significantly improves enterprise resilience (ER) among Chinese A‑share listed firms (2014–2023). This effect operates through two partially independent channels: (1) a defensive channel — DIT strengthens firms’ risk management capability, which improves stability and short-term recovery; and (2) an offensive channel — DIT builds core competitiveness, which supports post‑shock growth and strategic renewal. Financing constraints moderate these channels asymmetrically: they weaken the DIT → risk‑management pathway but strengthen the DIT → core‑competitiveness pathway. The overall effects are stronger in highly competitive industries and in non‑state‑owned enterprises.

Key Points

• Conceptual distinction: The paper distinguishes digital and intelligent transformation (DIT) — characterized by intelligent decision‑making and adaptive optimization (AI, big data, cloud) — from conventional digitalization that emphasizes process connectivity.
• H1 (baseline): DIT positively affects enterprise resilience (defensive + offensive dimensions).
• Dual mediators:
  • Risk management capability mediates the defensive effect of DIT on ER (shorter disruption impact, faster stabilization).
  • Core competitiveness mediates the offensive effect of DIT on ER (innovation, performance recovery and growth).
  • Both mediators produce partial mediation — DIT has direct and indirect effects on ER.
• Moderation by financing constraints:
  • Financing constraints reduce DIT’s positive effect on firms’ risk management capability (survival‑oriented resources are strained).
  • Financing constraints amplify DIT’s effect on core competitiveness (resource scarcity shifts firms toward development‑oriented, high‑value investments or reveals selection/efficiency effects).
  • Interpretation: coexistence of survival‑oriented and development‑oriented behavioral logics under resource scarcity.
• Heterogeneity: effects are larger in more competitive industries and in non‑state‑owned firms (private firms and competitive markets show stronger DIT→ER links).
• Robustness: authors use multiple empirical checks (fixed effects, mediation tests, IV estimation, robustness checks) to support findings.

Data & Methods

• Sample: Panel data of Chinese A‑share listed firms, 2014–2023.
• Measurement:
  • DIT: constructed via textual analysis of firm disclosures / corporate texts to capture digital and intelligent transformation activities (AI, big data, cloud, intelligent decision‑making emphasis).
  • Enterprise resilience (ER): operationalized as a multidimensional firm resilience measure (defensive: sustained operational capability; offensive: performance growth capability) following literature on multidimensional indices (authors construct an ER index accordingly).
  • Mediators: firm risk management capability and core competitiveness measured using firm variables (as described in the paper).
  • Moderator: financing constraints proxied by standard firm‑level financing constraint measures (authors test interactions).
• Estimation strategy:
  • Baseline: two‑way fixed effects panel regressions (firm and year fixed effects) to estimate DIT → ER.
  • Mediation: three‑step mediation tests to examine indirect effects through risk management capability and core competitiveness.
  • Endogeneity checks: instrumental variable (IV) estimation to address potential reverse causality / omitted variables.
  • Robustness: multiple robustness checks reported (alternative specifications and samples).
• Heterogeneity: subgroup analyses by industry competition intensity and ownership (SOE vs non‑SOE).

Implications for AI Economics

• Multi‑channel value of AI/digital investments: The paper shows AI and related intelligent technologies deliver both defensive (risk reduction, early warning) and offensive (competitiveness, innovation) value — models of AI investment returns should capture both channels rather than a single productivity effect.
• Financing frictions shape technological returns: Financing constraints alter the allocation of DIT benefits across risk management vs. capability building. Macro and micro models should incorporate financing frictions as moderators of technology adoption payoffs and endogenous reallocation between survival and growth strategies.
• Heterogeneous diffusion and welfare effects: Stronger DIT→ER effects in competitive industries and private firms imply uneven benefits across market structures; policy interventions (subsidies, fintech, targeted credit) could be needed to equalize resilience gains and prevent divergence.
• Policy design: To maximize social value from AI diffusion, policies should (a) support firms’ risk‑management capabilities (e.g., data‑sharing platforms, standards, cyber‑security incentives) to secure defensive resilience; (b) ease financing constraints for capability‑building investments (e.g., targeted R&D credit, venture/scale funding) to foster offensive resilience; and (c) tailor support by ownership and industry context.
• Research directions:
  • Incorporate moderated mediation structures in empirical models of AI adoption to capture conditional indirect effects.
  • Study thresholds/non‑linearities (costly overinvestment / inverted‑U risks mentioned in literature).
  • Model organizational frictions (resistance, skill gaps) and adjustment costs that may generate short‑term negative effects of rapid DIT.
  • Evaluate welfare and systemic implications (supply‑chain resilience, aggregate productivity resilience) when DIT adoption is heterogeneous across firms.

If you want, I can extract specific regression coefficients, mediation effect sizes, or the IV strategy from the full published manuscript once available (the provided version is an unedited pre‑publication copy).