Provinces that integrate digital-intelligence technologies more deeply show stronger agricultural resilience: a one‑unit rise in integration raises resilience by 0.0865, driven mainly by digital infrastructure and tech uptake; effects are largest where fiscal support is weak and financial regulation is tight.

Main Finding

Digital-intelligent integration significantly increases agricultural industry resilience (AIR) in China. Using provincial panel data (30 provinces, 2012–2022) the authors estimate that a one‑unit increase in their digital‑intelligence integration index raises AIR by 0.0865 units on average. Components matter: the paper finds large positive effects from digital infrastructure (coefficient = 0.6915) and from digital technology application (coefficient = 0.2038). The effect operates through at least three channels—promoting agricultural technological advancement, increasing industry diversification, and improving labor allocation efficiency—and is stronger in regions with lower fiscal agricultural support and tighter financial regulation.

Key Points

• Definition: "Digital‑intelligent integration" denotes the deep combination of digital technologies (IoT, cloud, big data, blockchain) with intelligent technologies (AI, machine learning, automation), emphasizing data-driven decisioning across the agricultural value chain.
• Main causal estimate: +1 unit in the digital‑intelligence index → +0.0865 in agricultural industry resilience (AIR).
• Component estimates:
  • Digital infrastructure → AIR (coef. = 0.6915)
  • Digital technology application → AIR (coef. = 0.2038) (Reported coefficients are from the paper; units reflect the authors' constructed indices.)
• Mechanisms identified (mediators):
  • Agricultural technological advancement (faster diffusion and adoption of new tech)
  • Industry diversification (broader crop/sector mix, deeper value‑chain integration)
  • Labor allocation efficiency (better matching, upskilling, cross‑regional labor flows)
• Heterogeneity: stronger digital‑intelligence effects where fiscal support for agriculture is relatively low and where financial regulation is relatively strict—suggesting digitalization may substitute for weak public support or interacts with institutional constraints.
• Contributions: extends literature from productivity/income outcomes to system‑level resilience; highlights institutional boundary conditions for digitalization benefits.

Data & Methods

• Data: Balanced panel of 30 Chinese provinces (excluding Hong Kong, Macau, Taiwan, Tibet), annual observations 2012–2022.
• Outcome: Agricultural Industry Resilience (AIR) — a constructed index intended to capture the sector’s ability to resist, recover, and adapt to shocks (paper constructs and discusses AIR; precise indicator components are defined in the full text).
• Key explanatory variable: Digital‑Intelligence Integration (DIL) — an index combining measures of digital infrastructure and digital technology application (details of construction in full paper).
• Empirical strategy:
  • Fixed‑effects panel regressions with province and year fixed effects to control for unobserved time‑invariant heterogeneity and common time shocks.
  • Controls: a set of standard covariates (economic and demographic controls; exact list in paper).
  • Mediation analysis: two‑step approach estimating DIL → mediator and mediator → AIR to identify channels (technological progress, industry diversity, labor allocation efficiency).
  • Heterogeneity tests: sub‑sample or interaction tests by fiscal support level for agriculture and by degree of financial regulation.
• Robustness: authors report robustness checks (full details in paper) to validate main estimates—e.g., alternative specifications and component decomposition.

Implications for AI Economics

• AI as resilience capital: This study treats AI and related intelligent technologies as part of a digital‑intelligence capital stock that raises sectoral resilience, not just short‑run productivity. That reframes AI investments as insurance‑like public and private goods (risk reduction, faster recovery).
• Complementarity with infrastructure and institutions: The much larger coefficient on digital infrastructure implies strong complementarities—AI value depends on physical/digital backbone. AI economics should therefore incorporate complementarity with infrastructure investments and institutional contexts (fiscal support, financial regulation).
• Mechanisms relevant to labor and adoption economics:
  • Labor market effects: AI-enabled platforms and automation change the matching process and skill requirements. Policy needs to consider upskilling and labor mobility.
  • Technology diffusion: AI accelerates knowledge diffusion and product/process innovation; modelling of diffusion dynamics should include platform-mediated effects.
• Policy design and welfare: Where fiscal support is weak, digital‑intelligence investments may play a quasi‑substitutive role—suggesting targeted public investments in digital infrastructure could be high‑return, especially in under‑supported regions. However, distributional impacts (who captures gains) and transition costs for displaced labor require economic evaluation.
• Research directions for AI economists:
  • Micro‑level causal identification of AI adoption effects on farm resilience and incomes (e.g., randomized rollouts of AI tools).
  • Measurement work: standardized indices for digital‑intelligence capital and resilience to enable cross‑country comparisons.
  • General equilibrium effects: how widespread AI adoption in agriculture affects input/output prices, rural labor markets, and food security.
  • Interaction with regulation: quantify how different financial and fiscal regimes modulate the returns to AI/digital investments.
  • Cost–benefit and distributional analyses that incorporate resilience (reduced downside risk) as an explicit welfare component.

If you want, I can extract the paper’s variable construction (exact components of the AIR and DIL indices), list the control variables used, or draft specific policy recommendations for governments or investors based on these findings.