Efficient logistics boost China's cross-border e-commerce, but only where AI is strong — provinces with higher AI development capture much larger gains from logistics efficiency, especially in the prosperous eastern region.

Main Finding

Higher logistics fulfilment efficiency (LEF) is positively associated with provincial cross‑border e‑commerce development (CBED), and this effect is amplified by stronger AI capability (AIDI). The LEF–AIDI complementarity is strongest in eastern provinces and remains positive but smaller in central and western regions.

Key Points

• LEF alone contributes positively and significantly to CBED at the province level.
• AI capability (measured by a four‑pillar AI Development Index, AIDI) strengthens the marginal effect of LEF on CBED through a positive interaction term.
• The complementarity between LEF and AIDI exhibits regional heterogeneity: largest in the more digitally mature eastern provinces, attenuated but still positive in central and western provinces.
• Results are robust to dynamic panel specification and endogeneity control via two‑step System GMM.
• Policy relevance: improving AI capability increases the payoff to investments in logistics efficiency, suggesting integrated digital‑logistics strategies raise provincial competitiveness in CBEC.
• Sustainability considerations are discussed in terms of operational efficiency gains (e.g., reduced waste, better asset utilization) rather than direct measurements of environmental outcomes.

Data & Methods

• Sample: Balanced panel of 31 Chinese provinces, 2017–2023 (N = 217 observations).
• Dependent variable: Customs‑based indicators of cross‑border e‑commerce development (CBED).
• Main independent variables:
  • LEF: Logistics fulfilment efficiency measured using a Super‑SBM Data Envelopment Analysis (DEA) approach.
  • AIDI: Four‑pillar AI Development Index capturing multi‑dimensional provincial AI capability.
• Empirical strategy:
  • Two‑step System Generalized Method of Moments (GMM) to address dynamic persistence and potential endogeneity.
  • Interaction term LEF × AIDI to test for complementarity.
  • Estimation run for the full sample and separately for eastern, central, and western regional subsamples to assess heterogeneity.
• Robustness: Dynamic panel methods and regional subsamples used to verify stability of findings (details on additional robustness checks not provided in the summary).

Implications for AI Economics

• Theoretical: Provides empirical evidence that digital technologies (AI capability) and logistics efficiency are complementary inputs for expanding international digital trade; supports models in which digital maturity raises the returns to physical/logistics investments.
• Policy:
  • Provinces should coordinate investments in AI capability with logistics upgrades to maximize CBEC gains, especially in less digitally mature regions where AI raises but does not eliminate the productivity gap.
  • Eastern provinces can expect higher marginal returns from combined AI and logistics investment; central and western provinces may need targeted programs (capacity building, incentives) to close the complementarity gap.
• Research directions:
  • Micro‑level studies to trace firm‑level mechanisms (how AI augments logistics operations and firm export behavior).
  • Causal inference via quasi‑experimental or program evaluation approaches to validate mechanisms.
  • Direct measurement of environmental impacts (e.g., emissions, energy use) to substantiate sustainability claims tied to operational efficiency gains.
• Cautions and limitations:
  • Provincial aggregates may mask within‑province heterogeneity and sectoral differences.
  • AIDI and LEF are index/efficiency constructs whose measurement choices can influence results; documented coefficients/magnitudes are necessary for assessing economic significance in policy design.