China’s national big-data pilot zones boost cities’ carbon‑efficiency by steering industrial upgrading, cleaner energy mixes and AI-enabled digital innovation; cross‑regional pilots and non–industry tax hubs see the largest gains, which spill over to neighbouring cities.

Main Finding

The establishment of National Big Data Comprehensive Pilot Zones (BDCPZs) significantly increases urban carbon unlocking efficiency. This improvement operates through four main mechanisms—industrial structure rationalization, energy structure optimization, enhanced digital technological innovation, and greater adoption of artificial intelligence (AI). Effects are stronger for cross-regional pilot zones and cities not dominated by industrial tax revenues, and BDCPZs generate positive spatial spillovers to neighboring cities.

Key Points

• Policy impact: BDCPZ designation causally raises carbon unlocking efficiency in prefecture-level Chinese cities.
• Mechanisms (four pathways):
• Industrial structure rationalization — BDCPZs promote shifts toward higher-value, less carbon-intensive activities.
• Energy structure optimization — local energy mixes become cleaner or more efficient under BDCPZ influence.
• Digital technological innovation — big data policies stimulate development and diffusion of digital tech that supports decarbonization.
• AI application — AI deployment enabled/accelerated by BDCPZs contributes directly to efficiency gains (e.g., process optimization, demand forecasting).
• Heterogeneity: Larger positive effects in cross-regional pilots versus single-region pilots; stronger in cities without heavy reliance on industrial tax bases.
• Spatial effects: BDCPZs produce significant positive spillovers on carbon unlocking efficiency in surrounding municipalities.
• Policy recommendation: Strengthen systematic, coordinated planning of BDCPZs to maximize carbon-unlocking benefits.

Data & Methods

• Data: Panel dataset of 281 prefecture-level cities from 2007 to 2023.
• Empirical strategy: Staggered difference-in-differences (DID) model exploiting phased BDCPZ implementation across cities to estimate the causal effect on carbon unlocking efficiency.
• Tests and extensions (reported): Mechanism analyses (mediation-style tests) for industry, energy, digital innovation, and AI channels; heterogeneity analyses across pilot types and fiscal/industrial structures; spatial econometric checks to identify spillovers to neighboring cities.
• Outcome: City-level measure of carbon unlocking efficiency (measured and analyzed over time; exact metric as used in the study).

Implications for AI Economics

• AI as a decarbonization lever: The study provides causal evidence that policies fostering big-data ecosystems accelerate AI adoption, which in turn improves carbon productivity—strengthening the case for viewing AI investment as part of green policy mixes.
• Market and investment signals: Positive local and spatial returns from BDCPZs imply potential aggregate social returns to public support for digital/AI infrastructure; investors and planners should factor spatial spillovers into cost–benefit assessments.
• Policy design: Cross-regional coordination of data and AI policy can magnify carbon-unlocking gains—standalone local pilots may be less effective. Fiscal and industrial characteristics matter for realized benefits, suggesting targeted support where industrial lock-in is weaker.
• Research directions: Quantify the relative contribution of AI versus other digital technologies to decarbonization; micro-level studies of firm behavior under BDCPZs; dynamic modeling of diffusion and spillovers to inform optimal placement and coordination of big-data/AI zones.