China’s green public procurement is linked to measurable falls in listed firms’ carbon intensity, and the effect is amplified when firms adopt AI or obtain green subsidies; state‑owned and high‑tech firms in digitally advanced regions benefit most.
The Impact of Government Green Procurement on Corporate Carbon Emission Reduction: A Dual Mediation Perspective of Artificial Intelligence and Green Finance
openalex
quasi_experimental
medium evidence
7/10 relevance
Summary only
summary available; pdf_status=paywall
DOI
Source
PDF
Green public procurement is associated with significant improvements in corporate carbon reduction among Chinese A‑share firms, and this effect is stronger for firms that adopt AI and receive government green subsidies, particularly for state‑owned, high‑tech firms in digitally advanced regions.
This study uses data of A-share listed companies in Shanghai and Shenzhen from 2020 to 2024. We manually collect green procurement lists from official government procurement websites and match them with firm samples. Employing the two-way fixed effects model and the Bootstrap method, this paper empirically examines the impact of green public procurement on corporate carbon reduction. The results show that green public procurement significantly improves firms’ carbon reduction performance. Mechanism analysis indicates that AI adoption and government green subsidies further strengthen this effect. Heterogeneity tests reveal that the impact is more pronounced for state-owned enterprises, high-tech firms and enterprises in regions with advanced digital economies. Accordingly, we propose suggestions including strengthening the driving role of green procurement, promoting coordination between green procurement and digital technology, optimising the allocation of green funds, and implementing targeted differentiated incentives. This research helps clarify the internal mechanism of green public procurement on carbon emission reduction performance and provides references for improving relevant practices in carbon emission reduction.
Summary
Main Finding
Green public procurement (GPP) significantly improves corporate carbon reduction performance for A-share listed firms in Shanghai and Shenzhen (2020–2024). The effect is amplified when firms adopt AI and when firms receive government green subsidies. The impact is stronger for state-owned enterprises, high‑tech firms, and firms located in regions with more advanced digital economies.
Key Points
- Data: Firm-level A-share sample (Shanghai & Shenzhen), 2020–2024; green procurement lists manually collected from official government procurement websites and matched to firms.
- Primary result: Firms matched to green procurement exhibit better carbon reduction performance (statistically significant).
- Mechanisms: AI adoption by firms and receipt of government green subsidies both strengthen the carbon-reduction effect of GPP (evidence of complementarities).
- Heterogeneity: Larger treatment effects for state-owned enterprises (SOEs), high‑tech firms, and firms in regions with advanced digital economies.
- Policy recommendations in the paper: strengthen the driving role of GPP, coordinate GPP with digital technologies, optimize allocation of green funds, and implement targeted/differentiated incentives.
Data & Methods
- Sample frame: A-share listed companies in Shanghai and Shenzhen, 2020–2024.
- Treatment measurement: Manual collection of government green procurement lists from official procurement websites and matching those procurements to firm samples.
- Empirical strategy: Two-way fixed effects (firm and time fixed effects) to estimate the impact of GPP on firm carbon reduction performance.
- Inference/robustness: Bootstrap method used to support statistical inference (robustness to sampling variation).
- Mechanism tests: Interaction/mediation-style analyses examining AI adoption and government green subsidies as channels that amplify the GPP effect.
- Heterogeneity analysis: Subsample tests by ownership (SOE vs non-SOE), firm technology status (high-tech vs others), and regional digital-economy development.
Implications for AI Economics
- Complementarity between AI and green policies: The finding that AI adoption strengthens GPP’s carbon-reduction impact highlights important complementarity between digital/AI capabilities and environmental policy. Models of policy effectiveness should incorporate technology adoption as a moderating factor.
- Diffusion and returns to AI investment: Results suggest additional private returns to AI adoption via enhanced ability to capture policy-driven demand (e.g., green procurement), implying heterogeneous incentives for AI uptake across firms and regions.
- Targeting and policy design: Policymakers can increase GPP effectiveness by pairing procurement with incentives or support for AI/digital adoption—especially in regions or firms lagging digitally—rather than treating procurement as a standalone instrument.
- Heterogeneity matters: SOEs, high‑tech firms, and digitally advanced regions respond more to GPP. Economic models and empirical work should allow for varying treatment effects by firm type and regional digital capacity.
- Empirical strategies and data: The study shows the value of combining administrative procurement data with firm-level outcomes. For future AI-economics research, collecting and matching procurement and technology-adoption data can reveal policy–technology interactions.
- Future research directions: investigate causal identification further (e.g., instrumentation or quasi-experiments for procurement exposure and AI adoption), quantify welfare/trade-offs (costs of AI adoption vs. emissions gains), and explore dynamic, long-run impacts of GPP on firm technology trajectories and market structure.
Assessment
Paper Typequasi_experimental
Evidence Strengthmedium — The use of firm and year fixed effects and bootstrap robustness checks on a multi-year panel strengthens causal inference relative to purely cross‑sectional correlations, and mechanism and heterogeneity tests provide supportive secondary evidence; however, the study appears to lack an exogenous source of variation (e.g., randomized assignment, policy discontinuity, or valid instrument) so residual time‑varying confounding and selection into green procurement remain plausible threats to causality.
Methods Rigormedium — Methodologically sound choices (manual matching of procurement lists, panel two‑way FE, bootstrap inference, mechanism and heterogeneity analyses) indicate careful empirical work, but the design does not demonstrate a convincing quasi‑experimental source of identification, and key robustness checks (pre‑trends, alternative identification like IV or natural experiment, clustering choices, measurement validation of carbon outcomes and AI adoption) are not reported in the summary.
SampleFirm‑level panel of A‑share listed companies in Shanghai and Shenzhen over 2020–2024, merged with manually collected green public procurement contract/lists from official government procurement websites; analysis uses firm‑year observations and includes indicators for firm AI adoption and receipt of government green subsidies; exact sample size and construction details not provided in the summary.
Themesgovernance adoption
IdentificationPanel two-way fixed effects on a firm-year panel of A‑share listed firms (2020–2024) with firms matched to manually collected green public procurement contract lists; robustness via bootstrap; mechanism tests using measures/indicators of firm AI adoption and receipt of government green subsidies and heterogeneity analyses across ownership, tech intensity, and regional digital economy levels. No clear exogenous shock, instrumental variable, or explicit difference‑in‑differences staggered policy design is described.
GeneralizabilityLimited to publicly listed A‑share firms in Shanghai and Shenzhen — may not generalize to private firms or SMEs., China‑specific public procurement and regulatory context — external validity to other countries is uncertain., Short sample period (2020–2024) that overlaps COVID‑19 and rapid digitalization trends, which may limit longer‑run inference., Manual procurement data may miss informal/uncatalogued contracts or have matching errors, affecting representativeness., AI adoption and carbon reduction measures are likely context‑dependent (measurement and reporting differences) and may not be comparable across settings.
Claims (8)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| This study uses data of A-share listed companies in Shanghai and Shenzhen from 2020 to 2024; we manually collect green procurement lists from official government procurement websites and match them with firm samples. Other | null_result | sample scope / data provenance (coverage of firms and years, data collection method) |
Reading fidelity
high
Study strength
high
|
not reported
|
| The paper employs the two-way fixed effects model and the Bootstrap method to empirically examine the impact of green public procurement on corporate carbon reduction. Other | null_result | methodological approach (statistical models used) |
Reading fidelity
high
Study strength
high
|
not reported
|
| Green public procurement significantly improves firms’ carbon reduction performance. Firm Productivity | positive | corporate carbon reduction performance (carbon emissions / carbon reduction) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| AI adoption further strengthens the positive effect of green public procurement on corporate carbon reduction. Firm Productivity | positive | corporate carbon reduction performance (moderated by AI adoption) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| Government green subsidies further strengthen the positive effect of green public procurement on corporate carbon reduction. Firm Productivity | positive | corporate carbon reduction performance (moderated by government green subsidies) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The impact of green public procurement on corporate carbon reduction is more pronounced for state-owned enterprises. Firm Productivity | positive | corporate carbon reduction performance (heterogeneous effect by ownership: state-owned vs. others) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The impact of green public procurement on corporate carbon reduction is more pronounced for high-tech firms. Firm Productivity | positive | corporate carbon reduction performance (heterogeneous effect by firm technology status) |
Reading fidelity
high
Study strength
medium
|
not reported
|
| The impact of green public procurement on corporate carbon reduction is more pronounced for enterprises in regions with advanced digital economies. Firm Productivity | positive | corporate carbon reduction performance (heterogeneous effect by regional digital economy level) |
Reading fidelity
high
Study strength
medium
|
not reported
|