How does the Belt and Road policy affect the level of green development? A quasi-natural experimental study considering the CO2 emission intensity of construction enterprises

The study examines whether and how China’s Belt and Road (B&R) policy influences the green development (GD) level of construction enterprises, a sector known for high energy use and significant CO2 emissions. In 2015 the Chinese government formally launched the B&R vision and later issued guidance to promote a green B&R, integrating GD into the initiative’s implementation. The paper addresses three research questions using panel data for 28 provincial administrative regions (2010–2020): (1) How does the B&R policy plan affect the GD of construction enterprises in provincial administrative areas? (2) Does the degree of regional development affect the GD level of construction enterprises in key areas? (3) Does research investment by government and enterprises affect the GD level of construction enterprises in key areas? The study’s contributions include constructing a fixed-effects model combined with DID and PSM-DID to assess policy impacts on construction-sector GD, introducing multiple control variables from an econometric perspective, broadening the research perspective for GD in construction enterprises, and providing policy evidence for advancing a green B&R.

Theoretical background focuses on two domains: (1) Green development (GD) of construction enterprises and (2) Assessment of the B&R policy. Prior GD research covers definitions, influencing factors (e.g., government behavior, industry scale, natural resources, urban construction development, R&D investment), drivers (business models, green production, public environmental needs, social development), and evaluation methods (indicator systems, entropy models, comprehensive weighting). While construction industry GD has gained attention due to its pollution footprint, most studies examine determinants rather than specific policy evaluations. B&R policy assessments span trade, overseas investment, and ecological/environmental impacts using various econometric approaches. Since 2017, policy emphasis has included green and sustainable development. However, few studies assess B&R’s impact on the GD of construction enterprises. This study fills that gap by using a PSM-DID fixed-effects framework to evaluate the B&R policy’s effect on construction-sector GD.

Design: Quasi-natural experiment using Differences-in-Differences (DID) and Propensity Score Matching DID (PSM-DID) within a provincial fixed-effects framework. Parallel trend assumption is graphically validated prior to policy implementation. Software: Stata 16.0. Sample: Panel data for 28 Chinese provincial-level administrative regions (excluding Tibet, Ningxia, Chongqing, Hong Kong, Macao, Taiwan) from 2010 to 2020. Policy timing: B&R officially implemented in 2015; pre-period 2010–2014; post-period 2015–2020. Treatment definition: 17 provincial regions planned as key B&R areas (e.g., Beijing, Shanghai, Tianjin) constitute the intervention group; remaining regions form the control group. Outcome (GD level): Proxied by CO2 emission intensity of construction enterprises (CCO2) in each province, computed from standard coal consumption (from China Statistical Yearbook) and converted to CO2 using national energy conversion coefficients (General Rules for the Calculation of the Comprehensive Use of Resources). Lower CO2 intensity indicates higher GD. Control variables: - R&D: Investment in technological input by industrial enterprises above scale. - Sti: Completed investment in sewage treatment projects (proxy for pollution control investment). - LP: Labour productivity of construction enterprises (computed using total construction output value). - UC: Number of undergraduate students in general higher education (education level). - Cgdp: Regional GDP per capita (development level). Data sources: China Statistical Yearbook (NBS, 2021a), China Environmental Statistical Yearbook (NBS, 2021b), China Energy Statistical Yearbook (NBS, 2021c). Models: Baseline DID: CCO2_it = α_i + α1 TREATED_i + α2 TIME_t + α3 (TREATED_i × TIME_t) + μ_it. Extended with controls: CCO2_it = α_i + α1 TREATED_i + α2 TIME_t + α3 (TREATED_i × TIME_t) + Σ α_j Control_jt + μ_it. Estimation details: Variables (non-dummies) are log-transformed (after multiplication scaling) for comparability. Robustness via PSM-DID: 1:1 nearest-neighbor matching using psmatch2 with CCO2 as dependent variable and the five controls as covariates; balance tests confirm matched samples have no systematic differences (standardized biases near zero; t-tests nonsignificant).

- Policy effect (DID): The interaction term TREATED × TIME is significantly negative at the 1% level across specifications, indicating the B&R policy curbed the upward trend of CO2 emission intensity among construction enterprises in key regions, improving GD. Magnitude: estimated reduction of approximately 0.346 units in CO2 emission intensity after policy implementation. - Controls (DID Table 3): R&D investment shows a significant negative association with CO2 intensity (1% level), implying R&D enhances GD by reducing emissions intensity. Cgdp (GDP per capita) and UC (higher-education undergraduates) are significantly positive (1%), indicating higher development level and education level are associated with higher CO2 intensity (hindering GD) in the sample period. Sti (sewage treatment investment) has a small positive effect on CO2 intensity (10% level in the full model). LP (labour productivity) is positive but statistically nonsignificant. - Robustness (PSM-DID Table 5): After matching, TREATED × TIME remains significantly negative at the 1% level (coefficients roughly −0.6645 to −0.6811), reaffirming that B&R implementation suppresses CO2 intensity growth in treated regions. - Parallel trends: Pre-2015 emission trajectories for treatment and control groups are similar; post-2015, treated regions exhibit a relative decline in the growth of CO2 intensity. Overall: B&R policy promotes GD in construction enterprises in key regions; R&D supports GD, while higher development level, education level (as measured), labour productivity, and pollution-control investment tend to hinder GD by being associated with higher CO2 intensity.

The results support Hypothesis 1: the B&R policy significantly improves the GD level of construction enterprises in planned key regions by curbing CO2 emission intensity growth. This aligns with broader findings that policy interventions can suppress construction-sector emissions. Regarding Hypothesis 2, higher regional development (Cgdp) is associated with higher CO2 intensity, suggesting economic expansion without sufficient green transformation can impede GD. Hypothesis 3 is conceptually supported: greater labour productivity corresponds with higher emissions intensity (in line with prior evidence from construction and cement sectors), though in this dataset the coefficient is not statistically significant in the full model. For Hypothesis 4, pollution-control investment (Sti) correlates with slightly higher emissions intensity, plausibly because expanded wastewater treatment involves energy use and on-site GHG generation, thereby reducing measured GD in the short term. Hypothesis 5 is supported: R&D investment reduces emissions intensity by fostering green technological innovation, enhancing cleaner production. Hypothesis 6 finds that the education level proxy (UC) is positively associated with emissions intensity, which contrasts with some literature; the authors attribute differences to the proxy used (regional undergraduate counts vs. targeted GD education), regional scope (China vs. other countries), and sector focus (construction industry). Collectively, the findings indicate that while B&R policy facilitates greener outcomes in construction, complementary measures—targeted R&D, sector-specific education, and efficiency improvements in pollution control—are necessary to sustain and deepen GD impacts.

Using DID and PSM-DID on panel data (2010–2020) for 28 Chinese provinces, the study concludes: (1) The B&R policy curbs the upward trend in construction-sector CO2 emission intensity in key regions, thereby improving GD. (2) Regional development level, education level (as proxied), labour productivity, and pollution-control investment are associated with higher emissions intensity, hindering GD, while R&D investment reduces emissions intensity, promoting GD. Policy implications: - Scale up R&D investment by both government and enterprises to drive green technological innovation in construction. - Expand the coverage of B&R planning and strengthen policy instruments that incentivize cleaner production across multiple dimensions. - Enhance GD-oriented education and training specific to the construction sector to raise green awareness and skills, and guide productivity improvements toward low-carbon outcomes. - Improve the energy efficiency and emissions performance of sewage and pollution-control processes to avoid rebound increases in measured CO2 intensity. The paper contributes econometric evidence from China’s construction industry to inform green B&R strategies and sectoral policymaking.

- Data coverage is limited to 28 provincial administrative regions due to availability; several regions (Tibet, Ningxia, Chongqing, Hong Kong, Macao, Taiwan) are excluded. - Important determinants of construction-enterprise GD—such as market maturity and firms’ GD willingness—are difficult to quantify and not included. - The education proxy (number of undergraduates) may not capture GD-specific human capital or quality of education. - Short- to medium-term assessment may not reflect longer-run dynamic adjustments. Future research could incorporate additional qualitative or simulated measures (e.g., computer simulation), richer firm-level data, and alternative proxies to refine mechanism identification.