The double-edged effects of regional environmental governance: evidence from policy tools of belt and road initiative

Abstract Regional environmental governance plays a pivotal role in global climate cooperation; however, the effectiveness of specific policy tools, especially their interactions, remains underexplored. The Belt and Road Initiative (BRI), with over 140 participating countries, provides a unique empirical context to assess how regional environmental governance affects carbon dynamics. Here, construct a novel dataset of 282 BRI agreements (2014–2024) and apply a multi-period difference-in-differences model, complemented by machine learning to quantify causal effects and identify synergistic interactions. Results indicate a dual impact: while BRI environmental governance significantly reduces carbon intensity, it simultaneously increases overall carbon footprints due to infrastructure expansion and trade growth. Further, there is substantial heterogeneity in the effectiveness of individual instruments, which demonstrates that certain policy combinations amplify mitigation in developed regions but exacerbate emissions in developing economies. These findings highlight the urgent need to tailor regional governance frameworks to local contexts and long-term sustainability goals. Our research makes three contributions: (1) it introduces a text-based empirical strategy to identify and classify environmental policy tools embedded in international agreements; (2) it provides the first integrated causal and machine-learning-based assessment of policy heterogeneity and synergy under the BRI framework; and (3) it advances methodological innovation in evaluating transnational environmental governance using multi-period panel data with staggered treatment.