Using the Stochastic Impacts by Regression on Population, Affluence, and Technology (STIRPAT) model and an unbalanced panel dataset of 128 countries covering 1990–2014, this study aims to examine the key impact factors (KIFs) of the global and regional carbon dioxide (CO₂) emissions and analyse the effectiveness of non-renewable and renewable energies. Given the potential cross-sectional dependence and slope heterogeneity, a series of econometric techniques allowing for cross-sectional dependence and slope heterogeneity is applied. The overall estimations imply that the KIFs at the global level are economic growth, followed by population size, non-renewable energy, and energy intensity in order of their impacts on CO₂ emissions; conversely, the KIFs at the regional level vary across different regions and estimators. The results also suggest that renewable energy can lead to a decline in CO₂ emissions at the global level. At the regional level, only for two regions (i.e., S. & Cent. America and Europe & Eurasia) renewable energy has a significant and negative effect on CO₂ emissions, which may be affected by the share of renewable energy consumption in the primary energy mix. Finally, the results indicate varied causality relationships among the variables across regions. Abbreviations: AMG: Augmented mean group; BP: British Petroleum; BRICS: Brazil, Russia, India, China, and South Africa; CCEMG: Common correlated effects mean group; CD: Cross-section dependence; CIPS: Cross-sectionally augmented Im, Pesaran, and Shin; CO₂: Carbon dioxide; PS: Population size; D-H: Dumitrescu-Hurlin; EI: Energy intensity; EU: European Union; EU-5: Germany, France, Italy, Spain, and the United Kingdom; Europe & Eurasia, Europe and Eurasia; GDP: Gross domestic product; IEA: International Energy Agency; KIF: Key impact factor; LM: Lagrange multiplier; Mtoe, Million tonnes oil equivalent; NRE: Non-renewable energy; RE: Renewable energy; S. & Cent. America, South and Central America; STIRPAT: Stochastic Impacts by Regression on Population, Affluence, and Technology; VECM: Vector error correction model; WDI: World Development Indicators

To explore how renewable energy reduces carbon dioxide (CO₂) emissions, this study decomposes the impacts of renewable energy on CO₂ emissions by employing the Generalized Divisia Index Method (GDIM), based on panel data of 25 countries along the Belt and Road Initiative (BRI) for the period 2005–2019. Also, we explore their decoupling relationships. The results indicate that: (1) The numbers of countries with an ideal strong decoupling state reach the highest point during 2010 and 2015. (2) The growth in renewable energy scale is the key driving force responsible for promoting CO₂ emissions in most countries along the BRI, while the main factor inhibiting CO₂ emissions is the carbon intensity of renewable energy. (3) The carbon mitigation effect of the increase in the proportion of renewable energy is considerably higher than that of technological improvements from renewables, and it will promote the transformation of the relationship between renewable energy and carbon emissions into a more ideal strong decoupling relationship. (4) The effect of carbon efficiency on emission is not significant at this stage, but it will be much more important in the future.