Assessing the spatio-temporal impacts of digital economy drivers on carbon emission intensity: An interactive geographically and temporally weighted regression framework

This study aims to precisely quantify the spatio-temporal impacts of specific digital economy (DE) drivers and their interactions on carbon emission intensity (CEI) across China. Utilizing a panel dataset from 30 Chinese provinces (2011-2022), this study first constructs a comprehensive composite digital economy index. The core of the methodology is a novel interactive geographically and temporally weighted regression model, which integrates interaction factors identified by the optimal parameter geographical detector into the geographically and temporally weighted regression model. The results indicate that (1) Among DE driving factors, software business revenue and enterprise R&D expenditure intensity exert the strongest explanatory power on CEI. (2) Interactions between DE driving factors augment their individual effects, with the interaction between proportion of digital industry employees and per capita telecommunication business volume exhibiting the strongest explanatory power. (3) The impacts of DE driving factors and their interactions on CEI exhibit pronounced spatial heterogeneity. Interactions between factors induce nonlinear changes in their individual effects, mainly reflected in the direction, intensity, and spatial scope of their impacts on CEI. These findings provide critical and actionable insights for designing regionally differentiated policies, enabling policymakers to harness the synergistic potential of the digital economy for targeted carbon mitigation, thereby supporting the achievement of China’s “Dual Carbon” goals. • The indicator system for a composite digital economy index (CDEI) is constructed. • We assess spatiotemporal characteristics of CDEI and carbon emission intensity (CEI). • The digital economy (DE) shows significant negative spatial correlations with CEI. • We propose a new interactive geographically and temporally weighted regression model. • Impacts of DE drivers and their interactions on CEI all show spatial heterogeneity.