Digital labour and the green transition: heterogeneous environmental effects of platform-based work

Purpose This study aims to investigate the environmental implications of digital labour expansion and whether the relationship between platform-based digital work and carbon emissions depends on occupational composition and electricity-generation structure. Specifically, it evaluates whether software-task specialisation and fossil-fuel electricity dependence condition the association between online labour intensity and per capita CO2 emissions. Design/methodology/approach The study constructs an unbalanced cross-country panel covering 116 countries over the period 2016–2023 by combining the Online Labour Index with macroeconomic, environmental and electricity-mix indicators from the World Development Indicators. Two-way fixed effects (TWFE) models are estimated to control for country and year heterogeneity. To address dynamic persistence and potential endogeneity, two-step system generalised method of moments (System-GMM) estimations with internal instruments are used. Findings Baseline TWFE results indicate no statistically significant average association between digital labour intensity and per capita CO2 emissions. Once occupational composition is incorporated; however, the interaction between digital labour intensity and software-task share is positive and statistically significant, indicating that the association becomes less negative as software specialisation increases. Fossil-fuel electricity dependence also positively moderates the digital labour-emissions relationship, while the insignificant three-way interaction suggests that electricity mix and occupational composition operate as distinct conditioning factors. System-GMM estimates confirm strong emissions persistence and reproduce the occupational-composition pattern. Originality/value This study introduces a labour-based measure of digitalisation into macro-environmental analysis and demonstrates that the environmental implications of digital transformation are conditional on occupational composition and electricity-generation structure. By highlighting the moderating role of occupational structure, the paper contributes to science and technology policy debates on how digital labour mobility, skills upgrading and clean energy transitions can be integrated to support sustainable development.