When Does Digitalization Accelerate the Green Transition? Asymmetric Evidence from a Heterogeneous G20 Panel

This paper examines whether digitalization is associated with stronger links between green finance (GF), renewable-energy innovation (REI), and the green low-carbon energy transition across 19 G20 economies over 2000–2025. A data-consistency audit corrects the low-carbon electricity input. The benchmark Green Low-Carbon Energy Transition index (GLCET_B) is a transparent four-pillar composite covering renewable deployment, power-sector decarbonization, energy efficiency, and coal exit; PCA-based, equal-weight, and indicator-exclusion variants serve as measurement-sensitivity checks. The empirical sequence distinguishes confirmatory evidence from exploratory diagnostics. For the preferred benchmark, the LREI × DIG interaction is positive in TWFE-DK (0.0631, p < 0.001) and remains significant under 999-replication country-level wild-cluster inference (p = 0.011), pooled Common Correlated Effects (wild p = 0.009), cluster-aware DML (wild p = 0.001), first differences, complete-observation sensitivity, and all leave-one-country-out samples. Heterogeneous-slope CCE Mean Group estimates remain positive (0.044–0.074) but imprecise, indicating an average rather than uniform association. By contrast, GF × DIG is not robust under conservative small-cluster inference. The subgroup evidence suggests that the LREI–digitalization complementarity is concentrated in emerging/lower-digitalization G20 economies, but this result is interpreted cautiously because the subgroup cluster counts are small and digitalization groups coincide with emerging/advanced status. All findings are interpreted as conditional panel associations rather than causal or long-run effects.