Do binding climate commitments decarbonize the power sector? Evidence from G20 economies after the Paris Agreement.

Abstract

Decarbonizing the power sector is central to achieving the Paris Agreement's temperature goals, yet rigorous evidence on whether binding climate commitments actually deliver measurable reductions in electricity-sector emissions remains scarce. This paper asks whether the formal adoption of binding climate commitments reduces power-sector emission intensity in G20 economies, and whether green bond issuance amplifies this effect. We construct an annual panel for 19 G20 countries over 2014-2024, combining satellite-based power-sector CO2 emissions, macroeconomic indicators, hand-coded policy adoption dates, and country-level green bond issuance. Exploiting the staggered adoption of Paris-aligned commitments between 2016 and 2022 as quasi-experimental variation, we apply heterogeneity-robust difference-in-differences and event-study estimators. We find that binding climate commitments reduce power-sector emission intensity by approximately 9-12%, with effects emerging immediately at adoption and persisting for at least seven years without significant attenuation. These reductions reflect genuine absolute emission cuts rather than slower GDP growth, providing evidence of strong decoupling in the G20 power sector. By contrast, green bond intensity does not exert a statistically significant incremental effect once policy adoption is accounted for, suggesting that regulatory commitment - rather than labelled financial flows is the primary driver of observed decarbonisation at current scales. Effects are systematically larger in high-income and high-emitting countries. The study offers one of the first post-Paris, sector-specific causal evaluations across the full G20, embeds green finance within a formal policy evaluation framework, and demonstrates the value of modern event-study designs for climate policy analysis.