Quantifying the Long-Run Carbon Reduction Associations of Coal Phasedown in China: Evidence from the Bootstrap ARDL Model

China’s pursuit of its dual-carbon targets—carbon peaking before 2030 and carbon neutrality before 2060—critically hinges on the structural decarbonization of its energy system. Utilizing a Bootstrap Autoregressive Distributed Lag (Bootstrap ARDL) bounds testing framework on macro-level time-series data spanning from 1965 to 2022, this study quantifies the long-run elasticities and short-run dynamic transmission mechanisms governing China’s carbon dioxide emissions. The empirical results establish a robust long-run cointegrated relationship among total CO2 emissions, coal consumption share, primary energy consumption, and per capita real GDP. Long-run elasticity quantifications indicate that primary energy consumption exhibits an elastic scale association of 1.108, while the coal consumption share shows a substantial carbon sensitivity of 0.487. These findings indicate that fuel switching away from coal dependence is a critical driver for decarbonization alongside capping the absolute energy scale. Conversely, the long-run parameters for per capita GDP and its quadratic term are not statistically significant; thus, the results do not provide evidence of an inverted U-shaped EKC within the specification used in this study, as the GDP terms are estimated conditional on primary energy scale and fuel structure. The error correction term (ECTt−1 = −0.373, p = 0.004) confirms that short-run disequilibrium adjusts back toward the long-run equilibrium trajectory at an annual rate of 37.3%. Based on these empirical insights, a policy roadmap incorporating enforced coal phasedowns, energy growth decoupling, and market-based Emissions Trading Scheme (ETS) expansions is proposed.