Environmental protection tax, digital transformation, and agricultural carbon emissions in China’s agricultural sector: evidence from provincial panel data

Introduction Against the backdrop of China’s Dual Carbon Goals and Rural Revitalization Strategy, clarifying the mechanisms through which Environmental Protection Tax (EPT) and digital transformation affect agricultural carbon emissions is significant for building sustainable food systems. Although the EPT does not directly tax the majority of agricultural producers due to the statutory exemption of agricultural non-point source pollution, it may still influence agricultural carbon emissions through three indirect channels (upstream cost pass-through, regulatory signaling and enforcement, and fiscal expenditure and technology diffusion), whose effectiveness is conditional on digital governance capacity. Methods This study constructs a balanced provincial panel dataset covering 31 provinces in mainland China from 2011 to 2023 (403 observations). Based on the theoretical framework of “Environmental Protection Tax, Digital Transformation, and Agricultural Carbon Emissions,” a two-way fixed effects model is employed as the baseline, supplemented by a multi-pronged identification strategy comprising lagged explanatory variables, a difference-in-differences design exploiting the 2018 Environmental Protection Tax Law as a quasi-natural experiment, a placebo test based on 1,000 random reassignments, and a Spatial Durbin Model with LeSage–Pace decomposition into direct, indirect, and total effects. Results Agricultural carbon emissions exhibit a pattern of “initial increase followed by stabilization in total volume, overall decline in intensity, and significant regional disparities.” A 1% increase in EPT intensity is associated with an approximately 0.05% decrease in agricultural carbon emission intensity, and a 1% increase in the Digital Financial Inclusion Index is associated with an approximately 0.07–0.08% decrease. The coefficient of the “EPT × Digital Transformation” interaction term is significantly negative, indicating that digital transformation amplifies the emission reduction effect of EPT; this synergy is particularly pronounced in eastern and central regions as well as major grain-producing regions, but limited in regions with weak digital infrastructure or livestock-dominated agricultural structures. The quasi-natural experiment yields a significant DID coefficient (−0.0254, p < 0.05), with the parallel trend test (joint F = 0.87, p = 0.493) and placebo test ( p = 0.046) jointly supporting a causal interpretation, while the Spatial Durbin Model reveals substantial cross-province spillovers that the within-province estimate fails to capture. Discussion In the digital era, relying solely on traditional environmental regulation is insufficient to fully unleash agricultural emission reduction potential. The coordinated design of Environmental Protection Tax with digital infrastructure, digital agriculture, and digital finance development represents a key pathway for achieving increased agricultural production, improved efficiency, and reduced emissions, as well as for building green and resilient food systems.