Marginal abatement cost of urban emissions under climate policy: Assessment and projection for China's 2030 climate target

• MAC serves as a valid and unbiased outcome indicator to evaluate climate policy. • LCCP raises MAC via green innovation and industrial upgrading, with city variations. • A hybrid approach predicts future trajectory of city-level MAC to 2030. • Using cost-intensity matrices to classify cities and their decarbonisation roadmaps. This study evaluates the effectiveness of Low-Carbon City Pilot programme and predicts its future trajectory through 2030, using the marginal abatement cost of CO 2 as the outcome indicator to address policy anticipation effects. Drawing on Chinese city-level panel data (2006–2019), we estimate and forecast marginal abatement cost through a hybrid approach combining machine learning regression, autoregressive integrated moving average and quadratic directional distance function. A Difference-in-Differences model is used to assess the policy's impact, while a cost-intensity matrix is employed to explore cities’ low-carbon transition stages and pathways to 2030. The results reveal that Low-Carbon City Pilot programme significantly increased the marginal abatement cost of CO 2 in pilot cities, driven by green technology innovation and industrial structure upgrading. These effects are more pronounced in developed, eastern, non-resource-based and non-industrial cities. Under the carbon peaking target, most cities will experience rising abatement costs, with widening inter-city divergence. By 2030, fewer than one-fifth of cities will achieve low-carbon transition, while the majority will remain in mid- or early-transition phrases. We recommend exploring emission reduction potential in central, western, and resource-based cities, addressing transition barriers, and formulating city-specific policies that account for unique underlying costs while enhancing synergies to avoid carbon leakage and worsening climate inequality.