Impact of government environmental attention on urban green economy efficiency: Evidence from text analysis and a three-stage network SBM model

The green economy, a model for economic, social, and environmental sustainability, is widely advocated globally. Green economic efficiency (GEE) evaluates green development effectiveness. However, existing research rarely treats government environmental attention (GEA)—a key environmental governance driver—as an independent factor shaping GEE, nor does it provide reliable GEA quantification or comprehensive GEE measurement methods. To fill these gaps, this study constructs a 33-environmental-keyword dictionary to quantify GEA via text analysis of 248 Chinese cities’ government work reports (2009–2020). Meanwhile, we develop a three-stage network SBM model under global benchmarking (with undesirable outputs) to measure GEE, integrating economic production, social development, and environmental governance subsystems to address traditional DEA model limitations. Based on panel data (2010–2021, 2,976 observations), our fixed-effects models show a 1-percentage-point increase in GEA significantly raises GEE by 7.386 percentage points. Heterogeneity analysis reveals stronger effects in low-GEE, secondary industry-dominant, and low-urbanization cities. This study enriches GEE driver theory, provides replicable measurement methods, and offers targeted evidence for governments promoting green development.