LAUSR

Chemical fertilizer is one of the most important input factors in agricultural production, but the excessive use of fertilizer inevitably leads to the loss of agricultural eco-efficiency (AEE). Therefore, it is necessary to explore the impact of fertilizer use intensity (FUI) on AEE. However, ordinary panel regression, based on the assumption of parameter homogeneity may yield biased estimation conclusions. In this regard, a panel quantile regression model (QRM) was constructed with the provincial panel data of China from 1978–2020 to test the difference and variation of this impact under heterogeneous conditions. The model was then combined with the spatial econometric model to explore the effect of the spatial lag factor. The results are as follows: (1) The QSM has unveiled a great improvement space for AEE that remains low overall, despite displaying a rising trend; the highest AEE is in the eastern region. (2) The FUI has a significant negative effect on AEE with the rise in quantiles, this negative effect tended towards weakening overall, although it rebounded slightly; it was stronger in areas with low AEE. It is necessary to consider the heterogeneous conditions in comparison with the average treatment effect of ordinary panel econometric regressions. (3) The impact of FUI shows significant variability in different economic sub-divisions and different sub-periods. (4) After considering the spatial effect of fertilizer use, the negative influence on local AEE had a faster decay rate as the quantile rose, but could produce a positive spatial spillover effect on AEE in neighboring areas. Local governments should dynamically adjust and optimize their fertilizer reduction and efficiency improvement policies according to the level and development stage of their AEE to establish a complete regional linked agroecological cooperation mechanism.