Green Fertilization Adoption and Current-Period Fertilizer Expenditure–Output Associations: Evidence from Chinese Apple Farmers

Green fertilization technologies (GFTs) have been widely promoted to support sustainable orchard production in China’s apple industry, yet their associations with current-period economic outcomes remain insufficiently understood. This study examines the relationship between GFT adoption and the fertilizer expenditure-based output ratio (FEOR) among 409 surveyed apple-growing households in Shandong Province, China, together with heterogeneity across subsidy conditions and technology components. FEOR is defined as physical apple output per CNY of fertilizer expenditure and is therefore a physical-to-monetary ratio rather than a measure of agronomic nutrient-use efficiency or overall farm profitability. Under the maintained assumptions of an endogenous switching regression (ESR) model, the estimated adopter-specific counterfactual difference is −1.185 log-points. Complementary estimators yield negative but substantially different estimates, and inference is subject to methodological limitations, including numerical instability and insufficient covariate balance and overlap in some alternative estimators. In a separate OLS interaction analysis, the average marginal adoption–FEOR association is −0.958 log-points when evaluated at no subsidy receipt, compared with +0.147 log-points, which is not statistically significant, when evaluated at subsidy receipt. Exploratory technology-specific regressions indicate a negative conditional association for organic fertilizer substitution, a positive conditional association for soil testing and formula fertilization after accounting for co-adoption, and no statistically significant conditional association for integrated water-fertilizer management. These technology-specific estimates are constrained by substantial co-adoption and possible selection bias. Separate regression-adjustment analyses indicate associations with higher fertilizer expenditure and lower physical output, describing the components of FEOR rather than identifying causal mechanisms. Importantly, the survey did not record the year of initial GFT adoption or adoption duration. The findings therefore describe contemporaneous fertilizer expenditure–output associations and cannot establish whether the observed differences represent temporary transition costs or changes in economic performance following adoption.