Agrochemicals, machines and biodiversity in Ghana: the ecological cost of land preparation methods of large-scale land acquisition by domestic entities

Purpose This study investigates how land preparation methods of large-scale land acquisitions by domestic entities (LSLADE)—specifically mechanization and agrochemicals—affect vegetation health and biodiversity in northern Ghana. Biodiversity outcomes are measured using satellite-derived vegetation indices: NDVI and GNDVI. Design/methodology/approach A two-round panel dataset of 398 smallholder farm households, fixed effects, random effects, Mundlak-adjusted models, control function approaches and a misclassification-robust local average treatment effect (MR-LATE) estimator to correct for selection bias and misreporting. Findings A 1% increase in LSLADE investment is associated with a 0.06–0.07% reduction in NDVI and a 0.10–0.11% reduction in GNDVI. Households using LSLADE services report 0.02–0.03 declines in both indices. MR-LATE estimates confirm significant 3% (NDVI) and 6% (GNDVI) declines, underscoring the adverse ecological effects of LSLADE-induced land clearing. Research limitations/implications The results may not be generalizable beyond the specific context of northern Ghana. Practical implications Policymakers should integrate agroecological methods and remote biodiversity monitoring into LSLADE policies to mitigate environmental trade-offs. Originality/value This is one of the first studies to assess biodiversity loss from LSLADE in Ghana using satellite indices and rigorous econometric approaches. It provides critical evidence for designing sustainable agricultural investment policies.