Answers are grounded in this paper and cite their sources.
This paper examines the relationship between foreign direct investment (FDI) and greenhouse gas (GHG) emissions in middle- and high-income countries. Particular attention is paid to income-group heterogeneity and the context of green technology. Using an unbalanced panel of 74 countries from 1975 to 2021, we estimate static and dynamic panel data models. The static analysis includes pooled OLS, random effects and two-way fixed effects specifications, incorporating income group and FDI-patent interactions. To address dynamic persistence and potential endogeneity, we also employ one-step and two-step system GMM estimators with lagged dependent variables and income-specific FDI effects. The results suggest a generally positive association between FDI and GHG emissions, although the magnitude and statistical significance of this relationship vary across different models. Dynamic system GMM estimations provide further evidence of a positive association between FDI and emissions. In the main interaction specifications, estimated effects are somewhat larger for middle-income economies than for high-income ones, though this difference is not consistent across all specifications and should be interpreted as suggestive rather than conclusive. Analysis of disaggregated green patent categories reveals substantial technological heterogeneity. Some patent classes are associated with a weaker FDI–emissions relationship, whereas others appear to strengthen it. In contrast, aggregate patent activity shows either positive or statistically insignificant relationships with emissions, suggesting that broad patent measures may obscure important technology-specific effects. Overall, FDI does not automatically lead to environmental improvement. Its consequences appear to vary with income-related structural conditions and the technological orientation of the host economy. These findings support broad policy caution: FDI attraction strategies should be evaluated together with environmental monitoring, and green innovation should be assessed at a more disaggregated technological level.
No takes yet. Share an insight, caveat, or question.
Mirzoyan et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: