Mapping Policy Pathways for FDI Resilience Under Climate Risk: An Econometric and Machine Learning Approach

This study investigates how physical climate risk reshapes foreign direct investment (FDI) flows and identifies policy pathways to enhance investment resilience. Integrating econometric and machine learning frameworks across 176 countries (1995–2023), we employ panel fixed effects, instrumental variables, and explainable machine learning to capture both causal relationships and nonlinear dynamics, with physical climate risk measured using the Global Climate Physical Risk Index (GCPRI). Three core findings emerge. First, higher climate risk significantly deters FDI, particularly in countries with weak adaptive capacity and low technological readiness. Second, the climate–FDI relationship is nonlinear: FDI remains resilient under moderate exposure but declines sharply beyond a critical threshold (GCPRI≈0.6). Third, counterfactual simulations reveal that technological innovation is the most potent driver of FDI resilience, outpacing governance and financial development, while coordinated reforms generate super‐additive gains. These findings demonstrate that climate resilience is a constructed capability. Strategic sequencing: beginning with technological upgrading, consolidating institutions, then integrating climate policy can transform climate exposure from an unmanageable threat into a governable risk.