Political polarization and policy uncertainty have become increasingly consequential for regional economic adjustment, yet their joint role in shaping socioeconomic resilience remains underdeveloped in the literature. This study advances the debate by conceptualizing regional resilience as the outcome of a multi-layer socioeconomic system in which external policy disturbances, institutional fragmentation, and structural adaptive capacity interact over time. Using balanced panel data for 16 Korean regions from 2004 to 2023, the analysis develops an integrated empirical framework that combines panel local projections, threshold estimation, structural moderation tests, dynamic robustness checks, and forward-looking machine-learning prediction. The results show that policy uncertainty is associated with lower regional socioeconomic resilience and that this effect persists over time. More importantly, political polarization does not simply accompany weaker resilience; it amplifies the transmission of uncertainty shocks, especially once institutional fragmentation crosses a critical threshold. Structural conditions further shape this process. Digital transformation, industrial diversification, and financial depth reduce vulnerability, whereas trade exposure intensifies it. These findings indicate that resilience is not determined by economic structure alone, nor by institutional instability in isolation. It emerges from the interaction between disturbance, amplification, and adaptive capacity within a regional system. The predictive analysis reinforces this interpretation. Variables identified as central in the econometric models also carry forward-looking information about future vulnerability states. This study therefore contributes not only by combining methods, but by linking explanation and prediction within a single systems-oriented account of regional resilience. The Korean case shows how institutional coherence and structural adaptability jointly condition resilience under uncertainty.
Yang et al. (Fri,) studied this question.