Understanding Economic Resilience Using New Quality Productivity Across Multi-Scale Spatial Locations: Machine-Based Spatio-Temporal Effects

Amid intensifying climate crises, widening inequalities, and geopolitical volatility, spatial economic resilience (SER) has become critical for regions facing systemic uncertainty. Traditional land-intensive productivity models prove increasingly untenable as spatial resources become finite and development space constrained. China’s new quality productivity (NQP) has emerged as a strategic response emphasizing innovation-driven structural renewal and territorial coordination. Conceptually, NQP is positioned as a SER-oriented strategy prioritizing adaptability, recoverability, and transformability. However, its actual associations remains theoretically overlooked and empirically untested, with existing research viewing it narrowly as technological upgrading while neglecting institutional dimensions, spatial dependencies, and multi-scalar heterogeneities. This study explores how NQP relates to SER from a spatio-temporal perspective: (1) How do the technological and institutional dimensions of NQP relate to SER? (2) What are the spatial patterns of NQP-SER associations across multi-scale locations? Employing XGBoost-SHAP, spatial generalized difference-in-differences, and Geographical Gaussian Process Regression across provincial, city, and enterprise scales in China, we find that NQP’s two dimensions relate to SER very differently. The technological–industrial dimension is the strongest predictor of SER at the provincial scale, exhibiting threshold-type, non-linear associations, while its predictive salience attenuates at the city and enterprise scales, where industrial structure and firm-specific fundamentals are more strongly associated with resilience. The institutional dimension, by contrast, is not positively associated with above-expectation resilience: once common shocks and provincial heterogeneity are absorbed, higher institutional policy intensity is negatively associated with SER, both within provinces and across neighbouring provinces. Spatially, provincial associations rely on coordination and interregional spillovers, while city associations concentrate in nodal clusters where the strength of association depends on capability–context alignment. The findings provide practical theoretical and analytical guidance for tailored policy-making in structurally diverse Global South facing ongoing uncertainty.