Impact of the uncertainty triad on US economic activity: a multi-resolution quantile machine learning framework

Climate, energy, and trade policy uncertainties form a triad that increasingly shapes US economic activity (EAC). This study investigates the association of climate policy uncertainty (CPU), energy-related uncertainty (EUI), and trade policy uncertainty (TPU) with US economic activity. It employs a novel CEEMDAN-based Multivariate Nyström Quantile-on-Quantile Kernel Regularized Least Squares (C-MN-QQKRLS) model, which extends traditional bivariate analysis, incorporates time-frequency asymmetries without absorbing the effects of other pertinent shocks, and addresses computational inefficiencies in conventional kernel designs. Results show that both EUI and TPU are positively associated with EAC in the short term, indicating that economic agents quickly adapt their strategies to changing circumstances. However, in the long term, the uncertainty triad and economic activity are negatively associated at their middle-to-high quantiles, suggesting that long-term uncertainty constrains growth. The persistent negative long-term CPU-EAC, EUI-EAC, and TPU-EAC relationships call for for immediate policy attention. These findings remain robust to macroeconomic controls, alternative variable measures, and nested estimator comparisons. The study highlights the importance of addressing climate-related challenges, achieving a balanced energy mix, and mitigating trade frictions. The evidence indicates that although temporary policy uncertainty shocks produce only short-term effects, persistent or long-term uncertainty imposes a structural constraint on US economic activity.