Climate change and the weather-energy Nexus in US energy demand

This study examines the long-run and short-run elasticities of U.S. residential and commercial final energy consumption with respect to historical variations in heating and cooling degree days and compares these weather sensitivities across sectors. It also investigates whether joint exposure to heating and cooling conditions has an additive or attenuating effect on annual energy consumption. Using annual U.S. data from 1960 to 2025, we employ a Fourier-augmented bootstrap autoregressive distributed lag (Fourier ARDL) framework to estimate weather–energy relationships while accommodating smooth structural change. We also use a Fourier-augmented Toda–Yamamoto procedure to examine directional predictive relationships. The results confirm long-run equilibrium relationships in both sectors. Heating degree days have positive, significant long-run effects on residential and commercial energy consumption. Cooling degree days become significant when we include the heating–cooling interaction term. The interaction coefficient is negative and significant, indicating that joint heating and cooling exposure has an attenuating, non-additive association with annual energy consumption. Because we include the interaction term, we interpret the estimated HDD and CDD coefficients as conditional effects rather than unconditional elasticities. Residential energy demand adjusts to disequilibrium approximately three times faster than commercial demand. The causality results indicate directional predictive relationships from HDD to commercial consumption and from CHDD to residential consumption. Overall, the findings provide evidence on historical weather–energy relationships and offer implications for demand forecasting, grid planning, and energy-efficiency policy, but do not constitute a formal counterfactual attribution of energy changes to climate change.