Production-based disaggregated analysis of energy consumption and CO 2 emission nexus: evidence from the USA by novel dynamic ARDL simulation approach.

The study investigates the effects of energy consumption on carbon dioxide (CO 2 ) emissions by focusing on production sources. In this context, the study focuses on the USA as the leading economy, includes monthly data between January 1973 and April 2022, and performs dynamic autoregressive distributed lag (ARDL) (DYNARDL) simulations. Besides, kernel-based regularized least squares (KRLS) and cointegrating regression approaches are applied for robustness checks. The results reveal that (i) there is cointegration between sub-components of the energy production and CO 2 emissions in the long run; (ii) fossil energy and nuclear energy production have an increasing effect on the CO 2 emissions in the both short and long run; (iii) renewable energy production has an increasing effect on the CO 2 emissions in the short run, but has a decreasing effect in the long run; (iv) negative (positive) shocks in the fossil energy production have a decreasing (increasing) effect on the CO 2 emissions, whereas negative (positive) shocks in the renewable energy production have an increasing (decreasing) effect on the CO 2 emissions in case of counterfactual shocks; (v) there is a casual-effect nexus between energy production sources and CO 2 emissions; and (vi) KRLS and cointegrating regression results validate the robustness of the DYNARDL simulation outcomes. Moreover, policy implications are discussed.