Revisiting the energy-economic growth nexus in low versus high-income economies: New evidence using nonlinear DCCE, CS-NARDL, and CS-NDL approaches

This study examines the two-way asymmetric energy-economic growth relationship using novel second-generation nonstationary panel approaches, namely nonlinear dynamic common correlated effects (nonlinear DCCE), cross-sectionally augmented nonlinear autoregressive distributed lag (CS-NARDL) and cross-sectionally augmented nonlinear distributed lag (CS-NDL) approaches. These models cover long panel data estimation issues of nonstationarity, heterogeneous slopes and cross-sectional dependence. However, their standard models are linear approaches and neglect asymmetric effects. Besides, economic variables might vary over time and react heterogeneously to shocks. Hence, this study fills this gap by innovating the linear models into nonlinear models to explore asymmetric relationships. The results’ robustness is checked with baseline models. This study compares the nexus in two contrasting income groups, that is, lower-income and high-income economies, from 1990 to 2019. Overall, energy consumption and gross domestic product (GDP) are positively linked in the long-run bidirectional nexus for both groups. Nevertheless, energy consumption has a weaker influence on the lower-income nations’ GDP, signifying lower energy efficiency. The favourable short-run impact of energy consumption decreases on lower-income countries’ GDP is noticeable, highlighting the importance of energy conservation. The reverse linkage's result, which reveals higher GDP accelerates energy consumption in both groups, reinforces energy conservation and the energy transition to renewables to ensure energy security. The upsurge in energy consumption due to the increased GDP is larger in lower-income economies, suggesting room for improvement in energy technologies. Furthermore, energy consumption and GDP are affected by their delayed effects. Therefore, authorities should always be alert to historical trends and potential business cycle fluctuations.