Assessing the relevance of the Granger non-causality test for energy policymaking: theoretical and empirical insights

This paper highlights the relevance of Granger non-causality tests in energy economics research, particularly for informing public policy decisions. While approaches such as CS-ARDL and estimators like AMG and CCEMG are widely used, they do not fully capture the predictive relationships between variables. To illustrate this, we revisit the findings of Irfan et al. (2023), who analyzed factors influencing energy transitions in G-7 and E−7 economies using Westerlund's (2007) cointegration method and CS-ARDL. Additionally, we incorporate data from Zhao et al. (2024) to estimate the relationships between artificial intelligence, GDP, trade, population, and energy efficiency using the CS-ARDL approach, complemented by Granger non-causality tests. Our results, in some cases, expand upon the evidence provided by Irfan et al. (2023), while in others, they suggest a different interpretation of key relationships. Specifically, we find that the mineral market does not exhibit significant predictive power over energy transition, whereas trade and economic growth contribute meaningfully to renewable energy development. Furthermore, using data from Zhao et al. (2024), we confirm that incorporating non-causality tests enhances the interpretation of CS-ARDL estimates, demonstrating that these tests provide valuable insights into the directionality of economic and energy relationships, which is important for policy formulation. These findings highlight the importance of integrating non-causality tests with traditional econometric methods to derive more robust and policy-relevant conclusions. • The importance of the Granger non-causality test was emphasized. • The CS-ARDL approach and AMG, CCEMG estimators may be insufficient for assessing variable predictive power. • Granger non-causality tests are recommended for renewable energy policy studies.