Energy consumption and real GDP in G-7: Multi-horizon causality testing in the presence of capital stock

- We examine causality among energy consumption, real GDP and capital stock in G-7.
- Two recent multi-horizon time series causality methods are employed.
- We take into account the presence of structural breaks in the form of trend changes.
- Crucial information on the dynamic interaction among the variables is uncovered.
- Real GDP dominates in anticipating energy consumption.

This paper applies two recent time series methods to re-examine the causal relationship among energy consumption, real GDP and capital stock in G-7 countries. These methods, the Dufour et al. [2006, Journal of Econometrics, 132:337-362] multiple horizon causality testing and the Hill [2007, Journal of Applied Econometrics, 22:747-765] sequential causality testing allow to test for (non)causality in a multivariate framework and can further reveal the time profile of causal effects, the presence of causation delays and the direct or indirect nature of the causal effects. Given the trending nature of the time series employed, we further take into account the presence of structural breaks in the form of trend changes. Our empirical results show that multi-horizon causality testing does uncover crucial information with respect to the dynamic interaction among energy consumption, real GDP and capital stock, while structural breaks do exist and appear to be critical for causality inference. In regard to causality direction, we find that real GDP dominates in anticipating energy consumption in G-7 countries.