Unlike previous energy consumption-economic growth studies, this study examines the relationship among energy consumption, economic growth, employment and gross fixed capital formation for 17 highly developed OECD countries by employing both the Toda-Yamamoto procedure which based on asymptotic critical values and the bootstrap-corrected causality test, since non-normality of the error term harms the validity of the Toda-Yamamoto procedure. This study finds that there is very small bias due to the assumption of normality. Furthermore using different information criterions, importance of lag length is tested. Findings indicate that selection of lag length is important for Denmark, Ireland, Norway and Spain. It is concluded that while there exists uni-directional causality running from energy consumption to real GDP for Japan, bi-directional causality is found for Italy, New Zealand, Norway and Spain. On the other hand, uni-directional causality from GDP to energy is found for Australia, Canada and Ireland whereas no causal nexus is found for all of other nine countries. Our analyses covering the sample periods imply that Japan, Italy, New Zealand, Norway and Spain should not follow energy conservation policy at the aggregated level, since the reduction of energy damages the economic growth. (C) 2012 Elsevier Ltd. All rights reserved.