This paper tests for the short and long-run relationship between economic growth, carbon dioxide (CO2) emissions and energy consumption, using the Environmental Kuznets Curve (EKC) by employing both the aggregated and disaggregated energy consumption data in Malaysia for the period 1980-2009. The Autoregressive Distributed Lag (ARDL) methodology and Johansen-Juselius maximum likelihood approach were used to test the cointegration relationship: and the Granger causality test, based on the vector error correction model (VECM), to test for causality. The study does not support an inverted U-shaped relationship (EKC) when aggregated energy consumption data was used. When data was disaggregated based on different energy sources such as oil, coal, gas and electricity, the study does show evidences of the EKC hypothesis. The long-run Granger causality test shows that there is bi-directional causality between economic growth and CO2 emissions, with coal, gas, electricity and oil consumption. This suggests that decreasing energy consumption such as coal, gas, electricity and oil appears to be an effective way to control CO2 emissions but simultaneously will hinder economic growth. Thus suitable policies related to the efficient consumption of energy resources and consumption of renewable sources are required. (C) 2013 Elsevier Ltd. All rights reserved.