This study attempts to examine empirically dynamic causal relationships between aggregate output, energy consumption, exports, capital and labour in the case of Turkey using the time series data for the period 1968-2008. This research tests the interrelationships between the variables using the bounds testing to cointegration procedure. The bounds test results indicate that there exists a long-run relationship between the variables in which the dependent variable is aggregate output. Within this study, three competing sets of hypotheses regarding the relationship between aggregate output, exports and energy consumption are tested. An augmented form of Granger causality analysis is conducted amongst the variables. In the long-run, causality runs interactively through the error correction term from labour, capital, exports and energy consumption to aggregate output. In the short-run, two important bilateral causalities were identified: between energy consumption and aggregate output, between exports and aggregate output. The short-run causality testing reveals further the existence of a unilateral causality running from exports to energy consumption too. The long-run relationship of aggregate output, energy consumption, exports, capital and labour equation is also checked for the parameter stability. The results also provide some important policy recommendations. (C) 2011 Elsevier Ltd. All rights reserved.