This paper explores the causal relationship between energy, emissions and income in Canada for the period 1960- 2005. This study explores these relationships using the Toda Yamamoto approach in a multivariate framework including labour and capital as auxiliary variables. We also test the hypothesis of indirect or multi-horizon Granger non-causality between these variables, since causal effects may occur more than one-period-ahead, as is assumed by the standard Granger non-causality test. We find that there is bi-directional direct causality between income and energy use in Canada, and no other channels of causality between the three variables. However, indirect Granger non-causality testing shows that there is bi-directional causality between all variables in the system. This result is contrary to other results in the literature, and has different implications for energy and environmental policy.