Energy Consumption-GDP Relationship in Iran: A Markov Switching Approach

The causal link between energy consumption and output is one of the most debated subjects in economics. A large literature has studied this link; however، the results are mixed. The sample period plays a crucial role and the results of causality tests appear to be sensitive to that. To get more consistent results we use a new method which allows for change in the casual links over the sample period. This method is based on a VAR model with time varying parameters، and these variations are governed by a Markov chain. In this paper، we use a Markov-Switching VAR model with two different relationships among the variables to examine the casual link between output and energy consumption in Iran، using annual data from 1346 to 1386. The results show that 1) GDP is a Granger cause of energy consumption in both cases، 2) this casual link is stronger in the 2nd case، 3) the causality is not constant over the sample period and it changes in 1368، and 4) energy consumption is a Granger cause of output only in the first case.