The paper analyses the importance of supply versus demand shocks on the global oil market from 1974 to 2017, using a parsimonious structural vector autoregressive moving average (SVARMA) model. The superior out-of-sample forecasting performance of the reduced form VARMA compared to VAR alternatives advocates the suitability of this framework. We specifically account for the changes in the oil market over three distinctive sub-periods — pre-moderation, great moderation and post-moderation periods, to provide a means of identifying the changing nature of shock transmission mechanism across times. Our findings shed some light on the effects of supply versus demand related oil shocks under different economic environment. Oil supply shocks explain large fraction of the movements in the global oil market in the pre- and post-moderation periods. The importance of global activity shock on oil price movements is obvious during the 2003-2008 boom period. The oil specific shock has an interesting transmission path on the global economic activity, where the global activity responded positively during the global economic expansion and negatively during economic contraction, emphasising the speculative nature of the oil shock.