This note presents a structural vector autoregressive (SVAR) model of the global oil market. The model identifies four types of shocks with different economic interpretations: oil supply shocks, oil-market-specific demand shocks, storage demand shocks and shocks to global economic growth. The historical decomposition of oil price fluctuations suggests that oil supply shocks were the dominant force during the 2014–15 oil price decline. Several examples illustrate the model’s usefulness for conditional forecasts of oil market variables under different scenarios for global GDP growth and oil consumption.