Forecasts of the real price of oil revisited: Do they beat the random walk?

In macroeconomic forecasting, the real price of oil is traditionally computed as the monthly average price of oil deflated by the price index. Consequently, the no-change forecast used to benchmark forecasts of the real price of crude oil is a monthly average price. We demonstrate that an alternative no-change forecast which reflects the random walk forecast from daily oil prices - the end-of-month price - is significantly more accurate in predicting the real price of oil up to one year ahead. We find that at the one-step-ahead prediction, all existing forecasts that outperform the monthly average no-change forecast perform worse than the end-of-month no-change forecast. The results call into question the usefulness of existing forecasting approaches for the real price of crude oil relative to naive forecasts.