The Convenience Yield and the Informational Content of the Oil Futures Price

Recent studies have shown that futures prices do not generally outperform naive no-change forecasts of spot prices, calling into question the usefulness of futures prices for forecasting purposes. However, such usefulness is predicated on the question of whether certain modeling strategies are able to yield more of the information found in futures prices. Applying a forecast-based approach, we study the extent to which alternative ways of modeling futures prices can reveal the extent of the information present in futures prices. Using weekly and monthly data, and futures of maturities of one to four months, we notably examine the out-of-sample predictability of futures prices over various forecast horizons, and in real-time, whereby parameters are updated prior to each sequential forecast. Our results with weekly data are particularly interesting. We find that models allowing for a time-varying convenience yield often produce considerably more precise forecasts over the three forecast horizons considered. Thus, more of the informational content of futures prices is attainable when both the price level and the distance of the latter from spot price are jointly considered, rather than when only the price level is considered. We also document that forecast performances improve with longer date-to-maturity futures, suggesting that the role of the convenience yield is greater when physical oil inventories are held for longer durations. Finally, we show that forecast accuracy is highest at the one year horizon, though the time-varying convenience models have a much higher accuracy than unit-root-based models even over the three and five-year horizons.