1. Karstanje, van der Wel, and van Dijk (2015) examine common factors in commodity futures curves over the last 20 years or so, finding that, on average, 62% of the variation in level at the short end of the curve can be explained by a combination of market-wide and sector factors and that 74% of the variation in slope and curvature is captured, on average, by common factors. In addition, it appears that the importance of common factors for levels has increased, possibly as a result of the so-called financialization of commodity markets since the 1990s.
2. For additional background on the posited reasons for expecting a return premium from investing in commodity futures, see Gorton and Rouwenhorst (2006).
3. To define roll yield, we compare the front-month contract to its comparable contract 12 months out the curve to determine whether each commodity is in backwardation or contango. Using a fixed 12-month distance between contracts gives a measure that is more homogeneous across different commodities, eliminates seasonality in prices, and significantly reduces the volatility of roll yields, and therefore portfolio turnover. To assess momentum, we compare the spot price today to the price 12 months ago.
4. Commodity prices tend to exhibit short-term momentum (prices that have gone up or down over the recent past continue the trend in the near future) in response to shocks in supply or demand. For example, the 2015 plunge in oil prices was largely a result of oversupply from fracking by US shale producers. In the United States, oil production rose from a previous low of 5 million barrels a day (mb/d) in 2008 to over 9 mb/d in 2015. According to the International Energy Agency, the implied global oil market surplus is expected to narrow significantly to 0.2 mb/d in fourth quarter 2016. Because shale producers have sunk costs in wells already drilled and will seek to make good on their investments, the delay in returning to a supply/demand balance will take time, creating persistence in the price of oil.
5. The three major sectors include energy, agriculture and livestock, and metals. Each sector takes one-third of the total index weight. The base weights for the commodities in each sector are determined by a liquidity measure proxy of the five-year moving average of dollar-volume traded.
6. The eligible contract may have open interest of at least 5% of the open interest of the nearby liquid contract. Contract eligibility is subject to annual review and adjustment as necessary for capacity. The contract roll takes place over five days beginning on the first Dow Jones RAFI Commodity Index business day of each month (i.e., on each roll day 20% of the incumbent contract is replaced with 20% of the new contract.)
Gorton, Gary, and K. Geert Rouwenhorst. 2006. “Facts and Fantasies about Commodity Futures.” Financial Analysts Journal, vol. 62, no. 2 (March/April):47–68.
Greer, Robert J. 2006. “Commodity Indexes for Real Return,” Chapter 5 in The Handbook of Inflation Hedging Investments, edited by Robert J. Greer. New York, NY: McGraw-Hill.
Karstanje, Dennis, Michel van der Wel, and Dick van Dijk. 2015. “Common Factors in Commodity Futures Curves.” University of Pennsylvania working paper (February 15).
Oldani, Chiara. 2008. Governing Global Derivatives: Challenges and Risks. New York, NY: Taylor & Francis Group.