LONDON Roll returns rather than spot price movements have been a much more significant source of profits and losses for commodity investors over almost any time horizon.
Yet most investors still formulate their strategy in terms of outright price moves rather than the spreads, missing out on the most important source of long-term performance.
Strategies with the most explicit focus on roll returns, such as the SummerHaven Dynamic Commodity Index offered by U.S. Commodity Funds, which had $432 million invested at the end of June, account for a tiny fraction of the $50 billion invested in all commodity-focused mutual funds or the nearly $200 billion invested in indexing strategies as a whole.
Over the 10-30 year horizons that most interest pension funds, roll returns may be the only sustainable source of returns from commodities.
But given the theoretical and empirical importance of roll returns, it is surprising how few index products have been designed specifically to maximize them.
In principle, it should be possible to offer investors a product which targeted roll returns specifically and is neutral to spot prices, by investing exclusively in spreads, with the aim of capturing scarcity premiums in the market while avoiding exposure to cyclical variations in commodity prices.
Most of physical trading houses already claim to operate this way (with positions in the spreads rather than the outright directional plays). But as investors question whether the commodity super-cycle has peaked or plateaued, it would make sense for other investors to contemplate the same strategy.
By now most investors are aware that it is not possible to achieve returns based on the spot price of WTI or a basket of commodities.
But few investors appreciate just how significant roll returns are. In practice, rolls dominate all other sources of performance over the medium and long term.
Charts 1 and 2 show spot prices and actual returns to investors from a long position in WTI futures or a diversified basket such as the Standard and Poor's Goldman Sachs Light Energy Index. In both cases, once roll profits and losses are taken into account, actual returns diverge sharply from the spot price.
Researchers have always emphasized the central importance of rolls. In the 1920s and 1930s, John Maynard Keynes highlighted the profits investors could reap from the "normal backwardation" in commodity futures.
More recently, Gary Gorton and Geert Rouwenhorst explored the "risk premium" commodity investors would have received between 1959 and 2004 from an equal-weighted long position in a basket of commodity futures, largely irrespective of spot prices, in their famous paper on "Facts and Fantasies in Commodity Futures".
Curiously, though, most commodity indices have been constructed with little reference to roll returns. The most popular Standard and Poor's Goldman Sachs Commodity Index and Dow Jones-UBS Commodity Index were both constructed on a production-weighted basis with modifications for liquidity and diversification.
As a result of the disappointing investment performance of both indices, especially the GSCI, since 2005, many banks are now marketing second and third-generation variants which seek to minimize roll losses and maximize roll gains.
But in their marketing materials, and method for selecting futures for inclusion, most indices continue to emphasize the potential for spot price appreciation. Roll returns are a distinctly second-order consideration.
Only the SummerHaven Dynamic Commodity Index (SDCI) focuses on selecting commodity futures primarily on their potential for generating roll returns.
According to its promoters, the SDCI selects 14 futures contracts with an equal-weighting from a potential universe of 27 eligible commodities. In the first stage, the seven commodities exhibiting the greatest backwardation (positive roll return) or least contango (negative roll return) are selected.
From the remaining 20 commodities, seven more are selected based on which show the greatest price momentum over the previous twelve months.
Roll returns are therefore the primary selection criteria for the SDCI.
"The SDCI is based on the notion that commodities with low inventories will tend to outperform commodities with high inventories, and that priced-based measures, such as futures basis and price momentum, can be used to help assess the current state of commodity inventories" SummerHaven argues.
In backtests, the SDCI has outperformed both the basic GSCI and DJUBS indices over the last decade.
Between August 2002 and September 2012, the SDCI provided average annual returns to investors of 18 percent, compared with around 7 percent for the two big benchmarks, according to SummerHaven (here).
Backtests should be handled with care. There is a danger of retrospectively designing the index to maximize past performance, employing a hindsight that is not available to investors.
There is also the danger past returns will be eroded as more investors try to exploit them. Index products which focus on roll returns (either as a primary or secondary consideration) have struggled to outperform in 2012.
Nonetheless, given the dominant part played by roll returns in investor returns, they will have to play an increasing role in index design and investment strategy in future. Maximizing roll returns, rather than spot price appreciation, should be the central preoccupation for investors.
(John Kemp is a Reuters market analyst. The views expressed are his own. To get his real-time views on the market, please join the Global Oil Forum)
(Editing by William Hardy)