LONDON: For the last two decades, the diversified approach to commodity investing encapsulated by the Dow Jones-UBS Commodity Index has outperformed the more production-weighted system employed by the Standard and Poor’s Goldman Sachs Commodity Index.
While the GSCI remains popular with asset managers because it is easier to beat, institutional investors concerned about absolute rather than relative performance, should insist on using the DJUBS rather than the GSCI as a benchmark.
The GSCI and DJUBS index families dominate the commodity indexing sector. Between them they account for almost all the institutional money devoted to indexing strategies, and they are the most widely employed benchmarks for long-only fund managers who use more active strategies to beat the market.
The two have long been rivals, with the GSCI more successful by a wide margin. In terms of the amount of money tracking them, institutional investors have favored the GSCI and its variants by a margin of around 2:1.
Ironically, the two index families are now offered by the same company, after McGraw Hill (which owns the Standard and Poor’s GSCI) and CME Group (which owns the DJUBS) formed a joint venture in July 2012. Ownership is split between McGraw Hill (73 percent), CME (24.4 percent) and Dow Jones (2.6 percent).
But in terms of performance, most investors and fund managers would achieve superior returns if they ditched the GSCI and its various derivatives and started tracking the DJUBS.
The DJUBS index has consistently outperformed the GSCI and its most popular variants over almost every time horizon in the last 20 years.
Between 1992 and 2012, the GSCI total return index achieved a compound annual return of 4.2 percent. The GSCI Light Energy index produced a compound return of just 3.1 percent. In contrast, the DJUBS achieved a compound return of 5.7 percent per year.
None of the indices has managed to match the returns on equities. Even after a decade in the doldrums, the broad-based S&P 500 equity index is still showing compound returns of 8.3 percent since 1992.
But there is clear evidence that investors tracking the DJUBS have done much better than their rivals who follow the GSCI.
The contrasting fortunes of PIMCO and the California Public Employees’ Retirement System (CalPERS) highlight the performance gap.
PIMCO’s giant Commodity Real Return Strategy Fund, which has more than $ 21 billion in assets, and is by far the largest mutual fund investing in commodities, uses the DJUBS as a benchmark. In the past five years the fund has eked out a small profit.
Its active managers have managed to generate just enough alpha to offset a small decline in the underlying index.
In contrast, CalPERS, with more than $ 3 billion invested in commodities, has lost money over the same period. Even its active management strategies have not been able to offset the heavy drag on performance from tracking the deeply underperforming GSCI.
The performance differences stem from the way in which the two weight exposure to different commodities, as well as some more minor differences in the way they roll positions forward to maintain a constant exposure as futures contracts expire.
The original paper on “Facts and Fantasies about Commodity Futures,” which did so much to popularize commodity investing among pension funds and other institutions, examined returns on an equal-weighted basket of commodity derivatives. In practice, almost no one has employed an equal weighted index, which would give an investor as much exposure to cocoa and sugar as to crude oil and natural gas.
According to its handbook, the GSCI is a “production-weighted index designed to reflect the relative significance of each of the constituent commodities to the world economy.”
In the first instance, weightings are determined by “world production quantities” for each commodity over the previous five years, with some subsequent adjustments.
The result is an index dominated by energy, with a much smaller exposure to industrial metals, agricultural produce and precious metals.
In 2012, the GSCI had an allocation of 67.5 percent to energy, and just 16.8 percent to agriculture, 8 percent to industrial metals and 3.2 percent to precious metals like platinum and gold.
The GSCI’s overall performance is driven by its over-whelming exposure to crude oil (both US light sweet crude futures and Brent) as well as refined products (which trade off crude).
In the 1990s and early 2000s, the heavy exposure to crude was a big benefit. Crude futures typically traded in backwardation. Investors were therefore usually selling a more expensive expiring contract to buy a cheaper one further forward, repeating the process over and over again, and riding the futures curve upward.
But from 2004, the structure of forward prices has flipped into contango, with prices for future delivery more expensive than the spot market.
The change has been most pronounced in US light sweet crude contracts. As a result the running yield on WTI-linked contracts has been negative most of the time and has acted as a significant drag on overall returns, given the high weighting (generally about 25 percent) WTI has in the index.
The DJUBS family has avoided some of these pitfalls by adopting a more diversified weighting, avoiding the excessive exposure to crude oil and WTI in particular that has bedeviled the GSCI. It has capped the maximum weighting for related group of commodities.
No single commodity may constitute over 15 percent of the total index. No single commodity together with its derivatives (such as crude oil, heating and gasoline) may account for more than 25 percent. And no related group of commodities (such as energy) may account for more than 33 percent, according to the handbook.
Capped weights are much closer to the spirit of the original research. And they have avoided the problems the GSCI has experienced with its over-exposure to the contango in crude oil markets.
Some of these features have been copied into the GSCI family.
Many investors now prefer to track a modified variant of the GSCI (such as the Reduced Energy, Light Energy or Ultra-Light Energy indices) which reduces overall exposure to crude and products. Even so, the DJUBS has generally offered superior performance over long periods of time.
Sell-side index operators are currently promoting a number of changes to commodity indexing methodology to pension funds disillusioned with the poor performance so far.
But the single biggest improvement many investors could make is to abandon the production-weighted approach encapsulated in the GSCI and embrace the more diversified approach epitomized by the DJUBS system.
The persistent popularity of the GSCI is something of a mystery. Its overexposure to crude oil and resulting underperformance has been well understood for some years.
One reason is that the GSCI’s production weighting is intuitively plausible. Most clients want more exposure to “important” commodities like crude oil than second-tier ones like cocoa. But the focus on economic significance rather than returns is a costly mistake.
Another less reputable reason is that the GSCI is easy to beat. The sources of its underperformance are well known, making it relatively easy for managers to outperform simply by underweighting WTI compared with the index and overweighting Brent.
Asset managers are generally rewarded by their ability to match or exceed the performance of the target index.
The GSCI is relatively easy to outperform, making everyone look good.
But investors should demand more than relative outperformance against a poorly performing index. For most institutions in the commodity sector, the first step to better returns should be to switch from the GSCI to the DJUBS as a benchmark.
Since the two are now owned by the same joint venture, it should be straightforward. No other change would guarantee such a simple uplift in returns.
— John Kemp is a Reuters market
analyst. The views expressed are his own.
Commodity investors should demand benchmark change
