(John Kemp is a Reuters market analyst. The views expressed are
By John Kemp
LONDON Jan 8 Commodities were the worst
performing asset class for the third year running in 2014.
Investors, including some of the world's largest pension
funds, have seen billions of dollars of wealth disappear as a
result of investing in commodity index products over the last
So it is essential to understand what went wrong to help
prevent a similar problem recurring in future.
"Facts and fantasies about commodity futures," first
published in 2004 by Gary Gorton and Geert Rouwenhorst, proved
one of the most influential research papers in 21st century
It provided the intellectual underpinning for the investment
boom in commodity derivatives which followed over the next eight
years until roughly 2012.
Gorton and Rouwenhorst concluded "the risk premium on
commodity futures is essentially the same as equities" and
better than bonds.
"In addition to offering high returns, the historical risk
of an investment in commodity futures has been relatively low"
and "they are an attractive asset class to diversify traditional
portfolios of stocks and bonds."
Yet all of those propositions have come under scrutiny as
returns on commodity index products have disappointed investors
over the last three years and in some cases longer.
Several high-profile investors and commodity index fund
operators have recently closed down their operations citing
returns which failed to match the complexity and risk involved
in running the programmes.
"Facts and fantasies" was based on an analysis of returns
that would have been available to an investor in an
equally-weighted index of commodity futures fully collateralised
by U.S. Treasury bonds between July 1959 and March 2004 (NBER
Working Paper 10595).
"Facts and fantasies," and similar papers written later by
others, played a pivotal role popularising investment in
commodities and making commodity indices respectable for a much
wider group of investors.
Previously, commodity investment was the preserve of
investors and hedge funds with a high appetite for risk and
willingness to endure volatility.
"Facts and fantasies" helped convince even conservative
investors, such as pension funds, that commodity derivatives,
especially indices, were a prudent addition to their portfolios.
Commodity derivatives were not just a directional bet on
boom-bust but an "asset class" that could be a source of
long-term returns across the business cycle.
Initially, the performance of commodity indices was in line
with the historical research, and even exceeded expectations.
Commodity indices soared between early 2002 and July 2008.
Hit-hard when the global financial crisis intensified in
third quarter of 2008, they staged a moderate comeback in 2009,
2010 and 2011. Since then, however, performance has been
consistently disappointing, as the attached charts illustrates
Between June 2004 and June 2014, the compound annual growth
rate (CAGR) for the S&P Goldman Sachs Commodity Index (GSCI) was
-1.8 percent. The Light Energy and Non-Energy versions of the
GSCI performed little better, eking out meagre returns of +1
percent and +2 percent per year respectively.
By Dec 23, however, returns on the GSCI averaged -3.7
percent per year since the middle of 2004, -1.3 percent for the
Light Energy version, and just +1.2 percent for the Non-Energy
Returns have been poor compared with stocks. The S&P 500
equity index achieved total returns of around +7 percent per
year between June 2004 and June 2014, increasing to about +7.9
percent by December 2014.
In practice, commodity derivatives have exhibited all of the
volatility of other asset classes (and often more) but none of
WHAT EXACTLY WENT WRONG?
The most widely invested commodity indices were the two
families known originally as the GSCI and the Dow Jones AIG
index. Both have changed ownership and been rebranded over time
and are now controlled by Dow Jones S&P Indices and Bloombeg
None of the most commonly tracked benchmarks is an exact
replica of the equal-weighted basket of commodity futures
analysed by Gorton and Rouwenhorst between 1959 and 2004.
For all sorts of reasons, not least the small scale of some
futures contracts, it is difficult to exactly replicate the
"Facts and fantasies" type index as an investable index in the
Most index families, but especially the main GSCI, are
heavily weighted towards petroleum futures (crude oil, gasoline
and distillate fuel oil), which tends to limit their
But the fact most commodity indices have produced similarly
disappointing returns since 2004, including variants with a much
lighter weighting towards crude oil and refined fuels, suggests
index composition and the process for rolling maturing contracts
forward on its own cannot explain the poor performance.
There is a tendency in the financial services industry to
celebrate successful products and try to quickly forget the
unsuccessful ones: why dwell on the failures of the past?
VANISHING RISK PREMIUM
The basic explanations for the poor performance of the
indices can be recounted easily enough. Index returns comprise
three components: (1) the spot price of the commodity; (2) the
yield from the Treasury securities used as collateral; and (3)
the roll return from swapping a position in maturing contracts
into longer dated ones.
The problem with spot prices is obvious given the fading of
the so-called "super-cycle" and the decline in prices for a
broad-range of commodities.
But the case for investing in commodities, and treating
commodity derivatives as an asset class, was never supposed to
rely on rising spot prices, a point which "Facts and fantasies"
The problem with the collateral yield is also obvious.
Near-zero official interest rates and low bond yields have
sapped reported commodity index returns.
In practice, many investors will have done better than
benchmark indices because they are invested in a range of bonds,
not just Treasuries. Nonetheless there is no denying that
returns on commodity-derivative related products have been
The real problem has come from the roll yield, or more
precisely the idea that the price of a basket of commodity
futures contracts should tend to rise over time to compensate
investors for assuming price risk from commodity producers.
There are various ways of characterising this type of
return. John Maynard Keynes called it "normal backwardation".
Gorton and Rouwenhorst call it a "risk premium".
Because of this risk premium, Gorton and Rouwenhorst showed
the holder of a fully collateralised basket of commodity futures
would have obtained a much higher rate of return than simply the
spot price of the commodities between 1959 and 2004.
Rather than compensating investors for taking a risk on the
future direction of prices, broad-based commodity indices have
actually charged them for the "privilege".
So what changed between 1959-2004 and 2004-2014? The most
obvious answer is that it was the popularity of commodities as
an asset class, and the influx of new investment as a result,
which transformed the way futures prices behaved.
Prior to 2004, and especially prior to the early 1990s,
commodity derivatives markets were small and illiquid. For much
of the 1959-2004 period, there were no futures markets for
important commodities like crude oil and U.S. natural gas.
In small and illiquid markets there might well have been a
substantial risk premium for investors but it seems to have
vanished once markets became more heavily traded. The 1959-2004
period may not have been very representative of the long-run
returns that investors can actually achieve by investing in
It would be useful to update "Facts and fantasies" to see if
the results are robust for the entire period from 1959 to 2014,
and if estimates for the risk premium are sensitive to the
exclusion of certain sub-periods (eg 1959-1979 or 1994-2014).
In an email, Gorton told Reuters he and Rouwenhorst are
working on an update of the original paper, since ten years have
passed, and plan to release in it in the usual academic way.
In the meantime, the mistake made by investors and the banks
which sold index products was to assume historic returns would
be available in future once commodities moved from a niche
product into the mainstream.
The experience of the past decade strongly suggests the
answer is No.
(Editing by William Hardy)