Absolutely not.
As Dylan Grice, then at Société Générale pointed out, their expected long-term real rate of return is not appealing. Here's our introduction to 2010's Société Générale's Dylan Grice-"Commodities: ‘Their Expected Long-Run Real Return is 0%’" (please ignore the supercilious "Well duh", I was in my haughty Valley Girl phase, better now):
Well duh.
Commodities are for tradin' not investin'.
Which makes one wonder how CalPERS and the other big institutions got snookered by Goldman Sachs into being "Long-Only Index Investors".
Do you, gentle reader, think for one minute that Goldman's crown jewel, Alaron Trading, just buys and socks the stuff away?
Of course not. Alaron makes directional bets, both long and short, to take advantage of the movement.
To a competent trader, volatility is your friend.
In the case of the grains the darn things are mean-reverting.
If wheat doubles in price, the acreage devoted to wheat goes up and prices come down. The substitution effects at the producer level are predictable if not timable:
better net profit for soybeans than corn? Beans it is boys!
In the metals and in energy the more important substitutions are at the user level. If a utility's cost of a BTU is cheaper when gas-fired, the coal orders slow down.
And over-arching everything is the point that Mr. Grice is making. Human beings are adaptable....
And from the CFA Institute's Enterprising Investor blog, July 8:
If you focus only on returns and covariances over a one-year investment horizon, you may conclude that commodities have no place in an investment portfolio. The efficiency of commodities improves dramatically over longer investment horizons, however, especially when using expected returns and maintaining historical serial dependencies.
We’ll demonstrate how allocations to commodities can change across investment horizon, especially when considering inflation. Our analysis suggests that investment professionals may need to take a more nuanced view of certain investments, especially real assets like commodities, when building portfolios.
This is the third in a series of posts about our CFA Institute Research Foundation paper. First, we demonstrated that serial correlation is present in various asset classes historically. Second, we discussed how the risk of equities can change according to investment horizon.
Historical Inefficiency of Commodities
Real assets such as commodities are often viewed as being inefficient within a larger opportunity set of choices and therefore commonly receive little (or no) allocation in common portfolio optimization routines like mean variance optimization (MVO). The historical inefficiency of commodities is documented in Exhibit 1, which includes the historical annualized returns for US cash, US bonds, US equities, and commodities from 1870 to 2023. The primary returns for US cash, US bonds, and US equities were obtained from the Jordà-Schularick-Taylor (JST) Macrohistory Database from 1872 (the earliest year the complete dataset is available) to 2020 (the last year available). We used the Ibbotson SBBI series for returns thereafter.
The commodity return series uses returns from Bank of Canada Commodity Price Index (BCPI) from 1872 to 1969 and the S&P GSCI Index from 1970 to 2023. The BCPI is a chain Fisher price index of the spot or transaction prices in US dollars of 26 commodities produced in Canada and sold in world markets. The GSCI — the first major investable commodity index — is broad-based and production weighted to represent the global commodity market beta.
We selected the GSCI due to its long history, similar component weights to the BCPI, and the fact that there are several publicly available investment products that can be used to roughly track its performance. These include the iShares exchange traded fund (ETF) GSG, which has an inception date of July 10, 2006. We used the two commodity index proxies primarily because of data availability (e.g., returns going back to 1872) and familiarity. The results from the analysis should be viewed with these limitations in mind.
Exhibit 1. Historical Standard Deviation and Geometric Returns for Asset Classes: 1872-2023.
Source: Jordà-Schularick-Taylor (JST) Macrohistory Database. Bank of Canada. Morningstar Direct. Authors’ calculations.
Commodities appear to be incredibly inefficient when compared to bills, bonds, and equities. For example, commodities have a lower return than bills or bonds, but significantly more risk. Alternatively, commodities have the same approximate annual standard deviation as equities, but the return is approximately 600 basis points (bps) lower. Based entirely on these values, allocations to commodities would be low in most optimization frameworks.
What this perspective ignores, though, is the potential long-term benefits of owning commodities, especially during periods of higher inflation. Exhibit 2 includes information about the average returns for bills, bonds, equities, and commodities, during different inflationary environments....
....MUCH MORE
If interested see also "From Boom to Bust: A Typology of Real Commodity Prices in the Long Run" Plus a Compendium of Dylan Grice at Société Générale, 2009-2012
Here's the Internet Archive with Dylan Grice Full "Cred and Credulity:A collection of Popular Delusions essays from 2009 to 2012".
2008's "Classic Paper: Returns from Commodity Futures" came to a similar first-pass approximation conclusion:
Over the long term, the average annualized excess returns (above the risk-free rate) on futures for individual commodities is approximately zero, and these returns are largely uncorrelated with one another. There is little evidence of long-term return persistence among individual commodity futures.
With a trading strategy of monthly rebalancing of a portfolio of commodities you can grind out abnormal returns, until you rebalance into a string of losers.
