As inflation concerns rise among investors, many are reconsidering their allocation to commodities, which have historically acted as an effective inflation hedge. In this analysis, we raise an additional question: Which is the more effective way to gain a commodity exposure: via the futures market or via the relevant equity sectors?
In this article, we take the energy, industrial metals, gold, grains and livestock sectors, and for each one we regress the monthly excess returns over three-month Treasury Bills of a tradable equity basket (“Equity”) and a tradable futures or forwards basket (“Futures”) against the relevant spot commodity (“Commodity”) and S&P 500 Index (“Market”), to test their sensitivity to each asset class. We do this for the full available historical data sample, as well as splitting the full period into three equal parts.
In each case, we find that commodity futures have tended to deliver a much more commodity-like exposure than commodity equities. We discuss the reasons for this in our conclusion.
Energy
Over the full period of 1989 – 2021, our energy equity basket exhibits more than twice the level of beta against the S&P 500 than against energy commodity spot returns. We can see why that is by looking at some aspects of the historical returns. For example, during the First Gulf War, while energy prices rapidly rallied by 80%, general weakness in equity markets meant that energy equities actually lost value.
Moreover, even after factoring in both equity market and commodity returns, the full-period R-squared coefficient for our energy equity basket, which measures the extent to which the variations in one set of data are determined by those in other, was still only 0.63, indicating a substantial amount of unexplained variation in the equity returns. This is likely due to uncompensated, company-specific risks.
By contrast, the futures basket is virtually uncorrelated with the S&P 500 and almost perfectly reflective of spot commodity prices. These outcomes have been robust across time periods.
Table 1: Regression model for energy investment options
| Sector | Investment | Period | Market Beta | Commodtity Beta | R Squared |
|---|---|---|---|---|---|
| Energy | Equity | Sep 1989 / Jun 2021 | 0.77 | 0.31 | 0.63 |
| Energy | Futures | Sep 1989 / Jun 2021 | -0.03 | 1.00 | 0.95 |
| Energy | Equity | Sep 1989 / Mar 2000 | 0.73 | 0.23 | 0.55 |
| Energy | Futures | Sep 1989 / Mar 2000 | -0.02 | 0.99 | 0.94 |
| Energy | Equity | Apr 2000 / Oct 2010 | 0.63 | 0.29 | 0.59 |
| Energy | Futures | Apr 2000 / Oct 2010 | 0.00 | 1.01 | 0.96 |
| Energy | Equity | Nov 2010 / Jun 2021 | 1.03 | 0.35 | 0.73 |
| Energy | Futures | Nov 2010 / Jun 2021 | -0.17 | 1.02 | 0.95 |
Source: Bloomberg. Equity, futures and spot returns are modeled by the S&P 500 Energy Index (S5ENRS), the S&P GSCI Energy Total Return Index (SPGSENTR), and the S&P GSCI Spot Energy Index (SPGSEN), respectively. Overall equity market returns are modeled by the S&P 500 Index (SPX). Equity indices used are total returns series including dividends.
Industrial Metals
The story has been similar with industrial metals.
The equity basket exhibits S&P 500 beta ranging from 0.90 to 1.11, whereas beta against industrial metals spot returns averaged 0.74 over the full period. Only in the last period were the exposure levels similar. And again, equity market and spot commodity risk leave a lot of variation unexplained in our basket of industrial metals equities: R-squared coefficients ranging from 0.42 to 0.67 suggest that choosing the wrong companies could lead to poor results—perhaps unsurprisingly in a sector facing particular scrutiny on environmental, social and governance (ESG) issues.
Here, too, commodity futures were virtually uncorrelated with the S&P 500, but almost perfectly correlated with spot industrial metal prices.
Table 2: Regression model for industrial metals investment options
| Sector | Investment | Period | Market Beta | Commodtity Beta | R Squared |
|---|---|---|---|---|---|
| Industrial Metals | Equity | Dec 1999 / Jun 2021 | 0.99 | 0.74 | 0.57 |
| Industrial Metals | Futures | Dec 1999 / Jun 2021 | 0.00 | 1.00 | 1.00 |
| Industrial Metals | Equity | Dec 1999 / Jan 2007 | 0.90 | 0.62 | 0.42 |
| Industrial Metals | Futures | Dec 1999 / Jan 2007 | 0.02 | 1.00 | 0.99 |
| Industrial Metals | Equity | Feb 2007 / Mar 2014 | 1.08 | 0.61 | 0.67 |
| Industrial Metals | Futures | Feb 2007 / Mar 2014 | -0.01 | 1.00 | 1.00 |
| Industrial Metals | Equity | Apr 2014 / Jun 2021 | 1.11 | 1.10 | 0.64 |
| Industrial Metals | Futures | Apr 2014 / Jun 2021 | 0.00 | 0.99 | 1.00 |
Source: Bloomberg. Equity, futures and spot returns are modeled by the S&P Metals & Mining Select Industry Total Return Index (SPSIMMTR), the S&P GSCI Industrial metals Total Return Index (SPGSINTR), and the S&P GSCI Spot Industrial Metals Index (SPGSIN), respectively. Overall equity market returns are modeled by the S&P 500 Index (SPX). Equity indices used are total returns series including dividends.
Gold
When we turn to gold, we see very different results.
Gold miners still exhibit meaningful idiosyncratic risk: their R-squared was 0.67 over the full period from 1993 to 2021. Unlike energy and industrial metals equity, however, they have had relatively low equity market beta (0.47, on average). This is largely due to gold acting rather like a bond, tending to rise in price as real interest rates decline; and to gold miners delivering substantially leveraged beta to gold spot returns: between 1993 and 2021, when gold went up 10%, gold miners went up 18.8%, on average. On the face of it, gold equities appear to deliver leveraged exposure to a commodity that is an excellent diversifier against the broad equity market.
Table 3: Regression model for gold investment options
| Sector | Investment | Period | Market Beta | Commodtity Beta | R Squared |
|---|---|---|---|---|---|
| Gold | Equity | Sep 1993 / May 2021 | 0.47 | 1.88 | 0.67 |
| Gold | Futures | Sep 1993 / May 2021 | 0.00 | 1.00 | 0.99 |
| Gold | Equity | Sep 1993 / Nov 2002 | 0.59 | 2.56 | 0.66 |
| Gold | Futures | Sep 1993 / Nov 2002 | 0.00 | 1.01 | 1.00 |
| Gold | Equity | Dec 2002 / Feb 2012 | 0.50 | 1.61 | 0.72 |
| Gold | Futures | Dec 2002 / Feb 2012 | 0.01 | 1.02 | 0.99 |
| Gold | Equity | Mar 2012 / May 2021 | 0.34 | 2.03 | 0.71 |
| Gold | Futures | Mar 2012 / May 2021 | -0.02 | 0.98 | 0.99 |
Source: Bloomberg. Equity, futures and spot returns are modeled by the NYSE Arca Gold Miners Index (GDM), the S&P GSCI Gold Total Return Index (SPGSGCTR), and the XAU Currency, respectively. Overall equity market returns are modeled by the S&P 500 Index (SPX). Equity indices used are total returns series including dividends.
A look at compound performance over time reveals what is hidden by these correlations, however. Neither our basket of gold equities nor the individual stock of one of the world’s largest gold miners has kept pace with gold itself. They tended to lose a lot more when the gold price weakened and gain a lot less when it strengthened. And over the full period, in months when the S&P 500 was down more than 5%, the average return of gold was 1% while the average return of gold equities was -2.3%.
Chart 1: Performance and diversification efficacies of gold equities and direct gold investment
Cumulative Log Returns
Source: Bloomberg. Gold Miners, Gold Direct and Barrick returns are modeled by the NYSE Arca Gold Miners Index (GDM), the S&P GSCI Gold Total Return Index (SPGSGCTR), and Barrick Gold Corporation (GOLD), respectively. Barrick is Equity indices used are total returns series including dividends .
In other words, the leverage generated by gold miners is real—but it has tended to be expressed on the downside rather than the upside. Investors might consider the more symmetrical and efficient leverage that can be gained through gold futures instead.
Grains and Livestock
It is difficult to find pure grains or livestock plays in the equity market. The major players are either privately held or generate diverse earnings streams that weaken their commodity exposure. To overcome this problem as best we can, we modeled grains equities with a 50/50 blend of Archer-Daniels-Midland and Bunge, and for livestock we simply used Tyson Foods—firms we regard as large enough and with sufficient track records for reliable statistical analysis.
Once again, we find much higher beta against equity markets than the relevant spot commodity prices, and during the 1980 to 1994 period, beta against commodities was even negative. This is somewhat expected, as even the firms we have selected are able to add value in various ways beyond simply growing or rearing their commodities and then selling them; they can also buy them and process them, which means their revenue streams may be negatively affected by the prices of the crops or herds they have bought and the demand cycle for their processed products. Furthermore, their equity market betas exhibit wide variation in the different periods and their R-squared coefficients are very low, reflecting the high level of idiosyncratic risk in these concentrated equity baskets.
Table 4: Regression model for grains and livestock investment options
Grains:
| Sector | Investment | Period | Market Beta | Commodtity Beta | R Squared |
|---|---|---|---|---|---|
| Grains | Equity | Jul 1980 / Jun 2021 | 0.82 | 0.09 | 0.25 |
| Grains | Futures | Jul 1980 / Jun 2021 | 0.02 | 0.93 | 0.93 |
| Grains | Equity | Jul 1980 / Feb 1994 | 1.24 | -0.09 | 0.44 |
| Grains | Futures | Jul 1980 / Feb 1994 | -0.01 | 0.89 | 0.88 |
| Grains | Equity | Mar 1994 / Oct 2007 | 0.36 | 0.09 | 0.05 |
| Grains | Futures | Mar 1994 / Oct 2007 | 0.06 | 0.91 | 0.91 |
| Grains | Equity | Nov 2007 / Jun 2021 | 0.78 | 0.18 | 0.32 |
| Grains | Futures | Nov 2007 / Jun 2021 | 0.02 | 0.97 | 0.97 |
Livestock:
| Sector | Investment | Period | Market Beta | Commodtity Beta | R Squared |
|---|---|---|---|---|---|
| Livestock | Equity | Dec 1984 / Jun 2021 | 0.84 | 0.05 | 0.16 |
| Livestock | Futures | Dec 1984 / Jun 2021 | 0.02 | 0.80 | 0.78 |
| Livestock | Equity | Dec 1984 / Jan 1997 | 0.95 | -0.28 | 0.19 |
| Livestock | Futures | Dec 1984 / Jan 1997 | 0.06 | 0.82 | 0.76 |
| Livestock | Equity | Feb 1997 / Mar 2009 | 0.85 | 0.31 | 0.17 |
| Livestock | Futures | Feb 1997 / Mar 2009 | -0.01 | 0.82 | 0.80 |
| Livestock | Equity | Apr 2009 / Jun 2021 | 0.65 | 0.08 | 0.12 |
| Livestock | Futures | Apr 2009 / Jun 2021 | 0.00 | 0.80 | 0.85 |
Source: Bloomberg. In the case of grains equity, futures and spot returns are modeled by a 50/50 monthly rebalanced portfolio of Archer-Daniels-Midland Co (ADM U.S. Equity) and Bunge Limited (BG U.S. Equity), the S&P GSCI Grains Total Return Index (SPGSGRTR), and the S&P GSCI Grains Spot Index (SPGSGR), respectively. In the case of livestock equity, futures and spot returns are modeled by Tyson Foods (TSN U.S. Equity), the S&P GSCI Livestock Total Return Index (SPGSLVTR), and the S&P GSCI Livestock Spot Index (SPGSLV), respectively. Overall equity market returns are modeled by S&P 500 Index (SPX). Equity indices used are total returns series including dividends.
Conclusion
Past performance suggests that commodity futures have been a more efficient way to create commodities exposure than commodity equities. We think there are fundamental reasons why this has been the case.
- Equity valuations reduce sensitivity to commodity price increases
Stock prices are primarily determined by discounted expected cash flows stretching far into the future, which tend to smooth out short- to medium-term commodity price spikes—not least because, when commodity prices rise, producers are expected to increase supply, and vice versa.
- Firms’ hedging positions reduce spot commodity exposures
Many commodity producers either sell a portion of their production through long-term fixed price contracts or engage in programs designed to hedge some of their exposure to volatile price movements. This may limit the benefit from rising prices and, moreover, producers tend to hedge more when their revenues have been declining, which is precisely the time when the industry in general is likely to be cutting production to support a price rebound.
- Inability to take advantage of supply shocks
During a demand shock, commodity producers may be in a position to respond with tactical leverage, temporarily stretching their production capacity to take advantage of rising prices. A supply shock, however, is by definition symptomatic of limited, compromised or insufficient production capacity—commodity prices will be rising, but producers will be unable to take advantage.
- Commodity firms are generally negatively impacted by a weakening dollar
Finally, major firms operate in various geographies, which means they incur local-currency wage and other operating costs while selling their commodities in U.S. dollars. Because they are priced in dollars, commodities tend to perform well when the dollar is weakening. But a weaker dollar means lower dollar revenue and higher local-currency costs for commodity producers—and therefore lower profits. For this reason, commodity equities tend to perform better while the dollar is strengthening, but in those conditions, investors are unlikely to need a hedge against higher commodity prices and inflation.
While the historical data comes down heavily in favor of commodity futures, some key commodity exposures—such as uranium, lithium or cobalt, which have no liquid derivative markets—can be attained only through equities. Investors may wish to consider isolating the commodity exposures by adding short positions in the broad equity market. Similarly, those who prefer equities might consider combining a broad equity market allocation with a commodity futures allocation: we believe that this is likely to generate similar or better commodity exposure than an allocation to commodity equities while consuming less of the investor’s equity bucket. This approach may also reduce exposure to ESG or idiosyncratic risks.