If you simply give attention to returns and covariances over a one-year investment horizon, you would possibly conclude that commodities haven’t any place in an investment portfolio. However, the efficiency of commodities improves dramatically over longer investment horizons, especially when expected returns are used and historical serial dependencies are maintained.
We show you ways commodity allocation can change over the investment horizon, especially when taking inflation into consideration. Our evaluation suggests that investment professionals have to take a more nuanced view of certain investments, especially real assets equivalent to commodities, when constructing portfolios.
Historical inefficiency of raw materials
Real assets equivalent to commodities are sometimes considered inefficient inside a wider range of decisions and subsequently typically receive little (or no) allocation in common portfolio optimization routines equivalent to mean-variance optimization (MVO). The historical inefficiency of commodities is documented in Figure 1, which incorporates historical annualized returns for U.S. money, U.S. bonds, U.S. stocks, and commodities from 1870 to 2023. The primary returns for U.S. money, U.S. bonds, and U.S. stocks were taken from the Jordà -Schularick-Taylor (JST) Macrohistory Database from 1872 (the earliest yr for which the total dataset is out there) to 2020 (the newest available yr). For returns thereafter, we used the Ibbotson SBBI series.
The commodity return series uses returns from the 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 sequence index from Fisher Price that calculates the spot or transaction prices in U.S. dollars of 26 commodities produced in Canada and sold on world markets. The GSCI – the primary major investable commodity index – is broad-based and production-weighted to represent the beta of the worldwide commodity market.
We selected the GSCI since it has an extended history, has similar component weights to the BCPI, and since there are several publicly available investment products that may roughly replicate its performance. These include the iShares Exchange Traded Fund (ETF) GSGwhose introduction date is July 10, 2006. We used the 2 commodity index proxies primarily resulting from data availability (e.g., returns going back to 1872) and familiarity. The results of the evaluation must be considered with these limitations in mind.
Figure 1. Historical standard deviation and geometric returns for asset classes: 1872–2023.
Commodities look like incredibly inefficient in comparison with bonds, debentures and stocks. For example, commodities have a lower return than bonds or debentures, but significantly higher risk. Alternatively, commodities have concerning the same annual standard deviation as stocks, however the return is about 600 basis points (bps) lower. Based on these values, commodity allocations could be small in most optimization frameworks.
What this angle ignores, nonetheless, are the potential long-term advantages of owning commodities, especially in periods of upper inflation. Figure 2 provides information on the common returns of bonds, securities, stocks and commodities in several inflation environments.
Figure 2. Average returns for asset classes in several inflation environments: 1872–2023.
.
We see that while commodities yield low returns when inflation is low, they perform significantly higher when inflation is high.
The correlation of commodities to inflation increases significantly over longer investment horizons, from about 0.2 for one-year periods to 0.6 for ten-year periods. In contrast, the correlation of stocks to inflation is just about -0.1 for one-year periods and about 0.2 for ten-year periods. In other words, specializing in the longer-term advantages of owning commodities and explicitly accounting for inflation can dramatically change the perceived efficiency of a portfolio optimization routine.
Allocation to raw materials
While inflation could be explicitly accounted for in certain sorts of optimizations, equivalent to “surplus” or liability-based optimizations, a possible problem with these models is that price changes in goods or services don’t necessarily move in lockstep with changes in financial markets. There may very well be lagged effects. For example, while financial markets can experience sudden changes in value, inflation has more of a latent effect: changes could be delayed and only change into apparent after years. Focusing on the correlation (or covariance) of inflation with a specific asset class, equivalent to equities, over one-year periods (e.g. calendar years) can hide potential long-term advantages.
To determine how the optimal commodity allocation would have modified depending on the investment horizon, we conducted a series of portfolio optimizations for investment horizons starting from one to 10 years in one-year increments. The optimal allocation was determined using a relentless relative risk aversion (CRRA), which risk-adjusts the cumulative asset growth over a given investment horizon.
The optimal allocations, which correspond to equity allocations from 5% to 100% in 5% increments, were determined based on the targeted risk aversion levels. We included 4 asset classes within the portfolio optimizations: bills, bonds, equities and commodities. Figure 3 comprises the optimal commodity allocations for every of the scenarios considered.
Figure 3. Optimal commodity allocation by asset definition, equity risk goal and investment period: 1872–2023.
Allocations to commodities remained near zero for virtually all equity allocation targets when wealth was defined in nominal returns (Panel A). In contrast, when wealth was defined in real terms (i.e., including inflation), allocations were relatively significant over longer investment horizons (Panel B). This was particularly true for investors looking for moderately conservative portfolios (e.g., ~40% equity allocation), where the optimal allocation to commodities could be around 20%. In other words, the perceived historical advantages of allocating to commodities varied considerably across asset definitions (nominal versus real) and assumed investment horizons (e.g., from one yr to 10 years).
Forward-looking expectations for commodity returns usually are not as bleak as historical long-term averages. For example, while commodities have historically underperformed equities by about 600 basis points on a risk-adjusted basis, the expected underperformance is closer to 200 basis points based on each PGIM Quantitative Solution’s fourth-quarter 2023 capital markets assumptions and the Horizon Actuarial Survey of 42 investment managers (with a give attention to 10-year returns).
We re-ran the portfolio optimizations using the identical historical time series but recentered the historical returns to match the expected returns for money, bonds, stocks, commodities, and inflation (3.6%, 5.4%, 8.4%, 6.1%, and a couple of.5%, respectively) and standard deviations (2.0%, 5.6%, 15.3%, 14.7%, and a couple of.0%, respectively). The optimal commodity allocations increased significantly no matter whether wealth is defined in nominal or real terms, as shown in Figure 4.
Figure 4. Optimal commodity allocation in keeping with asset definition, equity risk goal and investment period: Expected returns.
The optimal allocation to commodities is around 10% when the main focus is on nominal assets, whatever the investor’s equity risk objective or investment horizon, and around 20% or more when the main focus is on real assets. These results suggest that the potential advantages of an allocation to commodities are significantly higher when considering expected returns in comparison with historical returns.
Look beyond one-year returns and covariances
When considering the chance of an asset, it will be important to acknowledge that it isn’t at all times possible to capture its potential advantages by focusing only on returns and covariances over a one-year investment horizon. Asset classes equivalent to commodities have historically offered notable diversification advantages for longer-term investors concerned about inflation. It is essential for investment professionals to concentrate on these effects and the potential impact on optimal portfolios.
