Definitions within investments are a key detail and can lead to substantial differences in portfolio composition and returns. If the definition is too narrow, it can exclude target exposures; if it’s too broad, it can include unintended bets. Recent focus on public equity real assets has been intensified as investors look for liquidity, lower volatility and diversification in addition to their standard “alts” allocations.
The digital world requires a physical backbone.
The drive to invest in the AI boom naturally has led some investors to focus not only on the technology but also on the facilities and power required to support this growth.
Infrastructure may have been a relatively boring component within portfolio allocations as many investors saw the need for inflation protection, income generation and steady growth. When combined with Real Estate, Energy and Natural Resources, Infrastructure forms an exposure to a sleeve of liquid Real Asset equities that can be included alongside traditional equities and fixed income. Recently, these allocations have been increasing past historical averages, and some investors are re-examining the investment choices they made initially, presumably when investment funds that were available weren’t plentiful.
Examining the more popular exchange-traded funds (ETFs) that look to target global, listed infrastructure companies highlights the importance of design and definition, as the exposures of each are materially different. These differences may result in the unintended inclusion or exclusion of constituents that affect country concentrations and performance outcomes.
Comparing several popular ETFs shows underlying holdings of 75, 112 and 326, highlighting that the definition of “infrastructure” can carry a narrow or broad perspective. One of the key illustrations is NFRA, an ETF with ~$3B AUM and >325 constituents that tracks the STOXX Global Broad Infrastructure Index.
Included in its top holdings are three widely held telecom stocks, Deutsche Telekom (T-Mobile), AT&T and Verizon — names that are traditionally not considered infrastructure plays. While these companies do own and operate traditional infrastructure apparatuses like cell phone towers, their businesses are highly targeted to the consumer segment, relying on retail customers leasing phones, smart watches and data plans. These companies have showrooms full of accessories like phone cases, earbuds and wireless speakers and spend billions on naming rights for global sports venues and jersey sponsorships.
Including these names brings the correlation of this ETF closer to that of the global equity market, and investors looking for diversification aren’t necessarily getting what they may have intended. Additionally, including these common telecom stocks leads to an overweighted position if the investor owns the broader global market, as these names are prominent constituents in those indexes as well.
Investigating the reasoning for including these names in an infrastructure exposure showcases the importance of understanding the design and definitions of the index that the fund is tracking.
Buyer beware.
NFRA and the STOXX Global Broad Infrastructure Index that it tracks include Verizon, AT&T and Deutsche Telekom because >50% of these companies’ revenues come from “Communication: cable and satellite, data centers, wireless, wireless towers and wirelines.”
Compare this definition to that of the DJ Brookfield Infrastructure ETF, ticker TOLZ, which defines infrastructure as “pure play” and only includes communications companies focused on “development, lease, concession or management of broadcast/mobile towers, satellites and fiber optic/copper (excludes telecom services) cable.”
A third ETF, IGF, restricts its infrastructure holdings to a maximum of 75 companies that have GICS classifications of: Energy, Transportation and Utilities and tracks the S&P Global Infrastructure Index. This methodology also weights those three infrastructure clusters at 20% Energy, and 40% for Transportation and Utilities, a seemingly arbitrary design detail, perhaps included to reduce the risks that commodity prices have on the Energy sector. Despite capping the holdings at just 75, this ETF owns a couple stocks with unusually high betas relative to the other holdings. Vistra Corp. and Constellation Energy have betas to the S&P500 of 1.75 and 1.52 respectively, while the average beta of the other 73 holdings is .68. Including these 2 names can cause the index and fund to be much more volatile and experience potential outperformance but can also carry a higher risk of a more severe drawdown.
As a result, these funds all have a 10-year correlation to the broad global equity market of over 81% with the broadest of the infrastructure ETFs close to 90%. Investors looking for diversification may be surprised to learn that their infrastructure fund behaves more like their core equity allocation than they intended.
Design choices are plentiful; should industry classifications be derived from revenue, cashflow, EBITDA — should they be GICS-based? How are the sub-asset classes weighted? Most methodologies use market capitalization, but should these exposures be equal or even risk-weighted? Are there capping mechanisms in place at the sector, group, country, stock level? How often is the portfolio rebalanced, and what is the starting eligible universe — should emerging markets be included? Expanding the exposure to liquid real assets, including Natural Resources, Real Estate and Energy, significantly increases the choices investors need to make when selecting or creating a methodology to track.
The details of the definitions are a paramount decision in the design of an index. These details may drive the outcome of the exposure, and poor choices at the outset often leads to a rough journey on a bumpy road. Just as in the construction of roads, tunnels and bridges, portfolio construction can learn from the adage, “Measure twice and cut once.”
