We have long considered the cell tower business model as one of the most attractive, boasting consistent cash flow growth with low maintenance capital expenditures. As such, the cell tower REITs, consisting of Crown Castle (NYSE: CCI), American Tower (NYSE: AMT), and SBA Communications (NASDAQ: SBAC), produced average annualized total returns of 15% from 2007 to 2017, which compared to 5% for the MSCI U.S. REIT Index (Bloomberg: RMZ) and 8% for the S&P 500 (Bloomberg: SPX).
However, the model has come under fire recently, most notably due to the potential merger of two of their largest tenants, Sprint (NYSE: S) and T-Mobile (NASDAQ: TMUS). As such, the cell tower REITs have produced a total return of -7.2% from March 31, 2018 through May 24, which compares to the RMZ total return of +2.9%. We believe fundamentals are intact due in part to an expected massive ‘5G’ spending cycle by the carriers that will re-accelerate organic cash flow growth from numbers that are already above the REIT average.
Moore’s Law tells us that the number of transistors that can fit in a square inch doubles every year. Historically, that has applied to technology related to mobile phone performance and storage, but processor speed and storage is not the bottleneck holding back applications for mobile technology. Similar to the issue of developing charging stations to support electric vehicles, network infrastructure needs to be built out before increasingly data intensive applications can be deployed on mobile phones.
Each deployment of new equipment on hundreds of thousands of cell tower ‘sites’ takes significant time and money to become ubiquitous through what is called an ‘upgrade cycle’. While it is difficult to predict the timing of such spending, we are confident that the future of mobile technology will require increases in data speed and capacity. As the primary owners of network infrastructure, the cell tower REITs will be critical to meeting the future demand for data.
REITs’ Role In Media
Despite all of the advancement in technology, media distribution depends on real estate just as much as the first telegraph networks established in the 1800s. Shown in Figure 1, the telecom sector is split into five segments (network TV, other TV, internet, telephone, and radio), which are distributed via four primary networks. The cell tower and other telecom infrastructure REITs primarily own or lease parts of the wireless and broadcast networks (shaded blue and orange).
In recent years, Crown Castle has been increasing its exposure to the faster-growing parts of the wired network (shaded yellow) as well, which is made up of copper and fiber optic networks within metro areas (‘metro fiber’) and between metro areas (‘longhaul fiber’). The majority of the wired network is controlled by telecom operators such as AT&T (NYSE: T), Verizon (NYSE: VZ), Zayo (NYSE: ZAYO), and CenturyLink (NYSE: CTL), although there is one REIT (Uniti (NASDAQ:UNIT)) that is focused on the space. Satellite stations and equipment are primarily controlled by telecom operators, as well as more specialized satellite companies.
As shown in Figure 2, the majority of cell tower REIT revenue comes from macro towers (aka macro cells), the ubiquitous 50-300 ft tall steel structures on top of which are mounted rows of multiple 1-5 ft antennas and microwave dishes, often facing in three different directions. We expect few new builds of macro towers in the future due to ‘Not In My BackYard’ (or NIMBY) attitudes that pervade urban areas where network demands are the highest.
As such, the growth in U.S. macro tower revenues will come from the addition of new tenants (e.g. FirstNet, to be discussed later) or lease amendments to add more equipment (e.g. 4G densification or 5G deployment). Internationally, many countries are behind the U.S. in macro tower development, which will create higher organic growth through new tenants and builds.
Small cells, also known as micro towers, make up the second-largest portion of tower REIT revenues. Small cells are essentially macro towers without the tower; instead, equipment is often mounted on the sides of 10-30 ft utility poles, flat-roofed buildings, and other vertical structures. Small cells are a key part of handling capacity in highly populated areas.
Notably, the range of small cell nodes is only 200-1,000 ft per site, which necessitates the installation of multiple nodes to cover the same area as a macro site. According to S&P/SNL Kagan, there are currently 60,000 small cell sites deployed, rising to 450,000 by 2020, a 650% increase! In comparison, there are 150,000 macro towers as of December 31, 2017, and they are projected to rise by only 8%.
Whether a text message, phone call, or video is received via a macro tower or small cell, the data previously had to pass through copper and fiber optic cables from the source, just as the original telegraph messages or long distance phone calls did. Labeled as ‘Fiber optic fronthaul/backhaul’ in Figure 3, these cables run between a region’s towers and its primary Mobile Telephone Switching Office (or MTSO), the wireless equivalent of a wired telephone central office. From there, the data travels across ‘metro’ or ‘long-haul’ fiber lines to another MTSO, central office, or data center, which forwards it along to its final destination.
The Next G
In our March 2016 REIT Outlook, we stated, “Following VoLTE will be the 5G upgrade, though we could be five or more years away from it being a significant driver of lease amendments.” This forecast could prove conservative, as industry standards for 5G mobile applications are expected to be finalized soon, which suggests to us that the upgrade cycle will start sooner. The carriers have been on record saying that they expect to spend over $125 billion combined upgrading their networks to 5G.
Upon adoption of the standards, the mobile device makers will be able to install the chip necessary to connect to the 5G sites, depending on availability. Thus, at the earliest, mobile users could access 5G in select markets in late 2019 or 2020.
In addition to the production and deployment of the mobile devices, the type and amount of equipment on cell towers will change, resulting in ‘lease amendments’ and driving higher revenue for the cell tower REITs. In particular, Sprint and Verizon plan to use 2.5 GHz and mmWave (>37 GHz) spectrum, respectively, which compare to 600-800 MHz for most 4G networks. Due to the shorter area of propagation of higher frequency spectrum, many more sites will have to be deployed to cover the same geographic area.
Another issue with 5G, estimated to be 10 to 100 times as fast as 4G, is the carriers’ ability to monetize the massive investment to achieve such high speed and capacity. In other words, carriers will not be motivated to install 5G unless customers are demanding it, and customers won’t demand it unless there are applications that cannot be done without it. This is not a new issue, as there were similar ‘chicken-and-egg’ problems in the 2G to 3G and the 3G to 4G upgrade cycle.
For example, there was a massive strain on the networks due to demands from video streaming in 2011, which would not have been possible without the deployment of 4G and prevalence of smartphones. As a result, carriers had to change their mobile data pricing plans, eliminating unlimited data plans (except for Sprint, though with some limitations). Ironically, unlimited plans re-emerged in 2017 as a key customer acquisition tactic by the carriers, which was made possible by the spending in prior years.
A Premature Call On Market Maturity
Most experts, including ourselves, expected U.S. demand for mobile data to grow by about 25% in 2018 as the market matured, high but significantly lower than the 80% annual growth between 2010 and 2015. Instead, most experts are now forecasting 35-50% annual growth over the next five years.
As the cost of data has fallen and devices are equipped with more memory, mobile apps have increasingly begun to approach the complexity and functionality of full desktop software. Mobile phones have facilitated more data than voice traffic since 2009, so it was no surprise in early 2014 when mobile apps passed desktops as the primary way that Americans accessed the internet. Mobile apps are effectively displacing the web browser for most purposes, as smartphone streaming is now faster than browser-based streaming and as fast as streaming to apps on a smart TV or laptop.
Social media apps such as Facebook (NASDAQ: FB) are increasingly using streaming video, while Google (NASDAQ:GOOG) (NASDAQ:GOOGL) Maps, Yelp (NYSE:YELP), Waze, Uber (UBER) and digital assistants have been rebuilt around data-intense augmented reality, artificial intelligence, and geolocation-oriented features.
Another wave of data demand will come from the ‘Internet of Things’ (or IoT), which includes ‘smart’ cities and self-driving cars. Similar to how smartphones connected people, IoT will connect ‘things’ to each other and to people. As of February 2017, research firm Gartner expected the number of connected devices to increase from 8.4 billion to 25 billion by 2020, an increase of almost 200%! Without 5G, the network will not be able to handle the traffic, as 5G can accommodate 10 TIMES as many devices as 4G.
The Internet of Things has the ability to change how humans interact with technology. Smart appliances will tell users when the laundry is done or when they need to buy more eggs. Smart cities will have traffic lights that can change on their own based on the flow of traffic. Healthcare providers can monitor patients remotely via wearable (or even implanted) devices that will prevent unnecessary trips to the doctor or hospital, while saving lives by recognizing crises as they happen.
Ranchers can monitor each cow or chicken, while manufacturers can track product assembly and deploy robots. A bit further into the future, self-driving cars will be able to communicate with each other and process information quickly through ‘edge computing’, which would be a large benefit to cell tower and fiber owners.
A New Tenant
In some instances, a strain on a network can be the difference between life and death. Since it is not economical for a carrier to build out a dedicated network that would only be used in emergencies, the government has agreed to fund the development of such a network in partnership with one of the carriers.
Officially called the First Responder Network Authority, ‘FirstNet’ is the world’s first cellular communications network built exclusively for America’s 25,000 public safety officials. Calls for such a network were first made in the wake of 9/11, when extraordinary traffic overloaded the networks in lower Manhattan and made coordination between first responders almost impossible.
After 15 years, a 9/11 Commission endorsement, and a 2016 bidding process, Congress awarded AT&T $6 billion and the rights to 25 years of FirstNet revenues, on the condition that AT&T develop the cell sites and deploy them on towers, subject to certain deadlines. We believe this will result in $1-2 billion of net new annualized revenue for the tower REITs by 2022, a 7-15% increase over their projected revenues for 2018.
Risk Or Buying Opportunity
While the addition of a new tenant in the form of FirstNet is a positive for the sector, one of the perceived risks to cell towers is consolidation between carriers, given the low number of tenants. In April, Sprint and T-Mobile announced that they had agreed to a merger, which sent the cell tower REITs down by 6.6% on average for the month, compared to the RMZ total return of +1.4%. Part of the rationale for the merger was that the combined company would allow 35,000 duplicate site leases to expire, although it would incrementally add 10,000 new sites. However, the market’s reaction may not have been warranted.
First, Sprint and TMUS have tried to combine twice in the past but were unsuccessful. In addition, AT&T and TMUS attempted a merger and were successfully sued by the Department of Justice. Therefore approval of the merger is far from a sure thing.
Second, even if the merger were to go through, the impact to the REITs’ cash flow growth may not be as severe as the stock price reaction would suggest. For example, CCI derived only 5% of its 1Q18 annualized revenues from Sprint on towers where there are both Sprint and TMUS sites.
Additionally, the weighted average remaining lease term for Sprint sites was 5-7 years, so there would be little to no effect on CCI revenues for several years after closing the proposed transaction. Finally, a primary driver for the proposed merger is Sprint’s lack of capital to build out a competitive network. Thus, the annual capital spent by the combined company could be higher than it would have been if they were separate.
Given the mission critical nature of the real estate and long-term predictable growth profile, the investment appeal of cell tower REITs should be widespread. S&P announced in August 2017 that SBAC would join the S&P 500 Index, thereby placing all three cell tower REITs in the index. In addition, Vanguard has changed the benchmark index for its passive REIT funds to one that includes cell towers, which will require the purchase of billions in cell tower REIT equity.
When coupled with average projected three-year annualized AFFO growth of 10% versus the Green Street Advisors estimated REIT average of 4%, and a 2018 AFFO multiple of 19.5x as of May 24, 2018 versus the REIT average of 19.7x, we believe the sector will continue to attract both REIT and generalist investors.
Disclosure: I am/we are long CCI, AMT, SBAC.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.