American Tower owns cell phone towers. The wireless carriers, AT&T and Verizon, used to own their own cell phone towers. But it was an inefficient use of their capital. An AT&T tower would tend to only have AT&T equipment on it. The wireless carrier is responsible for leasing the land, the backup generators, taxes, maintenance, and all the other expenses that add up. The wireless carriers were earning little return on their investments in their own towers.
American Tower can make the same investment, but they can generate higher returns. American Tower leases tower space to other wireless carriers. A typical tower can accommodate up to 4 carriers. American Tower can take a low yielding investment and turn into a high-return business.
The major wireless carriers are happy to share tower space because the cost of their rent is minuscule to the cost of maintaining a tower.
Cell phone usage and demand are in a secular growth trend. American Tower sits in a prime position to benefit from this growth as a mission-critical high-value tower provider.
American Tower started paying a dividend in 2011 when it converted into a Real Estate Investment Trust (REIT). Since then, the company has grown its dividend at a compound annual growth rate of 30%.
We expect its dividend to continue to grow at a double-digit rate.
Dividend Safety for a REIT is different than the standard dividend growth stock. We kept the cash dividend payout ratio as a comparison, but the two most important metrics are the AFFO payout ratio and the long-term debt to EBITDA.
AFFO stands for Adjusted Funds from Operations and you can think of it as a REIT’s free cash flow. We want a REIT’s dividend to be well covered by its AFFO.
REITs must pay out over 90% of their profits to shareholders. A REIT can’t rely on retained earnings to reinvest and grow their business like a traditional C corp. A REIT needs to raise capital through Debt and Equity Issuance. Too much debt and the REIT has to issue more equity. The more equity you issue the higher your cost of capital. If your cost of capital exceeds your returns on capital, then a company is destroying value.
American Tower’s low net debt ratio means it likely still has access to low-cost debt capital.
Catalysts for Future Price Appreciation & Dividend Growth
It is cheap to build out a cellphone tower. American Tower’s average cost to build a tower in the U.S is $275,000. The average rent is $20,000 per year and the average operating expense is $12,000. This generates a 3% return. A tower's costs are fixed. Each new tenant added increases returns.
According to American Tower, their return with two tenants is 13% and close to 24% with three tenants. Because they charge the first tenant $20,000 and each subsequent tenant $30,000.
American Tower averages 1.9 tenants per tower. Adding tenants is competitive. In the U.S. there are 3 major wireless carriers and two other large tower companies. But the major carriers want a “blanket of coverage” for their customers. If American Tower can increase its average tenant per tower and/or increase rent per tenant, then shareholders should see increased operational cash flow and increased dividend payments.
Renting out tower space is a commodity service. There are two other large tower companies in the U.S. that wireless carriers can partner with. The costs to build a new tower site are low. Besides zoning laws, the barriers to entry for a new competitor are low and so should be the switching costs but they’re not.
Crown Castle once stated that it costs carriers $40,000 to remove their equipment from a cell tower. The cost should be the same at American Tower. This is a tangible switching cost. Swapping tower companies also has intangible switching costs.
Towers are the backbone for the smooth functioning of cell phone service. A cell tower can handle around 1,200 connections. The denser your network the more customers you can handle without compromising service. If a wireless carrier switches tower companies, the new tower company must replicate the same coverage. If the new tower company can't replicate the same coverage, then the wireless carrier risks alienating its customers with poor service. The rental cost is minuscule to the reputational and operational costs, so the carriers don’t switch. The churn rate at American Tower is below 1.5%.
Each new generation of wireless networks comes with equipment upgrades. The transition from the old network to the new network takes a long time. A wireless carrier will have overlapping equipment at each tower site to serve both networks. This increases the rent a tower company collects per wireless carrier. Tower companies also negotiate rent increases and future rent escalators during the rollout of the new network.
The 5th generation of wireless service is starting to rollout. Very few phone devices are 5G capable right now. But by 2025 Ericsson estimates that 45% of all mobile data traffic will be on 5G networks.
Wireless carriers need their existing 4G to overlap with their 5G service. In some places, the wireless carriers still need to service their old 3G network. 5G networks tend to need MIMO (Massive Input, Massive Output) antennas in suburban areas versus small cell towers in dense urban areas. MIMOs are wider and heavier. Antennas that take up more space and cost more to put on the towers. American Tower will charge the wireless carriers more for the 5G MIMOs while still collecting rent on the 4G equipment that needs to stay up.
Coinciding with the rollout of 5G is the rise of the Internet of Things. The data demand from cell phones is expected to grow at 30+% per year in the U.S. alone.
Now add a million more connected devices. The wireless carriers need to increase their infrastructure spending to keep up with this demand. That spending should flow through to tower companies like American Tower.
American Tower believes its international opportunity is like the U.S. 10+ years ago. Cell phone standards in the developing world are 5-10 years behind the U.S. Many international markets still haven’t migrated to 4G. American Tower benefits from the continued 3G to 4G roll out. Like the 5G rollout in the U.S., the international carriers will overlap their 3G and 4G equipment on towers. Then international carriers will start their 5G upgrade cycle. The international wireless carriers also need to increase their coverage density.
According to American Tower’s latest investor presentation, American Tower’s U.S. sites account for 23% of their total communication sites but generate about 53% of American Tower’s revenue.
The average client density per international tower is 1.5 clients versus 2.7 for U.S. tower sites.
Excluding currency fluctuations. We expect American Tower’s international communication sites to generate higher returns than its U.S. business because mobile device penetration and data demand should grow faster than in the U.S.
The internet of things and the growing demand for mobile data means more demand for cloud computing infrastructure, primarily data centers. To meet the growing data processing demand, a combo of large central data centers and smaller edge data centers are needed. Edge data centers are placed closer to the devices that need quicker turnaround time on data transmission. A natural location is the land leased by cellphone tower companies.
A year ago, American Tower purchased the Edge Data Center company Colo Atl and this year it is rolling out American Tower Edge in 6 cities.
Large data centers like Equinix create network effects. Each data center is a node of connections between data users and providers. The more nodes you have the lower the latency of data transmission which makes your network more valuable and more attractive to others that want to be a part of your network.
American Tower is the largest Tower Company in the U.S. with more than 170,000 sites. This is very attractive to potential clients that want to get their hardware and software closer to the end user. It lowers the cost of data transmission and latency. If American Tower can add edge data centers to most of its sites it can create a highly desirable and valuable edge data network.
Creating a cell tower site is relatively cheap and easy. The hardest part is getting the land and passing the zoning laws. If American Tower’s competitors can offer lower rent and if they’re willing to accept lower returns, then American Tower will need to lower its rent to maintain tenants and its returns will diminish. Diminishing returns on investment means we overpaid for American Tower at today’s prices.
We view this as a low risk. American Tower’s big advantage is it already has a large cell tower network and prime site locations. Replicating their coverage is almost impossible because of the regulatory burden it would entail.
A tower with one tenant generates a 3% return. A new tower site has to offer a large enough discount to make just the physical cost of moving sites worth it for wireless carrier. This doesn’t account for the intangible costs of disruption to a wireless carrier's network. The discount would cause the return on the tower site to be negative for several years. It also can take several years to add another tenant. So how long is a competitor willing to take a loss on their tower site before they break even?
Poor Capital Discipline
The bigger risk is poor capital discipline. American Tower doesn’t need a competitor to justify taking a lower return on investment. It simply needs to foster a corporate environment that focuses more on growth than economic returns.
If American Tower pursues growth for the sake of growth without regard to capital discipline and returns on capital, then we’ve overpaid for American Tower. Our investment is predicated on management maintaining capital discipline and generating strong economic returns on their tower investments.
The bulk of American Tower’s cell towers are located outside the U.S. The general risks of investing in other countries apply here.
Does the country follow the rule of law?
Does the country favor local industries? Could this lead to unlawful asset seizures?
Our base case estimate of fair value based on a discounted cash flow model is $297 per share.
A DCF model is just a tool to see if we’re paying a reasonable price given some basic assumptions. Our assumptions could be wrong. What we’re really investing in is a company with a history of disciplined capital allocation, in a prime position to benefit from the secular growth of mobile data usage, and with a potentially valuable option in the growth of edge computing. Given the underlying quality of this business, we believe that any minor errors in our assumptions will be bailed out by these factors in the long-term.