It’s not uncommon to see these pitches that are based on daydreams of, “What if you had bought Brand Name Stock X 50 years ago?”

And there’s a reason for that: They’re compelling, they help you to think about long time horizons, and they play off of a huge bias that we all have to ignore historical failures and fixate on winners.

That’s often called “survivorship bias,” and it’s one reason we daydream about buying IBM (IBM) or Apple (AAPL), companies that beat the market handily over 40 years, but don’t give a second thought to companies that disappointed investors and no longer exist, like Digital Equipment or Packard Bell or Gateway Computer (just to stay in the same industry).

When you talk about broad market investing returns, the survivorship bias fails us because you only look back at the historical for the stocks that now exist, failing to account for the ones who disappeared along the way… but it’s true when you think of specific companies as well — we are hard wired to think about the winners and ignore the fact that at one time they were part of a cadre of competitors that, in many cases, lost or were swallowed up or just disappeared along the way… and identifying the eventual winners in those early days would have probably required a lot more skill (or luck) than we like to admit.

And, of course, we all think that deep down, we’re smarter than the average bear — so of course we would have chosen Apple over Packard Bell or Gateway Computer, but we forget how much enthusiasm there was for these now-defunct (OK, almost defunct) companies at some points in time. Gateway, for example, was a beloved brand for a while and went public in a hugely popular IPO in 1993 at $15 a share, four years before Steve Jobs came back to Apple and eight years before the iPod was introduced. At the time, Apple was just a struggling computer maker steadily losing share to the Wintel cartel… 14 years later, in 2007, Gateway was acquired off the scrap heap by Acer, which itself has gone almost nowhere in 20 years, for less than $2 a share (coincidentally, Acer later also bought Packard Bell — which had the largest market share of any computer maker in 1995 — and for a little while it even made Apple-compatible computers when Apple toyed with licensing to other manufacturers to try to keep up with Microsoft in the mid-80s… but it is now just another discount brand that’s sold mostly in Europe and Africa)…

… so Gateway, which was certainly a more popular and faster-growing computer maker than Apple in 1993, generated 80% losses for shareholders over the next 14 years… while Apple went on to gain 3,000% by the time Gateway was acquired by Acer, which was still before the iPhone was released, and a total of more like 30,000% if you held it for another dozen years to today.

Which has nothing to do with this latest Motley Fool Rule Breakers pitch, of course, but I thought it was worth throwing in a little perspective before we dig too deep into the promises. Everything is obvious in hindsight, including Ray Kroc’s ridiculous idea to franchise a hamburger stand into every town in America.

So what’s the story being pitched by the Fool this time? Here’s a little taste of the ad:

“I just read that buying $2,250 worth of shares of McDonald’s in 1965 would have given you a $15.55 million stockpile right now.

“That’s enough to fund your retirement and then some.”

And the letter, which is signed by Shahin Dehestani at the Motley Fool but is an ad for David Gardner’s Rule Breakers service (that’s their growth-oriented “entry level” subscription, $59/yr), goes on to point out something you may have heard: That Ray Kroc didn’t really build McDonald’s financial might by franchising and selling milk shake machines… he did it by leasing properties to those franchisees.

More from the ad:

“It was a brilliant move that led Kroc to STOP signing ‘franchisees’ and START signing ‘tenants’ on land he began buying.

“Today, McDonald’s earns $30 billion from leasing grounds to franchisees… and has made a lot of money for its investors.

“But that’s not why I’m writing you today…

“The purpose of this email is to inform you that we’ve spotted one American company that’s at the forefront of an exploding tech revolution… and get this… they’re using the exact same playbook McDonald’s used to totally corner and dominate its market.”

OK, so it’s some kind of real estate investment. Probably a REIT, if it’s like other “McDonald’s of XXX” stocks that we’ve seen teased over the years. But which one?

Some clues from the spiel for you…

“… this company isn’t selling burgers. Instead…

* They’re growing fast in the $12.3 trillion 5G industry.
* They’ve cleverly bought 171,000 cell properties and are “landlords” to tech giants like Verizon, AT&T, Sprint, and many more.
* They’ve locked tenants into five-to-10-year contracts with renewal terms in place for the 5G rollout.
* They’re on track to build 10,000 more towers by the end of the year.”

Sound familiar? Notice that word “towers” in there? Yes, this is another pitch for a tower REIT — the companies (there are three big public ones and a few smaller “partial” ones) that own the towers and other locations used by the major mobile telecom companies to host their antennae and network equipment.

Most such companies have adopted Real Estate Investment Trust (REIT) status to avoid corporate income tax, which just means that they have to pass along 90% of what would be their taxable income to shareholders in the form of REIT dividends (which are taxable income to you, so the tax does eventually get paid).

This is an aside, but I should also note that in practice, REITs generally pay out more in dividends than they would have had in taxable income, taking advantage of depreciation accounting rules and the fact that real estate depreciation hits the books a lot faster than it hits the actual business (meaning, they can depreciate a tower over (just making the number up) 15 years, but it actually might have a useful life of 30 years… and because of the constant availability of new equity capital, they can use new cash to replace deteriorated assets instead of reserving cash for their eventual replacement). Sorry, I know accounting is boring, and I’m not a tax expert… just keep in mind that when teasers say a REIT is “legally obligated” to pay you dividends, that’s true to the extent that the REIT makes money… but the obligation does not necessarily mean they have to pay out dividends at the current (or higher) level, often the legal obligation (90% of taxable earnings) is well below the current dividend. Still, these are companies that are almost all explicitly set up to appeal to dividend-hungry investors, and management usually tries to maximize dividend payments — they know what their owners want.

But anyway, which tower REIT is he talking up here? Let me share one more bit of the ad first — it’s about the common argument that 5G will be great for the tower companies, because demand for their properties will increase as more 5G antennae need to be set up:

“… the key to winning rests in the hands of this company’s real estate and towers — without them, 5G will be reserved for the next bidder.

“But look, this isn’t a wild guess…

… because the same thing happened in 2009 when 4G made its debut.

“Back then, The Motley Fool’s co-founder David Gardner saw the pressing demand for new 4G towers… and quickly recommended this same stock to members of our Rule Breakers service.

“Those who listened made a stunning 491.1% return!

“But here’s the thing…

“Those gains are a small flesh wound compared to the large dent this stock could make in the coming months with 5G going mainstream.”

So yes, dear friends, this is American Tower (AMT). And this is when I feel both a little bit old, and a little bit sheepish… because that’s absolutely right, David Gardner was pitching this same stock almost a decade ago, I covered the first teaser I saw of the stock when he touted it as the “Ultimate Wireless Winner” in May of 2010… and I was a bit too put off by the rich valuation and didn’t buy it.

Back then, it was a $16 billion company that was just about to convert to REIT status, and hadn’t yet started to pay a dividend, and it was just beginning to expand into international towers. Today, it’s a $100 billion company (it’s the largest REIT in the world, I believe), it pays roughly a 1.7% dividend, and it has increased the dividend every single quarter since they started to pay regular dividends in April of 2012 when the REIT conversion was actually finalized (that’s roughly 24% annual dividend growth). The return for an investor who took the dividends in cash would be about 560% right now.

(Which means that in order for me to live with myself, I’d have to find something that I bought around that same time that also did very well — the only one that more or less fits the bill, I’m afraid, is Activision Blizzard (ATVI), which I bought that same month and would have returned over 400% if I held today, or 700% at the peak… but, of course, I was a dummy and didn’t hold onto it for the whole time, and in fact just started buying ATVI again this year. That’s enough to stop me from doing the Monty Python head-smash prayer, but not enough for celebration.)

So yes, American Tower has been a beast for a decade. And since it converted to a REIT and started paying dividends, the return has essentially been driven by the dividend growth (the stock has risen 20% a year and paid another 1-2% a year in dividends, and the dividend has risen about 24% a year), which is what almost always happens (that’s why you want REITs that have the potential for extraordinary dividend growth).

Yes, it’s exposed to 5G because 5G will require more antenna locations… and yet will still require 4G to fill in the gaps because mm-wave 5G signals won’t cover long distances and are easily impeded. The tower companies make a lot of their money through the magic of colocation — which basically just means they get to lease the same tower to more than one company. The original tenant covers most of the cost of the tower, but then they can lease space to three or four other tenants on that tower as well, since there’s often plenty of room for other antennae, and those subsequent leases generate much higher profit margins.

The open question, really, is how 5G will play out — it will require a lot more locations for what they call “small cells” — think locations on street corners instead of hulking towers alongside the highway, with power and fiber connections for backhaul connections to the network — and no one is quite sure whether this “colocation” model will work with small cell locations. Most of the tower companies, including American Tower, have built up their portfolio of small cell locations that they can offer to telecom companies, but that’s still a very small business compared to the existing tower business… and American Tower has focused much more doesn’t seem to believe the lower-margin small cell business is very attractive (unlike Crown Castle, which has bet big on small cells).

There’s some risk in the short term, because if 5G investment is relatively slow to take off, and the Sprint/T-mobile merger goes through, then there will be some cell locations that the combined Sprint/T-mobile might be able to drop. I’m not particularly worried about that, since they have long leases and Sprint and T-mobile use different equipment so will really need to maintain duplicative networks for an extended period of time, but that’s been one worry hanging over the three big cell tower companies. And it was one reason why I didn’t go “all in” on the tower companies a year or two ago, since I was hoping that merger would bring a dip in the shares (it didn’t, dammit).

The other big ones aside from American Tower, by the way, are SBA Communications (SBAC) and the aforementioned Crown Castle (CCI), both of which are huge and also among the 10 largest US REITs — CCI has a $57 billion market cap, SBAC $26 billion. I prefer and am invested in Crown Castle, mostly just because it is US-centric and has focused a lot on building up small cell (and fiber) capacity, but also because it pays a higher current yield (given the next anticipated dividend hike in a week or so, I’m guessing the forward yield should be about 3.5% — right now it’s 3.3%). CCI’s dividend is not growing as quickly as AMT’s, however — CCI grows the dividend by about 8-10% a year, AMT is still growing theirs by more like 20% a year.

Crown Castle has lagged AMT and SBAC in the year since I bought shares because of lower growth rates, partly due to the fact that SBAC and AMT are both expanding rapidly overseas while CCI has invested in small cell networks that aren’t going to pay off in the near term (though, of course, they hope this bet will be important longer-term), so I’ve been wrong on that, but we’ll see how it goes (CCI is up about 29% in a year, roughly half of the gains that AMT and SBAC have made). SBAC was the last to convert to REIT status and just initiated a dividend in August, they currently pay at about a 0.6% rate but we’ll see how that evolves and if they begin to grow the dividend meaningfully from here. All three have been pretty heavily teased from time to time as good plays on 5G by various newsletters, particularly over the past year or two.

All three companies were around for a long time before they converted to REIT status, by the way, and AMT has been the best performer and the most volatile for pretty much all of that time (unless you bought right at a pre-crash peak)… so your opinion about where we are in the market cycle might color which of the tower REITs looks most appealing to you — this is what the long-term chart looks like, for example:

But this is what it looks like if you bought near the peak of the dot-com bubble, when AMT was doing wildly better than SBAC or CCI:

And the outperformance for AMT is similarly less dramatic if you bought near the pre-crisis peak in September of 2008:

They do clearly tend to trade together, though AMT has pretty consistently outpaced the growth of CCI and SBAC. They’ve all always been very expensive compared to the average real estate investment, which is what deterred me from getting involved for a long time — so there’s some risk that if growth expectations die down, these three could all have pretty far to fall, and I think CCI should have more downside support than the others due to its higher dividend… but so far most such worries have been overstated in recent years.

What does the current valuation look like? REIT investors usually go by funds from operations (FFO, a cash flow measure) instead of earnings per share, and AMT currently trades at a price/trailing FFO ratio of about 26 (CCI is at 23, SBAC at 32). On a more traditional PE ratio, AMT is at 61X forward earnings, CCI 69X and SBAC 191X (SBAC went through an awful patch and still, I think, has some big losses they can write off for tax purposes, which depresses their reported earnings). Right now, analysts project that cash flow (or EBITDA, at least) will grow at a 4% rate for AMT, versus 6% for CCI and 8% for SBAC (using 2018 numbers and estimates for 2021)… in terms of revenue, the equivalent forecasted growth rates are 4.7% for AMT, 5.4% for CCI, and 6% for SBAC.

Which all makes me reconsider — even now, perhaps, I should be looking at AMT and its much higher historical growth rate and larger (and international) portfolio over CCI and its more comforting current yield, somewhat better valuation, and focus on 5G and small cells in the US… and maybe even considering SBAC, which could emerge to generate a real dividend eventually. But that’s been challenging because I’ve been dealing with the hang-up of judging REITs based on the dividend yield, underestimating AMT’s ability to grow the dividend, and struggling to justify a 1.7% yield for a mega-cap REIT. I don’t really want to chase those valuations with big bets after a huge run this year, but in retrospect I should have probably spread my investment across all three… and I expect all three will work out well over time, and will likely benefit from both a multi-year wave of 5G infrastructure investment around the world, and from continually depressed interest rates (which make their borrowing cheaper, and make their yields look more compelling for investors).

And incidentally, if you just like the yield and want a little growth “juice” for your portfolio, then if you buy just the broad Vanguard Real Estate ETF (VNQ) you’ll be getting a fair dose of tower REITs — AMT, CCI and SBAC combined make up about 15% of that index ETF (and sticking with the “high tech” theme, data center REITs are another 6% or so), and it should have a forward dividend yield of about 3.7%. Of course, you get shopping malls and office buildings and timber and apartment complexes and nursing homes along with all that, and maybe you don’t want that stuff… but we often overthink these things, and not everyone wants to spend their days obsessing over individual stock tickers (I do, of course, but I’m told I’m a little bit deranged).

REITs in general have been on a heckuva run as the sentiment shifted from “interest rates will rise and cause problems for REITs” to “interest rates will never rise, and may go to zero,” so do be mindful that all of these leveraged “asset” investments will be significantly influenced by changes in interest rate expectations. Cutting interest rates further, even to zero, might not spur economic growth at this point… but it would probably continue to boost asset values and make investors hungry for dividends that would have seemed almost inconceivably low a decade or two ago.

If you want just “high tech” REITs, by the way, there is also the Benchmark Data & Infrastructure Real Estate ETF (SRVR), which is mostly tower REITs and data center REITs with a few oddballs thrown in (mostly billboards)… it’s relatively expensive for a ETF, with a 0.6% expense ratio, but you do get some diversification in the space (American Tower and CCI account for about 29% of that very focused ETF, so if you just bought those two and Equinix (EQIX, the biggest data center REIT), which is another 15% of the ETF, you’d also get probably a pretty similar result and, thanks to the expense ratio, a slightly higher effective dividend yield… potato, potahto).

The other risks to the cell tower REITs in general come from the fact that they have really pressured the mobile telecom companies by keeping lease rates high, so a big part of the cost structure for Verizon and AT&T is tower rent… and that might backfire, since companies like Verizon and AT&T now pay out a pretty huge percentage of their revenue as lease payments to the REITs, which means they probably sometimes regret the fact that they spun out their tower networks to these REITs in the first place — which means they might continue to hedge their bets by building their own towers and small cells to reduce their reliance on the colocation tower owners. That has accelerated over the past year or so with AT&T and Verizon both bankrolling Tillman Infrastructure, a startup that’s building new towers for them, mostly as a way to pressure the existing tower owners (this won’t be a quick or dramatic change, AMT and CCI each own 40,000 tower locations in the US, but it is a sign of the potential competitive threat… which is meaningful in a market where there are only a few customers).

Add in the potential loss of Sprint or T-mobile, and if we end up without a strong and viable third mobile company in the end it might be that Verizon and AT&T have enough heft between them to put the pressure back on the tower owners and keep a lid on lease prices. And, of course, the situation in every other country is different, sometimes better and sometimes worse for the tower owners (AMT and SBAC rely a lot on growth from overseas, which sometimes hits snags — like AMT’s troubles last year in India as wireless companies consolidated there and they lost tenants — but is largely why they’ve grown faster, while CCI is really US-only).

So there you have it, dear friends — yes, American Tower is a pretty solid play on continuing demand for wireless data infrastructure, and it has been a fantastic investment since David Gardner first teased it a decade ago. It’s tough to swallow at a $100 billion market cap and with a rich valuation and low current yield… but I found it too tough to swallow in 2010 as well, and that was a mistake. Which is why, thankfully, you have to decide for yourself what you should do with your money — interested in American Tower and its relentless dividend growth as they build cell towers around the world? Prefer their near-peers like upstart turnaround SBAC, or higher-yielding CCI? Let us know with a comment below.

Disclosure: Among the stocks mentioned above, I own shares of Crown Castle, Activision Blizzard and Apple. I will not trade in any company covered for at least three days after publication, per Stock Gumshoe’s trading rules.