|Shares Out. (in M):||965||P/E||0.0x||0.0x|
|Market Cap (in $M):||408||P/FCF||0.0x||0.0x|
|Net Debt (in $M):||3,240||EBIT||0||0|
I’d like to re-litigate the short case for Clearwire (CLWR, $1.64), which has been written up twice before on VIC – in May of 2010 by buggs1815, with shares at $7.58, and again in Feb of 2011, by vandelay278, with shares at $5.00. One key valuation assumption shared in both write-ups is deeply flawed, and my own long thesis implies that the material issues raised in those earlier write-ups have either been addressed, are over-exaggerated and/or no longer apply given developments at the company over the last seven months.
Please note that I submitted this to VIC almost a month ago, and there have been several developments at the company and majority owner Sprint in the interim. None have materially altered my conclusions below, however. I’ll attempt to address those more recent events in the comments section at the end if asked. Much of the current angst around the company reflects that it is in the process of raising capital, but in the below I argue that fears of massive dilution are overblown due to, among other things, the probability of securing favorable vendor financing. And while I agree both the converts and 2015 notes could look appealing right now, I am comfortable with my position in the capital structure as a common equity holder.
I’d refer potential Clearwire longs to both short write-ups here on VIC to better understand the current high short interest. Know that much has changed about Clearwire since those previous short write-ups, however. They also both contain the following significant error.
Correcting the Record
With regards to a prominent valuation benchmark mentioned in both Mr. Buggs and Mr. Vandelay’s previous short cases for Clearwire, it is true that in October of 2009 at Auction 86, the FCC auctioned 1.77 bil MHz-POPs of spectrum in a similar band as Clearwire’s spectrum for the price of $0.011 per MHz-POP. But due to numerous key differences between that auctioned spectrum and that possessed by Clearwire, Auction 86 is an exceptionally poor valuation benchmark. Specifically:
The licenses offered in Auction 86 covered 78 areas. Licenses for 75 of those areas were up for bid because of previous “default, cancellation or termination.” In other words, the private market had already determined the licenses at that particular auction were of very low value. In addition, the licenses to be granted were only for a ten-year term, incumbents were already operating in the same bands, and the geographies covered were decidedly unappealing.
More specifically, 20% of the areas covered in those Auction 86 licenses were towns with populations of less than 100,000 – the smallest being Williston, ND, home to 26,000 people.
I will be the first to concede that the downloading habits of people living in small towns in North Dakota could make crack cocaine look like decaf. Nonetheless, the demographics of areas like the ones covered in this auction in no way compare to areas covered by Clearwire – a key factor in any relative valuation of spectrum.
Only two locations up for bid in Auction 86 even covered more than a million people – Miami and San Juan, Puerto Rico. The Miami license would likely not have come unencumbered in October 2009, and the value of a license in San Juan to an operator located anywhere other than Puerto Rico is likely indistinguishable from zero.
Other regions in this Auction included Guam, American Samoa, and three locations that, quite literally, plot right in the middle of the Gulf of Mexico.
Here is the link to the areas covered in Auction 86:
And here is the correct link to the appropriate FCC study:
As discussed later, valuing spectrum - even in the same band - using relative valuation approaches can be fraught with error, given the differences in populations, terms and encumbrances. The above is a textbook example.
I am also compelled to point out, however, that given the circumstances above, using a $0.01 per Mhz-POP benchmark from Auction 86 to either derive or anchor a valuation of Clearwire – as was done in those two prior short CLWR theses here on VIC – is inappropriate. I share both authors’ skepticism of the company’s previous assertions that its spectrum was worth between $0.50 to $1.00 per MHz-POP (or $23B to $46B in total), but to cite $0.01 as the relevant benchmark appears either naïve or disingenuous.
Regardless, it is also clearly wrong, and in my opinion could result in valuations and assumptions about Clearwire that should be disregarded accordingly.
My own thoughts on valuation on the company follow. I’ll summarize them upfront as follows:
Through the deployment of a network covering 135 million people, Clearwire has begun to gain significant momentum in the monetization of an asset that is unmatched in its industry – deep, wide, clean spectrum. Shares are temporarily depressed due to elevated uncertainty about an impending capital raise and a number of tactical issues which will soon be resolved. Once the current capital raise is in hand, Clearwire will be on the pathway to providing more data at a lower cost to wireless consumers than any other competitor in the U.S. - easily.
Shares now trade at $1.64. Using conservative assumptions bordering on the pessimistic, per share intrinsic value is likely between $4 and $6, depending on the eventual terms of a pending $1.0 billion capital raise. A more objective and analytically justifiable valuation of their spectrum puts shares closer to $16. That higher valuation has also been independently corroborated by the work of a well-respected value investor who appears to have been recently accumulating more shares.
Clearwire’s spectrum constitutes a margin of safety. There is a significant probability of large gains. New management is competent enough. The former interim CEO and now Chairman, John Stanton, bought $5 million worth of shares on the open market in August. Prior insider buying by Stanton as CEO of both Voicestream and Western Wireless resulted in phenomenal, multi-bagger returns. Several near-term catalysts exist.
Now, the long version.
Clearwire Corporation provides 4G mobile broadband services through a network that covers 135 million people in the U.S. The company’s network, and entirely IP-based and built on open standards, is monetizing an unparalleled spectrum asset which I believe constitutes an underappreciated competitive advantage.
Specifically, the company has over 45 billion MHz-POPs of wireless spectrum - deep, unencumbered, and averaging 160 MHz across the top 100 U.S. markets. Clearwire to date has spent approximately $20 billion to acquire that spectrum and build out its network. For some time the company has been touting the breadth of its spectrum as a competitive advantage, but only recently does it appear to have finally aligned its business model to capitalize.
After adjusting for the cost of spectrum leases, the company’s enterprise value is currently $6.5 billion, meaning a private market buyer could acquire the firm and all assets for 33% of what it cost to build them just a few years ago. On a conservative EV/MHz-POP basis, then, the market is currently valuing Clearwire at $0.14 – meaningfully below two formerly bankrupt satellite companies with spectrum that, it should be noted, has historically not merited a premium valuation.
Clearwire is now on the cusp of a significant transition. Once funding for additional capex is secured – more on this in a bit - Clearwire plans to deploy an overlay of “LTE-Advanced” technology over its existing WiMax network. In recent trials, the company has recorded download speeds of using the new technology greater than 100 Mbps - or 3x what Verizon has noted in tests of its own next generation LTE technology.
While this test speed is unlikely to hold up in wide scale commercialization, the point is that its spectrum enables Clearwire to do things that other operators cannot. Though the risk of technological change in this industry has historically been high, at this time, it appears technical barriers will be enough to insulate the company’s earnings power for at least the near- to medium-term.
So the pending network overlay will represent speed and capacity that potential competitors will be unable to match. Given looming spectrum constraints in the industry, Clearwire’s excess spectrum capacity should drive the further adoption of the company’s wholesale business model. Management believes the migration from WiMax to LTE-Advanced will be relatively elegant, costing $600 million and consisting primarily of software and line card upgrades.
In Q2, Clearwire lost $0.65 per share on revenue of $323 million – and there is no denying the company has an intimidating history of losses to date. Welcome to the world of start-up telcos. Concerns about cash burn, technology adoption, dilution, and/or debt often come with the territory, and CLWR is no exception. To be clear, it is in no way my intent to excuse the P&L, but I do think it’s valid to acknowledge where the company is in the corporate lifecycle in order to avoid nausea.
Prior to its March 2007 IPO, Clearwire racked up $200 million in losses on just $76 million in revenue over the first 9 months of ’06. That the company was still able to raise $600 million in the primary market was clearly more a testament to the Street’s infatuation with Craig McCaw than an endorsement of the company’s financials. So I think it’s appropriate to conclude that Clearwire went public too early. That said, to presume the future will resemble the past misses the whole point of capital expenditures.
Over the next three quarters, the P&L should begin to confirm what subscriber counts are already indicating – that the company has reached a critical transition point. Management stated on the Q2 call that it has enough cash for the next 12 months, and though I am projecting a later date, they expect the business to be EBITDA positive in the first quarter of 2012. I believe the recent appointment of the former COO/CFO as the new CEO and the presence of John Stanton as Chairman has bolstered management’s credibility. Stanton personally bought $5 million worth of Clearwire shares on the open market this August, and his track record of insider buying at Voicestream and Western Wireless was exceptional.
Clearwire ended Q2 having 7.7 million total subscribers, up sharply from 1.64 million subscribers in the second quarter of last year. Of that total, 1.3 million are retail customers - folks that walked into a retail wireless store and walked out with a new dongle. As in “laptop modem,” ladies. These retail customers accounted for 59% of the company’s revenue last quarter. The company intends to gradually de-emphasize the acquisition of retail clients in the future, however, and in late June outsourced the management of its network to Ericsson to focus on selling wholesale capacity service to other carriers – a significant opportunity discussed more in a bit. The CFO is confident the company can continue to accrete cash to the business through the retail channel for the foreseeable future.
Clearwire had 6.4 million subscribers in its wholesale business in Q2 – all of which effectively came from majority-owner Sprint. That wholesale subscriber base was 3.3 million at the end of last year. So, while Clearwire has faced some challenges over the years, the ability to load users onto their network is no longer one of them.
Clearwire now expects to end 2011 with approximately 10 million subscribers, with most of the new subscribers coming from its wholesale business. This is the second increase in subscriber guidance the company has given this year, and I estimate the company will end up with just over 21 million subscribers by the end of 2015.
The emergence of an attractive wholesale business at Clearwire is being obscured by a solid but otherwise unremarkable retail business, as well as a number of other tactical issues that will soon be resolved. Given the many variables involved, it is difficult to place high confidence in a traditional multi-year earnings model for the company, but the company’s spectrum asset nonetheless provides an appropriate margin of safety at recent share prices. The potential for significant appreciation in shares appears high.
I believe concerns about the dilutive impact of near-term funding are overblown, as demonstrated later. And while the company’s relationship with Sprint has exhibited strain in the recent past, the companies remain important partners whose conduct will if nothing else be bounded by a policy of mutually assured destruction. Sprint may or may not choose to fund Clearwire’s near-term capex as part of its Network Vision initiative to be announced in early October. No matter the decision, however, it will speak more to the future of Sprint than Clearwire, as I believe the latter has multiple funding options.
Larry Robbins of Glenview Capital Management is long CLWR equity and debt, and I believe his thesis is also noteworthy. As per his Q4 2010 letter:
“…Looking five years out to 2016, we expect Clearwire could generate as much as $3.4 billion of EBITDA, which should command a 7x multiple, yielding a target enterprise value of $24 billion, and an out-year target equity value of $17 per share resulting in a 28% CAGR over the next five years…simply put, we like Clearwire because:
a) there is significant downside support from asset value,
b) the investment to build those assets is substantially behind them, and
c) the wireless industry is at an inflection point in terms of demand for these assets.”
Nonetheless, for several reasons, Clearwire hasn’t garnered much respect the last few years. And candidly, until recently, it probably didn’t deserve it. But now I believe it does. Here is why.
Why Is This A Good Business?
Harbor no illusions - this will forever be a capital-intensive business in an industry driven by technological change and cutthroat competition from large, occasionally irrational incumbents. In the absolute sense, however, Clearwire is a better business now than it has ever been. It’s also the beneficiary of favorable timing in at least two ways.
First, the WiMax standard, which the company had previously bet its future on, is on the way to being relegated to Betamax status. The company has been fortunate that LTE technology has developed as quickly as it has over the past two years in particular – and that migration will be relatively cheap and easy.
Similarly, Clearwire will be approaching EBITDA breakeven on an unparalleled platform at a time when, as the FCC Chairman has stated, “…mobile broadband is being adopted faster than any computing platform in history, and could surpass all prior platforms in [its] potential to drive economic growth and opportunity."
The FCC projects that mobile data traffic in the U.S. will be 35 times higher in 2014 than it was in 2009. Cisco believes mobile traffic will increase by 82% a year through 2015. Cisco has also noted that a smartphone generates, on average, 24 times more wireless data than a plain vanilla cell phone; a tablet generates 122 times more wireless data than a feature phone; and a wireless laptop creates 515 times the wireless traffic as traditional cell phones.
Smartphone adoption rates paint a similar picture of rapid growth. Across the four major operators in the U.S., smartphones went from single digit penetration in 2006 to 40% in Q1 of this year. Verizon Wireless is projecting that smartphones will make up 50% of its postpaid subscriber base by year-end.
AT&T has also reported that the iPhone drove a 50-fold increase in data traffic on its network in just two years. By 2015, AT&T projects mobile data traffic on its network to reach eight to ten times what it was in 2010. In other words, in just the first five to seven weeks of 2015, AT&T expects to carry all of the mobile traffic volume it carried during 2010.
There is an argument to be made the federal government may be guilty of mismanaging spectrum, and that to the coming crunch may be of our own making. Nonetheless, I don’t care. Spectrum is a classic example of a good with relatively fixed/inelastic supply. Given strong anticipated demand, in several years, the industry will have significantly less available spectrum than it does know. That scarcity will enable Clearwire to extract attractive economic rents due to the excess spectrum capacity it possesses.
In short, there is a looming spectrum shortage, and Clearwire is uniquely positioned to exploit it.
The Wholesale Network Model
Again, Clearwire has two lines of business – retail and wholesale. Until recently, the company seemed content to butt heads with other operators to win over retail customers, spending an average of $313 to land each new subscriber in the second quarter. Now, however, it is shifting focus to emphasize wholesale - and positioning the retail business to effectively fund that transition.
Clearwire wants to aggregate the data traffic of other wireless carriers onto its network and effectively “wholesale” its excess capacity. It will earn revenue via multi-year contracts with specified minimum usage and payment terms, including a minimum charge per network device, in exchange for providing high capacity broadband service that is guaranteed to meet specific quality standards. Carrier customers will effectively be able to repackage and resell Clearwire’s capacity to their own subscribers.
It’s already providing wholesale service to Sprint under an agreement announced this past April for a minimum of $1 billion paid out through the end of 2012. Seizing an opportunity due to the floundering of LightSquared, Clearwire is apparently in early conversations with the other major U.S. carriers as well.
To understand the appeal of the wholesale model to other carriers, it’s important to understand this:
The key to profitability in the mobile wireless data industry is the capex required to add capacity.
Declining equipment pricing will help, as will consolidation. The most economically appealing way to defray capex costs, however, is through network sharing, sometimes called network hosting. By aggregating other carriers’ data traffic onto its own LTE-Advanced network and providing excess capacity cheaper than customers can deploy themselves, Clearwire intends to become the carrier’s carrier - or as the CFO recently put it, “the Switzerland of the broadband industry.”
So through its wholesale model Clearwire intends to offer other wireless operators a viable way to mitigate their considerable capex expenses while better controlling opex, too, both of which should lead to increased profitability for them.
Though the network wholesale model is a relatively new concept in the U.S., it has been at least partially validated by similar models being rolled out by wireless operator Yota in Russia – another WiMax-turned-LTE provider - the telco Mobyland in Poland, and Vodacom in South Africa as well. Closer to home, the satellite spectrum company Lightsquared has recently been a vocal advocate of this model as well. More on that company later.
Perhaps the most appealing aspect of the wholesale model from Clearwire’s perspective is that there are so few variable costs. For instance, it costs them nothing to add a new user to their network. (CPGA = 0.) LightSquared executives, for instance, have estimated the breakeven on a wholesale business is half that of building a traditional retail channel.
In a business that is based on wholesaling excess capacity, like the wholesale model, Clearwire will have a sustainable competitive advantage. Specifically, the company will be the lowest cost provider of bits and bytes delivered over the air.
Because it’s all IP-based, Clearwire’s network cost less to run and maintain than that of a legacy voice operator. The company can also make full use of the latest antenna and network architecture technologies, including wireless backhaul, to deliver a much lower cost per bit. Because of its large spectrum, it can deliver oodles more bits. So the marginal cost of providing additional capacity ends up close to nothing. This competitive advantage exists because the wholesale model focuses on already loaded, densely populated areas. The higher the subscriber density is in a given area, the more data traffic that results, and the more Clearwire’s own capex costs are spread over each bit. So they can charge a price low enough to entice carrier customers to use them, yet still maintain attractive margins. It’s also important to note that Clearwire was able to choose the specific type of LTE technology it desired to use – time division duplexing, or TDD – as opposed to the frequency division duplexing or FDD flavor of LTE that AT&T and Verizon were forced to adopt given their original spectrum allocations.
So at the risk of greatly simplifying, Clearwire’s flavor of LTE will give the company more efficient use of its spectrum asset relative to other operators. An given that asset already has the most capacity in the industry, the company’s position as the low cost provider is further bolstered by a significant technical barrier. At least for now. I suppose that somewhere near Silicon Valley, some punk in a hoodie could be figuring out how to bank-shot RF waves off redwood leaves, but for the time being, Clearwire’s technical barriers appear reasonably insulated from disruption.
To be clear, the company’s competitive advantage is not analytically provable yet, nor is it instrumental to my thesis. Shares in Clearwire are currently cheap enough to be appealing without a moat. I do think, however, that it represents an under-appreciated aspect of Clearwire’s future earnings power. And right now it costs nothing.
Why Are Shares Cheap?
Exclusive of macro concerns, tax-based selling by strategic partner Intel earlier this year, and/or constant speculation about the tenor of Clearwire’s relationship with Sprint, shares appear undervalued for a number of reasons that appear temporary in nature. Here are the most obvious – which shorts seem to cite, too.
1 – Funding and/or the threat of dilution.
Most notably, the market seems concerned about funding prospects for the company. This angst is way overblown. I’ll address the impacts of various funding scenarios on the per share intrinsic value of the company in a bit. For now, though, I would reiterate that (a) Clearwire will likely be raising $1 billion for capex purposes at some point over the next few quarters, (b) management has stated the company has enough cash to last through June of 2012, and (c) they have a number of funding options other than Sprint. More specifically, there are five: vendor financing, equity, debt, spectrum sales, and pre-payments associated with onboarding new wholesale customers.
Based on recent comments from the CFO, however, it seems best to assume a large portion of the capital will come through LTE vendors, with the remainder contributed through some combination of equity and debt. Any spectrum sales would be solely opportunistic, and pre-payments are too speculative at this point. As per the valuation section of this write-up, however, at recent prices the market seems to be pricing in a greater-than-$1 billion raise via straight equity sales. Anything less could be a positive catalyst for the shares.
The most important thing to know about Clearwire’s ability to secure vendor financing right now is this:
There are no other major LTE contracts currently up for grabs in the U.S.
If you presume, as I do, that LightSquared will forever remain an idea on a whiteboard, there will be no other major contracts in the immediate future, either. The competition between equipment providers to win the Clearwire LTE contract should be intense, and here’s to hoping it borders on the irrational. So in the most relevant near-term, Clearwire appears to be in an excellent position to negotiate attractive terms and conditions from vendor financing.
Ericsson and Alcatel-Lucent are clear leaders in the market today, and given Clearwire’s existing managed services relationship with Ericsson, it would not be a surprise to see Alcatel or current WiMax provider Samsung do something rash to secure the LTE contract. Nokia Siemens, too, might be more motivated than usual, given that its role in LightSquared’s planned network evaporated once it signed a network sharing deal with Sprint. I also believe Clearwire’s other current WiMax provider, Huawei could be desperate to bid on LTE business, too, given its recent success in Europe - and since the company was shut out last year on Sprint’s Network Vision contract. And Huawei has a reputation for generous equipment financing.
Regardless, this is all speculation. My point is that Clearwire should be able to extract quite favorable terms from vendors when it comes to financing. I also believe each of the above companies has previously extended vendor financing to companies in excess of the $1 billion Clearwire desires, so that size line should not be an issue.
Again, fears of massive dilution appear overblown.
2 - LightSquared.
The market appeared to interpret the signing of a network sharing agreement between LightSquared and Sprint early this summer as a rebuke of Clearwire by its largest wholesale customer. Overlooked, however, was the complete lack of any risk to Sprint in signing the deal, as well as the considerable technical challenges LightSquared must overcome to gain approval to provide service. There is a contentious and real GPS interference issue on the table currently which has provoked dire warnings from the GPS industry, DoD and the FAA.
Even if the company succeeds at navigating that challenge, the company may have another technological challenge to surmount in terms of latency in its network.
Latency is synonymous with delay, and it is a measure of the time it takes an IP packet to go from one point on a network to another. In satellite-based systems, latency has historically been high, which is problematic. The time it takes bits to literally travel 22,000 miles up and back again can negatively impact real-time data services like gaming and videoconferencing. To be fair, I believe LightSquared only plans to use satellite links for backhaul, but that, too, presents technical challenges. Regardless, even if the company has a solution, it seems rational to presume it could not charge as much or guarantee the same level of service to wholesale customers as delivered through a network like that of Clearwire. LightSquared is a high-latency solution in a low-latency world.
One more. Even if it can navigate these technical challenges, Clearwire’s LTE TDD technology provides it with much more efficient use of already wider channels. So if LightSquared is required to use guard bands to prevent GPS interference, the company is at further technical disadvantage.
Simply put, I am unconvinced that LightSquared will ever be a viable wholesale competitor to anyone, let alone Clearwire.
3 – Debt.
Yep. There’s over $4 billion of it right there on the balance sheet, plus another $2.6B more after capitalizing spectrum leases. But I’ll worry about this in three years. The company’s unadjusted D/E is .52 and its primary $2.7B due isn’t due until 2015, so it’s got time and breathing room. The company is arguably better positioned now to meet those obligations than ever before – which is why, of course, it decided to sell that debt last year to begin with.
4 – Inferior spectrum and/or technology.
Criticisms of the WiMax technology were valid, but that weakness was recently turned on its head with Clearwire’s decision to adopt LTE-Advanced.
With regards to spectrum quality - it is true that the 700MHz spectrum that AT&T and Verizon are using to build their LTE networks has historically been considered beachfront property. The relatively low frequency means that those networks can send data longer distances and penetrate buildings more easily than spectrum at higher frequencies, like that of Clearwire. In a traditional coverage-based model of service, then, almost all other spectrum is inferior, Clearwire’s included.
However, while Clearwire’s signals do not penetrate buildings well at all compared to other providers of retail data traffic, they don’t need to, either, in a wholesale model. The company is no longer trying to compete on the basis of the best retail coverage. It’s trying to increase other carriers’ capacity, and its spectrum is optimal for that. So between a de-emphasis on the retail business, a shift towards a capacity-based network, intra-building WiFi and multi-mode device chips, this is not a major concern to me looking ahead.
5 - Sprint.
Sprint is the largest shareholder of Clearwire, with a 54% voting interest and a 48% equity interest.
As per the Pardus Capital letter earlier this year, there is a perception in the market that Sprint won’t let Clearwire be successful, as it ultimately desires to buy-in the rest of the equity through a take-under. This smacks a bit too much of black helicopters to me.
Sprint no longer retains any board positions at Clearwire. Sprint has no input in, knowledge of, or control over any daily strategic or operational matters at Clearwire. It effectively just has a right of first refusal on any potential acquisitions.
Sprint also recognized the value of the wholesale model before Clearwire. I believe its agreement with LightSquared was primarily an attempt to defend an idea as opposed to a meaningful endorsement. Now, however, that potential partner appears to be unable to get out of the gate, and Sprint has its own financial constraints. That leaves Clearwire as the least expensive, most efficient way Sprint can have equity participation in a network hosting model.
I don’t doubt that is hard to swallow. But to theorize that Sprint would actively encourage the demise of its wholesale partner ignores a partnership enforced by revenue on one side and equity on the other.
What happens next between Sprint and Clearwire is subject to much speculation. The fate of both companies remains intertwined, though, which should be of some significance as this uncertainty is soon resolved.
6 – Competition.
Another perceived knock against Clearwire is that it will not be the first mover in 4G. I view this as another marginal issue that is being assigned more weight than it deserves given the aggregate uncertainty related to the five previous points.
This argument, to me, is a straw man. First, though, as a point of fact - Clearwire is already the first (and only) mover in the network wholesale business. And that will likely be a far bigger deal than the date on which another operator happens to land its first retail LTE customer. That aside, first mover advantage does not form any part of my own investment thesis in Clearwire. At all. And I’m not convinced it exists in telecom in any case.
It is absolutely true that competition on the retail side of Clearwire’s business will continue to intensify. But, again, the company isn’t trying to compete in retail anymore as much as it is trying to ramp the wholesale business. And while there are certainly other companies which would probably like to wholesale capacity, too, Clearwire’s cost structure and LTE TDD put it in a defensible position.
So how could you really kill this company?
Some sort of perfect storm of excessive dilution, technological woes/obsolescence and just plain old bad decisions could do it. But the easiest way to kill Clearwire would be to suddenly flood the market with a comparable amount of unencumbered spectrum of similar characteristics. Fortunately, it doesn’t exist. Nonetheless, spectral efficiency technologies are constantly improving, and this warrants watching.
It should also be noted that the FCC believes a national spectrum shortage in the medium-term is all but inevitable. As such, the agency has said it will make 500 MHz of spectrum available for mobile broadband use within the next nine years. A significant portion of that will potentially come from underutilized broadcast bands and the release of mobile satellite spectrum for terrestrial networks – as in the case of LightSquared. Again, that company’s recent woes are a good reminder of why its spectrum does not represent an immediate competitive threat to Clearwire. And given recent allegations of political patronage in the award of the LightSquared license, I suspect the FCC will approach any future auctions in bureaucratic super-slow-mo.
The sale of spectrum from private licensees who for whatever reason decide not to deploy a wireless network is worth keeping tabs on, too. AT&T, for instance, is in the process of closing on the purchase of Qualcomm’s 700 MHz portfolio. This does not strike me as a meaningful threat to Clearwire’s wholesale model, simply because there is no network equipment available anywhere that will work on it.
The other potential sellers to keep an eye on include SpectrumCo, a group of AWS spectrum license-holders, and Cox, which recently punted on building out its own 700 MHz network. Lastly, should AT&T have to jettison spectrum to close the T-Mobile deal, that, too, could merit attention.
In the end, however, I put a low probability on any of these scenarios meaningfully impacting Clearwire in the foreseeable future. That spectrum and network are unique.
The following discussion may border on the didactic, but I think it may be useful for the generalist, nonetheless.
Spectrum can be a challenging asset to value. It’s not unlike owning farmland – except, of course, that it’s invisible, constantly floats above your head and stretches coast-to-coast. The value of spectrum is also contingent, meaning it depends among other things on intended use, regional populations, and available technologies. The value of spectrum can also be dramatically affected by the supporting ecosystem – the number and depth of equipment manufacturers and handset providers – as well as usage restrictions, the presence of incumbent users, and existing interference.
To confidently value spectrum very accurately, then, requires a fluency in the sort of techno-babble that can make you pine to learn Latin. Fortunately, you don’t need to speak CPGA or build an elaborate model to know that Clearwire shares right now are very cheap. (To be clear, I did build a honking model, which I will post on SumZero, but in the end, I should have just gone fishing.)
So valuing spectrum as a stand-alone asset can be a challenge. Spectrum that is monetized via an all IP network like that at Clearwire also presents some additional relative valuation challenges. The metrics that have traditionally applied to voice-centric telcos are of significantly less utility on the data side but continue to dominate, anyway. I’ll rail against ARPU another time, but know that popular metrics fall short when it applied to just the wholesale business of Clearwire, where the optimal metric will one day be something like “average margin per Gigabyte.” Telco analysts are in need of some novel usage-based valuation models, but we’re just not there yet – particularly in terms of cost data disclosed by the operators.
In any case, I chose to value Clearwire two ways – on a relative valuation basis, and using EBITDA as a proxy in a DCF.
In relative valuations, the most common measure of the value of spectrum licenses is price per MHz-POP. It’s calculated by dividing the price paid for a license by the quantity of the population covered by the license and the amount of spectrum, measured in MHz (i.e. Price/(MHz*Population)).
So it takes into account the amount of bandwidth and the number of people who could potentially use it. What it completely fails to capture, though, are things like population density, propagation characteristics and licensing rules – hugely important characteristics that each play a significant role in spectrum valuation. For similar reasons, and as per those errant benchmarks I mentioned earlier, the dynamics of spectrum auctions may not be a means of true value discovery, either.
That said, ya gotta go with something.
What are Shares Worth?
There are at least five potential levers of mobile broadband revenues at Clearwire, including: (1) adoption of the company’s wholesale model, (2) increased penetration of smartphones, (3) adoption of tablets and mobile hotspots, (4) growth in devices like e-readers and machine-to-machine connections, and (5) increased per subscriber data usage across all platforms. So, in short, I assume Clearwire revenues will grow ~25% CAGR to $ 3.0 billion in 2015 driven by subscriber growth of ~43% CAGR, to 21.3 million subscribers.
Retail ARPU is expected to decline in my five-year forecast period, though the company maintains ARPU should stabilize in 2012. Wholesale revenue is expected to increase slightly as the company leverages its network to gain meaningful volume.
Near-term, I expect Clearwire to continue to report significant EBITDA losses, and estimate Adjusted-EBITDA will turn positive in 2013, ultimately increasing to $1.0 billion + in 2015. That would represent an EBITDA margin of 34%, towards the industry average. In contrast, the company believes it will be EBITDA positive in Q1 of 2012.
So given the high level of uncertainty around a wide range of variables here, I believe I’m using a conservative and arguably pessimistic approach.
Given the challenges in spectrum valuation, I believe that precedent transaction analysis is a more efficient way to proceed. I consider the most recent two industry transactions as reasonably good benchmarks for valuation purposes.
Precedent Transaction Analysis:
Deal 1: DBSD Spectrum Value Discount
MHz Owned 20
Population (MM) 300
Offer Amount ($MM) 1400 10%
$/MHz-POP $ 0.23 $0.21
Deal 2: Qualcomm Spectrum Value Discount
MHz Owned 12
Population (MM) 185
Offer Amount ($MM) 1925 30%
$/MHz-POP $0.87 $0.61
Average of Deals 1 and 2: $ 0.41
The DBSD/DISH acquisition was announced on Feb 1, 2011 and the Qualcomm/AT&T deal was on Dec 20, 2010.
I would also note that the transaction multiples above are at a discount to the 700 MHz valuation at Auction 73 (see below). I included further discounts of 10% and 30% based on the characteristics of each deal. Here are the specific factors I considered in assigning those discounts:
- 700 MHz spectrum is considered more valuable than 2.5+ GHz spectrum. One reason for that historically higher price in auction bids is that the superior propagation characteristics of low-band spectrum usually reduce the costs of a network build-out.
- In the AWS markets, I believe $0.20- $0.25/MHZ POP is a relatively consistent metric given the average of the small market bids in the 2006 AWS auction. (See below)
- The price in the DISH/DSBD deal was already depressed due to bankruptcy and related debt assumptions. I gave it another 10% haircut to make sure I was being conservative.
- AT&T’s acquisition seemed to be at a large premium, given it was buying unpaired spectrum for which there is presently no available equipment. That said, there may have been another angle to that premium related to the potential for interference with its own 700 MHz C-block licenses. Also, AT&T plans to deploy the spectrum by working with Qualcomm to integrate carrier aggregation technology, which could have tweaked the price. So, given all those factors, I discounted it by 30%.
Other Spectrum Transactions
Date Acquirer Seller Spectrum Band Price per MHz-POP
Oct-07 AT&T Aloha Partners 700 MHz $1.06
Feb-07 Clearwire BellSouth 2.5 GHz $0.18
May-06 NextWave WCS Wireless 2.3 GHz $0.13
I also checked historical auction valuations to make sure I was in the right ballpark for Clearwire. The tables below are effectively floor/ceiling checks - I was initially using them as bounds until I noticed all recent transactions were within the limits, anyway. Then it seemed to just make the most sense to use the latest two transaction multiples from above.
Historical: Weighted Average Price per MHz-POP - 700 MHz
Auction 44 -2002 Auction 49-2003 Auction 73 - 2008
0-250M $0.03 $0.03 $0.56
250-500M $0.02 $0.02 $0.83
500-750M $0.02 $0.03 $0.97
750-1MM $0.03 $0.06 $1.23
1-2MM $0.04 $0.04 $1.41
2MM+ $0.03 $0.03 $1.37
All $0.03 $0.03 $1.28
Historical Weighted Average Price per MHz Pops - AWS
Auction 66- 2006 Auction 78- 2008
0-0.12 $0.07 -
250-500M $0.13 $0.11
500-750M $0.17 -
750-1MM $0.20 -
1-2MM $0.33 -
2MM+ $0.65 $0.07
All Markets $0.54 $0.08
Comparison of Prices
License Size 700 MHz AWS
REAGs $0.76 $0.67
Eas $1.02 $0.47
CMAs $2.68 $0.40
RSAs $0.73 $0.12
POP 250k-1M $1.00 $0.17
POP <=250k $0.65 $0.13
Mean $1.29 $0.54
(Note: The acronyms above relate to the geographic distributions of the spectrum. REAGs – Regional Economic Area Groupings. Eas – Economic Areas. CMAs – Cellular Market Areas. RSAs – Rural Service Areas.)
Here are four factors that, within a given range, support a premium valuation for Clearwire:
1. Clearwire has significant spectrum of approximately 160 MHz covering a population of 135 million. Approximately 2/3rds of that population is in the top 100 U.S. markets. Additionally, there is an expected supply gap for spectrum in the industry, which in turn should push valuations higher - especially in the 700 MHz and BRS bands.
2. Many players are expecting to migrate to 4G, which simply might not be a possibility on their current bands. Sprint in particular is gaining a significant advantage using Clearwire’s deep and broad spectrum.
3. Clearwire is in a privileged position - not only does it have a huge amount of spectrum, but it was able decide whether to use TDD or FDD interfaces. TDD technologies are better suited to accommodate asymmetric data traffic, which has a much heavier downlink load, and is increasingly driven by streaming video and audio applications.
4. Given the looming spectrum shortage, Clearwire should be able to command above market price for spectrum sales, should they ever be pursued.
In contrast, here are three factors that, within a given range, support a discounted valuation for Clearwire:
1. A greenfield deployment at 2.5-2.7 GHz may need 2-4x more sites for the same coverage as lower bands.
2. Indoor penetration of 2.5-2.7 GHz would be extremely poor compared to 700 MHz.
3. Competition: Verizon has already launched its 4G service using LTE FDD technology. It is faster than the WiMAX technology currently used by Clearwire. AT&T is also expected to soon launch its 4G service that will also run on FDD. There is uncertainty around when Clearwire may be able to roll-out LTE Advanced as it would require a raise of $600M for capex. So it may not be before 2013 that Clearwire will be able to roll out LTE-Advanced, and the company may lose first mover advantage with regards to 4G.
Please note that, as per my earlier comments, while I would dispute the validity of each of these concerns, they could nonetheless impact a potential buyer’s perceptions of spectrum value. So I included them, regardless. (Also, please ping me in the comments if you're curious about my thoughts on the leased/owned mix of spectrum licenses. In short, I believe it's immaterial.)
Based on all the above, then, and assuming significant discounts, I believe Clearwire’s spectrum should be valued in the range below:
Scenario Pessimistic Base Optimistic
Spectrum Value (Mhz-POP) $0.21 $0.41 $0.61
Current Clearwire Spectrum (MHz-POP) 46500 46500 46500
Total Value $9,765 $18,995 $28,225
The value per share based on a MHz-POP basis is below:
Scenario Pessimistic Base Optimistic
Spectrum Value (Mhz-POP) $0.21 $0.41 $0.61
Current Clearwire Spectrum (MHz-POP) 46500 46500 46500
Total Value $9,765 $18,995 $28,225
Less: Net Debt 3,189 3,189 3,189
Less: Frictional Costs 195 380 564
Equity Value 6,381 15,426 24,472
Shares Outstanding 965 965 965
Price per Share $6.62 $15.99 $25.37
Estimated Equity Risk Premium 6.0%
Risk Free Rate 3.1%
Cost of Equity 15.1%
Nominal Marginal Cost of Debt 12.5%
Debt/Equity Ratio 0.52
Nominal Tax Rate 35.0%
Debt Cost of Capital (Tax-Adj) 8.1%
Cost of Equity 15.1%
Derivation of Terminal Multiple
EBITDA Multiple Range 5.0x
The risk-free rate is the 10-year US long term Treasury rate. While normally irrelevant, I plugged in an equity risk premium and Beta above to quickly derive a verifiable cost of capital. Clearwire’s cost of equity is ~15%, and since the nominal cost of debt is equivalent to the rate on the latest Dec. 2010 debt, using the current debt-to-equity of .52 results in a WACC of 11.5%.
For terminal value, I am assuming an EBITDA multiple of 5.0x. That represents a 15-30% discount to suitably relevant recent transactions which had EBITDA multiples of 6.0-7.0x, specifically:
Sprint Nextel/Virgin Media – 5.9x EBITDA
Sprint Nextel/iPCS – 7.9X EBITDA
(In $’000) 2011 2012 2013 2014 2015 TV
Total Revenue 1,228,014 1,783,371 2,159,082 2,650,704 3,013,644
Depreciation & Amortization 699,870 748,504 756,982 821,335 887,987
Proforma EBITDA (694,410) (125,401) 419,909 677,377 1,031,380
Tax Rate 35% 35% 35% 35% 35%
Post Tax EBITDA 670,397
Add: Depreciation Taxshield 244,954 261,976 264,944 287,467 310,796
Change in Working Capital 44,612 (45,205) 6,389 106,384 152,943
Capital Expenditures 485,688 624,180 496,589 530,141 542,456
Unlevered Free Cash Flow (979,755) (442,400) 181,875 328,319 285,794
Terminal Cash Flow 5,156,901
Free Cash Flow to Firm (979,755) (442,400) 181,875 328,319 5,442,695 5,156,901
Derivation of Fair Value
Gross Asset Value ($ 000) 5,524,732
Add: Cash & Equivalents 829
Less: Debt 3,950
Net Asset Value (NAV) 5,521,611
Diluted Shares 964.59
Fair Value per Share $5.72
So both the relative and fundamental back-of-the-envelope valuations triangulated reasonably well around a conservative value for Clearwire shares in the $6 range.
EBITDA Based $5.72
Average $/Share $ 6.17
Finally, I wanted to gauge the impact of the future $1.0 billion raise on the above. So, here is the per share valuation impact under three different scenarios.
a) Funding is 100% Equity
b) Funding is 50% Equity and 50% Debt (traditional or vendor financing)
c) Funding is 25% Equity and 75% Debt (traditional or vendor financing)
Assuming all the same assumptions carryover, here is the impact on the value per share of Clearwire given three different combinations of capital raise:
Valuation Share price at which CLWR raises funds
Method $ 1.00 $ 2.00 $ 3.00 $ 4.00 $ 5.00
DCF $ 2.75 $ 3.70 $ 4.17 $ 4.46 $ 4.65
MHz-POP $ 3.25 $ 4.36 $ 4.92 $ 5.25 $ 5.48
Average $ 3.00 $ 4.03 $ 4.54 $ 4.85 $ 5.06
50% Equity & 50% Debt/Vendor:
Valuation Share price at which CLWR raises funds
Method $ 1.00 $ 2.00 $ 3.00 $ 4.00 $ 5.00
DCF $ 3.77 $ 4.54 $ 4.88 $ 5.06 $ 5.18
MHz-POP $ 4.36 $ 5.25 $ 5.64 $ 5.86 $ 5.99
Average $ 4.06 $ 4.90 $ 5.26 $ 5.46 $ 5.59
25% Equity & 75% Debt/Vendor:
Valuation Share price at which CLWR raises funds
Method $ 1.00 $ 2.00 $ 3.00 $ 4.00 $ 5.00
DCF $ 4.59 $ 5.11 $ 5.32 $ 5.42 $ 5.49
MHz-POP $ 5.25 $ 5.86 $ 6.09 $ 6.21 $ 6.29
Average $ 4.92 $ 5.48 $ 5.70 $ 5.82 $ 5.89
Thus my earlier point that dilution fears, while a concern at the margin, are largely overblown. It’s also interesting to note that on a simple DCF basis, the current market price of Clearwire effectively assumes the company will raise more than a billion dollars, all in equity. This is unlikely.
So, to conclude, in the most pessimistic scenario, I believe Clearwire shares on a conservative basis are currently worth between $4 and $6, depending on the terms of the impending $1.0 billion raise. That value, again, is conservative.
While I am certainly not unbiased, I personally believe the “Base Case Scenario” above, which puts Clearwire’s true spectrum value at $0.41, is more realistic. If so, then shares are worth closer to $16 each. And for what it’s worth, you’ll recall that approximately $16 per share valuation was corroborated earlier in the previous quote from Glenview Capital Management. So the upside in shares if that first scenario proves too conservative will be significant.
- Sprint offers re-assurance of its relationship with Clearwire during its Q3 conference call tomorrow.
- Injection of capital via existing or new strategic partners, including DISH, cable companies and/or equipment vendors.
- Better-than-expected terms of the capital raise (i.e. significant vendor financing).
- Better than expected Q3 results. Independent of top-line growth, 700 Clearwire employees were transferred over to Ericsson on June 27th as per an outsourced WiMax managed services agreement.
- T, VZ, Metro or Leap become wholesale customers.
- AT&T makes favorable concessions in order to placate DOJ and close on T-Mobile.
- Short-squeeze based on any of the above.
|Entry||10/26/2011 12:30 AM|
Some additional thoughts I'd add to my original article above, for what they're worth:
1 - Here's a much briefer summary of why Clearwire's spectrum is so valuable:
"The bottom line on the spectrum is that Clearwire has approximately four times Verizon's spectrum, and it may be at least 50% more efficient, so that means six times. This makes all the difference in the world in terms of Clearwire's ability to compete and to do so with competitive pricing."
2 - This is a short post of mine on SA - "Thoughts on CLWR After Sprint's Analyst Day":
3 - Another recent SA post of mine, in which I responded to others' questions about the short case against CLWR, funding, why the technology matters, and LightSquared - "Q&A with Clearwire Longs and Shorts":
So, might be some more good background in the above, too. Thanks.