Save Access

There are many ways to parse the fights in the 700 MHz auction: incumbents v. new entrants, rural v. large incumbents, public safety v. commercial use, and the occassional suggestion by us in the public interest community. But, as I recently indicated elsewhere, an analysis of the band plan fight about large licenses v. small licenses reveals another interesting battle: Telcos v. Cable, with new entrants lining up with Telcos for large licenses and non-vertically integrated wireless carriers like T-Mobile aligning themselves with the cable-dominated consortium SpectrumCo.

What makes me believe license size in 700 MHz auction has become a new front in the fight between telcos and cable cable cos? Why has this new battleground emerged? And what are its implications?

See below . . . .

A quick perusal of Verizon's FCC filings in this docket show that Verizon has very few things it cares about in this auction, but has concentrated extensively on the surprisingly highly controversial issue of license size. Verizon is joined by Cingular (now AT&T Wireless) and with the potential new entrants DBS Wireless and the 4G Coalition.

On the other hand, the Comcast/TW/Cox/Sprint-Nextel consortium SpectrumCo LLC, favors predominantly small licenses. Indeed, it cares so much about this issue it submitted this white paper in favor of smaller licenses blocks all the way back in July before the comment period started. SpectrumCo position has attracted support not merely from the usual small rural providers, but from the wireless incumbents not vertically integrated with telcos: T-Mobile, MetroPCS and so forth. They are also supported by the “Balanced Consensus Plan,” which consists of an alliance of rural providers (which I generally refer to as “rural local exchange carriers” or RLECs — a term borrowed from the wired world) and mid-size wireless incumbents such as Alltel, MEtroPCS and US Cellular.

So what gives?

Here's my answer: the 700 MHZ auction has become the next front in the Cable v. Telco struggle for dominance. Verizon and AT&T want large license blocks for the same reason DBS and 4G want them, to construct a much needed national footprint of broadband-capable spectrum. SpectrumCo desperately wants to deny this, as do the incumbent wireless providers.

I freely admit my answer rests on speculation. I would love for those with the time and resources to do further digging and on this to tell me if they think I'm right or blowing smoke. But whatever the reason, the fact that Verizon and AT&T are on the same side as the 4G and DBS wireless guys and against the other incumbents, while SpectrumCo lines up with the pure wireless plays, seems to have blown by most folks not spending 24 hours a day breathing this stuff.

Why the 700 MHZ Battleground?

To understand this, we need to stop falling into the two biggest traps in analyzing the communications markets. People either want to think of wireless as entirely different from related markets, such as broadband and voice, or they want to think of them as all part of one big market and therefore fiercely competitive.

Us usual, reality turns out to be far too complex for such nice characterizations. What we have here are nested markets. If you ever studied set theory in math, you may recall that we can describe sets as having all sorts of different relationships with each other. Unsurprisingly, communications markets work the same way. Cable providers offer a variety of services and have a particular set of assets. So do telephone companies and everyone else involved in this incredibly complicated multiplayer game. Sometimes they compete, sometimes they ally. Sometimes they have nothing to do with each other. But as communications markets increasingly evolve, and as consumers increasingly start thinking (and companies increasingly market) in terms of total communications services rather than as individual unrelated services, market analysis gets very, very messy.

If it makes it easier, you can also think of it as a complicated ecology in which certain key players have evolved or moved into each other's niches. At one time, telephone companies and cable companies coexisted and did not compete. Now they do, on a variety of different fronts. In some areas, such as the delivery of residential video services, cable providers have a huge advantage. In other areas, such residential voice service, telephone companies have a huge advantage. What makes life even more interesting is that the regulatory environment makes it easier for cable operators to move into the voice market (because the law still requires telcos to complete voice calls originating on cable networks) than it is for voice operators to move into video (because it is much harder and costlier for telcos to get needed programming — particular in new programming services like video on demand — from cable cos).

Why Do The Telcos Want Large Licenses?

Why the new entrants want large licenses I have explained before. They need to build large enough footprints to challenge the incumbents and (in the case of DBS) overlap their existing service area enough to make “triple play” of video, voice and data possible. But if large licenses favor new entrants like DBS providers, why do incumbent telcos want large licenses as well?

The answer lies in the changing world of broadband and how it impacts the related communications markets. About 3-5 years ago, several things became obvious to anyone paying attention:

1) VOIP was going to happen, making it possible for cable operators to offer high-quality residential voice service without needing to go through the same expensive and blockable steps as competing local exchange carriers (CLECs). So even if the telcos won their unbundling and broadband deregulation fights against the CLECs (which they did), it wouldn't eliminate the threat of cable competition.

2) High speed internet access would become increasingly important as a residential service. Cable operators — who had upgraded their physical plants to hybrid-fiber coax to offer more channels in response to the possible threat from DBS — could use the same upgrade to offer high speed residential service. Again, even if telcos won their regulatory fights against ISPs, they could not exclude cable operators.

3) Because switching services is a huge hassle (i.e., the switching cost is high), customers are very unlikely to switch from one provider to another in order to save a few dollars a month. So a customer enticed to switch to cable for “triple play” is effectively gone forever — even if the phone service or data service turns out to be somewhat less good or more expensive after the introductory offer than what the telcos offer. The customer will stick rather than go through the hassle of switching back. Similarly, even if the phone company eventually offers its own triple play package, it will be too late to keep the customer that left for cable triple play earlier.

4) In other words, unless the telcos could find a way to counter the cable threat, they faced gradual erosion of their core residential voice monopolies.

On the plus side, because switching from phone service is a huge hassle (i.e., switching cost works in their favor here) the telcos had time. Nevertheless, anyone in the telco industry not in a state of denial recognized the need to react or face erosion down to just about nothing. So each telco considered how to react to the cable threat.

Qwest is such an undesirable target they didn't need to care. Outside of a handful of urban areas, such as Denver, Qwest faces huge costs for provision to poor markets. It's basically a huge Rural Local Exchange Carrier (RLEC) with a handful of “crown jewels” and a backbone that parallels the interstate highways. So Qwest did just about nothing, other than ramp up its DSL packages in cities.

Verizon, by contrast, faces intense cable competition — primarily from Comcast, but also from Cablevision, Time Warner and Cox. Verizon could not afford to do nothing.

So Verizon decided to become the compact network of excellence. Their service area includes the densely populated Northeast and high-income areas they picked up from GTE and MCI elsewhere in the country. Verizon therefore began investing in fiber, upgrading and maintaining quality on its wireless network, and selling off its rural systems where excellence costs too much to deliver.

AT&T opted to protect itself by shear size. It has become so huge, covering such a diversity of populations and geographies, that it will take a very long time to erode its market share significantly. Unfortunately for AT&T, its service area also includes some hugely expensive places to upgrade, service and maintain.

(BellSouth tried to bulk up on wireless by acquiring AT&T Wireless and leveraging its 2.5 GHz systems. But it ran out of time and surrendered to AT&T.)

Umm....O.K., but what's this got to do with the 700 MHz auction?

For both AT&T and Verizon, the 700 MHZ auction represents an opportunity to leverage existing assets and offer a set of services superior to that of their cable rivals. It also solves certain network problems. Both already offer national wireless networks and are beginning to bundle these into a “quadruple play” offering that they hope counters the cable “triple play” offering. 700 MHZ auction presents a way to boost their wireless quadruple play without cannibalizing existing business lines.

Consider Verizon. By 2010, the date when acquisition of the 700 MHZ licenses will start to make a difference (since it will take time to build the networks and they can't start operation until the broadcasters clear out in 2009), Verizon will offer FIOS speeds above 50 Mbps on its network (possibly higher). It could offer an additional national mobile package through 700 MHZ spectrum of 15 Mbps while not losing residential or business subscribers addicted to the higher speeds available via fiber. A Verizon national mobile network with those kind of speeds would present a huge challenge to cable operators — particular for the high-end residential customers and the business market customers that everyone covets.

The story for AT&T is a bit more complicated, because their copper network will still suck rucks unless they make the jump to fiber. But for them, national high-speed wireless may offer an even more powerful story. AT&T has a huge inventory of other wireless assets (huge spectrum available in 2.3 GHz, AWS, and other CMRS spectrum, even after divesting 2.5 GHz as part of their merger). Combined with the 700 MHZ licenses, AT&T could actually shift from offering service via copper to offering service via wireless — especially in those areas too expensive to upgrade profitably to fiber.

For both AT&T and Verizon, however, this means national footprint. At the least, it requires substantial overlapping coverage of their existing service areas. That's a heck of a lot easier to do if the FCC divides up the licenses primarily into a few sets of large licenses. By contrast, if the FCC adopts a plan similar to the AWS plan and supported by SpectrumCo: one set of large regional licenses and the rest in the smaller Economic Area (EA) and even smaller cellular market areas (CMAs), this becomes much more difficult. A multitude of small licenses makes it easier for rivals to block, increases expense of assembling a package, and ensures that critical spectrum will end up in the hands of commercial rivals. (More on this below.)

Accordingly, Verizon and AT&T would much rather duke it out with DBS and other new entrants for large licenses than try to accumulate enough small licenses to build a network capable of offering the kinds of broadband speeds and services that can challenge the cable operators (and the 4G networks of stand-alone wireless incumbents like T-Mobile and MetroPCS).

That Explains the telco position. Why do the cable guys in SpectrumCo take the opposite position?

The cable companies, by contrast, do not have the same set of assets or business considerations. For them, this is about expanding their services now that pure video has pretty much reached its limit, and about defending their core video monopoly from possible attack by telcos or from DBS. That means adding voice and data. Happily for them, they had to put that infrastructure in the ground anyway to expand capacity for digital tier to meet the potential threat from DBS. Now it pays huge dividends. As Brian Roberts observed after yet another “Comcastic” quarter (Q1 2007): “Triple play really is changing the company and it's the gift that keeps on giving, if you will.”

Unlike Verizon and AT&T, however, Comcast has no wireless network. It has repeatedly said it does not plan to enter the cell phone market (except in partnership with Sprint-Nextel). Yes, Comcast won huge amounts of AWS spectrum in the recent auction, but — as it keeps saying — it doesn't really have plans for that spectrum right now. In fact, as I have documented ad nauseam, Comcast basically entered the AWS auction to block DBS Wireless. After that, it picked up whatever licenses seemed handy on a theory that it could invest in the asset and just sit on it until they figured out how to make it pay. (Of course, this is rather the opposite of what classic auction theory tells us should happen. But the notion that the FCC or various ideologically motivated folks will change their minds in the face of reality seems so absurd that folks like Brian Roberts don't even try to disguise what they are talking about unless they are making a direct presentation to the FCC. It's friggin' pathetic....)

Anyway, to get back the point, the top cable companies that control SpectrumCo have no interest in acquiring a national footprint. On the other hand, they have a huge interest in stoping anyone from acquiring a national footprint. In a world with the spectrum divided up between a few sets of large reag licenses and a handful of CMAs to keep the rural guys happy, blocking both telcos and DBS Wireless and potential new players like our new Google Overlords becomes very expensive, even with the help of the existing wireless incumbents. OTOH, if SpectrumCo persuades the Commission to adopt a plan similar to the AWS plan, it becomes much easier to block a competitor from getting a national footprint — especially if the Commission continues to refuse to adopt other pro-competitive measures like anonymous bidding. All SpectrumCo and its wireless incumbent allies need to do is ensure that no one of the three (or more) potential competitors gets more than one or two of the REAGs (and not in geographic proximity). After that, the shear number of EA or CMA size licenses a potential new entrant would have to win makes it bloody certain that SpectrumCo and its allies (as well as interested bidders going after the smaller licenses in their own right) can block the telcos and potential new entrants from getting a national footprint — and ensure that enough spectrum ends up in their hands to make accumulation of a footprint through roaming agreements hideously expensive (more on that below).

But why are the wireless incumbents siding with SpectrumCo instead of their good buddies AT&T Wireless and Verizon Wireless?

Members of the spectrum oligopoly are not stupid. Comcast and the other cable members of SpectrumCo have gone to great length to explain to anyone who will listen that they do not intend to start their own mobile phone network in competition with the existing wireless incumbents. Sure, if the cable guys need to, they can eventually buy their way into the spectrum incumbent club by purchasing Sprint-Nextel or T-Mobile. But for the present SpectrumCo has made it as clear as possible that its members pose no competitive threat to the wireless incumbents.

By contrast, AT&T Wireless and Verizon Wireless would pose a huge competitive threat to the wireless incumbents if allowed to do so. The network they would construct with a national 700 MHZ footprint would out-class by an order of magnitude the existing wireless networks of the other wireless incumbents. In the past, when Verizon looked like it was poised to pull ahead of its sister wireless networks, it found itself targeted in one of the PCS auctions and forced to pay a huge premium for spectrum it wanted (this is why Verizon did not push back against anonymous bidding last time — they are much more interested in avoiding being a target than targeting others). The wireless incumbent club usually unites to keep out potential troublemakers, but it also works to keep any one member from getting too far ahead.

Unlike AT&T or Verizon, the other wireless incumbents do not have the assets to make the leap and offer a wireless “third pipe” themselves. They cannot expect to outbid AT&T and Verizon (and others) for a set of large licenses. They do not have fiber in the ground like Verizon or huge additional spectrum assets like AT&T. So a band plan that favors AT&T and Verizon (and potential new entrants) works against the wireless incumbents like T-Mobile and MetroPCS.

On the other hand, a plan with predominantly smaller licenses like the AWS auction plays to their strengths. The individual networks can focus on filling holes in their systems and acquiring other potentially valuable spectrum for future expansion/roaming negotiations. In addition, the wireless incumbents have a network of smaller allies and designated entities with whom they have significant material relationships. Combined with the reassurances from the cable cos that they will not offer a competitive threat, it makes perfect sense why the large wireless incumbents not associated with wireline telcos should side with SpectrumCo instead of with their usual buddies AT&T and Verizon. (Hey, it's nothing personal. It's just...business. YaknowwhatImean?)

Wait! What about the rural guys? Don't they need small licenses?

Rural guys certainly prefer small licenses, and with good reason. Small licenses let rural guys pick up licenses the big guys don't want. Like a lap dog begging for table scraps, the RLECs have great relationships with the major carriers. So the RLECs make a nice living building out cheap systems in high-cost rural areas and then doing roaming deals with the major carriers.

The question isn't whether to have no small licenses or lots. The question is how many smaller licenses v. how many larger ones. For one thing, the 30 MHz in the “lower 700 MHz” (channels 52-59) are already designated for mostly EA (“economic area”) and CMA (“cellular marketing area,” the smallest size) licenses. The existing lower band plan calls for auctioning 176 EA and 734 CMA licenses (as paired 6 MHz licenses) and one set of REAG licenses (12 MHz unpaired). So the big fight is really just about the 30 MHz in the upper 700 MHz band.

Even here, Martin's original band plan, supported by the telcos, the 4G Guys, and DBS Wireless, would have had 75% of the spectrum in large regional licenses and 25% in EA and CMA size licenses for rural guys. Since rural licenses rarely get challenged (the fights for CMA-size licenses happens in the wealthy areas), Martin argued this (combined with the lower 700 MHz licenses) would leave enough spectrum for the little guys.

But the rural guys argued that this wasn't enough. Their chief argument was that larger players shut out of the large licenses would fight them for CMA-size licenses as a consolation prize. The SpectrumCo plan (supported by the wireless choir) called for a “balanced mix” of licenses in the upper 700MHz as well as in the lower 700 MHz, modelled on the “highly succesfull” (from the cable and wireless incumbent point of view) AWS auction. Unsurprisingly, this argument once again proved succesfull in persuading Commissioner Adelstein (whom I think the world of, but who has a soft-spot for rural providers), who persuaded Commissioners Copps and McDowell to force Martin to roll the license size question over to the Further Notice. (It helped that members of Congress have weighed in on the side of SpectrumCo and the adorable little RLEC lap doggies.)

Couldn't folks put together a wireless network through roaming agreements like they did before?

Those who follow this argue that companies seeking national footprint will ultimately negotiate with each other for roaming rights the way they did in the mid-1990s. There are a bunch of problems with this argument.

The standard argument, made by Cramton in the above referenced Verizon filing refers to the huge cost and delay imposed on systems by making roaming agreements. As folks who study cellular deployment never tire of pointing out, the multiyear time lag between the major PCS auctions of the early/mid-1990s and the roll out of national plans in the late 1990s and their initial high prices flows in no small part from the very expensive process of aggregating roaming agreements.

This alone would give possible national new entrants serious heartburn. From the telco point of view, the more people who disconnect their telco wire and go just cable, the harder it is to remain competitive. Because, as I keep saying, switching cost makes people very reluctant to change even for better service at lower cost. If the telcos must spend years making the necessary roaming arrangements, a national wireless footprint that goves them an edge will come way too late.

The same thing applies even more powerfully to potential new entrants. The DBS guys already face the challenge of losing subscribers and an inability to pull video subscribers away from the big cable companies because they cannot offer a competing broadband (and voice) package. It's why Rupert Murdach has taken to calling DIRECTV the turd bird and is selling his interest in it to John Malone. The DBS providers also cannot afford to spend years trying to build a significant footprint through roaming. Assuming they could do so.

As for genuine new entrants, such as our much mooted Google Overlords, spending years and heapin' quantities of cash before even entering the market, only to find the market locked up by the existing incumbents, is simply out of the question. And again, this assumes they will actually find 700 MHz licensees willing to negotiate roaming deals at all.

Which brings me to reason number two why the “aggregate through roaming agreements” won't work. SpectrumCo and the wireless incumbents — the companies that will have the bulk of the needed spectrum post-auction — have no incentive to negotiate roaming agreements with potential rivals. To the contrary, they have every incentive to avoid cutting a deal. Again, people need to realize it is no longer the 1990s and communications markets have changed.

In the mid-1990s, no cellular company had market power and they all needed to develop a national footprint. This is the classic case in which all relevant players have incentive to work with each other. Even though the cellular players ultimately needed to compete with one another, their common interest in creating national networks and the lack of any signle-player advanatge made it possible to conclude mutual roaming agreements. In addition, companies with no interest in going national — but with lots of interest in getting big roaming fees from the large companies that wanted to go national — had every incentive to make deals with everyone who wanted to roam and no pressure to cut exclusive deals.

The current environment is very different. The large cellular companies have consolidated, and are likely to hold to bulk of necessary spectrum to build a substantial 700 MHz footprint. SpectrumCo will also likely grab a significant number of these licenses. If nothing else, SpectrumCo will try to acquire enough licenses to influence how the 700 MHz band evolves and block any challenge to the major cable cos in their core areas of dominance. And, where smaller companies hold licenses, they will have plenty of incentive to take a big fat payment from SpectrumCo or the stand alone cellular companies for exclusive deals that prohibit (or make very costly) deals with the telcos or potential new entrants.

In other words, we can expect post-700 MHz auction that significant spectrum holders will actively resist roaming agreements with the parties that want predominantly large licenses. This industry dynamic changes the universe significantly from the mid-1990s, when the only relevant market was the mobile phone market and the players all had similar incentives and limited market power. If it was an expensive and lengthy process to negotiate roaming agreements given the industry dynamic of the 1990s, it will be practically impossible to negotiate roaming agreements now — particularly to provide the kind of hyped-up broadband/4g services the telcos and the new entrants need to provide to compete effectively.

Yes, one can imagine hypothetical scenarios where a combination of telcos and potential new entrants win enough spectrum to create a viable national provider. But this strikes me as extremely unlikely unless the FCC also adopts other rule changes to facilitate new entry and prevent blocking or signalling by SpectrumCo and the wireless incumbents. Designating 30 MHz for open access and/or prohibiting incumbents from participating in the auction could also overcome this problem, by creating a reserve of spectrum for anyone to use and by ensuring it ends up in the hands of companies that have no incentive to block construction of a national footprint by rivals. Unsurprisingly, the telcos and the potential new entrants that want to run their own networks — like DBS Wireless — are adamantly opposed to adopting open access to offset the anticompetitie aspects of smaller licenses.

So, Harold, what's the best policy from a public interest perspective?

In my not-so-humble-opinion, the FCC needs a mix of remedies to open up the 700 MHz band in a way that will have a powerful, positive transformative effect on the wireless industry and the broadband industry. That means getting as many new competitors out there as possible, enabled to offer genuinely radical new services rather than just pumping up incumbents to offer more powerful versions of their own crippled captive networks. To do this, the FCC has to stop looking at the questions of licenses size, service rules, and auction rules as unrelated separate questions and think about how they all work together.

I think the best combination is a plan that favors large reag licenses but including the Frontline proposal (Proposal 4, figure 9 in the Further Notice). Without Frontline, I would support the “4G-DBS” plan (Proposal 1, Figure 6). I would reject the “SpectrumCo plan” (Proposal 2, figure 7) under any set of circumstances.

I also don't think that the license size issue solves all the new entrant problems, so I continue to support either preventing incumbents from bidding or providing some sort of credit for new entrants. I also think that the PISC open access proposal (30 MHz combined in upper and lower 700 MHz bands) provides the best mechanism for creating new competitors by allowing other wireless players (using both licensed and unlicensed spectrum) to boost their existing systems with choice spectrum and allow wholly new competitors to emerge by leasing spectrum (possibly to use in conjunction with other facilities — such as BPL).

At the end of the day, however, the important thing is to consider the synergistic impact of all of these proposals rather than treat them as unrelated. If the Commission decides to adopt the SpectrumCo license size band plan, it needs to determine how it will offset that advanatge to cable and wireless incumbents elsewhere. Similarly, if the Commission declines to take other steps to facilitate new entry, it needs to consider carefully whether it can afford to give SpectrumCo and the wireless incumbents their choice of band plan.

Stay tuned . . .