The Blog

Right now, DOJ’s Sharis Pozen is holding a press conference to announce that DOJ is suing to block the merger. This is great news that shouldn’t be surprising. It was a terrible deal in a host of ways.  But even without the deal we still have a problem.  Here’s why:

• The suggested merger shed light on the fact that we are heading towards – and may already have – a duopoly in the market for wireless access, with a yawning, insurmountable gap between the two big wireless carriers and everyone else;

• Even without the merger, there are insufficient protections in place for innovation in connection with these wireless networks. The existing unregulated duopoly will have ample incentive and ability to keep profit margins as high as possible by discriminating against uses and services these companies believe are undermining their business plans. The mere threat of this discrimination casts a cloud over investment in new ideas and new ways of making a living for all Americans.

Society reaps increasing economic returns from the existence of ubiquitous high-speed communications infrastructure, and it is appropriate to incentivize and support the creation of this infrastructure. But those social returns come because the benefits of ubiquitous, general-purpose, nondiscriminatory communications infrastructure spill over to all of us, not just to a few large companies. The complete discretion already enjoyed by AT&T and Verizon to capture economic rents and choose winners and losers from among the companies that use their networks to launch businesses of their own – and to raise their prices and set the terms and locations of their services at will – means that their private incentives are already not necessarily aligned with our nation’s social incentives.

The merger between AT&T and T-Mobile would have further solidified this duopoly by removing from the field a challenger that was providing an open platform to application developers, offering lower-price options for consumers, and taking on the duopoly in the policy realm in Washington. (And it would not have necessarily resulted in greater investment in infrastructure, capacity, or innovation by AT&T; indeed, as in the merger between Comcast and NBCU approved earlier this year, it is difficult to identify the public interest benefits of the combination.)

The Context for the AT&T/T-Mobile Merger

We have had a highly-concentrated market in wireless transmission for some time at the national level, and the merger would have increased that concentration. It would have given AT&T and Verizon, together, somewhere between 76% (all retail subscribers) and 82% (post-paid) of wireless subscribers nationwide and would have increased Herfindahl Hirschman Index (HHI) levels to 3198, far above the 2500 level that DOJ/FTC consider to be highly concentrated.

AT&T argued strongly that the relevant markets are local, and that there is plenty of competition in local markets because Metro PCS, Leap, US Cellular, and the (not-yet-operating and apparently doomed) LightSquared are also present. This is like asserting that my former hometown of Washington, DC has several football teams:  the Redskins, the Georgetown University team, and the Gonzaga High School team. It’s strange to say that the last two are substitutable for the first. AT&T and Verizon provide reliable nationwide service without extra roaming charges; the pre-paid players offer only unsubsidized handsets, routinely impose roaming charges, have puny data plans, and reach a much less affluent segment of the American public.

These companies have become consolidated for very good reasons. Wireless is a high-fixed-cost business; it costs a great deal to install towers, feed them with wires (and update those wires to fiber), and buy spectrum. Indeed, wireless access service has all the hallmarks of a natural monopoly, with its high up-front costs, sharply declining costs to add additional subscribers, and barriers to entry in the form of tower-siting approvals and licenses to use spectrum. Very few companies are able to achieve the scale necessary to make a go of it, and so they routinely combine rather than compete.

When the cellular phone emerged as a consumer product in the 1980s, it operated in 800 MHz frequencies, for which the FCC initially gave away two licenses for 40 MHz of spectrum in each of the 306 market areas in the United States – one to a wireless provider and one to a wired provider. Small-market licenses frustrated the buildup of viable nationwide wireless infrastructure; companies in urban areas only had a few voice channels, which wasn’t enough capacity to serve demand, and companies in rural areas couldn’t produce enough revenue to survive. No one could operate at the scale needed to make the business worthwhile.

The 1980s licensing process led, predictably, to quick consolidation and market-division agreements among the applicants. This desirable “beachfront” low-frequency spectrum went to the corporate ancestors of today’s AT&T and Verizon. It represents a significant windfall advantage to these companies that cannot be replicated by Sprint or T-Mobile, much less the pre-paid providers; there is, as a result, an enormous gap between AT&T and Verizon, on the one hand, and everyone else, on the other, in terms of subscribers, revenues, margins, and free cash flow.

This gap only increased following the FCC’s 2008 700 MHz auction, for which no spectrum caps were imposed. AT&T and Verizon collectively accounted for more than four-fifths of the auction proceeds, spending almost $20 billion; because it was clear to T-Mobile that the foreclosure value to the two giants of keeping a new competitor out of the arena would exceed any reasonable market value for this spectrum, and because the giants weren’t barred from the game even though they already had enormous holdings in desirable “beachfront” low-frequency spectrum, T-Mobile didn’t even enter the 700 MHz auction.

I want to underscore the importance of T-Mobile’s inability to get access to “beachfront” lower frequencies. Data rates go up with T-Mobile’s higher frequencies (1900 MHz PCS band, 1700-2100 MHz AWS band), but the distance data can travel goes down. So when you get into the gigahertz bands, which is where T-Mobile is today, you can indeed carry gigabits of information, but you might have to have cell towers at an impossible every 100 yards in order to do that. (This explains why Wi-Fi (high-frequency) speeds are faster than commercial wireless speeds but travel only short distances.)  This means that a carrier with “better” (lower) spectrum can build fewer base stations, which is a major cost advantage. These spectrum issues have created a yawning gap between T-Mobile (and Sprint) and the two big wireless carriers, with their broad, unchallenged holdings in the 700 MHz and 850 MHz bands.

Nonetheless, T-Mobile’s management told investors in January 2011 that the company was “a very good asset” that was generating positive annual free cash flow of between $2.5 billion and $3 billion a year, with a strong network architecture in which half its towers were already fed by fiber, with terrific smartphones, the best value proposition for consumers, great customer service, and with higher (and growing) margin on revenue than Sprint. Why the optimism? Management also said during that same call: “We’re absolutely positive and optimistic about [the] commercial option in [the 700 MHz] D block.” (“T-Mobile Investor Day – Transcript,” Jan. 20, 2011.) When the Administration appeared to take the possibility of auction for commercial use of the 700 MHz D Block off the table not long ago, Deutsche Telecom apparently could not see a path forward and decided a merger was the best route.

I do not believe any of the other pre-paid carriers exercise any pricing or service discipline on AT&T. I’m also worried about T-Mobile’s future.

I also suspect, given my observation of the cable industry, that Verizon Wireless and AT&T tacitly divide the market in California for wired services that they provide (“you take Sacramento, I’ll take LA”) and for the provision of backhaul middle-mile links to one another. I urge you to look into this matter; 95% of any wireless network is made up of wires, and Politico reported in late June that the independent backhaul business is threatened by an alleged reciprocal arrangement between AT&T and Verizon to provide infrastructure to connect each other’s wireless data traffic. (“AT&T-Verizon Pact Alarms Backhaul Provider,” June 21, 2011.)  It is beyond question that Verizon and AT&T are in a position to tacitly collude, divide markets, and protect their joint interest in pricing power over wireless services.

We Face A Duopoly for Wireless Services

This is a business that requires operating at scale in order to survive and thrive. Right now, only Verizon and AT&T have real scale in wireless. The suggested merger sheds light on the fact that we will have a duopoly in the market for wireless access, even if the merger itself never happens.

Let me explain. Even though most of AT&T’s and Verizon’s assets are actually wires, not wireless connections, both companies are losing money hand-over-fist on the wired side of their businesses. Pokey DSL high-speed Internet access, which operates over traditional copper phone wires, cannot compete with the cable distributors’ DOCSIS 3.0 connections; the differential is arguably more than Redskins v. Gonzaga High – it’s akin to rushing river v. water fountain. Wall Street investors can’t stand the long-term nature of the investments necessary for these telephone companies to install new fiber networks that would compete with DOCSIS 3.0, and so the telcos have dramatically backed off on these investments. Cable snaps up 90% of all new wired broadband subscribers these days; Americans love cable’s high-speed wired connections, and new high-data applications are continuing to drive this love affair.

This merger came at a time in which Americans are thirsty for high-data-rate applications. The mobile world – in which communications operate in a harsh world of interference and degrade sharply over distance – cannot compete with the wired world when it comes to data transmission. Each “wire” in a wireless network (each tower) has to serve 436 times as many homes as a cable network, which serves one home at a time; each wireless network spot has 1/37th the capacity as a cable wire; and there is vanishingly low interference inside a cable network.

So AT&T and Verizon have no choice but to focus entirely on the separate world of wireless, where they are still wringing out some profits. But they face enormous threats on the wireless side as well. Their margins for voice services are ten times higher than their margins for data services, but Americans love data services and data usage is skyrocketing. AT&T and Verizon have every incentive to ensure that their wireless operations keep data usage as low as possible for as long as possible by managing scarcity: imposing usage-based billing and avoiding installing fiber to their towers wherever they respectably can.

Usage-based billing allows AT&T and Verizon to reduce the attractiveness of potentially competing data services and other nonaffiliated products crossing their wireless networks; consumers won’t want to use competing services, even if they technically can, because they’ll be subject to large overage charges, cut-off of service, or other remedies imposed within the carrier’s discretion.

Avoiding capital-intensive fiber installations where possible will both please investors and continue to shape users’ expectations; we’ll be used to relatively slow, crippled, heavily-curated and compressed mobile services as the status quo.

Adoption of the LTE protocol will give AT&T and Verizon enhanced technical ability to charge premiums for their own or affiliated products and afford them “Quality of Service” treatment; LTE is a protocol that is optimized on billing.

All of this is rational; it is done in the service of keeping margins as high as possible so as to please Wall Street.

AT&T’s choice to merge with T-Mobile made sense; it is trying to force a natural monopoly, utility communications service (like water and electricity), with its extraordinarily high upfront costs and sharply declining cost curves, into a private, profit-making, Wall-Street-attractive service, and the only way to do that is to continue to scale, tightly ration capacity, price-discriminate, keep capital costs down, and eliminate competitive ideas – like the low-priced services, open development platforms, and policy positions pushed by T-Mobile in DC – that would undermine this mindset. Folding in T-Mobile held the potential for AT&T to add subscribers without adding to its employee headcount, and would have permitted it to possibly lower its ratio of employees to subscribers.

By the way, AT&T said that it needed to merge in order to address the flood of data across its network, but even doubling the available spectrum for wireless broadband wouldn’t change the laws of physics – we’d be out of capacity again in a year or so if usage growth isn’t stemmed. So the only way for AT&T and Verizon to keep on top of the situation is to maintain the market power that makes rationing the new normal for all American consumers.

All of this is to make one simple point:  AT&T and Verizon’s top priority must be to maintain and enhance their scale advantages and market power in the separate world of mobile communications in order to keep their margins high and please investors, and this merger serves that end.

The Duopoly Threatens Innovation

The future question is what regulatory environment will make sense given the utility, natural-monopoly nature of wireless transmission services. Someone is bound to notice that these actors are trying to pretend they are media companies when, in fact, they are more frequently seen by Americans as basic transport providers; what they are providing is access to the Internet.

The Internet Changes Everything. Let’s spell this out a bit. We used to assume that there was a necessary association between a particular form of infrastructure (like telephone and cable wires) and a particular functional capacity. So we assumed that each wire could do only one thing, and we had to have a separate network for each thing we wanted to do. This led to business models where a network owner was also the provider of whatever particular service—phone, cable, etc.—was carried over that particular kind of wire.

The Internet has completely overturned that assumption. The Internet is best understood as a collective agreement to use a particular language (the Internet Protocol) when connecting computing machines to telephone, fiber, and cable lines that are interconnected around the world. The innovation of this language was to allow computers or other devices connected to the Internet (including wireless handsets, televisions, fax machines, and TiVOs) to send and receive information of any kind via data streams over many different types of physical wires or fibers or wireless transmissions. The Internet Protocol can run over anything. And any different use (phone calls, television, news) can be communicated over the Internet Protocol. These uses may be provided by the network infrastructure owners, or they may be provided by other people (including any one of us).  Phone services can come from Skype—over the Internet. Video on demand can come from Amazon’s online movie rental store.  Television shows can come directly from the servers of users. And so on.

“The Internet” is thus not the same thing as Verizon’s or AT&T’s lines, fibers, or wireless connections. Though that infrastructure is important, these actors are merely providing one set of connections that allow users and businesses to connect to the constant, dynamic, decentralized interaction and communication using data that the Internet Protocol facilitates.

The Internet threatens vertically-integrated private carriers seeking high returns. Now, the Internet is taking over the functions of all of the communications networks we used to have. Each of the vertically-integrated network access providers in this country sees this change as a threat. AT&T and Verizon offer their own television services, music services, and premium content. The open Internet is the greatest competitor they have ever seen—precisely because it is not one competitor, but a general-purpose vehicle for thousands of entrepreneurs across the country, and most particularly in California, to offer innovative new products.

As things stand now, both of these dominant network access wireless providers has the freedom to act as an editor or gatekeeper for its own commercial purposes. They would like their services to be much more like cable programming than general-purpose communications—edited and constricted communications offerings. AT&T and Verizon have succeeded in persuading federal regulators that they should not be treated as communications providers, and see the future potential for vertically-integrated services that they control and monetize. They have a giant built-in conflict of interest.

Verizon and AT&T are fans of the cable distribution model. Although they’ve lost the battle on the wired side of their businesses to the cable companies, Verizon and AT&T are admirers of the cable distribution business model and the tens of billions of dollars it throws off in fees each year to programmers and distributors alike. The major Internet companies that support this merger (Microsoft, Facebook, Qualcomm, Oracle, Yahoo!, Avaya, Brocade, venture capital firms Kleiner Perkins Caufield & Byers and Sequoia Partners, and BlackBerry maker Research in Motion) are also fans of this model. You can think of Facebook as the ESPN of the wireless world: It’s an addictive, must-have channel that has deep exclusive relationships with its audience and the power to command prime, first-tier fees and placement from the distributors. AT&T needs Facebook, just as Time Warner Cable needs ESPN.

In this model, both programmers and distributors win. One loser is the consumer, who faces ever-increasing costs and homogenized choices that have been edited by a powerful distributor; another loser is the innovator – independent channels that don’t want to pay the freight to the distributor or otherwise meet whatever arbitrary conditions the dominant carrier imposes don’t get carried. (Notice that Al Jazeera can topple authoritarian regimes but cannot get carried by Comcast.)  Facebook could not have been born absent the regulatory regime that allowed for introduction of online applications without the permission of communications carriers; now that Facebook has achieved massive scale, it is willing to shut the door of innovation behind it.

To make the whole thing work, you need gatekeeper control at the distribution level so that price increases can be forced down, services can be tiered and channeled and charged-for reliably, and consumer expectations of an expensive, bundled service can be set and maintained. This merger, backed by a $20 billion non-recourse loan from JP Morgan, ensures that AT&T’s strength as a wireless distributor will stay in place.

The problem for America presented by deregulation in both the wired and the wireless worlds is that the increasing returns made possible by ubiquitous Internet access will be captured in the form of revenues that will go only in the pockets of a few very large companies. That kind of spillover won’t necessarily serve all of our interests. What we should want, instead, are increasing returns for society as a whole. Increasing returns for all Americans will be prompted by new ideas and new ways of making a living. Those ideas, in turn, will be facilitated by widespread, open, ubiquitous Internet access, which has been and will continue to be a playground for innovation unlike any the world has ever seen.

Instead, we have handed to the carriers the ability to decide what innovations make sense for America. We have created through deregulation a context for communications that is the equivalent of the “orderly marketplace” beloved of the early 20th-Century trusts and combinations.

The lack of network-operator competition we currently have, and the concomitant control the operators have, provides the opportunity for discrimination and gatekeeping on a mammoth scale when it comes to new data applications, new uses of these networks, and new devices. After-the-fact rationalizations for “management” of these networks (“discrimination” using a more neutral name) are so easy to craft. The real danger to innovation is the pervasive threat inherent in the ability to “manage.” An application developer unwilling to “partner” with the carrier cannot attract investment, because the network provider may degrade the functionality of the application at any time – imagine a highway designed to favor only particular kinds of cars at particular moments, and then imagine the frustration of an auto entrepreneur with a new kind of design ready for funding. Arbitrariness, by itself, is enormously threatening to innovation. The risk to American innovation is that almost anything – including discrimination for commercial reasons as well as viewpoint-related reasons – can fit within “reasonable network management.”

Because the AT&T/T-Mobile merger would have solidified AT&T’s and Verizon’s ability to dictate the business model for wireless communications in America, and would have ensured AT&T’s continued ability to decide what “programming” it “carries,” what devices it permits to attach to its networks, and on what conditions, it would have enhanced the arbitrariness of the mobile platform as an innovation ecosystem in America.

That can’t be good for the development of the new ideas and new ways of making a living that our country needs so much.