Antitrust & Competition Policy Blog

Editor: D. Daniel Sokol
University of Florida
Levin College of Law

Monday, May 23, 2011

Comments of Richard Brunell on At&t/T-Mobile Merger

Posted by Richard Brunell

The relevant geographic market may be an important consideration in evaluating the merger because if there is anational market for wireless services, then the merger reduces the number of players from 4 to 3[1] – suggesting a fairly strong presumption of illegality – and the potential remedy of divestitures to local or regional carriers in certain local markets would be ineffective because it would not replace the loss of a national competitor.

The relevant geographic markets are likely to be both local and national. While some competition for wireless services is local, other competition (among the national carriers) is primarily national, as illustrated by the billions of dollars spent in national advertising. In its acquisition of the regional carrier Centennial, AT&T claimed that “the predominant forces driving competition operate at the national level. . . . AT&T establishes its rate plans and pricing on a national basis . . . in response to competitive conditions and offerings at the national levels – primarily the plans offered by the other national carriers.”[2] AT&T explained that its plans were uniform throughout the country for efficiency and marketing reasons, and that “[v]ery infrequently,” it may offer a local promotion.[3] In contrast, in its current application to acquire T-Mobile, AT&T emphasizes the “the local nature of this marketplace,” but does not suggest that pricing of its service plans is done on anything other than a national basis. Rather, any local promotions appear to be limited to handsets and peripheral devices.[4]

In the past, the DOJ and FCC have considered only local geographic markets in wireless mergers, but that is because they have not previously reviewed a merger between two national carriers. Indeed, in recent wireless mergers, the DOJ has emphasized that “[t]he existence of local markets does not preclude the possibility of competitive effects in a broader geographic area, such as a regional or national area . . . .”[5] Insofar as competitive effects may occur on a national level, it is appropriate to define a relevant market that is national in scope. This is consistent with the revised Horizontal Merger Guidelines, which provide that “The hypothetical monopolist test . . . does not lead to a single relevant market. The Agencies may evaluate a merger in any relevant market satisfying the test, guided by the overarching principle that the purpose of defining the market and measuring market shares is to illuminate the evaluation of competitive effects.” U.S. Dept of Justice & Fed. Trade Comm’n, Horizontal Merger Guidelines § 4.1.1 (Aug. 19, 2010). It is also consistent with United States v. Grinnell Corp., 384 U.S. 563, 575-76 (1966), in which the Supreme Court held that the relevant geographic market for accredited central station protection services was national because it “reflect[ed] the reality of the way in which” the business was built and operated, even though the service was provided on a local basis.

Regional and local wireless carriers are not participants in the national market because they do not offer services on a national basis. (They offer national roaming but that does not mean a person located in a roaming area can become a subscriber.) Moreover, their local offerings do not appear to affect the national post-paid plans offered by the national carriers. Indeed, the regional carriers like MetroPCS and Leap/Cricket offer a different product and serve different market segments. They offer only pre-paid, non-contract services, and tend to serve customers with poor credit histories; they do not market to businesses. Further, smaller and regional carriers are limited in the competition they can provide to the national carriers because they lack brand names like those of the national carriers built up by years of intensive advertising, lack the array of smartphones offered by the national carriers, their networks are perceived to be of inferior quality, and they must depend on expensive roaming agreements with the national carriers. Accordingly, even if geographic markets are defined as local, the competitive significance of the local and regional carriers on the AT&T/T-Mobile combination is probably minimal.

-------------------------------------------------------------------------------- [1] Or, as the American Antitrust Institute (AAI) said, “more realistically, [from 4 to] 2 1/2, since the merger may have the effect of marginalizing Sprint as a competitor.” Letter from the American Antitrust Institute to Chairman Herb Kohl, May 16, 2011, available at [2] Merger of AT&T Inc. and Centennial Comm’cns Corp., Description of Transaction, Public Interest Showing and Related Demonstrations 28-29 (Nov. 21, 2008). AT&T stated it “focuses on the other national carriers in its competitive decision-making and does not consider Centennial in deciding on pricing and service offerings.” Id. at 37. AT&T made a similar claim when it acquired the regional carrier Dobson in 2007, explaining, “Where national competitive forces determine prices and the same products are offered nationwide at the same price, the relevant geographic market is national, rather than local.” Merger of AT&T Inc. and Dobson Comm’cns Corp., Description of Transaction, Public Interest Showing and Related Demonstrations 19 n.74 (July 13, 2007). [3] Description of AT&T/Centennial Transaction, Declaration of David Christopher Chief Marketing Officer ¶ 6. [4] See Acquisition of T-Mobile USA, Inc. by AT&T Inc., Description of Transaction, Public Interest Showing and Related Demonstrations 74 (April 21, 2011). [5] United States et al. v. Verizon Commc’ns Inc. and Alltel Corp., No. 1:08-cv-01878, Competitive Impact Statement at 7 n.2 (Oct. 30, 2008); United States et al. v. AT&T and Centennial Commc’ns Corp., No. 1:09-cv-01932, Competitive Impact Statement at 6 n.2 (Oct. 13, 2009) (same).

| Permalink

TrackBack URL for this entry:

Listed below are links to weblogs that reference Comments of Richard Brunell on At&t/T-Mobile Merger:


I've started writing what has become this comment as a comment to each of the posts that have been made so far -- but given that you discuss geographic market most directly, this seems an apt place to share these thoughts.

Whether the geographic market for mobile wireless telecommunications services is national or local is in many ways a $39 billion question -- in other ways (I think) it is a relatively unimportant question in that we are likely to reach the same outcome no matter whether our analytical starting point is local or national. In any event, it appears to be a misunderstood question, which I think can benefit from a more parameterized analysis.

Let me start with what the question is, and also what the different question that others seem to be answering is. Geographic market definition is not about where these services are used or are available. If AT&T, Leap, Cellular One, and Verizon are all offering the same service in a proximity to one another, at least all four of those carriers are in the same geographic market -- it doesn't matter whether their services are also available in other markets. The geographic market question focuses on where customers buy their service (in this case it is analyzed under section 4.2.2 of the HMGs, because carriers sell service by zip code -- as can seen by searching for service on any of their web pages), and how far they would travel to get the same service if faced with a SSNIP in their current proximity. Customers in Virginia are not going to go to Oklahoma to get their wireless service; and even if they wanted to, the carriers would bill them at Virginia rates, based on their billing address. It is fair to ask whether the carriers do offer meaningful localized pricing; it also might be reasonable to ask whether the zip-code-based price discrimination is problematic (more on this at the end of this comment). It is simply wrong, however, to say, under the modern approach (cites to not-so-modern Grinnell aside) that a service is part of a national geographic market because it is a national service.

Even to the extent that I am somewhat sympathetic to the view that the local-market approach is detached from reality (and I may be), at a practical level, it is hard to imagine litigating a case like this arguing a national market. No matter whether we start with a local geographic market definition, where more concentrated markets are divested, or a national geographic market where we whittle areas of competitive concern away, we are likely to end up in a similar position. Imagine: DOJ walks into court and says that this is a 4-to-3 merger in a national market. What does AT&T say? They tell the judge that they will -- and always have planned to -- offer up all the divestitures that DOJ might have required under a local-market analysis. AT&T has just turned this into an acquisition of a regional carrier, filling in gaps or weak areas in its network, or a 5-to-4 (or a 4-to-unconcentrated-3) in various local markets. It is hard to understand how DOJ could, putting aside whether it should, argue against that transaction -- especially given that it faces the burden of proving that this transaction, with the AT&T-proposed divestitures, likely substantially lessens competition.

The question that most seem to be addressing is about product market, not geographic market: are the mobile wireless telecommunications services offered by AT&T, Verizon, T-Mobile, and Sprint viewed by consumers are substitutes for regional and local services? This question is not about where the services are bought or sold (geographic market questions). It is about whether a consumer, facing a SSNIP for a national wireless service, would substitute to a regional or local service. The initial prices don't even need to be the same. If AT&T and Cellular One have $80/mo and $60/mo plans, respectively, that are comparable on everything except price, and AT&T raises its $80/mo plan to $84-88/mo (a 5-10% increase), would too many consumers divert to Cellular One's plan for this increase to be profitable to AT&T? If so, then these are in the same product market; if not, then national and regional services are discrete products. The scope of the network, that is, is an aspect of the network's (actual or perceived) quality; and similar products of different qualities may or may not be substitutes for one another. (We may also ask whether at some point 3G, pre-3G, and post-3G networks are all in same product market.)

This is not a distinction without meaning. IO economists will tell us that geographic and product market definition are part of the same inquiry: when will a SSNIP be profitable? Given this, we might think that whether national network scope is considered as part of either question doesn't matter, so long as both are considered. But it's more complicated that this. We need to ask this question at both stages: we need to ask in each local market whether consumers, facing a SSNIP, would substitute to another product sold in the same geographic area, as well as whether they would substitute to other carriers' services. It is entirely possible, for instance, that consumers in the southeastern United States are more willing to substitute to a regional carrier in response to a SSNIP than consumers in the mid-Atlantic are. Or that consumers in Albuquerque, Santa Fe, and Taos, NM are less willing to treat a regional carrier as a substitute for a national carrier. Or, to return to the argument that starting with either local or national geographic markets gets us to the same place: if we start with a SSNIP in a national market, which might be profitable on a product-market basis, there could be some geographies where it would not be profitable (and therefore not likely to substantially lessen competition) -- in those markets, therefore, the merger shouldn't challenged.

We cannot simply gloss over these distinctions by calling it a "national market" and being done with it. This process is complicated and data intensive. And it is certainly the case that given any (geographic, product) market pair in which a SSNIP would be profitable for the merged firm DOJ should seriously consider challenging the merger bounded by that pair -- but, conversely, in any areas where it would not be profitable, DOJ should be very hesitant to try to bring what would likely be an unsuccessful suit. Importantly, given the legal standard (DOJ needing to affirmatively prove a likelihood of substantially lessening competition), any challenge should be limited to the smallest (geographic, product) market pair in which such lessening is likely.

That all said, there are many peculiarities about this market that tie into this discussion, two of which I want to mention now. First, which most of us are familiar with, is the role of exclusive and locked handsets. Second is the market norm of localized sales, when Internet-based sales could easily be accommodated. It is interesting that most people still buy cell phones and service plans by going to their carrier's local store -- where, to activate the service, the clerk basically enters some information from the phone into an internet-connected computer. The iPhone is one of few phones that can be purchased and activated other than at a carrier's store. It is not difficult to imagine a world where I could purchase a phone at Best Buy in Virginia, log into a carrier in Oklahoma's web site and enter the phone's IMEI number to subscribe to a service plan offered by the Oklahoma-based carrier. That is, there's no technical reason why the market needs to remain locked into the zip-code-based geographic price discrimination model, in which case we would be operating under section 4.2.1 of the HMGs, and could be facing a nationwide geographic market. The obvious answer to why the market hasn't evolved in this direction (yet?) is that most of the carriers that would benefit from being able to offer service where they don't have physical stores (i.e., the regional carriers) don't have roaming contracts that would make this affordable. Given a world with a different roaming regime, it would be interesting to see whether the market evolves to a true nationwide geographic market.

But I'm getting beyond the scope of this comment. My point here is simply to say that the fact that some mobile wireless telecommunications services operate on national networks does not make this a nationwide geographic market -- rather, it is a market in which national- and regional-services, which may or may not be part of the same product market, are offered in local geographic markets.

Posted by: Gus Hurwitz | May 24, 2011 9:00:26 AM

Post a comment