A policy framework for spectrum allocation in mobile communications.

• Author: Beard, T. Randolph

1. INTRODUCTION II. FROM DE JURE TO DE FACTO SPECTRUM CAPS III. ANALYTICAL FRAMEWORK A. Theoretical Model B. Illustrations of the Theory IV. "MORE" SPECTRUM DOES NOT AFORTIORI MEAN "MORE" FIRMS A. Theoretical Evidence B. The Historical Evidence V. CONCLUSIONS I. INTRODUCTION

   Across the globe, the mobile communications revolution is well under way. From advanced economies such as the United States, to developing economies like India, mobile telecommunications, in both voice and data forms, is quickly becoming the communications technology of choice. In the United States, it took less than fifteen years for wireless telephones to move from a thinly consumed service to effective ubiquity. (1) In 2009, there were 285.6 million wireless accounts in the United States, which translates to roughly 1.1 accounts for every person over ten years of age or more. (2) Mobile communications has evolved well beyond voice technology to now include enhanced communications services such as text messages, e-mail, and broadband connectivity, which are, in fact, quickly becoming the dominant source of consumer value for mobile service. In the not-so-distant future, it is expected by some that mobile appliances, like the iPhone, will replace traditional computers for many consumers. (3) For many individuals and households, mobile broadband may be the Internet connection of choice. (4)

   This rise in wireless connections, as well as the rapidly increasing demand for data services over such connections, is a mixed blessing. On the one hand, it provides an enormous economic boon to consumers, businesses, and providers; on the other hand, however, it is beginning to test the capacity of networks to provide such services. As a result, the supply of available quality commercial spectrum is rapidly becoming exhausted. (5)

   Fortunately, this fact has not gone unrecognized by policymakers. FCC Chairman Julius Genachowski recently observed that "America is facing a looming spectrum crunch" (6) because "the United States does not have nearly enough spectrum to meet its medium- and long-term mobile broadband needs." (7) Perhaps the single most important proposal in the National Broadband Plan is to make 500 megahertz (MHz) of additional spectrum available by 2020 for the provision of mobile broadband services, with ideally 300 MHz of that spectrum being made available by 2015, (8) a vision that President Obama formally endorsed by Presidential Memorandum. (9)

   Allocating more spectrum to advanced mobile services (such as broadband) is widely viewed as a sensible, if not a necessary, public policy. (10) However, merely stating that more spectrum is to be allocated to commercial mobile services leaves some highly relevant details unresolved. There are (at least) two important questions that must be answered when increasing the supply of spectrum: (1) how much new spectrum is to be allocated; and, more importantly, (2) who gets it? (11) On the first question, as noted a moment ago, the FCC has proposed to increase substantially the spectrum available for mobile services. On the other hand, the FCC has sent clear signals that it is concerned with increasing industry concentration (despite the fact that it approved every auction and wireless merger to date) and, by implication, that it would prefer to allocate any new spectrum primarily to new entrants and possibly smaller incumbents, rather than the largest incumbent providers, in order to "deconcentrate" the industry. (12)

   The allocation of spectrum among firms is a complex issue for which economic theory can provide key insights. To this end, we provide in this Article a theoretical analysis of some of the relevant tradeoffs involved in allocating spectrum among service providers. Informally, our analysis contemplates two (theoretical) states of the world. In the first, a fixed amount of spectrum is divided among many firms so that each firm has a "little" spectrum. In the other, that same fixed amount of spectrum is divided among fewer firms, so that each firm has "much" spectrum. Incorporating the modeling assumption that a firm with a larger holding of spectrum can provide more advanced services due to greater capacity and throughput than a firm with less spectrum, (13) the theoretical tradeoff, all other things constant, is straightforward. In a setting with "many firms with little spectrum," there may be more price competition, but that competition takes place over relatively less advanced services. (14) In a setting with fewer firms with larger allotments of spectrum, there may be less price competition (due to the Cournot assumption), but that competition occurs over more advanced services (due to the relationship between spectrum and the capacity to offer such services). Lower prices are good, and higher quality is good, but, if quality requires large amounts of a fixed allotment of spectrum, then the two may not occur together (as a result of the modeling assumptions). Consequently, the policymaker is asked to trade off between potentially lower prices for less advanced services, and potentially higher prices for more advanced services. It is not apparent, at first glance, which situation is "better."

   Since, however, either of the above-described outcomes can be supported as "best" under certain conditions, linking the theory to the current structure of the mobile telecommunications industry is required in order to render policy-relevant conclusions. We believe this linkage, properly understood, can render fairly strong prescriptions. First, while additional firms can lead to lower prices (at least in the Cournot competition framework, which is the standard for regulatory policy), at some point additional firms have almost no effect on price. In fact, economic theory suggests that the price cuts resulting from additional firms rapidly diminish, and there is evidence to support this theoretical result. (15) Price cuts are mostly exhausted after about three to five rivals are present in the market. (16) From an empirical standpoint, FCC data indicate that at present about ninety-one percent of the population has access to four or more mobile providers. (17) Consequently, adding new competitors to the mobile industry is expected to have a small impact on prices. The gains from dividing spectrum into smaller parts in an effort to create more firms (e.g., using spectrum caps) are therefore likely to be very low, even under favorable conditions. (18) Alternately, the economic gains from having access to broadband services are typically viewed as being very large, so a policy of granting existing firms (which can maximize spillovers from their existing plant and customer relationships) sufficient spectrum to run scalable networks that support innovation in both applications and devices is likely to produce substantial economic benefits. Therefore, in effect we have a situation where price changes are no longer a part of the calculus, since adding firms is expected to have a small effect on price competition. As such, the tradeoff revealed by the theory is simply between less- or more-advanced services, and the best policy is clear.

   In large part, our analysis comports with the recommendations of the U.S. Department of Justice (DO J), which address not simply the issue of how much spectrum should be allocated to the industry, but the importance of how much spectrum is given to a single firm. In a letter to the FCC, the DOJ observes, "[s]tated simply, without access to sufficient spectrum a firm cannot provide state-of-the-art wireless broadband services." (19) The DOJ also addresses, to some extent, the details of allocation decisions, observing, "[t]he goal in assigning licenses to any such new spectrum designated for commercial services should be to ensure that it generates the greatest ultimate benefits to the consumers of those services." (20) Our economic model adopts this "greatest ultimate benefits" approach, which is standard economic fare.

   Another important insight from the theory is that policymakers do not get to choose the number of firms offering mobile telecommunications services simply through the government's spectrum allocation decisions. Stated another way, the conventional wisdom that "more" spectrum somehow a fortiori means "more" firms simply is not true. (21) As explained in Competition After Unbundling: Entry, Industry Structure and Convergence, (22) which was cited at length by the FCC in the Fourteenth CMRS Report, it is the supply- and demand-side economic conditions of the marketplace that determine the equilibrium number of firms. (23) Economics determines the viable number of providers, not the intentions of policymakers. (24) Mobile services cannot be supplied without spectrum, but having spectrum does not imply financial success. Spectrum is simply one input to production and cannot singularly determine the financial viability of firms offering mobile communications services. This is apparent from the history of the mobile industry, which has undergone consolidation over much of its history (and may be due for more). (25) This consolidation is merely the industry adjusting towards a sustainable structure--an equilibrium consisting of fewer firms than that licensed by the design of the early FCC spectrum allocation decisions. The reality is that, while we may want five, ten, or twenty mobile telephony service providers, the economics are unlikely to permit it. (26) Consequently, the heavy use of incumbent-exclusion policies (such as spectrum caps or other limitations on spectrum use by firms) may not result in more providers, but may instead lead simply to inefficient use of scarce spectrum resources. (27)

   Finally, we present empirical evidence shedding light on the relationship between industry concentration and the amount of spectrum licensed for commercial services. While many believe and claim that more spectrum will necessarily lead to more...