Julius Genachowski’s Speech at CES Part 2…

Posted on by

Most industry experts and pundits agree that the U.S. needs more spectrum in the hands of commercial mobile broadband providers.  To this end, Congress should be commended for their efforts to pass legislation to give the Federal Communications Commission the authority to hold voluntary incentive auctions to help facilitate the transfer of prime spectrum from broadcasters to mobile broadband providers.  As noted yesterday in Larry’s post, however, last week at CES FCC Chairman Julius Genachowski blasted a bill currently working its way through the House Energy and Commerce Committee because the proposed legislation would eliminate the FCC’s ability to preclude certain companies from participating in future auctions as well as to encumber spectrum with regulatory obligations above and beyond what FCC rules otherwise require.

The Chairman’s desire to keep spectrum out of the hands of the country’s largest and most commercially successful wireless companies stems from his fixation on industry concentration, which in turn depends on a fairly naïve view of the economics of telecommunications markets when industry concentration alone determines price.

I would like to pose an interesting question:  Is the overly simplistic economic view of the industry in which high industry concentration translates directly to high prices sufficient to warrant incumbent-exclusion rules in spectrum auctions?  Then answer is no, and we show why in a paper we published last year in Federal Communications Law Journal entitled A Policy Framework for Spectrum Allocation in Mobile Communications.

In that paper we looked at the question of whether the FCC’s desired incumbent-exclusion rules, whether formal or informal, would help or harm consumers.  Contrary to Mr. Genachowski’s apparent belief, we found that the incumbent-exclusion rules championed by Mr. Genachowski are unlikely to increase consumer welfare under current conditions.  The paper is based on a somewhat technical treatment of the issue, so my plan it to present the intuition of the analysis in this post.

Our paper provides a theoretical analysis of some of the relevant tradeoffs involved in allocating spectrum among service providers, with a particular focus on incumbent-exclusion rules (including such policies as spectrum caps).  There are two key assumptions driving the analysis.  First, we assume that the more competitors there are, the lower are prices.  In the economics lingo, this means we assume the firms act as Cournot competitors.  This assumption is consistent with Mr. Genachowski’s view of industry structure and performance, and makes the HHI a relevant measure of market power.  (Absent the Cournot assumption, the HHI says little to nothing about market power.)

Second, based on the technology of spectrum, we assume that more spectrum permits firms to offer more advanced services due to greater capacity and throughput.  This assumption is not controversial and has been made repeatedly by the FCC.

Given a fixed supply of spectrum, the problem of spectrum allocation becomes immediately apparent.   Say, for example, you had 500 MHz of spectrum.  You could, theoretically, divide the spectrum among 500 firms, giving each 1 MHz, thereby having a large number of firms (and thus low prices).  Of course, the firms could not do much if anything with so little spectrum, and even if they could all 500 would not survive financially.  At the other extreme, you could give all 500 MHz to a single firm.  By doing so, the firm could offer some highly advanced services, but given the Cournot assumption, it would do so at high prices.  The policy issue is, therefore, how to divide the spectrum up in the intermediate range between these two extremes.  With the maximization of consumer welfare as our measure of good policy (not auction revenue), the derived theoretical tradeoff is somewhat intuitive:  In a setting with many firms with little spectrum, there are low prices (by the Cournot assumption) but relatively less advanced services (by the technology assumption); however, in a setting with fewer firms with larger allotments of spectrum, there may be higher prices but also more advanced services.   Rules that block incumbent access to additional spectrum are usually motivated by price competition, but that price competition comes at the potential price of less advanced services.

Our analysis highlights several key components of the spectrum allocation decision.  First, an incumbent-exclusion rule is not “pro-entry,” but instead seeks to select one form (price cutting) of entry over another (quality improving).  Notably, as I read the commentary, the issue of spectrum exhaust is not so much about the number of competitors, but the effect of capacity constraints on the quality of services offered by wireless firms.  Ad nauseam, we are informed that the economic benefits of advanced wireless services are likely to be very high, but providers need more spectrum to provide such services.  If mobile providers are going to provide the high-quality broadband services many feel are essential for our economic, political and social well being, then providers (not the industry) need more spectrum.  With a fixed supply of spectrum, this obviously means fewer providers.

Second, with Cournot competition, the effect on price of adding more competitors is subject to diminishing marginal returns.  That is, most of the price reductions from entry occur with the first two or three firms.  (This fact forms the basis for the HHI thresholds in the Merger Guidelines.)  Thus, when the number of firms exceeds a few, the potential for sizeable competitive price effects is low.  Given that most U.S. consumers have access to four or more providers, the gains from additional entry are likely to be relatively small.  Even if the price effects are moderate, these effects must be weighed against gain in quality.

Given industry conditions (e.g., spectrum exhaust, multiple providers), we concluded that the incumbent-exclusion rules are unlikely to be helpful, where helpful is measured against the standard of consumer welfare.  (Notably, our measure of welfare is rigged in favor of such rules since we do not count the welfare derived from the more advanced services and ignore producer welfare).  In other words, we believe that under present conditions, the quality problem is more important than the price problem.

We make another very important point that is typically ignored in the policy debate.  That is, access to spectrum resources does not necessarily convey financial success, as spectrum is but one of many inputs necessary to provide service.  Policymakers may want more mobile providers and may be willing to throw spectrum at new entrants (or smaller incumbents) in an effort to make it so.  Unfortunately, just having access to spectrum does not imply that a firm can achieve financial success.  LightSquared is a case in point (and there are many other entities that have spectrum but have chosen not to provide services).  Moreover, history has shown that as spectrum resources have risen, the number of competitors has not.  Contrary to popular belief, more spectrum does not imply more competitors.

The figure below (reproduced from the paper) illustrates the relationship between the market shares of the largest mobile telephony firms and the total MHz of spectrum made available by the FCC to such firms over the period 1993 through 2009.  Total spectrum is shown by the shaded area in the figure and is rising over the entire time period.  In 1993, there was 50 MHz of spectrum used for mobile telephony.  Including all auctioned spectrum, this number rose to 361 MHz by 2009.

The Concentration Ratio, CRn, is used to measure industry concentration.  The Concentration Ratio is computed as the sum of the n largest firms in the market.  That is, CR2 measures the summed market shares of the two largest firms, and CR5 the market share of the five largest firms.  Both the CR2 and the CR5 are illustrated in the figure.  Finally, the average revenue per minute for mobile telephony is provided.  All data is computed at the national level.

The figure shows clearly the following.  First, the amount of spectrum has risen, yet industry concentration, as measured by the concentration ratio, has not declined.  Thus, historical evidence does not support the notion that more spectrum means a lower level of industry concentration.  Second, while concentration has risen over this interval, the price of mobile telephony has fallen consistently over the period.  Therefore, historical evidence also does not support the notion that higher concentration leads to higher prices.  The latter result has important implications for the theory.  If changes in concentration (or the number of firms) do not impact market performance, then the gains from an incumbent-exclusion rule are likely to be small and the net losses large.

It is worth noting that these data cover many years, and technology has evolved over the years.  As such, the trends in the figure are merely suggestive.  Nevertheless, the historical data cannot be ignored and, if considered, provide important insights for the economic value of incumbent-exclusion policies.

The tradeoff derived in the paper is intuitive.  Our particular interpretation of the facts is just that—a particular interpretation.  But, even if one sees the facts differently, the theoretical tradeoff remains valid and useful.  If incumbent firms are precluded from obtaining more spectrum, particularly successful firms serving large customer bases, then the quality of service will suffer.  Under existing conditions that include spectrum exhaust, an attempt to pump up the number of competitors through incumbent-exclusions rules, even assuming that doing so leads to more price competition, may not (and in our view is unlikely to) make American consumers better off.

Finally, for you deficit hawks, precluding the nation’s most successful wireless providers from bidding on spectrum will not increase the sale price of spectrum.  If anything, it will reduce the auction value of the spectrum.  More bidders in an auction can’t hurt.  As observed by noted economist Louis Phlips in The Economics of Imperfect Information (1989), “… total rent recovery is an increasing function of the number of participants. [ ] The more bidders there are, the more information is thus aggregated and the closer is the winning bid to the common value (p. 109).”  Certainly, throwing out those bidders prone to have very high valuations is unlikely to increase auction revenues, especially since they are being thrown out for fear of them winning.