One of the most important components in designing a successful spectrum auction is establishing the right license sizes - both the amount of spectrum devoted to each license and the geographic area it covers. As history shows, providing enough small blocks and a mix of geographic areas will help ensure the auction does not become a playing field for only Verizon and AT&T.
50 megahertz of prime spectrum nationwide is at stake as the FCC finalizes its AWS-3 auction service rules, and not surprisingly, Verizon and AT&T are advocating for big blocks covering big geographic areas. However, an auction with too many large blocks is a configuration that has the real potential to dissuade competitive carriers from participating in the auction. By contrast, smaller spectrum blocks with a mix of large and small geographic areas will provide the greatest opportunity for a wide variety of bidders to participate and obtain licenses that match their licensing needs.
In particular, T-Mobile would like to see the AWS-3 spectrum divided into as many 5x5 megahertz blocks as possible. This approach will bring in more bidders and increase auction revenues as a consequence. If the Commission nevertheless believes some bidders would benefit from the ability to purchase spectrum in a 10x10 megahertz block, it should include no more than one such large block in the AWS-3 band plan given the significant countervailing benefits of making the spectrum available in 5x5 megahertz blocks.
The FCC should also take into account the effect its block size decision will have on its subsequent mobile spectrum holdings proceeding. If, as T-Mobile has suggested, the FCC adopts spectrum aggregation limits in that proceeding, 5x5 megahertz blocks would give all carriers, including AT&T and Verizon, increased flexibility to conform their holdings to comply with such limits.
As with block sizes, smaller geographic license areas will likewise encourage greater auction participation and, as a result, will bring in higher revenues. Because AWS-3 spectrum will be used by many wireless carriers to complement their existing AWS-1 spectrum holdings, T-Mobile has argued that it makes sense to have a mix of license sizes that mirror, to the extent possible, the AWS-1 license areas - smaller cellular market areas (CMAs) and larger economic areas (EAs). The FCC is potentially also considering licensing spectrum over other geographic areas, like small Partial Economic Areas (PEAs). Regardless of the particular type of license areas the FCC chooses, it is imperative that there be a sufficient number of CMAs or other small geographic areas available in at least two 5x5 megahertz spectrum blocks so that competitive carriers have the opportunity to efficiently match their participation to their requirements.
In a recent blog post, AT&T asserted that the 2006 AWS-1 auction results support its demand for larger blocks and geographic areas for all or most of the licenses to be offered in the AWS-3 auction. AT&T is wrong for several reasons. First, with respect to license area, AT&T inappropriately compares the revenue generated from AWS-1 REAG licenses and CMA licenses to show that larger license sizes should be preferred. However, the FCC is not considering the use of REAGs in this auction. A more apt comparison is between the A and B Blocks in the AWS-1 auction – licenses for the same amount of spectrum but licensed on EA and CMA bases – precisely what is under consideration in this auction. Those results show little difference between revenue from EA ($0.43/MHz/POP) and CMA ($0.39/MHz/POP) licenses.
Second, AT&T argues that the “activity ratio” for REAGs was higher than it was for CMAs in the first bidding round. Its findings are unremarkable considering the number of licenses in each category of geographic area. Because there are only twelve REAGs as compared to 734 CMAs, it would be unreasonable to expect the same multiple of bidders for CMAs as REAGs.
Finally, with respect to license block size, AT&T fails to note the obvious comparison between identical license areas with different size spectrum blocks – specifically, the AWS-1 B and C Blocks, which were both licensed by EAs, but with different spectrum block sizes. The 10x10 megahertz B Block licenses sold for $.43 MHz/Pop while the 5x5 megahertz C Block licenses sold for $.51 MHz/Pop, approximately 15 percent greater revenues for the smaller spectrum blocks.
As the FCC considers this Goldilocks choice, it should keep in mind that neither bigger nor smaller is always better. Rather, if the spectrum block and geographic area sizes are just right, the auction will surely be a success.