Yesterday, in the FCC’s Incentive Auction proceeding, AT&T filed a computational study that seeks to provide insights into some fundamental and essential auction questions – namely, for any given clearing target, what might a successful auction look like in terms of how many broadcasters must participate and in what markets?  And, how much diversity is there in the solution sets necessary for a successful outcome and, how are adjacent markets correlated?

As important and interesting as these questions are, to date scant analysis on these questions has been offered.  The research submitted today, authored by Michael Kearns and Lili Dworkin of the University of Pennsylvania, will help address that void.

The study attempts to ask and answer a series of straight-forward mathematical questions like:  how many broadcasters need to participate in the New York City designated market area (DMA) to clear 84 MHz of spectrum?  How do those numbers change if you take the FCC’s published domain constraints into consideration?  If you assume an 84 MHz clearing target, how many markets, on average, will require some level of broadcaster participation?

Given the complexity of the auction, the answers are rarely black and white.  For any clearing target, there are almost unlimited variations on possible broadcaster participation and a very large number of possible repacking solutions.  But from the research, an outline of the scope of the possible begins to emerge.

For example, Dr. Kearns studied the minimum number of broadcasters that must be cleared to meet a range of nationwide clearing targets.  The reported results are optimistic in that they assume all broadcasters are willing to participate, whereas in reality we recognize that some certainly may not be.

Nevertheless, under the assumption of full participation, we learn that to clear 84 MHz of spectrum, a minimum of roughly 200 voluntary broadcaster exits are needed ignoring the domain constraints as identified by the FCC, and approximately 250 exits, if those constraints are considered.  We also learn that the gap between the two – approximately 50 participating broadcasters – is the rough “cost” of the currently identified domain constraints.

Dr. Kearns also studied large sets of feasible repacking solutions to understand the relative size of the challenge in specific DMAs.  From this we learn, not surprisingly, that New York, Philadelphia and Chicago are the most challenging DMAs, requiring the highest number of average clearings in the observed solution sets.  We also see some surprising entries on the list of the top 25 most challenging DMAs, including Milwaukee (#10, because of its proximity to the congested Chicago market) and various North and South Carolina markets because of their role in the challenging East Coast daisy chain.

As to the minimum number of markets that must participate, we learn that to clear 84 MHz the absolute (and highly unrealistic) minimum is 43 DMAs.  In other words, at least one solution set exists where there is exit from only 43 DMAs to achieve 84 MHz cleared nationwide.  More relevant to the auction, the average number of DMAs needed to clear 84 MHz is 74 (again, ignoring the domain constraints).  This suggests that to maintain flexibility in the auction, the FCC will need to look well beyond the top 30 DMAs.

Dr. Kearns also does analysis around the diversity of solution sets and the frequency at which individual broadcasters are cleared in observed solutions.  The former sheds light on which DMAs offer the most flexibility around which individual broadcasters must clear to reach a target (Denver is #1 on the diversity scale); the latter looks at individual broadcast interference profiles and how frequently the solution sets relied on clearing a specific broadcaster to achieve the desired result.

As important as what the paper does is what it does not do – the analysis takes no position on any of the important policy issues still pending in the proceeding.  For example, the study does not address likely participation. It simply accepts various assumption sets and tests against those assumptions for results.  Similarly, the study does not address the question of reverse auction pricing or what values the FCC should offer participating broadcasters.  But on that point, AT&T continues to believe that wide broadcaster participation is absolutely essential to success and that scoring on any criteria except interference profile will undermine participation.

In filing the study, we hope to shed some light on the nature and scope of the repacking challenges the FCC faces in the auction.  The auction challenges are without a doubt significant.  But the FCC has made significant progress on these challenges, as evidenced by the recently-released Order.  And we believe that the computational analysis submitted yesterday demonstrates that, as long as the financial incentives are attractive, a successful outcome is well within reach.


Share this