Tuesday, August 19, 2014

Here We Go Again: Don’t Eliminate the “Incentive” from the Incentive Auction

By Gregory J. Vogt, Visiting Fellow
This past June the FCC published the Incentive Auction Order, in conjunction with the Mobile Spectrum Holdings Order. Together, their intent is for broadcasters to volunteer spectrum in the “reverse” auction, to be then repurposed for mobile broadband use in the “forward” auction. The aim of the two-sided incentive auction is to achieve a significant win-win-win for consumers, government, and industry. But the forward auction design adopted in June, in my estimation, entails novel, risky, and complex elements that gives preferences to all bidders other than AT&T and Verizon, including to multi-billion dollar industry participants.
Sprint and T-Mobile filed petitions for reconsideration on August 11. T-Mobile, unsatisfied with the advantages it already achieved, filed a petition seeking changes to certain aspects of the forward auction design. Adopting bidding preferences such as those contained in the two June Orders creates a serious risk of producing lowered revenues, as I reported from an analysis in the FSF blog based on past auction experience. Lowering revenues will risk decreasing the amount of volunteered spectrum – the “incentive” in an incentive auction – in the same way as T-Mobile’s now-discredited Dynamic Market Rule, and thus should not be entertained. T-Mobile’s proposed changes would exacerbate these already serious downsides to the two Orders. Sprint now seeks more changes to the spectrum screen to further advantage itself.
In these two Orders, the FCC developed an extremely complicated structure which I describe only at a high level here. The structure reserved a portion of the 600 MHz spectrum for nationwide-provider bidders that do not possess at least 45 MHz of spectrum below 1 GHz spectrum in each geographic market. The maximum size of the “reserve” varies with the amount of broadcast spectrum available in a particular geographic market at the initial stage, 30 MHz for 100 MHz markets declining to 10 MHz for 40 MHz markets. The auction design establishes a two-component “trigger” which must be met before the reserve spectrum becomes available: a price/proceeds component and an expense component (legislatively mandated administrative expenses such as compensating broadcasters and FirstNet funding). Until both components of the “trigger” are met, any bidder may bid on all spectrum in a market. 
In its petition, T-Mobile argues that the amount of “market based reserve spectrum” is set too low and requests that the reserve be at least half of available spectrum. It also contested the formulation of the price/proceeds benchmark component of reserve spectrum trigger as unnecessary and subject to gaming.
The most important aim of the incentive auction legislation is to maximize the amount of broadcast spectrum repurposed for mobile broadband use. Although the FCC originally agreed with this conclusion, that goal is all but absent in the final rule, apparently side-tracked in the FCC’s effort to “fairly” allocate spectrum, with the reserve tipped heavily in favor of everyone but the top two market players. Although a last minute compromise preserved the ability of all the providers to obtain some spectrum in the forward auction, it is unknown whether the FCC struck the right “balance.” We won’t know until we see how much spectrum broadcasters actually volunteer. The good news is that, at least at this point, there are no vocal defections from the auction process from either broadcasters or potential bidders.
But the forward auction design is a far cry from a free-market model characteristic of most spectrum auctions, which are routinely designed to maximize auction receipts. The forward auction picks winners and losers unrelated to a sound economically-based competition model like that normally attendant to an accepted competition analysis. Here, unless “smaller” carriers are capacity-constrained, which no one has alleged or proved, there simply is no significant anti-competitive situation that needs to be corrected. The necessity of adopting a first-ever minimum six-year holding period for awarded reserve spectrum only reinforces the conclusion that the forward auction design lacks market-driven incentives. Why would the government preclude T-Mobile from selling reserve spectrum unless it has doubts either that T-Mobile was unable to obtain needed spectrum in the first place or the mechanism will lead to lower-than-market prices? T-Mobile’s reconsideration plea for additional reserve spectrum merely seeks to exacerbate this government-made market.
The request should be rejected for four additional reasons. 
First, T-Mobile’s proposal risks further depressing auction receipts. Achieving receipts maximization is critical to producing the incentive for broadcasters to volunteer spectrum. The incentive is already at risk: the original estimate of up to 120 MHz of volunteered spectrum has declined to what government now admits will likely only achieve 85 MHz. The incentive should not be diminished further.
Second, T-Mobile’s requested modifications are anti-competitive. By increasing the amount of the reserve, and eliminating the price/proceeds component of the trigger, smaller bidders could be able to obtain spectrum without fully competing even with each other (let alone AT&T and Verizon) because there would be far more in the reserve than the 20 MHz block most competitors say they need. Thus, bid prices for reserve spectrum could be less than for unreserved spectrum, a result I note here actually occurred in Canada, because smaller carriers can make strategic bids based on the knowledge that high bidders may be excluded from competing for the reserve and there would be more than enough reserve spectrum to go around.  Reduced costs for spectrum in turn could create a competitive imbalance in the mobile broadband market due to unbalanced cost structures.
Third, T-Mobile’s claim that AT&T and Verizon might game the trigger is speculative at best because the combination of blind-bidding and delayed creation of the reserve until the final stage rule is achieved, completely undermine any incentive to manipulate the trigger. A low bid by one large provider to avoid invocation of the trigger could simply be a white flag that ends up handing spectrum over to the other major player.
Fourth, although I agree with T-Mobile that setting a complicated price/proceeds trigger is relatively arbitrary and creates other risks to the bidding process, eliminating that component of the trigger would exacerbate disincentives to participate, not improve them.
As to the modified spectrum screen, the FCC added mobile broadband spectrum to achieve a more complete product market definition in transaction analyses. Sprint, which holds significant spectrum added to the screen for the first time, seeks to blunt this reform by asking the FCC to weight differently low-, mid-, and high-band spectrum, effectively straight-jacketing the FCC when using the screen. Although unrelated in a direct sense to the spectrum auction, the request does represent another attempt by competitors to skew competitive analysis in their favor. The FCC has already announced it will take into account the potential competitive impact of certain types of spectrum, such as that below 1 GHz. Since that decision alone unjustifiably favors Sprint, it should just declare victory without seeking further undeserved advantages.
Although I do not support the Rube Goldberg contraption that the forward auction design entails, or the complications to the new spectrum screen already adopted, the FCC should reject further manipulations the parties now request. The FCC must preserve the incentive to volunteer a maximum amount of broadcaster spectrum if there is to be a chance of achieving the win-win-win situation the incentive auction was intended to produce. And it should reject strict formula-based reviews of transactions, relying instead on a fact-based competitive analysis.