Unlicensed Operations in the 600 MHz Guard Bands: Potential Impact of Interference on the Outcome of the Incentive Auction

Abstract

In its June 2014 Report and Order on the 600 MHz incentive auction, which will repurpose TV broadcast spectrum for mobile broadband service, the FCC authorized the use of unlicensed devices in the post-auction guard bands. Although the FCC maintains that it will not be a problem, key stakeholders have argued that the operation of unlicensed devices in the 600 MHz guard bands could cause harmful interference to licensed mobile LTE services in nearby bands. In this paper, we analyze the potential for such interference and its implications for the incentive auction. We begin by critiquing the FCC’s technical analysis, which found that an unlicensed device using the 600 MHz guard bands would interfere with the operation of a licensed mobile device using LTE at 600 MHz if the two devices were less than 20 feet apart. Although that finding will give potential bidders serious pause, it understates the real problem. When we modify key FCC assumptions to reflect conditions at hand, the estimated interference range is 45 feet to 75 feet.Next, we analyze how harmful interference will reduce LTE network capacity and the corresponding market value of 600 MHz spectrum. Based on our analysis of an LTE network in a band similar to 600 MHz, we find that a 5 percent loss of capacity will lower the value of the affected spectrum by at least 9 percent; a 20 percent loss of capacity will lower its value by at least 43 percent; and a 35 percent loss of capacity will eliminate most (93 percent) of its value. Finally, we trace how the prospect of interference, through its adverse impact on the value of spectrum, could damage the incentive auction. Two features of the auction magnify the impact of the risk of interference. First, at the allocation stage, bidders will be unable to distinguish between those blocks that are subject to interference from unlicensed operations and those that are not. Thus, rational bidders will assume that all blocks are subject to interference. We estimate that the allocation-stage revenues will decrease by an amount ranging from 9 percent (for a 5 percent level of interference) to 93 percent (for a 35 percent level of interference). Using for purposes of illustration the FCC’s 84-megahertz band plan (84 megahertz of broadcast spectrum cleared and 70 megahertz repurposed for mobile broadband), with a 5 percent level of interference, total allocation-stage revenues will be reduced by $4 billion, to about $40.3 billion. At the 35 percent level, allocation-stage revenues for this plan will be reduced by about $41.2 billion, to $3.0 billion.Second, even though some of the revenue lost at the allocation stage will be recovered in the assignment round (when bidders will be able to identify individual blocks), the amount of TV spectrum cleared depends solely on the revenue generated in the allocation stage, which must cover total clearing costs. Thus, the prospect of interference, by reducing allocation-stage revenues, will limit how much spectrum even makes it to the assignment round. For example, at the 10 percent level of interference, allocation-stage revenues will fall short of required clearing costs for the FCC’s five largest band plans. Under that scenario, the largest feasible plan would be the illustrative band plan, which would clear 84 megahertz and repurpose 70 megahertz for mobile broadband. This is 50 megahertz less than would be repurposed under the FCC’s largest band plan (144 megahertz cleared, 120 megahertz repurposed), and some of it would be subject to interference. We estimate that every 10 megahertz of broadcast spectrum that is not repurposed for mobile broadband represents at least a $60 billion loss in consumer welfare. Thus, with only a 10 percent level of interference, the best possible outcome (70 megahertz repurposed) would represent at least a $300 billion loss in consumer welfare relative to the FCC’s largest band plan (120 megahertz repurposed).