Working Paper No. 14-16:
Efficient Spectrum Reallocation With Hold-Ups and Without Nirvana
Date Posted: May 2014
Abstract (below) | Full text (most recent) on SSRN
In 2010, the U.S. Federal Communications Commission (FCC) determined that up to 20 television channels should be shifted to mobile services. If successful, the reform could generate over $1 trillion in social gains. To achieve these efficiencies, regulators rejected traditional tools, which would have terminated existing wireless licenses, as too contentious. Instead, they chose to create a two-sided auction in which incumbent TV licensees state their offer prices to exit (broadcasting), being paid from the winning bids for mobile licenses (granted access to the reallocated TV spectrum). The FCC conducts this “incentive auction” and, importantly, unilaterally reassigns TV channels for stations remaining on the air. These involuntary transfers are designed to eliminate station hold-ups, allowing new blocks of contiguous spectrum (more valuable than scattered frequencies of equivalent bandwidth) to be made available for mobile licenses. A rival policy option – “overlay licenses” – would grandfather existing TV stations and then auction new licenses permitting liberal use of the TV Band (i.e., not restricted to broadcasting). Overlay winners could then bargain with TV stations to make more bandwidth available for mobile broadband. Policy makers rejected this approach on the grounds that incumbent TV licensees would engage in strategic hold-up, pre-empting efficient deals. But so, too, are hold-ups endemic in the design and implementation of incentive auctions; to asymmetrically ignore these costs is to commit the Nirvana Fallacy. This paper offers a template to avoid the Fallacy, evaluating transaction costs across the rival policy regimes. A first order quantification suggests that the costs of incentive auctions have been substantially under-estimated by regulators, perhaps reflecting confirmation bias and a principal-agent conflict in the assessment of policy options.