Spectrum in the 3.5 GHz band is expected to drive the first wave of fifth-generation (or “5G”) wireless broadband services, in large part because many countries have focused on harmonizing this spectrum for 5G use on an international basis. International harmonization of 5G spectrum in these frequencies would enable faster speeds, lower cost devices, and international roaming.
The Federal Communications Commission (FCC) and the domestic mobile industry both recognize the benefits of revising outdating regulations governing the 3.5 GHz band that would hinder the use of this spectrum for 5G in the United States. When the current regulations were adopted, the FCC envisioned that the spectrum would be used like Wi-Fi, with hotels and coffee shops installing antennas indoors, and manufacturers and utility companies using it for industrial wireless systems (e.g., private video surveillance of a warehouse). So it decided to assign licenses using Census Tracts, which often cover only a handful of city blocks in urban areas. (There are 2,167 different census tracts in the City of New York alone, some with “no population, no workers, and next-to-no visitors” and at least nine in “Queens that are entirely, or almost entirely, cemeteries.”) At the time the FCC believed assigning a separate license for each Census Tract aligned well with its vision for the band—a vision that’s been superseded by the international community’s designation of the 3.5 GHz band as a core band for mobile 5G services.
The “IIoT,” a coalition of utility companies (e.g., railroads, ports, and electric utilities), is arguing that the FCC should still assign roughly half of the licensed 3.5 GHz spectrum using Census Tracts despite the urgent need for 5G spectrum. Their plan would be a disaster for the deployment of 5G networks in the United States.
The entire mobile industry—from rural and regional to nationwide providers—is in rare agreement that assigning a different spectrum license to each of the more than 74,000 Census Tracts is technically and administratively infeasible for 5G deployments.
Licensing 5G spectrum by Census Tracts would create a nightmare on every street. The Census Bureau purposely designs Census Tract boundaries to follow visible features, such as streets, roads, and highways. Census maps show that streets are far and away the primary type of boundary for defining Census Tracts in urban areas. For example, in most of lower Manhattan, every other North-South street (and in some areas every North-South street) is used as a Census Tract boundary.
The use of city streets works well for census takers, who can readily identify the relevant census boundaries when canvassing a neighborhood on foot. But it makes no sense for wireless networks, which are typically deployed on street poles (and along other roads and highways) for the purpose of serving motorists on both sides of the street—not just the cars going in one direction. Licensing mobile spectrum using boundaries designed for census takers would require a wireless provider to acquire at least two licenses to install even one antenna on many urban streets and even on many highways.
Due to urban Census Tracts’ extraordinarily small size and reliance on streets as a primary boundary type, using these areas to license 5G spectrum would promote the tragedy of the anti-commons—the division of licenses among so many different owners that they cannot be used efficiently—a tragedy that spectrum auctions were intended to solve. This tragedy plagued the FCC’s efforts to license the first cellular band by comparative hearing and then by lottery. The resulting distribution of licenses to a massive array of different owners delayed service to the public and hindered roaming until mobile providers were able to overcome the inevitable holdout problems and assemble larger and more efficient spectrum footprints. Licensing 3.5 GHz spectrum using Census Tracts would risk repeating this history.
These inefficiencies are what the Census Tract holdouts appear to be counting on. The IIoT’s members apparently think that making 3.5 GHz spectrum less attractive for 5G will allow them to buy licenses on the cheap so they can self-provision their own, private networks and avoid paying for commercial wireless services like the general public.