The state that lost its chair after the music stopped
Two wireless ISPs increasing their coverage areas may lead to a $416 million decrease in Michigan's allocation.
On a first look, the new FCC National Broadband Map seems to be a step in the right direction. For example, in Alaska, a known problem area, the number of locations and the estimated amount of money allocated increase significantly. But Michigan is another story. Michigan has 71,139 fewer Unserved locations in the new map versus the old one, by far the biggest decrease in the 50 states. It’s an 18% drop in the number of Unserved locations, second only to South Carolina among the sizable states. And it leads to a estimated decrease of $416 million that Michigan will receive from the BEAD program, also by far the biggest decrease in the 50 states. The reason is big increases in the coverage areas filed by a few wireless providers.
As most readers know, NTIA will allocate the $42.5 billion to the states based on how many Unserved Broadband Serviceable Locations (BSLs) each state has as a percentage of the national total. Each Unserved location is worth about $4,895 so the numbers add up quick. Small changes to a state’s Unserved number can lead to big changes to their piece of the $42.5 billion, particularly when the change is bigger than the average of the other states. The average state added 15,856 Unserved locations in the new map; Michigan lost 71,139.
Normally, we’d applaud when the number of Unserved locations decreases — it means more people have access to broadband. The number of Unserved locations has been falling for years. If that’s what’s happening in Michigan — great. This release is different. The challenge process and the looming allocation means every state is rushing to make sure all their Unserved locations are added to the map.
Let’s look at what changed in Michigan. First, 130,182 locations went from Unserved in the previous version of the map to Served or Underserved in this version. That’s second only to Texas (a much larger state) in terms of the raw number. The 130,000 decrease in Unserved is partially offset by successful challenges Michigan filed with the FCC which added about 60,000 Unserved locations. So the net change in Unserved is -130,000 + 60,000 = -70,000 (using round numbers).
More than two thirds of the locations that gained broadband coverage according the ISP filings are now Served or Underserved by a fixed wireless provider. Two ISPs — Mercury Broadband and Point Broadband — each filed with offerings that made almost 40,000 locations per ISP move from Unserved to Underserved or Served.
Let’s look at some randomly selected examples. First up is an address in Marion, Michigan. In the previous filing period (right screen shot), Point Broadband filed this address as 100 Mbps download throughput and 20 Mbps upload throughput over unlicensed fixed wireless, which would make the location eligible for BEAD funding. In this version, they kept the unlicensed filing, but added a licensed fixed wireless filing at 50/10, which makes the location ineligible for BEAD funding.
When you plug this address into their website, the maximum throughput you could purchase is 50 Mbps download. The upload throughput isn’t mentioned. There is no mention of “speeds may vary”. It is well-known that ISPs are allowed to report the “maximum advertised” download throughput, not the actual throughput the subscriber will receive. In the case of some technologies, like cable and fiber, I’ve argued there likely isn’t a big difference between “maximum advertised” and “maximum actual” throughput. But fixed wireless is different. Coverage depends on the technology of the radios, antenna height, spectrum used, tree coverage, and other things. The only disclaimer on real throughput versus advertised throughput is the second to last paragraph in the contract. I don’t doubt some of these locations can get 50/10. I do doubt all of them can. And in this case hundreds of millions of dollars are at stake.
For this particular provider, there’s also the question of licensed versus unlicensed spectrum. In their previous filing they were using unlicensed spectrum. Now they’re filing with licensed spectrum. (Caveat: I’m not an expert in the FCC licensing systems). It does look like they have some spectrum licenses, but they appear to be largely point-to-point microwave licenses. Fixed wireless ISPs also provided a bunch of supporting documentation to the FCC in this filing, including link budgets, model parameters, and clutter models. I can’t tell — and am not sure if it’s possible to know — which spectrum is being used for which coverage filings.
Let’s get to another example, this time in Tekonsha, Michigan. In this case, Mercury Broadband started offering 150/20 fixed wireless with licensed spectrum. Their website says service may be available but they need to “confirm your Internet coverage with other tools and a free site survey.” I think this is a perfectly logical signup flow, and there might be houses that can get 150/20 broadband from this provider. But not all of them will be able to. By considering them Served, we’re ignoring the houses that actually aren’t.
(I’m not anti-wireless. I do think that the way coverage is reported to the FCC overstates the coverage available to homes in some cases. These cases, when they’re the offering that is the difference between Unserved and Served, can be enormously consequential, as they are here. I’d be eager to work with anyone to come up with more granular coverage models.)
At this point, many people know the cycle that these National Broadband Maps will take: the FCC will publish a new version of the “Fabric” (the BSLs). Challenges will be accepted against the Fabric. Then ISPs will report coverage areas against the Fabric. Then they’ll release the new map version and challenges will commence against the coverage areas filed by the ISPs. Those get adjudicated in a lengthy process that is not particularly friendly to the challenger, and the whole process repeats. There are some “challenges are accepted on a rolling basis” type statements that get made, but at its core the process is quite linear— err, circular.
But here, it appears the music has stopped with this new map version. While I think the FCC can accept coverage challenges to the new map, I’m sure they won’t commit to adjudicating a coverage challenge before NTIA’s allocation deadline. This doesn’t mean the allocation should be delayed. But it does mean that everyone involved should work quickly to adjudicate challenges and assess the fairness of the allocation rules. The way this is set up is not fair to Michigan, to the tune of $400 million.