Conexon recently reported on the FCC “upgrading” homes from served to unserved without really verifying the upgrade. Their take on the move is pretty straight forward…
At the end of the year, the FCC released data that it knows to be inaccurate, which will damage the lives and livelihoods of millions of our fellow citizens who live and work in rural America. In its publication of eligible census blocks for the Connect America Fund (CAF) auction, the FCC excluded 432,302 rural homes and businesses in areas that previously had been eligible to receive public support for broadband service.
The vast majority of these areas had been determined by the FCC and the telephone industry to be too costly for the telephone companies to serve with broadband – even with subsidies – so the FCC initially decided to auction financial support for these remote areas. And yet, in the closing days of 2017, the FCC removed 30% of all the eligible rural locations from the CAF auction by applying the filter of newly released data on the availability of broadband service.
The FCC is relying on data that purports to show that, between June 2015 and December 2016, broadband was newly made available to some of the most remote areas of the country covering 432,302 locations. And that all this new rural construction occurred without any federal universal support or any obligation to provide service.
Of course, no such thing happened. It is fake data. 432,302 rural homes and businesses now have phantom broadband service.
Turns out the problem is self-reported data…
What is the data that the FCC is using to make such a critical decision? The self-reported Form 477 data from ISPs regarding their maximum advertised speed in each census block where they provide service. When I was at the FCC, not a single economist who worked in my group believed the data to be accurate or reliable. The FCC’s misapplication of the data in the CAF auction will have consequences for millions of rural Americans.
The infirmities of the FCC’s methodology include:
The FCC does little to verify the accuracy of the self-reported data. Even a simple search of ISP websites reveals that the 477 data reported by many ISPs do not match the advertised speeds. As I wrote last year, a check of the 37 wireless internet service providers claiming 100 Mbps speeds showed that 35 of the 37 did not even advertise such speeds on their websites.
If an ISP claims to advertise internet service availability of 10/1 Mbps to a single locationin a census block, the entire census block has been removed from the auction.
The maximum advertised speed is not the same as the delivered speed. So, when an ISP reports that it advertises 10/1 Mbps service, but only provides 4/1 Mbps service, the census block is still dropped from the auction. Worse still, since some technologies require site surveys to determine availability of service, just advertising 10/1 Mbps availability does not mean actual availability.
10/1 service isn’t broadband, even by the FCC’s anemic standards.
The FCC refuses to allow challenges to this inaccurate and unreliable data, because it can’t be bothered.
The article clearly has its bias – but a theme is emerging this week. Yesterday, I reported on a Harvard study that tried to compare broadband pricing but was unable because that data isn’t collected. Today we’re hearing about unreliable data from the Form 477. What I’ve heard about the 477 forms is that they are a pain to fill out. Unfortunately, as the article points out, the contents of the form matter. Communities qualify or don’t qualify for potential funding for infrastructure based on the forms. It would be nice to have a community feedback loop. Can you image federal policy based on customer self-reported data? In the world of crowd-sourcing it seems like it could be done. Maybe there’s a balance to be made.