The RDOF I mess spells the end for RDOF II
By now, everyone in the telecom space (and a lot of people outside it) have heard about the problems with the Rural Digital Opportunity Fund Phase I auction. The auction funded road medians because of underlying data errors. $850 million was awarded to SpaceX which would have provided service anyway. Indeed, just the presence and aggressive bidding of satellite providers with no incremental cost pushed down prices so much that it’s unclear ISPs can make the math work and actually deliver. A well-publicized example is bidders were allowed to bid with wireless service at the same prioritization as fiber optic.
It’s actually a less-talked about problem (though not unnoticed) with RDOF I that spells the end of RDOF II: giving subsidies to ISPs to bring service to areas that already are — or should be — well served. As a bit of background, the FCC split the RDOF auction into two phases. The first phase (just completed) “will target homes and business in census blocks that are entirely unserved”. The second phase “will cover locations in census blocks that are partially served, as well as locations not funded in Phase I”. The FCC broadband maps don’t know which Census Blocks are partially served, so by dividing it in two, Ajit Pai’s was able to speed up Phase I.
Even Phase I, which again is supposed to be census blocks that are entirely unserved, has obvious situations where the area wasn’t in need of a subsidy but got one anyway. I’ve identified 15,781 block groups representing $300 million in funding that might not need subsidies. The criteria for this is areas where, according to the Census’s ACS survey, more than 50% of households use broadband in their home, and more than 85% of the block group has access according to FCC. These areas are mostly suburban communities, many of which would be able to call the local ISP and get service.
For example, this golf resort identified by Free Press has an unserved block that makes it eligible for RDOF subsidies, even though 86% of the block group uses broadband, and 99% of the block group has access to it.
A random selection using the criteria above shows a suburban community outside Honolulu, a suburban community surrounding New Brighton HS in Beaver Co., PA, and a suburban community in Champaign Co IL (all shown below).
Do we really need to give scarce, and regressively-collected, federal subsidies to whatever data problem or anomaly led a block in these areas to be reported as unserved? Certainly not.
I’m harping on this because it will get way worse in RDOF Phase II. Whereas there is an allegedly completely unserved Census block that made these examples eligible for subsidies in Phase I, Phase II will by design be funding census blocks that are currently partially served. We’re planning to subsidize an ISP to, in some cases, drag a wire to the next house on the block.
The new broadband maps that will give us data about partially served blocks don’t yet exist. The FCC is working on it as part of their Digital Opportunity Data Collection proceeding, after USTelecom, er…Congress passed a law telling them exactly how to build the new maps. There’s a lot of excitement about the new maps, but I’m less sanguine. I’m concerned that spaghetti shapefiles submitted by the providers may not line up well with the underlying “Fabric” of eligible locations, and that the algorithmic selection of “primary residential locations” from satellite imagery and messy local data may inflate the amount of partially served areas.
Indeed just today, Midwest Energy & Communications, after studying an area where they’d been awarded subsidies as part of the Rural Broadband Experiment, asked the FCC to revise the number of eligible locations down from the Commission-estimated 421 to 382 — a difference of 9%. The 382 locations is actually a revision upwards from their previous estimate of 364 locations. The additional locations are from “multiple dwellings on a single parcel where, for instance, a home and a separate office facility is provided, or may be provided, broadband service by MEC” (footnote 4). They conclude, “the minor discrepancy in actual locations, based upon no fault of MEC, should not preclude MEC from retaining its full RBE award…”. It’s exactly this type of “eligible location” inflation and over-estimation that is concerning on a national scale.
Digital redlining, where an ISP won’t provide service to an area because its residents can’t or won’t pay fees high enough to justify the expense, is an important problem. If you believe, as I do, in universal and affordable service, we need a policy solution because it hasn’t and won’t fix itself. RDOF Phase II is not the solution. As happened in Phase I, Phase II would find local broadband providers competing against unrealistic upstarts and satellite providers. One solution is to use Title II authority to end digital redlining, even providing a subsidy if an ISP is bringing service to a legitimately hard to reach household.