The great and not as great of the AAIFAA
Digital equity and inclusion take center stage, but an unfortunate reliance on reverse auctions.
I love Congressman Clyburn’s Accessible, Affordable Internet for All Act. Let me count the ways. (There are a couple of way in which I love it less, and I’ll get to those later). If a version of this bill becomes law as part of President Biden’s Build Back Better infrastructure plan, it will make a significant impact in closing the Homework Gap through affordability, inclusion, and deployment.
First, it’s important to understand the context in which this legislation was introduced. It was summer 2020. Joe Biden hadn’t yet won the presidency. RDOF hadn’t yet made a mockery of the reverse auction system. And it seemed unlikely that Democrats would ultimately control the Senate and therefore all three branches of government.
Digital Equity and Inclusion and Other Provisions
Before we get to infrastructure deployment , which unfairly receives most of the attention, I think it’s important to start with the accessibility and inclusion aspects of the proposed law. The law allocates $1.312 billion for a broad array of digital equity planning and implementation programs, digital literacy, workforce training, and awareness of existing programs. At $10 per month, or even $0, there would still be households without broadband service, and digital equity and inclusion programs are an important part of a practical — and not just academic — solution.
On affordability, the law seeks to add $9 billion to a program with a similar shape to the currently-being-implemented Emergency Broadband Benefit. It would also help finish the implementation of the FCC’s National Verifier technology which assesses applicant eligibility. These provisions, the rest of the bill, and an FCC interested in a successful Lifeline program, provide a meaningful chance of improving the dismal 25% take-up rate in the Lifeline program.
A quick note here that the ISPs are anxious to be left out of the paperwork of reimbursement from the FCC, and instead have consumers get money directly in the form of some kind of restricted debit or credit card. There are promising pilot initiatives in the SNAP program to extend those benefits to be more credit card-like with online capability, though it hasn’t been widely adopted yet. If successful in SNAP, it seems like a natural and agreeable extension to use it for these broadband programs.
The section on pricing data transparency is no less important. The FCC collects limited data on the cost of broadband, a perilous position given the monopolistic shape of our broadband market -- 38% of census blocks with broadband have access to only one broadband provider, and another 36% only have access to two providers. Open Technology Institute’s research shows the average cost of broadband is $68.38 per month, far above the $10 per month that most advocates think would be affordable for low-income Americans. Let’s be honest about it: most of these deployment subsidy programs are extending monopolies into new areas. Even if you’re not worried about the cost today, it is hard to argue we shouldn’t have data on the prices consumers are being asked to pay.
I have to quibble with the Sec 2002(a)(5) about network outage transparency though I think the intent is right. It states the FCC shall collect “data necessary to assess the resiliency of the broadband internet access service network in the event of a natural disaster or emergency.” A few months ago a “fiber cut” in Brooklyn brought Verizon FiOS to a crawl for a large part of the mid-Atlantic. My understanding is because internet is an “information” service and not a “communications” service, they are not required to report to the FCC about what happened. My best guess is that as Verizon tried to route traffic around the fiber cut, they overloaded other routers not designed to handle the extra traffic. Regardless, they should be required to say what happened (without giving away information that could be valuable to attackers, of course) and what they plan to do about it so it doesn’t happen again.
Deployment and Auctions
The core of the deployment section is $80 billion — 75% to be disbursed by the FCC, and 25% by the states. All of it in reverse auctions. When conducting the auctions, the FCC and the states are obligated to first hold an auction only for gigabit speeds; only after an area is not bid on at the gigabit level would it be available for an auction at lower speeds. Within each auction, there has to be a “preference” for service to areas currently without access to 25/3 service. There are secondary preferences for projects with non-federal matching funds, areas with persistent poverty, projects that would provide open-access, among a few others.
It’s the reliance on the reverse auction that is the biggest problem of this bill. The landmark 2017 FCC study anticipated needing $80 billion to bring FTTP to every household not served by 25/3 broadband, which was 14% using the data at the time. Even after the FCC produces new maps, the number of unserved will likely be below 10% today. The RDOF cost model anticipated $26 billion to serve the 4.4% we know to be unserved (only committing $9.2 billion in the auction). The point is $80 billion is plenty of money.
Why, with access to all the capital that is needed, are we going to use a model that subsidies private companies to extend a monopoly into a new area? For example, Boone County, Kentucky is putting in $13.6 million towards a FTTP network; Cincinnati Bell is putting in $30 million. After a few contractual concessions, the ISP still owns everything. What if Boone County could afford, through this bill, to fund the whole thing? That would look like what West Des Moines is doing: $40 million to own their fiber network, leasing access to Google Fiber and any other ISP that wants to use it. When money is the scarce resource, it makes sense to use reverse auctions and prioritize saving money. If future-proof, affordable, excellent broadband is the goal, local ownership and open access should be the priority.
Part of the problem with the reverse auction is we don’t really know what’s going to come out the other end. I think there’s a good chance if you re-auctioned the RDOF areas with these rules (and with the FCC’s RDOF experience) you’d get a good result. It gets a lot more complicated for partially served census blocks, and places with existing service.
Partially served Census blocks, which are what the FCC is identifying with new maps, already have one, sometimes two, broadband providers on the block. The ““““incumbent”””” (who has not been providing service to the rest of the block for many years now) is likely keep bidding down to $0 to keep competitors out. It will be frustrating if it takes the threat of competition and $0 to finally bring service to an area that really never needed a subsidy.
One more point on the auctions: If we want this money to put fiber into the ground to be used for service to households, businesses, anchor institutions, as wireless backhaul, satellite basestations, etc—I’m still confused why we just don’t say that. Holding a gigabit-only auction is clearly trying to achieve that outcome. While the FCC would be unlikely to repeat the blunder of allowing fixed wireless to bid at the gigabit tier, the requirement of “technological neutrality” seems to invite this type of problem in the future. I’d suggest find/replace “at least 1 gigabit” with “fiber optic cable to the home capable of speeds not less than 1 gigabit symmetric”.
Speeds
An important debate has broken out about whether 100/100 is the right definition for broadband. The bill defines “low-tier” broadband as anything between 25/3 (under which is unserved) and 100/100 Mbps symmetric. Sec 723(b)(1) only allows the federal funds to be expended on unserved or low-tier areas. Proponents of that standard argue that 100/100 all but ensures fiber will be used to meet that need, and with a one-time infusion of capital, we should be building future-proof, once-and-done, better-than-good-enough networks. Detractors argue that 100/100 will bring large swaths of the country into scope as lacking good broadband and that currently unserved areas will fall to the back of the line after easier-to-upgrade areas.
While the law does preference bidding for areas unserved at 25/3, it allows “low-tier” areas (less than 100/100) to be included in the same auction. Cable internet—which remains the most common deployment in suburban areas—has speeds like 987/35 and would qualify for subsidies in this program. This threshold makes 175.8 million Americans eligible for these subsidies using currently available Form 477 data, many of which are cable internet where 35Mbps upload is likely sufficient, even working from home.
I see logic in getting the whole country on fiber internet with this once-in-a-generation infrastructure plan. However, the USF high-cost program gives an ISP a subsidy because they wouldn’t otherwise serve an area. That’s not the case here. I don’t think taxpayers are getting a fair deal subsidizing a suburban ISP to upgrade a cable network to fiber. To take on these upgrade costs, taxpayers should get more in return, for example: long term affordability commitments, open access requirements, local ownership of the wire, etc.
It’s important to remember that the Accessible, Affordable, Internet for All Act was introduced and passed by the House in the Summer of 2020 in a time of divided government. If it becomes law, it will be a landmark achievement in elevating digital equity and inclusion relative to deployment, which is long overdue. Unfortunately, the drafters didn’t have a crystal ball to foresee RDOF embarrass itself so tremendously. With one chance to remake broadband accessibility, we should strive for only the best.
_____
I’m grateful to CommunityNetworks for their series on the bill which was a helpful starting point for analysis.