If you’ve ever moved to a new neighborhood or city, you might have had a moment where you wondered about the internet service available at your new house or apartment. If you’re a bit obsessed with internet accessibility as I am, you might have even used the FCC’s broadband availability map to check the “official” record. If you did, the picture you got back from the FCC was likely an unrecognizable jumble. Both independent analysts and the Federal Communications Commission itself have acknowledged that the maps likely overstate the number of internet providers available to the average American.
The problem is so pronounced that Congress passed a law in March 2020 requiring the FCC to correct the maps. The commission is now in the midst of implementing a new mapping system to comply with the law—and is facing tremendous pressure to deliver the final product quickly. However, in this rush to fix the maps, the FCC’s new understanding of the digital divide might end up being only marginally better than the old maps.
The FCC data says 4.4 percent of Americans don’t have access to high-speed fixed terrestrial broadband (a fancy way to say internet through a wire). But we know the true number is higher because the FCC maps overcount access: If one house in a Census block is served, the commission assumes every house is.
George Ford of the Phoenix Center, a center-right think tank, did a thorough independent statistical analysis and found that there may be “4 million homes said to have broadband that may not.” Put another way, he found the FCC data overstates access by 3.3 percentage points—1.9 points in urban areas and 11.5 points in rural areas. By this estimate, 7.7 percent of Americans lack access to high-speed broadband.
And billions of dollars in federal broadband subsidies are based on this inaccurate data. When a block is incorrectly considered served, then some of its residents will have access to neither broadband nor federal funds to help bring broadband service. That’s why, with the pandemic forcing many to rely on the internet at home for work and school, politicians from both sides of the aisle have blamed the maps for the persistent digital divide.
“The [FCC 2020 Broadband Deployment Report] claims that 98% of West Virginians have access to fixed broadband or mobile LTE coverage despite the fact that I already proved these maps wrong,” Sen. Joe Manchin wrote to then-FCC Chairman Ajit Pai in May. He continued, “Every West Virginian knows that it is impossible to get reliable broadband access in many areas of our state. … We cannot close the digital divide until we resolve the data divide between the FCC maps and the actual service on the ground.”
While the maps are a problem, it’s important to remember that wholly unserved areas aren’t in dispute—only partially served blocks are. Still, the maps have delayed some funding efforts to close the homework gap. For instance, in August 2019 the FCC launched the Rural Digital Opportunity Fund, a $20 billion effort to bring broadband to everyone who currently lacks access. However, an inability to identify partially served areas forced them to split the program in two parts: Phase I was completed in November 2020, allocating $9.2 billion to completely unserved areas. Phase II is still waiting for new maps.
Discussions about how to address the map deficiency have been ongoing for years. In 2019, after a long history of “scuttling efforts to improve broadband mapping,” an industry group of internet service providers led by USTelecom decided to “lead the charge” for better maps. They advocated for the creation of a Broadband Serviceable Location Fabric—a rooftop-level national map of every location where broadband could be installed. The plan is for the FCC to combine these with similarly granular maps from local ISPs to create national broadband accessibility maps on a house-by-house basis.
In March 2020, the House and Senate passed the Broadband DATA Act unanimously and President Donald Trump signed it into law. The requirements of the bill largely follow the suggestions made by USTelecom. But the Broadband DATA Act got off to a slow start. Congress didn’t fund the effort until December, and the process didn’t start in earnest until President Joe Biden appointed Jessica Rosenworcel as acting chairwoman of the FCC in January.
Some states—Georgia, Maine, Pennsylvania—haven’t been waiting for the FCC. Georgia started developing its own broadband map in 2019, partnering with local governments and ISPs. It was published in June 2020. The FCC has the aditional challenge of coordinating the new national map with work that’s already been done at the state and local level.
Lawmakers seem to think the new data will itself fix the homework gap. At a recent hearing of the Senate Commerce Committee, Chair Maria Cantwell and ranking member Roger Wicker found common ground on the need for the FCC to hurry up and deliver the new maps. And Wicker has since followed up with a request for an oversight hearing to pin down the FCC on a timeline.
But we need Congress and the FCC to improve data for all the factors that prevent Americans from using broadband service—including affordability and the speeds you’ll actually get in rural areas.
Former FCC Chairman Ajit Pai liked to tout that the number of Americans without broadband access “fell to 14.5 million” or 4.4 percent, “a 46% decrease from the end of 2016” through 2019. But the census’s American Community Survey shows 35.9 million American households (approximately 93.7 million people) lack broadband service, which is 29.2 percent of all households and a decrease of a mere 7.6 percent since 2016.
Offering every American broadband won’t close the homework gap if they can’t afford it or don’t sign up. Research from New America’s Open Technology Institute 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. (New America is a partner with Slate and Arizona State University in Future Tense.) The FCC collects only limited data on the cost of broadband, a perilous position given the monopolistic shape of our broadband market—38 percent of census blocks with broadband have access to only one broadband provider, and another 36 percent only have access to two providers.
The Biden administration and congressional leadership appear to understand this important dynamic. The president’s infrastruture proposal specifically calls out price transparency, competition, and affordability as important components of actually closing the digital divide. Rep. James E. Clyburn’s Accessible, Affordable, Internet for All Act would implement price transparency by requiring the FCC to collect data on the cost of broadband. This bill doesn’t prescribe a technology solution to pricing data collection but rather directs the FCC to figure it out.
One approach might be to pair pricing data with speed data. The FCC does collect actual speed data collected through routers as part of its Measuring Broadband America report. Unfortunately the sample size in the report is a paltry 2,931 households. For example, there are just 21 participants in West Virginia and 66 participants on Cincinnati Bell DSL. In a market estimated to have revenues of $118 billion, we can surely afford to expand the sample size of this promising program to measure the impact of these federal subsidy programs.
This is all the more important because of the potential challenges with Congress’ plan for new maps, which could not be more different from how the FCC currently collects this data. Both the advanced technology used to identify rooftop locations and the maps submitted by the ISPs will need to be thoroughly tested against existing data for accuracy. Whereas currently the FCC data sits on top of the census’s information, in the future, everything will be based on new “serviceable locations” identified by this fabric—which means that the government has to move with care and caution. If you replace the foundation of your house, you don’t want to rip out the existing foundation all at once, then try to build in a new one. The house will fall. Instead, you take pieces of the foundation out, and carefully build in new ones.
To understand why careful implementation of the new maps is critical, we need to understand the plan beyond a superficial level. CostQuest Associates, a consulting firm and recognized industry leader in broadband mapping, presented a comprehensive approach to Congress in 2019. Its plan is to identify rooftops from satellite imagery, then add “parcel and land attribute data, address data, and other sources” to identify serviceable locations. CostQuest acknowledges the obvious data challenges of standardizing parcel and tax assessor data across every county and jurisdiction. This new technology is impressive and has the potential to be successful given complete and accurate data, but standardizing all that data at a national level is a nightmarish and gargantuan task.
Take this example above from the presentation. The technology will need to differentiate between houses that need broadband and other structures that don’t. A small error—such as identifying multiple buildings from this farm as eligible when only the farmhouse is—may not seem like a big deal. But aggregated across entire counties, it could significantly distort our understanding of the homework gap and the funding needed to close it.
Congress also directed the FCC to change how ISPs report their coverage areas. Instead of reporting whether they provide service to a census block (the smallest level of geography tracked by the census—roughly a city block of people), the ISPs will submit shapefiles that outline their coverage area.
Then, to create the maps, the FCC will overlay the ISPs’ shapefiles on top of the fabric’s serviceable locations. If the location is inside a coverage polygon, that location is served; if not, it’s not served. That makes sense but has one important flaw: Both data sets have to use the same underlying “reference data.” That is, they have to agree on where the serviceable location is on a plot of land.
In its testimony before Congress, CostQuest Associates highlights this as a potential issue that needs to be considered. It provides a hypothetical ISP’s coverage area (light-green polygons) that is based on the addresses and roads they serve. But if the fabric places the rooftop locations far off the roads (the green dots), the ISP’s polygon may not actually show them as served.
These are big changes that need to be implemented and tested carefully, not rushed. Even USTelecom, the industry group of ISPs, describes this project as “massive” and urges the FCC to make “iterative progress.” In one positive sign, in March, the FCC released a form asking consumers who have called their ISP and been denied service (which is called digital redlining, and not surprisingly happens disproportionately in low-income communities) to raise their hand. This might be the beginning of an iterivative approach by the FCC to unlock funding opportunities for unserved areas in advance of completed new maps.
There’s no doubt that the FCC’s broadband maps overstate access to broadband, particularly in rural areas. However, political pressure to quickly deliver the maps does not serve the shared interest of closing the homework gap as expeditiously as possible. To close the gap, Congress and the FCC should focus on measuring affordability and equity issues, in addition to a thorough, tested, and transparent revamp of the coverage maps.
Future Tense is a partnership of Slate, New America, and Arizona State University that examines emerging technologies, public policy, and society.
